THE TO'AGA SITE THREE MILLENNIA OF / POLYNESIAN OCCUPATION 2. THE MANU'A ISLANDS, I~~~~~~~~~N AMERICAN SAMOA _,I~~~~~~ - - -~ ~ ~ ~ ~ ~ ~ ~ ~ F View of the completed 1987 main excavation at Totaga. Cover photo: View of Ofu Island coastline fronting the To'aga site. The volcanic plug at Fa'ala'aga is visible in the distance. (Photo by P.V. Kirch) ISBN 1-882744-01-2 THE TO'AGA SITE THREE MILLENNIA OF POLYNESIAN OCCUPATION IN THE MANU'A ISLANDS, AMERICAN SAMOA R V. KIRCH AND T. L. HUNT EDITORS Number 51 Contributions of the University of Califomia Archaeological Research Facility Berkeley Primary funding forthis research has been provided by the Historic Preservation Office of the Department of Parks and Recreation, Government of Amenican Samoa, through grants from the U. S. National Park Service, Archaeological Assistance Division, Department of the Interior. However, the contents and opinions expressed do not necessarily reflect the views or policies of the Department of the Interior. Library of Congress Catalog Card Number 93-72699 ISBN 1-882744-01-2 © 1993 by the Regents of the University of California Archaeological Research Facility University of California at Berkeley Printed in the United States of America All rights reserved. No part of this publication may be reproduced or transmitted in any fonn orby any means, electronic or mechanical, including photocopy, recording, or any information storage or retrieval system, without permission in writing from the publisher. Dedicated to the memory of JOHN KNEUBUHL (1920-1992) Citizen of Two Cultures Tagaloa e, tawnuli ai, Tagaloafiamalolo; E mapu i le lagi Tuli mai vasa; Ta fili'a i peau a lao. Fea le nu'u na lua'i tupu? Manua-tele na mua'i tupu. Tagaloa, who sits at the helm, Tagaloa desires to rest; Tulifrom the ocean must rest in the heavens; These waves below affright my breast. Where is the land whichfirst upsrang? Great Manu'afirst uprose. (Mead 1930:152) Contents Preface ix Chapter 1. INTRODUCHON AND RESEARCH DESIGN. T. L. Hunt and P. V. Kirch Background to Samoan Archaeology Design of the Research References Cited 1 2 5 6 Chapter 2. OFU ISLAND AND THE TO'AGA SITE: DYNAMICS OF THE NATURAL AND CULTURAL ENVIRONMENT. P. V. Kirch 9 Geology and Geomorphology of Ofu Resources of the Island The Cultural and Social Landscape To'aga: Land Use and Vegetation Patterns References Cited 11 12 17 18 21 Chapter 3. SURFACE ARCHAEOLOGICAL FEATURES OF THE TO'AGA AREA. T. L. Hunt References Cited 23 30 Chapter 4. THE TOAGA SITE: MODELLING THE MORPHODYNAMICS OF THE LAND-SEA INTERFACE. P. V. Kirch 31 Geomorphology of the To'aga Coastal Terrace Morphodynamic Processes: Sea Levels and Subsidence Sediment Budgets: Sources and Modes of Deposition Modelling the Morphodynamics of the To'aga Coastal Terrace in the Mid- to Late-Holocene To'aga: Expectations for Archaeological Site Formation Processes References Cited 32 33 38 38 40 41 Chapter 5. EXCAVATIONS AT THE TOAGA SITE (AS-13-1). P. V. Kirch and T. L. Hunt 43 Introduction 1986 Test Excavation The 1987 Excavations The 1989 Transect Excavations Summary and Conclusions References Cited 43 46 46 57 82 83 Chapter 6. RADIOCARBON CHRONOLOGY OF THE TO'AGA SITEa. P. V. Kirch Radiocarbon Dating Procedures: Corrections and Calibrations The Radiocarbon Corpus from To'aga v 85 86 87 Discussion References Cited 89 91 Chapter 7. A GEOARCHAEOLOGICAL ANALYSIS OF SEDIMENT SAMPLES FROM THE TO'AGA SITE EXCAVATIONS. P. V. Kirch, E. Manning, and J. Tyler 93 Introduction 93 94 98 111 114 Methods Results Summary and Conclusions References Cited Chapter 8. NON-MARINE MOLLUSCS FROM THE TO'AGA SITE SEDIMENTS AND THEIR IMPLICATIONS FOR ENVIRONMENTAL CHANGE. P. V. Kirch 115 Introduction 115 116 116 Material and Methods Systematic Review Results 116 120 Conclusion References Cited 120 Clupter 9. THE TO'AGA CERAMICS. T. L. Hunt and C. Eficelens 123 The Assemblage Analytic Procedures Results of Macroscopic Analysis Ceramic Compositional Microanalysis Conclusions References Cited 123 125 128 137 145 148 Chapter 10. SAND TEMPER IN PREHISTORIC POTSHERDS FROM THE TO'AGA SITE. W. R. Dickinson To'aga Temper Variants 151 Temper Grain Types Profuse Basaltic Temper Sparse Basaltic Temper Feldspathic Basaltic Temper Mixed Temper Sand 151 152 152 153 154 154 Discussion: Temper Comparisons References Cited 155 156 Chapter I 1. NON4ERallP THE T(YAGA sum, En. 157 j_157 - ~~~~~~~~~~~~~~~~~~~~~1 Ornaments Miscellaneous Artifacts Worked Shell Unretouched Lithics References Cited 162 164 164 165 166 Chapter 12. CHEMICAL CHARACTERIZATION AND PROVENANCE OF MANUA ADZ MATERIAL USING A NON-DESTRUCTIVE X-RAY FLOURESCENCE TECHNIQUE. Marshall Weisler Intduction 167 167 168 170 175 185 185 185 The Collections Methods Results Summary and Conclusions Acknowledgements References Cited Chapter 13. FAUNAL ASSEMBLAGES FROM THE TOAGA SITE. Lisa Nagaoka 189 189 190 197 208 208 214 214 214 Methods Results Analysis of Bulk Samples Temporal Trends in the To'aga Assemblage Regional Comparisons Conclusions Acknowledgements References Cited Chapter 14. BIRD BONES FROM THE TO'AGA SITE: PREHISTORIC LOSS OF SEABIRDS AND NEGAPODES. David W. Steadman Intduction 217 217 217 226 226 226 Species Accounts Discussion Acknowledgements References Cited Chapter 15. SYNTHESIS AND INTERPRETATIONS. P. V. Kirch and T. L. Hunt 229 Cultural Sequence and Chronology Geoarchaeology and Landscape Change 230 232 Ancestral Polynesian Culture Subsistence Patterns Inter-island Contacts and Exchange The To'aga Site: Cultural Resource Management Considerations References Cited 236 240 243 243 246 vii CONTRIBUTORS William R. Dickinson, Department of Geosciences, University of Arizona, Tucson, AZ 85721 Conrad Erkelens, Department of Anthropology, University of Hawaii, Honolulu, HI 96822 Terry L. Hunt, Department of Anthropology, University of Hawaii, Honolulu, HI 96822 Patrick V. Kirch, Department of Anthropology, University of California, Berkeley, CA 94720 Elizabeth Manning, Department of Anthropology, University of California, Berkeley, CA 94720 Lisa Nagaoka, Department of Anthropology, University of Washington, Seattle, WA 98195 David W. Steadman, Biological Survey, New York State Museum, Albany, NY 12230 Jason Tyler, Department of Anthropology, University of Washington, Seattle, WA 98195 Marshall Weisler, Department of Anthropology, University of California, Berkeley, CA 94720 PREFACE In 1985 I received a request from te Historic Preservation Officer of American Samoa, Mr. Stan Sorensen, to submit a proposal for archaeological survey work in American Samoa. Realizing that the Manu'a Islands, lying at the extreme eastern end of the Samoan archipelago, were the most neglected part of the Samoan group from the viewpoint of prehistory, I suggested that we might undertake a reconnaissance of the three islands-Ta!u, Ofu, and Olosega-that make up Manu'a. A previous survey in 1962, by Y. Sinoto and W. Kikuchi of the Bernice P. Bishop Museum, had revealed the presence of various surface archaeological sites, but had failed to establish any significant time depth for the prehistoric occupation of these islands. The proposed reconnaissance was carried by Terry Hunt and myself from June to August, 1986 and resulted in the discovery of the first pottery-bearing sites in the Manu'a group, on Ta'u and Ofu, as well as a range of later prehistoric monuments and artifacts. (Tbe results of the 1986 reconnaissance survey were subsequently published by Hunt and myself in The Journal ofthe Polynesian Society, Vol. 97, pages 153-83, 1988.) From the results of our initial survey, it appeared that ceramic-bearing deposits exposed by a Works Department landfill at To'aga on Ofu Island were particularly promising for further archaeological investigations. Thus, in 1987 we proposed to the Historic Preservation Office a second phase of work at To'aga. In this second phase, we recommended a series of systematic subsurface test excavations in order to determine more accurately the nature and extent of the To'aga site deposits. This second field season in 1987 was again directed by Hunt and myself, with the assistance of Jason Tyler and Jean Gehman. (Preliminary results of this work were published in Archaeology in Oceania, Volume 25, pages 1-15, 1990). The 1987 excavations revealed the presence of a deep, well-stratified prehistoric record in the To'aga area, with subsurface deposits spanning most-if not allof Manu'an prehistory. These strata were so extensive, however, that we were not able to determine their limits in the 1987 field season. Therefore, a third season was proposed to define more precisely the full subsurface extent of the site. This work was carried out in the summer of 1989, with fieldwork co-directed by Hunt and me, and with the assistance of a team of students from the University of Hawaii and the University of California at Berkeley. This monograph constitutes the final report of our thre field seasons at the To'aga site. While our core objectives were those of site survey and inventory (with the aim of preparing a National Register of Historic Places nomination for To'aga), we have endeavored to go well beyond these minimal goals in thle present work. Thus, we have carried out extensive analyses of much of the excavated material, including faunal remains, ceramics, and basalt artifacts. In addition, we have paid considerable attention to the geomorphological and geoarchaeological problems of site formation at To'aga, noting that thee also impinge on cultural resource manage1X ment considerations. We trust that these efforts contribute significantly to our knowledge of Western Polynesian prehistory as well as to site survey and inventory in American Samoa. During the course of our tree expeditions to Manu'a we have received the support and assistance of a great many individuals and organizations. Our primary source of funding has been the Historic Preservation Office of the Department of Parks and Recreation, Government of American Samoa, through grants from the U. S. National Park Service, Archaeological Assistance Division. Additional support has been provided by the Burke Museum of the University of Washington (for the 1986 and 1987 seasons), and the Archaeological Research Facility of the University of California at Berkeley (for the 1989 season). We are particularly grateful to Stan Sorensen, the Historic Preservation Officer of American Samoa, for his interest and support from the inception of this project. Anne Sauter of the Archaeological Research Facility at Berkeley assisted in various administrative matters. Tanya Smith carefully edited and produced this final monograph. Two respective district governors of Manu'a have lent their support to the project: High Chief Aolaolagi Soli and High Chief Tufele Li'a. We are also pleased to record the support of the people of Ofu Village, especially High Chief Misa'alefua, High Talking Chief Faoa, Liulamaga Ta'ilele, Manu'a Peau, and Sina Peau. In 1987, Tito and Margaret Malae assisted greatly with housing and other arrangments. Likewise, in 1989 Manu'a Peau and her family provided housing, hospitality, and a sense of being at home with family. Our field crews of Ivala Live, Fuave'a Ta'ilele, Tillis Thompson, Pauesi Malo, Eleloi Misa'alefua, Ele'ele Utuone, Paulo Su'e, and Opetaia Fa'amita were not only dedicated workers, but interested co-investigators. We also thank Jason Tyler, Jean Gehman, Elizabeth Manning, Conrad Erkelens, Melissa Kirkendal, Robert Holsen, Lisa Nagaoka, and Ann Rowberg, who assisted in the 1987 and 1989 field- work. Prof. Roger C. Green of the University of Auckland kindly reviewed the entire manuscript, and we are most grateful for his insightful comments. Throughout all field seasons, John and Dorothy Kneubuhl ofTutuila generously opened their home to us and to our assistants, and helped us in countless ways. John's profound knowledge of Samoan culture and history was a constant inspiron. His visit to our field site in 1989, arriving with only a suitcase crammed full of two large legs of lamb, mint sauce, and a bottle of good scotch whiskey, was an event not soon to be forgotten. It was with profound sadness that we learned of his death in Tutuila in February of 1992. As a small token of ouresteem forJoln we dedicate this volume to his memory. Parick V. Kirch Berkee,January 1991 x~~~~~~~~. .^ LIST OF FIGURES 3 1.1. Map of the Samoan archipelago. 2.1. Map of the Manu'a Islands. 2.2. View of the exposed volcanic plugs at Fa'ala'aga from the beach at To'aga. 2.3. Sketch of basalt dikes exposed in the road cut across the Le'olo Ridge at Fa'ala'aga. 2.4. Map of Ofu Island. 2.5. Distribution of major floral dominants along Transect 7 at To'aga. 10 13 13 15 20 3.1. Map of the southern portion of Ofu Island showing site AS-13-1. 3.2. Plan of round-ended stone house foundation. 3.3. Large volcanic boulder at the base of the talus slope with twelve artificial grinding facets. 27 3.4. Plan and cross section of the Tui Ofu well at Muli'ulu. 29 3.5. Plan and cross section of the Tui Ofu Tia, stone mound/tomb at Muli'ulu. 25 25 4.1. A generalized transect across the coastal plain at To'aga. 35 4.2. A model for the sediment budget at To'aga, showing terrestrial and marine inputs. 4.3. A model of shoreline transgression and regression. 4.4. Map of the southwestern Pacific region showing locations with evidence for midto late-Holocene higher sea-level stands. 4.5. Time-elevation plot of sea levels in Polynesian archipelagoes 4.6. Time trends in four key variables affecting the morphodynamics of the To'aga coastal plain. 40 4.7. Retrodiction of the transgression-regression sequence along the To'aga coastline. 33 5.1. Example of the excavation recording form used in the 1987 and 1989 To'aga field seasons. 45 5.2. Elevation profile along the 1987 excavation transect. 48 5.3. Map of the 1987 excavation area, showing the locations of excavation units. 5.4. Excavation in progress in the 1987 main trench. 5.5. Stratigraphic section of the southwest face of the 1987 main excavation (Units 5-9). 50 5.6. Stratigraphic section of the south and west faces of Unit 3. 5.7. Stratigraphic section of the north and east faces of Unit 10. 5.8. Schematic map of the southern coastal flat of Ofu Island, from To'aga to Fa'ala'aga. 59 5.9. Elevation profile along Transect 3. 5.10. Stratigraphic profile of the east face of Unit 27. 5.11. Elevation profile along Transect 5. 5.12. Stratigraphic section of the west face of Unit 17. 5.13. Exploded stratigraphic section of all faces of Unit 17. 5.14. Stratigraphic section of the west face of Unit 16. 5.15. Stratigraphic section of the west face of Unit 15 5.16. Stratigraphic correlations between Units 15, 16, and 17 on Transect 5. 5.17. Elevation profile along Transect 7. 5.18. Stratigraphic section of the west face of Unit 19. 5.19. Stratigraphic section of the east and south faces of Unit 19. 5.20. Elevation profile along Transect 9. 28 35 36 37 39 47 49 52 53 60 61 62 63 64 65 66 68 69 70 71 72 5.21. Expansion of Unit 20 into Unit 23 5.22. Stratigraphic section of the west and north faces of Units 20/23. 5.23. View into Units 20/23. 5.24. Stratigraphic section of the south face of Unit 21. 5.25. Stratigraphic section of the north face of Unit 22. 5.26. Stratigraphic correlations between Units 20/23, 21, and 22. 5.27. Elevation profile along Transect 11. 5.28. Stratigraphic section of the west face of Unit 26. 5.29. Elevation profile along Transect 17. 5.30. Stratigraphic section of the west face of Unit 24. 5.31. Stratigraphic section of the north face of Unit 25. 6.1. Plot of radiocarbon age determinations from the To'aga site (cal B.P. at 1 sigma). 73 74 75 76 77 78 79 80 81 81 82 90 7.1. Photomicrograph (lOX) of the 1 phi size fraction of a modem control sample of calcareous beach sand from Transect 7. 7.2. Terminology adopted for the To'aga site sediments. 7.3. Photomicrograph of a typical "salt-and-pepper" sand with mixed calcareous and 95 97 volcanic-lithic lithology. 99 7.4. Summary diagram for sediment samples from Unit 15, Transect 5. 7.5. Histogram plots of grain size distributions for sediment samples from Unit 15, Transect 5. 102 7.6. Photomicrographs (lOX) of the 0 phi size fractions of sediment samples from Unit 15, Transect 5. 7.7. Summary diagram for sediment samples from Unit 19, Transect 7. 7.8. Histogram plots of grain size distributions for sediment samples from Unit 19, Transect 7. 106 7.9. Summary diagram for sediment samples from Unit 23, Transect 9. 7.10. Photomicrographs (lOX) of the 0 phi size fractions of sediment samples from Unit 23, Transect 9. 7.11. Histogram plots of grain size distributions for sediment samples from Unit 23, Transect 9. 111 7.12. Summary diagrams of grain size and lithology for Units 21, 22, and 23, Transect 9. 112 7.13. Summary diagrams of grain size and lithology for Units 24 and 25, Transect 17. 113 7.14. Histogram plots of grain size distributions for sediment samples from Unit 24, Transect 17. 114 8.1. Photomicrographs of terrestrial gastropods recovered from the To'aga site sediments. 9.1. Plot of exterior sherd color (hue with value and chroma) for the late period. 9.2. Plot of exterior sherd color (hue with value and chroma) for the middle period. 9.3. Plot of exterior sherd color (hue with value and chroma) for the early period. 9.4. Thickness and temper size histogram for the To'aga ceramic assemblage. 9.5. Thickness histogram for late period ceramics. 9.6. Thickness histogram for middle period ceramics. 9.7. Thickness histogram for early period ceramics. 9.8. Cross sections of selected rim sherds from the To'aga assemblage. 9.9. Elemental spectrum of clay composition for Sherd 22 from To'aga. 9.10. Dendrogram of Ofu pottery and clay samples using the Average Linkage (between groups) method. xii 101 103 104 107 109 117 130 130 131 135 137 138 139 140 141 142 9.11. Dendrogram of Ofu pottery and clay samples using Ward's method. 9.12. First and second discriminant function scores for analysis based on cluster (1-4) as grouping variable. 9.13. First and second discriminant function scores for analysis based on time period. 143 145 145 11.1. Basalt adzes from the To'aga site. 11.2. Turbo-shell fishhooks from the To'aga site. 11.3. Turbo-shell fishhooks and fishhook fragments from the 1987 excavation at To'aga. 11.4. Miscellaneous artifacts from the To'aga site. 11.5. Miscellaneous artifacts from the To'aga site. 12.1. Map of the Tatagamatau quarry complex, Tutuila Island, showing the location of analyzed source rock samples. 12.2. A comparison of source rocks analyzed by non-destructive x-ray fluorescence. 12.3. A wide dispersion of data points results when ratios of rubidium, strontium, and yttrium are used to assign specimens to possible sources. 12.4. Artifacts and source rocks plotted by ratios of zirconium, strontium, and niobium resulting in many specimens plotting within the Tatagamatau source envelope. 159 160 161 163 165 171 172 175 182 197 13.1. Relative frequency of fish taxa from the To'aga site, including Diodontidae. 198 13.2. Relative frequency of fish taxa from the To'aga site, excluding Diodontidae. 13.3. Percentage composition of fish faunal assemblages from major Westem Polynesian and Fijian sites. 205 14.1. Map of the Pacific Islands showing localities mentioned in the text. 14.2. Comparison of megapode bones. 219 223 15.1. Schematic map of the To'aga area showing inferred extent of ceramic-bearing deposits. xiii 233 LIST OF TABLES 2.1. Environmental characteristics of the Manu'a Islands. 11 3.1. Surface archaeological features at To'aga. 26 7.1. Analytic data for To'aga control sediment samples (1989). 7.2. American Samoa, To'aga site 13-AS-I sediment analysis (1987). 7.3. Analytic data for Unit 15 sediment samples. 7.4. Analytic data for Unit 16 sediment samples. 7.5. Analytic data for Unit 17 sediment samples. 7.6. Analytic data for Unit 18 sediment samples. 7.7. Analytic data for Unit 19 sediment samples. 7.8. Analytic data for Unit 21 sediment samples. 7.9. Analytic data for Unit 22 sediment samples. 7.10. Analytic data for Unit 23 sediment samples. 7.11. Analytic data for Unit 24 sediment samples (micro-constituents). 7.12. Analytic data for Unit 24 sediment samples. 7.13. Analytic data for Unit 25 sediment samples. 95 96 100 100 101 104 105 105 107 108 108 110 110 8.1. Non-marine molluscs from 1987 Main Trench. 8.2. Non-marine molluscs from Unit 3. 119 119 9.1. Ceramics from the 1987 excavation units. 9.2. Ceramics from the 1989 excavation units. 9.3. Frequency of body and rim sherds. 9.4. Frequency of sherds by temper size modes. 9.5. Frequency of sherds by "Preferred" orientation of inclusions relative to vessel walls. 9.6. Frequency of sherds by oxidation-reduction pattern in cross section. 9.7. Frequency of sherds by exterior Mohs hardness. 9.8. Frequency of sherds by interior Mohs hardness. 9.9. Frequency of sherds by exterior surface treatment. 9.10. Frequency of sherds by interior surface treatment. 9.11. Frequency of sherds by interior anvil casts. 9.12. Frequency of sherds by exterior paddle marks. 9.13. Frequency of sherds by organic residue. 9.14. Sherd thickness statistics by analytic time periods. 9.15. Sherds and Ofu clays selected for SEM/EDS clay elemental and sand temper petrographic analyses. 126 127 128 129 132 10.1. Frequency percentages of ferromagnesian mineral grains and basaltic volcanic lithic fragments in sherds containing "profuse basaltic temper." 153 10.2. Frequency percentages of ferromagnesian mineral grains and basaltic volcanic lithic fragments in sherds containing "sparse basaltic temper." xiv 132 133 133 134 134 134 135 136 136 144 154 10.3. Frequency percentages of plagioclase feldspar and opaque iron oxide mineral grains and basaltic volcanic lithic fragments in sherds containing "feldspathic basaltic temper." 155 10.4. Frequency percentages of calcareous grains and silicate-oxide mineral grains and lithic fragments for sherds containing "mixed temper sand." 12.1. Artifacts analyzed by EDXRF. 12.2. Evaluation of analytical accuracy and precision for U.S.G.S. Standard RGM-1. 12.3. Variations of oxides and elements from the Mo'omomi adz quarry, Moloka'i, Hawaiian Islands. 12.4. Geochemistry of Tatagamatau source rock (fused disks). 12.5. Geochemistry of Mako ridge source rock (fused disks). 12.6. Geochemistry of Fa'ala'aga source rock (fused disks). 12.7. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (fused disks). 179 12.8. Geochemistry of Tatagamatau source rock (whole specimen). 12.9. Geochemistry of Mako ridge source rock (whole specimen). 12.10. Geochemistry of Fa'ala'aga source rock (whole specimen). 12.11. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (whole specimen). 12.12. Geochemistry of To'aga and Ofu Island artifacts (whole specimen) 13.1. Summary of To'aga faunal remains. 13.2. Fish fauna from the 1987 field season (NISP). 13.3. Fish fauna from the 1989 field season (NISP). 13.4. To'aga fish fauna: main trench, Units 1,4-9. 13.5. To'aga fish fauna: transect 5, Units 15/29/30 (NISP). 13.6. To'aga fish fauna: transect 9, Units 20/23 (NISP). 13.7. Modern Samoan fishing methods (after Hill 1986). 13.8. Non-fish vertebrate fauna for 1987 and 1989 excavations. 13.9. To'aga non-fish vertebrate fauna. 13.10. To'aga invertebrate data: transect 9, Units 20/23. 13.11. To'aga invertebrate data: 1987 main trench, Units 1,4-9. 13.12. To'aga invertebrate data: transect 5, Units 15/29/30. 13.13. Unit 20/23 bulk sample analysis. 13.14. Unit 30 bulk sample analysis. 13.15. Taxa represented in bulk sample from Units 20/23 (NISP). 13.16. Taxa represented in bulk sample from Layer II, Unit 30 (NISP). 13.17. Density of identified fish bone from Layer IIIA/B, Unit 20/23. 13.18. Density of identified fish bone from Layer II, Unit 30. 13.19. Summary of recovery and quantification techniques for Western Polynesian faunal analyses. 13.20. Summary of Western Polynesian vertebrate faunal assemblages (NISP). 14.1. Birds from the To'aga site. 14.2. Measurements (in mm) of the femur and ulna of Megapodius. 14.3. Tarsal length (in mm) from skins of selected subspecies of Megapodiusfreycinet. xv 155 169 173 174 176 177 178 180 181 182 183 184 190 191 192 193 194 195 196 199 200 201 203 205 207 207 208 209 210 210 212 213 218 224 225 10.3. Frequency percentages of plagioclase feldspar and opaque iron oxide mineral grains and basaltic volcanic lithic fragments in sherds containing "feldspathic basaltic temper." 155 10.4. Frequency percentages of calcareous grains and silicate-oxide mineral grains and lithic fragments for sherds containing "mixed temper sand." 12.1. Artifacts analyzed by EDXRF. 12.2. Evaluation of analytical accuracy and precision for U.S.G.S. Standard RGM-1. 12.3. Variations of oxides and elements from the Mo'omomi adz quarry, Moloka'i, Hawaiian Islands. 12.4. Geochemistry of Tatagamatau source rock (fused disks). 12.5. Geochemistry of Mako ridge source rock (fused disks). 12.6. Geochemistry of Fa'ala'aga source rock (fused disks). 12.7. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (fused disks). 179 12.8. Geochemistry of Tatagamatau source rock (whole specimen). 12.9. Geochemistry of Mako ridge source rock (whole specimen). 12.10. Geochemistry of Fa'ala'aga source rock (whole specimen). 12.11. Geochemistry of Fa'ala'aga, Mako ridge, and Tatagamatau source rock (whole specimen). 12.12. Geochemistry of To'aga and Ofu Island artifacts (whole specimen) 13.1. Summary of To'aga faunal remains. 13.2. Fish fauna from the 1987 field season (NISP). 13.3. Fish fauna from the 1989 field season (NISP). 13.4. To'aga fish fauna: main trench, Units 1,4-9. 13.5. To'aga fish fauna: transect 5, Units 15/29/30 (NISP). 13.6. To'aga fish fauna: transect 9, Units 20/23 (NISP). 13.7. Modern Samoan fishing methods (after Hill 1986). 13.8. Non-fish vertebrate fauna for 1987 and 1989 excavations. 13.9. To'aga non-fish vertebrate fauna. 13.10. To'aga invertebrate data: transect 9, Units 20/23. 13.11. To'aga invertebrate data: 1987 main trench, Units 1,4-9. 13.12. To'aga invertebrate data: transect 5, Units 15/29/30. 13.13. Unit 20/23 bulk sample analysis. 13.14. Unit 30 bulk sample analysis. 13.15. Taxa represented in bulk sample from Units 20/23 (NISP). 13.16. Taxa represented in bulk sample from Layer II, Unit 30 (NISP). 13.17. Density of identified fish bone from Layer IIIA/B, Unit 20/23. 13.18. Density of identified fish bone from Layer II, Unit 30. 13.19. Summary of recovery and quantification techniques for Western Polynesian faunal analyses. 13.20. Summary of Western Polynesian vertebrate faunal assemblages (NISP). 14.1. Birds from the To'aga site. 14.2. Measurements (in mm) of the femur and ulna of Megapodius. 14.3. Tarsal length (in mm) from skins of selected subspecies of Megapodiusfreycinet. xv 155 169 173 174 176 177 178 180 181 182 183 184 190 191 192 193 194 195 196 199 200 201 203 205 207 207 208 209 210 210 212 213 218 224 225 1 INTRODUCTION AND RESEARCH DESIGN PATRICK V. KIRCH AND TERRY L. HUNT 1THE WESTERN POLYNESLAN ISLANDs-and particularly the large Tongan and Samoan archipelagoes-have long occupied a central focus in scholarly endeavors to decode the history of Polynesian origins and dispersals. Burrows (1939) recognized and defined the distinctive cultural pattems that set the Westem Polynesian societies apart from those of Eastem Polynesia. With the advent of modem stratigraphic excavation and of radiocarbon dating in Polynesia during the 1950s, it became apparent that Tonga and Samoa had been settled considerably earlier than any other Polynesian archipelagoes. Furthermore, pottery assemblages from both Tonga and Samoa had obvious relationships with early ceramic complexes in Melanesia to the west (Golson 1962). Thus, when Suggs attempted his pioneering synthesis of Polynesian prehistory, he could rightly claim that "the islands of Westem Polynesia were the earliest occupied of all the Polynesian triangle and the source of all subsequent settlements in Polynesia" (1960:101). As archaeological knowledge of the Westem Polynesian archipelagoes continued to accumulate, the relationship between the early Tongan and Samoan materials occasioned debate and controversy (e.g. Green 1967, 1968; Groube 1971). The presence of early (ca. 1200 B.C.) dentate-stamped, Lapita-style pottery in Tonga and, in Samoa of only later (ca. 200 B.C. tO A.D. 300) Polynesian Plain Ware assemblages appeared to support the hypothesis of Tonga as the original Polynesia "homeland." In 1973, the discovery of a submerged Lapita site at Mulifanua off the 'Upolu coast in Westem Samoa (Jennings 1974) brought the Samoan sequenc back to a comparable antiquity to tat of Tonga. Subsequently, ealy Lapita materials were recovered from a number ofthe smaller Westem Polynesian islands, including Futuna (Kirch 1981; Frimigacci 1990), 'Uvea, and Niuatoputapu (Kirch 1988b). These discoveries indicate that the entire Westem Polynesian region was rapidly explored and colonized by Lapita people during the penultimate centries of te second millennium B.C. In our current conception, the adjacent islands and archipelagoes of Western Polynesia constituted a homeland region in which a distinctive Ancestral Polynesian Culture emerged out of its immediate Lapita ancestor during the period from about 1000 to 500 B.C. (Kirch 1984; Kirch and Green 1987). Presumably, the processes of cultural change and differentiation during this time period were not uniform across all islands. Regular inter-island contact through exchange, however, kept each local community from becoming isolated and facilitated the spread of innovations. As Kirch and Green (1987) have argued, the reconstruction 2 The To'aga Site of Ancestral Polynesian Culture-not as a static entity, but as a dynamic and changing configuration over the course of 500 or more years-is a high priority for Polynesian prehistory. Only by understanding the technology, economy, settlement pattems, and socio-political organization of this ancestral culture can we provide a secure baseline for studying the subsequent development and diversification of later Polynesian groups throughout the vast Polynesian triangle. Reconstruction of Ancestral Polynesian Culture is essential if we are to disentangle those traits and institutions which are shared retentions, from those which are later innovations in various island societies. The present monograph represents a modest contribution toward the ultimate goal of tracing the development of Ancestral Polynesian Culture out of its Lapita roots. The To'aga site, situated on the tiny island of Ofu in the Manu'a Group of American Samoa, spans virtually the entire threemillennium-long sequence of Samoa, but is especially rich in well-stratified materials dating to the period from ca. 3200 to 1900 B.P. Our research at To'aga was initiated in order to meet certain cultural resource management (CRM) demands, as part of an archaeological survey and inventory of American Samoan sites under the auspices of the Office of Historic Preservation of the Govenmment of American Samoa (Pago Pago). Fortunately, the goals of this CRM project happily meshed with the objectives of academically oriented archaeological research, providing the opportmity to add to our knowledge of the early phases of Westem Polynesian prehistory while at the same time addressing contemporary historic preservation and land use management concems. The To'aga site was discovered during our 1986 reconnaissance survey of the Manu'a Islands (Hunt and Kirch 1987), one of two sites containing pottery and thus dating to the earliest period of Samoan prehistory. In our subsequent 1987 and 1989 field seasons we concentrated on the extensive To'aga site, seeking to define the spatial extent of its deeply buried deposits, its chronology, stratigraphy, material culture, faunal assemblages, and other aspects of intra-site variation. As our investigations progressed, it became apparent that To'aga was of critical importance to understanding the first millennium of Samoan- and Westem Polynesian-prehistory. No other site presently known in the Samoan archipelago encapsulates such a continuous stratigraphic record, nor one so rich in artifactual and faunal materials. Although the scope and extent of our excavations had to be limited by the CRM-funded nature of our project, we have nonetheless made every effort to push the analysis of our materials farther than is usual in such reconnaissance surveys. For example, using a transect sampling methodology, combined with detailed sediment analysis and radiocarbon dating, we have been able to test a model of shoreline progradation linked with a mid-Holocene higher sea level. This model has considerable geoarchaeological implications for the formation and subsurface burial of early Polynesian sites thrughout the tropical central Pacific region. We have also pushed the analysis of the archaeological record of Ancestral Polynesian material culture at the To'aga site through detailed studies of pottery (including analysis of temper and of the chemical composition of clays), adzes (including EDXRF analysis of basalts), and other artifact classes. Likewise, the samples of faunal materials have been subjected to intensive studies, including extinct and extirpated avifaunal remains, molluscan and fish faunal assemblages, and terrestrial (synanthropic) land snails. In the fourteen chapters that follow, we and our collaborators present the detailed results of our field and laboratory investigations of the To'aga site. BACKGROUND TO SAMOAN ARCHAEOLOGY Although Buck (1930) reported on stonework and adzes, the first modem archaeological effort in Samoa was that of Golson (n.d., 1962) in 1957, resulting both in a general account of the range of field monuments and in the discovery of prehistoric pottery dated to the first century A.D. at Vailele, 'Upolu. The latter discovery was particularly significant in the then-emerging picture of Polynesian origins as rooted in an earlier Melanesian ceramic complex beginning to be known by the term "Lapita" (Suggs 1961; Golson 1971; see also Kirch 1988a). Following Golson's lead, a major archaeological program was organized for Westem Samoa under the direction of Introduction and Research Design Roger C. Green (University of Auckland), with funding provided trugh the Bishop Museum's Polynesian Culture History Program. Between 1962 and 1967, this project brought seventeen archaeologists from nine institutions to Westem Samoa for a coordinated series of investigations including surface surveys and excavations of sites spanning early settlement to the historic period. Published in two large volumes (Green and Davidson 1969, 1974), the results of this project are a landmark in Polynesian archaeology. Subsequent to the conclusion of the Westem Samoa project in 1967, but in time to be incorporated in the second volume of results (Green and Davidson 1974), an accidental discovery of classically decorated Lapita pottery at Mulifanua, 'Upolu extended the Samoan sequence back to the beginning of the first millennium B.C. (fig. 1.1). Furthermore, the geomorphological context of this find-a submerged site capped by nearly one meter of reef rock-demonstrated that tectonically induced changes in the Samoan landscape could have significant implications for regional prehis- tory. In the 1970s, a project headed by Jesse Jennings of the University of Utah completed two seasons of archaeological research on 'Upolu 3 Island and on the adjacent small islet of Manono, concentrating on settlement pattem surveys and excavation of both plain ware and aceramic sites (Jennings et al. 1976; Jennings and Holmer 1980). Of particular note is their work on settlement patterns in which they propose the concept of the "household unit" as an analytical category (Jennings et al. 1982). As a result of these various projects, the outline of a prehistoric cultural sequence for Westem Samoa is reasonably well attested (Davidson 1979). In conventional outline, this sequence begins with the occupation of the archipelago between about 1200-1000 B.C. by makers of classic, dentate-stamped Lapita pottery (represented by the presently submerged "Ferry-Berth Site" at Mulifanua). The first millennium B.C. witnessed changes in the composition of Samoan ceramic assemblages, particularly the loss of decoration and of more complex vessel shapes, ending with Polynesian Plain Ware assemblages around the time of Christ. Stone adzes and other aspects of material culture also changed with the pottery, and this sequence as a whole is viewed as documenting the development of an Ancestral Polynesian Culture out of an older Eastem Lapita culture (Kirch and Green 1987). Figure 1.1 The Samoan Archipelago. 4 The To'aga Site Ceramnics ceased to be manufactued by about A.D. 300, and e paucity of archaeological materials during the next thousand years or so prompted Davidson (1979) to tenn this the "Dark Ages" of Westem Polynesian prehistory. In the final millennium of the Samoan sequence a number of new developmenas are evidenced in the archaeological record, especially the construction of several new fonns of field monuments, such as the star mound and ridge-top forifications. These large constructions are believed to reflect the rise of powerful chiefdoms competing for land and resources. The Samoan culurl sequence briefly oudined above parallels-in many key aspects-the sequences developed by archaeologists for other Westem Polynesian islands, such as Tongatapu (Poulsen 1987), Niuatoputapu (Kirch 1988a), Futuna (Kirch 1981; Frimigacci 1990), and 'Uvea (Frimigacci, Siorat and Vienne 1984). In contrast with Westem Samoa, the archaeology and prehistory of American Samoa are less well known. The first modem archaeological survey in Amercan Samoa was caried out by W. Kikuchi (1963, 1964) on Tutuila and Aunu'u Islands in 196162. Kikuchi's work provided an overview of the main kinds of surface sites, but was neither intensive nor systenatic in its coverage. In 1962, Kikuchi and Y. Sinoto of the Bishop Museum extended the survey to the Manu'a Group and conducted minor test excavations on both Ta'u and Tutuila Islands. Disappointed by their results, Sinoto decided to abandon the Samnoan project, and shifted his locus of field work to the Marquesas Islands of Eastem Polynesia (Emory and Sinoto 1965). Prior to 1980, the only other major field project conducted in American Samoa was that of Janet Frost (1978), who carred out limited test excavations at seven sites on Tutuila. There have also been severl limited cultual resource surveys, carried out under contract to the National Park Service, the U.S. Anmy Corps of Engineer, and the Department ofPublic Works (Ladd and Moris 1970, Kikuchi, Palama, and Silva 1975; Silva and Palama 1975; McCoy 1977; Atxhns 1987; Hunt 1987). In 1980, J. Clark compiled a summary of all recorded archaeological sites for the American Samoan Historic Preservadon Commission, based on ree weeks in the field checking the locations and status of many of these sites (Clark 1980). Clark summed up the status of archaeological survey in Amercan Samoa as of 1980 in these words: "the amount of land that has been intensively and systematically covered is small indeed" (1980:11). Over the course of the past decade, archeological knowledge of Amercan Samoa has increased substantially due to the efforts of te Historc Preservadon Office ofthe Deparment of Pariks and Recreation, Government of American Samoa. With funds made available by the U.S. National Park Service (NPS), the Historic Preservation Officer has commissioned a variety of archaeological reconnaissance and intensive survey projects in order to compile an inventory of significant sites on Tutuila and in the Manu'a Group. Given the rapid pace of economic development in American Samoa, many of the sites revealed through such survey projects have proved to be threatened by current or projected land use practices. On Tutuila Island, a substantial multi-year project was devoted to the recording and detailed study of the Tataga-matau basalt adz quarry site, situated on a complex of ridges inland of Leone Village (Best, Leach, and Witter 1989; Leach and Witter 1987; Leach and Witter 1990). This site is of interest not only for its intrinsic importance to understading the prehistory of Tutuila, but because the adzes produced at this major quany have been shown to have been distributed very widely throughout the Westem Pacific (Best, Sheppard, Green, and Parker 1992). Also on Tutuila, Jeff Clark and his associates have carried out a series of surveys and test excavations which have greatly amplified our understanding of the archaeological resources of American Samoa (Clark 1989, in press; Clar and Herdrich 1988, in press). When the senior author was asked by te American Samoan Historic Preservation Officer to consider undertaking an archaeological survey in the archipelago, we decided to focus on the thenneglected Manu'a Group. The selection of the Manu'a Group as the focus of our project was motivated by several factors. First, the larger island ofTutuila was already receiving substantial attention from several other archaeological field teams, leaving Manu'a as a continuing lacuna. Second, because we were particuladiy interested in seeking ceramicbearing sites dating to the first third of Samoan prehistory, we preferred to focus on several smaller islands where survey and subsurface testing could be concentrated on likely areas of early occupation. Introduction and Research Design Third, as the most easterly and somewhat isolated section of the Samoan archipelago, the prehistory and archaeology of Manu'a could conceivably exhibit significant differences from that of Tutuila, Upolu, and Savai'i. While such differnces were predictable, primary fieldwoik would be necessary for their documentation. DESIGN OF THE RESEARCH Over the course of three field seasons, our research focus in Manu'a evolved from extensive reconnaissance survey of the surface archaeology to intensive subsurface examnination of buried archaeological resources. Our field investigations began in 1986 with archaeological reconmaissance of he three islands of the Manu'a Group: Ta'u, Olosega, and Ofu (Hunt and Kirch 1987, 1988). In 1986 ceramicbearing sites were discovered at To'aga, Ofu Island and at Ta'u Village, Ta!u Lsland. In the 1987 field season, we focused on an intensive survey and systematic excavation of the site at To'aga (Site AS13-1). Our msults from 1987 (Kirch et al. 1989, 1990) revealed a deeply stratified site containing a long and continuous sequence of ceramics and dating to more than 3,000 yearS B.P. (Kirch et al. 1990). A site of this significance for either American or Westem Samoa was previously unknown. A third season at To'aga was therefore designed to determine the naure and extent of bured archaeological remains for the entire coastal flat of southem Ofu (extending from To'aga to Fa'ala'aga). Thus, the primary objectives of our 1987 and 1989 fieldwork were to determine the nature, significance, and spatial and stratigrphic extent of the deposits at site AS-13-1. This fundamental step was accomplished primarily for the purpose of cultural resource management, including preservation and public interpretation. The data collected will be used to nominate site AS-13-1 to the National Register of Historic Places and to assure its preservation in the face of any future development plans. While our Manu&a Project was designed first and foremost to address the CRM concerns of the Amercan Samnoa Historic Preservation Office, we also regarded the project as an opportunity to tackle several major research problems of Samoan archaeology and prehistory. Especially in the 1987 and 1989 field seasons, when we concentrated on the well- 5 stratified and extensive subsurface deposits at To'aga, it was possible to design the field and laboratory sttegies to address the following researh problems: 1. A major objective was the establishment of a temporal framework and prehistoric sequence for the Manu'a Islands. Prior to the commencement of our project, it was uncertain whether the Manu'a Group would prove to have been colonized at approximately the same time as the other Samoan Islands, or whether the main changes and trends in te Manu'a sequence (such as te timing of ceramnic change and eventual cessation of pottery manufacture) would parallel ose on Tutuila, Upolu, and Savai'i. 2. A second research goal was to determine the nature and magnitude of environmental change during the period of prehistoric Polynesian occupation of the Manu'a Islands. Interdisciplinary research on other central Pacific islands over the past two decades had shown that human activities frequently have resulted in major changes to the vegetation, fauna, and landforms of island ecosystms (e.g., Kirch 1984:123-51; Kirch 1988b:247-50-, Kirch and Yen 1982; Bayliss-Smith et al. 1988:1243). At te same time, natural environmental changes such as sea-level fluctuations, also were known to have affected site distribution and the geomorphology of coastal lowlands. At te To'aga site, we had te opponunity to address such issues of human-induced and natnal environmental changes over a 3,000+ year sequence. In particular, we were interested in determining the geomorphological history of te To'aga coastal plain, a narrow strip of intensively used land (see chapter 2, Kirch). To this end, we developed a morphodynamic model of landfonn changes at To'aga (chapter 4, Kirch) which we were able to test on te detailed data of site satigraphy (chapter 5, Kirch and Hunt), radiocarbon chronology (chapter 6, Kirch), and sedimentological analysis of fte excavated deposits (chapter 7, Kirch, Manning, and Tyler). 3. A third major issue concerned the reconstuction of certin aspects of Ancestral Polynesian Culture, especially te nature of its settement pattems and subsistence economy, which were rather poorly evidenced on archaeological criteria (as opposed to histonrcal linguistic reconstructions, see Kirch 1984:53-67). Because the To'aga site spans the entire firSt millennium B.C.-the period during which Ancestrl Polynesian Culture developed out of its 6 The To'aga Site Lapita ancestor-and because the site's calcareous sandy deposits preserve a wide range of organic materials, this was again an excellent situation in which to tackle this researh issue. In particular, we expended considerable effort in the analysis of the extensive suite of vertebrate and invertebrate faunal matetals rwcovered from the excavations (see chapter 13, Nagaoka and chapter 14, Steadman). 4. A more specific research topic is the explanation of ceramic change in Westem Polynesia, including aspects of technology, formal variation in vessel shape and function, and te eventual disappearance of pottery. The To'aga site, with a particulariy long sequence of strtified deposits, yielded a cermnic assemblage spaning virtually the entire period of pottery manufacture in Samoa. We fterefore decided to focus on a detailed examination of these ceramics, from several analytical perspectives, in order to refine our understanding of ceramic change in the Samoan archipelago (see chapter 9, Hunt and Erkelens; and chapter 10, Dickinson). 5. A fifth research issue focussed on te role of inter-island exchange of such matenal items as ceranics and basalt adzes. Oceanic prehistorians have increasingly come to realize that such exchange was extremely important in maintaining contacts between island societies (e.g. Kirch 1988a, b; Hunt 1989). In the case of materials frxm To'aga, we resolved to investigate the role of inter-island exchange trugh an intensive study of the ceramic and basalt adz artifact assemblages (see chapter 9, Hunt and Erelens; and chapter 12, Weisler). In sum, our field and laboratory studies of the To'aga site was oriented by these five research issues, combined with the prblems of site delineation and significance-assessment dictated by the CRM nature of our conta with the American Samoa Historic Preservation Office. Because the funds for fieldwork were limited, we knew that we would be unable to carry out extensive subsurface excavations or exposures of large areas. lTus, our field strategy had to be designed to obtain the kinds of data relevant to the research issues cited above in the most costeffective manner. In addition to intensive survey of surface archaeological featumr located throughout the Toaga area (Hunt, chapter 3), we concentaed on a series of systematic trsct test excavations following methods elaborated by Kirch in previous fieldwork on Ticopia and Niuatoputapu (Kirch and Yen 1982; Kirch 1988b). These unsect excavations allowed us to define the spatial extent ofthe site, especially the deeply buried, pottery-bearing deposits, to gain an overview of the statigraphic sequence and to obtain sizable samples of ceramics, other portable artifacts, and vertebrate and invertebrate faunal remains. Furtfer details of our sampling and excavation metiodology are provided in chapter 5. As the various chapters to follow demonstrate, the To'aga site is a unique and highly significant archaeological resource, encapsulating dtree millennia of Samoan prehistory. As the pace of development and land use change quickens in American Samoa, this site is likely to come under increased dteat. It is our hope that in addition to contributing to our understanding of Polynesian and Samoan prehistory, this volume will serve to heighten awareness concerning this valuable archaeological resource. The To'aga site deserves to be carefully protected and its non-replaceable resources managed so that future generations will have the opportunity to fuirter advance our knowledge and understanding of the past REFERENCES CITED Athens, J. S. 1987. Archaeological survey of Leone Bay, American Samoa. Unpublished manuscipt, Office of Historic Preservation, Pago Pago. Bayliss-Smith, T. P., R. Bedford, H. Brookfield, and M. Latham 1988. Islands, Islanders, and the World: The Colonial and Post-Colonial Experience ofEastern Fiji. Cambridge: Cambridge University Press. Best, S. B., H. Leach, and D. Witter 1989. Report on the second phase of fieldwork at the TatagaMatau Site, American Samoa, July-August 1988. Unpublished manuscript, Office of Historic Preservation, Pago Pago. Best, S., P. Sheppard, R. Green, and R. Parier 1992. Necromancing the stone: Archaeologists and adzes in Samoa. Journal of the Polynesian Society 101:45-85. Buck, P. H. (Te Rangi Hiroa) 1930. Samoan Materill Culture. Honolulu: Bemice P. Bishop Museum Bulletin 75. Burrows, E. G. 1939. Westem Polynesia: A study in cultural differentiation Ethnological Studies 7. Gothenburg. Introduction and Research Design Clark, J. T. 1980. Historc preservation in Amencan Samoa: Program evaluation and archaeological site inventory. Unpublished manuscript, Deparnment of Antrpology, Bishop Museum, Honolulu. . 1989. The Eastem Tutuila archaeological project: 1988 final report. Unpublished manuscript, Office of Historic Preservation. Pago Pago. -. m press. Radiocarbon dates frm American Samoa. Radiocarbon. Clark, J. T., and D. J. Herdrich 1988. The Eastem Tutuila archaeological project: 1986 final report. Unpublished manuscript, Office of Historic Preservation, Pago Pago. in press. Prehistoric settlement system in Eastem Tutila, American Samoa, Journal ofthe Polynesian Society. Davidson, J. 1979. Samoa and Tonga. IN J. Jennings, ed., The Prehistory ofPolynesia, pp. 82-109. Cambridge: Harvard University Press. Emory, K. P., and Y. Sinoto 1965. Preliminary report on the archaeological investigations in Polynesia. Mimeographed report on file, Department of Anthropology, Bishop Museum, Honolulu. Frimigacci, D. 1990. Aux Temps de la Terre Noire: Ethno-ArchAologie des les Futuna etAlofi. Paris: Peeters. Frimigacci, D., J. P. Siorat, and B. Vienne 1984. Inventaire et Fouille des Sites Archeologiques etEthnohistoriques de l'lle d'Uvea. Documents Provisoires, ORSTOM, Centre de Noumea, New Caledonia. Frost, J. 1978. Archaeological investigations on Tutuila Island, American Samoa. Unpublished Ph.D. diss., University of Oregon. Golson, J. 1962. Report on New Zealand, Westem Polynesia, New Caledonia, and Fiji. Asian Perspectives 5:166-80. -. 1971. Lapita ware and its transfonnations. IN R. Green and M. Kelly, eds., Studies in Oceanic Culture History, pp. 67-76. Pacific Andtopological Records 12. Honolulu: Bernice P. Bishop Museum nd. Report to TRIPP on archaeological fieldwork in Samoa and Tonga. Unpublished manuscript, University of Auckland. Green, R. C. 1967. The mmediate origins of the 7 Polynesians. IN G. A. Highland, et al., eds., Polynesian Culture History, pp. 21540. Bemice P. Bishop Museum Special Publication No. 56. Honolulu. .1968. West Polynesian prehistory. IN I. Yawata and Y. H. Sinoto, eds., Prehistoric Culture in Oceania, pp. 99-110. Honolulu: B. P. Bishop Museum. Green, R. C., and J. M. Davidson, eds. 1969. Archaeology in Western Samoa, Vol. 1. Auckland Institute and Museum Bulletin 6. 1974. Archaeology in Western Samoa, Vol. 2. Auckland Institute and Museum Bulletin 7. Groube, L. 1971. Tonga, Lapita pottery, and Polynesian origins. Journal of the Polynesian Society 80:278-316. Hunt, T. L. 1987. Archaeological survey and assessment of the proposed Fiti'uta Airport site, Ta'u Island, Manu'a Group, Amercan Samoa. Unpublished manuscipt, Depailrent ofPublic Woiks, Pago Pago. 1989. Lapita ceramic exchange in the Mussau Islands, Papua New Guinea. Unpublished Ph.D. diss., University of Washington, Seatde. Hunt, T. L., and P. V. Kirch 1987. An archaeological reconnaissance of the Manu'a Islands, Amercan Samoa. Unpublished manuscipt, Office of Historic Preservation, Pago Pago. 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal of the Polynesian Society 97:153-83. Jennings, J. D. 1974. The Ferry-Berth Site, Mulifanua District, Upolu. IN R.D. Green and J. Davidson, eds., Archaeology in Western Samoa, Vol. 2 pp. 176-78. Aukland Institute and Museum Bulletin 7. Jennings, J. D., R. N. Holmer, J. C. Janetski, and H. L. Smith 1976. Excavations on Upolu, Western Samoa. Pacific Anthrpological Records 25. Honolulu: Bemice P. Bishop Museum Jennings, J. D., and R. N. Holmer 1980. Archaeological Excavations in Western Samoa. Pacific Anthropological Records 32. Honolulu: Bernice P. Bishop Museum Jennings, J. D., R. N. Holmer, and G. Jackmond 1982. Samoan village patems: Four examples. Journal of the Polynesian Society 91:81-102. Kikuchi, W. 1963. Archaeological surface ruins in American Samoa. Unpublished M.A. thesis, 8 The To'aga Site University of Hawaii, Honolulu. -. 1964. Petroglyphs in Amencan Samoa. Journal of the Polynesian Society 73:163-66. Kikuchi, W., Palama, S., and T. Silva 1975. Archaeological reconnaissance survey, proposed Ta'u Harbor at Fusi and quarry site between Fusi and Fagamoto Ta'u Lsland, Manu'a Group, Amencan Samoa. Unpublished manuscript, Deparlment of Anthropolgy, University of Hawaii, Honolulu. Kirch, P. V. 1981. Lapitoid settlements of Futuna and Alofi, Westem Polynesia. Archaeology in Oceania 16:127-43. 1984. The Evoluion of the Polynesian Chieldoms. Cambrdge: Cambridge University Press. 1988a. A brief history of Lapita archaeology. IN P. V. Kirch and T. L. Hunt, eds., Archaeology of the Lapita Cultural Conmpex: A Critical Review, pp. 1-8. Seatte: Tlomas Burke Memorial Washington State Museum Research Repoit No. 5. -. 1988b. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Seatle. Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evolution in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler 1990. An ancestal Polynesian occupation site at To'aga, Ofu Island, American Samoa. Archaeology in Oceania 25:1-15. Kirch, P. V., T. L. Hun, and J.Tyler 1989. A radiocarbon sequence from te To'aga Site, Ofu Isand, American Samoa. Radiocarbon 31:7-13. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory andEcology of a Polynesian Outlier. Honolulu: Bernice P. Bishop Museum Bulletin 238. Ladd, E. J., and D. K. Moris 1970. Archaeological and ecological survey of 'Olovalu Crater, Island of Tutuila, American Samoa. U.S. National Pa& Service. Unpublished manuscript. Leach, H. M., and D. C. Witter 1987. Tataga-matau 'rediscovered.' New Zealand Journal of Archaeology 9:33-54. . 1990. Further investigations at the Tataga-matau Site, American Samoa. New Zealand Journal ofArchaeology 12:51-83. McCoy, P. C. 1977. Cultural reconnaissance survey, Auasi Harbor Project, Auasi, Tutuila Island, Amercan Samnoa. Unpublished manuscript, Department of Anthrpology, University of Hawaii, Honolulu. Poulsen, J. 1987. Early Tongan Prehistory: The Lapita Period on Tongatapu and its Relaionships. 2 vols. Terra Ausalis 12. Canberra: Australian Natonal University. Silva, T., and S. Palama 1975. Archaeological reconnaissance survey, proposed shoreline and highway improvements, Tutuila Island, and Aunu'u Boat Harbor, Aunu'u Island, American Samoa. Unpublished manuscrpt, Archaeological Research Center Hawaii. Suggs, R C. 1960. The Island Civilizaions of Polynesia. New York: Mentor Books. . 1961. Archaeology of Nuku Hiva, Marquesas Islands, French Polynesia. New York: Anthrpological Papers of the American Museum of Natural History 49(1). OFU ISLAND AND THE TO'AGA SITE: DYNAMICS OF THE NATURAL AND CULTURAL ENVIRONMENT PATRICK V. KIRCH Every Manuan thinksfirst ofhis own village, then ofhis island, then ofManua as a whole-a tiny archipelago set offfrom all the rest of Samoa. There may be rivalry between villages; there have even been warsinternecine conflicts between Fitiuta and the three villages ofTau and between OfQu and Olosega-but against the outside world Manua presents a solidfront. Mead (1930:51) THE MANU'A ISLANDS, ISOLATED from the main Samoan islands by 100 kn of oftenturblent ocean, form a geographically and culturally disinctive cluster at the eastem extreme of the archipelago (fig. 2.1). Being closest to Tutuila, Manu'ans had the greatest interaction with the occupants of that island (a pattern that continues today with Manu'a and Tutuila comprising the Terntory of American Samoa). Although sharing in most respects the dassic characteristics of Samoan culture, Manu'ans nonetheless regard themselves as different and distinctive. Mead (1930:9) commented on this distinctiveness, for example in the lack of emphasis on "war, its paraphemalia, its ritual, and its gods." She opined that "the chief historical value of Manua lies in her easterly and isolated position, offering a valuable check upon cultural traits which are intrusive in westem Samoa" (1930:9). Such isolation, however, was relative and should not be overemphasized; as our archaeological investigations have revealed, the prehistory of Ofu shares much in common with the westerly islands of Samoa, and prehistoric inter-island contacts can be documented in the transport of basalt adzes (see Weisler, chapter 12), and perhaps of pottery as well (see Hunt and Erkelens, chapter 9). Thus, despite considerable isolation, Manu'a was never a "closed system." The ihree islands of the Manu'a Group-Ta'u, Olosega, and Ofu-form an intervisible cluster and comprised a political unity in ancient times. Ta'u is by far the largest island (table 2.1) and dominated the group politically, being the seat of the Tui Manu'a paramount line of chiefs. Olosega and Ofu are practically adjacent to each other, separated only by a narnow and shallow strait, now spanmed by a concrete causeway. They are similar in size, Ofu being slightly smaller (3.4 kM2), but rising to a loftier summit (at 638 m) than Olosega. Ofu and Olosega are visually exquisite, their youthful volcanic profiles dthust abruptly out of dte Pacific, rising to basalt pinnacles often shrouded in misL Narrow ribbons of blindingly white coral sand fonn a no-man's-land between the rank green of tropical vegetation and the blue-green mosaic of coral reefs. Sheer cliffs towering hundreds of meters above To'aga, Sili, and other villages provide nesting sites for the white-tailed tropic birds that ride the trade winds battering the islands' mass. 10 The To'aga Site Figure 2.1 Map of the Manu'a Isands. Perhaps the most striking feature of Ofu's environment is its precipitous topography, and the scarcity of flat land. No less than ninety-one percent of the total island area consists of slopes greater than 30 degrees. Due to its geological youth, the steep sides of the volcanic cone have not yet been significantly incised by streams, the principal weathering being major landslides resulting from faulting, and creating sheer cliffs. The only level terrain is found around the margins of the island, at the critical land-sea interface, where a narrow coastal terrace has been constructed primarily from coral sand, with the addition of some talus and colluvium from the slopes inland. This coastal zone, ranging from 50 to 150 meters wide, is unevenly distributed, being found along the westem and southeastem coasts but not along the north coast, except for a small strip at the northeastem point. The To'aga site occupies the longest span of coastal terace land on the island, along the southeastem shoreline. This coastal terrace was a critical microenvironment for the Polynesian occupants of Ofu, first as the setting for their villages, and second as a major zone of garden land. More importantly, the narrow coastal terrace is a highly dynamic environment. Understanding the morphodynanics of this land-sea interface is essential not only to reconstructing the prehistoric sequence of the To'aga site itself, but more broadly for an understanding of Manu'an prehistory. In this chapter, I outline the main characteristics and resources of the Ofu Island environment that were significant to the prehistoric Polynesian inhabitants with their horticultural and fishing economy. I also pay considerable attention to those aspects of the contemporary environment-both physical and biotic-that inform us about the dynamics of environmental change. Polynesian archaeologists have now abandoned an earlier perspective that viewed island ecosystems merely as static backdrops to cultural developments. Increasingly, archaeological investigations throughout Polynesia are yielding striking evidence of ecological change: forest clearance and vegetational succession, extinction and extirpation of birds and other biota, alluviation of valley floors, sea level changes, and so forth (e.g., Kirch 1982, 1984; Steadman 1989; Flenley and King 1984; Olson and James 1984; McGlone 1983). Some of these changes resulted from natural processes; others reflect the impact of human populations themselves, colonizing isolated and biologicallyvulnerable islands for the first time (Fosberg 1963). The prehistory encapsulated within the To'aga site is as much a history of environmental dynamics as of cultural changes in pottery Wpes or adz forms. The Natural and Cultural Environment 11 Table 2.1 Environmental Characteristics of the Manu'a Islands Characteristic Ta'u Olosega Ofu Area (kIn) Highest Point (m) Area<30% slope (%) Coastline (kan) Population (1980) 28.5 965 41 32.5 1,146 4.5 494 10 13.3 340 3.4 638 9 10.4 254 GEOLOGY AND GEOMORPHOLOGY OF OFU ISLAND Initial studies of Ofu geology by Friedlander (1910), Daly (1924), and Stearns (1944) were based on short reconmaissance surveys. Although Steams spent only one or two days in Manu'a, he was able to produce "a remarkably accurate geologic sketch map" (Stice and McCoy 1968:429). The most intensive study to date was carried out by Stice and McCoy (1968), whose report and geologic map of Ofu we have drawn upon for the following summary. Subsequent work by Natland (1980) and McDougall (1985) has corrected certain errors made by Stice and McCoy in dating the island's geological formations. Ofu and Olosega comprise an integral geological complex "of volcanic cones that have been buried by lava flows from two coalescing shields" (Stice and McCoy 1968:443). At least five cones make up the Asaga Formation; these include a "composite cone exposed in the cliff behind To'aga," as well as "an explosion breccia cone with an associated intrusive plug at Fatuaga Point" (1968:443-44). The latter plug is exposed as the visually impressive spire at the eastern point of Ofu (fig. 2.2). Stice and McCoy believed that the Asaga Formation was of lower Pliocene age, although this has now been shown to be incorrect (see discussion below). All of these Asaga Fornation cones were subsequently buried by lava of the Tuafanua Formation, from the A'ofa and Sili coalescing shields. "After summit collapse of the shields, volcanism decreased so that a sea cliff about 300 feet high was cut around the islands" (1968:456). In the To'aga area, the high cliff seems to have resulted from a combination of faulting and sea erosion. Finally, renewed volcanism resulted in the construction of the tuff cone of Nu'utele Islet, as well as in several hawaiite and olivine basalt flows that filled in deeply eroded stream valleys on westem Ofu. Whereas Stice and McCoy (1968) regarded the geological history of Ofu to have extended over a fairly long period, from the lower Pliocene through to the late Pleistocene, recent work by Natland (1980) and McDougall (1985) has confirmed that the island is very recent in age. McDougall reports that "these volcanoes are quite youthful, confirmed by unpublished K-Ar ages from this laboratory [Austalian National University] averaging 0.3 Ma for Ofu/Olosega and less than 0.1 Ma for Ta'u" (1985:318). As Natland argues, "there can be little doubt that both Tau and Ofu-Olosega are substantially younger than the shield volcanoes of Tutuila, which were extinct and extensively eroded before drowning of the Pleistocene reefs" (1980:721). Indeed, the Samoan chain illustrates a typical instance of linear, "hot-spot" progressive volcanism, with the islands increasing in age from east to west. (Recent, renewed volcanism on Savai'i is evidently due to the proximity of the westem end of the archipelago to the Pacific Plate margin.) The youthful age of Ofu and Olosega is of considerable importance to the geomorphology and geoarchaeology of the To'aga site, as we shall argue in greater detail in chapter 4. This is because the islands are 12 The To'aga Site still tectonically unstable, due to point-loading on the oceanic crust, and thus are undergoing a phase of subsidence (Menard 1986). The rocks of the island consist primarily of olivine basalts, hawaiites, and ankaramites which were extruded as both pahoehoe and aa lava flows, interbedded with various pyroclastic tuffs and breccias. Intrusive rocks consist primarily of dikes, and of the plug at Fatuaga Point, a "hypabyssal intrusion of ankaramite" (1968:449). A major swarm of near vertical dikes rns through the central spine of Ofu, and is clearly visible in the cliffs behind the To'aga site. Stice and McCoy report that "the razorback ridge of eastem Ofu is the topographic expression of a dike complex about 400 feet wide. The dikes are nearly vertical ... Most are dense basalt, although olivine basalt, ankaramite, and feldspar-phyric basalt also are present" (1968:449). We were particularly interested in this dike complex as a possible source of dense, fine-grained basalt that could have been exploited by the prehistoric occupants of Ofu for manufacture of adzes or other flake tools. The recent road cutting at Fa'ala'aga provided an opporunity to examine this dike cluster at close hand, and to collect samples for XRF compositional analysis (see Weisler, chapter 12). As shown in figure 2.3, the dikes are closely spaced, cutting trugh older, weathered basalts of the Asaga Formation. On the edges of several of these dikes, where they came into contact with the older basalts, we observed glassy "chills" of low-silica volcanic glass ("obsidian"), up to 2 cm thick. Although of poor quality, such volcanic glass could have been exploited for the production of small flakes. The appearance of local volcanic glass flakes in archaeological contexts in the Manu'a Group is thus a distinct possibility. We did not observe other possible sources of fine-grained basalts suitable for adz manufacture, except for a rather weathered dike exposure on the edge of Mako Ridge at about 350-400 m elevation (also sampled for XRF analysis). It is entirely possible, however, that prehistoric Manu'ans exploited small exposures of suitable basalt or hawaiite for adz manufacture. We are not aware of any adz quaries in the Manu'a Group, but this negative evidence certainly does not preclude local adz manufacture. There has been relatively little stream erosion on Ofu, and the radial drainage pattem is poorly developed. A number of shallow valleys drain to the north and to the west from the slopes of Tumutumu Mountain. All of these are intennit- tent, flowing only after heavy rains. "'The stream valleys are all youthful and nowhere exceed 50 feet in depth" (Stice and McCoy 1968:455). The main erosional forces at work on Ofu since the cessation of volcanism have been marine wave attack and mass wasting, especially landslides. As noted above, an extensive sea cliff more than 100 m high "was carved into the island by the sea" (Stice and McCoy 1968:455) during the late Pleistocene to early Holocene.. These high cliffs, such as the one behind the To'aga site, "originated by faulting and/or foundering," but were certainly extended and modified by marine erosion. Landslides continue to be active erosional forces, contributing to the talus screes that border the inland edge of the coastal terrace. "Individual blocks also work loose from the cliff face and fall, fonning talus slopes that extend almost continuously around the islands at the base of these cliffs" (1968:455). One such massive block fell between our 1987 and 1989 field seasons: measuring at least 4 m in diameter, this giant boulder crashed through the banana and breadfruit orchards near the western edge of To'aga to land by the roadside, a daily reminder of the geological dynamism of Ofu. Similar large blocks dot the surface of the To'aga site, and one of these was modified in late prehistory for use as an adz grinding stone (see Hunt, chapter 3). For the prehistoric occupants of the To'aga site, such falling boulders posed a continual hazard. As Stice and McCoy relate, a local legend tells of a young girl who "was killed by a large block that rolled across the reef at Sili, where she was fishing" (1968:456). We have no reason to doubt the veracity of this story. RESOURCES OF THE ISLAND For sive rootaso arboricubw-. b taditionally upon inten- bphoriculture, and on s more than an abstract The Natural and Cultural Enviromnent Figure 2.2 13 View of the exposed volcanic plugs at Fa'ala'aga, from the beach at To'aga. The high mountain on the right is the summit of Olosega Island. constant, as rain is abundant and well distributed, 0 Figure 2.3 1 2m Sketch of basalt dikes exposed in the road cut across the Le'olo Ridge at Fa'ala'aga. and as the islands are surrounded by the ocean, it follows that the atmosphere is moist at all times and seasons" (1930:17). Indeed, relative humidity usually ranges between 80 and 86 percent, with the temperature varying only between 25.7-26.20C. Buxton also observed the effects of such humidity on humans: "the perspiring entomologist soon learns to recognize that his comfort depends entirely on the wind, for the warm damp air does not cool him unless it is in motion" (1930:18). Green (1969:4) echoed these remarks from the viewpoint of the field archaeologist, quoting Curry (1962) to the effect that the Samoan climate is at "the upper limit of thermal comfort for half-naked men at rest." concept. The Samoan archipelago lies within the humid tropics. Buxton, who spent much time studying the natural history of Samoa, summed up the climate thusly: "As the temperature is nearly There is, however, a distinct and significant seasonality to the Samoan climate, expressed most clearly in patterns of rainfall and of wind direction. We have no rain gauge records for Ofu itself, 14 The To'aga Site but the records for Tutuila Airport (Nakamura 1984, table 1) are probably similar, indicating a total annual average of 124 inches [3100 mm] (see also Coulter 1941:10-11). (Buxton gives 2,738 mm [107.8 in] for Apia [1930:16].) There is a distinctly drier season from about June through September, in which the monthly rainfall averages about 6 or 7 inches [175 mm], and a wet season from October through May, with montly averages of from 11-14 inches [350 mm]. Even during the "dry season," however, torrential downpours may occur, as beset us all too frequently during the 1989 fieldwork at To'aga. As elsewhere in Western Polynesia, this seasonality played an important role in scheduling the agricultural calendar, particularly the clearing and planting of swidden cultivations (Kirch 1978, forthcoming). In some years, the dry season may be particularly acute or lengthy, and the resulting drought can significantly reduce agricultural yields. During the wet season, the opposite may occur, with torrential rains and serious flooding. "Some floods are associated with hurricanes and tropical storms, but flooding can occur at other times as well" (Nakamura 1984:3). The dry season also corresponds approximately to the period of prevailing trade winds (from about April through September). During October to March the winds are more variable, and westerly reversals occur. This latter period was important to early Polynesian voyagers, as it allowed exploratory voyages from west to east (Finney 1985; Irwin 1981, 1992). In addition to the annual pattem of seasonality, there are stochastically recurring environmental hazards that seriously affect Samoan life. We have already mentioned the problem of periodic drought, which can inhibit agricultural production. Even more severe are the hurricanes or tropical cyclones that periodically lash the islands. Coulter notes that "Samoa suffers hurricanes at irregular intervals during the hotter season" (1941:12). Visher (1925:27, table 6) indicates an average annual frequency of two to three hurricanes in the Samoan area, although not all of these are equally intense, nor do their paths always cross the Manu'a Group. When a hurricane bears directly down on the islands, however, the effects are often devastating-to crops and orchards, to houses, and to human life itself. The most recent hurricane to lash Manu'a occurred early in 1987, between our first and second field seasons. Ta'u Island was particularly hard hit, with virtually every house destroyed; the area was declared a Federal disaster area. Flying in to Ta'u in June of 1987, one was struck by the devestation of the forest cover of the central volcanic cone which had not yet recovered. Aside from their obvious significance to the prehistoric Samoan population, hurrcanes have considerable archaeological importance as agents of landscape transformation, and of site formation as well as destruction. The torrential rains unleashed during these events can cause severe flooding and result in major landslides and in the deposition of colluvium and alluvium. Storm surges and high energy waves are capable of moving large quantities of sand and larger clastics (up to boulder size) in the coastal zone. During our 1986 reconnaissance survey on Ta'u Island, we observed a massive rampart of coral cobbles and shingles at Saua which had been thrown up during a previous storm surge. We shall give further consideration to these high-energy processes below, and elsewhere in this monograph (see especially chapters 4 and 6), as they obviously played a significant role in the geomorphological history of the To'aga site. Ofu soils are young and undeveloped, a reflection of the island's geological youth. The soils have been mapped and described by Nakamura (1984). Most of the steep interior is covered with "Ofu silty clay," a deep, well-drained soil formed in volcanic materials. In the steeper areas (slopes 40 to 70 percent) these silty clays are covered in forest. Where slopes range between 15 and 40 percent, the Ofu silty clays provide the main gardening soils. These gardening areas are confined to two zones on the westem slopes, and to one area on the northern slope of Tumutumu Mountain (fig. 2.4). The very steep slopes and talus regions lying inland of To'aga, and along the northem coast, are described as "Fagasa familyLithic Hapludolls-Rock outcrop association, very steep" (Nakamura 1984:1 1). At the inland edge of the coastal terrace are found strips of "Aua very stony silty clay loam" (1984:10), described as "very deep, well drained soil ... on talus slopes . . . formed in colluvium and alluvium derived dominantly from basic igneous rock." These areas The Natural and Cultural Environment 15 5 ; 0 'I z 0 N- a 0 - I *1 LI" - r44 R- ct :,31 N 0 I 4o 0D ift 0 N. 13 I E CcmJl 0 0 0! 0 4c CY 0 *4 C. S 0 I O-~ j4 z 16 The To'aga Site are used for subsistence gardening, primarily of tree crops such as breadfruit and bananas. Finally, the coastal terrace itself, as at To'aga, consists from the pedological viewpoint of "Ngedebus mucky sand" (Nakamura 1984:15). This is a "somewhat excessively drained soil ... derived from coral and sea shells." The natural vegetation of Manu'a has been significantly modified in the lower elevations by three millennia of human land use. The coastal terrace, as well as the less precipitous mountain slopes, comprise a mosaic of coconut stands, breadfruit and banana orchards, and aroid gardens interspersed with second growth (further discussion of coastal terrace vegetation below). On the higher and steeper slopes, however, the original ain forest vegetation persists. Yuncker, who studied the flora of Manu'a, lists of total of 421 species for the three islands, including mosses and pteridophytes as well as flowering plants (1945:4). The Samoan root-tuber and tree crop complex, of course, is wholly adventive to the archipelago, having been introduced and established by the early colonists. The terrestrial fauna of the Manu'a Group is very restricted in vertebrates, somewhat richer in invertebrates (especially land molluscs and insects). The only indigenous mammal is the fruit bat, Pteropus samoensis. These are commonly seen soaring over the forest canopy high above the To'aga site, especially at dusk. The diminutive Pacific rat, Ratus exulans, was introduced by early Polynesian settlers, and its bones are common in the archaeological deposits at To'aga (see Nagaoka, chapter 13). Also purposively introduced by the Polynesians were the domestic pig (Sus scrofa) and dog (Canisfamiliaris). The richest diversity of vertebrates is among the birds, both native land birds and nesting seabirds (Waading 1982). In the higher elevation forests on Ofu are found the lupe or Pacific Pigeon (Ducula pac4fica) and the Crimson-Crowned Fruit Dove or manutagi (Ptilinopus porphyraceus); both are occasionaUy taken for food. White-Collared Kingfishers, ti'otala (Halcyon chloris manuae), of which there is a distinct Manu'a subspecies, are frequently perched high on poles or telephone wire along the coast. Common in the coastal bush at To'aga and elsewhere on the island is the ve'a or Banded Rail (Gallirallus philippensis), often seen making its characteristic headlong dash across the dirt road from the security of one patch of undergrowth to another. The iao or Wattled Honeyeater (Foulehaio carunculata) is abundant in the banana groves at To'aga, where it feeds on banana flower nectar. This bird is sometimes hunted by young Samoan boys, who then pluck and roast the tiny carcasses over a fire in the bush. (Having sampled this delicacy myself, I can attest that while there is little flesh, it is sweet and a delicious complement to roasted bananas.) Also present on Ofu is the Polynesian Starling, mitivao (Aplonis tabuensis manuae). A number of seabirds, and some migratory species, also nest on Ofu. The White-Tailed Tropic Bird, tava'e (Phaethon lepturus), nests in the high cliffs towering over the To'aga site. These elegant birds, soaring high ovetiead, were always a wonderful visual diversion from the perspective of a grimy test pit. A seasonal migrant, the Golden Plover or tuli (Pluvialis dominicafulva) is also seen along the coast at To'aga. The contemporary avifauna of Manu'a is only an impoverished remnant of the pre-human bird life. Archaeological excavations throughout Polynesia have revealed a significant pattem of bird extinctions and extirpations due to humaninduced forest clearance and habitat destruction, and to direct predation (Olson and James 1984; Steadman 1989; Steadman, Pahlavan, and Kirch 1990; Steadman and Kirch 1990). As reported by Steadman in chapter 14, the To'aga excavations added further evidence of this widespread pattem of avifaunal extinction and extirpation. These faunal depletions are just one aspect of the dynamic nature of Polynesian ecosystems within the span of human occupation. The only other vertebrates indigenous to the island are a number of lizard species in the families Geckonidae and Scincidae. There is a rich endemic and indigenous insect fauna, but this is of little archaeological relevance. More important from the viewpoint of the prehistorian are the land snails, which include a number of endemic and indigenous taxa, as well as several species which have been introduced by humans, prehistorically as well as after European contact. The Natural and Cultural Environmen Among the important families represented in Samoa are Partulidae, Assimineidae, Tomatellinidae, Helicinidae, and Microcystinae. Land snails frequently preserve well in archaeological deposits, including those at To'aga, and are excellent indicators of microenvironmental change. Several synanthropic 'garden-snail' species, including Lamellaxis gracillis, were inadvertently transported around the Pacific basin by prehistoric people, presumably with crop plants and adhering soil. Their appearances in archaeological contexts are therefore important signals of habitat modification, and indirect evidence for horticulture. In chapter 8, I present an analysis of land snails recovered from the To'aga excavations, and their implications for environmental change at the site. Finally, we cannot ignore the marine environment, so crucial to the indigenous Samoan economy. Ofu is surrounded by a fringing reef, widest and most sheltered on the western side (opposite Ofu Village). The reef is a complex mosaic of micro-habitats and a source of shellfish and fish for the human population. A diverse array of molluscs occupy the reef flat and algal crest, including various bivalves such as Periglypta reticulata, Tridacna maxima, Hippopus hippopus, and Asaphis violascens, and gastropods such as Trochus maculatus, Turbo setosus, Nerita spp., Cypraea spp., Drupa spp., Thais armigera, and Conus spp. Many of these species were heavily exploited by the occupants of the To'aga site and occur in dense concentrations in the midden deposits (see Nagaoka, chapter 13). Also present on the reef are spiny lobsters (Panulirus sp.), sea slugs (holothurians), sea urchins (echinoderms), octopus, and various edible seaweeds. Approximately 800 species of inshore fishes occur around Ofu (Jordan and Seale 1906) and are still taken by the Samoans with a variety of traditional fishing strategies, using spears, nets, hooks, and other gear. Farrell's description of fishing activities in Westem Samoa is equally appropriate for Ofu: In the lagoon itself there is some activity at almost any time, day or night. Hundreds of yards from shore, near the edge of the reef, women hunt for sea foods between the breaking of the larger waves. Fishermen in small canoes (paopao) equipped with goggles 17 and spear search the placid waters inshore, while in shallower waters nets are tossed to enshroud passing shoals of small fish. Further out, midway between the shore and the reef, fish traps are built of hard-riven, close-spaced stakes and wire netting, or of mounds of coral rock. Into these traps fish are driven by 'beaters'; later the catch is shared (1962:179). Among the fish commonly taken are jacks (Caranx spp.), parrot fish (Scarus spp.), wrasses (Labridae), and acanthruids. Bones of these fishes also occur in great frequency in the To'aga archaeological deposits. The open sea beyond the reef is less heavily exploited but is the zone of the prized tunas (Scombridae) and flying fish. Marine turtles (Chelonia mydas and Eretmochelys imbricata) are rarely sighted in the waters off Ofu today but must have nested on the island's sand beaches in substantial numbers prior to early Polynesian settlement. The bones of these turtles are one of the most commonly occurring taxa in the To'aga faunal assemblages. THE CULTURAL AND SOCIAL LANDSCAPE As in all Polynesian islands, the landscape of Ofu is culturally and socially ordered. This includes the system of land use, the pattem of land tenure, and the village settlement pattem and its intemal structure, all of which have evolved over several millennia. One goal of our archaeological investigations in Manu'a has been to contribute to an understanding of how this distinctive cultural and social landscape has developed over the course of prehistory. A brief description of some of the key aspects of the contemporary landscape is therefore apposite, as an ethnographic reference or 'endpoint.' Farrell eloquently evokes the essence of the Samoan settlement pattem of "villages ... strung like beads unevenly along the thread of the coastline" (1962:177). Mead elaborates: "The cliff behind, the sea before it, defines the ground plan of a Manuan village, which may spread out in either direction as far as the land permits" (1930:45). On Ofu today, there is only one such village (with two named sectors, Ofu and Alaufau), 18 The To'aga Site strung along the westem coastal terrace. TIhis village lies conveniently between the widest and most sheltered expanse of reef flat and lagoon (protected by Nu'utele and Nuusilaelae Islets), and te largest expanse of arable mountain slope inland. Previously, the zone of coastal villages extended around most of the southem and southeastem parts of th island, including the To'aga area (see Hunt, chapter 3). There is also limited evidence that some occupation may have extended inland into the intensive gardening zone in late prehistorc times. Although in the past decade or two most houses on Ofu have been rebuilt using westem materialslargely concrete with corrugated roofs-the ground plan of the village remains essentially traditional. The central focus is the malae or "village green" where important ceremonies and feasts may be held, the church, and the guest house of the high chief (which functions also as a council house for the fono). The individual households, strips of land that extend in principle from the beach to the mountain slope, spread out on either side of fte sandy roadway. Traditionally, a household had thee main structures: a guest house, the main dwelling, and fte cookhouse (Handy and Handy 1924; Buck 1930:897). Mead aptly descrbed the Manu'an guest houses "which stand by the sea, are round and high-ircles of posts about four feet high topped by a twenty-foot thathed circular cone. They stand upon a foundation of small stones which rises in slightly higher, narrowing concentric circles, each terrace about five or six inches high, and edged by larger stones" (1930:46). The main dwelling or sleeping house was traditionally the long house (fale o'o), with rounded ends. The cookhouse or shed is thefale wnu, literally house which shelters the wnu or earth oven (Buck 1930:13). These are situated farthest from the guest houses: "a small shack supported on four pillars and roofed, not with sugar cane thatch but with mats woven of palm leaves" (Mead 1930:48). The layout of house types within the household complex reflects a distinctive social and symbolic structure to the village organization. The main axis of orientation is perpendicular to the beach, the interface of land and sea, and thus extends in two directions: i tai, toward the sea, and i uta, inland. The seaward direction is higher ranked, associated with chiefs and persons of status; irnland is lower ranked, associated with prduction and with the economic basis of society. Mead describes the seaward-irdand dichotomy thusly: The term i tai (towards the sea) stands for the optimum position; the village on the seashore, the house on the sea side of the village, the place of honor in the front of the house. And as the trails lead back from the village over narrow stils, through stony piaces, swamp pices, into deep gulches, and up slippery inclines; so the channel marks the way out to sea (1930:50). This structural dichotomy between land and sea, and the way in which it organizes the spatial stuctue of the village, is a widespemad-and therefore probably ancient-pattern within Westem Polynesia, including Fiji. Salins (1976:3745, fig. 6), for example, has described this structure for Fiji, and in another paper (1981) has outlined some of the cultural associations between the sea and high-ranking chiefs. The Polynesian Outlier of Tikopia (Firth 1936; Kirch and Yen 1982) likewise has a characteristic spatial organization very much like that described above for Samoa, with household units differentiated along a seaward-landward axis, canoe houses toward the beach, cookhouses inland, with the main dwelling mediating between. (The Tikopian dwelling itself is divided into seaward:male and landward:female divisions; see Firth [1936].) Similarly, on Niuatoputapu Lsland in northem Tonga, the prehistoric coastal villages appear to follow this pattem (Kirch 1988). And in Fuuna Island, Burrows (1936) describes a village structure very reminiscent of the Manu'an situation. In sum, this kind of cultural and symbolic ordering of space along a seaward-landward axis has a wide distribution in Westem Polynesia, and arguably has a deep prehistory in the region as a "stucture of the long run"'(Braudel 1980). Our tansect excavations at the To'aga site provided some evidence of a coastal village settement pattem ta may also have been organized in this characteristic Westem Polynesian model some 2,500 years ago. TO'AGA: LAND USE AND VEGETATION PATTERNS While the fonmation of the coastal terrace at To'aga was primarily a geomorphological phenomenon controlled by such extemal factors as sea level change, ectonics, and their effect on sediment budget (see Kirch, chapter 4), humans have also The Natural and Cultal Environment played an active role in the evolution of this landfonm. Inchatr5, for exampl, we s argue on the basis of our igraphic data that the input of teerial sediments onto the coastl terace increased after human colonization due to forest clearance and agnicultral activities on the talus slope and mounain. PeNaps the most obvious effect of humans, however, results frm the pate s of land use and vegetaion on the coastal terrace, for the whole zone constitutes an an ogenic, managed environment from the phytogeographic percive. The coastal terrace of Ofu provides the only flat land on the island, the setting for both villages and for certain Inds of intsive horticultura and aiboncultual production. The development of this pattem of intensive land use is one prbem that we have attempted to addrss in our archaeological investigations at To'aga As backgound to this study, it is necessary to charactenize te pattens of land use and vegetaton found on the Toaga coastal terrace today. Vegetion Transects Yuncker's (1945) study of the Manu'an flora enumerated the plant sources ofthe islands, but neglected patterns ofplant distribution. In 1987 and 1989 we reoorded the horizontal distribution of dominant species along three of our trasc across the To'aga coastal terce that had been cleared for archaeological subsurface sampling (see Kirch and Hunt, chapter S for futer details of these trnsets). One of xse, Trnsect 7, is graphically depicted in figure 25. Although there are minor differences in the distribution of species, the same overll pattem is evident in all trsec . Whisler (1980) described and ilustrated most of the species recorded in these tran s. Three main vegetation zones can be discerned, from seaward to ladward: 1. Strand Vegetation. Beginning at the highwater maik and extending to the seaward edge of the coastal road is a zone dominated by halophytic, liUttoral species. Ovehganing the beach are Scaevola taccada and Messerschmidia argentea, with the vines Canalia maritima and Jpomoea pes-caprae trailing over the sands. Larger trees sunnounting the beach ridge inlude Barringtonia asiatica (traditonally used for fish poisoning), Hernandia nyphae4folia, Cocos nucifera, and Terminala samoensis. Pandaiws tectorius and Hibiscus 19 tiliaceous shrubs dominate the iner edge of this zone, lning the sandy radway. 2. Arborlulural-Horticultural Zone. Commencing on the inland side of the road, and extending acrmss the width of the coastal terrace to the base ofthe talus slope is the main zone of economic plants. Coconut palms dot the area, especially toward the seaward halfofthe zone, where the soil is sandier. The badfruit tree, Artocarpus altilis, commences not far frm the roadway and is the main upper story dominant across the terace. Under and between the beadfruit and coconuts are planted a number of fruit and root crops, the most important being Ewnusa banas and the large arid, Alocasia macrrorhiza. Some taro, Colocasia esculent and the historically inrduced American amid Xanthosoma saggitolia, are also found, although in lower frequency. The Alocasia aroids are often densely planted in clearings which, after cropping, ae secondarily planted in bananas. In sone areas, the undertory beneath the coconut palms and badfruit trees is a tangle of second growth shubs, dominated by Hibiscus tiliaceous and Macaranga st4ulosa. Othr useful trees occuring less frequently though the zone include mosooi (Cananga odorata), the flowers of which are used for scening coconut oil, andfisoa (Colubrina asiatica), which has medicinal value and can be used as a soap substitute (Whistler 1980:41). This zone of tree, fruit, and root crps also exhibits a high frequency of feral or natralized species which are commonly cultivated in Oceanic agricultural systems, and which unquestionably are pesent in the Toaga area as suvivals frm an eailier phase of (presumably) more intensive culdvation. These include the t plant, Cordylinefruicosum, the arrowrot, Tacca leontopetaloides, the bitter yam, Dioscorea budbfera (which twines in gat abundance over fte trees and shrubs of this zone), and in lesser quantities, the a'a or Pueraria lobata. All of tese plants are mcognized by the Ofu people as having edible subternean roots or tubers and are regarded as potential famine resources. They are part of the complex of plarts a Barru (1965) has termed "witnss of the past," indicators of eatlier cultivation praices in the Pacfic islands. 3. Talus Slope Vegetation. The main zone of economic plants terminates abrupdy at the base of e talus slope, strewn with large volcanic boulders. The To'aga Site 20 4-4 Alocasia macrorrhiza i Colubrina asiatica Morinda citrifolia -4 Cordy/ine fruticosum Artocarpus a/tilis Macaranga sp. i 4I Tacca leontopetal/odes 4~ ~ ~ ~ 4-Dioscorea bulbifera Musa hybrids Hibiscus tiliaceous Cocos nucifera Pandanus tectorius Barringtonia asiatica @Hemandia nymphaeifo/ia e MassarsMhmida arnantna Scaevola tacc4ada 8 * 0 - 4 -- v -4 . 10 *\ # i D. ARBORICULTURAL ZONE fi Figure 2.5 Distribution of major floral dominants along Transect 7 at To'aga A few breadfruit trees and the occasional banana plant may be found extending a few meters up the slope, but the main dominant here is Hibicus tiliaceous, which fonns a dense tangle over the boulders, and larger forest trees such as Erythrina varigata and others. Land Use and Site Formation Although no longer a locus of pennanent village habitation, the To'aga coastal terrace remains an important zone of intensive horticulture and arboriculture for the Ofu Island population, as indicated by the analysis of vegetation patterns. Indeed, on an island where flat land comprises less than nine percent of the total area, the economic importance of this coastal terrace to the human population cannot be overemphasized. It is certain that the pattem of intensive cultivation of this area extended back at least into late prehistory, but it remains an archaeological problem to deternine just when this pattem first developed. The pattem of land use in prehistory must also have had consequences for archaeological site formation processes. For example, the clearance of indigenous forest cover on the talus slope above the flat would have exposed unstable soil and rock, and thus accelerated erosion and deposition of colluvial sediment onto the coastal terrace. Cultivation on the flat itself would have resulted in a continual reworking of the upper soil layer (through the actions of digging sticks as well as The Natural and Cultural Environmen thrugh floral-turbation by plant roots and tubers). The mixing of terruginous and calcareous sediments thugh cultivation would have created a well-drained, highly fertile edaphic medium which itself was probably more suitable for root crop cultivation than either the heavy colluvial clays, or the calcareous sands themselves (see Kirch and Yen 1982). In the chapters to follow, we pay particular attention to several lines of evidence that point to the gradual, historical development of the intensive land use pattem at To'aga. These lines include evidence for changes in the rate of deposition of colluvial sediments, stratigraphic evidence of buried soil surfaces, evidence of reworking of soils, and the presence of several species of synantrpic land snails that are markers of Oceanic horticultural activities. REFERENCES CITED Barrau, J. 1965. Histoire et prdhistoire horticoles de l'Ocdanie tropical. Journal de la Sociite des Ocdanistes 21:55-78. Braudel, F. 1980. On History. Chicago: University of Chicago Press. Buck, P. 1930. Samoan Material Culture. Honolulu: Bernice P. Bishop Museum Bulletin 75. Burrows, E. 1936. Ethnology of Futuna. Honolulu: Bernice P. Bishop Museum Bulletin 138. Buxton, P. A. 1930. Description of the environment. Insects of Samoa, Part IX, Fasc. 1. London: The British Museum. Coulter, J. W. 1941. Land Utilization in American Samoa. Bemice P. Bishop Museum Bulletin 170. Honolulu. Curry, L. 1962. Weather and climate. IN J. W. Fox and K. B. Cumberland, eds., Western Samoa: Land, Life, and Agriculture in Tropical Polynesia, pp. 48-62. Christchurch: Whitcombe and Tombs. Daly, R. A. 1924. The geology of American Samoa. Carnegie Institution Publicaton 340:93-152. Farrell, B. H. 1962. The village and its agrculture. IN J. W. Fox and K. B. Cumberland, eds., Western Samoa: Land, Life, and Agriculture in Tropical Polynesia, pp. 177-238. Christchurch: Whitcombe and Tombs. Finney, B. 1985. Anomalous westerlies, El Niflo, 21 and the colonization of Polynesia. American Anthropologist 87:9-26. Firth, R. 1936. We The Tikopia. London: George Allen and Unwin. Flenley, J., and M. King 1984. Late quatemary pollen records from Easter Island. Nature 307:47-50. Fosberg, R. 1963. Disturbance in island ecosystems. IN J. L. Gressitt, ed., Pacific Basin Biogeography, pp. 557-61. Honolulu: Bishop Museum Press. Friedlander, I. 1910. Beitrage zur der Samoa Inseln. K. Bayer. Akad. Wiss., Math-Phys. El., Munich, Band 27:358-69. Green, R. C. 1969. Archaeological investigation of Westem Samoan prehistory. IN R. C. Green and J. Davidson, eds., Archaeology in Western Samoa, Vol. 1, pp. 3-11. Bulletin of the Auckland Institute and Museum, Auckland. Handy, E. S. C., and W. C. Handy 1924. Samoan House Building, Cooking, and Tattooing. Bemice P. Bishop Museum Bulletin 15. Honolulu. Hiroa, Te Rangi (P. H. Buck) 1930. Samoan Material Culture. Honolulu: Bemice P. Bishop Museum Bulletin 75. Irwin, G. 1981. How Lapita lost its pots. Journal of the Polynesian Society 90:481-94. 1992. The Prehistoric Exploration and colonisation of the Pacific. Cambridge: Cambridge University Press. Jordan, D. S., and A. Seale 1906. The Fishes of Samoa. Bureau of Fisheres Bulletin 25, pp. 173455. Washington D.C.: Govemnment Printing Office. Kirch, P. V. 1978. Indigenous agriculture on 'Uvea (Westem Polynesia). Economic Botany 32: 157-81. 1982. The impact of the prehistoric Polynesians on the Hawaiian ecosystem. Pacific Science 36:1-14. 1983. Man's role in modifying tropical and subtropical Polynesian ecosystems. Archaeology in Oceania 18:26-3 1. 1984. The Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge University Press. 1988. Niuatoputapu: The Prehistory of a 22 The To'aga Site Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Seattle: Burke Museum. . forthcoming. The Wet and the Dry: Irrigation and Agricultural Intensification in Polynesia. Chicago: University of Chicago Press. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Honolulu: Bemice P. Bishop Museum Bulletin 238. McDougall, I. 1985. Age and evolution of the volcanoes of Tutuila, American Samoa. Pacific Science 39:311-20. McGlone, M. 1983. The Polynesian deforestation of New Zealand: A preliminary synthesis. Archaeology in Oceania 18:11-25. Mead, M. 1930. Social Organization of Manua. Honolulu: Bernice P. Bishop Museum Bulletin 76. Menard, H. W. 1986. Islands. New York: Scientific American Library. Nakamura, S. 1984. Soil Survey ofAmerican Samoa. Soil Conservation Service, U.S. Deparunent of Agriculture. Washington D.C.: Govermnent Printing Office. Natland, J. H. 1980. The progression of volcanism in the Samoan linear volcanic chain. American Journal of Science 280-A:709-35. Olson, S., and H. James 1984. The role of Polynesians in the extinction of the avifauna of the Hawaiian Islands. IN P. S. Martin and R. G. Kline, eds., Quaternary Extinctions, pp. 768-80. Tucson: University of Arizona Press. Sahlins, M. D. 1976. Culture and Practical Reason. Chicago: University of Chicago Press. . 1981. The stranger-king, or Dumezil among the Fijians. Journal of Pacific History 16:107-32. Steadman, D. 1989. Extinction of birds in Eastem Polynesia: A review of the record, and comparisons with other island groups. Journal of Archaeological Science 16:177205. Steadman, D. W., D. S. Pahlavan, and P. V. Kirch 1990. Extinction, biogeography, and human exploitation of birds on Tikopia and Anuta, Polynesian outliers in the Solomon Islands. Bishop Museum Occasional Papers 30:11853. Steadman, D. W., and P. V. Kirch 1990. Prehistoric extinction of birds on Mangaia, Cook Islands, Polynesia. Proceedings of the National Academy ofSciences U.S.A. 87:9605-9609. Steams, H. T. 1944. Geology of the Samoan Islands. Bulletin of the Geological Society of America 55:1279-1332. Stice, G. D., and F. W. McCoy, Jr. 1968. The geology of the Manu'a Islands, Samoa. Pacific Science 22:427-57. Visher, S. S. 1925. Tropical Cyclones of the Pacific. Bernice P. Bishop Museum Bulletin 20. Honolulu. Wading, D. 1982. Birds of Fiji, Tonga, and Samoa. Wellington: Millwood Press. Whistler, W. A. 1980. Coastal Flowers of the Tropical Pacific. Honolulu: Oriental Publishing Company. Yuncker, T. G. 1945. Plants of the Manua Islands. Honolulu: Bernice P. Bishop Museum Bulletin 184. 3 SURFACE ARCHAEOLOGICAL FEATURES OF TO'AGA T. L. HuNT boundaries rather an by the relative density of ~T'HE SAmoAN LANscAPE has been shaped by a long history ofhuman setlement and resource arifacts or architecturl remains. Thus, Site AS-13-1 exploitadon. Such evidence of settlement and land comprises the area from the shoreline to the steep use, both past and present, is distributed more-or-less slopes and cliffs, the coastal ter-ace described in continuously across the island landscapes of Samoa. detail in chapter 2. Isolated arfifacts, feaus, sites, and site complexes cover much of the Samoan Islands and vary signifiField Methods cantly in relative density (e.g., Davidson 1974:242). A systematic survey of the surface features at The continuous nature of archaeological remains, To'aga was accomplished using tanscts set frm the often expressed in terms of "non-site archaeology" beach to the steep slopes of the interior cliffs. In 1986 requires documenting the distributional pattems of material culture in space (Thomas 1974; Dunnell and a baseline was set out along the coastal road with Dancey 1983). This is in contrast to isolating "sites" lateral trans about 30 meters in width made at as dense clusters and ignoring the lower density every 100-meter interval. This resulted in 18 survey distributions that infonn upon a varety of peistoric transects that covered approximately 4.4 hectares of activities and their spadal distribution. The "non-site" the coastal terrace. Shrub vegetation, often dense, was cleared from these transects. Without extensive approach has proven useful elsewhere in Oceania clearing of ground cover and other vegetation, some where the vestige of entire settlement-subsistence areas had very poor surface visibility. This means patterns have been documented (e.g., Green 1980; surface features were undoubtedly missed, even in Kirch and Yen 1982; Weisler and Kirch 1985). transects cleared of some vegetation. Ideally, entire archaeological lamdscapes (spatial AU surface feaues and artifacts discovered on pattems) must be recorded in detail, although this the were recorded for provenience, deunsects ultimate objective must usually be met though the gradual accumulation of survey data over many years scrbed, and many were also mapped using a tape and compass. Other features encountered during fieldof effort work (in all three seasons) were wecorded in the same This chapter describes the surface features way. Fmally, well-known features or monumental recorded along systematic tansects on the coastal constructions were located with local guides and tenrace of To'aga (Site AS-13-1). As elsewhere in recorded as well. Manu'a, the "site" was defined by natural landscape 24 The To'aga Site Trnsects oriented perpendicular to the shoreline traverse the greatest range of geomorphic diversity that may correlate with variation in settlement, land use, and age of surface features. Surface data acquired along systematic transects can be used to produce maps (with interpolaffons as desired) that reveal pattem and structure in the archaeological remains. In short, this kind of systematic strategy provides a means to record, in some detail, the spatial configuration and variety of archaeological remains while producing an informative, albeit partial, picture of the broader archaeological landscape. Results Much of the coastal land of the south coast of Ofu Island, from the areas known properly as To'aga to Fa'ala'aga, appears to have a near continuous distribution of archaeological remains on the surface. In most areas where vegetation pennitted visibility of the ground surface, evidence of habitation and land use was present. Figure 3.1 shows the location of archaeological surface features recorded in the survey transects. These features, shown by consecutive numbers, are listed and described in table 3.1. The surface features recorded include basalt and coral boulder alignments usually of curved and oval form (house foundations) with coral and basalt pebble paving ('il'ili, e.g., figure 3.2); small oval or rectangular boulder alignments that appear to mark burials or are simply the partial remnants of house foundations; a massive, fine-grained talus boulder used for multiple (twelve) grinding basins/ surfaces (figure 3.3); pits with associated boulder slab lining in several places, probably the remains from the production of fennented breadfruit (lua'i mast); and long (inland-seaward oriented) single course alignments which appear to have served (and may continue to serve) as land boundary markers. Particularly noteworthy is a complex that includes the locally well-known Tui Ofu Well (Feature 23, figure 3.4) and Tui Ofu Tia (monumental tomb, Feature 24, figure 3.5). Situated near the eastem end of the coastal flat of southem Ofu (near Fa'ala'aga), these monumental constnrctions are traditionally associated with the highranking title Tui Ofu ("king of Ofu"). The Tui Ofu Well is a relatively elaborate construction of waterwom basalt boulders arranged in a rounded form that includes a sloped concourse as well as a small paved court around the excavated shaft. The depth of the shaft from the sunrounding court is 2.1 m. The Tui Ofu Tia (tomb) comprises a crudely terraced mound of basalt boulders (both rough and waterwom) that is set in against the massive talus and steep slope rising on the inland side of this area. A small pit lies among the boulder rubble on the uppermost terrace of the mound. No artifacts or other cultural materials were observed at the structure, except that omamental plants, such as crotons, have been planted around the base of the mound. The survey along transects has yielded data on only part of the surface record at To'aga. Other features will be found with more intensive coverage of the area. A more complete survey of surface remains at To'aga will require extensive clearing of vegetation which restricts access to parts of the coastal lands or obscures visibility. The geomorphic evidence (described in chapters 2, 4 and 5) suggests that the present surface at To'aga is less than 1000 year B.P., and very recent in some areas where colluvial and eolian (i.e., primarily calcareous beach sediment) deposition continues. Thus the surface evidence represents late prehistoric and historic activities on the coastal terrace. This observation is supported by excavation and geomorphic data (Kirch and Hunt, chapter 5). Many of the surface features are probably pene-contemporaneous, at least in archaeological time-frames. The surface features at To'aga are similar in their formal variability to those documented elsewhere in Manu'a (Hunt and Kirch 1988) and other islands of Samoa (e.g., Clark and Herdrch 1988; Davidson 1974; Green and Davidson 1969, 1974; Jennings et al. 1976, 1982; Jennings and Holmer 1980). The architectural remains present at To'aga do not reflect the full range known for Swface Archaeological Features Map of the southrn portion of Ofu Island, from To'aga to Fa'aia'aga, showing the location of surface archaeological features (1-26) of site AS-13-1. Figure 3.1 0 so 1 2m AA A Figure 3.2 25 Plan of round-ended stone house foundation (Feature 19); the shaded area is paved with coral and basalt gravel ('ili'ili). The To'aga Site 26 53sq * 19 ) WI) 6 (o s 1-i V-4 .1 x Q m C-i C> In S r. S v) o O; Q 0 ) _ _ esn " n Q o '.4 10 s C) S 5 P-400 0. (:.mm 9 x V-4 P-4 x lqt Cq C. x x 63 C)04aC> tn tnmW)0 J: 46(7A Aen V-4 v-4 0 W-4NW S i S K iS C = = 2 ai 7i ° c V-4 s Cf) qqt cc; Q xo .o t- co co el 0~~~~~~~~~~~~~~~~~~~~~~~~~0 Eo 0~~~~~~~~~~~~~~~~0 co > ] :t;;88r8r 6~~~~~~~~~~~~~~~c en ut °E6Ew wt To v :s I) Q *8 . - Po P 0I . wosoo _ _ = 3 0 ItC : m(mA *_& a c go _ m_ cb 5 go _ , 3 ° __t_ ... _ I _ ^__~~~~~~~.0. aca£a.'O .*m .*m.*m*m1S ... a m ms m m m mm m ON __ __e - N N m im Wm m N (4 N N Surface Archaeological Featwres 27 _~~~~~~~~~~ _~~~~~~~~~~~~~~~~~b 4)_ 4) 0E 4) 0 4) oU _ j~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 4) :U 4); 28 The To'aga Site A A A' r,. Figure 3.4 Plan and cross section of the Tui Ofu well at Muli'ulu. Surface Archaeological Features A- 29 -A' 2r A n A' Figure 3.5 Plan and cross section of the the Tui Ofu Tia, stone mound/tomb at Muli'ulu. 30 The To'aga Site Samoa but do include forns linked to domestic activities. Only the Tui Ofu Well and Tui Ofu Tia represent structures of specialized function, indeed those associated with social rank. The available evidence suggests a redundant pattem of associated features which can be interpreted as the remains of domestic compounds. The distribution of such habitation suggests dispersed settlement organized on the basis of descent groups, as represented in Samoan socioeconomic organization documented ethnographically (Mead 1930). The distribution of structural features also reveals late prehistoric and historic settlement situated primarily on the stabilized dune ridge. The surface and excavation evidence (Kirch and Hunt, chapter 5) suggests that occupation has centered on the high ground of stabilized sand dune ridges. As progradation of the To'aga coastal land occurred, settlements moved seaward. The archaeological evidence on today's relatively young surface reflects the last phase of habitation at To'aga. Thus, the setdement pattem marks continuity over To'aga's pre- and post-contact history. The primary focus of fieldwork at To'aga has been the definition and sampling of the subsurface deposits of the coastal terrace. Additional intensive surface survey is necessary to complete the picture that has emerged so far. REFERENCES CITED Clark, J. T., and D. J. Herdrich 1988. The Eastern Tutuila archaeological project: 1986 final report Unpublished manuscript, Office of Historic Preservation, Pago Pago. Dunnell, R. C., and W. S. Dancey 1983. The siteless survey: A regional scale data collec- tion strategy. Advances in Archaeological Method and Theory 6:267-87. Davidson, J. M. 1974. Samoan structural remains and settlement patterns. IN R. C. Green and J. M. Davidson, eds., Archaeology in Western Samoa, Vol. 1. Auckland Institute and Museum Bulletin 6. Green, R. C., and J. M. Davidson, eds. 1969. Archaeology in Western Samoa, Vol. 1. Auckland Institute and Museum Bulletin 6. Green, R. C., and J. M. Davidson, eds. 1974. Archaeology in Western Samoa, Vol. 2. Auckland Institute and Museum Bulletin 7. Hunt, T. L., and P. V. Kirch 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal of the Polynesian Society 97:153-83. Jennings, J. D., R. N. Holmer, J. C. Janetski, and H. L. Smith 1976. Excavations on Upolu, Western Samoa. Pacific Anthropological Records 25. Honolulu: Bernice P. Bishop Museum. Jennings, J. D., and R. N. Holmer, eds. 1980. Archaeological Excavations in Western Samoa. Pacific Anthropological Records 32. Honolulu: Bernice P. Bishop Museum. Jennings, J. D., R. N. Holmer, and G. Jackmond 1982. Samoan village pattems: Four examples. Journal of the Polynesian Society 91:81-102. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Bernice P. Bishop Museum Bulletin 238. Honolulu. Mead, M. 1930. Social Organization ofManua. Bernice P. Bishop Museum Bulletin 76. Honolulu. Thomas, D. H. 1974. Predicting the Past. New York: Holt, Reinhart, and Winston. Weisler, M., and P. V. Kirch 1985. The structure of settlement space in a Polynesian chiefdom: Kawela, Moloka'i, Hawaiian Islands. New Zealand Journal of Archaeology 7:129-58. 4 THE TO'AGA SITE: MODELUNG THE MORPHODYNAMICS OF THE LAND-SEA INTERFACE PATRICK V. KIRCH IN CHAP 2 wE described the economic and cultual significance of the narrow coastal terrace of Ofu which-a~the only flat land on the island-is thus the main locus of settlement and a major zone of subsistence prduction. This coastal tenrace, which the To'aga site exemplifies well, must be undertood as a dynamic geomorphological entity lying at the cdtical land-sea interface. McLean has aptly described the imporance of such land-sea interfacial zones on small tropical islands: The atactiveness of the coastal lowlands for setlement purposes is obvious.... On many islands they comprise dte only flat country. It is easy to move over and easy to build on. Soils, though naurally of low fertlity and moise statis, are easily worked. The sandy subsate, beach sand and beachrock are readily available for such uses as building matials, graves and road 'metal.' That the coastal flats are multipurpose and multi-utilized resources goes without saying. Moreover, their location at the land-sea edge permits optimal access to a complete range of terrestrial and imaine environments and rsources: hilislope, valley and forest; reef, lagoon and fish. They function as convenient bases to exploit the local environs. They also function as verandahs to the worlds outside (1980:.129-30). In their study of the prehistory and ecology of Tikopia, Kirch and Yen (1982) similarly argued for the importance of unraveling the morphodynamic histories of tese interfacial enviromnents: among these [environments] are the coastal/ marine interface, with implications for access to rsources of reef and sea, and that of calcareous plains and volcanic hills, with implications for erosion and agricultural developmenL ... Both of these interfaces have been active zones of landscape change, with long-tern implications for human adaptation (1982:17). In particular, the Tikopia case revealed the agricultural significance of human-induced modifications to the coastal zone: 'te plain and hill interface pmvides the opporunity for the mixing of volcanic and calcareous soils to form edaphic media, more favorable for the cultivation of root crops than either separate type" (Kirch and Yen 1982:17). Prior experence in Tikopia (Kirch and Yen 1982), Niuatoputapu (Kirch 1988), and other small tropical islands had sensitized us to the necessity of a geomorphologically informed approach to the archaeological study of the To'aga site. This was critical both to an undemanding of depositional and site-formation processes, as well as for contribudng to our knowledge of the evolution of the island's larger The To'aga Site 32 settlement paterns and production systems. This chapter focuses on several aspects of the local envinnt at To'aga which played key roles in te morphodynamic evolution of the site. Chief anong these are the ging configuration of sea level in the mid-Holocene, and the tectonic instability of the Manu'a Group. Equally sigficant, however, is the role of humans tiemselves in shaping landscape through physical manipulation of soil and rock, forwst clearance, the inroduction of a "portnantea" biota (cf. Crosby 1986), and the continual manipulation of the coastal enviro to best suit human needs. My aim here is to outline a model of these morphodynamic processes that may then be tsted agains the empirical evidence revealed thrugh our field and laboratory investigations. GEOMORPHOLOGY OF THE TO'AGA COASTAL TERRACE The key geomorphological features of the coastal terrace at To'aga can best be descnbed along a generalized transe ruing from the reef inland across the flat up to the steep hillslope and cliff as in figure 4.1. (Particular, surveyed transects are illustraed and described in chapter 5, along with the statigraphic evidence for temporal development) At the seaward end is the reef flat, source of te calcarous sediment (composed of the detritus of coral molluscs, and other reef-dwelling organisms) of which the coastal terrace is primarily constructed. This reef flat is a mosaic of microenvirornents, including surge channels, the 'lithotamrnium ridge' at the seaward edge, large coral heads, and sandy patches. A careful exanination of the beach slope reveals several clues as to the dynamic processes presently at wor along the land-sea interface. At the foot of the beach slope in various places are exposed layers of 'beach rock' (Wiens 1962:64-67) made up of sand and coral rubble cemented with calcium carbonate (CaC03). Such beach rock deposits can fonn fairly quickly under tidal conditions of continual wetting and drying witfin active beach ridges. Their exposure along the To'aga beach front, however, signals a cument phase of coastal erosion and net sediment loss (or lateral transfer). This evidence is reinforced by an examination of the vegetation line at the top of the present beach ridge crest, where large trees (such as Cocos, Cordia, and Hernandia) have been undercut and eroded by wave action. Moving inland, one mounts the crest of the present beach ridge or benn, lying between the beach slope and the sandy road. This nanrow zone is covered with a thick tangle of vegetation, dominated by Pandanus tectorius, Barringtonia asiatica, Hernandia nymphaeifolia, Scaevola taccada, Messerschmidia argentia, Cocos nucifera, and other species that anchor the loose, unconsolidated calcareous sands. Test pits dug into this beach ridge revealed only a thin (ca. 5 cm) organic A horizon directly overlying the young parent sands. The absence of archaeological deposits suggests that the present beach ridge is of no great antiquity. Crossing the road and plunging thrugh a narrow band of Pandanus and Hibiscus tliaceous shrubs, one enters the main extent of the coastal tenrace, a zone of intensive economic utilization though arboriculture and root crop gardening (see Kirch, chapter 2). The ground at first slopes down slightly from the beach ridge, levels out, and then begins to rise again gradually toward te talus slopes inland. As one moves inland from the road through this zone, the soil gradually becomes less sandy and more clayey, reflecting the increased contribution of volcanic colluvium which has been eroded frxm the hillslopes and added to the calcareous sands that form the main component of the substrate. The mixing of tese calcareous and volcanic sedimentsthrogh the continual action of human cultivationhas made this zone of particular edaphic value for Samoan horficulture. This is also the main zone of archaeological feaures, both surface and subsurface, as described in chapters 3 and 5. Moving farter irland toward the talus slope, one begins to note volcanic cobbles and boulders stewn over the land surface, signals of the dynamic instability of the 500-m-high cliffs that tower over the To'aga flaL Some of these boulders are several meters in diameter and can fall with considerable destructive force. [be ground surface closest to the talus is entirely gravelley clay-loam with no calcareous component evident Rather abruptly, one arrives at the base ofthe talus itself, an imposing and unstable jumble ofboulders rising steeply toward the cliffs, over which grows a tangle of Hibiscus tiliaceous and scattered forest trees such as Erythrina varigaa. In sum, this geomorphological traverse across Morphodynamics of the Land-Sea Inerface STRAND VEGETATION ARBORICULTUR4L ZONE / C_lluvium 33 Erosion Scarp i :: .: : : :: .:: Unconsolidated Calcareous Sands >,Exposed Beach Rock :. Coral Reef Platform- Figure 4.1 AgeneralizedtasectacrossthecoastalplainatTo'aga,showingthedistributionofcolluvialandcacareous sediments, major vegetation associations, and geonnorphological featu . To'aga coastal terrace reveals several key feaues: (1) the terrace is constructed primarily of marine biogenic sediments (calcareous sands and larger clastics); (2) one finds a progressively greater contribudon of terrigenous sediments as one approaches the volcanic mass; and (3) there is some evidence for curnt coastal erosion and maine transgression at the present time. These are features that must be accounted for in any model of the morphodynamics of the To'aga site over the past several thousand years. the MORPHODYNAMIC PROCESSES: SEA LEVELS AND SUBSIDENCE As the archaeological deposits of the To'aga site are integral sedimentary components of the coastal terrace, any model of site formation processes must first account for the geomorphological processes by which the terrace was formed. Lying at the critical and highly dynamic land-sea interface, any consideration of the morphodynamics of this zone must begin with exanination of controlling processes in shoreline formation. Among the most important of these are: (1) glacio-eustatic sea-level changes; and, (2) the local tectonic situation which-as we an fshall argue-is one of subsidence. The temporal penod with which we are concemed is the mid- to late-Holocene, from about 6 kyr B.P. to fte preet This time span covers some 3 kyr piior to the human colonization of Ofu as well as the subsequent interval during which humans have added their input to the development of the local lanscape. Before examining the evidence for reladve sealevel change during this period, it is essential to inrduce a few key concepts regarding coastal change on small islands. In this I have drawn primarily from Chappell (1982) and McLean (1980b). The construction of a coastal tenrace such as that at To'aga results from progradaion, the "progressive formation of new land by sedimentation inespective of fte tendency of sea level movement" (Chappell 1982:71). Sea-level changes themselves, whether due to glacio-eustacy or tectonic movements, or both, are inportant as controlling factors for the sediment budget, but alone ftey do not provide a sufficient model of prgradation. As Chappell emphasizes, ". sea level changes alone cannot be used to account for coastal changes. In fact, for the last 6000 years, the sedimentary budget is the more important factor' (1982:71). The sediment budget can be thought of as te net sum of sediment input from both terrestrial . . 34 The To'aga Site (talus, colluvium, etc.) and marine biogenic (calcareous sands and coral detritus) sources, minus the loss of sediment from transport (fig. 4.2). In modelling the Toaga coastal terrace formation, therefore, we need to pay particular atention to anges that would either increase or decrease the production of sediment from te ial or marine sources. Figure 4.3 graphicaly portrays the dynamic effects of dcanges in relative sea level (due either to glacio-eustatic or tectonic change), combined with increases or decreases in sediment budget. In this diagram, sediment budget is indicated along the x axis, and relative sea-level change along the y axis. The heavy diagonal line separates transgression or coastal rera, frm regression or coastal advance. Holocene Sea Levels in dhe South Pacific A rapid rise in sea levels following the end of the Pleistocene is a global phenomenon that has been widely rognized (Fairbridge 1961; Shepard 1963). More controversial-because they depend upon a complexity of local conditons and processes-have been the details of the eustatic sea-level curve in the mid- to late-Holocene, especialy the matter of whether there have been higher-than-present stands. Bloom (1980, 1983) modelled some of the global diversity in these Holocene curves and suggested that a +1-2 m stand existed in the south Pacific region during ts period. Substantial geomorphic and radiometic evidence from a variety of islands now supports this intepetation of a +1-2 m high sea level during the period between about 4-2 kyr B.P. In Fiji, for examnple, Nunn (1990:304) concluded that the coasts "experienced a middle to late Holocene sea-level maximum some 1-2 m above prset mean sea level." Recent work by Miyata et al. (1990) also supports this finding. Similar results are presented by Ash (1987) for Viti Levu Island, and Yonkura et al. (1988) report evidence for a +1.7 m stand between 3400-2900 B.P. on Mangaia Island in te southem Cooks. In French Polynesia, Pirazzoli and Montaggioni (1986,1988; Montaggioni and Pirazzoli 1984) describe evidence from various islands for a MSL between +0.8 and 1.0 m beginning about 6-5.5 kyr B.P. and lasting as late as 1.2 kyr B.P. In Westem Samoa, Rodda and his colleagues (1986; Sugimura et al. 1988) summarize varous evidence for Holocene higher stands. Isla (1989:361-63) discusses a comparable range of evidence for several Pacific Islands. In figure 4.4, the geographic distibudon ofmid- to late-Holocene higher sea-level stands is plotted along with associated radiometric ages. Figure 4.5 shows a timeelevation plot of sea levels in various Polynesian archipelagoes over the past 5 kyr B.P. This widespread and consistent patem of radiometrically dated shoreline features provides strong evidence for a higher sea level ranging between about +1-2 m over fte southwestem Pacific, from at least 5 kyr B.P. and lasting until sometime between 2-1 kyr B.P. After about 2 kyr B.P., sea level fell (pedaps faidiy rapidly) to its present position. TIis mid- to late-Holocene sea-level curve thus provides one important dimension for a model of coastal terrace formation at the To'aga site. Subsidence in the Samoan Archipelago Most ofthe volcanic archipelagoes of the centrl Pacific region are aiayed lieiy along ageprogression sequences emanating from a "hot spot" or magma plume on the floor of the Pacific Plate (Menard 1986). As the Pacific Plate gradually migmtes west or nortwest, active volcanic islands move off te hot spot, and a new island is formed. Thus, the islands in such an archipelago comprise a "plume trace" of inceasig age from east to west Classic instances of this pattem are the HawaiiEmperor chain and the Society Islands. The Samoan archipelago presents a somewhat more complicated geological picture, primarily due to reent volcanism on Savaii Island which, at the westem end of the chain, theoretically shodd be the oldest island (Menard 1986:186-87). This unusual siution confounded geologists for many years (Dana 1849; Daly 1924; Steams 1944; Natland 1980). Recently, however, K-Ar dating of rocks from various Samoan islands has confirmed the basic pattem of an east to west age pogression; the mean K-Ar ages are 0.1 myr for Ta'u, 0.3 myr for Ofu-Olosega, 1.26 myr for Tutuila, and 2.2 myr for Upolu (McDougall 1985:318-19). Savai'i is presumably the oldest island but has had renewed volcanism in recent times. McDougall concluded ta "Upolu, Tuuila, and the Manu'a Islands comprise a hot spot or plume trace on te Pacific plate, similar to the Hawaiian and Society Islands chains." Nevertheless, the Morphodynamics of the Land-Sea Interface MARINE INPUT TERRIGENOUS INPUT Landslides Mass Wasting | Surface Flow SEDIMENT BUDGET H SrghEnergye Biogenic Volcanic Clastics TRANSPORT LOSS Colluvium M(oral) Clastics Calcareous Sand Reef Detritus Soil Figure 4.2 A model for the sediment budget at To'aga, showing terrestial and marine inputs. dominane of youthful volcanism on Savai'i remains an egngma and may indeed reflect major rejuvenescence related to deformation of the Pacific Plate adjacent to the Tonga Trench (1985:319; see also Menard 1986:186-87). This linear age-progression and the fact that Sea Level + I - 0 0 E E .0 co .' co Sea Level Figure 4.3 35 - A model of shoreline transgression and regression as a function of relative changes in sea levels and in sediment bugt (after Chappell 1982). Manu'a Islands are sll in a youfl stage of geological evolution are essential factors in understading the local tectonic siion of Ofu Island. The rapid constuction of a volcanic mass on the oceanic crust of the Pacific Plate results in a phase of subsidence, due to point loading on the thn and flexible crst (Menard 1986:165-69). Such subsidence and associated crustal defornation is well documented for the younger Hawaiian Islands. The island of Hawai'i appears to be subsiding at an average me of about 4.8 mm/yr. For Ofu Lsland itself, we are not aware of geological documentation of subsidence at the present time, although we have strong indirect reasons to believe that this is the case. The young age of te island (only 0.3 myr) would itself suggest that subsidence and crstal deformation (which lag behind volcanism) have not yet weached equilibrium. The evidence for active erosion of the island's coastline, descrbed above, stmngly supports this interprtation of active subsidence. Further evidence is provided by the reefs along the To'aga area which display active coral head growth and lack the solution-pitted and eroded reef platforms typical of islands that have been teconically stable during the Holocene. These observations suggest that Ofu and Olosega islands are still in a phase of subsidence due to point loading on the oceanic crust resulting from their initial volcanic construction. Elsewhere in Samoa, dramatic evidence for rapid subsidence was provided by the chance 36 The To'aga Site o *1 a) 0o Is I 0 I- a) 0C) .- 8t 04 0 c .o ° bo Io kI ,.I qq *s Morphodynamics of the Land-Sea Interface SOCIETY ISLANDS o Scilly " Mopelia * Maupiti v Tupai v Bora Bora & Raiatea A Huahine * Moorea o Tahiti Nui o Tahiti Iti m +2- 1.04 +1- _~ .~~ _s ,_ p L~~~~~o so m.s.I.- I 5 2 5~~~~~~I 1 0 - km t~~~~ t ' " '~P-M ~~~~~~ ~ ~ 37 lwpgf_ 1-164 1* K)4PW4 -oo, .NM r~~~~~~~~~~~~~~~~~66 09 I .I II 3I 3 2 4I I I I 8 I 6 N.W.TUAMOTU 5 4 m * Makatea * Rangirioa v Takapoto * Mataiva o Arutua a Kaukura o Apataki +1 s0. -1.94 0 0 mm P-6 -- :0:~ 1*64 4 ~0: V 5. GAMBIER ISLANDS a Temoe m +1 COOK ISLANDS o Mangaia o Aitutaki Rarotonga E.TUAMOTU w Vahitahi 20 Reao - 104W m.s.l9. 4 3 2 1 I I .+ 1.04 WAO404 B..4 I I I I I a 4 2 a 5 I I 6 Age (Kyr) Figure 4.5 Time-elevation plot of sea levels in Polynesian archipelagoes (source: Spencer 1989, fig. 13). I. 38 The To'aga Site discovery of the Mulifanua archaeological site (Green and Davidson 1974; Green and Richards and Gren 1989). Dating to 3251 ± 155 1975; L B.P., Mulifanua is a Lapita-pottery bearing deposit now situated -1.5 m below mean sea level, and further capped by 0.75 m of reef rock (Leach and Green 1989:324-26). The site was accidentally discovered by dredging for a ferry berth Given the site represents a village occupation on a former beach or coast terrace, between 2.6-3 m of subsidence is indicated, or an approximate subsidence rae of I mAlyr. (Davis [1928:249-53] also discusses evidence for the submergence ofTutuila Island.) In constructing a morphodynamic model for the developnent ofthe To'aga cotal tenrace, we assume that Ofu has been undergoing continual subsidence trughout fte Holocene. Although no precise rate can be empirically determined at this time, we may use a woddng rate of 1 mAkyr based on the Mulifanua siuation. (It is possible that the aul rate of subsidence on Ofu could exceed this value.) although this increases as one moves landward. The primary source oftenestrial sediment is the steep cliff which towers over the site, continually depositing talus and rockfall onto the coastal teface. Some finer grained sediments wash over and down the cliff during heavy rains, but there is no major alluvial contribution of sediment Thus mass wasting (principally landslides) and sheet erosion are the major agents of terigenous sediment deposidon. The sediment budget of the Toaga geomorphic system would have fluctuated during the Holocene, as the generation of biogenic sediments in particular would vary with sea-level regimes. During periods of rapid sea-level rise, when corals are actively growing below mean sea level, the generaton of sediment would be substantially reduced. When sea level dropped or was stable for a perod of dme, coral growth would have caught up with sea level and would have been exposed to wave action, resulting in erosion and te generation of calcareous sediment SEDIMENT BUDGETS: SOURCES AND MODES OF DEPOSITION MODELLING THE MORPHODYNAMICS OF THE TO'AGA COASTAL TERRACE IN THE MID- TO LATE-HOLOCENE In addion to the relative changes in sea level resulting frm a combination of glacio-eustatic and tectonic processes, we need to consider both the sources of sediment which contributed to the costruction of the To'aga coastal terrace and the agents responsible for the deposition of these sediments. The bulk of the Toaga coastal terrace is constructed of calcareous sands and larger clastics and chunks of reef conglomerate) of (coral marine biogenic origin. These biogenic sediments are produced on the reef thogh a number of mechanisms, including wave action and biologic processes (such as the generation of sand by parrot fish and other species that rasp and grind coral to extract algae). These biogenic sediments are transorted landward over te reef flat by wave action. Episodes of high-energy storm surges, associated with the cyclones perodically lash the island, are extremely important for the rapid accumulaion of large quantities of sand and ofthe larger size component of coral boulders and cobbles. The contribudon of terigeos sediments to the coastal terace appears to have been less significant, e The variables essential to a model of the formation of the coastal terace at To'aga in the mid- to late-Holocene have been reviewed. In figure 4.6 these varables are diagrmmed along the same temporal axis. Figure 4.6A shows the Holocene glacio-eustic rise in sea level, reaching a +1-2 m maxnimum between about 4-2 kyr B.P. Prior to about 5 kyr B.P., the rapid rise in sea level would have continually drowned the shoreline, and e formation of a stable coastal terrace would not have been possible. Rather, the sea would have encrached direcly against the island's volcanic mass, creating the dramatic sea cliffbehind the To'aga site, as described by Stice and McCoy (1968). Only after a maximum sea level was achieved, between ca. 4-2 kyr B.P., could a coastal terrace have begun to have formed along the base of the cliffs. At the sane time, we presume ta Ofu has been subsiding at a rte of about 1 m/kyr, as depicted in figure 4.6B. Tberefore one takes into account these two controlling processes to determine the probable net change in relative sea level as graphed in figure 4.6C. This Morphodynamics of the Lad-Sea Interface penod as the coral reef flroing the To'aga area was exposed to wave action, especially stonn sures. During the past I kyr B.P., as sea levels again blized and local teconic subsidence coninued, te sediment budget would again have decreased, so that a phase of regression recommenced. This is consistent with our field observations of active shoreline erosion at the present time along the To'aga site. This model of the fonmation of the To'aga site can also be diagrmmed as a temporal trajectory along the two axes shown earlier in figure 4.3. In figure 4.7, we have plotted a rtDdiction of the most probable transgression-regression sequence for the To'aga shoreline over the course of the past 6 kyr B.P. In this diagram, Xt y-axis represents the net sea-level change resulting from the combinaion of both glacio-eustatic and local tectonic effects, while fte x-axis represents fte changing sediment budget. This model provides a wolking hypothesis for te fomiation of e To'aga coastal terrace and its archaeological deposits and may be tested against fte strtgraphic and radionetric evidence denived +2 A 0-2 EUSTATIC SEA LEVEL - B SUBSIDENCE -. ~~~NET SEA LEVEL L D+l 5 4 3 2 1 te 0 KYR B.P. Figure 4.6 Time trends in four key variables affecting the morphodynamics of the To'aga coastal plain. te most likely period during predicts which the To'aga terrace could have prograded through the rapid deposition of biogenic sediments would have been between ca. 2-1 kyr B.P., when the eustatcally contolled sea level dropped from its mid-Holocene maximum down to modem levels. Depending upon the extent of this drop (1-2 m), the actual sea level on Ofu would have been either stable, or there may have been a slight fall. This would have resulted from te eustatic sea-level fall, off-setting the continual effects of tectonic subsidence. As shown in figure 4.6D, the biogenic sediment budget would increase significantly during this graph of trasect excavations at various points along site. Several predictions may be generated based on the model: (1) The earliest archaeological deposits should be located adjacent to the cliffs and should date no earlier than about 4 kyr B.P. (2) These early sediments are likely to have a higher component of volcanic clastics, because talus matenial would still have been readily available for erosion by wave action. (3) There should be a fairly rapid or even abrupt episode of coastal progradation beginning sometime after about 2 kyr B.P. and ending by about 1 kyr B.P. The sediments deposited during this interval would consist almost wholly of from our program ~~SEDIMENT BUDGET I 6 39 marine biogenics. This model was not developed prior to the actual fieldwork, but evolved during the courae of our investigations, as a dialectic between field observa- tions and theoreical exercises. Nonetheless, it is crucial to stress that the model in no way depends on our archaeological data; it is wholly independent, deriving from cumnt geological and geomorhological knowledge of coastal process in the southwest Pacific. To briefly anticipate the results of our field and laboratory studies presented in chapters 5, 6, and 7, we shall see dt the predictions of the model developed here are substantially bome out 40 The To'aga Site Relative Sea Level + I .0- ._0 E 1 a) E 0 0 CO C) Sea Level Figure 4.7 Retoiction of the tansgrssion-regression sequence along the To'aga coastline fron 5 kyr B.P. to the peent. TO'AGA: SOME FURTHER EXPECTATIONS FOR ARCHAEOLOGICAL SITE FORMATION PROCESSES Schiffer has avenred that "during human occupation of regions, natural processes, influenced by cultual behavior, have created an ever-changing landscape that the investigator perceives at just one point in time. The contemporary region is a complex, the-dimensional mosaic consisting of natural sediments, vegetation, modem artifacts and settlements, and archaeological remains. In order to find sites and, especially, to understand how settlements functioned in regional systems, one must endeavor to infer or reconstruct changes in the landscape" (1987:261). In this chapter, I have endeavored to model some of the key processes which contributed to the formation and modification of archaeological deposits in the To'aga area. Lying at a fragile interface between land and sea, the To'aga coastal terrace is subject to envirornental influences of several kinds, and can potentially undergo rapid changes if any of these inputs vary. This interfacial zone has also been subject to intensive human use which has implications for both te formation and modification of the archaeological record. Among the most important implications of our model of morphodynamics of the coastal terrace is that the very existence of this interfacial microenvirorunent depends upon a sediment budget which has fluctuated highly during the mid- to late-Holocene. Since the archaeological deposits at To'aga are an integral part of the geomorphological structure of this zone, it is essential that we understand these processes of coastal ten-ace construction and erosion. To quote Schiffer again, "from the standpoint of survey [and excavation] design, the archaeologist needs to know where deposition and erosion are occurring and where they have occunred during the period of human occupation of the region" (1987:257). Our model of relative sea level and sedimentary budgets suggests that the To'aga coastal ten-ace could not have begun to form or stabilize until after the Holocene sea-level maximum of about 5-3 kyr B.P. Given that the Westem Polynesian region was first colonized by the makers of Lapita pottery at about 3.4-3.2 kyr B.P. (Kirch and Green 1987; Kirch and Hunt 1988), the area of coastal terrace available for initial establishment of human habitations would have been very restricted, probably consisting of little more than a beach ridge lying directly under the steep cliff. As the coastal terrace began to rapidly prograde after about 2 kyr B.P., the area available for intensive land use would have increased substantially. Such progradation, however, combined with deposition of colluvium, probably would bury the earlier occupation deposits deeply. Thus, our morphodynamic model also implies that the earliest archaeological deposits will be the most difficult to locate, being situated fartest inland and possibly buried under substantial depths of colluvium. The model predicts, therefore, that archaeological deposits dating to the earliest phase of Samoan prehistory-the phase marlced by te presence of ceramic assemblages-are not likely to be exposed on present ground surfaces or to be encountered by surface survey alone, no matter how intensive. Indeed, the initial archaeological reconnaissance of the island by Sinoto and Kikuchi [Emory and Sinoto 1963] failed to locate a single site Morphodynamics of the Lnd-Sea Interface of this early period. The model also has implications for the longertenn managemen of archaeological sites or 'cultual resources' of the To'aga area. We explore these implications in more detail in chapter 15, but briefly note here that if, as our model predicts, the To'aga area is now undergoing a phase of active erosion and shoreline transgression, the archaeological sites of this area will in time be eroded away. How quickly these sites will be theatened by erosion depends on several factors, especially the rate of subsidence and of eustatic sea-level rise. The problem of 'global waming' and associated sea-level ises (Geophysics Study Committee 1990), however, could potentially rsult in an acceleration of coastal erosion in the To'aga area. ACKNOWLEDGEMENTS I am particularly grateful to Joanna Ellison of the Depanment of Geography, University of California at Berkeley, for kindly supplying me with vadous references on sea-level change in the southwest Pacific, for the use of figure 4.4, and for her comments on an earlier draft of this chapter. REFERENCES CITED Ash, J. 1987. Holocene sea levels in norerm Viti Levu, Fiji. New Zealand Journal of Geology and Geophysics 30:431-35. Bloom, A. L. 1980. Late quatemary sea level change on south Pacific coasts: A study in tectonic diversity. IN N-A. Momer, ed., Earth Rheology, Isostasy and Eustasy, pp. 505-516. New York: John Wiley and Sons. 1983. Sea level and coastal changes. IN H. E. Wright, ed., Late Quaternary Enviroinents of the United States, vol. 2, The Holocene, pp. 4253. Minneapolis: University of Minnesota Press. Chappell, J. 1982. Sea levels and sediments: Some features ofthe context of coastal archaeological sites in the tropics. Archaeology in Oceania 17:69-78. Crosby, A. 1986. Ecological Imperialism. Cambridge: Cambridge University Press. Daly, R. A. 1924. The geology of American Samoa. Carnegie Institution Publication 41 340:93-152. Dana, J. D. 1849. Geology. Vol. 10 of U. S. Exploring Expedition Durng the Years 18381842 Under Cmmand of Charles Wilkes, USN. Washington D.C.: Govenmnent Printing Office. Davis, W. M. 1928. The Coral ReefProblem. Amercan Geophysical Society Special Publication No. 9. New Yoik. Ellison, J. 1989. Pollen analysis of mangrove sediments as a sea level indicator. Assessment from Tongatapu, Tonga. Palaeogeography, Palaeocibnatology, Pakaeoecology 74:327-41. Fairtmidge, R. W. 1961. Eustatic changes in sea level. Physics and Chemistry of the Earth 4:99185. Farrell, B. H. 1962. The village and its agriculture. IN J. W. Fox and K. B. Cumberland, eds., Western Samoa: Land, Life, and Agriculture in Tropical Polynesia, pp. 177-238. Christchurch: Whitcmbe and Tombs. Friedlander, I. 1910. Beitrage zur der Samoa Inseln. K. Bayer. Akad. Wiss., Math-Phys. El., Munich, Band 27:358-69. Geophysics Study Committee, 1990. Sea-Level Change. Studies in Geophysics. Washington D.C.: National Academy Press. Green, R. C., and J. M. Davidson, eds. 1974. Archaeology in Western Samoa, Vol. 11. Bulletin of the Aucldand Institute and Museum. Green, R. C., and H. G. Richards 1975. Lapita pottery and a lower sea level in westem Samoa. Pac#fc Science 29:309-315. Isla, F. I. 1989. Holocene sea-level fluctuadon in the southem hemisphere. Quaternary Science Reviews 8:359-68. Kirch, P. V. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memoral Washington State Museum Monograph No. 5. Seatte: Burke Museum. Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evoludon in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., and T. L. Hunt, 1988. The spatial and temporal boundaries of Lapita. IN P. V. Kirch and T. L. Hunt, eds., Archaeology of the Lapita Cultural Complex: A Critical Review, pp. 9-31. Thomas Buike Memorial Washington State Museum Research Report No. 5. Seattle: Burke 42 The To'aga Site Musewn. Kirdc, P. V., and D. E. Yen 1982. Tikopia: The Pre/udtory and Ecology of a Polynesian Outlier. Honolulu: Benice P. Bishop Museum Buletin 238. Leach, H., and R. C. Green 1989. New infonnadon for the Ferry Berth Site, Mulifanua, Westem Samoa. Journal of the Polynesian Society 98:319-29. McDougall, I. 1985. Age and evoluton of te volcanos of Tutuila, American Samoa. Pac#fc Science 39:311-20. McLean, R. F. 1980a. The land-sea interface of small tropical islands: Morphodynamics and man. IN H. Brookfield, ed., PopulationEnvironment Relaions in Tropical Islands: The Case ofEastern Fiji. MAB Tehnical Note No. 13. Paris: UNESCO. -. 1980b. Spatial and temporal vadability of extenal physical controls on small island ecosystems. IN H. Brokfield, ed., PopulaionEnvirovnent Rela s in Tropical Islands: The Case ofEastern Fiji. MAB Technical Note No. 13. Paris: UNESCO. Menard, H. W. 1986. Islands. New Yoik: Scientific American Libray. Miyata, T., Y. Maeda, E. Matsumoto, Y. Matsushima, P. Rodda, A. Sugimum, and H. Kayanne 1990. Evidence for a Holocene high sea-level stand, Vanua Levu, Fiji. Quaternary Research 33:352-59. Montaggioni, L F., and P. A. Pirazzoli 1984. TMe significance of exposed coral conglomerates from Fenh Polynesia (Pacific Ocean) as indicators of recnt rladve sea-level changes. Coral Reefs 3:29-42. Nakamura, S. 1984. Soil Survey ofAmerican Samoa. Soil Conservation Service, U.S. Department of Agriculue. Washington: Government Printing Office. Natand, J. H. 1980. The pgion of volcanism in the Samoan linear volcanic chain. American Journal of Science 280-A:709-735. Nunn, P. D. 1990. Coastal processes and landfonns of Fiji: Their bearing on Holocene sea-level changes in the south and west Pacific. Journal of Coastal Research 6:279-3 10. Pirzzoli, P. A., and L. F. Montaggioni 1986. Late Holoce sea-level cages in the northwest Tuamotu Iands, French Polynesia. Quaternary Research 25:350-68. 1988. Holocene sea-level changes in French Polynesia. Paleogeography, Paleoclimatology, Paleoecology 68:153-75. Rodda, P. 1988. Visit to Westem Samoa with HIPAC team. IN N. Yonekura, ed., Sea Level Change and Tectonics in the Middle Pac(i1c. Report of the HIPAC Project in 1986 and 1987, pp. 85-90. Deparunent of Geogrphy, University of Tokyo. Shiffr, K B. 1987. Formation Processes of the Archaeological Record. Albuquerque: University of New Mexico Press. Schofield, J. C. 1969. Notes on late Quatemary sea levels, Fiji and Ramtonga. New Zealad Journal of Geology and Geophysics 13:199206. Shepard, F. P. 1963. Thirty-five thousand years of sea level. IN Essays in Marne Geology. Los Angeles: University of Califonia Press. Stearns, H. T. 1944. Geology of the Samoan Islands. Bulletin of the Geological Society ofAmerica 55:1279-1332. Stice, G. D., and F. W. McCoy, Jr. 1968. The geology of the Manu'a Isands, Sanoa. Pacfic Science 22:427-57. Sugimum, A., Y. Maeha, Y. Matshima, and P. Rodda 1988. Further report on sea-level investigation in Westeem Samoa. IN N. Yonra, ed., Sea Level Change and Tectonics in the Middle Pacific. Report of the HIPAC Project in 1986 and 19879 pp. 79-84. Department of Geogray, University ofTokyo. Visher, S. S. 1925. Tropical Cyclones of the Pacifc. Bernice P. Bishop Museum Bulletin 20. Honolulu. Wiens, H. J. 1962. Atoll Environment and Ecology. New Haven: Yale University Press. Yonekur N., T. Ishii, Y. Saito, Y. Maeda, Y. Matsushima, E. Matsunoto, and H. Kayanne 1988. Holocen fringing reefs and sea-level change in Mangaia Island, Southern Cook Islands. Palaeogeography, Palaeoclitology, Palaeoecology 68:177-88. 5 EXCAVATIONS AT THE TO'AGA SITE (AS-13-1) PATRICK V. KIRCH AND T. L. HUNT INTRODUCTION NTrr of fte Manu'a of the archaeolknowledge in 1986, our Project ogy and prehistory of the Manu'a Islands was lImited to a few late prehistoric architectal sites and to surface finds of stone adzes and flakes (Kikuchi 1963; Emory and Sinoto 1965; Clark 1980). No wellstratified occupation sites were known, and there was no established sequence of occupation phases extending back to the Ancestral Polynesian period, as had been developed for Westem Samoa (Green and Davidson 1969, 1974). bdeed, previous archaeological reconmaissance in Manu'a had failed to locate any sites that contained prehistonc pottery and thus might date to the first millennium B.C., based on comparisons with the Samoan sequence as defined for Upolu Island. A major goal ofour 1986 reconnaissance survey of Ta'u, Olosega, and Cfiu islands was thus to determine whether the appaent absence of ceramicbearing sites was simply an artifact of archaeological sampling, or whether such early perod sites were tnuly absent in the Manu'a Group. The surface find of a Polynesian Plain Ware sherd in Ta'u Village on the first day of our 1986 reconnaissance was followed up by test excavations, revealing the presence of subsurface occupation deposits dating to the Ancestral Polynesian period (Hunt and Kirch 1988). This discovery was sufficient to establish that the prehistory of the Manu'a Islands PRIOR Tro Th coMmEN would in general terms parallel that of the larger and better known Westem Samoan group. Moving from Ta'u to Ofu Island, we subsequently discovered a pottery-bearing site at To'aga on the southem coast, where the Public Wo*ks Department had bulldozed a landfill dump and thereby exposed subsurface deposits. This area was within the site AS-13-1 defined by aark (1980). We canied out a limited test excavation at the To'aga landfill in 1986, eovering pottery and other artifacts in situ (see below). Because of the significance of the To'aga materals for elucidating Manu'a prehistory, we expanded our excavations in 1987, revealing that the site was far more extensive than originally thoughL A third season of trnsect test excavations was carried out in 1989 to assess the full area of this extensive and deeply staified site. In this chapter, we present the details of these three excavation seasons, including the specific excavadon pxcedures used, as well as the srtigraphy and depositional sequences revealed. Subsequent chapters provide detailed analyses of the ceramics, other portable artifacts, and faunal materals, as well as discussions of radiocarbon chronology and site geomorphology. Field Methods: General Comments Some aspects of field methodology which were constant in all field seasons may be brefly summa- rized. The prncipal sampling strategy used was that 44 The To'aga Site of "systematic trasec " (Redman 1974; Kirch and Yen 1982; Kirch 1988), with 1r-m excavation units generally spaced at 10 m intervals. We oriented these transets perpendicuarly to the coastine in order to provide geomorphic proffles across the coastal flat All trae pfeswee canied out to the reefflat and coI to mean sea leveL Usng a tlescopc level and stadia md, a srface elevato prfile was obtained for each trmuect. Observan were also made of soil vanation and vegetation associations along the trasect, as thse data indicated cuntland use and provided im clues as to the geomorphic history of the site. This strategy of acquiring data (surface and excavation) along systematic transects has the advantage of producing results that can be mapped and interpolated. Whenever possible, excavation proceeded according to natral stigraphy, although arbitray subdivisions were made within thick strata Individual excavation blocks (refened to as "spits') were designated within strata, but never cross-cut stratigraphic boundaries. Detailed rwcords were maintained on aarz recording forms (fig. 5.1), on which horizontal feames were drawn to scale and all localized finds were plotted according to x, y, and z coo.rinates. Aftrexcavation, measuwd stratgrapic pofiles were dawn for each excavation unit, and all sta were described in terms of thickness, boundary, morphology, color, lithology, cultual conte, and other caracteristics. Color designations are from fte Munsell soil color charts (Munsell 1988). Most proffile descriptions were done by Kirch to assure consistency (in 1989, Hunt described several units after Kirch left the field to initiate a project in the CDook ls). Following the drawing and description of each profile, a senres of sediment samples were taken; sediment samples were taken from within stata, never cross-cuttng stratigraphic boundaiies. These samples were subsequently analyzed in the University of California, Berkeley, geacaeology laboratory (see Kirch, Manning, and Tyler, chapter 7). All excavated sediment was sieved through 0.25 inch mesh screens. Although smaller mesh (particularly 0.125 inch mesh) would have enhanced the recovery of minute faunal remains, we opted against tfhis screening stategy in the inreests of covering a greater area duing this testing phase. Given our primary objective of detemuning t areal extent and nature of the subsurface deposits at To'aga, sufficient areal sanpling was judged to be a more important consideation than complete faunal recovery. During screening, however, we made perodic checks of the small size fraction passing thrgh the 0.25-inch mesh screens and were generally satisfied that the bulk of the faunal materals was being retained. In addition, bulk sediment samples were taken from strata in the areal excavation blocks, as well as from some of the individual test pits (see Nagaoka, chapter 13). These samples allowed us to assess the frequency of minute faunal remains. All vertebrate and invertebrate faunal materials caught in the sieves were bagged and shipped back to the laboratory for identification and analysis. Preliminary sorting of the vertebrate fauna, as well as detailed identifications of the avifaunal remains, were carried out by Dr. David Steadman of the New York State Museum (see chapter 14). Furfter analysis of the non-bird vertebrates and of the invertebrate fauna was undertaken by Lisa Nagaoka of the University of Washington (see chapter 13). The numbering of stratigraphic units (distinguished by roman numerals) is based primarily on sedimentological (rather than culturl) critera Each sedimentological-lithological unit that was determined to have been deposited either as a single event, or as several events all representing the samne source and mode of deposition, was designated as a "layer." Subunits withn these sedimentological layers, including cultural occupadons, are designated by letters. Hence, an occupation episode that is wholly incorporated within a calcaeous sand beach ridge deposit, Layer III for example, might be designated as Layer HIB, with e lturally sterile, but lithologically identical sands above and below designated Layers hIA and IIIC. Thus, our layer designations emphasize fte geomorphological site formation processes rather than simply the pesence or absence of cultural materials. In describing the lithology of beach rdge depositional units, we paid special attention to admixture of volcanic lithic grains with the dominant calcareous grains. When volcanic grains are present, these give the sediment a 'salt-and-pepper' appearance. The signficance of this 'salt-and-pepper' lithology is that it reflects the availabiity of a volcanic sediment source from which sand grains Excavation I 45 x '' PM. 89 MNa 'a Project, 199 Sieve& DIy cI4.et 0 &zt Sit As-i- A"11111m 'r-41s0o5 S'IddaW Spit F ; 1 Df~~ Depths: Surfae Ciiatum 0O .L .. .. .; .t End Ieve -P I N yI__ m Mea En*cm|!r | Mean StartZ= *_cm- COMMENTS (Note sediment char~teristics, color; T disturbances; samples taken; speclproblems): 0 (J STl &SALT' (tu44AaM foft(6 -tVt tA~8ftC-4 .~~.)rL -~~~~~ CliAkColF - u S .b^t i.* OL~ jA B,e~ $ - * , 6ete8)M}(~s mVAIO 7 468( 74S,5 A ht5RP) LAAYE tc.'oAn 1' &OlCIWA I;"? -SVAItrat msScr''So6( F'S -T AAz VAN,*w - _ Vw--To Ftw&V Rr'.AA-TtO%Z d Re lb o ~~~~~~~~~~CJ #Ai ,r~u lm 1-5.Ud'I¢°~W/ OLA. CbCgAC Ceix P PC"' Ct- 5LIRP1,^ e c: fS 5)g fw'A Nr.v T ->psli,o# ShApSIOL. 7-*A7VST65CmA- - 0 P 6C&AC-46 Recorder: C. mw,ae-ls P I ALJW1 L -m Figure 5.1 Example of the excavation recording form used in the is for level 5 of Unit 20 on Transect 9. 1987 and 1989 To'aga field seasons. This form 46 The To'aga Site were produced by wave action along the fonner shoreline. Today, fte sandy beaches fng To'aga are almost pwely calcares. Volcanic grains are presen only in a few spots, where volcanic headlands p gh the coastal terrace and are exposed to wave action. The frequent presence of 'salt-and-pepper' lithologies in deeply buried deposits exposed by our excavation indicates a landform stage whm the coastal ten-ace was much narrower (prior to progradaton), and when substantially more volcanic material was available for incorportion into the sediment budget Thus, along most transects, there is a tran from deeper, earlier 'salt-and-pepper' sand lithology to pudely calcareous lithology, reflec coastalion and removal of most volcan:ic source material from the sediment system. In chapter 7 we use point counting of sediment sampes to furtxr document the precise qutities of volcanic lithic grains in strata within various excavaton units. containing coral and basalt pebbles and cobbles, shell midden (mostly Turbo spp.), and ceramics (four thickware sherds orly). Layer II rached a maximum depth of 135 cm below surface. The contact with Layer II is very diffuse. A single adiocarbon date (Beta-19742) on Turbo shell yielded a conventional radiocarbon age of 2350 ± 50 B.P. (cal 28 B.C.-A.D. 108 at one s ard deviation; see chapter 6). Layer m: No culturl material was pent in the white (10 YR 8/2) calcareous sand of this layer. Excavation of test Unit A reached a maximum depth of 160 cm below surface. Subsquent woik in 1987 revealed ta these major stratigraphic zones fit well within the geomorphological sequence for the To'aga flat as a whole (see below). Artifacts collected frxm the bulldozed area of the To'aga landfill and from Unit A were described and illustrated by Hunt and Kirch (1988:169-76) and are included here in chapter 1 1. 1986 TEST EXCAVATION THE 1987 EXCAVATIONS TIe Toaga site (AS-13-1) was initially discovered duing te 1986 naissance survey, when examination of a deep bulldozer cutting made by the Public Woiks Department for a sanitary landfill distbed a buned cultual deposit contning Polynesian PaiOn Ware cermnics (Hunt and Kirch 1988:168). The bulldozercutting lies atthe edge of a fan ofmassive talus bouldes. It appears that the landfill site is at the inland-most location possible on this paricular section of the coastal flat A single test pit (1 m2) was placed dirty adjacent (3 m west) to the bulldozed area of the To'aga landfill. This locale had mained undistubed by the bulldozer activities. Excavation of the test pit (designated Unit A) revealed the following major stata: Layer I: This dark (2.5 Y 2V0) mucky sand mixed with coliuvial clay is eniched with organic maer (contemporary A soil horizon). It contains coral pebbles, shell midden (mosdy Turbo spp.), and waterwom basalt pebbles and cobbles. Layer I reached a maximum depth of 60-70 cm below surface. The contactwith Layer 1is diffuse over a 1-3 cmzone. Layer U: This layer had motted pale yellow to grayish brown (2.5 Y 7i2 & 512) calcareous sand The 1986 surface collections from the landfill site, combined with the limited results frm the test excavation, revealed the presence of early Polynesian occupation deposits in the To'aga area. Along with the discovery of pottery at Ta'u Village on Tau Island (Hunt and Kirch 1988), this was the first record of an early phase of occupation in the Marn'a Islands. With e concurence ofthe Amedcan Samoa Historic Preservation Officer, we therefore determined ta a major objective for a second season of fieldwork in Manu'a should be more intensive investigation of the To'aga site, with the aplicaton of subsurface systemadc transect tting in order to determine whether undistubd potteiy-bearing occupation deposits were pnt in the vicinity of te landfill site. Tibs second phase of work was caried out in 1987. Excavation Procedures Returning to the To'aga site in 1987, we focused on the coastal flat immediately northeast ofthe ladfill which had not been disurbed by bulldozing (fig. 5.3). As can be seen in the transect profle in figure 5.2, this flat is about 125 m wide from the base of the steep colluvial-alus slope to the Eavatio 47 oD CL) OC E 00E 0 a. *. Beach ca. : Beach ca. 2000 B.P. 0 o 25m 350BP VERTICAL SCALE X4 Figure 5.2 Elevation profile along the 1987 excavato transect, showing the positions of the main excavato and 3, in relation to geomoaphic and pedologic feat. shoreline. Using the systematic tranct strtegy developed by Kirch for sub-surface sampling in similar coastal settings in Tikopia and prese Niuatoputapu islands (Kirch and Yen 1982; Kirch 1988), we laid out a transect baseline across the Toaga coastal fiat, extending at a location 60 m of the Ofu landfill. The first ree units excavated along the transect (at 0, 15, and 45 m from the base of the talus slope) revealed a complex and deep cultual stratigraphy in Unit 1, but only shallow cultual deposits overlying calcareous beach sands in the more seaward Units 2 and 3. A shovel test at 105 m fardher seaward along this transect, near the crst of the modem beach ridge, revealed a total absence of culural deposits, with only calcareous sand. These tests thus demonstated that the oldest cultunal deposits were to be found close to the base most of of the steep talus and volcanic cliff, and the coastal flat consisted of culturally sterile coral sands and reef detritus, which had been deposited during seaward progradation during the past 2-3 kyr northe B.P. Following these iniitial tansect tests, Unit 1 was expanded into a larger excavation in order to effectively sample the deep stratigraphic sequence, including the in-situ deposits of Polynesian Plain Ware ceramics. Units 4-9 were excavated, joining with Unit 1 to fonn a T-shaped trench as shown in figures 5.3 and 5.4. All units were dug trugh the Layer II calcareous sand deposit containing pottery, while Units and 6 were caried deeper into under- and Units 2 lying Layers HI and IV (see Stratigraphy, below). The third stage of our 1987 excavation strategy was to detennine the lateral extent of the early pottery-bearing deposits southwest and northast of the main excavation, parallel to the base of the talus. Unit 10 was thus laid out 45 m southwest of the baseline transect, as close to the base of the talus as feasible (acmally set in among several massive rockfall boulders). This revealed a stigraphy similar to that in the main trench, including deeply buried cultural deposits with Polynesian Plain Ware and one fine, thin-ware sherd. Unit 1 1 was then laid out 45 m nonheast ofUnit 1. Unit 11 revealed ceramic-bearing deposits, but these were tnncated by a large, deep pit (probably a late prehistodc ua'i masi or breadfruit-fermentation pit), and so an adjacent square, Unit 14, was opened to clarify the stratigraphy. In order to get as close as possible to the base of the talus, another test was laid out 15 m northwestof Unit 11, atthe footofthetalus and designated Unit 12. In Unit 12, a massive deposit of colluvium and large angular boulders had to be penetated to a depth of 1.8-1.9 m before we were able to reach a thin deposit of calcareous sand containing four thin, fine-tempered potsherds (one rim and four body sherds). These test units indicated that early, pottery-bearing cultual deposits at To'aga extend over a distance of at least 105 m northeast of the Ofu landfill, in a narrow zone at the base of the colluvial-talus slope. A final t, Unit 13, was opened in the centerof 48 The To'aga Site 636 CM .5 I- 0 I- z z E z 0 Uf) G) C: 0 F4 -0 0. I- z l a) .8 0 -4 L.. 0 Iz A CD o c 0 X 0 (a As 'I U 01 Lt CM I- Ct) I- z z *s g Iza d) '3> )& Vol 0 n C 4._ Co bO I) 0 ol: 0 O- /.:N z -S O.. : 6:2 .... 1eI wI Ecavations Figure 5.4 49 Excavation in progress in the 1987 main trench. Note the large talus boulder on the left. P. Kirch is recording the stratigraphy of the south face of Unit 1. a low pavement of waterwom pebbles (such pavements are called 'ili'ili in Samoan) apparently marking a later prehistoiic house floor. Our objective in excavating Unit 13 was to obtain a sample of this later prehistoric midden to contrast with the older, ceramic-associated assemblage. Stratigraphy of the 1987 Main Excavation Although the stratigraphy of the 1987 excavations varied from unit to unit, the most complete depositional sequence was revealed in the main trench excavations (Units 1, 4-9), but was reflected as well in Units 10, 11, 12, and 14. Seaward of the main trnch, Units 2 and 3 displayed simpler stratigraphic profiles, resulting from the later progradation of calcareous beach sands. The westem profile of the 1987 main trench is fsown in figure 5.5, in which all of te principal depositional units are represented. The stradgraphic units follow: Layer IA: The upper 15-20 cm portion of the upper colluvium (10 YR 3/1.5) found here has been heavily reworked by gardening. Various planting pits or depressions are detectable in te secdon. Layer IB: This is a massive deposit of reddish brown colluvium (10 YR 312), very compact, with no internal lensing or bedding evident. No charcoal flecking was observed. The deposit incorporates numerous angular to subangular weathered volcanic lithic fragments in the gravel-to-pebble size range. A small, lens-shaped pocket of slightly darker soil containing charcoal flecking was noted within Layer IB (designated Feature 1) and may represent a garden bum feature. This would suggest that Layer IB accumulated gradually and that the land surface was internittendy gardened during its deposidon. Layer IC: The basal 15-20 cm of the upper colluvium (10 YR 2/1) makes up this sm. It is somewhat darker than e overlying Layer IB, containing dispersed flecks and chunks (5-10 mm) of charcoal. The deposit is associated with an earth oven feature in Unit 9, and the charcoal (which is concentrated in a zone around the oven) appears to 50 The To'aga Site ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~. fi 0 ,| Il0_, _ S *,b;''j]. I- - t (f _ W 1t_I- 5 -0\'' i CD~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C t -D LO W (w\ a 1ll @ _/z1 l, W--< tt X |; @ i \ rib -C _'t'S 1- ox t § :'-\ I k l 0 °l| < j~~~~~~~~~~~~~~~~~~~~~~- Z~~~~~~~~~~~~~~~~~~ - lo. .!~~~~~~~~~~\1 .E0 U; '~~~~~~~~~~~~~olmnSmpeiu o ±.1'm Z ~~~~ Iii ~~~~~ ° 0Csh. I 0 IIgN ~~~~~~~~~~~~~ I-~ o E- z ColumnSampleCut~~~~~~~~~~b ci c~~ Depth in Meters c~ . |a Exavations have derived from oven rake outs. Some thick, coarse-tmpered pottery was also recovered from this deposit. It does not appear to have been a permanent occupation, however, and may represent a short-term or intermittent occupation such as a field shelter on a former land surface. Layer H: This unit with its subdivisions represents a period of active deposition of a calcareous beach ridge. Layer HA-1: A thin zone at the top of Layer HA, slightly darker and organically enriched (10 YR 512), this stratum contains anthrpophilic land snails and represents a phase of stabilizaton and vegetation of the Layer II sandy beach ridge. Some occupation in the vicinity is suggested by the presence of a few thick, coarse-tempered sherds and a thin scatter of marine shell midden. Layer HA: This deposit contains loose, calcareous beach sand (10 YR 7/2), not compacted or cemented and lacking culul materials. Although most of the sediment consists of calcareous materials, there is a subordinate quantity of basaltic lithic grains, giving the sand a distinct 'salt-and-pepper' appearanc. This indicates ta at the time of deposition there were exposed volcanic headlands in the vicinity of the beach, providing a source for the basaltic sand grains. Layer IIB: This principal pottery-bearing deposit represents a perod during which the surface of the actively accumulatng sandy beach ridge was occupied. Lithologically, Layer IIB is similar to HA and IIC, but with the addition of organic/cultural materials due to human occupation making it both daiker (7.5 YR 4/2) and more compacted. The deposit contains shell and bone midden, large quantities of small sea urchin spines and test fragments, ceramic sherds (prmarily of thick, coarsetempered ware), and other artifacts. It also contains antrpophilic land snails. The deposit is nonconcentrated and probably accumulated over a fairly brief span of time. Layer [IC: This layer, the basal component of the Layer II beach ridge consists, as with HA, of a 'saltand-pepper' lithology with dominant calcareous grains and a subordinate quantity ofvolcanic lithic fragments (10 YR 5/2). Toward the base of this deposit are numerous large coral cobbles and some angular volcanic cobbles, along with branch coral fingers and coral rubble. This material indicates a 51 relatively high energy deposidonal environmaet, such as storm activity along an exposed beach frnt. Thirteen sherds ofthin, fine-tempered ware were present, although te deposit showed no evidence of being an in-situ occupation locale. Layer II: Massive silty-clay colluvium (7.5 YR 3- 4/2) with some incorporated subanguar lithic fragments makes up this layer. Occasional charcoal flecks are present, pariucularly near te top of the deposit. Nine thin, fine-tempered potsherds were also incorporated in the deposit. It appears to represent a single depositional event, resulting from a combination of mass-wasting ad fluvial spot of terigenous sediment from the colluvial slope above te site. A lower zone, designated Layer IIIB, incorporates some sand mixed frm Layer IV, presumably at the time of deposition. The presence of charcoal flecking indicates burning of this slope prior to fte deposition of fte sediment, pediaps due to gardening or oter human disturbance. Layer IV: Here one encounters a mixed deposit of fine-grained calcareous sand and reddish silt-clay (7.5 YR 5/4), apparendy culturally sterile. Layer V: This basal deposit of 'salt-and-pepper' sand (10 YR 8/2) includes marine shell and reef detritus; fairly compact, and showing the initial stages of cementation probably due to frequent ground-water wetting. The deposit represents an active beach ridge depositional environment. It yielded two sherds of thin, fine-tempered ware, which may have derived from a primary occupation locus behind the beach (which would now be bured under at least 10 m of colluvium). Excavation Units 2 and 3 The stratigraphies of Units 2 and 3, seaward of the main excavation, were essentially identical; the section for Unit 3 (fig. 5.6) is typical: Layer I: This organically enriched, silty-clay loam is presently under cultivation. Lithologically, the deposit is a mixture of calcareous sand (77%), silt, and clay-sized terigenous grains, the latter deriving from sheet wash erosion from the colluvial slope. Various planting depressions filled with loose, rewoiked soil are visible in the section. ITh deposit also contains fragments of marine shell midden from occupation in the vicinity. An earth oven feature was exposed in the west face of the unit The To'aga Site 52 I UNIT 3: SOUTH I UNIT 3: WEST Adz Section I 0.0 0 4'0 40 2 0.5 c 0. 0 1.0 Figure 5.6 Stratigraphic section of the south and west faces of Unit 3. "C" indicates coral, and "V" indicates volcanic. Layer II: Here one encounters loose, calcareous beach sand which is culturally sterile. This deposit is composed wholly of calcareous grains and lacks the 'salt-and-pepper' combination of calcareous and basaltic grains found in the Layer II deposit in the main excavation This absence indicates that by the time of deposition of Layer 11 in Unit 3, the coastline had prograded sufficiently to buq the volcanic headlands. Excavation Unit 10 As shown in figure 5.3, Unit 10 was located 45 m southwest of the 1987 baseline transect. The pit was situated in a small clear space between several very large boulders, at the base of the steep talus. Our aim here was to get as close to the talus as possible, in the hope of exposing deeply buried cultural deposits. This unit revealed a staigraphic sequence similar to that in the main excavation trench. The north and east faces of Unit 10 are shown in figure 5.7. The strata are descnrbed below: Layer IA: This upper pordon of tfie colluvium (5 YR 3/2) is presently gardened and tius loose and reworked. Its silty clay loam includes angular lithic gravel. The contact with Layer IB is diffuse over a 2-3 cm zone. Layer IB: This very compact deposit of clayey colluvium (5 YR 4/4) contains subangular lithic fragments. Occasional charcoal flecks are prsent There is no evidence that this zone had been reworked by gardening, and it probably represents a single depositional event. The contact with Layer IC is diffuse over a 2-3 cm zone. Layer IC: The basal porton of the colluvium (5 YR 312), this deposit is similar to IA, and may represent an older land surface that was gardened. It is less compact than Layer IB and contains some anthropophilic land snails (e.g., Lamellaxis gracilis). Also, some disturbances penetrate down into Layer II. The contact with Layer nI is distinct. Layer HA-1: This thin, discontinuous deposit of calcareous sand is enriched with organic matter and stained gray (color 5 YR 4/1). Land snails of several genera are present, indicating the fonner presence of a stable land surface with vegetative cover. This deposit is a former paleosol fonned on the old beach ridge represented by Layer H. It reflects a phase of stability capping the earlier phase of beach ridge accumulation. This paleosol must have formed after pgration of the shoreline removed the sotve of cal os sediment that had resulted in he deposiion ofthe deeper Layer ldeposits. The cntact with Layer IIA is diffuse and gradational over a 3-4 cm zone. Excavations UNIT 10: NORTH 53 UNIT 10: EAST 0.0 0.5 at 0 0 1.0 c.0. 0 a 0 -1.5 Stratigraphic section of the north and east faces of Unit 10. 2.0 Layer IIA: A culturlly sterile deposit of 'salt-andpepper' sand (7.5 YR 7/2), this stratum is predominantly calcareous but incorporates some volcanic grains indicating the exposure of volcanic headlands near the site at the time of deposition. The deposit is fairly loose and was probably developed rapidly after the abandonment of the underlying Layer UB occupation. The contact with Layer IIB is very irregular but relatively sharp. Layer HIB: This is a ceramic-bearing occupation deposit (7.5 YR 5/2). The parent lithology is essentially identical to Layer IHA. This layer represents a period of occupation on the beach ridge while te active deposition of sandy sediment continued, primarily through aeolian action. The deposit is stained grey with finely dispersed organic material but is very sparse in midden (not nearly so concentrated as Layer IIB in the main excavation trench). The contact with Layer UC is diffuse. Layer UC: The mixture of calcareous sand and reddish clay (7.5 YR 5/4) found here prbably results from fluvial/hydraulic tnsprt of the clay from the talus upslope. It is culurally sterile. The contact with Layer IIn is sharp but iregular. Layer m: This compacted deposit of reddish brown clay-silt (5 YR 3/4) includes no larger lithic fragments, suggesting a fluvial rather than masswasdng mode of deposition. No charcoal flecks were observed. This deposit appears to represent a single depositonal event, perhaps associated with a 54 The To'aga Site storm or cyclone. The contact with Layer IV is somewhat diffuse and mixed. Layer IV: The 'salt-and-pepper' beach sand (10 YR 812) of this layer incorporates larger coral cobbles and waterwom branch coral fingers. The layer represents a fairly exposed, high-energy beach nidge depositional event. Apart from shell and bone midden, Unit 10 yielded little artifacul material. A few small flakes of obsidian were wecovered from Layer IIA-1, and a single thinware body sherd (5.7 mm thick) with an orange (5 YR 5/6) slipped exterior came from Layer IIB. Excavation Units 11114 As shown in figure 5.3, Units 11/14 were located 45 m east of te baseline transect. This location is among massive talus boulders where the ground surface is relatively level. Excavation of Unit 11 revealed a large pit feature over the entire unit which extended from the lower colluvium (ca. 55 cm below surface), cutting down through calcareous sand layers to 2.25 m below surface. The fill of this pit contained 181 thickware sherds and 4 thinware sherds (2 with red-slipped exterors). Field observations indicated ta pottery came from the lower deposits adjacent to the pit and not frm the pit fill itself. As nonnal stratigraphy for this area could not be discemed with the intnasion of the deep pit, Unit l 1 was expanded into arectangulartrench (1 by 2 m) with the excavation of Unit 14. Unit 14 exposed an undistrbed sraigraphic sequence, as follows: Layer IA: The upper 20-30 cm of colluvium found here was dark brown (10 YR 312) and had been reworked by gardening. This layer contains some charcoal flecking (probably modem) and no pottery. Layer IB: This massive lower colluvium is very poorly sorted and includes angular volcanic rock fragments with no evident bedding or lenses. No charcoal or other cultual materials were observed in this dark brown (10 YR 312) layer. Layer IC: This sm contained a dark grayishbrown (10 YR 3/1) mixture of clayey colluvium and calcareous sand with abundant midden, including shell, bone, pottery (309 sherds, with 304 [98%] thick-coarse, and 5 [2%] tiinware), and charcoal. This deposit was the source of the fill of the large pit described for Unit 1 1, but in Unit 14, it made up a discrete area between 54-75 cm below surface. This layer appears to represent a depositional event contemporaneous with Layer IC in the main trench (see above). Layer U: This calcareous sand deposit is enriched with organic matter and some eluviation of clay particles from the clay-rich Layer IC above. Layer II is brown (10 YR 5/3) and contains pottery (15 thickware sherds) and a relatively sparse amount of shell midden. This stratum is comparable to Layer II in the main area excavation. Layer Im: Calcareous sand (10 YR 7-812) with waterwom coral cobbles and unwom branch coral makes up this layer. It contains little cultural material and appears to represent a high energy depositional environment. Layer III compares in lithology and content to Layer IIC in the main trench. Layer IV: This brown (7.5 YR 4/2) mixture of silty-clay colluvium and calcareous sand has some organic matter. Few larger lithic cobbles or boulders are present, suggesting a sediment that probably represents sorting with surficial flow of the terrigenous source, and wind (saltational) ransport of the calcareous (beach ridge) source. A small hearth comprised of waterwom coral boulders surrounding a shallow ash concentration was excavated at the surface of this layer in Unit 14. Five thinware sherds were ecovered from this stratigraphic boundary. Excavation was discontinued at the surface of Layer IV due to time constraints at the end of dte 1987 season. Excavation Unit 12 Unit 12 was located 16 m north and 45 m east of Unit I inthe maintrench (see figure 5.3). he unit is situated on the colluvial slope amidst huge talus boulders. Excavation of this unit (chosen to penetrate the subsurface deposits as close to the talus slope as possible) proved to be hazardous given te large boulders protruding from the unstable sidewalls of the pit, especially as the clayey sediment dried and became friable. The stratigraphy of Unit 12 was as follows: Layer IA: The upper colluvium (0-20/30 cm below surface) consists of an organically enriched, A horizon soil. The layer is dark brown (10 YR 3/1) Exavtions and has been rewoiked by recent gardening. No cultural material was observed. Layer IB: The massive lower colluvium, dark brown (10 YR 3/2) in color, consists of very poorly sorted, large, angular rock fragments, with cobbleand boulder-sized clastics, in a clay-silt matrix. No bedding or lenses of finer-grained sediment were observed, nor was any cultural material noted. Layer IC: This is the base of the colluvial deposit, distinguished by an incmase in fine-grained sediment The boundary with Layer IB is gradual, reflecfing a continuous icreae ofclay and silt conte toward fte base of Layer I. This day-loam aains some calcamous sand (thoroughly intermixed) and but these are reangular volcanic rck frag stricted to cobble-sized. Charcoal flecks and waterwom coral gravel are present in small amounts. Small amounts of bone and shell were recovered from this layer, but pottery was absent. Layer I extends from the surface to 2 m below surface. Layer IC-1: This is a brown (10 YR 3/3) clay-loam with calcareous sand thoroughly intermixed. Layer IC- I can be distinguished by both its greater calcareous sand content and the presence of pottery. The pottery in this layer is predominantly thinware (25 sherds, with 2 thick-coarse and 23 thinware) and includes duee direct rims, one of them with red slip and an impressed lip. Shell and charcoal were recovered in small quantities. The deepest (2.6 m) portions of this layer had no cultural materials. The stratigraphy of Unit 12 was almost entirely colluvial in origin. The upper pordons, free of cultural materials, represent one or more rapid depositional events, i.e., mass wasting of material from the nearby steep volcanic cliffs. The clay-loam of the lowest portions of the sequence represents gradual accumulation of colluvial sheet wash during a perod of occupation. Calcareous sand entered the deposit through wind trasport (saltation) from the active beach ridge, which was then much more closely adjacent to this location. , Excavation Unit 13 Located 45 m east and 37 m south from Unit 1 in the main trench, Unit 13 was near the center of a stone-faced, waterwom gravel-filled ('ili'ili) mound typical of Samoan house foundation construction. The gravel mound measures 15 by 16 m across its 55 center and is nearly round in plan view. The mound is elevated approximately 25-30 cm above the surrounding organically enriched and stable sandy substrate. Excavation of Unit 13 revealed the following stratigraphy: Layer I: This deposit of waterwom coral gravel ('ili'ili) is mixed with a small prporton of waterwom volcanic gravel and with a finer, daick organically enriched, sandy loam matrix (7.5 YR 2/ 0). A bottle glass sherd was excavated from this layer, possibly indicating occupation of the house mound into the Historic period. Layer IB: This is a lighter colored (10 YR 5t2) sediment of the same texture as Layer I. The color difference is due to a lower percentage of organic matter. Ash, shell, and numerous sea urchin spines were mixed into the gravel paving stones. Excavation of the lower part (3048 cm below surface) of this stratum exposed a small pordon of a human burial cut from the floor surface (Layer I) of the house mound and extending into Layer II below. Only bones from the feet were recovered, the remainder of the burial lying outside of the excavation uit. Layer H: This stratum contains culrally sterile, white (10 YR 8/2). calcareous sand. Excavation was completed with the testing of this layer to 68 cm below surface. Radiocarbon Chronology of the 1987 Excavations Seven samples of charcoal and shell from the 1987 excavation units at To'aga provide the basis for a radiocarbon chronology of the deposits in the vicinity of the Ofu landfill. The 1987 dates were reported in full by Kirch, Hunt, and Tyler (1989) and are discussed in detail in chapter 6. The oldest dates are those for samples Beta25035 and -25673, both frm Layer V of the main trench, the basal calcareous beach sand which yielded two thin, fine-tempered sherds. Both samples consisted of unweathered marine shell which were probably deposited at approximately the same time as the sherds (details of these samples are provided in Kirch, Hunt, and Tyler 1989:11-12; see also Kirch, chapter 6). The samples yielded ovedapping ages at two standard deviations which indicates a calibrated time range of between 3700-3300 cal B.P. for the deposition of dtis beach sand which 56 The To'aga Site contained the thin, fine-tempered ware pottery sherds. While this age range is certainly early for human occupation in Westem Polynesia, it is not out-of-line with the earliest known dates for Lapita sites in the region (Kirch and Hunt 1988; Kirch 1988, table 48). Three samples are in direct association with the Polynesian Plain Ware assemblage (Beta-25034, -26464, and -25033), denrving from th Layer II occupation in the main trench and from its corelated deposit in Unit 10. All three samples agree well in age and indicate a calibraed time range of between 2500-1900 cal B.P. for Xt main Polynian Plain Ware occupation. A terminus post quem for the use of ceramics at the To'aga site is provided by sample Beta-26463, which derives from the base of an aceramic occupation deposit which statigraphically postdates Layer n in the main trench. This sample yielded a calibrated age of 1389-1287 cal B.P. The final sample (Beta-26465) was obtained from the base of the aceramnic house mound tested by Unit 13 and yielded a calibrated age of 1122-950 cal B.P. Since the platform is consructed on the present land surface, it is evident that the sequence of coastal progradation and of colluvial deposition in the vicinity of th main excavation had stabilized by the beginning of the second millenniUM A.D. The Depositional Sequence in the 1987 Main Trench The depositional sequence in the 1987 main trench excavation at To'aga can be summadzed as te following series of stages: Stage 1 (3700-3300 Cal B.P.): The calcareous sand beach represented by Layer V of the main trench was formed at a time when the island's shoreline was close to the base of the talus slope. Thus, volcanic headlands were exposed to active wave erosion, yielding mixed calcareous-basaltic lithology sands. Human occupation in the vicinity of the beach (presumably now buried under a considerable depth of colluvium and talus) is suggested by the presence of tin fine-tempered ware sherds. Stage 2 (> 2500 cal B.P.): The older beach ridge was buried at this time by terigenous silt-clay from upslope, with erosion initiated in part by human clearance of the vegetation with fire, indicated by charcoal flecking in the erosional deposit. Stage 3a (ca. 2500 cal B.P.): Accumulation of the calcareous beach ridge was renewed during this stage, while the shoreline was still close to the talus slope (revealed again by mixed calcareousbasaltic grain suites). Exposure to the active shoreline is indicated by the large coral cobbles and reef detritus, representing one or more storm events. Stage 3b (2500-1900 cal B.P.): Humans inhabited the active beach ridge surface (Layer IIB of the main trench) during this time, resulting in the deposition of ceramics and midden. Stage 4 (2500-1900 cal B.P.): In this stage, the beach ridge was abandoned as an occupation locus, and the accumulation of the beach ridge (Layer HA) continued. Stage 5 (ca. 1900 cal B.P.): The old beach ridge surface stabilized during this stage. This is indicated by the formation of an A soil horizon and the deposition of anhopophilic garden snails (Layer IIA-I of the main trench). The stabilization presumably resulted from seaward progradation of the shoreline, thus removing the immediate source of calcareous sediment. At roughly this time, the basal sands represented by Layer II in Units 2 and 3 were deposited. Stage 6 (< 1900 cal B.P.): During Stage 6 there was additional erosion and deposition of temgenous sediments from upslope in the area of the main trench. The presence of ceramics and an earth oven indicate a brief occupation event. Stage 7 (< 1900-1000 Cal B.P.): The progradation of the shoreline continued to its present position about 125 m seaward of the main excavation locus. Gradual deposition of fine-grained terigenous sediments over the surface of the newly formed coastal flat and the rewoiing of these deposits through gardening activities also occunred during this stage. Stage 8 (ca. 1000- 100 cal B.P.): Human habitation dispersed across the expanded coastal flat during Stage 8, represented by the 'ili'ili pebble pavement tested by Unit 13 and by other surface cultural features. in sum, the depositional sequence revealed in the 1987 excavations begins late in the second millennium B.C. with a coastal teriace only a few tens of meters wide, and with a reef platform substantially wider than at present. This geomorphic Excavations situation fits well with evidence for a +1-2 m higher stand of the sea in the southwest Pacific at ca. 3 kyr B.P., as argued in chapter4. By about 1900 B.P., the shoreline had begun to rapidly prograde, associated with a drop in sea level to its present stand and with exposure of the reef crest leading to a higher rate of calcareous sediment production. The development of a coastal terrace more than 100 m wide was accompanied by the addition of fine terigenous sediments due to erosion and sheet-wash of the higher colluvial slopes, forming a highly prducfive zone for intensive cultivation and habitation. lherefore, the morphodynamic model of coastal terace fonnation outlined above in chapter 4 is closely supported by the stratigraphic and radiometric data from the 1987 excavations Swnmary f the 1987 Excavaions The program of excavations carried out in 1987 was successful in establishing the presence of undisturbed, well-stratified prehistonic occupation deposits in the To'aga area. The basal occupation layers contained Polynesian Plain Ware ceramics, including a fine, thinware pottery dating to the midfirst millennium B.C. The 1987 excavations also established a sequence of geomorphological change in the To'aga area, indicating substantial coastal progradation and consequently deep burial of the pottery-bearing deposits under recent calcareous dune ridge and colluvial slope-wash deposits. By the close of the 1987 season, it had become clear that the subsurface archaeological resources of the To'aga area-especially te deeply buried, pottery-bearing deposits-were far more extensive than originally thought. The 1987 excavation units had revealed the presence of culural deposits over an area of at least 4,000 i2, but without reaching the horizontal boundaries of these layers. Because of the great significance of the To'aga site for Samoan prehistory, it was clearly essential to establish the full areal extent of the subsurface archaeological features at this extensive and well-staified site. A third phase of excavations was therefore required, applying the systematic tansect strategy of subsurface testing over the whole extent of the southeem coastal plain on Ofu Island, extending from te Ofu landfill site and adjacent 1987 excavations eastward towards Fa'ala'aga. Potentially, this entire coastal 57 strip neady 2 km long and varying between 50 to 100 m wide, might contain subsurface archaeological deposits; this could only be determined drugh subsurface testing. Such a testing prgrmn was proposed to the Historic Preservation Office of American Sanoa in 1988 and implemented by a third field phase of the Manu'a Project in 1989. THE 1989 TRANSECT EXCAVATIONS 1989 Excavation Procedures The prmary objectives of the 1989 field season at To'aga included-mong other things-the areal definition of the fuill extent of subsurface archaeological deposits witin the coastal terrace and the furter elaboration of a morphodynamic model of site fonnation processes. These goals dictated the continued use of a systemadc tract sanpling strategy, with excavation transects spaced along the full extent of fte coastal terrace (as far to the northeast as Fa'ala'aga). In order to maintain accurate spatial control on the test units to be excavated along these ransects, our first task was to establish a baseline. This was staked out at 100 m intervals, following the course of the dirt road that paallels the shoreline trughout te To'aga area. A zero or origin point was established where this baseline bisected tX 1987 trmnsect (which was designated Transect 1); from this 0-point, the baseline extended 300 m to the southwest, and 1500 m to the norheast. Each 100-m interval along the baseline was identified as a potential transect-intersection point for subsurface testing, and tansects were labeled fron 1 to 19. Transects 2-3 extended southwest from the origin point (Transect 1), while tansects 4-19 extended to the norheasL Choice of specific trans for test excavation was dictated by several factors, including permission from local landowners, locations of gardens (we did not wish to disturb active Alocasia garden sites), presence of 'sacred' sites (in particular, the Tui Ofu tomb complex), as well as te usual constraints of time and budget. Over the course of the field season, we were able to excavate test units along six of the potential sampling tansects: Transects 3 (200 m W), 5 (100 m E), 7 (300 m E), 9 (500 m E), 11 (700 m E), and 17 (1300 m E). The locations of these transects, and their relationship to te baseline, am 58 The To'aga Site diagrammed in figure 5.8. (Note that in this figure the scales for the baseline and te trinsects are different.) These trants thus cover a total area extending 1.5 kn from southwest to nonheast, and in every case across te lateral extent of the coastal terrace. Prior to the commencement of excavation, each selected traect was cleared of obscuring vegetation, and a transe baseline staked out at 10-m intervals. A continuous elevation profile along the Utnsect was then surveyed using a telescopic level and stadia rod. These surveyed proffles provided valuable infonnation on the geomorphic structure of the coastal terrace. In order to ascenain the relationships between surface and subsurface geomorphic features and modem sea level, each tansect profile was also surveyed down to the shoreline, and out onto the reef fiat. A total of 16 -iM2 units were excavated along the six tranwcs listed above. Virtually all units extended to at least 1.5 m below surface, and some units reached depths as great as 2.9 m before culturally sterile sediments were exposed. Thus fte total excavated volume was approximately 32 m3. Excavation units were numbered sequentially in the order that they were iniidated, beginning with Unit 15 and ending with Unit 30. In the following pages, the 1989 excavations are described by trasect, beginning with Transect 3 at the southwestem end of the coastal terrace and progressing towards the norheast to Transect 17. Transect 3 Transect 3 is situated 200 m to the west of the 1987 excavation area (Transect 1); between Transect 3 and Transect 1 lies the bulldozed landfill site where pottery was first discovered during the 1986 reconnaissance survey. The surveyed elevation profile of Transect 3 is shown in figure 5.9, with the location of the single st, Unit 27, excavated along this transect. The ground surface along Transect 3 is composed of clayey colluvium with numerous volcanic cobbles and boulders scattered over te surface. The vegetation consists predominantly of cultivated bananas. Excavation Unit 27 This test excavation is situated 42 m inland of the road. The profile of the east face is shown in figure 5.10. The following strata were identified: Layer I: This dari brown colluvial clay contains angular and subangular basalt cobbles. Charcoal flecks were present thughout in low frequency. Layer I: An orange-brown-stained calcareous sand makes up this layer. The lower part ofthis stm contains many coral and basalt pebbles and cobbles. The distribution of ese larger clastics suggests a storm or other high-energy depositional event, followed by a more gradual accumulation of the finer-grained sands in the upper portion of the deposit. Layer MIIA: The brown-to-tan calcareous sand in this statum exhibits some "peppenrng" of volcanic grains. The straum also contains many angular and subangular basalt cobbles and boulders, shell, and coral rubble. Layer hIB: This layer is composed of brown sandy clay with angular to subangular basalt cobbles, rounded coral cobbles, and shell. The clay component of this sand is higher than in Layer lIA, although the boundary separating the two strata is diffuse and unclear. Layer IIIB produced the bulk of the cultural material in Unit 27, between about 175214 cm below surface. This cultual materal included more than 6 kg of shell midden, 120 fish bones, and some turtle (see Nagaoka, chapter 13). The lower part of Layer EB was sterile, and excavation was discontinued at 225 cm below surface. Unit 27 yielded a variety of prehistorc arifacts. iMese include pottery, pieces ofobsidian (probably natural), basalt (from Layers H and IIA), and an adz butt fragment from Layer IRA. Layer I did not contain any artifacts. A total of 95 sherds was recovered, with the following strafigraphic distribution: Layer 11, 5 thickware sherds; Layer IIA, 80 sherds (5 [23%] thinware, 57 [71%] thickware); Layer HIB, 10 sherds (3 [309] thinware, 7 [70%] thickware). No radiocarbon dates were processed from Unit 27. On the basis of the ceramics recovered, however, the occupation in Layer HsB could date to avationh 59 q~~~~~~~~~~~~~~~Y LL~~~~~~~~~~~~~~~~~~~~~~~~~ 0 ~~~~~0- LL~~~~~~~~i -l) I *U, S ~~~~ gioesuleji 4 \%*%%N%~~~j<(I) 1;ISUSJ1L V~~~~~~~~~~~~~~~~~~ > co ) 0) LL S ~ ~ ~ ~ ~ ~ ~ E0 0~~ EM0 0~~~~~~i The To'aga Site 60 w 0. 0 4 0 W cnD -l 0) CD~~C) C 2co CM C A -92 C~~~~~ 0 0 CD)0 0 w Q. x wCO 12 LI. LI. -3 -4 -5 Figure 5.9 t' w 0 10 20 m TRAN[SECT 3 Elevation profile along Transect 3, showing the location of Unit 27 in relation to geomorphic features. approximately 1000 B.C. Only a single occupation phase is indicated (in Layers IIIA and IIIB), followed by fte accumulation of calcareous sands, mixed with increasing quantities of termgenous, claysilt sediment in the higher parts of the stratigraphic column. The concentration oflarge clastics in the lower part of Layer II is probably the result of a high-energy storm event. In the most recent depositional phase, the locality has been capped by 60-70 cm of colluvium. Transect 5 Transect 5 lies 100 m east of the 1987 baseline which incorporates the main excavation trench (Units 1, 4-9) and Units 2 and 3. In 1989, Transect 5 was largely in second growth vegetation, with some Alocasia and bananas near the talus slope. The elevation profile along Traect 5 is shown in figure 5.11. Excavations along Transect 5 commenced with three units, designated Units 15, 16, and 17, spaced 10 m apart After these units had been completed and their statigraphic proffles correlated, it became apparent that the most inland test, Unit 15, had exposed a deeply buried, thin cultnal deposit containing a few sherds of thin, orange-slipped pottery. Seeking to sample more of this early deposit, another excavation (Unit 28) was opened 10 m farther inland, where the talus slope began to rise steeply. However, in this pit the massive upper colluvial deposit extended filly 2.4 m deep, making exposure of the underlying calcareous sandy strata that contained cultual materials exceedingly difficult. (Nonetheless, we did manage to carry the excavation down to 3.6 m, a risky matter in an unreinforced lxI m pit!) We therefore retmned to the Unit 15 excavation, and expanded it with 1I-m extensions north (Unit 29) and west (Unit 30). The excavation of these extensions was complicated by the partial collapse of the expanded pit walls at one point. Notwithstanding this setback, we carried te excavation of these extensions down to 2.5 m, thereby exposing a larger sample of the early culural deposits. Our persistence in seeking an enlarged sample of this deeply buried material was justified subsequently by the msults of 14C dating of a charcoal sample from Layer nI of Unit 28. The result of 3257-2879 cal B.P. is the oldest date for unquestionably in-situ cultural materal obtained from the To'aga site. Indeed, this sanple is penecontemporaneous with the submerged Lapita site of Mulifanua on Upolu Island. Excavation Unit 17 Unit 17 was the most seaward excavation along TrnsectS, situated 56 m inland from the baseline Excavations here in an ashy, brown (10 YR 5/3), sandy matix. This appears to be a tick, artificialy laid house floor deposit and contains well-preserved shell midden. Four post molds visible in te exploded section (fig. 5.13) are filled with Layer II material and were cut thrugh the undedying Layer III pavements. In addition, a hearth visible in the east and south profiles lies at the contact of Layers II and UNIT 27: EAST FACE 0.0 0.5 Ill. ~~II ' " 1.0 0 .' 0 0 a 0. 1.5 '): ok ) .,,O Figure 5.10 ~~~Sherds :? D 61 2.0 Stratigraphic profile of the east face of Unit 27. along the road. A detailed profile of te west face is shown in figure 5.12, and an 'exploded' profile of the entire unit in figure 5.13. This unit was of particular note because of the exposure of a sequence of gravel house floor pavements ('iliill) in Layer III, associated with several post molds and a hearth. The stratigraphy as wecorded on the west face was as follows: Layer I: This 'greasy' midden deposit of sandy clay loam contains much dispersed coral 'iliili gravel and shll middenm It has a very dark gray color (10 YR 3/1). Ihecontact with Layer is gradational over 2-3 an. Layer fl: A compact deposit of 'ili'ili gravel occurs Layer m: This thick deposit is made up of dutee successive ili'ili pavements, laid directly over each other. The paving matenals range from small waterwom coral pebbles (ca 0.5-1 cm diameter) down to very coarse coral sand (1-2 mm diamneter). There are discontinuous bands of black ash and fine charcoal trughout the deposit as well, probably deriving frm hearth rake-out events. The lowest pavement (Layer HIC) lies dircly on Layer IVA, with a very sharp, abrupt, flat contac All of the pavements were pale brown (10 YR 7/3). The individual sub-components were: IIA, faidty clean coral gravel; HIB, coral gravel mixed with sand and ashy beds; and HIC, coarse coral sand and gravel in fine beds. These ree gravel deposits presumably represent successive repavements of a house floor. Buck (1930:68) describes the traditional Samoan practice of "levelling off the upper surface of the platform within the house and covering it with small stones ('iitiI). ....he larger coral gravel was picked out on the beach and carried up in baskets. It often took some time to get a sufficient quantity. A Samoan woman pickdng over the coral gravel might remind one of a woman of a higher culture selecting a carpet." Layer IVA: This deposit consists of strong brown (7.5 YR 5/6) calcareous sand with a mixture of fine volcanic silt-clay. The admixture of silt-clay suggests surficial alluvial deposition of fine-grained sediment emanating from one of the colluvial fans inland, perhaps after a torrential rain. Some charcoal flecking is also present Layer IVB: The grayish brown (10 YR 5/2) calcareous sand of this deposit has small dispersed charcoal flecks. The deposit had been distubed by the deposition of Layer IVA, and is therefore discontinuous. It represents an old paleosol horizon on top of the beach ridge following stabilization. The contact with Layer V is iregular and diffuse. Layer V: This is a massive, structureless deposit of The To'aga Site 62 w E 6 4 C 0 -W co 2 0ccn -J0 (0) 0) CO) cD .) -j Cu co (0) C* L( aI') D _: c Lll co)I I qm 0) D LL w w 0- (0--4 O I 2 -t 20 0 & Figure 5.11 a . 40m a. TRANSECT 5 Elevation profile along Transect 5, showing the locations of Units 28, 15, 16, and 17 in relation to geomophic features. fine to medium-graired calcareous sand (white, 10 YR 812) with a minor component of volcanic lithic fragments. It is cultally sterile. The contact with Layer VI is gradational. Layer VI: This layer contains coarse-grained calcareous sand (white, 10 YR 812) with a greater frequency of volcanic lithic grins. The deposit also incorporates two distinct bands of fist-sized coral cobbles. Layer VII: Composed of fine-grained sand with a distinctly 'salt-and-pepper' mixture of calcareous and volcanic grains, this layer maiked te base of the excavation which was reached at 225 cm below surface. Unit 17 did not produce a large quantity of cultural materials. One bone of Sus scrofa from Layer I was the only identifiable vertebrate faunal specimen. Shell midden was present in Layers I-Ill, with the highest concentration in Layer III (see chapter 13 for details). Unit 17 yielded a single fishhook fragment (the point of a jabbing hook) from Layer II. No pottery was recovered from this unit, suggesting deposition of dtese strata after the abandornent of pottery production in Manu'a. Excavation Unit 16 Unit 16 was positioned 10 m farther inland along Transect 5 from Unit 17. The straigraphic profile of the west face, shown in figure 5.14, follows: Layer IA: This is the A horizon-garden soil. On the surface one finds much leaf litter with numerous anthropophilic snails in the genera Subulina, Lamellaxis, Succinea, and Pleuropoma. Layer IB: The very dark gray (10 YR 3/1) sandy loam here is a cultural midden deposit. Parent materal consists primarily of calcareous sand, with a minor component of fine clay-silt. The dar1k gray color results from the inclusion of much finely dispersed charcoal and other organic material. The layer contains much shell midden (9.4 kg) and some dispersed 'ili'ili gravel. The contact with Layer IC is gradational. Layer IC: The very daik grayish brown (10 YR 3/ 2) calcareous sand in this stratum has some fine charoal flecking. The layer appears to represent an Exavataons UNIT 17: WEST FACE 0.0 * :. I X . - ; ; f...','0,.- 0~~~~~~~~~~~ 0.5 WBWA 0 1.0 c ._ 0 a Branch Coral * VI. 1.5 Coral CoWes . ' ' * Figure 5.12 *. . VII..~~~~*.. 2.0 Stratigraphic section of the west face of Unit 17. old stable surface (paleosol A horizon) of the former beach ridge. The surface is somewhat disturbed and rewoiked by the overlying Layer IB occupation. The contact with Layer II is irnegular, and root casts penetat from Layer IC into Layer II, indicative of vegetation on the former stable surface. Layer II: This white (10 YR 8/2), massive, stucureless deposit of very fine-grained calcareous sand incorporates some fine volcanic lithic fragments. At the base of Layer II is a zone of fist- to head-sized, waterwom coral cobbles indicative of a high-energy stonn event. The contact with Layer Ill is very unclear and gradational. Layer m: This is a very pale brown (10 YR 7/3), 63 extremely non-concentrated culturl deposit. It yielded a few sherds during excavation, and one sherd was visible in situ in the west wail during profile recording (see figure 5.14). Several volcanic fire-altered oven stones were also noted in this layer. Except for its slightly darker color, this deposit is lithologically hardly distinguishable from Layer II. It probably represents the old beach slope fronting either the Layer IIB or Layer UD occupadons in Unit 15. The lower contact with Layer IV is gradational and unclear. Layer IV: The white (10 YR 8/2) fine grained calcareous sand in this deposit contains volcanic lithic grains ('salt-and-pepper' sand). Unit 16 produced considerable culual material, including 741 fish bones, as well as rat, turtle, and bird bones. Shell midden was most heavily concentrated in Layer I (9.4 kg), with lesser quantities in the deeper layers. Full details of the faunal analysis are presented in chapter 13. Unit 16 yielded a worked Tridacna shell (possibly an adz preform) and a shell scraper from Layer IB. Artifacts from Layer III include a shell bracelet fragment, a Turbo shell fishhook fragrnent, obsidian (including a red and black banded specimen), and basalt flakes. In all, forty-four sherds was recovered from Unit 16. These include twelve thickware sherds in Layers I and H, one of which (in Layer I) has a parallel-rib, paddle-impressed exterior surface. Layer III yielded twelve (37%) thinware sherds and twenty (63%) thickware sherds. Excavation Units 15129130 As indicated above, Unit 15 was excavated along with Units 16 and 17 on Trnsect 5. When a deeply buried, thin cultural deposit containing thin, red-slipped pottery was encountered, we decided to expand Unit 15 by excavating two adjacent squares, designated Units 29 and 30. The straigraphy described below, and shown in.figure 5.15, is that of the west face of Unit 15. Layer IA: The upper 10 cm of Layer I is composed of dark reddish brown (2.5 YR 3/4) colluvium. This upper zone is slightly humic, with an A soil horizon at the surface. The deposit is a poorly sorted, very rubbly mixture of volcanic sand and clay, full of angular to subangular gravel and small cobbles. This deposit is the tongue of a colluvial fan emanating 64 The To'aga Site UNIT 17 I w N E S I 0.0 0.5 co 0 0 0 1.0 c ~ ~ ~ ~ Po tm ld c 1.5 VI 2.0 V. Figure 5.13 Exploded stratigraphic section of all faces of Unit 17, showing postmolds and hearth associated with Layer II. from a small hanging valley above the Le'olo cliff. Layer IA grades into IB. Layer IB: Although almost identical to Layer IA, this layer lacks the organic, humic component. The lower 5-10 cm of IB is noticeably more clayey, with a lower frequency of larger clastics. Layer II: This 'greasy' midden deposit is a very dark gray (5 YR 3/1). The parent material is primarily calcareous sand with some silt-clay admixture. The dark color and greasy textu result from the incorporation of much finely dispersed ash, charcoal, and other organic material. The deposit contains abundant fist-sized, volcanic, fire-altered oven stones. Shell midden is also present (3 kg) but is somewhat chalky and chemically degraded. The contact with Layer IB is straight and abrupt; that with Layer lIIA-I is irregular, varying from abrupt to gradational. The high carbon content and the presence of oven stones suggest that this layer represents a cookhouse activity area. Layer IIA-1: The reddish-yellow (7.5 YR 6/6) calcareous sand of this deposit has pockets of pinkish-gray sand (7.5 YR 6/2). In addition to considerable mottling, various irregulanties ad disturbaces appantly resulted from the Layer occupation on an old, stabilized beach ridge surface represented by this layer. Root casts extend from IIIA-l into iIIA. The contact with Layer IIIA is shaip but highly irregular. This layer is interpreted as a formerly vegetated, stabilized dune surface (paleosol). Layer IIA: This sterile deposit of white (10 YR 8/ 2) calcareous sand is well-sorted with a mediumsized (2 phi) mode and has a minor admixture of volcanic lithic grains ('salt-and-pepper'). The contact with Layer IUB is very sharp but highly irregular ('wiggly'). Layer IUB: This pale brown (10 YR 6/3), very non-concentrated occupation deposit is indicated primarily by its slightly darker color and the Excavations UNIT 16: WEST FACE ...S ,.IB: 0.0 r Voicanlci Boule 0.5 IC . I * * . .U Root Mold II1.0 to 0 ._ a 0 Coral Cobbles a 1.5 (9 Vocanic Oven Stones * IV* . 2.0 6 Iv. Water Table 2.5 Figure 5.14 Stratigraphic section of the west face of Unit 16. Black squares indicate sediment sampling locations. Thinware Layer II Layer IIIA Layer IIIB Layer IIID 3 8 7 19 5% 24% 13% 38% 65 presence of some marine shell midden. The deposit is medium- to coarse-grained, and less well-sorted than Layer IIIA. The lower contact with Layer IIIC is irregular and slightly more gradational than the top contact. Lithologically, this is the same as Layer IIIA. Layer mc: The white (10 YR 8(2). calcareous sand of this stratum is essentially identical to Layer IIIA but contains a higher frequency of coarse grains (I phi). Layer IUD: In Unit 15, this deposit appeared only discontinuously, as pockets in the west face of the section. These were pale brown (10 YR 6/ 3), very similar to the Layer IIIB deposit. In the east face of Unit 15, the zone was continuous. This is the deposit that yielded thin, fine-tempered, red-slipped pottery, together with thickware, during excavation. Layer IV: The coarse- to very coarse-grained, white (10 YR 8(2) calcareous sand that is found here is poorly sorted, having a considerable mixture of volcanic lithic grains ('salt-andpepper' appearance) and some small volcanic pebbles. The deposit also contains considerable quantities of branch coral fingers (water rolled) and coral pebbles. The presence of these larger grained sediments and the lack of sorting suggest that this was a relatively high energy, exposed beach depositional environment. Excavation of Units 15(29/30 yielded an of array artifacts that include two shell fishhook tabs (Layers II and IIID), cut pearlshell (Layer II), a shell bracelet fragment (Layer IIIB), an unfinished fishhook (Layer IIIB), and numerous flakes of basalt and obsidian (Layers II, IIIB, and IIID). A total of 200 sherds was recovered from these units. The distribution of these sherds by layer was as follows: Thickware 58 95% 26 76% 48 87% 31 62% Total 61 34 55 50 The To'aga Site 66 UNIT 15: WEST FACE .Trkdacnai ,'Oven . 0 , Stones .I' A.- U/ _L~fB> ! mc. 0 .IV v -k Figure 5.15 v Q M. 0 , Stratigraphic section of the west face of Unit 15. Black squares indicate sediment sampling locations. Two thickware direct rims from Layer HIA have impressed lips, which appear to have been made with a carved paddle. Red slip is resticted to thinware (2 sherds only) from Layer II). As the frequencies above indicate, there is a generl trend toward increasing relative frequency of thickware in proporton to thinware over time in this unit. Excavation Unit 28 Unit 28 was excavated 10 m farther inland from Unit 15 in an effort to trace the deep Layer IIID occupation in Unit 15, inland under te steeply nsing colluvial fan. Clearly, the depth of colluvial overburden which would have to be removed was substantial, making this the most inland position in which excavation could be attempted witout the use of heavy machinery. The stratigraphy described below is that of the norh face. Layer I: This massive deposit of dark reddishbrown (25 YR 312) clayey colluvium is between 2.2-2.4 m thick and contains numerous, large, angular volcanic cobbles and boulders. No bedding or lenses were present, although clay content increases toward the base of the stratum. Free of cultural material, this colluvial deposit represents one or more events of mass wasting (landslides) from the adjacent volcanic cliffs. Layer HA: This discontinuous zone (pockets) of calcareous sand has been discolored (7.5 YR 6/6) by eluviation of clay from the colluvium above (Layer I). A small amount of cultural material was recovered, including pottery, shell and bone midden, and charcoal. Layer TIB: The white (10 YR 812) calcareous sand in this stratum has waterwom coral cobbles and a few coral boulders. This layer contains a minor admixture of volcanic sand grains ('salt-and-pepper' lithology) as well as cultural material. Layer IC: The pale brown (10 YR 6/3) calcareous sand of Layer HC is similar to IIB in origin, but is distinguished by a darker color explained by te addition of organic matter and some clay, resulting from human occupation. Its relatively abundant cultural material includes midden ad pottery (see below). Layer IIC of Unit 28 is comparable to Layer IlD of Units 15129/30. Layer mI: This layer is a white (10 YR 812). coarsegrained calcareous sand with abundant waterwom coral cobbles and fresh (unwom) brnch coral fingers. Layer III yielded less cultural material than IIC and clearly is culurally sterile at its lower depths (3.3-3.4 m below surface). Artifacts from Unit 28 came primarily from Layer IIC. These include worked shell, basalt flakes, obsidian flakes, an echinoid-spine abrader, and duee Cypraea shell dorsa found artificially nested together, which may be parts of an octopus lure apparatus (see Kirch, chapter 1 1). A total of 127 sherds was recovered, with the following stratigraphic distribution: Excavatio Thinware Layer IIA Layer IIB Layer IIC 7 51 30% 49% decline in thinware in proportion to thickware the overll decrease in density of ceramics. Included in the above sherd counts are two thinware rims from Layer IIB, and nine thinware and five thickware rms from UC. One of the thickware rims has an impressed lip. Eight thinware sherds, all from Layer IIC, carry a red or orange slip. Since the ceramic assemblage from Layer HB dates to 1308-930 cal B.C. (Beta-35601), Layer UC must be of equal or possibly greater antiquity. This suggests that in the Manu'a Islands, te ceramic trnsition from Early Eastem Lapita (calTying dentate-stamped decoration) to thin, plainware must have occurred relatively rapidly at the end of the second millennium B.C. The over time is notable, as is Radiocarbon Dates from Transect S Three 14C age deterninations were obtained from Transect 5 samples. Ihe oldest date (Beta35601) comes from charcoal at the Layer IVIll interface in Unit 28, associated with the early, thin, red-slipped pottery. This sanple has an age of 32572879 cal B.P. at one standard deviation. A sample of marine shell midden (Turbo setosus) from a cookhouse activity zone, Layer II in Unit 15, is associated with thick, coarse-tempered pottery. This yielded an age of 1631-1477 cal B.P. at one standard deviation (Beta-35924). The third sample (Beta35600) was obtained from the Layer Ill 'ili'ili pavement in Unit 17. This sample, which had no ceramics associated, yielded an age of 1256-1007 cal B.P. at one standard deviation. Further details of these radiocarbon dates and their implications are provided in chapter 6. Transect5: Sunmary of the Depositional Sequence The correlation diagram of the Transect 5 excavation units (fig. 5.16) also shows the relationship of these units to contemporary shoreline and sea level features. The general depositional sequence along Transect 5 can be reconstructed as follows: Stage 1 (> 3200-2800 cal B.P.): A narrow coastal Thickware 100% 1 16 70% 52 51% 67 Total 1 23 103 bench at the base of the steep talus was fomied. The active shoreline at this time would have been in the vicinity of Units 16-17, considerably inland of the modem shoreline. The 'salt-and-pepper' lithology of the beach ridge sediments indicates exposure of volcanic headlands along the coasline, providing a source of volcanic lithic grains. In addition, the presence of larger clastics (coral cobbles, branch coral fingers) indicates a fairly high energy shoreline. Stage 2 (ca. 3200-2800 cal B.P.): Humans began to 1, resulting in non-concentrated midden deposits that contain thin, fine-tempered, orange- or red-slipped pottery in Units 28 and 15129/30. The main area of occupation was probably farther inland from Unit 28, and thus is now deeply bured under talus rockfall and colluvium. The deposits exposed in Units 28 and 15/29/30 appear to represent the seaward periphery of such an occupation, down the slope of the former beach ridge toward the old shoreline. Archaeological exposure of the putative main occupation zone would require the use of heavy machinery, since as much as 5-15 m overburden of boulder talus and colluvium would probably have to be removed. Stage 3 (ca. 2800-2000 cal B.P.?): Deposition of calcareous sands onto the beach ridge continued, with significant seaward progradation of the shoreline occuning late in this phase which resulted in the deposition of the basal deposits in Units 16 and 17. During this stage, a second occupation phase resulted in the midden deposit of Layer IIIB in Units occupy the narrw bench formed during Stage 15129/30. Stage 4 (ca. 2000-1600 cal B.P. ?): A stabilized land surface formed during these four centuries over the now wider and prograded coastal terrace, marked by the paleosol horizon (Layer IIIA-I in Units 15/29/30; Layer IC in Unit 16; Layer IVB in Unit 17). Stage SA (ca. 1600-1400 cal B.P.): The stabilized coastal terrace in the vicinity of Units 16 and 15/29/ 30 was occupied during this terminal phase of ceramic manufacture and use on Ofu Island. The 68 The To'aga Site UNIT 15 UNIT 16 UNIT 17 0.0 - 0.5 1.0 1.5 e 0 0 2 2.0 *r 2.5 3.0 3.5 Figure 5.16 Stratigraphic correlations between Units 15, 16, and 17 on Transect 5. Layer II deposit in Units 15129/30 is interpreted as a cookhouse activity area, while the contemporaneous Layer lB in Unit 16 is a concentrated midden. Stage SB (ca. 1300-1000 cal B.P.): Acermic occupation on the coastal terrace in the vicinity of Unit 17 resulted in the construction of a low house mound fonned by several successive gravel ('ili'ili) pavements. Stage 6 (< 1000 cal B.P.): A tongue of clay-silt colluvium was deposited out onto the coastal ten-ace, probably due to increased up-slope forest clearance, agriculural acdvity, and subsequent erosion. At this time the coastal terrace was used for tree-cropping and shiffing cultivadon, continuing into te present era. stratigraphic sequence along Trnsect 5 and its corelation to modem sea level, as well as the key topographic features of the contemporary coastline (reef flat, active beach ridge) also provide key evidence for testing the morphodynamic model developed in chapter 4. As indicated in figure 5.16, the oldest cultural deposit at TransectS is the Layer IIID occupation in Unit 15, yielding thin, finetempered, orange-slipped pottery, between 180-200 cm below surface. In Unit 28, this occupation appeared at the base of Layer II, between 290-300 cm below surface, and was 14C dated to 3257-2879 The cal B.P. When corelated to modem sea level, this occupation zone lies at virtually the same elevation as mean high tide and about 1.8 m above the reef flat. These elevation relationships provide incontrovertible evidence for tectonic subsidence, as the in situ cultural matenials in Units 15 and 28 were cleatly deposited on a narrow terrace or beach ridge that must have been at least 1 m, and more likely 2 m, above te sea level at 3 kyr B.P. Given a +1-2 m high sea level stand at 3 kyr B.P., this means ftat fte To'aga site has undergone between 2-3 m of tectonic subsidence over fte past three thousand years, as suggested by the morphodynamic model developed in chapter 4. In other words, if we retrodict the elevation of te early beach-ridge occupation at 3 kyr B.P. to +3 m (based on our hypothesized subsidence rate of 1 m/kyr), this would put the occupation surface between 1-2 m above the 3 kyr B.P. sea level (which itself was +1-2 m above the present level). The altemative hypothesis-that there was no tectonic subsidenc-would require the deposition of the occupation deposit under water, a physical impossibility given the sedimentological evidence. In short, the Transect 5 stratigraphic profile strongly confilrs several key elements of the morphodynamic model of coastal terrace fonration developed in chapter 4. Excavation Transect 7 Transect 7 is situated 300 m east of the 1987 excavation area. The coastal terrace here is fairly narrow, only about 65 m from the crest of the active beach ridge to the base of the steep talus slope. The elevation profile along Transect 7 is shown in figure 5.17 (a vegetation transect is given in chapter 2, fig. 2.5). The area was primarily in banana gardens under coconut and breadfruit trees in 1989. Two tests were excavated along this transect, 10 m aparL Unit 18 was located at the base ofthe talus slope, while Unit 19 was situated on the flat depression in the center of the arboricultural zone. Unit 18 was excavated to 1.4 m below surface, at which depth the presence of large basalt boulders forced us to tenninate the excavation. Unit 19, on the other hand, was taken to a depth of 2.2 m, and a thin, nonconcentrated cultural deposit was exposed near the base. Excavadon Unit 18 The following strata were discemible in Unit 18: 69 Layer IA: This massive deposit of colluvial clay has numerous inclusions of subangular to angular gravel and pebbles. It is weak red to dusky red (2.5 YR 3-412) and very compact and sticky. Layer IB: This layer contains the same material as Layer IA, but with an admixture of snall quantities of calcareous sand grains. Some charcoal flecking (arge chunks) was observed in a 10 cm thick zone between Layers IA and lB. The boundary between IA and IB is gradational over this zone of charcoal flecking. The contact with underlying Layer II is diffuse and inregular, over a 2-3 cm zone. Layer II: Fie-grained calcareous sand with some clay admixture due to eluviation from Layer IB makes up Layer l. This pale brown (10 YR 6/3) layer incorpomtes massive volcanic cobbles and boudders which forced a temiination ofthe excavation. Excavation Unit 19 Unit 19 was positioned 45 m inland of the coastal road, in a relatively level area, sunrounded by a few large talus boulders. The stratigraphy of the west face is described below and is illustrated in figure 5.18: 6 w 0. 4 c w co CIO a Co CC W D E Cu 2 O2 < co IV,, 0) 9- :L-l a) I-1 4-0 c c :3 111 0 c 0 -2 -m -4 -6J w w cr. 10 0 20m TRANSECT 7 I Figure 5.17 Elevation profile along Transect 7, showing the locations of Units 18 and 19 in relation to geomorphic features. The To'aga Site 70 UNIT 19: WEST FACE . .~~IB InIY Turtio ~ VI Figure 5.18 Stratigraphic section of the west face of Unit 19. Black squares indicate sediment sampling locations. Layer IA: This dark reddish brown (5 YR 3/2), poorly sorted colluvial deposit has some subangular volcanic pebbles, but the general absence of gravel suggests deposidon of the sand to clay-sized particles by sheet wash at the margin of a colluvial fan. Layer IA grades into IB over a 5 cm zone. Layer IB: This strum contains reddish-brown (5 YR 5/3), fine clay mixed with calcareous sand. The calcareous sand component increases with depth. The contact with Layer II is sharp and only slightly irregular. Layer U: The very pale brown (10 YR 7/3), very fine-grained calcareous sand of this layer is cultur- ally stenle. This deposit appears to represent a faidy long period of beach ridge stability, probably under vegetation. (The very pale brown color is virtually identical to that of beach ridge deposits under Pannus and Hibiscus just inland of the modem beach.) The contact with Layer III is diffuse and irregular. Layer IH: This stratulm is composed of pale brown (10 YR 6/3), medium- to fine-grained calcareous sand which is well-sorted with scattered coral and volcanic pebbles and cobbles. Layer III is barely distinguishable from Layers n and IVA. The contact with Layer IVA is very irregular. Layer IIn appears to represent another phase of beach ridge stability. Layer IVA: The white (10 YR 8/2), medium- to fine-grained calcareous sand of this layer is wellsorted, lacks larger inclusions, and is cultually sterile. Layer IVB: This layer with its very pale brown (10 YR 7/3), medium- to fine-grained calcareous sand, probably represents a period of beach ridge stability. Layer V: The color of this cultural deposit ranges from dark grayish-brown (10 YR 4/2) to grayishbrown (10 YR 512). This non-concentrated midden consists of medium- to fine-grained calcareous sand stained with charcoal and ash. There are considerable quantities of dispersed charcoal flecks and charred wood fragments, some quite large (5-10 cm diameter). A Turbo shell fishhook was excavated from Layer V at 164 cm below surface. No pottery was present The contacts with Layer IVB and with VI are both quite irregular but fairly distict. Layer VI: This layer contains white (10 YR 8/2), coarse- to fine-grained calcareous sand and is cultally sterile. Excavation was terminated at 220 cm below surface. An auger was then used to core to a depth of 385 cm below surface, at which point the water table was encountered. No cultual materials were observed in the augured sediments. Unit 19 was notable also for the presence of a large pit, sectioned in the east and south faces of the square, as shown in figure 5.19. The pit, which was cut from Layer I down trugh Layers II and HI into Layer IV, has a mixed fill incorporating Layer IB sediment. The pit is straight-sided, with a flat base. There is no evidence of burning or use as an oven pit, nor was the feature filled with midden or trash. In shape and size, the pit is consistent with Iua'i masi, subterranea silos for the fernentation and Excavation 71 UNIT 19 I E I S I 0.0 0 L. 0 Figure 5.19 Stratigraphic section of the east and south faces of Unit 19, showing the large pit cut from Layer I into Layers II-IV. 0 -1.0 *E 0 -2.0 storage of breadfruit paste or masi (Buck 1930:13233; Kirch 1984:132-35; Cox 1980). This function is also suggested by the presence of large volcanic cobbles in the pit fill; such cobbles are used to cap lua'i masi pits after filling with breadfruit. The To'aga area is dotted with shallow circular depressions, ranging from 1-2 m in diameter, which our infonmants described as fonner masi pits. Unit 19 produced a modest faunal sample (see Nagaoka, chapter 13), and a Turbo shell fishhook from Layer V. No other arfifacts or pottery were present. Although charcoal and Turbo shell samples were collected from Layer V for radiocarbon dating, these have not been processed due to budgetary limitations. We are thus uncertain wheter te absence of pottery in Layer V reflects sampling enror, or whether this deposit post-dates the cessation of pottery use on Ofu (i.e., after about 1500 cal B.P.). Transect 9 Transect 9 was positioned 500 m east of the baseline origin point. Inland of the road, the coastal tenrace here was in dense banana plantations under breadfruit and coconut. No surface archaeological features were evident, other thwa a filled-in lua'i masi pit (1.5 m diameter, 30 cm deep depression) between Units 21 and 22. The elevation profile along Trnsect 9 is shown in figure 5.20. We initially laid out three units for subsurface sampling along Transect 9: Units 20, 21, and 22, situated at 10 m intervals. Unit 20 was posidoned at the base ofthe talus slope, with Units 21 and 22 on the flat between the talus and the road. Unit 20 penentated a thick series of cultual deposits between about 1-2.6 m below surface, yielding an aray of thin, fine-tempered ceramics (some with notched iims) and thickware, as well as fistooks, onaman, and odter antifacts. We therefore decided t expand this test in order to enlarge our sample of the comparatively rare cultual materals recovered from these deposits. Unit 23 was dterefore opened adjacent to Unit 20 on the north (inland) side (fig. 5.21). Thus a total of 4 m2 was excavated along Transect 9. The To'aga Site 72 6 C 0 1C - cO 2 2 0c 'IW I-0 C - 2 co -L L F-J w w Cc -4 -6 Figure 5.20 \lc u 1U 20m rI,T A XTOVrP A :ia 11AtIobrikl Elevation profile along Transect 9, showing the locations of Units 20/23, 21, and 22 in relation to geomorphic features. Excavation Units 20/23 The stratigraphy of this 1x2 m excavation was recorded along the west and north faces, as depicted in figure 5.22, and described below: Layer IA: This stratum consists of dusky red (2.5 YR 312) colluvium; the upper 10 cm is humic garden soil. The very poorly sorted, compact, sticky, very plastic sand-to-clay material has abundant subangular volcanic gravel- to cobble-sized inclusions. The contact with Layer IB is gradational. Layer IB: This dark reddish-brown (5 YR 3/2) massive colluvial deposit is poorly sorted and has abundant larger volcanic clastics. Calcareous sand grains and occasional charcoal flecks are dispersed th}oughout. The deposit also contains a low frequency of dispersed, chalky shell midden (due to chemical decomposition in the humic soil matrix) and occasional waterwom MiMi'ili coral pebbles. These cultural materials suggest some low intensity use of the area during the perod of colluvial deposition. This deposition was presumably gradual and incremental, and not a single event. The contact with Layer IC is gradational over a 2 cm zone. Layer IC: The dark reddish-brown (5 YR 3/3), compact, sticky silty-clay of this stratum lacks larger pebble-sized inclusions. The clay appears to represent a single depositional event, presumably result- ing from sheet wash (fluvial mode of deposition) following heavy rains. The contact with Layer IIA is abrupt and slightly irregular. Layer IHA: This stratum is composed of dark reddish-brown to reddish-brown (5 YR 34/4), coarse- to fine-grained, poody sorted, calcareous sand mixed with fine clay-silt and is marked by abundant charcoal flecking. This appears to represent a paleosol or fonmer stable surface of the beach ridge. The contact with Layer IIB is gradational. Layer UB: This layer contains yellowish-red (5 YR 5/6) to reddish-yellow (7.5 YR 7/6) well-sorted, medium- to coarse-grained, structureless calcareous sand mixed with reddish clay and is culturally sterile. The clay content decreases with depth, with the base of the layer consisting of nearly pure, mediumgrained sand. The contact with Layer IIIA is abrupt and irregular. Layer MA: This cultural midden deposit is very daik gray to very dark grayish-brown (10 YR 3/1-2) and has a sandy loam texture. The dark coloration of the calcareous sand matrix results from a high carbon and organic content. This layer was associated with a large earth oven feature, from which a 14C sample was obtained (see below). The oven and adjacent sediment contained heavy concentrations (more than 2 kg) of the spines of the large slatepencil sea urchin (Heterocentrotus mamnillatus) and Excavations Figure 5.21 73 Expansion of Unit 20 into Unit 23; removal of the thick colluvial deposit in progress. of fire-altered volcanic oven stones. This stratum clearly represents a cookhouse or food preparation activity area The contact with Layer IIIB is gradational. Layer IUB: In this layer, a dark brown to brown (10 YR 4/3), thick, cultural midden deposit of calcareous sandy loam incorporates numerous, large, subangular, volcanic cobbles. Near the top of the layer is a distinct lens of compact grayish-white ash, about 1 cm thick. This ash lens was discontinuous over the lx 2m excavation unit. The ash lens is capped by 1-2 cm of clean white beach sand, probably an artificially deposited house floor. Layer ilB contned a heavy concentration of shell and bone midden, ceramics, and other portable artifacts. The contact with Layer IIIC is gradational over 5-10 cm. Layer mC: Subangular volcanic pebbles and cobbles are contained in the light yellowish-brown (10 YR 6/4) calcareous sand midden of this stratum. The matrix consists of coarse- to medium-grained calcareous sand. This layer is essentially a lower facies of Layer IIIB, distinguished primarily by a lower concentration of midden and organic staining. The contact with Layer IV is gradational. Layer IV: The white (10 YR 8/2), coarse- to medium-grained, 'salt-and-pepper' sand (calcareous grains with an admixture of volcanic liftic grains) of Layer IV contains abundant large clastics consisting of branch coral fingers, and waterwom volcanic pebbles and gravel. It is culturally sterile. An earth oven was exposed in Layer IIIA of Unit 23, as noted above, and is illustrated in figure 74 The To'aga Site UNIT 20 UNIT 23 w UNIT 23 N w IA I X I 1 1 0.0 1v 1 IN 0 0 0 LBE I K' K'~~~~~~~Ah K 1.0 c en 2.0 v~~~~~~ .2.0 C3~~~~ Figure 5.22 Stratigraphic section of the west and north faces of Units 20f23. Black squares indicate sediment sampling locations. "C" indicates coral; "V" indicates volcanic. 5.23. The oven consisted of a circular pit, with a diameter in excess of 80 cm, and a depth of 38 cm from the rim to the base. The oven pit fill consisted of fire-cracked volcanic oven stones (fist-sized) with a few larger stones at the base, charcoal, and ash. Both the oven fill and the sunTounding midden contained large quantities of echinodenm spines and Layer IIB Layer IIIA Layer IIIB Layer IIIC Thinware 5 10% 8% 8 50 21% 32 28% test fragments, particularly of the slate-pencil sea urchin (Heterocentrotus mammllats), and several smaller echinoid species. Units 20/23 yielded a large sample of pottery and a number of noteworthy portable artifacts. A total of 499 sherds was recovered, with the following stratigraphic distribution: Thickware 45 90% 87 92% 191 79% 81 72% Total 50 95 241 113 Excavations Figure 5.23 75 View into Units 20/23, showing circular pit or postmold features filled with Layer IIB sediment, exposed at the contact with Layer IIIC. These frequencies confinn a decrease in thinware relative to thickware also noted from other tmnsects (see above). The sherd assemblage from Units 20/23 includes one Wtckware rim from Layer IIB, six thickware rims and one thinware rim from Layer hIA, twelve thickware and six thinware rims from Layer HIB, and five thinware rims from Layer HIC. None of the thickware rims have impressed lips, and all were recovered from Layer IIIB. Of 499 sherds, 15 are red-slipped (on thinware only) and were recovered from Layers IIIA, IIIB, and IIIC. One small thinware body sherd from Layer IIIC has incised lines on the exterior, but no pattem is discemible. Several Turbo-shell fishhooks were recovered from Layer III, including one with a distinctly bent shank similar to early Eastem Polynesian fonns (see Kirch, chapter 1 1 for further discussion). The butt of a plano-convex section adz (Samoan Type V) was excavated from Layer IIIB. Layer IIIC produced an unfinished ring of Tridacna shell which evidently broke in half during the process of manufacture. At the contact of Layers IIIB and IIIC, we also recovered an abrading stone designed for the manufacture of small Conus-shell beads. All of these finds are further described and illustrated in chapter 11. Excavation Unit 21 Unit 21 was positioned 10 m seaward of Units 20/23, at essentially the saime elevation above sea level. The profile of the south face is illustrated in 76 The To'aga Site figure 5.24, and described below: Layer IA: This dusky red to weak red (2.5 YR 34/ 2) massive colluvial deposit is pnimanly clay with some large volcanic cobbles and talus boulders and some smaller subangular volcanic inclusions. It is very sticky and plastic. Small quantities of chalky shell midden are dispersed thrughout the deposit. The contact with Layer IB is very gradational over 10-15 cm. Layer IB: In this mixture of fine clay and calcareous sand, the sand content increases with depth. Some shell midden was noted near the top of the deposit, but the layer becomes culturally sterle near the bottom. There are many discretely dispersed flecks and chunks of wood charcoal, probably deposited with the clay, and suggestive of forest clearance/burning up slope. The contact with Layer IIA is quite irregular and somewhat diffuse. Layer HA: This stratum is composed of very pale brown (10 YR 7/3) calcareous sand. There is a minor lithological component of volcanic lithic grains. Layer UB: The sandy midden deposit of Layer IIB is grayish-brown to dark grayish-brown (10 YR 4-5/ 2). Contacts with both Layers IIA and IIC are irregular (wavy) but fairly distinct. This deposit contains large quantities of Heterocentrotus mamillas spines, some fire-altered volcanic stones, shell midden, bone, and ceramic sherds. This layer appears to correlate with the thick Layer III midden deposit in Units 20/23, being a seaward 'pinching out' of this occupation zone in the direction of the former beach slope. Layer IIC: The white (10 YR 8/2), fine-grained, calcareous sand of this stratum has a minor component of volcanic lithic grains. The layer is culturally sterile and represents a beach slope depositional environment. The Layer IIB midden yielded a large faunal assemblage, including 13.1 kg of marine shell, 1,803 fish bones (including a large number of spines of the puffer fish Diodon hystrix), rat, bird, and turtle (see Nagaoka, chapter 13 for further details). Unit 21 yielded several artifacts of worked shell and a shell omnament (see Kirch, chapter 11), all from Layer IIB. Pottery from IIB includes three (15%) thinware and 17 (85%) thickware sherds. No rims, decorated sherds, or red-slipped sherds were present. UNIT 21: S FACE 0.0 I7IhIA nic Slab Turb l IA /+I -ItE 2 i/+ -0.5 1.0 to Volcanic f~~~/, :7| I._ 0 0. 0 -1.5 ImA HA . .ER. -2.0 Oven Stones 0 * A2 uc 2.5 Coral ,:~ .~~~~ K- Figure 5.24 Stratigraphic section of the south face of Unit 21. Black squares indicate sediment sampling locations. Excavation Unit 22 Unit 22 was positioned 10 m further seaward from Unit 21, and at a slightly higher ground Excavation elevation (on the inland slope of the present beach ridge). The noith face of Unit 22 is illustated in figure 5.25 and described below: Layer I: The very daik grayish-brown (10 YR 3/2), sandy loam of this sm has a 'greasy' texture. The deposit consists primarily of calcareous sand with a minor component of fine silt-clay. Dispersed trughout is chalky shell midden and some 'ili'ili gravel. The deposit becomes daiker toward the base, with increasing charcoal and ash content. The contact with Layer II is diffuse and irregular. Layer U: Considerable coral and basalt waterwom 'ili'ili gravel is dispersed throughout this brown (7.5 YR 5/4) sandy midden. There is some charcoal flecking, especially in the bottom of a shallow pit feature at the base of the deposit. The contact with Layer III is sharp but irregular. Layer m: The white (10 YR 8/2), medium-grained, calcareous beach sand of this stratum with its minor component of volcanic lithic grains is culturally sterile. 77 No ceramics were present in this unit. Radiocarbon Dates from Transect 9 Three 14C age detenminations were obtained from Transect 9, all on samples excavated from Unit 23. Charcoal from the large earth oven in Layer IIA (Beta-35602) yielded an age range of 2845-2612 cal B.P. at one standard deviation. This sample was in direct association with a ceramic assemblage composed of 8% thinware and 92% tickware. A second sample (Beta-35603) of dispersed charcoal flecks from d sratigraphically older Layer IIB cultual deposit yielded an age range of 2917-2382 Cal B.P. at one stadard deviation. A large single valve of Tridacna maxima shell (Beta-35604) from Layer IIlB was also dated, yielding an age range of 2444-2289 cal B.P. at one standard deviation. All dte samples from Unit 23 ovetlap at one standard deviation, suggesting that the deposition of Layers IIIB and IIIA occurred fairly rapidly, possibly over a span of only one or two hundred years. Furter details on all samples are provided in chapter 6. UNIT 22: N FACE rt.r~F-.-- . 1..r. nic Swnmary of the Transect 9 Depositional Sequence I 0.0 ni V 16 c Tur60 olcani V~~~~ 0.5 .SCharcoal * Concentration O) * . ~ . >.1.. . ~ % . .= . 1.0 . . b Charcoal mI Figure 5.25 1.5 Stratigraphic section of the no rth face of Unit 22. Figure 5.26 is a correlation diagram of the strata exposed in the Transect 9 excavation units. This diagram also shows the elevational relationships between these strata and the modem sea level and reef-shoreline features. Evidence from the Transect 9 excavation data suggests the following depositional sequence: Stage 1 (> 2800 Cal B.P.): A narrow coastal terrace or bench with a mixed calcareous-volcanic lithic sand lithology was fonned. At this time the active shoreline was probably in the vicinity of Units 2122. Stage 2 (ca 2800-2300 cal B.P.): The occupaion of the nanow coastal terrace formed in Stage 1 resulted in the accumulation of the thick pottery-bearing Layer III strata in Units 20/23. The Layer IIB deposit in Unit 21 represents a seaward diminution of this occupation down the former beach slope. The shoreline itself was presumably no more thm 10-20 m funher seaward of Unit 21 at this stage. Stage 3 (ca 2300-1900 cal B.P. ?): The rapid progradation of the coastal terrace was initiated by an increase in the biogenic/calcareous sediment The To'aga Site 78 0 &fi v- CNJ CN4 ci ,Higher Berm IVegetation Line 2 o2 -1 -2 Zone Ceramic Zone - Post- 1900 BP Progradation e tv ¢ Approx. Shoreline at 3000 BP -3 Figure 5.26 Sigraphic correlations between Units 20/23, 21, and 22 shown in relation to a schematic elevation Transect 9. The horizontal dimension along this transect is schematic and arbitrary, while the vertical dimension is expressed in meters above and below mean high water. profile of budget. Ihis resulted in the sterile calcareous Layer UB sand deposit in Units 20q23 (conrelated with Layer HA in Unit 21 and Layer HI in Unit 22), capping the earlier pottery-bearing occupation. During this stage, the shoreline prograded substantially, and active accumulation of calcareous sediments in the vicinity of Units 20-21-22 subsequently ceased. Stage 4 (ca 1900 cal B.P. ?): Following progradation of the shoreline, the coastal terrace in the vicinity of the excavation units was stabilized, with the surface covered by vegetation. This is indicated by the stable paleosol surface represented by Layer HA in Units 20123. Stage 5 (< 1900 cal B.P.): At this time colluvial clays and larger volcanic clasfics were deposited onto the coastal tenrace. The presence of charcoal in these clays suggests forest cleane and buming on the talus slope inland of the site. These human activities thus initiated increased rates of erosion. Stage 6 (< 1900 cal B.P.): The mixed colluvialcalcareous soil of the coastal terrace was used for subsistence gardening and for dispersed habitation during Stage 6. In sum, the depositional sequence of Transect 9 closely replicates that of the 1987 excavation locality, and that of Transect 5, as described above. The elevational data from Transect 9 also fully support the morphodynamic model outlined in chapter 4, and argued above for Transect 5. As can be seen in the correlation diagram (fig. 5.26), the Stage 2 occupation is no more than I m above the modem high water mark, and some 3.5 m below the modem beach ridge. Given a + 1-2 m higher sea level stand at 2-3 kyr B.P., the Stage 2 habitations would have been awash if they were not at a higher elevation than at present. Retrodicting a subsidence rate of about 1 m/kyr puts these habitation surfaces at 1-2 m above the then sea level, which also fits the lithological-sedimentological data (see Kirch, Manning, and Tyler, chapter 7). Thus, the incorporation into our morphodynamic model of a tectonically induced subsidence rate for Ofu Island of ca. I m/kyr is supported by the Tansect 9 stratigraphic data. Transect 11 Transect 1 1 lies 700 m east of the baseline origin-point (the 1987 excavation locality). Here the coastal tenrace is quite broad and flat, neatly 100 m wide from te present shoreline to the base of the talus slope. The elevation profile is depicted in figure 5.27. Only a single excavation, Unit 26, was located here, some 60 m inland from the baseline. The soil along Transect I I is more sandy than at Transects 9 or 5, lacking the colluvial clay component except very close to the talus slope. The vegetation consists of tree crops (coconut and Excavation 79 w -a 3 'C - ) cc co 0 m m O-1- 0 4 I 0 -2 0 I CO -3 U. w -4 0 10 20m TRANSECT 11 Figure 5.27 Elevation profile along Transect 11, showing the location of Unit 26 in relation to geomorphic features. breadfruit) with bananas, but aroids are absent, reflecting tfie sandy, edaphic conditions. The stratigraphy of the west face of Unit 26, shown in figure 5.28, follows: Layer IA: The dark grayish-brown (10 YR 3/2) sandy loam in this layer is culturally sterile. A thick root mass was present in the upper 5 cm. This zone contains abundant shells of several terestrial molluscs, including Subulina, Lamellaxis, and Pleuropoma. The contact with Layer IB is gradational over 3-4 cm. Layer IB: This very dark gray (10 YR 3/1), nonconcentrated midden deposit is lithologically identical to Layer IA. There is some charcoal flecking, and the texture is slightly 'greasier' than Layer IA due to carbon and organic material. Some fist-sized, volcanic fire-altered stones were noted. The contact with Layer IHA is iregular but distinct Layer HA: This stratum contains a white (10 YR 8/ 2), massive, stnuctureless deposit of cultually sterile calcareous, medium-grained sand. There is an irregular, diffuse contact with Layer IIB. Layer JIB: The pale brown (10 YR 6/3), faintly darker calcareous sand in this zone is apparently an old, stabilized A horizon paleosol within the beach ridge depositional sequence represented by Layer II. Layer UC: The white (10 YR 8/2), coarse-grained, culturally sterile sand ofthis layer has a 'salt-andpepper' lithology (mixture of calcareous and volcanic grains). Little cultural material was recovered durng the excavation of Unit 26. Layer lB, the major source, yielded slightly more than 2 kg of shell midden. No pottery was present in any of the layers. One broken Turbo shell fishhook was recovered at the Layer IB/ IIA contact. No radiocarbon dates were obtained from this unit. The Layer IIB deposit may possibly correlate with the phase of progradation and subsequent coastal terrace stability at ca. 1900 cal B.P. identified fiom other trnsects. This could not be verfied in the absence of radiometric dates, however. If a ceramic period occupation were ever present in the vicinity of Transect 1 1, it would have had to have been inland of Unit 26, and therefore now deeply buried beneath the steep talus rockfall. Transect 17 This was the most easterly of our 1989 excavation transects, situated 1300 m east of the baseline origin-point. Transect 17 mns just slightly west of the low gap over the island's central ridge at Fa'ala'aga. At this easterly end of the To'aga coastal terrace, the reef flat is quite narrow; thus the coastline is exposed to greater wave energy, particularly during storms and cyclones. The beach fronting Transect 17 consists of coarse sand and gravel, rather tha the finer-grined sands typical of locations farther to the west Also,.the modem beach sediments at Fa'ala'aga contain a high percentage of 80 .Xe: .:~ (9.: The To'aga Site Excavation Unit 24 UNIT 26: W FACE ......1 . .1.. . 1 1* ~ * ~ ~ . . . . 0 . ... _ ..110.-. Figure 5.28 Straigraphic section of the west fact of Unit 26. Black squares indicate sediment sampling locations. volcanic clastics, derived from the exposed volcanic headland at the eastem tip of Ofu (a control sample of this modem beach sediment is described in chapter 7). These particular geomorphological features are likewise reflected in te litology of te two units, 24 and 25, excavated along Trnsect 17. The elevation profile along the trnset is shown in figure 5.29, indicating the positions of the two excavated pits just seaward of the talus slope. Unit 24 was situated 1 14 m inland of the baseline (along the coastal road) and about 10 m from te base of tX talus slope. The talus here consists entirely of large subangular boulders, without a colluvial clay component. The ground surface at Unit 24 was sandy loam with a cover of coconut and breadfruit. The stratigraphy of the north face of Unit 24, shown in figure 5.30, follows: Layer IA: This stratum contains dark brown (10 YR 3/3) sandy loam (A honzon) developed under Hibiscus tiliaceus, banana, breadfiuit, and other vegetation. The parent material consists wholly of coarse- to medium-grained calcareous sand. The contact with Layer IB is irregular and slightly diffuse. Layer IB: Light yellowish-brown (10 YR 6/4), coarse- to medium-grained, calcareous sand with a very minor volcanic component makes up this stratum. Also, there is slight humic staining. The contact with Layer IC is gradational. Layer IC: This white (10 YR 812). very coarse- to coare-grained sand contains lenses of waterwom coral, volcanic gravel, and waterwom branch coral fingers. The sand is dominantly calcareous but with a strong secondary mode of volcanic lithic grains. TMe contact with Layer ID is fairly sharp and regular. Layer ID: Coral, gravel, and rubble with waterwom branch coral fragments, in a matrix of coarse-grained sand make up this stratum which appears to represent a single, high-energy depositional event. The contact with Layer IE is sharp and distinct. Layer IE: The coarse- to medium-grained 'salt-andpepper' sand of this layer has scattered waterwom coral and volcanic gravel- to cobble-sized clastics. Layer IF: This stratum contains volcanic, lithic, coarse-grained sand with a slightly dominant calcareous mode. The deposit incorporates large cobbles of waterwom basalt and cobble-sized coral head shingles. It represents a very high-energy beach, with constant input of volcanic source materials to the sediment budget. The depositional sequence in Unit 24 reveals a gradual transition from a very high-energy, exposed beach directly at the base of the talus slope, then to a lower-energy beach resulting from progradation, and finally to the formation of a stable coastal tenrace Excavatio, 81 W-0 W 0 0 0- cor :30 6 U- e 4 I- 6.2 -2 -Jw U-. W a: o -4 20 0 40m Elevation profile along Transect 17, showing the locations of Units 24 and 25 in relation to geomorphic features. Figure 5.29 UNIT 24: W FACE with soil development under vegetation. Three ceranic body sherds (1 thickware, 2 thinware) were recovered from Layer LB of Unit 24. One of these is particularly unusual in that it is thick, coametempered pottery yet with a very clear red-slipped exterior surface. While thickware is obviously contemporaneous with early thinware assemblages, ilufing those witi red slipping, this is the only thickware sherd from the To'aga site with a red slip. 0.0 *IA Root Mold LB 0.5 '. TRANSECT 17 0 O 2 IC o a Excavation Unit 25 4._ 9 ~E Unit 25 was positioned 10 m seaward of Unit 24. The stratigraphy of the north face, shown in figure 5.31, is: Layer IA: The very dark grayish-brown (10 YR 3/ 2) sandy loam of this stratum has some subangular, fist-sized, volcanic stone inclusions (not fire altered). The contact with Layer LB is gradational over 2-4 cm. Figure 5.30 St secto of dte west fae of Unit 24. Black squares indicate sediment sampling "V" i lbcatio. "4C" indicae coal; vocanic. Layer IB: The black (10 YR 2/1) sandy loam of this layer is a very 'greasy' textued cultural deposit of calcareous sand with a heavy carbon and organic materal component Shell midden and fire-altered volcanic stones are present. The contact with Layer IHA is fairly disdnct and slightly inregular. Layer IC: The daik grayish-brown (10 YR 4/2) to very dark grayish-brown (10 YR 312) sandy loam of this deposit appears to be identical to Layer IA. It probably represents a fairly long period of coastal 82 The To'aga Site SUMMARY AND CONCLUSIONS UNIT 25: N FACE 0.0 IA 4* TB :o.: 1.5 Base:of Excavation at2.1 ma. U 25Graveli Bl;/ squares * . of Coral and Branch Coral O * o ° / - ° * %'0'' , ~~~~1.5 ***Base of Excavation at 2. 10 m Figure 5.31 Sltlga_ secd of th ot fac of Unit 25. Black squaresidct sediment sampling loations. terrace stability under vegetation, with gradual accumulation. The contact with Layer IIA is distinct and slightily irregular. Layer HlA: This layer contains pale brown (10 YR 6/3), coarse-graied sand with a marked 'salt-andpepper' appearance due to a heavy mixture of volcanic grains. In the northeast comer of the unit there is a facies of sand matrix incorporafing waterrolled coral and volcanic gravelpebbles. The contact with Layer IEB is sharp but slightly irregular. Layer JIB: The white (10 YR 812), coarse-grained, 'salt-and-pepper' sand ofthis stratum has a lens of water-rolled coral rubble and gravel (incorporating branch coral fingers) at about 145 cm below surface. The heavy carbon content ofLayer IB is suggestive of cookhouse activity. This interpretation is also supported by the presence of fire-altenred volcanic oven stones. No pottery was present in Unit 25. A detailed discussion of the results of our duee seasons of excavations at the To'aga site is best defened to te concluding chapter of this volume (see Kirch and Hunt, chapter 15). Here we confine ourselves to a brief summary ofseveral key points: 1. Areal extent qfthe To'aga site. One of our pnmary objectives during the 1989 excavation season was to determine the full horizontal extent of the subsurface cultural deposits at To'aga. Through the use of systematic tnsect sampling, we have been able to ascertain that buiied cultural deposits dating as early as ca. 3000 cal B.P. extend contnuously from Transect 3 to Transect 9, a distance of some 700 m. Between Transects 9 and Ilthe density of cultual materials declines significantly, but some culturl materals were present in Unit 26 along Trauect 1 1. Transt 17 also yielded a very low density of culual materials. We were restricted by the local landowners from testing the 600-m gap between Transects 11 and 17 because of the presence of the sacred-and rather feard-Tui Ofu tomb and other surface cultural remains in this area. Based on our transect results, we can now state ta subsurface archaeological deposits are present thrughout the entire To'aga coastal flat as far as Fa'ala'aga, but that the main pottety-bearing, concentrated midden deposits are restricted to the area between Transects 3 and 9. The bured deposits are concentrated in a narrow band, extending no more than about 30 m seaward from the present base of the talus slope. Shovel tests in vanous localities in the present beach ridge (on both sides of the road) revealed no archaeological materials. We are also certain that cultual deposits extend inland under the steep talus for an unknown extent. The great thickness of e talus and colluvium, however, prevents testing of these deposits without the aid of heavy machinery which was not available to us. It would probably require fte removal of as much as 15 m of massive talus-olluvial overburden to expose some of these buried occupation deposits. 2. Testing o.f the morphodynnic model. The morphodynamic model of coastal terrace fonnation outlined in chapter 4 has been substantively supported by our stratigraphic and radiometric data from the various tansect excavations. In particular, we have been able to confirm that Ofu Island has been Excavations undergoing subsidence at a rate of 1-2 m/kyr during the late Holocene. Rapid progradation and fonnation of the coastal terrace occurred during the period 1900-1000 cal B.P., largely due to an increase in the marine-biogenic sediment budget. Thus, occupation deposits dating earlier than 2000 cal B.P. are confined to a very narrow, fonner bench or beach ridge situated at the base of the talus slope. 3. The O.fu cultural sequence. The cultural materials recovered from our varous transect excavations also provide the basis for outlining a cultural sequence for Ofu Island. This sequence commences at about 3000 cal B.P. with colonization of the island by makers of a ceramic complex including both thin and thickware vessels. While no classic dentate-stamped Lapita pottery has been recovered at To'aga yet, the early ceramic assemblage from AS-13-1 fits well within the Lapitoid ceramic series (see Kirch 1988). Other portable artifacts, such as Type V basalt adzes and shell annbands, are also similar to Early Eastem Lapita assemblages from Tonga and Fiji. Over the period from 3000-1900 cal B.P., the To'aga site was continuously occupied, and the ceramic assemblages reveal a gradual shift in dominance from thinware to thickware. At around 1900 cal B.P., a rapid progradation of the shoreline commenced, resulting in the construction of the present coastal terrace. Also, the deposition of terrigenous sediments onto the coastal terrace increased, creating a prime zone for horticultural and arboricultural activities. During the past two millennia, the To'aga coastal terrace has been heavily utilized for such subsistence pursuits, although habitations dispersed over the terrace continued to be occupied, evidenced by such features as the 'ili'ili house mound tested by Unit 13. In chapter 15 we discuss this sequence in greater detail, and compare it with the sequence defined for Westem Samoa. REFERENCES CITED Buck, P. H. [Te Rangi Hiroa] 1930. Samoan Material Culture. Honolulu: Bemice P. Bishop Museum Bulletin 75. Clark, J. T. 1980. Historic preservation in American Samoa: Program evaluation and archaeological 83 site inventory. Report prepared for the Amencan Samoa Government. Anthrpology Deparment, Bernice P. Bishop Museum, Honolulu. Cox, P. A. 1980. Masi and tanu 'eli: Ancient Polynesian tehnologies for the prservation and concealment of food. Pacfic Tropical Botanical Garden Bulletin 10:81-93. Emory, K. P., and Y. H. Sinoto 1965. Preliminary report on the archaeological investigations in Polynesia. Report prepared for the National Science Foundation. Anthropology Department, Bemice P. Bishop Museum, Honolulu. Green, R. C., and J. M. Davidson, eds. 1969. Archaeology in Western Samoa, Vol. 1. Bulletin of the Auckdand Insfitute and Museum. . 1974. Archaeology in Western Samoa, Vol. 11. Bulletin of the Auckland Institute and Museum. Hunt, T. L., and P. V. Kirch 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal of the Polynesian Society 97:153-83. Kikuchi, W. K. 1963. Archaeological surface ruins in American Samoa. Unpublished M.A. thesis, University of Hawaii, Honolulu. Kirch, P. V. 1984. The Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge University Press. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memoral Washington State Museum Monograph No. 5. Seattle. Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A radiocarbon sequence from the To'aga site, Ofu Island, American Samnoa. Radiocarbon 31:713. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Honolulu.: Bemice P. Bishop Museum Bulletin 238. Munsell Soil Color Charts 1988. Munsell Color, MacBeth Division of Kollmorgen Instruments Corporation. Baltimore. Redman, C. L. 1974. Archaeological Sampling Strategies. Reading, Mass.: Addison-Wesley Module in Antwpology, No. 55. 6 RADIOCARBON CHRONOLOGY OF THE TO'AGA SITE PATRICK V. KIRCH T HE AGE OF ARCHAEOLOGIcAL deposits at To'aga may be roughly estimated by typological comparison of te excavated cermnics and basalt adzes with dated assemblages of similar artifacts from Westem Samoa (Green and Davidson 1974), and from elsewhere in Westem Polynesia (e.g., Kirch 1988; Poulsen 1987). Such comparisons suggest that prehistoric occupation at To'aga spanned much of the first millennium B.C. and on into te Christian era Typological cross-dating is always suspect, however, as it depends upon the assumption that stylistically similar artifacts frm different islands were used contemporaneously, and thus ignores the possibilities of cultral lag or of independent rates ofchange. For example, there is no particular reason to assume, a prioin, that the change from Early Eastem Lapita pottery (characterized by dentate stamnped designs) to Late Eastem Lapita should have occurred at the same time trughout the Samoan archipelago. hIneed, the specific aspects of ceramic change may themselves have varied spatially over the Samoan Islands. Likewise, the cessation of ceramic manufacture and use was probably not a contemporaneous event trughout Westem Polynesia, as indicated by relatively late dates for pottery from Niuatoputapu (Kirch 1988:13942). In order to obtain a more precise chronology for the To'aga site, and an independent assessment ofthe age of the varous arfifact assemblages sampled from the To'aga deposits, fourteen samples of charcoal and shell were radiocaibn dated. This chapter psnts provenience and sample details for each of ftese dates, and discusses some of the broader implications of the To'aga site chmnology for Samoan prehistory. An initial assessment of the age of potterybearing deposits at To'aga was obained by dating a sample of Turbo shell midden obtained from the 1986 reconnaissance test pit adjacent to the landfill dump. This sample (Beta-19742) and another pottery-associated sample from Ta'u Island (Beta1974 1) were first reported by Hunt and Kirch (1987). The sanples were close in age, with a weighted average of 2605 ± 56.6 B.P., caibraed to 360-50 cal B.C. aftr corrcion for marine rervoir effects. Hunt and Kirch noted that these dates from Manu'a were "in close contemporaneity with corrected radiocarbn assays of wood charcoal associated with stylistically and technologically similar ceramic assemblages from tUpolu Island in Westem Sanoa" (1987:417). At the conclusion of the 1987 field season, an additional seven samples (six shell, one charcoal) were submitted to Beta Analytic for 14C assay. These dates were reported by Kirch, Hunt, and Tyler (1989), and provided "a statigphically consistent chrnologic sequence for 86 The To'aga Site human occupation of Ofu Island, spaning the period from ca. 3700-3300 B.P. up to the modem era" (1989:10). At the end of the 1989 excavation season, six more samples (two shell, four charcoal) were submitted to Beta Analytic for 14C age assessment These 14C ages ar reported here fortfhfirst time, along with their contexual details. The 1989 samples all proved to be irnally consistent, both saigraphically and in comparison with the 1987 14C series. In swn, a suite of fourteen radiocarbon dates are now avalable for the To'aga site, ranging in age from the second millanium B.C. up to fte end of fte first millennium A.D. All dates are stratigraphically consistent, and in aggregate provide a firm basis for establishing the chrnlogy of human occupation and of geomorphological change at To'aga. Because this series spans viually the entire "ceramnic period" of Samoan prehistory, it is also of considerable significance for the cultural history of the archipelago as a whole. RADIOCARBON DATING PROCEDURES: CORRECTIONS AND CAUBRATIONS hi selecting samples for radiocarbon dating, emphasis was placed on material which was in firm stratigraphic context (with no apparent distubance or inversion), and in obvious association with artifacts or featur. When possible, charcoal samples were obtained from hearth or oven features, although in four cases finely dispersed charcoal flecks had to be dated in order to ascertain the age of features of particular interest. The charcoal samples consisted of wood charcoal (species not determined) mixed in some cases with carbonized fragments of coconut endocarp (Cocos nucifera). With the exception of two shell sanples from the stratigrahically oldest beach deposits, the shell selected for 14C dating was culurlly-modified midden material about which there was no uncertainty as to culturl associations. All dated shell is of marine origin, consisting of shallow-water, reefdwelling gastrpods and bivalves. Whenever possible, we seleced the species Turbo setosus for dating, as a further control on consistency of results. This reef gastpod is one of the dominant components of the shell midden assemblages from the To'aga site (see Nagaoka, chapter 13). All 14C measurements on samples from Toaga were performed by Beta Analytic, Ic. Shell samples were pretrated by etching away the outer layers with dilute acid, to remove any adhering contaminants. The smples were then atacked with further acid to produce carbon dioxide, which was used as Xt carbon source for dating. The charcoal samples were first picked manually for rootlets, and then given a series of acid, alkali, and acid ings to remove carbonates and humic acids; they were then rinsed repeatedly to neutality. One sample (Beta-35600), consisting of very finely dispersed organics (chamoal and ofter organic material), was picked for rootlets and then dispersed in hot acid to eliminate carbonates. All benzene syntheses and countings proceeded nonnally (M. Tamers, pers. comms., 21 July 1988; 6, 13 March 1990). Four samples (three of charcoal and the organic sample) had only small quantities of carbon remaining after ptreatment, and thus were accorded extended counting times (four times the normal duration) in order to reduce statistical eror as much as practicable. The 13C/14C ratios were measured for all samples to establish 13C adjusted, "conventional 14C ages" following the wecommendations of Stuiver and Polach (1977). The nonadjusted "'Libby ages," conventional 13C adjusted ages, and a13C values for all samples are presented below. Calibration of conventional 14C ages for secular effect, and in the case of marine shell samples for reservoir effect as well, were made following the rece work of Stuiver, Pearson, and Braziunas (1986) for shell samples, and of Stuiver and Becker (1986) for fte tenestrial charcoal and organic samples. All calibrations and probability estimates were obtained with the use of the CALIB microcomputer program (Rev. 2.0) developed by Stuiver and Reimer (1986). As discussed by Kirch and Hunt (1988b:23), the dating of marine shell samples from southwestem Pacific archaeological sites has posed some problems of calibration due to the "reservoir effect" of older carbon present in the wodrd's oceans. Archaeologists have recognized that marine shell samples frequently give ages somewhat older than those of charcoal from identical archaeological contexts, and several attempts have been made to develop correction factors for Pacific island shell dates (Gillespie and Polach 1979; Gillespie and Swadling 1979; Jansen 1984; Atxhns 1986; Kirch 1989:139). These Radiocarbon Chronology 87 attempts suggested that southwest Pacific samples of shell from shallow reef envirornents ranged frm about 400-550 years older than the actual age. Recently, Stuiver, Pearson, and Braziunas (1986) produced a first, approximate, global model of txse reservoir effects for both an "'upper mixed layer" of the ocean (relevant to the To'aga samples) and for the deep ocean. Their model takes into account time-dependent changes in 14C activity, and allows for local geographic variations in reservoir effect (e.g., te effect ofupwelling from the deep ocean along continental shelves). The local geographic variations are accounted for by a A-R correction factor. Ideally, such a A-R factor should be independently determined for Samoa by 14C assay of modem, pre-bomb marine shell. As yet, however, this has not been possible. We therefore have used as a A-R correction factor a weighted average of empirically determined A-R values from several mid-Pacific islands, specifically Eniwetok, Hawaii, and Tahiti and Mo'orea in the Society group (Stuiver, Pearson, and Braziunas 1986, table 1). This pooled A-R value of 100 ± 24 represents a reasonable working value for tropical, central Pacific island samples of shallow-water marine organisms. We have also used this A-R value of 100 ± 24 in calibrating radiocarbon dates from a wide range of Lapita sites (Kirch and Hunt 1988a, b). It is significant that following calibration of the marine shell samples according to the above procedure, these are entirely consistent with the calibrated charcoal dates. This is shown for example by samples Beta-35603 and -35604, both from the same excavation level of Unit 28 in Transect 9, which overlap at one standard deviation. (hat the charcoal sample appears to be slightly older may reflect the use of wood from a tree that had been growing for a century or more prior to cutting and buming.) face type, along with the D13C value in paits per thousand. Calibrations for each date are given at one standard deviation; values within parenteses indicate intercepts on the calibration curve. THE RADIOCARBON CORPUS FROM TO'AGA Beta-25035. 3370 ± 70 B.P.; 3820 ± 70 B.P.; a13C = + 2.4 %/o. The fourteen 14C age determinations from To'aga are listed below, by laboratory number, with all relevant information on provenience, associations, samnple details, and calibrations. The uncorrected "Libby age" is given first, followed by the a13C correCted "conventional 14C age" in bold- Marine shell (2 specimens: Asaphis violascens and Lunella cincerea, total weight 48 g) from 1987 main excavation trench, Unit 6, Layer V, 314 cm below surface. Both specimens retained their surface coloration and were not water-rolled, thus indicating deposition in Layer V soon after death. These shells consist of naturally deposited marine Beta-19742. 1890 ± 50 B.P.; 2350 ± 50 B.P.; a3C= + 2.9 0/o0. Marine shell (Turbo setosus) from the initial test pit excavation adjacent to the Ofu landfill dump (Layer D, Level 10). The Turbo shell consisted of midden refuse in an organicaly-enriched midden deposit, with calcareous sand matrix. The sample was direcly associated with small quantities of thick, coarse-tempered pottery. Cal 28 B.C. (A.D. 45) A.D. 108 at 1 a; cal B.P. 1977 (1905) 1842 at 1 a. Beta-25033. 2190 ± 80 B.P.; 2640 ± 80 B.P.; al3C = + 2.3 %/o. Marine shell (Turbo setosus, 71 g) from 1987 main excavation trench, Unit 6, Layer IIA-1. The sample consisted of cultrally deposited shell midden in direct association with an earth oven feature, and with small quantities of Polynesian Plainware ceramics. Cal B.C. 362 (244) 145 at 1 a; cal B.P. 2311 (2193) 2094 at 1 a. Beta-25034. 2120 ± 80 B.P.; 2570 ± 80 B.P.; al3C = + 2.5 %oo. Marine shell (Turbo setosus, 70 g) from 1987 main excavation trench, Unit 6, Layer IIB. The sample consisted of culurally-deposited shell midden from the pnincipal occupation deposit in Layer II, in association with Polynesian Plainware ceramics and with Turbo-shell fishhooks. Cal B.C. 295 (161) 58 at 1 a; cal B.P. 2244 (2110)2007 at 1 a. 88 The To'aga Site cultural food refuse) in a calcareous shell (rather beach deposit also containing isolated tn-fine ware ceramic sherds. Cal B.C. 1765 (1682) 1600 at 1 a; cal B.P. 3714 (3631) 3549 at 1 a. Beta-25673. 3170 ± 80 B.P.; 3620 80 B.P.; al3C = 0/oo. + 2.2 Marine shell (Phalium sp.,45 g) from 1987 main excavation trench, Unit 1, Layer V, 290 cm below surface. The dated specimen was a single whole gastpod, not waterwom and retaining original surface coloration, thus indicating rapid deposition in the Layer V beach deposit soon after death. Layer V contained isolated thin, fine-tempered ceramic sherds. Cal B.C. 1526 (1441) 1377 at 1 ai; cal B.P. 3475 (3390) 3326 at 1 a. Beta-26463. 1460 ± 50 B.P.; 1910 + 2.5 50 B.P.; al3C = %1oo. Marine shell (Turbo setosus, 72 g) from Unit 3 of the 1987 excavation tranet; in Layer 11,40-70 cm below surface. The dated specimen was a single large Turbo shell, with the apenural margin displaying chipping due to cultual removal of the operculum (presumably to extract the edible soft pails). The specimen was stradgraphically situated at the basal contact of the cultural, aceramic midden deposit and the underlying sterile calcareous beach sand. The specimen's context thus indicates a time period after the commencement of rapid progradation of the To'aga coastal terrace. Cal A.D. 561 (620) 663 at 1 a; cal B.P. 1389 (1330) 1287 at 1 Beta-26464. 2660 ± 140 B.P.; 2620 140 B.P.; a13C --27.8 %1oo. Charcoal flecks from 1987 excavation Unit 10, Layer HB, at 70-80 cm below surface. The charcoal was in association with Polynesian Plainware in the Layer IIB occupation deposit. The field sample weighed ca. 1 g, yielding 0.2 g of carbon after pretreatment in the laboratory; the sample was accorded extended counting time. Cal B.C. 967 (801) 454 at 1 a; cal B.P. 2916 (2750) 2403 at 1 a. Beta-2646S. 1160 ± 70 B.P.; 1600 70 B.P.; a13C = + 2.0 %1oo. Marine shell (Turbo setosus, 66.4 g) from 1987 excavation Unit 13, Layer III, 35-45 cm below surface. The dated sample consisted of 1 nearly complete shell and 2 small fragmentary shells, all displaying cultually-induced fractures and chipping (possibly from ardfact manufactre). The Turbo shells were in direct association with an aceramic cultural midden near the base of a pebble-paved house platfoim, in which Unit 13 was excavated. Cal A.D. 828 (914) 1000 at 1 a; cal B.P. 1122 (1036) 950 at 1 a. Beta-35600. 1200 ± 70 B.P.; 1190 ± 70 B.P.; a13C = - 26.1 %/o. Finely dispersed charcoal and ash ("organics') from Trnsect 5, Unit 17,53 cm below surface in the southeast comer of the unit. The field sample weighed 510 g. The charcoal and ash were dispersed in the interstices of a gravel ('ili'iIi) pavement layer (see chapter 5), presumably a dwelling house floor. Cal A.D. 694 (781, 789, 805, 821, 829, 839, 862) 943 at 1 a; cal B.P. 1256 (1169, 1161,1145, 1129, 1121, 1111, 1088) 1007 at 1 a;. Beta-35601. 2950 ± 110 B.P.; 2900 ± 110 B.P.; al3C = - 27.8 %1o. Charcoal flecks from Transect 5, Unit 28, base of Layer 11,290-300 cm below surface. The field sample weighed 12 g, and produced 0.6 g of carbon after laboratory pretreatment; it was accorded extended counting time. The charcoal was from a non-concentrated midden deposit containing several sherds of a red-slipped, thin, fine-tempered ceramic ware. Cal B.C. 1308 (1188, 1184, 1127, 1126, 1107, 1105, 1083, 1059, 1054) 930 at 1 a; cal B.P. 3257 (3137, 3133,3076, 3075, 3056,3054, 3032, 3008, 3003) 2879 at 1 a. Beta-35602. 2660 ±100 B.P.; 2630 ± 100 B.P.; a13C = - 26.9 0/oo. Charcoal in an ash and sediment matrix (470 g) fm Transect 9, Unit 23. The charcoal was contained within an earth oven feaure cut from the upper pan of Layer HIA into Layer IlIB, which contained predominantly thickware pottery, with a small quantity of thinware. Cal B.C. 896 (803) 663 at 1 a; cal B.P. 2845 (2752) 2612 at 1 a. Radiocarbon Chronology 89 Beta-35603. 2660 ± 170 B.P.; 2600 ± 170 B.P.; a13C stratigraphically consistent series of dates s = - 28.4 0/oo. the period from ca. 3700 to 1000 cal B.P. In figure 6.1 these dates are plotted as calibraed age ranges at one standard deviation, in chronological order. The cultual associations of the samples are indicated on the right-hand side of the diagram. The eau¶iest two dates (Beta-25035 and -25673) are frm the Layer V beach deposit of te 1987 main trench excavation and are associated with thin, fine-tempered ceramics. Sanple Beta-35601, also associated with this early type of pottery, is from an in situ cultrl deposit. A suite of four dates (Beta-35602, -26464, -35603, and -35604) is associated with assemblages having significant frequencies of thin, fine-tempered pottery, fishhooks, and other artifacts as well as coarse-tempered thickware. Another suite of four dates (Beta-25033, -25034, -19742, and -35924) is in association with predominantly thick, coarsetempered pottery, of the sort known in Westem Samoa as "Polynesian Plainware" (Green 1974). Some thinware, however, continues to be represented in these later deposits. These ages indicate a time span of ca. 24001500 cal B.P. for the phase of dominant use of thick pottery. The final three dates (Beta-26463, -35600, and -26465) are from aceramic cultural contexts, and all are younger than 1500 cal B.P. Although we have no 14C dates from To'aga younger than about 1000 Cal B.P., it is lkely that the site continued to be occupied at least until European contact, based on historic artifacts and varous surface archaeological indications (house mounds, grindstones, masi pits, etc.). DISCUSSION Charcoal (70g) from Tnsect9, Unit 23, from the base of Layer IUB, 190-206 cm below surface. The field sample of 70 g yielded 0.15 g of carbon after labortory patment, and was accorded extended counting time. The charcoal flecks were collected from a culturl occupation deposit containing thin, fine-tempered pottery, including some rnms with notched decortion. This sample was from the same depositional context as shell sample Beta35604. Cal B.C. 968 (797) 433 at 1 a; cal B.P. 2917 (2746) 2382 at 1 a. Beta-35604. 2330 ± 80 B.P.; 2770 ± 80 B.P.; D13C = + 1.7 0/0o. Marine shell (Tridacna maxima, 455 g) from Transect 9, Unit 23, Layer IIB, 198 cm below surface. The dated specimen consisted of a single valve of T. maxina, showing no signs of water abrasion or rounding; the shell was presumably collected live for food and deposited directly in the cultual matrix. It was in direct association with ceramics and other portable artifacts, in the sane stratigraphic position as charcoal sample Beta35603. Cal B.C. 495 (385) 340 at 1 a;; cal B.P. 2444 (2334) 2289 at 1 a. Beta-35924. 1640 ± 70 B.P.; 2100 ± 70 B.P.; a13C = + 2.7 /co. Marine shell (Turbo setosus, 400 g) from Transect 5, Unit 15, Layer II, 60-70 cm below surface. The dated sample consisted of numerous cultually modified (chipped and broken) pieces of T. setosus, representing food refuse and probably also detritus from the manufacture of Turbo shell fishhooks. The shell was dispersed in a dark grey, 'greasy' (carbon-nich) deposit with much fire-cracked y a cookhouse culural deposit. The rock, deposit also contained thick, coarse-tempered ceranics. This is the youngest 14C age from To'aga in direct association with pottery. Cal A.D. 319 (410) 473 at 1 a; cal B.P. 1631 (1540) 1477 at 1 a. The To'aga Sequence The fourteen samples described above provide a The radiocarbon sequence from the To'aga site is the largest suite of dates from any single ardhaeological site in the Samoan archipelago, and provides an imporant baseline not only for establishing the cultural chronology of the Manu'a Islands, but for comparisons with sequences from Tutuila and Westem Samoa. An initial radiocarbon and stratigraphic sequence for the Samoan archipelago was established by Green and Davidson (1974) on the basis of forty-five dates from a number of sites on 'Upolu and from two sites on Savai'i. The University of Utah Samnoan Archaeological Program (Jennings and Holmer 1980) obtained another thirty-four 14C dates from dteir excavations of sites on 'Upolu and 90 The To'aga Site B-26465 -4 * B-35600 4 w o-. B-26483 -4 B-35924 __) '-4 B-19742 i 4 B-25034 i- 4 B-25033 i ;--- i B-35603 i I B-26464 i- I--4 '- B-35604 w B-35602 B-35*01 B-25673 '4|B-25035 I I 2 3 4 I 1 KYR CAL B.P. Figwre 6.1 Plot of radiocarbon age determinations from the To'aga site (cal B.P. at 1 sigma). Manono Islet. These seventy-nine dates from Westem Samoa provide a firm basis for an 'absolute' chronology. In American Samoa, the situation is less well developed. Frost (1978) obtained five radiocarbon dates from tee sites (AS-21-1, AS-21-2, AS-25-1) excavated by her. One sample, from Tulauta Village (AS-21-1), yielded a very early date of 810 ± 140 B.C. along with a significandy later date of A.D. 1320 ± 70. The other sites all dated to the last thousand years. Best et al. (1989) reported seven radiocarbon dates from the Tataga-matau adz quarry on Tutuila, indicating a sequence of use spaning the past 1000 years. Clark (in press) reviews an additional fifteen radiocarbon dates, largely from Aoa and Leone valleys on Tutuila Island. These include a date of 3389-2749 cal B.P. for initial occupation of the Aoa Valley, as well as very late dates for the use of ceramnics. The late dates associated with ceramics have caused Clark (in press) to argue that "pottery was not abandoned uniformly and wholesale trughout the archipelago." Dating Initial Hwnan Colonization On archaeological (ceramic) critera, t oldest known site in te Samoan archipelago is the submerged Early Eastem Lapita occupation at Mulifanu'a, Upolu (Green and Davidson 1974; Leach and Green 1989). A sample of marne shell Radiocabon Chronology (NZ-1958) from this site was originally reportd as having an age of 2980 ± 80 B.P. (Green and Richards 1975:317). More recently, Leach and Green have indicated that the correct convendonal 14C age for this sample should be 3251 ± 155 B.P. (1989:319). They suggest that the "most probable age" for this sample is 3116 cal B.P. (1 166 cal B.C.), after taking into accou te marine resevoir effect (Using the 100 ± 24 A-R value applied to the To'aga samples, the Mulifanua date would be calibrated to 1280-800 B.C.) This is consistent with other dates on Early Eastem Lapita assemblages from Fiji, Tongatapu, and Niuatoputapu (Kirch 1988, table 48), suggesting initial Lapita colonization of the Fiji-Westem Polynesian region by about cal 1200 B.C. (see also Kirch and Hunt 1988a, b). The earliest two 14C detemninations from the To'aga site, from the Layer V beach deposit in the 1987 main trench excavation, range between about 3700-3300 B.P. Although small numbers of thin, fine-ware sherds are present in this beach deposit, we must be cautious in interpreting these dates, because the deposidonal context is not specifically cultural. As argued in chapter 5, it is likely that the sherds derive from an in situ occupation situated on a beach ridge somewhat inland of the Layer V beach, and now buried under several meters of talus and colluvium. That the shell samples used for dating showed no signs of water-rolling or prolonged weathering is noteworthy, indicating that they were deposited in Layer V soon after death. It is conceivable, however, that the sherds became incorporated into the beach deposit somewhat later in time. Although these Layer V dates are not inconsistent with the earliest radiocarbon ages for Lapita colonization in the region, they do fall at the early end. Sample Beta-35601, from a deep test unit directly against the talus slope, is unquestionably in primary, cultural context and is associated with thin, fine-tempered ware. It dates to 3257-2870 cal B.P. (1308-930 cal B.C.). Thus, we can be certain that Ofu Island, and the To'aga site, were settled by the end of the second millennium B.C. as part of te proess of discovery and colonization of the FijiWestem Polynesian region by Lapita populations (Kirch and Green 1987). Calibrating the Morphodynamic Model A second objective of some importance in our 91 work has been to calibrate th sequence of shoreline change and progradation esponsible for the creation of the To'aga coastal ten-ace. Based on the model developed in chapter 4 (Kirch), it was suggested that the coastal tenace would not have begun to prograde significantly until sometime after about 2000 cal B.P., when te Holocene sea-level maximum began to drop to its modem level. This hypothesis is substantially confinned by the series of 14C dates from To'aga which reveals at the coastal terrace was very narrow and confined to a zone adjacent to the high cliffs until sometime early in the second millennium A.D. Sample Beta-26463, at 1389-1287 cal B.P., comes from a straigraphic context ta postdates fte onset of significant prograation of fte coastal terrace. The Samoan Ceramic Sequence Fmally, the To'aga radiocarbon suite provides a finn chronology for the changes in the ceramic sequence revealed by our excavations and analyses (see chapter 5). Although classic dentate-stamped Early Eastem Lapita pottery was not recovered from To'aga, the red- and orange-slipped thin, finetempered ware recovered from the deepest units would appear to be contemporary with the Mulifanu'a Lapita assemblage from rUpolu. The manufacture and use of thin, fine-tempered ware, some decorated with rim notching, spwanned the period from about 2800-2400 cal B.P. This was followed by a phase in which the manufacture and use of thicker, coarse-tenpered pottery became dominant between about 2400-1500 cal B.P. The use of pottery appears to have been discontinued entirely by about 1500 cal B.P., thus matching closely the chronology for 'Upolu derived by Green and Davidson (1974). Furtr details of this ceramic chronology are presented in chapter 9 by Hunt and Erkelens. REFERENCES CITED Athens, S. 1986. Archaeological investigations at the Tarague Beach site, Guam. Report prepared for the Department of the Air Force. San Francisco, Califomia. Best, S., H. Leach, and D. Witter 1989. Report on the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August 92 The To'aga Site 1988. Department of Anthropology, University of Otago, New Zealand. Clarlc, J. T. 1989. The Eastem Tutila archaeological project: 1988 final report Report prepared for the Govenmnent of American Samoa. in press. Radiocarbon dates from American Samoa. Radiocarbon. Frst, J. 1978. Archaeological investigations on Tutuila Island, American Samoa. Unpublished Ph.D. dissetation, University of Oregon, Eugene. Gillespie, R., and H. A. Polach 1979. The suitability of marine shells for radiocarbon dating of Australian prehistory. IN R. Berger and H. Suess, eds., Proceedings of the Ninth International Coniference on Radiocarbon Dating, pp. 404-421. Berkeley: University of Califomia Press. Gillespie, R., and P. Swadling 1979. Marine shells give reliable ages for middens. Search 10:9293. Green, R. C., and J. M. Davidson 1974. A radiocarbon and stratigraphic sequence for Samoa. IN R. C. Green and J. M. Davidson, eds., Archaeology in Western Samoa, Vol. 11, pp. 212-24. Auckland stitute and Museum Bulletin 7. Hunt, T. L., and P. V. Kirch 1987. Radiocarbon dates from two coastal sites in the Manu'a Group, American Samoa. Radiocarbon 29:41719. Jansen, H. S. 1984. Inshtie of Nuclear Sciences INS-R-328: Radiocarbon Datngfor Contributors. Lower Hut, New Zealand: New Zealand Institute of Nuclear Sciences. Jennings, J. D., and R. N. Holmer 1980. Archaeological Excavations in Western Samoa. Pacific Antrpological Records 32. Honolulu: Bemice P. Bishop Museum. Kirch, P. V. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Seattle. Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evolution in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., and T. L. Hunt 1988a. Radiocarbon dates from the Mussau Islands and the Lapita colonization of the southwestem Pacific. Radiocarbon 30:161-69. 1988b. The spatial and temporal boundaries of Lapita. IN P. V. Kirch and T. L. Hun, eds., Archaeology of the Lapita Cultural Complex: A Critical Review, pp. 9-32. Burke Memorial Museum Research Report No. 5. Seattle. Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A radiocarbon sequence from the To'aga site, Ofu Island, American Samoa. Radiocarbon 31:713. Leach, H., and R. C. Green 1989. New information for the Ferry Berth Site, Mulifanua, Westem Samoa. Journal of the Polynesian Society 98:319-29. Poulsen, J. 1987. Early Tongan Prehistory: The Lapita Period on Tongatapu and its Relonships. 2 vols. Terra Australis 12. Deparunent of Prehistory, Australian National University. Canberra. Stuiver, M., and B. Becker 1986. High-precision decadal calibraion of the radiocarbon time scale, A.D. 1950-2500 B.C., Radiocarbon 28(2B):863-910. Stuiver, M., G. W. Pearson, and T. Braziunas 1986. Radiocarbon age calibration of marine samples back to 9000 cal B.P. Radiocarbon 28(2B):9801021. Stuiver, M., and H. A. Polach 1977. Discussion: Reporting of 14C data. Radiocarbon 19:355-33. Stuiver, M., and P. J. Reimer 1986. A computer program for radiocarbon age calibration. Radiocarbon 28(2B): 1022-1030. 7 A GEOARCHAEOLOGICAL ANALYSIS OF SEDIMENT SAMPLES FROM THE TO'AGA SITE EXCAVATIONS PATRIcK V. KIRCH, ELIZABETH MANNING, AND JASON TYLER INTRODUCTION A MAJOR THUSr OF OUR research at the To'aga site has been the investigation of the geomorphological and sedimentary context of the site's extensive buried archaeological deposits. Such an approach is essential both for the underading of the "formation processes" by which the site was created (Schiffer 1987), and for the broader issues of landscape change in relation to human ecology on the Manu'a Islands. An understanding of the site's geomorphology also relates directly to cultural resource management concems for, as we shall argue, there are reasons to believe that te To'aga coastal terrace is now in a stage of active coastal erosion. While the deeply bunred archaeological deposits at To'aga are not immediately theatened by shoreline regression, they could become endangered should this process intensify, particularly through the effects of global warming and sea-level rise (Geophysics Study Committee 1990). In chapter 4, a model of the morphodynamics of the To'aga coastal terrace was deduced from the general pattem of Holocene sea-level regimes in the southwestem Pacific, and from certain features of Samoan geological history, especially evidence of tectonic subsidence. The stratigraphic results of our transect excavations, presented in detail in chapter 5, substantially confinned this model, while the radiometric chronology presented in chapter 6 provided a temporal calibration of the stages of coastal terrace formation. In this chapter, we focus on another line of evidence for testing the morphodynamic model: that of geoarchaeological analysis of sediment samples obtained from the stratigraphic profiles of the varous transect excavation units. Our approach here follows long-established principles of sedimentology (Blatt et al. 1972; Folk 1974; Krumbein and Sloss 1963; Twenhofel 1950, Reineck and Singh 1980) in the context of a geoarchaeological perspective (Stein and Fan-and 1985; Shackley 1975). In particular, our aim is to augment our field stratigraphic descriptions and interpretations through laboratory analysis of sediment samples, using a "life history" approach to sediment interpretation. Any sediment, for example a sand layer in one of our transect units, can be understood in tenns of four "life history" stages. Each sediment has: (1) a source; (2) a tansport history; (3) an environment of deposition; and, (4) post-depositional alterations (Stein 1985:6-7). These stages may be infenred through the application of various analytical techniques including particle-size analysis; grain lithology and mineralogy; pointcounting of grain composition; chemical and physical detennination of carbonates, organic matter, and pH; colorimetric analysis; and oher techniques. 94 The To'aga Site Our geoarchaeological investigation of sediment samples from To'aga was conducted in two phases. In 1987, a series of archaeological and control samples was obtained to check aspects of our statigraphic interpretation. ibese samples were analyzed by Tyler at the University of Washington geoarchaeology laboratory, following procedures outlined by Stein (1985). In 1989, a much more extensive set of samples (both archaeological and contrls) was obtained from our trLansect excavations. These 1989 samples were analyzed at the University of Califonia, Berkeley, Archaeological Research Facility by Mmaning, under the direction of Kirch. The analytical methods applied to the 1989 samples differed slightly from those used in 1987, largely because in the latter case we were able to define more precisely the specific research questions to be addiessed by laboratory analysis. For examnple, in 1989 we recognized that time-onsuming pipette determination of the fine-fcion (< 4 phi) was unnecessary to determine the environment of deposition for these predominately coarse-grained, calcareous sediments. lherefore, we concentrated our efforts on mechanical sieving and textual analysis of the pebble- to fine-sand-sized components, with additional determination of grain lithology (basalt versus carbonates) in the -1 to 1 phi (¢) size ranges. In other aspects, however, such as pH, organic matter, and carbonate determination, our 1987 and 1989 laboratory tecniques were identical. As indicated above, Tyler and Manming were responsible for the laboratory work upon which this chapter is based, while Kirch authored the text and is responsible for the final interpretations and data presentation. directly into heavy plastic bags and sealed for shipment back to the laboratory. As discussed by Stein (1985:7-9, fig. 1), sampling strategies can vary depending upon the questions to be asked in laboratory analysis. Since we were interested pimarily in depositional events, our samples were taken from within individual strata, avoiding boundares or contacts between layers. Where large clasdcs were present (e.g., > -2 0) which could not be adequately sampled, these were noted in stratigraphic descriptions. Subsampling in the laboratory, for various analytic tests, was caried out with the use of a Jones sample splitter. A total of 32 sediment samples (including 3 controls) was obtained and analyzed in the 1987 season. Another 70 samples (including 5 controls) were taken and analyzed from the 1989 excavations. Of the 1989 samples, those from Units 15, 19, and 23 were accorded a full analysis, while samples from other units were analyzed only for color, pH, parficle size, and lithologic composition. Controls Several control samples were obtained in the To'aga area from modem depositional envimrnments, in order to assist in the interpretation of the archaeological sediments. These included samples of active colluvial material from the slope inland of the 1987 excavations, sand from contemporary beach ridges, sand from te beach in front of Transect 7 (fig. 7.1), and a high-energy, sand-and-gravel beach at Fa'ala'aga. Details of tese control samrles are provided in tables 7.1 and 7.2. Particle-Size Analysis METHODS A primary aim of laboratory analysis was to Field Sampling interpret the environment of deposition of sediments. Sediment samples were systematically taken from all units following th completion of excavation, in conjunction with the drawing and description of stratigraphic profiles. In both 1987 and 1989, the exact positions of sampling blocks were recorded on the satigraphic profiles, and these have been indicated on the various section drawings reproduced in chapter 5. Sample blocks, which usually measured 10 x 15 x 10 cm, were cut into the cleaned section walls with a trwel. Samples were placed This requires a knowledge of the particle-size distribution, since the size ranges and degree of sorting of a sediment reflect the energy levels in that envirnment. For example, low-energy beaches will be characterized by well-sorted (stmngly unimodal) sands in the medium- to very fine-grained size range. Higher energy beaches, on the other hand, display less sorting and a higher frequency of coarse-grained to granule-pebble-sized clastics. Similatly, colluvial sediments deposited by mass wasting can be expected to be very poorly sorted, exhibiting a ful size Geoarchaeological Analysis of Sediment Samples Figure 7.1 Photomicrograph (lOX) of the 1 phi size fraction of a modem control sample of calcareous beach sand from Transect 7. grains together (Kunze 1965). Organic materials were removed by pretreatment with hydrogen peroxide, H202 (Jackson 1969). A 1:1 solution of sediment and distilled water was prepared, to which H202 was gradually added until frothing and foaming subsided. The beaker was then placed in a water bath at 80 OC, with additional H202 added range from silts and clays up through very large clastics (cobbles to boulders). Unconsolidated sands generally required no pretreatment prior to parficle size analysis. The colluvial samples, however, required pretreatment for the removal of organics, and in some cases, for iron oxides which would otherwise bind individual Table 7.1 Analytic Data for To'aga Control Sediment Samples (1989) Sample No. 89-131 89-132 89-133 89-134 89-135 95 % Gravel Location Fa'ala'aga Beach Fa'ala'aga Beach Fa'ala'aga Beach T-7, A horizon, beach ridge T-7, beach sand % Sand % Textural Class Silt pH % Calc % Basalt 2.3 43.0 97.4 1.7 97.7 56.9 2.5 97.0 0 <0.1 <0.1 1.3 S sG G S 9.01 8.68 8.95 8.28 97 90 89 100 3 10 0.2 99.8 0 S 8.64 100 0 11 0 96 The To'aga Site N _- 0 SW a en Q 'N C! N., Q en 1-- 1-.. m m C4 WI) t- W) en C4 w 04 04 04 N 04 >, 0>, - . >4 tn 4n 00 t t V-4 7-4 "a U >4 lw i :t t: en C14 'r_oo 4)0 en CtO >4 t t_ W 04 -O: 04 -o r4 No t-o lq vx IS CN 00 O g t', . v :sco E-4 t (N en 0 E W) C4 -4 er *4 >.0 I. 1: s' cI(NI r) r) - 0> 004 > r-4 - _ >0 "t >rX O; oo 1. < 0 0 0x ~c o o\ O ~o t_ 0%O-t W) C1 ~o C14"I "-4 - N-4 c t 00 t "t 'tt enON t- 00 x- t- qt cr 0 as Cu :t -. O WO 44 ) O>4 W w W tn --- odo40' e 0 - 1- W: Rr)tfL0 C4 ^k ;soo-seno '-qo*o4 0 _;~W C sfu00x N 00 s*;o-en*r in r*6 ^ t _- \0 > () _ 4 6 t - - 0-0 In a a 1014 >, Q 00 !qQ en V- m 4 N \0 o T--4 m r- ON t ONi ON m r xur ON N t) ~o CS _o-x ON oN W) xh on xn C C1 C -4 47 00 n C1 ot W) en r vx cr 00 r- u N N a m 00 (Vs e 1 - ~o e W- ~u :3f 00 i 06~ 60 06 6t:0 _ tt oo (NI C,' Cu V- o c;O en V W o,- Q xy4 6~ 6O0 60 e tn en _ot e >o _" o Q w00 V-4 *O 9 0 0 0 t 0*-->_ N-oa_t R W o m o 8 o _ _ W~~~~~~~~1 0 o& 0_ 00N 0oen ~o Wo ON _ r:6 0 0 0 t~- 0 0r 00 00 00 ro_ Os00_V- om tn C) - ~o_ O _ N 0 QN eon r ro t _ tn CZ ON 0_t ON r-4 o0o0 t \O00'iO~~~~~~~~l 00e O\O\NI 0-- 00 .o N 8 vo Wo e %v _oo - o7Nr e t\eOo% O N N Co O t O oO t oO N -_ _ X O O ON 0 cr o e xO\O x0t cr \ cO Ft ON _ ox x o'r r0 f-400en oo ON t- t r- O 00 (In en t N 06 li t--: r- W) 00 t- O oO O- 0 st0 0 c.) g0 .o U C.) u co D g ci c) i i M U 0-4 011, . .o oN 4-0 m u -- c > ._ CS '-4 -4- - C 4..# in .:) ::) 4-0 .9.4 .W = .9-= , .4 _ _= _ V-4 - m -- - > - -, .1.4 ."" t fi 4 0. _ S:S:4 .4 = o o o_ 3>3V4 & - (A 1- t v4 *s:. _ *.= 4"4 *- * _ -P .* *- .4 z §4 . rAZI 0 Geoarchaelogical Analysis of Sedime SWples Pretreated samples were mechanically shaken through a nested set of geological reens ofmesh sies -2 to 4 # (plus pan) for fifteen minues. Weights of each subsample were deteimined with a high-precision digital electic balnce. The 1987 samples were futher analyzed by the pipette metod to determine the particle size distribution of the fine frtion; this was not done for the 1989 samples. Following mechanical sieving and weighing, the percentages and cumulative percentages of all* classes were calculated, as was the sand:mud atio (mud is defined as all material > 4 *, i.e., silt plus clay). Tese stastics permit the textual classification ofthe sediments according to the system outlined by Folk (1974:27-30, table 1). This system utlizes a triangular diagram with apices of gravel, sand, and mud, reproduced here as figure 72. The dominant mode is indicated by a capital letter, supplemented by lower case letters for secondary modes. For exanple, a sandy gravel is designated sG, while a slightly gravelly muddy sand would be untl all reaction had ceased. The sanple was then cooled, rinsed witfi distilled water, and centnfuged at 1700 rpm for fifteen minutes. Superatant liquid was thn decaned, and the ctifuging procedure repeated until the soludon was dlear. The metd for removal of inn oxides was as follows: sediment which had lready been pretrat for removal of organics was placed in a 250 ml centrifuge tube to which about 200 ml of citrte buffer soludon (sodium citrate dihydrate; sodium bicarbonate NaHCO3; sodium chloride NaCI) was added. Samples were then warmed to 80 OC in a water bath within the fume hood. Four grams of sodium dithionite (Na2S204) were slowly added while stiring coantdy for one minute. Following a fifteen-minute digestion period, 10 ml of sated sodium chlonde (NaCI) was added to flocculate te sample, which was then centrifuged at 1700 rpm for fifteen minutes; the supematant liquid was decanted and discarded. The procedure was repeated until samples appeared light gey or white in color, indicating the mnoval of mostor all offe free imn oxides. Figure 7.2 (g)mS. Gravel (>2n11im) Terminology adopted for the To'aga site sediments, in terms of the relative contribution of gravel, sand, and mud size fractions, expresd asa trary plot 8(1 (Souce: G-dinerandDackombe 1983:108). Z\ MUDl)Y a a tGRAVEL | a (mG) a< 0 4. ...i A GRAVELLY \7 MUDDY \ GRAVELLY MUD sIGhTLY - v SuIIY G;RAVELLY SANDY - tf (ginS) Mtl) ((Wg)M)h Hf --Iv Irl-v xx SAND (gM) GRAVLLY MUD - SLIGIITL Y GRAVELLY MUI)DY SANI) l(igmS) IC I SLIGHTLY GRAVELLY SAND ((g) S) A IL / UMUI(M)/ SANDY MUD (sM) MUDDY SAND (mS)\ SAND (S\ 0 Mlud (<0.06 mm) I :9 97 a I 1: 1 9:1 Sand: mud ratio Sand (0.06-2mm) 98 The To'aga Site pH Point-Counting (Lithology and Micro-artifacts) The pH (acidity, neutrality, alkalinity) of samples was detennined witfi a Mettler automatic pH meter. A 20 g sample was p din a 1:1 solution with distfilled water, and standards set to pH 7 and 10. Tlree trials were made for each sample, with the repoited value being the average of these readings The pH values from the To'aga sediments are of intres prmarily for the assessmentof preservadon of organic culural materials, such as bone and shell faunal remains. pH values for the upper colluvial deposits, or of contemporary A soil horizons with humic materials, tended to range between 6.4-8. Deeper sediments, generally calcareous, had more alkalin pH values ranging frxm 8 to 9.5. This alkalinity strngly favors the preservation of bone and shell, reflected in the generally excellent condition of tX faunal assemblage. In contast, the relative neutrality of the shallower deposits probably reflects th presence of humic acids. This is evident in te "chalky" nate of shell midden in these upper During the field description of sedimentary units, we observed that the more deeply buiied and older staa tended to have sands with a mixed calcareous-volcanic lithology, designated 'salt-andpepper' sands (fig. 7.3), whereas higher and more strata. Organic Matter and Carbonates Ihe p=sence of organic matter and carbonates in sanples was determined by the 'loss-on-ignition" method (Dean 1974; Stein 1984). This is based on the principle that organic materials will begin to ignite at 200 OC and will bum completely when the temperue raches 550 OC. Calcium carbonate (CaCO3) evolves to carbon dioxide gas when heated to 800 OC and is eliminated at 850 OC. Thus te loss-on-ignition procedure involves the controlled burning of samples with precision weighing before ignitions after 500 °C, and after 1000 OC bums. All samples were processed in a Thermolyne muffle fumace. Color Sediment colors were recorded in situ during the description of profiles, using the Munsell Soil Color charts, and have been reported in chapter 5. In addition, the colors of laboratory sediment samples wecorded both dry and moist, also using the Munsell system. were rcent sands were largely or wholly calcareous. These differences were believed to reflect the process ofcoastal progradation and termce formation, which gradually removed sources of volcanic sediment from the To'aga coastal sediment budget. These sources originally would have been volcanic headlands and talus boulden which when exposed to high-energy wave action would generate volcaniclithic sand grains. In order to quantify more precisely these field observations, we undertook controlled point-counting of samnples in the laboratory, following procedures first outined by Galehouse (1971). The -1, 0, and 1 + size classes were selected for point-counting following mechanical sieving. Subsamples of these + classes were mounted on petrographic glass slides, which could then be examined under a Nikon stereozoom microscope. An average of 100 grains was counted for each sample, with grains being classified as calcareous (consisting of coral sand, foraminifera, marine shell fragments, etc.) or volcanic lithic (basalt or similar materials). These counts were then calculated as percentages. For Units 15, 19, and 23, we also used the pointcounting method to determine the frequencies of the following classes of micro-artifactual or faunal materials: charcoal, marine shell, bone, sea urchin, and terrestial gastrpods. (In order for marine shell to be counted as a micro-artifact constituent, it had to have sharp or fractured edges, as opposed to waterworn or rolled edges, the latter indicating a natural constituent of the sand matrix.) The terrestrial gastropods are of particular interest as they consist primarily of synanthrpic species; these are the subject of a separate and detailed analysis in chapter 8. RESULTS The 1987 Excavations Thirty-two sediment samples were obtained from the 1987 excavations (including fhree control Geoarchaeological Analysis of Sediment Samples Figure 7.3 99 Photomicrograph of a typical "salt-and-pepper" sand with mixed calcareous and volcanic-lithic lithology; from Layer IF of Unit 24 (1 phi size fraction; lOX). samples). Samples were taken from Units 2 and 3 along the initial transect, from Units 5, 6, and 9 in the main trench, and from Unit 10. Analytical data from all of these samples are reported in table 7.2. The sedimentary sequence in the main excavation trench is summanzed by samples 13-23 listed in table 7.2. The colluvial units (Layers I and III) are similar in their higher frequencies of silt and clay, contrasting markedly with Layers II and IV which are dominated by sand-sized particles. Similar contrasts are expressed in the percent of calcium carbonates in these strata. The 1989 Transects Transect S Units 15, 16, and 17 were selected for sediment analysis along Transect 5, with a complete set of analyses for Unit 15, and particle-size and lithological analyses only for Units 16 and 17. Complete analyfical data for these units are presented in tables 7.3 to 7.5. The sedimentary sequence in Unit 15 is graphically depicted in figure 7.4, and a more detailed graphic presentation of the grain-size distribution results is provided in figure 7.5. In figure 7.4 several overall trends are apparent. The lower sediments (Layers Ill to IV) are dominated by sands, with significant gravel components confined to the upper deposits (Layers I and II). This change reflects the shift from a littoral depositional environment to a mass wasting (colluvial) depositional environment. The pH shifts from neutral to somewhat alkaline between Layers IA and IIIA and remains alkaline to the base of the section. Organic matter is greatest in Layers IA and II. The shift from littoral to ternigenous mass wasting depositional environments is also clearly indicated in the percent of CaCO3 and in the lithological composition of basalt and calcareous grains determined by point-counting of the 1 0 size 100 The To'aga Site 00 NO00 0 S 0 Q zt C4 >- >4 >- _- t._. _. "-4 RQI tv 0'r)0 Qo OO0 sri o. - 00 N- 00 0; 00I NN0C I' qA to z >0 e 00 cir ad r- (1 - e 0 w0 od)00 04 rC 0 A ~ -C.I Q "4H 2 -1 00 -H e "a o 0oo - 0 R o 0t r w ON C C O C bON --4-b i cl m Cu 2 0 Y r-'%o ON * Nt t O-_N -4 CR -O R ON ON 0000 ON t u V" 014 0' IGIN 00 00 O CIA ON C u *CI0 4- co E0tc; Cu *s: Cn Cu c, w) - N w) r- c7 o 00000000 XN X0 i. Iw ;a -4 Cu e4 V- tn V-- 0 ONO 0* PC U qtC ) ,t"TC C) r- NO 00 ON 4 N \0 £ o CN GII r-G ooCK a en Ce 0 NN 0 VN cI O "-4 d ~0.t-ONt--00 (V Wm cf \0 0 r-c- <t), ONt00 _. C4) OO c\-C\ C\b 'w Ev c o m0--c _e 0<-4 _> -. CX w riA N-00OGN C m - m 0000 00 &.. - - -- - 0.4 00 ON0 0 ,it )RI~R ON 00 0000 000000 00 0000 0 _-4 00 00 00 C0oo Geoarchaeological Analysis of Sdet Se 101 s Table 7.5 Analytic Data for Unit 17 Sediment Samples Munsefl Color 1+ Sample 9 No. Layer Gravel 89-66 89-67 89-68 89-69 89-70 89-71 89-72 89-73 89-74 89-75 I II IIIA IIIB IIIC IVA IVB V VI VII GRAIN SIZE 49.1 80.7 83.5 83.4 63.0 1.9 1.4 2.3 11.5 0.4 % Sand % Textural Silt Class 47.8 17.4 14.6 15.1 35.4 95.9 94.9 97.5 88.4 99.5 % Basalt Dry 7.84 8.42 8.51 8.47 8.46 8.48 8.57 8.54 8.90 9.50 99 100 100 99 100 99 76 100 100 93 1 0 0 1 0 1 24 0 0 7 7.5YR 3/2 7.5YR 5/2 10YR 5/3 7.5YR 5/2 10YR 5/3 7.5YR 5/8 7.5YR 6/2 10YR 8/3 7.5YR 8/4 7.5YR 8/2 ORGANIC MATTER pH 8 sG G G G sG S S S gS S 3.1 1.8 1.9 1.5 1.6 2.1 3.6 0.2 0.04 0.05 pH % Calc 9 o i 5 10% CARBONATES 0 I I Moist 5YR 2.5/2 7.5YR 5/2 7.5YR 5/2 7.5YR 4/2 10YR 5/3 7.5YR 5/6 7.5YR 5/2 10YR 8/3 7.5YR 8/4 7.5YR 8/2 LITHOLOGY (1+) 100% _- MA-1 TRANSECT 5, UNIT 15 Figure 7.4 Summary diag of grain size, pH, organic matte, carbonates, apd lithology for sdiment samples from Unit 15, Transect 5. The To'aga Site 102 2 IA 30 50 50 0 0 -~~~~~MIII 0 80 70 70 0~~~~~m 50 ~~~mc L X t 50 0 -2 -1 0 I 2 GRAN SIZE Figure 7.5 3 44 'P' t Hstogram plots of grain size distribu dons for sedinent samples from Unit 15, Transect 5. class. PhotomicrograPhs of the 0 + size frcions for sediment samples from Unit 15 are siown in figure 7.6; tese illusae the vadations in lithology between the various cultl and natrl deposits. The detailed grain-size distnbutions shown in figue 7.5 likewise document changes in the environment of deposition. The basal sand, Layer IV, consists of rather poorly sorted sands (dominant mode = coarse sand) with a small granule-sized component, suggestive of a relatively high-enrgy beach environment In Layen HID to lIIA-1 there is a steady shift to well sorted sands dominated by particles in the medium sand range (2 0) with granule-sized particles generally lacking. This shift is presumably correlated with the progradation of the active beach seaward of the Unit 15 locus. The Layer II midden deposit displays a very poorly sorted sediment, reflecting the incorporation of cultual materials (including oven stones and other large clastics) into a medium-sized sand matrix. Layers lB and IA are also poorly sorted, with dominant modes in he pebble-to-cobble size ranges (-2 to -I O). These upper sediments are typical of young colluvium denived from mass wasting of the cliffs and volcanic slopes immediately inland of te site. In sum, the Unit 15 sedimentary sequence reveals the following geomorphological evolution. The intmdal enviromnent of deposidon (at ca. 3,000 Cal B.P., prior to human colonizadon) was a highenergy beach, indicating an active shoreline very close to Unit 15. The source of sediment was primarily the calcareous reef flat, with te minor addition of volcanic lithic grains derived from exposed headlands and/or talus rockfall exposed to wave actdon. During the period of human occupation represented by Layers HID to IIIA-l (ca. 30001600 cal B.P.), the Unit 15 locality shifted to a beach ridge depositional enviromnent, with trasport of calcareous sediment primarily by aeolian processes (e.g., saltation of grains inland). Following the Layer H midden occupation on top of a vegetated beach ridge, the Unit 15 locality began to be covered by terrigenous, poorly sorted, volcanic sediments deriving from the steep talus and colluvial slopes inland. The encraclhnent of xse volcanic sediments probably reflects increasing slope instability due to forest clearance and agricultual modifica- tions. Geoarchaeological Analysis of Sediment Samples Figure 7.6 103 Photomicrographs (lOX) of the 0 phi size fractions of sediment samples trom Unit 15, Transect 5: A, Layer IB; B, Layer II; C, Layer IIIA; D, Layer IIA-1; E, Layer IIIB; F, Layer IIIB; G, Layer IIID; and H, Layer IV. (e.s. = echinoid spine; Lb. = fishbone; s.g. = synanthropic gastropod). 104 The To'aga Site Transect 7 Sediment analyses were perfonned on samples from Units 18 and 19 along Transect 7. Tables 7.6 and 7.7 summanze the detailed analytical results. A graphical summaiy of the Unit 19 sedimentary sequence is shown in figure 7.7, and the detailed grain-size distnbutions are depicted in figure 7.8. The overall sequence minors ta described above for Unit 15, with such features as the shift from neutral to alkaline pH, the rapid decease in organic matter, and the relative contributions of basalt and Transect 9 Units 21, 22, and 23 were selected for sediment analysis along Transect 9. Tables 7.8 to 7.10 present Table 7.6 Analytic Data for Unit 18 Sediment Samples 1+ Sample % No. Layer Gravel 89-86 89-87 89-88 43.8 27.3 1.1 IA IB II GRAIN SIZE 0 50 IB 50.1 64.9 95.6 7 9 7.8 3.3 msG gmS (m)S 7.42 8.46 8.63 95 99 Munseil Color % Basalt Dry 63 5 1 5YR 3/3 IOYR 3/4 IOYR 5/4 37 CARBONATES 0 50 I 100% No Data 44 0 IA 40 Sand Moist 5YR 3/3 10YR 3/4 5YR 4/6 LITHOLOGY (1+) o so 40 I 4p TRANSECT 7, UNIT 19 Figure 7.7 100% a Calcareous IVB VI 6.1 pH % Calc I m V % Textural Silt Class ORGANIC MATTER pH 100% IATG % Sand i1~ carbonate sediment grains. Layers VI to II are all reladvely well-sorted sands dominated by a mediumsand (2 0) mode, with the exception of Layer V, in which some larger (-1 and -2 0) clastics represent the incorporation of cultural materials into the sand matrix. The Layer IA colluvium is typically poorly sorted. Summary diagram of grain size, pH, organic matter, carbonates, and lithology for sediment samples from Unit 19, Transect 7. 4 Geoarchaeological Analysis of Sediment Swpks 00 t00 00 lx IY od o t.o 0 Qr: 00 oo 00tt00 -- >00I 0 40 1- V- " - -4 "au Qrtn[o a u x00 000 i ( eir-: 2m gO ', 00 a Ck (2 -N 6 tr; t 105 A a ) 00 Ia -4 -4 --_ 000 r- Os Q4> I Q en d 4- 0 2 N It C4 I'- - t- - - CU ce ~u cr4 0 - C I 0 0 r-00 N% o t) m 00 0000 0X 0% 0000 ICu cr, Eu r-:00 vo m m tl 00 00 1 00 00 'a 00 00 co "a CO) C "A CT C F* to trxocraxo aN *a: *so co %,wI c' c 6 enCq4-4 - - " 0| >e as ON as q |t t 00 en in c r-4 -4 _s 0oe 1-m -m -t N P0% 0 - --~ 0 13 m S . > > > &o 04 _0 04 a 0 6- S< > S1 WI ON o _- N en m o x ON ON ON ON O ON oEoao b oo o o Oa o00 0 00 00 00 o-~ VI The To'aga Site 106 the detailed analytical results. A graphic swumary of the Unit 23 deposidonal sequence is shown in figure 7.9, and the detailed grain-size distributions are illuste in figure 7.1 1. otomicrographs of the 00 size fraction for sediment samples from Unit 23 are shown in figure 7.10. As seen in figure 7.9, the overall pattem of E2 .2 U .~ 30 so 0 60 I 70 IVA 0 60 l . 0- 0 ~~~~~IVB soV -2 ' -1 1 ' 2 ' 3 ' 4 ' PN' GRAIN SIZE Figure 7.8 (#) Hisgram plots of grain size distributions for sediment samples from Unit 19, Transect 7. change is much like that described above for Transects 5 and 7. Of particular note is the prsence of basalt grains in the basal sands (Layers IV and IIIC), reflecting the 'salt-and-pepper' lithology observed in the field. The detailed grn-sze histograms (figure 7.1 1) indicate that Layer IV is a coarse-to-medium-grained sand with a 'tail' extending into the larger clastic size ranges, suggestive of a higher-energy beach depositional environment In Layers MIB and IIA, the larger dastics are cultual materials (such as oven stones, marine shells, and large sea urhin spines) which were added to a coarse-to-medium-grained sand matrix. Layer HA, intepreted in the field as a paleosol horizon, is less well sorted than the underlying sands and begins to incorporate fine-grained terigenous sediments. The Layers IC to IA colluvial deposits are all very poody sorted, with ineasing quantities of terrgenous (basalt) materals. Figue 7.12 presents a graphical summary of grain-size distribution and lithology for Units 21, 22, and 23. These results clearly reflect the temporal shift from (1) high-energy beach depositional environments in the basal levels of Units 23 and 21 to (2) beach ridge depositional environments with human occupations, to (3) the encroachment of terigenous colluvial deposits in the vicinity of Unit 23. A micro-artifact constituent analysis of sediment samples from Unit 23 is presented in table 7.1 1, with frequencies for five categores of culural material indicated by grain size (-1 to 10 size classes). Charcoal was represented only in the upper colluvial deposit Layer IB, presumably reflecting humaninduced buming up-slope, associated with agricultral activities. Other cultural materials, mainly sea urchin and bone, were confined to Layers IIA to IUB. Transect 17 Sediment analyses were performed for samples from Units 24 and 25 along Transect 17. Analytical results are summarzed in tables 7.12 and 7.13. In figure 7.13, the results of grain-size distribution and lithology are presented for both Units 24 and 25. A shift from high to lower energy depositional environments is indicated by the grain-size distribution, while a decreasing contribution of volcanic sand grains is reflected in the 10 lithology. Geoarchaeological Analysis of Sediment Sampks GRAIN SIZE pH ORGANIC MATTER 5 10 CARBONATES 107 LITHOLOGY (10) 15 TRANSECT 9, UNIT 23 Figure 7.9 diapm of grain size, pH, organic matter, carbonates, and lithology for sediment sanples Unit 23, Transect 9. Summary from Table 7.9 Analytic Data for Unit 22 Sediment Samples Sample No. Layer % Gravel % Sand 89-111 89-112 89-113 I II III 15.5 31.5 0.1 79.8 67.0 99.6 This indices the effects of coastal progradation, and the gradual eliminadon of volcanic boulders and headlands as a source of sand sediment. detailed grain-size distribution for Unit 24 clearly reflects the changing depositional environments at Transect 17 (figure 7.14). Layers IF hough IC are indicative of very high-energy beaches with substantial components in the granulepebble-cobble size modes (O to -2 O). Layer ID, in The % Textural Silt Class 4.7 1.5 0.2 gS sG S pH 8.36 8.64 8.84 1¢ % % Basalt Calc 100 100 100 0 0 0 particular, probably represents a major stonn event, possibly associated with a tropical cyclone. Layer IB reflects a rapid shift to a lower energy beach, dominated by coarse-to-medium-grained sands, while the upper Layer IA reflects the development of a stable beach rdge depositional environment There was no significant encroachment of colluvium over Transect 17. It is clear that for most of its late Holocene The To'aga Site 108 Iid 0 0 *9 r5o 94 .- a I a - I a I 1 I I C'4 I I I I I I a I I I I I I I I 0 I I I U) t_ - > t .- Cfi I ln tn W t a I Ia I a - I WI WI WI hia ;, I ~oc * He | *. a I a 0 a I sn" 0 m 0%sr4n te '8It 0 C14 %o 0^ 0o00 "4 C J %O0 2 ;~aco la q i N~ E- 00 00 00 V- riC5,, 6 "-4 | cl s. I a . "4^ e4 -IL - N I I I I I I 6 "4 *a as c; U) *sA m en 00 00 00 00 0 a) *= 3U 0 I I I I a I I I I I I I a I I I I I I I a (NI I I I I a I I I I I I en I I I . I' I 000000Oa V) o)n _4V 4C o I 0+ *;. Q 0a" x x o *0>N -D""0 o I 8 i -ItV - *Eoo 'I 6< ma4 DI as ;5 f<< c; a- 4 0- rt0m o cl 4 14 6 4fiw v- %t V-4 en 6< m C; 6 m < C t V- %4 oa "04 >, 8 0-~-~ e t- m u > -N 0x 00 00 lid m u "-4 aao 000000000 00 00 8 x0 oo an "-4 0 "-4 00 0 o4 00 0 0 x4 "-4 00 a 00 * Geoarchaeological Analysis of Sediment Samples 109 I I. I I. I Figure 7.10 Photomicrographs (lOX) of the 0 phi size fractions of sediment samples from Unit 23, Transect 9: A, Layer IB; B, Layer IC; C, Layer IIA; D, Layer IIB; E, Layer IIIA; F, Layer IIIB; G, Layer IIIC; and H, Layer IV. The To'aga Site 110 I" It 0 Li Ig &W 0 0 - _B -- ~u 0 I- 2 cn 0 v Q en CZ r-0000 00 C4 0 9114 >~~~~~~~~~~~~~~~~~ o~~ ~~~~~~~~~~~~~~~~~= '-4 V--4 -4 -4 *N t - Cu en Q '.4 1-m W) Cd C'i '0e c o 0 en o _t >-, Cd 0 W) >4 t- W) 00 it en en * , 00 I- I---- C4r:400 od Q C) b-C> so C1 U) v ;._ 4) 112 ~aN - - .Ude u exi r CF oo CD ON Ch 0000000en00000 Ono \0 0-00 C00 C N IT m £ r- 0 0x 0x 0x 0o 0o od eu CX V1) Li N) *s: 00 Ul (jA t-0 *& C 00 00 i2 N co r cnCA 1-4 co Cu s N x00N N 0 0CC b0t co *a '0N U) Q lqt'-(14 IOCN00 QON \ -C1 co PC 00 00 rj, C1400 \OC1 000C 00 o00 CI40WIr) bi : ^ 0 'U 'Ut 4) NW 04) Nx 'r) \or' eNcNw 0000 00 00 CO o .< m U O.4-40- ;~ a e -4e N V--N c oo~ 00 00 00 - X0 00 00 X0 00 C0 00 '_~ 40 40 30 _ 0 ~L Geoarchaeological Analysis of Sedcm Sanples I history, Transect 17 has been an area of high-energy beach deposition, which conelates with the lack of significant cal deposits. The stable coastal ten-ace probably did not extend into this Fa'alapaga area unil the last 1-1.5 kyr B.P. I~~I SUMMARY AND CONCLUSIONS IA I0 I o- 0 40 0 IC HIA . lA m _r_ 50 0 50 50 o~~~~ rm 0 -2 '-I 'PAOIs 2 ~~~IV 3' GRAIN SIZE (4) Figure 7.11 Hisg plots of grain size distributions for sediment samples from Unit 23, Transect 19. 111 Th results of detailed sdiment analyses frm the To'aga excavations provide a set ofindependen data with which to t th morphodynamic model of coasal change developed in chapter 4. These rsts ar highly consistent between tasects and confirm that dtere has been a significant shift in source, mode, and environment of deposition at To'aga which conmlates with shoreine prg ion and formation of the coasal terrace. Pidor to, and at the time of initial human colonization of Ofu, the shoreline along the southem pail of the island was much closer to the volcanic cliffs, in the vicinity of the present surficial contact between the talus slope and coastal terrace. These basal deposits are consistently reflective of high-eergy beacrs, and the admixture of basalt with calcareous grains indicates ta volcanic headlands along with the coral reef provided sources of sediment A shift to low-energy beach ridge depositional environments by 1900 cal B.P. reflects coastal progradation, as predicted by our morphodynamic model. Subsequently, inrased deposition of young, pootly sorted, temgenous colluvial sedimets onto the calcareous coastal terrace suggests inreased human disuibance (burning and agricultural activity) on the interor volcanic slopes after about 1900 cal B.P. Further implications of ftese results are explored in grater detail in chapter 15. 112 The To'aga Site UNIT 21 UNIT 23 100 UNIT 22 100 0 LA IA IB IB IC HA I1A fIB LB RC I MLI mA GRAIN SIZE (%) mc wV INLAND 46 0I IA IA IB IB IC IIA IIA fB fiB 'IC 50 100 SEAWARD 50 0 100 I Calcareous II Calcareous m "Salt and Pepper" Sands MfA MB mc IV Figwe 7.12 "Salt and Pepper" Sands GRAIN LITHOLOGY I (%) Summary diagrams of grain size and lithology for Units 21, 22, and 23, Transect 9. GeoarchaeologicalAnalysis of Sediment Swmp1es UNIT 25 UNIT 24 GRAIN SIZE GAIN SIZE LllTHOLOGY (1) 0 50 100% 0 IA IA IB IB IC IC ID HIA is LIf) EF IIB INLAND Figure 7.13 -' * 50 LITHOLOGY (14) 100% 0 50 100% SEAWARD Summary diagrams of grain size and lithology for Units 24 and 25, Transect 17. 113 The To'aga Site 114 Q '5| t~~ E3 A siH Jill~' I 40 r ~~ O~ 70 - 40 IC so 40 50 L: IE 0 so phological Field Manual. London: Allen and Unwin. Dean, W. E., Jr. 1974. Detennination of carbonate and organic matter in calcareous sedimentary rocks by loss-on-ignition: Comparison with other methods. Journal ofSedimentary Petrology 44:242-48. Folk, R. L. 1974. Petrology ofSedimentary Rocks. Austin: Hemnphll Publishing Co. Galehouse, J. S. 1971. Point counting. IN R. Carver, ed., Procedures in Sedimentary Petrology, pp. 385-407. New York: Wiley and Sons. Geophysics Study Committee 1990. Sea Level Change. Studies in Geophysics. Washington D.C.: National Academy Press. Jadcson, M. L. 1969. Soil Chemical Analysis: An Advanced Course. 2nd ed., Department of Soil Science, University of Wisconsin, Madison. Krumbein, W. C., and L. L. Sloss 1963. Stratgraphy and Sedirnentation. San Frandsco: Freeman. Kunze, G. W. 1965. Pretreatment for mineralogical analysis. IN C. A. Black, et al., eds., Methods of SoilAnalysis, pp. 569-77. Madison: American Society of Agronomists. Reineck, H. E., and I. B. Singh 1980. Depositional Sedimentary Environments. New Yodk: Springer-Verlag. IF 0 -2 -1-I F 0 1 GRAIN SIZE Figure 7.14 3 2 4N (#) Hisgraam plots of grain size distributions for sediment samples from Unit 24, Transect 17. REFERENCES CITED Blatt, H., G. Middleton, and R Murray 1972. Origin of Sedimentary Rocks. New Jersey: Prentice Hall. Dackombe, R V., and V. Garliner,1983. Geomor- Schiffer, M. B. 1987. Formation Processes of the Archaeological Record. Albuquerque: University of New Mexico Press. Shackley, M. L. 1975. ArchaeologicalSediments. New York: Wlley and Sons. Stein, J. K. 1984. Organic matter and carbonates in archaeological sites. Journal of Field Archaeology 11:239-46. 1985. Interpreting sediments in cultural settings. IN J. K. Stein and W. R. Farrand, eds., Archaeological Sediments in Context, pp. 5-20. Orono: Center for the Study of Early Man, University of Maine. Stein, J. K., and W. R. Farrand 1985. Archaeological Sediments in Context. Orono: Center for the Study of Eai¶y Man, University of Maine. Twenhofel, W. H. 1950. Principles of Sedimentation. New York: McGraw-Hil. 8 NON-MARINE MOLLUSCS FROM THE TO'AGA SITE SEDIMENTS AND THEIR IMPLICATIONS FOR ENVIRONMENTAL CHANGE PATRIcK V. KIRCH INTRODUCTION NON-MARINE MOLLUScs, or "land snails," can occurin high frequencies at archaeological sites where deposidonal environments are favorable to their preservation (e.g., calcareous sands such as at the To'aga site). In general, these snails are diminutve (often in fte 1-3 mm size range), and their presence in archaeological deposits results from natual ratherthan cultual deposition. Since land snails are highly sensidve to microenvironmental conditions such as vegetation cover, changing fruencies of snail shells in sratdfied archaeological deposits can provide a proxy measure ofenvironmental change. Snail analysis for the purposes of envionmental reconstruction has been carided out in English and European archaeology for many years (Evans 1972), but has been relatively neglected in the Pacific. Beginng in the 1970s, Kirch and Christensen began to apply snail analysis in Oceanic arhaeology with fruitful results (Kirch 1975, 1989; Christensen and Kirch 1981, 1986; Christensen 1983). bse studies have demonstrated that nonmarine molluscs in Pacific archaeological sites can provide significant infoimation with regard to (1) local environmental change (especially in vegetaton), and (2) the intrducdon of exotic biota by preistoric humans. Local enviromnental changes are demostrated trugh changing assemblages of endemic and indigenous snail species which are sensitive to microenvironmental conditions. The introduction of exodc biota, on the other hand, is signalled by the presence of one or more species of synanthropic (or anthropophilic) snails that are closely associated with human habitations, gardens, and disuibed environments. Pacific malacologists recognized some time ago that a number of widely disseminated snail species in te Pacific had almost cenainly been dispersed trugh prehistoric human agency. C. Montague Cooke, who spent a lifetime reearching Pacific snail faunas, wrote that '"ere is no doubt that about a dozen species were carned by Polynesians in their migrations. At least four species were carried by the latter to the Hawaiian Islands. These for the most part are minute species of snails that are always found in situations just above high-water mark and are fairly uniformly distributed wherever Polynesians live" (Cooke 1926:2279). Among these synanthropic snail species are Gastrocopta pediculus, La,nellidea pusilla, Liardetia samoensas, Assiminea niida, and Lmnellaxis gracilis. Because tese widely dispersed species are highly characteristic of atolls, they have sometimes been described as an "atoll fauna" (Reigle 1964; Harry 1966), although they are just as common in lowland, antrpogenic habitats on high islands. During fte 1987 excavations at To'aga, I observed te presence of non-marine mollusc shells in a number of sntaigraphic contexts, and therefore undertook systematic sampling of fte 1987 main trench and several oter test units for snail shells. 116 The To'aga Site This chapter presents the results of analysis of these samples and discusses the implications of our results for paleoenvironerntal change and for the inrduction of exotic biota at To'aga. MATERIAL AND METHODS Sediment samnples for non-marine mollusc analysis were bained from two columns at site AS13-1, one from the main trench of the 1987 excavation and the other from Unit 3. The main trch column was taken from the cleaned west face of Unit 9, and fte nine column samples spanned the stratigraphic sequence from Layer IIC thrugh Layer IIA1. No land snail shells were observed below Layer IIC or in the acidic colluvial sediments of Layer I. The position of the land snail column in the main tench is depicted on the main stratigraphic section shown in figure 5.5. For the most panl, the methods used in the analysis of non-manne molluscs at To'aga follow those developed and reported by Christensen and Kirch (1986:55-56) for use in Hawaiian archaeological sites. Sediment samples for snail extraction were taken as continuous columns. Prior to sieving, the sediment samples were air-dried, te weighed and measured volumetrically in order to assure quandtative comparability between samples. In the laboratory, the sediment samples were wet-sieved through 4, 2, and 1-mm mesh nested geological screens, and the washed residues air-dried. Snail shells were then extacted by hand-sorting under a lOX binocular microscope. Snail shells were identified by referrng to published taxonomic monographs and to reference material in the collections of the Benice P. Bishop Museum (Honolulu). In the next section the various species of nonmarine molluscs present in the To'aga samples are listed in taxonomic order, with remarks on their distribution and ecology. This is followed by the presentation of quantitative data on snail species frequency from the two stratigraphic columns. Family Assimineidae Assuimnea cf. nitida (Pease) The genus Assuimnea includes a number of estuarine and standline-dwelling species (Abbou 1958). The most abundant land-snail species in the To'aga site sediments is an Assuimnea (fig. 8.1) which very closely resembles A. nida (Pease). This species is known to be widely distibuted from Soufteast Asia eastwards into Oceania. Family Achatinellidae Lamnellidea pusilla (Gould) Lamellidea pusilla is (fig. 8.1) one of several species in the Achadnellidae that have extensive distributions throughout the inner Pacific. Cooke and Kondo remarked that "'there is little doubt that the wide distribution of L. pusilla is due to human agency. It was probably tnsported from island to island during the Polynesian migrations" (1960:188). The species is distributed from the Maranas and Palau groups in the westem Pacific and as far eastward as Mangareva (Cooke and Kondo 1960: fig. 81). Family Pupillidae Gastrocopta pediculus (Shuttleworh) This small pupilHid (fig. 8. 1) is also abu tin the To'aga samples. Pilsbry made the following remarks on the distribution and geographic origin of Gastrocopta pediculus: Adaptability to life around habitations has, no doubt, led to the vast Polynesian distribution of G. pedculus. I infer that it has been carried from island to island, sticking to native impedimenta, cocoanuts, or other food mateials, in the thousand years or more of inter-island canoe voyages of the Polynesians. Its original habitat may have been somewhere between the Philippines and New Caledonia (1916-18:140-41). Generally distributed in coastal environments, G. pediculus shows a strong liking for coconut groves, often being found in large numbers "under leaves and sprouting cocoanuts" (Pilsbry 1916-18:148). SYSTEMATIC REVIEW Family Succineidae Family Helicinidae Pkuropoma sp. A few specimens of Pleuropoma are present in e To'aga samples. It has not been possible to identify these to species level. Succinea sp. Only a single example of this species was encountered in the To'aga site sediments, in Unit 3, even though living Succinea sp. were observed in leaf liUtter samples in the arboricultural zone. Non-Marine Molluscs : :9!Ek v 500rt7% ° igure 8.1 ? 3 mm Photomicrographs of terrestrial gastropods recovered from the To'aga site sediments: a, Sinployea cf. allecta; b, Gastrocopta pediculus, c, Lamellidea pusilla; d, Assirrinea cf. nitida; e, Lamellaxis gracilis. 117 118 The To'aga Site Family Caropidae Sinployea cf. allecta (Cox) A few specimens of a caropid snail of the genus Sinployea are present in several samples (fig. 8.1). This is most likely S. allecta, although the closely related species S. clausa (Solem) is also present in the Manu'a Group. Solem (1982:127) has described a subspecies, Sinployea allecta tauensis, from Ta'u Island which closely matches the remains found at the To'aga site. These species are apparently all lowland dwellers, primarily in leaf litter. Family Helicarionidae Liardetia samoensis (Mousson) This species occurs rarely in the To'aga samples. Liardetia samoensis is the most widely distributed member of the genus, ranging from te Bismarcks eastward to the Marquesas (Solem 1959:96). Baker, who monographed the zonitid snails of the Pacific, observed that the genus Liardetia "contains species that have been widely disseminated, probably trugh the agency of man" (1938:12), among them Liardetia samoensis. Solem (1959) regards the probable center of distribution of Liardetia as Indonesia. Family Subulinidae Lamellaxis gracilis (Hutton) This distinctive species (fig. 8.1) is reasonably well represented in te To'aga samples from both the 1987 main trench and from Unit 3. Pilsbry (1906) remaked L. gracilis is "probably the most widely distributed land snail in the world," and suggested that its dispersal throughout the inner Pacific may have been due to the Polynesians. Other malacologists (Cooke 1934; Solem 1978) regarded its wide distribution as a result of modem commerce. Christensen and Kirch (1981) demonstrated, however, that Lamellaxis gracilis was present on the Polynesian outlier of Tikopia by 900 B.C., where Lapita colonists presumably introduced it inadvertently. Subsequently, specimens of L. gracilis have been identified from prehistorc contexts in Tonga (Niuatoputapu, in association with Lapita sites, Kirch 1988:233-34), the Marquesas (Rolett 1989, table 5.14), and Hawai'i (Christensen and Kirch 1986:60). The dissemination of this synanthrmpic species probably began with the Lapita expansion and continued with the Polynesian dispersal into the far eastem Pacific. RESULTS The frequencies of non-marine mollusc species in the various sample units from the 1987 main excavation, and from Unit 3, are presented in tables 8.1 and 8.2. In addition to the raw frequency counts per sample, the tables provide infornation on sample weight and volume, and on te density of snail shells per cubic liter of sediment. The nine sediment samples from the main trench (Unit 9) spanned te Layer II depositional sequence as follows: Samples 1-3, Layer IIC; Samples 4-6, Layer UB, the main occupation horizon; Samples 78, Layer IIA; and Sample 9, Layer IIA-l. Several significant pattems are evident from te data presented in table 8.1. First, the assemblage is dominated by a suite of synanthropic species, particularly Assiminea cf. nitida, Gastrocopta pediculus, Lamellidea pusilla, and Lamellaxis gracilis, but with Liardetia samoensis also represented. Indeed, the dominance of synanthropic species in xse samples is remarkable, greater than any oter archaeological example that I am aware of from the Pacific. The two indigenous/endemic species present, Pleuropoma sp. and Sinployea sp., are represented only in very limited numbers. The temporal distribution of species is also striking. The oldest sample contains only one species, Assiminea cf. nitida. This species is joined by Lamellidea pusilla and Gastrocopta pediculus in Sample 2, and by Lamellaxis gracilis and Liardetia samoensis in Sample 3. Thus, by the time of occupation represented in Layer IIB, ca. 2500 cal B.P., five species of synanthropic, human-transported land snails had been introduced to, and become established on, the coastal terrace of Ofu Island. Because Lamellads gracilis is particularly associated with human gardening sites, its appearance by 2500 cal B.P. is important in tenns of the economic prehistory of the To'aga site. As can be seen from table 8. 1, the density of non-marine molluscs is extremely low (8 snaiLs/I3) in the lower part of Layer IIC, rising rapidly to a peak density of 400 snails/I3 in the middle of Layer IIB, the occupation horizon. Shell density then drops Non-Mari'ne Molluscs 119 Table 8.1 Non-marine Molluscs from the 1987 Main Trench, To'aga Site Sample No. Taxon 1 2 3 4 5 6 7 8 9 Pleuropoma sp. Assiminea sp. Lamnellidea pusilla 3 5 4 9 1 17 26 26 1 19 12 37 51 10 40 1 59 5 17 1 46 4 18 2 47 1 8 3 84 14 19 6 75 269 9 110 400 8 90 321 4 18 65 1 5 76 276 73 203 3 64 Gastrocopta pediculus Sinployea sp. Liardetia samoensis Lamellaxis gracilis Total snails counted 3 8 Snails/13* * 197 11 128 366 standardized density, adjusted for sample volumes. somewhat in Layer IHA, but rises again to 366 mails/13 in the Layer IIA-I deposit. This distribution is sigficant, since it correlates well with the sedimentological history of the column and is precisely the sort of overall density distribution that would be expected with the establishment of human habitations in Layer IIB and of a stable vegetated soil surface represented by Layer IIA-1. The land snail data for Unit 3 are presented in table 8.2. The overall pattem resembles that in the Table 8.2 Non-marine Molluscs from Unit 3, To'aga Site Sample No. Taxon Pleuropoma sp. Assiminea sp. Lamellidea pusilla 2 10 2 Gastrocopta pediculus 6 Succinea sp. Sinployea sp. Liardetia samoensis 3 Lamellaxis gracilis Total snails counted 21 71 Snails/13* * 3 4 1 27 6 9 44 3 7 1 2 8 52 184 10 66 280 5 1 110 18 27 1 4 1 13 175 735 standardized density, adjusted for suample volumes. 6 15 3 18 68 120 The To'aga Site main trench, except that the suite of synanthropic snails is already established at the base of the column (Sample 2). The increasing density of land-snail shells and the dominance of synanthropic species suggest the establishment of a stable, anthropogenic vegetation over the To'aga coastal terrace following human occupation of the island. CONCLUSION The analysis of non-marine molluscs from the To'aga site provides important conroborative evidence for several interpretations made elsewhere in this volume. That the snail assemblages at To'aga are dominated by a suite of synanthropic species (the so-called "atoll fauna') known to have been widely disseminated by human agency highlights our interpretation of the To'aga coastal terrace as a strongly anthropogenic habitat. The absence of endemicfindigenous species below Layer IIC in the main trench supports our interpretation of the coastal terrace as a narrow, exposed, calcareous depositional environment prior to human colonization. After human settlement and the establishment of crop plants and other introduced vegetation at the beginning of the first millennium B.C., the coastal terrace began to sustain an adventive snail fauna of synanthrpic species. The high concentration of these snails in Layer HA- I is also consistent with our interpretation of this deposit as a fonner stable, vegetated A1 horizon paleosol. The snail assemblages from To'aga further document of the early spread of synanthropic species into central Polynesia. Such species as Lamellaxis gracilis, associated with Polynesian gardening environments, were already known to have been present in Lapita contexts (Christensen and Kirch 1986; Kirch 1988). The presence of this and four other species at To'aga reflects the propensity of early Pacific colonizers to create "transported landscapes." REFERENCES CITED Abbott, R. T. 1958. The gastropod genus Assiminea in the Philippines. Proceedings ofthe Academy of Natural Sciences ofPhiladelphia 110:21378. Baker, H. B. 1938. Zonitid Snailsfrom Pacific Islands, Part 1. Bemice P. Bishop Museum Bulletin 158. Honolulu. Christensen, C. C. 1983. Analysis of land snails. IN Archaeological Investigations of the WaimeaKawaihae-Mudlane Road Corridor, Island of Hawaii: An Interdisciplinary Study of an Environmental Transect, eds. J. T. ClaIk and P. V. Kirch, pp. 449-71. Departnental Report 831, Anthropology Departnent, Honolulu: Bernice P. Bishop Museum. Christensen, C. C., and P. V. Kirch 1981. Nonmarine mollusks from archaeological sites on Tikopia, Southeastem Solomon Islands. Pacific Science 35:75-88. . 1986. Nonmarine mollusks and ecological change at Barbers Point, O'ahu, Hawai'i. Bishop Museum Occasional Papers 26:52-80. Cooke, C. M., Jr. 1926. Notes on Pacific land snails. Proceedings ofthe Third Pan-Pacific Science Congress, pp. 2276-84. . 1934 Land shells of Makatea. Occasional Papers of the Bernice P. Bishop Museum 10:111. Cooke, C. Montague, Jr., and Y. Kondo 1960. Revision of Tornatellinidae and Achatinellidae (Gastropoda, Pulmonata). Honolulu: Bemice P. Bishop Museum Bulletin 221. Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press. Harry, H. A. 1966. Land snails of Ulithi Atoll, Caroline Islands: A study of snails accidentally distributed by man. Pacific Science 20:212-23. Kirch, P. V. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Seattle: Burke Museum. -. 1989. Non-marine molluscs from the rockshelter sediments. IN Prehistoric Hawaiian Occupation in the Anahulu Valley, O'ahu Island: Excavations in Three Inland Rockshelters, ed. P. V. Kirch, pp. 73-82. Contributions of the University of Califomia Archaeological Research Facility No. 47. Berkeley. Pilsbry, H. A. 1906. Manual of Conchology, Second Series: Pulmonata, Vol. XVIII, Achatinidae: Stenogyrinae and Coeliaxinae. Philadelphia: Academy of Natural Sciences. . 1916-18. Manual of Conchology, Second Series: Non-Marine MoUuscs Pubdonata, Vol. XXIV, Pupillidae (Gastrocopdnae). Philadelphia: Academy of Natua Sciences. Rdgle N.J. 1964. Nonmaine mollusks of Rogeap AUlMaral slanids Pac#c Science 18:126-29. Rolct, B. V. 1989. Hanai: Changing subsistence and ecology in the prehistory of Tahuata (Marquesas Islands, French Polynesia). Un- 121 published PtLD. dissertation, Yale University. Solem, A. 1959. Systemadcs and Zoogeography of the Land and Fresh-water Mollusca of dte New Hebrides. Fieldiana: Zoology 43. Chicago. . 1982. Endodontid Land Snailsfrom Pacfic Islands, Part l, Familes Punctidae and Charopidae, Zoogeography. Qiicago: Field Museum of Natuml History. 9 THE TO'AGA CERAMICS T. L. HUNT AND C. L. ERKELENS CsAmi HAvE PLAYED a ciitical role in understdig prehistory in Samoa and West Polynesia. They are usually well peserved, archaeologically visible, and carry a large amount of infoimation on variation in style (temporal and spatial), technology, function, and raw material. While ceramics have proven useful for making culture-iistorical inferences about Samoa and the region, ftey also present several interesting problems to be resolved in their own righlt First, what is the nature of ceramic variability (temporal and spatial) in Samoa? Second, what kinds of change (stylistic, techoogicalimaterial, and functional) occuned over the duation of ceramic prodution in Samoa? And third, why did ceramics disapear in Samoa after an appoximate thousand-year sequence of production? In this chapter we describe the ceramic assemblage from the To'aga excavations and begin to address hese questions through analyses of ceramic clay composition, technology, function, and style. THE ASSEMBLAGE The 1986 excavations on Ofu and neighboring Ta'u produced a total of 147 sherds (32 sherds from To'aga), and were analyzed and reported in Hunt and Kirch (1988:169-71). [be 1987 field season added 1464 sherds to the To'aga assemblage. These sherds were described by Kirch et al. (1990:7-8) and are further analyzed here. Excavations in 1989 provided an additional 938 sherds to the previous total. In sum, hre seasons of fieldwoik atTo'aga (AS-13-1) have yielded a total ceramic assemblage of 2434 sherds. The To'aga ceramic assemblage is significant in several respects. The sherds come from a deep, well-stratified site dated with several radiocarbon determinatons (see chapters 5 and 6). The assemblage spans the full duration known for ceramic production in Samoa, and consists of quantities of the pottery distinguished as "Ihickware" and "'thinware" (see Green 1974; Holmer 1980, Clark and Herdrich 1988). Pottery from To'aga, Samoa's eastemnmost ceramic-bearing site, dates to the firt millennium B.C. Finally, the To'aga assemblage is among the largest excavated in Samoa-only the SU-SA-3 (Green 1974) assemblage from Vpolu is larger-and thus provides an adequate sample to assess several dimensions of variability. The entire ceramic assemblage has been cataloged, with individual sherds enumerated. In the most general terms, the assemblage can be divided into broad classes of thickware and thinware sherds. This division of wares is based on sherd thickness, temper size, and paste texture. Such distinctions are qualitative (and somewhat impresionistic), however, and do not accurately reflect the range of variability present in the assemblage,-hence the need for detailed analysis. Vessel parts include only direct rims (no necks) and body serds, indicative of a single class ofvessel forms comprising open, round-based bowls. While occurring in only minor 124 The Tolaga Site frupencies, classes ofdecoration include red- slipped, carved paddle-imprsed, incised, and rm sherds with notched, impesd, or crenelated lips. Frequencies of sherds by these broad classes in stratigraphc context are sunmarized in tables 9.1-2. Detailed analysis of ceramics is extremely labor intesive. Observations and measurements on individual sherds can require as much as five to ten minutes eacL This problem requires selecting a representative sample from the larger assemblage. In the case of To'aga, we chose excavation units with largersamples. Also, pottery was selected from excavation units to span the full temporal sequence represented at To'aga. This stagy enabled us to assess change across the fu sequence of pottery manufacture at Toaga. Sherd samples were drawn from the main areal excavation of 1987 (units 1 and 4-9, see Kirch et al. 1990 and chapters) and from units 27 and 28 excavated in 1990. The total sample selected for intensive analysis is 737 sherds. The main areal excavation of 1987 (units l and 4-9) contained a large number of sherds (n=527) within a well-dated, statified context The primary ceramic-bearing occupation layers are dated (Beta25033, 25034) to an averaged coneted range of 306-138 B.C. at one stard deviation (see chapter 6 and Kirch et al. 1990). Units 27 and 28, excavated in 1989, were selected because they provided large samples from early contexts (see chapters 5 and 6). Based on radiocarbon dates and statigraphic correlations, the sheris for intensive analysis can be divided into early, middle, and late periods for comparative prposes. The early ceramics, from layer III in units 1, 5-7, 9, and 28, range from approximately 1250-500 B.C. Middle period ceramics, from layer I (B and C) in units 1, 4-7, 9, 27, and 28, range from about 500 B.C. to the beginning of the Chuistian era. The late period sample, from layer II (A) in units 4-9 and 27, dates from the time of Christ and may span the first 200-300 years A.D. These are not "ceramic periods" or "hases," but simply represen a three-pan division devised to analyze change in the sample. The pol for intensive analysis was designed to assess variability in raw materials, technology, style, and fuction. In addition to recoding provenience (by unit, stratum, and level) and catalog number for each sherd, the following analytic protocol was used and observations or measurements coded for analysis (using SPSS- PC+; Nonrsis 1986): 1) Exterior surface Etment: 0. eroded ("missing data") 1. plain (untextred) 2. wiped (striations present) 3. puddled 4. slipped 5. carved paddle impressed 6. residue obscuring ("missing data") 2) terior surface Ltatment: (same criteria as above) 3) Orientation ofinclusions ("preference') relative to vessel walls: 1. indeterminate (i.e. no definitive long axis to gais) 2. random orientation 3. paalel 4. perpendicular 4) Interior anvil casts: 0. indeterminate ("missing data") 1. present 2. absent 5) Exterior paddle marks casts: 0. indeterminate ("missing data") 1. present 2. absent 6) Exterior hardness (Mohs scale) 7) Interior hardness (Mohs scale) 8) Exterior surface color (Munsell): 0. erded ("missing data") 1. 75R (& value & chroma for all modes) 6. 7.5YR 2. SR 7. 10YR 3. IOR 4.2.5YR 8.2.5Y 9.5Y 5.5YR 9) Interior surface color (Munsell): (recorded same as above) 10) Oxidation/reduction pattern (using pattern template) 11) Organic residue: 0. absent 1. interior 2. exterior 3. both surfaces 12) Weight of sherd (grams) Ceramics 13) Mean sherd thickness (mm, 3 measuements/3) 14) Variance in sherd thickness (mm, maximum minus minimum value) 15) Size modality of temper (Wentworth scale using grain size template, sherd viewed under lOX magnification): 0. canrKot determine temper size modality ("miss- ing data") 1. granule (-1 phi; >2.0 mm) 2. bimodal granule, very coase sand 3. very coae sand (0 phi; >1.0 mm) 4. bimodal very coarse, medium sand 5. medium sand (2 phi; >0.25 mm) 6. bimodal medium, fine sand 7. fine sand (3 phi; >0.125 mm) 8. bimodal fine, very fine sand 9. very fine sand (4 phi; >0.0625 mm) 16) Rank order of temper by material in hand specimen only (fist temper by rank, second, tIrd); materials code: 0. indeterminate ("missing data") 1. black tachyte 2. green olivine 3. gray basalt 4. clear tanslucent crystals (quartz) 5. calcaous sand 6. ferous peds 7. opaque feldspathic crystals 17) Decoration technique (note: carved paddleimpressed and red slip are included under the dimension of surface treatment): 0. undecorated (or eroded, "'missing data") 1. tool impressed 2. incised Rims Only 18) Rim course (angle to central vertical axis): 1.direc 2. inverted 3. evented 19) Cross-section of lip: 1.onded 2. square 3. poned >2) Rim profile (degree of thickening): 1. noe, parallel walls 2. tined 3. thickened, exterior 4. thiickted,inteior 125 21) Mean thickness of rim at lip (mm, 3 measures/ 3) 22) Variance in thickness of rim at lip (mm, maximum minus minimum value) 23) Estimated orifice diameter (10 cm intervals) ANALYTIC PROCEDURES Observation of exterior and interior surfaces allows one to determine the final finishing metxhds used on vessels prior to firing. Where sherds have an eroded exterior or interior surface, such observations could not be made and were coded as "missing data." Plain surfaces were simply smooth, with no other finishing techniques evident. Wiped surfaces were recognized by fine, parallel striations that were made in the clay pror to firing. Puddled surfaces, sometimes called "self-slipped" (Rye 1981:57), are those formed by wetting the surface, thus bringing the finest clay particles to the surface of the paste. This technique is recognized by textural differences on the sherd surface and in cross-section. Slip, a thin surface coating created from a fluid suspension of clay in water, commonly has a different color than the body (Rye 1981:41). Also, slip diffesfrom the sherd body in texture. Carved, paddle-impressed surface treatment is usually produced with secondary forming using a paddle and anvil. The textured (patterned) paddle leaves a raised design on the vessel surface. The "preferred" orientation of sand temper inclusions relates to techniques of primary forming (Rye 1981). Slab building and coiling can be detected, in part, by differences in orientation. A definitive long axis for the particles is necessary to observe orientation. Interior anvil casts and exterior paddle (carved and plain) maiks were noted as present/absent These traits reveal the use of the paddle and anvil technique in secondary forming in the ceramic production sequence. Hardness was measured on the interior and exterior surfaces of sherds by a scratch test using the Mohs scale. Hardness, measured on this ordinal scale, may relate to ceramic strength, raw materials, firing, and post-depositional diagenesis. Sherd color was measured on the exterior and interior surfaces using the Munsell Soil Color Chart (1988). Ceramic color, while complex, reflects 126 The Tolaga Site Table 9.1 Ceramics from the 1987 Excavation Units Unit Layer Thin n (%) Thick n (%) Rims Other Sherds Red Slip Temporal Analytic Period Total 1 IIB IIC III 2 (3) 0 (0) 2 (67) 66 (97) 0 (0) 1 (33) 10 1 0 Middle Middle Early 68 1 3 4 IIA IIB 0 (0) 4 (44) 7 (100) 5 (56) 0 0 Late Middle 7 9 5 IIA IIB III 4 (14) 12 (27) 8 (100) 24 (86) 32 (73) 0 (0) 0 4 3 Late Middle Early 28 44 8 6 IIA IIB IIC III 2 (8) 2 (13) 17 (74) 4 (80) 22 (92) 13 (87) 6 (26) 1 (20) 3 0 0 2 Late Middle Middle Early 24 15 23 5 7 IIA IIB IIC III 1 5 5 4 (13) (71) (83) (80) 7 (87) 2 (29) 1 (17) 1 (20) 1 0 0 .1 Late Middle Middle Early 8 7 6 5 Late 27 Late Middle Early 69 22 3 Late Middle 4 1 Late Middle Middle 8 98 2 Early 20 Late Middle 182 95 8 IIA 0 (0) 27 (100) 0 9 IIA IIB III 2 (3) 2 (9) 1 (33) 67 (97) 20 (91) 2 (67) 11 IIA IIB 1 (25) 1 (50) 3 (75) 0 (0) 0 0 11 IIA IIB IIC 3 (38) 8 (8) 1 (50) 5 (62) 90 (92) 1 (50) 2 11 0 12 III 20 (100) 0 3 14 IIA IIB 7 (4) 10 (11) 175 (96) 85 (89) 10 0 128 (15) 663 (85) (62) 10 Totals* * Rims, slipped, and decorated 0 Incised (1) 0 1 Notched rim (1) 791 sherds included with thick/thin sherd counts; eroded sherds not included in total. Ceramics 127 Table 9.2 Ceramics from the 1989 Excavation Units Temporal Thin Unit Layer 15 23 24 9 (90) 22 (73) 13 (87) 2 (50) (0) (0) 12 (40) 9 (100) 3 (100) 18 (60) IIIC 4 (7) 12 (9) 10 (22) IIA IIB IIIA IIIB I 0 0 Rims Slip 1 0 2 0 Incised (1) 0 53 (93) 120 (91) 36 (78) 4 10 10 1 (25) 2 (12) 3 (75) 14 (88) 0 IIIC 8 (13) 32 (35) 17 (28) 54 (87) 60 (65) 44 (72) 10 3 IB 1 (33) 2 (67) 0 II (0) (8) 3 (30) IIIA IIIB IIA IIB IIC 29 II IIIA 5 (100) 0 5 57 (92) 7 (70) 3 1 (100) 16 (70) 52 (50) (0) 12 (100) 2 (22) 0 7 (78) 1 Impressed (6) 0 4 2 1 Analytic Period Total Middle Early Early Early 10 30 Late Middle 9 3 30 Early 3 0 0 (0) 7 (30) 51 (50) Other Sherds 3 2 1 IIA IIIB 4 57 132 Middle Middle 4 16 Early Early 46 1 Early 62 92 61 1 Middle 3 Late Middle Early 62 Decorated (4) 5 10 0 Late 1 2 13 Middle Middle 23 103 1 Middle Early 12 9 34 39 0 9 Impressed (1) II 2 (6) 32 (94) 28 (72) 1 Middle 11 (28) 6 Early 198 674 (68) Totals* 15 Early Early Early IIIA 30 * 1 Red 2 27 28 (%) (10) II III 21 n 8 (27) 2 (13) 2 (50) II IIIA IIIB IIID 16 20 n (%) Thick 872 Rims, slipped, and decorated sherds included within thick/thin sherd counts; eroded sherds not included in total. 128 The To'aga Site vaniability in raw material, pyrotechnology, and use. The oxidation/reduction pattern was recorded for sherd cross-sections using an inductively derived pattern template (e.g., Rye 1981:116). These patterns are indicators of the atmosphere and temperature of firing (Rye 1981:115-18). The presence or absence of organic (carbonaceous) residue was recorded. Residues supply important clues of ceramic function. Variation in sherd thickness has proven significant in Samoan pottery studies. Sherd thickness is estimated as a mean value from three measurements on different parts of the sherd. This mean value is a more reliable measure of sherd thickness than a single measurement (Barry 1978). Three thickness measures also provide a range expressing the variance in sherd thickness. Variance measures the uniformity or evenness of sherd thickness. Temper was examined in terms of its rank by raw material as estimated from hand specimens only. Also, the size modality of the sand temper grains was recorded using a template produced with sand samples of varying phi sizes on the Wentworth scale. Sherds were viewed under lOX magnification. Dickinson's study of sand temper petrography offers a much more detailed and reliable (some materials are difficult to determine in hand specimen alone) analysis of a selection of sherds from the To'aga assemblage (see chapter 10). For sherds with one or more intact surfaces, decoration technique was recorded as tool-impressed, incised, or undecorated (plain). Carved paddle-impressed and red slip, while decorative techniques, are included under the dimension of surface treatment. Rim sherds were identified in terms of their course relative to the central vertical axis of the pot (e.g., direct, everted). The cross-section of the lip (form) and the profile of the rim, or the degree of thickening, were also classified. Finally, the thickness of each rim was measured, and orifice diameters were estimated. Of the 737 sherds sampled for intensive analysis, 583 sherds retained both interior and exterior surfaces enabling measurement or observation of the characteristics listed above. The results of intensive macroscopic analysis are summarized and discussed below. These results, together with compositional results (below and chapter 10), provide the basis for reconstructing aspects of ceramic raw material use, production technology, style, and function. RESULTS OF MACROSCOPIC ANALYSIS Many of the results used for intensive analysis in tables 9.3-13. Specific characteristics that relate to material use, forming techniques, pyrotechnology, style, and function are discussed below. The analysis of sand temper composition and grain size modality (see table 9.4) revealed the following: temper ranges in size from grains measuring approximately 2.5 4 mm (i.e., granules are summarized - Table 9.3 Frequency of Body and Rim sherds Frequency Percent Cum. Percent Body Rim 683 54 92.7 7.3 92.7 100.0 Total 737 100.0 Sherd (Valid cases = 737; Missing cases = 0) Ceramics 129 Table 9.4 Frequency of Sherds by Temper Size Mode(s) Temper Size Modes Frequency Percent Cum. Percent granule granule, very coarse very coarse very coarse, medium medium medium, fine fine 30 189 140 114 81 91 92 4.1 25.6 19.0 15.5 11.0 12.3 12.5 4.1 29.7 48.7 64.2 75.2 87.5 100.0 Total 737 100.0 (Valid cases = 737; Missing cases = 0) < -1 phi) in many sherds, to particles visible only under magnification (i.e., fine to very fine sand, > 3 phi). No attempt was made to observe modes of finer (e.g., silt) particles. Most of the sherds in the sample assemblage (86%) had a mixture of sands and contained more than one temper compositional class. Thicker sherds tended to have coarser temper and thinner sherds tended to have finer temper. Temper size correlates to temper composition (x2 = 237.21, df= 9, p > .0001) in that well-rounded, calcareous sand grains ame smaller than the very coarse (often angular) sand class (O phi). Grain size and temper composition are also associated with sherd thickness. Calcareous sad temper occurs most often in thinware (73% of this temper occurs in thinware). Yet, thinware cannot be described as predominantly calcareous sand tempered because only 26% of the thinware has a ped anceof calcareous sand temper. On the oter hand the largest tempers, grains of glassy black tm~hyte, dull grey basaltic, rounded olivine, pale p feldspars, clear crystalline, and red ferrous, rounded grains were resented in the smallest size classes as well. Color variability was assessed by plotting values in ascateplot to check for trends or grouping tendenies. Bivariate plots for a three dimensional cassification were achieved by plotting Munsell hue against value-chroma (figures 9.1-3; e.g., see Bishop et aL 1988). These plots reveal that color variability in the To'aga ceramics is comparable for all hree time periods and for both thick- and thinwares. Sherds range from red (lOR) to yellow-red (lOYR) and cover an array ofvalue and chroma The majority of the assemblage is reddish brown or gray in color. Early ceramics include red-slipped sherds (10 R 5/6 and 2.5 YR 5/6) distinctive in coloration from the west of the assemblage. Although analysis showed that interior and exterior surface colors have similar distribution, the conrelation on individual sherds was poor (Pearson's r = 0.143, p > 0.001). Oxidation/reduction patterns (table 95) play some part in this correlation, since 24% of the sherds have interior surfaces that were not fully oxidized as compared with only 3% for the exterior surfaces. As a result, more sherds have interior surfaces that are darker than their exterior surfaces. Most of the sherds (68%), however, were fully oxidized during firing. Following analysis, the entire assemblage was inspected to ceck for further variation in color (or other unrecorded differences in temper, etc.). Five sherds, from units 15129/30 were anomalous in color. These sherds (5 YR 6/6 and 7.5 YR 6/3) are described respectively as reddish yellow and light brown in the Munsell system. The former group of thickware sherds (5 YR 6/6) is similar in color to the predominate color of the sberds from Upolu Island designated "Falemoa Tan" by Holmer (1980:114). Sherd hardness, measured on the Mohs scale, 130 The Tolaga Site I I I I I C I I I I I I 1 6- 1 3 1 E 7B C H U K 1 1 4.5*815 327I I I 1 I I 27 I 36 I 45 I 1 I I I I 63 54 I I VALUE AND CHROMA Figure 9.1 Plot of exterior sherd color (hue with value and chroma) for the late period (note to figures 9.1-3: number of cases in each position is shown in sequential order by 1-9; A-; 10-35 cases; and *, more than 36 cases). I I I 5 I I I I 1 * 22 I I I 1 2 6- I K 1 B H 0 U 3 1 4J Q* 4.5- *1 4 1 1 3I I 2 I 27 9 I 3 36 I 4 I 45 5 54 9 I 6 I 63 VALUE AND CHROMA Figure 9.2 Plot of exterior sherd color (hue with value and chroma) for the middle period (see note to fig. 9.1). I I Ceramics I I 1 I I I I I I I I 131 I 1 X 11 6- 1 1 IR 1D H 2 U 1 1 B 4.5R7 2 3I * 2 * 27 3I I * * * I 36 4 * * I 5 * 54 45 * I 6 * - lF 63 VALUE AND CHROMA Figure 9.3 Plot of exterior sherd color (hue with value and chroma) for the early period (see note to fig. 9.1). ranged from 2.0 to 8.0 (tables 9.6-7). This is a remaikable range ofhardness, but reflects the difficulty of measuring paste hardness with a scratch test in sherds with abunant temper (which is consistently hard). Ceramic tiles manufactured from colluvial clay samples and fired in an open fire (for appoximately fifteen minutes), or in the muffle furnace (5000C for fifteen minutes) were uniformly 3 in hardness. There is a strong correlation (r = 0.8467, p > .0001) between the interior and exterior hardness of individual sherds. The sample analyzed (n = 737) had 583 sherds which retained both surfaces (i.e., non-eroded), allowing the measurement of mean thickness for each sherd. Results show that sherds range from 4.20 mm to 16.97 mm in thickness. The total sample mean sherd thickness is 9.42 mm (a = 2.84), however the distribution is not normal, but has hree modes (figure 9.4). The late portion of the sample (n = 196) with 140 measurable sherds, ranges from 5.60 mm to 14.55 mm in thickness. mean sherd thickness is late sample 10.14 mm (a = 1.77) (figure The middle period sample (n = 41 1) with 323 9.5). measurable sherds, ranges from 4.20mm to 16.97 mm in mean sherd thickness. The sample mean thickness is 9.97 mm (a = 3.10) (figure 9.6). The early ceramics (n = 130) had 120 measurable sherds with a mean thickness of 7.09 mm (a = 1.68). These sherds range in thickness from 4.30 mm to 14.52 mm in a somewhat normally distributed range of measurements (kurtosis 4.357, skewness 1.783), but skewed toward the timner sherds (figure 9.7). Differences in sherd thickness are summanzed by temporal-analytic periods in table 9.14. Rims for the entire To'aga assemblage were analyzed and identified to class (see protocol). These forms are illustrated in figure 9.8. One large reconstructed sherd provided a measurable portion of a rim (9% of the estimated total) so that the diameter is reconstructed as 48 cm. About 89% of the rims are oriented 900 to a central vertical axis of the vessel. Other rims have angles approximating 800. All rim courses are direct, with the majority (80%) having a squared lip cross section. The remaining rims (20%) are rounded in cross section. The only decorated rims are from early contexts, comprising approximately 7% of the total. The lips of these decorated rims are impressed with narrow tools forming U-shaped notches, repeating parallel lines perpendicular to the rim, or in one case, oblique to the rim. A rim with a crenelated lip was also recovered. Other decorated sherds are small in number. Only 30 slipped sherds and 1 1 other decorated 132 The To'aga Site Table 9.5 Frequency of Sherds by "Preferred" Orientation of Inclusions Relative to Vessel Walls Inclusion Orientation Frequency Percent Cum. Percent Indeterminate Random Parallel 199 458 80 27.0 62.1 10.9 27.0 89.1 100.0 Total 737 100.0 (Valid cases = 737; Missing cases = 0) Table 9.6 Frequency of Sherds by Oxidation-Reduction Pattern in Cross-Section Pattern Fully oxidized Core oxidized Ext. oxidized it. surf. reduced Fully reducd Ext. surf. reduced Surfs. rediced Ext. reduced Ext. surf.reduced int. reduced Total Percent Valid Percent Cum. Percent 380 51.6 67.9 67.9 10 68 1.4 9.2 7.5 1.8 12.1 9.8 69.6 81.8 91.6 Frequency 55 9 14 2 7 1.2 1.6 93.2 1.9 .3 .9 2.5 .4 1.3 95.7 96.1 97.3 100.0 15 2.0 2.7 177 24.0 Missing 737 100.0 100.0 (Valid cases = 560; Missing cases = 177) Ceramics Table 9.7 Frequency of Sherds by Exterior Hardness (Mohs Scale) Exterior Hardness Frequency Percent Valid Percent Cum. Percent 2 3 4 5 6 7 8 Eroded 31 113 109 154 145 53 18 114 4.2 15.3 14.8 20.9 19.7 7.2 2.4 15.5 5.0 18.1 17.5 24.7 23.3 8.5 2.9 Missing 5.0 23.1 40.6 65.3 88.6 97.1 100.0 Total 737 100.0 100.0 (Valid cases = 623; Missing cases = 114) Table 9.8 Frequency of Sherds by Interior Hardness (Mohs Scale) Frequency Percent Valid Percent Cum. Percent 2 3 4 5 6 7 8 Eroded 24 110 122 141 158 52 31 99 3.3 14.9 16.6 19.1 21.4 7.1 4.2 13.4 3.8 17.2 19.1 22.1 24.8 8.2 4.8 Missing 3.8 21.0 40.1 62.2 87.0 95.1 100.0 Total 737 100.0 100.0 Interior Hardness (Valid cases = 638; Missing cases = 99) 133 134 The To'aga Site Table 9.9 Frequency of Sherds by Exterior Surface Treatment Surface Treatment Frequency Percent Valid Percent Cum. Percent Plain Wiped Puddled Slipped Eroded 128 13 481 4 111 17.4 1.8 65.3 .5 15.0 20.4 2.1 76.8 .6 Missing 20.4 22.5 99.4 100.0 Total 737 100.0 100.0 (Valid cases = 626; Missing cases = 111) Table 9.10 Frequency of Sherds by Interior Surface Treatment Surface Treatment Valid Percent Cum. Percent 18.9 2.1 77.3 1.4 0.3 18.9 21.0 98.3 99.7 100.0 Frequency Percent Residue Eroded 124 14 508 9 2 80 16.8 1.9 68.9 1.2 0.3 10.9 Missing Total 737 100.0 100.0 Plain Wiped Puddled Slipped (Valid cases = 657; Missing cases = 80) Table 9.11 Frequency of Sherds by Interior Anvil Casts Anvil Casts Frequency Percent Valid Percent Cum. Percent Absent Present Indeterminate 85 297 355 11.5 40.3 48.2 22.3 77.7 Missing 22.3 100.0 Total 737 100.0 100.0 (Valid cases = 382; Missing cases = 355) Cerenics Table 9.12 Frequency of Sherds by Exterior Paddle Marks Frequency Percent Valid Percent Cum. Percent Absent Present Indeterminate 54 282 401 7.3 38.3 54.4 16.1 83.9 Missing 16.1 100.0 Total 737 100.0 100.0 Paddle Marks (Valid cases = 336; Missing cases = 401) courht 25 _~ granule, _~ 20 - - 10 -- 5 - 09A v. coarse, medium medium, fine a= 15 v. coarse 4 5 6 7 8 9 10 11 12 fine, v. fine 13 14 15 16 thickness (mm) Figur 9.4 Thickiess and temper size histogram for the Toaga ceramic assemblage (n = 538). Temper classes are indited by tew variable shading of the histam bars. 135 The To'aga Site 136 N d E * U, *IC2 0 oo: o! oqfio v 00 -401 N 00"()o *aC Ch _ at o ; 00 e 6 6 0 00 (%.- x 0 0 8._o m EI" 0 N I (A 00 C' ;Ob 0* W ~u " 0%. 0 si U _m N N .2~~~~~~C e 6 9 6 00 I% 0 3 11 U 00 O r ~~~~F co * oa a, .0 to *r W 0 U~~~~~~) 0 W ml 1 0% §t , w 0% 0 0~ as N co es 0 ON 8o 0%. o xo O _- t.- x~ 0 a, .0 T- s Ceramics 137 count 25 20 15 10 5 0 4 5 6 Figure 9.5 7 8 10 11 12 9 thickness (mm) 13 14 15 16 Thickness histogram for late period ceramics (n = 196). sherds were recovered. The color of the slip is red (10 R 5/6 and 2.5 YR 4/6). This slip occurred on both the interior and exterior of the sherds and was found on both rim and body sherds. Three body herds from the assemblage are decorated. Two sherds have incised lines, although both are small making the overall patterns indistinct. The third is a body sherd with parallel-ribbed, carved paddle essions on the exterior surface of the vessel. This sherd was from an early context (unit 20, layer IB), and compares to other paddle-impressed sherds known from Westem Samoa (Green and Davidson 1969:pl. 17). All sherds recovered during excavations were examined for residue in the field, (i.e., prior to any cleaning). Ten sherds with substantial quantities of carbonized residue were discovered; seven with residue on the interior and exterior, and three on the interior alone. This residue has yet to be identified but is probably the result of cooking starchy foods (see Hill et al. 1985). CERAMIC COMPOSITIONAL MICROANALYSIS Analysis of macroscopic ceramic traits supports many research objectives, especially those examined here. Documenting raw material vanability requires additional work, especially with respect to temper and clay of the ceramic fabric. The temper component has been analyzed and discussed by Dicdinson (chapter 10). Here we present the compositional microanalysis of thee clay samples and the clay portion of the ceramics from To'aga. These results allow us to address questions of clay variability (as a part of technology) and the potential for ceramic exchange in Samoa. A sample of twenty-nine sherds was chosen for their visual differences in thickness, temper, and paste in hand section. Also, this variety of sherds came from excavation contexts that could be inferred to be of different ages (table 9.1). Age differences 138 The To'aga Site count 30 25 20 15 10 5 0 4 5 6 Figure 9.6 7 8 9 10 11 12 thickness (inm) Thickness histogram correspond to the analytic divisions made for the larger sample (i.e., early, middle, and late). The sherd samples were also analyzed by Dickinson (chapter 10) for their sand temper petrography. Examination of the To'aga ceramics by both sand temper petrography and elemental analysis takes advantage of the strengths of each approach (Hunt 1988). In addition to the shends, three clay samples were collected from colluvium near the base of the cliff at To'aga on transects 1, 5, and 9. These clays were fired in a furnace (at 500(C) for fifteen minutes to produce ceramic tiles resembling sherds. One of these clay samples (from Transect 9) was prepared in the laboratory as a fired ceramic tile (sherd). It was also analyzed by Dickinson (chapter 10) to compare "self-tempered" sherd petrography. The elemental microanalysis was accomplished using an energy-dispersive spectrometer (EDS) integrated with a scanning electron microscope for middle 13 period ceramics (n 14 = 15 16 411). (SEM). SEM/EDS microanalysis is described in this chapter, and these results are integrated with those from the petrographic analysis. The distinct advantage of the SEM/EDS is in the selectivity afforded by the microscope component of the instrunent. Using the SEM in conjunction with an x-ray analyzer, it is possible to characterize the clay matrix alone, or individual inclusions, slips, and residues (e.g., Hunt 1989). Analyses described here were conducted by one of us (TLH) on a JEOL model JSM-840A SEM fitted with a Tracor Northem energy-dispersive x-ray detector housed at the University of Washington. Selective elemental microanalysis of pottery is possible by the coupling of an x-ray analyzer with the SEM. In the simplest terms, dte SEM provides a source of electrons of appropriate energy that impinge on a sample and cause the emission of xrays. The x-rays emitted have energies and relative abundances that reflect the elemental composition of 139 Cereaics count 20 15 10 5 0 4 5 6 Figure 9.7 7 8 9 10 11 12 thickness (inm) 13 14 15 16 Thickness histogram for early period ceramics (n = 130). the sample. The characteristic x-rays are deted by a lithium-drifted silicon Si(Li) crystal that-together with electic amplifiers and signal processors- collects and electrically sorts all the energies from the x-rays emitted. Under nonnal operating conditions, elements with atomic numbers above 10 (Na = 11), and below 100 (Es = 99) are deteable. The conversion of x-ray emissions into a compositional spectrum (figure 9.9) and potential quantitative data is achieved through a series of electic components described in some detail by Postek et al. (1980) and Goldstein et al. (1981:22224). Qualitative and quantitative analysis of the xray spectrum for the composition of a particular sample is complex, yet well understood (Goldstein et al. 1981:275-392). As in other recent studies, (e.g., Dunnell and Hunt 1990, Graves et al. 1990, Hunt 1989), quantitative analysis of To'aga ceramic clays used the ZAF conection method (see Goldstein et al. 1981:308). The final values calculated are quantities of elements (by weight and atomic percents) present on the cross-section surface of the sample at the point/area impinged by the electr beam. A goodness of fit between those elements quantified and stand intensities is evaluated by a chi-square test This test provides an objective criterion to evaluate the goodness of fit for the peak-fitting algorithms used in each particular application (Goldsteinet al. 1981:411-12). The analystcan judge the success or failure of x-ray collection for goodness of fit for a particular specimen on statistical criteria for each spectrum and quantitative analysis generated. All these features are available through the Tracor Northem software used. Minimum elemental detection limits for energydispersive microanalysis are below 0.1% under ideal settings, and typically less than 19%, with a relative precision of 1-5% throughout the elemental range detected (Hunt 1989). Rice (1987:375) notes the general concentration range for x-ray analyses as 140 The Tolaga Site L J~ ~ ~ o --w. - I.If i -f- A~~~~~~1-'O E ~~~~~~t _- o ' F;, K~~~~~~~~~~~ _ 8 I X~~~~~~~~~~~~~~~I Ceramics Figure 9.9 Elemental spectrum of clay composition for sherd 22 firn To'aga (Univ. of Washington SEM, 27 Dec. 1989, 1OKV, l200X). detection of major, minor, and trace elements (>100 ppm) Specimens were viewed on the CRT of the SEM while undergoing x-ray analysis. An area that appeared to be only day was selected at low magnification (30- 1OX) for analysis. This area was isolated by a step-wise increase in the magnification, allowing careful inspection of the region for inclusions or other anomalies. X-rays were collected from areas that appeared at higher magnification to be clay matrix only. All other specific sample ation and analytic procedures follow hose described in detail by Hunt (1989:155-59). Twelve major and minor elements, Na, Mg, Al, Si, P, Cl, K, Ca, Ti, Cr, Mn, and Fe, were selected for analysis. These multivariate data for thirty-two specimens (sherds and To'aga clays) form the data matrix used to search for compositional structure in the Toaga ceramics. Quantitative Analysis of Clay Elemental Data The goal of analysis of elemental data is orgaor reducing compositional variability into archaeologically meaningful groups (Armold et al. 1991; Harbottle 1976:42). Ideally, such groups represent discrete clay sources, and thus, the minimum number of production locales represented in the ceramic assemblage. While individual clay sources are not always distinguishable, the method has the potential to sort clays from different regions. nizing 141 Islands offer an especially good setting for sorting ceramic provenance because they vary in age and geologic orgin. In addition, clays come from small drainages and do not mix as in continental deposits. Sorting the clay elemental data matrix into meaningful groups generally requires deductive tests using multivanate statistics (e.g., Bishop and Neff 1989; Davis 1986). Such analyses are usually directed at infernng a probable number of distinguishable clays represented in an assemblage. Distinguishable clays might be used to infer production locales for prehistoric pottery. Some studies are successful at linking prehistoric pottery to specific (known) clay sources, either on quantitative criteria that range from ordinal tests to multivariate ones (e.g., Neff et al. 1988; Topping and MacKenzie 1988). Yet, as Arnold et al. (1991:85) have pointed out, individual clay deposits may not be easily distinguished, and in Oceania an island-or even a group of islands-as a unit of geographic space may form a single "source" in compositional space. The To'aga compositional data matrix was analyzed for its grouping tendency with two different clustering algorithms: average linkage between groups and Ward's method (Norusis 1986). These algorithms are agglomeradve and build groups or clusters from the individual specimens (sherds and clays). Given the problems of evaluating cluster dendrograms, a rigorous solution is in the use of different algorithms. Then, only those clusters which arise independently in different analyses are considered valid, or accurately descriptive (Alenderfer and Blashfield 1984:65; Dunnell 1983:146; Sokal and Sneath 1963:166). This strategy will work in data sets in which clay groups (i.e., not necessarily individual "sources", see Arnold et al. 1991) are chemically distinctive and can be detected by statistical measures. Discriminant function analysis was also used to examine compositional stnrcture in the To'aga data. Discriminant function analysis offers a deductive tool for testing structure in a multivanate data set (Bishop and Neff 1989; Davis 1986; Hunt 1989). As Bishop and Neff (1989) imply, ceramic compositional data sets are complexly multivariate and require going beyond an inductive search for structure (see also Harbottle 1976). 142 The To'aga Site Compositional Results Results of these clustering procedures were plotted as dendrograms (figures 9.10-1 1). From these dendrograms, comparable clusters of sherds (and clay sample tiles) occurring in both solutions can be deduced. Comparison of the Ward and average linkage between group cluster dendrograms reveals four identical clusters. Table 9.15 provides summary data and results for the sample. Cluster assignments from the two solutions (1-4, and two cases unassigned) were analyzed and plotted against first and second discriminant functions. The plot (figure 9.12) of discriminant function scores illustrates the distribution of the groups in multidimensional space. Cluster 4 includes seven sherds of both thickand thinware as well as the three clay samples collected from To'aga colluvium. This match suggests that local colluvial clay from To'aga was used in some pottery manufacture. The remaining three (1-3) clusters represent clay compositional groups as yet unmatched to samples from Ofu, or elsewhere. These unmatched clays are similar in composition to those of the local colluvial clay sample, and may come from other unknown sources/ source areas on Ofu, elsewhere in Manu'a, or beyond. A determination of local versus exotic provenance for the unmatched sherds would be premature; additional sampling of clays is necessary. Sherds from other islands in Samoa should also be tested for their compositional similarity to those of Rescaled Distance Cluster Combine C A S E Seq Label 20 22 05 25 09 21 23 06 15 26 27 08 24 07 14 17 16 13 19 29 18 11 12 30 03 04 02 01 10 32 31 Figure 9.10 20 22 5 25 9 21 23 6 15 26 27 8 24 7 14 17 16 13 19 28 18 11 12 29 3 4 2 1 10 31 30 0 5 10 I i I I i 15 20 i Il I i 25 I I I I I Dendrogram of Ofu pottery and clay samples using the Average Linkage (between groups) method. Ceramics 143 Rescaled Distance Cluster Combine C A SE Label Seq 20 22 05 25 09 24 29 18 13 19 14 17 16 21 23 06 15 26 27 08 07 10 32 31 03 04 02 11 12 30 01 0 5 10 15 i I i iI I 20 25 I 20 22 5 25 9 24 28 18 13 19 14 17 16 21 23 6 15 26 27 8 7 10 31 30 3 4 2 11 12 29 1 Figure 9.11 Denrgm of Ofu pottery and clay samples using Ward's method. the To'aga assemblage. The association of clay compositional groups with thickware, thinware, red-slipped ware, the paddle-impressed sherd, and the three colluvial clay samples from To'aga (table 9.15) shows that thickand thinware cannot be separated compositionally. All compositional groups are represented in thickand thinware. Discrhminant function analysis, using ware as the grouping variable, confirmed this observation. Scatterplots revealed little separation along the first and second discriminant functions (figure 9.13). The red-slipped pottery (n = 2) falls into groups 2 and 3, although a larger sample is needed to assess compositional variation in this class. The To'aga colluvial clays (in group 4) match sherds of thickware, thinware, and the carved paddle-impressed sherd. Compansons of ware with temper groups (table 9.15) identified by Dickinson (chapter 10) reveal that all four temper groups are represented in thickware. The profuse basaltic temper is found only in thickware, for this sample. Thinware contains sparse basaltic (temper group 2), feldspathic (temper group 3), and mixed (temper group 4) tempers (see chapter 10). The two red-slipped sherds in the sample have sparse basaltic temper. The paddle-impressed sherd and the analyzed "self-tempered" clay sample (from Transect 9 colluvium) have the mixed (temper group 4) temper, including calcareous sand. Calcareous sand in the colluvium suggests saltational transport of grains from the coast over the previously shorter distance to the colluvial deposits where mixing could 144 The To'aga Site Table 9.15 Sherds and Ofu Clays Selected for SEM/EDS Clay Elemental and Sand Temper Petrographic Analyses Specimen No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Provenience Class Unit6 IIa Unit 6 IIa Unit 6 Ila Unit6 IIa Unit 6 Ilb Unit 6 Ilb Unit 6 lIb Unit 6 HIb Unit 6 IIb Unit 6 Ilc Unit 6 IIc Unit 6 IIc Unit 20 Ilb Unit 20 HIb Unit 20 Ilb Unit 20 IIIa Unit 20 I11a Unit 20 I11a Unit 20 1I1b Unit 20 IlIb Unit 20 T11b Unit 20 IIIa Unit 20 IIIa Unit 20 IHIb Unit 20 Tub Unit 20 IlIc Unit 20 IHIc Unit 20 HlIc Unit 24 II Transect 9 Transect 5 Transect 1 Thick Thick Thick Thick Thick Thick Thick Thick Thick Thick Thick Thick Thick Thick Thick Thin Thick Thick Temper Group Clay Cluster 1 1 1 1 1 2 2 4 3? 2 4 4 2 1 3 2 3 3 2 2 2 2 2 4 4 4 2 2 2 4 4 4 4 1 3 4? 4 1 4 4 4 2 2 3 2 2 2 2 1 3 1 3 4? 1 3 3 Thin' Thin Thin2 Thick Thick Thin3 Thin Thin Thick Thin Thick4 Clay Clay Clay 1 Square rim with an impressed lip 2 Red-slipped 3 Carved paddle-impressed (parallel rib motif) 4 "Thickware" with red-slipped exterior - 4 - - 2 4 4 4 Period Late Late Late Late Middle Middle Middle Middle Middle Early Early Early Middle Middle Middle Early Early Early Early Early Early Early Early Early Early Early Early Early Early? --------- Ceramics x x x 4.0 4 1 * 1 13 .0 2 11 3 3 3*2 2 2 44 4 4 *4 4 * 2 2 4 -4.0 + x X x -6.0 -4.0 -2.0 x 4.0 6.0 First and second discriminant function scores for analysis based on cluster (1-4) as grouping variable; * indicates group center, # indicates unclassified as to cluster. Figure 9.12 x 2.0 .0 I x I x 4.0 3 3 3323 3 31 332*3 1 3 * 2 * 333 3 2 .0 1 -4.0 x Y. A- I -6. 0 -4.0 -2.0 Figure 9.13 X I I .0 2.0 I 4.0 ..AV Profuse basaltic temper (Dickinson's group 1) is associated with cluster 4 (that includes the three "self-tempered" colluvial clays) more than expected by chance alone (expected = 1.7, observed = 4). Furthermore, temper groups 1, 2, and 4 occur with clay compositionally indistinct from the To'aga colluvium samples. All temper groups are associated with clay group 1. It is also noteworthy ta the overall association is otherwise somewhat well dispersed. This observation suggests that in many of the specimens, temper and clay composition vary independently. Compositional groups tabulated by their analytic time penods show a pattern of decline. Compositional variation reflected by temper groups reveals a similar pattern of decline. The early time period (1250-500 B.C.) ceramics fall into all fourclay compositional groups, reflecting the greatest variety of clay (source) use. One of the clay sources in use during the early period is the local colluvial clay from To'aga, and it is in both thick- and thinware. The middle period (500 B.C.- A.D. 0) may show a decline to time clay groups, and does not include the colluvial clay from To'aga. The late (A.D. 0-300?) ceramics are only represented by the To'aga colluvial clay in thickware samples. A decline in compositional variability reflects the general simplification and homogenization of the total To'aga assemblage with time. This potential trend, however, could simply result from the smaller samples in the middle and late time periods. Additional samples must be analyzed to test a hypothesis of change in the compositional variability in the To'aga sequence. CONCLUSIONS 6.0 First and second discriminant function scores for analysis based on time period (early [1], middle [2], and late [3]) as grouping variable; * indicates group center. occur natualy. Temper with calcareous sand could also represent purposefil mixing on the part of ancient potters. The association of clay compositional groups temper groups (table 9.15) defined by Dickinson reveals some maiked correspondence. with 145 Ceramics provide a critical source of information for building chronologies and inferring cultural relatedness because they vary in style. Recent advances in physico-chemical and archaeological analyses (Rice 1987) open the door to many new questions in ceramic studies. In this study, we focused on ceramics in terms of material composition and provenance, production technology, style, and function. The To'aga ceramic assemblage is particularly valuable for this kind of detailed study. The assemblage is large and comes from a stratified site where a long chronology of pottery production can be delineated. In the discussion that follows, we 146 The Tolaga Site offer some partial answers to the questions posed at the beginning of this chapter. Ceramic Provenance and Production Technology The day microanalytic and sand temper petrographic results provide a basis for several conclusions and new hypotheses conceming ceramic provenance and production technology. Thickware, thinware, and a carved paddle-impressed sherd from To'aga can be inferred to be of local production, using colluvial "self-tempered" clay source(s) from Leolo Ridge on Ofu. Such clay(s) could be gathered near the base of the cliff at To'aga, immediately adjacent to the prehistoric occupation. Processing of such clay appears to have been minimal. The colluvial source(s) accounts for the greatest amount of pottery in the EDS/petrographic sample. One class of pottery, the red-slipped (thinware and one thick, red-slipped sherd), does not match the local clays as presently known. The red-slipped pottery in the EDS/petro-graphic sample also contains only sparse basaltic temper (group 2). These distinctions may suggest an exotic provenance for red-slipped ware that arrived on Of through inter-island exchange. This hypothesis requires a larger sample to test further. In sum, the available compositional evidence suggests at least four hypotheses that may be confirmed or falsified with additional research: 1) The decline in compositional groups (both temper and clay groups) merely reflects sample size differences for the time periods (cf. Grayson 1984, 1989; Kintigh 1989). This is, in a sense, the null hypothesis suggesting that with larger samples, the association of time period and clay or temper group will become more even (random in the statistical sense). 2) Local colluvial clay(s) provided a source for most pottery production at To'aga. Such clay(s) underwent little, if any, processing by potters. In most cases the colluvial clays could be described as "self-tempered." 3) Red-slipped pottery does not conform to the clay compositional data known for To'aga (clays and pottery). This ware, and others of similar clay-temper composition may be exotic to Ofu, and represent inter-island exchange. 4) The To'aga ceramic sequence is marked by a decline in the diversity of clays used in production (which in this case is not a product of sample size effects). This decline reflects change in the use or availability of the clay sources. Such a trend might also denote a decline in exchange, including that from other islands (see Hunt 1989; Kirch 1988, 1990). Reconstruction of production technology is supported by the macroscopic ceramic analysis outlined above. Vessel form (bowls) and the observation of some laminar fracturing in the sherds point to slab-building as the primary forming technique. The analysis of orientation angle was designed to provide evidence for primary forming (table 9.5). However, due to the absence of grains with a definitive long-axis, and perhaps the difficulty of determining "random," "indeterminate," or "preferred orientation," the vestige of possible slabbuilding is not reflected in attempts to analyze particle orientation. Secondary forming is indicated by paddle impressions (visible on 17% of the sherds) and anvil marks (present on 23% of the sheeds). Two sherds show the unmistakable impression of a finger used for the same purpose. The majority of the surface treatment is puddling (77%), a finishing technique using water and wiping to bring the finest clay particles to the surface of the paste. Wiping was also evident, occurring commonly on the rim sherds (tables 9.9-10). Approximately 25% of the sherds in the sample display a pattern of incomplete oxidation adjacent to the interior surface in contrast to 4% with incomplete oxidation at the exterior surface. Only 2% of the sherds show little or no oxidation present About 67% of the sherds are completely oxidized. Based on this, and other evidence described (hardness and comparison of experimentally fired-clay tiles), we suggest that pottery was fired in open conditions of temperatures reaching approximately 500-600( C. The fact that interior surfaces were darker (less oxidized), suggests that bowls were placed up-sidedown for firing. This technique is similar to some documented edtnographically in Melanesia, where pottery is still made in many locations (e.g., Irwin 1985; May and Tuckson 1982). Explaining the abandonment of pottery production in Samoa remains unresolved. Our To'aga analyses show that diversity of material use may Ceramics have declined over the period of ceramic production. It could be hypothesized that changes in raw materials, for example the use of "self-tempered" colluvial clays, resulted in a ceramic product of marginal quality. This hypothesis requires additional study (e.g. see Feathers 1990). Styk The To'aga assemblage is simple in form and carries very little decoration. Vessel parts present (direct rims and body sherds only) indicate tat only forms of unr cted orifice (bowls) were produced. There is no evidence in the To'aga assemblage of globular pots, jars, plates, or other complex vessel forms. Decorative attributes are restricted to impressing and notching on the lip, red-slip, carved paddle impression, and incision. Such a short roster departs dramatically from assemblages of comparable age from Mulifanua, 'Upolu, and from assemblages in Tonga and Fiji. Style can be defined for analytic purposes in archaeology as traits ta are free to vary independent of function (Dunnell 1978). This definition emphasizes style as governed by stochastic processes, and distributional frequencies that behave accordingly. In this perspective, thickness might be treated as a "stylistic" tait. Sherd thickness has received much attention in previous attempts to understand diachronic change in Samoan ceramics (Claik and Herdrich 1988; Green 1974; Hunt and Kirch 1988; Jennings and Holmer 1980, Kirch et al. 1990). The changing (declining) frequencies of thinware, in particular, might be a reflection of homology ("style"), and its independence from functional constraints. The frequency distributions (see tables 9.1-2) of thinware (defined as <7.5 mm) and thickware (>75 mm) from To'aga allow the following conclusions: 1) Thickware is present in the earliest deposits, and its abundance over time is relatively stable. 2) Thinware is never dominant in the assemblage but occurs in roughly equal percentages to thickware in the earliest deposits. 3) The presence of thinware declines in real and relative values over time but persists perhaps as long as pottery production itself. 4) Pottery declines in abundance early in the ChIstian era and then its production disappears 147 entirely. The evidence firm Western Samoa is similar in many, but not all, rspects. In spite of early dates for To'aga (i.e., contemporaneous with the Mulifanua Lapita site), no dentate-stamped Lapita pottery is known for this site, or elsewhere in Manu'a. This absence may be paralleled in the cases from Tikopia (Kirch and Yen 1982) and Anuta (Kirch and Rosendahl 1973) where assemblages of pottery date to early times, yet do not share the degree of decoration known elsewhere in the southwestern Pacific. Perhaps this reflects isolation from a larger intensland network that shared ideas of designs, or pots themselves. Manu'a may have simply been far enough away to incur such isolation from other islands of Samoa, Tonga, and Fiji. Green (1974) proposed a sequence of ceramic change for Samoa. His chronological analysis from Upolu was based on a short occupation sequence, with radiocarbon dates ranging from 1840 ± 100 B.P. (GaK-1441) in the lowest cultural layer (V) to 1800 ± 80 B.P. (GaK-1341) in the layer (IV) above (Green 1974:115). These dates overlap at one standard deviation. When calibrated and averaged together, these two dates yield a calibrated age range at one standard deviation of A.D. 1 17-254 (Stuiver and Reimer 1986). Based on his analysis of over 7400 sherds, Green (1974:130) concluded that, "thin and thick ware sherds occur in association in both layers" and that over time the end is for thickware to predominate but not totally replace thinware (1974:248). Inspection of Holmer's (1980:116) data, and his comparison to other Western Samoan assemblages, reveals a similar trend for thin- and thickware frequencies. Function Vessel function in the To'aga assemblage is suggested by fonn and the presence of residues. rhe single vessel form (bowls) might have served functions of storage, cooking, and serving dishes. Carbonaceous residues suggest cooking, at least in a small number of the vessels. Microanalysis for concentrations of phosphorus (P) was performed on three sherds in an experimental effort (Dunnell and Hunt 1990). These and other test case results were varied, and revealed that functional inferences based on P concentrations in pottery are unreliable 148 The Tolaga Site (Dunnell and Hunt 1990). The ceramics from To'aga are among the best studied in the Samoan Ilands. Addressing difficult issues beyond questions of chronology and cultural affinities demand detailed studies as we have attempted here. With regard to the questions posed at the start of this chapter, our study contributes to answers that will necessarily come from several studies ofcomparable detail and scope conducted with assemblages from throughout Samoa and the larger region. REFERENCES CITED Aldenderfer, M. S., and R. K. Blashfield 1984. Cluster Analysis. Beverly Hills: Sage Publications. Arnold, D. E., H. Neff, and R. L. Bishop 1991. Compositional analysis and "sources" of pottery: An ehnoarchaeological approach. American Anthropologist 93:70-90. Barry, B. A. 1978. Errors in Practical Measurement in Science, Engineering, and Technology. New York: John Wiley & Sons. Bishop, R. L., and H. Neff 1989. Compositional data analysis in archaeology. Archaeological Chemistry IV:57-86. Bishop, R L., V. Canouts, S. P. DeAtley, A. Qoyawayma, and C. W. Aikins 1988. The formation of ceramic analytic groups: Hopi pottery production and exchange, A.C. 13001600. Journal ofField Archaeology 15:317-37. Clark, J., and D. Herdrich 1988. The Eastern Tutuila archaeological project 1986, final report Prepared for the government of American Samoa, Office of Historic Preservation, Dept of Parks and Recreation. Pago Pago, American Samoa. Davis, J. C. 1986. Statistics and Data Analysis in Geology. Second Edition, New York: John Wiley & Sons. Dunnell, R C. 1978. Style and function: A fundamental dichotomy. American Antiquity 43:192- 202. 1983. Aspects of the spatial structure of the Mayo Site (15-JO- 14), Johnson County, Kentucky. IN R. C. Dunnell and D. K. Grayson, eds., Lulu Linear Punctated: Essays in Honor of George Irving Quimby, pp. 109- 165. Anthropological Papers No. 72, Museum of Anthropology, University of Michigan, Ann Arbor. Dunnell, R C., and T. L. Hunt 1990. Elemental composition and inference of ceramic vessel function. CurrentAnthropology 31(3):330-36. Feathers, J. K 1990. Explaining the evolution of prehistoric ceramics in southeastern Missouri. Ph.D. dissertation, University of Washington, Anm Arbor: University Microfilms. Goldstein, J. I., D. E. Newbury, P. Echlin, D. C. Joy, C. Fiori, and E. Lifshin 1981. Scanning Electron Microscopy and X-ray Microanalysis. New York: Plenum Press. Graves, M., T. L. Hunt, and D. Moore 1990. Ceramic production in Mariana Islands: Explaining change and diversity in prehistoric interaction and exchange. Asian Perspectives 29:21133. Grayson, D. K 1984. Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas. New York: Academic Press. 1989. Sample size and relative abundance in archaeological analysis: Illustrations from spiral fractures and seriation. IN R. D. Leonard and G. T. Jones, eds., pp.79-84,Quandfying Diversity in Archaeology. Cambridge: Cambridge University Press. Green, R. C. 1974. Excavation of the prehistoric occupation of SU-SA-3. IN R. C. Green and J. Davidson, eds., Archaeology in Western Samoa, Vol. II, pp. 108-154. Bulletin of the Auckland Institute and Museum No. 7. Gren, R. C., and J. Davidson, eds., 1969 Archaeology in Western Samoa, Vol. I. Bulleting of the Auckland Institute and Museum No. 6. Harbottle, G. 1976. Activation analysis in archaeology. Radiochemistry 3:33-72. Hill, H. E., J. Evans, and M. Card 1985. Organic residues on 3000-year-old potsherds from Natunuku, Fiji. New Zealand Journal of Archaeology 7:125-28. Holmer, R 1980. Samoan ceramic analysis. IN J. Jennings and R. Holmer, eds., Archaeological Excavations in Western Samoa, pp. 104-116. Pacific Anthrpological Records 32. Bernice P. Bishop Museum, Honolulu. Hunt, T. L. 1988. Lapita ceramic technological and compositional studies: A critical review. IN P. Ceramics V. Kirch and T. L. Hunt, eds., Archaeology of the Lapita Cultural Complex: A Critical Review. pp. 48-59. Thomas Burke Memorial Washington State Museum Research Report No. 5. Seattle. . 1989. Lapita Ceramic Exchange in the Mussau Islands, Papua New Guinea. Ph.D. dissertation, University of Washington, Ann Arbor: University Microfilms. Hunt, T. L., and P. V. Kirch 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal of the Polynesian Society 97(2):153-83. Irwin, G. J. 1985. The Emergence ofMailu. Terra Austalis, No. 10, Department of Prehistory, Research School of Pacific Studies. Canberra: Australian National University. Jennings, J., and R. Holmer 1980. Archaeological Excavations in Western Samoa. Pacific Andtpological Records 32. Bernice P. Bishop Museum, Honolulu. Kintigh, K. W. 1989. Sample size, significance, and measures of diversity. IN R D. Leonard and G. T. Jones, eds., Quantifying Diversity in Archaeology, pp. 25-36. Cambridge: Cambridge University Press. Kirch, P. V. 1988. Long-distance exchange and island colonization. Norwegian Archaeological Review 21:103-117. . 1990. Specialization and exchange in the Lapita Complex of Oceania (1600-500 B.C.). Asian Perspectives 29:117-33. Kirch, P. V., and P. Rosendahl 1973. Archaeological investigation of Anuta. IN D. E. Yen and J. Gordon, eds., Anuta: A Polynesian Outlier in the Solonon Islands, pp. 25-108. Pacific Anthropological Records No. 21 Honolulu: Bernice P. Bishop Museum 149 Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Bishop Museum Bulletin 238, Honolulu. Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler 1990. An ancestral Polynesian occupation site at To'aga, OfN Island, Ameuican Samoa. Archaeology in Oceania 25(1):1-15. May, P., and M. Tuckson 1982. The Traditional Pottery ofPapua New Guinea. Sydney: Bay Books. Neff, H., Bishop, R. L., and Arnold, D. E. 1988. Reconstructing ceramic production from ceramic compositional data: A Guatemalan example. Journal of Field Archaeology 15:33948. Norusis, M. J. 1986. SPSS/PC+for the IBM PC/XT/ AT. Chicago: SPSS Inc. Postek, M. T., K. S. Howard, A. H. Johnson, and K L. McMichael 1980. Scanning Electron Microscopy: A Student's Handbook. Ladd Research Industries, Inc. Rice, P. M. 1987. Pottery Analysis: A Sourcebook. Chicago: University of Chicago Press. Rye, 0. S. 1981. Pottery Technology: Principles and Reconstruction. Washington, D.C.: Taraxacum. Sokal, R. R., and H. A. Sneath 1963. Principles of Numerical Taxonomy. San Francisco: W. H. Freeman. Stuiver, M., and P. Reimer 1986. A computer program for radiocarbon age calibration. Radiocarbon 28:1022-1030. Topping, P. G., and A. B. MacKenzie 1988. A test of the use of neutron activation analysis for clay source characterization. Archaeometry 30:92101. 10 SAND TEMPER IN PREHISTORIC POTSHERDS FROM THE TO'AGA SITE W. R. D cKINSON TWENTY-NI RETATIVE sherds selected by T. L. Hunt from the collection of artifacts excavated at the To'aga site near the south coast of Ofu Island in the Manu'a Group of American Samoa were examined pergraically in thin section. As there is no current rason to suspect that any of the To'aga ceramic ware was made elsewhere, the purpose of the study was to provide baseline information about Manu'a temper sands. All the sherds examined contain volcanic sand as temper, aithough this basaltic detritus is mixed with calcareous grains derived from reef sources in some of the sherds. As would be predicted for Manu'a and other parts of Samoa, the volcanic sand is typical of the oceanic basalt tempers common to intra-oceanic Pacific archipelagoes (Dickinson and Shutler 1968, 1971, 1979). Several variants of temper sand are present in different sets of sherds, and available information is inadequate to pinpoint their ive sources. All could p ly have been collected on Ofu or nearby Olosega Island, but derivation of some from Ta'u in Manu'a or even elsewhere in Samoa is not precluded by the petrgric data. Their petrologic compatibility with Samoan lavas, however, and their overall resemblance to the spectrum of basaltic tempera studied to date from Samoa, makes importation from outside Samoa quite unlikely. As none of the temper variants are identical to tempers known from Tutuila or Upolu, all are regarded provisionally as indigenous To'aga temper, with the proviso that petrographic evidence alone cannot indicate how far afield ancient potters may have gone in their seach for suitable clay and temper within Manu'a. TO'AGA TEMPER VARIANTS The following variants ofbasaltic temper sand are all pr in varying numbers of To'aga sherds, and each is described in detail in subsequent passages: (a) Profise Basaltic Temper: Seven sherds (1-5, 7, 14) contain ferromagnesian basaltic sand so abundant that it forms 50-60 percent ofthe sherd bodies. The proportions of grain ypes in six (1-5, 7) of the sherds are statistically indiuis e, but the seventh (14) contains a related volcanic sand of slightly different composition. (b) Sparse Basaltic Temper: Twelve sherds (6, 10, 13, 16, 19-23, 27-29) contain sparse feremnagnesian basaltic sand of somewhat different composition and texture. The mineral and lithic grains form only 5-15 percent of the sherd bodies, but about a third (25-40 percent) of the temper grains used may have been fragments ofbroken pottery. (c) Feldpathic Basaltic Temper: Three sherds (15, 17, 18) contain feldspathic basaltic sand, which 152 The To'aga Site fonns a normal prport (30-40 percen) of the sherd bodies and is both mineralogically and texturt the two ferroma ally disinc from an variants of Toaga tper. Another sherd (9) contins a similar but sparser temper sa (A15 i percent of body) that appears to be a hybrid sand with a significant e of detritus from the kinds of bedrock sources that yielded the feromagesia basaltic sands. (d) Mixed Temper Sand: Seven sherds (8, 11, 12, 24-26, 30) contain mixed teper sands composed of both basaltic and reefdetitus in varying poportions. Fcrwmagnesian mineral grai and basaltic litc fragmens do the volcanic sand components of the mixed tempers, but their proportions are highly variable and plagioclase feldspar grains are al pesent in some sherds. temmed "slabwor') display blocky to prismatic subhedral pyroxene crystals iergrown with aggregated and multiply twined plagioclase crystals of slabmike aspect. Lithic fragments of intermediate grain size (here termed "latiwotk) display disorented mosaics oftwinned plagioclase laths with equant and largely anhedral pyroxene grains within their interstices. As all gradations are seemingly present between "slabworkW and 'lathwodc" and 'tachylitic" grains, the distinction made among hem is useful in a qualitatve sense only. Groundmass iron oxides in lithic fragmes range from equant or granularto skelet or elongate in form without apparent regard to other aspes of internal texture. TEMPER GRAIN TYPES The fenomagnesian volcanic sands in sherds with profuse basaltic temper are moderately sorted assemblages of subangular to subounded grains with a texture suggestive of alluvial orgin. Unmistable rounding of the edges of many gains indicates n rally occuming sand, rather than artificially cnmshed aggregate, and local ravine streams may have provided the sources of the temper. As would be expected for such a setting, lithic fragmens are generally but not uniformly larger than mineral grains. Abundance of subangular silty basalt detritus within the clayey paste in which the temper sand is imbedded suggests that potters collected naturally tempered sandy clay. This circumstance may account for the superabundance of temper sand in proportions higher than typically encountered in Pacific Island sherds. Proportions of grain types in most sherds containing this alluvial temper are quite consistent (table I0.1): half pyroxene, a quarter lithic fragments, a fifth olivine, and a trace of opaque iron oxides. All micrhenocrysts in lithic fragments are pyroxene and olivine. As the inteal texture of lithic fragments is somewhat variable, being ree-quarters 'lathwok" in two sherds (1, 7) and two-thirds "slabwok" in four others (2, 5), collecting sites were evidently closely related but not identical. Nevertheless, the average temper composition (table 10.1) for the six sherds in which proportions of grain types are essentially the same is taken here to be the best estimate, petrlogically speaking, of proportions of consituents for characteristic To'aga temper. The non-calcareous components of all the To'aga temper types are without exception composed of mineral grains and lithic fragments derived entirely from basaltic bedrock sources, eiter lavas or pyroclastic deposits, together mi some cases with fragments of broken pottery. The mineral grains, originally phenocrysts or micrphenocrysts in basalt, include clinopyroxene, olivine, opaque iron oxides (magnetite anid/or ilmenite), and plagioclase feldspar. The lithic fr s, represing aphanitic groundmass of basaltic lava or tephra, display a spectrum of internal textures reflecting an inherent range of constituent grain sizes. Microphanocrysts in lithic fragments include all the mineral species that were also p t in the temper sands as separate mineral gams Routine distinction between pyroxene and olivine in thin section was based upon key diagnostic features visible for each grin, and their identifications were checked by observations ofoptic axial angle and birefringence on suitably oriented grains. Pyroxene grains generally display faint green tints, and many show either cleavage or prismatic shapes. Untinted olivine grains are brighter in plane light, and many are altered along edges and fIracues to bright reddish iddingsite. Most basalt lithic fragments have an intergranuar internal texture, although the finest gained (here tenmed "tachylitic") ae intersertal with plagioclase microlites set in black basaltic glass (tachylite). The coarsest grained (here PROFUSE BASALTIC TEMPER Sand Temper in Prehistoric Potsherds 153 Table 10.1 Frequency Percentages of Ferromagnesian Mineral Grains' and Basaltic Volcanic Lithic Fragments (VRF) in Sherds Containing "Profuse Basaltic Temper" Sherd 1 2 3 4 5 7 Ave 14 n Py 01 Fe VRF Py(Py+01) 130 160 210 105 185 260 51 50 53 51 50 54 52 36 21 23 20 17 24 23 21 14 2 2 1 4 4 1 2 3 26 25 26 28 22 22 25 47 0.71 0.68 0.73 0.75 0.78 0.71 0.71 0.72 'Py, clinopyroxene; 01, olivine; Fe, opaque iron oxides Note: n=number of grains counted in each sherd and average (Ave) composition is calculated for sherds 1-5 plus 7 but not 14. A seventh sherd (14) contains distinctly more lithic volcanic sand (table 10.1), although its pyroxene/olivine ratio is very close similar to that of the other sherds. Lithic fragments, mostly "lathwork," are also more irregular in shape and some are microvesicular. Curved re-entrants on some lithic fragments and the presence of a few grains of microvesicular brownish basaltic glass suggest a pyroclastic component lacking in the more characteristic six sherds whose tempers were probably derived entirely from bedrock lava sources. SPARSE BASALTIC TEMPER The ferromagnesian volcanic sands in sherds with sparse basaltic temper are well sorted aggregates of subrounded grains with a texture suggestive of beach origin. The lack of finer grained grit within the clayey paste suggests that potters added artificial temper to clay bodies. Dark angular blotches within the clayey paste are probably ghosts of broken pottery fragments also added as pain of the tempering process. Although recognition of this grog constituent is equivocal in some sherds owing to indistinct outlines of the pottery fragments, its presence may account for the low overall proportion of volcanic sand, which amounts alone to much sparser temper than typically encountered in Pacific Island sherds. Frequency percentages of grain types are highly variable for different sherds (table 10.2), but so few grains are present in each sherd (average of only 20 per sherd in 10 of the sherds) that the statistical significance of individual counts is questionable. Consequently, all grains (250 total) were summed from all sherds counted to yield net frequency percentages, but net and average temper compositions are almost identical (table 10.2). The fact that both measures of bulk composition are similar to values for the single sherd (22) containing the most grains (60) gives confidence that either measure is a valid estimate of the overall temper composition. The sparse basaltic temper, probably beach sand, is less pyroxenic and more lithic than the average composition of the dominant alluvial variant of the profuse basaltic temper, but grain proportions closely resemble those in the more lithic variant of alluvial sand. In general, differences are not great enough to suggest wholly different provenance except for the contrast between stream and beach collecting sites. Proportons of lithic grain types are quite variable from sherd to sherd, but all three types are present in subequal amounts within the suite of twelve sherds as a whole. The net pyroxene/olivine ratio (0.63) is only slightly lower than in the alluvial sands (0.72), and may have been reduced marginally by preferential cleaving of pyroxene grains and winnowing of resulting cleavage fragments during prolonged reworking in a beach environment. 154 The To'aga Site Table 10.2 Frequency Percentages of Ferromagnesian Mineral Grains and Basaltic Volcanic Lithic Fragments (VRF) in Sherds Containing "Sparse Basaltic Temper" Sherd 6 10 16 19 20 21 22 23 27 28 29 Net Ave n Py 01 Fe VRF 20 10 15 30 10 10 60 25 30 20 20 250' 5 40 47 33 30 40 42 32 40 25 25 34 33 10 20 33 13 20 20 20 20 40 5 15 20 20 5 10 7 4 10 10 2 8 3 5 5 4 6 80 20 13 50 40 30 36 40 17 55 55 42 41 'Summation of n for 11 sherds listed. Note: n=number of grains counted in each sherd (note that net and average compositions are essentially the same). Sherd 13 too weatherd to allow accurate count. FELDSPATHIC BASALTIC TEMPER The volcanic sand in sherds (15, 17, 18) containing feldspathic basaltic temper essentially lacks fenrmagnesian mineral grains (table 10.3), and nearly all microphenocrysts in lithic fragments are plagioclase rather than pyroxene or olivine. One olivine grain is present, however, in one sherd (18), and one olivine microphenocryst is present in another (15). Lithic fragments are consistently larger than separate plagioclase mineral grains, although the two are present in about the same frequency (table 10.3), and the sand overall is only moderately sorted. Most lithic fragments are "tachylitic," many have smoothly curved re-entrants typical of tephra clasts, and some are microvesicular. The textural features of the sand jointly suggest that scoiaceous basaltic as, possibly reworked locally, was added as artificial temper to the clay body by potters having some selective aim in using such a tempering material The feldspathic basaltic temper shows no compositional overlap with the fenomagnesian basaltic tempers, but the geograhic separa- tion of their respective sources need not have been great. As if to underscore that point, one sherd (9) contains well-sorted and subrounded temper, probably a beach sand, that apparently represents a mixture of feldspathic and ferromagnesian volcanic sands. This anomalous sherd has an apparently sparse temper (- .15% of body) but also includes a few fragments ofbroken pottery as part of its overall temper component. MIXED TEMPER SAND The volcanic sands in sherds containing an admixture of reef-derived calcareous grains (15-75 percent) are highly variable in mineralogical composition (table 10.4). Although all are dominantly ferromagnesian volcanic sands, nearly half contain feldspathic components as well. Coupled with the presence of the calcareous grains, the good sorting and rounding of the sands is diagnostic of coastal origin on beaches where mixing of detritus from multiple sources is to be expected. Proportions of temper sand vary from 10-25 percent (8, 11, 12, 24) 155 Sand Temper in Prehistoric Potsherds Table 10.3 Frequency Percentages of Plagioclase Feldspar (P1) and Opaque Iron Oxide (Fe) Mineral Grains and Basaltic Volcanic Lithics Fragments (VRF) in Sherds Containing "Feldspathic Basaltic Temper" Sherd 15 17 18 Ave 9 n P1 Fe VRF 60 35 125 43 50 56 50 18 8 6 4 6 7 49 44 40 44 59 60 Py 01 - - - - - - - - 13 3 Note: n=number of grains counted in each sherd and average (Ave) composition is calculated for sherds 15 plus 17-18 but not 9 (Py and 01 are clinopyroxene and olivine mineral grains in sherd 9). Table 10.4 Frequency Percentages of Calcareous Grains (calc), Silicate-Oxide Mineral Grains, and Lithic Fragments for Sherds Containing "Mixed Temper Sand" Sherd 8 11 12 24 25 26 30 Calc n Py 01 Fe P1 VRF 13 22 20 15 44 18 77 65 35 12 30 70 160 60 32 14 33 27 21 22 39 32 7 17 23 17 18 14 3 7 25 - - 1 1 2 1 5 20 33 72 25 50 60 54 25 Note: n=number of non-calcareous grains counted in each sherd (percentages reported sum to 100 exclusive of calcareous grains) to 40-60 percent (25, 26, 30), the "tachylitic" variety of lithic fragments form about half the lithic popula- DISCUSSION: TEMPER COMPARISONS mon, and the overall pyroxene/olivine ratio ('0.60) is similar to that in the other sherds thought to contain beach sand temper. Computation of an average or net composition for the sherds containing mixed sand temper would be meaningless, given their inherent compositional variability, but their volcanic sands fit broadly within the spectrum of temper types present in other sherds. Sherds containing only sparse mixed temper sand also contain fragments of broken pottery in uncertain amounts. Although each of the To'aga temper types has clear distinguishing characteristics, compositional links argue that they form a related temper suite that is presumably indigenous to Manu'a. The two ferromagnesian basaltic tempers have contrasting textures tat reflect different sedimentological origins such as stream and beach sands, but the same grain types are present in both in slightly different proportons. The admixture of similar fenomagnesian constituents in one ofthe sherds containing 156 The To'aga Site feldspathic basaltic temper suggests hat bedrock sources for the fenomagnesian and feldspathic volcanic sands exist not far apart This inference is strengtned by the observation that mixed beach sands containing calcareous grains contain a varied spectrum of fenomagnesian and feldspathic constituents. Prehistoric sherds examined previously from Upolu (Dickinson 1969,1974,1976) contain generally similar basaltic temper sands composed of the same basic grain types, but none of the 'Upolu tempers is identical in texture or composition to the To'aga tempers. Fenromagnesian basaltic tempers from 'Upolu commonly contain a higher proportion oflithic fragments, typically have a higher ratio of olivine to pyroxene, and generally contain a subordinate proportion of brownish basaltic glass particles not present in To'aga tempers. Moreover, Upolu ferromagnesian tempers sherds with a contain well-sorted coastal sands texturally unlike the stream sands evident in well tempered To'aga sherds. Feldspathic basaltic tempers from 'Upolu are broadly similar to their To'aga analogs, but internal textures of lithic fragments differ in being coarser gained in the lJpolu sherds studied to date. Feldspathic trachytic tempers present in all available Upolu collections apparently have no counterparts at To'aga. The generic resemblance of all Samoan temper types examined to date pennits the strong inference that the tempers in To'aga sherds are indigenous to Samoa. On balance, there is no reason to suppose on petrographic evidence that any of the To'aga temper types were derived from sites elsewhere in Samoa. The fragments ofbroken pottery present in about half the To'aga sherds are not common constituents ofPacific Island wares, but do occur in sherds from the Ryukyu Islands, Palau, and the Nan Madol site on Ponape (Dickinson and Shutler 1979; Dickinson et al. 1990). Their presence at To'aga presumably reflects a common paucity of suitable local temper sand, rather than any close cultural relationship between the Caroline Islands region and Samoa. REFERENCES CITED Dickinson, W. R 1969. Temper sands in prehistoric potsherds from Vailele and Falefa. IN R. C. Green and J. M. Davidson, eds., Archaeology in Western Samo, Vol. I: pp. 271-73, Auckland Institute Museum Bulletin 6. . 1974. Temper sands in shends from Mulifanua and comparison with similar tempers at Vailele and Sasoa'a (Falefa). IN R. C. Green and J. M. Davidson, eds., Archaeology in Western Samoa, Vol. II: pp. 179-80, Auckland Institute Museum Bulletin 7. 1976. Mineralogy and petrology of sand tempers in sherds from the Feriy Beth site, Paradise site, and Jane's Camp. IN J. D. Jennings, R N. Holmer, J. C. Janetski, and H. L. Smith, eds., Excavatios on 'Upolu, Western Samo, pp. 99-103. Pacific Anthrpological Records 25. Honolulu: Bernice P. Bishop Museum. Dickinson, W. R, and R. Shutler, Jr. 1968. Insular sand tempers of prehistoric pottery from the southwest Pacific. IN I. Yawata and Y. H. Sinoto, eds., Prehistoric Culture in Oceania, pp. 29-37. Honolulu: Bishop Museum Press. 1971. Temper sands in prehistoric pottery of the Pacific islands. Archaeology and Physical Anthropology in Oceania 6:191-203. -.1979. Petrography of sand tempers in Pacific islands potsherds. Geological Society of America Bulletin 90 (Part I, summary):993-95; and Part II (microfiche), No. 1 1, Card 1, pp. 1644-1701. Dickinson, W. R. J. Takayamna, E. A. Snow, and R. Shuder, Jr. 1990. Sand temper of probable Fijian origin in prehistoric potsherds from Tuvalu. Antiquity 64:307-312. 11 NON-CERAMIC PORTABLE ARTIFACTS FROM TO'AGA PATRiCK V. KIRCH A soOM ThE SBsrANTiALquatities of pottery above in chapter 9, the To'aga site excavations yielded a small but typologically diverse assemblage of non-ceramic portable artifacts. Because the site's alkaine, calcareous depositional environent (particularly in the lowerlevels) favors the preservation of bone, shell, and sea urchin, a variety of artifacts made from these organic materials was recovered, in addition to objects of basalt and coral. This contasts with most early Samoan archaeological sites, such as Vailele or Sasoa'a on Upolu (Green and Davidson 1969, 1974), in which the acidic soils did not preserve a wide range of materials. Prior to our work at To'aga, only the Potsa and Falemoa sites on Manono Islet (Janetski 1980) had yielded a significant array of artifacts of shell, bone, and sea-urchin spine in association with Samoan ceramics. Thus, our knowledge of early Samoan material culture was largely restricted to basalt adzes, non-retouched lithics, and ceramics (Green 1974). This was in contrast to the situation with sites of comparable age in Tonga, where excavations on Tongatapu (Poulsen 1987), Niuatoputapu (Kirch 1988), and Ha'apai (Dye 1987) had produced a diverse array of material culture dating to the Ancestral Polynesian period. Hence, the To'aga artifact assemblage, described in full below, significantly expands our knowledge of the Samoan variants of Ancestral Polynesian material culture in the first millennium B.C. The non-ceramic artifacts from To'aga are described below according to broad functional classes in general use by Polynesian archaeologists. Comparisons are also made between the To'aga assemblage and other assemblages from Ancestral Polynesian period sites in Samoa and elsewhere in Westem Polynesia. STONE ADZES Six adzes which were either whole or sufficiently intact to be classified were excavated, primarily from pottery-bearing contexts. These adzes are classified according to the system devised by Green and Davidson (1969) for adzes from Westem Samoa. In addition, we recovered five small flakes with ground or polished surfaces, which appear to have been derived from adzes during use or bevel resharpening. Most of these diagnostic specimens were petro-chemically analyzed by M. Weisler in order to determine the range in quarry sources utilized. Weisler used the nondestructive XRF technique and presents the results of his study in chapter 12. From Layer HA-I in Unit 9 we excavated a finely ground and polished, complete adz of Samoan Type V (Green and Davidson 1969:24-26). The adz is of a very fine-grained, light grey basalt or andesite, and most of the original flaking scars have been removed by extensive polishing (fig. 11.1, c). 158 The Tolaga Site The bevel is curved, and the poll shows distinct battering, indicating use as a hammer while hafted. ile adz is 136.8 mm long, 54 mm wide, and 35.2 mm thick at the midpoint It weighs 422 g. Type V adzes are commonly associated with plainware ceramics both in Samoa (Green 1974) and in other early Western Polynesian contexts (Kirch 1988:192, 203). An incomplete section of another Type V adz, consisting of the bevel to the midsection, was found in Layer IIIB of Unit 20 (fig. 11.1, a). The adz is of a light-grey, fine-grained basalt The bevel is curved and very highly polished, while other paris exhibit remnant flake scars. The plano-convex section is rather high. The incomplete length of the adz is 78.4 mm; the width, 38.2 mm; and the thickness at midsection. 29.1 mm. Another partial adz ofType V, consisting of the butt to midsection, was found in Layer IIIB of Unit 23 (fig. I 1.1, b). Made of greyish basalt, the adz has a low (flattened) plano-convex cross section. In plan view, it also distinctly narrows toward the butt. The front and sides are partially ground and polished, but some flake scars remain. The midsection break displays considerable battering, indicating that the specimen was used as a hammerstone after breaking. The incomplete length is 68.1 mm; its width at the butt, 29.0 mm; the width at midsection, 53.3 mm; and the thickness at midsection, 27.3 mm. A small adz of fine-grained basalt was recovered from the disturbed landfill site at To'aga during the 1986 reconnaissance. The adz has a sub-triangular cross section, and thus would be classified as Type VI in the Green and Davidson (1969) system. However, it has been well ground on the front, removing the original flaked ridge (and thus rounding off the apex of the triangle). Hence, in some respects, the adz resembles a Type V. A rather battered remnant section of an adz, possibly of Type V or another type with a subquadrangular section, was excavated from Layer iIIA of Unit 27. This specimen is of dark grey basalt and has polished front and back surfaces. The butt is largely intact, but the artifact has been heavily battered from use as a hammerstone. The thickness is 25.4 mm, and the incomplete length, 75.4 mm. From Unit 3, in an aceramic depositional context we recovered the midsection of a partially ground, fine-grained basalt or andesite adz with trapezoidal cross section, probably of Samoan type IV (Green and Davidson 1969:24). The midsection is 24.2 mm thick, with the width ranging from 30.6 to 50.5 mm. Petrohemical analysis by non-destructive XRF, described further in chapter 12, suggests that this adz was manufactured at the large Tatagamatau quarry site on Tutuila Island (Best et aL 1989, 1992; Leach and Witter 1987, 1990). This is noteworthy, since most of the Manu'a adzes assignable to the Tatagamatau quarry were surface finds, also of trpezoidal sectioned types typical oflater Samoan prehistory. This adz from Unit 3 is associated with a 14C date of 1389-1287 cal B.P., Which indicates that adzes from the Tatagamatau quarry were being distributed as far as the Manu'a Group by at least the mid-first millennium A.D. In Layer HIlA of Unit 27 we excavated a flaked, tabular piece of daik gray basalt, extensively flaked, but retaining some cortex on one surface. The flake which measures 63.9 by 58.6 mm, and is 16.4 mm thick, may be a large decortication or trimming flake from adz manufacture. In addition to the large diagnostic specimens described above, we excavated five small flakes, each with one or more ground and polished facets. These are from Units 16, 17, 20, and 22 and all derive from adzes, either from use or resharpening. Four flakes were analyzed by XRF (see Weisler, chapter 12). Two of these can be ascribed to te Tatagamatau quarry site on Tutuila Island. SHELL ADZ A small adz of heavy shell, possibly Cassis sp., was found in association with plainware pottery at the landfill site during the 1986 reconnaissance. The adz is rectangular in shape with a slightly rounded bevel. Shell adzes are very me in Samoa and may have been restricted to the earlier ceramic period. Buck (1930:353-54) records only two shell adzes in the Bishop Museum collection from Samoa. HAMMERSTONES Two hammerstones, both from Layer IIA-1 of Unit 9 in the main trench, were excavated. One is an ovoid cobble of porphyritic igneous stone (with abundant feldspars). It is 30 mm thick, has a diameter of 93-105 mm, flat sides, and distinct Non-Ceramic Portable Artifacts 0 5 cm C Figure 11.1 Basalt adzes from the ToWaga site: a, bevel section from Unit 20, Layer IIIB; b, butt section from Unit 23, Layer hUB; c, complete Type V adz from Unit 9, Layer HA-I (drawings by J. Ogden). 159 160 The Tolaga Site pecking or damage along the margins. One face appears to be ground smooth, peitaps during use as an abrading or polishing stone. The second specimen is an elongate basalt cobble, beach-worn, with pecking damage on the broader end. The cobble measures 147 mm long (max. width 70 mm), and the damaged surface has an area of 17.1 by 26.9 mm. shell fragments. Thus, the total fishhook assemblage from To'aga now stands at twenty-eight whole or incomplete specimens, not including tabs and unfinished fragments. This is by far the largest assemblage of prehistoric fishing gear recovered from Samoa and is a major addition to our knowledge of early Polynesian fishing. The To'aga fishhook assemblage is remarkably uniform in size and morphology, with only minor variations. The hooks were all manufactured from the body whorls of Turbo setosus, a gastropod common on the reef edge of Ofu Island. The various midden deposits contained large quantities of T. setosus shell, some of which was probably manufacture debris (see Nagaoka, chapter 13). One worked fragment from Layer IIB of the 1987 trench, probably an unfinished hook tab, was of the larger and less commonly occurring species Turbo narmoratus. Examples of the hooks are illustrated in figures 11.2 and 11.3. They are small, andrather delicate, and were probably used to take smaller reef fish. The complete hooks have shank heights ranging from 13.1 to 30.4 mm. Hook widths range from 10.1 to >20.8 mm. Most hooks appear to have been FISHING GEAR Turbo-Shell Fishhooks Samoan archaeological sites have been notoriously poor in the preservation of bone or shell artifacts, and only a few specimens of fishing gear have ever been excavated (Green and Davidson 1969, p1. 23; Janetski 1980). The same has been true of other Western Polynesian sites in Tonga and Futuna (Kirch and Dye 1979). In our 1986 test excavation at To'aga, two fragments of small Turboshell one-piece fishhooks were recovered (Hunt and Kirch 1988:175, fig. 8, b-c). In 1987 the expanded excavations yielded four neatly complete hooks and fourteen hook fragments. In 1989, we recovered an additional eight hooks or hook fragments, and a large number of prepared tabs and unfinished Turbo 1. f--l - b a c A- d f e 9 0 L Figure 11.2 a 5cm .- a . Turbo-shell fishhooks from the To'aga site: a, roughed-out fishhook tab from Unit 21, Layer IIB; b, well-ground fishhook tab from Unit 23, Layer IIIB; c, ground and perforated tab from Unit 30, Layer II; d, unfinished fishhook from Unit 15, Layer U; e, head and shank fixxn Unit 27, Layer ILA; f, complete hook from Unit 23, Layer IIIB; and g, hook with sharply inturned shank and head from Unit 20, Layer IIIB (drawings by J. Ogden). Non-Ceramic Portable Artifacts 2: 3 4 161 5cm U] 11 -1 Figure 11.3 Turbo-shell fishhooks and fishhook fragments from the 1987 excavation at ToWaga. rotating in form, although one hook is technically of the jabbing variety. The bends have an 'O' or 'U' shape. Three hooks have a distinctive in-curved or "bent" shank, strongly reminiscent of some early Marquesan hooks (Suggs 1961:8 1, fig. 26). Two specimens have an inner shank knob, presumably to assist in line attachment. Several other shanks have small notches or grooves on the outer shank face, also for line attachment. Turbo-Shell Fishhook Tabs The manufacture of fishhooks from Turbo setosus shell is well attested in the To'aga site by the presence of numerous preforms or tabs roughed out of the body whorls of this gastropod as well as by worked shell fragments and many kinds of abrading tools (see below). Several examples of fishhook tabs are illustrated in figure 11.2. Of particular note is a specimen from Unit 23, Layer IIIB, which has been carefully shaped and fully ground on the exterior surface and around the margins (fig. 11.2, b). This tab measures 21.7 by 16.4 mm. Another specimen (fig. '1.2, c), from Layer II of Unit 30, represents yet a further stage in manufacture, with the entire center of the tab removed by drilling and filing. This specimen measures 19.8 by 13.4 mm. These tabs indicate that the reduction procedure for the manufacture of Turbo-shell hooks at To'aga was as follows: (1) a tab was first roughed out of the body whorl of Turbo; (2) this roughout was then ground flat on the exterior surface and carefully shaped by grinding around the margins; (3) the interior was then removed by drilling and filing; and (4) the gap between the shank and point was opened last by cutting and filing. Sinoto (1967:353, table 3) remarks that "simple drilling" and "chipping and filing" were the methods used by early Marquesans in hook manufacture. Thus, not only the forns of the To'aga hooks, but thle specific manufacture methods, are consistent with the Marquesan hooks for which the To'aga specimens may have been 162 The To'aga Site prototypes. During the 1987 excavations we did not make a special effort to distinguish Turbo-shell tabs from Turbo-shell midden or worked debris, and no exact counts are therefore available. For the 1989 materials, however, shaped tabs were carefully separated from the shell midden during the laboratory study of faunal materials by L. Nagaoka. The following counts by unit indicate the frequency of such prepared tabs: 2 tabs Unit 15: 1 tab Unit 16: 7 tabs Unit 20: 11 tabs Unit21: Unit 23: 9 tabs Unit 30: 1 tab Cypraea-Shell Caps The caps or dorsa of large Cypraea shells (especially C. tigris) comprised one component of the Samoan octopus lure. Buck (1930:434-38, fig. 257, pl. XLI, B) describes and illustrates this apparatus. Three such dorsa were recovered together in Layer III of Unit 28 and were possibly part of such an octopus lure rig. ABRADING TOOLS Coral Abrader From Layer III of Unit 11 we recovered a tabular shaped abrader of Porites sp. coral. The abraded facet has a surface area measuring 50 x 60 mm. Echinoid-Spine Abraders The long spines of the slate-pencil sea-urchin (Heterocentrotus mammillatus) have a natural abrasiveness and thus were used throughout most of Polynesia to manufacture fishhooks and other objects of shell and bone. Two such abraders were excavated from Layer IIA-1 in the main trench. Both have distally abraded facets at an angle to the longitudinal axis, as do the abraders reported by Janetski (1980, fig. 43, g-i) from the early Falemoa site in Western Samoa. A complete spine which has been slightly faceted at the distal tip was excavated in Layer II of Unit 15. From Layer IIIC in Unit 23 we recovered a sea urchin spine abrader which had been distally abraded to a point (circular section), presumably from use as a drill in the manufacture of Turbo-shell fishhooks (fig. 11.4, a). The tip only of a circularly abraded echinoid spine was also found in Layer IIB of Unit 20. A particularly interesting echinoid abrader was found in Layer IIB of Unit 28, and is illustrated in figure 1 1.4, j. This spine, 73.2 mm long, has been equally reduced on two sides from the distal end to form a thin, saw-like blade. It would appear that this blade edge was purposefully produced in order to cut shell or bone objects. Also from Unit 28 (Layer III) was a small fragment of sea urchin spine which was abraded laterally to form a flat surface. All of these abraders were likely used to manufacture the Turbo hooks and other artifacts of shell. Shell-Bead Abrader An abrader of Porites coral, specifically adapted for grinding small Conus-shell beads, was recovered from Layer IIIB of Unit 23 (fig. 11.4, h). The abrader consists of a naturally waterworn coral pebble (68.6 by 46.4 mm and 19 mm thick) which has been flattened on one face by grinding. In the center of this face is a single depression or "cupule" with a diameter of 9.1 mm, about 1-2 mm deep. This depression has a central "nipple" which results from positioning a Conus-shell spire in the depression, and then using the abrader to grind the shell against a larger grindstone. Such specialized Conusshell bead abraders had been reported from Vanuatu (Garanger 1972) and from Vanikoro in the Santa Cruz Islands of eastern Melanesia (Kirch 1983:102104, fig. 16), but were previously unknown from Western Polynesia. Recently, however, Sand (pers. comm.) excavated such an abrader from the Asipani Lapita site on Futuna Island. ORNAMENTS Conus-Shell Beads From Layer IIB in the main trench are two delicate beads of Conus sp., very well ground, with diameters of 5.6 and 5.9 mm, and thicknesses of 1.9 and 2.1 mm (fig. 11.5). A slightly larger bead or ring of Conus, complete and very well ground (fig. 11.4, g), was found in Layer IIIC of Unit 20. This has an extemal diameter of 15.8 mm and is 2.4mm thick. Conus-Shell Rings Layer IIB in the 1987 main trench produced two fragments of larger Conus sp. rings, very well ground, with original diameters of about 50 mm (fig. Non-Ceramic Portabk Artifacts Ia 163 f e cK (d :a -M 9 I b -10-11 I11 I I 1 i 0 Figure 11.4 5cm Miscellaneus artifacts from the To'aga site: a, echninoid spine abrader from Unit 23, Layer HIC; b, bone point from Unit 27, Layer IIIB; c, Conus-shell ring fragment from Unit 29, Layer ITIB; d, Conusshell ring fragment from Unit 29, layer IUB; e, drilled shark's tooth from Unit 21, Layer HI; f, Conusshell bead from Unit 30, Layer II; g, Conus-shell bead from Unit 20, Layer IIIC; h, coral abrading stone for grinding shell beads, from Unit 23, Layer IB; i, unfinishedTridacna-shell ring from Unit 23, Layer IIIC; and j, echinoid-spine abrader from Unit 28, Layer UB (drawings by J. Ogden). 11.5). One fragment had been sharpened to a point after breaking. Layer II in Unit I I produced a fragment of a large shell ring or anmband, made either of a large species of Conms, or possibly of Tridacna. The ring fragment is 7.3 by 11.6 mm in thicness, and has a reconstructed diameter of about 70 mm (fig. 11.5). A similar annband fragment from the Falemoa site is illustrated by Janetski (1980: fig. 45, b). From Layer Im of Unit 16 we recovered a fragment of a Conus-shell ring with a cross section measuring 3.5 by 4.8 mm, and a reconstructed diameter of about 35 mm. Layer rIB of Unit 29 produced another Conus-shell ring fragment (fig. 1 1.4, d) with a roughly rectangular 164 The To'aga Site cross section (5.3 by 4.8 mm), which would have had an original diameter of about 30 mm. Based on Buck's extensive compilation of Samoan material culture (1930), Conus-shell rings were not a part of the Samoan ornamental repertoire in historic times. Indeed, they were probably associated only with the early ceramic period. Bone Point From Layer IIIB of Unit 27 was recovered a facet bone "point" of unknown function (fig. 11.4, b). The bone, of either dog or pig, has been carefully faceted to a chisel-like tip, across which were abraded a series of fine grooves. WORKED SHELL Ufinished Tridacna-Shell Ring Approximately one-half of a Tridacna-shell ring which broke during the process of manufacture was found in Layer IIIC of Unit 23 (fig. 11.4, i). The Tridacna valve incorporates part of the hinge. It was worked by chipping and pecking to create a central perforation. Presumably the artifact broke during this chipping process, prior to the initiation of grinding. The specimen has an outer diameter of 73 mm, and the central perforation is 15 mm in diameter. Nerita-Shell Beads From Layer IIB in the 1987 main trench were two Nerita sp. shells with artificial perforations in the basal whorl, perhaps for stringing as beads. A third specimen was found in Unit 21. Buck (1930:638) mentions the use of sea shells as beads but does not illustrate examples or provide further details. Gastropod Bead A small gastropod (species unknown) from Layer II of Unit 30 has had both the spire and basal whorl removed by grinding (fig. 11.4, f), leaving only the midsection of the shell as a bead. It has a diameter of 14.0 mm. Echinoid-Spine Bead In Layer IIIC of Unit 20 we found a unique bead made from a section ofHeterocentrotus mammilat spine which was double-drilled to form a central perforation. The bead measures 13.5 mm in diameter and is 10.5 mm thick. MISCELLANEOUS ARTIFACTS Drilled Shark's Tooth A small shark's tooth (14.7 mm high) was found in Layer IIB of Unit 21 (fig. 11.4, e). This has been drilled (hole diameter 2.6 mm), presumably in order to lash the tooth to a handle. A large piece of Tridacna shell (possibly from T. gigas) which has been chipped around the edges to a roughly rectangular shape (measuring 170 by 135 mm) was found in Layer IIIB of Unit 29. This may have been intended as a Tridacna adz preform or may have been for the manufacture of some other object, such as a shell ring. From Layer III of Unit 28 we recovered two matching pieces of worked Conus shell. These consist of part of the main body whorl, with a cut and beveled edge near the spire. These are presumably rejected material resulting from the removal of a large Conus spire, as a part of the manufacture process for Conus-shell rings. The chipped basal whorl section of a species of Trochus or Textus shell was found in Unit 21. This may have been intended to be a ring or armband. Various small pieces of worked shell were recovered throughout the excavations. Most of these are of Turbo spp. and relate to fishhook manufacture. In the 1987 excavated material these were not distinguished from the Turbo-shell midden. In 1989, however, all worked Turbo shell was carefully segregated during faunal analysis by L. Nagaoka, yielding the following frequencies by excavation unit: 2 specimen(s) Unit 15 5 Unit 16 1 Unit 17 14 Unit 20 3 Unit 21 1 Unit 22 9 Unit 23 2 Unit 25 1 Unit 26 5 Unit 28 1 Unit 30 Five specimens of worked pearl shell (Pinctada sp.), were also recovered from Units I 1 22, 23, 29, and 30. A triangular-shaped specimen from Layer Non-Ceramic Portable Artifacts Figure 11.5 165 Miscellaneous artifacts from the To'aga site: left to right, Conus-shell ring fragments, Conus-shell beads, and echinoid-spine abraders. 111B of Unit 23 is of special interest because it shows distinct filing or cutting marks on all three margins. This piece, measuring 24.9 by 16.9 mm, is probably detritus from the manufacture of some other object, rather than a preform. UNRETOUCHED LITHICS Basalt Flakes Flakes of basalt were surprisingly uncommon in the Totaga excavations. During the 1989 excavations, when particular attention was paid toward the recovery of such lithics during screening, only eighteen flakes were noted. Eleven of these are from Unit 23 [Layer III], suggesting that this may have been a locus of basalt flaking activity. The other flakes are from Unit 28 (three flakes) and Unit 29 (four flakes). Most of these are rather small and could derive from adz use, although they do not show polished surfaces. Obsidian Flakes A number of very small flakes of an opaque, black, low-silica volcanic glass or obsidian were found from various excavation contexts. Most of these are less than 5 mm in size. As the dike complex of Leolo Ridge overlooking the To'aga site has many glassy chills along the dike margins (see chapter 2), it is most probable that these "flakes" are natural and simply derive from the talus rockfall above the site. One small core from Layer IIIC of Unit 23 is completely different, however, from the other obsidian specimens. This is of a reddish-brown color, with black spots and banding. The "core" measures 12.2 by 13.3 mm. Its geological provenance is unknown, but it is very likely an import to Ofu. REFERENCES CITED Best, S., H. Leach, and D. Witter 1989. Report on 166 The To'aga Site the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August 1988. Department of Antrpology, University of Otago, Dunedin. Bests S., P. Sheppard, R. Green, and R. Parker 1992. Necromancing the stone: Archaeologists and adzes in Samoa. Journal of the Polynesian Society 101:45-85. Buck, P. H. [Te Rangi Hiroa] 1930. Samoan Material Culture. Bernice P. Bishop Museum Bulletin 75. Honolulu. Dye, T. S. 1987. Social and cultural change in the prehistory of the ancestral Polynesian homeland. Unpublished Ph.D. dissertation, Yale University, New Haven. Emory, K. P., W. J. Bonk, and Y. H. Sinoto 1959. Hawaiian Archaeology: Fishhooks. Bernice P. Bishop Museum Special Publication No. 47. Honolulu: Bishop Museum Press. Garanger, J. 1972. Arch!ologie des NouvellesHebrides. Publications de la Sociat des Ocdanistes No. 30. Paris. Green, R. C. 1974. A review of portable artifacts from Westem Samoa. IN R. C. Green and J. Davidson, eds., Archaeology in Western Samoa, Vol. II, pp. 108-154. Auckland Institute and Museum Bulletin 7. Green, R. C., and J. Davidson 1969. Descrpton and classification of Samoan adzes. IN R. C. Green and J. Davidson, eds., Archaeology in Western Samoa, Vol. I, pp. 21-32. Auckland Institute and Museum Bulletin 6. Green, R. C., and J. Davidson, eds. 1969. Archaeology in Western Samoa, Vol. 1. Bulletin of the Auckland Institute and Museum. . 1974. Archaeology in Western Samoa, Vol. II. Bulletin of the Auckland Institute and Museum. Janetski, J. 1980. Shell, bone, coral, and urchin spine artifacts. IN J. D. Jennings and R. N. Holmer, eds., Archaeological Excavations in Western Samoa, pp. 123-31. Pacific Andhpological Records 32. Honolulu: Bishop Museum. Kirch, P. V. 1983. An archaeological exploration of Vanikoro, Santa Cruz Islands, Eastern Melanesia. New Zealand Journal ofArchaeology 5:69-113. -. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memonal Washington State Museum Monograph No. 5. Seattle. Kirch, P. V., and T. Dye 1979. Ethnoarchaeology and the development of Polynesian fishing strategies. Journal of the Polynesian Society 88:53-76. Leach, H., and D. C. Witter 1987. Tataga Matau rediscovered. New Zealand Journal of Archaeology 9:33-54. . 1990. Further investigations at the Tatagamatau site, American Samoa. New Zeaad Journal ofArchaeology 12:51-83. Poulsen, J. 1987. Early Tongan Prehistory: The Lapita Period on Tongatapu and Its Relationships. 2 vols. Terra Australis 12. Canberra: Department of Prehistory, Australian National University. Sinoto, Y. H. 1967. Artifacts from excavated sites in the Hawaiian, Marquesas, and Society Islands: A comparative study. IN G. A. Highland et al., eds., Polynesian Culture History: Essays in Honor ofKenneth P. Emory, pp. 341-62. Bernice P. Bishop Museum Special Publication No. 56. Honolulu: Bishop Museum Press. Suggs, R C. 1961. The Archaeology ofNuku Hiva, Marquesas Islands, French Polynesia. Anthropological Paprs of the American Museum of Natural Hisory 49(1). New York 12 CHEMICAL CHARACTERIZATION AND PROVENANCE OF MANU'A ADZ MATERIAL USING A NON-DESTRUCTIVE X-RAY FLUORESCENCE TECHNIQUE MARSHALL WEISLER INTRODUCTION THE COLONIZATION SiRATEGIEs employed during the settlement of Polynesia and the subsequent diversification of island societies are key issues in Oceanic prehistory. Yet it is only recently that lsearchers have begun to amass empirical evidence of inter-island communication-throughout island sequences -that undoubtedly influenced the direcdon and tempo of post-settlement island histories. I refer here to recent finds of Tongan pottery in the Cook Islands (Walter and Dickinson 1989:465), of Samoan adz material in Fiji (Best 1984; Best et al. 1992) and the Cook Islands (Weisler 1993), and of inter-island transport of volcanic glass and finegrained basalt in Hawaii during late prehistory (Weisler 1990; et al. In prep. [For a review of Polynesian basalt adz provenance studies, see Weisler In press a.]). Isolation has been invoked as '"the most fundamental" mechanism of divergence in Polynesia (Kirch and Green 1987:440), but "how . . . different extremes of isolation have influenced the evolution of human diversity from island to islandand perhaps even on the same island" (Terrell 1986:122-23) remains a problem that must be addressed for each island sequence. Dissimilarities may also originate because of continuing contact (see Terrell 1986:147). Isolation may have contrib- uted to regional variations in portable artifacts, architecture, and language in New Zealand (Prickett 1982), Society Islands (Emory 1933), and Hawaii (Kirch 1985, 1990). In contrast, inter-island communication is inferred by parallel histories in subsistence and technological transformations in the northemrn and southemrn Marquesas (Rolett 1990:363) and similar changes in ceramic styles and adz evolution in the Samoa-Tonga-Fiji area at similar times imply continuous inter-island contact (Davidson 1977, 1978, 1979). Tracking the spread and subsequent communication of prehistoric Polynesian societies has been difficult without the widespread occurrence of pottery and obsidian, artifact classes that have proven especially useful for demonstrating interaction in the southwest Pacific. Without inferring contact from similar styles of fishhooks, adzes, and architecture, Polynesian archaeologists are left with few items to track empirically the intra- and interisland movement of things. Connecting stylistic nodes, however, may only reflect convergent evolution (Kirch and Green 1987). I may risk being accused of taking a hard-line empiricist view of culture contact and change, but I think it crucial to confront the problem withseparating stylistic from functional dimensions (Dunnell 1971:26-30) in relation to documenting the movement of exotic raw 168 The To'aga Site materials and finished objects (e.g., peairlshell and industrial stone) between and within island groups. The spatial and temporal dimensions of contact spheres (Irwin 1990:92; Walter 1990; Weisler In press b), can be delimited more accurately using exotic materials than by connecting artifact styles. Consequently, the distribution of exotic materials provides a better framework for evaluating the role of isolation and communication in shaping historical developments of island societies. This preliminary effort to document inter-island communication during the prehistory of the Manu'a Islands should take into account the region's geography. These islands are situated 100 km east of the most important source of adz material in central Polynesia, the Tatagamatau adz quarry, Tutuila (Leach and Witter 1987, 1990). The distance and often turbulent ocean conditions between Tutuila and the Manu'a Group may have reduced the frequency of contacts (Hunt and Kirch 1988:155). THE COLLECTIONS Two questions are addressed by XRF analysis of the Ta'u and Ofu island assemblages and source material: (1) Did any artifacts originate at the Tatagamatau adz quarry complex on Tutuila Island located 100 km to the west?; and (2) What was the geochemical variation of adz material and unmodified basalt flakes from Manu'a sites? Although it seemed unlikely that I would be able to assign a provenance to all artifacts, I could determine if specific geochemical groups were correlated with certain tool classes. Interaction spheres could be delineated by identifying similar rock at different sites and, perhaps, different temporal periods. An additional goal of the XRF analysis was to generate geochemical data for the important Tatagamatau quarry complex on Tutuila and two local To'aga sources that may have been used. The Archaeological Sample The archaeological contexts of the Ta'u and Ofu island assemblages are described elsewhere (Hunt and Kirch 1988; Kirch et al. 1990; chapters 5 and 11, this volume) and are not reiterated here. The artifacts selected for compositional analysis consisted of all whole or fragmentary adzes, all flakes exhibiting polished surfaces and assumed to be adz fragments, and a sample of unmodified basalt flakes which represented the macroscopic vanriability present. Table 12.1 presents characteristics of the analyzed artifacts. Specimens ranged from 1.2 to 421.9 gmins and 22.08 to 136.30 mm in length. The thinnest sample was 3.53 mm. Source Material During their 1987 field season, Kirch and Hunt collected nine samples of source rock from the Tatagamatau quarry complex on Tutuila (Best et al. 1989; Leach and Witter 1987). On Ofu Island, they collected four samples from dike swarms at Mako Ridge and three samples from Fa'ala'aga. The Tatagamatau quarry complex consists of numerous cut-and-fill terraces, stone-working areas, pits, and possible fortifications along several ridges and scree slopes which together may cover more than 110 acres (Best et al. 1989; Leach and Witter 1987, 1990). The raw material derives from a dike complex at the foot of a waterfall at the head of a small gulch (P. Kirch, personal communication 1990, Leach and Witter 1987:39). Fine-grained dike rock may have been removed from dikes but most likely was prised from the soil and rock scree of slopes and from streambeds. Although I have not visited Tatagamatau, the geological setting appears to be similar to the Waiahole quarry on windward O'ahu (Dye et al. 1985), albeit on a much smaller scale. Two collections of source rock were made by Kirch and Hunt during a visit on July 3, 1989. Eight samples, consisting of large primary or secondary flakes with obvious bulbs of percussion, were collected at three locations along the main spur of the complex which trends NE-SW, and an additional flake was retrieved from Area 3 (figure 12.1). The exterior color of the source rocks varied from dark brown (1OYR 3/3), dark gray (10YR 4/1), to dark grayish-brown (2.5Y 4/2) due to weathering and contact with lateritic soil. Fresh saw-cut surfaces are mostly dark bluish-gray (Gley card, SB 4/1) grading to dark gray (7.5R 4/0 and 2.5Y 4/0). Even prior to geochemical analysis, it was not likely that nine samples would document the total variability of quarry rock from so large a quarry, but this collection did contribute significantly towards developing a strategy for collecting additional material. Chemical Characterization of Manu'a Adz Material Table 12.1 Basalt Artifacts Analyzed by EDXRF Lab Number 89-22* 89-25 89-26* 89-31 89-30* 89-28 90-30* 90-33 89-18 89-34* 89-33* 89-32 89-19 89-17* 90-22 90-23 90-42 90-46 9045 90-38 90-27* 90-37 90-41 90-34 90-48 90-29 90-36 9040 9043 9049 9044 90-25 90-24* 90-26 90-31 90-39 90-47 90-35 * = Artifact Number Weight (g) Fiti'uta 1 11-1-S-4 11-2-S- 1 11-2-S-2 11-2-S-3 11-2-S-8 11-2-9 11-52-S-9 13-S-3 13-S-5 13-S-6 13-S-9 13-1-S-36 13-2-3-1-1 13-1-9-2-4 13-1-13-87 13-1-16-1 13-1-16-11 13-1-16-11-81 13-1-16-12-83 13-1-20-4-118 13-1-20-5-121 13-1-20-5-122 13-1-20-5-123 13-1-20-6-139 13-1-20-6-140 13-1-20-6-141 13-1-20-7-143 13-1-20-9-147 13-1-21-7-151 13-1-22-2-171 13-1-22-2-173 13-1-22-2-174 13-1-23-7-194 13-1-27-5-39 13-1-27-5-40 13-1-27-5-41 13-1-27-6-42 Probable source is Tatagamatau. 24.8 70.9 41.4 35.3 21.4 33.1 60.9 133.1 48.4 44.7 33.8 44.4 63.9 63.4 421.9 0.7 5.5 1.5 127.9 20.5 3.4 1.2 4.4 11.3 11.4 128.4 56.7 3.2 20.6 59.3 3.7 2.1 0.4 211.4 89.4 128.1 2.1 9.5 Description Quadrangular, reworked adz Quadrangular adz Quadrangular adz, front section Quadrangular adz, butt section Quadrangular adz, front section Quadrangular adz, front section Quadrangular adz, butt section Quadrangular adz, front section Quadrangular adz, front section Quadranglulkar (?) adz, front (?) section Quadrangular adz, front section Quadrangular adz Plano-convex adz Quadrangular adz, midsection Plano-convex adz Adz flake Adz fragment Unmodified flake Retouched flake Retouched flake Adz flake Unmodified flake Unmodified flake JUnmnodified flake Unmodified flake Plano-convex adz, front section Retouched flake Unmodified flake Unmodified flake Used flake Unmodified flake Adz flake Adz flake Plano-convex (?) adz Quadrangular (?) adz Quadrangular (?) adz, butt (?) section Unmodified flake Unmodified flake 169 170 The To'aga Site Situated at about 350 m elevation and exposed by a recent road-cut, the Mako Ridge dike swarm on Ofu consists of relatively dense basalt. Weathering has smoothed the normally angular, loose dike rock into sub-rounded cobbles which lie buried within a thick soil matrix. It is doubtful that this material was used for making stone tools, but four samples were collected for chemical analysis. The weathered exterior surface was gray in color (7.5YR 5/0) and fresh breaks, dark gray (2.5Y 4/0). A large concentration of dikes is found along a steep ridge towards the east end of Ofu. The Fa'ala'aga dike swarm (Stice and McCoy 1968) may have been a more important source of local, medium-grained basalt during prehistory. Angular rock from the eroding dikes form a large scree slope which descends to the To'aga coastal zone where stone is readily available (see Kirch, chapter 2). Three fresh samples were removed from dikes exposed by a modem road-cut (see fig. 12.2) to provide the general chemical composition of the dike swarm although many dozens of dikes may exist each with varying chemical compositions. Surface color of these rock samples is very dark gray (2.5Y 3/0) with fresh breaks being gray (2.5Y 5/0). METHODS Because this specific technique of non-destructive, energy-dispersive XRF had not been used previously with adz material from Oceania, each individual source sample was divided and analyzed by two techniques: destructive XRF using pelletized samples and non-destructive XRF using whole specimens. Destructive XRF analyzes a wider range of elements, and results are fully quantitative for most elements and comparable to other data sets now being developed for adz source rock in Polynesia The non-destructive technique focused on those elements that are more reliably detected with this procedure, and results presented are considered here as semi-quantitative. Destructive XRF Source samples from Tatagamatau, Mako Ridge, and Fa'ala'aga were analyzed by Dr. Peter R. Hooper at the Department of Geology, Washington State University, on 16 and 17 October 1989. The equipment and operating procedure are paraphrased from Hooper and Johnson (1987; this manuscript is available from Dr. Hooper or myself). Twentyseven elements were analyzed on an automatic Rigaku 3370 spectrometer. For the oxides, SiO2, A1203, TiO2, FeO*, MnO, CaO, MgO, K20, Na20, and P205, the weight percents are presented. For the other seventeen elements, the weight percent or ppm (parts-per-million) of the element are provided. All elements are analyzed on a single 2:1 lithium tetraborate:rock powder fused disk. Each element analysis is fully corrected for line interference and matrix effects of all other analyzed elements. The results are normalized and printed out with total iron expressed as FeO*. Whole rock samples were reduced with a tungsten carbide jaw crusher to small chips generally about 1 cm in average size. Chips were hand picked and reduced further in a Tema swingmill shatterbox with tungsten carbide surfaces where they were milled for two minutes. Rock powder weighing 3.5 grams was mixed with pure lithium tetraborate and emptied into a 34.9 mm-intemal-diameter, graphite crucible. The samples were fused in a muffle furnace at 1000'C. After grinding, the samples were fused a second time. The lower, flat surfaces of the fused disks were ground on coarse (240) grit and fine (600) grit. Then, they were washed in an ultrasonic cleaner, dried, and loaded into the XRF spectrometer. The x-ray intensity of each element from unknown samples was measured and compared to values of two beads from each of eight international rock standards (U.S. Geological Survey [U.S.G.S.] standards PCC- 1, BCR-1, BIR-1, DNC- 1, W-2, AGV-1, GSP-1, and G-2). Standards were run and recalibrated after processing between one and two hundred unknown samples. U.S.G.S. BCRP-84 and GSP-1 were used as internal standards and routinely run after every twenty-eight samples to check instrument performance. Drift between standard calibrations was due almost entirely to iron which produced slightly higher values for other elements presented as oxides, weight %. When thirty analyses of the U.S.G.S. standard BCRP were compared to accepted values, the precision or standard deviation of oxide values to known measures is 0.22% ppm. Values reported for thirteen trace elements in ppm were below 6% ppm at one standard deviation, except barium (16.2%) and vanadium (10.0%o). MThe highest precision was achieved with Rb, Sr, Zr, Y, Chemical Characterization of Manu'a Adz Material Figure 12.1 171 Map of Tatagamatau quarry showing locations of source samples (after Best et al. 1989, fig. 21). Oa,Cu, and Zn. Ni, Cr, Sc, V, and Ba should be xgarded as semi-quantitative below the 30 ppm kvel. Rb, Sr, Zr, Nb, and Y have satisfactory Non-destructive Energy Dispersive XRF precisions and accuracies down to one to three ppm, while Y and Nb could be measured to 0.1 ppm. Evaluation of accuracy suggested that variation between different samples of standard powder or onhomogeneity resulting from sample preparation is greater than inaccuracies caused by inadequate Portions or splits of all source samples from Tatagamatau, Mako Ridge, and Fa'ala'aga were analyzed by Weisler along with thirty-eight artifacts during four runs in late 1989 and early 1990. Sample preparation consisted of submerging specimens matrix and interference correction. 172 The To'aga Site 70T Mako (x) 0 Tatagamata[ 0 1" Co X A Tatagamatau 50 subsource L. Mako (x) A 60 s s A Fa'ala'aga (x) A Fa'ala'aga (X) =ubsource 40 N 40 40 50 50 60 60 70 70 80 80 9 0 90 100 10 110 110 120 120 130 Nb/Sr X 1000 Figure 122 A compaison of source rocks analyzed by non-destuctive x-ray flourescence using whole specimans and XRF with fused disks (indicated by italic type). in a sonic bath of distilled water for up to one hour then air drying. Some artifacts, deeply stained with lateritic soil, were scrubbed with brushes as well. Carbonate encrustations on a few artifacts were removed with a 10% solution of HCI and then rinsed in distilled water. The EDXRF instrument is limited to the maximum weight and size of specimen that can be analyzed. Although the opening of the sample holder where the x-ray beam is directed through to the specimen is 32 mm in diameter, careful placement of the artifact or rock in the EDXRF instrument may accommodate samples (as in this analysis) up to 421 gmns and 136 mm long. Molding clay can be used to secure specimens on the tray which can hold up to twenty samples. Each artifact was carefully examined to locate the flattest surface for analysis that could be accommodated within the space parameters of the sample chamber (see illustration in Bouey 1991:fig. 5). Source samples were cut to size and only fresh saw-cut surfaces were analyzed. Laboratory facilities and equipment were provided by the Department of Geology and Geophysics, University of Califomrnia, Berkeley. The same equipment and nearly identical operating conditions were used as reported by Hughes (1986:esp. 25-30) in his comprehensive study of California and Oregon artifactual and source obsidian. The XRF spectrometer consisted of a Spectrace 440 energy dispersive machine and 572 power supply (50kV, lmA), 534-1 pulsed tube control, 588 bias/protection module, 514 pulse processor or amplifier, Tracor Northemrn 1221 100 mHz ADC converter, and a Tracor Northern 2000 computer based analyzer with an LSI-11 microcomputer (Hughes 1986:25). The Si(Li) solid state detector with 144 eV resolution (FWHM) at 5.9 keV in a 30 mm squared area was used for detecting all x-ray intensities. For analysis of trace elements in the midz energy range, a rhodium (Rh) x-ray tube was used for primary x-ray excitation at 30.0 keV, .20 mA pulsed, with a .05 mm Rh primary beam filter in an air path at 300 seconds livetime. Analytical lines used for analysis were Ni (Ka), Cu (Ka), Zn (Ka), Ga (Ka), Pb (Lb), Th (La), Rb (Ka), Sr (Ka), Y (Ka), Zr (Ka), and Nb (Ka). For rare earth elements, an Am241 100 mCi radioscope source was used in the 20-60 keV range at 300 seconds livetime (Hughes 1986:26), and analytical values were derived from the Ka lines of Cs, Ba, La, Ce, Pr, Nd, and Sm. The x-ray tube was operated at 15.0 kV, .40 mA pulsed, with an Al primary beam filter in a vacuum path at 200 seconds livetime for Ka lines of Fe, Mn, and The elemental data as reported represent one analysis per specimen. While Bouey (1991) has demonstrated some variability with multiple analyses of the same obsidian specimens when values are presented in strictly quantitative (ppm) data, these Ti. Chemical Characterization of Manu'a Adz Material effects are minimized when presenting data as ratios (see, for example, Jack and Carmichael 1969). The XRF technique has its greatest accuracy in detecting elements in the mid-z range (e.g., bidium, strontium, yttrium, zirconium, and niobium) which was confirmnned by this study for both destructive and non-destructive sample preparations analyzed on their respective instruments. It is fortunamte that these elements are of particular interest to igneous petrologists (Cox et al. 1979:332). For pressed pellets or fused disks these particular mid-z elements have satisfactory precisions and accuracies down to 1 3 ppm, while niobium and yttrium could probably be measured to 0.1 ppm (Hooper and Johnson 1987). In basalt, mid-z elements are also present in sufficient concentrations to be easily measured, and detection in whole, unaltered specimens ranges from 10 ppm to 100% with an accuracy of ± 2 5% under favorable conditions (Parkes 1986:153). Lighter elements are not only harder to detect and measure by XRF, but they are usually not present in great abundance. Iron, titanium, and magnesium values, reported here, were detected under vacuum. Lead (Pb), although detected during the mid-z analysis, was not used to discriminate sources or characterize artifacts due to its presence in low abundance and its susceptibility to atmospheric contamination (Flanagan 1969:82). Rock standards have been used in routine XRF ru- - - analyses for about twenty-five years (Flanagan 1969), and currently at least 272 geostandards are in use worldwide (Govindaraju 1989). Standards are important for calibrating the XRF instrumentation for matrix effect corrections and for monitoring the precision and accuracy during analysis (Germanique and Briand 1985). Using standards thought to be close in composition to the unknowns limits the efficacy of matrix correction programs to a restricted range of elements. Conversely, many different standards provide a greater range of elements and values for calibrating the program and evaluating unknowns. Therefore, more than ten internal standards were used to calibrate the Spectrace 440 machine used in this study. U.S.G.S. standard RGM-1, a rhyolite from Glass Mountain, California, was used to monitor precision and accuracy during analysis and the results are presented in table 12.2. Ppm values reported by Govindaraju (1984, 1989) ar preferred "working values" which are the average of at least forty results from more than four techniques of analysis (1989:7). The accuracy and precision values reported for this study in ppm and selected ratios are reasonably close to accepted standard "working values." Due to its cryptocrystalline texture and homogeneous distribution of elemental abundances, it is not surprising that obsidian has garnered most analytical attention in XRF studies. Large grain sizes can Table 12.2 Evaluation of Analytical Accuracy and Precision for U.S.G.S. Standard RGM-1 Rb Sr Zr Y Nb Pb Th Zr/Sr Nb/Sr Govindaraju Govindaraju 1984 1989 155 100 200 25 9.4 21 15 2 0.094 149 108 219 25 8.9 24 15.1 2.03 0.082 173 This study (n=4) Precision Accuracy 159.9 107.3 227.2 27.7 10.3 28.6 20.4 2.12 0.096 156.3-163 106-108 225.3-229.9 25.9-29.2 9-11.7 27.1-31.1 18.9-22.1 Govindaraju (1984, 1989) pressed powder samples; this study, whole specimens analyzed. 174 The Tolaga Site distort XRF analysis of whole samples. Pressed pellet samples are prepared by reducing whole rock to a fine powder estimated to be 90% less than 400 mesh (Bice 1980:19) or ca. 40 microns. However, very fine-grained basalt may have more than 130 grains per mm (ca. 160 microns per grain). Table 12.3 Variation of Oxides and Elements from the Mo'omomi Basalt Quarry, Hawaiian Islands While this grain-size is early four times larger than pressed pellet samples, it has not been demonstrated that grains of this size adversely effect XRF analysis. This subject should be investigated further. The distribution of elemental abundances within a single basalt specimen may not be as homogeneous as obsidian, but a comparison of the date in table 12.3 is quite instructive. Here, the distribution of selected elements within a quarry (1.8 hectares in size) are indeed quite regular, and perhaps this is more so for individual rocks or artifacts. Another factor which can affect XRF analysis is an uneven specimen surface. Analyzing adz material has a distinct advantage, however, especially over bifacially flaked obsidian artifacts, because, almost by definition, most adz surfaces are extremely flat and, in many cases, are ground to a near mirror-like finish. Recalling that fused disks are finished by 240 (coarse) and 600 (fine) grit prior to XRF analysis (Hooper and Johnson 1987), examined under 10-40X magnification source rock from eight west Moloka'i Island basalt quarries whose material exhibited a range of textures and phenocryst sizes and densities. These samples were polished with 600 grit mesh and compared to the adz material in the present study. The prepared specimens had uniform, smooth surfaces and occasional striations formed by disintegrating phenocrysts that had been trapped between the rock and grinding plate. Although the artifacts were not as uniform in contour and had many more striations, portions of the artifacts were as smooth-if not more finely polished-4han the prepared specimens. Therefore, careful selection of artifact surfaces for EDXRF analysis may limit the amount of analytical distortion caused by uneven sample surfaces. To reduce or eliminate the effects of uneven sample surface, many researchers have advocated presenting elemental abundance values as ratios. "In spite of variations in effective sample surface of randomly broken pieces or loosely packed grains, relative intensities may be very precisely determined" (Jack and Carmichael 1969:30; see also Range Mean n=10 Oxide (weight %) 46.16±+0.12 SIO2 15.72 0.04 AL203 4.17 0.01 TIO2 0.20 14.21 FEO* 0.19 0.01 MNO 8.53 0.03 CAO 6.32 0.08 MGO 0.92 0.02 K20 3.52 0.07 NA20 0.62 0.01 P205 ± ± ± ± ± ± ± ± Element (ppm) Ni 68.30 1.90 5.40 1.69 Cr 18.50 2.58 Sc 302.20 7.70 V 215.50 20.69 Ba 14.40 0.92 Rb 786.00 6.62 Sr Zr 272.50 2.91 Y 37.80 0.98 27.09 1.19 Nb 23.30 1.79 Ga 11.80 5.62 Cu Zn 133.30 1.90 4.30 1.35 Pb 18.80 9.05 La 62.00 9.89 Ce 0.80 0.98 Th 45.9646.36 15.66-15.81 4.157-4.183 13.91-14.04 0.178-0.202 8.50-8.58 6.18-6.42 0.90-0.95 3.40-3.67 0.613-0.634 ± 64-70 3-8 15-23 283-310 187-258 12-15 775-798 268-278 37-40 25.0-28.9 21-26 4-21 ± 131-137 ± ± ± ± ± ± ± ± ± ± ± ± ± ± 2-6 2-31 49-80 0-3 *total iron Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University Stross et al. 1968:82). Sheets et al. (1990:149-50) concur that errors introduced by sample size and shape are largely cancelled by "the use of abundance ratios of elements having nearly the same energy (e.g., Rb, Sr, and Zr)." These observations were 175 Chemical Characterization of Manula Adz Material taken into consideration for the present study. Ever since the pioneering work of Parks and Tieh (1966), selection of elements for obsidian analysis has been fairly standard. Regarding data reduction, however, two "camps" have emerged somewhat recently. For reasons cited above, elemental abundances are usually presented in semiquantitative ratio form including temrnary diagrams (e.g., Best 1984; Jack and Carmichael 1969; Shackley 1988; Stross et al. 1968) or scatter plots (see fig. 12.3, this study; Jack and Carmichael 1969). Hughes (1986, 1988a, b) suggests that simple bivariate plots of zirconium and rubidium in ppm are sufficient for distinguishing raw material and determining artifact provenance in some regions. However, Bouey (1991) demonstrated that multiple analyses of the same specimens produced "widely divergent determinations in ppm concentrations," while ratio level presentation reduces this error significantly (cf. Jack and Carmichael 1969; Sheets et al. 1990). Hughes (1986) and Walter (1990) have applied statistical clustering programs to define geochemical groups. Multivariate statistics are quite appropriate for some problems, however their use eliminates any consideration of the elemental abundances (either in ppm or at a nominal level) as having any geological value. For example, on west Moloka'i, Hawaiian Islands, relatively high values of Y signal one particular quarry (David Clague, personal communication, 1989) and, simple bivariate plots of silica and total alkalis can be very informative for determining sources and characterizing geochemical groups. Statistical clustering techniques should only be used after the data have been examined for geological information that can inform on the "provenance environment," that is, the geology of the suspected interaction sphere. Conversely, groups created by statistical manipulation are an end in themselves. RESULTS The results of the fused disk XRF analysis of sixteen source rocks from Tatagamatau, Mako Ridge, and Fa'ala'aga are presented in tables 12.4-6 and summarized in table 12.7. According to a recent comprehensive and international evaluation of the systematics of igneous rocks (Le Maitre 1989), the Tatagamatau and Mako Ridge rocks are classified .125T 0 .115 .105 .095 .085 CO, Tatagamatau Fa'ala'aga (x) subsource .075 A/ 0 ..0 D f~- r * A .065 0 0 .055 A 0 * Mako (W) .045 I . -I .045 .055 .065 .075 .085 Y/Sr Figure 12.3 A wide dispersion of data points results when ratios of rubidium, strontium, and yttrium are used to assign specimens to possible sources. chemically as hawaiites and the Fa'ala'aga rock as basalt. Rock names were assigned by plotting silica values between total alkalis. Between sources, marked differences are found with the oxides, A1203, FeO, CaO, and P205, and most trace elements. The samples from the Tatagamatau source reveal important intra-source variability between Areas 1 and 3. Oxide values for AI203, MnO, Na20, and P205 demonstrate marked differences as well as the trace elements Ni, Cr, and Cu. Until additional samples collected from several areas of the 110-acre Tatagamatau quaTrry complex are analyzed and the geochemical variability of source rock is understood in greater detail, Ni and Cr may well be significant trace elements for demonstrating intra-source variability; that is, at least between Areas I and 3. 176 The To'aga Site U _ t- Ct. .l NO tn e qt qC C-4- Wo C4 Co o r "4 Rt 4 en i t- -i cn ooo %enao%0 %A o4 r- t- t- m X e-O m W)0i A ^ ml C14 1", ° ° ""8 tx T ,r § A"'EDn . 0\ w C2 q0 - O-" t __) 9 00 o (C: cne A co X Xo? -4 o, t-Oo 1¢ V- 00 Ir) V"4 %o t o 8 7- "4 A (n "4 "M4 0 "I0"-4.rr 0a ( t_4 00 ON Q£ N en %n qt - ."-4 en fo - oi 00C)C in C1I4 t O 00 r q en _en 4 i*b CD - A " 4n -4 teIc- _) m nx w x) _- M x "o V0% PC w _4 - , a 4E. t- N 'v O- NA xob ICoooo %At} xNb£l_ or_ WI--,, en 00 N -O Oso 0 CoS co en)" "400 0 ~ie~ en 8v4 C 00%o 00 000% o "4 w f~'% n-t 0 X 4) %0 p.1. as N e4 * - qt *) X 0% 00 8 N wi a- W) X 00 00 _4 wr t- t..: w V 4 t-) qt e VC'4 z 0% A O* " e4 0 O-> Bo "I 00 Eco0 "- 8 W) t-- r 00 O C -g S m t- mW F mO W) -n -OO-~o "-4 "-4 U) 00 l m CNqt 00 t en 00 N en NW^ o4 x ) o o m Ch " en t _ o 0t o 0 Rt o N 4 o \0 C4 t oW 0 t-X t4 Wo tn ON O "400C~~0~%00 . _rT ^ . _ C o0 -j _ on o en tl-W J; e en N N O El> en n qt 0 cx oor ;x 14o7A 0x (sOut ;^xe " en W-O I :a r- 0% 0% _4*j 00 %0t %n c4 c4 4 4 0D 0 _V- o 0% 0%N. 00I "40 0 o O- to Xou >. _ t -O en LL A t o iu i izso eq ONI w o t"9 a ne w TO 4 *i;t * s; Chemical Characterization of Manu'a Adz Material 177 Table 12.5 Geochemistry of Mako Ridge Source Rock (Fused Disks) Sample No. 89-1 89-2 89-3 89-4 Oxide (Weight %) SiO2 46.48 A1203 18.22 TiO2 3.552 FeO* 12.10 MnO 0.21 CaO 6.65 4.44 MgO K20 1.89 Na20 3.45 P205 0.988 Total 97.980 49.84 17.22 3.238 10.68 0.193 8.42 4.41 1.79 4.18 0.897 100.868 43.44 19.72 3.886 13.20 0.231 5.50 4.42 1.98 2.55 1.070 95.997 48.12 18.33 3.473 0.208 7.29 4.24 1.37 3.84 0.960 99.311 0 0 0 0 8 19 246 471 52 481 469 61 104.8 39 0 188 7 62 144 8 1 12 204 450 14 856 429 58 89.3 35 Element (PPM) Ni 0 Cr 3 Sc 23 V Ba Rb Sr Zr Y Nb Ga Cu Zn Pb La Ce Th 228 433 46 660 433 13 206 421 39 883 397 97 96.3 36 0 163 7 86 141 7 82.7 30 2 154 3 68 121 7 46 11.48 0 167 6 87 161 6 Range Mean 43.44-49.84 17.22-19.72 3.238-3.886 10.68-13.20 0.193-0.210 5.50-7.29 4.24-4.44 1.37-1.98 2.55-4.18 0.897-1.070 46.97 18.35 3.537 11.87 0.211 6.966 4.38 1.76 3.51 0.979 98.533 0 0-8 12-23 204-246 421-471 14-52 481-883 397-469 46-97 82.7-104.8 0 30-39 0-2 154-188 3-7 62-87 121-161 6-8 3 16.75 221 443.75 37.75 720 432 65.5 93.275 35 0.5 168 5.75 75.75 141.75 7 * total iron Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University, 16-17 October 1989. EDXRF analysis of Tatagamatau, Mako Ridge, and Fa'ala'aga source rock are presented in tables 12.8-10; mean and range for sources are summarized in table 12.11. Elemental values are presented with one standard deviation which represents the uncertainty of counting statistics at 300 seconds livetime. A comparison of EDXRF and fused disk samples for source means and the Tatagamatau source envelopes is illustrated in figure 12.2. Zr/Sr values are very comparable, whereas Nb/Sr ratios are higher for fused disk values. This probably corresponds to the greater detection efficiency (machine sensitivity) for niobium by the Rigaku spectrometer (Hooper and Johnson 1989); or it could be a sample thickness problem since niobium x-rays are excited as deep as 5-7 mm into the specimen (David Clague, written 178 The To'aga Site Table 12.6 Geochemistry of Fa'ala'aga Source Rock (Fused Disks) Sample No. 89-14 Range Mean 89-15 89-16 0.674 99.808 46.69 13.79 5.426 12.93 0.178 10.56 4.8 1.73 3.13 0.684 99.918 46.79 13.80 5.335 13.06 0.174 10.75 4.76 1.60 2.97 0.627 99.866 46.69-46.89 13.68-13.80 5.335-5.426 12.93-13.06 0.172-0.178 10.54-10.75 4.72-4.80 1.60-1.73 2.97-3.13 0.627-0.684 46.79 13.76 5.391 13.01 0.175 10.62 3.16 1.67 3.03 0.662 98.268 Element (PPM) Ni 57 24 Cr 21 Sc V 344 Ba 300 Rb 45 644 Sr Zr 323 34 Y Nb 68 28 Ga 104 Cu Zn 128 5 Pb La 40 Ce 115 4 ITh 88 25 25 357 305 46 642 330 36 71.1 30 117 138 7 45 108 6 72 38 19 359 307 42 651 313 35 66.6 29 101 129 6 31 104 3 57-88 24-38 19-25 344-359 300-307 42-46 642-651 313-330 34-36 66.6-71.1 28-30 101-117 128-138 5-7 31-45 104-115 3-6 72.3 29 21.7 353.3 304 44.3 645.7 322 35 68.57 29 107.3 131.7 6 38.7 109 4.3 Oxide (Weight %) 46.89 SiO2 A1203 TiO2 FeO* MnO CaO MgO K20 Na20 P205 Total * 13.68 5.412 13.04 0.172 10.54 4.72 1.68 3.00 total iron Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University, 16-17 October 1989. communication, 1990). The larger source envelope for EDXRF may relate to the greater variability in specimen surface. After selecting elements for analysis (based on analytical precision, accuracy, and sufficient elemental concentrations), assigning artifacts to source was facilitated initially by trial and error. Figure 12.3 illustrates a wide dispersion of data points with most artifacts plotting outside the source envelope when Rb/Sr and Y/Sr ratios are used. Figure 12.4, however, using ratios of Zr/Sr and Nb/Sr, is much mome useful for assigning artifacts to the Tatagamatau source. The source envelope is delimited by taking into account the variability in analytical precision. Chemical Characterization of Manu'a Adz Material 179 Table 12.7 Geochemistry of Fa'ala'aga, Mako Ridge, and Tatagamatau Source Rock (Fused Disks) Fa'ala'aga (n=3) Range Mean Mako Ridge (n=4) Range Mean Tatagamatau (n=8) Tatagamatau sub-source Range Mean (n=1) 50.02 15.45 3.413 12.68 0.177 7.63 4.69 1.58 4.02 0.800 49.32 13.86 3.649 12.40 0.167 7.80 7.06 1.79 3.45 0.710 Oxide (Weight %) SiO2 A1203 46.69-46.89 13.68-13.80 TiO2 5.335-5.426 FeO* MnO CaO 12.93-13.06 0.172-0.178 10.54-10.75 4.72-4.80 1.60-1.73 2.97-3.13 0.627-0.684 MgO K20 Na20 P205 Element (PPM) Ni 57-88 Cr 24-38 Sc 19-25 V 344-359 Ba 300-307 Rb 42-46 Sr 642-651 Zr 313-330 Y 34-36 Nb 66.6-71.1 Ga 28-30 Cu 101-117 Zn 128-138 Pb 5-7 La 31-45 Ce 104-115 Th 3-6 46.79 13.76 5.391 13.01 0.175 10.62 3.16 1.67 3.03 0.662 43.44-49.84 46.97 17.22-19.72 3.238-3.886 10.68-13.20 0.193-0.210 5.50-7.29 4.244.44 1.37-1.98 2.554.18 0.897-1.070 18.35 3.537 11.87 0.211 6.966 4.38 1.76 3.51 0.979 49.56-50.56 15.35-15.62 3.385-3.467 12.34-12.90 0.173-0.179 7.56-7.70 4.55-4.77 1.54-1.62 3.95-4.09 0.783-0.814 0 0 0 0 -,,.--J 694-708 348-362 45-49 52.4-56.0 27-31 0-9 165-178 4-8 13-50 88-110 3-7 1.1 18.8 207.8 290.5 42.9 702.9 355.6 47.6 54.5 29.4 1.5 171.3 6.1 35.3 100.3 5.25 72.3 645.7 322 35 68.57 29 107.3 131.7 6 38.7 109 4.3 1t-'.JL 1 .0O 481-883 720 432 65.5 93.3 35 0.5 168 5.8 75.8 141.8 7 397469 46-97 82.7-104.8 30-39 0-2 154-188 3-7 62-87 121-161 6-8 173 181 27 233 331 50 683 325 37 47.9 26 31 153 5 43 99 2 * = total iron Analyst: Dr. Peter R. Hooper, Dept. of Geology, Washington State University, 16-17 October 1989. Nine artifacts fall within the envelope and four others are very close. Taken together, fifty percent of the adzes and other artifacts with one or more polished surfaces (adz flakes) can be assigned to the Tatagamatau quarry on Tutuila. Weathered surfaces of these specimens were gray (2.5Y 5/0; 2.5YR 5/0; 7.5YR 5/0), dark gray (2.5Y 4/0), to very dark gray (7.5R 3/0; 2.5Y 3/i, 7.5Y 3/0). Freshly broken surfaces on two specimens were very dark gray (2.5Y 3/0; 7.5Y 3/0). Additional samples from this quarry will undoubtedly define a much larger source envelope since the Tatagamatau sub-source is well outside the limits of the eight samples used to define the geochemical dimensions of the quarry. This underscores the need to collect sufficient samples to define the geochemical variability of adz quarry sources. The diagonal line in figure 12.4 separates + ur *i 6 + +1 The To'aga Site 180 0 I I eq M en.ON en.en M It tn.C. M -4 M In C4 r go mg 12, WI V-4 en 67,e C4 .s V-48 . _- en x ~o , C14 m V- -4 qt n en CD t -4 Q 4 % ent en , t- en C1 V-4 66 en en %n 0 e etiM t- 0 9Cs +1M t.o 1- te +1 ~o 0 in 00 ^o 4 e en t^ .- n7 -4xv V oN ON x - 4en ot4 o, C4 +1 (: m o" g} +l ,t° +l ON tn eq +l. +l + %n W r- + +1 WI)x 6 od eneW) in in ~oC' +l ' q CDC ( q 0 t- m (D t- >o t*4 F_ C. +D w C. :;V- IC .6 ,4 -H +1 m 'H +l t+l 00 C% m o V-4 C114 V O 00 6 C-4 Go Cf oo +1 s0 0 c>WqN t t- o Is v- +1 +H w' *o e4 TNO t- V4 en oo1 ql -V 0 :aco [.. F-4 eq. WI, C1 .2 00 ,H +l +l r eq oq. 'n:.H+~ t Mt 0e F* a +l en +1 06 C>r- em ok +l +l 't00 ,N t.. m ° C4 + oo:~ cr I00 no4 en ,--. * 0 o,,oen - cr W- W) 6 +1 t-: o6 T-4 _ ,t > +l 0 0 O-4 o ot C>W -4. cr) -t o) o o Q M + * +l;+It+l I1 0 r- 1* er r.8°H en cli * -H +.l , +l qq c5s + cl os t *1 W) V) in WIt- Q M o m o m; .1 + +1oo1 r- C> - .. WI O' 0 n ON -- t_ _ i * 0 It 0 0 14t +1 6 +1 00 W) oP-4 ,N.I 0o/ +I tn Af ·0 en o It=,,~ ko 00 C1 ot A m +lv" .H j" X- C It C114 0 p -2IL,20= Vt .H+I ._ C. q1t C; X=t, qqtC4 C1 c+1 kn o V) t 14 * (I +1 cri 00 ~=,=~ +1 10 0 in tn I Chemical Characterization of Manu'a Adz Material O XD tn CN 00 W N 00 00 e 0 ren rren C .4 It m CN b E V- en ,q N t- O~N D N-v00 "- tmO m 0\NXN0 M -- N 6 o x0 ~o 00 V0 OQ- 181 * ofi a uN 'r(..if0 tno ur 0% F-~t ,it re x- N "- - e. Ce i 0_ 'R r-W N I'0 - 06t % er- ( Oe -- _4 -O 6 ,^ 0 0 r- le U, co ._ 41 ' +00 U 414444cq N- 44414 0 0; -o N t n Cf) CN m 41 i ON W- V Ch o r- eni Wo< i X 't W 'o00 c _ +I - 't +1 + vo +1 +1 _- t_ O C1 _~ t- N V-4 en..4 C_ te m )N 00 Q + 'H _. c'4 V- .H .Hooi oof. ~o r,. . '0 I- c~ 00 0% c~ Is am' O9 4~ 0 c~ o .,t cn en 0 t.4 -v" et)o -4 . 1414m - X00 o 00 O t l(" t"'4v c-4 4 14141 414n .H.H 0H 4 41 - /00 * +4 H00 Sf Co- 0 m z Oa ONOO ON d+ W) 01 41~ Eas ,; fi · q * rt en -cr 4-141 tn W~ .H V- > 8'Hc,, ON aN 0 ) * 0 H .H en m -'0 414 H 'H 00' Vt- t- -- q ~t r 4 004s M It IR 0C0000 o.0 +141 41 Cx00 i en c m so t- 00 "_ - - N ON et qq 11- oN 00 cn W) 0% -_ ON * 00 (Ni h 04 41 (Ni4VNI H .H C'4 v'~ x- * N e "i4 cV'l cN I.H * (14"t r * cf CN _ ) It M 0 N 414+1 r~ 4141414141414 '0 o v * - wi -r x V At o 0 41 cs 0 (4 v r- It /3 o. .7.. P v. 2 w z m 1: (n t¢ >. z 0u 9 u u ;zQtaC < 182 The To'aga Site 70. Tatagamatau source rock from adz material deriving . 00. 8| - () / ~~~A . ·0 *c* from the dike sources of Fa'ala'aga and Mako Ridge on Ofu and from all unmodified flakes. The source for the flakes was probably the scree slope below the Mako M Tatagamatau X_ ss subsource.D ": A N **A Fa'ala'aga R) . *|~~~~~~~~~ subsor:e * * 40 Fa'ala'aga dike swarm. There are indeed several flakes that plot close to this source mean. The dike *source rock and flakes recovered from the site are all medium- to coarse-grained, and eight specimens (61.5% of the total flakes) have one or more natu- dike rock *.0~~~~~~~ ~rally flat surfaces characteristic of dike rocks. 50 60 ~70 80 ~~Nb/Sr 1~ margin is smoothed from use. These medium~distal to coarse-grained rocks were probably not conducive X Figure 12A.4 Several flakes, however, exhibit retouching and one Anifas and source rocks plotted by ratios of zirconiu, strontium, and niobium resulting in many specimns ploing widin the Tatagamatau source envelope. to adz manufacture, and it is perhaps noteworthy that all fine-grained rocks were manufactured into, or are Table 12.10 Geochemistry of Fa'ala'aga Source Rock (Whole Specimen) Sample No. (PPM) Oxide TiO2 FeO MnO Element Ni Ba Rb Sr Zr Y Nb Ga Cu Zn Pb La Ce Nd Cs Fr Th Co As 89-14 Range Mean 89-15 89-16 50476.8 ± 358 50757.1 ± 366.8 168791.1 ± 1153.7 171686.3 ± 1191.4 2391.4 ± 44.6 2396 ± 45.5 47928.1 ± 334 156480.4 ± 1047.1 1978.4 ± 40 47928.1-50757.1 156480.4-171686.3 1978.4-2396 49720.7 165652.6 2255.3 89.7 ± 7.1 233.3 ± 6.1 42.9 ± 2.4 704.8 ± 6.3 341.7 ± 5.2 39.1 ± 2.7 52.3 ± 3.6 33.9 ± 2.5 134.4 ± 5.7 115.9 ± 4.6 7.6 ± 1.8 26.8 ± 3.2 91.2 ± 3.9 34.4 ± 3.9 0 9.2 ± 3.5 0 0 1.6 + 0.4 84.6-103.2 230.3-234.4 92.5 232.67 45.57 690.57 354.43 38.4 54.7 29.9 135 132.2 7.33 35.53 80.9 39.63 0 8.47 0 0 1.7 84.6 ± 7.9 234.4 ± 6.1 45.4 ± 2.4 691.6 ± 6.2 362.4 ± 5.3 40 ± 2.8 51.8 ± 3.7 25.7 ± 2.7 135.9 ± 6.5 143.1 ± 5.5 7 ± 1.8 40.1 ± 3.3 70.2 ± 3.9 43.7 ± 3.9 0 8.3 ± 3.5 0 0 1.7 + 0.5 103.2 ± 8.1 230.3 ± 7.1 48.4 ± 2.6 675.3 ± 6.6 359.2 ± 5.7 36.1 ± 3 60 ± 4 30.1 ± 2.6 134.7 ± 6.4 137.6 ± 5.4 7.4 ± 2 39t.7 ± 3.9 81.3 ± 4.5 40.8 ± 4.4 0 7.9 ± 4 0 0 1.8 ± 0.4 Analyst: Marshall Weisler, September 1989. 42.9-48A.4 675.3-704.8 341.7-362.4 36.1-40 51.8-60 25.7-33.9 134.4-135.9 115.9-143.1 7-7.6 26.8-40.1 70.2-91.2 34.4-43.7 0 7.9-9.2 0 0 1.6-1.8 Chemical Characterization of Manus'a Adz Material 183 Table 12.11 Geochemistry of Fa'ala'aga, Mako Ridge, and Tatagamatau Source Rock (Whole Specimen) Mako Ridge (n=4) Fa'ala'aga (n=3) (PPM) Oxide i02* FeO* MnO* Element Ni Ba Rb Sr Zr Y Nb Ga Cu Zn Pb La Ce Nd Cs Fr Th Co As *x Mean Range Mean Range Tatagamatau (n=8) Range Mean (n=l) 3A.414 3.033-3.393 13.65-14.78 0.238-0.285 3.147 14.33 0.258 3.160-3.490 14.60-16.85 0.206-0.243 3.320 16.00 0.230 nd nd 340.0 40.4 791.9 477.7 66.1 71.2 33.0 16.2 168.7 7.8 57.3 121.8 68.8 0.6 11.6 3.5 nd 1.6 nd 209.2-262.4 36.4-46.0 718.1-769.7 376.4-409.6 47.1-53.1 41.0-47.3 27.3-36.8 19.1-95.3 159.0-191.1 0.0-13.1 28.3-41.4 70.5-84.5 34.1-52.3 nd 0.0-19.4 0.0-13.1 nd 0.8-2.1 nd 226.7 40.7 750.3 395.8 49.7 44.4 30.8 42.0 177.7 8.0 33.9 75.7 45.1 nd 8.9 1.6 nd 1.3 4.793-5.076 15.65-17.17 0.198-0.240 4.972 16.57 0.226 84.6-103.3 230.3-234.4 42.9-48.4 675.3-704.8 341.7-362.4 36.1-40.0 51.8-60.0 25.7-33.9 134.4-135.9 115.9-143.1 7.0-7.6 26.8-47.6 70.2-81.3 34.443.7 92.5 232.6 45.6 690.6 354.4 38.4 54.7 29.9 135.0 132.2 7.3 38.0 77.6 39.6 nd nd 7.9-9.2 nd nd 1.6-1.8 8.5 nd 327.1-354.2 14.6-52.2 551.6-959.7 466.1-497.1 48.2-94.8 66.4-75.3 32.1-33.9 8.3-26.4 161.6-178.3 5.9-9.8 47.2-66.5 115.2-131.3 60.2-75.3 0.0-2.4 10.1-15.1 0.0-13.8 nd 1.7 1.2-2.2 nd Tatagamatau sub-source 15.55 0.220 188.0 259.2 56.5 738.6 357.0 38.8 32.7 28.3 61.2 149.2 nd 33.1 75.9 38.3 nd nd nd nd 1.5 1i4 Analyst: Marshall Weisler, September 1989. the by-products of, adz production. Geochemical data for all artifacts are presented in table 12.12. N'me artifacts assigned to the Tatagamatau adz quarry complex on Tutuila are from three archaeological sites on Tau Island and two localities within the large To'aga coastal habitation area (see table 12.1). Unformmnately, only three artifacts are from excavated contexts, and the remaining are surface finds. All stylistically diagnostic specimens are quadrangular-sectioned adzes or fragments dating to the late prehistoric penriod (Green 1974; Green and Davidson 1969). By connecting the six sites with adz material that originated from the Tatagamatau quarry, we can document several nodes of an interaction sphere between the islands of Tutuila, Ta'u, and Ofu. These.results seem quite promising for additional provenance studies of the kind employed here. 184 The To'aga Site Table 12.12 Geochemistry of To'aga and Ofu Island Basalt Artifacts (Whole Specimen) Trace Element Concentrations (PPM) Artifact Number Fiti'uta 1 11-1-S-4 11-2-S-1 11-2-S-2 11-2-S-3 11-2-S-8 11-2-9 11-52-S-9 13-S-3 13-S-5 13-S-6 13-S-9 13-1-S-36 13-1-3-1-1 13-1-9-2-4 13-1-13-87 13-1-16-1-84 13-1-16-11 13-1-16-11-81 13-1-16-12-83 13-1-20-4-118 13-1-20-5-121 13-1-20-5-122 13-1-20-5-123 13-1-20-6-139 13-1-20-6-140 13-1-20-6-141 13-1-20-7-143 13-1-20-9-147 13-1-21-7-151 13-1-22-2-171 13-1-22-2-173 13-1-22-2-174 13-1-23-7-194 13-1-27-5-39 13-1-27-5-40 13-1-27-5-41 13-1-27-6-42 Lab Number 89-22 89-25 89-26 89-31 89-30 89-28 90-30 90-33 89-18 89-34 89-33 89-32 89-19 89-17 90-22 90-23 90-42 90-46 9045 90-38 90-27 90-37 9041 90-34 9048 90-29 90-36 9040 9043 9049 90-44 90-25 90-24 90-26 90-31 90-39 9047 90-35 Pb 11.3 :e 1.3 11.1 ± 2 11.0 ± 1.6 7.7 + 1.7 11.1 ± 1.7 0.0 10.4 ± 2.1 8.9 ± 1.8 6.5 ± 1.6 11.7 ± 1.8 6.6 ± 2 7.6 ± 2 5.3 ± 1.9 9.2 ± 1.6 0.0 12.0 ± 2.1 8.3 + 1.8 9.6 ± 2 9.6 + 2.2 9.7 ± 1.8 6.9 ± 1.7 10.9 ± 2.1 0.0 9.1 ± 1.7 8.0 ± 1.8 7.4 ± 1.6 11.8 ± 2 9.1 ± 1.7 11.4 ± 1.9 8.7 ±t 2 8.8 ± 1.7 8.7 ± 1.9 12.8 ± 2.4 7.6 ± 1.9 8.6 ± 1.9 7.6 ± 1.6 7.1 ± 1.9 9.1 ± 1.8 Th 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Rb Sr Y Zr Nb 65.4 ± 2.4 68.1 ± 6 52.4 ± 2.2 49.0 ± 2.2 50.5 ± 2.3 30.2 ± 2.1 57.8 ± 2.4 59.3 ± 2.5 48.7 ± 2.3 49.0 ±t 2.2 43.0 ± 2.3 46.3 ± 2.4 38.9 ± 2.4 58.9 ± 2.4 48.9 ± 2.5 26.4 ± 2.2 31.5 ± 2.3 32.5 ± 2.2 40.8 ± 2.6 37.3 ± 2.3 48.5 + 2.3 33.5 ±t 2.3 38.7 ± 2.4 37.2 ± 2.2 40.1 ± 2.4 46.4 ± 2.1 31.1 ±t 2.3 35.6 + 2.2 832.5 ± 6.3 724.0 ± 6 812.8 ± 6.1 782.3 ± 6 816.8 ± 6.2 578.0 ± 5.3 815.7 ± 6.4 750.7 ± 6.2 823.3 ± 6.5 824.4 ± 6.1 772.1 ± 6.2 691.2 ± 6 635.8 ± 6.2 854.1 ± 6.5 894.5 ± 7.2 660.0 ± 5.8 545.5 ± 5.6 638.5 ± 5.8 856.3 ± 7.5 967.8 ± 7.3 772.8 + 6.2 691.7 ± 6 733.0 ± 6.7 700.9 + 5.9 739.8 ± 6.5 724.4 ± 5.5 758.6 ± 6.6 694.4 ± 6.2 969.8 ± 7.9 524.1 ± 5.5 786.8 ± 6.3 722.4 ± 7 712.3 ± 6.8 689.7 ± 5.6 793.6 ± 6.2 729.7 ± 5.7 676.3 + 5.7 703.6 ± 6.4 52.0 ±: 2.7 51.9 ± 2.7 49.6 ± 2.6 41.1 ± 2.6 40.3 ± 2.6 39.2 ± 2.6 51.1 ± 2.8 48.0 ± 2.8 43.7 ± 2.7 50.4 ± 2.6 44.0 ± 2.7 34.7 ± 2.7 39.1 ± 2.9 50.9 ± 2.7 49.0 ± 3 31.6 ± 2.6 33.6 ± 2.8 32.4 ± 2.6 35.3 + 3.1 34.7 + 2.6 40.8 ± 2.7 36.8 ± 2.6 29.3 ± 2.8 33.9 ± 2.6 33.2 ± 2.7 35.7 ± 2.4 37.0 + 2.8 27.9 ± 2.7 44.7 ± 3.1 24.4 ± 2.8 37.5 ± 2.7 45.2 ± 3.1 49.6 ± 3.1 47.1 ± 2.6 50.2 ± 2.7 50.0 ± 2.6 30.8 ± 2.5 40.7 ± 2.8 438.8 ± 5.2 449.1 ±t 5.4 397.6 ± 5 369.4 ± 4.9 422.8 ± 5.2 339.1 ± 4.8 427.4 ± 5.4 457.0 ± 5.5 406.9 ± 5.3 438.0 ± 5.1 413.1 ± 5.3 322.5 ± 5 365.1 ± 5.6 464.6 ± 5.5 527.1 ± 8.2 292.1 ± 4.8 329.7 ± 5.1 293.2 ± 4.8 333.8 ± 5.7 331.0 ± 5.2 386.0 ± 5.1 317.7 ± 4.9 321.3 ± 5.4 324.5 ± 4.9 335.7 ± 5.2 368.9 ± 4.6 342.1 ±t 5.3 314.2 + 5.1 401.7 ± 6 264.6 ± 4.9 385.8 ± 5.2 335.4 ± 5.7 361.9 ± 5.8 415.3 ± 5 455.4 ± 5.4 440.0 ± 5.1 298.1 ± 4.6 392.7 ± 5.6 50.5 ± 3.3 51.1 ± 3.4 49.4 ± 3.2 43.2 + 3.2 46.9 ± 3.3 36.7 ± 3.3 49.9 ± 3.5 53.1 + 3.5 44.3 ± 3.4 50.9 ± 3.3 43.9 ± 3.4 37.4 ± 3.4 53.0 ± 3.8 54.3 ± 3.4 64.5 ± 3.8 27.8 ± 3.3 34.1 ± 3.6 49.7 ± 3.5 53.0 ± 4 51.1 ± 3.5 44.4 + 3.4 48.3 ± 3.5 58.5 ± 3.8 56.6 ± 3.5 52.1 ± 3.7 36.0 ± 3 61.7 ± 3.8 52.0 ± 3.6 63.4 ±4 41.4 ± 3.7 49.6 ± 3.5 40.8 ± 3.9 43.6 ± 3.9 45.4 ± 3.3 53.7 ± 3.4 47.3 ± 3.2 48.9 ± 3.3 43.0 ± 3.6 69.0 ± 49.8 + 52.3 ± 46.1 ± 41.5 ± 48.2 ± 49.6 ± 50.5 ± 38.1 ± 43.4 ± 2.8 2.5 2.4 2.7 2.7 2.2 2.3 2.2 2.2 2.4 Analyst: Marshall Weisler, September 1989 to Febr,-uy 1990. Chemical Characterization of Manula Adz Material SUMMARY AND CONCLUSIONS Defining the spatial and temporal dimensions of intra- and inter-island communication is an important precursor for evaluating historical developments on insular landscapes. Because Polynesia generally lacks the "footprints of pottery" throughout sequences of most island groups, and large quantities of obsidian are absent in this geologic province, finegrained basalt manufactured into adzes and widely distributed must of necessity formnn an important basis for tracking regional interaction. Knowledge of the "provenance environment" or geologic features and their chemical signatures are essential aspects for identifying and predicting the occurrence of finegrained, stone-tool-quality basalt. Two material characterization and provenance techniques have been described and evaluated, and geochemical analysis is argued to be the most profitable for longterm and regional-scale distributional studies as: (1) the results are reproducible; (2) instrument operating conditions can be reported in full facilitating comparison of regional databases; (3) identification of elements is not subject to human error as with thinsection descriptions; (4) elemental abundances can be reported with precision and accuracy values for specimens and standards; and (5) geochemical sampling locales on specimens more closely represent the population rather than petrographic thinsections which are limited by two-dimensional surfaces. The efficacy of non-destructive x-ray fluorescence spectroscopy has been demonstrated and its continued use for distributional studies of Polynesian adz material is warranted. Although the technique, as applied here, is limited to elements that can be analyzed with high precision and accuracy (those in the mid-z range), future applications should seek to expand these limits. A major benefit of EDXRF is the wide range of specimen sizes that can be analyzed without destruction. In this study, artifacts ranged from as small as 2.1 grams up to 421.9 grams with lengths of 22 to 136 mm. Accommodating a wide range of specimen sizes can be advantageous when analyzing assemblages with a high proportion of small adz flakes, a common situation with most collections of adz material. In fact, with nearly 1000 m2 of archaeological excavations on the island of Moloka'i, small adz flakes outnumber whole or stylistically diagnostic adzes by more than 10 to 1. Moreover, most adzes are surface 185 finds whereas adz flakes are more often from excavated contexts. Application of the non-destructive, energydispersive technique to source matenrial and artifact assemblages from several sites on Ta'u and Ofu islands has permitted delineation of an interaction sphere during late prehistory that links habitation complexes on these islands to the large adz quarry complex at Tatagamatau, Tutuila Island, some 100 kmn to the west. Local dike rock from Ofu Island is chemically similar to most of the non-polished and unground flakes. Geochemical analysis has shown that these medium- to coarse-grained rocks were not fashioned into adzes, but this material was restricted to a few retouched, "awl-like" tools and one flake with use-wear along a distal margin. The vast majority, however, were unmodified. EDXRF is not the answer to all provenance studies, but the results reported here suggest that the technique merits further use and refinement with adz material from Polynesia and possibly throughout Oceania (e.g., Weisler 1993; Weisler et al. In prep.). As the ownership and preservation of archaeological sites and museum collections become increasingly the focus of controversy between the scientific and native communities, discovering more about the past without destroying the evidence itself, may help to bridge our common goals. ACKNOWLEDGMENTS I thank David Clague (Director, Hawaii Volcanoes Observatory), Joachim Hampel (Department of Geology, University of Califomrnia, Berkeley), M. S. Shackley (Lowie Museum, University of Califomrnia, Berkeley), Pat Kirch, and Roger Green for comments on an earlier version of this paper. I am grateful to Joachim Hampel who was especially helpful during my XRF analyses, and I thank Pat Kirch and Terry Hunt for inviting me to participate in the Manu'a archaeological project. REFERENCES CITED Best, S. 1984. Lakeba: The prehistory of a Fijian island. Unpublished Ph.D. dissertation, Department of Anthropology, University of Auckland. Best, S., H. Leach, and D. Witter. 1989. Report on the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August 186 The To'aga Site 1988. Report on file, Historic Preservation Office, American Samoa. Best, S., P. Sheppard, R. Green, and R. Parker 1992. Necromancing the stone: Archaeologists and adzes in Samoa. Journal of the Polynesian Society 101:45-85. Bice, D. 1980. XRF data storage/matrix correction programs. Ms. on file, Department of Geology and Geophysics, University of California, Berkeley. Bouey, P. 1991. Recognizing the limits of archaeological applications of non-destructive energydispersive x-ray fluorescence analysis of obsidians. Materials Research Society Symposiwn Proceedings 185, Materials Issues in Art and Archaeology ll. pp. 309-320. Pittsburgh. Cox, K., J. Bell, and R. Pankhurst 1979. The Interpretation ofIgneous Rocks. London: George Allen and Unwin. Davidson, J. 1977. Westem Polynesia and Fiji: Prehistoric contact, diffusion and differentiation in adjacent archipelagos. World Archaeology 9:82-94. .1978. Western Polynesia and Fiji: The archaeological evidence. Manlakind 11:383-90. . 1979. Samoa and Tonga. IN J. Jennings, ed., The Prehistory of Polynesia, pp. 82-109. Cambridge: Harvard University Press. Dunnell, R 1971. Systematics in Prehistory. The Free Press, New York. Dye, T., M. Weisler, and M. Riford 1985. Adz quarries on Moloka'i and Oahu, Hawaiian Islands. Report on file, Historic Sites Section, Department of Land and Natural Resources, Honolulu, Hawaii. Emory, K. 1933. Stone Remains in the Society Islands. Bernice P. Bishop Museum Bulletin 116. Honolulu. Flanagan, F. 1969. U. S. Geological Survey standards-11. First compilation of data for the new U.S.G.S. rocks. Geochimica et Cosmochimnica Acta 33:81-120. Gemianique, C., and B. Briand 1985. XRF determination of Zr, Nb, Y, Sr, Rb, Zn, Pb, in fifteen international geochemical reference samples. Geostandards Newsletter 9:31-34. Govindaraju, K. 1984. 1984 compilation of working values and sample description for 170 international reference samples of mainly silicate rocks and minerals. Geostandards Newsletter Special Issue 8. . 1989. 1989 compilation of working values and sample description for 272 geostandards. Geostandards Newsletter Special Issue 13:1113. Green, R. 1974. Review of portable artifacts from Westemrn Samoa. IN R Green and J. Davidson, eds., Archaeology in Western Samoa Vol. II, pp. 245-75. Bulletin of the Auckland Institute and Museum 7. Green, R., and J. Davidson, eds. 1969. Archaeology in Western Samoa. Bulletin of the Auckland Institute and Museum 6. Auckland, New Zealand. Hooper, P., and D. Johnson 1987. Major and trace element analyses of rocks and minerals by automatic x-ray spectrometry. Ms. on file, Department of Geology, Washington State University, Pullman. Hughes, R. 1986. Diachronic variability in obsidian procurement patterns in northeastern California and southcentral Oregon. University ofCalifornia Publications in Anthropology 17. Berkeley: University of California Press. . 1988a. Archaeological significance of geochemical contrasts among southwestern New Mexico obsidians. The Texas Journal of Science 40:297-307. -. 1988b. The Coso volcanic field reexamined: Implications for obsidian sourcing and hydration dating research. Geoarchaeology 3:253-65. Hunt, T., and P. Kirch 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal of the Polynesian Society 97:153-83. Irwin, G. 1990. Human colonisation and change in the remote Pacific. Current Anthropology 31:90-94. Jack, R., and I. Carmichael 1969. The chemical 'fingerprinting' of acid volcanic rocks. Special Report 100:17-32. California Division of Mines and Geology, Sacramento. Kirch, P. 1985. Feathered Gods and Fishhooks: An Introduction to Hawaiian Archaeology and Prehistory. Honolulu: University of Hawaii Press. . 1990. Regional variation and local style: A neglected dimension in Hawaiian prehistory. Chemical Characterization of Manu'a Adz Material Pacific Studies 13:41-54. Kirch, P. V., and R. Green 1987. History, phylogeny, and evolution in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., T. Hunt, L. Nagaoka, and J. Tyler. 1990. An ancestral Polynesian occupation site at To'aga, Ofu Island, American Samoa. Archaeology in Oceania 25:1-15. Leach, H., and D. Witter 1987. Tataga-matau "rediscovered." New Zealand Journal of Archaeology 9:33-54. . 1990. Further investigations at the Tataga-matau site, American Samoa. New Zealand Journal of Archaeology 12:51-83. Le Maitre, R. ed. 1989. A Classification ofIgneous Rocks and Glossary of Termns, Recommendations of the International Union of Geological Sciences Subcommission on the Systematic of Igneous Rocks. Oxford: Blackwell Scientific Publications. Parkes, P. 1986. Current Scientific Techniques in Archaeology. London: Croom Helm. Parks, G., and T. Tieh 1966. Identifying the geographical source of artifact obsidian. Nature 211:289-90. Prickett, N., ed. 1982. The First Thousand Years: Regional Perspectives in New Zealand Archaeology. New Zealand Archaeological Association, Palmerston North, New Zealand. Rolett, B. 1989. Hanamiai changing subsistence and ecology in the prehistory ofTahuata (Marquesas Islands, French Polynesia). Unpublished Ph.D. dissertation, Yale University, New Haven. Shackley, M. 1988. Sources of archaeological obsidian in the Southwest: An archaeological, petrological, and geochemical study. American Antiquity 53:752-72. Sheets, P., K. Hirth, F. Lange, F. Stross, F. Asaro, and H. Michel 1990. Obsidian sources and elemental analyses of artifacts in southemrn Mesoamerica and the Northem Intermediate Area. American Antiquity 55:144-58. 187 Slice, G., and F. McCoy 1968. The geology of the Manu'a Islands, Samoa. Pacific Science XXII:427-57. Stross, F., J. Weaver, G. Wyld, R. Heizer, and J. Graham 1968. Analysis of American obsidians by x-ray fluorescence and neutron activation analysis. Contributions of the University of California Archaeological Research Facility 5:59-79. Berkeley. Terrell, J. 1986. Prehistory in the Pacific Islands. Cambridge: Cambridge University Press. Walter, R. 1990. The southern Cook Islands in eastern Polynesian prehistory. Unpublished Ph.D. dissertation, University of Auckland, New Zealand. Walter, R., and W. Dickinson 1989. A ceramic sherd from Ma'uke in the southern Cook Islands. Journal of the Polynesian Society 98:465-70. Weaver, J., and F. Stross 1965. Analysis by x-ray fluorescence of some American obsidians. Contributions of the University ofCalifornia Archaeological Research Facility 1:89-93. Berkeley. Weisler, M. 1990. Sources and sourcing of volcanic glass in Hawai'i: Implications for exchange studies. Archaeology in Oceania 25:16-23. . 1993. Long-distance interaction in prehistoric Polynesia: Three case studies. Unpublished Ph.D. dissertation,University of California, Berkeley. In press a. Provenance studies of Polynesian adze material: A review and suggestions for improving regional data bases. Asian Perspectives 32(1). In press b. The settlement of marginal Polynesia: New evidence from Henderson Island. Journal of Field Archaeology 24(1). Weisler, M., P. Kirch, and J. Endicott In press. The Mata'are basalt source: Implications for prehistoric interaction studies in the Cook Islands. Journal of the Polynesian Society. Weisler, M., J. Sinton, and Y. Sinoto In prep. Mfh Polynesian basalt artifact geochemical database. 13 FAUNAL ASSEMBLAGES FROM THE TO'AGA SITE LISA NAGAOKA THE CALCAREOUS SAND depositional environment at To'aga favored the preservation of faunal matenial (table 13.1). Over 165 kilograms of invertebrate remains were recovered, represening more than forty families. The To'aga fish-bone sample-the largest in Western Polynesia contains 2,196 identified bones across twenty-two taxa. Pig, chicken, rat, marine mammal, urle, and bird comprse the 687 bones of the non-fish vertebrate sample. Each component of the faunal assemblage is described in detail below. The problem of recovery bias is addressed here through the analysis of bulk sediment samples from the To'aga excavation. Generally, the use of smallersized screens increases the size of the faunal sample and the number of taxa recovered. The To'aga bulk samples were sieved through different screen sizes to determine the effects of screen size on the composition of the faunal assemblage. Cuwent knowledge of Western Polynesian subsistence practices is limited since few zooarcheological studies have been conducted in the region. In this context, the To'aga faunal assemblage is important for adding new information to our understanding of regional subsistence trends. The long temporal sequence at To'aga allows for an assessment of changing subsistence patterns. Despite the small sample of Western Polynesian sites, comparisons of the To'aga assemblage with other regional faunal assemblages may yield information about subsistence patterns. METHODS The faunal remains were recovered by dryscreening all excavated earth (except the clayey colluvial sediment) through 1/4" mesh. To determine the feasibility of screening the colluvium, Layer I of Unit 20 was screened rough 1/4" screens. Only one poody preserved Turbo shell was recovered from the 0.8 m3 sieved. A decision was made not to screen the colluvial layer in other units because of the difficulty in dry-screening the matrix, and because of the low density and poor preservation of faunal and other material in this clayey deposit (See chapter 7 for further discussion of the pH and other aspects of the To'aga site sediments.) During the 1987 field season, most of the shell recovered was identified, weighed, and discarded in the field. Voucher samples and the remaining unidentified shell and bone were shipped back to the laboratory. All faunal materials recovered from the 1989 field season were washed in the field and returned to the laboratory for identification and analysis. Fish remains were identified to the family level using reference collections from the Bishop Museum, and the personal collections of Patrick Kirch (U.C. Beikeley), Melinda Allen (University of 190 The To'aga Site Table 13.1 Summary of To'aga Site Faunal Remains Faunal Class Tota Shel (kg) Identified Shell (kg) Tota Non-fish (NISP) Tota Fish (NISP) Identified Fish (NISP) Washington) and the author. Reference collections from the Bishop Museum were used to identify rat, dog, pig, manine mammal, and marine turtle. The bird component was identified by David Steadman of the New York State Museum (see chapter 14). Although many Pacific faunal analysts use MNI (minimum number of individuals) to quantify vertebrate remains (e.g., Leach 1986; Anderson 1986; Green 1986), we use NISP (number of identified specimens) for the To'aga vertebrate assemblage. The problems with both measures have been discussed extensively elsewhere (see Chaplin 1971; Grayson 1979, 1984; Payne 1972). NISP was chosen here because the effects of aggregation make MNI an inconsistent measure. Although the problem of interdependence affects NISP, the measure is constant across aggregation units. The invertebrate faunal component was identified using standard shell identification guides (Abbot and Dance 1986; Hinton 1972; Eisenberg 1981). Tentative identifications were confirmed using reference collections at the Burike Museum in Seattle and at the Bishop Museum in Honolulu. Invertebrate remains were quantified by weight. As with MNI and NISP, use of weight has its drawbacks because of variations in size and density of different shell taxa. For example, Tridacna maxima, the giant clam, has a very dense shell, and one large individual may weigh more than 1 kg. On the other hand, shells such as limpets (Patellidae) are very light, so that many individuals may account for a small Excavation Units 1-14 15-30 Total 50.291 50.291 118.367 115.669 168.658 165.960 322 365 687 3462 6062 9524 723 1473 2196 amount of weight. As a result, heavy shells may be overrepresented and light shells undernepresented in any sample. Post-depositional alterations such as leaching and fossilization also may distort shell weight RESULTS The To'aga faunal data are presented in three categories: the vertebrate component, which is subdivided into fish and nonfish, and the invertebrate component Of the thirty excavation units, complete data by stratigraphic layer from thre areal excavations (1987 Main Trench, Units 20/23, and Units 15/ 29/30) are presented in the text for comparison. These units were chosen to represent the site because they comprise a larger sample than the individual units. Data for the remaining excavation units are presented only in the summary tables for the separate faunal categories. The complete faunal data by stratigraphic layers from all excavation units are available from the author on request. Fish Remains The To'aga excavations yielded 2,196 identified fish bones representing twenty taxa (tables 13.2-6). Although this is the largest archaeological fish bone assemblage from Western Polynesia, an average of only 73 bones were identified for each excavation unit. Acanthuridae (surgeonfish), Diodontidae FawuaIAsbages 191 Table 13.2 Fish Fauna Recovered from the 1987 Excavations (NISP) Excavations Units Taxa 1, 4-9 3 2 10 12 11 13 14 Total - 4 9 328 80 -- 1 13 67 2 -- 2 15 67 1 6 -- 1 5 -- 1 3 -- -- 6 -- -- -- 2 -- -- 2 38 32 28 -- -- -- -- -- 1 17 1 -- -- -- -- 2 - -- -- -- 7 1 -- -- 1 Diodontidae Holocentridae 289 64 1 Acanthuridae 48 47 24 22 24 15 9 5 9 8 5 3 2 2 1 1 -- -- 4 -- -- 1 1 -- -- Seffanidae Scaridae Carangidae Balistidae Muraenidae Labridae Ostaciidae Lutjanidae Aulostomidae Congridae Elasmobranchii Lethrinidae Belonidae Kyphosidae Sphyraenidae Scombridae Bothidae TOTAL IDENTIFIED UNIDENTIFIED TOTAL 4 -- 15 -- 3 1 -- -- - -- ~ -- -- - -- -- -- -- 4 577 2003 16 2580 20 (spiny puffers), Holocentridae (squinelfish), Serranidae (groupersfcods) and Scaridae (paotfish) comprise appoximately 78% of the identified fish remains (fig. 13.1). These taxa are usually the most almiant across time and space at the To'aga site. 1 7 45 1 52 13 12 11 8 7 5 3 2 1 1 1 1 2 --- -- 1 19 2 -- -- -- -- -- 38 190 228 -- 1 -- 1 1 -- -- --- -- -- --- 3 2 5 1 10 19 29 82 464 546 722 2739 3461 The stucture and composition of the Toaga fish-bone data pwbably reflect a conbination of methodological, environmental, and cultuml factors. Methodological factors include recovely bias and problems in identification and quantification. Bias 192 The To'aga Site 3ma 4 >! 8 oo ) (71, w m xn N n x Ch m m _N O _ N _ o _ o _ o N - : C- ', N ct 4 U) O F m _ _xo x %D on o vo It N m 2 m ~o o vo t _4 _ I mQ 7 _; tn Q2 1. t S: ON m 9 X0 m e4 C4 e N V-4 V4 -_ _ 0 _- a _ 1. !4 - oo A r)o 00 9- V- : : : - N : - 00 V- 00 W 0 e4 la A0 II "t tn 6 e _- I I I I - - I I I I t ,m A "-4 II C CU _-4 kn Cie;IaC en ;O. - I I tI I II "I : II : I I : I I I I : I : I I - a - : !s) Cu 0x - "-4 en lqt I I C4 ON I.t I I -4 (I "14 00 0 ) 00 .o o Wo 0 T-4 1 e> - C4 V-4 00 ar - - V-4 a "4 a *;0 - v a a a : : ~ a C V-4 en Cf) dl~t - en en - - | U 0~ -- 4] 3V3 8 -4 - 0 ; "-4 tn r Faunal Assemblages 193 Table 13.4 Fish Fauna from the 1987 Main Trench (Units 1, 4-9) Taxa Diodontidae Holocentridae Acanthuridae Serranidae Scaridae Balistidae Carangidae Muraenidae Labridae Lutjanidae Aulostomidae Ostraciidae Congridae Elasmobranchii Lethrinidae Belonidae Kyphosidae TOTAL IDENTIFIED UNIDENTIFIED TOTAL HA-I 4 3 3 5 2 4 2 --- 2 2 --- --- Layers HA HB 25 17 13 11 1 8 4 4 3 1 1 1 154 25 20 17 21 10 10 4 2 5 4 3 1 HC Total 106 19 12 14 1 2 6 7 2 1 3 1 3 1 289 64 48 47 25 24 22 15 9 9 8 5 5 3 2 2 1 578 2003 2581 ----- 1 2 --- --- 2 --- 1 1 --- --- --- ----- 28 191 219 93 535 628 --- in the recovery process is shown to affect the sample size and the taxa represented in the assemblage (see "Bulk Samples" section). Problems in the identificaion process include the quality of the reference collection, which can limit the accuracy of the identifications and the number of taxa represented. For the To'aga assemblage, several distinctive mouth parts could not be identified using the reference collection at hand. With a better reference collecton, subfamily identifications may also be possible. Another methodological problem is the inclusion of "special bones" in the NISP count. A few taxa are identified mainly by special bones that can number up to 300 per individual, thus greatly inflating he NISP count. This is especially true for 1 279 827 1106 178 450 628 Diodontidae, which can have more than 250 spines per individual, and to a lesser extent for Ostraciidae, Elasmobanchii, and Balistidae. Of the 923 Diodontidae bones identified at To'aga, 901 were spines and only 22 were mouth parts, most being concentrated in Unit 21 and in the 1987 main excavation trench. If the Diodontidae spines are removed from the NISP count, the ranking of diodonts drops from one to thirteen, and the shape of the graph changes (fig. 13.2). Although the presence of this poisonous fish may seem odd, its remains are common in middens across the Pacific (e.g., Allen 1990; Butler 1987; Masse 1989). Moreover, the fish is still eaten by some modem Pacific populations (Bagnis 1972, Masse 1986). 194 The To'aga Site Table 13.5 Fish Fauna from Transect 5, Units 15/29/30 (NISP) Layers Taxa H HA-1 Diodontidae Lutjanidae Labridae 9 9 11 3 3 3 6 Ostraciidae --- Lethrinidae Muraenidae IIIB 9 8 4 HID Total --- 10 3 2 2 1 16 15 7 5 8 5 1 3 2 --- --- --- --- --- --- Carangidae --- --- 1 Mullidae Balistidae --- --- 3 1 2 2 --- --- --- AcanthuridM Serranidae Holocentridae Scaridae Kyphosidae Bothidae Congridae TOTAL IDENTIFIED UNIDENTIFIED TOTAL --- 1 --- 1 --- 1 ----- --- 1 1 --- --- --- --- --- --- 48 201 249 3 16 19 42 250 292 1 67 297 364 The structure of the To'aga fish-bone data also may reflect natural distributions and abundances of fish taxa. Most of the To'aga assemblage can be classified as inshore fishes, although a few families such as Serranidae, Lutjanidae, and Carangidae cover a wide range of habitats. Reef ecosystems are generally more diverse and have a higher pwductivity rate than open ocean environments; therefore, the abundance of inshore versus pelagic fish may reflect the natural diversity of the different environments. Fishing strategies may also be reflected in the 35 32 23 18 15 10 9 4 4 3 2 2 2 1 1 1 160 764 924 fish-bone data. The fishhooks recovered from the site (see Kirch, chapter 1 1) may have been used to catch serranids, holocentrids, and lutjanids, but probably not scarids or acanthunds which are more likely to be caught by netting or spearing. A comparison of modem Samoan reef exploitation (Hill 1986) and the To'aga fish data shows that the most abundant taxa in the archaeological assemblage can be caught by several fishing techniques (table 13.7). These taxa may have had more opportunity to be caught than taxa for which only one stsegy was used. FawalAssemblages 195 Table 13.6 Fish Fauna from Transect 9 (Units 20/23) (NISP) Layers Taxa Acanthuridae Serranidae Diodontidae Muraenidae Holocentridae Scaridae Labridae Ostraciidae Carangidae Lutjanidae Congridae Aulostomidae Lethrinidae Balistidae Scombridae Elasmobranchii Bothidae Kyphosidae Mullidae TOTAL IDENTIFIED UNIDENTIFIED TOTAL IIB HIA HIIB uic IV Total 3 2 3 3 1 2 30 18 --- 53 5 14 37 13 9 --- 53 3 23 3 14 --- 1 14 --- 1 15 3 9 9 5 --- 1 1 2 2 2 1 12 9 2 44 --- 38 8 9 1 1 27 25 6 --- 6 13 2 --- --- --- 22 22 14 12 2 --- 7 5 1 2 3 --- --- --- 1 4 2 - 16 63 79 Non-Fish Vertebrate Remtains The non-fish vertebrate sample of 687 bones is small, averaging only 23 bones per excavation unit (tables 13.8, 13.9). About half the sample consists of Rattus exulans, the Pacific Rat, with nearly half the rat bones coming from Layer II of the 1987 Main ----- 5 4 3 1 --- 1 1 1 34 191 90 161 610 801 195 6 1 1 10 7 414 17 46 348 1294 504 63 1642 Trench. Bird bones were the second most abundant, with 139 bones. Steadman of analysis the 72 identified bird bones presents in chapteran14. Fifty-six marine turtle-bone fragments were scattered throughout the excavations with one-third of the sample concentrated in Layer IIIB of Unit 20, dating to about 2900-2400 B.P. From this same time period, 196 The To'aga Site Table 13.7 Modern Samoan Fishing Methods (after Hill 1986) Taxa Gleaning Linefishing Diving/ Spearing Gillnetting Throw- netting Holocentridae Serranidae Acanthuridae Mullidae Lutjanidae Muraenidae Letbrinidae Carangidae Scaridae Mugilidae Labridae three-fourths of the marine mammal bones were recovered from Layer HIC of Unit 15. Another concentration of marine mammal bones was associated with the 'ili'ili paving found in Layers I and II of Unit 22. No fruit bat bones (Pteropus sp.) were identified from the site, although the fruit bat is present on Ofu Island today. Of the domesticated animals, only 1 pig (Sus scrofa) tooth and 15 chicken (Gallus gallus) bones were identified. The chicken bones are concentrated in Layer IIIB of Unit 20123. Generally, the non-fish vertebrate bones were very fragmented and difficult to identify to species or even class. Thus, 44 bones were placed in the "general vertebrate" category and 40 in the "general mammal" category. Many of the bones placed in the mammal category may be either pig or dog, but a distinction between the two could not be made. Invertebrate Remains As is true for most Pacific island faunal assemblages, the invertebrate component dominated the To'aga faunal assemblage. The densest concentra- tions of shell midden were recovered from layers that dated to two periods of time and contained eitlr 'ili'ili paving or features inteipreted as food preparation areas. For the period of 2500-1900 cal B.P., the instances of concentrated midden are dispersed across the site. Unit 28, Layer RC of Transect 5 had a shell density of over 7.8 kg/n3. Along Transect 9, about 27 kg of shell midden with a density of 11.9 kg/M3 were recovered from Layer III of Unit 20123 (table 13.10), along with one-third of the marine turtle remains for the site and half the chicken bones. The upper portion of this layer contained a large earth oven. The extension of Layer IIn into Unit 21 contained over 13 kg of shell midden (8.2 kg/M3). The densest midden in the 1987 Main Trench, in Layers IIB and UC, also dated to this time period (table 13.11). In Transect 5, Layer II of Unit 15129/30, interpreted as a cookhouse activity area dating to the period 1641-1477 cal B.P., contained nearly 12 kilograms of midden, a density of 7.4 kg/M3 (table 13.12). This layer is contemporaneous with Layer I of Unit 16 of the same transect which was associated with a dispersed distribution of 'ili'iMi gravel and Fauna! Assemblages 197 45- 40- 30-f .-- t 25- ------- f- E - ------ t D~ 20f- 15-t-------tio-It-. . I t 5- , ----------------------------------------------------------------------------------- rl r.l ,mm I I I II I . I I . 1 I -- I I 1 nv h. . - . 2 3 I 4 5 6 7 8 rl ,§ tl owu Oman . 9 10 11 12 13 14 15 16 17 is 19 20 21 22 I Tuaa Figure 13.1 Relative frequency of fish taxa from the To'aga site, including Diodontidae. contained the densest concentration of midden in the site (13.4 kg/m3). A midden with 7.9 kg/n3 density was also present in Layer IIC of Unit 28, Transect 5. Only a few families make up the majority of the invertebrate assemblage. Over 76% of the 165 kilograms of identified shell consisted of three families, Turbinidae, Trochidae, and Tridacnidae, with Turbo setosus by far the most abundant species. Besides the shell, more than 14 kilograms of slatepecil sea urchin (Heterocentronus mamilaus), comprising over 8.5% of the invertebrates, were recovered from the site. Most of the sea urchins were concentrated in Units 20-24 along Transect 9; about half were associated with the earth oven in Layer III, Units 20/23. The rank order of the invertebrate taxa varies little across time and space. Turbinidae is by far the major taxon in the assemblage with Echinoidea, Trochidae, Tridacnidae, Conidae, Cypraeidae, Muricidae, and Neritidae as secondary taxa. The remaining thirty-seven taxa are minor components, contributing less than 1% each to the assemblage. This high diversity may reflect both cultural and environmental factors. Food choice in foraging often reflects the natural abundance and distribution of resources. However, some of the most abundant taxa in the assemblage, such as Turbinidae, Tridacnidae, Echinoidea, and Conidae, were also used as raw material in the manufacture of artifacts. The abundance of these taxa therefore may reflect these dual uses and species may have been selected disproportionately to their natural distributions. A comparison of natural and archaeological invertebrate distributions through modem marine survey infonmation would be useful in sorting out the influence of environment versus cultural effects on the invertebrate taxa represented archaeologically at To'aga. ANALYSIS OF BULK SAMPLES Archaeologists screen sediments in order to increase comparability within and between sites by systematically sampling the archaeological record. The To'aga Site 198 OWN-01 zu- f- 18-4 16-4 -----1-- 144- --------i ci) 12-It 10-fr 1-.. - 01) 8- 4-0 I........ ." . .. . . .. 6X::.......... : 1 : I : 1 1 1 " - I - I . ................................................. "' - .. . - - I - - I - - I - I - I .-=.-. C 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17- 18- 19- 20 21 - 22 Taxa Figure 132 Relative frequency of fish taxa from the Todaga site, excluding Diodontidae. Recovery methods can greatly influence the kinds and the amount of matenal retrieved from excavation (Grayson 1984). As with other sampling techniques, the size of the screen used is determined by the research problem. Pacific island archaeology has been oriented toward the recovery of artifacts, such as pottery, that can be readily recovered by 1/4" screens. Unfortunately, the consistent use of 1/4" screens does not always sufficiently sample other classes of archaeological material such as smaller faunal remains. Experiments on the differential recovery of faunal material show that screen size affects the sample size and the number and kind of taxa represented (Thomas 1969; Casteel 1972; Butler 1987). Larger screen sizes bias the sample toward taxa with larger body sizes. The use of smaller screens increases the sample size and retrieves smaller taxa that would otherwise be lost though the larger screens. To determine how our recovery methods influenced the composition of the To'aga faunal assemblage, bulk samples of ca. 5 kilograms (ca. 500 cm3) was taken from Layer IIIA/B of Units 20123 and from Layer II of Unit 30. These layers were chosen because they contained dense midden concentrations. Thus the two samples may not represent the site in general since the recovery rate may be less for areas with a lower midden density. The bulk samples were wet-sieved thrugh window screen in the field to reduce the bulk of the sediment for shipping. They were dry-screened in the laboratory though nested -3 phi (8mm), -2 phi (4mm), -I phi (2mm), and 0 phi (Imm) geological sieves. The contents from each phi size were separated into gross categories (rocks, coral, shell, sea urchin, crab, and bone) and weighed (tables 13.13 and 13.14). The bone was then identified and quantified using NISP. The recovery of bone was affected by screen size more than shell. Almost all the bone was recovered in phi sizes -2 and smaller. The shell recovered by screen sizes less than -3 phi was difficult to identify. Furthennore, the weight of the Fawual Assemblages I 0 F- 0% .o fC S N1 Cl 6 c1 I N i tt Stn t- %n C N t t "< I I It I : > -4 t t N- I I I I 00 ' C% II I z II a 0o I - 13 iI cd _l "14 - Is 00 " II I 0 0) Nw t- I ' N -4 e4 "4 'Cuo A oo .00 C- 0 - as X II I F* 0- E "4 00 I "-4 I II I en "4 II I II I0i Al C Clq I II I I II I I I II I II 0 II I I II II I 9-4 V-4 %O C4 t- "4 II I 4) "4 C! en 0ON I II I II I eq II I ClCl Cl Cl4 (4 e4 cO, Cu Cl go en eq II I II I C4 II I II I "-4 iI I I II I II I II I II 0 00 II I "-4 Cl1 "-4 ON V-4 II a SrI I I 4) N 40 * "4 II "-4 2 5-a, I cuoo Cl4 Cl 1^ - a- N8 1 i Cu/ Il I m _ Cl- it I I 1 "-4 " 4. '0 -4 Il I T-4 I~" v N "-4 e a en . tN F _ N 00 C4 "4 0 A - : 00 Cl V-4 - WI Cl N V asd N so "4 Co 0 co cn I q m II~l 0 "_4 _- Ft Cl11 0 V) " -4 a ON "-4 N _4 -4 F0 199 i I t1i, I i Cl4 11 1 V-4 W " V" 4 01 1 V- 9-4 kn C "4 l41 I I I en .o V4 "-4 "4 No 0% Cl4 II 1 1 7 II I -4 V-4 L 00 t" M !F-o m A4 ! (.- !! "I s CZ> 0, c %) % Q -t 14 to x Cs* _F- bC - 2 2 _ -s m 2 0 ,It 200 The Todaga Site Table 13.9 Non-Fish Vertebrate Fauna from the To'aga Site A. 1987 Main Trench, Units 1, 4-9 Layers Taxa HA-1 hA HIB HIC III Total 9 63 -- 5 -- 13 1 1 78 28 1 14 -- -- 88 2 21 2 -- 2 45 -- 1 188 3 54 3 3 251 RaNus exulans Mamnmal Bird Marine turtle Vertebrate TOTAL -- 14 113 -- -- 1 -- B. Transect 5, Units 15/29/30 Layers HIA-1 IIIO HIC HID Total 2 22 12 1 -- -- -- -- -----2 -- 18 2 2 1 --25 -- -- -- 1 11 42 6 4 47 8 20 2 11 8 11 107 IIB IIIA Layers HIB HIC IV Total 5 2 4 5 20 2 -- -- -- 7 4 18 -- -- 2 3 1 -- -- -- -- 34 25 3 34 4 8 15 22 2 85 II Taxa Rattus exulans Mammal Marine mammal Gallus gallus Bird Marine turtle Vertebrate TOTAL 10 8 --- 4 3 -- 25 -- 13 C. Transect 9, Units 20/23 Taxa Rattus exulans Mammal Gallus gallus Bird Marine turtle Vertebrate TOTAL -- 1 6 --- 12 -- 2 1 2 11 Faunal Assemblages Table 13.10 Invertebrate Fauna from Transect 9 (Units 20/23) (weight in grams) Layers Taxa GASTROPODA Patellidae Trochus maculatus Trochus niloticus Trochus spp. Tectus pyramis Turbo crassus Turbo setosus Turbo spp. Turbo operculae Astrea stellare Lunella cinereus Nerita albicilla Nerita picea Nerita plicata Nerita polita Nerita spp. Neritina spp. Tectarius grandinatus Cerithium nodulosum Cerithium spp. Clypeomorus spp. Strombus mutablis Strombus spp. Hipponix spp. Cypraea annulus Cypraea arabica Cypraea caputserpentis Cypraea mappa Cypraea moneta Cypraea tigris Cypraea spp. Natica spp. Tonna spp. Cassidae Cymatium nicobarium Cymnatium spp. Bursa granularis Bursa spp. Drupa ricina Drupa morum Drupa spp. Morula uva Nassa spp. Thais armigera Thais tuberosa (continued next page) I IUB 32.5 0.3 118.4 --- --- 48.2 263.9 8.3 89.9 276.4 792.8 114.6 138.1 2.0 0.4 HLiA --- 235.8 1.2 3.9 21.1 932.9 1561.0 140.8 228.7 7.5 1.0 HIB HIC 0.2 642.6 0.5 23.8 46.5 1734.6 3980.6 135.1 1632.0 34.1 9.1 0.4 394.6 IV Total 0.9 1.2 --- 4.7 --- --- 500.2 2275.0 125.3 894.4 8.9 11.8 58.4 100.9 0.7 10.6 --- 1.1 --- 1.0 5.2 1.0 0.6 3.3 2.2 0.9 1.2 1.5 4.3 2.8 23.2 47.4 38.0 2.7 29.3 44.9 11.4 -- 1.2 8.0 5.6 0.1 1.6 7.5 54.8 19.0 1.7 21.9 15.1 4.7 26.4 --- --- --- 7.7 9.2 16.0 13.7 0.6 2.4 1.3 12.7 6.0 14.4 19.2 68.7 9.0 3.1 6.5 29.9 89.9 62.0 66.3 80.4 23.8 57.6 206.6 --- --- --- 51.5 --- --- 9.8 9.4 --- --- 2.2 37.1 11.3 9.0 --- 41.3 --- --- --- ------- 10.9 141.9 10.9 2.0 473.2 0.5 0.5 6.6 30.2 7.5 --- 50.6 --- --- 11.0 --- --- 11.0 1.0 --- --- 3.3 --- 2.5 4.2 70.1 17.1 0.6 --- 2.4 --- 15.9 --- --- 13.1 --- --- 1.1 7.6 63.3 107.9 56.9 0.1 1.6 68.5 78.6 4.1 33.0 74.3 22.0 52.0 19.2 101.6 171.7 175.4 6.3 3.6 5.9 --- 1425.1 1.7 32.4 67.6 3550.7 8974.2 524.8 2993.7 52.5 22.3 18.0 8.1 67.8 --- 4.0 19.6 10.2 12.9 6.9 25.4 --- 16.9 91.6 29.6 --- 40.3 5.9 1.4 1.2 0.8 48.0 - --- --- 2.5 12.8 89.7 70.9 19.4 6.9 51.8 1.2 16.9 238.5 37.7 201 202 The To'aga Site Table 13.10 (continued) Layers Taxa Th'a-is spp. Cantharus undosa Nassarius spp. Vasum ceramicum Conus argus Conus chaldeus Conus eburneus Conus cf. maculifera Conus spp. Bulla spp. Dolabella spp. Melampus spp. Pythia spp. PELEYCYPODA Mytilidae Isognomon spp. Chama spp. Chlamys spp. Cardiidae Periglypta reticulata Tridacna maxima Quidnipagus palatam Scutarcopagia scobinata Trapezium oblongum Asaphis violaceus Pinna spp. I IIB LiA fIB HIC --- 6.3 27.0 0.3 12.6 46.9 5.1 2.9 22.4 43.4 10.9 1.6 40.9 --- --- --- --- --- -- - -1.2 2.3 20.3 17.1 16.7 6.7 14.5 38.2 13.9 19.8 1.3 --- --- 0.5 --- 27.4 0.2 --- --- 0.4 1.6 110.1 1.4 0.8 11.9 --- --- --- --- 20.0 --- 6.7 0.7 23.9 --- 4.8 --- --- --- 2.2 15.1 14.9 4.7 17.9 10.6 9.0 --- 3.4 81.1 --- 55.2 3.2 5.7 28.0 6.8 1.9 1.6 7.2 1.8 --- 1.1 --- --- --- --- 283.5 18.1 5.1 13.7 12.5 234.4 70.3 75.4 ---35.4 0.3 2.2 ------- --- ---- 147.1 18.3 5.5 21.3 3.3 --- IV 3A ----- 3.6 11.1 Total 113.7 36.2 11.5 117.4 38.2 16.2 41.3 0.5 348.0 23.1 12.0 66.8 21.2 52.5 2.3 8.3 16.9 14.9 0.8 205.4 31.8 29.8 1.8 2.6 0.2 7.6 34.1 1.8 5.5 ----- 822.3 132.6 121.5 13.7 61.5 0.5 --- 919.8 6.8 3052.9 30.7 3742.2 39.3 395.5 54.4 Unidentified 14.2 48.8 116.4 38.4 67.9 285.7 TOTAL (g) 672.5 2676.5 7281.7 13703.0 5875.7 256.7 30466.1 1.90 0.35 0.85 3.15 ECHINOIDEA CRUSTACEA VOLUME (m3) DENSrrY (kg/m3) 25.0 0.40 18.20 0.95 14.24 0.90 6.53 0.30 0.86 8169.5 133.0 5.30 5.75 Faunal Assemblages 203 Table 13.11 Invertebrate Fauna from the 1987 Main Trench, Units 1, 4-9 (weight in grams) Layers Taxa GASTROPODA Haliotis ovina Patellidae Trochus maculatus Trochus spp. Turbo crassus Turbo setosus Turbo spp. Turbo operculae Astrea stellare Astrea rhodostoma Lunella cinereus Nerita plicata Nerita polita Nerita spp. Cerithium nodulosum Cerithiidae Strombus cf. maculatus Strombus cf. mutablis Strombus spp. Hipponix conicus Cypraea arabica Cypraea annulus Cypraea caputserpentis Cypraea mappa Cypraea noneta Cypraea cf. tigris Cypraea spp. Policines spp. Cymatium nicobarium Cymatfidae Tonnidae Bursidae Drupa grossolaria Drupa ricina Drupa spp. Nassa spp. Thais armigera Thais spp. Muricidae Cantharus undosus Nassarius spp. Vasum ceramicum Conidae Bulla sp. Melampus fasciatus Melampidae Pythia scarabeus (continued next page) IC HA-I HA UB UC 13.9 2.0 66.8 145.4 136.9 20.0 65.0 71.3 383.1 1274.5 667.7 973.1 4646.3 2850.6 235.0 313.1 189.5 872.3 1116.1 2045.2 -- --- --- 4.3 7.6 --- --- --- 3.2 0.3 6.7 20.0 --- --- 2.0 15.0 1.2 9.2 9.0 --- --- 11.0 --- 3.0 --- --- --- --- 1.0 11.6 2.6 0.9 7.9 65.4 20.0 34.9 60.0 2.2 15.3 25.6 --- 1.0 0.4 53.4 5.9 25.2 69.8 1.1 --- 1.5 15.2 --- 276.0 3.7 24.1 4.2 19.5 --- --- 2.3 1.6 13.7 6.5 5.9 1.4 48.8 70.0 183.0 2.7 --- --- --- --- --- 0.9 79.2 1.6 0.5 --- --- --- --- --- --- 8.1 4.5 8.7 16.8 43.4 166.5 2.7 65.0 1.4 1.0 1.3 14.9 2.4 15.9 31.6 24.3 9.7 255.4 0.7 6.8 2.0 17.2 16.1 1.7 13.2 3.9 5.0 --- --- 185.8 32.0 18.1 8.4 3.3 0.5 4.1 --- --- 4.6 13.9 3.1 406.3 162.3 2892.8 12.9 --- --- 1.1 20.5 1.0 44.9 1.1 0.7 4.3 9.0 --- --- --- Total 5.0 --- --- IV --- ---- 36.7 5.0 20.0 498.0 60.0 2072.3 --215.2 347.3 HI 11.5 1.1 0.5 15.0 9.3 14.5 1.4 --- 11.0 64.4 48.7 9.5 0.2 30.0 2.2 1.8 285.0 5.5 6.4 68.1 82.5 1.2 1.2 37.0 4.8 0.9 1.5 10788.1 992.9 4403.5 24.8 5.0 13.0 1.9 51.6 139.1 186.5 48.8 149.0 5.0 2.2 45.2 25.6 20.5 10.0 24.3 42.3 0.4 660.2 5.0 19.5 49.3 22.9 64.0 1.7 8.8 10.2 1.4 64.7 2.7 617.2 7.6 6.4 141.8 162.0 13.2 2.8 93.0 1.5 The To'aga Site 204 Table 13.11 (continued) Layers Taxa PELECYPODA Anadara sp. Arca spp. Mytilidae Isognomon spp. Chama spp. Codakia divergens Gafrarium spp. Lucinidae Periglypta reticulata Tridacna maxims Hippopus hippopus Quidnipagus palatam Scutarcopagia scobinata Tellinidae Asaphis violaceus IC HA-1 HA IHB HC HI IV Total 1.0 6.0 226.6 0.7 31.1 0.9 65.0 10.0 59.8 --- --- 1.0 --- --- --- --- --- --- --- ----- 2.0 75.4 69.5 --- --- --- 38.5 --30.0 6.0 40.2 0.7 1.1 0.9 1.0 --- --- --- --- --- --- --- --- --- --- --- --- --- --- ------- ----- 3.0 303.5 22.3 --- --- --- --------- 0.3 3.3 8.0 3.1 1.9 10.0 ECHINOIDEA --- CRUSTACEA --- 21.6 4.8 TOTAL 88.0 3862.7 VOLUME (i3) DENSITY (kg/m3) 0.10 0.88 0.65 5.94 --- 65.0 7.0 --- --- --- --- --- --- 0.6 1698.8 215.0 13.2 10.9 75.0 --- 59.2 1320.4 121.4 10.3 7.5 57.0 0.9 --- 1.4 ----------------- 32.8 5.1 27.5 28.4 52.2 1.9 17.9 0.3 5.3 152.0 45.8 2949.7 10305.2 9255.7 429.8 51.5 26942.6 1.85 5.57 1.85 5.00 1.75 0.25 1.05 0.05 8.20 3.29 --- --- 0.95 3.10 shell recovered by the smaller screens added rela- tively little to the shell recovered from the larger screens. The bulk samples show that the size of the vertebrate sample greatly incmases as the screen size decreases. Only one unidentifiable bone was recovered from the -3 phi screen. The -2 phi screen recovered only a fraction of the material recovered from the -I and 0 phi screens (tables 13.15 and 13.16). Comparisons of the density of identifiable fish bone obtained from the bulk samples to that from the excavation unit illustrate the amount of material being lost trugh the 1/4" screens (tables 13.17 and 13.18). While the standardization of the volume to a cubic meter exaggerates the recovery rate for the bulk samples, it shows that a significant amount of bone may be lost through 1/4" screens. The smaller screen sizes also increase the sample's richness through the addition of new taxa. --- 40.7 ----- 4.8 --- --- 3385.7 336.4 25.7 26.5 150.0 5.4 Balistidae and lizard were not recovered in either excavation unit. Along with Balistidae, four other fish families (Ostraciidae, Muraenidae, Carangidae, and Apogonidae) were added to the Unit 30 data through fine screening. Most of these are smallbodied taxa with small diagnostic skeletal elements that are less likely to be recovered by 1/4" screens. Although the To'aga fish sample is the largest in Westem Polynesia, the analysis of the bulk samples shows that sample size, taxonomic richness, and thus sample representativeness can greatly increase through the consistent use of smaller screens and bulk samples. This point is especially relevant for Pacific island archaeology where the vertebrate samples from most sites have been small. Because the representativeness of the 1/4" sample is suspect, the validity of interpretations based on measures of diversity, such as richness (the number of taxa present) and evenness (the distribution of abundance Faunal Assemblages Table 13.12 Invertebrate Fauna from Transect 5, Units 15/29/30 (weight in grams) Layers Taxa GASTROPODA Haliotis spp. Patellidae Trochus maculatus Trochus niloticus Tectus pyramis Turbo crassus Turbo setosus Turbo spp. Turbo operculae Astrea stellare Lunella cinereus Nerita albicilla Nerita picea Nerita plicata Nerita polita Nerita spp. Cerithium nodulosum Cerithium columna Cerithium spp. Clypeomorus spp. Strombus mutablis Strombus spp. Hipponix conicus Hipponix sp. Cypraea annulus Cypraea caputserpentis Cypraea eburneus Cypraea mappa Cypraea mauritania Cypraea moneta Cypraea tigris Cypraea spp. Policines spp. Naticidae Tonna spp. Cymatium nicobarium Cymatium spp. Bursa granularis Bursa spp. Drupa ricina Drupa morum Drupa rubusidaceus Drupa spp. Morula sp. (continued next page) H HIA-1 HIUB HID Total 2.7 0.4 95.4 1.3 1.9 151.6 872.8 138.9 396.5 57.4 7.8 2.7 3.4 684.2 43.8 14.3 1330.2 5665.1 372.7 3434.5 83.1 11.1 4.0 10.7 45.4 123.6 91.5 100.2 2.6 14.4 35.5 10.4 52.1 8.3 10.6 2.5 54.1 9.8 92.9 0.8 27.6 0.9 383.7 3.1 6.7 --- --- --- 0.8 504.5 42.5 12.4 988.0 4213.8 172.8 2721.2 21.3 2.9 4.0 1.9 29.1 36.0 23.7 56.8 --43.4 2.2 40.9 --- ----- --- 45.0 170.0 13.6 153.6 3.2 145.6 408.5 47.4 163.2 1.2 0.4 ----- --- --- --- 8.5 13.8 82.6 51.7 17.4 2.6 9.2 8.6 5.2 11.2 2.9 1.8 0.3 8.6 --- 0.3 2.5 3.9 16.1 26.0 --- --- --- 5.0 19.9 --29.8 5.4 7.8 1.7 37.5 9.8 24.8 0.8 10.1 --- 0.2 6.5 5.2 8.4 --1.1 --- 0.5 --- 2.7 --- --- ----1.3 --0.8 --- 1.0 0.5 6.7 --- --- 3.8 --- 63.5 --- --- 17.5 --- --- 241.6 9.8 --- 0.9 22.6 2.3 1.8 0.1 7.3 --- 0.3 11.6 7.2 6.2 42.5 5.8 34.9 29.0 13.6 8.3 2.8 ------- ----- ----- 109.7 0.8 4.6 5.4 3.5 4.5 ----- --- --- 9.7 --- 9.7 6.8 --- --- --- --- --- --- 1.1 --- --- 4.5 1.2 17.1 18.0 10.7 42.5 5.8 54.3 35.8 13.6 13.9 4.0 205 206 The To'aga Site Table 13.12 (continued) Layers Taxa Nassa sp. Thais armigera Thais tuberosa Thais spp. Cantharus undosa Nassarius gaudiosis Latirus filamentosa Vasum ceramicum Conus chaldeus Conus eburneus Conus spp. Terebra sp. Bulla spp. Dolabella spp. Pythia spp. Melampus spp. II HIA-1 IB IIID Total 2.3 88.2 111.7 124.5 --7.6 1.0 --- 12.9 25.4 1.9 --- 5.5 156.0 142.0 193.5 34.3 0.6 66.8 53.4 0.5 12.0 401.5 5.7 29.0 1.4 51.4 14.9 11.0 --- --- --- 66.8 42.2 --- --- --- ----0.4 3.6 ----15.0 2.2 60.2 17.4 33.6 21.4 0.6 --7.6 0.5 0.8 36.6 --- --- --- 0.8 7.2 5.3 --- --- --- 0.4 0.2 3.0 3.5 48.0 10.2 0.7 1.2 --- 9.7 1.6 14.8 1.4 26.4 11.2 349.5 5.7 15.7 1.4 --- 1.0 10.0 --- PELECYPODA Arca spp. Mytilidae Isognomon spp. Chama spp. Codakia spp. Periglypta reticulata Tridacna maxima Hippopus hippopus Quidnipagus palatam Scutarcopagia scobinata Trapezium oblongum Asaphis violaceus --- --- 26.9 1170.1 --- 7.7 1.0 0.3 11.0 215.2 --- --- --- 28.9 51.2 4.2 50.7 1.6 2.4 4.2 13.1 13.6 3.6 10.3 1.1 7.1 3.6 34.6 1.7 30.1 0.3 41.5 1703.8 13.6 38.3 77.0 4.2 60.4 POLYPLACOPHORA ECHINOIDEA CRUSTACEA --- --- --- 72.6 17.7 5.0 1.0 28.7 11.1 5.3 215.6 8.8 5.3 321.9 38.6 Unidentified 198.4 19.4 30.5 162.4 410.7 TOTAL (g) 11853.2 514.6 1326.0 3212.5 16906.3 1.60 7.41 0.10 5.15 1.50 0.88 1.85 1.74 5.05 3.35 VOLUME (in3) DENSITY (kg/m3) 1.3 8.9 0.3 2.7 ----- --- --- 3.6 310.8 --- 1.5 Fauna! Assemblages Table 13.13 Units 20/23 Bulk Sample Analysis (weight in grams) Screen Size Class >-3 -2 Rock Coral 522.5 234.4 Shell 56.6 30.7 34.1 19.6 Crab Bone TOTAL 8.6 0.4 950.2 -10 1 133.6 103.3 2.9 136.5 2.5 105.8 --- 1.0 104.5 Table 13.14 Unit 30 Bulk Sample Analysis (weight in grams) Screen Size Class >-3 Rock Shell 138.4 102.2 57.8 Bone --- Coral TOTAL 299.4 -2 49.7 68.5 24.5 0.8 143.5 -10i 10.9 30.4 30.4 3.9 14.8 2.5 32.9 values) is also questionable (Gordon 1991). Ideally, a faunal assemblage should reflect the larger target population of the archaeological record, not simply te archaeological recovery techniques used. TEMPORAL TRENDS IN THE 207 change which has been described for some Pacific island sites is a quantitative shift from the exploitation of wild or naturally occuning resources to a dominance on horticultural production (e.g., the Tikopia case documented by Kirch and Yen [1982] or the Mangaia case described by Steadman and Kirch [1990]). Temporal increases in the frequency of pig, dog, and chicken and decreases in wild vertebrate taxa such as birds, turtle, and marine mammal are taken to indicate this trend. In contrast, the character of the To'aga assemblage changes little over time and does not strongly reflect this kind of shift. Wild taxa are found thmughout the site, and the sample of domesticated animals is too small to draw any firm conclusions. Although much of the wild taxa (especially the birds) are represented in early contexts, most of the chicken bone is also found in those early layers. Thus, there is no clear cut shift from one type of resource use to the other. A corollary of the wild to domesticated fauna hypothesis is the reduction of marine resources with increasing reliance on horticulture (e.g., Janetski 1976, 1980; Kirch 1982, 1988). Resource exploitation and environmental degradation by humans are also suggested to contribute to the decline in marine resources, with decreases in the density of shellfish used to support this hypothesis. Invertebrate density varies at To'aga, increasing then decreasing over time (tables 13.10-13.12). However, the use of density measures may be misleading since changes in density may result from other factor, such as changing rates of sedimentation or shifts in settlement pattern. In sum, the composition of the To'aga assemblage changes little over time. The invertebrate assemblage best illustrates this with a few taxa dominating the assemblage across time and space. A similar trend appears to be evident for the fish assemblage as well. At present, the cause of this pattern is not evident. Some possible causes include the exploitation of naturally abundant taxa from a temporally stable environment, a lack of change in subsistence practices, or a combination of both factors. TO'AGA ASSEMBLAGE REGIONAL COMPARISONS One goal of faunal analysis is the description and interpretation of temporal change in prehistoric subsistence patterns. A pattern of subsistence Comparisons of faunal assemblages from different areas or islands allow for the assessment of 208 The To'aga Site Table 13.15 Vertebrate Taxa Represented in the Bulk Sample From Units 20/23 (NISP) Screen Size Taxa -24 -10i Balistidae 7 5 Ostraciidae Serranidae 3 3 Labridae 1 Holocentridae 1 4 2 2 2 Diodontidae 1 Muraenidhae 1 1 Acanthuridae Scaridae 1 2 Rattus sp. Bird 1 Lizards TOTAL IDENTIFIED 1 18 UNIDENTIFIED 10 196 TOTAL 11 214 * 456 475 Not found in regular 0.25 inch screened samples from this excavation unit. regional trends. To'aga may be compared with other assemblages from well-documented sites in Western Samoa (Green and Davidson 1969, 1974; Janetski 1976, 1980; Lohse 1980; Smith 1976), Tonga (Kirch 1988; Poulsen 1987), and Fiji (Best 1981, 1984; Hunt 1980; Kay 1984). First, the issue of data comparability is addressed to determine the quality of the regional data base. Differences in recovery, identification, and quantification techniques can seriously affect the comparability of data across assemblages (Butler 1988; Nagaoka 1988). If data sets are not comparable, differences between them may reflect methodological rather than regional differences. Once these issues have been addressed, the faunal data are then examined for the faunal 1 19 invertebrate, fish, and non-fish vertebrate categories. For Western Polynesian faunal assemblages, recovery and quantification techniques vary considerably across sites (table 13.19). As was shown in the analysis of the bulk samples from To'aga, screen size influences the kind and the size of the faunal sample. Screen size differences can even change data at a nominal level since smaller screen sizes add taxa. Quarter-inch screens have been the most commonly used although for some sites screen size was not reported. In other cases, several screen sizes were used, but when and where the different sizes were used was not reported. This lack of information makes it difficult to evaluate the comparability of the data. Faunal Assemblages 209 Table 13.16 Vertebrate Taxa Represented in the Bulk Sample from Layer II, Unit 30 (NISP) Screen Size Taxa -20 -10 0 - Balistidae 6 Diodontidae 8 Labridae 2 7 4 Ostraciidae 5 Serranidae Muraenidae 3 2 Scaridae 1 Lutjanidae 1 Carangidae 1 Apogonidhae 1 Rattus sp. 2 3 Lizard 3 1 TOTAL IDENTIFIED 4 34 13 UNIDENTIFIED 18 246 532 TOTAL 22 280 546 * Not found in 0.25 inch screened samples from this excavation unit. NISP was the common technique for quantificadon of the vertebrate component, except for the Fiji sites where MNI was used. For the invertebrates, weight was used except for Lakeba and Naigani. These differences in quantification may not be as severe as screen size differences since, in many cases, there is little difference between quantification techniques at an ordinal level (Grayson 1984; Jaretski 1980). If the data are considered in terms of the rankings of taxa, it may still be possible to make valid comparisons. In the identification process, differences in the reference collections and the diagnostic elements used can also affect the data. Kirch (1988) and Best (1984) noted that their fish reference collections inadequate, limiting the number of possible taxa that could be identified. This problem also exists for the To'aga assemblage. Publication of reference collections and the elements used would help evaluate how these factors have biased the data. Some differences among the faunal samples may be due to the range of diagnostic elements used to identify the assemblages, especially for the fish assemblages. Kirch, Poulsen, and Best used mainly the premaxilla, dentary, and special bones for their fish identifications. For the To'aga assemblage three additional elements, the articular, maxilla, and quadrate, were used. This increased the size of the To'aga sample about fifteen percent and added two were taxa. 210 The To'aga Site Table 13.17 Density of Identified Fish Bone from Layer III A/B, Units 20/23 No. of Identified Fish Excavation Unit Bulk Samunple Sample Volume (m3) Density (NISP/m3) 221 1.3 170 34 0.0005 68,000 Given the problems in data comparability of the Western Polynesian faunal assemblages, only general comparisons between the data sets can be made. The issue of comparability is important for future faunal work in the area. Ideally, a faunal data base would be created in which data from different sites could be easily assimilated into one body of knowledge with new data continually adding to our knowledge of subsistence patterns. Fish For many of the Western Polynesian sites either little faunal material was recovered or the data are poorly reported. The data on fish bones from Western Samoa consists of brief notes on their presence in the sites. Janetski (1976, 1980) mentions 10 fish bones identified from Potusa and an unknown quantity from Falemoa and Jane's Camp. Over 174 grams of fish bone were recovered from Lotofaga (Davidson 1969), but no other data are presented. The best reported and largest samples of fish come from Tongatapu, Niuatoputapu, Lakeba, and To'aga. Kirch (1988) recovered a sample of 231 NISP across 11 taxa from Niuatoputapu. From Tongatapu, Poulsen (1987) identified 15 fish taxa containing 179 NISP. Lakeba produced 323 MI or 1782 NISP, and 21 taxa from the four sites for which the fish component was analyzed (Best 1984). The To'aga assemblage is comparable in size to Lakeba with 2196 NISP and 22 taxa represented. The most abundant fish taxa are inshore/reef fishes, and a few taxa make up the majority of the assemblage (fig. 13.3). The cause of this distribution of fish taxa may be cultural (fishing strategies, food preferences) or environmental (natural abundances and distributions). Unfortunately, the biases created by the recovery techniques, differential preservation, and identifiability may have influenced these distributions. The most common fish families across sites are Scaridae, Lethrinidae, Serranidae, Acanthuridae, and Diodontidae. The dominance of taxa, such as Scaridae, Lethrinidae, and Diodontidae, may be due to preservation and identification bias as much as subsistence patterns. The diagnostic elements of these taxa are very robust and easily identified. It is Table 13.18 Density of Identified Fish Bone from Layer II, Unit 30 No. of Identified Fish Excavation Unit Bulk Samunple 29 52 Sample Volume (m3) 0.8 0.0005 Density (NISP/M3) 36 104,000 Faunal Assemblages 211 To'aga To'aga -a Other (22.0%)- riranidae (17.7%) Other (28.0%). (42.1%) Scaridae (6.6%)- -Acanthuridae (18.2%) Muraenidae Hobcentridae (8.1 %)- (6.0%)I Carangidae (5.2%)Acanthuridae (10.7%)- Fblocentridae (13.7%) Scaridae (112%) '-Serranidae (10.4%) Lakeba Lakeba 101/7/196 101/7/197 Other (6.6%) Balistidae (5.3%) Other (16.7/o)- Scaridae (25.0%) Serranidae (6.6%%) (39.5%) Diodontidae (6.6% Acanthuridae (6.7%)Labridae (10.0%) Lethrinidae (15.0%) Diodontidae (1 1.7%)- Lethrinidae (35.5%) LBalistidae (15.0%) Niuatoputapu Tongatapu Other (22.3°h)- Scaridae (24.6%) Other (1 7.7h)- Acanthuridae (7.8°h)- Labridae (7.8h)- -Lethrinidae (17.3%) Serranidae (8.4%)- Diodontidae (9.5%)- Figure 13.3 '-Gireilidae (10.1%) Elasmobranchii (10.0°h)- Diodontidae (22.5%)-' Percentage composition of fish faunal assemblages from major Western Polynesian and Fijian sites; the two diagrams for the To'aga site are with and without Diondontidae. (42.0%) The To'aga Site 212 Table 13.19 Summary of Recovery and Quantification Techniques for Western Polynesian Faunal Analyses Vertebrate Invertebrate Screen Size Quantification Quantification Site SAMOA Potusa Falemoa Jane's Camp Lotofaga TONGA Niuatoputapu Tongatapu FUI Lakeba Naigani Yanuca Reference 1/4" 1/2", 3/16" 1/4" P/A, NISP P/A, NISP P/A, NISP P/A, Weight Weight Weight Weight Weight Jennings et al. 1980 Jennings et al. 1980 Jennings et al. 1976 Green & Davidson 1969, 1974 1/4" 1/4" NISP NISP Weight Weight Kirch 1989 Poulsen 1987 2.5, 5, 9mm 2.5, 3.5, 7.1mm N/A MNI MNI MNI MNI NISP N/A N/A, information not available P/A, presence/absence therefore more likely that these taxa will be preserved and identified than taxa with less robust and distinctive skeletal elements. The quality of the fish reference collections has also influenced the presence or absence of taxa in the assemblages. Based on edthoarchaeological data, Mullidae and Pomacentridae are among the most abundant fish caught on Niuatoputapu, but none are recorded archaeologically (Kirch and Dye 1979). Kirch (1988:225) suggests that the differences between the modem and archaeological assemblages may be due to the poor quality of the reference collection. The lack of an adequate reference collection is also a factor in the composition of the Lakeba (Best 1981:497) and To'aga data. Finally the use of 1/4" screens may have resulted in the absence of Pomacentridae from te Niuatoputapu archaeological assemblage since these fish have small diagnostic elements. --- Best 1984 Best 1981; Kay 1984 Hunt 1980 NISP, number of identified specimens MNI, minimum number of individuals Tongatapu and To'aga. Fruit bat was recovered from the Fijian sites, from Falemoa, and from Niuatoputapu. The 'wild' vertebrate fauna tend to be from earlier instead of later sites. Some of the largest amounts of turtle and bird were from the Lapita sites, TO-2, NT-90, 101M7/196 and 1017/197. Chicken is the most common of the three domesticated animals. Dog and pig are less abundant, possibly due to the problem of distinguishing between tex two species when the bone is fragmented. The evidence for the presence of the pig, dog, and chicken from initial colonization is scant. Pig is present throughout the Lotofaga sequence, but the basal date of the site is about A.D. 1000. At Tongatapu, only chicken is found in the early site, TO-2. Dog and chicken, as well as a bone tentatively identified as pig, were recovered from the Lapita layers of Yanuca. The best evidence for dte early presence of all three domesticated animals comes from NT-90, on Niuatoputapu. Non-Fish Vertebrates Invertebrates Compared to the fish, the non-fish vertebrate sample is smaller, but better reported (table 13.20). Rat, bird, and marine turtle are found at most sites. Marine mammal was identified at only two sites, The invertebrate component comprises a large proportion of Western Polynesian faunal assemblages. The most abundant taxa vary across sites, Fawsal Assemblages 213 Table 13.20 Summary of Western Polynesian Vertebrate Faunal Assemblages (NISP) Site To'aga Jane's Camp Falemoa Potusa Lotofaga Locus A Locus B Locus C Tongatapu TO-I TO-2 TO-3 TO-4 TO-5 TO-6 Total Niuatoputapu NT-90 NT-91 NT-93 NT-100 Pig 1 2 37 p p p NT-163 Total Lakeba 1o0n/196 1o0n1/97 ioin/47 olin/i32 Chicken 15 p --- 380 Marine Turtle Bird 56 25 139 p p p p p p p Fish 27 p p 3 1 74 2 198 505 12 1 3 16 189 202 4 294 9 2 1 p 10 >39g >74g >61g 17 404 19 1 18 15 474 125 109 42 2 73 167 504 2 72 43 1 1 1 4 27 36 179 71 31 10 1 --- 3 7 3 6 7 1 37 16 --- 6 --- 10 --- --- 2 1 3 1 7 15 46 --- 3 14 3 26 92 19 2 1 3 --- --- 17 20 --- 11 2 1 11 4 1 --- 19 33 1 2196 87 p p p 3 1 93 28 8 22 3 ioin/135 4 5 52 2 69 1 --- 12 34 231 76 60 180 7 1 iOin/2b 1o0n//66 Total Naigani Yanuca Marine Mammal p 47 7 2 Fruit Bat Rat p 10 NT-l10 NT-112 NT-113 NT-135 Dog 4 4 1 1 5 1 2 P = present in unknown quantities 21 2 20 3 4 11 5 2 61 12 9 72. 4 11 323 18 23 214 The To'aga Site but the dominant taxa tend to reflect the marine environment near the site. For example, the exploitation of a sheltered lagoon is reflected in the abundance of bivalves in the Tongatapu assemblage. Other sites contain mainly Turbo or other taxa reflecting the exploitation of a fringing reef environment. Changes in the dominant taxa at Niuatoputatpu, Tongatapu, and Lakeba are also used to indicate changing marine environments. While it appears that the most abundant taxa are good indicators of environment, the influence of environmental versus cultural factors still needs to be determined. As at To'aga, the bulk of the invertebrate assemblages is concentrated in a few taxa. Whether this uneven distribution reflects the exploitation ofnaturally abundant taxa or cultural preferences will need to addressed in future studies. CONCLUSIONS The analysis of the To'aga faunal assemblage has generated a number ofnew questions. Despite the time depth represented at the To'aga site, there is a striking lack of change in the resources exploited. A few taxa comprise a large percentage of the fish and invertebrate components of the assemblage. The overall pattem of high diversity may reflect the exploitation of naturally abundant taxa or culturally preferred taxa. Population studies of marine environments off To'aga would be useful for creating a baseline of natural distributions which could then be compared to the archaeological data. Addressing these and other faunal questions requires data robust enough to compile into a cumulative data base and to use at a level higher than nominal. The To'aga analysis has shown that methodological biases introduced during excavation and analysis can severely affect the data, reducing its robustness. Thus, interpretations must be made cautiously. The problems created by these biases are compounded when data are compiled from different sites into a regional data base. Variability in and among data sets may be attributed to differences in recovery or analytical techniques rather than prehistoric cultural patterning. Faunal analysis can be a useful tool for understanding subsistence practices, an important aspect of prehistoric culture. Its utility in future studies, however, depends upon the commitment of faunal analysts and archaeologists to create quality faunal data. ACKNOWLEDGEMENTS M. S. Allen, P. V. Kirch, and the Bishop Museum kindly allowed the use of their reference collections. V. L. Butler assisted in the fish identifications. E. Marshall at the Burke Museum and R Kawamoto at the Bishop Museum assisted in confirming shell identifications. E. Gordon, K. Stark, and P. Kirch read the numerous versions of this chapter. Funding was provided by Sigma Xi Grants-in-Aid of Research for the development of the author's fish reference collection. REFERENCES CITED Abbot, R. T., and S. P. Dance 1986. Compendium of Seashells. Amen can Malacologists, Melboume, Florida Allen, M. S. 1990. Excavations at the Ureia site, Aitutaki, Cook Islands: Preliminary results. Archaeology in Oceania 25:24-37. Anderson, A. 1986. Mahinga ika o te moana: Selection in the pre-European fish catch of southeem New Zealand. In A. Anderson, ed., Traditional Fishing in the Pacific: Ethnographical and Archaeological Papersfrom the 15th Pacific Science Congress, pp. 151-66. Pacific Anthropological Records 37. Honolulu: Bernice P. Bishop Museum. Best, S. 1981. Excavations at Site VL21/5, Naigani Island, Fiji, a preliminary report. Department of Anthropology, University of Auckland. -. 1984. Lakeba: The prehistory of a Fijian Island. Ph.D. dissertation. University of Auckland. Butler, V. 1987. Fish remains. IN D. E. Lewarch, ed., The Duwamish No. I Site: 1986 Data Recovery, pp. 1-37. URS Corporation and BOAS Inc., Seattle. 1988. Lapita fishing strategies: The faunal analysis. IN P. V. Kirch and T. L. Hunt, eds., Archaeology of the Lapita Cultural Complex: A Critical Review, pp. 99-116. Thomas Burke Memorial Washington State Museum Research Report No. 5, Burke Museum, Seattle. Casteel, R. 1972. Some of the biases in recovery of archeological faunal remains. Proceedings of the Prehistoric Society 36:382-88. Faunal Assemblages Chaplin, R. E. 1971. The Study ofAnimal Bones from Archaeological Sites. New York: Seminar Press. Eisenberg, J. M. 1981. A Collector's Guide to Seashells ofthe World. New York: McGrawHill. Gordon, E. 1991. Differential faunal recovery and models of subsistence change: A Hawaiian example. Paper presented at the 1991 Annual Meeting, Society for American Archaeology, New Orleans, Louisiana Grayson, D. K. 1979. On the quantification of vertebrate archaeofaunas. IN M. B. Schiffer, ed., Advances in Archaeological Method and Theory, vol 2, pp. 199-237. New York: Academic Press. Grayson, D. K. 1984. Quantitative Zooarchaeology. New York: Academic Press. Green, R. C. 1986. Lapita fishing: The evidence from sites in the Reef/Santa Cruz group, southeast Solomons. IN A. Anderson, ed., Traditional Fishing in the Pacific: Ethnographical and Archaeological Papersfrom the 15th Pacific Science Congress, pp. 1 19-36. Pacific Anthropological Records 37. Bishop Museum, Honolulu. Green, R. C., and J. Davidson 1969. Archaeology in Western Samoa, Vol. 1. Auckland Institute and Museum Bulletin 6. 1974. Archaeology in Western Samoa, Vol. 2. Auckland Institute and Museum Bulletin 7. H. Hill, B. 1978. The use of nearshore marine life as afood resource by American Samoans. Unpublished MA. thesis, University of Hawaii, Honolulu. Hinton, A. G. 1972. Shells ofNew Guinea and the Central Indo-Pacific. Rutland (Vermont): Tuttle. Hunt, T. L. 1980. Toward Fiji's past: Archaeological research on southwestern VidtLevu. Unpublished M.A. thesis. University of Auckland. Janetski, J. C. 1976. Dietary remains from Jane's Camp-a midden site. IN Jesse Jennings, ed., Excavations on Upolu Western Samoa, pp. 7582. Pacific Anthropological Records 25, Department of Anthropology, Bishop Museum, Honolulu. 1980. Analysis of dietary remains from Potusa 215 and Falemoa. IN J. Jennings and R. Holmer, eds., Archaeological Excavations in Western Samoa, pp. 117-22. Pacific Anthropological Records 32, Department of Anthropology, Bishop Museum, Honolulu. Kay, R. M. A. 1984. Analysis ofarchaeological materialfroom Naigani. Unpublished M.A. thesis. University of Auckland. Kirch, P. V. 1976. Ethno-archaeological investigations in Futuna and 'Uvea (Western Polynesia): A preliminary report. Journal of the Polynesian Society 85:27-69. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Burke Museum, Seattle. Kirch, P. V., and T. S. Dye 1979. Edmoarchaeology and the development of Polynesian fishing strategies. Journal of the Polynesian Society 88:53-76. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Bishop Museum Bulletin 238. Bishop Museum, Honolulu. Leach, F. 1986. A method for the analysis of Pacific Island fishbone assemblage and an associated database management system. Journal of Archaeological Science 13:147-59. Lohse, E. 1980. Excavations on Manono Islet. IN J. Jennings and R. N. Holmer, eds., Archaeological Excavations in Western Samoa, pp. 21-32. Pacific Anthropological Records 32, Department of Anthropology, Bishop Museum, Honolulu. Masse, B. 1989. The archaeology and ecology of fishing in the Belau Islands, Micronesia. Unpublished Ph.D. dissertation, Department of Anthropology, Southern Illinois University, Carbondale. Nagaoka, L. 1988. Lapita subsistence: The evidence of non-fish archaeofaunal remains. IN P. V. Kirch and T. L. Hunt, eds., Archaeology of the Lapita Cultural Complex: A Critical Review, pp. 117-134. Thomas Burke Washington State Memorial Museum Research Report No. 5. Seattle. Payne, S. 1972. On the interpretation of bone samples from archaeological sites. IN E. S. Higgs, ed., Papers in Economic Prehistory, pp. 216 The To'aga Site 65-81. Cambridge: Cambridge University Press. Poulsen, J. 1987. Early Tongan Prehistory: The Lapita Period on Tongatapu and its Relaionships, Vols. 1 and 2. Terra Ausais 12. Department of Prehistory, Research School of Pacific Studies, Australia National University, Canberra. Smith H. L. 1976. Jane's Camp (SUFl-I). IN J. Jennings, ed., Excavations on Upolu, Western Samoa, pp. 61-74. Pacific Anthropological Records 25. Department of Antrpology, Bishop Museum, Honolulu. Thomas, D. H. 1969. Great Basin hunting pattems: A quantitative method for treating faunal remains. American Antiquity 34:393401. 14 BIRD BONES FROM THE TO'AGA SITE: PREHISTORIC LOSS OF SEABIRDS AND MEGAPODES DAVID W. STEADMAN INTRODUCTION A S PART OF A LONG-TERM program to reconstrUct the natural distribution and diversity of birdlife in the South Pacific (figure 14.1), I have sought bird bones from archaeological sites in many different archipelagos. Until recently, the islands of Samoa have not been represented in this data base. In 1987, T. L. Hunt and P. V. Kirch conducted excavations at the To'aga site (AS-13-1), Ofu Island, American Samoa yielding a small sample of bird bones representng at least six taxa (Steadman 1990). This first glimpse ofthe prehistoric avifauna of Samoa showed that at least two species (a shearwater and petrel) had been lost on Ofu since the first anrival of humans more than 3000 years ago. Hunt and Kirch expanded the excavations at To'aga in 1989. This paper reports the entire bird bone assemblage from both the 1987 and 1989 field seasons. The sample of bird bones from To'aga now consists of at least fifteen taxa (table 14.1) and provides a more thorough, although still far from complete, portrayal of the birdlife of ancient Samoa. The bird bones from To'aga have been catalogued in the University of Washington Burke Museum (UWBM) Fossil Bird Collection. Comparative skeletal or oological specimens are from UWBM, the British Museum (Natural History) (BMNH), the New York State Museum (NYSM), and the United States National Museum of Natural History, Smithsonian Institution (USNM). hi the species accounts that follow, "Unit" refers to meter square excavations (designated with Arabic numerals). Roman numerals refer to stratigraphic layers. For details of the stratigraphy, chronology, material culture, and non-bird faunal assemblages ofthe To'aga site, see Hunt and Kirch (1988), Kiivh et al. (1989, 1990), and various chapters in this volume. Unless stated otherwise, the modem distribudons of species within American Samoa are taken from the excellent surveys conducted in 1975-1976 by Amerson et al. (1982ab) and in 1986 by Engbring and Ramsey (1989). Modem distributions for elsewhere in Polynesia are taken from Pratt et al. (1987). The prehistoric records are from Steadman (1989a) and from unpublished data from my recent research; islands preceded by an asterisk (*) represent extirpated populations. SPECIES ACCOUNTS Order Procellariiformes Family Procellanidae Pwffinuspacificus (Wedge-tailed Shearwater) MATERIAL. Humerus (UWBM 1680), T9/ 500E, Unit 21, fiB. Five ulnae (UWBM 1244, 1251, 218 The To'aga Site Table 14.1 Bird Bones from the To'aga Site Taxa Number of Identified Bones Seabirds *Puffinus pacificus *Puffnus iherminieri *Puffinus griseus *Pterodroma rostrata *Pterodroma sp., size of P. heraldica (*)Procellariidae sp. *Sula sula Fregata sp. Anous stolidus Gygis alba Steminae sp. Landbirds Egretta sacra Numenius tahitiensis (M) Gallus gallus (I) *Megapodius sp. Gallicolumba stairii TOTAL Total Seabirds (species/bones) Total Landbirds (species/bones) Total Native Landbirds, without IM (species/bones) % of Bones from Extirpated Species (without IM) 11 2 15 6 2 9 1 2 1 1 1 1 1 16 2 3 15n4 10/51 5/23 4/7 85% I=Introduced species M=Migrant Species * Extirpated on Ofu (*) represents extirpated species, but not necessarily different from those already listed. 1256, 1678, 1679), Unit 9, IIB; Unit 5, IIC; Unit 14, llIa,4; 19/500E, Unit 21, IIB (2 bones). Radius (UWBM 1664), T9/500E, Unit 20, IIIB. Carpometacarpus (UWBM 1630), T3I200W. Unit 27, fI1A. Tibiotarsus (UWBM 1642), T5/100E, Unit 28, HIB. Two pedal phalanges (UWBM 1246, 1248), Unit 4, IIB; Unit 1, IIB. REMARKS. This tropical shearwater is rarely noted at sea today in American Samoa It may nest on Pola Islet (off Tutuila) and on Ta'u, but this has not been confirmed. There are no previous rcords from Ofu. It is widespread in Polynesia today, although nesting islands are few. Other Polynesian archaeological records of PwJnuspac~lcus are from *Ua Huka and *Tahuata (Marquesas), *HuAhine (Society Islands), *Lifuka and * Eua (Tonga), and Tikopia and Anuta (Solomon Islands). Puojinus lherminieri (Audubon's Shearwater) MATERIAL. 2 ulnae (UWBM 1651, 1671), T9/500E, Unit 23, IIB; T9/500E, Unit 21, IIB. REMARKS. Within American Samoa, Puffinus iherminueri nests only on Tau, where at least 200 birds breed in the cloud forest, and on Tutila (colony size unknown). It is uncommon at sea and has not been recorded previously on Ofu. Like P. pacificus, P. lherminieri nests today on relatively few islands scattered across Polynesia. Other Polynesian archaeological records of P. Iherminieri are from *Nuku Hiva, Ua Pou, *Ua Huka, *Hiva Ok, and *Tahuata (Marquesas), *Huahine (Society Prehistoric Loss ofSeabirds and Megapodes I I 0C C C,5 N 0 o 4. *0 J . 0~~~ z~~~~~~~ o w -J ~~~~~~~~~~-J C)( 00J *. (4 t zes* C X * 4 z~~~~~ 0 -0 Z .00 I-.a~~~~~~~~ a . : Y 0 J - u0 Z '0 w 0 Z ;'= - 4 0I | *oa a-J 4 uot~~~~~~~~~~; J z~~~~~ 0. - - i z z w Z 9~~~~~~~~~c *r. 219 220 The Tolaga Site Islands), *Mangaia (Cook Islands), Eua (Tonga), and Tikopia and *Anuta (Solomon Islands). Puffinus griseus (Sooty Shearwater) MATERIAL. Sternum (UWBM 1681), T9/ 500E, Unit 21, IIB. Coracoid (UWBM 1645), T5/ 100E, Unit 29, RIB. Scapula (UWBM 1245), Unit 9, IIB. Three humeri (UWBM 1641, 1643, 1659), T5/100E, Unit 29,1; T5/100E, Unit 28, IIB; Unit 20, RB. Three ulnae (UWBM 1240, 1241, 1647), Unit 1 1, I-8 (2 bones); T5/100E, Unit 30,1H. Carpometacarpus (UWBM 1259), Unit 7, hA. Manus digit (UWBM 1260), Unit 7, HA. Pelvis (UWBM 1674), T9/500E, Unit 21, IIB. Three tibiotarsi (UWBM 1252, 1253, 1652), Unit 6, hA (2 bones); T9/500E, Unit 23, IRA. REMARKS. As reported by Steadman (1990), these bones are larger than those of any species of shearwater that resides today in tropical Polynesia. They agree in all osteological details with bones of Puffinus griseus, a species that probably migrates through the Samoan region today (Harrison 1983:260, 420, Pratt et al. 1987:55), although there are no records from American Samoa. The Sooty Shearwater nests today only on islands off New Zealand, southern Australia, and extreme southern South America (Harrison 1983:260, 420). Most of the nesting localities of P. griseus are temperate or subantarctic, although in Australia nesting occurs at least as far north as 320 40' S on the subtropical Broughton Islands. Three possible explanations for the unexpected presence of Paufinus griseus on Ofu were proposed by Steadman (1990): (1) the bones represent migrant birds taken at sea by fishermen; (2) P. griseus once resided on Ofu and, like other procellarids, was extirpated through predation in their nesting burrows by humans and rats; (3) the bones represent an extinct, resident shearwater that differs specifically or subspecifically from modem P. griseiss but is osteologically very similar. Regarding the first explanation, I am aware of no ethnographic accounts of Polynesians capturing seabirds while fishing. The data now available, particularly that fifteen bones (nine different skeletal elements) are osteologically indistinguishable from modem skeletons of P. griseus, suggest that the second explanation may be corect, even though the oceanographic conditions near Ofu are wanmer than those of the wannest localities where P. griseus nests today. That this large shearwater (or an osteologically indistinguishable subspecies [explanation 3 is not necessarily incompatible with explanation 2]) was once resident rather than migrant on Ofu is indicated by the presence at To'aga of two bones of P. griseus (UWBM 1645,1653) that, based upon porosity of the external surface, are from volant juveniles unlikely to have dispersed far from their place of birth. The fonner residency of P. griseus on Ofu is supported furtxr by the abun e of its bones at To'aga (15 of 74, or 20% of identifiable bones). Sooty Shearwaters are represented as well among the few bird bones from a Lapita site on *Niuatoputapu (Tonga), southwest of American Samoa (Steadman 1990). Pterodroma rostrata (Tahiti Petrel) MATERIAL. Two mandibles (UWBM 1250, 1682), Unit 4, RB; T9/500E, Unit 21, IIB. Scapula (UWBM 1684), 19/500E, Unit 21, IIB. Humenrs (UWBM 1657), T9/500E, Unit 20, IIB. Ulna (UWBM 1242), Unit 5, HA. Tarsometatarsus (UWBM 1247), Unit 4, IIB. REMARKS. This large petrel nests today at a few widely scattered localities in te Marquesas Islands, Tahiti, New Caledonia, and American Samoa. Its presence on the latter is based upon a colony of about 500 individuals discovered in 1976 in the montarn cloud forest of Ta'u, a single nesting bird discovered on Tutuila in 1986, and a few birds heard on Olosega in 1986. An onshore record in 1972 from Taveuni (Clunie et al. 1978) suggests th Pterodroma rostrata may nest in Fiji as well. Elsewhere in Polynesia, P. rostrata has been idendfied from archaeological sites on *Ua Huka and Tahuata (Marquesas), *Huahine (Society Islands), *Aitutaki (Cook Islands), *Lifuka and *Eua (Tonga), Tikopia (Solomon Islands), and New Caledonia. Pterodroma sp. (unknown petrel, size of P. heraldica) MATERIAL Radius (UWBM 1670), T9/500E, Unit 21, RB. Femur (UWBM 1676), Unit 21, RIB. REMARKS. These specimens, although not adequate for species-level identification, represent a species ofPterodroma in the approximate size rae Prehistoric Loss of Seabirds and Megapodes of P. heraldica (Herald Petrel, which often is considered conspecific with P. arminjonana) or perhaps the slightly larger P. exlerna (Juan Fernandez Petrel), both of which are significantly smaller than P. rostrata. In American Samoa, P. eterna is known ordy from uncommon sightings at sea, while P. heraldica nests on TaVu. Regardless of species-level identification, UWBM 1670 and 1676 indicate that a species of petrel other than P. rostrata probably once nested on Ofu. Procelladiidae sp. (unknown petrel/shearwater) REMARKS. These specimens cannot be distinguished from those of Fregata minor or F. ariel, both of which visit but do not nest on Ofu today (total of thirteen birds counted in 1975-76; still recorded as a visitor in 1986). The bones ofF. minor males and F. ariel females are similar in size and difficult to distinguish from each other (Steadman et al. 1990). Both species of Fregata occur fairly commonly in Polynesian archaeological sites. Order Charadriifonues Family Laridae MATERIAL. Coracoid (UWBM 1633), T5/ 100E, Unit 16,11. Five ulnae (UWBM 1243,1258, 1634, 1662, 1669), Unit 9, IIB; Unit 5, IIB; T5/ 100E, Unit 16, I; T9/500E, Unit 20, IIIB; T9/500E, Unit 21, fIB. Radius (UWBM 1249), Unit 4, IIB. Two carpometacarpi (UWBM 1239, 1677), Unit 9, UB; T9/500E, Unit 21, IIB. REMARKS. These fragmentary bones cannot be identified beyond the family level. They do not represent a taxon separate from those already identified. Order Pelecaniformes Family Sulidae Sula sula (Red-footed Booby) 221 Anous stolidus (Brown Noddy) MATERIAL. Humerus (UWBM 1668),T9/ 500E, TP 21, IIB. REMARKS. This tern is common and widespread today in American Samoa as well as most of Polynesia. The Ofu population was ca. 500 in 197576 and was not accurately estimated in 1986. Other Polynesian archaeological records of Anous stolidus are from Henderson (Pitcairn Group); Ua Pou, Ua Huka, and Tahuata (Marquesas); Huahine (Society Islands); Mangaia and Aitutaki (Cook Islands); Niuatoputapu, Lifuka, and 'Eua (Tonga); 'Upolu (Western Samoa); Tikopia and Anuta (Solomon Islands); and Mussau (Papua New Guinea). MATERIAL. Radius (UWBM 1656), T9/500E, Unit 20, IIIC. REMARKS. Sula sula nested on Ofu as recently as 1975-76 (only twenty-five birds), but the small colony no longer existed in 1986. This booby still nested in 1986 on Tutuila and Rose islands. It is widespread in tropical oceans. Other Polynesian archaeological records of S. sula are from *Henderson (Pitcaim Group); *Ua Huka, *Hiva Oa, and *Tahuata (Marquesas); *Huahine (Society Islands); *Aitutaki (Cook Islands); Niuatoputapu (Tonga); and *Tikopia and *Anuta (Solomon MATERIAL. Humerus (UWBM 1631), T5/ 100E, Unit 16,I. REMARKS. This distinctive tern is common and widespread today in most of Polynesia including American Samoa. The Ofu population was ca. 100 in 1975-76, and at least 500 in 1986. Other Polynesian archaeological records of Gygis alba are from Henderson Island (Pitcairn Group), Ua Huka and Tahuata (Marquesas), Huahine (Society Islands), Mangaia (Cook Islands), and Niuatoputapu and 'Eua Islands). (Tonga). Gygis alba (Common Fairy-Tern) Sterninae sp. (unknown tern) Family Fregatidae Fregata sp. (unknown frigatebird) MATERIAL. Ulna (UWBM 1257), Unit 14, llb-5. MATERIAL. Humerus (UWBM 1254), Unit 14, lIa-4. Ulna (UWBM 1675), T9/500E, Unit 21, BB. REMARKS. This eroded, fragmentary specimen represents a tern smaller than either of the above species. UWBM 1257 is approximately te 222 The To'aga Site size of the ulna in Sterna swnatrana orAnous minus, both of which occur today in American Samoa, although only the latter nests on Ofu (just ten birds in 1975-76, and perhaps about the same in 1986). Order Clconiifonmes Family Ardeidae Egretta sacra (Pacific Reef-Heron) MATERIAL. Coracoid (UWBM 1655), T9/ 500E, Unit 22, 1. REMARKS. Egreta sacra resides throughout American Samoa and most of Polynesia today. It is uncommon on Ofu, where only dark-phase birds have been recorded. Other archaeological records are from Nuku Hiva and Ua Huka (Marquesas), Huahine (Society Islands), Mangaia and Aitutaki (Cook Islands), and 'Eua (Tonga). Order Charadriiformes Family Scolopacidae Nwnenius tahidensis (Bristle-thighed Curlew) MATERIAL. Corcoid (UWBM 1635, 1636; originally believed to be two separate bones), T5/ 100E, Unit 15, IIID. REMARKS. This widespread but rather uncommon migrant shorebird has not been recorded previously from Ofu, although undoubtedly it still occurs there occasionally. American Samoan records ofNwnenius tahlitensis are confined to Tutuila, Ta'u, Swains, Rose, and Olosega. Other Polynesian archaeological records for N. tahidensis are from Henderson Island (Pitcaim Group), Ua Huka (Marquesas), Huahine (Society Islands), Mangaia and Aitutaki (Cook Islands), and Tikopia (Solomon Islands). During the autumn wing molt, thirteen of twenty-nine individuals ofN. tahtidensis captured on Laysan (Leeward Hawaiian Islands) were flightless (Marks et al. 1990). This adaptation would seem to be viable only in a predator-free setting. It may have led to reductions in the distribution and abundance of N. ahitiensis following the human colonization of Polynesia. Order Galliformes Family Megapodiidae Megapodius sp. (Unknown Megapode) MATERIAL. Ulna (UWBM 1637), T5/100E, Unit 15, HID. Femur (UWBM 1649), Unit 30, HID. REMARKS. These two fragmentary specimens (fig. 14.2) are most similar quantitatively and qualitatively to modem specimens of Megapodius freycinet, a widespread species that now reaches its eastern limit in Vanuatu. UWBM 1649 is indistinguishable qualitatively from the femur in modem M. freycinet, while UWBM 1637 differs from the ulna of M. freycinet in having a slightly deeper sulcus radialis and in lacking a diagonal trough on the cranial surface of tuberculum carpale. Like the archaeological specimens of Megapodiusfreycinet from *Tikopia (Solomon Islands; Steadman et al. 1990), the specimens of Megapodius from To'aga are at the extreme small end of the size range ofM.freycinet (table 14.2). Among living subspecies of M.freycinet, there is no indication that individual body size decreases eastward in Oceania. For example, the tarsus (in skins) of the Vanuatu population is not smaller han that from Australia (table 14.3). The bones from To'aga are larger than in M. pritchardi (confined to Niuafo'ou, Tonga) and M. laperouse (found orny in Palau and Mariana Islands), but much smaller than in two extinct species recently described from late Holocene archaeological and paleontological sitesM. molistructor of New Caledonia and Lifuka and M. alimentum of Lifuka and 'Eua (Balouet and Olson 1989; Steadman 1989b, pers. obs.). Megapodius pritchardi is the only species of megapode that survives east of Vanuatu. Both To'aga specimens of Megapodius are from one of the site's deepest and oldest strata, Layer IIID in units 15/29/30. This suggests that megapodes may have been lost from Ofu not long after prehistoric colonization of the island. These bones provide the first unequivocal, well-documented record of a megapode from Samoa. There is, however, complicated historical evidence that a megapode, described as M. stain by Gray (1861), may have existed in tbe mid-i 800s on Upolu or Savai'i (Western Samoa). Gray (1861) also described M. burnabyi from Prehistoric Loss of Seabirds and Megapodes 223 C% A Figure 14.2 B Comparison of megapode bones: ulna in ventral aspect (A, B) and femur in lateral aspect (C, D). A, C. Megapodius cf.freycinet, archaeological specimens (UWBM 1637, 1649), To'aga site, Ofu, American Samoa. B, D. M. freycinet, modem specimen, USNM 557015, male, Halmahera, Northern Moluccas. Scale = 20 mm. Ha'apai (Tonga). The true identities of M. stairi and Mf. burnabyi are uncertain because only a single egg -ver was collected of each, and these two eggs cannot be assigned unequivocally to M. pritchardi, M.freycinet, or any other species of megapode (Steadman 1991). One or both of these eggs may represent the same species as the bones from Ofu, which also are at the very lower limit of the size range of M. freycinet and larger than, or at the uppermost size limit of, M. pritchardi. The original, handwritten data slip with the holotypical egg of Megapodius burnabyi in the British Museum notes that this egg was "called the chief s egg' as they are only allowed to eat them." Such a chiefly tabu might suggest rarity of the bird and a knowledge that overexploitation, which probably was the cause of the rarity, would eventu- ally lead to extinction. Alternatively, being called the "chief's egg" could suggest that megapode eggs were prestigious trade items brought to Ha'apai from another island. An extensive exchange network operated among Fiji, Tonga, and Samoa in late prehistoric and early historic times (Kirch 1984:23842; 1988:257-60). The Samoan voyages recorded by Stair (1895) included Fiji and Tonga as well as much of East Polynesia. Bennett (1862:247) noted that the nesting grounds of M. pritchardi on Niuafo'ou were "under the protection of the king or chief, and by his permission only can the birds or eggs be procured." Even if megapodes of the Fiji/Tonga/Samoa region were confined by the nineteenth century to Niuafo'ou, they would have been known to Tongans in Ha'apai, as well as to Samoans. Unless additional evidence comes forth, neither 224 The To'aga Site Table 14.2 Measurements (in mm) of the Femur and Ulna of Megapodius, with Mean, Range, and Sample Size Specimen Locality Megapodious sp. UWBM 1637 (U) Samoa: Ofu: To'aga Site Megapodius sp. UWBM 1649 (U) Samoa: Ofu: To'aga Site M. f cf. eremita Lab #: 66, 68 (U) New Ireland M. f freycinet USNM 556995-557002 557016, 557017 (M) N. Moluccas: Halmahera M. f freycinet USNM 557006-557008 6.5 1 7.7 1 C D E 4.4 1 3.3 1 6.8 1 4.9 1 8.4 1 7.3 1 8.2 7.6-9.1 10 3.8 5.0 4.7-5.3 3.3-4.1 10 10 7.4 7.0-7.9 N. Moluccas: Halmahera 8.2 7.6-8.8 10 8.3 7.7-8.8 10 3.7 4.9 4.6-5.4 3.5-4.1 10 10 7.3 7.0-7.8 10 Sulawesi 7.2 6.9-7.5 2 7.4 7.3-7.6 2 3.2 4.3 4.3 3.2-3.3 2 2 6.5 6.5 2 8.1 1 8.0 1 5.2 1 4.1 1 7.4 1 8.3 8.0-8.6 2 8.3 8.2-8.4 2 5.0 3.8 4.9-5.0 3.6-4.0 2 2 7.8 7.6-7.9 2 5.6 5.4-5.7 2 6.0 5.9-6.1 2 2.8 3.4 3.3-3.5 2.8-2.9 2 2 5.2 5.0-5.3 2 USNM 226175, 226176 (F, M) M. f cumingi BM(NHl) 1862.2.10.2 (U) Philippines: island unknown M. f nicobariensis USNM 19686, 19700 (M) Nicobar Islands M. pritchardi tJSNM 319633, 319634 (IJ) Tonga: Niuafo'ou M. laperouse USNM unnumbered (U) N. Mariana Islands: Rota: Payapai Cave M. alimenturm UWBM 2100 (IJ) Tonga: 'Eua: 'Anatu (Ground-Dove Cave) F = Female, M = Male, U = Sex unknown Column Headings: A. Femur: width at deepest proximo-lateral muscle scar B. Femur: depth at deepest proximo-lateral muscle scar C. Ulna: minimum width of shaft D. Ulna: minimum depth of shaft E. Ulna: width of distal end B 8.2 7.5-8.7 10 557010-557013, 557019-557022 (F) M. f gilberti A 3.4 - 1 2.5 1 6.0 1 4.4 1 10 5.4 1 Prehistoric Loss of Seabirds and Megapodes 225 Table 14.3 Tarsal Length (in mm) from Skins of Selected Subspecies of Megapodius freycinet, with Mean, Range, and Sample Size. Based upon Specimens from BM(NH). Subspecies Tarsal Length M. f layardi Vanuatu: Santo, Vate (F, 5M) 73.8 (71-76), 6 M. f eremita Solomon Is.: San Cristobal, Guadacanal, Bouganville 72.2 (67-77), 6 M. f eremita Papua New Guinea: New Britain (3F, 2M, 3U) 67.1 (62-71), 8 (3F, 2M, tI M. f yorki Australia: Queensland (2F, 4M, 4tJ) 71.6 (70-74), 10 F = female, M = male, U = sex unknown Megapodius stairi orM. burnabyi should be regarded as certain records of indigenous populations ofmegapodes in nineteenth century Samoa or Ha'apai, although this cannot be ruled out. The survival of M. pritchardi on Niuafo'ou has been due to chiefly control of exploiting eggs and birds at the conspicuous nest mounds, as first described by Bennett (1862). The people of Niuafo'ou must have realized that conserving megapodes, which probably occurred nowhere else in the region, would help to maintain their share of commerce in the SamoaTonga-Fiji trade network. Family Phasianidae Gallus gallus (Chicken) MATERIAL. Sternum (UWBM 1261), Unit 11, 1. Coracoid (UWBM 1654), T9/500E, Unit 23, IIIB. Scapula (UWBM 1663), T9/500E, Unit 20, HIB. Ulna (UWBM 1255), Unit 14, Illa-4. Radius (UJWBM 1660), T9/500E, Unit 20, IhB. Two pelves (UWBM 1665, 1686, 1688; the last two originally believed to be separate bones), T9/500E, Unit 20, RIB; T9/500E, Unit 21, IIB. Two femora (UWBM 1666, 1667), T9/500E, Unit 20, hIB. Four tibiotarsi (IJWBM 1632, 1644, 1658, 1661), T5/100E, Unit 16, 1; T5/1OOE, Unit 29, IIIB; T9/500E, Unit 20, IIB; T9/500E, Unit 20, IhB. Two tarsometatarsi (UWBM 1646, 1683), Unit 29, rIB; Unit 21, II. REMARKS. Feral and/or domestic populations of Gallus gallus occur neatly throughout Polynesia, including all inhabited Samoan islands. All chickens recorded on Ofu in 1986 were near human habitation and not from deep within forests. This non-native species has been found thugh virtually all of Polynesia in archaeological sites of any age, except that it is absent from all sites on Henderson Island (Schubel and Steadman 1989). Chicken bones occur throughout the To'aga sequence, but are especially well represented in the Layer IIIB occupation in Units 20/23 along Transect 9: This indicates that G. gallus was a commonly eaten bird during the Ancestral Polynesian phase, ca. 2500 yr B.P. 226 The To'aga Site Order Columbiformes Family Columbidae Gallicolwnba stairii (Shy Ground-Dove) MATERIAL. Humerus (UWBM 1638), T5/ 100E, Unit 15,111D. Ulna (UWBM 1658), Unit 30, IIID. Tarsometatarsus (UWBM 1690), T5/100E, Unit 29, IIID. REMARKS. This species occurs only in very old deposits at To'aga, primarily Layer IIID in Units 29/30. Gallicolwnba stairii is extremely rare on Ofu today, with a very roughly estimated 100 birds surviving in 1975-76. Only two or three grounddoves were seen on Ofu during the 1986 surveys; no population estimate was made. Within American Samoa, only perhaps on Olosega does another small population of G. stairii survive. Similar declines or losses of populations of G. stairii have occurred in Tonga, where the only other archaeological record of G. stairii is from *'Eua. DISCUSSION Virtually all of the bird bones from To'aga are broken, often with both articular ends missing. Most of these breaks are not fresh, although often it is difficult to distinguish whether human or sedimentary processes have caused the breakage. A few bones are rounded, suggesting some post-mortem sedimentary transport. The breakage and rounding might indicate that the calcareous sands at To'aga represent a somewhat higher energy deposit than the calcareous sands at certain other Polynesian archaeological sites, such as Hane (Ua Huka, Marquesas) or Tongoleleka (Lifuka, Tonga). Among the shearwaters and petrels, however, a systematic butchering technique is suggested by the fairly consistent pattern of both ends of the humerus, ulna, and tibiotarsus being broken off. Two of the chicken bones had been chewed by rats. None of the bird bones seems to have been modified into recognizable artifacts. Among indigenous, resident species recorded from the To'aga site, five of ten seabirds and one of three landbirds are extirpated on Ofu (table 14.1). At least two of the surviving species (Sterninae sp., Gallicolwnba stairit) exist today on Ofu only in very small, threatened populations. Should these species be lost from Ofu, the proportion of bones of extir- pated species at To'aga would increase from 85% (table 14.1) to 93%. The majority of bird bones from To'aga (46 of 74, or 62%) are of at least five species ofpetrels or shearwaters (table 14.1), none of which nests on OfN today (Amerson et al. 1982a:90). Only two of hse species (Audubon's Shearwater and Tahiti Petrel) are known certainly to nest today anywhere in American Samoa. As seems to be case throughout Polynesia (Steadman 1989a, Dye and Steadman 1990), the island of Ofu had a diverse and probably abundant seabird fauna when humans first arrived. In the case of Ofu, not a single species of procellariid has survived the three millennia of human occupation. Compared to avian assemblages from other Polynesian archaeological sites, the dominance of procellariids at To'aga would characterize a fairly early site, i.e., one that dates to within the first thousand years of human occupation (Dye and Steadman 1990). When compared to sites that seem to represent the initial human occupation of an island, however, such as the Hane site (Ua Huka, Marquesas; Steadman 1989a) orTongoleleka site (Lifuka, Tonga; Steadman 1989b), the To'aga site's lower percentage and diversity of bones from native landbirds and higher percentage of chicken bones would suggest that this site may not represent the first 500 years of human occupation of Ofu. This suggestion is compatible with the radiocarbon chronology at To'aga which indicates occupation of the site from about 2800 to 1900 yr B.P. (see Kirch, chapter 6). Although the bird bones from To'aga provide much new data on the prehistoric distribution of Samoan birds, a bone sample about an order of magnitude larger would be necessary to provide a fairly complete picture of the past birdlife of Ofu. That six of the fifteen taxa of birds from To'aga are represented by only a single bone indicates that more species await discovery if a larger bone sample were available. The 1987 sample of twenty-three bones yielded six taxa, which increased to fifteen taxa with the addition of fifty-one bones from the 1989 excavations. The point of diminishing returns is difficult to determine, however, from comparison with assemblages from other sites. For example, thirty-five species of birds are represented in a sample of ca. 350 bones from the Fa'ahia site (Huahine, Society Islands; Steadman and Pahlavan Prehistoric Loss of Seabirds and Megapodes 1992), while thirty-nine species of birds ae represented in a sample of ca. 12,000 bones from the Hane site (Ua Huka, Marquesas) (Steadman 1989a, pers. obs.). Twelve of the species from Hane are known from either one or two bones. Another gauge of the incompleteness of the To'aga avifauna is the low percentage of species in the modem avifauna of Ofu tha are represented at the site. In this case, only two of sixteen possible species of resident landbirds (12.5%) ar represented. The To'aga site has given us an intriguing introduction to Samoa's prehistoric birdlife. Our understanding of the relationship between native birds and the first human inhabitants of Samoa undoubtedly will improve as more early archaeological sites are discovered and carefully excavated. ACKNOWLEDGEMENTS This research was supported in part by National Science Foundation grant BSR-8607535. I than T. L. Hunt and P. V. Kirch for inviting me to study the bird bones from To'aga. L. Nagaoka and D. S. Pahlavan assisted in curation. S. L. Olson and M. P. Walters kindly shared information on megapodes. For comments on the manuscript, I thank T. L. Hunt, P. V. Kirch, N. G. Miller, S. L. Olson, and M. P. Walters. This is contribution number 670 of the New York State Museum and Science Service. REFERENCES CITED Amerson, A. B., Jr., W. A. Whistler, and T. D. Schwaner 1982a. Wildlife and Wildlife Habitat ofAmerican Samoa. 1. Environment and Ecology. U.S. Dept. of Interior, Fish and Wildlife Service, Washington D.C. . 1982b. Wild!ife and Wild0lfe Habitat ofAmerican Samoa. H. Accounts ofFlora and Fauna. U. S. Dept. of Interior, Fish and Wildlife Service, Washington D.C. Balouet, J. C., and S. L. Olson 1989. Fossil birds from late Quatemary deposits in New Caledonia. Smithsonian Contributions to Zoology 469. Bennett, G. 1862. [Letter from Dr. G. Bennett on Didunculus, Tallegalla, and Megapodius.] Proceedings of the Zoological Society of London 1862:24648. Clunie, F., F. C. Kinsky, and J. A. F. Jenkins 1978. 227 New bird records from the Fiji Archipelago. Notornis 25:118-27. Dye, T., and D. W. Steadman 1990. Polynesian ancestors and their animal world. American Scientist 78:207-215. Engbring, J., and F. L. Ramsey 1989. A 1986 Survey of the Forest Birds of American Samoa. U.S. Dept. of Interior, Fish and Wildlife Service, Washington D.C. Gray, G. R 1861. [pub. April 1862] List of species composing the family Megapodiidae, with descriptions of new species, and some account of the habits ofthe species. Proceedings of the Zoological Society ofLondon 1861:288-96. Hanison, P. 1983. Seabirds: An Identification Guide. Houghton Mifflin Co., Boston. Hunt, T. L., and P. V. Kirch 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal of the Polynesian Society 97:153-83. Kirch, P. V. 1984. The Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge University Press. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Kirch, P. V., T. L. Hunt, and J. Tyler 1989. A radiocarbon sequence from the To'aga site, Ofu Island, American Samoa. Radiocarbon 31:713. Kirch, P. V., T. L. Hunt, L. Nagaoka, and J. Tyler 1990. An ancestral Polynesian occupation site at To'aga, Ofu Island, American Samoa. Archaeology in Oceania 25:1-15. Marks, J. S., R. L. Redmond, P. Hendricks, R. B. Clapp, and R. E. Gill, Jr. 1990. Notes on longevity and Sightlessness in Bristle-thighed Curlews. Auk 107:779-81. Pratt, H. D., P. L. Bruner, and D. G. Berrett 1987. A Field Guide to the Birds ofHawaii and the Tropical Pacific. Princeton: Princeton University Press. Schubel, S. E., and D. W. Steadman 1989. More bird bones from Polynesian archaeological sites on Henderson Island, Pitcaim Group, South Pacific. Atoll Research Bulletin 325:1-18. Stair, J. B. 1895. Flotsam and jetsam from the great Ocean; or, summary of early Samoan voyages and settlements. Journal of the Polynesian 228 The Tolaga Site Society 4:99-131. Steadman, D. W. 1989a. Extinction of birds in Eastern Polynesia: A review of the record, and comparisons with other Pacific island groups. Journal ofArchaeological Science 16:177-205. 1989b. New species and records ofbirds (Aves: Megapodiidae, Columbidae) from an archeological site on Lifuka, Tonga. Proceedings of the Biological Society of Washington 102:53752. . 1990. Archaeological bird bones from Ofu [Manu'a, American Samoa]: Extirpation of shearwaters and petrels. Archaeology in Oceania 25:14-15. 1991. The identity and taxonomic status of Megapodius stairi and M. burnabyi (Aves: Megapodiidae). Proceedings of the Biological Society of Washington 104:870-77. Steadman, D. W., and D. S. Pahlavan 1992. Prehistoric exploitation and extinction of birds on Huahine, Society Islands, French Polynesia. Geoarchaeology 7:449-83. Steadman, D. W., D. S. Pahlavan, and P. V. Kirch 1990. Extinction, biogeography, and human exploitation of birds on Tikopia and Anuta, Polynesian outliers in the Solomon Islands. Occasional Papers of the Bishop Museum 30:118-53. Honolulu: Bemice P. Bishop Museum. 15w SYNTHESIS AND INTERPRETATIONS P. V. KIRCH AND T. L. HUNT PRIOR TO THE MANU'A PROJECIr in 1986, our knowledge of the archaeology and prehistory ofthe most easterly Samoan islands-Ta'u, Olosega, and Ofu-was rudimentary indeed. Limited reconnaissance survey had revealed the presence of stoneand-earth monuments, especially house platfonns, typical of Samoan settlement pattems on the larger islands (Kikuchi 1963). Attempts at excavation on Ta'u Island by a Bishop Museum expedition in 1962 (in one cave site and two open "cooking-house sites') were disappointing (Emory and Sinoto 1965:40-48), and no stratified materals were discovered that might be used to outline a cultual sequence. Although the Bishop Museum team collected 201 basalt adzes and coconut graters from Manu'a, these were all from surface contexts (Emory and Sinoto 1965, table 2). In short, at the commencement of our project, knowledge of Manu'an archaeology was limited to bdef descriptions of urface sites and stone tools, with no time depth. Against this background, our 1986 reconnaissance work thrughout the Manu'a Group yielded a number of significant advances (Hunt and Kirch 1988). The number and range of surface monuments were extended considerably. Peitaps most importantly, several stratified sites were discovered, two of which on test excavation yielded ceramic assemblages radiocarbon dated to between 1950-1850 cal B.P. (Hunt and Kirch 1987). These results confirned our initial expectations that the Manu'a Group would prove to have a deep prehistoric sequence extending back to the Ancestral Polynesian phase, quite probably in parallel with the sequence defined for Westem Samoa (Green and Davidson 1969, 1974). The preceding fourteen chapters have presented the detailed results of two subsequent seasons of concentated field and laboratory research at the most promising of the ceranic-bearing sites discovered during the 1986 reconnaissance: the To'aga site on Ofu Island. In 1987 and again in 1989 we carried out a progrwn of systematic subsurface sampling of te To'aga area, producing an areal and stratigraphic definition of what is currently the largest and most deeply stratified early site known for the entire Samnoan archipelago. Although our research design was of necessity orented first and forenost to considerations of culrl resource management (i.e., the spatial definifion of the site and assessment of its significance), we have also been able to use this research opportunity to expand on certain aspects of our knowledge of Manu'an-and indeed general Samoan-prehistory. For example, we have fonnulated and tested a model of morphodynamic laape change witfi implications for other coastal sites dtoughout the archipelago. Similarly, our various analyses of artifactual and faunal materials 230 The To'aga Site from the systematic tra excavations have enhanced the weconstncdons of early Samoan material cultu, subience economy, and interisland exchange. In this concluding chapter, we attempt to integrate these significant new results from the To'aga site with the existing reconstrucdons and interetations of Samoan prehistory, deriving primarily from work in Westem Samoa (Green and Davidson 1969, 1974; Jennings et al. 1976; Jennings and Holmer 1980) and to a limited extent from rcent work Tutila Island (e.g., Claik and Herdrich 1988; Cark 1989; Leach and Witter 1987; Best et al. 1989). In presenting this synthesis, we are acutely aware that our fieldwork at To'aga has barly begun to tap te immense archaeological potential of this large and complex site. Our excavated sample of 31 m2, while sufficient to give some idea of the extent and range of subsurface deposits and assemblages, represents a very smaU porton of the estimated 21,000 m2 of buried, ceramic-bearing deposits present in the To'aga area. It will take a much larger effort, including the application of time-consuming and costdy horizontal excavation methods, to begin to exploit fully the potential ofthis site to reveal unknown aspects of the Samoan past. on CHRONOLOGY AND CULTURAL SEQUENCE The suite of fourteen radiocarbon age determinations from the To'aga site (see Kirch, chapter 6) constitutes the largest set of dates from a single excavation locality anywhere in the Sanoan archipelago. These 14C dates define a two-millennium long sequence of coastal tenace formation and occupation beginning ca. 3600 cal B.P. and continuing witout pause up to ca. 1000 cal B.P. Although no radiocarbon dates younger than about 1000 cal B.P. were obtained durng our fieldwork, this does not necessarily imply site abandonment dunng the last millenium. Rather, the distribution of radiocarbon dates from To'aga reflects our emphasis on testing and dating the earlier, deeply buried occupation deposits, as opposed to later surface features. Given the presence of vanous aceramic ocwpation pavements and mounds ('ili'ili pavements), grinding sues, bua'i masi breadfruit storage pits, other surface featues dtoughout the To'aga area, and the presence of historic artifacts (see Hunt, chapter 3), it seems likely that To'aga was in fact continuously occupied by prehistoric Polynesians for a full hee millennia ndeed, this is the only site ecorded to date in the archipelago that appears to encapsulate dre entire prehistoric record of the islands, from initial settlement to historc contact The iming of initial human settlement in Westem Polynesia has been a matter of some contention over the years and is direcdy relevant to cunent debates over the rate of dispersal and colonization of the southwestem Pacific by the makers of Lapita pottery (Kirch and Hunt 1988a,b; Spriggs 1990). Kirch (1988:244, table 48) summarized the radiocarbon evidence for initial settlement of Fiji, Tonga, and Samoa, based on such Lapita sites as Natunuku, Yanuca, Lakeba (Site 197), Naigani, To. 2, and Mulifanua. "Almost all of these [14C] ages cluster between about the twelfth and ninth centuies B.C." (1988:244). The ordy site in Samoa which has yielded dentate-stanped Early Eastem Lapita pottery is Mulifanua, which has a single radiocarbon age of 1280-800 Cal B.C. (based on a A-R value of 100 ± 24, and correcion for the oceanic reservoir effect; see Leach and Green 1989). The To'aga site provides additional new evidence for the timing of early, if not iuntial, settlement. (This qualification is necessary, because we cannot be certain whether our systematic trascts actually exposed the earliest occupation deposits in the To'aga area. It is a distinct possibility that earlier strata are preserved under the deep colluvium and talus at te inland edge of the site, which could not be penetrated without the use of heavy mechanical equipment.) Our oldest 14C sample in direct, unquestionable association with ceramics is Beta-35601 from Unit 28, at 1308-930 cal B.C. This age is essentially identical with that from the Mulifanua site and entirely consistent with the suite of dates from other early Westem Polynesian localities. Ihe radiocarbon sample was associated with thinware and thickware pottery, although no dentate-stanped sherds were present. The absence of dentatestamping in this early context may simply be a function of sampling error (only 8 percent of te Mulifanua cermnics were decorated). Alternatively, it is possible that a stratum bearing dentate-stamped pottery lies furter inland, under the impenetrable Synthesis and Interpretations talus. A third possibility is that e decoraton of pottery with dentate-stamping had ceased prior to the colonization of Ofu. Only furter woik at the To'aga site will be able to discriminate among these altemative hypotheses. Despite the absence of "clasic" dentatestamped pottery, it is clear that Ofu Island and the Manu'a Group were colonized within the same general time penod, between ca. 1200-900 cal B.C., as were Eariy Eastem Lapita sites in Westem Samoa, Niuatoputapu, Tongatapu, Futuna, and Fiji. Given that Manu'a lies at the eastem extreme of the Westem Polynesia region, this finding is quite significant. It implies that the Lapita colonization of the entire Fiji-Westem Polynesian region was accomplished rapidly, with no appreciable lag between sites at the westem and eastem boundaries. The settlement of Manu'a by the close of te second millennium B.C. also implies that Lapita populations were poised at the threshold of the vast eastem Pacific, thus raising again te question of wheher ftere was tuly a "long pause" between fte setdement of Westem and Eastem Polynesia (Irwin 198 1; Kirch 1986; Terrell 1986). If highly successful and rapidly advancing island colonizers had moved as far to the east as Manu'a by 1000 B.C., it seems strange that they did not continue eastward into central Polynesia (dte Cook, Society, and Austral archipelagoes) within the next two or thee hundred years. Yet, at present we have no confinned human habitaion sites in central Polynesia dated to the first millennium B.C. (The earliest sites are stfill those in te Marquesas, probably dating to as early as 400200 B.C., although this remains controversial, see Kirch 1986.) In sum, while our radiocarbon dates from To'aga confirm the pattem of rapid colonization of Westem Polynesia by the end of the second millennium B.C., they also heighten the controversy surrounding te subsequent phase of human settlement of the eastem Pacific. Another chronological issue of some concem within Westem Polynesia has been the timing of the cessation of pottery manufacture and use. Poulsen's initial claims for a long sequence of pottery manufacte in Tongatapu have now been revised, and ceranic use appears to have ceased somewhere between about 400 cal B.C. and cal A.D. 50 (Poulsen 1987:83; see also Groube 1971). Based on a large number of 14C dates from several pottery-bearing 231 sites on 'Upolu, Green and Davidson (1974) put the date of pottery disappearance in Samoa after A.D. 300, somewhat later than the Tongatapu sequence. On Niuatoputapu Island, situated between Tonga and Samoa, radiocarbon dates frm Sites NT-93 and -100, containing plainwares, may indicate the persistence of ceramic manufacture and use as late as A.D. 800-900 (Kirch 1988:142, 246). Thus, while tere is a consistent pattem of pottery decline and eventual loss throughout Westem Polynesia, the timing of this process may not be contemporaneous in all locations. At To'aga, the youngest radiocarbon date in direct association with pottery (predominately coarse, thickware) is Beta-35924, at cal A.D. 319473. Sanple Beta-26463, which comes from the base of an acermnic midden deposit (in Unit 3), dates tO Cal A.D. 561-663, while Beta-35600 from an aceramic 'ili'ili pavement, dates to cal A.D. 694-943. Thus, the radiometric evidence from To'aga indicates that the cessation of pottery manufacture and use in the Manu'a Group occurred during te fifth-sixth centuries A.D. If Green and Davidson's dating of pottery cessation in 'Upolu to approximately A.D. 300 iscorrect, there was a lag of 100-200 years in the Manu'a Group. Such chronological differences are certainly not surprising, especially in light of the known ethnographic differences between Manu'a and the rest of Samoa in historic times (Mead 1930, see Kirch, chapter 2). The radiometric chronology developed for the To'aga site now allows the Manu'a Group to be incorporated into a cultual sequence for the Samoan archipelago as a whole. The initial settlement of Manu'a was pene-contemporneous with that at Mulifanua, suggesting that all islands of the archipelago were settled faidy rapidly at the close of the second milleMium B.C. Parallel changes in material culture during the following 1500 years are also evidenced (see furiher discussion below), with the cessation of ceramic manufacture in Manu'a possibly lagging behind 'Upolu by one or two centures. One chronological issue not addressed at To'aga is the development of large monumental architecture (such as star mounds and tombs). (We were not pennitted by the landowners to excavate in or near the Tui Ofu monumental complex situated within the To'aga area.) We suspect, however, that as in Westem Sanoa, these features will prove to date to the last 232 The To'aga Site millenniun of Manu'a prehistory. GEOARCHAEOLOGY AND LANDSCAPE CHANGE A major thstof ourrsearhprogam at To'aga has been the application of a geoarchaeological and geomorphological approach, in order to address questions of site fomnation prcesses. Given that most of fte archaeological deposits and features at To'aga are deeply buried, a geomorphological approach was essential even for such basic objectives as location and subsurface mapping of archaeological deposits. Our aim, however, has been to go furthr and to explore the evidence for both natural and culatal processes of landscape change at To'aga over the thee millennia represented in the prhistonc sequence. To this end, a morphodynamic model for the formation of the To'aga coastal terrace was developed (see Kirch, chapter 4) and has been explicitly tested on (1) stratigraphic relationships (see Kirch and Hunt, chapter 5), (2) radiocarbon dates (see Kirch, chapter 6), and (3) sedimentological analyses (see Kirch, Marning, and Tyler, chapter 7). Here we briefly summarize te major results of this effort, with particular attention to the broader implications for Westem Polynesian prehistory. One specific result of our strategy of systematic trasect excavations at To'aga was the areal definition of te extent of subsurface archaeological deposits dating to the ceramic period. Although our excavation sample is admitedly small, the highly consistent distribution of pottery-bearing deposits in a narrow zone at the base of the talus slope allows us to predict with considerable accuracy the probable extent of these subsurface deposits. These early occupation layers are all associated with former beach-ridge environments, at a time when the shoreline was much closer to the cliff and talus, and when sea level was apparently at a +1-2 m stand. The infenred distribution of pottery-bearing occupation deposits (dating to between ca. 3000-1900 cal B.P.) is plotted in figure 15.1, which also depicts the probable location of the Ofu shoreline during the first milleniuM B.C. It is important to stress, however, that we were unable to determine precisely the inland boundary of these pottery-bearing deposits, as this lies somewhere under the modem talus slope. Ling heavy mechanical equipmnen, it was impossible for us to penerate fte compact mass of colluvium and talus boulders (estimated to be as thick as 10-15 m in places) which caps the inland portons of these early layers. Probably these deposits do not extend any farther inland than 2040 m from our inland-most test units. Based on the distribution depicted in figure 15.1, te total area of buried subsurface archaeological deposits containing ceramics covers a minimum of 21,000 M2, and a maximum of 35,000 M2. By any Polynesian standards, this is a large site with considerable potential for future horizontal excavations. As noted earlier in this volume (see Kirch, chapter 2), the coastal terace nnning along the southem side of Ofu from To'aga to Fa'ala'aga comprises an important land use and resource zone for the island's Polynesian occupants. Given the island's very steep and rugged topography, this stretch of flat land is one of the few areas suitable for intensive arborcultur and for habitation. Yet, our geoarchaeological studies at To'aga clearly demonstrate that this coastal tenrace is a highly dynamic accumulation form which developed into its present configuration only within the past 2,000 years-well after initial human colonization of the island. At the time of initial settlement, the coastal tenrace was very nanrow, consisting of little more than a beach ridge situated directly beneath fte steep oveihanging cliffs and talus. A phase of active progradation of this shoreline did not commence until about 1.9 kyr B.P. This prgradation appean to have been initiated by a rapid fall in sea level (presumably a eustatic fall which is evidenced over much of fte southwestem Pacific at this time, see chapter 4) equalling or slightly exceeding the rate of local subsidence, and thus exposing fte reef flat to erosion and storm surges. The littoral, calcareous contribution to the sediment budget was thereby increased, producing a sufficient volume of sediment to prograde the shoreline between 40-100 meters from its location prior to 1.9 kyr B.P. Following th classification of "accumulation forms" proposed by Zenkovich (1971:95-97, fig. 4.1), the To'aga tenrace is an attached form, specifically type c, a "'tefface formed by the infiling of a concavity (supplied laterally)." In this case, the concavity was formed by the marine cliffinland ofthe site, which was deepest at the southwestem end of the coastal stip. The infilling Synthesis and Interpretations 233 C) \ *I a 4* .0 *bO 8 4o H0 *r 0 C 9o w EC0 _w [I 0 Ct 0 E 0 0 .O 234 The To'aga Site and progradation of the coastal terrace proceeded from southwest to norheast This sequence of infifling is confirmed by our tansect excavations and conrelates withi the early pottery-bearing deposits (daing to 3-1.9 ky B.P.) which are confined to th southwesem part of the To'aga coastal tenre. The norteastem end of the terrace, in the vicinity of Fa'ala'aga Cransect 17), remained a high-energy beach until relatively recently. The morphodynamic model developed and tested for the To'aga site has wider implications for coastal archaeology elsewhere in Samoa and, indeed, on other volcanic oceanic islands (i.e., ose situated on te Pacific Plate). In Manu'a, we would predict that coastl terraces on Olosega and Ta'u islands (for example, the Faga and Saua aeas on Ta'u Island) will prove to have similar dynamic geomorphological histories to that evidenced at To'aga. lTus, any effort to discover occupation sites dating to the period between ca. 3-2 kyr B.P. will rquire the use of subsurface testing to locate the early beach ridge environments that are prdictably buried under later prograded sediments. We believe that it is likely that such bured sites do exist on Olosega and Ta'u. Indeed, a ceramic-bearing site in precisely this kind of bured beach ridge envirnment was identified by Hunt and Kirch at the base of the marine cliff behind Ta'u Village during our 1986 rconnaissance survey (1988:166-67). The morphodynamic model of the To'aga site is also likely to apply, pediaps with minor modifications, to Tutuila Island. (One modification to the model which may be required is the rate of subsidence. Tutuila is somewhat older than Manu'a, and thus may already have passed trugh its phase of rapid subsidence due to point-loading of the thin oceanic crust.) Claik and Herdrich (1988) demonstrat that the 'Aoa Valey has been substantially infilled since initial human occupation, indicated by a pottery-bearing site situated along the interior edge of the modem valley floor and by other geomorphological signs of a former shoreline now well inland of the present coast. They listed several hypotheses to account for this sequence of infilling (1988:17375), including "a lowering of sea level" of about 1-2 m. We would suggest that this hypothesis is the most likely of the alternaives they list, although the contribution of terrigenous sediment to the 'Aoa sediment budget (resulting from human-induced forest clearance and erosion in the interior valley slopes) was doubdess greater ta at To'aga. In any event, the failure of early archaeological efforts to locate ceramic-bearing sites on Tubuila (e.g., Kikuchi 1963; Emory and Sinoto 1965; Frost 1978) was due to the lack of a geomorphologically informed approach to site discovery. It is probable that most if not all of the valley floors as well as the coastal tenrc ofTutuila Lsland have undergone significant infifling and progradation during the past two millennia Thus, early coastal sites are unlikely to be exposed on the surface. It will be essential to woik out local geomorphological sequences of infilling and progradation as an integral part of archaeological survey in these cases. We also predict that similar sequences of progradation and burial of early archaeological sites will be found trughout many of the volcanic high islands of centr Polynesia, particularly in the southem Cooks, Society Islands, and probably the Austals. These are all "hot spot" linear volcanic chains situated on te Pacific Plate, which typically undergo point-loading induced subsidence (Menard 1986:95-99). Steams (1978:286), for example, points to geological and geomorphological evidence for rapid subsidence of the Society Islands during the late Pleistocene (and probably continuing into the Holocene). Furtermore, there is now considerable radiometrically dated evidence for a +1-2 m higher sea level at ca. 4-2 kyr B.P. throughout French Polynesia with a rapid fall to modem level after about 2 kyr B.P. (see Kirch, chapter 4). In short, both ofthe major controing processes that resulted in coastal terrace construcdon on Ofu (subsidence and sea level change) were probably also operating in central Polynesia. This suggests that early occupation sites will not be easy to discover using conventional archaeological surface survey methods. Indeed, the presence of buried, and even partially submerged, archaeological deposits is known for Huahine (Sinoto 1979) and for Mo'orea (Green et al. 1967) in the Society Islands. On Mo'omea, Lepofsky recently discovered anaerobically prerved coconuts buried under 2 m of recent alluvium in the interor of the Opunohu Valley (Lepofsky, Haries, and Kellun 1992). These problems of site survey and discovery raise serious issues for Eastem Polynesian prehistory. Specifically, given the probability of deep Synthesis and Interpretations burial of early coastal sites, te archaeological record for central Polynesia as curnently defined is likely to be highly biased toward later prehistoric sites. It is entirely possible that the early phases of occupation in the Societies and Cooks have yet to be discovered. Just how far back in time we may be able to extend these chronologies after a program of geomorphologically informed, systematic subsurface sampling is launched is impossible to say. It seems conceivable, however, that the "long pause" currently identified between the settlement of Westem Polynesia and the movement of people into Eastern Polynesia (Irwin 1981; Kirch 1986; Terrell 1986) may prove to be an artifact of archaeological sampling bias. Only a concerted effort to work out the late Holocene dynamics of coastal landforms in central Eastem Polynesia, and to search for early sites in these depositional contexts, will provide a definitive answer to this problem. Although the sequence of coastal tenrace accumulation at To'aga was largely controlled rough the interaction of sea level change and subsidence, it would be a mistake to attribute the entire pattem of landscape change to natural processes. At To'aga, as throughout much of te tropical Pacific, humans have played a major role in modifying and shaping island environments (see Kirch 1982, 1983, 1984:123-51). One process which is probably due largely, if not wholly, to human interference in the Ofu ecosystem is the erosion and deposition of substantial volumes of colluvium. Our wansect excavations revealed a consistent pattem over most of the To'aga area (except at Transect 17 which has an almost exclusively boulder talus underlying the narrow nidge) of increasing rates of colluvial deposition onto the coastal terrace after about 2 kyr B.P. This deposition occurred as a series of small, overlapping colluvial fans emanating out of intermittent watercourses or small ravines inland of the site. These fans cap the early calcareous beach-ridge deposits containing pottery-bearing occupations, and in our inland-most test units frequently exceeded 1 m in depth. The fans rapidly pinch out as they extend onto the coastal tenrace, with the larger angular clastics decreasing in frequency, and thin "tongues" of fine-grained silt and clay extending out onto the coastal flat. That humans played a key role in initiating this incrased rate of erosion and deposition of colluvium after 235 about 2 kyr B.P. is suggested by the presence of charcoal in most of these colluvial sediments. Natural forest fires are an extremely rare occurrnce in te humid tropics, and thus the presence of charcoal flecking in these sediments is almost certainly a signal of human buming (see Kirch and Yen 1982:154, 351-52 for a discussion of this phenomenon on Tikopia Island). This buming most likely was associated with forest clearance for agriculture, specifically shifting cultivation of root and tuber crops on the steep hillsides inland of the site. Once the native forest was disturbed and opened up, erosion of the youthfil volcanic soils on the steep slopes would have increased drnadcally. The effects of this erosion and deposition of colluvium on the newly prograded coastal terrace were by no means negative from the human viewpoint of potential land use. Rather, the addition of highly fertile, young volcanic sediments to the welldrained calcareous terrace created a mixed edaphic environment that was more suited to the cultivation of tuber and tree crops (such as Dioscorea yams, Alocasia aroids, breadfruit, and coconut), than either the volcanic or calcareous sediments by themselves. Hence, as the coastal terrace itself prograded and expanded in area after about 2 kyr B.P., its potential as a zone of intensive agricultural production was significantly enhanced by the human-induced deposition of volcanic sediments. We can only speculate that the creation of this coastal zone, edaphically well-suited to intensive agriculture as a result of human actions, may have occurred at a time when the island's population was likely to have been increasing after a millennium or so of settlement. Here too, it may not be at all coincidental that the traditional site of chiefly power on Ofu Island-the Tui Ofu monument complex-is situated in the approximate center of this rich resource zone (see Hunt, chapter 3). The emerging chiefly polity of the island could be expected to have exercised its hegemony by seizing control of this newly created and highly productive resource zone. Other indications of human impacts on te Ofu Island ecosystem in the archaeological record at To'aga include te extinction or extirpation of bird populations (see Steadman, chapter 14) and the intduction of adventive species of terrestrial molluscs associated with Polynesian horficulture (see Kirch, chapter 8). Consideration of this evi- 236 The To'aga Site dence, however, will be deferred to the discussion of te prehistoric subsistence systen in a later section of this chapter. Befoe closing this discussion of the sequence of landscape change at the To'aga site some brief comparisons with other documented sequences frm tropical Polynesia are wamnted. The pattems of change that dramatically wansformed the lowland environment of To'aga between 3 kyr B.P. and e present are not unique to Ofu Island. Similar sequences have been atested for other islands in the southwestem Pacific, including Tikopia (Kirch and Yen 1982), Aneityum (Spriggs 1986), Lakeba (Hughes et al. 1979; Bayliss-Smith et al. 1988), Futuna (Kirch 1975, forthcoming), Mangaia (Kirch et al. 1992) and Niuatoputapu (Kirch 1988). The specific similarties include: (1) the progradation of cosal lowlands between ca 3-1.5 kyr B.P., presumably due to sea-level changes in the area; (2) humaninduced forest clearance and erosion resulting from shifting cultivation; (3) deposition of alluvial and colluvial sediments in lowland landforms; and (4) edaphic enhancement of lowland environments as a consequence of (3). Also attested in these sequences are human impacts on the endemic and indigenous biota. In short, the To'aga site adds yet another case to the growing catalog of significant human modifications to the island ecosystems of the central Pacific. ANCESTRAL POLYNESIAN CULTURE The derivation of Polynesian cultures from a Lapita ancestor is now well attested in the archaeological sequences of Westem Polynesia. Kirch (1984) and Kirch and Green (1987) have used the tenn "Ancestral Polynesian Culture" (or "Society," depending upon the frame of reference) for the culture which emerged in the archipelagoes of Westem Polynesia during the middle of the first millennium B.C. The reconstruction of Ancestral Polynesian Culture (which doubtess was not unifom troughout the region) is arguably a key task for Polynesian prehistorians, because it provides the baseline against which subsequent cultural divergence, evolution, or transformation can be measured. Such reconstruction may be attempted from at least three different, but complementary, sources of evidence: comparative ehnography, historical linguistics, and archaeology. Green (1986) has oudined this "tiangulation" approach in gater detail using the example of Ancestal Polynesian settlement systans. The To'aga site is potenially a major source of archaeological evidence for the reconstruction of Ancestral Polynesian Culture, because the site's stratigraphic sequence spans the whole of the first millennium B.C., the period duMg which tis Ancestral culue emerged out of its Lapita predecessor. Because our objectives and temporal-fiscal limitations necessitated a testing stategy of excavation, our present evidence from To'aga is primarily in the realms of material culure and of faunal material bearing on dte subsistence economy. In the fuure, however, expanded horizontal excavations at To'aga may reveal much new evidence regarding the structure and spatial organization of Ancestl Polynesian settlements. Ceramic studies have been an important part of understanding prehistory and culure change in Samoa and West Polynesia. The assemblage from the To'aga excavations is especially significant as it is large, excavated from a well-stratified, well-dat deposit represents the full duration of pottery manufacture in Samoa, and has been studied in detail (chapter 15). Green (1974) perfonned the first detailed analysis of Samoan ceramics from the SU-Sa-3 site at Sasoa'a, Upolu. He defined one "type" with two varieties of plainware: a coarse-tempered thickware, and a fine-tempered thinware. Green (1974) followed a methodology for classification developed in the culture-historical paradigm of American archaeology (i.e., especially in the work of J. Ford, A. Kreiger, and I. Rouse; see Dunnell 1986a, 1986b). Contrary to crticisms from proponents of the New Archaeology (1950s-60s), classification fonnulated by culture historians is deductive, problem-oriented, and based largely on a paradigmatic structure (Dunnell 1971). Culture historians understood the explanatory meaning of variability-their methodology was founded in a materialist ontology (Dunnell 1986b). The debate between Ford and Spaulding in the mid-1950s is illustrative of the contrasts between culture history and the aspirations of the New Archaeology (see Dunnell 1986b). Following the culture historians' lead, Geen (1974) defined two classes of pottery by the Synthesis and Interpretations combination of dimensions in a paradigmatic structre (see Dunnell 1971) for thickness and temper size. Thick- and thinware, as vanreties of one type (in the type-variety system; see Dunnell 1986b: 174), are ideational units or classes. Ideational units "are tools of our construction the purpose of which is to allow us to recognize and describe these things about which empirical claims are made" (Dunnell 1986b: 151). Classes are based on intentional definitions, where a specific set of features fonns the necessary and sufficient conditions formembership in a unit (Dunnell 1971:16). Once Geen (1974) defined classes for analytic purposes, he described the empirical variability within them. His descriptions include observations such as color, paste texture, temper composition, and surface treatments-attributes that were not part of the defining criteria. Dunnell (1971) has distinguished definitions from descrptions as the contrast between class and group. Classes comprise lists of criteria; groups are sets of things (Dunnell 1986b:181). Thus, defining classes and describing groups (empirical entities), which he called "categories," characterizes Green's analytic approach. In his analysis, Green (1974) attempted to use the vessel, rather an sherds, as the basic counting unit comprising the archaeological assemblage. This innovative approach to establish a minimum number of vessels based on similarity in sherd color, temper, and other attributes has been attempted by others (e.g., Brose 1970, Sullivan 1983). However, as Feathers (1990:139) explains, attempts to transfonn sherds into vessels are met with some problems. And, such attempts are not always necessary, since variability between assemblages will be reflected in attributes of sherds as well as vessels. As Feathers (1990:13940) puts it, "sherd information is not inferior to vessel infornation. It is just different And because of the problematic representation of vessels archaeologically, sherd data is [sic.] te best source of assemblage infonnation." Green's (1974) minimum number of vessels approach does not lessen the usefulness of his analysis. While later statistical analyses (e.g., Clark and Herdrich 1988; Hunt and Eikelens, chapter 9) revealed problems in the intuitive definition of thick- and thinware, Green's rich descriptions provide data suitable for comparative analysis. In a subsequent study, Smith (1976) took 237 assemblages from the Mulifanua Lapita site, Jane's Canp, and the Paradise site for ceramic analysis. Smith used principal components analysis, a multivariate grouping procedure, to (1) evaluate Green's conclusions, (2) compare Lapita with later Samoan cermnics, and (3) in his words, "attempt a meaningful and useful classification of the pesent Samoan ceramic material using a wide range of both stylistic and techological variables" (1976:83-84). Focusing on attributes that appeared to vary temporally, Smith (1976:86) initially soiled sherds into three kinds based on thickness and paste texture: a thick coarse-textured (tempered?) ware; a thinner, finertextured ware; and an extremely fine textured ware. The distinctions of thickness and paste texture (temper size?) appear identical to those made by Green (1974), with the addition of a finer ware (no doubt due to earlier ceramics in te sample). Smith's (1976) principal component analysis, like Green's pottery descriptions, is a means to delineate the empirical varability of groups. Smith's principal components analysis is not a classificatory tool, however, but a method that analyzes the grouping tendency of the sherds studied. His results show that pottery from these assemblages is relatively homogeneous with respect to the attributes of paste and color. The variables of paste and color are represented in components I and n of his analysis. A third component includes thickness, exterior and interior evenness (variance in thickness?), "filler type" (temper composition?), and "filler size" (temper size?) (Smith 1976:90). While Smith (1976:90) states that "it does not appear possible to interpret this combination of variables in any meaningful fashion," these variables clearly relate to the criteria used to define thick- and thinwares (e.g., Green 1974; Hunt and Erkelens, chapter 9). In analytic terms, Smith (1976) conflates the distinction of group and clas (Dumell 1971). He has not created a classification, but has only shown the grouping tendency of sherds from three assemblages. Urdike paradigmatic classes, groups (as statistical summaries) change with the addition of every new case (sherd or assemblage). Smith (1976:92) proposes that his groups forn the basis for a new "typology" of Samoan ceramics. He points out, however, that his "types" would serve as descriptive categories only as they do not correspond 238 The To'aga Site to either chronology or spatial distribution. Smith (1976: 93) implicitly recognizes the problem of confladng class and group as he notes that the discreteness of "'types" will disappear with additonal analyses of Samnoan ceramics (see Dunnell 1971, especially pp. 87-110, figure 8). Clearly, analysis of groups does not provide the basis for a useful classification. Istead statisdcal grouping techniques allow one to examine variability among defined classes. Holmer (1980) re-analyzed sherds from Mulifanua, Jane's Camp, and te Paradise site with sherds from two new sites, Potusa and Falemoa, both on Manono Lsland. He attempted factor analysis on unspecified variables-apparently similar or identical to those used by Smith (1976). When factor analysis failed to produce groups (clusters), a new strategy was attempted. Holmer subjectively (implicitly) classified sherds into seven "types." Varables were then selected for observation/ measurement, and data from the analysis of sherds was used in discriminant funcdon analysis. In this analytic procedure, Holmer (1980) simply confirms (statistically) his "subjective" sorting criteria (implicit a priori class definitions). As with Smith's (1976) analysis, the ideational (classes) and empirical entities (groups) are confused. Holmer (1980) is thus left with 'lypes" that will constantly change with every new case analyzed. These problems with confusing group and class, evident in Smith (1976) and Holmer's (1980) work, are not merely an academic issue. The conflation of these observational and analydc steps is the reason why Smith and Holmer did not succeed in producing classificatory systems, contrary to their stated objectives. This is because: (1) Analysis of objects is necessarily based on an a priori (most often implicit) classificatory system (i.e., to observe "x" is to observe "x" as a case of something). Grouping procedures are data manipulations performed on unanalyzed and implicit classifications; they are inductive and often formulated as "problem-free." (2) Groups are based on empirical sets (e.g., statistical summaries) that change with every case added, thus an object cannot be assigned to a pre-existing group without altering the "definition" of the unit (3) Grouping provides descriptions of phenomena, but not definitions stipulating necessary and sufficient conditions for membership (althoughpost hoc definitions might be extracted). Group membership is based on similarity which vanes in degree. This means that individual objects in a group may share many, some, or in extreme cases, no traits in common (see Dunnell 1971:fig. 8). Clark and Herdrich (1988) recently called attention to the problem of Green's (1974) original formulation of coarse-tempered thickware and finetempered thinware. They point to the lack of explicit crteria for what is thick and thin, fine or coarse. Clark and Herdrich (1988) illustrate the variability of these dimensions in an assemblage from 'Aoa, eastem Tutuila. While preliminary in nature, their examination of ceramics points to the importance of classification as a means to document change in te Sanoan sequence. The To'aga ceramic study reported in this volume is the most intensive analysis of assemblage from Samoa to date. The analytic protocol was designed to support the definition of numerous classes (i.e., from two or more of the dimensions) deduced to address particular research questions of the assemblage. While the To'aga assemblage is especially well-studied, much wor remains to answer the larger questions posed by Hunt and Erkelens (chapter 9) for ceramic evolution. Results from To'aga show tha thickware is present in the earliest deposits, and its abundance (in actual numbers) over time is relatively stable. Thinware is never clearly dominant at To'aga. Thinware declines in real and relative abundance over time but persists perhaps as long as pottery production itself. At To'aga, as elsewhere in Samoa, pottery declines in abundance early in the Christian era and then disappears entirely. Results from ceramic compositional analyses show that the bulk of pottery, including both thickand thinwares, and some carved paddle-impressed ceramics, was produced locally with colluvial "selftempered" clay source(s) from Leolo Ridge on Ofu. Red-slipped pottery from To'aga does not match the local colluvial clays presently known. Based on this, the red-slipped ware may have an exotic provenance, amving on Ofu through inter-island exchange. Finally, results also suggest that the diversity of raw materials declined over time. Such a pattem if substantiated with furher work, suggests changes in availability and/or procurement of raw material. A decline and eventual end to inter-island exchange of Synthesis and Interpretations ceramics might also be hypothesized. The To'aga assemblage, like most others from Sanoa, is simple in foni and includes very little decoration. Only bowls are represented. Decoration is restricted to impressing and notching on te lip, and on body sherds, red-slip, carved paddle-impression, and incision. This simplicity stands in marked contrast to Lapita assemblages of comparable age from Mulifanua, 'Upolu, and assemblages known from Fuuna, Tonga, and Fiji. Isolation of Manu'a from communities beyond Samoa might account for this stylistic divergence. The To'aga excavations also yielded a small but important set of stone adzes in association with the ceramics just discussed. Significantly, these are all variants of the plano-convex sectioned Type V adz described by Green and Davidson (1969). A dominance of Type V adzes in ceramic-bearing contexts in Westem Samoa was noted by Green (1974:257-58). The Manu'a results confirm this pattem for the eastem part of the Samoan archipelago. Indeed, Type V appears to have been a widespread and common form throughout the Ancestral Polynesian region, given the presence of this form in sites in Futuna (Kirch 1981), Niutoputapu (Kirch 1988:192), and Tongatapu (Poulsen 1987:170). Type V was dropped frm the Samoan adz inventory early in the first millennium A.D., and our surface collections from Manu'a are dominated by adzes with quadrangular or trapezoidal cross sections. One of these trapezoidal forms was recovered from a late, aceramic depositional context in Unit 3; this particular specimen appears to have been manufactured at the Tatagamatau quarry on Tutila Island (see Weisler, chapter 12). Pottery and stone adzes are the best documented classes of Ancestral Polynesian portable artifacts. For both of these artifacts, the development of uniquely Ancestral Polynesian forms out of Lapita prototypes has been archaeologically demonstrated (e.g., Green 1971, 1974). This is not the case, however, with another important Polynesian artifact class: the one-piece fishhook. When archaeologists first commenced stratigraphic excavations in the Eastern Polynesian archipelagoes of Hawaii, the Marquesas, Societies, and New Zealand in the 1950s and 60s, fishhooks proved to be among tX most ubiquitous arfifacts. Indeed, in the absence of pottery, Polynesian archaeologists applied their skills 239 at classification and seriation to fishing gear, in an effort to establish chronological sequences (e.g., Emory, Bonk, and Sinoto 1959; Suggs 1961). Thlus, when efforts were directed at the Westem Polynesian islands of Tonga and Samoa, the initial expectation was that similarly rich assemblages of fishing gear would be recovered. These expecations were quickly thwarted by an almost complete absence of fishhooks in Westem Polynesian sites. Poulsen recovered only one "certain specimen" of one-piece hook in his excavations of six sites on Tongatapu (1987:186), while the major Samoan archaeological programme of Green and Davidson (1969, 1974) recovered but a single fragment of a Turbo-shell hook from the Lotofaga midden (Green and Davidson 1969, pl. 23). Kirch's excavations on Niutaputapu fared only slightly better, with four onepiece hooks out of thirteen sites sampled (1988:204, fig. 124). The extreme paucity of fishing gear in Westem Polynesian sites-in contrast with the typically high density of fishhooks in Eastem Polynesian sitesraised a number of questions. Given the extensive scope of Westem Polynesian excavations, sampling enfor alone can be ruled out. Rather, it appears that te rarity of hooks in early Westem Polynesian sites (i.e., those dating to the Ancestral Polynesian period) is an accurate reflection of the relative unimportance of angling gear (Green 1986:131). The faunal assemblages from these sites, however, clearly indicate that inshore fishing was a major component of the subsistence economy. Were hooks being made of pershable materials (such as wood) and hence not preserved in the archaeological record? Were other fishing strategies, such as netting, spearing, or poisoning, prefened over angling by Ancestal Polynesian fishermen? To'aga is the first Ancestral Polynesian site to produce a large assemblage of one-piece fishing gear, and thus demonstrates anoter kind of variability in eady Polynesian culture. As described in chapter 11, a total of twenty-eight whole or partial Turbo-shell hooks were recovered from our excavations, along with another thirty-one prefonns or tabs. This is a fishhook density level much more in keeping with Eastem Polynesian sites. Why should To'aga produce such an assemblage of one-piece hooks when other Samoan and Westem Polynesian sites are devoid of these aftifacts? We suggest that 240 the The To'aga Site answer lies in the differenidal marine environ- ments of the vanous islands. The Westem Samoan islands of 'Upolu and Savai'i, as well as Niuatoputapu and Tongatapu, are all characterized by extensive barrier reef and lagoon ecosystens. In these kinds of coastal environments, the most effective fishing stategies are usually those involving nets (seines, dip nets, nets used with weirs, and other techniques). This was well documented, for example, in ethoarchaeological studies of contenporary fishing on Niuatoputapu Island (Kirch and Dye 1979; Dye 1980). In contastL the marine environment of Ofu is that of a relatively nanow fringing reef, lacking a broad protected lagoon. In such fringing reef environments, angling becomes a far more significant fishing stategy, to exploit the dominant fish populations of the reef crest and outer slope. The hypothesis that a greater emphasis on angling gear is conrelated with frnging mef (as opposed to barrier reef-lagoon) environments receives some support from fte general Oceanic picre of archaeological fishing gear distribution. For exanple, both Tikopia (Kirch and Yen 1982) and Anuta (Kirch and Rosendahl 1973)-small high islands with narrow fringing reefs-yielded high frquencies of one-piece fishhooks in teir archaeological sites. The same is true of Hawai'i and the Marquesas, where reefs are finging or even lacking altogether. On the other hand, the Society Islands which have extensive lagoons have produced relatively low densities of fishing gear in comparison with other Eastem Polynesian sites. Consequently, we would argue that the Manu'a Group is one area within Westem Polynesia where the marne ecological conditions favored the use of angling gear. The Turbo-shell fishhook assemblage from To'aga is of some interest from a morphologicalstylistic perspective, in addition to its ecologicalfuncdonal implications. As noted in chapter 11, several of the hooks exhibit morphological features similar to those in eatly Eastem Polynesian fishhooks. These include the strongly incurved or "bent" shank and the single-notched, line-lashing devices. Thus, the To'aga hooks can madily be identified as a "prototype" stage from which the greater diversity of Eastem Polynesian forms was subsequently developed. THE SUBSISTENCE ECONOMY OF EARLY SAMOA The To'aga excavations produced one of te largest and best preserved faunal assemblages ever recovered from a Westem Polynesian site: 10,209 vertebrate bones and approximately 166.5 kg of invertebrate materials. Largely due to poor prsrvation, most previously excavated Samoan sites yielded very poor faunal collections. In Westem Samnoa, only the late prehistoric Lotofaga midden reported by Davidson (1969) and the three coastal sites analyzed by Janetski (1976, 1980)-ceramicbearing Potusa, Falemoa, and Jane's Camp-have well-preserved vertebrate and invertebrate faunal materials. lTus, the To'aga materials provide the first extensive sample of fauna frm well-stratified and dated contexts spanning the first half of the Samoan sequence. The faunal data have been presented and analyzed by Nagaoka and Steadman in chapters 13 and 14, respectively. Here we expand on their analyses with several general observations and with comparisons to other Samoan and Westem Polynesian sites. One problem that has concerned archaeologists in Westem Polynesia is whether the Polynesian triad of domestic animals-pig, dog, and chicken-was introduced at the time of initial settlement and colonization (Groube 1971; Hunt 1981; Best 1984; Kirch 1979, 1988). At To'aga, only the chicken (Gallus galus) is well represented in our faunal suites. Chicken is actually the most frequent bird species represented in the avifaunal material, with 16 NISP (see Steadman, chapter 14). Chicken bones were especially well represented in the Layer III deposits in Units 20123, dating to ca. 2800-2300 cal B.P. Pig, however, is unambiguously represented only in later contexts (in Layer I of Unit 17). Some of the unidentifiable mammal bone from eadier strata may indeed be of pig or dog-or both-so that the absence of pig and dog in early contexts is not certain. Nonetheless, given the large vertebrate faunal sample and excellent preservation, it is cerain that neither of these domestic animals was ever present in large numbers at the To'aga site. Another adventive species introduced (presumably as an inadvertent "stowaway" on voyaging canoes) at the time of initial colonization is the Synthesis and Interpretations Polynesian or Pacific rat, Rattus exulans. This species is ubiquitous in the To'aga strata and occurs in the earliest dated depositional context (Layer HID in Units l5129/30). Tate (1951) discusses the very widespread dispersal of this synanthropic species. Another group of human-intrduced organisms appearing in the early To'aga deposits is the set of five synantrpic terrestrial molluscs discussed in detail in chapter 8: Assiminea cf. nitda, Lanellidea pusilla, Gastrocopta pediculus, Liardetia samoensis, and Lauellaxis gracilis. These species are closely commensal with humans, their preferred habitats being gardens and disurbed environments adjacent to habitation sites. The species are all minutevisible to the human eye only on close inspectionand can only have been tansported inadvertently. Many of these species have also been identified from early, Lapita-associated archaeological contexts on Niuatoputapu Island (Kirch 1988:233-35) and on Tikopia (Kirch and Yen 1982:308-309). As discussed in chapter 8, the most likely mechanism for their inter-island transfer, and intrduction to the Manu'a Islands, was with economic plants and adhering soil media. In this regard, these snails provide indirect evidence for early plant introductions to the island. Indeed, in the absence of direct ethnobotanical evidence for cultigens, the suite of synanthDpic snails is the best clue that the early Polynesian colonists intrduced a complex of economic plants to the island, along with the domestic chicken (and possibly also pigs or dogs). Future excavations at To'aga should test this hypothesis through identification of charcoal, carbonized parenchyma, and other carbonized plant materials from earth ovens, hearths, and other stratigraphic contexts. Recent developments in the identification of such carbonized materals by J. Hather (Institute of Archaeology, London; pers. comm., 1991) and others, not available to us at the time the To'aga excavations were undertaken, now make the possibilities for such paleoethnobotanical studies possible. Two kinds of larger marine animals are represented in the vertebrate faunal collections: sea urtes and unidentified marine mammal. The sea trtles probably consist mostly (if not exclusively) of the Green Sea Turtle, Chelonia mydas, but definite identifications on the post-cranial skeleton are virtually impossible. Turtle bones were fairly common and were especially frequent in Layer IIIB 241 of Units 20123 and Layer IIB of Unit 19. The marine mammal bone is most likely from one or more species of porpoise. A relatively high frequency (18 NISP) of marine mammal bone was found in the early Layer IIIC deposit in Units 15/29/ 30. The bird bones are of particular interest, for they reveal exinctions and extirpations consistent with a pattem of avifaunal change from early sites thoughout Polynesia (Steadman 1989; Steadman, Pahlavan, and Kirch 1990, Steadman and Kirch 1990). An unexpected discovery was the presence of two Megapodius sp. bones from Layer IIID of Units 15/ 29/30, the oldest dated layer at the site. Megapodes were not fonnerly known to have been present in Samoa, and this find thus represents an eastern extension of the prehistoric range of this taxon. Given the restriction of this taxon to the earliest stratum, it is likely that the species was rapidly overexploited-to the point of extinction-by the early colonizers of Ofu. Also striking is the presence of bones of six species of seabirds which no longer occur on Ofu Island, including Puwinus pacificus, Puinus lherninieri, Puffinus griseus, Pterodroma rostrata, Pterodroma sp., and Sula sula. The loss of these species from the island within the span of human occupation most likely reflects both direct predation by humans and habitat disturbance. Ninety-four percent of the vertebrate fauna from To'aga consists of fishbone, of which 2,229 NISP were identifiable to family-level taxa (see Nagaoka, chapter 13). Although there are frequency differences in taxa witiin different excavation units, there is remarkable consistency overall in the rank-order dominance of particular fishes. Four families dominate the faunal assemblages: Diodontidae, Serranidae, Acanthuridae, and Holocentridae. These families include numerous species, most of which occur on the reef flat or immediately off the reef edge. They may be taken with a variety of fishing strategies including netting, spearing, poisoning, and angling. The acanthurids, holocentrids, and serranids especially, can be taken with hook-and-line, and it is very likely that the small one-piece Turboshell fishhooks recovered from the To'aga site were used to capture these taxa. Of considerable interest is the high frequency of Diodontidae (primarly Diodon hystrix), the porcupinefish. These fishes are known to carry tetradontoxin which can cause severe 242 The To'aga Site illness or even death when ingested by humans. That such a dangerous fish should dominate the To'aga faunal assemblages is cunous, although not inconsistent with pattems in other early Pacific sites (see Green 1986:132). InTikopia, Diodon hystrix was extremely plentiful in the early middens (Kirch and Yen 1982: 292, table 42), as it was also in the Early Eastem Lapita site of NT-90 on Niuatoputapu Island (Kirch 1988:223, table 29). A second tier of fish taxa, in terms of fteir rankorder abundances, comprses the following families: Scaiidae (parrotfish), Carangidae (jacks), Labridae (wrsses), Lutjanidae (snappers), Muraenidae (moray eels), Balistidae (triggerfish), and Osaciidae (boxfish). Again, these are all inshore, reef or reef edge fishes, represented by a large number of species. A varety of fishing strategies were doubtless employed to take these fishes. A number of other taxa are less commonly represented among te fish fauna assemblages from various excavation units. These are again prmarily inshore fishes, but several exanples from the family Scombridae (nas and mackerels) are present. This is significant, for it does indicate the prctice of pelagic fishing, probably with pearl-shell trolling lures. Trolling for tuna, however, was clearly a minor fishing strategy in terms of its contribution to the total fish catch. In terms of sheer bulk, invertebrates (and especially molluscs) comprise the majority of the faual materials from the To'aga site. (Nonetheless, their contribution of meat to the prehistoric Samoan diet was probably less than that of fish.) In tenns of rank-order abundances based on weighLt a few taxa dominate the assemblages. Consistently the most abundant species is the reef gastropod Turbo setosus; the closely related species T. crassus is also quite common. Only slightly less common is the bivalve Tridacna maxima, which occupies the reef platform. Turbo spp. gastropods comprise on average about 62 percent of the invertebrate faunal suite from the To'aga midden deposits, while Tridacna bivalves constitute anodter 7 percent Other commonly represented taxa include: Trochus maculus, Tectus pyramis, Cypraea spp., Conus spp., Vasum ceramicwn, Cerithium noduloswn, Strombus maculatu, Thais armigera, Asaphis violaseus, and Nerita spp. All of these molluscs occur on the reef platform and reef crest fronting the To'aga site. Sea urchins of several species are also repre- sented in the To'aga middens. The smaller-spined taxa are doubdess underrepresented in our samples, because the spines usually are not retained in the 0.25-inch mesh sieves that we employed. Some indication of their presence was provided by the bulk samples and micro-aftifact analyses of selected sediments (see Kirch, Manning, and Tyler, chapter 7; Nagaoka, chapter 13). The Layer IIIA/IhB occupation in Units 20/23 was noteworthy for an unusually dense concentration of the large slatepencil sea urchin (Heterocentrotus mamllatus). More an 6 kg of these spines and test fragments were overed from these strata, pardy in associadon with an earth oven feature. In sum, the To'aga excavations have provided significant new information on which to base reconstnuctions of Ancestral Polynesian subsistence economy, and of the impacts of these early island colonists to the biota of remote Pacific islands. An economic strategy integrating brad-spectrum exploitation of natural faunal resources (marine and tenrestrial) with agricultural producdon is indicated by the To'aga evidence, reinforcing reconstmctions based on other early Fijian and Westem Polynesian sites (Kirch 1984; Kirch and Green 1987). The presence of an extinct or extirpated species of megapode, of six species of extirpated seabirds, and of marine twutle, all in the earliest deposits at To'aga (especially in Layer IhB of Units 20/23), suggest that iniidal exploitation of the island's larger faunal resources may have exceeded te capacity of these naural populations to survive or reproduce under the pressures ofintensive human predation. In addition to these early impacts on fte naltura biota, however, te Polynesians radically altered the To'aga area, transforming te coastal environment in particular into a ftoroughly anthropogenic landscape. The purposive inroduction of domestic animals and economic plants, and dte inadvertent introduction of rats, terrestrial snails, and other organisms were the first stages in the conversion of the Ofu ecosystem into a cultumal landscape capable of supporting a dense human population. Following progradation of fte coastal plain after about 1900 B.P., te To'aga area was developed into a highly intensive arboricultural production zone, dominated by economic plants of Polynesian inroduction (coconut, breadfruit, aroids, yams, arrowroot, and othes). Even the steep volcanic slopes inland were cleared Synthesis and Interprdatios 243 of native forest and converted to zones of shifting cultivation. Such interior slope modification resulted as well in incrased rates of soil erosion and deposition onto the coastal flats, enhancing the edaphic condition of e latter zone for cmp production. In this regard, the To'aga data add another instance in the rapidly accumulating repertoire of archaeological evidence for prehistoric human transfonnation of Pacific island environments (Bayliss-Smith et al. 1988; Kirch 1983; Steadman 1989). Hunt and Eikelens (chapter 15) conclude that red-slipped ware, and perhaps other pottery made of clay distinctive from the locally Ikown colluvial sources on Ofu, may reflect imports to the island. The decline in compositional diversity hypothesized for the ceramic assemblage might also indicate that exchange diminished in importance over time. Additional research on sherds from Manu'a and elsewhere in the region is necessary, however, to fully test hypotheses for inter-island ceramic exchange. INTER-ISLAND CONTACTS THE TO'AGA SITE: AND EXCHANGE CULTURAL RESOURCE MANAGEMENT CONSIDERATIONS Archaeological research in fte Fiji-Westem Polynesia region has produced some evidence for inter-island contacts and exchange (e.g., Best 1984; Davidson 1977; Kirch 1988). While inter-archipelago contacts and exchange are known ethohistorically, extensive exchange of ceramics and other materials appears to have occurred in the earliest perod of the region's history. Best (1984), for example, shows that a substantial propordon of early ceramics on Lakeba was imported to the island. While te Tatagamatau basalt quany site on Tutuila (Leach and Witter 1987; Best et al. 1989) is a likely center for adz export, little is known about potential pattems of exchange in Sanoa and the quany's place in a regional system. Excavations at To'aga produced a range of materials for which provenance can be deduced or inferred. Compositional analysis ofvolcanic rock lithics and adzes, temper and clay of ceramics, and raw materials from potential sources offer a means to explore questions of prehistoric exchange. Weisler (chapter 12) used non-destmctive XRF analysis on ardfacts and potential source rocks from Manu'a and Tatagamatau. His results show that the composition of rock in finished adzes (and in flakes from adzes showing polish) of fine-grained basalt cluster with those from Tatagamatau on Tutuila. In contrast, a relatively coarse-grained dike stone from Fa'ala'aga on Ofu Island is represented only in debitage and simple flake tools. These results suggest that perhaps much, or all local stone was not suitable for adz producdon. Exchange, at least with Tutuila some 100 kn to the west, brought adzes and probably other materials to Manu'a. As he initial objective of the Manu'a Project was the identification of archaeological sites for purposes of cultural resource management-under contract to the Historc Preservation Office of the Government of American Samoa-it is appropriate to conclude this monograph with a discussion of the significance of the To'aga site, the current status of the site, and the potential impacts which may hreaten its integrity in future years. Site AS-13-1 is unquestionably one of the most significant archaeological sites yet discovered in American Samoa, and indeed, in the Samoan archipelago as a whole. Within American Samoa, it certainly ranks with the extensive Tatagamatau adz quanry complex on Tutuila in tenns of its potential to yield infonnation on the prehistory of the archipelago. Some specific aspects of the To'aga site that collectively contribute to its archaeological significance are enumerated below: 1. Site AS-13-1 incorporates the largest continuous area of subsurface archaeological deposits dating to the ceramic phase of Samoan prehistory of any site yet discovered in the archipelago. These deposits are estimated to cover between 21,000 and 35,000 m2 and appear to represent a seres of domestic household units. 2. Site AS-13-1 is well statified, and thus has the potential to yield a finely-detailed chronological sequence of cultral change for the Manu'a Islands. The remarkably deep stratification in parts of the To'aga site contains three or more meters of culural deposits. Because of this stratigraphic record, the 244 The Toaga Site To'aga site presens excellent opporunities for recovering a detailed sequence of artifacal, faunal, and settlement information. 3. Site AS-13-1 spans virully the entire prehistoric sequence of the Samoan archipelago. Initial occupation of the site began around the close of the second millennium B.C., contemporaneous with Xt Mulifanua Lapita site on 'Upolu Island. The stigraphic wecord from ca. 3000 B.C. tO A.D. 800 has been well documerned by our systematic transect excavations, detailed in this monograph. Other archaeological feaus dating to the last one thousand years are also present in the area, altough they have not yet been intensively investigated or radiocartbon dated. No other single site locality in American Samoa has yet produced such a continuous occupation sequence spanning the whole of regional prehistory. 4. The preservation of both ardfacts and faunal materals in Site AS-13-1 is excellent, especially in the deeper stratigraphic units, where calcareous (alkaline) sedimentary conditions prevail. The majodty of Sanoan archaeological sites are characterzed by acidic soil conditions which do not favor the preservation of such organic materals as bone, shell, or sea urchin spines. In such acidic contexts, cultal materials are usually limited to pottery and stone artifacts. At To'aga, the excellent preservation conditions yield not only ceramic and stone artifacts, but extensive assemblages of bone and shell faunal materials, as well as arfifacts of shell, bone, and sea urchin spine. As a result, our knowledge of early Samoan materal culture and economy has been significantly expanded by te materials from Site AS-13-1. A particular example is the complex of Turbo-shell fishing gear, which for te first time has given us some in-depth information on Samoan angling strategies in the first millennium B.C. Similarly, the faunal assemblages from the To'aga site are the largest-in terms of both numbers and taxonomic richness-from any site yet excavated in the archipelago. 5. Site AS-13-1 also incorporates a number of features of considerable cultural significance to te people of Ofu Island. In particular, To'aga is the taditional seat of the Tui Ofu chiefship, represented by the Tui Ofu well and burial mound (see Hunt, chapter 3). These monuments are held in considerable awe by the people of Ofu Island and are directly conected to a body of oral tradions (see Mead 1930). Togefter, all of te aspects of Site AS-13-1 enumerated above combine to make this archaeological complex one of the most significant cultural resources in American Samoa, and indeed, in the Samoan Lslands as a whole. The site has already yielded much imporant new infonnation on te prehistory of the Manu'a Group and the Samoan archipelago, and its potential has hardly been tapped. Because of its outsanding significance, we have nominated Site AS-13-1 to the National Register of Historic Places, in conjunction with the Historic Preservation Office of the Govenunent of American Samoa. From te culurl resource management viewpoint, it is important to assess any potential hreats to te To'aga site. As described in chapter 2, the present mode of land use over most of the site is subsistence gardening in a more-or-less traditional manner (arboriculture and Alocasia aroid swiddening). This relatively low intensity land use does not seriously heaten the integrity of the site, other than for minor impacts to surface archaeological features such as 'ili'il pavements or lua'i masi pits. There has already been some significant damage to the site, however, trugh the construction of the Public Woiks Department landfill at the southwestem edge of the site. (Ironically, it was this landfill that led to the original discovery of surface potterybearing deposits during the 1986 reconnaissance survey.) It appears that most of the bulldozed landfill pit lay outside of the area of deeply stratified archaeological deposits, but these were disturbed along the inner edge of te bulldozer cut. While the 1986 bulldozing probably did not greatly impact the total site area, it is extremely important that no furxer expansion of this landfill operation occur without prior consultation with the Historic Preservation Office. It should be possible to plan for future landfill needs by situating such landfill pits in the seaward portions of the To'aga coastal fiat tiat do not contain subsurface archaeological deposits. We strongly recommend ta prior to any future landfill operations, the Public Woiks Deparument consult with the Historc Preservation Officer and arrange for limited test excavations to assure that landfill bulldozing take place outside of the zone of buried Synthesis and Interpretations archaeological features. We are unaware, at present, of any other planned developments or constuction in the To'aga area, but if such projects arise, ftey could also have te potential to threaten the integrity of the site. For this reason, it is important that AS-13-1 be placed on the National Register of Historic Places, and that the American Sanoa Historic Preservation Officer attempt to monitor any proposed land use actions in the vicinity of the site. Over the longer tern, te entire archaeological complex at To'aga could be serously thratened by natural coastal erosion. According to the morphodynamic model developed in chapter 4, and tested through various field and laboratory observations, it would appear tha the southem Ofu coastline may have entered a phase of sea level rnsgression. This appears to be reflected along the modem coasdine by active erosion of the beach ridge and associated vegetation line, and by exposure of beach rock. Given that the subsurface archaeological deposits are situated well inland of the present shoreline (between about 40 to 60 m, depending upon the particular locality), such erosion is not at present a serious concem. However, should this transgression phase continue or intensify, for example as a result of global wanming and consequent sea level rise (Geophysics Study Committee 1990), the integrity of Site AS-13-1 would ultimately be affected. Continued tectonic subsidence of Ofu Island may itself eventually result in the natural erosion and destruction of the site, but this pocess could be seriously intensified and quickened by rapid sea level rise. Obviously, these are not problems requiring immediate attention, but they should not be wholly ignored either. Fnally, we wish to conclude by briefly drawing attention to some further research possibilities at the To'aga site. While we have been able to use the opporunity of subsurface testing and site survey at AS-13-1 to addess a number of research problems in Samoan archaeology, our investigation of the To'aga site in 1987 and 1989 must be regarded as no more than a pioneering phase. This site has enormous potential to add to our knowledge and undersanding of the prehistoiy not only of Samoa, but of the Westem Polynesian region as a whole. The site is also of sufficient size, with an estimated 21,00035,000 m2 of stratified deposits dating to the ceramic 245 phase of Samoan prehistory, that even a large scale excavation program would not remove more than a small percentage of the total area, leaving the majority of the site as an "archaeological bank" for future research. We suggest that the next logical phase of archaeological research at To'aga might be to employ a horizontal excavation strategy to expose one or more larger areas (on the order of 100- 150 m2) within the zone of subsurface deposits dating to the ceramic phase of Samoan prehistory. Our systematic transect sampling suggests that this zone is made up of clusters of domesfic or household residential units, probably consisting of series of dwelling, cookhouse, and possibly other special purpose activity areas. Horizontal exposure of one or more of xse residential units could provide the first clear picture of the setflement layout and spatial anangement of an early, Ancestral Polynesian community. Such a project would be of considerable interest not only for Samoan prehistory, but for expanding our knowledge of Ancestral Polynesian culture, which is a critical baseline for the development of later variants of Polynesian culture oughout the whole Polynesian triangle (Kirch and Green 1987). Obviously, such an expanded excavation program would also need to be combined with oher kinds of rsearch objectives and methods. For example, Weisler's trial application of the nondestructive XRF method of chaacterizing and sourcing basalt artifacts (see chapter 12) could be followed up with a more intensive study, with much potential to reveal patterns of long-distance exchange at all periods of Manu'an prehistory. Similarly, the varation in ceramnics noted at To'aga, such as the distinction between fine thinware and coarse thickware, could be explored along avenues other than just chronological changes in frequency. It may be that these ceramic wares reflect functional, or social, pattems in early Samoan society. These can only be explored through the use ofhorizontal excavation strategies in which the distribution of ceramics can be closely mapped in comparison to the spatial layout of households. These suggestions are not meant to be an exhaustive catalog of research problems that might be addressed at the To'aga site but simply some possible research directions that we feel would 246 The To'aga Site contnbute significantly to cunrent issues in Polynesian archaeology and prehistory. We are pleased that our pioneering phase of research at AS-13-1 has already been able to yield inporant new insights on Sanoan prehistory and look forward to the contributions dt future work in the deeply stratified sands of To'aga will doubtless bring. REFERENCES CITED Bayliss-Smith, T., R. Bedford, H. Bookfield, and M. Latham 1988. Islands, Isainders and the World. Cambridge: Cambridge University Press. Best, S. B. 1984 Lakeba: The prehistory of a Fijian island. Unpublished PhD. dissertation, University of Auckland. Best, S., H. Leach, and D. Witter 1989. Report on the second phase of fieldwork at the Tatagamatau site, American Samoa, July-August 1988. Department of Anthropology, University of Otago. Brose, D. S. 1970. The Archaeology of Summer Island: Changing Settlement Pattern Systems in Northern Lake Michigan. Anthropological Papers No. 41, University of Michigan Museum of Anthropology, Ann Arbor. Clark, J. T. 1989. The Eastem Tutuila archaeological project: 1988 final report Report prepared for the Government of American Samoa. Department of Sociology-Anthropology, North Dakota State University, Fargo. Clark, J. T., and D. J. Herdrich 1988. The Eastem Tutuila archaeological project, 1986 final report Report prepared for the Government of Amencan Samoa. Department of SociologyAnthrpology, North Dakota State University, Fargo. Davidson, J. M. 1977. Westem Polynesia and Fiji: Prehistoric contact, diffusion and differentiation in adjacent archipelagos. WorldArchaeology 9:82-94. Dunnell, R C. 1971. Systatics in Prehistory. New York: The Free Press. 1986a. Five decades of American archaeology. IN D. J. Meltzer, D. D. Fowler, and J. A. Sabloff, eds., American Archaeology Past and Future. Washington D.C.: Smithsonian istitution Press. 1986b. Medtodological issues in Americanist artifact classification. Advances in Archaeological Method and Theory 9:149-297. Dye, T. S. 1983. Fish and fishing on Niuatoputapu, Tonga. Oceania 53:242-71. Emory, K. P., W. Bonk Jr., and Y. H. Sinoto 1959. Hawaiian Archaeology: Fishhooks. Bemice P. Bishop Museum Special Publication 47. Honolulu. Emory, K. P., and Y. H. Sinoto 1965. Prliminary report on the archaeological investigations in Polynesia. Report prepared for the National Science Foundation. Department of Anthrpology, Bemice P. Bishop Museum, Honolulu. Feathers, J. K. 1990. Explaining the evolution of prehistoric ceramics in southeastern Missouri. Ph.D. dissertation, University of Washington, University Microfilms, Ann Arbor. Frost, J. 1978. Archaeological investigations on Tutuila Island, American Samoa. Unpublished PhD. dissertation, University of Oregon. Geophysics Study Committee, 1990. Sea-Level Change. Studies in Geophysics. Washington D.C.: National Academy Press. Green, R. C. 1971. Evidence for the development of te early Polynesian adz kit. New Zealnd Archaeological Association Newsletter 14:1244. 1974. Excavation of the prehistoric occupations of Su-Sa-3. IN R. C. Green and J. Davidson, eds., Archaeology in Western Samoa, Volume II. Bulletin of the Auckland Institute and Museum. 1986a. Lapita fishing: The evidence of site SERF-2 from the Main Reef Islands, Santa Cruz Group, Solomons. IN A. Anderson, ed., Traditional Fishing in the Pacific, pp. 119-35. Pacific Anthropological Records 37. Honolulu: Benice P. Bishop Museum. 1986b. Some basic components of the ancestral Polynesian settlement system: Building blocks for more complex Polynesian societies. IN P. V. Kirch, ed., Island Sociedes: Archaeological Approaches to Evolution and Transformation, pp. 50-54. Cambridge: Cambridge University Press. Green, R. C., and J. M. Davidson, eds., 1969. Archaeology in Western Samoa, Vol. 1. Bulletin of the Auckland Institute and Museum. 1974. Archaeology in Western Samoa, Vol. 11. Synthesis and Interpretations Bulletin of the Auckland Institute and Museum. Gren, R. C., K. Gren, K Rappapoit, A. Rappaport, and J. Davidson 1967. Archaeology on the Island ofMo'orea, French Polynesia. Antropological Papers of the Amedcan Museum of Natual History 51(2). New Yor. Groube, L. 1971. Tonga, Lapita potery, and Polynesian oiigins. Journal of the Polynesian Society 80:278-316. Holmer, R 1980. Samoan ceramic analysis. IN J. Jennings and R. Holmer, eds., Archaeological Excavations in Western Samoa, pp. 104-116. Pacific Anthropological Records 32. Honolulu: Bemice P. Bishop Museum. Hughes, P., G. Hope, M. Latham, and M. Brookfield 1979. Prehistoric man-induced degradation of te Lakeba landscape: Evidence from two inland swanps. IN H. Brookfield, ed., Lakeba: Environmental Change, Population Dynamics, and Resource Use, pp. 93-110. Paris: UNESCO. Hunt, T. L. 1981. New evidence for early horticulhur in Fiji. Journal of the Polynesian Society 90:259-69. Hunt, T. L., and P. V. Kirch 1987. Radiocaftn dates from two coastal sites in the Manu'a Group, American Samoa. Radiocarbon 29:41719. -. 1988. An archaeological survey of the Manu'a Islands, American Samoa. Journal ofthe Polynesian Society 97:153-83. Irwin G. 1981. How Lapita lost its pots: The question of continuity in the colonisation of Oceania. Journal ofthe Polynesian Society 90:481-94. Janetski, J. C. 1976. Dietary remains from Jane's Camp-a midden site. IN Jesse Jennings, ed., Excavations on 'Upolu Western Samoa, pp. 7582. Pacific Antrpological Records 25, Department of Antrpology, Honolulu: Bernice P. Bishop Museum. 1980. Analysis of dietary remains from Potusa and Falemoa. IN J. Jennings and R Holmer, eds., Archaeological Excavations in Western Samoa, pp. 117-22. Pacific Anthrpological Records 32, Deparument of Antropology, Honolulu: Bemice P. Bishop Museum. Jennings, J. D., and R. N. Holmer, eds., 1980. Archaeological Excavations in Western Samoa. 247 Pacific Anthropological Records 32. B. P. Bishop Museum, Honolulu. Jennings, J. D., R. N. Holmer, J. C. Janetski, and H. L. Smith 1976. Excavations on 'Upolu, Western Samo. Pacific Anthropological Records 25. Dept of Anthropology, Honolulu: Benice P. Bishop Museum. Kikuchi, W. K. 1963. Archaeological surface ruins in American Samo. Unpublished M.A. thesis, University of Hawaii, Honolulu. Kirch, P. V. 1975. Cultural adation and ecology in Western Polynesia. Unpublished Ph.D. disseration, Yale University, New Haven. 1979. Ecology and subsistence. IN J. Jennings, ed., The Prehistory ofPolynesia, pp. 286-307. Cambridge: Harvard University Press. 1981. Lapitoid settlements of Futuna and Alofi, Westem Polynesia. Archaeology in Oceania 16:127-43. 1982. The impact of the prehistoric Polynesians on the Hawaiian ecosystem. Pacific Science 36:1-14. 1983. Man's role in modifying tropical and subtropical Polynesian ecosystems. Archaeology in Oceania 18:26-31. 1984. The Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge University Press. . 1986. Retirning East Polynesian prehistory. Journal of the Polynesian Society 95:940. 1988. Niuatoputapu: The Prehistory of a Polynesian Chiefdom. Burke Museum Monograph No. 5, Seattle. . forthcoming, The Wet and the Dry: Irrigation and Agricultural Intensilcation in Polynesia. Chicago: University of Chicago Press. Kirch, P. V., and T. S. Dye 1979. Etnarchaeology and the development of Polynesian fishing strategies. Journal of the Polynesian Society 88:53-76. Kirch, P. V., J. Flenley, D. W. Steadman, F. Lainont, and S. Dawson 1992. Prehistoric human impacts on an island ecosystem: Mangaia, Central Polynesia. National Geographic Research and Exploration 8:166-79. Kirch, P. V., and R. C. Green 1987. History, phylogeny, and evolution in Polynesia. Current Anthropology 28:431-56. Kirch, P. V., and T. L. Hunt 1988a. Radiocarbon 248 The Toaga Site dates from the Mussau Islands and the Lapita colonization of the southwestem Pacific. Radiocarbon 30:161-69. 1988b. The spadal and temporal boundaries of Lapita IN P. V. Kirch and T. L. Hunt eds., Archaeology ofthe Lapita Cultural Complex: A Critical Review, pp. 9-32. Burke Memorial Museum Research Report No. 5. Seattle. Kirch, P. V., and P. Rosendahl 1973. Archaeological investigation of Anuta. IN D. E. Yen and J. Gordon, eds., Anuta: A Polynesian Outlier in the SolomnIslands, pp. 25-108. Pacific Antropological Records 21. Honolulu: Benice P. Bishop Museum. Kirch, P. V., and D. E. Yen 1982. Tikopia: The Prehistory and Ecology of a Polynesian Outlier. Benice P. Bishop Museum Bulletin 238. Honolulu. Leach, H., and R. C. Green 1989. New information for the Ferry Berth Site, Mulifanua, Westem Samoa. Journal ofthe Polynesian Society 98:319-29. Leach, H. M., and D. Witter 1987. Tataga-matau "'ediscovered."' NewZealandJournalof Archaeology 9:33-54. Lepofsky, D., H. C. Harries, and M. Kellum 1992. Early coconuts on Mo'orea Island, French Polynesia Journal of the Polynesian Society 101:299-308. Mead, M. 1930. Social Organization ofManua. Bernice P. Bishop Museum Bulletin 76. Honolulu. Menard, H. W. 1986. Islands. New Yori: Scientific American Library. Poulsen, J. 1987. Early Tongan Prehistory: The Lapita Period on Tongatapu and its Relaionships, Vols. 1 and 2. Terra Austalis 12. Department of Prehistory, Research School of Pacific Studies, Australia National University, Canberra. Sinoto, Y. H. 1979. Excavations on Huahine, French Polynesia Pacific Studies 3:1-40. Smith, H. L. 1976. Jane's Camp (SUF1-1). IN J. Jennings, ed., Excavations on 'Upolu, Western Samoa, pp. 61-74. Pacific Anthropological Records 25. Department of Anthropology, Honolulu: Bernice P. Bishop Museum. Spriggs, M. 1986. Landscape, land use, and political transformation in southeem Melanesia. IN P. V. Kirch, ed., Island Societies: Archaeological Approaches to Evolution and Transformaon, pp. 6-19. Camnbridge: Cambridge University Press. -. 1990. Dating Lapita: Another view. IN M. Spniggs, ed., Lapita Design, Form and Composition, pp. 6-27. Occasional Papers in Prehistory No. 19. Departnent of Prehistory, Australian National University, Canberm Steadman, D. W. 1989. Exinction of birds in eastem Polynesia: A review of the record, and comparsons with other Pacific island groups. Journal ofArchaeological Science 16:177-205. Steadman, D. W., and P. V. Kirch 1990. Prehistoric extinction of birds on Mangaia, Cook Islands, Polynesia Proceedings of the U. S. National Academy of Sciences 87:9605-9609. Steadman, D. W., D. S. Pahlavan, and P. V. Kirch 1990. Exdtion, biogeography, and human exploitation of birds on Tikopia and Anuta, Polynesian outliers in the Solomon Islands. Bishop Museum Occasional Papers 30:118-53. Honolulu. Steams, H. T. 1978. Comparison of the geology of the Society and the Hawaiian islands. Bernice P. Bishop Museum Occasional Papers 24(14):275-90. Suggs, R C. 1961. Archaeology of Nuku Hiva, Marquesas Islanus, French Polynesia. Antrpological Papers of te American Museum of Natal History 49(1). New Yoik. Sullivan, A P., III 1893. Storage, non-edible resource processing, and the interpretaton of sherd and lithic scatters in Sonoran Desert lowlands. Journal of Field Archaeology 21:101-114. Tate, G. H. H. 1951. The Rodents ofAustralia and New Guinea. American Museum of Natural History Bulletin 97:187-430. Terrell, J. 1986. Prehistory in the Pacific Islands. Camnbridge: Cambridge University Press. Zenkovich V. P. 1971. A theory of the development of accumulation fonms in the coastal zone. IN J. A. Steers, ed., Introduction to Coastline Development, pp. 94-116. Cambridge: M. I. T. Press. .z Ing 0 am M.w Exploradons on the Makran Coast, Pakistan: A Search for Paradise George F. Dales and Carl P. Lipo. Drawing upon the diary of his 1960 expedition to the remote Makran coast, Dales recounts the search for evidence of commercial and cultural contacts between the ancient Indus civilization (2500 to 1900 B.C.) and the Near East. 1992. paper, 288 pp., 96 plates, 68 figures, 11 tables, and 1 foldout map, No. 50 24.00 The Archaeology and Ethnohistory of Fort Ross, Cal#fornia Vol. 1. K. G. Lightfoot, T. A. Wake, and A. M. Schiff. This new series on the archaeology and ethnohistory of Fort Ross examines the responses of Native Americans to Russian mercantile adctivities in northern California. This volume focuses on the interactions of Native Californians with the Russians and the Native Alaskans. 1991. paper, 250 pp., 33 maps, 12 appendices, 6 pp. of illustrations, No. 49 $18.00 Current Directons in Calfornia Obsidian Studies. Editor Richard E. Hughes, with contributions by Jonathon E. Ericson, Christopher M. Stevenson and Barry E. Scheetz, M. C. Hall and R. J. Jackson, Robert L. Bettinger, Thomas M. Origer, David A. Fredrickson and Mark E. Basgall. State-of-theart research on sourcing and hydration rate studies in California. 1990. 126 pp., No. 48 $14.00 Prehistoric Hawaiuan Occupation in the Anahulu Valky, O'Ahu Island: Excavations in Three Inland Rockshelters. Editor Patrick Kirch, with contributions by Terry L. Hunt, Sara Collins, Melinda S. Allen and Gail M. Murakami. Impact of Hawaiian occupation, circa A.D. 1300, on the local environment of the Anahulu Valley. Detailed studies of rockshelter sediments, archaeobotanical remains, charcoal, and landsnails. 1989. 130 pp., 30 plates, No.47 $12.00 Prehistoric Hunter-Gatherers of Shelter Island, New York: An Archaeological Study of the Mashomack Preserve. Authors Kent G. Lightfoot, Robert Kalin and James Moore. Case study of prehistoric subsistence and settlement patterns of Shelter Island, New York. The authors evaluate whether or not horticulture and sedentary lifeways were adopted widely by coastal hunter-gatherers. 1987. 224 pp., No.46 $9.00 Methods in Artfact Analysis: A Study of Upper Paleolithic Burins. Author Richard N. Dreiman. 1979. 79 pp., No.42 $5.00 Studks in Ancknt Mesoamerica, IV. Editor John A. Graham, with contributions by M. Johnson, E.M. Shook, M.P. Hatch, J.K. Donaldson, P. Mathews, D.M. Pendergast, D.C. Pring, D.S. Rice and P.M. Rice. Collection of papers on the archaeology, architecture and epigraphy of the Maya and the Olmec. 1979. 277 pp., 14 plates, No. 41 $8.00 An Archaeological Assay on Dry Creek, Sonoma County, California. Authors M.A. Baumhoff and Robert I. Orlins. 1979. 244 pp., No. 40 Il $8.00 For a complete list of titles and ordering information, please write to: Administrator, Archaeological Research Facility, Anthropology Department, University of Califomia, Berkeley, CA 94720. A