Lapps and Labyrinths Saami Prehistory, Colonization and Cultural Resilience Noel D. Broadbent With contribution by Jan Stora Lapps and Labyrinths is a detailed analysis of Saami prehistory from 5000 B.C. to A.D. 1500 along 500 kilometers of the Bothnian coast in northern Sweden. The Saami were highly specialized seal hunters who also practiced animal husbandry, farming and metallurgy in ways analogous to the Norse. In the early fourteenth century they were assimilated by the Swedish state, Christianized and driven inland where many later became nomadic reindeer herders. Their land-uses, place- names, technologies and spiritual ideas have strongly impacted north Swedish society and left an indelible, and yet little appreciated, imprint on Nordic culture. Contributions to Circumpolar Anthropology 8 For information and to order volumes in the Contributions to Circumpolar Anthropology series, please contact: Department of Anthropology National Museum of Natural History Smithsonian Institution 10th and Constitution Avenue, NW MRC 112 Washington, D.C. 20013-7012 Phone: (202) 633-1887 Fax: (202) 357-2684 www.mnh.si.edu/arctic Lapps and Labyrinths Saami Prehistory, Colonization and Cultural Resilience Lapps and Labyrinths Saami Prehistory, Colonization and Cultural Resilience NOEL D. BROADBENT WITH CONTRIBUTION BY JAN STORA fArctic C-enter Smithsonian Institution Scholarly Press WASHINGTON, D.C. 20I0 Published by the ARCTIC STUDIES CENTER Department of Anthropology National Museum of Natural History Smithsonian Institution P.O. Box 37012, MRC 112 Washington, D.C. 20013-7012 www.mnh . si .edu/arctic In cooperation with SMITHSONIAN INSTITUTION SCHOLARLY PRESS P.O. Box 37012, MRC 957 Washington, D.C. 20013-7012 www.scholarlypress.si.edu Copyright © 2010 by the Smithsonian Institution AU rights reseived. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of the publisher. Cover image: A painting of a bear burial marked by an antler-sheathed Saami spear. Ossian Elgstrom, 1930 (Norrbotten Museum). Colors used on cover inspired by the Saami flag. Library of Congress Cataloging-in-Publication Data: Broadbent, Noel. Lapps and labyrinths : Saami prehistory, colonization and cultural resilience / Noel D. Broadbent ; with contribution by Jan Stora. p. cm. "Published by Arctic Studies Center, National Museum of Natural History Smithsonian Institution ... in cooperation with Smithsonian Institution Scholarly Press." Includes bibliographical references and index. ISBN 978-0-9788460-6-0 (paperback : alk. paper) i. Sami (European people)—Sweden—History. 2. Sami (European people) — Sweden—Antiquities. 3. Coastal archaeology—Sweden. 4. Sweden—Antiquities. 5. Antiquities, Prehistoric—Sweden. I. Arctic Studies Center (National Museum of Natural History). II. Title. DL641.L35B76 2010 948.5'oo49455—dc22 2009028011 ISBN-13: 978-0-9788460-6-0 ISBN-io: 0-9788460-6-0 Printed in the United States of America @ The paper used in this publication meets the minimum requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992. Contents xi DIRECTOR'S NOTE William W. Fitzhugh xiii FOREWORD Inger Zachrisson XV PREFACE xvii ACKNOWLEDGEMENTS xix LIST OF FIGURES xxv LIST OF TABLES 1 1 INTRODUCTION AND NARRATIVE CONTEXT 2 Vikings, Dwarves and Giants 3 What Is the Cultural Identity of the Saami? 4 The Testimony of a Bear . . . 4 The Power of Historical Narratives 7 Archaeology and the Welfare State 9 The Means to an End 9 Notes 2 II CULTURE AND ECOLOGY II The Historical-Ecological Setting II The Importance of the Coastal Zone 13 The Saami People 19 Human-Environmental Relations 22 Glacial Topography and Settlement 23 Climate, Ice and Seal Oil 24 Periodicity and Settlement Cycles 24 World Systems and Northern Ecology 26 Punctuated Sedentism 26 Resilience Theory 27 Five Hypotheses 3 29 PREHISTORIC FOUNDATIONS 29 The Early and Late Stone Ages 31 Rock Art 32 New Ideologies, New Technologies 33 The Early and Late Metal Ages 35 Cairn Graves 36 Iron 36 Animal Husbandry and Cultivation 37 Artifacts in Context 40 Summary /g^HSO/vJJwN. 40 Economy / \ ( riAKObZOIO^J 4o Settlement 41 Technology — 41 Ideology and Religion 41 Conclusions 4 43 THE RESEARCH STRATEGY 43 Surveys 43 Site Investigations 44 Elevations 46 Excavations 46 Sampling 46 Soil Chemistry 46 Animal Osteology 47 Charcoal Analysis 47 Macrofossils 47 Radiocarbon Dates 47 Metallurgy 47 Shoreline Displacement 49 Lichen Growth on Uplifted Beaches 51 A Glossaiy of Archaeological Features 52 Huts 53 Cairns 53 Caches 53 Alignments 53 Circles f 54 Labyrinths ' 54 Compass Roses 54 "Russian" Ovens 54 Shooting Blinds 54 Engravings 5 57 EXCAVATIONS AND ANALYSES 57 Regional History 59 Bjuroklubb and Bjuron 60 Jungfrugraven, Site 79 63 Site 64 63 Site 65 63 Site 67 65 Site 68 65 Site 70 66 Jungfruhamn, Sites 138-139 66 Huts A, B and C 68 Lappsandberget, Site 144 72 Grundskatan, Site 78 72 Feature i, Hut 73 Feature 2, Hut 75 Feature 3, Hut 77 Feature 4, Hut (Bear Burial) 83 Feature 5, Hut 83 Feature 6, Hut 83 Feature 7, Hut 84 Phosphate Mapping 85 Feature 8, Cairn 85 Feature 9, Cairn 85 Feature 10, Baking Oven ("Russian" Oven) 86 Feature 11, Hut 87 Feature 12, Hut 87 Feature 13, Hut 88 Feature 14, Hut and Labyrinth 89 Feature 15, Pit 90 Feature 16, Stone Circle 91 Feature 17, Cache and Hut 91 Summary 92 Stora Fjaderagg Island 94 Archaeology of Stora Fjaderagg Island 95 Hut A 96 Hut B 105 Hut C 108 Hut D 109 Circular Features 112 Gamla Hamnen 113 Summary Discussion 113 Comparative Analysis of the Osteological Material (by ]an Stora) 113 Harp Seal Populations 121 Snoan Island 122 Site 49 123 Site 92 123 Site 53 124 Conclusions 124 Stor-Rebben Island 126 Hut A 127 Hut B 127 Hut C 128 Hut D 130 Summary 131 Hornslandsudde 132 Feature 5, Hut 134 Features 12-14, Double Hut and Enclosure 136 Feature 15, Hut 136 Features 19 and 25, Stone Alignments 136 Site 132 138 Caches 139 Summary and Discussion 6 141 CHRONOLOGY AND CULTURE 141 Source Criticism 142 Site Comparisons 144 The Start Dates 144 The End Dates , ^siiTTiTTiprv pnrl Oi'sni'^'^inriLi 1 1 111 idx y fXii\A. H-J J i\j Li 145 What Do the Artifacts Tell Us? 147 The Medieval Transition 149 Cycles of Change 7 THE ARCHAEOLOGICAL ROADMAP Architecture Site Structure 156 Storage and Surplus Corrals and Fences 158 Seals, Seasonality and Animal Husbandry Place-Name Evidence of Reindeer Husbandry 160 Pastoralism and Heterarchy 163 Discussion 165 Metallurgy 166 Three Iron Forges 166 Bjuroklubb 67 166 Grundskatan 78, Hut 11 Hom^l^inrl'snrlHp Sitp tto1 IWl llOlCtl 1V_1 l3 Ll\-it_i j kJ 1 1 1 168 Smithing and Shamanism 171 Volundr and the North 171 Conclusions 8 173 RITUALS AND RELIGION 176 The Wild and the Domesticated 177 The Bear Burial 180 The Grundskatan Find 181 The Setting and Social Context 181 The Economy and Timing 181 The Dwelling as the Saami Cosmos 183 The Tree of Death and Rebirth 183 Conclusions 184 Circular Sacrificial Features 187 Hornslandet and Yttre Bergon, Halsingland 188 Lappsandberget, Vasterbotten 188 Jungfrugraven, Vasterbotten 189 Stora Ejaderagg, Vasterbotten 189 Grundskatan, Vasterbotten 189 Overall Chronology 190 Social Context 191 Siunmary 193 PLACE-NAMES AND CHURCH TOWNS 193 Place-Names 197 The Geography of "Lapp" Place-Names 199 Lapp Place-Names in Sweden 201 Conclusions about Place-Names 201 A Model for Coastal Settlement in Lovanger 204 The Church Town 207 Assimilation and Change 208 Summary and Discussion 211 THE LABYRINTH AND THE BEAR 217 SYNTHESIS 220 Summary of Results 222 Einal Thoughts 225 APPENDIX i: Osteological Material 229 APPENDIX 2: Radiocarbon and AMS Dates 231 REFERENCES 257 INDEX 269 ABOUT THE AUTHOR Director's Note Scandinavian archaeology has a long and revered history leading back to the foundation of scientific archaeology pioneered by Christian Thomson, a Dane who devised the three-age classification system (Stone, Bronze, Iron Ages) and Oscar Montelius, a Svv'ede who first de- veloped the seriation method of relative dating based on style-change through time. In the 1940s Gutorm Gjessing, a Norwegian, was one of the first to begin promoting social interpretation of archaeological remains, a view later developed by Frederik Barth, a Norwegian, by developing the anthropological theory of social boundaries as expressed in visible signaling of material culture, style and design. In Lapps and Labyrinths Noel Broadbent, an American who lived and taught for many years in Sweden, carries this tradition of archaeological innovation into the problematic field of historical ethnicity - in this case the social and territorial history of the Swedish Saami. I visited many of the sites discussed in this book with Noel in 1984, before their significance had become obvious from his excavations of the past decade. Like Noel, I spent many years conduct- ing "boulder-field" archaeology in a similar subarctic environment, central and northern Labrador. I had found this work exceedingly frustrating because the corrosive nature of subarctic soils and transient nature of the sites resulted in poor artifact preservation and recovery. One was often left with elaborate maps of sites and structures of a people whose culture and identity remained un- known or conjectural. One could easily describe the architectural forms, but determining who they were was frequently elusive. Broadbent 's careful excavation techniques and ingenious analytical methods have turned the archaeology ofboulder-field sites from a confusing conundrum to a coher- ent picture that overturns a century of conventional archaeological wisdom about Saami origins, settlement systems and adaptation. For the first time the ubiquitous but inscrutable boulder sites lining the raised beaches and terraces of Sweden's northern Baltic coast sites have been shown to be the remains not of recent Germanic pioneers but of people who must have been ethnic Saami - but Saami living a very different life than known from historical records. Integrating archaeological finds with an array of anthropological data, place-names, history, religion, geography and ecology, Broadbent has produced a revolutionary new synthesis that indi- cates a former, long-term Saami occupancy of the North Swedish coast and outlines a model of cul- ture change and acculturation stimulated by the northward advance of Germanic-Swedish farmers and fishermen. Rather then viewing this history as one of ethnic confrontation and geographic and political isolation - processes that have characterized Saami relations with the Swedish state during recent centuries and continue today - Broadbent reconstructs a Late Medieval period characterized XI by processes of accommodation, cultural exchange and demographic mixing. Only later did insti- tutionalized nation-state policies begin to exclude Saami rights from traditional coastal territories and resources. In time those policies resulted in the re-definition of Saami ethnicity and identity into the reindeer herder of the upper river valleys, interior lakes and mountain zones where most Saami had lived exclusively since the 1700s. Noel Broadbent's research raises many questions that call for further study. More data are needed from other regions of the Baltic coast; correlations are needed between archaeological re- mains and Lappish place-names in Sweden south of the study area. Relations between traditional Saami shamanic religion, bear cults and medieval Christian practices need exploring. Coastal and interior archaeological sites need more comparative study. Broadbent's work lays out a new para- digm that powerfully calls into question the established version of Swedish and Saami histories as separate and apart; it sets forth a new conception of social history for the North Baltic, and perhaps even the greater North Nordic region, in which the Saami have to be seen as more important play- ers in the history of their respective modern states than previously accorded through history and ethnology. For these reasons this work should be of interest not only to archaeologists and culture histo- rians of northern regions but to students of anthropology, history, linguistics, political science and native studies. It is a work in the broadest of anthropological tradition and breaks new ground in the application of archaeological and anthropological methods to issues of modern concern. While dealing with the history of a small Saami population in a restricted area of the northwestern Baltic, the historical situation that transpired following the appearance of newcomers in their lands has been experienced by many indigenous peoples around the world. In this sense Broadbent's Lapps and Labyrinths has broad application and demonstrates the value of anthropological studies for bal- ancing the dominance of history in native studies. This work is in the best tradition of Scandinavian archaeology and breaks new ground for science and society. William W. Fitzhugh, Director Arctic Studies Center Smithsonian Institution xii DIRECTOR'S NOTE Foreword Saami bear burials seem to retain their magic, even today. They are eye-openers, being such explicit expressions of Saami identity. They clearly tell that the Saami were here! This insight gave Noel Broadbent a new direction in his research - one that he had not expected. The same befell me in 1970. As the archaeologist in charge at Vasterbotten County Museum in Umea, northern Sweden, I was commissioned to investigate a bone find on an island in Lake Storuman in Lapland. This turned out to be a well-preserved bear burial, not older than 250 years. Up to then I was specialised in the metal techniques of the Scandinavian Viking Age. But just like Noel, I became fascinated by Saami cultural history. A totally new world opened up before me. I still have that fascination, and it surprises me that more Swedish archaeologists have not been affected in the same way. Since then, I have worked with Saami archaeological material and its relationship to Scandinavian culture, initially with material from undisputed Saami areas - that is, the inland regions north of the River Angermanalven. But in 1982, I realised that a cemetery from the elev- enth and twelfth centuries, much further south at Vivallen in northwestern Harjedalen (excavated in 1913), must also be Saami - nobody was more astonished than myself Two hearths from the ninth and thirteenth centuries typical of Saami huts were found nearby and we excavated a hut foundation as well as other remains from the eleventh century. This led to an interdisciplinary project of early Saami culture and its relationship to the Scandinavian (Germanic/Nordic) culture in the central part of the Scandinavian Peninsula. This project showed that the Saami had been there at least 2,000 years and had extended south of there to about the 60th parallel. The book Moten i gmnsland: samer och germaner i Mellanskandinavien (Encounters in Border Country: Saami and Germanic Peoples in Central Scandinavia) was published by the Swedish Museum of National Antiquities in 1997. The history of the Saami in central Scandinavia has also become important regarding one of the most extensive court cases in Sweden, about the rights of the Saami in Harjedalen to let their reindeer graze in winter on private land. The Saami lost their case in 1996 and on appeal in 2002, but this case has recently (2009) been accepted by the European court of justice. For over a century, the dominant standpoint had been that the Saami had only relatively recently immigrated into this area and that they had not reached their southernmost territories until the eighteenth century. A newer view, and the role of the Saami in Swedish prehistory, is possible today. This is in reality a return to views prevalent for most ofthe nineteenth century, namely that the Saami have a very long history, not only in northern, but also in central Scandinavia. All available source materials lead to XIII the conclusion that Saami cultures emerged out of local hunter-gatherer cultures, just as Noel has argued for coastal Vasterbotten. New genetic research tells its own story, showing that there are many connections between today's Saami and the first people to arrive in Scandinavia more than 10,000 years ago. Swedish archaeology developed out of the seventeenth century goal of demonstrating our national greatness. Unfortunately, there are still attitudes within Swedish archaeology that can be characterized as ethnocentric, nationalistic and chauvinistic. The ethnic pluralism that once existed in Sweden has all too often been overlooked in favour of a one-sided focus on "Swedish" prehis- tory. As a consequence, people without their own written histories are often left defenceless in the courts. Many Swedish archaeologists still look upon Saami culture as static and inferior - even as non-definable. If the archaeological material does not coincide with historically known Saami cul- ture, they find it hard to imagine that it can be Saami. It will be a long time before the "new" ways of looking at the role of the Saami in the historical process filters through to the government and the general public. The museums - and even more so, the schools - have much to do in this regard. Even in local archaeological exhibits in northern Sweden the Saami are most often presented as an exotic minority of the eighteenth and nineteenth centuries, while the visitor gets the impression that it is the Scandinavian farmer whose culture extends back to the Stone Age - not the other way around. The research work mentioned above has up till now almost totally concentrated on mountain and forest areas in the inland of Scandinavia. The Bothnian coast(land) in Sweden and its Saami connections were on the whole unknown. The research results by Professor Broadbent and his col- leagues are therefore a minor revolution in our knowledge of the Saami past. Lapps and Labyrinths presents new interpretations of Saami prehistory in Sweden as well as innovative interdisciplinary methods and theoretical approaches to the study of ethnicity in archaeology. A large number of sur- veys, excavations and analyses are put into the context of long-term ecological and cultural changes. The study of Lapp place-names has not previously been the object of this kind of project and opens the door to much future archaeological research. Lapps and Labyrinths is easy to read, and the hypotheses and conclusions are well argued in simple to understand language. The excellent illustrations are in no small measure part of the book's impact - they are pedagogical enough to be accessible to the broader public and for use in schools. To conclude, Lapps and Labyrinths is an important contribution to Saami and Swedish pre- history, to the history of northern Europe and to indigenous studies everywhere. Inger Zach risson , Ph.D. Associate Professor ofArchaeology and Curator emerita Museum of National Antiquities Stockholm xiv FOREWORD Preface This book is the culmination of an academic journey that started in 1979 when I completed my doctoral dissertation at Uppsala University, followed by a seven-year stint as the Director of the Center for Arctic Cultural Research at Umea University. My own academic ambitions took a back seat when I moved to Washington, D.C., in 1990, and for six years I managed other people's research at the National Science Foundation (NSF). Three individuals that deserve my appreciation for those amazing years are the late Dr. Peter Wilkniss (Director, Polar Programs), Dr. lerry Brown (Head, Arctic Section), and Dr. Robert Corell (Director, Geoscience). In 1996, 1 was awarded the Chair ofArchaeology at Umea (later merged with Saami Studies) and for seven years commuted between Umea and Washington, D.C. Many exciting things hap- pened during my tenure, one of which was the Northern Crossroads (Motm i norr) Project. This Bank of Sweden-funded project energized the department on multiple fronts and provided full- time salaries for a post-doctoral position and nine doctoral students (including one in Stockholm and one in Lund), but unfortunately, once again, left me with little research time of my own. I chose to finally resign that position late in 2003, a decision that ironically brought my own original research ambitions back to life. With the generous support of NSF starting in 2004, I could move forward at last, and this has resulted in the book you now hold in your hands. I came to know and work with many Nordic Saami and Saami organizations such as the Nordic Saami Council and the Swedish Saami National Organization (SSR), and through the NSF Office of Polar Programs, I was able to travel widely in Alaska. In the latter instance, I would like to acknowledge Julie Kitka, President of the Alaska Federation of Natives; Dr. Ray Barnhardt and Dr. Oscar Kawagley, University of Alaska, Fairbanks; Caleb Pungowiyi (former president of the Inuit Circumpolar Conference); Amy Craver, Alaska Native Science Commission (now with the National Park Service); and Patricia Cochran, Alaska Native Science Commission (former president of the Inuit Circumpolar Conference). Thanks to these contacts my research started to evolve into a deeper appreciation of the meaning of knowledge to the people of the North. This is where my journey has led me. My scholarly life has literally been divided by worlds apart, both geographically and spiritually, and I have found peace in the wisdom that has now brought me full circle, and to a more meaningful completion of this project than I had originally envisioned. I trust this can serve not only the interests of the archaeological community and the development of the field, but of those who are invested in understanding the past - teachers, policy makers and all who live in the North. XV Pronunciation and Saami Orthography Swedish has three letters at the end of the alphabet that are unfamiliar to English speakers: "a" is pronounced like o in /ore (long) and like o in yonder (short); "a" is pronounced like ai in fair (long) and like e in best (short); "o"is pronounced like eu in the French deux, and before an r like u infur, or like an e in her. The "o" in Swedish is written as "0" in Danish and Norwegian and a can be written as aa in Danish. In order to avoid confusion regarding Saami spellings, which are often Swedish versions of Saami words in older literature (e.g., Manker i960), Lule Saami orthography, as sum- marized in Rydving (1995), has been used. Exceptions to this are quotations or references to North Saami sources, such as sijdda (South Saami) versus siida (North Saami). xvi PREFACE Acknowledgements There are many people I would like to thank in Sweden: Professor Bertil Almgren at Uppsala University, who had so welcomed me to the Department of Nordic and Comparative Archaeol- ogy, and saw me through the completion ofmy dissertation; the late Docent Hans Christians- son, who brought me into the Swedish North; Professor Evert Baudou, who encouraged me to join the Department of Archaeology in Umea and was instrumental in the establishment of the Center for Arctic Cultural Research; Professor Lars Beckman, the late Umea University Rector; Umea University Director Dr. Dan Brandstrom, later Director of the Bank of Sweden Tercentenary Fvmd; the late Professor Hakan Linderholm, M.D., with whom I helped organize the 7th International Conference on Circumpolar Health - all of whom truly understood the value of international re- search in Sweden. Academic colleagues for whom I also feel a special fondness are Professor Ake Hyenstrand, Professor Phebe Fjellstrom and Professor Mats Maimer. These old friends are gone now, but not forgotten. I wish to also acknowledge Carina Lahti (the heart and soul of the Department of Ar- chaeology in Umea), Professor Bozena Werbart, Professor Birgit Arrhenius, Docent Anders Carls- son, Docent Patrik Lantto and Margareta Axelson, stalwart friends over the years. Dr. Lana Troy (Professor of Egyptology in Uppsala) has been my de facto editor-at-large, whipping me into gram- matical shape and maintaining my intellectual rigor. Film-maker and photographer Boris Ersson (Lulea) has been a great collaborator in recent years. Docent Inger Zachrisson (Swedish National Historical Museum) is an icon of Saami archaeology in Sweden, having stood her ground when most others did not. I am honored by her foreword. Together with Dr. Inga-Maria Mulk (former director of the Ajtte Mountain and Saami Museum, lokkmokk) and Dr. Ingela Bergman (director of the Silver Museum in Arjeplog), she has laid the groundwork for this project and others like it. The wonderful years and productive collaboration at the Center for Arctic Cultural Research were made possible by Professor Roger Kvist and Dr. Rabbe Sjoberg. Both made invaluable contribu- tions to this book through their research in the Seal Hunting Cultures Project, and Rabbe rendered my field drawings into fine ink illustrations. Elaine Reiter, while a student at Northern Virginia Community College, transformed many of the figures into Photoshop® masterpieces. My interns: Kim Consroe at George Washington University, Jacquelyn Graham at the University of Minnesota, together with Dr. Katherine Rusk helped produce great reconstructions and maps. Intern Aza Der- man. The Bronx, helped with my bibliography using EndNote®. Dr. Pam Stern (formerly of Sterling College) gave early critiques of my manuscript and has made many helpful suggestions; and Cara XVII Seitchek and especially Ginger Strader (Smithsonian Scholarly Publications Manager) worked on the final edits. Unless otherwise noted, all translations, figures, tables, and photographs and con- clusions are by the author, who is solely responsible for their content. The aerial photographs are published with the permission of the Swedish Department of Defense. Special thanks naturally go to my colleagues Britta Wennstedt Edvinger, and Jan Stora (Os- teological Research Laboratory, Stockholm University). Britta and Kjell Edvinger (Arkeologicentrum i Skandinavien AB) facilitated new excavations at Grundskatan in 2004 and an international field school at Hornslandsudde in 2005. Britta and I have co-authored a number of articles through this project and her knowledge of Saami prehistory has been invaluable. Her contributions are particularly important in Chapter 8 (Rituals and Religion). Jan has done wonders with the animal bones, bringing many aspects of the investigation to life. He has made a major contribution to this book and all the osteological tables, figures and analyses are his work, none of which have been published before. Eva Hjarthner-Holdar, Director of the Geoarkeologiska Laboratoriet, and her colleagues in Uppsala (Eva Grandin, Emma Gronberg and Daniel Andersson) have advanced our knowledge of northern iron working through their analytical reports. Imogen Poole (Utrecht) and David Black (University of Western Michigan) have revealed the identities of trees hidden in bits of charcoal, Johan Linderholm, soil chemistry, and Karin Viklund, macrofossils (both Umea University). Thanks also to Ulf Lundstrom (Skelleftea Museum) for his inspirational manuscript on Saami place-names and eskers in Skelleftea. There are a number of people to thank here at the Smithsonian, many of whom have read my texts: Dr. Dan Rogers, Dr. Don Ortner, Dr. Bill Honeychurch (now at Yale), Dr. Torben Rick, Dr. Mary Jo Arnoldi, Dr. Bruno Frolich, Dr. Candace Greene, Dr. Mindy Zeder, Ann Kaupp and Kathleen Adio. Marcia Bakry helped with the illustrations, as did Beatrix Arendt. My sincere gratitude also goes to my colleagues of many years at the Arctic Studies Center: Dr. Bill Fitzhugh, Dr. Stephen Loring and Dr. Igor Krupnik. For the financial support that made this all possible, the National Science Foundation's Arctic Social Sciences Program and its Director Anna Kerttula. These years at the Department of Anthropology and the National Museum of Natural History have been among the most enjoyable and stimulating ofmy career. Finally last, but hardly least, thanks to my patient wife Elaine and our daughter Rosanna, who was born in Uppsala and has been on Swedish archaeological sites before she could walk, and Greg Lavallee, who has been my in situ information technology expert. Thank you one and all! Noel D. Broadbent Washington D.C. XVlil ACKNOWLEDGEMENTS List of Figures Fi^Jure I. View of sealer's hut, Stora Fiaderagg Island. xxviii FijJure 2. Front plate of Sven Nilssoris The Aboriginal Population ofNordic Scandinavia. 6 FijJure J Eskimo igloo and artifacts (Nilsson 1866). 7 FiiJure 4. The Circumpolar Region and the area of study. 10 Fi^Jure 5. Main river drainages of Sweden into the Gulf of Bothnia. 12 Fi^Jure 6. Provinces and counties of Sweden. 13 Figure 7. Carta Marina map by Olaus Magnus from 1539. 14 Figure 8. Saami summer camp in North Norway. 15 Fi^Jure 9- Present Saami territory and language areas. 16 Fi^Jure 10. Map based on a description of the Nordic region by Adam of Bremen. 17 Figure II. Visual cross section of North Swedish biogeography. 20 Fi^Jure 12. Map showing the oak limit, alpine and Bothnian ice regions. 21 Fi^Jure J Numbers of seals taken by county and by length of coastline. 22 Fi^Jure 14. Map showing the mtDNA of voles (and viruses) in northern Sweden. 23 Fi^Jure 15- Cycles of moose populations in the taiga and cycles of coastal settlement. 26 Fi^Jure 16. Environmental changes, stray finds, site aggregations and agrarian indications in Vasterbotten. 28 Figure 17- Stone artifact types from Lundfors. 31 Fi^Jure 18. Lundfors inlet and the locations of the sites and lines of seal nets. 31 Fi^Jure 19. An "x-ray" rock carving at Stornorrfors in coastal Vasterbotten. 32 Fi^Jure 20. Unused thick-butted flint adzes from the Bjurselet site. 33 Fi^Jure 21. A wooden spear shaft from Ostra Abyn in Bygdea parish. 33 Fi^Jure 22. Map showing the spread of related technologies into the Nordic region. 34 Fi^Jure 23. Sketch of both large and small grave cairns in Vasterbotten. 35 Fi^Jure 24. Distribution of asbestos-tempered vessels. 36 FijJure 25- Ananino mold for a socketed axe. 37 Fi^Jure 26. Iron Age artifacts from coastal Vasterbotten. 38 Fi^Jure 27. Two decorated Viking Age Saami skis from Vasterbotten. 39 Fi|Jure 28. Eskers, sites mentioned in the text, and economic zones. 40 Fi^Jure 29. North to south transect along the Bothnian coast. 42 Fi^Jure 30. Distribution of Iron Age huts along the Bothnian coast. 44 xix Fi^^ure 31- Detail of Gulf of Bothnia and the locations of the investigated sites. 44 Fi^'lire 32. Aerial view of the uplifted fishing harbor at Bjuroklubb. 45 Fi^Jure 33- Watermark at Ratan, 40 km north of the Umea. 48 Fij^ure 34. Shoreline displacement curve for Vasterbotten. 49 Fi^Jure 35- Rhizocarpon geographicum lichens. (Lovanger Church and Bjuroklubb) 50 Fi^Jure 36. Maximum diameters of R. geographicum on seven beaches. 51 Fi^Jure 37- Hut foundation at the Grundskatan site (Site 78, Fiut 3). 52 Fi^Jure 38. Small cache beside a hut on Stora Fjaderagg Island (Sites 31-32). 53 Fi^Jure 39- Stone alignment at the Grundskatan site. 53 Fi^Jure 40. Labyrinth at Ratan in Vasterbotten. 54 Fi^Jure 41. Compass rose on Snoan Island. 54 Fi^Jure 42. Large intact Russian Oven in Osterbotten, Finland. 54 Fi^Jure 43- Watermark at Ratan in Vasterbotten carved during the reign of Gustav III. 55 Fi^Jure 44. Engraved names, ownership marks and dates carved by sealers from Osterbotten. 55 Fi^Jure 45. Aerial view of Stor-Rebben Island. 56 Fi^Jure 46. The Ume and Pite Lappmarks and Skellefte and Lovanger parishes. 58 Fi^Jure 47. Lovanger church town. 59 Fi^Jure 48. Woodcut of Bjuroklubb from 1555. 59 Fi^Jure 49. Aerial view of Bjuroklubb. 61 Fi^Jure 50. Map of site locales on Bjuron and the 10 m elevation. 61 Fi^Jure 51- Photo at Jungfrugraven. 62 Fi^Jure 52. Map showing the walls and cairn stones at lungfrugraven. 62 Fi^Jure 53- Aerial view of Site 64. (Bjuroklubb) 63 Fi^Jure 54. Sketch map of Site 65. (Bjuroklubb) 63 Fi^Jure 55- Sketch map of Site 67. (Bjuroklubb) 64 Fi^Jure 56. Plan of excavation of Hut, Site 67. (Bjuroklubb) 64 Fi^Jure 57- Plan of hearth excavated near Site 67. (Bjuroklubb) 65 Fi^Jure 58. Sketch map of hut row, Site 68. (Bjuroklubb) 65 Fi^Jure 59- Map of Site 70 and three adjacent huts and wall alignments. (Bjuroklubb) 65 Fi^Jure 60. Phosphate sampling of Site 138. (Jungfruhamn) 66 Fi^Jure 61. Detailed maps of Huts A and B. (Jungfruhamn) 67 Fi^Jure 62. Profile drawn diagonally through Hut A hearth. (Jungfruhamn) 67 Fi^Jure 63. Map of Site 144 and excavated areas. (Lappsandberget) 68 Fi^Jure 64. Exposed surface within the stone circle. (Lappsandberget) 69 Fi^Jure 65. Close-up of one of three lichens growing on the stone circle. (Lappsandberget) 69 Fi^Jure 66. Map of stone circle showing soil deposits. (Lappsandberget) 70 Fi^Jure 67. Aerial view of wave-washed moraine beaches at the Grundskatan site. 70 Fi^Jure 68. Map showing archaeological features at the Grundskatan site. 71 Fi^Jure 69. Aerial view of the Grundskatan site. 71 Fi^Jure 70. Map showing Hut i. (Grundskatan) 72 Fi^Jure 71- Maps and profiles of Hut 2. (Grundskatan) 74 Fi^Jure 72. Map of Hut 3 with section of a small posthole. (Grundskatan) 75 XX LIST OF FIGURES t mre / J Map of the central area of the Grundskatan site with feature numbers. 77 Fif t ^ure 74. Map and profile of Hut 4. (Grundskatan) 77 nire Photograph of Hut 4 during excavation. (Grundskatan) 78 Fu c ^ure 76. Map of the cairn and bear bone finds in the southeast corner of Hut 4. (Grundskatan) 79 Fi^Jure 77. Calibrated dates of the hearth and bear bones in Hut 4. (Grundskatan) 79 FitJure 78. Map of southeast corner of Hut 4. (Grundskatan) 80 Vk c mre 79- Map of southeast corner of Hut 4 showing location of bear bone deposition. (Grundskatan) 80 Fi^Jure 80. Three profiles through Hut 4. (Grundskatan) 81 Fi^mre 81. Map of Hut 5. (Grundskatan) 82 Fi^Jure 82. Map of Hut 6. (Grundskatan) 83 Fi£Jure 83. Map of Hut 7. (Grundskatan) 83 Fi^mre 84. Phosphate map. (Grundskatan) 84 Fi^Jure 85. Map of Feature 8. (Grundskatan) 84 Fi^Jure 86. Map of Feature 9. (Grundskatan) 85 Fi^mre 87. Photo of Russian Oven (Feature 10) at Grundskatan. 85 Fi^mre 88. Map of Hut II and profiles through the hearth. (Grundskatan) 86 Fi^Jure 89. Map of Hut 12 and profile of the hearth. (Grundskatan) 86 Fi^Jure go. Map of Hut 13. (Grundskatan) 87 Fi^Jure 91. Calibrated radiocarbon dates from Hut 13. (Grundskatan) 87 Fi^mre 92. Map of labyrinth and hut wall at Grvuidskatan, Feature 14. 88 FifJure J J Profile of the hearth under the labyrinth in Feature 14. (Grvuidskatan) 88 Fi^mre 94. Greg Lavallee standing in Feature k. (Grundskatan) 89 Fi^mre Map of Feature 16. (Grundskatan) 90 Fi^Jure 96. Photo and drawing of Feature 17. (Grundskatan) 90 Fi^Jure 97. Radiocarbon dates from the Grundskatan site. 91 Fi^Jure 98. Three-dimensional rendition of topography of Stora Fiaderagg Island and map showing locations of investigated huts and labyrinths. 93 Fi^Jure 99. Map of Sites 31-33. (Stora Fiaderagg) 94 Fi^Jure 100. Map of Hut A. (Stora Fjaderagg) 94 Fi^Jure lOI. Map of Hut B excavation and locations of profiles. (Stora Fjaderagg) 95 Fi^Jure 102. Anatomical representation by numbers of identified specimens. (Stora Fjaderagg) lOI Fi^Jure 103. Anatomical representation in the floor area. (Stora Fjaderagg) 103 Figure 104. Anatomical representation in the hearth. (Stora Fjaderagg) 103 Fi^Jure los- Anatomical representation in the storage pit. (Stora Fjaderagg) 10^ Fi^Jure 106. Map of Hut C with alignments, huts and a storage cache. (Stora Fjaderagg) 107 Fi^Jure 107. Map of Hut D. (Stora Fjaderagg) 108 Fi^Jure 108. Sketch map of Site 33. (Stora Fjaderagg) no Fi^Jure 109. Photos of features at Site 33. (Stora Fjaderagg) no Fi^Jure no. Maps of huts (A-B) by Gamla Hamnen. III Fi^Jure III. Map of Gamla Hamnen. III LIST OF FIGURES Fi^Jure 112. Aerial view of Gamla Hamnen. 112 Fi^Jure 113. Calibrated dates of huts at Sites 31-32. (Stora Fjaderagg) 113 Fi^Jure 114. Level of epiphyseal fusion of finger and toe bones in seals. (Stora Fjaderagg) 114 FijJure 115. Size comparison of two temporal bones from Stora Fjaderagg. 114 Figure 116. Anatomical representation of seal bones in Hut B. (Stora Fjaderagg) 118 Fi^Jure 117. Anatomical representation of seal bones in Hut B, by weight. (Stora Fjaderagg' Fi^Jure 118. Anatomical representation of seal bones in Hut B and Hut D according to MAU (minimum anatomical units). (Stora Fjaderagg) 119 Fi^Jure 119. Three-dimensional rendition of Snoan Island with radiocarbon dates. 120 Fi^Jure 120. Map of Snoan Island. 121 Fi^Jure 121. Map of Site 49. (Snoan Island) 121 Fi^Jure 122. Photo of Hut B with distinctive hearth. (Snoan Island) 121 Fi^Jure 123. Map of fishing harbor, Site 92. (Snoan Island) 122 Fi^Jure 124. Map of fishing harbor area, Site 53. (Snoan Island) 123 Fi^Jure 125. Calibrated dates from Snoan Island. 123 Fi^Jure 126. Photo of Hut A. (Stor-Rebben Island) 124 r Fi^Jure 127. Map showing investigated huts on Stor-Rebben Island. 124 Fi^Jure 128. Ann Wastesson and Ann-Christin Nilsson excavating. (Stor-Rebben Island) I2SJ Fi^Jure 129. Map of Hut A at Stor-Rebben. 125 Fi^Jure 130. Profile of trench through Hut A. (Stor-Rebben Island) 125 Fi^Jure 131. Map of Hut B and three labyrinths. (Stor-Rebben Island) 126 Fi^Jure 132. Excavation unit in Hut B. (Stor-Rebben Island) 126 Figure 133- Profile of the hearth in Hut B. (Stor-Rebben Island) 126 Fi^Jure 134. Map of Hut C. (Stor-Rebben Island) 127 Fi^Jure 135. Calibrated radiocarbon dates of huts on Stor-Rebben Island. 128 Fi^Jure 136. Map showing location of Hornslandsudde, Sites 119 and 1^2. 129 Fi^Jure 137- Hornslandsudde site area based on photogrammetry (Eriksson 1975). 130 Fi^Jure 138. Hut 5, showing double wall lines. (Hornslandsudde) 132 Fi^Jure 139. Excavated area in Hut 5. (Hornslandsudde) 132 Fi^Jure 140. Trench through Feature 13. (Hornslandsudde) 133 Fi^Jure 141. Section through connecting wall between Huts 12-13. (Hornslandsudde) 133 Fi^Jure 142. Features 12-14 showing soil samples. (Hornslandsudde) 133 Fi^Jure 143. Excavation of Hut 15. (Hornslandsudde) 134 Figure 144. Map of sections of two stone alignments. (Hornslandsudde) 134 Fi^Jure 145. Site 132, Hut A. (Hornslandsudde) 135 Fi^Jure 146. Site 132, Hut B. (Hornslandsudde) 135 Figure 147. Map of Site 132 showing Hut A, Hut B and a cache. (Hornslandsudde) 135 Figure 148. Cairn by Site 132, Hut B. (Hornslandsudde) 136J Fi^Jure 149. Reconstruction of Hut B, lean-to and storage cairn at Site 132. (Hornslandsudde) 136 Fi^Jure 150. Cache #4 next to bedrock outcrop. (Hornslandsudde) 137 Fi^Jure 151. Cache #1 in wave-washed moraine field. (Hornslandsudde) 137 Fi^Jure 152. Map of cache locations. Sites 119 and 132. (Hornslandsudde) 137 xxii LIST OF FIGURES Jure 153- Calibrated dates from Hornslandsudde. 137 Fi{^ure 154. Map of excavation of Huts 12-13 Hornslandsudde, Site 119. 138 Fi|'ure 155- Diagram illustrating the elevations of radiocarbon dates. 140 Fi|Jure 156. Representative sample of 24 C-14 dates. 143 Fi^Jure 157- Numbers of dates of Bothnian sealing huts. 144 Fi^Jure 158. Graphic distributions of the latest sealing hut at Grundskatan, a farmstead locale and a Saami hearth from Bjuroldubb. 148 FijJure 159. Calibrated dates of the latest sealing hut at Grundskatan, the Bole farmstead site, Gamla Hamnen on Stora Fjaderagg Island and a Saami hearth on Bjuroklubb. 148 Figure 160. Chronology of Bothnian sealing huts. 150 Fi^Jure 161. Lengths and widths of Bothnian huts. 152 Fi^Jure 162. Hut forms by radiocarbon date. 152 FijJure 163. Saami goat hut. 152 Fi^Jure 164. Coastal hut forms in North Norway and the Kola Peninsula. 153 Fi^Jure 165. Clustering of huts at Grundskatan. 154 Fi^Jure 166. Clustering of huts at Stora Fjaderagg. 155 Fi^Jure 167. Clustering of huts at Hornslandsudde. 155 Fi^Jure 168. Typical clustering of 5 dwellings. 156 Fi^Jure 169. Reconstruction of site complex near Site 70 on Bjuron. 156 Fi^Jure 170. Various possible hut forms and constructions at Grundskatan. 157 Fi^Jure 171. Numbers of place-names with the prefix ren (reindeer) and rmgard (reindeer corral). 160 Fi^Jure 172. Bjuroklubb 67. 167 Fi^Jure 173- Vent in Bjuroklubb 67. 167 Fi^Jure 174. Map of Hut II at Grundskatan. 168 Fi^Jure 175- Reconstruction of Huts 12-1}. (Hornslandsudde) 168 Fi^Jure 176. Detail of forging hearth in Hut 12, Hornslandsudde. 169 Fi^Jure 177. Slag from Grundskatan Site 78, Hut 11. 169 Fi^Jure 178. Iron scales, ropey iron and sphericals from Grundskatan Site 78, Hut 11. 169 Fi^Jure 179. Furnace clay from Bjuroklubb 67. 169 Fi^Jure 180. Map showing general distribution of documented bear burials and bone depositions in Sweden and Norway. 172 Fi^Jure 181. A Saami at an offer site with a stone seite and antlers. 175 Fi^Jure 182. Animal offerings at 357 sites. 176 Fi^Jure 183. Find-places of bear burials and bone depositions in Norway and Sweden. 178 Fi^Jure 184. Percentages of bear burials and bone depositions by landscape feature. 179 Fi^Jure 185. Bear burials by date. 179 Fi^Jure 186. Bear bones from Hut 4 at Grundskatan. 180 Fi^Jure 187. Map of circular sacrificial sites in Sweden and Norway. 185 Fi^Jure 188. Photo of a circular feature at Gagsmark. 187 Fi^Jure 189. Circular sacrificial feature. Site 133, Hornslandet. 188 Fi^Jure 190. Circular features in their social context. 190 Fi^Jure 191. Late Iron Age coastline and place-names referred to in the text. 192 LIST OF FIGURES XXIli Fi^Jure 192. Map of some Saami place-names and characteristic Lapp place-names near Skelletftea and Umea. 194 Fi^Jure 193. Numbers of place-names with the prefix "Lapp" in Upper Norrland. 195 Fi^Jure 194. Map of Lovanger church town. 197 Fi^Jure 195. Distribution of 1,147 Lapp place-names in Sweden. 200 Fi^Jure 196. Numbers of place-names with the prefix "Lapp" by county. 200 Fi^Jure 197. Map and profile drawing of the earth wall at Broange. 203 Fi^Jure 198. Sixteenth-century seal-netting areas of Lovanger parish. 205 Figure 199. Map of villages and coastal economic zones. 205 Figure 200. Reconstructed sijdda village territories in Lovanger. 207 FijJure 201. Map of Nordic labyrinths. 210 Fi^Jure 202. Displacement of the Saami from the Bothnian coastland. 212 Fi^Jure 203. Halfbuilt labyrinth at Svarthallviken on Bjuron. 212 Fi^Jure 204. Saint Olav statue from Enanger Church in Halsingland. 214 Fi^Jure 205. The labyrinth and the Bronze Age cairn at lavre. 215 XXIV LIST OF FIGURES List of Tables Table i. Chronological periods in Sweden and Norway. 30 Table 2. Investigated sites. 45 Table 3. Calculated rates of shoreline displacement in Vasterbotten County. 48 Table 4. Hut 67: Osteological finds. 64 Table 5. Site 138. Species by weight. 67 Table 6. Site 138. Species by fragment. 67 Table 7. Site 138, Hut A: Anatomical distribution of burned bone fragments from seals. 68 Table 8. Site 138, Hut B: Anatomical distribution of burned bone fragments. 68 Table 9. Grundskatan, Hut i: Anatomical distribution of bones. 73 Table 10. Grundskatan, Hut i: Macrofossils. 73 Table 11. Grundskatan, Hut 2: Macrofossils. 75 Table 12. Grundskatan, Hut 3: Bones (fragments). 76 Table 13. Grundskatan, Hut 3: Anatomical representation for ringed and indeterminate seals on the floor and in the hearth, burned bones (fragments). 76 Table 14. Grundskatan, Hut 4: Identified species based on numbers of fragments and weights for burned bones in hearth. 78 Table 15. Bear ( Ursus arctos) bones by weight. 81 Table 16. Bear bones by fragment. 82 Table 17. Bear bones by anatomy. 82 Table 18. Sites 31-32, Hut A: Faunal remains by weight. 95 Table 19. Sites 31-32, Hut B: Bones recovered in different areas, by weight. 96 Table 20. Sites 31-32, Hut B: Species by weight. 97 Table 21. Sites 31-32, Hut B: Species, based on numbers of identified specimens. 97 Table 22. Sites 31-32, Hut B: Species in the floor area. 97 Table 23. Sites 31-32, Hut B: Species in the hearth. 98 Table 24. Sites 31-32, Hut B: Species in the i m^ pit. 98 Table 25. Sites 31-32, Hut B: Species and skeletal elements, unburned bones. 98 Table 26. Sites 31-32, Hut B: Species and skeletal elements, charred bones. 99 Table 27. Sites 31-32, Hut B: Species and skeletal elements, burned bones. 100 Table 28. Sites 31-32, Hut B: Species and skeletal elements, unburned bones. 102 XXV Table 7 Ci Sites 31--32, Hut B: Auatomiral renreseritatinri foT seals bv MAIJ (rninirnum anatomical unit). 102 Table j^- Sites 31- j^' Hut B: Anatomical renTesentation foT seals 104 Table j'-- Sites 31- J ' Hut B: Anatomicnl representation for seals by weight. 104 Table 7 9 Sites 31- J ' Hut B: T n t; Table Sites 31- J ' Hut B: Burned bones, anatomical representation for seals by weight. Table T.A Sites 31- J ' Hut B: Unburned bones, anatomical representation for seals. 106 Table 2 CJJ- Sites 31--^2J^' Hut B: Unburned bones, anatomical representation for seals, by weight. 106 Table Sites 31--32, Hut B: r^Vi^^VTPr] linrip*^ ;^ri^itnTnirpl TpnrpQpnt^tinn For <^:p^i1.. F'fofCBSor naieritao; J , af K. VcwivBk, , (if K. Vii 1 Ortlfborn; Ltd. iif V^-i.-f- urioa.; LTmp. cl R. A... ,r » in PhiladclpLiii: Tli- \ ; N&turwtM. In Marburi-' lb, Oewh. a. Alterth ii. 'uunlc* I BeUlngfors eaait f K. NordBiJ. Ord^ Liut. af K. \Va»n Ord,, Kommend. af K. D. Danocbrogs Ord. nllqv.-Akad., 1 T'hvaloffT Kall^k. i f.iio ! ikila Vallc Tiberiii.i 't r i.oDdon membr. bi'in r I am Mayn. GOrlll/ - r pbya. Turki Nor:ad. Cn-a. Leop. Carol. I, i-i^iiion; Tbe Aeadcmy of Natural ! / BerllDi WeUer-GflseUsch. in Kj..i. I r...ui^ Slllsk. pro Fauna «t I'lora Si.i'iety of Loadnu, Coircup. fOrsta BANDET. S t en a 1 d e r n. Aiidra I'pplagaii, MID AVMABKNIliGAR Ol 11 TILLAGG, SAMT 6;TID K-VPITLET (OM BKOKSKUl.TUREN!! IIRIENTM.1 URSPUlTIGj OMAEBETADT. KliR^EDD MEU XVl.OGRAHER OCH EN MANGD LlTOGHAFIERAl'E UGLUEK. Figure 2. Front plate of Vol. I (The Stone Age) of Sven Nilsson's The Aboriginal Population of Nordic Scandinavia: An Attempt at Comparative Ethnography and a Contribution to Humankind's Evolutionary History (second edition, 1866). In the 1980s, new ideas started developing in Norway inspired by social anthropology. Knut Odner and Bjornar Olsen, influenced by the theories of Frederik Barth (1969), began character- izing Saami ethnogenesis as the result of cultural interactions, including collaboration and conflict (Odner 1983; cf Olsen 1994) - more specifically, local hunters in North Norway and Finland with metal-producing agrarian groups from central and eastern Russia (Jorgensen 1986; j0rgensen and Olsen 1988). Hansen and Olsen (2004:36-42) pushed this process of "cultural consolidation" back to the Early Metal Age/Bronze Age. 6 CHAPTER 1 Against this background, I have been struck by how much Saami and Nordic cultures had in common in northern Swe- den, and how much the Saami had contributed to Swedish cul- ture. Many Saami offer rites and ritual practices, for in- stance, mirror Germanic Iron Age practices, and vice versa. Hunting and land-use patterns are remarkably similar through- out the Stone, Bronze and Iron Ages. The northern hunter- gatherer past is even expressed in historic times by the orga- nization of farmsteads on the Bothnian coast (Roeck-Hansen 2002) and through beliefs about forest spirits (Rathje 2001:162-174). There is abundant evidence of long-term continuity in coastal Vasterbotten. Cultures can change relatively quickly but still retain cognitive structures that transcend race, economy, material culture and even religion (Goody 1977; Banton 1981; Lloyd 1990; D'Andrade 1995). The ways northern societies related to their environments, both natural and human, are therefore central to our understanding of the origins of identities (Berkes and Folke 1998; Berkes 1999; Ingold 2000). Oral histories, stories, myths and sagas are the verbal repositories of these memories. The bear rites and stories about dwarves and Stalo giants are examples of this phenom- enon. They are so-called longue duree manifestations of cultural identity (cf Braudel 1949; Thomas 1996; Redman and Kinzig 2003). Long-term continuities have also been described as the "cultural trajectories" of local societies, and these can sometimes converge with those of majority societies because ofcommon interests (Wolf 1997:23). The Saami have, nevertheless, been viewed as a dilemma and as peripheral in Nordic ar- chaeology. There have been both racial and evolutionary overtones in the discussions and, in spite of the social anthropological interpretations regarding the origins of ethnic/cultural identities, the narratives inevitably break down to minority and majority power relations (Eriksen 1993). Archaeology and the Welfare State While the methods of modern archaeology are the same everywhere, there is a fundamental dif- ference between European archaeology and the "archaeology as anthropology" paradigm that characterizes North American archaeology. Anthropological archaeology grew out of the colonial experience of Europeans in North and South America, Africa, Australia and Asia. European ar- chaeology has focused instead on national identities and has had the political goal of asserting the origins and histories of those societies. Social interpretations, particularly since the 1970s, have Figure ]. Plate VIII showing artifactsfrom Sweden, Ireland, Mexico, Greenland and Pennsylvania together with an Eskimo igloo (Sven Nilsson 1866). INTRODUCTION AND NARRATIVE CONTEXT 7 consequently often been based on European political, economic and sociological theory (cf. Gid- dens 1977) and especially Neo-Marxism, including World Systems Theory (Wallerstein 1974) and the French Annales School of Historiography (cf Braudel 1949; Bourdieu 1977). Although I have found these theories very useful in my own research, they have also led many Nordic scholars, in my opinion, to an overemphasis on the explanatory importance of conflicts and crises, a perspective that still dominates interpretations of northern prehistory. Curiously, although northern Sweden has clearly been subjected to internal colonialism (cf Loeffler 2005), this has manifested itself politically as the victimization of the Swedish settlers, hydroelectric-, forest- and mine workers by the state, and who resent the outflow of capital to Stockholm more than the plight of the Saami. But, even in its most benign forms, nationalistic archaeology has rendered the rights of mi- nority societies problematic. This was complicated in Sweden by the development of the welfare state system in the 1930s. Racial hygiene (eugenics) and education were among the core principles of this social engineering effort. Sweden had already been the first country in Europe to establish a State Institute of Racial Biology [1921]. The Sterilization Act, implemented in 1941 and affecting some 63,000 Swedish citizens, mostly women, was only discontinued in 1975, 30 years after WWII (Broberg and Roll-Hansen 2005; Broberg and Tyden 2005:77-149). As early as 1913, special nomadic schools were established for Saami children to help preserve the nomadic lifestyle and perpetuate the stereotype of the Saami as primitive, vulnerable and con- sequently needing the protection of the state (Lundmark 1998, 2002:40-41). It was believed that the Saami were even physically predisposed for reindeer nomadism. These ideas have been referred to as "a Lapp shall remain a Lapp" policy (Lundmark 2002:63-75). The Germans, while occupying Norway from 1940, seem to have shared this patronizing view, allowing the Saami, for example, to continue with trans-border herding (Lantto 2005). These social goals have had, and still have, major political, economic and cultural consequences. "Real" (nomadic reindeer herding) Saami were put under state protection and Saami land was, and still is, state-owned. The remaining 90% of the Saami in Sweden, the non-reindeer-owners, have been viewed as assimilated (cf Morkenstam 1999). Ironically, in spite of the protectionist ideology for the herders and international solidarity regarding human rights and the support of archaeology in developing countries through the Swed- ish International Development Agency, the Swedish government still refuses to acknowledge the Saami, including the herders, as indigenous people in accordance with United Nation policies (ILO nr. 169, 1989). Convention 169 comprises the principles, guidelines and obligations for the protec- tion of indigenous peoples, including their institutions, properties, lands and culture. Indigenous people are defined under these provisions as people having their origins in ethnic groups that lived in the country when national borders were formed, and who have wholly or partly retained their social, economic, cultural and political organizations. The situation has been exacerbated by an unreflective projection of majority culture in Nordic museums (cf Goodnow and Akman 2008). In her study regarding Saami representations in Swed- ish and Finnish museums, Janet E. Levy found while Nordic continuities with the Viking past are still presented as an almost foregone conclusion, there is at the same time an earnest desire to "de- politicize" archaeology and to not discuss ethnicity (Levy 2006). Unfortunately, this non-committal position de facto relegates the question of Saami origins to that of a prehistoric wilderness, almost the same attitude toward these people and their "unoccupied" lands as expressed by the Swedish Crown in 1328 (cf Steckzen 1964:119-128). 8 CHAPTER 1 Intellectual polarity can go both ways but the majority (master) narrative always has a distinct advantage on the national stage (cf. Zachrisson 2004). This study aims to create a more balanced picture of prehistory. While it should be obvious that one cannot take identities ofthe present and un- critically apply them to the past, it is equally obvious that these identities have pasts and we have an obligation to understand their origins, particularly when these affect policies today (cf Ojala 2006). The Means to an End This book is intended to be accessible to a broad professional readership and an international pub- lic who have an interest in archaeology, climate change, Nordic prehistory and history, cultural resiliency and northern indigenous issues. One immediate goal has been to publish the original archaeological data from the project, but this is also very much a book of ideas. I have long argued that the Nordic region is one of the best places in the world to study long-term cultural interactions, not only because of the meeting of two great cultural-ecological systems, but because of the socio- economic complexities not often seen this far north (Broadbent 2000). This is also a highly interdisciplinary study. Like the issue of Saami ethnogenesis itself, it is often at the "borderlands" that the greatest dynamics of systems occur, be they cultural and/or ecological, and where the causes and effects of change are most easily observed and understood (Broadbent 1997). Archaeological data are always fragmentary and to overcome these limitations I have turned to the ecology of lichens, geophysics, chemical analyses, organization of northern church towns, place-name distributions, tax records and so forth in order to fill in the many blanks. From this web of information emerges a picture of the past much like a photograph in a developing tray. This picture is the subject of the book. The narrative is the interpretation of this picture and its origins in 7,000 years of prehistory (cf. Carlsson 1998). Historian Eric R. Wolf has best described this search for a past: We can no longer be content writing the histoiy of the victorious elites, or with documenting the subjugation of dominated ethnic groups . . . we thus need to uncover the "people without his- tory . . ." (iggy.xvi) Of course, there are no people without histories of their own, and I trust I have moved us closer toward illuminating the Saami past through this interdisciplinary archaeological effort. The last chapter is a synthesis of the conclusions of the study and offers some perspectives on what we can learn from prehistory about resiliency and the value of cultural diversity in contemporary society. Notes 1. The term circumpolar refers to the northern lands encompassing the Arctic Ocean and hav- ing Arctic and Sub-Arctic climates, plants, animals and cultures. 2. A large iron ring with the oldest Nordic law code written in runes was found spiked to the door of Forsa Church in Halsingland, northern Sweden. It was probably taken from a pre-Christian cult site or temple in the region ("Ringen fran Forsa." Catalogue edited by Ian Lundell and Lars Nylander, Halsinglands museum-Hudiksvall 2005). INTRODUCTION AND NARRATIVE CONTEXT 9 Figure 4. The Circumpolar Region. The Nordic region is the northwestern corner ofthe Eurasian continent and a long- term meeting-ground ofcircumpolar and European plants, animals and peoples. The capital cities of Oslo, Stockholm and Helsinki coincide with the Goth parallel and the ecological boundary between these two worlds. Culture and Ecology The Historical-Ecological Setting Scandinavia includes the kingdoms of Norway, Sweden and Denmark. The Nordic countries by definition also include Finland, Iceland, Greenland and the Faroe Islands in the North Atlantic. The Nordic region is a meeting ground of two great human-environmental systems: the North Eu- ropean and the circumpolar worlds. The borderland between these two biogeographical regions in Sweden is called Limes noniandicus, the Norrland (North Land) border, and coincides roughly with the Dal River (Dalalven) and the Goth parallel, which runs just north of the capital cities of Oslo and Stockholm and south of Helsinki (Figure 4). Human history in Sweden and Norway plays out against the background of major envi- ronmental changes on land and in the seas within the confines of a ca. 400 km wide peninsula between the North Atlantic and the Baltic Sea. The northern boreal forest (taiga) extends eastward from Norway and Eurasia to the Bering Straits and across Alaska to Canada. The high latitudes and environments make the Nordic North an indisputable part of the circumpolar world. Many indigenous peoples, such as the Saami, developed cultural adaptations and survival strategies that are unique to these northern forest, tundra and maritime environments (Gjessing 1944). The Importance of the Coastal Zone The significance of the Bothnian coast in Saami prehistory has been little discussed in Sweden, even though the idea is not new. Birger Steckzen (1964), an historian and archivist, speculated that the disappearance of the Saami from this coastland was due to brutal taxation by the Birkarls, state- sanctioned tax collectors operating in the fourteenth century. He was well aware of the Iron Age hut sites, but none of this material had been excavated. Gustaf Hallstrom, who had mapped many of the coastal sites in Vasterbotten, had commented on their similarities to the so-called "Stalo huts" in the sub-alpine regions of Sweden (1949:76). In 1965, Skelleftea Museum director Ernst Wester- lund rejected the idea of coastal Saami in Vasterbotten, rebutting Steckzen (1964), and ascribing the sites to seasonal sealers and fishermen from the south, or even to Swedish settlers (Westerlund 1965). The idea of a "local" culture has also been argued (cf Rathje 2001). Westerlund's view has been a commonly held opinion (cf Westin 1962; Norman 1993; Lindstrom and Olofsson 1993), and there has been little interest, and even an aversion, by any but a handful of Swedish archaeologists in arguing otherwise. 11 In other respects, the maritime connection has de- fined Nordic prehistory more than any other single environ- mental or geographic factor. The Nordic Peninsula is cir- cumscribed by the North Atlan- tic and Baltic seas, marked by large lake systems and cross-cut by rivers and river valleys. Prox- imity to water routes was one of the key elements conveying economic advantages in Eurasia and the European Peninsula as a World System (Wolf 1997:31). The Norwegian and the Both- nian coasts were major cultural meeting grounds and coastal waters were the best way to travel to and from the north in both countries; traveling along the shore of the Swedish coast rather than by land made it 11. -J . . Figure ^. Main river drainages of Sweden into the Gulfof Bothnia.possible to avoid 12 ma]or river & j & j j j crossings (Figure 5). In addition to the direct effects on Nordic climate, marine food resources laid the founda- tions for larger hunter-gatherer populations and more settled lives that opened the door to cultiva- tion and animal husbandry at these latitudes. Coastal zones were also critically important because of the reliability and accessibility of their resources, including birds, shellfish, fish and fuel that all the members of these communities, including the elderly and the young, could gather on a daily basis. One of the most fascinating aspects of Bothnian economy is that it involved high arctic ice hunting (cf Gustafsson 1971, 1988, 1990; Westerberg 1988; Kvist 1987, 1988, 1990; Nystrom 1988, 2000; Nilsson 1989; Edlund 1989; Olsson 1990). Seal hunting also has a long history in the region (Ekman 1910:222-260; Hamalainen 1930; Clark 1946; Forsten 1972; Broadbent 1979; Stora and Lougas 2005), and there is a rich Swedish vocabulary referring to seals and sea ice (Edlund 1989, 2000). Ice hunting continued into historic times, and for three months of the year a large percent- age of the male population of "Norrabotten" (the Mare Botnicum or Bay of Bothnia) was recorded as going out on the ice. As many as 15,000 seals were taken in one year in the sixteenth century (Westin 1962; Tegengren 1965; Kvist 1991). Olaus Magnus, the last Catholic Archbishop of Sweden, illustrated the organized mercantile hunting of seals by northern peasants on his Carta Marina map from 1539 (Figure 7). 12 CHAPTER 2 Figure 6. (left) Provinces of Sweden, (right) Counties of Sweden. The county system (Idn) was established in 16]^ and superseded the older provincial system of administration. The county system is used with reference to archaeological sites, place-names and other data presented in this study. The Saami People The Saami are the people of Sdpmi (Saamiland). This name Saami (alt. Sami) is recorded from the thirteenth century but is probably much older (Zachrisson I997a:i74). The Saami were also called, sometimes derogatively, "Lapps." The name "Saami" occurs in all Saami languages and, whenever possible, this self-designation will be used in this text, although the term "Lapp" is employed con- cerning administrative areas, place-names or historical documents. The Saami number approximately 80,000 people today. About 17,000 Saami live in Sweden, of whom 2,000 are involved in reindeer herding. Their ancestors were referred to in the oldest written sources as the Fenni, and as Skrid-finnar (Ski-Finns). In Norway, Lapland is still referred to as "Finnmark." In 1673 Johannes Schefferus speculated that the Saami had even come from Finland: The Lapps derive without doubt their originfrom the Finns, were horn among them, hut had been driven away and exiled from Finland . . . after leaving their homeland they withdrew as exiles and this is the origin of the name. They did not themselves use this name [Lapp] . . . but that of theirformer countrymen, the Finns. (iGj^^o) The Roman historian Tacitus had described the Fenni as early as A.D. 98: CULTURE AND ECOLOGY 13 Figure 7. Carta Marina map by Olaus Magnusfrom 3539. This portion ofthe map shows the North Bothnian coast, hunters spearing gray seals and sleds pulled by reindeer and horses. A Saami woman north ofthe Ume River in Vdsterhotten is milking a reindeer. Original in the Carolina Rediviva Library in Uppsala. Hie Fenni are astonishingly wild and horribly poor. They have no arms [weapons], no horses, no homes. They eat grass, dress in skins and sleep on the ground. Their only hope is in their arrows, which, for lack of iron, they tip with bone. The same hunt provides food for men and women alike; for the women go everywhere with the men and claim a share in securing the prey. The only way they can protect their babies against wild beasts orfoul weather is to hide them under a makeshift network of branches. This is the hovel to which young men come back; this is where 14 CHAPTER 2 the old must die. And yet they count their lot happier than that of others who groan over field labor, sweat over house building, or hazard their own or other men's fortunes in the wild lottery of hope and fear. They care for nobody, man or god, and have gained the ultimate release: they have nothing to prayfor. What comes after them is the stuffoffables—Hellusii and Oxiones with thefaces and features of men, but the bodies and limbs ofanimals. On such unverifiable stories I will express no opinion. P. Cornelius Tacitus, Germania Translated by H. Mattingly (1948:140) This account paints a dismal picture and one should view such descriptions with skepticism (Whitaker 1980). In fact, the characterization of hunter-gatherers or nomadic people as wretched is quite common (Sahlins 1972; Cribb 1991). The expression fattig lapp "poor Lapp" still has this col- loquial meaning in Scandinavia. The ethnonym Lapp has an east Nordic origin and probably derives from the Finnish Lappi (Lappalainen means "Laplander") - or may even derive from the word for a patch of cloth or hide (lapp). In Russia, the name lop and lops is known from chronicles dating to ca. A.D. 1000 (Hansen and Olsen 2004:49-50). According to von Diiben (1873:376), the term also has reference to magic and witchcraft. The ethnonym Lapp appears in the Gesta Danorum (History of Denmark), written by Saxo Grammaticus in A.D. 1190 (Schefferus 1673:42), and also in the Icelandic Orkneyinga Saga, writ- ten down as a revised text in the 1200s (Zachrisson i997a:i59). The place-name Lappi has been associated with Saami settlements at more than 575 places all over Finland (Itkonen 1951:33). The term lappar was first used in an official Swedish document in the Talje statutes from September 5, 1328. By my count, there are 1,147 place-names with the prefix Lapp in Sweden, based on the National Land Survey place-name registry (Chapter 9, Figure 195). These place-names have thus far been given little attention in Sweden but are invaluable sources regarding former Saami territory. The Saami speak nine languages ofthe Finno-Urgric (Uralic) language family (CoUinder 1953; Nickel 1990) (Figure 9), and their mtDNA and Y chromosomes indicate that they are an ancient European population (Beckman 1996: Tambets et al. 2004). The Finno-Ugrian languages probably spread from central Russia and the southern Urals to the Nordic region as early as the Stone Age, and Figure 8. Saami summer camp in North Norway. Photo: Fred Ivar Utsi Klemetsen. CULTURE AND ECOLOGY 15 the Finnish language developed later through contacts with other Finno-Ugrian speakers from the south Baltic. This linguistic dif- ferentiation is believed to have occurred during the Bronze Age/ Arctic Bronze Age, some 4,000 years ago (Strade 1997:183; Car- pelan et al. 2001). However, on the basis of "Paleo-Lappish sub- strate" words, such as the word for seal (morsa), Aikio (2004) ar- gues for a later date. Saami territory in Norway and the Kola Peninsula in Rus- sia extends along a thousand kilometers of northern coasts (Figure 9). According to the Icelandic and Norwegian Sagas from 1100-1200S, the Saami also lived as far south as Hadeland, some 20 km north of Oslo, and throughout southern Norrland and as far south as Svealand in Sweden (Zachrisson 1987:26, i997a:i7i). The latter is even implied by mtDNA at the iconic Late Iron Age boat burial site of Tuna in Alsike between Uppsala and Stock- holm (Gotherstrom 2001:25-26). The ancestors of the Saami were also known as far south as the Western Dvina (Daugava) River in Latvia (Itkonen 1947; Eidlitz-Kuoljok 1991:32) (Figure 29). In 1911, Johan Turi, a Swedish Saami reindeer herder and author ofone of the first indigenous ethnographies ever written, Muittalus Samid Birra (A Book on Lappish Life), wrote the following regarding Saami origins and settlement: Figure g. Present Saami territory and language areas. Asterisk shows the main area ofstudy. Map adaptedfrom Nickel (iggo). It has not been said that the Lapp camefrom somewhere else. The Lapp was settled all over Lap- land and the Lapp lived on the seacoast and there were no other dwellers besides themselves, and that was a good timefor the Lapps. And the Lapp also lived everywhere on the Swedish side and there were no settlers anywhere; the Lapps did not know there were other people besides themselves. {1911:2} This view is reiterated by most Saami scholars today (cf Kuoljok 1996; Haetta 2002; Solbakk 2004; Lehtola 2004), although Sdpmi in Sweden and Finland is invariably shown as coinciding with inland herding areas. The Germanic speaking groups of Sweden and Norway were referred to as Svear and Nor- dmenn in sources such as Adam of Bremen's History of the Archbishops of Hamburg- Bremen (ca. A.D. 1081). The Svear settled only as far north as 63°N on the Bothnian coast and as far as Jamtland and Harjedalen in the interior. The Nordmenn occupied the coast of Norway up to Malangen at about 70°N (Odner 1983). Orosius (Othere from Halogaland), who lived in North Norway in the late ninth century, described the Fenni, Terfenni and the Bjarmi - all Finno-Ugrian speakers - in his account to the English King Alfred the Great (Bosworth 1885). In the Historia Norvegiae (A.D. 1195), 16 CHAPTER 2 mention is made of another group, the Kveni, who hved on tlie north Bothnian shores. Egil's Saga, written down in the 1200s, also mentions Kvenland (Paulsson and Edwards 1976). Adam of Bre- men referred to these Kveni as "Amazons," but the designation was probably his confusion over kvenerlandet (the land of the Kvens) with kvinderlandet, which means "the land of the women" in Danish (Lund 1978:68). It has, nevertheless, been speculated that the "Amazon" description related to large-scale sealing expeditions that left Bothnian farming settlements populated by only women and children for three to four months of each year (Tegengren 1965). This curious tale originated among travelers going north who had actually encountered these villages ofwomen (according to Adam). Adam had also written that the people north of the Svear were ruled by a woman. Lillian Rathje (2001) has emphasized the important role that women had in northern agrarian economies while the men were off hunting and fishing. The Ski-Finns are identified as living south and west of the Amazons and north of the Svear (Figure 10). This description is fairly accurate, and Kvenland could actually thus refer to Finland (lulku 1986). Amazons notwithstanding, there is little doubt that seal hunting was ofconsiderable economic I ,\KTi rF.ST NORTH OCEAN OF DARKNESS AND ICE £LEiaillO£RNCA I GREENLAND THE NORTHERN OCEAN ^ f^^T—\ M-i'-iwi^G6TLAMD *S^ -tiAMAZONS^A-LANER -r^-^ \ LAMER to V-J X "SACHSEN "VM.«a.i..rt KURLAND BALTIC SEA P MME.RANEftE SKIJTEP. TURKEa o^ard RUSSIA POLAND NULER. ""TjTIVsWTRAVEL' Figure 10. Map based on a description ofthe Nordic region by Adam of Bremen (ca. A.D. 1081). The Saami are referred to as Skridfinner (Ski-Finns) and Hved in Halsingland. Tlie "Amazons" lived on the Baltic coast ofwhat was probably Finland. Even Vinland is also mentioned in this early account. Map adaptedfivm Lund (igyS). CULTURE AND ECOLOGY 17 importance from the earliest times and into the modern era. Swedish Bothnian seahng sites from the Iron Age have been previously discussed by a number of archaeologists and historians: Hallstrom (1942, 1949). Varenius (1964, 1978), Westin (1962), Steckzen (1964), Westberg (1964), Westerlund (1965), Nilsson (1989), Norman (1993), and Rathje (2001). Four cultural entities have been considered: the Kvens, the Finns, the Saami (Lapps) and Germanic Iron Age settlers (Scandinavian speakers). Othere from Halogaland (ca. A.D. 900), Adam of Bremen (ca. A.D. 1081), the Historia Nor- wegiae (ca. A.D. 1195) and Egil's Saga (ca. A.D. 1200) state that the North Bothnian groups were Finno-Ugric, not Germanic peoples. Adam wrote that Halsingland was Saami territory. During those times Halsingland extended around the Bay of Bothnia and down to the Ule River in Finland. The northern limit of Germanic settlement in Sweden extended only as far as Arnas in Anger- manland (Figure 12). This is the northernmost limit for Iron Age forts, rune stones, tumulus cemeter- ies, Nordic long houses and the kdung maritime defense system (Westerdahl 1989). Diagnostic Late Iron Age settlement place-names, such as those with the name elements -sta, -hem and vin-, are not found north ofthere either (Edlund 1988). Nordic linguists have given considerable attention to these settlement names (Franzen 1939; Holm 1949; Stahl 1976; Wahlberg 2003). The name hem means "home" today, but originally referred to a village settlement and was often a first, or landnam, name. The term landnam is associated with the first landings by the Norse in Iceland, and this coloniza- tion phase usually leaves a distinctive environmental signature on vegetation surrounding northern settlements because of grazing, burning, removal of trees and the introduction of cultivated plants and weeds (cf Hicks 1988, 1994; Aronsson 1991; Karlsson 2006). Vin probably referred to grazing land and also indicates that this was a homestead site. Sta originally referred to boat-landing places but was later applied to towns that had developed because there were harbors at these places [stad means "town" today). All three names date back to the period A.D. 800 to iioo in Sweden. These names concentrate in Jamtland, Medelpad and Angermanland in northern Sweden (Selinge 1997). The name elements hole and mark are, by contrast, common in northern Vasterbotten and are associated with the expansion of Scandinavian farmers starting in the fourteenth century (Holm 1949:94). In an analysis of the Kven issue regarding potential recognition under ILO 169 (the Interna- tional Labour Organization's 1989 Indigenous and Tribal Peoples Convention), Swedish historian Lennart Lundmark came to the conclusion that the Kvens were not an indigenous cultural group at all, but were Finnish middlemen for the Novgorod fur trade (Lundmark 2007). They were later formally recognized by the Swedish state as Saami tax collectors, the Birkarls. Thomas Wallerstrom, a Swedish historian and archaeologist, had earlier reached the same conclusion (1995). Assuming they are right, and I think they are, the Kvens can be discounted as having anything to do with the Iron Age sealing sites. The Kvens seem to have had their operations and farms in the northernmost parts ofthe Bay of Bothnia and on the Swedish coast only as far south as the Skellefte-Byske Rivers, an area with many Finnish connections (Fjellstrom 1988). Having eliminated the Kvens from serious consideration, the remaining alternative for these seal hunters is that of an indigenous, but non-Germanic culture. One complication in all of this is that these two cultural-linguistic entities, the Finno-Ugric and the Germanic, with languages that are as far apart as any on earth, existed side by side for thou- sands ofyears. This situation is nothing like North America, where the sudden impacts of European colonialism starting 500 years ago transformed whole continents, and whose diseases, animals and plants almost annihilated indigenous populations (Crosby 1986). In the Nordic region there was, on the contrary, continuous interaction and a slow process of give-and-take. There is considerable 18 CHAPTER 2 evidence that there were mostly peaceful and symbiotic relations between Germanic, Finnish and Saami groups (Mundal 1996; Zachrisson i997a:22i-234; Olsen 2000; Price 2000). Cultures with such a long history of interaction inevitably overlap in numerous ways. It is also likely that these pre-literate people were multilingual. This implies that instead of sharp boundaries between them we are in reality dealing with overlapping cultural and linguistic topographies. Although people certainly have identities independently of outsiders, this meeting of cultures was itself a potent force of ethnogenesis (Barth 1969, 1994), and a key to the formation of many distinctive Saami ethnic markers as seen by dress, rituals and other practices (cf. Kleppe 1977; Reymert 1980; Baudou 1987; Zachrisson 1987b; Zachrisson i997a:i89-22o). Metal artifacts, such as brooches, were produced in many different regions and were readily incorporated into Saami dress. Christian crosses and symbols, as well as decorative styles (e.g., Nordic plaited band designs), were also used by the Saami. The concept ofcultural clines, comparable to population studies, and "seamless cultural topog- raphies" (Caulkins 2001:121), helps conceptualize Saami prehistory as an integrated part of Nordic prehistory. This not only de-marginalizes the narrative, it connects the discourse to the formation of all northern cultural identities. The Early Metal Age, which corresponds to the Scandinavian Bronze Age and the so-called 'Arctic Bronze Age" (Tallgren 1949; Bakka 1976), was the period during which Saami ethnicity (as we know it) is believed by many archaeologists to have developed through contacts between "Proto-Saami" and other groups (cf Carpelan 1975b; Baudou 1987; Olsen 1994). Human-Environmental Relations The Nordic region is situated in the far northwestern corner of the Eurasian continent and stretches some 2,000 km between latitudes 54°N to 72°N, as far north as the North Slope of Alaska, Baffin Island in Canada and northern Siberia. This shoulder of land juts far out into the North Atlantic where it intercepts the warm and salty waters of the Gulf Stream. The climate-mitigating influences of the Gulf Stream have had profound effects on ecology and settlement. Agrarian economies were made possible farther north and earlier than anywhere else in the Arctic and the rich Sub-Arctic waters have been among the most productive fishing regions of the world (Dunbar 1954, 1968). Seals, salmon, whitefish, cod, herring and shellfish have drawn people to these shores from the earliest times, and most Nordic populations still live within 20 km of the coast. The Scandinavian Peninsula is furthermore separated from Finland for most of its length by the Baltic Sea and the Gulf of Bothnia. This body of water cuts deeply into the continent and extends about 1,400 km from north to south. It separates the two regions by 100 km or more of open water today, but had once been twice as wide. The Baltic has changed from being an ice-dammed freshwater lake 10,000 years ago, to an open arm of the sea, back to a freshwater lake as a result of the many rivers that flow into it, only to once again become a sea (Berglund et al. 1994). Today the Baltic is an intermediate brackish body of water. The most productive fishing-, hunting- and shellfish-collecting periods were when the Baltic was most open to the Atlantic Ocean, and this occurred during the Litorina Period (named after a saltwater mollusk) between approximately 5000 B.C. and 1000 B.C. Even today, powerful strokes of saline water into the Baltic have had immediate and positive effects on plankton, shellfish and seal populations lasting for up to five years (cf Segerstrale 1957:757). Sea level rises due to worldwide oceanic events would have had even longer-term effects. CULTURE AND ECOLOGY 19 Figure n. A visual cross section of north Swedish biogeography. Clockwisefrom top left: alpine birch forest and mountains, headwaters of the Skellefie River interior pine heath forest (with trapping pitsfor moose or reindeer), ice cover on the Bay of Bothnia in April, the Bay of Bothnia with wave-washed moraine beaches, interior lake and esker landscape. Most of the vegetation above 6o°N in Sweden is classified as northern boreal forest with a sub-alpine birch region in the mountain foothills to the west. The boreal forest environment brought characteristic animals such as bears, moose, beavers, lynxes, wolverines, pine martens, otters, black grouses and hazel hens. During the warm Atlantic climate period, ca. 5000-2500 B.C., there were more deciduous trees such as birch, alder and even elm, and this growth greatly 20 CHAPTER 2 stimulated beaver and moose populations, who fed on their leaves and bark. These are the most commonly found animal bones on archaeological sites in the interior (Ekman and Iregren 1984). The pine forests had changed into a mixture of pine and spruce forests by around 1000 B.C. (Engel- mark 1976; Berglund et al. 1994). Black spruce (Picea abies) is an eastern hybrid and grows on wet- ter ground at the expense of alder and birch. Colder and wetter conditions during the Sub-Atlantic period, starting around 2,400 years ago, led to an expansion of reindeer into the southern forests. Bogs increased and covered much greater land areas than earlier. Although moose, beaver, reindeer and seals were important food resources, freshwater fish were also of great economic value. Salmon, perch, whitefish, pike, burbot (a freshwater cod), trout and char were fished in the rivers and lakes. Starting in the thirteenth century, the Catholic Church asserted ownership of northern salmon rivers. Taxes and salaries were paid using fish, and most parish churches and homesteads in Lapland were established by fish-rich lakes (Campbell 1948). Medieval (Hanseatic) herring fisheries became a basis for many Bothnian coastal villages and towns. The Gulf of Bothnia is a virtual heat reservoir in the fall that pulls temperature gradients northward into lines parallel with the shore- lines. But, unlike the North Atlantic coast, it freezes over in winter and creates a negative temperature anom- aly in the spring and summer. The Bothnian coast is ice-bound for up to 200 days per year north of 63°N, which also coincides with a change from a mountainous to a flat coastal topography in Sweden (cf. Angstrom 1968; Helmfrid 1994). This was, not surprisingly, the northern limit for Germanic (Iron Age) agrarian settle- ment in Sweden, and therefore an im- portant cultural-ecological boundary (Figure 12). Although during warmer periods these ice boundaries could have varied, the coastal topography kept the ice locked into the Bay of Bothnia because of the narrow bottle- neck of the Kvarken between Sweden and Finland. The 100-day ice ex- tended down to the Hornsland Pen- insula in Gavleborg County, which is the most prominent peninsula south Figure 12. Map showing the oak limit, alpine and Bothnian ice regions and Bjuroklubb. CoaStal topography the northern Hmit of Germanic Iron Age setttlement. waS also leSS broken and with fewer CULTURE AND ECOLOGY 21 7000 6000 5000 4000 3000 2000 1000 Norrbotten Vasterbotten Vasternorrland Gauleborg Uppsala islands south of there. Hornslandsudde was therefore an ideally suited place to investigate the southern margins of Saami sealing. Hunting statistics from 1902 to 1906 (Ekman 1910:262) show the re- lationship between numbers of seals killed as well as numbers of seals killed per 10 km of Bothnian coast (Figure 13). These figures reflect the relationship between ice conditions and topography. The more broken and island-rich the coastline, the more stable the coastal ice which, together with snow depth, conferred an advantage for ice-breeding seals and the survival of immature seals that could drown if the ice melted before they could swim. Vasterbotten was a core area for sealing, and Bjuron, with many Iron Age sites, projects far out into the Bay of Bothnia where it intercepts the counter-clockwise gyre that opens up as a drift-ice channel close to shore. Specialized boat-based seal hunting expeditions from Oster- botten in Finland, referred to as the Falan (Cnieff 1757; Gustafsson 1971), sometimes put ashore when the channel was blocked. Many sealers carved their names on the rocks of Svartallviken on Bjuron (Chapter 4, Figure 44). Norrbotten Vasterbotten Vasternorrland Gavleborg Uppsala Figure i?. (top) Numbers ofseals taken by county, north to south, (bottom) Numbers ofseals taken by length ofcoastline (based on Ekman 1^10:262). Glacial Topography and Settlement Nordic topography has been sculpted and planed down by the advances and retreats of glacial ice. Sweden and Finland consequently have low rolling landscapes, and Norway, which has a more mountainous topography and softer bedrock, deeply cut fjords. Northern Sweden is also transected by numerous northwest to southeast running rivers, and Finland is speckled by lakes dammed up by a huge terminal moraine that runs north of Helsinki. Crossing the land in the direction of glacial retreat are long alignments of moraine deposits called "eskers" that mark the courses of former rivers in the ice. Eskers formed natural travel routes for both animals and humans across the lakes and low and boggy expanses of the interior, and are natural connectors between the interior and the coastlands. Most archaeological sites in northern Sweden are found on or near them. Elongated teardrop glacial deposits, called drumlins from the Irish "little hill," formed settlement islands in interior lake and lowland areas, as well as coastal hills and off-shore islands where sealing and fishing could be based. 22 CHAPTER 2 The Nordic region was de-glaciated starting about 15,000 years ago and by 8,000 years ago northern interior Sweden was completely ice free (Bergstrom 1968; Bergman et al. 2004). This was one of the last places in Europe to be colonized by humans. The process of glacial melting was so rapid that vast lowland regions of the North Euro- pean plain were inundated (Bjork 1995; Christian- sen 1995; Gornitz 2009b). It is likely that many of the western and southern pioneer colonizers of Scandinavia were displaced hunters and gath- erers from what is now the bottom of the North Sea between England, the Shetland Islands and the Netherlands (cf. Fischer 1995; Knutsson 2005). This flooding, and related sudden events, seems to have peaked around 6500-6000 B.C. (cf Bjork et al. 1996; Lamb 1977, 1995) and marks a major push into northern Sweden from the south, as seen by the introduction of micro-blade technology (cf Olofsson 1995, 2003). As the North European hunters and the animals they hunted moved into Scandinavia from the southwest, other human groups and animals moved in from the southeast and east, exploring the newly opened lands before them (cf Nunez 1997; Carpelan 2001; Bergman et al. 2004; Knutsson 2005). The mtDNA of every- thing from voles to bears shows these two groups met up about halfway down in Sweden, or at about latitude 63°N (Jaarola et al. 1999) (Figure 14). Saami mtDNA and Y chromosomes extend southward in an analogous distribution (Tambets et al. 2004). The crust ofthe earth was not only flooded, it was depressed by the weight ofmore than a ki- lometer of ice and is still compensating following the release of that burden (called isostatic uplift). Land uplift is nearly a meter per century today in Vasterbotten (Bergqvist 1977), one of the highest rates in the world, and was perhaps as much as 10 m per year immediately following deglaciation. Shoreline displacement thus provides a unique opportunity for investigating this prehistory (Chap- ter 4, Figure 34). Climate, Ice and Seal Oil Nordic climate is classified in "Group D" - continental/microthermal - on the Koppen climate classification system, with local maritime influences (Angstrom 1968). In winter the region is positioned between high pressure areas in Siberia and Atlantic low pressure centers, resulting in warm southwesterly winds. Inland lakes are frozen from about the beginning of November until Figure 14. Map showing tht mtDNA of voles (and viruses) in Sweden (Jaarola et al. 3999 j. The migration of terrestrial animals into the region following deglaciation is a good analoguefor human colonization. Tlxe cultures, plants and animals ofthe Nordic North have always beeti a mixture ofthese components, onefrom the south and one from the east, one European and the other circumpolar. CULTURE AND ECOLOGY 23 the middle of May. The Bay of Bothnia freezes over from the middle of November and usually breaks up by May 20. Snow cover averages 160 to 180 days per year. Precipitation on the outer coast is quite low, but just a few kilometers inland increases to 500 mm per year. During June there is an average of 350 hours of clear sunshine, and in December there are only about 18 hours of sunlight. The average growing season, which corresponds to 3°C, lasts only 160-170 days per year. The total amount of solar energy at 6o°N is only a small percentage of what it is farther south, and survival requires taking advantage offood resources during very short growing and breeding seasons. Farm- ers must keep their livestock alive for six months or more without access to pasturage, and hunters manage on scarce resources during a very long winter and a late spring. Marine mammal oil (seals in the Baltic) was the main source of lighting for circumpolar hunter-gatherers, and seal oil was one of the most important commodities across the Nordic region long before whale oil became commercially important in Europe. Periodicity and Settlement Cycles As defined thus far, three environmental perspectives provide focus in this study: the long-term ecological perspective, the coastal or maritime perspective and the regional perspective. Because of the seasonal fluctuations of both terrestrial and marine resources, northern hunting peoples had to depend on combinations of ecosystems for survival. Depending on geography, this could be done through annual rounds or through sustained uses of particular regions. Economic efforts in northern Sweden, as in most regions of the world, were concentrated during times of maximum productivity, including spawning, breeding and growing seasons (cf Ekman 1910:455; Campbell 1948:81-211). Hunter-gatherer/fisher societies were also dependent on longer-term resource fluc- tuations. It is now generally accepted, however, that far from always living in perfect harmony with the environment even these societies could over-exploit and even wipe out game, particularly when engaged in extensive trade. Over-exploitation is considered by some historians as having been a fac- tor in the transition to nomadic reindeer herding and farming, and over-grazing is still a problem in the interior (cf Hultblad 1968; Lundmark 1982). Long-term human-environmental dependencies can be analyzed at the landscape level using the concept of historical ecology (Crumley 1994); this relates large-scale environmental changes to the level of human interaction and impacts on local ecosystems. Interactions on much larger scales have been discussed by archaeologists through World Systems Theory, which encompasses global climate change (Kristiansen 1994; Chase-Dunn and Hall 1997; Kardulius 1999; Hornberg and Crumley 2007). World Systems and Northern Ecology The Circumpolar Region has not generally been discussed as a World System, and yet it is one of the great cultural and ecological regions in which vast trading networks, as well as even more integrated economic systems, were operating. One of the first great world industries was the Rus- sian fur trade (Wolf 1997:158-194; Crowell 1997). This is also a region in which early bronze and iron metallurgy developed (Chernykh 1992; Hjarthner-Holdar 1993; Khlobystin 2005; Kuz'minykh 2006). Trade had always played a part in Nordic prehistory but was operating on a Eurasian scale from at least the Bronze Age/Early Metal Age and reached a culmination during the Viking Period around A.D. 1000 (cf. Fedorova 2002). World Systems Theory emphasizes the redistribution of 24 CHAPTER 2 resources, such as furs for metal, between a periphery and a core. The "exporting" and the "receiv- ing" societies underwent recurring periods of expansion and regression due to environmental as well as economic cycles (Wallerstein 1974; Modelski and Thompson 1999). Pulses of these kinds are particularly common in arctic regions, especially when there was heavy exploitation of resources through over-hunting, or when agriculture had been carried out at the limits of its sustainability (cf Martin 1982; McGovern 1988). The brackish Gulf of Bothnia is a relatively species-poor marine system, although influxes of ocean water through global warming and worldwide sea level rises probably dramatically raised productivity. These warm periods are indirectly measurable through shoreline transgressions (flooding) in the south and are seen as increased beach erosion in the north. But they are also seen through settlement accumulations at these levels, hunters taking advantage of increases in seal and fish populations, and by agrarians expanding their settlements northward. These fluctuations are very pronounced on the Vasterbotten coast (Chapter 3, Figure 16). It is well known that Bothnian ice conditions were critical for the hunting of ice-breeding seals, particularly the ringed seal (Ekman 1910; Gustafsson 1990). Hunting statistics (Svensson 1904; Hult 1943; Soderberg 1974a, b) and tax lists from the sixteenth century (Kvist 1988, 1990, 1991), suggest that heavy ice conditions led to fewer hunting bounties being paid out and diminished profits in the seal oil trade. In general, the ringed seals were more widely distributed and harder to find when ice cover was widespread, and the other seal species moved southward. Warmer winter conditions seem to have had the opposite effect. Less ice meant that the seals were more concen- trated in the north and sought out the stable ice of archipelagos. The seals came to the hunters. This circumstance is described in Chukotka in Russia (Krupnik 1993), eastern Canada (McLaren 1958) and Greenland (Vibe 1970). Reduced ice cover and rises in temperatures at these latitudes also had major effects on terrestrial growing periods, as seen by tree pollen, insects and macrofossils on the North Bothnian coast (Broadbent 1979:158-169). The most common response by hunter-gatherers to major fluctuations of these kinds has been settlement mobility, with moves to coastal regions and then back to inland territories when conditions changed. Whole regions could be largely aban- doned for up to a generation between periods of bounty (cf Mine and Smith 1989). Igor Krupnik has also described this reciprocal settlement phenomenon, involving reindeer hunting and herding in the interior and whaling on the coasts in the Eurasian Arctic (Krupnik 1993:194-197). Shoreline erosion data from the seal hunting sites of the Lundfors complex in Vasterbotten indicate, for ex- ample, that these sites were only occupied over a 35-year period, or about one generation, and then abandoned (Broadbent 1979:30). The same phenomenon could also be true of interior settlements in northern Sweden, as moose were also subject to natural population swings and were highly vulnerable to predation (cf Ekman 1910:30-38; Markgren 1974; Peterson et al. 2003). Boreal mam- mals generally beneflt from cold and dry winters that limit parasites, ground ice and snow depth (Sugden 1982:114-127; Pruit 1978). This implies that there could have been a negative correlation between peaks in moose populations in the interior and the accessibility of seals on the coast. This possibility has considerable consequences for interpreting how the region may have been used by hunter-gatherers, but even more so regarding the role that animal husbandry played in adapting to environmental and cultural changes during subsequent periods. Husbandry offered a viable al- ternative to both natural and human-caused declines in resources as well as changes in territorial accessibility. CULTURE AND ECOLOGY 25 Punctuated Sedentism It is known that the Saami took precau- tions to maintain fishing waters by shifting to different lakes every few years (Ekman 1910:460-462), but longer-term regional shifts in settlement could have been forced on by the combined effects of ecological changes and over-hunting (cf. Krupnik 1993:156-159). Abandonment of the interior for up to a generation would have also helped keep moose, beaver and fish populations vi- able by allowing them to recover from preda- tion. As shown by historic accounts, moose and beaver had almost been completely wiped out by north Swedish settlers within a few centuries (Hiilphers 1789). Trapping pits were made illegal in 1864 in an effort to restore moose stocks, and beaver had to be reintroduced from Canada. I have described this long-term coast plus interior settlement model as "punctuated sedentism" (Broad- bent 2004) (Figure 15). A system of periodic or generational depopulations of the interior may have been the very reason why moose and beaver populations survived thousands of years of predation within this narrow peninsula of boreal forest. Rich coastal hunting resources offered an alternative, but also underwent cycles, and rapid landscape changes due to shoreline displacement may have necessitated decadal reorganiza- tions. For a semi-sedentary regional system of this kind to work, however, these groups would have needed access to both coastal and interior regions. When access to either one of them was blocked by other groups for any length of time, it would no longer be possible to fall back on these alternative sources of livelihood. That is probably one reason why reindeer and sheep/goat pastoralism was so readily adopted by these northern hunter-gatherers. Obviously fishing was of great importance, but there was fierce competition for fishing waters in the interior as well. Most northern strife recorded by the courts related to fishing waters (cf Campbell 1948:231-236). Resilience Theory Successful long-term adaptive strategies, as proposed here, can be characterized using the concept of resilience, the amount of change a system can withstand and still retain its structure. Resiliency builds on a system's ability to maintain diversity, flexibility and opportunity under changing condi- tions (Redman and Kinzig 2003:13). Cultural heterogeneity is the equivalent of species diversity, and the loss of diversity, while often increasing efficiency, also leads to vulnerabilities. Resilience theory (Flolling et al. 2002; HoUing and Gunderson 2002; Redman and Kinzig 2003) encompasses many themes of interest in this study: the diversity of "Saaminess," the pulses observed in coastal PUNCTUATED SEDENTISM Figure jj. Cycles ofmoose populations in the taiga, and long-term cycles ofsettlement on the Bothnian coast. Moose/predatorfigures cover a 40 year period (Peterson et al. 200]). 26 CHAPTER 2 settlements and, perhaps most important, the fact that the Saami have survived as a viable cultural and economic presence for thousands of years in northern Sweden. The persistence and survival of indigenous societies is starting to be recognized, as is the fact that indigenous cultures have greatly influenced settler societies (Murray 2004:1-16). The now-classic book Finns and Ter-Finns. Ethnic Processes in Northernmost Fenno-Scandinavia (translated title) by Knut Odner (1983) builds on Fredrik Earth's interpretations in which Saami ethnic identity is essentially defined as a process created through contacts with others (Earth 1969). This definition, and the dynamic concept of adaptive change through resiliency, can be extended to the idea of cultures as clines or seamless topographies, rather than entities with sharp boundaries or as patchworks of cultures (Caulkins 2001). Clines are also a way ofmapping cultural landscapes, and this approach will be applied to artifacts, sealers' huts and place-names throughout this book. As already noted, cultural and environmental change tends to be episodic and punctuated in time and space rather than incremental. Adaptive change occurs in two social dimensions according to resilience theory, one based on the amount of stored "capital" or energy and the second on the degree of connectedness of social systems (Holling and Gunderson 2002; Holling et al. 2002). The more complex, more specialized, larger and more connected a system, the more brittle and less resilient to major ecological or economic changes it becomes. In theory, all systems neverthe- less eventually reorganize. This release of capital, social as well as economic, is followed by a new growth period. When applied to the archaeological material presented here, one can readily identify some of these trends. Starting in the Early Metal Age/Eronze Age, there was a major reorganization ofnorthern hunter-gatherer societies. This process accelerated during the Late Metal Age/Iron Age. Northern trade networks evolved into economic dependencies on the scale of World Systems and were finally absorbed into state-level systems during the Early Medieval Period. These trends and the character of Saami resiliency as seen in the Eothnian coastal zone are major themes of interest. As an overall starting point, five hypotheses have been formulated to structure my analysis. Five Hypotheses 1) The Saami are an indigenous people with roots going back at least 7,000 years in northern coastal Vasterbotten. 2) There are two major cultural-ecological regions in Sweden, the Circumpolar and the Euro- pean. During the Iron Age the Germanic agrarian settlement boundary coincided with the 63rd parallel on the Eothnian coast. 3) Proto-Saami, Proto-Finnish and Proto-Germanic societies (for lack of better terms) had been in close contact for thousands of years and were heterogeneous and overlapping. 4) Coastal and interior settlement during the Stone Age in northern Sweden occurred in semi- sedentary cycles relating to peaks and declines in terrestrial and marine resources. Animal husbandry changed this pattern and contributed to sedentism as well as to nomadism. 5) Northern Sweden was part of a World System of trade and information exchange that had been in existence since the Early Stone Age. CULTURE AND ECOLOGY 27 Figure 16. Diagram ofenvironmental changes, strayfinds, site aggixgations and agrarian indications in Vasterbotten. Peaks and settlement cycles on the coast appear to coincide with warm periods as reflected by Baltic transgressions (Fairbridge 1963; Berglund icj6^; Digerfeldt igyy, Miller igyg). A warmer climate would have encouraged noHhern agrarian expansion, and sealing and agriculture have gone hand in hand in the region since at least 2000/2^}/^ B.C. Tliere are at least five horizons in Vasterbotten during which coastal settlement concentrated (uncal/cal): ^000/^goo, ^400/4^00, 2yoo/}4yo, 2000/2^^4. and iooo/i2]o B.C. There are alsofour periods ofagrarian expansion: 20oo/2^]4 B.C., 1000/12^0 B.C., A.D. 400/6]^ and A.D. iooo/io}g. Climate data in Northern Europe indicate warm periods between 2600-igoo B.C., 1^00-^00 B.C., A.D. o-^^o and A.D. goo-1^00. Cold and cool periods are recordedfor igoo-1^00 B.C., 500 B.C.-o, A.D. ^^o-goo and A.D. 3300-1850. Prehistoric Foundations The Early and Late Stone Ages This overview characterizes the long-term adaptive strategies of local societies as well as the envi- ronmental and cultural influences that have affected change in coastal Vasterbotten. The general prehistory of the region is described by Broadbent (1979, 1982), and of the Metal Ages by Serning (i960), Forsberg (1999), Rathje (2001) and Liedgren and lohansson (2005). Environmental changes following de-glaciation, including temperatures and forest succession, salinity and evidence of ag- riculture, are summarized in Figure 16. Nordic prehistory has been divided into three periods since first proposed in 1836: the Stone-, Bronze- and Iron Ages (Stenberger 1964; Burenhult 1999). A somewhat different terminology is used in North Norway and Finland (cf Carpelan 1975b; Olsen 1994) that is particularly useful because it encompasses the concept of the Saami Iron Age (dates according to Olsen 1994:14). The dates of the Saami Iron Age (1-1500 A.D.) coincide with the material in this study and are therefore referred to as the Late Metal Age in this text, unless referring to a specific Iron Age period (Table i). Calibrated dates in the text are listed together with uncalibrated dates (uncal/cal). The salty Litorina Sea (ca. 5000/5600-2000/2400 B.C.) was a time of maximum produc- tivity in the Baltic. The Atlantic climate was 3-5°C warmer than today; there was a greater overall terrestrial biomass in the north with more deciduous trees and the animals that feed on them, such as moose and beaver (Berglund et al. 1994:12-17). Specialized adaptations to coastal and inland resources developed with larger and more permanent settlements, increasing social and cultural complexity and new symbolic expressions and exchange networks. Sedentism and semi-sedentism were the hallmarks of most of the Eurasian world during this time (cf Hall 1999:8). This was an era of demographic and social consolidation and the establishment of the regional societies that were the foundation of many cultures throughout Eurasia. The Flurkmark site (ca. 5000-2500 B.C.) near Umea in Vasterbotten is a huge settlement area that extends over 300 m on several sandy terraces about 95 m above current sea level on what was once an inlet of the Litorina Sea (Lundberg and Ylinen 1997; Sjogren 1997; Lundgren 2001; Stora 2002; Broadbent 2003). The site was connected by an esker to the Vindel River valley. This site is a superb example of the early, sustained and intensive use of the Bothnian coast and the natural routes that connected the coast to the interior and other regions. Raw materials used for tool-making include Baltic flint, basalt, quartzite, quartz, jasper and other stone types from widespread western, southern and eastern sources. Red ocher, which was used in both graves and rock art in northern 29 Table 1. Chronological periods in Sweden and Norway (uncalibrated). SWEDEN NORTH NORWAY Mesolithic Neolithic Bronze Age Early Iron Age Late Iron Age 400 B.C.- 400 A.D 400-1100 A.D. 4500-1800 B.C 1800-400 B.C. 13,000-4500 B.C. Early Stone Age Late Stone Age Early Metal Age Late Metal Age (Saami Iron Age) 10,000-4500 B.C. 4500-1800 B.C. 1800 B.C.-l A.D. 1-1500 A.D. Medieval Period 1 100-1500 A.D. Sweden and Finland, was found in thick deposits on the site. The animal bones include those of ringed- and possibly gray seals, moose, beaver and a variety of small mammals, fish and birds (Stora 2002). Ringed seals were the most common species, followed by beaver. Pike was the most common fish, and ducks the most abundant of birds. Spring, summer, fall and winter are indicated by this material. Several trapping pits had been dug on the site in much later times. (The site was discov- ered because a transverse-based arrowhead of Early Metal Age type was found there.) This is still good area for moose hunting. Large settlements of these types are found in considerable numbers at similarly high elevations along the Bothnian shores, from Norrbotten County in the north to the Stockholm region in the south (cf Lindgren 2004). The Lundfors Site (3400/4200 B.C.) is about 100 km north of Flurkmark and lies at 78 m above current sea level. It consists of seven separate settlement concentrations on an inlet south of the Skellefte River estuary (Figures 18). This site exemplifies the collective hunting strategies that were widely employed on the coast and in the interior to catch terrestrial and maritime animals (Broadbent 1979:174-198). The Lundfors economy was based on the capture of ringed seals using long net systems, known to be the most effective during the dark fall months (cf. Hamalainen 1930). Ringed seals would enter these bays and swim up rivers, gorging on fish in the fall; their blubber and pelts were at their maximum at this time of year (Ekman 1910:252; Holm 1921; Helle 1974; Broadbent 1979:187). The use of nets to catch ringed seals - other species are usually too large to be caught this way - was still the main sealing method in Vasterbotten in the sixteenth century (Kvist 1991). The same wide-meshed nets could be used to catch salmon and beaver (Ekman 1910:218). In addition to seals, bones of moose, beaver, small game, fish and shellfish were identified and these show that local islands, inlets and valleys were also exploited. A perforated digging stick weight (otherwise typical of Finland) was also found (Figure 17), indicating plant (especially root) collection (Broad- bent 1986; Broadbent 1979:171-173)- The seal nets were probably made of braided roots, willow or other plant fibers, although only the net weights have survived. Huge Stone Age net systems for both seals and salmon have been found preserved in Finland at the Fori site (Luho 1954) and at Kierikki (Koivonen and Makkonen 1998). Stone technology at Lundfors was based on quartz that was quarried locally, ground schist adzes of Finnish types, and slate knives and projectiles, some ofwhich were made of red slate from the Swed- ish alpine regions. Red ochre was found in clumps and produced by burning on the site (Broadbent 30 CHAPTER 3 Figure ly. Stone artifact typesfrom Lundfors (not to scale): a perforated digging stick weight, a ground slate knifefor butchering seals and two slate pendants. Figure 18. Lundfors inlet and the locations of the sites and lines of nets that were used to catch ringed seals in the fall. Hundreds of net weights werefound on the site and on the bottom of the jormer inlet that now lies y8 m above sea level. 1979:134-135). A rock painting of a moose is recorded behind the site but has never been relocated. Massive amounts of fire- cracked stones link this settlement to the indigenous boiling stone (and non-ceramic) technology of interior northern Sweden, where dwellings were built up by mounds of fire-cracked rocks (Lundberg 1997). The Lundfors settlement area covers 12,000 m^ and was almost certainly semi-sedentary. A detailed analysis of beach erosion shows, nevertheless, that the whole complex was used over only a 35-year period, or one gen- eration (Broadbent 1979:30). A comparative analysis of 13 sites from both coastal and interior Norrland, dating to the period 7000 B.C. to A.D. i, challenges the idea that these types of sites were only temporary "aggregation camps" (Kack 2009). The Lundfors settlement model repeats itself even at the great rock art site at Namforsen in coastal Angerman- land, which dates to the Late Neolithic- Bronze Age. Some 590 transverse-based projectiles, 6 kg of asbestos-tempered and textile-impressed ceramics, grooved stone clubs, large amounts of slag and even part of an iron furnace, were found at the adjacent Stallverk settlement site (Kack 2009:50-94). The rock art was produced by a local population with widespread social networks (Kack 2009:154-184), and this ar- tifact material is closely associated with the most formative period of Saami ethnogen- esis throughout the Nordic region. Rock Art Moose remained the central icon in rock art and represented the core of indigenous identity in both coastal and interior north- ern Sweden until the end of the rock art period about 600-400 B.C. The moose as a symbol, especially in the South Saami PREHISTORIC FOUNDATIONS 31 region, harkens back to the roots of these societies in the boreal forest zone and was widespread in northern Eurasia (Hult- krantz 1964; Ramqvist 1992; Zachrisson 1997a; Fanden 2002). The brown bear also persisted as an important symbol and was one of the most significant wild animals in the Saami pantheon (Backman and Hultkrantz 1978; Helskog 1988; Eds- man 1994; Mebius 2003). The largest rock art site on the Vasterbotten coast is at Stor- norrfors on the Ume River (Ramqvist et al. 1985). Moose figures dominate, includ- ing "x-ray" images (Figure 19), which are also seen at the aforementioned rock art site at Namforsen (Hallstrom i960). New Ideologies, New Technologies The Late Middle Neolithic (1800/2250 B.C.) has been identified as a climate optimum in northern Europe that led to an expansion of agriculture and husbandry northward throughout the Nordic region (cf Graslund 1980; Sjovold 1982). The Sub-Boreal Period, 2400-400 B.C., was a time ofextreme dryness in western Eurasia (Mayewski et al. 1993; Mayewski and White 2002). The desertification of the Russian steppes led, among other things, to the northern and western expansion of herders and herding technologies into the boreal forest zone (cf. Toynbee 1934; Gimbutas 1965; Frachetti 2008). These societies introduced new burial practices, religious and social ideologies and, equally important, metal technologies (Chernykh 1992). The Bjurselet site (1900/2400 B.C.) is one of 10 locales in coastal Vasterbotten characterized by caches of thick-butted flint adzes of Baltic types, useful for working wood but also kept as grave goods (Christiansson and Knutsson 1989). Over 295 adzes have been found, 175 from Bjurselet alone (Figure 20). This technology is associated with the spread of the Boat Axe (Battle Axe) Cul- ture, also referred to as the Single Grave or Corded Ware cultures in Sweden, Denmark and Finland and the Fatjanovo culture in Russia (cf Maimer 2002). The largest find locales in Vasterbotten are by rivers and estuaries, and the smaller concentra- tions are in coastal valleys. Pollen analysis has indicated that barley was cultivated and some sheep/ goats were kept (Konigsson 1968), but ringed seal bones are by far the most common animal bones found at Bjurselet, and net weights like those at Lundfors were found at the Kusmark site (Lepik- saar 1975; Broadbent 1982). A thick-butted flint axe was even found in a grave cairn in Vasterbotten (Huggert 2001) and the Battle Axe Culture people are known to have cremated their dead. There is a continuation of these burial practices into the Early Metal Age. As regards metallurgy, early copper finds have been found in northern Sweden and Finland dating to as early as 3900 B.P. (Huggert 1996). This was, in other words, a major turning point in north Swedish prehistory and established the foundation for subsequent cultural development. Figure ig. An "x-ray" rock caning al Stomonfors in coastal Vasterbotten. Tlie image shows a life-line or a spear, the sectioning ofthe body, and two points perhaps representing the testicles ofan immature bull. 32 CHAPTER 3 Figure 20. Unused thick-butted Jlint adzesfrom the Bjurselet site in Vdsterbotten. Courtesy Antikvariskt-topografiska arkivet, the Nnfio)!((/ Heritage Board. Stockhohn. Figure 21. A notched wooden spear shaftfrom Ostra Ahyn in Bygded Parish, coastal Vasterhotten, was poUen dated to the Bronze Age (Westin 1962 j. The pattern looks like stitching and is uncannily similar to Saami pewter wire designs on cloth and leather (courtesy Antikvariskt- topografiska arkivet, the National Heritage Board, Stockholm). The Early and Late Metal Ages The Early Metal Age of the Nordic North, during which there was an intensification of east- ern influences, is characterized by asbestos-tempered and textile-impressed pottery, transverse- based projectiles, grooved stone hammers and Russian "Ananino" bronze artifacts and molds (Carpelan 1975a; Bakka 1976; Chernykh 1992; Olsen 1994; Hjarthner-Holdar and Risberg 2001) (refer Figure 22). Transverse-based projectiles, also called straight-based, originated in the southern steppe regions of Russia and have been found in Proto-Scythian Timber Grave burials in the Lower Volga basin (Gimbutas 1965:544). They are associated with the spread of powerful, probably composite, bows. The projectiles date to approximately 1700-500 B.C. in Sweden (Baudou 1992:43), but finds have been made in North Norway dating to as early as ca. 2200 B.C. (Helskog Thrash 1983). Asbestos fiber-tempered clay vessels may have functioned as portable furnaces for metal work- ing (Hulthen 1991). The grooved clubs were probably used as sledge hammers and are also found in South and Middle Russia (Indreko 1956:58-59). Asbestos (chrysotil) deposits have been found far upstream on the Skellefte River at Ruopsok near Lake Hornavan in northern interior Sweden (Hulthen 1991:14) and in East Einland. This metallurgy, which also has many shamanistic aspects, is central to the development of Saami culture (cf Carpelan 1975a; Jorgensen and Olsen 1988). Whole complexes of pits in coastal Norrbotten were used for the rendering of seal oil (Lun- din 1992). They have also been documented in coastal Vasterhotten (Forsberg 1999) and date to the PREHISTORIC FOUNDATIONS 33 Figure 22. Map showing the spread of related technologies into the Nordic regionfrom the Russian steppes and the Ural Mountains, Late Stone Age/Early Metal Age. Late Bronze Age/Early Metal Age. The intensification of sealing and processing, together with large numbers of trapping pits in the interior during the Late Metal Age/Saami Iron Age, is undoubtedly connected to intensified trade (cf Selinge 1974; Spang 1997). North Swedish hunter-gatherers were be- coming part of a large-scale trading network that would reach its maximum during the Viking Period. The Fahlmark site in northern Vasterbotten lies at 39 m above sea level, dates to ca. 1300 B.C., and has three finished and 35 rough-outs of quartzite and basalt projectiles (Arwill 1975). A bundled cache of similarly worked projectiles was found at Vannas near the confluence of the Ume and Vindel rivers. A fragment of a point made of Baltic flint, a grooved stone club, and a red slate pro- jectile of Sunderoy type, a contemporary artifact type first described in North Norway (Gjessing 1942:172; Baudou 1977:30-31; Olsen 1994:106), was found at the Kaddis settlement at the 40 m elevation near Umea (Broadbent 1984). There are some 16 additional projectile finds in this Swed- ish coastal region, including one made of Russian flint at Bjurselet (Huggert 1984) and another at the Flurkmark site. A bifacial projectile or knife of this type was even found on the Lundfors site (Broadbent 1979:237). Prastsjodiket is a stone grave setting that had been placed on an island near the Kaddis settle- ment during the Early Metal Age. The grave goods consist of a quartz knife, transverse-based ar- rowheads of Baltic flint, Russian flint, basalt and quartz, and a bronze spear. The cremated bones oftwo young individuals were found (Lundberg 2001). Like Flurkmark, which is dated 3,000 years earlier, this find reflects a diverse mixture of raw materials and the far-reaching networks that con- nected this region to the outside world. There are almost 600 whole or fragmentary arrowheads with transverse-bases at the rock art site of Namforsen in coastal Angermanland (Forsberg 2001; Kack 2001, 2009). The sheer 34 CHAPTER 3 numbers of these characteristic artifacts in these coastal contexts imply that these people were no different from the cairn builders (cf Bolin 1999 for the same conclusion regarding Angerman- land). A transverse-based arrowhead of quartzite was actually found in a Bronze Age cairn grave in Angermanland (Baudou 1968), and this funerary association is paralleled by the Prastsjodiket grave in Vasterbotten. Bronze is an alloy of copper and tin and required access to these two rare metals. Nearly all of it had to be brought in as raw materials, or as finished products. There are two primary Euro- pean sources of these metals, one in the Ural Mountains of Russia and the other in central Europe. Copper, tin and gold helped solidify the power of a social hierarchy that had started to take form in the Late Middle Neolithic in Sweden (Maimer 2002). A hierarchical social system, with chiefs and big men, was sustained by the redistribution of metals for agricultural products, furs, hides and amber (Welinder 1977). Vast trade networks and alliances were reinforced and validated by religious symbols as seen in south Scandinavian rock art, such as sun disks, weapons and boats, some of which have been found at Namforsen, as well as grave rituals and sacrificial practices (Kristiansen 1994; Larsson 1999). Cairn Graves There are more than 550 grave cairns in coastal Vasterbotten. Most large cairns (38%) are at the 35 m elevation, corresponding to the Bronze Age shoreline, and smaller cairns are found at Iron Age levels, or 10-20 m above sea level (Figure 23). Most ofthe cairns occupy coastal headlands and islands, but others lie in inner fjords where settlements were located (Bergvall and Salander 1996; Forsberg 1999). Baudou (1968) noted that the cairns in Angermanland seem to divide the coast into segments 5-15 km in length, probably corresponding to social territories or clans. I have proposed the same clustering (Broadbent 1982:117-121), as did Bergvall and Salander (1996) in northern Vasterbotten. The cairn graves probably also mimicked their circular and oval house forms. These northern graves had been built by local populations who had taken up southern Scandinavian ritu- als, including cremation, but can hardly have been building monuments to Bronze Age chiefs, as often argued in southern Scandinavia (Broadbent 1983). One of the oldest theories about the end of this large cairn building era is that climate dete- rioration had caused the Scandinavians to withdraw, and that this vacuum was filled by the Saami (Hallstrom 1929). The Sub-Atlantic Period, starting at 650 B.C, was indeed a period of climate deterioration, with lower average temperatures and increased precipitation (Lamb 1977, 1995)- Granlund (1932) first recognized these changes in northern Sweden through water-logged levels in bogs called "recurrence" surfaces. There were drops in tree lines and glacial advances (Nesje and Figure 23. Sketch ofboth large and small grave cairns in Vasterbotten by A. F. Ekdahl. iSij. PREHISTORIC FOUNDATIONS 35 Kvamme 1989). Cultivation com- pletely disappeared on the north- ern Vasterbotten coast at this time (Engelniark 1976). These changes in Bothnian coastal settlement are best seen as an overall adaptive response by indige- nous populations that took the form of greater mobility, a pattern that oc- curred at exactly the same time in northern interior Sweden (Forsberg 1985; Mulk 1994:229-251). Olsen described this same phenomenon in northern Norway, with a greater focus on interior and alpine regions and an increase in reindeer hunting. These changes dated from ca. 900 B.C. to A.D. I and coincided with the Kjelmoy ceramic phase (Olsen 1994:106-124). A ••••Am ASBESTOS-TEMPERED CERAMICS Figure 24. The distribution ofasbestos-tempered vessels (after Baudou igg^ with additional points). Arctic Bronze Age aiiifacts and the Bothnian ski, a probableforerunner of the Saami ski, have almost identical distributions (Broadbent 1^82:14^,146). Iron Iron deposits are formed as precipi- tates in lakes and in bogs and could be easily scooped up into boats or through the ice in winter. The wide- spread availability of iron ore in Sweden thus completely side-stepped the old bronze redistribution system. Eva Hjarthner-Holdar (1993) has described the introduction of iron technology into the Sweden and Finland as closely associated with the boreal forest and as "the first great egalitarian movement in Europe" (Hjarthner-Holdar, pers. comm. 2006). Harrsjobacken and Hamptjarn near Burea in Vasterbotten are good examples of this pattern. Textile-impressed and asbestos-tempered Kjelmoy type pottery was found, as well as iron slag, in a forging pit at Harrsjobacken. Hamptjarn is a nearby small cairn cemetery, as is the site of Nedre Back, with one grave containing bear phalanges. An oval cooking pit at Harrsjobacken was radiocarbon-dated to A.D. 79-245 (Sundqvist et al. 1992). Testing of the soils for lipids shows that seal oil had been rendered there as well. Asbestos-tempered pottery of Kjelmoy type and textile- stamped pottery, with flakes characteristic of the production of transverse-based arrowheads, were also found at the Savar site, which has been interpreted as a semi-sedentary sealing settlement (Sanden 1995). Animal Husbandry and Cultivation Pollen diagrams from northern coastal Sweden indicate sporadic cultivation during the Late Mid- dle Neolithic or around 2000/2400 B.C. and from 600 to 400 B.C. (Konigsson 1968; Huttunen 36 CHAPTER 3 and Tolonen 1972; Engelmark 1976; Wal- lin 1994). Sheep/goat bones have been found at Bjurselet, Kaddis and Fahlmark (Lepiksaar 1975; Broadbent 1984). There is a 900-year gap in the agrarian footprint in Vasterbotten from ca. 400 B.C. to 500 A.D., however, followed by yet another gap in the seventh century, particularly on the Lule and Torne rivers. Wallin (1995) inter- prets the agrarian discontinuities as a re- sult of frequent shifts in the small coastal field systems, but they were also undoubt- edly due to climate deterioration during the Sub-Atlantic Period and a worldwide climate deterioration in the seventh cen- tury (Lamb 1977; Nesje and Kvamme 1989; Roberts 2009). When conditions were favorable and cereals or livestock could be obtained, cul- tivation and animal husbandry were un- doubtedly part of a northern indigenous economic strategy during the Early Metal Age and continuing through the Saami/ Late Iron Age. Asbestos pottery of Kjelm0y type was found with some charred barley in northern Angermanland (Lindkvist 1994:98), and hair-tempered textile ceram- ics have been found at the Bjurselet site (Sanden 1995). Textile-impressed ceramics, which arose between the Oka and Middle Volga, accompanied early metallurgy and possibly agriculture to Karelia and eastern Finland (Lavento 2001). Bones of domestic animals, especially sheep and goats, have been identified at Halla, Ra-Inget and Stallverket by Namforsen in Angermanland, and this evidence shows that animal husbandry and even cultiva- tion were part of the "package" of both Proto-Saami and later Saami culture in northern Sweden. Early Saami Iron Age cultivation has likewise been documented in Norway (Bergst0l 2008). Artifacts in Context Although we don't have any iron implements from this earliest period, there are some 25 Iron Age artifacts from the coastal area under study. One remarkable archaeological find was a bun- dle of decorative bronze pins and brooches, dating to ca. A.D. 350, at Storkage. They derived from the South Baltic Finno-Ugric region (Estonia) (Hjarne 1917; Serning 1960:21-24). These items were brought to the north, only to be buried on a small coastal river bank. This find resembles the Saami metal offerings of later times in which valuable, and especially exotic, orna- ments were put into bogs or lakes. Figure 25. A soapstone moldfor a socketed axe of Russian Ananino type. Courtesy Antikvariskt-topografiska arkivet, the National Heritage Board. Stockholm. PREHISTORIC FOUNDATIONS 37 Figure 26. Iron Age artifactsfrom coastal Vasterbotten with parallels at Saami offer sites {boxes). Metal artifact distributions in interior indicated by shaded areas. Fifty percent of the coastal metal artifacts have parallels at Saami offer sites in the interior. Among these are iron arrowheads and oval and horseshoe-shaped brooches (Serning i960; Pos- sum 2006) (Figure 26). Bronze bells and a silver ring from the Island of Stora Fjaderagg, some 14 km offshore, have parallels at the Saami offer sites, as does a crescent-shaped earring from Tame in Byske parish (Serning 1960:36-49). An odd fork-like object, possibly for shamanistic divination, was found at Skramtrask, near Lundfors, and has a parallel at the offer site of Gratrask (Serning 1960:147). A decorated knife from Karelia was found at Kataselet on the Byske River (Broadbent 1982:170). Archaeological investigations of sacrificial/offer sites in the interior, which were known from oral traditions, have sometimes rendered great numbers of finds (Serning 1956; Zachrisson 1984). The most common artifacts derive from Finland, Russia and the Baltic, but many coins witness contacts with Norway, Germany and Britain. The objects consist of brooches. 38 CHAPTER 3 pendants, clasps, buckles of pewter, bronze and silver, silver coins and iron arrowheads. The finds date to A.D. 700-1400, with most sites dating to A.D. 900-1100 (Zachrisson 1984; Wallerstrom 2000; Mulk and Bayless-Smith 2006). Seven metal artifacts (28%) from the coastal region are from small grave cairns. These crema- tions are parallels to the so-called "inland lake graves," "forest graves" or "hunting ground graves" found in lamtland and Harjedalen in Southern Lapland. In evaluations of the evidence regarding the hunting ground graves in Jamtland and Harjedalen, and Trondelag in Norway, Sundstrom (1997:21- 27), Gollwitzer (1997:27-33), Zachrisson (1997a: 195-200) and Bergstol (2008) argue that these inland graves are of Saami origin. In the case ofthe Vivallen cemetery there is, however, physical an- thropological evidence ofintermarriage (Iregren 1997:84-98; Alexandersen 1997:99-116). Liedgren and Johansson (2005) describe 15 graves in Upper Norrland that can date to the Early Iron Age. Most graves are known from the coast, where there is no evidence whatsoever of Germanic settlement. These authors state, somewhat ambiguously, that the "burial customs" were probably not of Saami origin, and that the metal artifacts in Lapland were due to the Saami dependency on trade for iron objects. As will be discussed in Chapter 9, there is ample evidence that the Saami were themselves able iron metallurgists and, except for arrowheads, most Lapland finds are non-utilitarian. It is very likely that many Swedish Saami groups practiced cremation in the first centuries A.D. (cf Zachrisson I997a:i95-i97) and this pattern is certainly seen at the large cemeteries at Smalnaset and Krankmartenhogen in Harjedalen in southern Lapland (Ambrosiani et al. 1984). These cremation graves, including triangular stone settings, are Germanic in form, but offerings of reindeer, moose and bear, in addition to their locations far outside any known Germanic settlement areas, indicate that they are the graves of what has been circumspectly described as "local hunting cultures." With some individual exceptions, the northern coastal graves should be interpreted in the same light as the other "hunters' graves," as those of local populations (cf Forsberg 1999:251-285; Rathje 2001:91-118; Fossum 2006:89-99). Albeit in close contact with their neighbors, these so- cieties could logically only have been Saami in origin. Wooden artifacts shed some additional light on the identities of these people in coastal Vasterbotten, including two nearly identically decorated skis from Kloverfors and Bygde- trask (Serning 1960:62) (Figure 27). Both have Viking Period pollen dates, substantiated by C-14 (Astrom and Norberg 1984). Their plaited-band and ribbon ornamentation is typical of Southern Saami decorative styles (Holmqvist 1936). Vaster- botten is, in fact, the core area of the Bothnian ski type, which is Figure 27. Two decorated Viking Age Saami skisfrom Vasterbotten (after Seniing igGoiiSo.zyS). PREHISTORIC FOUNDATIONS 39 the basis of the Saanii ski (Manker 1971; Astrom and Norberg 1984). These ski types, as pointed out by Manker (1968), together with finds ofsewn boats (Westerdahl 1988), are evidence of a long- standing Saami presence in the North Bothnian region, as do trapping pits with Lapp place-names (Broadbent 1982:90-97). Summary On the basis of this overview, the following generalizations can be formulated with respect to con- tinuity on the northern Vasterbotten coast: Economy 1) Sealing was a major activity during all periods. Subsistence sealing was combined with terrestrial hunting, fishing and gathering and with animal husbandry/farming from ca. 1900/2340 B.C. Moose-hunting, beaver-hunting and seal-hunting were organized col- lectively through the use of trapping and netting systems. Hunting took on an almost mercantile focus starting in the Early Metal Age, as seen by boiling pit complexes and trapping pit systems, and this was intensified during the Late Metal Age. A special form of seal oil rendering pit, more trench-like, but also containing fire-cracked rocks, has been described by Rathje (2001:134-136) and dates to the seventh century A.D. 2) Mixed hunting and herding societ- ies established territories that di- vided the coast into socio-economic units encompassing bays and is- lands in 5-15 km wide segments. 3) Trade was extensive during all periods. Both raw materials and artifacts of stone, bronze and iron were regularly circulated within an area extending from Norway in the west, the Urals in the east, and to southernmost Scandina- via, Finland and the Baltic region. This can be seen as part of a World System of information and object- exchanges with roots going back to the Early Stone Age. Settlement i) Most coastal sites align with the larger eskers and river valleys. The important archaeological sites of Fahlmark, Harrsjobacken, Lund- fors and many other sites are found along natural inland travel routes. Figure 28. Eskers, sites mentioned in the text, and the economic zones that connected the coastlands with the interior. 40 CHAPTER 3 2) Larger and more permanent coastal settlements concentrated in inner f]ords and bays where there was the best micro-climate and, in the case of husbandry, grazing-lands. Technology 1) Quartz/slate and boiling stone technology was identical on the coast and interior ca. 5000 B.C.-2000 B.C. 2) A "stimulus package" of new technological knowledge (including metallurgy) reached the northern coast ca. 2400-1800 B.C. 3) Transverse-based projectiles of quartz, quartzite, flint, and fluted slate projectile types, asbestos-tempered ceramics, Ananino bronzes and iron technology were associated with sealing, animal husbandry and even cultivation on the Bothnian coast during the period 1600 B.C.-A.D. 100. 4) Iron working was being practiced from the late first century A.D. This world was served by mobility and the exploitation of new ecological niches. New cultural identities were being formed in an ever-widening circle that came to encompass most of northern Scandinavia and Finland. Ideology and Religion 1) The iconography of coastal and interior rock art reflects circumpolar cosmologies that in- volved mediation with the northern animal world, especially moose and bears. Shamanism was also manifested through transformative technologies, especially metal working. 2) New burial rites, including cremations, and offer practices were introduced to the Bothnian coast from the south and east. Cairn graves were built by local people to mark their coastal territories during the Early Metal Age and continued until the influences of Christianity took hold in the late Viking Period. Conclusions These data demonstrate that all the social, economic, technological and symbolic components as- sociated with the process of Saami ethnogenesis were present in coastal Vasterbotten, and there is little difference in these respects from North Norway, interior Sweden or Finnish Lapland (c£ Car- pelan 1975a; Forsberg 1992; Hansen and Olsen 2004:36-42). In nearly all regards, there is a congruence of cultural characteristics, artifacts and landscape geography, beginning in the deep past and continuing into later prehistory. This is a historical tra- jectory, as conceived of by Wolf (1997:3-24), that defines adaptive strategies at the local level, the regional level and the super-regional (World System) level. The coast was intimately connected to the interior by natural west-east travel routes, by social interdependencies, and through the exploitation ofcombined terrestrial and maritime resources. These patterns were strongly manifested during the Early Metal Age. These criteria set the stage for and regionally contextualize the narrative of coastal settlement during the Late Metal (Iron) Age and Medieval Periods. This material offers an alternative to the interpretation of northern cultures as the results of continuous population incursions. The focus is shifted to indigenous societies undergoing transformations through frequent interactions with the outside world. The ethnogenesis of the coastal Saami is part of this process within a frame- work of the meeting of eastern and southern networks at cultural-ecological borderlands. PREHISTORIC FOUNDATIONS 41 Figure 2C). The north to south transect along the Bothnian coast: (A) Stor-Rebbeu. (B) Hornslandsudde, (C) The northern Umit ofGermanic settlement. Tlte dark shaded area marks current Saami territoiy. and the Ughter shaded area is the probableformer extent ofSaami territory. The Research Strategy This section presents the methods, surveys, excavations and results obtained during the proj- ect. Each site is presented individually and includes artifacts, radiocarbon and AMS dates, archaeo-zoological and botanical analyses, soil-chemical and other results. Archaeological investigations were carried out at nine locales. Thirty-one huts at 12 elevations between 3 and 20 m above sea level were excavated. A glossary of the main types of archaeological features provides an overview of these and other constructions. The core area of Bjuron is presented first, followed by the sites of Stora Fjaderagg, Snoan, Stor-Rebben and Hornslandsudde. Surveys Surveys conducted by the Swedish National Heritage Board list numerous Bothnian locales lower than 20 m above sea level. There are two principal contexts: sites on raised beaches more than 10 m above sea level and harbor sites less than 10 m above sea level. Both types of locales were investi- gated to determine their chronological relationships and to address questions regarding continuity from the Iron Age into medieval times. Seal hunting sites were closely associated with islands and promontories where hunters could establish camps, oversee nets or kill seals on land, in the water and on the ice. Lower-lying hut foundations were associated with fishing locales, especially harbor basins (Norman 1993). Similar sites with temporary huts and shelters have been documented in southern and western coastal areas of Sweden and Norway (Magnus 1974; Atterman 1977). Excava- tions of the North Bothnian sites were first carried out by the author in 1987-1988, with follow-up excavations and analysis in 2004-2006. The excavations at Hornslandsudde were made possible by Britta Wennstedt Edvinger and Kjell Edvinger (Arkeologicentrum i Skandinavien AB). h has long been recognized that hut foundations found between 10 and 20 m above sea level were associated with Iron Age shorelines (Hallstrom 1942, 1949; Westin 1962; Varenius 1964, 1978; Westberg 1964; Steckzen 1964; Westerlund 1965). According to the national survey, there are 497 of these huts: 74 huts (14%) in Norrbotten, 219 huts (44%) in Vasterbotten, 136 huts (27%) in Vasternorrland and 68 huts (14%) in Gavleborg County (Norman 1993:194-195). Sixty percent are found in the two northernmost counties, of which 75% are in Vasterbotten County. Site Investigations The first goal of the project was to obtain as representative a sample of sites, features and artifacts as possible. The second goal was twofold: to investigate sites at different elevations above sea level 43 and to sample sites along a north-south transect. This approach facilitated com- parisons of different time periods within different regions and constituted a form of horizontal stratigraphy and chorology. The northernmost investigations were carried out on Stor-Rebben Island, Pitea Municipality, in Norrbotten County at 65° 11' N, 21° 56' E. This site complex is the largest of its kind this far north. Bjuroklubb is located about 80 km south of Stor-Rebben and is 200 km north of Arnas in Anger- manland County, the northern limit for Germanic Iron Age settlement and settle- ment place-names. The southernmost exca- vations were carried out at Hornslandsudde in Halsingland at 61° 37' N, 17° 29' E. Hornslandsudde is ca. 460 km south of Stor-Rebben as the bird flies, and 300 km north of Stockholm. This is the largest site of its kind this far south and is comparable in size to Grundskatan and Stora Fjaderagg in Vasterbotten and Stor-Rebben in Norrbotten. The association of this site with "Lapp" place-names and its oral history make this locale of special interest to the project. Surveys of Saami settlements in Halsingland were previously carried out by Wennstedt Edvinger and Ulfhjelm (2004). The Hornslandsudde site is otherwise within a region of well- known Germanic Iron Age settlement (cf. Liedgren 1992). Some of the largest sites are located near Bjuroklubb on Bjuron Island in Lovanger parish in Vasterbotten. This is a poor region as far as salmon fishing, or even agricul- ture, are concerned and the parish is not associated with any larger river systems. This contrasts with what is known regarding typical Germanic settlement areas in southern and middle Norrland. Elevations One of the most important aspects of the project has been the determination of accurate elevations of features and beach levels. This was accomplished through the use of on-site surveying equipment, including laser total stations. Daily sea levels were obtained by automatic "mariographic" stations. Handheld sighting levels were also used to deter- mine elevations above mean sea level. Figure jo. Distribution of Iron Age huts along the Bothnian coast, from Norrbotten County in the north to Gdvleborg County in the south (total: ^gj). Figure ]i. Detail ofGulfof Bothnia and river drainages showing the locations of Iron Age sites excavated within the project: (1) Stor-Rebben, (2 ) Bjuroklubb and Bjuron, (j ) Jungfhihamn, (4-6) Grundskatan, (y) Stora Fjaderagg, (8) Snoan, (g) Hornslandsudde. 44 CHAPTER 4 Table 2. Investigated sites, the distances between site areas, elevations sannpled at each site and numbers of radiocarbon dates from each site. INTER-SITE SAMPLE SITES (N-S) DISTANCES (KM) ELEVATIONS (M) DATES Stor-Rebben 16, 13, 12, 7 4 Bjuroklubb (67) 75 17 1 Site 70, Jungfruhamn (138, 139) 2.4 12,15,16 4 Grundskatan 2.4 16,15,14,13,12 15 Stora Fjaderagg 76 20, 19, 15, 13, 8 6 Snoan 73 15, 5, 3 6 Hornslandsudde 172 20,18,16,13 7 Because of the effects of land rise, the harbor basins were short-hved and boat houses and cottages had to be moved numerous times. When possible, historical records were used for dating such sites, but when these were uncertain, harbor basin thresholds were determined. Excavations were also carried out on any huts or related features by these basins in order to obtain radiocarbon dates. Threshold levels for harbors were determined at Jungfruhamn on Bjuron, at Gamla Hamnen Figure 32. Aerial view of the uplijied fishing harbor at BjiuvMubb (refer Figure '^o). THE RESEARCH STRATEGY 45 on Stora Fjaderagg Island and at two harbors on Snoan Island. In addition to survey information provided by the Swedish National Heritage Board and local county authorities, independent surveys were conducted at all site areas, resulting in a greatly expanded corpus of archaeological features. Excavations The outer coastal landscape is characterized by exposed bedrock, wave-washed moraine and shal- low soils. Preservation of organic material is poor. Excavations were therefore most often directed toward hearths, which were usually the only preserved cultural deposits. On the positive side, it is rare that one has so many visible architectural features. The use of beach cobbles for construction resulted in a rich inventory of often intact and fully exposed dwellings, stone alignments, cairns and other features. Although the find elevations provide a chronological foothold, since features cannot be older than their contemporary shorelines, many constructions could have been built long afterward. Because of this source-critical problem, considerable efforts were put into additional ways of dating them. Radiocarbon dates were obtained from hearths, but for most other categories of construc- tions there is nothing to date but the stones themselves. Our application of lichenometry (lichen growth measurement) was intended to overcome this problem. These methods have always been used in combination with shoreline displacement dates (maximum ages), radiocarbon chronol- ogy of adjacent features and historical data. Lichen dates are always considered as minimum ages. Rock weathering made it possible to determine if stones had been overturned during construction, thus reducing the risk that lichens had only been moved from beaches as opposed to colonizing a newly exposed surface, although lichens don't normally survive being moved (Benedict 1985). Rock- weathering was quantified using the Schmidt Test Hammer (Sjoberg 1987). Sampling Excavation was based on a sampling strategy intended to recover charcoal, macrofossils, animal bones and artifacts from different elevations above sea level. Hearths and different hut forms were chosen at different elevations, from the highest-lying to the lowest-lying, at each given site. Hearths were first tested using a soil auger and subsequently chosen for excavation based on the presence or absence of organic materials. All soils were screened using 4 mm mesh or less, and sometimes whole hearths were removed for flotation in the lab. Soil Chemistry Phosphate enrichment of site soils, due to human or animal defecation and urination and animal carcasses, was analyzed using laboratory methods by Johan Linderholm at the Environmental Archaeology Laboratory at Umea University. Later comparative field tests were conducted by the author using the La Motte Soil kit. Phosphate analysis (P°) is based on the extraction of organic phosphates from soil samples using a weak (2%) solution of citric acid and measured colorimetri- cally. Phosphate content is defined as mg/P^Oyioo g dry soil. The La Motte system is based on a colorimetric scale indicating Low, Medium or High levels of soil enrichment. Animal Osteology All animal bones from the project were analyzed by Dr. Jan Stora of the Osteological Research Laboratory of Stockholm University (Stora 2002, 2005, 2008). When possible, bone fragments were 46 CHAPTER 4 identified as to class, species, bone or bone fragment and side. Because of fragmentation, only a few metrics were possible. The identified fragments were individually counted and weighed, while unidentified fragments were only counted. Quantification was based on the numbers of individual fragments (NISP, or numbers of identified specimens). Refitted fragments were counted as single specimens, as were fragments found together. The reference collection is the comparative collection of the Osteological Research Laboratory at Stockholm University. Charcoal Analysis Charcoal from hearths was analyzed by Dr. hiiogen Poole in association with the National Her- barium Nederland, Universiteit Utrecht Branch, the Netherlands, and by Mr. David Black, M.A. University of Western Michigan. Macrofossils Macrofossils were analyzed by Karin Viklund, Environmental Archaeology Laboratory at Umea University. Radiocarbon Dates Radiocarbon and AMS dates were processed at three laboratories: Stockholm (St), Uppsala (Ua) and Beta-analytic (Beta). Dates were calibrated using OxCal 3.10 and 4.0. The uncalibrated dates are listed together with calibrated dates (uncal/cal). All dates are listed in Appendix 2. Metallurgy Analysis of slag and residues from hearths relating to iron working was conducted by the Geo- archaeology Laboratory (GAL) of the National Heritage Board, under the direction of Dr. Eva Hjarthner-Holdar. Analysis was carried out by Eva Grandin, Eva Hjarthner-Holdar, Emma Gron- berg and David Andersson. Shoreline Displacement The starting points for uplift curves are old watermarks that had been carved by students of Carl von Linne (Linnaeus) in the 1700s (Broadbent 1979:199-201; Nordlund 2001). The theory at that time was that the sea was evaporating, not that the land was rising. It was believed that Earth goes through three stages: one of flooding (the Biblical Flood), the intermediate stage we are experienc- ing today, followed by conflagration, when all the seas have evaporated. Although these scholars were unaware of the idea of glaciation, their watermarks are an invaluable contribution to geophys- ics and archaeology. Stone Age sites in northern Sweden are found on old beaches that have been uplifted 120 m or more. Settlements are distributed in time at different elevations like steps in a stairwell. Be- cause of the flatness of north Swedish topography, however, several meters of difference between site elevations can lead to a kilometer or more of horizontal separation between them. Coastal archaeological sites in northern Sweden therefore have little overlap and chronological resolution is exceptional. Increases and decreases in worldwide sea levels due to climate changes have either accelerated the effects of land-rise by lowering sea levels or stabilized the shorelines by keeping pace with the uplift. During warm periods, prehistoric settlements and artifacts tended to collect in greater numbers at stabilized levels because of slowdowns in displacement. During cold periods, THE RESEARCH STRATEGY 47 Figure y. Watemtark at Ratan, 40 km north of the Umea. The oldest mark was catved by Anders Chydenius, a student of Carl Linnaeus, in ly^c). An automatic sea level (mariographic ) station is housed in the concrete huildine. the opposite took place and sites and arti- facts were more widely spread across the landscape. These patterns are particularly evident during the Litorina Period when sea level rises were greatest. The shoreline displacement curve for Vasterbotten is based on a watermark from 1749, radiocarbon dates, archaeo- logical finds and pollen zones (Broadbent 1979:204-211). These points were fitted in a least squares curve-fitting analysis using an exponential model simulating crustal rebound. (The physical model is that of a steel spring releasing and contracting in a viscous solution.) The resulting shoreline displacement model for the past 2,000 years has been verified by counting annual lake sediments (varves) in uplifted lake ba- sins in coastal Vasterbotten (Segerstrom and Renberg 1986). This shoreline curve is illustrated in Figure 34 and includes four new points from the current study. The highest rates are in the north near Storkage, and some- what lower rates occur to the south of there. This downward tilt (slowing) toward the south can be seen by the calibrated ra- diocarbon dates from Stora Fjaderagg and Snoan. Estimated sea level fluctuations (based on Miller and Robertsson 1979) are shown as a dashed line. It should be remembered that dates of settlements and place-names using el- evations are maximum possible ages; that is to say, they cannot be older than their contemporary shorelines (have been under Table 3. Calculated rates of shoreline displacement in Vasterbotten County. YEARS (UNCAL/CAL) 5000/6100 B.C. 4000/4900 B.C. 3000/3800 B.C. 2000/2600 B.C. 1000/1450 B.C. 0/-200 A.D. 500/300 A.D. 1000/950 A.D. 1500/1400 AD 1900-Present ELEVATIONS UPLIFT RATES (M) (M/YR) 119 2.9/2.4 92 2.4/1.9 70 2.0/1.7 51 1.8/1.5 35 1.6/1.3 20 1.4/1.1 15 1.2/0.98 9.6 1.1/0.93 4.5 0.98/0.87 0.67 0.92/0.81 48 CHAPTER 4 500 1000 1500 2000 AD Figure Shoreline displacement curvefor Vasterbotten. Dashed line shows eustatic (sea level) Jluctuations. Squares indicate radiocarbon datesfrom Iron Age huts. water). Because of the rapidity of uplift in Vasterbotten there was no extended inundation of sites although some could have been temporarily flooded during storms. Exceptions to this are ship- wrecks, such as the Ava boat in Lovanger (Jansson 1981). Lichen Growth on UpHfted Beaches One of the most daunting challenges of the project has been the fact that most constructions consist of exposed piles of rocks. Only dwellings with hearths could be radiocarbon-dated, and shoreline dates provide only maximum possible ages. This means that labyrinths and other con- structions were essentially undatable. Following a suggestion by ecological botanist Christer Nils- son, I pursued the idea of using lichen growth on stones for dating cultural features on uplifted beaches. The lichenometric method was developed in the 1950s and has been mostly used in moun- tain environments for the dating of glacial moraine (Beschel 1950; Benedict 1967, 1985; Locke et al. 1980; Karlen 1975; Topham 1977). The most often used species is the gray-green crustose group Rhizocarpon geographicum and R. alpicola. Crustose lichens grow symmetrically, and thallas (lichen body) diameters can be used to estimate age once their growth rate has been determined. Thallus diameters can vary on the same specimen because of competition or other impediments to growth, but a maximum diameter where there are no impediments always corresponds to the maximum age of the lichen. Age can be calculated indirectly through lichen colonization on previously dated THE RESEARCH STRATEGY 49 surfaces, such as headstones, and in the Bothnian region on rocks on uphfted shores (Figure 35). Using the previously calculated rates of shoreline displacement and an analysis of headstones at Bygdea cemetery, it was possible to calculate growth rates for this coastal region. Lichen growth proceeds in two phases, a short period of very rapid growth, the so-called great period, during which the surface is colonized, and thereafter a constant rate of linear growth. The rapid coloni- zation rate for this region was de- termined at Bygdea (Broadbent 1987a), and linear growth rates at Grundskatan, Bjuroklubb, Stora Fjaderagg and Snoan (Broadbent 1987c) (Figure 36). The hypothesis behind this approach is that follow- ing colonization of uplifted beaches, lichen sizes will corre- late with their elevations above sea level. For every meter above sea level, the original colonizers will have added approximately a century of growth (Broadbent and Bergqvist 1986; Broadbent 1987c). The largest (oldest) li- chens at every meter above sea level were measured at numerous sites including those mentioned above. Bjuroklubb is, for example, characterized by eight fairly straight exposed cobble terraces extending up to 15 m above sea level. Lichens were measured at 14 levels starting at colonization, which was observed at the 1.28 m level. The largest (oldest) lichens ranged in diameter from 19 mm at the lowest level to 358 mm at 14.29 m above sea level. The correlation coefficient between elevations and maximum diameters was 0.99 at this site, 0.98 at Grundskatan and Stora Fjaderagg, and 0.94 on Snoan Island. On the basis of this data, individual linear regression equations were calculated for each site. The growth equation for estimating lichen age, based on the combined data from Bjuroklubb and Grundskatan (22 points), is: Y (age) = 152 -1- 3.47 x (diameter in mm). This means that a lichen diameter of 100 mm is 152 + 347, or 499 years old. The standard deviation is 31 years, so the range Figure Rhizocarpon geographicum lichens growing on a monument by Lovanger Church and a 33 cm diameter Uchen growing on a beach boulder at an elevation of i4.2g m above sea level at Bjuroklubb. 50 CHAPTER 4 m.a.s.l. u 12 10 • • • • A.D. soo 700 900 1100 1300 1500 1700 1900 10 15 20 25 30 35 uo cm Figure }6. Maximum diameters ofR. geographiciim at elevations up to 35 m above sea level on seven beaches between latitudes 6]°N-6fN (uplift between S.^-Q.i mm year). Linear growth was obsewed at all sites. with one standard deviation is 468-530 years, or A.D. 1478-1540. Equations were calculated for each individual site, but varied only slightly along the Bothnian coast, which indicates that there were minimal micro-environmental differences between locales. A lichen date is a minimum age, which means that an archaeological feature, such as a labyrinth, is at least as old as the lichens growing on it. Its elevation above sea level provides a maximum age, so there is a means for bracketing the feature in time. Lichen growth also makes it possible to determine relative differences in construction periods as some surfaces may have large lichens growing on them, whereas other areas that may have been built later or were disturbed, will have smaller ones. The best approach in using lichenometry is to combine all the chronological information about a site, including the overall find context (cf Broadbent i987d:43-45). Historical data, elevations above sea level, rock-weathering and proximity to other dated features were used to evaluate the results. A Glossary of Archaeological Features A number of characteristic stone constructions have been documented within the project. They were made using dry-wall methods and, although sometimes chinked using pebbles and often in- corporating boulders, neither mortar nor bricks were used to build them. Each site has been desig- nated by a number assigned by the National Heritage Board (Riksantikvariecimbetet or Raa). Because THE RESEARCH STRATEGY 51 of the variety of constructions at each site, these are referred to as "Archaeological Features" or as "Features" (anlaggningar or anl.), and identified by a number or letter corresponding to the order in which they were investigated. Huts Bothnian huts (Swedish tomtningar) are shelters or foundations consisting oflow cobble walls, usu- ally less than 30 cm in height, with cleared and level floor surfaces (Figure 37). Their walls are ca. 0.3-1.0 m in thickness. There can be one or more entrances. The foundations can be open-sided, but most doorways measure less than a meter in width and were often well set, sometimes with sills. Average floor size is 3 x 4 m, although some floors are smaller and some are over 6 m in length and 4 m in width. The foundations usually occur in clusters of two to nine constructions and can be built in rows with shared curved or straight walls. Most are rounded-rectangular in shape, although oval and rectilinear shapes are found at nearly all levels. Secondary features include internal chambers with rounded walls, small cairns built into walls, small well-built pits in floors, and external wall lines that could have served as parts of enclosures. There is no certain evidence of internal roof supports, such as postholes or post supports, and it can be assumed that the roofs were supported by wooden frames embedded in or braced against the cobble walls. Most huts had central hearths, and these normally measure i m in diameter and are recognizable from traces of charcoal, fire-cracked rocks and burned bones. Most lack stone rings, although many have one or several larger stones by them. Hearths in dwellings were possibly bordered by wooden frames that are no longer preserved. Hearth deposits are shallow, 10-20 cm in thickness, but can have lenses indicating multiple uses. Other bowl-shaped hearths associated with iron working were found near the rear walls of huts. Hearths are also found adjacent to and in front of huts and are often identifi- able by vegetation and lichen growth on the otherwise sterile cobble surfaces. Figure jy. Hutfoundation at the Grundskatan site (Site 78, Hut jj. 52 CHAPTER 4 Cairns Cairns can be roughly divided into stone constructions of less than i meter in diameter and those that are greater in size, usually 3 m to 4 m in diameter. The largest cairns can be graves, sea markers or food storage caches. Cairns can occur singly or, in the case of the smaller cairns, in tight groups or in lines. These small cairns frequently occur as "fields" of post-supports by fishing sites and har- bors and were used in connection with the drying and repairing of nets (Swedish gistgdrd) (refer Figure 112). Caches Caches are storage places for perishable items and lack any traces of burning or burned bones. They consist of cairns or subsurface stone chambers of different sizes. They often occur near, or even in, huts and are often found in open boul- der fields. They were frequently built next to boulders and bedrock outcrops (Figure 38). Figure ]8. Small cache beside a hut on Stora Fjaderiigg Island (Sites 31-32)- Figure }g. Stone alignment at the Grundskatan site. Alignments Alignments are lines of stones that are not dwelling constructions or net-drying post supports. Some are probably associated with en- closures for livestock. These kinds of alignments are documented elsewhere in Scandinavia (stenstrdngar) (Lindqvist 1968; Myhre 1972; Carlsson 1979). Other alignments occur in long parallel lines with spaced openings and were probably used for snaring birds, as seems to have been the case at the site of Horn- slandsudde (Figure 144). Some parallel stone align- ments are set at right angles to former shorelines and mark boat-landing places. Circles These features have been described regarding Saami sacrificial sites (Vorren 1985, 1987; Vorren and Erik- sen 1993; Wennstedt Edvinger and Broabent 2006). They consist of a single circle, concentric rings of stones or a wall enclosure, and often have a central cairn. Single circles can also be tent rings, distin- guishable by their hearths and entrances (refer to Figures 109, 187). THE RESEARCH STRATEGY 53 Labyrinths These are spirahng stone construc- tions made up of single hues of stones forming walkways toward a center point (Figure 40). Bothnian labyrinths normally emanate from a central cross (refer to Chapter 10). They are most often associated with fishing sites (Kraft 1977). Compass Roses These elegant features are small stone settings with eight arms and N-S and E-W alignments. They were built to af- ford compass bearings on land and were probably practical ways to orient fishing and sailing activities (Figure 41). "Russian" Ovens These features are free-standing stone chambers with openings toward the front. They sometimes have a lintel stone. During the eighteenth and nine- teenth centuries, they were used by the Russian Navy for baking bread on shore, and are therefore frequently re- ferred to as Russian Ovens (Figure 42). Shooting Blinds These blinds (Swedish jaktskaror or gomslen) consist of short lines or piles of stones on beaches or by ponds where hunters could hide themselves. They are associated with firearms and the shooting of seals or birds. Engravings Engravings have been recorded at nu- merous places and consist of water- marks carved during the eighteenth and nineteenth centuries as well as names, dates and ownership marks 54 Figure 41. Compass rose on Sudan Island. Figure 42. Large intact Russian Oven in Osterbotten, Finland. CHAPTER 4 carved by seal hunters and fishermen. A large number of engravings dating between 1797 and 1915 were docu- mented at Svarthallviken on Bjuron in Vasterbotten (Figure 44). Figure 43. Watermark at Ratan in Vasterbotten carved during the reign of Gustav III. Figure 44. Engraved names, ownership marks and dates carved by sealersfrom Osterbotten in Finland at Svarthallviken near Bjurdklubb. THE RESEARCH STRATEGY 55 Figure 45. Aerial view of Stor-Rebben Island in the Pitea archipelago, Norrbotten County. Locations of investigated huts indicated. Excavations and Analyses Regional History The main region of the study is Skelleftea Municipahty in Vasterbotten County and Skelleftea and Lovanger parishes. The name "Vasterbotten" dates to 1454 and refers to the western (vaster) side of the basin (botten) of the Bay of Bothnia (Agren 1969:45) (Figure 47). Coastal Vasterbotten is charac- terized by a plain with elevations of less than 50 m with some undulating hilly country of 50 to 100 m elevations. The area south of the Skellefte River, especially the Lovanger region, has numerous old fault lines forming more rectilinear relief with low fjords and narrow bays. In the Telje statutes of September 5, 1328, Knut Jonsson wrote that the northernmost parts of Helsingland up to the Ule River in Finland should be settled and cultivated, but that the Lapps should not be prevented from hunting there. The first official mention of the "Lappmarks" (Lap- lands) is from 1340. King Magnus Eriksson's government declared that this area was bona vacantia (wilderness), "not known to be occupied by many people," and was now subject to the Helsinge Law. The land was made available for settlement to Christian folk and to those who would convert, and land-title would be granted to them (Sommarstrom 1966). The eastern boundaries of Lapland in Vasterbotten, about 100 km inland, were established as late as 1752. This boundary was strength- ened in 1865 as the so-called Agricultural Limit, intended to separate Saami territories from Swed- ish and Finnish agrarian expansion. The use of "Lapland" as the demarcation of Sdpmi is therefore misleading for archeological as well as linguistic purposes. Lapland, as shown on maps in most sources, is best viewed as state-mandated territory. Johannes Schefferus stated that although Lappish (Saami) territory did not in his time en- compass the Bothnian coasts, it did so before the time of Damianus a Goes (1502 -1574). He also quotes Olaus Petri Niurenius, a priest and rector in Umea in Vasterbotten (1580-1645), who stated: "the Lapps formerly had their camps on the Bothnian coast but had been driven away from there" (Schefferus 1673:50). Swedish colonization of the northern coasts was almost certainly underway in the late thir- teenth century and by 1316 Uppsala Cathedral had claimed ownership of salmon fisheries on the Ume River (Olofsson 1962). This is about the same time that Norwegian settlement expanded in Finnmark (1307) and Sweden tried, unsuccessfully, to gain control of these territories as well (cf Odner 1983; Urbariczyk 1992). 57 The vagueness of the new boundary with Russia through the Treaty of Oreshek/Noteborg in 1323 made it imperative for Sweden to occupy the north Bothnian coastlands. These northern peoples were aheady taxed differently than Swedish settlers to the south, in Medel- pad and Angermanland. Under the Halsingelagen, the old Pro- vincial legal system, Swedish peasants were expected to pay taxes and support the ledung, the ship-based defense system. People living to the north of Angermanland were, by con- trast, to pay two skins for every bow (a taxable adult who could span a bow and thereby hunt), and were not expected to contribute to the ledung. In addition to this policy, the Birkarlar, Bothnian merchants of Finnish extraction, were given special privileges to tax the Saami on behalf of the Crown, especially the fur trade (Steckzen 1964:119-128). King Magnus Eriksson's Municipal Law furthermore designated Stockholm as the market for all sur- pluses, which effectively put a lid on independent mercantile activities in the North. The city of Stockholm was fully established by Birger Jarl in about 1300 and thereafter became the capital of Swedish commerce. The Saami in interior Vasterbotten, speakers of the Ume Saami language, were evidently taxed differently than other Saami under the Birkarl system. According to the Lundii Adiscriptio Lapponiae (1670), the Uhmalappar (Ume Lapps) and the Narrlappar (North Lapps) were quite dif- ferent. Fjellstrom (1987) postulated that taxation of the Ume Saami was more directly tied to the Swedish Crown because they were konungslappar (the King's Lapps). This suggests that they had been "appropriated" by the Crown at a very early stage. They had probably already been forced out of the coastal areas which were so important for medieval maritime communications, fishing and trade. This appropriation of territory was logically paralleled by a special tax burden, both as a part of the process of displacement and a means for controlling future settlement and economic activ- ity. Colonization continued throughout the fifteenth century, and by 1519 Bjuroklubb had become a major center for the Swedish herring fisheries (Magnus 1555; Olofsson 1962). Eight medieval churches (Umea, Bygdea, Lovanger, Skelleftea, Pitea, Lulea, Kalix and Tornea) were established. Of these, only Bygdea and Lovanger were not situated by larger rivers. Skelleftea and Lovanger parishes were first established under the Archbishop in Uppsala, Jakob Ulfsson, in 1340 {Skellopt cum capella Lavanger). The present stone church at Lovanger dates in all probability to 1507 and was dedicated to Saint Anna, but the church town could be much older (Figures 47, 194). Historical sources indicate that there had been competition for the church Figure 46. The Ume and Pile Lappmarks and Skellejte and Lovanger parishes in the Province of Vasterbotten in the fifteenth century. The Ume and Pile designations refer to the two Saami languages that align with these rivers. Inset from Figure 7. 58 CHAPTER 5 Figure 47. Lovanger church town. Each cabin along this street belonged to a householdfrom the same village, and the street points in the direction of the village. The church town was a microcosm of regional demography. site between the villages of Lovanger and Mangbyn (Hedquist 1949:276-277). The area of Mang- byn/Broange is interesting because it contains archaeological remains dating to the Late Iron Age (Broadbent and Rathje 2001), and church sites of these kinds usually piggy-backed on already es- tablished market places. Bjuroklubb and Bjuron Bjuroklubb point on Bjuron, at latitude 64° 28' N, longitude 21° 34' E, rises 47 m above sea level and is best known from an account published by Olaus Magnus in 1555 (Figure 48). He had vis- ited the area in 1519 and described the point as "a cliff in the sea, of local people called Bjuraklubb, whose high prominent crown appears from a distance to sail- ors as consisting of a crown with three points." He goes on to describe this vision as lead- ing sailors to safe harbor. Once closer, one could observe great quantities of fish drying on the black rocks along the shore [Swedish Svarthallorna]. These fish, according to Magnus, were consumed locally, traded for ce- reals since local crops rarely rip- ened because of the cold, or "as delicacies for those who lived in Figure 48. Woodcut of Bjuroklubbfrom 2555. The image shows herring drying on the roclcs at Svarthdllviken, the rocky point with three crowns, a medieval ship and seals in the Bay of Bothnia (Oluus Magnus 155^, Book 11:88). EXCAVATIONS AND ANALYSES 59 the interior, and through an exchange for fish one could obtain the riches of the forest in the form of expensive furs" (Book II: 88-89). Much of the Bjuroklubb point consists of exposed granite, gneiss bedrock and wave-washed moraine beaches. The soils consist of shallow gravel and sandy sediments with weak podsolic profiles although in some areas on Bjuron large sand dunes are found. Vegetation consists primarily of pine heaths with dry blueberry (vaccinium) type ground cover and lichen vegetation is abundant, especially reindeer lichens. Birch, alder and spruce woods predominate on marshy and low-lying ground. The area is a nature reserve and there are rare plants of mountain ^'ail type. The name Bjuroklubb, which literally means "beaver island point," derives from the name of this former island (and beaver hunting areas) and the historic harbor and point. The oldest harbor basin on the island, and adjacent to Svarthallorna, is lungfruhamn. Its threshold was determined to be 2.59 m, meaning that it had to have been abandoned by 1656 (Broadbent i989b:26). The harbor and chapel at Bjuroklubb dates to 1658 and was thereafter used into the twentieth century (Figure 32). Even the name Jungfruhamn was probably secondary, however, and derives from what may have been the original name of Bjuroklubb, which was jungfrun (The Virgin) (Wennstedt 1988:25). The lungfru name was often used to describe dangerous coasts and rocks and was probably a taboo word referring to the female demons that caused shipwrecks and other mischief The name also re- ferred to the narrow vaginal shape of Jungfruhamn harbor and was an obscene reference, according to one source quoted by Wennstedt (1988:25). The name Jungfrun was, in any case, to be avoided in everyday conversation, as was the case in many taboos, and was therefore probably replaced with the neutral administrative name it has today, Bjuroklubb (Wennstedt 1988:26). Against this background, the so-called Jungfrugraven (The Virgin's Grave), two boulder walls enclosing a cairn, becomes interesting as a potentially pre-Christian construction. Although the lichens on its stones date it to A.D. 1532-1604 at the latest, its elevation (10 m) associates it with the Viking Period. We have interpreted this feature as a sacrificial site and contemporary with other cir- cular sacrificial sites on Bjuron that are discussed in Chapter 8 (cf Wennstedt Edvinger and Broad- bent 2006). The name Jungfru suggests the place was menacing. This was perhaps due in equal measure to the dangers for mariners as to the many highly visible pre-Christian constructions in the area. Excavations in the vicinity of Bjuroklubb were conducted at Site 67 (Bjuroklubb), Site 70, Site 78 (Grundskatan), Sites 138 and 139 (near Jvmgfruhamn) and Site 144 on Lappsandberget. Soil samples were additionally collected at Site 79 (Jungfrugraven) and mapping was carried out at Sites 64, 65 and 68 (Figures 49, 50). Jungfrugraven, Site 79 This construction consists of a ca. 13 m wide and 17.5 m long oval enclosure encompassing a stone cairn (Figures 51, 52). The walls are 1-1.5 m wide and 0.75 m high. One straight wall runs parallel with the inlet and a curved wall 21 m in length connects with it, leaving i m wide entrances at its south and north ends. A 7 x 8 m horseshoe-shaped cairn is situated at the north end ofthe enclosure. A small pile of stones with an upright and engraved cross lies in its center. The largest lichens on the walls and the elevation above sea level bracket this construction to the time period A.D. 950-1604. Phosphate mapping was carried out within and around the enclosure and revealed no enrichment, as would be expected at a fishing harbor. Lichen growth on the central cairn displays two patterns: the innermost area has relatively large lichens, while the outermost lichens are very small (Figure 52). 60 CHAPTER 5 GRUNDSKATAN V Y/JUNGFRUHAMNW LAPPSANDBERGETV BJUROKLLIBB'S Figitre 49. Aerial view of Bjurdkhtbb faciitg south. Sites mentioned in text shown. Figure 50. Map ofsite locales on Bjuron and the 10 m elevation corresponding to ca. A.D. 950. EXCAVATIONS AND ANALYSES Figure 52. Photo of Inga-Maria Midk and Tim Bayliss-Smith in the Jungfrugraven enclosure. This indicates that the horseshoe-shaped cairn is relatively recent, and that the stones in its outer ring had been taken from the center of the cairn. This horseshoe shape is shown in A. F. Ekdahl's documentation from 1827, but the small pile of stones with the engraved cross is not shown in this drawing. Lichen growth confirms that the small cairn and cross had been set up in the early twentieth century. Cemetery enclosures of this type are known from the Medieval Period, for instance on Hol- mon Island. What distinguishes lungfrugraven is its cairn, which was not the foundation of a cha- pel, but something more akin to Saami offer sites in North Norway. These cairns and enclosures are associated with hunting and fishing sites (Vorren and Eriksen 1993). Although the Jungfrugraven enclosure is larger than the Norwegian sites, it co- incides by form, location and chronology to this material and, most significantly, to the cultural context of the Iron Age huts adjacent to it (refer to discussion in Chapter 8). JUNGFRUGRAVEN boulder wall Figure 52. Map showing the wall lines and displaced cairn stones at Jungfrugraven. The original cairn could have had supported an idol. 62 CHAPTER 5 yure 53. Aerial view of Site 64. SITE 65 (37.4m) / 5m \ \ 3S // ill Site 64 Site 64 is a large recon- structed 7 X 5.5 m rectangu- lar hut with walls measuring 90 cm in height and up to 1.5 m in width (Figure 53). The floor measures ca. 5.5 m in length and 4 m in width. A doorway faces southwest toward the mainland. This large dwelling lies higher than Site 65, ca. 17 ni above sea level, and was probably contemporary with Site 67, or A.D. 450-650. Site 65 Site 65 is a cluster of four huts in a well-protected de- pression at 14-15 m above sea level (Figure 54). Three huts lie within 3 m of each other and a fourth hut stands 37.4 m to the south- east on the edge of a steep waterlogged ravine. Three of the huts have hearths that are mostly overgrown. The floors measure 4 x 4.5 m, 4 x 4 m and 3 x 2.8 m respectively. The smallest hut lacks a hearth and may have been used for storage. A second small round foundation measuring 3 x 3 m is located within 35 m of this complex and could have been a "goat hut" (refer to Chapter 7). No excavations were undertaken at this site, but these huts are interpreted as analogous to and contemporary with the hut groups at Grundskatan, Site 78. They probably date to A.D. 700-1000. Figure 54. Sketch map of Site 65 with three huts in a duster, and a fourth hut on the edge ofa ravine 37.4 to the southwest. Site 67 Site 67 is a solitary construction and lies on a 17 m high saddle of land with access to both sides of the island (Figure 55). The hut is kidney shaped and measures 6x5m with a floor area of 4.5 x 2.6 m. A 75 cm wide entranceway runs through its southwest wall. Nine soil phosphate samples were collected from the floor and range from 30 P° to 194 P°, which is relatively high. Small amounts of burned and unburned bones were also obtained. One fragment was identified as a seal humerus. A 3 cm thick deposit of sooty soil was found adjacent to the back (north) wall. The burned area ex- tended into the wall and into an opening that appears to be a chimney or vent ofsome kind (Figures EXCAVATIONS AND ANALYSES 63 SITE 67 'v'vhut \N 17.08->r:^^ Russian Ovens 16.53 5 m Figure 55. Sketch map of Site 6y with associated features. An oval hearth lies about 100 in to the south and downslopefrom these features. 56, 173). This arrangement is unlike most dwell- ings, but has close parallels to huts at Grundska- tan (Hut 11) and at Hornslandsudde (Hut 12). In all three cases, the hearths were associated with iron slag and furnaces. Figure 56. Plan ofexcavation of Hut, Site 67. Finds 22 pieces of slag (207 g) The slag from the Hut 67 was analyzed by the Geoarkeologiska Laboratoriet. (GAL) in Uppsala and determined to be the by-products of forging (Grandin et al. 2005). This mate- rial is discussed in detail in Chapter 7. One radiocarbon date was obtained from this site: (i485±7o B.P.), which calibrates to A.D. 467- 648. The structure stood just under the top of the rise and had some protection from westerly winds, but its elevation also provided a good draft for a furnace. Phosphate enrichment and seal bone suggests that this was also a dwelling. This hut was probably associated with nearby dwellings, Sites 65 and 68, but these have not been dated. The age of this site makes it contemporary with the oldest phase of the Grundskatan settlement on the opposite end of Bjuron. Three additional fea- tures were recorded within 15 m of the hut. A small cairn, possibly a cache, lies on a rock outcrop, and two stone constructions, one fairly recognizable as a Russian Oven, stand parallel with each other below the outcrop. Both ovens have traces of charcoal in them and can be assumed to date to the Russian invasions of the early 1700s. A small hearth was also discovered partly exposed in the pathway leading up to Site 67 (Figure 57). It is roughly oval in shape and consists of a selection of stones 20-30 cm in size, with one larger flat stone measuring 50 cm across. The surrounding sandy soil was stone free and Table 4. Hut 67: Osteological finds. ANATOMY SEAL? UNIDENTIFIED TOTAL Humerus 1 1 Ossa longa 4 4 Indeterminate 7 7 Total 1 11 12 64 CHAPTER 5 Figure 57. Plan ofhearth excavated near Site 67. Figure 58. Sketch map ofhut row. Site 68. red-burned. Some 6.8 g of burned and unburned bones were found, and a radiocarbon date of the charcoal (Beta 191232) rendered an age of: 2}o±40 (A.D. cal. 1641-1953). The bone is from a large ungulate, probably a reindeer, suggesting the hearth was connected with Saami in the area, as is noted in historic accounts of Bjuroklubb (cf. Wennstedt 1988). Site 68 This site is in an area of completely exposed wave-washed moraine beaches lying to the east of the previous site. A row of five disturbed floors with six dividing walls lies on a terrace at the 14 m level (Figure 58). Traces of four, possibly five, hearths are indicated by fire-cracked rocks. One hut has a small storage pit that is identical to that observed in a group of row houses at Grundskatan Site 67, Hut I. As the row huts at both sites are similar in appearance, and are at the same elevations above sea level, it can be assumed that they were contemporary. The radiocarbon date from Hut i at Grundskatan is: ii6o±7o B.P. or A.D. cal. 779-968. SITE 70 14.30 13.30 10m Figure 59. Map of Site 70 and three adjacent huts and wall alignments. Site 70 Site 70, at 64° 27' N, 21° 35' E, consists ofa group of 10 hut foun- dations between ca. 10 and 13 m above sea level (Figure 59). This EXCAVATIONS AND ANALYSES 65 complex was situated on the north shore of Bjuron Island and faces Bjuroklubb. A charcoal sample was obtained from a hearth in the middle of the complex, as well as 4.6 g of seal bone. The charcoal was dated to: io20±6o B.P. (A.D. cal. 902-1149) (Beta-196485) and this is consistent with the elevation of the site. Foundations of three huts with adjoining wall alignments were mapped in an area near Site 70 and at a somewhat higher elevation (Figure 170). These features had not been previously registered and were very overgrown, so much so they could barely be relocated in 2005. There are two adjoined huts with floors measuring 3x4m and 3 x 3.5 m, and a third simpler construction measuring 3 x 3.5 m. What makes these huts especially interesting are the stone alignments/walls that measure up to 10 m in length. Based on their elevation and proximity to Site 70, these features date sometime between A. D. 900 and 1150. Jungfruhamn, Sites 138-139 Huts A, B and C Two huts (Hut A and Hut B) were sampled at Site 138 (Figures 60-61). Their elevation above sea level is 15 m. A third hut (C) was shovel tested, but not excavated. It lies in the woods some 20 m away (Site 139) and radiocarbon-dated to: i7io±i25 (A.D. cal. 139-526) (St. 11909). Hut A is rectangular and measures ca. 7.5 x 5.5 m. The hut has two entrances on the opposite walls facing north and south. The floor area measures 5 x 3 m. A i m^ test pit was excavated to sample the hearth. A diagonal profile was drawn through the pit showing a layer of 10-20 cm burned soil with bone and charcoal (Figure 62). Two dates were obtained: (St. 11176) 985+70 B.P. (A.D. cal. 990-1155), and (Beta 196490) i2io±50 B. P. (A.D. cal. 722-887). Some 58.8 g of burned bone was found in the hearth. Hut B is rectangu- lar and measures 4 x 5.5 m. The floor area measures ca. 3 x 3.5 m. A 75 cm wide entrance opens I • 0-99 • 100-119 (8) 120-149 #>150 PHOSPHATE Figure 60. Phosphate sampUng of Site 1^8 conducted by Johan Linderholm. Values above 120 P° were obtained around Hut B. 66 CHAPTER 5 Figure 61. Detailed maps of Huts A (right) and B (left). toward the south/southwest and faces the inlet. A i pit was excavated to sample the hearth and one radiocar- bon date was obtained: (St. 11177) i300±i30 B.P. (A.D. cal. 636-886). Finds White (burned) flint chip (15mm) Gray flint chip (10 mm) Red brown flint chip (10 mm) 2 soapstone/asbestos (?) slivers (45 mm) All the bones were burned and were identified as ringed seal, undifferentiated seal, sheep/goat, a large ungulate, and a bird of uniden- tified species. Most of the material was found in Hut B, which was also more varied compared to Hut A. Table 5. Site 138: Species by weight (g). Table 6. Site 138: Species by fragment. SPECIES HUT A HUT B TOTAL SPECIES HUT A HUT B TOTAL Ringed seal 0.18 0.18 Ringed seal 1 1 Seal 4.43 0.12 4.55 Seal 3 1 4 Sheep/goat 0.53 0.53 Sheep/goat 1 1 Large ungulate 2.31 2.31 Large ungulate 2 2 Bird 0.52 0.52 Bird 9 9 Indeterminate 53.92 133.58 187.5 Indeterminate 436 903 1339 Total 58.88 136.71 195.59 Total 440 916 1356 A hearth N Figure 62. Profile drawn diagonally through Hut A hearth. Osteological Material The bone material from the two hearths weighs 195.59 §• EXCAVATIONS AND ANALYSES 67 The anatomical breakdown of this material is difficult because of fragmentation and low number of species. It should be noted, however, that there were bones of the cranium, tibia and radius in Hut A, bones that are uncommon at, for example, Grundskatan. The radius is from a large adult seal. In addition, there is also a fragment of a tibia from a goat/sheep one to two years of age. Among the indeterminate bones, there is a dominance of long bones and vertebral frag- ments, probably from seals. In Hut B, a femoral fragment from a ringed seal and a seal phalanx were identified, as well as a fragment of a large ungulate shoulder blade and humerus. These appear to have been from the same animal (or cuts of meat), either an adult cow/moose or rein- deer. The indeterminate bones appear to have mostly derived from seals, although some of the heavier long bones probably emanated from a large ungulate. There are also numerous small bones from birds. Table 7. Site 138, Hut A: Anatomical distribution of burned bone fragments from seals. ANATOMY Cranium Radius Tibia Total SEAL Table 8. Site 138, Hut B: Anatomical distribution of burned bone fragments. ANATOMY Scapula Humerus Femur Ph3 post Ossa longa Total RINGED SEAL SEAL LARGE UNGULATE BIRD TOTAL 1 1 1 1 9 13 Lappsandberget, Site 144 The goal of the investigation of Site 144 was to document a cir- cular stone feature on Lappsand- berget found during survey in the late 1980s and later regis- tered by the county. This rocky hill rises up to ca. 25 m above sea level. Excavation involved removal of vegetation and ex- posure of the circle and a 13 m^ area outside of the circle (Fig- ures 63, 64). The soil was sandy and barely covered the bedrock. Darker brown soil was observed in four patches within this circle. Twenty soil samples were taken. LAPPSANDBERGET 0.30 lichen (8cm) 0.23 0.40 lichen ( 1 1cm excavation 1 m -0.15 Figure 63. Map of Site J44 and excavated areas. 68 CHAPTER 5 Figure 64. Exposed surface within the stone circle. (Note depression due to a pbmdering attempt.) 13 from inside the circle and 6 from outside the circle. Three additional samples were taken from the dark soil deposits. A metal detector was also used but revealed no metal debris. The stone circle measures ca. 2.70 x 2.70 m and consists of some 50 stones 20-45 size. Four- teen stones of comparable sizes were found within the circle. A 70-80 cm wide and ca. 10 cm deep depression was observed in the center of the circle and was likely to have been the result of a plundering attempt that pushed these stones aside. Three li- chens of Rhizocarpon geographi- cum measuring 80-110 mm in diameter were observed on two in situ stones in the circle, and on a disturbed stone within the circle (Figure 65). These lichens date to A.D. 1480-1583 and pro- vide a minimum date for the feature. The elevation above sea level (-25 m) is equivalent to ca. 300 B.C., but this locale was undoubtedly chosen because it overlooked the settlements and the sea. This feature does not appear to have been a grave and no traces of charcoal or bones were found in it. The brown soil deposits represent some kind of organic enrichment, and the phosphorus samples support this conclusion. The surrounding soil is very low in phosphorus by comparison. Nitrogen levels were high from within the circle, and this enrichment can potentially derive from organic sources such as blood, flesh and bone (Figure 66) The place-name itself suggests a Saami context, as do oral-historical accounts of Saami living on Bjuron (Wennstedt 1988), as well as the oval hearth of Saami type near Site 67 on Bjuroklubb. The age of the feature is most probably within the range of A.D. 1000-1600. Figure 65. Close-up ofone of three lichens growing on the stone circle. EXCAVATIONS AND ANALYSES 69 GRUNDSKATAN Ox o o ' c^ i9 C - -14- ' ^ - 13--" , -.12 -- ' 14. ^ 13 12 11 10 ~8. "7, ~" 5 - 100 m Figure 68. Map showing archaeological features at the Grundskatan site. A. area with numerous hunting blinds and the Russian Oven (#io). B. location ofthe labyrinth ("#14). C, location of the bear burial (#4). D, location of the circular feature (#17). E, location of the large pit (#1^). F, location ofHut 2. Figure 69. Aerial view ofthe Grundskatan site with huts and hearths clearly visible at elevations between 12 m to 14 m above present sea level. A labyrinth can be seen near the center of the photo. EXCAVATIONS AND ANALYSES Grundskatan, Site 78 Grundskatan (Grundskataraften) lies on Bjuron at 64° 28' N, 21° 35' E. The site was described as early as 1882 but was first mapped in the 1940s by Carl Holm and Gustaf Hallstrom (Hallstrom 1942:250-257). There are more than 40 hut foundations, (Swedish tomtningar) , cairns and other features, as well as a stone labyrinth and a Russian Oven (Figures 67, 68, 69). The site is situated on the northern half of a NNW-SSE oriented drumlin. The moraine has been wave-washed and consists of gravel and boulders mostly less than i.o m in diameter. The exposed beach toward the east has nine terrace formations up to 16.5 m above sea level. There is boggy and water-logged land north of the site. Grundskatan once formed a small peninsula when the sea level 10 m higher and was an ideal base for sealing. Feature 1, Hut This hut is one of nine partly overgrown dwellings aligned in a single row between 13 m and 14 m above sea level (Figure 70). Ten cobble walls, 0.5-1.0 m in width, separate the floors. The huts were built so that a beach ridge forms the major portion of the back wall. Three huts have northeast- facing walls with door openings that measure 0.5-1.0 m in width. Five of the floors have central hearths. Hut i was in the middle of the row, was the most intact of these huts, and was therefore chosen for excavation. The floor area was completely overgrown by mosses, lichens, grass and brush. The floor measures 4.0 x 3.0 m and the hearth measures i.o m in diameter. The hearth was a round ashy deposit with a lens-shaped cross section up to 10 cm thick. There were no larger stones around it or beneath it, but there were a number of fire-cracked stones in it. A small ca. 20 cm wide and 15 cm deep cylindrical pit was found in the southwestern corner of the floor near the back wall. This straight-sided pit was tightly packed with smaller pebbles. An identical pit was observed in a Figure 70. (top) Map showing Hut 1 and profile excavated through the floor. Jltis hut is one of approximately 9 built up against the same beach ridge, (bottom) Cross section of the hut and beach ridse. BEACH RIDGE 13.19 hearth 13»5 BEACH RIDGE Storage pit hearth entrance 1 m 72 CHAPTER 5 row of disturbed huts at Bjuroklubb Site 68 at about the same elevation. It had probably been for storage of some kind, but was completely sterile. Finds I gray flint chip Chronology One charcoal sample from the hearth was radiocarbon dated to: 1160+70 B.P. (St. 10787). Calibrated date: A.D. cal. 779-968. Table 9. Grundskatan, Hut 1: Anatomical distribution of bones. ANATOMY SEAL LARGE UNGULATE BIRD TOTAL Cranium V caud McV Coxae Talus? Mt Phi post Ossa I Total Table 10. Grundskatan, Hut 1: 2 2 1 3 2 1 1 2 12 lacrofossils. CARBONIZED Chenopodium (goosefoot) Stellearia graminea (lesser stitchwort) Rubus idaeus (raspberry) /\rcfostap/iy/os iva-ursi (bearberry) Empetrum (crowberry) Picea (spruce) Vaccinium (blueberry) Osteological Material Seal bone {Phoca sp.) dominated and included parts of the vertebrae, back- bone, front and rear flippers and frag- ments that probably derived from the skull. One foot bone (a talus) from the rear flipper of an adult seal, two large ungulate heel bones (cow, moose, or reindeer), and two bird bones (Aves sp.) were also found. 336 bone frag- ments could not be identified. Macrofossil Analysis Seven seeds were identified, four of which were berries, and one of cheno- podium, an edible plant. Their pres- ence in the hearth suggest summer or fall. Crowberry and blueberry seeds were also foimd in the Hut 2 hearth (below). NON-CARBONIZED Feature 2, Hut Feature 2 consists of a totally overgrown hut with a floor area measurmg 4.0 x 4.0 ' m. Its elevation is 16 m above sea level (Figure 71). This hut is one in a row of three huts on the opposite end of the point and facing west. Two inwardly curving cobble walls 0.5 to i.o m in width give the hut a horseshoe shape with a southwesterly facing entrance. These structures had also been built up against a beach ridge. The floor is a natural pebble surface that had been cleared of larger stones. An area of ashy EXCAVATIONS AND ANALYSES 73 HUT 2 74 CHAPTER 5 soil measuring 1.7 m in diameter lay midway between the rear of the hut and its entranceway. Two additional floor depressions of approximately the same size lie parallel to the hut. Not enough un- Table 11. Grundskatan, Hut 2: Macrofossils. CARBONIZED NON-CARBONIZED Arctostaphylos iva-ursi (bearberry) 1 3 Vaccinium (blueberry) 3 contaminated charcoal was obtained for a date, but the form suggests these huts were contemporary with Hut I. Macrofossil Analysis. Two types of berries were flotated from the hearth: bearberry and blueberry. Feature J, Hut Feature 3 lies on level ground at ca. 16 m above sea level. It is a rounded rectangular cobble founda- tion with an outer measurement of 7 x 7 m and a floor area of 4 x 4 m (Figure 72). The foundation is i.o to 2.0 meters wide and up to 30 cm in height. The hut has a clearly marked entranceway in Figure 72. Map of Hut 3 with section ofa small posthole near the center of the hut. EXCAVATIONS AND ANALYSES 75 the southeast wall. At the rear of the hut is a 2.0 to 3.0 meter wide and relatively level layer of stones that may have supported a platform. The hearth lies in the center of the main floor in front of this platform as an irregular burned surface measuring 1.4 x 1.6 m. These deposits were i to 6 cm in thickness. There were no stones around the hearth, but several large stones look like they might have served as seats or served as "tables." One possible posthole was found between the hearth and the platform foundation, but was not large enough to have been a roof support. Finds Two white calcified flint chips (less than 5 mm) One pebble (whetstone.^) 23 pieces of iron slag (1-30 mm), 80 g. Table 12. Grundskatan, Hut 3: Bones (fragments). SPECIES HEARTH FLOOR TOTAL Ringed seal 1 4 5 Seal 65 64 129 Hare 1 1 Bird 1 1 Indeterminate 327 65 392 Total 393 135 528 Osteological Material More than 500 fragments of burned bone were found in the hearth and on the floor that derived from ringed seal {Phoca hispida), seal {Phoca sp.) and hare (Lepiis sp.). All of the skull bones were found on the floor together with most of the bones from the front flippers. Other bones from the front flippers and rear flippers, wrist, etc. were from the hearth. Chronology Two radiocarbon dates were obtained from the hearth: i50o±ioo B.P. (St. 11907) and i205±70 B.P. (St. 11906). The calibrated dates are: A.D. cal. 435-643, 695-894. Charcoal Pinus sp. (pine) Table 13. Grundskatan, Hut 3: Anatomical representation for ringed- and indeterminate seals on the floor and in the hearth, burned bones (fragments). ELEMENT HEARTH FLOOR TOTAL Cranium 1 28 29 Dentes 1 1 V caud 1 1 2 Gestae 3 3 Radius 1 Cr+i 3 4 C2 1 C3 1 1 Mcl 2 Moll 1 MclV 1 Mc 6 Phi ant 2 9 Ph2 ant 11 11 Ph3 ant 8 8 Fibula 1 Calcaneus 1 1 Mtl 1 Mtll 1 1 Mt 2 1 3 Mp 2 2 Phi post 20 3 23 Ph2 post 5 5 Ph3 post 5 2 7 Ph post 9 9 Sesamoidea 1 1 Total 66 68 134 76 CHAPTER 5 GRUNDSKATAN r 16 ® 07 / ..P ; 13g -":._ 11 14 «2)o) 17 O /' ; 6 50 m 15.56 turf ^ 1 g carbon m 16.07 (\J\x 15.71 [y] gravel hut wall hearth hut wall Figure 73. Map 0/ tlie central area of the Grundskatan site with feature numbers. Figure 74. Map and profile ofHut 4. Feature 4, Hut (Bear Burial) Hut 4 is situated at ca. 16 m above sea level and clusters together with three nearby dwellings and other foundations at the same elevation (Figures 73, 74). The outside mea- surements are 7.7 x 7.0 m and the floor area measures 4.0 x 3.0 m. The cobble walls average i.o m in width and 0.30-0.50 m in height. A ca. I m wide entranceway is observable in the north wall of the structure. A second opening is on the opposite (south) wall, but is more irregular and was probably disturbed when a cairn was constructed in the south- east corner of the dwelling. A i.o m diameter central hearth and an ir- regular area of sooty soil that could have been the result of secondary use were exposed on the floor. This hearth was not delimited by stones, although several large stones were found beside it. As will be described EXCAVATIONS AND ANALYSES 77 in more detail regarding the bear burial, the cairn that had been built directly on the floor had prob- ably made the hut uninhabitable. Chronology Charcoal was obtained from the hearth and rendered a radiocarbon date: iiio±no B.P. (A.D. cal. 780-1020) (St. 10785). A small indeterminate bone fragment found in a charcoal sample from the floor was AMS-dated. It proved to be older than the hearth, i5oo±40 B.P. (A.D. 536-621) (Beta- 210236), and was contemporary with Hut 3. Soils from under the walls of Hut 4 showed phospho- rus enrichment in conjunction with several fire-cracked rocks, indicating the existence of cultural deposits that predated this hut foundation. Test pits between Huts 3, 6 and 4 also produced fire- cracked rocks. Hut 3 was thus both partly older than and contemporary with Hut 4. Two char- coal samples that had been collected in Table 14. Grundskatan, Hut 4: Identified species based on numbers of fragments (NISP) and weigfits (g) for burned bones in heartfi. 1987 were dated in 2006: (Beta-196486) i90±40 B.P. and (Beta 196487) 420±B.P, but were almost certainly contaminated. SPECIES ISP WEIGHT (G) Ringed seal Seal Bird Indeterminate 3 42 2 338 1.52 20.37 0.05 38.8 Osteological Material Burned bones from the hearth (385 frag- ments) weighing 60.74 g were identified as ringed seal {Phoca hispida), seal (Phoca sp.) and bird {Aves sp.). Total 385 60.74 78 Figure 75. Photograph ofHut ^ during excavation. Viewfrom west. CHAPTER 5 Figure y6. Map of the cairn and bear bone finds in the southeast comer oj Hut 4. A Cairn A trench was opened across the floor across the hearth and expanded to encompass the cairn (Figures 75, 76). The cairn measured ca. 3 x 3 m and was ca. 15 cm higher than the foundation. It was sectioned and excavated down to sterile gravel. A concentration of bones within an area of ca. i.o X 0.5 m was exposed directly beneath the cairn stones in the southeast corner of the hut. The Bear bone 1080±40BP n^rih lllO^llORp ^^Jfltfi^ aiBQGilAD 500CalAD lOOOCalAD Cahbrateddate I5OOQ1IAD Figure 77. Calibrated dates of the hearth and bear bones in Hut 4. bones were concentrated in a 10 cm thick layer between 15.71 and 15.74 m above sea level. This was approximately 10 cm below the former ground surface. At that time of discovery, the appearance of these bones was so dissimilar to the bones in the hearth, they were judged as being unrelated. They EXCAVATIONS AND ANALYSES 79 Im M H PHOSPHORUS Figure 78. Map 0/southeast comer ofHut 4. Phosphorus enrichment in the area of the hear bone deposition and under the hut walls. Fire-cracked stones were alsofound under the walls indicating an older settlement surface. X 1031 V 400 OQ OQoQ ^—. 1 5. hi c ^ o /' / X 100 V 400 wall line Figure 79. Map ofsoutheast corner ofHut 4 showing location of bear bone deposition. were subsequently AMS-dated to: 1080+45 ^-P- (Ua-18930), indicating that the bear bones were contemporary with the hearth. This date cahbrates to A.D. cal. 898-1014. Their median values are: A.D. cal. 912 and A.D. cal. 958. Phosphorus sampling was undertaken in 2005 to additionally delineate the bone deposi- tion. The La Motte system is based on a colorimetric scale indicating "Low," "Medium" or "High" levels of soil enrichment. Based on comparisons with previously analyzed phosphate samples on the site by Johan Linderholm (Broadbent i987b:57), "Low" phosphorus levels correspond to 0-90 P', "Medium" to 100-150 P° and "High" to >i50 P°. The off-site values for the site were low using both methods. The highest measured phosphate enrichment on the site was 209 P° and the aver- age for the huts was 108 P". A sample from the southwest corner of Hut 4 measured 119 P°. The La Motte readings from Hut 4 are uniformly low except for two samples from the exact area where the bear bones had been lying when excavated in 1987 and in two samples from adjacent areas under the walls of the hut (Figures 78, 79, 80). This simple field test has thus rendered consistent information regarding the location of the bone deposition in the hut and has also confirmed that there had been an earlier occupation surface. The whole cairn was subsequently excavated and a tree stump in the foundation was removed. Although the tree had disturbed the outer part of the 80 CHAPTER 5 X 1 00Y400 UNDISTURBED SOIL PROFILE X100Y395 SANDY SOIL AND GRAVEL IINIXCW Al 1 I) X101.80Y400 HUT FLOOR ^esf X10I.80Y395 Jl V' BEAR BONES X1()4Y4()0 I5.<.2 I.. I A XI04Y395 Figure 80. Three profiles through Hut 4. cairn and foundation, the boulders that had been incorporated into the construction prevented any disturbance inside of the dwelhng. The cairn was therefore mostly intact, as was the integrity of the bone find. Bear Bones A total of 66 bones weighing 949 g were found, of which 275 g could be identified as to skeletal element. The morphology and structure of the bone fragments indicate that they originated from the same individual. Judging from the sizes of the largest bones and the thickness of the cortex of the long bone fragments, they derived from an adult bear. Phosphorus enrichment was noted in this corner, as compared with the rest of the hut floor, and this localized buffering of soil acidity probably helped preserve the bones. The soils are otherwise quite acidic with pH values of 4.9 to 5.2. The bones lay in three separate piles surrounded by scattered fragments. In one pile lay parts of the cranium, both halves of the lower jaw, a few fragments of teeth and numerous small frag- ments of long bones. The second pile consisted of three larger long bones, and a third pile consisted of tooth fragments. There were no phalanges. Some of the bones were partly charred, dark colored and heavily fragmented. The tooth fragments also exhibit traces of heating (charring and discolor- ing) and the crowns had been broken due to heat. It is clear from the excavation that the bones had not been burned at the place of deposition. Table 15. Bear (Ursus arctos) bones by weight (g). SPECIES TOTAL Bear Indeterminate Total 275.4 673.68 949.08 Bones and Teeth Although no skeletal elements lay in correct anatomical po- sition, the deposition was not without structured elements. Most bone fragments had been placed without any specific order in the largest heap, but some larger leg bones, 2 tibia EXCAVATIONS AND ANALYSES 81 Table 16. Bear bones by fragment. SPECIES NO. FRAGMENTS WEIGHT (G) Table 17. ANATOMY Bear bones by anatomy. TOTAL Dear do 97R /in IVIdllUIUUId Indeterminate 6* 673.68 Dentes 50 Q/i Q ns .Uo Humerus 1 Radius 1 *all indeterminate fragments not counted Ulna 1 Tibia 2 Long bone fragments 6 Total 66 fragments, a radius, an ulna and a larger indeterminate fragment of a long bone, had been placed next to each other in a separate pile. A number of tooth fragments had been placed in a third concentration. There had been a conscious sorting of the bones. Marks on the bones provide some insights into how the bear carcass had been handled prior to burial. The radius and the ulna exhibit fresh frac- ture patterns, indicating that they had been broken or cracked shortly after the death of the bear when the carcass was in a fresh state. One of the charred fragments exhibits cut marks that show that the carcass had been slaughtered and severed prior to burning. Charcoal (From sooty deposits by the bear bones) Yew (Taxus sp.) Heather (Erica sp.) Birch/Alder (Bctulaceae) Angiosperms (>i3 small twigs) Conifers (5 small twigs and wood remains). The charcoal analysis from the area of the bear grave has pro- duced some unexpected results. While most of the identified plants are typical of the area, the find of yew is totally out of its normal range. Yew grows in southern Sweden and this find could suggest that it had been part of a bow or had ritual mean- ing (refer to Chapter 8). Figure 81. Map ofHut 5. 82 CHAPTER 5 Figure 82. Map ofHut 6. Feature 5, Hut This oval hut was previously un- registered. It lies between 15-16 m above sea level and was built up against a beach ridge (Figure 81). It is one of three structures along the same beach ridge and probably belongs to the same complex of dwellings. It mea- sures 5 X 6 ni with a floor area of 4x4m. Ashy deposits indicate a hearth area, but no bones or charcoal samples were obtained. Finds 2 gray flint chips Feature 6, Hut This dwelling lies 10 m south I of Huts 3 and 4. It consists of a rectangular foundation measur- ing 6x5m with a floor area of 4.5 X 3 m (Figure 82). There is an opening toward the south, that was probably an entranceway, but there was possibly a second entranceway on the eastern short end of the dwelling. There was a small central hearth but there were no intact deposits. A large wall extension runs out of and parallel to the west wall of the dwelling, extending 4 m, and then run- ning parallel with the southern wall for 8 m. This construction is interpreted as an addition to the dwelling, possibly a covered shed or porch. The wall also extends toward the west and could also be part of a stone alignment. Feature y, Hut This foundation is located 5 m SSW of Hut 5 and on the same beach ridge. It mea- sures 4.5 X 5 m with a floor area of 4 x 4 m Figure 83. Map ofHut 7. EXCAVATIONS AND ANALYSES 83 '' 7^^ i 1—.— ' 0S ^—— < i « 1 9 • > . ^ 1 —• — < ® • • « ^ < 1 ( 1 ( —•— ( 1 i • ^ — » — • 0-99 •99-119 ® 120-149 #>150P'' Figure 84. Phosphate map showing enrichment to the southeast ofthe huts and near stone alignments. (Figure 83). There was no hearth but an entranceway toward the SE. It is probably the foundation of a storage shed. Phosphate Mapping Fifty-three phosphate samples were taken at 10 m intervals in an area of 50 x 150 m (7500 m^) (Figure 84). These were analyzed using standard laboratory methods. The samples showed some enrichment in the area between Huts 3, 4 and 6 but displayed the greatest concentrations to the east and south of the dwellings in a 25 m wide band that runs from the area of Feature 8 (a cairn), and around several stone align- ments 20 m to the east of Hut 3. This pattern indicates organic enrichment near the cairns and what may possibly have been a livestock enclosure area. Finds of sheep/goat and other ungulate bones at five hut sites show that animal husbandry was indeed practiced by these seal hunters. Manure would create phosphate concentrations such as those seen at this site (greater than 150 P°) (cf Aronsson 1991). What is curious about this enrichment is that it is not inside a potential enclosure, but around it. This could mean that if a fence had Figure 8^. Map ofFeature , 84 CHAPTER 5 Figure 86. Map of Feature 9. Figure 87. Photo of Russian Oven (Feature 10) at Grundskatan. been supported by these stones it was intended to keep animals out, not in. This could thus have been a temporary holding pen used for marking or milk- ing, or was perhaps a small gar- den plot. Feature 8, Cairn This cairn is one of three that extend in a line to the north of Huts 3-7. The cairn measures ca. 3.5 X 3 m, is ca. 50 cm in height, and has a central cham- ber measuring i x 0.5 m (Figure 85). The opening is well con- structed rather than the result of plundering. There were no finds of burned bones or char- coal. The feature is interpreted as a cache. Feature 9, Cairn This cairn is almost identical to Feature 8. It measures 4 by 4 m and was built of cobbles and larger stones around a boulder (Figure 86). The height of the cairn is 50 cm and its central pit measures 60 by 85 cm. There were no traces of carbon, bone, or other organic residues. It is interpreted as a cache. Feature 10, Baking Oven ("Russian" Oven) A distinctive oven with thick charcoal deposits lies at ca. 10 m above sea level and off to the side of the main site area (Figure 87). The soil was flotated, but revealed little except charcoal. The radio- carbon date (St. 10785) indicates an age of less than 250 years, which would be the 1700s, a period of Russian invasions in Vasterbotten. There are two similar ovens near Site 67, but these were not dated. Two radiocarbon dates were obtained from an oven on Snoan Island, however, both indicat- ing the fifteenth century, Site 92. This means that although these features were probably used for baking, they were not all associated with the Russian period. EXCAVATIONS AND ANALYSES 85 14.10 w A E 0.5m Feature ii, Hut Feature ii is a rectangular hut foundation measuring ca. 7 x 5 m (Figure 88). The eleva- tion above sea level is 14.0 m. The walls of the hut measure i.o m in width and 0.3 m in height. The floor measures 5.0x3.5 m. Azm wide entrance opens toward the southeast. A third parallel cobble wall forms a small addi- tional room. This wall is 4.5 m long and i.o m wide. Another narrow rectangular foundation, possibly a storage shed, lies several meters be- hind Hut II. Hearth deposits were found just beneath the humus and extended 10-20 cm below the surface, and somewhat off-center to- ward the back wall. Unfortunately, a tree had grown in the hearth area and several large roots had penetrated the deposits. Finds 300 g of iron slag The find material consists of both homo- geneous slag and slag with melted stones, red-burned clay and rusted iron. A techni- cal analysis was performed by Grandin et al. (2005). The slag includes varying proportions of wiistite, olivine and glass flux. Fine-grained magnetite was also observed. Drops of metal- lic iron indicate that the iron content had been high. There had also been a good supply of ox- ygen. The slag derived from forging, and iron scales, the result of hammering, were picked up in the hearth using a magnet. A single ra- diocarbon date of charcoal (St. 11170) gave an age determination of ii75± 100 B.P. (A.D cal. 723-972). No animal bones were found. Charcoal Betula sp. (birch) Pinus sp. (pine) Angiosperm Conifer Figure 88. Map ofHut 11 and profiles through the hearth area. Figure 8cj. Map ofHut 12 and profile of the hearth. 86 CHAPTER 5 Feature 12, Hut Feature 12 is a horseshoe-shaped hut foundation measuring ca. 8 x 6.5 m. The elevation is 13 m (Figure 89). The floor area measures 3.5 x 3.5 m. A single i.o m wide entrance faces north. An oval hearth measuring ca. 1x2m had been partially disturbed. One half of the hearth was excavated and rendered animal bone (17.06 g, 391 fragments) and charcoal. Chronology One radiocarbon date was obtained: (St. 11171) 1430+ no B.P. (A.D. cal. 437-760). Finds None Osteological Material None of the bones were identifiable. Hut 13c 12(XW0BP ^*Ji^^.^ Riif Hp im'^IIORP M^^^^^. Hut 13b 88at80BP 1 1 1 1 1 1 1 1 1 CalBaCalAD 500CalAD lOOOCalAD Calibrated date ISOOCalAD Figure 93. Calibrated radiocarbon datesfrom Hut ij. Feature 13, Hut Hut 13 lies at ca. 12 m above sea level and had been built up against a beach ridge. It measures 6.5 X 5 m (Figure 90). A well-marked entrance faces east toward the beach. A well-constructed oval chamber occupies almost one half of the floor space, presumably for storage. Hearth deposits were found in the rear of the structure and three carbon samples were collected: (Beta-198488), i200±40 B.P., (St. 11908), io45±iio B.P. and (St. 11172) 88o±8o B.P. These calibrate respectively to: A.D. cal. 777-884, 880-1155, 1044-1220. EXCAVATIONS AND ANALYSES 87 Artifacts None Feature 14, Hut and Labyrinth Feature 14 is a labyrinth mea- suring 8x9m. The labyrinth has 10 rows of stones and is situ- ated at approximately 13 m above sea level. The entrance to the labyrinth faces north/northwest. The labyrinth overlies a hut wall measuring 6 m across (Figure 92). The exposed interior of the floor area is 4 m across. A 2 x I m trench and a 50 x 50 cm square were excavated across the labyrinth stones in order to ex- pose the hearth beneath it. The hearth deposits extended to 20 cm below surface (Figure 93). 14.03 13.59 profile 13.47 excavation 13.56 / 5 m Figure 92. Map of labyrinth and hut wall at Grundskatan, Feature 14. The labyrinth had been deliberately built on top of the hut and wall stones had been usedfor its construction. 88 CHAPTER 5 Chronology This excavation rendered two radiocarbon dates: (St.11173) ii45±ioo B.P. (A.D. cal. 776-990) and (St. 11174) iooo±i85 B.P. (A.D. cal. 870-1231). This complex feature established the chrono- logical relationship between Iron Age huts and labyrinths in the region. It appears that the stones from the walls were, in fact, used to construct the labyrinth. The hut dates to the same period as the majority of huts at Grundskatan, the Late Iron Age. The labyrinth had a maximum lichen growth of 90-95 mm, giving the feature a minimum age of A.D. 1505 to 1523 ±31, using the formula in Chapter 4. Osteological Material Some 16.04 g of burned bone (205 fragments) were found in the hearth but were unidentifiable. Feature 15, Pit Feature 15 is a very large pit measuring 10 by 12 m across and 2.0 m in depth (Figure 94). The pit was dug at the highest level of the drumlin at 17 m above sea level. A surrounding earth wall aver- ages 50 cm in height. A 4 m long and 50 cm wide trench was excavated through the north side of the wall to investigate its construction. The trench extended down to the former ground level where a thin charcoal layer was found. The radiocarbon date for the charcoal layer is: (St. 11175) 67o± 245 B.P. (A.D. cal. 1033-1467). The median date is A.D. 1295. An older sample was analyzed in 2006 (Beta-196498), 320±40 B.P. (A.D. cal. 1515-1641), but is probably contaminated as were other im- properly stored samples. Although the oldest date has an exceptionally wide range, it does show EXCAVATIONS AND ANALYSES 89 that the pit can be contemporary with the huts. Ap- proximately half way down the inside of the pit there is a narrow ledge or offset. This suggests there could have been a construction, possibly a floor, at that level. There is no side access to this pit which shows it was not a tar-rendering pit. It is an unlikely place to dig a large solitary hunting pit, and is interpreted as most probably associated with seal oil production during the period A.D. 1033-1500 (1641.^). Feature 16, Stone Circle Feature 16 is a 30 cm high cobble oval located ca. 10 m northwest of Hut 3 (Figure 95). It has a single open- ing facing south and a small central cairn. It measures 6 m in length and 3 m in width. The central cairn is approximately 80 cm in diameter. This enigmatic construction bears some resemblance to the so-called Jungfrugraven, located less than 2 km away. Its size and proximity to Huts 3-6 suggests it had a ritual function. The elevation above sea level, 16 m, gives it a maximum age of ca. A.D. 400. In 2006, a previously missed 220 mm diameter specimen of Rhizocarpon geographicum was discovered by Tim Bayliss-Smith on the inner wall of the construction. The stone with the lichen sat securely wedged in the wall and showed no evidence ofhaving been moved since construction. On the basis oflichen growth curves specifically developed for Bjuroklubb and Grundskatan, this lichen is calculated as being 916 years old and dates to A.D. io34±3i (with B.P. = 1950), which is almost identical to the radiocarbon- and AMS-datings of the adjacent bear burial and hearth in Hut 4. Figure 95. Map of Feature 16. 13.00 Figure 96. (left) Photo of Feature ly.froni southwest, (right) Drawing ofFeature 17. 90 CHAPTER 5 Feature ij, Cache and Hut This double feature consists of a 4 m wide cairn with a well-made central chamber about 1.5 m wide and i m deep (Figure 96). It lies at the 13 m level and is adjacent to Huts 13 and 14. It is tightly packed with smaller stones. The cairn was built together with a round wall foundation measuring 4.5 by 5 m, with a 2.4 by 2 m floor area, and a i m wide entranceway. The floor is level and has no traces of a hearth. A similarly built small hut stands near a cluster of4 dwellings about 40 m to the east. Feature 17 is interpreted as a livestock (goat or reindeer) hut and a cache. Summary • Nineteen archaeological features were investigated at Grundskatan: eleven huts, six mis- cellaneous features (four cairns, a labyrinth and several stone alignments), a large pit and a Russian Oven. There is an assortment of hut forms, including round, rectangular and square shapes, structures with internal storage rooms, platforms and porch-like exten- sions. These occur in groups and in rows. • Nine radiocarbon and two AMS dates were obtained from hearths, of which three cali- brated with 95% probability to ca. A. D. 330-700. Seven ofthe dates fall within the eleventh century. • The three oldest dates were obtained from Huts 3, 4 and 12, and the youngest dates were from Huts 13 and 14. There is a spread within individual hearths indicating multiple occu- pations. The medians of the 11 dates range between A.D. cal. 542-1019. Three samples had HLit4e 150Qfe4OBP Hut 3a ISOOtlOOBP Hut 12 1430illOBP Hut 3b 120&=70BP Hut 13c 1200i^BP Hutu 1175tl(]0BP Hutl 1160t70BP Hut 14a 1145tlOOBP Hut 14b lOOOfclSSBP Hut 13a im5±110BP Hut 13b 88at80BP I I I I I I I I lOOOCalBC 500CalBCCalBC/CalAD 500CalAD lOOOCalAD 1500CalAD 2000CalAD Calibrated date Figure 97. Radiocarbon dates from the Grundskatan site. EXCAVATIONS AND ANALYSES 91 probably been contaminated by improper storage over fifteen years at room temperatures (Beta-196486, Beta-196487, Beta-196489). • The date of the large pit (Feature 15) indicates that it was potentially contemporary with the huts. • The Russian Oven radiocarbon dated to the eighteenth century and there are a number of hunting blinds etc. at lower levels on the beach that are not described here, but are also judged as being of relatively recent date. • The labyrinth at Grundskatan (Feature 14) lichen dates to the early sixteenth century. This Late Medieval date is most plausible and borne out by other dated labyrinths in the project (Broadbent and Sjoberg 1990), as well as the historic context of Bjuroldubb. Most signifi- cant in this particular instance, however, is the fact that there is a stratigraphic association between a sealing hut and a labyrinth. The symbolic meaning of this super-positioning of stone constructions is discussed in Chapter 10. • A stone circle (Feature 16) was dated using lichenometry to A.D. io}^±}i, which is con- sistent with the dates of the hearth and bear bones in Hut 4. A full discussion of the stone circles is given in Chapter 8 and also in Wennstedt Edvinger and Broadbent (2006). • Iron working is evidenced by slag in Huts 3 and 11. Hut 11 contained an iron furnace, which is discussed in Chapter 7. • The most unusual find at Grundskatan is the bear burial in Hut 4, which dates to the elev- enth century and the main period of site occupation. A detailed discussion of this find is presented in Chapter 8. • The bone material from the hearths derives from five huts. There was approximately 700 g of bone, and the identified species or classes of animals were: bear, ringed seal, indetermi- nate seals, hare, large ungulates (moose/cattle or reindeer), as well as an indeterminate bird. • The anatomical representation of bones from Hut 3 indicates that a selection of seal parts had been taken to the hut. Of the seal bones from the extremities, both front and rear flippers dominate. Bones from the cranium and backbone are few. Age determination, on the basis of closure of the epiphyses, shows the bones from adult seals dominate. Two phalanges have changes indicative of a high age. The same selection of bones was present in Huts 3 and 4. One long bone fragment of seal has cut marks. • Macrofossils from Huts i, 2 and 3 indicate berry harvesting, presumably in the fall, as well as an assortment of other plants. • Charcoal analysis indicated that pine, birch, alder and rowan were burned. A find of yew from the Hut 4 is very unusual and can be the remains of a bow or relate to Eurasian sha- manism (refer to Chapter 8). Stora Fjaderagg Island Stora Fjaderagg Island is located ca. 14 km from the mainland and 3 km northeast of Holmon Island (63° 48' N, 21° 00' F). The island is roughly triangular in shape and measures 1.4 km in length and 1.2 km in width. It is the highest and oldest island in the Holmon Island group with an elevation of ca. 22 m (Figure 98). Its vegetation resembles the sub-alpine region with heaths, stunted stands of pine and spruce and exposed moraine and bedrock. Several larger ponds, and its location in the Bay of Bothnia, have made it a stopover for some 83 species of 92 CHAPTER 5 STORA FJADERAGG ISLAND GAMLA HAMNEN Figure 98. (top) Three-dimensional rendition oftopography of Stora Fjaderagg Island. Vertical scale exaggerated. Seenfrom the southeast. GIS by Katherine Rusk, (bottom) Map ofisland with locations of investigated huts and labyrinths. birds. The odd name probably derives from fara, which means "to travel," or might refer to danger or the appearance of the island. Historical sources from the sixteenth and seventeenth centuries describe the island as the village territory of Holmon islanders, but it was also used by seasonal herring and salmon fisher- men, and sealers from the mainland and Finland. During the most intensive fishing seasons more than 100 people are recorded as living there. It was even possible to grow potatoes. A small chapel had been built in 1729 by fisherman from Nykarleby in Finland (Jirlow 1930). According to local oral history, the first settlers of Holmon were the "Fisher-Lapps" Hakan, Kerstop and Klemmet. These names (Hakansson, Christiern and Clemmeth) were still common in the sixteenth century (Sandstrom 1988:138). Other place-names on the island derive from these personal names, such as "Klemmetsgrundet" (Klemmet's Reef) and "Clemmets Myra" (Klemmet 's Bog). Their original farmstead sites are still known. EXCAVATIONS AND ANALYSES 93 STORA FJADERAGG Figure 99. Map of Sites ^1-33 with hut locations and forms, pits, stone circles and excavations. Sandstrom judged the references to Fisher-Lapps to be credible, hi fact, residents of Holmon were called Lapps by other villagers in the region, as were the people of Stocke and Ranea on the mainland (Sandstrom 1988:138). This reference is reinforced by a unique Saami practice on Hol- mon ofmaking ropes using roots, a technique also known in Ranea (Sandstrom 1988:139). Finally, it can be mentioned that the Orrskar cemetery on Llolmon (with a squarish stone enclosure) was referred to as a "Viking nest" because of the heathen practices, including a sacrifices, that were said to have occurred there (Sandstrom 1988:138). This was a chapel site but, like lungfrugraven, could have an older pre-Christian rit- ual association. Archaeology of Stora Fjaderagg Island A concentration of 35 huts lies on the eastern side of the island (Sites 31-32). This array of huts and features faces southeast and all have excellent views of the surrounding seas; they are situated without reference to any harbor basins (Figures 98, 99). Excavations were carried out at 20.5 m, 19 m, 15 m and Figure loo. Map of Hut A. 94 CHAPTER 5 13 m above sea level. An area of ii stone circles between 7 m and 9 m above sea level was also documented (Site 33). In addition, two small dwellings near the Old Harbor (Gamla Hamnen) were excavated (Site 34). They lie at ca. 8 m above sea level and overlook the harbor basin (Figures 110, III, 112). Five labyrinths were lichen dated by Rabbe Sjoberg and range in age from A.D. 1525-1664 (Broadbent and Sjoberg 1990:295). Hut A Hut A is located at 20.5 m above sea level. It has an inverted "G" form with an extended entranceway facing south (Figure 100). The feature measures ca. 8 x 5 m and has a floor area measuring 3x4 m. The walls are ca. i m wide and 0.30-0.50 m high. The entranceway is 1.5 m wide and the most distinctive aspect of the hut is its unusual form. A i m^ pit was excavated in the hearth area. Table 18. Sites 31-32, Hut A: Faunal remains by weight (g). BONES BURNED (G) BURNED, NISP Indeterminate seal 0.18 2 Indeterminate seal 10.91 Not counted Total 11.09 Chronology 13.8 g of charcoal was obtained. This sample (St. 11181) produced one radiocarbon date: ioi5±ioo B.P. (A.D. cal. 898-1155). Figure 101. Map ofHut B excavation and locations ofprofiles. (A) Section through the hearth in the center of the dwelling. (B) Shows the hearth depression. (C) Cross section of the dwelling. Finds None Osteological Material Bone was recovered (11.09 §) ^^^^ two specimens could be identified, both frag- ments of metatarsal bones from the rear flippers of an adult seal. EXCAVATIONS AND ANALYSES 95 Hut B Hut B is situated at 19 m above sea level. The heavily overgrown hut is approximately 11 m in diam- eter with a floor area of 7 x 6 m (Figure loi). The walls are thick, up to 2.0 m wide, and there is no obvious entrance. The hearth was excavated as a unit (a i m^ pit) and this sampling area was then expanded to cover 19 m^. Finds Red clay/slag (iron furnace wall) Gray flint chip (11 mm) Gray flint chip (less than 10 mm) Gray flint chip (less than 10 mm) Slag (21x23 rni'n) Slag (34x23 mm) Fragment of whetstone (36 x 13 x 15 mm) Chronology Charcoal was recovered (18.7 g), and two radiocarbon dates were obtained: (St. 11900), i66o±70 B.P. (A.D. cal. 259-529), and (St. 11182), i235±3i5 (A.D. cal. 465-1154). The medians are: A.D. cal. 386 and 779. Osteological Material (by Jan Stord) A total of 7.9 kg of bones were recovered from the hut. Approximately 0.6 kg was either un- burned or charred. 5.2 kg of bones were recovered from the floor area and 3.2 kg from the hearth. The floor area contained a slightly smaller proportion of charred fragments than the hearth and the pit (according to weight). The bones came from the original i m^ excavation pit, the hearth, the floor area and the profile wall. Most of the bones from Hut B could only be identified as unspecified seal. The most common species was harp seal, followed by ringed seal, cattle, sheep/goat and duck. The representation of species is similar on the floor area and in the hearth, while the pit contained only bones of seals. There are also some differences in species representation between the burned and unburned bones. The minimum numbers of individuals for the identified species are: 4 harp seals, 2 ringed seals, Table 19. Sites 31-32, Hut B: Bones recovered in different areas, by weight (g). BONES PIT HEARTH SECTION* FLOOR TOTAL Burned 29.29 2181.84 752.7 4946.93 7910.76 Unburned 0.99 11.55 8.57 15.81 36.92 Charred 2.86 248.47 40.7 276.77 568.8 Soil sample (Indeterminate fragments) 1029.65 Total 33.14 2441.86 801.97 5239.51 9546.13 * from hearth. 95 CHAPTER 5 Table 20. Sites 31-32, Hut B: Species by weight (g). SPECIES BURNED UNBURNED CHARRED SOIL SAMPLE TOTAL Harp seal Ringed seal Indet. seal Cattle Sheep/goat Large ungulate Indet. duck Bird Bird? Indeterminate Not analyzed (indeterminate) Total * from hearth. 22.31 10.41 1986.1 0.42 1.82 0.15 0.2 0.27 5889.08 7910.76 0.44 8.82 24.96 2.7 35.92 54.74 510.06 4 568.8 1029.65 1029.65 77.49 19.23 2521.12 2.7 0.42 1.82 0.15 0.2 0.27 5893.08 1029.65 9546.13 Table 21. Sites 31-32, Hut B: Species, NISP. SPECIES BURNED UNBURNED CHARRED TOTAL Harp seal 9 Ringed seal 4 Indeterminate seal 3474 Cattle Sheep/goat 1 Large ungulate 1 Indeterminate duck 1 Bird 1 Bird? 2 Total 3493 1 3 15 1 7 623 20 630 17 7 4112 1 1 1 1 1 2 4143 Table 22. Sites 31-32, Hut B: Species in the floor area, NISP. SPECIES BURNED UNBURNED CHARRED TOTAL Harp seal Ringed seal Indet. seal Cattle Bird? Indeterminate Total 5 3 2580 2 162 2752 4 255 5 264 9 3 2841 1 2 167 3023 EXCAVATIONS AND ANALYSES 97 Table 23. Sites 31-32, Hut B: Species in the hearth (including the finds from the profile), NISP. SPECIES BURNED UNBURNED CHARRED TOTAL Harp seal 4 13 8 Ringed seal 1 2 3 Indet. seal 884 9 364 1257 Sheep/goat 1 1 Large ungulate 1 . 1 Indet. duck 1 1 Bird 1 1 Indeterminate 109 109 Total 1002 12 367 1381 Table 24. Sites 31-32, Hut B: Species in the 1 m2 pit, NISP. SPECIES BURNED UNBURNED CHARRED TOTAL Ringed seal 1 1 Indet. seal 10 4 14 Table 25. Sites 31-32, Hut B: Species and skeletal elements, unburned bones. ELEMENT HARP SEAL INDET SEAL RINGED SEAL CATTLE TOTAL Cranium Mandibula Hyoideum Atlas Vertebrae Gestae Cr+i Coxae Femur Talus Calcaneus T4 Mtv Ph3 post 1 Total 20 CHAPTER 5 Table 26 Sites 31 -32 Hut B- Soecies and skeletal elements, rharrpd hnnp^ ELEMENT HARP SEAL INDET. SEAL TOTAL Cranium 5 19 24 Mandibula 1 4 5 Atlas 3 3 Axis 1 1 V caud 8 8 Vertebrae 510 510 Costae 3 3 Humerus 1 1 Cr+i 2 2 Cu 1 1 CI 1 1 C2 1 1 C3 1 1 Mc II 1 1 Mc V 2 2 Phi ant 2 2 Ph2 ant 4 4 Ph3 ant 1 1 Coxae 4 4 Sacrum 2 2 Femur 1 7 8 Patella 1 1 Fibula 2 2 Talus 7 7 Calcaneus 2 2 Tc 2 2 Tl 1 1 T2 3 3 T3 4 4 T4 2 2 Mt 1 2 2 Mt II 4 4 Mt V 3 3 Mt 1 1 Phi post 6 6 Ph2 post 1 1 Ph3 post 4 4 Total 7 623 630 EXCAVATIONS AND ANALYSES 99 Table 27. Sites 31-32, Hut B: Species and skeletal elements, HARP RINGED INDET. SHEEP/ LARGE ELEMENT SEAL SEAL SEAL GOAT UNGULATE burned INDET. DUCK bones. BIRD BIRD? TOTAL Cranium 5 2 92 99 Mandibula 1 23 24 Dentes 1 8 9 Hyoideum 9 , 9 Atlas 2 2 Axis 14 14 V cerv 4 4 V caud 196 196 Vthor 1 1 Vertebrae 506 506 Costae 5 1 6 Cartil. Costae 1 1 Scapula 1 1 Humerus 2 2 3 Ulna 1 1 Carpometacarpus 1 1 2 Cr+i 4 4 C2 1 1 C4 1 1 Mc 1 6 6 Mc II 2 2 Mc IV 1 1 2 Mc 5 5 Phi ant 14 14 Ph2 ant 5 5 Ph3 ant 6 6 Ph ant 2 2 Coxae 31 31 Sacrum 2 2 Femur 15 15 Patella 4 4 Tibia 1 37 38 Fibula 1 19 20 Talus 22 22 Calcaneus 9 9 Tc 30 30 Tl 23 23 T2 54 54 T3 43 43 T4 27 27 Mt 1 33 33 Mt II 42 42 Mt III 30 30 Mt IV 22 22 Mt V 45 45 Mt 193 193 Mp 1 1 Phi post 469 459 Ph2 post 301 Ph3 post 310 310 Ph post 198 198 Tarsi/carpi 127 127 Sesamoidea 472 472 Baculum 3 3 Ossa longa 2 1 3 Total 9 4 3474 1 2 1 1 2 3494 100 CHAPTER 5 I cow, I sheep/goat and i duck. If the seal bones are treated as one unit regardless of species, the material contains bones from at least 21 different individuals. Epiphyseal fusion data of the first toe bone (Phi post) indicate that sixteen ofthem were adults while the other five were sub-adults. Some of the bones of both harp seals and ringed seals exhibit skeletal lesions characteristic of old age, indicating that there are at least a few very old adults among them. The other species identified in Hut B are represented by one individual each. There are noteworthy differences in the anatomical representation of seals between the dif- ferent find contexts and between the burned, charred and unburned bones. The differences are obvious, according to both the number of specimens and the weight distributions. The burned bones exhibit a higher representation from the rear flippers. The charred fragments are mostly from the vertebral column, while the unburned bones are from the cranium (and rear flippers). The bones identified as deriving from cattle or large ungulates come from the upper extremities, i.e. the meatiest parts of the animals, while the bones from sheep/goats are from the meat-poor lower ex- tremities. One femoral fragment comes from a calf These finds probably represent food resources brought to the island. Eight seal bones exhibited marks associated with butchery. One element is from a front flipper, two elements are from a rear extremity, and five are from rear flippers. Seven ofthe elements exhibit chop marks and two have superficial cut marks. One fragment exhibits both types of marks. The character of the marks indicates rather crude partitioning techniques using heavy tools, probably axes. Poor preservation of the bone surfaces may, however, mask the true frequencies of lighter cut marks. The differences in anatomical representation are not related to the different numbers of bones in each anatomical region. A comparison of anatomical units of seals shows that there was Stora Fjaderagg Hut B, NISP Total (N=4136) Unburned (N=19) Charred (N=630) Burned (N=3487) 0% 20% 40% 60% 80% 100% Cranium HTeeth DVertebral column Ei]Ribcage B Front extremity QFrontflipper [D Rear extremity Q] Rear flipper E3 Front/rear flipper Figure 102. Anatomical representation by NISP (excluding ^yi sesamoids, j bacula and 2 long bone fragments). EXCAVATIONS AND ANALYSES 101 Table 28. Sites 31-32, Hut B: Species and skeletal elements, unburned bones. ANATOMY ELEMENT HARP SEAL INDET. SEAL RINGED SEAL CATTLE TOTAL Cranium Backbone Rib cage Front flipper Rear extremity Rear flipper Cranium Mandibula Hyoideum Atlas Vertebrae Costae Cr+i Coxae Femur Talus Calcaneus T4 MtV Ph3 post 1 Total 20 Table 29. Sites 31-32, Hut B: Anatomical representation for seals by MAD (minimum anatomical unit). Burned, charred and unburned bones not separated. ANATOMY STORA FJADERAGG B (N=60) Cranium 6 Front extremity 3 Front flipper 4 Rear extremity 5 Rear flipper 42 Total 60 a clear preference for rear flippers. This comparison takes into account the number of skeletal ele- ments in each body region. The bones from seals represent a minimum of 6o different anatomical units; 42 of these are rear flippers. Note that this comparison excludes the vertebral column (see discussion, this chapter). There are also differences in anatomical representation in the different areas of the hut. The burned fragments exhibit relatively great differences between the hearth and the floor area; the hearth contained a larger proportion of burned fragments from the vertebral column as compared 102 CHAPTER 5 Stora Fjaderagg Hut B, Floor. NISP Unburned (N=6) Charred (N=259) Burned (N=2588) 0% 20% 40% 60% 80% 100% Cranium BTeeth DVertebral column Rib cage SFront extremity B Front flipper HI Rear extremity IHRear flipper Front/rear flipper Stora Fjaderagg Hut B, Hearth. NISP Unburned (N=12) Charred (N=367) Burned (N=889) 0% 20% 40% 60% 80% 100% Cranium Teeth DVertebral column E3 Rib cage HFront extremity QFront flipper [B Rear extremity ffl Rear flipper EBFront/rear flipper Stora Fjaderagg Hut B, pit. NISP Unburned (N=1) Charred (N=4) Burned (N=10) 0% 20% 40% 60% 80% 100% {Cranium DVertebral column BRear flipper Figure lO]. Anatomical representation in the floor area according to NISP (excluding 472 sesamoids, j bacula and 2 long bone fragments) . Figure 104. Anatomical representation in the hearth by NISP (excluding 4^2 sesamoids, j bacula and 2 long bone fi'aginents). Figure 105. Anatomical representation in the storage pit by NISP (excluding 4J2 sesamoids, ] bacida and 2 long bonefragments). EXCAVATIONS AND ANALYSES 103 Table 30. Sites 31-32, Hut B: Anatomical representation for seals, NISP. ANATOMY BURNED UNBURNED CHARRED TOTAL Cranium+teeth 132 7 29 168 Vert. Column 725 2 524 1251 Rib cage 6 1 3 10 Front extr. 13 1 1 15 Rear extr. 108 2 15 125 Front flipper 48 1 16 65 Rear flipper 1851 5 42 1898 Front or rear flipper 127 127 Sesamoidea, baculum 475 475 Long bone fragments 2 2 Total 3487 19 630 4136 Table 31. Sites 31-32, Hut B: Anatomical representation for seals by weight (g). ANATOMY BURNED UNBURNED CHARRED TOTAL Cranium+teeth 132.22 12.80 79.34 224.36 Vert. Column 382.51 6.72 368.52 757.75 Rib cage 5.11 3.09 3.83 12.03 Front extr. 5.26 0.17 1.45 6.88 Rear extr. 172.03 1.77 33.15 206.95 Front flipper 36.98 0.62 24.40 62.00 Rear flipper 1145.54 9.05 54.11 1208.7 Front/rear flipper 60.13 60.13 Sesamoidea, baculum 80.60 80.60 Long bone fragments 1.71 1.71 Total 2018.82 34.22 564.80 2617.84 with the floor area, which mainly contained bones from the rear flippers. The charred fragments displayed a similar anatomical representation for seals on the floor and in the hearth, with most bone fragments from the vertebral column. The unburned fragments were dominated by fragments from the rear flippers in the floor area, and cranial fragments in the hearth. The small number of unburned fragments probably makes the comparison somewhat unreliable, however. The anatomi- cal representation in the pit is more restricted than the on floor area and in the hearth, but this 104 CHAPTER 5 Table 32. Sites 31-32, Hut B: Burned bones, anatomical representation for seals, NISP. ANATOMY HEARTH PIT FLOOR TOTAL Cranium 67 56 123 Teeth 3 6 9 Vertebral column 410 1 314 725 Rib cage 4 2 5 Front extremity 2 11 13 Front flipper 20 28 48 Rear extremity 73 35 108 Rear flipper 255 9 1587 1851 Front/rear flipper 18 109 127 Long bone fragments 2 2 Sesamoids, baculum 37 438 475 Total 889 10 2588 3487 Table 33. Sites 31-32, Hut B: Burned bones, anatonnical representation for seals by weigfit (g:). ANATOMY HEARTH PIT FLOOR TOTAL Cranium 72.72 57.52 130.34 Teetfi 1.04 0.84 1.88 Vertebral column 247.30 0.09 135.12 382.51 Rib cage 4.00 1.11 5.11 Front extremity 0.94 4.32 5.26 Front flipper 15.33 21.65 35.98 Rear extremity 121.99 45.77 158.75 Rear flipper 190.68 3.99 950.87 1145.54 Front/rear flipper 7.92 52.21 60.13 Long bone fragments 1.71 1.71 Sesamoids, baculum 7.26 73.34 80.50 Total 569.18 4.08 1345.56 2018.82 comparison is also affected by the small mimbers of fragments. It general, it appears from this material that there was some form of spatial organization and related activities in Hut B. HutC Hut C is located at 15.5 m above sea level It measures ca. 5.5 x 5.5 m and the floor area is ca. 3 X 3 m (Figure 106). The walls are heavily overgrown and measure ca. i.o m in width. A single EXCAVATIONS AND ANALYSES 105 Table 34. Sites 31-32, Hut B: Unburned bones, anatomical representation for seals, NISP. ANATOMY HEARTH PIT FLOOR TOTAL Cranium 5 117 Vertebral column 2 2 Rib cage 1 1 Front extremity 1 1 Front flipper 1 1 Rear extremity 2 2 Rear flipper 2 3 5 Total 12 1 6 19 Table 35. Sites 31-32, Hut B: Unburned bones, anatomical representation for seals by weighit (g). ANATOMY HEARTH PIT FLOOR TOTAL Cranium Vertebral column Rib cage Front extremity Front flipper Rear extremity Rear flipper Total 10.93 6.72 0.17 1.77 0,53 20.00 0.99 0.88 8.52 0.99 13.11 12.80 5.72 3.09 0.17 0.62 1.77 9.05 34.22 3.09 0.62 Table 36. Sites 31-32, Hut B: Charred bones, anatomical representation for seals, NISP. ANATOMY HEARTH PIT FLOOR TOTAL Cranium 6 23 29 Vertebral column 333 1 190 524 Rib cage 1 2 3 Front extremity 1 1 Front flipper 7 9 16 Rear extremity 3 12 15 Rear flipper 17 3 22 42 Total 367 4 259 630 CHAPTER 5 Table 37. Sites 31-32, Hut B: Charred bones, anatomical representation for seals by weight (g). ANATOMY HEARTH PIT FLOOR TOTAL Cranium 35.26 44.08 79.34 Vertebral column 200.62 0.48 167.42 368.52 Rib cage 0.61 3.22 3.83 Front extremity 1.45 1.45 Front flipper 12.12 12.28 24.40 Rear extremity 9.30 23.85 33.15 Rear flipper 31.26 2.38 20.47 54.11 Total 289.17 2.86 272.77 564.80 Table 38. Sites 31-32, Hut B: Bones exhibiting nnarks of butchery. All fragments from the floor area. ANATOMY ELEMENT MARK Front flipper Mcl Cut marks at distal epiphysis Rear extremity Tibia Chop mark at distal end Tibia Chop mark on diaphysis Rear flipper Tc Chop mark and cut mark Tc Chop mark T2 Chop mark T2 Chop mark Mtlll Chop mark Figure 106. Map of Hut C with alignmmts. huts and a storage cache. EXCAVATIONS AND ANALYSES 107 entrance, measuring i.o m in width, faces south- east. The excavated hut is one of three construc- tions with external wall alignments, a depression and a cairn. One of the huts resembles the inverted "G" shape of Hut A and is roughly contemporary with it. This appears to be a small cluster of con- temporary structures. Finds One chip of gray flint was found and measures 15 mm in length. Chronology A I m^ pit was excavated in the hearth and 15.7 g of charcoal was found. One radiocarbon date (St. 11183) was obtained: 955+75 B.P. (A.D. cal. 1018-1163). Osteological Material No bone was recovered. Figure loy. Map ofHut D. HutD Hut D is situated at the 13 m elevation. It is roughly oval in shape and has two opposite-lying en- trances facing east and west. The features measures ca. 6 x 4.5 m (Figure 107). The floor area mea- sures ca. 4 X 3 m. The doors are ca. 0.75 m wide. A i x 0.50 m pit was excavated in the hearth. This rendered 205.8 g of bone. Charcoal was found (10.6 g) and radiocarbon dated (St. 11184): 1110+145 B.P. (A.D. cal. 714-1036). Table 39. Sites 31-32, Hut D: Identified species, NISP. SPECIES HEARTH SAMPLE 1 SAMPLE 2 SAMPLE3 SAMPLE 4 TOTAL Indeterminate seal 1.67 27.95 1.23 1.03 0.08 31.95 Indeterminate 93.43 12.2 6.81 6.25 2.06 120.75 Total 95.1 40.15 8.04 7.28 2.14 152.71 Finds One gray flint chip Red brown clay furnace fragment (1.5 cm) Osteological Material The hearth soils and 4 additional soil samples were collected from Hut D. 108 CHAPTER 5 Table 40. Sites 31-32, Hut D: Anatomical representation for indeterminate seals, burned bones, NISP. ELEMENT HEARTH SAMPLE 1 SAMPLE 2 SAMPLE 3 SAMPLE 4 TOTAL Cranium 12 3 V caud 5 5 Vertebrae 1 1 C2 1 1 Tc 1 1 Tl 1 1 T2 • 12 1 4 Mtl ^ 3 3 Mt II 1 1 Mt III 1 1 MtV 2 1 3 Mt 1 7 8 Phi post 20 1 4 25 Ph2 post 8 8 Ph3 post 4 4 Tibia 1 1 Ph post 1 3 4 Sesamoidea 21 2 1 24 Total 6 81 5 5 1 98 Table 41. Sites 31-32, Hut D: Anatomical representation for seals by MAU (minimum anatomical unit). ANATOMY STORA FJADERAGG D (N=6) Cranium Front extremity Front flipper Rear extremity Rear flipper The bones derive from seals of inde- terminate species, and the rear flippers were most common. The minimum number of in- dividuals is two (based on three rear flippers). Epiphyseal fusion of the toe bones indicates that one ofthem was an adult and the other, a sub-adult. One element from the rear flipper (Mtll) exhibits cut marks on the diaphysis. Total Circular Features A group of ten ring-shaped stone settings measuring 3 to 5 m in diameter is recorded in the archaeological surveys (Site 33). The rings are on two beach ridges within an area of ca. 50 X 112 m between 7 and 9 m above sea level (Figures 108). The largest feature is an oval wall measuring 0.75 m to i m in height, with a diam- eter of 4.9 X 3.7 m (Figure 109). The wall is ca. 50 cm in width and incorporates a large unusual EXCAVATIONS AND ANALYSES 109 looking boulder. Fifteen meters to the northwest of this feature are two simpler stone rings, one mea- suring 3.5 X 4.1 m in diameter, and a smaller ring measuring 2x3m. Fifty meters to the northeast, and on a terrace at the 9 m elevation, is an additional cluster of stone constructions, including a small cairn of stones lying up against a boulder. Five meters to the north- west are three smaller rings. One of these consists of three small chambers. Five meters above them is a 5.8 m long cobble wall, an oval depression measuring 4 x 3.3 m. hearth circle enclosure 4|p cairn by a boulder ^'"-^^ , ^'.^-l^ multiple circles circle ^ V' v.* circle depression * *-\ < alignment depression y 10 m Figure 108. Sketch map of Site with stone circles and related constructions and depressions. 110 CHAPTER 5 GAMLA HAMNEN 7 m excavations » * compass rose S m 4 m LABYRINTH .v^ (1525-1595) i> pond shooting blind CHAPEL ^ n (1729-1910) 8m -7m 7 • net-drying y ; caims BASIN 3 m and a stone ring measuring 4 X 37 m. The complex faces in the same direction as the settle- ment as a whole, southeast. The shore level suggests a date anal- ogous to the radiocarbon dates from Huts B-D., approximately A.D. 900-1200. Only two of the stone rings can be seen as dwellings (tent rings). The other circular features do not appear to have had any practical func- tions. These features parallel other finds documented within the project that are interpreted as Saami in origin (cf Wennst- edt Edvinger and Broadbent 2006). The most remarkable fea- ture at the site is the cobble-built enclosure and a large boulder with odd eye-like holes (Figure 109). According to early writ- ten sources (Leem 1767), these circular constructions could be Figure ni. Map ofGamla Hamnen. EXCAVATIONS AND ANALYSES 111 Table 42. Stora Fjaderagg: Bones from Gamla Hamnen. SPECIES WEIGHT (G) NISP ELEMENT Large ungulate 22.28 2 1 centrotarsale; 1 long bone fragment Indeterminate seal 1.66 1 Part of the claw of the third toe bone, Ph3 posterior Total 23.94 3 covered over to prevent dogs or predators from desecrating the offerings. These interpretations are discussed in Chapter 8. Gamla Hamnen Two small overgrown huts lying at the ca. 8 m level above the Gamla Hamnen area were in- vestigated. Hut A is an irregular oval hut measuring ca. 4 x 3.5 m. The low walls measure ca. 0.50 m in width. A small hearth was found in the northwest corner of the hut. Hut B is also irregu- lar feature and measuring ca. 4 X 4.5 m (Figures no, iii). Finds None , - , Figure 112. Aerial view oj Gamla Hamnen, looking east. Chronology One radiocarbon date was obtained from Hut B (St. 11901): 3io±70 B.P. (A.D. cal. 1490-1650). Osteological Material All bones were unburned: one tarsal bone and a long bone fragment of a large ungulate (cattle or moose) and the claw (nail) of a third toe bone of an indeterminate seal were identified. The harbor basin has a threshold elevation of 2.54 m above present sea level, which dates to ca. A.D. 1650. A beach ridge at its entrance is 50 cm higher than this level and is probably the result of storm surges. The chapel by the basin dates to 1729. As a whole, the radiocarbon date from the huts and the lichen dates of the labyrinths, suggest that the harbor could date as early as 1490 to 1650. As noted earlier, Sandstrom (1988:136-141) has undertaken an extensive analysis ofthe eleva- tions of place-names on Holmon and found they start at the 6 m level, or ca. A.D. 1300. 112 CHAPTER 5 Summary Discussion The highest-lying huts and the stone circles from Stora Fjaderagg Island date to the period ca. A.D. 200 to 1200. This range is fully consistent with the other inves- tigated sites with similar eleva- tions. The harbor and associated features date to the late sixteenth century, in line with the licheno- metric dating of the labyrinths on the island. The artifacts from the huts show evidence of iron working, both slag and furnace clay. This material was found in Huts B and C, ranging in age from 1660 to 1015 B.P. Flint chips from strike- a-lights were found in Huts B, C and D. The two stray finds from the island, a silver ring and bronze bells, date to the late Viking Pe- riod and are contemporary with the huts (cf Serning 1960:150). In addition to the seal bones discussed below, there were finds of cattle, sheep/goat and duck bones. This evidence of husbandry parallels the finds of sheep/goat bones from Jungfruhamn Site 138. Even stone alignments, as found at Bjuron Site 70 and Grundskatan Site 78, were in evi- dence at Hut C on Stora Fjaderagg, and are seen at other huts on the island as well. The circular features, combined with the oral history regarding Lapp settlers on Holmon and Saami traditions in rope- and net making, make the island a key locale in this analysis of coastal Saami culture in Vasterbotten. Comparative Analysis of the Osteological Material (by Jan Stora) Harp Seal Populations Earlier analyses have shown that prehistoric harp seals in the Baltic exhibited a smaller body size than extant ones from the Atlantic (Stora and Ericson 2004; Stora and Lougas 2005). Due to the high level of fragmentation it has been difficult to document osteometric data from Stora Fjaderagg. The (greatest) diagonal breadth ofpars mastoideus of two temporal bones (42.7 and 37.15 mm) are the only measurements providing information on the sizes of the harp seals. One of the elements comes from a rather large adult individual, in fact larger than most individuals from the HutBl 1660t70BP HntB2 1235±315BP HilD niOtl45BP Hut A lOIStlOOBP HutC 95St75BP aXJOCalBC KlOOCalBC GilBC/CalAD laXtalAD 2000QilAD Calibrated date 12(XralAD l4(KK:aIAD 16()0CalAD ISIXKTalAD 2(K)()CalAD C'dibrated dale Figure ii}. Calibrated dates ofhuts at Sites p-p. Below, calibrated date ofhut by Gamla Hamnen. EXCAVATIONS AND ANALYSES 113 Stora Fjaderagg Hut B. Level of epiphyseal fusion of finger- and toebones in seals Ph 2 post Ph 1 post Ph1 ant 0% 20% 40% 60% 80% 100% Ph1 ant Ph 1 post Ph 2 post Unfused 2 48 27 min fusion 1 Fused 8 157 133 Fused in In fusion Unfused Figure 114. Level ofepiphyseal fiision offinger and toe bones in seals. The unfiised elements most probably derive from sub-adult seals while thefused bones comefrom adult seals (Aging according to Stord 2001a). Size distribution of Bronze Age, Iron Age harp seals from the Baltic and extant adult harp seals from the Atlantic 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 (Greatest) diagonal breadth of pars mastoideus (mm) Q] Bronze Age Iron Age Stora Fjaderagg Modern adult Figure 315. Size comparison oftwo temporal bones (37-15 and 42.'/ mm ) from Stora Fjaderagg with bonesfrom the Aland Islands and Estonia, and extant harp sealsfrom the Atlantic (Modifiedfrom Stord and Lougas 200^). Bronze Age and Iron Age. The size nevertheless corresponds to that of the smallest adults of extant harp seals from the Atlantic. A small temporal bone from Stora Fjaderagg probably de- rives from a sub-adult individual. Approximately 31% of the material by weight could be identi- fied as to species or class of animals. Bones from ringed seals {Phoca hispida), harp seals {Phoca 114 CHAPTER 5 Table 43. Stora Fjaderagg: Combined species by weight (g). Material in soil sannples excluded. SPECIES BURNED UNBURNED CHARRED TOTAL Ringed seal 10.41 8.82 19.23 Harp seal 22.31 0.44 54.74 77.49 Indeterminate seal 2018.24 26.62 510.06 2554.92 Cattle 2.7 2.7 Large ungulate 1.82 22.28 24.1 Sheep/goat 0.42 0.42 Indeterminate duck 0.15 0.15 Bird 0.2 0.2 Bird? 0.27 0.27 Indeterminate 6020.74 4 6024.74 Total 8074.56 60.86 568.8 8704.22 Table 44. Stora Fjaderagg: Combined species, NISP (indeterminate fragments not counted). Material in soil samples excluded. SPECIES BURNED UNBURNED CHARRED TOTAL Ringed seal 4 3 7 Harp seal 9 17 17 Indeterminate seal 3574 16 623 4213 Cattle 1 1 Large ungulate 12 3 Sheep/goat 1 1 Indet. duck 1 1 Bird 1 1 Bird? 2 2 Total 3593 23 630 4246 groenlandica) and indeterminate seals (Phoca sp.) by far dominate the material, together with solitary occurrences of cattle (Bos taunis), sheep/goat [Ovis aries/Capra hircus) and bird bones. There are only minor differences in the species compositions among the burned, charred and unburned bones. Some bones could be identified as to class of animals or group of animals only. The bones identified as large ungulates may originate from moose (Alces alces), cattle or reindeer (Bos taurus, Rangifer tamndus). The possibility of horse (Equus caballus) can, with some certainty, be excluded. EXCAVATIONS AND ANALYSES 115 SPECIES Harp seal Ringed seal Seal sp. Sheep/goat Large ungulate Indeterminate duck Bird Bird? Indeterminate Total The osteological analysis of the finds from Stora Fjaderagg has given new insights into the activities at these coastal sites. This is the first time that harp seal has been iden- tified at such a location, and this is rather surprising. Previously, only ringed seals have been identified on sites this far north dating to Late Iron Age or Medieval period. Ethno- graphic records shovv? that the ringed seal was also important in historic times, together with gray seal (Hali- choerus grypus). The latter species is surprisingly rare in archaeological records from the Bothnian Sea and the Bay of Bothnia (see Ekman and Iregren 1984; Ukkonen 2002). At present, it seems that the hunting for gray seals has a rather recent history. The discovery of harp seals from Stora Fjaderagg is chrono- logically one of the youngest in the Bothnian Sea. In the most compre- hensive survey of faunal remains from archaeological sites in north- ern Sweden published in 1984, no finds of harp seal were reported (Ekman and Iregren 1984). How- ever, harp seal bones were later identified at other sites. Some finds of harp seals have, for example, been reported from the Neolithic sites of Lillberget in Norrbotten (Halen 1994; Wal- lander 1992) and Bjurselet in Vasterbotten (Lepik- saar 1975), and in recent years the species has been identified at several coastal sites in the southern (and middle) coastal areas of the Bothnian Sea, i.e. Bjastamon in Angermanland (Olson et al. 2008) and Frakenronningen in Gastrikland (Holm 2006). The harp seal is very common in coastal site refuse faunas dating to the Neolithic period in the Baltic Sea (e.g. Stora and Ericson 2004). Table 45. Stora Fjaderagg: Combined burned bones by species, NISP (indeterminate fragments not counted). HUT A HUT B HUT D TOTAL 9 4 3474 1 1 1 1 2 271 + 3764+ 98 98 9 4 3574 1 1 1 1 2 271 + 3864+ Table 46. Stora Fjaderagg: Combined unburned bones by species, NISP. SPECIES HUT B OLD HARBOR TOTAL Harp seal Ringed seal Seal sp. Cattle Large ungulate Total 1 3 15 1 20 1 3 16 1 2 23 Table 47. Stora Fjaderagg: Combined charred bones by species, NISP. SPECIES HUT B Harp seal Seal sp. Undetermined Total 7 623 5 635 116 CHAPTER 5 The number of finds of harp seal in the Baltic decreases after the Neolithic Period. New studies suggest, however, that the species may have been present in the region more recently. Two sub-fossil harp seal skeletons on the Finnish West Coast of the Bothnian Bay have been radiocar- bon dated to the Bronze Age (Ukkonen 2002). Another find of harp seal in Finland was recovered in a cairn dating to the Early Iron Age (Makivuoti 1986). More numerous finds in archaeological contexts have also been reported in the Baltic. Iron Age finds of the harp seal have been reported from the Aland Islands, the Estonian Islands, the Islands of Gotland, Oland and Bornholm (Stora and Lougas 2005 and references therein). The finds from Stora Fjaderagg are contemporary with these finds. The size of one adult harp seal at Stora Fjaderagg is amongst the largest found com- pared with other Iron Age harp seals from the Baltic. The finds from Stora Fjaderagg provide some support that remnants of large Neolithic Period harp seal populations were still present in different areas of the Baltic Sea during the Late Iron Age. The uniqueness of the harp seal finds at Stora Fjaderagg makes it difficult to generalize about hunting patterns. The harp seal is obviously the most common seal found in Hut B. The minimum number of individuals for ringed seals is two. However, the minimum number of seals of all spe- cies in Hut B is at least 21. In contrast to the most often solitary ringed seal, the harp seal is both migratory and gregarious. It breeds in late winter/early spring and is unable to keep breathing holes open in fast ice. This latter behavior is similar to that of the gray seal. The behavioral patterns of the harp seal are rather different from those of the ringed seal and in Neolithic times hunting pat- terns differed for the two species (Stora 2001b). The behavioral patterns almost certainly affected hunting strategies during the Iron Age, but unfortunately it has not been possible to pinpoint sea- sonality using only the seal bones from Stora Fjaderagg. If behavioral patterns of the harp seal are considered, the most suitable period for hunting would have been during periods with open water, the summer and fall. The harp seals in Hut B were found together with bones of a duck; the latter would have also been in the area during the warmer seasons of the year. Whether the Stora Fjaderagg find is unique remains to be seen. Hut B is in several respects different from other studied hut structures of similar date and in corresponding locations. The presence of harp seal is certainly significant and the amount of faunal remains in Hut B exceeds that of all the other examined hut structures (e.g. Stora and Broadbent 2001). It appears that not all body parts of the seals were transported to Hut B. Bones from the rear flippers clearly dominate the assemblage. Interestingly enough, the meat of the flippers has often been considered the tasti- est by hunters, and sometimes even the only edible parts of seals. At Stora Fjaderagg there was an obvious preference for rear flippers, while the front flippers were uncommon. Hut D on Stora Fjaderagg shows a similar (although not as obvious) preference for the same anatomical parts as that seen in Hut B. The vertebral column is also fairly well represented, especially among the charred remains. Due to the level of fragmentation it has not been possible to estimate the minimum number of anatomical units for the different sections of the vertebral column (cervical, thoracic, lumbar and caudal). It is possible that the different regions of the vertebral column are not represented in similar proportions in Hut B. Due to the high level of fragmentation, most fragments have been identified as indeterminate vertebrae only. Considering a complete seal, it would be expected that thoracic vertebrae would be best represented. Also noteworthy is the fact that most vertebral frag- ments, identified to a specific region, are caudal vertebrae and cervical vertebrae also seem to be EXCAVATIONS AND ANALYSES 117 well represented. The second cervical vertebra, the axis, is represented by at least lo individuals. Thoracic and lumbar vertebrae are not absent, but they appear to be underrepresented by com- parison with the caudal and cervical regions. The few fragments of ribs at Stora Fjaderagg are an indication that this was the case, as the ribs are anatomically associated with thoracic vertebrae. It is also possible that rib fragments are rare because this part of the seal not was processed in Hut B. Indeed, the caudal vertebrae and rear flippers may have been be connected if the rear body parts of the seals were handled as a single unit. It seems plausible that only parts of seals were brought to Hut B for preparation and consump- tion. The character of the cut marks from slaughter indicates rather crude methods for partition- ing the body parts. The anatomical parts of seals that are missing in Hut B (as compared with the numbers of rear flippers), are the crania, front extremities and front flippers, thoracic vertebrae, ribs and rear extremities; these parts were taken somewhere else. The missing parts at Hut B roughly comprise the articulated trunk. One aspect that needs more attention is a possible bias in preserva- tion. The burned fragments exhibit a clear dominance of bones from the rear flippers, while the charred fragments were dominated by vertebral fragments. This indicates that the vertebrae were subjected to a lower level of burning. Many vertebral fragments are burned, but the proportion of charred fragments is higher as compared to other anatomical units. The few unburned fragments came from several different anatomical regions, but they are too few to highlight further. The lower level of burning of vertebrae may also be because they were more deeply embedded in soft tissue than the more superficial bones, e.g. the toe bones of the rear flippers. If the vertebrae (and other more embedded bones) were not burned to the same extent as the rear flippers, this could mean these bones are underrepresented. This possibility has to be taken into account, but most probably Stora Fjaderagg, Hut B. Anatomical representation for seals, NISP 2000 1600 1200 800 400 > O o 0) O) ni o Si S X i_ re o <3> O. a. a. a. ns Q. a. 1 = U. (Q 0) re o T3 o E re (0 0) c o Burned (N=3487)Q] Unburned (N=19)H Charred (N=630) Figure ii6. Anatomical representation ofseal bones in Hut B. NISP. 118 CHAPTER 5 Stora Fjaderagg, Hut B. Anatomical representation for seals by weight (g) 1400 1 Burned (201 8.82g) Unburned (34.22g) Charred (564.80g) Figure iiy. Anatomical representation ofseal bones in Hut B, by weight (g). Stora Fjaderagg, anatomical representation for seal, MAU Stora Fjaderagg D (N=6) Stora Fjaderagg B (N=39) I -I 1 \ \ \ 0% 20% 40% 60% 80% 100% Cranium ! Front extremity Q] Front flipper B Rear extremity QRear flipper Figure ii8. Anatomical representation ofseal bones in Hut B and Hut D according to minimum anatomical units. EXCAVATIONS AND ANALYSES 119 cannot explain the observed patterns of anatomical representation for seals at Stora Fjaderagg. The absence of finger bones from the fore flippers cannot be explained in this manner. The finger bones (and also the cranial bones) were not embedded in thick layers of soft tissue. The faunal material from Hut B has highlighted the complicated taphonomic history of this kind of assemblage. It is obviously one of the more intriguing faunal assemblages from such a location and time period. It is of interest that there are differences in anatomical representation in Hut B between the hearth and the floor area. This has some implications regarding the character of carcass utilization. If the archaeological excavations had targeted either the floor area or the hearth, the results would have been different. Additionally, the analysis has shown that it is important to relate the faunal remains as closely as possible to the overall find context. To conclude, the faunal assemblage from Stora Fjaderagg Island has provided new insights on a number of issues: 1) For the first time harp seal (Phoca groenlandica) has been identified in such a location and time period this far north. The ringed seal is not the most common species in Hut B, in- dicating that seal hunting patterns during the Iron Age may have been more varied than previously believed. 2) The amount of faunal remains in Hut B exceeds that of any other previously analyzed coastal hut structure. Hut B contains the body parts from at least 21 different seals indi- cating a planned utilization of the seals and use of the hut structure. The meat of the 60 anatomical regions identified in Hut B would have been a considerable food resource. 3) The anatomical representation of seals indicates that selected body parts were brought to Hut B for specific processing and probably consumption. The large parts of the seals that are missing in Hut B indicate strategies beyond immediate use. The missing parts ofthe 21 Figure iig. Three-dimensional rendition ofSnoan Island with radiocarbon dates. 120 CHAPTER 5 SNOAN Figure 120. Map of Snoan Island, Umea Municipality. Vasterbotten. Investigated areas indicated by rectangles. SITE 49 individuals represent a considerable amount of meat, blubber and skins. 4) Differences in anatomical representation in the find contexts of Hut B indicate spatial organization. The differences observed between the burned and charred bones, as well as the differences between the hearth and the floor, have been established for a rather large faunal material. This suggests a repeated behavioral pattern inside the hut. 5) All ofthe above indicates that Hut B must be regarded as a rather permanent structure in a well-organized hunting complex. The activities that produced this faunal assemblage were also well planned. Stora Fjaderagg contains many kinds of material remains indicative of the importance of the island over a long time period of time. Snoan Island Snoan Island is outermost in the Snoan archipelago at 63° 29' N, 20° 53' E, and is ca. 7 km from the mainland. The island consists of two parallel glacial ridges and is arrowhead shaped (Figures 119, 120). It measures 2.8 km in length, 0.9 km in width, and rises up to 17 m above sea level. Snoan Island was surveyed by the National Heritage Board in 1981 and an excellent account of this work was published by Lofgren and Olsson (1983). Five locales with huts, eight labyrinths and an equal number of Figure 121. Map of Site 49. Figure 122. Photo oj Hut B with distinctive hearth, facing east. EXCAVATIONS AND ANALYSES 121 compass roses were recorded. The main hut con- centration of some 50 structures hes above the 8 m curve within an area of 650 x 100 m (Site 49). The first historical account/map of the island dates to 1646. The name was written as Snodan and means "barren." Nine fishing sheds and a chapel were recorded in 1821 and there was culti- vatable land in the central part of the island. The goal ofmy investigation was to obtain charcoal for radiocarbon dates, animal bones and artifacts from different shoreline eleva- tions. Two huts were investigated at the 10-15 m elevation, Site 49. Extensive mapping was undertaken at Sites 53 and 92, which are Late Medieval fishing harbors between 3-8 m above present sea level. Two dates were obtained from a Russian Oven at the 5 m level, and two dates were obtained from a hut by a harbor basin at the 3-5 m levels. Site 49 These oval huts form a cluster offeatures near the highest part of the island. Fourteen huts cluster above the 10 m level (Figure 121). Five of the huts face west, three east, one north and two south. Nine ofthe huts have hearths. Hut 49A measures 7x8m and has an inner measurement of 5 x 6 m. The entrance is 2-3 m wide. Hut 49B has the same measurements. Hut 49B has a large built- up hearth measuring 2x3m that nearly fills the floor (Figure 122). Charcoal was recovered from Hut 49B and dated: (Ua-1323), 255±ioo B.P. This recent date suggests that the hearth was contami- nated by later re-use. One artifact was found, a white-gray burned flint piece measuring 22 x 15 mm. Hut 49A (Ua-1322) rendered a date of 1150+100 B.P. (A.D. cal. 775-988), which is con- sistent with its elevation. Burned bone was found in both huts. In Hut A, a large ungulate was iden- tified. In Hut B seal and hare bones were identi- fied. There was only 16 g of bone from the two huts, including those of a large ungulate (moose or reindeer), two rear flipper bones, including one Table 48. Stor-Rebben: Bones in Huts A and B. SPECIES HUT A HUT B TOTAL Seal 2 2 Large ungulate 1 1 Hare 1 1 Indeterminate. 64-I-* 64* Total 65+* 3 68* * not all small fragments counted labyrinth (1538) c::c SITE 92 8 m E7^'" 1 C boathouses 3 m • net drying • .• i cairns bog I Huts QU . t N 5 m 8 m 3 m I \ 2 Q c(3i"Russian Oven" >^ fiire-cracked 25 \ * rocks — — 2.75 m (threshold) \ m Figure 323. Map offishing harbor, Site 92. 122 CHAPTER 5 SITE 53 Figure 324. Map offishing harbor area. Site 53. tiit49A llSOtlOOH' Hut 53A 735±120BP Oven92B47aiS)5BP Hut 53B 445tl05BP Oven92A43Qt95H' CaBQCMAD 500CalAD KXiralAD ISOOCalAD 2000CalAD Calibrated date Figure 125. Calibrated datesfrom Sndaii Island. harbor shows them to pre-date the Russian invasions. The labyrinth dates to A.D. i538±35. Sjoberg measured Hchen g that dated them to A.D. 1388-1816 (Broadbent and Sjoberg 1990:295) from an adult seal, and an adult hare. Although this is a small sample, this pattern is most simi- lar to that of Grundskatan. Site 92 Site Area II is located in the cen- ter of the island and by a narrow former inlet. Numerous features lie on the west side of the inlet and follow the 5 m elevation (Figure 123). The entrance to the inlet has a threshold of 2.75 m above sea level, which means it could not have been used after ca. 1700. The stone features consist of U-shaped stone struc- tures that open toward the for- mer inlet. They were probably boat slips and it is possible they once had boat houses attached to them. There are also many small cairns that supported posts for drying nets. A number of small hut-like shelters, fire-cracked rocks and a Russian Oven were found as well. Two radiocar- bon dates were obtained for the oven:. (St. 11902): 43o±95 B.P. (A.D. cal. 1413-1630) and (St. 11903) 47o±95 B.P. (A.D. cal. 1320-1618). These are among the few radiocarbon-dated Rus- sian Ovens in the region and their association with a fishing 5 m level dates to ca. A.D. 1450 and the rowth on seven labyrinths on the island Site 53 Site 53 is a complex of small net-drying post cairns, hut-like enclosures, boat slips, a labyrinth and a compass rose (Figure 124). This site (Area III) follows the 3-5 m elevation and had probably replaced Site 93 when it became too shallow. Two features were sampled for charcoal and rendered two dates. EXCAVATIONS AND ANALYSES 123 Feature A, situated at 5-6 m above sea level, is a hut and produced one radiocarbon date: St. 11904) 735+120 B.P. (A.D. cal. 1167-1392). Feature B, located just above the 3.0 m shoreline, and rendered (St. 11905) 445±i05 B.P. (A.D. cal. 1401-1631). . These dates are consistent with the elevations. The net-drying cairns lie lower than 3.0 m and show the use of this harbor after 1600. These two har- bors are very similar to the harbors on Stora Fjaderagg and at Jungfruhamn. Conclusions The investigations on the Island of Snoan focused on three sites: the highest lying area, 10-15 m.a.s.l., and two harbor sites at 5 and 3 m.a.s.l. The sealing huts date with some certainty to the Viking Period. It is clear that from A.D. 1300 fishing was the principal activity on the island. The two dates from a "Russian" Oven are a unique result and show that bread was baked on the island hundreds of years prior to the Russian invasions. Stor-Rebben Island Stor-Rebben Island is located in the Pitea archipelago at 65° 11' N, 21° 56' E in the County of Nor- rbotten, about 5 kilometers from the mainland and 90 km north of Bjuroklubb. The island is Figure 126. Photo of Hut A, [1 | looking east. Figure izy. Map showing investigated huts on Stor-Rebben Island. 124 CHAPTER 5 Figure 128. Ann Wastesson and Ann-Christin Nilsson excavating a trench in Hut A. In the background, left to right, the author. Sture Berghmd and Rabbe Sjoberg. Photo by Nils Ogren. Figure 129. Map ofHut A on Stor-Rebben. area of fire-cracked rocks: 1 1 1 ; 1 1 1 1 dlJ humus pvl carbon I I sooty soil [o] soil sample 4 3 2 PI sooty, sandy soil Q sterile Figure ijo. Profile ofi trench through Hut A. rectangular in shape, 1.6 km long and 0.85 km wide. A sandy spit fans out toward the mainland (Figures 45, 127). According to Swedish historian Birger Steckzen the name rebben probably de- rives from the Saami words ruebpe or riebpe, which means a stony overgrown hill with brushy veg- etation. Reference is also given the term ruobba which means a rocky hilltop (Steckzen 1964:232). This is certainly an accurate description of Stor-Rebben Island (Figure 126). This site was chosen for comparative purposes and is the northernmost of these hut localities to be investigated. The island rises up to 17 m above sea level and dwellings are found on beach terraces at three main levels: 16 m, 13 m and 7 m above sea level. Claes Varenius had described the island in two articles in connection with archaeological surveys in the 1960s and 70s (Varenius 1964, 1978). He identified 27 huts, four of them double huts, and 11 huts with central hearths. He also recorded six labyrinths and a compass rose. Five of the labyrinths occur together with 3 hut groups: two of them at the 16 m level, two at the 13 m level and two at the 7 m level. There are four clusters of huts consisting of two to four structures each, just above the 16 m contour. At about the 14 m contour there are seven clusters, consisting of two to three huts and three labyrinths. Finally, at the 7 m level are the remains of a dwelling with EXCAVATIONS AND ANALYSES 125 ^' labyrinth 13.09 N t HUTB 13.62 \ 13 m beach ridge 5 m labyrinth 13.10 ^13.60 labyrinth Figure iji. Map oj Hut B and three labyrinths. a chimney and steps. Many of these features have only partial walls and are obviously dis- turbed. The most intact dwellings with ar- chaeological potential, those with charcoal or bones, were chosen for excavation. Figure i}2. Excavation unit in Hut B. Figure Profile of the heaHh in Hut B showing two lenses. Hut A Hut A is the larger of two structures at the i6 m level. It lies in a depression between a bedrock outcropping and a gravel ridge (Figures 126, 128, 129,130). The hut consists of an oval foundation measuring 6 x 8 m with a floor area of approximately 5x6m. There are two possible entrances at both short ends of the construction. Although no delimited hearth was distinguished per se, the floor had a thick layer of charcoal and fire-cracked rocks concentrated in the middle of the floor. Two radiocarbon dates were obtained (St. 11910 and St. 11178): (i845±i35 B.P. and i494±7o B.P.). These two dates calibrate respectively to 23 B.C. to A.D. 340 (median A.D. 172) and A.D. 443 to 643 (median A.D. 558). The first date range is the oldest radiocarbon date in the project. Against the 126 CHAPTER 5 Figure 134. Map oj Hut C. background of shoreline displacement at Stor-Rebben, and the calculated age of the 16 ni shoreline; the true age for the hut is probably A.D. 200-300. Old charcoal is an obvious risk, but these dates are not totally unreasonable. No bones could be identified from the dwelling. Carbonized seeds were found and identified as coming from crowberry bushes. These berries were also found in hearths at Grundskatan and Stora Fjaderagg. A gray flint chip was recovered, as well as 14 tiny slag fragments. The iron slag is the northernmost evidence of iron working. Hut B Hut B is one of two structures at 13.5 m above sea level (Figure 131). The form of these huts is rounded-rectangular and they have front and rear entranceways. One wall circles onto a small chamber that was probably a storage space. A central hearth rendered charcoal (St. 11179): i045±7o B.P. (A.D. cal. 880-1155). The median age is A.D. 985. The hearth displayed two ashy lenses sepa- rated by a sterile layer of sand (Figure 133). The radiocarbon date is from the upper level. This evi- dence shows that the dwelling was not used on only one occasion, and the range of dates from this site indicates - like the dates from Hut A - repeated use over hundreds of years. Iron slag was also found in this hearth, as well as a flint chip from a strike-a-light. Twenty-one bone fragments were recovered, but were small and unidentifiable. Three labyrinths lay within 15 m of these huts, one of which abuts the wall of a dwelling. As will be discussed in Chapter 10, it is not likely that the huts and the labyrinths were contemporary. Hut C Hut C is one of four structures situated at 13.5 m above sea level. These huts are rectangular in shape and three have central hearths (Figure 134). The huts have entrances through their shortest EXCAVATIONS AND ANALYSES 127 HutAl 1845±130BP HLitA2 1494t70BP HutB IMSfcllOBP HutC 945±110BP lOOOCalBC CalBC/CalAD Calibrated date lOOOCalAD 2000CalAD Figure 335. Calibrated radiocarbon dates ofhuts on Stor-Rebben Island. walls and at right angles to the shore. The floor areas are 3.0 x 4.0 m. There is a small storage cairn in one of the walls. The hearth in the most intact of the three huts was excavated. It produced charcoal and 230 burned bone fragments. One bone was identified as coming from a reindeer or goat/sheep. The radiocarbon date (St. 11180) is: 945±iio B.P. (A.D. cal. 999-1212). The median is A.D. 1092. Two grey flint chips and a steatite piece with a groove and five slag pieces measuring 10- 32 mm were found. HutD This dwelling differs completely from the others, and is of medieval or historic date. It has the base of a brick chimney and stairs. In addition to burned bone, there were flint fragments and old iron nails. A possible tripod leg of pottery was also found. The bones derived from goat/ sheep, reindeer and an ungulate. The 7 m elevation renders a maximum date of ca. A.D. 1200. This coincides with the thirteenth century colonization of the region and founding of parishes in Pitea (Axelson 1989). Chronology The location of this island is consistent with many other Bothnian Iron Age sealing sites, such as Stora Fjaderagg in Vasterbotten. Four radiocarbon dates were obtained and the period of site use ranges from A.D. 200 to 1200. The oldest dwellings cluster between 13 and 16 m above sea level. The huts are also comparable by size and form to the Vasterbotten material, and also include storage facilities and cairns. The frequency of labyrinths on Stor- Rebben is comparable to Snoan and Stora Fjaderagg. Unfortunately, none of them could be lichen dated. It is also likely that stones from the hut walls were used to build the labyrinths and this can best be seen in the area of Hut B. A partial hut is actually abutted by a labyrinth, and a second laby- rinth lies just below it. There is an excellent Table 49. Stor-Rebben: Unburned bones by fragment. SPECIES HUT C HUT D TOTAL Sheep/goat Reindeer Ungulate Indeterminate Total 128 CHAPTER 5 Table 50. Stor-Rebben: Anatomical breakdown of bones. SHEEP/ ANATOMY ELEMENT GOAT REINDEER UNGULATE TOTAL Rib cage Costae 1 1 Front extremity Humerus 1 1 Rear extremity Femur 11 2 Tibia 1 1 Total 2 1 2 9 example of the relationship of older prehistoric features, in this case a Bronze Age grave cairn, and labyrinths at the nearby mainland site of lavre. The largest lichens on the labyrinth stones measure 155 mm and date to A.D. cal. i299±3o (refer Chapter 10). Finds Hut A: A gray flint chip, 14 tiny slag fragments, crowberry seeds {Empetrum) Hut B: Iron slag, a flint chip Hut C: Two grey flint chips, and a steatite piece with a groove and five slag pieces measuring 10-32 mm. Hut D: Flint chips, as well as iron nails and a possible unglazed tripod leg of pottery. Gray flint chips were found in all four huts and were probably the byproducts of strike-a-lights. Iron slag was found in Huts A, B and C. The frequency of iron slag in these coastal huts suggests that iron working was a common practice. The steatite piece from Hut C could be associated with metallurgy, and has a parallel in a find from Site 138 at Jungfruhamn. Osteological Material The bones from the Stor-Rebben hearths emanate from sheep/ goat, reindeer and an unidenti- fied ungulate. All of these were unburned. Only one bone could be identified from Hut C, a hu- merus from an ungulate, pre- sumably reindeer or sheep/goat. The size and shape is most con- sistent with that of a sub-adult reindeer. All the rest of the iden- tifiable material came from Hut D. In Hut D, a femur of an adult and relatively large reindeer was identified together with a thighFigure ij6. Map showing location of Hornslandsudde. Sites ug and i}2. EXCAVATIONS AND ANALYSES 129 bone and a fibula of a sheep/goat. There are also rib bone fragments. It appears as though meat- rich body parts had been brought to the island as food. While preservation of the burned material is poor, it is surprising that not one seal or fish bone was identified. But, because of the location, sealing and fishing had to have been the primary reason for these huts. Equally remarkable are the finds of sheep/goat and reindeer bones. These sealers may have had brought food with them, although animals could have also been grazed on the islands. Summary In conclusion, Stor-Rebben is fully comparable to the other sites in the project. The island reveals both an older hut-based hunting economy, dating to circa A.D. 200-1100, and a younger fishing- based economy from ca. 1300. lolin Kraft cites some evidence that Bothnian labyrinths could have been associated with Saami magic, but he also noted most were associated with fishing sites and of younger date (Kraft 1977). The Saami place-name Rebben coincides with the oral histories at Stora Fjaderagg/Holmon and Hornslandsudde that identify these places as Lappish/Saami camps. The Stor-Rebben investigation, although limited in scope, has provided important data regarding the chronology, technology, economy and place-name context of these Iron Age sites, including the transition to the historic period. Figure 137. Site area based on photogrammetry (Eriksson 1975^. Investigated areas in 200^ are shown in boxes. The archaeological features (hutfoundations) are situated between 12 and 25 m above present sea level. Huts were sampled at lyn. 16m. 18m. and 20m above sea level. 130 CHAPTER 5 Hornslandsudde Hornslandsudde (Site 119) was chosen for comparative purposes and is located only about 300 km north of Stockholm (Figure 136). The investigated region is in Gavleborg County, Hudiks- vall Municipality, Rogsta parish in the province of Halsingland at 61° 37'N, 17° 29' E. The site is located in an area of largely exposed wave-washed moraine beaches up to 25 m above sea level. The beaches have distinctive terrace formations. Shoreline displacement is currently 0.75 mm per year. Vegetation consists primarily of pine heaths with dry blueberry (vaccinium) type ground cover. Lichen vegetation is abundant, especially reindeer lichens, and a large herd of reindeer were brought to the area during the harsh winter of 2006-2007. Hornslandsudde has stands of very old pine showing traces of burning from a forest fire in 1888. The Hornslandsudde site was first published by Westberg (1964). Bjorn Ambrosiani (1971) excavated some hut foundations and stone alignments in 1966. Photogrammetry-based mapping was carried out in 1966 and 1973 (Eriksson 1975). A new site survey was reported by lonsson (1985) (Figure 137). Westberg relates the oral history of the Hornslandsudde area as follows: According to tradition, which is still preserved among the older population who practiced fishing at Hdlick's fishing village ca 2 km west of Hornslandsudde, the dwelling sites on the point derive from a fishing people of Lappish (Saami) origin. (1^6/^:2^) It should be noted that sealing was referred to as "seal fishing" in the Bothnian region [sjalfiske). The place-names Lappmon and Lappmoherget (Lapp Sand and Lapp Sand Mountain) are very close to this site (Westberg 1964:24). Westberg identified Hornslandsudde as a historically documented sealing place, which is also indicated in the place-name Sjallhallorna (Seal Rocks). According to a map for Rogsta parish from 1799, a fishing site on Hornslandet was named Lap- phack. Westberg has provided an excellent overview of the history of fishing sites in the area and the islands of Kuggorarna, Balson and Hastholmen (Westberg 1964). The provincial law, "Hel- singelagen," stipulated that one-tenth of all fowl, wild game, fish, moose and bears were to be paid in taxes, as well as every fifteenth salmon and fifteenth pound of herring, seals and gray squirrels. The priest and the Church would divide this equally (Westberg 1964:36). In 1545, Gustav Vasa initiated taxation of all fishing on the Bothnian coast. In the late 1500s, fishermen from the newly founded town of Hudiksvall established many fishing places in the region, although Agon was an exclusive herring fishery of the highly organized Gavle (i.e., urban) fisherman. The fishing harbors and labyrinths dating to the 1600s and 1700s at Kuggoren and Balson are close parallels to those of, for example, Bjuron in Vasterbotten. Salmon and seals were caught using net systems on poles at a number of locations and seals were also hunted on the ice. The Hornsland Peninsula had formerly been an island and although some lakes provided freshwater fishing, agriculture was extremely marginal. Westberg suggested that this provided a coastal sanctuary for hunter-gatherers in an otherwise Germanic- settled region. Wennstedt Edvinger and I have documented several Saami circular sacrificial features on Hornslandet, one on Yttre Bergon that is similar to an enclosure on Stora Fjaderagg and another by Arnoviken (Wennstedt Edvinger and Broadbent 2006). Neither was associated with harbor basins. Both sites are near Iron Age graves and hut areas. EXCAVATIONS AND ANALYSES 131 Figure ijg. Excavated area in Hut 5 (Jackelyn Graham). Figure 1^8. Hut ^from east showing double wall-lines. The excavations by Ambrosiani in 1966 did not produce any artifacts, radiocarbon samples or osteo- logical results. New excavations were therefore under- taken in 2005 to better determine the ages ofthe huts and recover any organic remains. According to Jonsson (1985), Site 119 at Hornslandsudde has 48 hut foundations and some 60 stone alignments. The alignments are very distinctive long and low walls of stones with openings that were probably used to catch forest birds using snares. In 2005, we investi- gated two alignments and four hut foundations at 20 m, 18 m, 16 m and 13 m above sea level (Figure 144). A second site area about 75 m to the east of Site 119 was also sampled. This site (Site 132) consists of three features: two hut floors below a rocky cliff to the north, and a circular storage cairn. They lie at about the 20 m elevation. Finally, 11 storage caches and cairns have been registered at the two sites. Feature 5, Hut Feature 5 is a rounded-rectangular foundation measuring ca. 3.5 x 6 m and walls measuring 0.2- 0.3 m height (Figures 138, 139). The hut is situated at 18 m above present sea level. This hut shows secondary walls indicating re-use. A hearth deposit with charcoal, burned bone and slag was found in the middle of the floor and investigated. The hearth was ca. i m in diameter and lacked a stone circle or any demarcation. The hearth is less than i m in diameter and, as seen by the profiles, less than 20 cm deep. The soil in the hearth was brown in color and the surrounding area had differing shades of brown. Some charcoal was found in and around the hearth, and samples were taken for radiocarbon dating. An abundance of highly fragmented bones was also found within and around the hearth. Fire-cracked rock (ca. i liter) was also found. Finds Iron slag (120 g) Six red clay pieces (furnace walls) (0.6 to 1.2 cm) One iron fragment (i.o cm). 132 CHAPTER 5 Figure 540. Trench through Feature 13 (foreground) from the east. (See Figure 154. j Figure 141. Section through connecting wall between Huts 12-1} where ruminant tooth fragments werefound. PHOSPHORUS Low L/M Medium Feature 12 Feature 14 Feature 13 @ ® Osteological Material 375 bone fragments (60 g) The bone material could only be identified as long bones from mammals. Chronology Charcoal from the deepest part of the hearth was radiocarbon dated, but had neverthe- less been contaminated: i40±40 B.P. (Beta 217790). This error is undoubtedly due to the forest fire of 1888. Because of this prob- lem, a bone was submitted to the Svedberg Laboratory in Uppsala for an accelerator dat- ing (Ua-32857), which produced an age of i390±30 B.P. (A.D. cal. 623-664). Figure 142. Features 12-14 showing soil samples 1-21. EXCAVATIONS AND ANALYSES 133 Conclusions Hut 5 dates to the seventh century, although double walls suggest it had been rebuilt. Of special interest is the fact that iron slag, fired clay and an iron piece, indicative of iron working, were found. This material is described separately, including analysis results by Andersson (2007), in Chapter 7. Features 12-14, Double Hut and Enclosure Features 12 and 13 are a double hut situated at 16 m above sea level. Feature 12 is rectangular in shape and measures 3.5 x 4.5 m with walls 0.1 to 0.2 m high. Feature 13 is approximately rectangular and measures 4x4m with walls 0.3 to 0.5 m high. Feature 14 is an irregular enclosure with lower walls measuring 5x7m (Figure 142). As far as could be determined, it probably was an enclosure. Only slight traces of burning were found in the Feature 13 hearth, not enough for a radiocarbon date. Ad- jacent to the hearth was a large cobble, which could have been used as a seat or as an anvil. A possible whetstone was found adjacent to it. A trench was run to cross-cut both hut floors and the wall between them (Figure 141). In ad- dition, the south end ofthe foundation, which protruded out about 30 cm, was sectioned (Figure 142). Fragments of a tooth from a ru- minant (sheep/goat or reindeer) were found at a depth of 20 cm. Hut 12 was found to have a much more uneven floor littered with a num- ber of medium-sized cobbles. A hearth was found up against the north wall, and this fea- ture was full of of iron slag and furnace clay. A magnet produced quantities of iron scales and sphericals that were by-products of iron smithing. In addition, a channel penetrated the wall on the right side of the hearth. Finds Iron slag, ca. 20, 1-5 cm (i.i kg) Iron scales, ca. 100, < i cm, (Gog) Osteological Material Tooth enamel from a small ruminant (sheep/ goat or reindeer); (numerous fragments (22g) from the same tooth, from maxillary). Figure 143. Excavation ofHut lyfrom south. FEATURE 25 FEATURE 19 N i [! test pit t A % e 1 S 1 «i 2m \±\ J e Figure 144. Map ofsections oftwo stone alignments (Jacquelyn Graham). 134 CHAPTER 5 chronology Two AMS dates were obtained: (Beta-209908) i57o± 40 B.P. (A.D. cal. 434-537), and (Beta- 207939) i820±40 B.P. (A.D. cal. 136-236). The hearth in Hut 12 was clearly used for iron working. The hearth measured ca. i m in di- ameter and lacked a stone circle. The channel-like opening through the wall next to the hearth was probably an air duct for a bellows and is a parallel to the wall opening documented at Bjuroklubb Site 67. No bones were preserved in the hearths of either Huts 12 or 13, although enamel from a ruminant tooth was found in the wall between the huts. Twenty-one soil samples taken across the area of these huts showed very low levels of phosphorus enrichment (0-50 P°). The only high phosphorus level is sample 11 from the corner of the dwelling, Hut 13. This is most likely a consequence of very poor preservation conditions and the shallow deposits at the site. The three features, 12, 13 and 14, form a unit consisting of a smithy, a dwelling/workshop and a possible enclosure for livestock. Figure i^y. Map oj Site i]2 showing Hut A. Hut B and a Figure 146. Site 1^2, Hut B.from southwest. cache. EXCAVATIONS AND ANALYSES 135 Figure 149. Reconstruction ofHut B, lean-to and storage cairn at Site 132 (Jacquelyn Graham). Figure 148. Cairn by Site 132, Hut B. Feature 15, Hut Feature 15 is a rounded-rectangular hut, measur- ing 4x5m and with walls 0.2 m high. It is situ- ated at an elevation of 12.5 m above present sea level. Hut 15 was trenched across its midline in order to expose the hearth (Figure 143). A shallow ashy deposit was subsequently found in the center of the floor. An additional i.o x 0.50 m unit was excavated to expose this feature. The trench was dug down to 35 cm below the surface across the floor. Only tiny bone chips and small carbon pieces were found in this hut. Even this carbon sample (Beta 210237) proved to be contaminated: 26o±7o B.P.. This hut was examined because of its elevation above the 12 m shoreline. No animal bone was obtained for analysis. It is the lowest- lying dwelling we investigated and is adjacent to the area of stone alignments. Features 19 and 25, Stone Alignments Ten soil samples were collected beneath cairns along two stone alignments. Features 19 and 25 (Fig- ure 144). All the phosphorus samples were low (0-50 P°). Potassium levels, suggestive ofburning, were high in all the test pits. This probably reflects the forest fire that had burned over the area in 1888. No reliable charcoal samples were obtained. Site 132 Two hut floors with small hearths and a storage cairn lie approximately 75 m to the northeast of the main hut site. The elevation is ca. 20 m above present sea level. According to a footnote in the survey records from 1982, a local informant believed these were connected with the military signal station that had been established nearby in World War II. Both huts, although not overgrown with lichens or vegetation, had small hearths with charcoal that could be collected among the loose beach stones. Burned animal bone was found in Hut B. Hut A, the smaller of the two huts, is lo- cated up against an outcropping and is bounded on the west by a wave-washed cobble beach. This 136 CHAPTER 5 Figure 250. Cache #4 next to bedrock outcrop, from south. Figure 253. Cache #1 in wave-washed moraine field, from west. hut measures ca. 4 x 4 m and contains a hearth (Figures 145-147). Charcoal was submitted for analysis but was only par- tially carbonized and was the by-product of later site use. This proved to be the case (Beta-217789): 210+40 B.P. Hut B is larger than, and located to the south of, Hut A. h measures ca. 6 x 4 m. The northern walls for both huts are bedrock/boulder outcrop- pings. The stones from Hut B form a pat- tern suggesting that two structures had stood there. One is an ovoid form with a cleared floor surface and was probably a bent-frame hut. The hearth was outside of this hut and up against the boulder. A sec- ond line of stones marks the line of a pos- sible lean-to (Figure 149). An open cairn measuring ca. 2 x 2 m in diameter lies Figure 252. Map ofcache locations, Sites iig and i}2. Hut 12a 182at40BP ^M^^ Hut 12b 1570±40BP -•l^- Huts 139aJ30BP ^ Hut 132B 113(M)BP ^ 1 , < < < 1 , , , < 1 , , , , 1 , , < < 500CalBC CalBOCalAD 500CalAD lOOOCalAD CalibiBted date Figure 253. Calibrated datesfrom Hornslandsudde. EXCAVATIONS AND ANALYSES 137 ca. 3 m to the south of Hut B. Its proximity to the dwelhngs impHes that it had served for storage and is interpreted as a meat cache. Chronology Charcoal was found in Hut B (Beta-210238) and dated to: 1130 ±40 B.P. (A.D. cal. 881-980). Osteological Material Hut B also contained bones that were found to be the proximal epiphysis of the second metatarsal bone from a large seal, probably a gray or harp seal, and a long bone fragment. Site 132 dates to the Viking Period. The seal bone found in Hut B is consistent with the inter- pretation of the site as a place for marine hunting. The cairn is interpreted as a storage cairn. This locale was well protected from winds and had good access to the former shoreline. Caches Five small stone caches were identified at the foot of a rocky outcropping less than 10 m to the north of the upper dwelling area (Huts 1-5) (Figures 150-152). These features appear as small pits and niches by the bedrock outcrop. A larger and deeper cache was found in a boulder field (Cache #1) in an area 60 m to the east of Hut 5. This well-constructed cache measured i.o to 1.5 m across and was 1.2 m deep. Additional caches were also documented on the site. MunsL-il Sol! Culoi: Ml = lOyr'Tvl Gray M2 - l()yr/6/6 Brownish Yellow M.i - !()yr/7/l l ight Gray V14 - lOyr/5/2 Grayish Brown M5 - lOyivfi/I Light Brownish Gray M6 = IOyr/5/1 Gray M7 = l()yr/5'2 Grayish Brown MS - 1 Oyr/5/2 Grayish Brown N t w Profile A Profile B 1 1 s = Slag ® ^ Carbon Sample - 20cm 3,5 Liters of Fire Cracked Rock Wall Floor Wall Figure 154. Map ofexcavation of Huts 12-1] at Homslandsudde. Site iig (Jacquelyn Graham). 138 CHAPTER 5 Summary and Discussion The results of the investigation show that the oldest huts at Site 119 date to the Early Iron Age. Site 132 dates to the Viking Period. Most of the bones recovered appear to be seal bone, although this material is very fragmentary. The identification of an adult gray or harp seal at Site 132 is interest- ing and corresponds to finds from the island of Stora Fjaderagg in Vasterbotten. This suggests open-water hunting. The huts are similar in size and form to the Vasterbotten material. Width is fairly uniform, 3-5 m, and length, 3-8 m, is more variable. The huts lie in roughly four groups ranging between 12-25 ^ above sea level and form loose clusters aligned parallel with the shorelines and facing south and south/southeast. Seven huts cluster along the 12 m level, four to five huts at the 14 m level, seven to nine huts at the 16 m level, five to seven huts at the 17 m level and two to five huts at the 20 m level and higher. It is not certain that the highest lying features are the oldest; net drying post cairns at the highest elevations suggest medieval or historic fishing. Site 132, Hut B, consists of three elements: a small hut, a hearth outside ofthe hut and a wind- break. Near the hut is a storage cache. A reconstruction gives an impression of what this dwelling site looked like (Figure 149). The distinctive storage caches are another close parallel to the Vaster- botten material. Some caches are adjacent to huts while others could have been more communal. The numerous stone alignments run at right angles to the shoreline and have been interpreted by others as net-drying features or boat slips (cf Westberg 1964:28). While these lines extend from 8 to 17 m above sea level, they lie in an area west of the main hut clusters and are not necessarily contemporary with them. Two stone alignments (19, 25) were sampled, but unfortunately the forest fire in 1888 left carbon over the whole site surface, thus contaminating superficial cultural layers. Soil samples indicate high levels of potassium on the site, probably a reflection of this fire. There are many similarities to the Vasterbotten region sites by chronology, function, cluster- ing, storage, and even by ritual (cf Wennstedt Edvinger and Broadbent 2006). On this basis, this archaeological material, together with the place-name evidence, speaks in favor of the theory that these had been hunter-gatherer (Saami) sites. But unlike Vasterbotten, there is a parallel Early Nor- dic Iron Age complex of house terraces, grave mounds etc. in the region (cf Liedgren 1992). These settlements in Rogsta parish were situated ca. 15 km to the north and above the 15 m level. Interest- ingly enough, this complex seems to have disappeared around A.D. 600. This raises the fascinating issue of the co-existence of two groups in the region, one a hunter-gatherer/herding/trading group of Saami and the other an agrarian-based/trading Germanic community. The Saami may even have been especially drawn to this region because ofGermanic settlement, and offered their hunting and healing skills and even specialized iron working. Conversely, the two groups could have occupied different ecological niches in the coastal zone with long-term Saami interactions in two familiar forms forms, first as hunter-gatherers and later as reindeer herders, but were probably also living in ways analogous to the Germanic farmers and fishermen. EXCAVATIONS AND ANALYSES 139 25 20 c 15 re > 0) 10 LU 5 500 1000 1500 Median date AD. 2000 Figure J55. Diagram illustrating the elevations of radiocarbon dates on the Bothnian shores and the gap between Iron Age sealing sites and medieval fisheries. Chronology and Culture This chapter provides an overview of the radiocarbon and AMS dates from huts between 20 m and 3 m above sea level. As noted earlier, this material encompasses 31 huts at 9 locales and 12 elevations. This sampling facilitates comparisons at different elevations above sea level at the same locales and comparisons between regions. There are five major questions: 1) When were these huts first used? 2) When were these huts last used.-* 3) Were there differences between regions? 4) What is the chronological relationship between the sealers' huts and medieval fisheries? 5) Are there patterns of use reflecting economic or environmental cycles? Although one standard deviation (68% probability) was used for the descriptions of individual site dates, the following analysis is based on two standard deviations (95% probabilities) of44 dates. Source Criticism There is always a risk that samples have been contaminated by old wood or recent forest fires. Charcoal analysis has indicated that pine was the most common fuel, although birch and alder were also burned. The Hornslandsudde site area has stands ofsome of the oldest pines in Sweden, and even dry pine branches can be several hundred years old. This said, elevations above sea level can be used to reject dates that are clearly older than their contemporary shorelines. There are only two potential dates in this latter category: Stor-Rebben Ai and lungfruhamn C, although they still fall within one standard deviation of their elevation dates. They are therefore considered as probable. In the case of Hornslandsudde, the shoreline association of Hut 12 is confirmed by the fact that the slag found in the hearth had been water-rolled (Andersson 2007). Two recent dates were obtained from carbon samples collected in 2004 from the partially disturbed cairn in Hut 4 at Grundskatan. These are so recent they must be rejected. A sample that was improperly stored from Snoan 49B is also suspect, but is perhaps evidence that the hearth had been re-used during historic times. Three dates from Huts 5, 15 and 132A at Hornslandsudde were clearly contaminated by recent forest fires and unusable. Finally, due to the earlier limitations on charcoal amounts necessary for dates, some results have very large standard deviations, namely Stora Fjaderagg B 141 (±315), Grundskatan 14 (±185) and Grundskatan 15 (±245). These are not terribly useful except within the framework of the material as a whole. A related question is whether or not the huts had been built no more than 1-2 m higher than their contemporary shorelines. This depended on the topography of each site, but a correlation coefficient of 0.73 between elevation and date shows that this assumption is largely true. While radiocarbon dates have a high correlation with elevation dates, the re-occupation of older hut areas during the tenth and eleventh centuries is also evident. Huts from this period lie as much as 20 m above sea level. Huts by harbor basins are much more shore-bound, as can be expected, although these huts can also lie on higher ground. For instance, a hut dating to the sev- enteenth century on Stora Fjaderagg Island lies 5 m higher than the harbor basin at Gamla Ham- nen. In spite of these concerns, consistent results have been obtained from different localities and elevations along the transect Site Comparisons The individual calibrated dates were sorted by increments for each site in order to see how they were spatially distributed (Table 51). This does not show how many dates were obtained at a given site, but which time periods are represented. Three periods dominate: A.D. 100-400, A.D. 400-600 and A.D. 800-1000. Together these horizons comprise 81% of the total. Looking at the material as a whole (Figure 156), the dates appear to cluster in four succes- sive steps. The oldest cluster ranges from 1845 to 1660 B.P. The next cluster is from 1570 to 1390 B.P., the third is from 1300 to iiio B.P and the fourth is from 1045 to 880 B.P. The weighted averages of these clusters and their ranges within two standard deviations are shown in Table 52. The two final age groups are separated by only 55 years and correspond to the early and late Viking Period. This gap is probably not significant. There is a significant gap of 170 years, how- ever, between the second and third groups, or between A.D. 608 and A.D. 778. Groups i and 2 are separated by 105 years, which is also significant. These numbers support the conclusion that the clustering is real. Table 51. Chronological horizons by site. SITES (N-S) 100-400 400-600 600-800 800-1000 1000-1200 Stor-Rebben Bjuroklubb; Site 70 Jungfruhamn Grundskatan Stora Fjaderagg Snoan Hornslandsudde X X X X X X X X X X X X X X X X X X X X X 19% 29% 10% 33% 10% 142 CHAPTER 6 Stor-RebbenAl lS45j403BP Homsktbudde 12a 1820±40BP- JungftuhannC 17iatl2§BP StauFjackragiBl 1660t70BP Hm^kuickidct 12b 157(MBP QiDTdstoT3a ISOOtlOOBP Stor-RebbenA2 14)5±70BP Bjuroklubb67 1485i:70BP Hmislancbiidde 5a 139ftt3(BP Ju^iilTami B 130Qtl3()BP JLn¥Ji\iliamiA2 12iat50BP Gruicfekatan3b 120St70BP aunciskatanl3c 120CWOBP Gruiicbkatan 1 1 ll75tl(X)BP Cfuxfekatan 1 lieOtDBP Sn6an49A ILSitiaBP Hon-Blanckidcie 132b ll3ftbttBP StoraFjaderaggD lllOtWSBP Q\in(bkata]i4c 1080*45 BP StoraFjaderaggA 1015tl00BP Stor-RebbenC 99StllOBP StoraFjadera^C 955t75BP Quixfekatan 13b 88(M0BP lOOOCkBC 5(XX:alBCCalBC/CalAD500CalAD lOOOCaJAD 1500CalAD 2000QilAD Calibrated date Figure 256. Representative sample 0/24 dates in orderfrom the oldest (top) to the youngest. The datesform a step-like progression. Table 52. Clusters of hut dates (weighted averages). C-14 CLUSTERS AVERAGES (2 S.D.) 1) 1845-1660 B.P. 1776±33B.P. A.D. 134-334 2) 1570-1390 B.P. 1513±26 B.P. A.D. 439-608 3) 1300-1110 B.P. 1175±20B.P. A.D. 778-941 4) 1045-880 B.P. 979±30 B.P. A.D. 996-1154 CHRONOLOGY AND CULTURE The Start Dates The luimbers of dates by century show that this form of hut-based seahng began in the early second century A.D., had a peak in the sixth century, a sUght decUne in the seventh and eighth centuries, reached its greatest peaks in the ninth through eleventh centuries, and declined sharply in the twelfth century. The End Dates The youngest dates at the five main sites range between 1130 and 880 B.P. One of the Grundskatan dates has a large standard deviation, but even discount- ing this, four dates calibrate to the thir- teenth century at the latest. The year A.D. 1279 (discounting St. 11174) can be taken as the best measure with 95% probability for the terminus ad quern of this North Bothnian hunting system. The difficulties of obtaining uncontam- inated dates at Hornslandsudde have led to only one date from this latter period. Figure 157. Numbers ofdates of Bothnian sealing huts (median values of calibrated dates). Summary and Discussion Hut-based Bothnian sealing began in the late first or early second century A.D. and all sites had been abandoned by A.D. 1279, although most sites came into disuse in the late twelfth century. Comparisons along the north to south transect show the same occupation trends. There are few differences between sites or between regions; they were part of the same phenomenon. Clusters of radiocarbon dates fall into three main periods: A.D. 100-400, A.D. 400-600 and A.D. 800-1000. There were peaks in the sixth and the ninth through eleventh centuries A.D., and declines in the seventh and the late thirteenth centuries A.D. that correspond to widespread settlement regressions in the Nordic region and the Little Ice Age. The Stalo hut sites in alpine areas of Sweden and Norway (Manker i960; Kjellstrom 1974; Storli 1991; Mulk 1994; Bergstol 2004; Liedgren et al. 2007; Bergman and Liedgren et al. 2008) Table 53. End dates of the sealing sites. SITE DATES RANGES (2 S.D. Stor-Rebben Jungfruhamn Grundskatan Grundskatan Stora Fjaderagg Hornslandsudde 945±110 B.P. (St.11180) 985±70 B.P. (St.lll76) 1000±185 B.P. (St.lll74) 880±80 B.P. (St.lll72) 1015+100 B.P. (St.lll81) 1130±40 B.P. (Beta-210238) A.D. 886-1278 A.D. 885-1214 A.D. 656-1305 A.D. 1013-1279 A.D. 788-1221 A.D. 779-994 144 CHAPTER 6 offer interesting parallels to the Bothnian sealing sites. Both were based on the specialized exploi- tation of marginal environments. The ecology of the alpine regions is remarkably similar to the outer coasts, especially in regard to vegetation. The Stalo huts are found from the Tornetrask area of northernmost Sweden to Frostviken in Jamtland in South Lapland. There are some 500 regis- tered huts in Sweden, about the same number as the Bothnian huts. They vary in size and shape from oval to rectangular and often occur in clusters of three to seven sod foundations, frequently aligned in straight or bowed lines. Bone caches, hearths and pits are common. Their hearths are often indistinct but can have a passjo/boassjo stone by the hearth. There are few artifacts, mostly flint or quartz chips, whetstones and iron slag. These huts were associated with seasonal reindeer hunting and/or herding, although some appear to be more permanent. Radiocarbon dates have ranged from A.D. 400-1600, but a recent source-critical analysis of 22 huts at 12 sites shows them to mostly date to A.D. 640-1180 (Liedgren et al. 2007). These dates overlap with the main period of the Bothnian huts and were probably results of the same historical and economic processes affect- ing larger regions of the Nordic North. Radiocarbon dates of oval and rectangular Saami hearths in the forest lands of Southern Lapland add a further dimension to this discussion (Hedman 2003). These date to as early as A.D. i, but mostly date to the Viking Period. They were associated with Saami kata/goahte huts or tents. Rock-filled hearths were associated with the more permanent katas. These hearths occur in alignments of up to 10 hearths, probably accumulations of two to three tents or huts at a time representing sijddas, or family hunting groups (Bergman 1991; Bergman and Liedgren et al. 2008). During the Viking Period there was a change in site locales from the river valleys, where they coincide with Stone Age sites and finds of asbestos pottery and transverse-based projectiles, to reindeer grazing areas by bogs, springs, streams and small lakes. This is interpreted as a change from a hunting-, gathering- and fishing economy to a semi-nomadic herding economy, and a shift from collective to individual forms of property and animal ownership (Hedman 2003; Bergman and Liedgren et al. 2008). There are a number of artifacts with parallels in Saami offer sites, including coins, hack silver and weights. There is also iron slag indicative of forging (Hedman 2003:161-189). While these finds confirm the associations of the hearths with Saami metal offer sites, the hearths did not disappear in the fourteenth century, but instead increased in numbers. This can mean only one thing: Saami population density increased in the interior. This is probably the beginnings of larger winter villages as described by Tegengren for Kemi Lapmark (cf Tegen- gren 1952; Bergman 1991; Mulk 1994)- This material reflects a number of trends of relevance to the North Bothnian coastal mate- rial. The first is the fact that this manifestation of Saami settlement within the sijdda system goes back to A.D. i, and that this pattern coincides with Stone and Early Metal Age settlements, which is reflective of long-term continuity. The second is the expansion of this system, including linear alignments ofhearths that relate reindeer pastoralism, trade and the accumulation ofwealth during the period A.D. 700-1100. And lastly, there was an increase in hearth density in the interior from A.D. 1300. As discussed in the following chapters, these patterns are a result of larger-scale changes in the Nordic region relating to state formation. What Do the Artifacts Tell Us? Inga Serning (1956, i960) has discussed the Iron Age artifact material of Upper Norrland within the context of Saami offer sites and these studies constitute major points of reference for understanding CHRONOLOGY AND CULTURE 145 coastal chronology. Inger Zachrisson (1984) has published additional material in this Iron Age con- text, as has Thomas Wallerstrom (2000), Lillian Rathje (2001) and Sven-Donald Hedman (2003). Coastal artifacts from the period before A.D. 600, including the Storkage find (that can be an offer site) from ca. A.D. 350 and the Javre grave find, a wheel-shaped ornament from a grave cairn in southern coastal Norrbotten (Broadbent 1982:154-255), are mostly of East Baltic, Finnish or central Russian (Volga-Kama) origin. The chronology of the Germanic longhouse settlement at Gene at the northern limit of Ger- manic settlement on the Bothnian coast shows it was occupied continuously from ca. A.D. 100/200 to 500/600. This settlement was abandoned around 500/600 along with a widespread regression in the cultural landscape, which is also seen in north and southwest Norway, the islands of Oland and Gotland, Ostergotland and the Malar Valley (Ramqvist 1983:194). In Medelpad and in Halsingland (and the Hornsland region), there is an almost identical pattern with farmsteads from the second century A.D. (Broadbent 1985) and widespread settlement and landscape abandonment at ca. A.D. 600. Although the region still had farmsteads, population did not rebound until Late Viking and medieval times, A.D. 1100-1300 (Liedgren 1992:191-219). Trade was intense during the Early Iron Age especially when the Germanic chiefdoms of the Mid-Nordic region (Angermanland, Medelpad, Jamtland, Trondelag and Osterbotten) reached their peak. This trade brought Roman goods north- ward and trapping pits were dug by the thousands in the interior to harvest reindeer and moose (Selinge 1974; Spang 1997). Artifacts dating to A.D. 800-900 that are of Scandinavian origin are few in number in the northern coastal zone, but a pair of round brooches was found in a grave containing the bones of an adult and a child at Obbola near Umea. These objects derived from southern Scandinavia but the use of these types in pairs was more of a Finnish custom and the grave form is typical of the Vasterbotten coastland. A contemporary find from a grave from Luopa in Osterbotten contained buttons/bells of the same types found on Stora Fjaderagg Island (Christiansson 1969:197-210). During the period A.D. looo-iioo, the artifacts were mostly of eastern origin. This is also the main period of the Saami metal offer/deposition sites in the interior. As offerings, these beauti- ful and rare objects sanctified the relationships of northern peoples to their own gods, and to the Nordic gods, when this seemed expedient. The 1200s mark a major change and Western European finds became more common, presumably because ofGerman (Hanseatic) trade. The 1300s were the effective end ofthe metal offerings, although some sites contain objects from later periods (Serning i960: 67-94: Zachrisson 1984:119). The main period of Bothnian seal hunting sites conforms well to the main period of the metal offer sites and the Stalo huts, ca. A.D. 800-1100, and declines at the same time, ca. A.D. 1250-1350. The coastal connection to the enormous geography of Saami trade, including the goods found in Saami graves throughout the Nordic North and in local graves on the Bothnian coast, entails that Bothnian seal hunting must be viewed in the same super-regional context. Most au- thors (e.g., Serning 1956, Fjellstrom 1985, Hansson and Olsen 2004) have related the changes in eastern and western trade items to middlemen from Finnish Karelia, Russian Novgorod, Birka and Sigtuna in Sweden, Gotland until 1361 (when Gotland fell to the Danes), the rise of Hansa fish- eries, particularly through mercantile centers such as Bergen and Vagen in Norway, and the fur market in Tornea on the Finnish border. Lars-Ivar Hansen (1990) and Thomas Wallerstrom (1995, 2000) have expanded the documentation of these historically known forces in northern trade and 146 CHAPTER 6 mercantilism, including hack silver, a primitive from of currency. Hack silver and part of a scale were found at the Saami offer site of Unna Saiva (Serning 1956). These middlemen groups were undoubtedly important, but Serning proposed a different perspective. She suggested that the Lapps, "who had lived by Lake Ladoga and on the shores of the White Sea and far down in southernmost Finland with all the natural connections to the east from whence many of these objects came, had surely passed trade goods directly from Lappish to Lappish hands" (Serning 1956:105). Serning also discussed the circumpolar shamanistic context of this material, including evidence of Saami drums during the Viking Period. The fact of the matter is that the east-west bands of metal artifacts and metal offer sites in northern Sweden track along the same valleys and eskers as sites from earlier periods (cf. Figure 192). The constellations of objects, technologies and ideologies are remarkably parallel to those of the Stone and Bronze Ages, including connections to Finland, the Baltic and southern and central Russia. In other words, this network and redistribution pattern of goods and social obligations was not new but had been an established network among local societies for hun- dreds if not thousands of years. The chronological results from the Bothnian coast add an entirely new dimension to the nar- rative of the Saami trade: direct and independent involvement by the North Bothnian Saami now seems highly probable. The rise to power of Finnish/ Kvennish and other middlemen and tribute collectors is actually one reason why the coastal Saami sites were abandoned, as argued by Steckzen (1964), not the other way around. But the political power of emerging states and the consolidation of that power through the Christian Church are what ultimately changed the balance of trade and social relations. Grundberg (2006) describes this as the Europeanization of the north. According to resilience theory, the thirteenth century was a period of release of social, cultural and economic capital, and the fourteenth century was the reorganization of these northern societies into the forms we have today. This convergence of social, political, religious, economic, epidemiological and climatic factors made this one of the major turning points in North European history. This further entailed the formation ofnew polities within Saami society itself, and new sources of internal com- petition for limited resources. The Medieval Transition An important question is that of the relationship between the establishment of the North Both- nian fishing sites and the abandonment of the seal hunting sites. While they occupy completely different places in the coastal landscape, they probably overlapped in time. Although I have previ- ously argued that the sealing sites were largely abandoned by then (Broadbent 2006), there was a transition of sorts between these two kinds of economies. The dates of harbors on Snoan and Stora Fjaderagg indicate that these were established during the period A.D. 1200-1300. This agrees with the chronology of other harbors and centers along the Bothnian coast, in particular the har- bors of Kyrkesviken in Angermanland (A.D. 1220-1300), Saint Olafs hamn in Halsingland (A.D. 1300-1620) and Kyrkudden in Norrbotten (A.D. 1300-1620) (Grundberg 2001, 2006). Another interesting aspect of this medieval material concerns the labyrinths that were closely associated with fishing sites. The lichenometric datings obtained in the project bracket them into the time period A.D. 1300-1850. The labyrinths and compass roses on Stora Fjaderagg Island date to A.D. 1456-1660. The lichen-dated labyrinths from Snoan Island date to A.D. 1388-1816 (Broadbent and Sjoberg 1990). CHRONOLOGY AND CULTURE 147 30 1000BP Q, R' 800BP 8" ID Q. rD rD I O'3 600BP 400BP 200BP OBP -200BP mndskatan 13ii 3A lOOOCalAD oleS mm ISnSan ! 'Srt6ah 92B ISOOCalAD Calibrated date 2000CalAD Figure 15S. Graphic distributions (2 s.d.) of the latest sealing hut at Grundskatan. a farmstead locale and a Saami hearthfrom Bjurdklithb. rnnvkkatan HR SSQi^^ORP ^Ali^^ Bole 1 760i40BP SmanS^A 'R'ytVmP • Bole 2 61 li35BP Snoan92A 47ftt95RP ^ ' Snoan92B 430±95BP ^ nii^. Oinii Hamien ^IfttlOSRP A Oval Hearth (Site 67) 230i40BP ^ —A _i_ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 CalBQCalAD 500CalAD lOOOCalAD 1500CalAD 2000Ca]AD Cahlxated date Figure 359. Calibrated dates (2 s.d.) ofthe latest sealing hut at Grundskatan, the Bolefarmstead site, Gamla Hamnen on Stora Fjaderdgg Island and a Saami hearth on Bjurdklubb. CHAPTER 6 The youngest date from Grimdskatan 13 statistically overlaps the oldest harbor date from Snoan Island. Rathje (2001, 2005) excavated farmsteads with brick fireplaces and field stones at Bole in Lovanger in Vasterbotten and obtained three dates that are shown in Figure 159. They overlap with Snoan 53A. The place-name evidence also speaks in favor of this interpretation. The frequency of "Lapp" place-names on the coast is presented in Chapter 9. Even the oral histories of places like Holmon and Hornslandsudde speak of the Saami transition to permanent farmsteads and probably to fishing (cf Sandstrom 1988). Cycles of Change The material presented here shows cyclical changes, and although it is not possible to relate these variations solely to temperature, the fluctuations in sealing sites coincide with global and European climate and environmental cycles. The cold and wet Sub-Atlantic period at 600 B.C. ended about A.D. 270, which is a Global Climate Boundary according to Lamb (1977). Erom A.D. 270 to 450, the global climate became warmer and drier and between A.D. 600 and 690 the climate cooled considerably (Lamb 1977; Stuiver and Kra 1986; Denton and Karlen 1973). This was marked by agrarian settlement regressions throughout the Nordic region. The climate started warming again in the eighth century and reached a maximum (the Medieval Warm Period) around A.D. 1100-1150, but it started to cool by A.D. 1200. It also became wetter and glaciers started expanding, as did sea ice around Iceland (Granlund 1932; Lamb 1977; Stuiver and Kra 1986). The Little Ice Age (LIA) began around A.D. 1300 and lasted until 1850. Starting in 1315, there were widespread crop failures and famines in Europe. Erom 1300-1350, fishing replaced cereal crops as the main food source in Iceland, and 1408 was the last record of Norse settlement in Greenland (Grove 1988; Eagen 2001). Warmer and more stable temperatures certainly facilitated farming and animal husbandry, and lessened ice conditions may have created exceptional opportunities for sealing. The finds of harp seals on Stora Ejaderagg show that such was the case during the Viking Period. Other ice- dependent seals, such as the ringed seal, may have been highly concentrated in near-shore areas. The Bothnian hut sites were, nevertheless, completely abandoned in the late thirteenth century with the onset of the LIA. Using tax records, Kvist (1988:89) has documented the dramatic declines in Bothnian sealing during even later LIA periods in Osterbotten. As in Iceland, fishing became much more important than farming and the Church mandated fish as a necessary part of household economy. Eor many Saami, reindeer herding took on new significance, and the colder climate was probably in their favor; as for the Christianized Saami/Swedish coastal settlers of Norrland, fishing likewise gained in importance. CHRONOLOGY AND CULTURE 149 HORNS. SNOAN STORAF. GRUND. /BJUR./JUNGF. STOR-R A.D. Figure 160. Chronology oj Bothnian sealing huts. The Archaeological Roadmap Architecture Bothnian hut foundations occur in a variety of forms and varied configurations. Except for huts with iron forges, hearths were centrally placed. These fireplaces often have one or more larger cobbles near them, but are generally without stone fillings or rings. Huts without hearths were probably not dwellings and were rather used for storage of equipment, food supplies, etc. A special type of hut measuring on average 3x3m has been interpreted as a shed and is usually found near dwellings. They are comparable in size to get-kdtas (goat huts) among the Forest Lapps of Sweden (Manker 1968:204; Stoor 1991). These small huts lack hearths but have doors and were intended to keep animals warm and safe from predators at night (Figure 163). A sample of 61 dwelling floors on the coast indicates that floor size varied between 3.5 and 6.1 m in length and between 2.9 and 5.0 m in width. Mean length is A^rj+o.'yj m and mean width is 3.73+0.47 m. Length was more vari- able than width (Figure 161). This is logical as structural size could be most easily expanded on either short end, as opposed to broadening width (cf. Liedgren et al. 2007). In only one instance has a posthole been identified and this was too small to have been a roof support. The roof and wall supports therefore probably consisted of internal frames lodged against or set into the stone foundations. Except for the smallest temporary dwellings that might have had skin coverings, walls and roofs were probably constructed of timber, grass sod, skins, birch bark and combinations of these materials. Drift timber was readily available on the coasts because of the continuous outflow of rivers through the forested interior and into the Gulf of Bothnia. The use of timber in construc- tions has been documented very early in coastal Finland (cf Ranta 2002) and in northern Sweden at sites such as Lillberget (4200 B.C.) in coastal Norrbotten (Halen 1994). Timber huts are also well documented among the Forest Saami (Manker 1968). It must be assumed that this building tradition has a long indigenous history in northern Sweden. Bothnian coastal foundations range in shape from round to oval, and square to rectangular. When viewed in terms of the elevations above sea level, the oval forms are the highest lying, and their antiquity is borne out by the radiocarbon dates from Stora Fjaderagg, Stor-Rebben, Horn- slandsudde and on Bjuron (A.D. 200-600). The oval huts are, nevertheless, found at all levels. Rectangular and square foundations are not as frequent at the highest elevations and are most commonly found at levels dating to A.D. 700-1200. There is a greater variety of forms after A.D. 700, including "inverted Gs", row houses and dwellings with internal chambers, platforms, porches or lean-tos. Row houses consist of both rectilinear and curved-wall constructions with up to nine 151 Figure 161. Lengths and widths ofBothnian huts. ROUND/OVAL SQUARE/RECTANGULAR 100 200 300 400 500 600 700 800 900 1000 1100 1200 ADcal (1 sd) Figure 162. Hutform by calibrated radiocarbon date. The oval/round huts and the rectilinear huts overlap by chronology and are often found together Tlie oldest lying huts are oval inform and their antiquity is confirmed by radiocarbon chronology. Tlxe rectilinear huts were most common during the Viking Period. Tent rings are fixquentlyfound at the lowest levels. Figure i6j. Saami goat hut used to keep both goats and reindeer safe during the night (Manker ig68). Severalfoundationsfor these small huts were documented at the Grundskatan site, one of which was attached to a storage cairn with a central chamber separate rooms (Figure 160). A final category of dwelling is that marked by simple stone circles with evidence of hearths (tent rings). These circles are also found at higher elevations but are most frequent below the 10 m elevation. Following in- terpretations from North Norway, oval and round huts are of Saami design, whereas the rectilinear constructions could be the result of interactions with non-Saami. Grydeland has characterized the shift from oval to square dwellings as the transi- tion from a hunter-gatherer/pastoralist economy to a farmstead/fishing economy at Kvaenangen in North Norway, a fjord with three coastal siidas. These changes occurred during the period A.D. 1200-1700. The settlements nevertheless go back Figure 164. (clockwisefrom left) Saami hut with cattle in North Norway (Leem iy6y); coastal hutfoundations at Kvceangen, North Norway (Grydeland 2001:2^); Saami fishing village ofsod huts on Kildin Island, Kola Peninsula (Jan Huyghen Linschoten 1594-1595)- to A.D. 700 (Grydeland 2001:65). These huts are very similar to the North Bothnian huts, including the inverted "G" forms (Grydeland 2001:25) (Figure 164). Odner (1992, fig. 47) has documented similar kinds of forms and arrangements of dwellings in Varanger that are contemporary with the Bothnian sites. Sheep/goat bones have also been dated to A.D. 1000 in Varanger (Schanche 2000), and comparable settlements have been documented by Andersen farther south at Ofoten dating to the Late Iron Age. Even the cultivation of barley has been demonstrated at Ofoten (Andersen 1992: Hansen and Olsen 2004:197). Grydeland (2001:61) has seen this change in hut forms as a shift from collective to individual property ownership, a change that was formalized by the church and state, and there are even implications regarding gender relations; women lost status, and there were dramatic population increases. Site Structure Since North Bothnian coastal features were constructed over a long period of time, some dwellings and areas were inevitably re-used. In spite of this, rough clusters of structures are still visible at the same elevation levels. Although somewhat impressionistic, the huts can be grouped into shore-level clusters at the larger sites. THE ARCHAEOLOGICAL ROADMAP 153 1) At Stor-Rebben there are clusters at ca. 13 m and 16 m. 2) At Hornslandsudde there are clusters at ca. 12.5 m, 14.5 m, 16 m, 18 m and 20 m. 3) At Stora Fjaderagg Island there are clusters at ca. 11 m, 13 m, 15 m and 20 m. 4) At Grundskatan there are clusters at ca. 12 m, 14 m and 15 m. At Stor-Rebben there are four clusters of two to three huts at the 16 m level and six clusters of two to three huts down to the 13 m level. At Grundskatan there are five groups of five to seven dwell- ings; at Stora Fjaderagg there are five clusters of five to nine huts each, and at Hornslandsudde there are five to six clusters oftwo to nine huts at each level (Figures 165-167). While there is good reason to believe that these clusters represent repeated visits over time, the overall pattern suggests that they consisted of contemporary household groups of 15-25 people. A cluster of three to five house- holds was probably the norm and this is, in fact, the most commonly seen number of huts at the smallest locales (cf Figure 168). Some of these dwellings could have also been occupied year-round, such as Hut B on Stora Fjaderagg Island. Based on the organization of historically known sealing expeditions in the North Bothnian region, each hunting team consisted of the male members of households, and the teams ranged in size from five to eight households. In the fall and winter, teams of up to eight men in two boats would use 20-30 nets for catching ringed seals (Hamalainen 1930). Saami seal hunting on the Kola Peninsula, salmon fishing on the Tana River in North Nor- way and even bear hunts were similarly organized (cf Fellman 1906; Ekman 1910:252; Grydeland 2001:79). Collective efforts can also be assumed regarding the use of trapping pit systems on land that required a good deal of labor to dig and maintain. As will be discussed regarding tax records from the sixteenth century and the organization of church towns in northern Sweden (Chapter 9), this clustering offers a blueprint of territorial organization as well. GRUNDSKATAN o '.cO -e 5-7 Figure 165. Clustering of huts at Grundskatan. LABYRINTH'' (fjg) ' 5-7 "GOAT" HUTS „„ 154 CHAPTER 7 Figure 166. Clustering oj huts at Stora Fjdderagg. Square indicates area of circularfeatures. Figure i6j. Clustering ofhuts at Hornslandsudde. Numerous stone aUgnments lie at right angles to the shorelines in the center and southwest comer of the mapped area. THE ARCHAEOLOGICAL ROADMAP 155 caim / Figure 168. A small cluster of dwellings, middens and associated features dating to the period A.D. cal. 800-1200 at the Grundskatan site. This constellation corresponds to the Saami sijdda, a group of households living and working together. Storage and Surplus Storage pits were identified at all sites. They range from small lined pits in hut floors to chambers measuring a meter or more across. Large cairns or pits were observed near dwellings at most sites, often with piles of stones that had been removed to empty them and to subsequently cover supplies or foodstuffs. The largest of these cairns were documented at the Grundska- tan site. They consist of stone constructions measuring up to 4 m in diameter and run in a line between two clusters of dwell- ings (Figures 85, 86). These types of arrangements can be as- sumed to have been communal facilities. A large well-made pit in a moraine boulder field between Sites 119 and 132 at Hornslandsudde was also probably a shared storage facility of this kind (Figuie 169). Historic Saami storage features are well known and consist of above-ground structures, earth cellars, cairns and stone chambers used for storage of milk products, fish and meat (Manker 1968:203; Ruong 1969:128-130; Val- tonen 2006:64-74). There were also different types of wooden constructions, including small huts, as well as out-buildings and sheds. Foundations of these types of sheds or storehouses, huts without hearths, are com- mon on the Bothnian sites. While reindeer domestication has been given great significance regarding northern societies, the significance of storage has been little discussed. Both activities in many respects represent the same goal of securing and con- trolling the distribution of re- sources. The environment itself was a storehouse for immedi- ate returns and pastoralism was "storage on the hoof." Caches rep- resent the social appropriation of these resources (Ingold 1983). Storage can consist of household supplies, emergency stores and fixed-point storage, but larger depots imply community investments. Caches are thus important expressions of shared social space, access and distribution. They are a form of "resource husbandry." Storage also fosters sedentism and was a precondition for trade and the integration of local hunter-gatherer societies into wider systems of exchange and redistribution (Ingold 1983). Collective hunting and storage efforts further generated a need for leadership and Figure 269. Reconstruction of site complex near Site 70 on Bjuron. 156 CHAPTER 7 coordination. The abundance oflarge and small storage facilities at the Bothnian sealing sites should therefore be considered as one of the most significant socio-economic indicators at these locales. The osteological results from Stora Fjaderagg show that large numbers of seals were taken and processed beyond immediate consumption needs. A substantial pit (measuring lo x 12 m) at Grundskatan (Figure 9) is interpreted as a large-scale blubber-rendering basin analogous to those known from the White Sea region (cf Tegengren 1965). Taken together with the dwellings, the caches and blubber- rendering pits are evidence of the extensive exploitation of resources and the appropriation of these resources by local communities for systematic intercourse with the outside world. Corrals and Fences Stone alignments were documented at Site 70 and Grundskatan on Bjuron, on Stora Fjaderagg and at Hornslandsudde (Figures 84, 106, 142, 169). These walls were all attached to dwellings or were near to them and bear a resemblance to the field, corral and pasture walls (Swedish stenstrangar) known from Gotland, Southwest Norway and elsewhere (cf Lindqvist 1968; lV[yhrei972; Carlsson 1979). Phosphate testing of these features at Grundskatan and Hornslandsudde did not reveal any enrichment as compared with surrounding areas; in fact at Grundskatan there were higher val- ues outside the enclosure. Based on ethnographic accounts of Saami settlements, fences could be erected as corrals, as well as to keep reindeer offhuts and out ofgarden plots (Ruong 1969:130; Kjell- strom 2000:88). They were also used for milking and as temporary holding (marking) compounds. Obviously, most of these were seasonal hunting sites and not suited as permanent farmsteads, but they do reveal the existence of animal husbandry in the region, and place-names provide valuable clues as to where the more permanent settlements were located. Unfortunately, the likelihood of actually finding preserved wooden structures and fences is small but not impossible in boggy areas. Figure, lyo. Various possible hutforms and constructions at Grundskatan: (A) sod construction, (B) beam and timber construction. (C) bent-frame construction. (D) timber storehouse, (E) goat hut, (F) corral/fence. (G) stone cache. (H) meat rack. Based on KjeUstrom (20oo:8S-u'^). THE ARCHAEOLOGICAL ROADMAP 157 Seals, Seasonality and Animal Husbandry The finds ofanimal bones and teeth from the individual sites are presented in Chapter 5. From this material it is possible to discern general patterns relating to seasonality, hunting and animal hus- bandry. Preservation conditions were very poor and varied considerably from site to site, therefore this material only represents what happened to be preserved in hearths rather than what may have been a much more varied array of activities. Based on the numbers of identified specimens (NISP), a total of 4,539 bones were identified from two species of seals, a brown bear, sheep/goat, reindeer, cattle, hare, ducks, birds and large ungulates, probably moose. Ninety-eight percent of the bones are from seals. Of these, only 3% could be identified to species: 52% were from harp seals, which were found at two sites, and 48% from ringed seals, which were found at three sites. The harp seal bones were mostly from Hut B on Stora Fjaderagg Island. This material has been discussed in detail by Ian Stora in Chapter 5. His conclusions are: i) North Bothnian sealing during the Iron Age was more varied than previously believed, and 2) Hut B has to be regarded as a more or less permanent structure in a well-organized hunting complex. Seal bones were found at all sites except Stor-Rebben, which can only be explained as the result of poor preservation. The seal bones display considerable variability as regards anatomical representation in different huts. Of 4,439 bones, 34% derived from the cranium, vertebrae and rib cage, and 55% were from the extremities, particularly the flippers. The latter are often considered to be delicacies by seal hunters. The best-preserved seal bones from Stora Fjaderagg Hut B also show selectivity within the dwelling and that large numbers of seals were processed elsewhere. This sug- gests that the bones from the hearths represent "disposable meals" tossed into hearths, rather than what may have been prepared in other ways. Ringed seals are known to have been hunted on the ice of late winter, although there is also ample evidence that these smaller seals were caught using nets in the dark fall months. Seal skins, blubber and meat were of the best quality during this time of year. Female ringed seals fast in the spring, and unless the goal was to obtain their cubs, they were thin and generally sank quickly. Until firearms were used, ice hunting of ringed seals focused on single breathing holes, required great patience and skill, and the use ofdogs to find them in pressure ridges. It can be assumed that this species of seal was hunted during both the fall and the spring, but seal netting in the fall was more efficient and rewarding. The age determinations of the seal bones indicate mostly adults and sub-adults, consistent with seal netting, but this can be biased as younger individuals would not have been as well preserved. Harp seals do not breed in fast ice and are considered an open-water species that was hunted in the summer or fall (Stora 2001b). Eighteen bones of birds were found at three sites, two of which are from ducks. These ani- mals were probably killed during the summer or fall. Finds ofhare bone suggest winter hunting, as they were most easily trapped in the snow (Kjellstrom 1995:55, 212-273). Bones of larger ungulates, moose or reindeer, were foiuid in several hearths and can reflect fall or late winter hunting. On the whole, the seals, birds and small mammals indicate site use during the late winter, spring, sum- mer and fall. The bear bones derived from one individual, and based on what is known about bear hunting, this animal was most likely killed in late winter (Zachrisson and Iregren 1974:79-83). Some dwellings had apparently been occupied year-round. In spite of this, not one fish bone was identified from any site. Seventeen bones and one tooth fragment derive from large and small ungulates (cattle, sheep/ goat, reindeer and moose). These bones were found at five different sites from the northernmost to 158 CHAPTER 7 Table 54. Finds and dates of domesticated animals. SITE LATITUDE FIND DATE (B.P.) Stor-Rebben A,B Jungfruhamn, A,B Stora Fjaderagg B Hornslandsudde 13 65°N 64°N 63°N 6rN reindeer/sheep/goat sheep/goat sheep/goat, cattle ruminant (tooth) 945±110 B.P. 985±70 B.P., 1210±50 B.P., 1300±130 B.P. 1560±70 BP, 1235±315 B.P. 1820±40 B.P., 1570±40 B.P. the southernmost locales. The presence of domesticated animals agrees with the evidence that there were livestock enclosures of some kind associated with the dwellings. The anatomical representa- tion of these bones shows that 94% derived from meat-rich cuts, primarily the legs. This can mean that meat was brought to the sites for consumption rather than having been slaughtered there. The evidence of corrals and the place-names relating to reindeer corrals in the coastal region suggests that these animals had not been kept on sealing sites as sources of meat, rather as sources of milk products or possibly even as transport animals. Islands are also ideal places for keeping livestock; there was access to good fodder, including reindeer lichens, and the animals could never roam far from the settlements. This practice was used extensively by northern farmers who routinely transported their animals to islands during fishing seasons. Evidence of heavy grazing can still be observed on Stora Fjaderagg Island. As a general conclusion, sealing was the main focus of these sites during both the spring and fall hunting and netting seasons. These activities were probably organized by households from local communities. The presence of domesticated animals suggests that there were more than men living at these settlements. The hunting and processing activities were relatively large-scale and well organized. Flotation of hearth soils from two huts at Grundskatan revealed that berries had been collected, including raspberries, bilberries, blueberries, crowberries and bearberries. Additionally, goosefoot and stitchwort were identified. These finds are clear proof of late summer-fall activities. This is further evidence that women were present on these sites and all of these plants are known to have been used by the Saami (Viklund 2005). Place-Name Evidence of Reindeer Husbandry The frequency of place-names referring to remdeer husbandry in coastal Vasterbotten is both note- worthy and useful from the archaeological perspective. As early as the 1940s Holm (1949:143-145) had drawn attention to the Lovanger parish place-names Rengdrdtsjdrn (Reindeer Corral Lake) and Rengdrdsmyr (Reindeer Corral Bog) and commented on the frequency of the place-name rerigdrd in the region. He quoted Israel Ruong, a well-known Saami scholar and reindeer expert who com- mented, "the information that the word rengdrd occurs in the coastland is . . . of great interest, and implies that reindeer husbandry and its intensive Forest Saami form . . . occurred there" (Holm 1949:145). There are, in fact, 74 place-names referring to reindeer in Skelleftea Municipality, 37 of which refer to rengdrd of which 16 are the place-name Rengdrdsmyren (Reindeer Corral Bog). The Swedish word ren means reindeer and gdrd means an enclosure, a yard or a farmyard. In this con- text, like the word rengdrde, it probably refers to a reindeer corral. THE ARCHAEOLOGICAL ROADMAP 159 Reindeer herds belonging to the Vindel and Ume Saami villages still graze in the Vasterbotten coast- land, including Bjuroklubb. Several nomadic herding routes lead directly to Bjuron. Cladonia rangiferina, the main winter fodder of reindeer, grovv^s abundantly on the outer coast and along the eskers leading there. Mixed small-scale intensive reindeer husbandry survived into historic times among the Forest Saami peo- ple in Norrbotten, Vasterbotten and Angermanland. These Saami were not nomadic and based their settled existence on a mixture of fishing, hunting and the keeping of reindeer, goats and even cattle (Manker 1968). They were speakers of the Ume Saami language in Vasterbotten. They are remarkably similar to the Bothnian Saami 700 years earlier, and linguistically the Ume Saami language they speak, which has words for seals, is hard to explain without having once extended down to the coast. Interest- ingly enough, Olaus Magnus' map from 1539 (Figure 7), not only shows a Saami woman milking a reindeer upstream from Umea, but includes the captions renaval (reindeer husbandry) and, just north of Lovanger, rensby (reindeer village). Pastoralism and Heterarchy The archaeological evidence has shown that the Bothnian dwellings occur in groups together with storage cairns and caches, fence lines and stone circles that are interpreted as ritual features. North- ern huts and doorways normally face south, although this varied depending on topography. On the coasts doorways most often faced shorelines. It should be noted that the actual cardinal directions of north and south, as expressed by most indigenous informants, often had more of a symbolic significance than a magnetic reality, however, and "north" also referred to the area behind the dwelling and opposite the front entrance. According to the Chukchi this orientation was important ". . . in sacrificing, the odor of the hearth of the house standing in the wrong position might reach the sacrificial fire of the preceding house and taint its fire and fire-tools" (Bogoras 1909:613). Saami settlement clusters, sijddas, corresponded to the basic unit of most hunter-gatherers, the band, or a small group of nuclear families living and working together. Such a unit was mobile and less vulnerable than a single family and yet not large enough to overtax local fuel and game. The Saami sijdda was a flexible system typical of hunter-gatherer subsistence groups, although it was not incompatible with small-scale herding (Graburn and Strong 1975; Ingold 1978; Storli 1991). Annual sijdda/siida territories were roughly circular in inland areas, but changed into long rib- bons running parallel with the river valleys with the transition to nomadic herding (Vorren 1968). 80 1 70 60 50 H 40 30 20 10 Ren A ll,ri-,[ljiA Rengard # .^^ J> J' J" <^ > Figure 171. Numbers ofplace-names with the prefix ren (reindeer) and place-names with rengard {reindeer corrals) hy municipality between Norrbotten and northern Ansermanland. 160 CHAPTER 7 Norwegian coastal Saami siidas were more like circular inland territories and enveloped coastal fjords and islands (cf. Bjorklund 1985; Grydeland 200i:i8j. Numerous archaeologists have proposed that the Saami siida or sijdda was the basic unit of prehistoric settlement, and that the clustering of dwellings in the alpine and forest regions reflect this (cf Bergman 1991; Mulk 1994; Hedman 2003). This settlement arrangement is described in detail by Mulk (1994:216-221). The dwellings were arranged in rows or bowed groupings of three to five huts. A "courtyard" {sjallo) lies in front of and to the "south" of the huts. Directly behind the huts were domestic features, such as earth ovens and caches, and to the north ofthem were offer sites. Each dwelling represents a family and two to five families generally lived together. These families were not necessarily related and membership was flexible. Larger gatherings of sijddas into winter villages, which seem to have originated in the sixteenth century, are called vuobme or dalvadis. Each sijdda also had its sacred ground, a mountain or an unusual stone called a seite, a name that probably derives from the same root as sijdda. The Stalo huts in the alpine regions and hearths that have been connected with Saami kdtas in the for- est regions of northern Sweden, Norway and Finland have been described as either single sijddas or groups of sijddas. Settlements in the forested interior of Finland (cf. Tanner 1929; Tegengren 1952), as well as the coastal Saami sites of north Norway (cf Bjorklund 1985; Odner 1992; Grydeland 2001), follow the same pattern. Bogoras (1909:612) described analogous settlements among the reindeer herding and mari- time hunting Chukchi of northeastern Siberia. The Chukchi camp usually included two to three families, and the whole number of inhabitants was 10-15 people. Camps of four, five or six families formed a slight minority, and a camp of ten houses was almost impossible except for special rea- sons, like the temporary camps at trading places. Bergman and Liedgren et al. (2008) have recently discussed the kinship and residence pattern of the Swedish alpine regions ca. A.D. 1000 with a particular focus on the linear alignments of huts that became quite distinctive during the period A.D. 640-1150. In their study of historically documented dwellings in the Arjeplog region, huts 6.5 m in length could house up to 10 people [comparable to the large oval Bothnian huts], and the smallest documented hut, measuring 3.2 m in diameter, housed an elderly couple. The sijddas con- sisted of two to six households (Bergman and Liedgren et al. 2008:104). Although there is a clear understanding that sijddas were not strictly based on kinship, the authors argue that these linear alignments of huts were an expression of lineages under "great stress" (Bergman and Liedgren et al. 2008:107). The stress factor is identified as the transition from hunting to pastoralism and the need to affirm the security of the "core social unit," the sijdda. The main forces of change were internal, according to the authors, although they also assign significance to trade. Comparable linear align- ments of huts are, in fact, described by Bogoras and these lines ofhuts are indeed related to herding (1909:612-614). Reindeer husbandry quite simply created a need for aggregating both herds and herders. Poor herders with only a few animals kept together for a few months and dispersed just as easily. Wealthy herders needed more than their family to manage their animals and distant rela- tives or strangers camped together with them. These huts were arranged in lines based on hierarchy. The "chief" of the camp was referred to as the "one in the front house" among other references, including "the strongest one." The other tent occupants were called "camp neighbors" or "that of the rear house." The eldest of the brothers, or his son, had preference over the others in this lineup. The position of the front house is the first on the right side of the line ofhouses (Bogoras 1909:613). THE ARCHAEOLOGICAL ROADMAP 161 The owner of the herd, or largest part of it, was in charge of the pastures, the days of slaughtering, ceremonies and sacrifices. This hierarchy was socially enforced as no one could join a camp without permission, even at temporary camps (Bogoras 1909:614). Comparable alignments of dwellings, and during the same time period, are seen on the North Bothnian coast (refer Figure 160) and can relate both to the organized labor of herding and to col- lective hunting (cf Tanner 1979:73-107). This was a time of exceptional prosperity within Nordic society. The Saami elites, "Finn-Kings," are referred to by Snorre and in other sagas, and 21 curious defensive Mangerom type "manors" are found along the Finnmark and Kola coasts (Fiansen and Olsen 2004:214-220). According to Storli (1991) and Urbariczyk (1992:213-215), there were alliances between Saami big men and Norse chiefs, and most relationships were mutually beneficial. The long lines ofhearths in the interior and alpine regions of Sweden almost certainly coincide with the acqui- sition of wealth through intensified trade and the consolidation of labor. Metal objects were given as offerings at numerous locales in Swedish Lapland, and reindeer were sacrificed at hundreds of sites. Both Mulk (2006a) and Odner (1992) have argued that the offerings of metal objects were in- tended to maintain internal social solidarity by taking wealth out of circulation. This is reminiscent of the Bergman and Liedgren et al. (2008) argument regarding hut alignments, but as Zachrisson (1984:108, i987b:6i-68) has observed, these offer sites may actually have been inspired by Norse practices. They were expressions of alignment with the Norse gods and Norse society, not attempts to downplay their own social hierarchies. Their cessation might likewise relate to the eradication of Norse religion. CoUinder (printed in Manker 1957:51) expressed similar ideas: Lappish heathenism was syncretistic. It took up many Scandinavian beliefsfrom different times, perhaps asfar back as the Bronze Age. As the Saami becamefamiliar with Christianityfrom the goos and later, they borrowed from it. Much of their religion was magic, distinguished by crass needs, and their contacts with Christianity could have actually strengthened, but not weakened this in their cults. The Scandinavians were successfid and so were their cults. The Saami were wise to take up Scandinavian offer practices without abandoning their own. During the period 900-1200, many Nordic and European kings took up Christianity as a way of consolidating power. By the end of the tenth century, OlafTrygvesson had converted Norway and vowed to put to death all that refused to accept the new faith. Vladimir had converted Russia in A.D. 988, and Boleslav the Brave converted Poland in A.D. 999. In Sweden, Ansgar had spread the faith at Birka as early as A.D. 829, but Sweden took another 300 years to become fully (more or less) Christian, and the Archbishop in Uppsala and the five bishops of southern Sweden were first in place at the end of the iioos. Erik (later Saint Erik) was martyred in Uppsala over the issue in A.D. 1160. The Church provided the new common denominator of state formation. Needless to say, the Saami seemed to have largely abandoned their borrowed Norse religious practices by the thirteenth century because they were no longer likely to have any benefits for them, spiritually or otherwise, and thereafter quickly aligned with Christianity. Christian symbolism was already embedded in their personal adornments through trade with the west (Vagen, Bergen, Trondheim, Lofoten) and the east (Novgorod, Ladoga) (Urbahczyk 1992). This is, to my mind, the most plausible reason why the metal offer sites, which were inspired by Oden's Law in Norse religion, were so expeditiously abandoned. Interestingly enough, the extended hearth and hut alignments cease at the same time. 162 CHAPTER 7 The bear rite's distinctive religious expression of social solidarity continued, by contrast, into the nineteenth century. Discussion The sijdda system was, as in most band societies, a highly flexible and resilient combination of per- sonal and household autonomy, as argued by Odner (1992, 2000), and social solidarity as argued by many authors (cf Gjessing 1955; Grydeland 2001:67). It was undoubtedly both of these things, however, and this suggests a heterarchy, different contemporary frameworks of social relations de- pending on context (Ehrenreich, Crumley and Levy 1995). The egalitarian model, while typical of small-scale societies, does not mean there were no differences in status and wealth (cf Zachrisson I997a:i44-i48). The social obligations within such a heterarchy entailed the responsibility not to dispose of wealth, but to redistribute it, including in the form of offerings on behalf of the com- munity, all of which nevertheless enhanced prestige. Saami society had been forged from thousands of years of spiritual and economic transac- tions. The diversity of their graves, perhaps more than any other archaeological manifestation, bears witness to this fact. The impacts of the Christian Church, by contrast, especially after joining forces with the Swedish state following the Reformation in the sixteenth century, were far more disrup- tive than the beginnings of pastoralism. This was the "end ofdrum time," as described by Rydving (1995). Sapmi itself was becoming state property. Nordic linguists have long speculated that the Saami acquired "packages" of knowledge, including terminologies, from Germanic and Karelian agrarians and metal workers. The Saami, it turns out, had borrowed as many as 3,000 Scandinavian words, most of which relate to skill sets. Ingold (2000:312-338) has reasoned that this was not technology in the modern sense, rather skills that were embedded in daily life. Many of the Scandinavian loan words are believed to have been acquired on the Norwegian coast in connection with boat building, fishing and so forth but what is most remarkable is how many words relate to farming and animal husbandry. These are not random words but whole systems of terms, including names for domestic animals, corrals, sheds, farms, fields, animal products and equipment. This list gives some examples of these loans (from Wiklund 1947:57-61; Collinder 1953:53-69): The Norse terms are set in brackets. A cow is called a kussa [kyr] in Saami, an ox is called a vuoksa [oxi], a calf is a called galbe [kalfr], a goat is called a kaihtsa [geit] and even cats are called gatto [kotta]. The same is true of agrarian byproducts: wool is ullo [ull], milk is mielke [mjolk], and cheese is vuosta [ostr] and so on. The word for "farm" is garde [gar5r], "sheep byre" is fiekse [fahus] and "field" is akkr [akr]. Even the word "tame" was probably bor- rowed, tames [tamr]. There are Finnish loan-words as well, such as the word for "flour," jo/fo [jauho] and "beer," vuola [olut]. Based on runic inscriptions it has been possible to document the periods during which these Scandinavian words were acquired. The inscriptions date this process through regular changes in Scandinavian "sound-laws." The oldest runic inscription in Sweden dates to about A.D. 200, the Roman Iron Age (Stenberger 1964:373). Altogether there are about 50 inscriptions of "Urnordic" (Ancient Nordic) age. There are over 3.500 rune stones in Sweden, the northernmost of which is in Halsingland (Brink et al. 1994). The first linguistic horizon used by Nordic philologists is called Primitive or Archaic Scan- dinavian. The theory is that if sound-laws are not observed in the loan-words, the words date to THE ARCHAEOLOGICAL ROADMAP 163 before these changes. The change of ai to a before "h" is seen in the runic inscriptions from the fifth century (Skold 1979). Another change about the same time is that of au to o before "h". An example of a word supposedly borrowed from Scandinavian includes sai'va meaning fresh water, water in a river or lake and corresponding to Old Norse sjdr, sjor (Skold 1979:108). There are as many as 500 Archaic Scandinavian words in all Saami languages (CoUinder 1953:61). Late Archaic Scandinavian ended about A.D. 800 and was followed by Common Scandinavian, in which the Scandinavian languages are divided into western and eastern dialects. The chronology thus spans from A.D. 200 to 800. Most scholars believe that the process of word acquisition by the Saami occurred on the Norwegian coast. From a linguistic point ofview, we can come to the conclusion that theforbearers ofthe Lapps and the Scandinavians met in northern Norway about the time ofthe birth ofChrist, and it's probable that the Lapps were there when the Scandinavians arrived. (Skold igygnii) Although the Norwegian context is not in doubt regarding linguistic influences, Karl Wiklund suggested that these were not just borrowed words but that the ancestors of the Saami were directly involved in animal husbandry and farming: What we are confronted with is a class ofSaami who during ancient Nordic [Archaic Scandina- vian] times, presumably around the birth of Christ, in addition to hunting and fishing (possibly reindeer herding), supported themselves through aform ofanimal husbandry andfarming which was at about the same level and in the same region as the Norse. (Wiklund ig^y.Go) Wiklund's ideas were criticized because he had no proof at the time. It is clear today, however, that Wiklund was correct. Archaeological fieldwork in the 1980s revealed that there were Germanic enclaves during the Early Iron Age in both southern and middle Norrland as well; but even more relevant is the fact that we now know that the Saami in these regions practiced husbandry and even cultivation in the ways that Wiklund had proposed for Norway. Aronsson has found this idea plau- sible even for Pite Lappmark (2005) regarding the many formative connections between reindeer husbandry and settled farming (cf Khazanov 1984). Recent genetic analyses of Eurasian reindeer show, furthermore, that reindeer herding did not spread to Scandinavia from Siberia but they were domesticated independently in many different regions (Roed et al. 2008). The first Nordic loan words relating to husbandry, herding and farming were undoubtedly acquired much farther south than previously believed, perhaps as far south as Svealand in Sweden, where there were fur markets such as the famous, still-operating distmg market in Uppsala (Magnus 1555, book 4, ch. 7:182-183). Saami words were also borrowed by Germanic speakers and these relate to transportation, trade and hunting. Professor Olavi Korhonen (1982, 1988) has examined Saami terms relating to boats and boat building, which was a special Saami skill set, as well as words relating to sealing and the Saami use of dogs for finding ringed seal dens in the ice. The Nordic community, as attested by the Norse sagas, was also highly respectful and even fearful of Saami healing and witchcraft and they borrowed the Saami word noaidi, which means "shaman" or "healer." A related phenomenon involves the use of taboo words in connection with hunting. These were words that were used by Swedish hunters to hide their intentions from the game as the animal would never be mentioned 164 CHAPTER 7 Table 55. Finds of iron slag, clay and iron furnaces. SITE LATITUDE FINDS DATE (B.P.) Stora Fjaderagg B Stora Fjaderagg D Hornslandsudde 12 Hornslandsudde 5 Grundskatan 3 Grundskatan 11 Stor-Rebben A,B,C Bjurdklubb 57 6rN 65°N 64°N 63°N Slag Forge, iron, copper Slag Forge Slag, furnace clay Furnace clay Forge, slag Slag, furnace clay, iron 1845±135 945±110 1485±70 1500±100, 1205±70 1175±100 1660±70, 1235±315 1110±145 1820±40, 1570±40 1390±30 by its real name. The Saami word alge and its variants, including Ume Saami word alggie, was bor- rowed by Bothnian sealers. It means "son" and was also the Ume Saami word for "seal." Another Ume Saami word that was widely adopted by Bothnian sealers as a taboo word is mdrssie or morsd, which means "fiance" (Edlund 2000). Metallurgy Iron metallurgy was one of the most formative elements of Saami ethnogenesis. Iron slag has been found at five of the sites along the Bothnian coast, from Stor-Rebben in the north to Hornslan- dudde in the south. Iron forges were documented at Bjuroklubb 67, Grundskatan 78 and Horns- landsudde 119. Slag was found in seven additional hearths, and furnace clay has been documented in four hearths. A forging pit has been previously documented together with K)elm0y ceramics at Harrsjobacken in Lovanger (Sundqvist et al. 1992), and the forging sites on Bjuron can be viewed as continuations of these activities. The radiocarbon date of Harrsjobacken (A.D. 79-245) overlaps those of Stor-Rebben, Stora Fjaderagg, as well as Hornslandsudde. While the detailed technological analyses of this material are important, of equal interest are the shamanistic and symbolic aspects of metal working and, in this study, their connections to Saami bear ceremonialism and offer practices. Fire was the most transformative form of technology available in prehistory. Metallurgy is thus more than a technology, it is magic. Metallurgy is defined as the extraction of metals from their ores and modification of metals for use. It is a form of pyrotechnology, the use of heat for ma- nipulating raw materials (Hodges 1970). In this sense it is an extension of the methods employed by earlier hunters and gatherers to anneal stone to make them easier to flake, or ofquarrying using fire and water to shatter rock or to manipulate minerals, as has been documented at the Lundfors site in Vasterbotten, 3400/4200 B.C. (Broadbent 1979:99-108). The oldest documented iron working in Sweden is found in a band across the 60th parallel from the Malardalen region in the east to the Swedish west coast (Hjarthner-Holdar 1993). This can be connected to early metallurgy in Finland and emanates from the same regions as textile- (Lovozero) and striated ceramics, as well as Malardal- and Ananino axes, that is, from the Volga bend to the Ural Mountains in Russia (Hjarthner-Holdar 1993:17-29). Iron working was possibly underway in Russia as early as 1800 B.C. (Chernykh 1992; Khlobystin 2005) and spread to Sweden THE ARCHAEOLOGICAL ROADMAP 165 by at least 800-500 B.C. Asbestos-tempered ceramics are closely associated with this technology in north Nordic regions and date to ca. 900 B.C-A.D. i in Norway and somewhat later in Sweden (Hulthen 1991; Olsen 1994:101-108). Iron technology represents a major economic breakthrough because iron ore was readily available locally, especially through bog iron precipitates (limonites) and also because there were vast pine forests that could provide the fuel necessary to produce and manipulate it. Even winter was an advantage as iron deposits could be easily scooped up and transported through the stable platforms of frozen lakes and bogs. Sweden was to later become one of the major exporters ofhigh quality iron from sources in the so-called jarnbaraland (iron-bearing land) of southern Norrland, as described in many early texts, including that of Saxo Grammaticus in the 1190 (Hyenstrand 1974). During the period A.D. 400-600, large amounts of iron were produced in the forests of Dalarna and Jamtland, presumably for trade (Hyenstrand 1974; Magnusson 1986). This activity shows the same drastic decline as elsewhere in the Nordic region during the seventh century, but rebounded, albeit on a smaller household scale, from A.D. 1000. This is also where the so-called forest or hunting graves are found and these are most likely Saami in origin (c£ Hvarfner 1957; Sundstrom 1997:21-27; Gollwitzer 1997:27-33; Zachrisson i997a:i95-20o; Bergstol 2008). Most of the slag has been found on the shores of lakes and rivers and coincide with the distributions of Stone Age settlements. Three Iron Forges Bjuroklubb 67 Bjuroklubb 67 (Chapter 5) is a solitary dwelling that lies just below the top of a ridge facing west on what was once a small island. The elevation is 17 m above sea level. A radiocarbon date was obtained from charcoal (i485±70 B.P.), which calibrates to A.D. 460-640. The find material consists of 22 pieces of iron slag and clay. This material was analyzed by Lena Grandin, Eva Hjarthner-Holdar and Emma Gronberg at the Geoarkeologiskt Laboratorium ofthe Swedish National Heritage Board (2005). A heavily corroded cylindrical piece of iron measuring 28 x 8 x 8 mm and weighing 2.8 g was examined, but its exact composition could not be ascertained. Slag samples were determined to consist of complex combinations of slag and silica, some of which could have derived from melted sand in the soil, and some of which seems to have been added on purpose to improve smithing. Silica reduces oxidation of the metal and helps to weld iron to steel. There were also traces of cop- per that might have been used to decorate objects (Grandin et al. 2005:5). In general, the slag is homogeneous reduction slag containing both magnetite and wiistite that had been produced in a highly oxygenated environment. It was the result of secondary smithing using billets of the quality of so-called Fellujarn or Kode types (cf. Andersson 2007:6). An analysis of the silty clay furnace wall material shows it had been tempered with sand and had been heated to ii5o°C. The hearth soil was swept using a magnet in 2008 and both plano-convex hammer scales and iron sphericals were obtained, both by-products of smithing (Figure 178). At the time of excavation, it was believed that wall stones had fallen into the hut, but these were probably anvils. Grundskatan 78, Hut 11 Excavations at the Grundskatan site (Chapter 5) also produced evidence of forging in a hut. The form of the hut was very similar to that at Bjuroklubb, a rounded-rectangular foundation with low 166 CHAPTER 7 Figure lyi. Bjurdklubb 67 showing area of sooty soil and vent in Figure ij]. Close-up of the vent in Bjuroklubb 67. the rear wall ofthe hut. cobble walls measviring 7x5 m. A radiocarbon date of 1175+100 B.P. (A.D. 720-980) was obtained from the hearth. No animal bones were found and 300 grams of iron slag were collected. This ma- terial consists ofboth homogeneous slag, slag with melted stones, red-burned clay and rusted iron. The slag includes varying proportions of wiistite, olivine laminates and glass. Two pieces derived from the same smithing hearth cake. Fine-grained magnetite and drops of metallic iron occur and the iron content is high. There was a good supply of oxygen. The slag derives from secondary forg- ing (Grandin et al. 2005), and iron scales, the results of hammering, as well as sphericals were also picked up in the hearth using a magnet. Wood charcoal from the hearth was varied and included Betula sp. (birch), Pinus sp. (pine) and conifers (pine or spruce). Hornslandsudde, Site 119 Features 12 and 13 are a double hut situated at 16 m above sea level. Two dates were obtained: i57o±40 B.P. (A.D. cal. 430-540) and 1820+40 B.P. (A.D. cal. 130-240). The hearth measured ca. I X I m and was ca. 20 cm deep. Some i.i kg of slag was found in or near the hearth and a hundred or more hammer scales (60 g) were collected. Charcoal derived from pine (Pinus sp.). Some slag and furnace clay was water-rolled, which confirms the oldest radiocarbon date of this feature (An- dersson 2007:3). THE ARCHAEOLOGICAL ROADMAP 167 Figure J75. Reconstruction of Huts 12-ij. Technical analyses were performed by Daniel Andersson of the Geoarkeologiskt Laborato- rium of the Swedish National Heritage Board (2007). The slag derived from smithing hearth cakes that had probably been formed on the bottom and sides of the pit, which is somewhat unusual as they usually only stick to the walls below the air inlet and rest on a bed of fuel. Differences in mineral composition from the same area suggest that both oxygen and temperatures were rather unstable. The hammer scales are generally comprised of iron oxides and the sphericals of glass and dendritic wiistite. High quality iron was found, including one small metallic piece of steel, and three corroded pieces of iron. This find suggests the manufacturing ofedged tools. Additional slag, burned clay and a small rolled thin iron sheet measuring i x 0.3 cm were found in Hut 5, which dates to i390±30 B.P. (A.D. cal. 465-779). Feature 5 is a rounded-rectangular hut foundation mea- suring ca. 3.5 X 6 m with cobble walls measuring 0.2-0.3 m height. No traces of a smithing hearth were found in this dwelling, and it is probable that the slag had been collected from Hut 12. Iron slag (120 g) and six red clay pieces (0.6 to 1.2 cm) were found in the hearth together with 375 bone fragments (60 g) and one iron fragment (i.o cm). The large number of burned bones and a central hearth supports the interpretation that this hut was a normal dwelling, not a forge. Smithing and Shamanism Three shallow smithing hearths have been identified in huts ranging in age from A.D. 130 to 980. The oldest dated material is from Hornslandsudde and the youngest from Grundskatan. This spread suggests that iron working had been carried out the entire time these sealing sites were used, over 800 years. The three forging sites are very similar and consist of simple huts of compa- rable sizes and constructions to the dwellings. Bjuroklubb 67 had seal bones and phosphate enrich- ment indicative of a normal dwelling, and Huts 12-13 were a double hut, a smithy and a dwelling with a livestock enclosure. All three smithies were adjacent to or directly within settlements. The smithing hearths were small and shallow and quite comparable to what is known about clay-lined bowl furnaces, which average 30-80 cm across and 12-45 depth (cf. Martens 1988:70-85; Hjarthner-Holdar 1993:94-101). The floors of the forging huts also have numerous anvil stones. 168 CHAPTER 7 Munsell Soil Color: Ml = l()yv/6/l Gray M2 = lOyr/6/6 Brownish Yellow M.? -= lOyr/7/1 Light Oray M4 1 Oyi/5/2 Grayish Brown M5 l()yr/6/2 Light Brownish Gray M6 = lOyr/5/1 Gray M7 = 1 Oyr/5/2 Grayish Brown M8 - 1 Oyr/5/2 Grayish Rrown Profile A I inexcaviiled Profile B s = Slag (g) ^ Carbon Sample - 20cm 3.5 Liters of Fire Cracked Rock Figure 176. Detail offorging hearth in Hut 12, Homslandsudde. S= slag, X=AMS dates. Figure lyS. Iron scales, ropey iron and sphericalsfrom Figure ijg. Furnace clayfrom Bjurokluhb 67 (Grandin et Grundskatan Site yS, Hut 11. Spherical = 1 mm. al. 2005:14). THE ARCHAEOLOGICAL ROADMAP 169 This coastal smithing was evidently carried out for domestic uses and the manufacturing of edged tools (Andersson 2007:6). According to Grandin et al. (2005:8-9), the material is not com- parable to the site of Lappnaset (Englund et al. 1996) or Kyrkesviken in northern Angermanland (Kresten 1999), where primary smithing had taken place. It is nevertheless comparable to the mate- rial from a smithing hearth at Lill-Mosjon, Grundsunda, Angermanland, which dates to 25oo±65 B.P. (Englund 2000). Although no primary iron production has been found in the Lovanger area, there are sources of bog iron in the Nolbyn-Mangbyn area (Granlund 1943). The iron on the sites came in the form of billets, but only one billet of the spade-shaped type produced in Middle and Southern Norrland has been found north of Angermanland, and this was in northern Finland (Liedgren and lohansson 2005:290). The Bothnian sealers were clearly very familiar with the intricacies of small-scale iron work- ing. They could produce high quality objects, including decorated items. To gauge temperatures by color, the smithies were in huts with special venting systems involving channels running through rear walls; one of these was probably for a bellows operated outside of the hut. It is quite possible that they were in use during all sealing seasons, both the fall and spring months for ringed seals, as well as during the summers when harp seals were hunted. It is also notable that slag nodules and even furnace clay have been found in so many hearths without furnaces. For example, 23 small iron slag nodules were found in the hearth of Hut 3 at Grundskatan, which is adjacent to and partly contemporary with Hut 4 (with the bear burial) and the circular sacrificial feature at this site. Slag was also found in Huts A, B and G at Stor-Rebben, Huts B and D on Stora Fjaderagg and the aforementioned Hut 5 at Hornslandudde. Slag was similarly found in oval Saami hearths in Norrbotten (Hedman 2003:161-189) and in the hearths of Stalo huts. Mulk (1994:177-184) has, for instance, recorded iron slag from two Stalo huts at Suollakavalta and a hearth in Singi. These sites date to ca. 1000-1200. Similar sites have been documented in southeast Norway (Narmo 2000). Iron blanks, rivets, tongs and other tools together with iron fragments are otherwise documented from Stalo huts (cf Mulk 1994). Iron blanks and rods, as well as a crucible and asbestos wares, are recorded at other sites, such as Halla (nos. 869-870) in Asele, Lake Overuman, Tarna, Norrvik, Paulundsvallen in Lycksele, Rappasundet in Arjeplog, Landsjarv, Sorviken and Varghalsen (Zach- risson 1976:71). Zachrisson has also observed that slag is found on Saami sites and has analyzed some, including parts of a small plano-convex cake measuring 10 cm in diameter. Four of these pieces were verified by the technical department of the Museum of National Antiquities and two of them, from Gafsele in Asele Parish and Stallverket on the Angermanland River (Zachrisson 1976:129). In a more recent overview of the evidence, Zachrisson (2006) describes likely Saami iron working sites in jamtland, Harjedalen, northern Dalarna, western Halsingland and Medelpad, as well as iron objects with distinctive Saami markings. Iron slag has likewise been documented in North Norway, for example in a probable sha- man's hut at Vapsgieddi (Grydeland 2001:37-42). While some sites are smithies, most of them are not, and this raises the very real possibility that slag was deliberately put in hearths because of its magical, especially transformative, properties. Metamorphosis was an empirical reality and the hearth was a sacred place (Qvigstad 1926:321). Mats Burstrom (1990) has argued that slag had been deliberately placed in Iron Age graves in southern Norrland, both as grave goods and as fill. This ritual connection in Scandinavia has been discussed more recently against the background of comparative ethnographic analogies in Africa (Haaland 2004; cf Haaland et al. 2004). 170 CHAPTER 7 European and Norse mythology is full of the myths, legends and folktales that build on the magic and rituals of the smith and iron working (Haaland 2004:11; cf Green 2002). Hedeager (2001) pursued this idea regarding the Norse figure Volundr, who could change shape as a shaman to mediate between humans and the gods. She also discusses Regin from the Volsunga Saga, who is a liminal figure and a dwarf (Hedeager 2001:492). Although a person to be respected and even feared, in Germanic society the smith was nevertheless only working at it part-time, and this was not necessarily a high status profession (cf Haaland 2004:14). Volundr and the North The master smith Volundr/Weland is described in the Older Edda from ca. 1000, which is one of the oldest of the Icelandic texts. He was originally a dwarf or from a family of dwarves or elves, but he was completely human and with human emotions. His tale is prefaced by a brief description of his background: "The Finn King had three sons, Slagfinn, Egil and Vohindr, who traveled on skis and hunted reindeer ..." (Bseksted 1970:228). The tale then goes on about how Volundr, having made a magic sword and 700 rings of red gold that he tied to his forge, was robbed by King Nidud and his soldiers, who wore chain mail. Meanwhile, Volundr, on returning home from bear hunting, was captured and tied up. He had his leg tendons cut, but took to the sky using wings he had forged (Baeksted 1970:229). Most intriguing about this story are the references to the "Finn King," skis, reindeer, dwarves and even a bear. Volundr's forge was on an island. The Volundr allusions point northward, and there are valid reasons for taking them seriously. For one thing, there is now credible archaeological evi- dence for Finn Kings (Hansen and Olsen 2004:214-220), and the hunting of reindeer on skis could only have taken place in the north. The reference to dwarves, who knew magic and were devious, can refer to the meeting ofthe Saami and the Norse as proposed by Nilsson's early study ofthe saga litera- ture (cf Nilsson 1866) and, of course, the Saami stories about Stalo giants. The shamanistic context in Norse religion is also expressed through the Seic)r rituals, which involves female divination, as seen among the Saami (Dubois 1999:121-138; Price 2002:91-328). Amanda Green (2002) has related the ritualistic value of iron to the Germanic practice of offering weapons and animals in bogs and watery cult places, which we know the Saami practiced. There is likewise a strong gender component to met- allurgy in which the forge is seen as a womb and symbolizes fertility (Herbert 1993). Conclusions The chronology, architecture and organization of North Bothnian settlements coincide extremely well with what is known about other Saami settlements and cultural manifestations throughout the Nordic region. Although dwelling constructions differ by region, they reflect comparable social and resource exploitation strategies. They were closely connected with pastoralism, trade and intense cultural in- teractions. The closest Saami architectural parallels to the North Bothnian material that I know of are in northern coastal Norway. The combinations and even the transitions from oval to rectilinear structures could similarly relate to contacts between the Saami and other groups, or simply reflect different seasonal needs. Technology, it must be assumed, was a shared interest and yet another arena ofcollaboration between Saami and non- Saami. Ritual behavior, the connection between metallurgy, metal offerings in bogs and shamanism, seems to have belonged to the realm ofcommon ground. As will be discussed in the next chapter, the transformative properties ofthe bear, and the significance of the bear rites, relate directly to the dwelling, the hearth, fertility and to Saami social identity. THE ARCHAEOLOGICAL ROADMAP 171 1^J' BEAR BURIALS & BONE DEPOSITIONS Figure 180. Map showing general distribution ofdocumented bear burials and bone depositions in Sweden and Nonvay. Icon based on an imagefrom a Saami drum (Kjellstrom and Rydving ic)88:26). Bear burials in Sweden concentrate in South Sdpmi, to which Grundskatan belongs. Bone depositions are more common in North Sdpmi. Photo inset ofbear skull find from Jdmtland. Courtesy Antikvariskt-topog-afiska arkivet, the National Heritage Board, Stockholm. Rituals and Religion Most religious interpretations of Saanii archaeological material are based on written ac-counts from the i6oos and 1700s. Swedish Lord High Chancellor Magnus Gabriel De la Gardie commissioned studies by Swedish priests who were to then turn over their materi- als to lohannes Schefferus, author of Lapponia in 1673. Priests also collected information on religious practices in Norway. These efforts were intended to define, and then overpower, Saami witchcraft, magic and heathen beliefs and practices (Manker 1957:9). Information on Saami sacred sites can be sought from many sources: physical traces in the landscape, place-names, traditional knowledge and written accounts. Each source has its limita- tions, however. Landscape impacts were often minimal and building materials mostly perishable. Most constructions were made of wood, brush, sod and birch bark. Place-names, which once iden- tified locales, have in many areas been replaced by names from the linguistic majority. In other instances, Saami place-names have been lost along with the disappearance of the Saami language and changed land uses. The written sources are relatively recent and were recorded by non-Saami. It is probable that the Saami were loath to reveal the locations of sacred sites, sacred place-names or practices, either because this would shame the sites and weaken the power of their traditions, or simply to avoid punishment, prison and even execution (Lundius 1905:32, Olsen 1910:7 ff ). Saami sacred sites were often landforms such as mountains, lakes, islands, points and pen- insulas, caves, crevices, cliffs, ridges, ledges, water divides, rapids, waterfalls, springs and streams (Qvigstad 1926; Manker 1957). These were places where power was concentrated. These powers consisted of the spirits of ancestors and different categories of helping and protective beings that maintained different classes of animals. Special rules applied regarding the interaction ofhumans and these powers. It was at these kinds of sites one could seek contact with spiritual forces. Ac- cording to Saami traditional beliefs, to die was to wander in the underworld (Hogstrom 1747:210). The underworld was also the home of dead relatives (Backman 1975). These people lived a paral- lel existence and even walked upside down with their feet against those of the living (Lundius 1675:6). All of these spiritual entities received offerings at places where conditions for contact were favorable. Offerings to dead relatives could occur near graves or at other locales, especially on spe- cial platforms near settlements or in the natural landscape where there were transitions between worlds. Sacrifices and offerings often occurred at places that were associated with game. Offerings were made to Tjaetsiealmaj, "the water man," for fishing luck on the shores of lakes or in the water. Offerings were made at a kill site to Liejpiealmaj, "the alder man." Inside the hut, offerings were 173 made to the female deities Maadteraahka and her daughters Saarahka, Joeksaahka and Oksaahka - overseeing ah that is female, including menstruation and childbirth. Offerings were made daily. Each entity had its own special place in the hut. Under the hut floor resided Jaemiehaahka, "death woman," who controlled the distribution of the vital powers between the living and the dead. Traces of Saami ritual practice have mostly disappeared, but some constructions have sur- vived. These can be graves, stone circles or enclosures, mound-like constructions and different types of sacrificial platforms or cairns. Sacrificial idols were often made of wood, but in some instances were made of stone on or near the offer site. Rich deposits of bone and antler together with blood, fat and flesh gave rise to both lush vegetation and characteristic plant types (cf Manker 1957:123; Wennstedt Edvinger and Winka 2001:108). Sacred sites were used for both "bloody sacri- fices" and metal offerings. Of the bloody offerings, bone and antler could survive, but seldom any other visible indications. Archaeological investigations of offer sites, which were known from oral traditions, have rendered astounding numbers of artifacts (Hallstrom 1932; Serning 1956; Zachris- son 1984). A single site can contain hundreds of objects from a wide geographic region. The objects consist of brooches, pendants, clasps and buckles of pewter, bronze and silver, silver coins and iron arrowheads. Coins and ornaments are usually perforated. There can also be considerable amounts of bone from many species: reindeer, cattle, horses, sheep/goats, pigs, fish, birds (including swans and roosters), bears, dogs, wolves and cats (Manker 1957:45-46). The most common day-to-day offerings were ordinary items: bits of food, reindeer milk, and tobacco or vodka (Mebius 2003; cf Jordan 2003). While Saami sacred sites can consist of a number of different features or none at all, two types are of particular interest in this study. One form is a circular sacrificial feature and the other manifestation is the bear burial. Circular stone features of these kinds have been known for over a century in Norway and have now been documented in coastal Sweden. This new material has been presented by Wennstedt Edvinger and Broadbent (2006). An analysis of the Grundskatan bear burial is based on 42 comparable bear burials and bear bone depositions in Norway and Sweden. The bear burial is one of the most powerful manifestations imaginable of Saami identity and terri- tory. In order to avoid confusion regarding Nordic grave types, the term "burial" is used instead of "grave" to describe the interment of bear bones. The use of this term encompasses the act of cover- ing the remains and the rituals connected with it. An offering is a symbolic gesture, and as most of the animals had been consumed prior to being offered, they are not, strictly speaking, sacrifices. Interestingly enough, the creatures that were not normally consumed, for example fur-bearing ani- mals, were sacrificed with their bones intact. Some animals were obtained specifically for sacrifices, such as house cats and even horses (Manker 1957:46). Manker (1957:10-11) has defined nine traits of Saami religion: 1) All nature was animated and forces of nature, and even illness and death were personified by gods. 2) Reindeer herding, hunting and fishing had specific gods and guardians. 3) Power and danger were connected with specific locales. 4) Gods were worshiped in the form of unusual stones, cliffs or wooden idols. 5) Cults and rituals most often had utilitarian motives, such as good luck in fishing, hunting, herding and health. 174 CHAPTER 8 6) There was no priesthood, and every family used sacred drums for their spiritual needs. 7) The drum was the primary instrument of Saami cults. 8) The most distinctive of the Saami cults relates to bears. 9) There were a number of female taboos regarding offer sites, hunting, the handling of the drum, etc., but special goddesses (family or kin spirits) were connected with the dwelling, childbirth and small children. There are also a number of special terms that applied to sacred sites, three ofwhich are rele- vant here. The term seite and its variants refer in part to an idol, usually an unusual natural stone and also the place of this idol. The South Saami term bissie has three meanings: the concept of sacred, a sacred offer site, and the offering itself The term ahka or akka refers to the female goddesses, inclvid- ing the mother goddess Maadterahka and her three daughters (L^stadius 1838-1845: Manker 1957:13 ff ). The Saami cosmos had two, possibly three, levels: the upper world (including the heavens) and the underworld. These are, in any case, parallel worlds and their boimdaries can also be defined by the land and the sea, or by a mountain top and the sky. These worlds were united through the world axis. Physical representations of the supernatural world, such as idols, were part of everyday life. Human graves, however, were to be avoided (Stora 1971; Mulk and Bayliss-Smith 20o6b:25-29). Acts of communication with Maadterahka took place through routine domestic observances, small-scale offerings, larger-scale seasonal offerings and, when necessary, through shamanistic intercession (Mulk and Bayliss-Smith 2006:91). There was also a hierarchy of offer sites and sacri- ficial sites that related to different social settings. The primary setting was that of the family and the family dwelling; the second setting was the territory used by a local band or the sijdda: and the RITUALS AND RELIGION 175 third setting was the region used by related bands, referred to as the vuobme or tjiellde (Swedish, lappby). These units are comparable in size to hunter-gatherer bands; a family normally averages five people, a band 25 people or three to five families, and a viable biological and linguistic group of about 500 people (cf Lee and Daly 1999)- Comparable numbers have been obtained from cameral records for the Forest Saami of lokkmokk parish (Kvist and Wheelersburg 1997). Ernst Manker published more than 500 sacred sites in 1957. Eleven of these sites are known for their rich finds of metal objects. A twelfth major site has been added since then (Zachrisson 1984). These sites date primarily to the period A.D. 700-1400, but there are also sporadic finds of quartzite, slate and asbestos-tempered pottery. There is thus every reason to believe that these offer practices did not originate during the Late Iron Age, but have deep roots extending back 4,000 years or more (Manker 1957:52). The Wild and the Domesticated While Saami shamanism related to mediation between hunters and prey, there was also a strong pas- toral element that involved tame and domesticated animals, herds and pasturelands. Saami ideology thereby embodied a dualism of dependency and control (cf Ingold 1986; Hamayon 1996:76-89). Seventy-six percent of the animal bones from offer sites derive from combinations of wild and do- mesticated reindeer and 8% derive from livestock, especially sheep and goats. An offering is a ritual- ized form of communication with the gods, and it is evident that Saami gods could be satisfied by this mixture of wild and domesticated creatures. As quoted in Manker (1957:44-45): Randulf (1723) described live sacrifices of horses, goats, dogs and cats at sites with wooden idols in North Norway; Eorbus (1727) mentioned sheep or goats being sacrificed to Beifwe, the sun god in Swedish Lap- land; Kildal (1730) described the offering of spirits, tobacco, cheese, porridge {Saaraahka's porridge), calves, sheep, lambs, goats, pigs, cats and roosters to Maad- teraahka and her three daugh- ters; Hogstrom (1746) in Lule Lapmark mentioned sheep and goats among other animals sac- rificed; Leem (1767), writing of Finnmark, mentioned that sheep and other livestock together with milk and cheese, but seldom seals, were offered. While many of these finds date to the historic period, Manker has presented ev- idence that livestock had been of- fered during the Viking Period: goats were found at seven sites and cattle at three sites. Reindeer and goats were, in fact, among the most common offer animals (Manker 1960:46, 76). Figure 182. Aniinal offerings at j^j sites (based on Manker ig^y.p). 176 CHAPTER 8 The Bear Burial The bear was one of the most important symbols of Saami society. While bears were the largest and most dangerous predators in the Nordic region and were revered as such, their spiritual sig- nificance among circumpolar peoples like the Saami related in greater measure to their humanlike attributes, including body proportions, particularly when skinned, their upright and sitting stances, footprints, omnivorous diets, feces, cleverness and even emotional behavior, including crying and masturbation (Edsman 1994:20). Added to these qualities is the bear's ability to hibernate, to sur- vive without eating, and then seemingly rise from the dead in the spring (Hallowell 1926:149). The bear was a sacred animal in all Saami areas and bear hunting was a sacred undertaking (Backman and Hultkrantz 1978:83). Saami bear rituals were first documented by Danish, Norwegian and Swedish priests in the seventeenth and eighteenth centuries (Niurenius 1645; Rheen 1671; Graan 1672; Thurenius 1724; Hogstrom 1747; Fjellstrom 1755; Leem 1767). A comprehensive synthesis and analysis of the large corpus of original source materials, as well as published literature on Saami and Finnish bear ceremonies, is given in Manker (1957) and Edsman (1994). The brown bear (Ursus arctos arctos) was a co-migrator with humans into the Scandinavian Peninsula from the south and east following the rapid deglaciation ofthe region between 8,000 and 10,000 years ago. Bear figures appear as portable art (Carpelan 1977) and images of bears in early northern rock art, together with figures of birds, fish, reindeer and humans, have close parallels on Saami drums (Helskog 1988). It has been speculated that the drums may have replaced rock art as a shamanistic medium (Helskog 1988:110-112). After reindeer, bears were the most common animals portrayed on these drums (Kjellstrom and Rydving 1988). Twenty Nordic sources from 1631 to the nineteenth century describe Saami bear burials. Zachrisson and Iregren (1974) summarize these accounts and describe nine Swedish bear burial finds, as well as 20 finds of bear bones of "special character" (i.e., bear bone depositions, includ- ing skulls). The latter are more common in north Lapland and Norway (Zachrisson and Iregren 1974:38). Although of seemingly different character, the motivation for a burial and for a bone depo- sition was the same, the need to show respect for the bear and for renewal (Mebius 2003:108-110). Mulk and Iregren (1995) published an additional bear burial from a dwelling site at Karats near Jok- kmokk in Lapland. A study encompassing 30 finds from North Norway was published by Myrstad (1996), and another about Spildra, an island with nine bear burials/bone depositions in Norway, was published by Bjorklund and Grydeland (2001). Altogether, some 43 burial sites are recorded in Norway and Sweden. An additional find had been made at Onbacken in Halsingland in 1923 (Liedgren 1985). According to the excavator, Gustaf Hallstrom, bear bones and a complete skull with teeth were found in the southeast corner of an Early Iron Age terrace house and not far from some graves. This is a "typical bear grave of Southern Lappish type" (Liedgren 1985:340). This parallels the Grundskatan find and shows that the Saami were directly involved in spiritual interactions with Germanic farmers in Halsingland. Saami rituals took place at many locations in the landscape and within a hierarchy of space, from the mountaintops to the hearths. Manker sorted 342 offer sites by topography. Forty-four percent were by springs, waterfalls, rapids, lakes, islands and points. He also noted that islands were important because they were isolated and protected (Manker 1957:23-28). A majority of the bear burials/depositions (73%) are also associated with water and had been placed on islands or on points. All but two of the Norwegian finds were on the Norwegian coast. RITUALS AND RELIGION 177 The construction ofbear burials, like dwellings, reflects the availability of local raw materials (e.g., cairns in the mountains and on the coasts and inhumations with earth or log coverings in interior and forested areas). Most bear burials and depositions (48%), especially in Norway, were found in fissures, under boulders or in caves. It has been pointed out that these are the places where bears live, their dens, and also where there were openings to other worlds (Myrstad 1996:66-67). Some 17% were in cists or stone circles, 6% were in earth mounds, and 8% were under cairns. Al- though not common, the bones could be charred, as seen at Grundskatan (see also Paulson 1963), and a burned surface was observed under a bark layer at the bear burial site at Karats (Mulk and Iregren 1995). Of the 29 bear burials and depositions with multiple skeletal parts, 22 of them (78%) had some or most of the bones chopped, broken and split. Because of the ritualistic significance of the bear burial, only a selection of bones seems to have been, in practice, necessary. The common lack of phalanges shows that the bear had been flayed with the claws attached. According to Saami tra- dition, bear claws, which could have been removed as amulets, contained vdki, the essence of the power of the animal (DuBois 1999:105). Even the bear skull, which was of special significance in bear ceremonialism (Hallowell 1926:135), was not necessarily put into the burials and could have been removed for other purposes: 32% lacked skulls and 31% of the skulls were fragmentary. Only 38% had teeth present. Long bones, by contrast, were nearly always present in both depositions and burials. These bones contained the most marrow, were highly prized as food, and were powerful symbols of the life force of the animal (Edsman 1994:20). Seventeen radiocarbon dates of bear burials/depositions range from A.D. 200 to 1800 (Zach- risson and Iregren 1974; Mulk and Iregren 1995; Myrstad 1996). There are two apparent spikes: A.D. 800-1200 and A.D. 1600-1800. The first spike corresponds to the Viking Age and the Gr- undskatan find, and the second to the "end of drum time" during which Lutheran priests cracked down on Saami religion. 20 18 16 14 12 % 10 8 6 4 2 h LI 178 Figure 18]. Find-places ofbear burials and bone depositions in Norway and Sweden {N=^8}. CHAPTER 8 30 1 25 20 % 15 10 5 i i Bogs Coasts Hills Interior Islands Lal i\^ 8 3.0 Fallan 2 1.5 Rilirnn 4 1.0 Sunnana 2 3.0 Nolbyn 6 5.5 Krakanger 11 1.0 Kashnip1 \ i_J O kJ w 1 \^ 5 1.0 Vastana 4 1.5 Selet 9 1.0 Uttersjon 6 4.0 Risbole 3 1.0 Bole 3 1.0 1 ;inn\/il