Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The organization of microcore technology in the Canadian southern interior plateau Greaves, Sheila 1991

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A1 G73.pdf [ 19.08MB ]
Metadata
JSON: 831-1.0100348.json
JSON-LD: 831-1.0100348-ld.json
RDF/XML (Pretty): 831-1.0100348-rdf.xml
RDF/JSON: 831-1.0100348-rdf.json
Turtle: 831-1.0100348-turtle.txt
N-Triples: 831-1.0100348-rdf-ntriples.txt
Original Record: 831-1.0100348-source.json
Full Text
831-1.0100348-fulltext.txt
Citation
831-1.0100348.ris

Full Text

THE ORGANIZATION OF MICROCORE TECHNOLOGY IN THE CANADIAN SOUTHERN INTERIOR PLATEAU By SHEILA GREAVES B.A., The University of British Columbia, 1970 M.A., The University of Calgary, 1981 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Anthropology and Sociology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1991 ©Sheila Greaves, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date "j"*" DE-6 (2/88) i i ABSTRACT The purpose of this research is to construct and test a model of the organization of microcore technology, a standardized core technology, within the subsistence-settlement system of prehistoric, semi-sedentary hunter-gatherers. The study of technological organization involves investigation of why a society selects particular tool designs, and how i t structures the manufacture, use, maintenance and discard of tools and associated debitage across the landscape. The model tested here associates the use of microcore technology with a design for a maintainable and transportable tool assemblage which conserves li t h i c material, and with a regional distribution focused on residential camps as the locus of microcore manufacture and microblade production and use. The model is tested through a comparative case study of archaeological tools and debitage from microlithic and non-microlithic sites in two upland valleys in the British Columbia Southern Interior Plateau. Research hypotheses and corresponding test implications are evaluated with data and analyses relating to core reduction and tool production stages, to tool use, and to activity area patterning within the sites. Results of hypothesis testing indicate that the model only partially explains the role of this particular standardized core technology in the study areas. Microcore technology is found to be associated with high residential and logistical mobility; a transportable, expediently-used tool assemblage; and the conservation of a specific raw material in one valley. Thus, this research proposes that microcore technology was a standardized technology which was variable in design goals and distribution, even within the same geographically and ethnographically defined region. i i i T&PT.H car CQNTKNTB ABSTRACT . . . . i i LIST OF TABLES v i LIST OF FIGURES ix ACKNOWLEDGEMENT X CHAPTER 1: INTRODUCTION 1 CHAPTER 2: RESEARCH DESIGN Theoretical Framework 8 The Organization of L i t h i c Technology 11 Definition 11 Resource Procurement Strategies..... 12 Technological Strategies 15 Standardized Core Technology as an Organizational Strategy 22 Introduction 22 Design Goals of Standardized Core Technology 25 Distribution of Standardized Core Technology 31 Discussion 33 Model of the Organization of Microcore Technology 35 Definition of Microcore Technology 35 Goals of the Model 36 Salient Features of the Model 37 Technological Organization Goals 38 Distribution of Microcore Technology 39 Application of the Model to the Study Areas 42 Research Hypotheses 43 CHAPTER 3: THE STUDY AREAS Biophysical Environment 48 Geographic Setting 48 Paleoenvironmental Setting 50 Fraser Glaciation and Subsequent Deglaciation 50 Holocene Climatic and vegetational Sequence 53 Modern Environmental Setting 58 Climate 58 Drainage ..58 F l o r a l Resources 60 Faunal Resources 66 L i t h i c Resources 70 Cultural Environment ..72 Synthesis of Prehistoric Subsistence-Settlement System.... 72 Regional Cultural-Historical Sequence 72 Local Prehistoric Subsistence-Settlement Pattern 78 Microcore Technology i n the Southern Interior Plateau 86 Introduction 86 Chronological and Geographic Distribution 86 iv Ethnic Affiliation and Microcore Technology 91 Subsistence-Settlement Pattern and Microcore Technology 93 The Use of Microblades 95 Discussion 98 Synthesis of Ethnographic subsistence-Settlement System 99 Introduction 99 Subsistence-Settlement System in the Study Areas 103 Discussion 115 CHAPTER 4 : THE ARCHAEOLOGICAL DATA BASE Artifact Descriptions 122 Introduction 122 Lithic Material Types 124 Debitage Classification 124 Non-microlithic Debitage 125 Microlithic Debitage 125 Tool Classification 128 Formed Tools 131 Expedient Tools 131 Microlithic Tools 134 Site Descriptions 134 Introduction 134 Upper Hat Creek Valley 139 Microlithic Sites 141 Non-microlithic Sites 148 Highland Valley 150 Microlithic Sites 151 Non-microlithic Sites 155 Summary of Site Data Base 158 Discussion 159 CHAPTER 5: ANALYTICAL METHODS Debitage Analysis 162 Non-microlithic Debitage 162 Introduction 162 Selection of Attributes 164 Definition of Manufacturing Stages 168 Results 177 Microlithic Debitage 182 Introduction 182 Implications of Previous Research 184 Experimental Replication 186 Summary of Experimental Microblade Production 188 Selection of Attributes 190 Results 203 Manufacturing Typology for Microoores 206 Tool Analysis 210 Introduction 210 Attribute Selection and Recording 211 Employable Unit Typology 216 Results 218 Activity Area Analysis 221 V Introduction 221 Method 225 Description of Activity Areas 228 Activity Area Clusters 228 Discussion 243 Settlement Types 247 Method 247 Results 247 Discussion 254 CHAPTER 6: TEST OF MODEL Evaluation of Research Hypotheses 258 Hypothesis 1 258 Hypothesis 2 267 Hypothesis 3 277 Hypothesis 4 283 Conclusions 299 CHAPTER 7: SUMMARY AND DISCUSSION Summary 306 Introduction 306 Results of Hypothesis Testing 310 Evaluation of Model 316 Discussion 321 BIBLIOGRAPHY . 329 APPENDIX I. PLANTS, MAMMALS AND BIRDS IN STUDY AREAS 348 APPENDIX II. ACTIVITY AREAS IN STUDY SITES 354 vi LIST OF TABLES 1. Characteristics of biogeoclimatic zones in the study areas 62 2. Cultural-historical affiliation of Upper Hat Creek Valley sites...79 3. Cultural-historical affiliation of Highland Valley sites 82 4. Interior Salish subsistence technology I l l 5. Inferred ethnographic subsistence-settlement pattern in uplands.. 120 6. Site selection criteria 137 7. Summary data for the study sites 138 8. Artifact frequency counts across study sites 143 9. Artifact percentage counts across study sites 144 10. Frequency and percentage counts of lithic raw material type 145 11. Discriminating values for debitage classes on platform-remnant bearing flakes 175 12. Discriminating values for debitage classes on shatter 175 13. Frequencies of debitage assigned to manufacturing stages... 176 14. Percentages of debitage assigned to manufacturing stages 178 15. Mann-Whitney two-sample tests on debitage class percentages grouped by previous site classification 180 16. Mann-Whitney two-sample tests on manuacturing stages grouped by previous site classification 180 17. Attribute values of experimentally produced microblades in order of detachment from microcore 193 18. Measures of central tendency and dispersion for early stage of microblade production 194 19. Measures of central tendency and dispersion for late stage of microblade production 194 20. Mann-Whitney two-sample tests on attributes measured on experimental microblades grouped by stage 196 21. Experimental microblade attribute principal components analysis factor loadings 196 22. Classification of experimental microblades by clusters 201 23. Measures of central tendency and dispersion for two clusters of experimental microblades 201 24. Mann-Whitney two-sample tests on attributes measured on experimental microblades grouped by cluster 202 25. Classification probabilities for experimental microblade sample..202 26. Attributes and values recorded on Employable Units 213 27. Correlation of attribute values with motion 215 28. Correlation of attribute values with worked material 215 29. Employable Unit types 217 30. Frequency counts of Employable Unit types 219 31. Frequency counts of Employable Unit types, worked material types, and tool:EU ratio 220 32. Activity area descriptive data 229 33. Debitage data for activity area cluster analysis grouped by cluster 233 34. Tool data for activity area cluster analysis grouped by cluster..236 35. Mean proportion of activity area attributes grouped by Ward's cluster analysis 239 36. Mean size and artifact density of activity area clusters 240 v i i 37. KrusJcal-Wallis tests on activity area attributes grouped by Ward's cluster analysis 240 38. Activity area measures used to discriminate among sites 244 39. Activity area cluster membership re-ordered by settlement cluster membership 248 40. Site characteristics re-ordered by settlement cluster membership 250 41. Settlement type frequencies in study sample 251 42. Summary of settlement types by biogeoclimatic zone 256 43. Frequency counts of residential camps, field camps and stations in study sample ..259 44. Frequency counts of settlement types in Upper Hat Creek Valley...259 45. Frequency counts of settlement types in Highland Valley 259 46. Frequency counts of residential camps and field camps in Upper Hat Creek Valley 262 47. Frequency counts of residential camps and field camps in Highland Valley 262 48. Comparison of archaeological subsistence-settlement pattern win inferred ethnographic subsistence-settlement pattern 264 49. Presence/absence of EU types on morphological tool types 268 50. Presence/absence of EU types on morphological types in microlithic sites 268 51. Presence/absence of EU types on morphological types in non-microlithic sites 268 52. Percentage counts of EU types on morphological tool types 270 53. Percentage counts of morphological tool types that are versatile.273 54. Chi-square test for significance of association between technological type and versatile tools 273 55. Percentage counts of the most frequent EU types 275 56. Frequency counts of activity area types 275 57. Chi-square test for significance of association between technological type and activity area type 275 58. Median of weight and size of complete formed tools 278 59. Mann-Whitney two-sample tests on weight 278 60. Mann-Whitney two-sample tests on size 278 61. Frequency counts of microlithic artifacts 280 62. Frequency and percentage counts of microcore types 282 63. Microblade/microcore ratio 284 64. Chi-square test for significance of association between technological type and reduction stage 286 65. Percentage counts of debitage types in manufacturing stages 286 66. Mann-Whitney two-sample tests on reduction stage 288 67. Mann-Whitney two-sample tests on reduction stage in Upper Hat Creek Valley sites 288 68. Mann-Whitney two-sample tests on reduction stage in Highland Valley sites 288 69. Percentage counts of debitage reduction stages in Upper Hat Creek Valley sites 289 70. Percentage counts of debitage reduction stages in Highland Valley sites 289 71. Summary of unstandardized core measurements 290 v i i i 72. Chi-square test for significance of association between technological type and flake tool type 292 73. Chi-square test for significance of association between technological type and flake tool type in Upper Hat Creek Valley sites 292 74. Chi-square test for significance of association between technological type and flake tool type in Highland valley sites..292 75. Chi-square test for significance of association between technological type and bifacial tool type 294 76. Chi-square test for significance of association between technological type and bifacial tool type in Upper Hat Creek Valley sites 294 77. Chi-square test for significance of association between technological type and bifacial tool type in Highland Valley sites 294 78. Chi-square test for significance of association between technological type and expedient tools 296 79. Percentage of bifacial thinning flakes in study sample 296 80. Percentage of bifacial thinning flakes in Upper Hat Creek Valley sites 298 81. Percentage of bifacial thinning flakes in Highland Valley sites..298 82. Percentage distribution of lithic types in microlithic artifacts.300 83. Percentage distribution of lithic types in a l l artifacts 300 84. Species names of plant resources in the study areas 349 85. Mammals with ranges overlapping the study areas 351 86. Birds of potential economic value with breeding ranges overlapping the study areas 352 LIST OP FIGURES 1. Location of study areas 49 2. Hat Creek Valley drainage system 59 3. Highland Valley drainage system ...61 4. Language groups in study areas 102 5. Non-microlithic debitage 126 6. Non-microlithic debitage 127 7. Microlithic debitage ...129 8. Formed tools 132 9. Expedient tools 133 10. Expedient and microlithic tools 135 11. Location of Upper Hat Creek Valley study sites 142 12. Location of Highland Valley study sites 152 13. Attributes measured on platform-remnant bearing flakes and shatter 166 14. Boxplot of maximum dimension by number of dorsal scars in Upper Hat Creek Valley pilot study sites 170 15. Boxplot of maximum dimension by number of dorsal scars in Highland Valley pilot study sites 171 16. Boxplot of maximum dimension by amount of cortex in Upper Hat Creek Valley pilot study sites 173 17. Boxplot of maximum dimension by amount of cortex in Highland Valley pilot study sites 174 18. Experimentally produced microblades and exhausted core 189 19. Attributes measured on experimental microblades 192 20. ward's cluster analysis of experimental microblades 200 21. Microcore attributes 207 22. Microcore typology 208 23. Employable unit on a microblade 212 24. Ward's cluster analysis of activity area attributes 232 25. Complete linkage cluster analysis of site attributes 249 26. Activity areas in Upper Hat Creek Valley microlithic sites: EeRilO; EeRj49; EeRj55 355 27. Activity areas in Upper Hat Creek Valley microlithic sites: EeRj56; EeRj62; EeRj60 356 28. Activity areas in Upper Hat Creek Valley non-microlithic sites: EeRj8; EeRj20; EeRj42 357 29. Activity areas in Upper Hat Creek Valley non-microlithic sites: EeRj64; EeRjlOO; EeRk52 358 30. Activity areas in Highland Valley microlithic sites: EcRg2AA; EcRg2CC; EcRg4C 359 31. Activity areas in Highland Valley microlithic sites: EdRglB; EcRg4J; EdRglA 360 32. Activity areas in Highland Valley non-microlithic sites: EcRg4A; EcRg4B; EcRg4D 361 33. Activity areas in Highland Valley non-microlithic sites: EcRg4E; EdRg5; EdRg6 362 X ACKNOWLEDGEMENTS The successful completion of this dissertation owes much to the patience and generosity of faculty, staff and fellow students, who provided the help essential for an out-of-town student. The members of my Advisory Committee, Dr. D.L. Pokotylo, Dr. R.G. Matson and Dr. A. Stryd, provided guidance, encouragement, and support throughout my program. I also thank Dr. D. Aberle and Dr. M. Blake who provided valuable critiques while serving as interim committee members. I am particularly grateful to Dr. Pokotylo, my research supervisor, whose research on microcore technology influenced the direction of my own studies, and who provided me with an opportunity to do f i e l d work i n the Plateau. Dr. Matson's pioneering research into the multivariate analysis of archaeological data forms a s t a t i s t i c a l basis for this research. Finally, I thank Dr. J. Ryder, Dr. B. Hayden, Dr B. Alfred, and Dr. P. Smith, who served on my University Defence Cccmittee, and provided additional insights i n their areas of expertise. The University of B r i t i s h Columbia, through the Graduate Fellowship program, as well as the Government of B r i t i s h Columbia, provided financial support. As well, a S.S.H.R.C. grant to Dr. Pokotylo and employment through the Museum of Anthropology ensured additional funding. Research space was generously provided by Dr. P. Schledermann and M. Robinson, present director, of the Arctic Institute of North America. Alan Hoover and Shelley Reid of the Royal B r i t i s h Columbia Museum, Dr. Stryd of Areas Associates, Steve Lawhead and Dr. Pokotylo expedited the transfer of art i f a c t s . Dr. K. McCullough, of the Arctic Institute of North America, assumed responsibility for the artifacts i n Calgary. Dr. Pokotylo, Dr. Stryd, Sylvia Albright and Diana Alexander provided generous access to unpublished data and interpretations. The majority of figures were prepared by the late Moira Irvine, and completed by Marilyn Croot. Photographic expertise was provided by Joyce Johnson and Tom Jack. Ann Stevenson contributed her e d i t o r i a l and computer expertise. I also thank Alexander Mackie and Dr. Pokotylo for their experimental microblade production. Finally, preparation of the f i n a l product was greatly assisted by Ross Goodwin and Dr. Schledermann of the Arctic Institute of North America. My appreciation also goes to Laura Finsten, Evelyn Legare, Karen McCullough, Peter Schledermann, Greg Schwann, Ann Stevenson, Mary Ann Tisdale, Wendy Walker, and Anne Underhill for advice, inspiration, technical aid, hospitality, and, most of a l l , for friendship. I also thank my " substitute mothers": Lyla Bernard, Nancy Benedict, Marleen Spi t t a l and, especially, the late Elizabeth "Grammy" Barry, who made my daughter Elizabeth's time away from home a happy experience. My husband Tom has participated i n a l l stages of this research. He counted flakes, commented on countless drafts, helped with s t a t i s t i c a l problems, and sacrificed evenings, weekends and holidays to ensure that I had the most important quality of a l l - time. I thank him and my daughter Elizabeth for always providing the love and encouragement which enabled me to make i t through one more draft. I dedicate this dissertation to the memory of the late Moira Irvine, i n recognition of her substantial contribution to student research projects carried out throughout B r i t i s h Columbia. She provided me, and other graduate students at the University of B r i t i s h Columbia, with l o g i s t i c a l expertise, technical s k i l l s and, not least of a l l , a "friendly ear". 1 CHAPTER I INTRODUCTION The goal of this research is to examine the organization of microcore technology within the subsistence-settlement system of prehistoric hunter-gatherers. The study of technological organization involves investigation of why a society selects particular tool designs, and how i t structures the manufacture, use, maintenance, and discard of tools and tool manufacturing debris across the landscape. Tool design and distributional patterns reflect the organizational needs and constraints of a society's subsistence-settlement system, and its relationship to the biophysical environment. Although microcore technology is well represented throughout western North America, archaeologists s t i l l lack a testable model of the actual uses of microblades, the method and trajectory of manufacture, and the organization of microcore technology within regional subsistence-settlement systems. In this research, a model is constructed which deals with the technological, functional and spatial aspects of microcore technology in hunter-gatherer adaptations. A test of the model is carried out through a comparative case study of both tools and debitage from microlithic and non-microlithic sites from two areas in the British Columbia southern Interior Plateau. Microcore technology belongs to a subset of core technologies classified as standardized or prepared. Standardized core technologies are defined as the production of flakes or blades from a core with a platform prepared and regularly maintained by specific procedures. Until recently, the investigation of standardized core technologies was restricted to descriptions of the technological procedures involved and their place in 2 local cultvtral-historical sequences. However, a shift in focus to the organization of technology has directed investigation to other parameters associated with standardized core technologies in a variety of archaeological settings. Recently, researchers have associated the organization of standardized core technologies with a number of characteristics: a limited resource base (Torrence 1983; Johnson 1987a); a high level of time stress; (Torrence 1983); high residential mobility (Koldehoff 1987); a specialized function (Johnson 1987a); a range of functions (Clark 1987) and maximum cultural complexity (Morrow 1987). However, the research upon which these assumptions are based was carried out almost exclusively on sedentary, often complex societies. The limited amount of research which investigated residentially mobile hunter-gatherers has indicated that standardized core technology is associated with a need for specialized tools (Custer 1987) and the conservation of raw material (Parry and Kelly 1987). Although this recent research has laid the foundation for the formation of a general model of the organization of core technology among prehistoric groups, the bulk of i t was based on single site studies and a sampled artifact assemblage. In order to carry out further research in this direction, several issues relating specifically to mobile hunter-gatherer subsistence-settlement systems must be addressed: the actual use of the products of standardized core technology; the material implications of the organization of technology in relation to the complete lithic assemblage; and the regional distribution of standardized core technology. The distribution of microcore technology in the southern Interior Plateau of British Columbia has received little analytical attention as to its 3 significance in prehistoric human adaptation. Most of the research on the significance of microcore technology in this region has focused on description of its technological attributes and its place within local cultural-historical sequences (Sanger 1967, 1968, 1970b; Donahue 1975, 1977; Wilmeth 1971, 1978). Although the dating of microlithic components, in both riverine and upland sites, is variable, the trend is to accept Sanger's (1967, 1969) original argument that microcore technology is not significant after 3500 B.P. (Richards and Rousseau 1988; Hayden et al. 1987). Thus, microcore technology is viewed as representative of a residentially mobile forager resource procurement strategy (Richards and Rousseau 1988; Hayden et al. 1987). However, initial investigations of the resource procurement strategy associated with microlithic components suggest that microcore technology is present both at base camps and limited activity sites, and may represent features of a collector resource procurement strategy (Ludowicz 1983). This research project is based upon the proposition that the use of microcore technology is an organizational response to constraints or problems offered by the biophysical environment. This investigation consists of a case study of microcore technology in an region known to have been inhabited protohistorically by seasonally sedentary hunter-gatherers. The two areas selected for study, Upper Hat Creek Valley and Highland valley, have a variety of seasonally available resources that were collected and processed during short-term occupations of residential camps and field camps. Previous archaeological and ethnographic research has shown that microcore technology is present in both these valleys, and there are some appreciable differences and similarities in both in the resource base and 4 appreciable differences and similarities in both in the resource base and inferred prehistoric subsistence-settlement patterns. In order to determine the uses of microblades and the role played by microcore technology within the regional subsistence-settlement system, assemblages from microlithic and non-microlithic sites in both valleys are described, evaluated and compared. Few previous technological studies have attempted to document and explicate the full range of lithic production strategies employed by any specific cultural group and to examine how lithic procurement, production, and maintenance activities were integrated into the subsistence-settlement systems. Clarification of the relationship between these parameters and standardized core technology will contribute to the formation of a general model about the relationship between core technologies in general and the cultures that used them. Chapter II presents the research design of this study. The theoretical framework is based on the assumptions and definitions of cultural ecology, which assumes that lithic technology is a selected adaptation to biophysical environmental constraints. A method for elucidating the organizational properties of the society under study is developed, following the recent approach of archaeologists who view technology as a system of strategies which is organized to cope with problems posed by the environment and which has predictable material consequences both in the design of the tools and in their distribution. In this study, organization of technology refers to the interrelationships among chipped stone tool manufacturing, utilization, recycling and discarding activities in a prehistoric context. The archaeological implications of viewing microcore technology as one aspect of the technological organization, integrated with other parameters of the 5 role played by microcore technology in the adaptations of semi-sedentary hunter-gatherers is developed, along with research hypotheses and corresponding test implications, to evaluate the validity of the model. Chapter III presents an overview of the biophysical and cultural environments of the study areas: Upper Hat Creek Valley and Highland Valley. The section on the biophysical environment includes a current interpretation of the paleoenvironment and a description of the modern environment, including climate, water sources, vegetation, fauna, and lithic resources. The cultural environment section provides a brief synopsis of the prehistoric culture-historical sequence in the southern Interior Plateau of British Columbia, followed by a description and interpretation of the archaeological data from the study areas in relation to the regional picture. The implications of previous research into microcore technology in this region are discussed and evaluated. The final cultural section provides an interpretation, compiled from ethnohistoric observations, reconstructions and recent ethnographic research, of the subsistence-settlement system of the aboriginal inhabitants of the southern Interior Plateau at the time of contact. The last section in the chapter presents an interpretation of the prehistoric subsistence-settlement system of the inhabitants of the valley during the last 4500 years. Chapter IV describes the data base. The first section presents the morphological artifact classification scheme. The second section discusses site selection criteria, and describes each site with particular reference to biophysical location and characteristics, size, assemblage contents, density, features, radiocarbon dates, cultural affiliation, and previous interpretations of site function in relation to local subsistence and 6 settlement patterns. Chapter V describes the development and application of the analytical methods used to investigate the nature and distribution of activities associated with microcore technology. In order to have a comparative data base on which to test the research hypotheses, the study sites are re-classified using the same criteria, by the following methods: analysis of debitage reduction stages, analysis of tool use, and analysis of activity area patterning. The first section deals with the methodology for deriving the relative importance of sequential manufacturing stages from selected attributes on non-microlithic debitage. The second section presents and discusses the methods and results of replicative experiments in microcore technology. The initial goal of this particular segment of the study is to develop a manufacturing stage typology for microblades and microcores. The third section presents the methodology for the use-wear analysis developed in this research, which attempts to determine the range of uses for each tool, and whether or not the tools were multi-purpose. The fourth section describes the methodology for the activity area analysis applied in this study. The final section integrates and interprets the results of the three analytical methods outlined in the previous chapter, in order to construct a settlement typology for the study sites. Chapter VI presents the methods used to determine the validity of the test implications for the three research hypotheses. Each hypothesis is evaluated and the implications for the model are discussed. The final chapter assesses the methodological and substantive results of the investigation. The research model of microcore technological organization is summarized, evaluated, and modified using the results of the 7 hypothesis tests conducted in Chapter VI. The success of the major research methods is evaluated/ and recommendations are made for the direction of future research in this region. Contributions are made toward a general model of the organizational roles of standardized core technology in hunter-gatherer subsistence-settlement systems. 8 CHAPTER II RESEARCH DESIGN Theoretical Framework The theoretical foundation of this research rests on the assumptions and definitions of cultural ecology, which is the study of human beings and their relationships to their biophysical and social environments. Cultural ecologists assume that some of the variability in human behaviour is caused by variability in the biophysical environment. Steward (1955) was one of the first to develop the theory of cultural ecology by partitioning the biophysical environment and culture into specific features to examine their interrelationships. He viewed the environment not simply as setting limits on cultural behaviour but rather as closely related to a l l aspects of culture. According to this view, the technology and economy of a society are most directly related to the environment, while social and political organization may be less directly related, and religion, rituals, arts and myths are relatively free of direct environmental influence. Evolutionary theory, with its concepts of adaptedness and adaptation, provides an underlying framework for ecological studies. According to the cultural ecological paradigm, adaptedness is best understood as the selection of a valid set of culturally transmitted solutions to basic ecological problems, such as the procurement of food and shelter in a given environment (Jochim 1981). Thus, adaptation is viewed as the process of modifying these solutions, in response to changing environmental parameters or internally initiated processes (Kirch 1980). An adaptive strategy will 9 change in response to changes in the environment as well as to changes in the system itself (Kelly 1983). As a process, adaptation depends on a source of variability within each culture in order that responsive changes in behaviour may occur. In ecological terms, cultural evolution, in the sense of adaptedness to specific environments, is defined as change resulting from the differential persistence of variability in human behaviour, in the context of environmental selection pressures (Kirch 1980). According to this theoretical base, culture is assumed to be the primary means of adaptation to both the biophysical and social environments. Culture consists of learned patterns of behaviour, motives and strategies for survival, which provide one means of adaptation (Jochim 1981). An evolutionary theory of adaptation should account for both continuity and change in human behaviour, and must generate rules to explain and predict behavioural changes within groups with specific characteristics under stated sets of conditions (Alland 1975). In order to study the process of change, archaeologists view culture as a system or a group of components or variables interrelated such that a change in one produces a change in al l others, and the interrelationships among variables are as significant as the variables themselves (Jochim 1981). A systemic, adaptive paradigm of culture has been explicitly elaborated by Binford (1962), who adopted White's (1959) concept of culture as people's extrasomatic means of adaptation. Binford's (1962, 1964, 1968) theoretical contribution also included a focus on the definition and explication of variability and a recognition of the relevance of biophysical environmental variability. An important component of the cultural system is the economic subsystem, which includes information about the resources of a territory, and a 10 strategy for assigning a measure of the relative importance of each resource, as well as allocating and coordinating the level of effort directed toward the procurement and distribution of each resource (Clarke 1968). Variation in economic subsystems occurs primarily in the resource procurement strategies, or the manner in which resources are obtained, and in the mobility strategies, or the way in which groups position themselves across the landscape in order to procure resources. These land use strategies are related to the characteristics of both the biophysical resources and the preferences of the social group exploiting them. The technological component constitutes a major portion of the set of economic strategies selected by prehistoric hunter-gatherers, and may be more responsive to adaptive change in subsistence-settlement systems. In recent approaches to the interpretation of the remains of prehistoric economic subsystems, archaeologists have viewed technology as a strategy which is organized and has predictable material consequences (Nelson 1989). The procurement of raw material, and the design, production, and use of lithic tools may vary, depending upon the way in which technology is organized in any given group. Thus, tools used for the same purpose may assume different morphological forms and may be manufactured by different methods, and in different locations, according to how the society organizes its subsistence-settlement system. An understanding of how technology is organized is crucial to determining how and why technological change occurs. In addition, Binford (1980, 1982) stressed the importance of a regional perspective in order to delineate and explicate the total range of variability present within the economic subsystem of any particular group. Lithic tools and debitage are assumed to be a remnant of the 11 economic subsystem concerned with the procurement/ processing/ and consumption of resources/ and constitute material evidence for the echnological organization of the economic subsystem. Lithic tools and debitage preserve well and retain physical indications of the techniques of reduction/ resharpening and refurbishing. Therefore/ a description and explication of the variation in design and distribution of lithic tools and debitage in a regional sample of archaeological sites may provide evidence of the differential manufacture/ use and discard of tools, and of the variation in the performance and spatial distribution of these tasks. The Organization of Lithic Technology Definition Technological organization is defined as "the way in which a society designs its tools and arranges tool production/ use, and maintenance, so that the tools constitute an effective response to the problems faced by the society in its daily interactions with the environment" (Koldehoff 1987: 154). This view of technology as a system of strategies focuses the investigation of technological change on a set of cultural behaviours by which human groups adjust to changing biophysical and social environments (Binford 1979). Changing environmental parameters occur through space as well as time; for example, a group may change its technological organization from season to season as the resource procurement strategy changes from residential to logistical mobility. 12 Resource Procurement Strategies Recent research into the technology of hunter-gatherers has demonstrated that the two variables most influencing the organization of technology are the group's mobility and i t s resource procurement schedule (Binf rd 1977, 1978a, 1979, 1980; Jochim 1981; Torrence 1983). A useful device for examining the economic strategies of hunter-gatherers i s Binford's (1980) model of resource procurement strategies which focuses on these two variables. According to this model, foragers use high residential mobility i n order to be close to resources which are procured daily i n small amounts on an encounter basis. Processing and consumption of resources occurs daily, and food storage i s lacking or minimal. Residential mobility i s viewed as an organizational response to a homogeneous biophysical environment, and produces two archaeological s i t e types: base camps, called residential camps i n this study, and resource acquisition locations (Binford 1980). A residential camp i s defined as the locus of most preparation, processing and consumption of edible resources, as well as manufacturing and maintenance a c t i v i t i e s (Binford 1980). V a r i a b i l i t y i n the contents of the residential sites should r e f l e c t seasonal a c t i v i t i e s and duration of occupation. Locations should have low archaeological v i s i b i l i t y due to the small amount of resources procured and the limited amount of f i e l d processing. At the opposite end of the continuum, i n this model, are collectors who procure and process food resources for storage, i n addition to supplying daily requirements. Logistically organized work parties, which may consist of small groups of adults or family units, move to specific resources for a short period. Binford (1980) interpreted collecting as an adaptive response to the problem posed by widely spaced and simultaneously available c r i t i c a l 13 resources. In addition, Binford (1983) suggested that storage is a successful subsistence procurement strategy only i f the food resource is reliable and highly aggregated. As well as base, or residential, camps, collectors may occupy a variety of other sites, also known as special-purpose sites: field camps or temporary residences for logistical task groups; resource acquisition and processing locations; stations where special-purpose task groups gather to collect information about future resources; and caches to provide temporary storage for resources in transit from locations to residential camps (Binford 1980). Seasonality and the nature of the specific resource being collected may affect the composition of the assemblages of al l five site types. In addition, locations may contain evidence of the large-scale procurement and/or processing of resources, for example, roasting ovens. According to Binford's model, settlement systems which include a logistically organized component should be characterized by a high degree of variability among assemblages which is related to special-purpose functions. However, site assemblages other than base camps should be internally consistent in structure and content, because occupation time is very short and the activity is only a portion of a resource procurement strategy. Archaeological evidence indicates that the use of logistical strategies increased throughout the prehistoric occupation of North America (Kelly and Todd 1988). However, most ethnohistoric groups practised a subsistence-settlement strategy which involved a combination of residential and logistical mobility strategies. Residential and logistical variability are not opposing principles but organizational alternatives which may have been employed in different combinations in different settings. For example, a 14 group may have relied on stored resources for part of the year, and resorted to foraging for the remainder of the year. During the early historic period in the southern Interior Plateau, extended families occupied a winter base camp for several months, and relied on stored food and logistically organized hunting parties. In the spring the same group became residentially mobile, although logistically organized work parties continued to operate from field camps. Or, a group may have been residentially mobile but stored a portion of each season's resources near the residential camp. For example, the Tahltan combined residential moves with logistically-organized work parties, storing edible resources near summer villages, near early fall camps, and near fall and winter camps (Albright 1984). Recent analyses of site assemblages have focused on the problem of distinguishing between sites occupied primarily as residences and sites occupied for task-specific activities. Pokotylo (1978) identified residential and special-purpose sites in the southern Interior Plateau on the basis of tool assemblage content and debitage traits. Assemblages containing high frequencies and diversity of tool types are also characterized by high correlations of complete projectile points, projectile point tips, biface fragments, microblades and unifacially retouched flakes (Pokotylo 1978). The debitage results from a wide range of manufacturing activities. Pokotylo (1978) interpreted these sites as representative of intensive occupation involving a wide range of activities. Other assemblages containing only fragments of projectile points and bifaces are characterized by debitage primarily from the late stages of tool manufacture and maintenance. These were interpreted as limited activity or special purpose sites (Pokotylo 1978). Finally, Chatters (1987) compared archaeological 15 assemblages derived from residential camps, representative of a residentially mobile resource procurement strategy, with those derived from field camps, representative of a logistically mobile resource procurement strategy. Chatters (1987) found that residential camps are larger, contain a less diverse tool assemblage, and display less within-site type similarity of tools. Technological strategies The technological strategies, or decisions related to which tools and techniques to use, provide additional solutions to problems of resource procurement. Choices made about how, when and where to direct procurement activities can have a major influence on the nature of the technology. Technological choices may be frequently understood by reference to their implications for procurement reliability and efficiency (Jochim 1981). In addition, particularly for collectors who must store at least a portion of the resources procured, reliable and efficient technology for processing is essential (Schalk 1977). An emphasis on pocurement reliability may be most adaptive in those contexts associated with significant spatial and seasonal variability, and with risk in the natural and social environments (Jochim 1981). The higher the risk, the more oriented toward security rather than efficiency the technological strategy may be (Torrence 1983). In addition, societies relying on stored resources limit the majority of their procurement and processing activities to a portion of the year, and will probably want to maximize returns by emphasizing reliable methods of procurement and processing. 16 Efficiency can be defined as the successful completion of a task in the shortest time possible, and can be measured in terms of time savings or labour savings (Jochim 1981). Time efficiency is particularly important when a resource is only available for brief periods of time. For example, roots and berries are at their best for only a few weeks each year. In addition, migratory resources, such as salmon and birds, are in a particular region for a limited time. Another context favouring the selection of time efficient technology is the restriction of conditions suitable for resource procurement or processing. For example, salmon can be dried for winter storage only during dry, windy weather. The second type of procurement cost is labour. Technology can be used to increase labour efficiency either by reducing the amount of individual exertion required or the number of workers (Jochim 1981). Tools can be made more durable, thus reducing the amount of time required to manufacture and/or repair them, or easier and faster methods of tool manufacture can be developed (Bleed 1986; Jochim 1981). The need to ensure access to reliable and efficient tools while maintaining a subsistence-settlement strategy based on high residential and/or logistical mobility may be resolved by technological strategies which solve the problem of the incongruence between sources of lithic raw material and the locations where tools must be used (Bamforth 1986; Binford 1979; Keeley 1982; Parry and Kelly 1987). Binford (1977, 1979) distinguished between two types of technological strategies: curated strategies which involve the planned manufacture, use, maintenance and recycling of tools, and expedient strategies which incorporate the situationally or unplanned manufacture, use and discard of tools. Nelson (1989) recently clarified this distinction by defining both curation and expediency as planned strategies, 17 and opportunism as unplanned or the result of inadequate anticipation. According to Nelson (1989), opportunistic designs correspond to the minimal tool capable of performing the required task; these tools have very littl e impact on overall tool kit design and will probably not occur at residential or field camps because the appropriate raw material for crura ted or expedient tool types will be available there. Little effort is invested in either design or manufacture. Nelson (1989) also suggested that the material implications of opportunistic designs are difficult, i f not impossible, to predict because they are produced by the needs of the moment, the condition of the tool kit and the type of raw material available. Microcore technology can not be considered an opportunistic design because of the requirements for suitable raw material, specific manufacturing tools, and a suitable manufacturing location and time. Thus, opportunistic designs will not be considered further in this study. Both curation and expediency depend upon anticipation of the conditions of future activities with respect to the availability of tools, raw materials and time (Nelson 1989). derated tools are probably characterized by a high level of investment in manufacture and maintenance, and a well-developed storage and caching strategy (Binford 1979). Expedient tools may exhibit very little expenditure in manufacture, no investment in maintenance, and infrequent specialization (Binford (1979). In both these definitions, there is an untested assumption that only tools requiring a high level of energy or skill will be curated, specialized tools, while those tools manufactured expediently will not be either curated or specialized. These assumptions should be tested on each archaeological assemblage under study to determine, i f possible, which tools are, in fact 18 "curated" for future use and which are used "expediently". Another dimension to curation which has not been explored in the archaeological literature is the preparation of various types of cores which are then carried on foraging or collecting trips, or from one residential base to another, in anticipation of the production of tools for future use. Nunamiut informants commonly referred to carrying prepared cores into the field; these were discoidal in shape and used to produce butchering flakes, which were later modified into scrapers. The exhausted core was also reworked into a scraper (Binford 1979). In addition, Western Desert aborigines prepared cores at quarry sites, and carried them to a residential camp for further reduction into flakes and hafted tools (Gould 1978). Again, an important factor in whether or not cores are curated may be the amount of preparation time involved; that is, standardized cores which require more effort and a higher skill level may be more likely to be curated for future reduction than unstandardized cores. This study uses Nelson's definition of expediency because i t solves an apparent contradiction in the use of cores as being both curated (planned) and expedient (unplanned). The definition of curation is also expanded here to include cores as well as tools. Curation is the manufacture of tools in advance of use, or the preparation of cores in advance of flake or blade production and subsequent transport; both require a high level of energy input. The tools and cores are subsequently transported to the location of use or reduction. Curation may be employed in varying degrees with respect to spec fic tools and tool kits depending on the conditions to which this strategy is a response. Curation can be used to solve the problem of incongruity between availability of tools or tool manufacturing materials, 19 and the location of tool using activities (Bamforth 1986; Binford 1979; Keeley 1982; Parry and Kelly 1987). Additionally, curation may solve the problem of coping with time stress, associated with the procurement of mobile resources or resources with short periods of availability (Ebert 1986: Torrence 1983). Expediency is a technological strategy which involves the manufacture of tools, using a low level of energy input, at places where they will be used and discarded. Technological expediency may be dependent on knowledge of and access to adequate supplies of raw material, on enough time to produce the tools, or on a low level of mobility. The raw materials should be readily available or should have been previously transported and cached (Parry and Kelly 1987). The location should have been occupied for a fairly long time or regularly reoccupied, in order that use of the stored material takes place. In addition, there must be sufficient time available to manufacture tools (Torrence 1983). A sedentary community could have stockpiled raw material, and could have performed most daily tasks with expediently-produced flakes (Koldehoff 1987). Torrance's (1983) model explains in part the distinction between curated and expedient technology by suggesting that time stress is a major factor affecting technological organization. Those subsistence strategies which rely on a small number of relatively mobile, or seasonally limited resources, are characterized as "time-stressed", that is, the organization of time for the procurement and/or processing of these resources is critical for the success of this particular type of adaptation. In these situations, i f task efficiency is critical and i f a large number of specialized tools are more efficient, then standardized core technologies may have been used 20 to produce them. Other technological solutions to high tine stress may include the use of composite tools, a dependence on curated tools, and the manufacture of specialized tools from organic materials (Torrence 1983). In those subsistence strategies where the emphasis is placed on relatively abundant resources, the need to be time efficient may less critical. Dependence on a greater range of subsistence resources, in association with more time available for procurement and processing, may be related to the manufacture and use of generalized tools, presumably, although not necessarily, from unstandardized cores. It is important not to perceive curation and expediency as dichotomized systems, but as options that probably suit different conditions within a set of adaptive strategies. Curated and expedient plans can be interwoven. For example, cores may be prepared in advance of reduction, curated and reduced at another location in the seasonal round. The flakes or blades removed thus become expedient tools, manufactured, used and discarded at the same location. Alternately, flakes and blades may be hafted and the composite tools will be curated for future use. It is important not to assume, without testing in the archaeological record, that expediently-produced tools will also be expediently-used and discarded tools. Binford (1972, 1973, 1977) predicted that, under conditions of low curation, sites used for different activities will have high inter-assemblage variability because most of the tools will be discarded at the sites where they were used. Under conditions of high curation, sites which are used for different purposes will probably manifest low inter-assemblage variability because most of the tools used there will be resharpened or recycled and taken to the next location. Of course, these predictions derive 21 from assumptions that the majority of curated artifacts are specialized tools, and that few tools are worn out at each site. Hayden (1987) pointed out that i f tools wear out or break then they will be left at the site where last used, regardless of the type of site; thus the proportion of tool types present in residential sites will s t i l l reflect the activities which occurred there. Special activity sites should s t i l l have distinctive assemblages because wide ranges of activities do not occur there. The effects of curation are probably negligible in sedentary village sites, but may be more pronounced in special activity sites (Hayden 1987). The majority of broken stone tools are probably not returned to the residential base camp site for repair; exceptions may be high investment and/or complex tools, such as hafted bifaces (Hayden 1976, 1987). In these cases, the broken stone fragments would have been discarded at the site of use. Other important factors influencing the disposal of tools may be: location of primary use; potential recycling value; complexity of repair; effort invested in original manufacture; fragility; size; and degree to which tool parts break into small pieces (Ammerman and Feldman 1974; Hayden 1987). Another material implication of technological expediency may be low investment in retouch, because expedient tools are made and used when and where the need occurs, and discarded when worn. Expediency has been associated primarily with the production and use of unretouched and marginally retouched flakes (Parry and Kelly 1987; Johnson 1987b). In addition, a high percentage of flakes should show evidence of use-wear (Johnson 1987b). Finally, cores in various stages of reduction should be located at residential sites where expedient tool production is important (Nelson 1989). 22 Archaeologists have focused on two parameters which may reflect the technological organizational strategy of the group: the design of tools and the distribution of tools and debitage across the landscape (Nelson 1989). Lithic tools can be designed to emphasize reliability (Bleed 1986; elly 1988), maintainability (Binford 1979; Bleed 1986; Kelly 1988; Shott 1986), transportability (Kelly 1988; Parry and Kelly 1987; Kelly and Todd 1988; Shott 1986), and conservation of raw material (Bamforth 1986). These may al l have material implications for tool form, core form, reduction technique, and maintenance technique. In addition, the differential acquisition, transport, caching, and reduction of cobbles, as well as the manufacture, maintenance and use of tools may produce variable distributions of these sequential stages, both within individual sites and in the regional settlement pattern. The section below will discuss the technological organization of standardized core technology by hunter-gatherers, in relation to these technological goals and their archaeological implications. Standardized Core Technology as an Orcranizational Strategy Introduction Core technology is defined as the production of flakes or blades for tools, and can be standardized or unstandardized (Johnson 1987a). Standardized cores are "those in which the platform is formed and maintained by specific procedures, excluding the occasional edge grinding which is evident on some amorphous cores" (Johnson 1987a:2). Examples of standardized core technologies include large blade technology, microcore technology and bifacial technology. The use of standardized core technology increases the 23 amount of production time per artifact but ensures the production of a flake or blade of a predictable size and shape, and conserves raw material. In addition, some standardized core technologies, for example, those involved in the production of large blades and those based on the wedge-shaped microcore, are characterized as having an inflexible manufacturing trajectory. Once the preparation of the core platform and face is complete, i t is almost impossible to use the core for any other purpose, without major reworking. However, other standardized core technologies, particularly bifacial, are characterized by a very flexible manufacturing trajectory and a product which can be used as either a tool or a core (Kelly 1988). Unstandardized, or amorphous, core technology is characterized by a complete lack of preforming or preparation, and a lack of intentional control over the form of the resultant flakes (Parry and Kelly 1987). Unstandardized core technology involves less production time, and no resharpening or rejuvenation time because i t is easier to produce a fresh flake with a sharp edge than to resharpen a dull flake. Standardized core technology has been associated faypothetically with a limited resource base (Torrence 193), a high level of time stress (Torrence 1983), high residential mobility (Koldehoff 1987; Johnson 1987b; Parry and Kelly 1987), a specialized function (Johnson 1987a; McNemey 1987) and maximum cultural complexity (Morrow 1987). Unstandardized core technology has been associated hypothetically with abundant raw material (Parry and Kelly 1987; Custer 1987), a diverse resource base (Torrence 1983), low residential mobility (Parry and Kelly 1987), high logistical mobility (Custer 1987), and a lack of time stress (Torrence 1983). Archaeologists tend to equate standardized core technology with the production of curated, 24 specialized tools, and unstandardized core technology with the production of expedient, generalized tools (Parry and Kelly 1987). Furthermore, until recently (cf. Parry and Kelly 1987), the use of standardized core technologies was associated almost exclusively with residentially based, often complex societies, while the use of unstandardized core technologies was associated with residentially mobile hunter-gatherers. McNerney (1987) associated standardized core technology with a sedentary subsistence-settlement system because the product is expedient but standardized for a unique purpose. Placement of retouch on blades was examined and interpreted, according to Odell's (1981) criteria, as the result of light duty cutting and scraping. Morrow (1987) also examined sites occupied by a sedentary society, and suggested that standardized core technology is associated with the predictable, constant needs for specialized tools. Koldehoff (1987) also proposed that the standardized core technologies located at the permanent residential sites investigated may have been the product of specialists. Clark (1987) proposed that large blade production is efficient only when there is sufficient centralization in the economy to allow craft specialization, and also associated this particular technology with the conservation of high quality imported raw material. Johnson (1987a) stated that the majority of prepared core technologies produce a tool designed for a single, highly specialized purpose, and suggested that the main advantage of prepared core technology is not conservation of raw material, but the production of a large number of blanks of a predictable shape and size. Although Torrence (1983) did not specifically address the question of standardized core technology, she did suggest that a low stress subsistence strategy based on a large number of 25 relatively abundant resources should be associated with the use of generalized tools which require less manufacturing time. In order to understand the significance of the use of standardized core technology, we must consider the place of standardized tools in the overall organization of stone-based technology. The following is an examination of standardized core technology in relation to the goals of technological production, and the hypothetical archaeological implications of these goals for the strategies of tool design, production, use and discard patterns of tools and debitage, adapted from Nelson (1989). Design Goals of Standardized Core Technology 1. Reliability Bleed (1986) defined a reliable design as one which always works when needed. Factors in the manufacture of a reliable design include standard replacement parts and a sturdy construction (Bleed 1986). Although reliable designs may reduce task performance time, there is a corresponding increase in manufacture and maintenance time, and in the amount of raw materials required (Nelson 1989). Reliable designs may be responses to situations where time is limited; the resource location is unpredictable; and the procurement strategy is primarily logistical (Bleed 1986; Binford 1978a; Nelson 1989; Kuhn 1989). Reliable technology is probably almost exclusively associated with hunting, and does not appear to offer advantages for processing and tool manufacturing activities. The material implications of reliable designs include a predictable replacement part produced most economically by a standardized core technique, and the production of a sturdy haft with secure fittings (Shott 26 1984/ 1986). The discarded replacement parts should display similar use-wear patterns because the use of the tool does not change. In addition/ similar use-wear patterns should occur on different formal tool types in assemblages designed for reliability (Nelson 1989). Finally/ the occurrence of a higher percentage of complete tools and a lower percentage of retouched/ resharpened tools may indicate an emphasis on replacing tools before they wear out. 2. Maintainability A maintainable tool is defined as one which functions under a variety of conditions and is usable even i f broken (Bleed 1986). Maintainability is probably the optimal design for generalized tasks which have continuous need but unpredictable schedules and generally low failure costs (Bleed 1986). Maintainable tools should be modular, so that broken parts can be easily removed and replaced (Bleed 1986). In addition, maintainable tools should be simpler, lighter and more portable than reliable ones (Nelson 1989). Acceding to this definition, maintainable designs can either be flexible or versatile (Bleed 1986). A flexible tool class can be used in a wide range of tasks (Shott 1986). A versatile tool is a multi-purpose tool, used sequentially in different tasks (Shott 1986). For both flexible and versatile tools, time invested in manufacture is made worthwhile by the advantage of always having a functioning tool. This advantage is important where the timing and place of use cannot always be anticipated, but where exploitation of a range of resources and occurrence of a variety of activities is anticipated (Nelson 1989). A second advantage of flexible and versatile designs is the simplification of tool assemblages. If groups using high residential mobility as a subsistence strategy must maintain limited 27 tool inventories, they may use multi-purpose tools. It also seems likely that efficient use of processing time for handling stored resources is important, especially where resources are available for a limited time or spoil quickly. Maintainable modular tools, like reliable tools, also depend on replacement parts of a standard shape and size. However, whereas reliable tools are always used for the same purpose, maintainable tools are used for a series of different purposes. In order that replacement modules be of a standard size and shape to f i t into the haft, a method for producing these must be available, i.e. standardized core technology. In addition, a simple repair kit must be maintained because tool forms must be altered (Bleed 1986). In prehistoric lithic technologies, this kit might include bone and stone hammers, prepared cores, and resins (Bleed 1986). In addition, the bifacial or disc core is considered to be a flexible design (Binford 1979; Morrow 1987; Parry and Kelly 1987; Kelly 1988) because a variety of flake forms, to be used as tools, can be produced. The core itself may be used throughout the reduction sequence in a variety of tasks (Kelly 1988). If maintainability is a factor in prehistoric tool kit design, then certain material implications can be searched for in archaeological assemblages. Since the form of the working edge or edges is changed in order that different tasks may be performed, or by the performance of different tasks, the various parts used in a flexible tool will display different patterns of wear. Thus, a flexible tool kit will contain tool classes which each display several different wear patterns. On the other hand, in a versatile tool kit, individual tools should display multiple working edges or surfaces with evidence of several different use-wear patterns (Nelson 28 1989). 3. Transportability Others have investigated the relationship between the type of core technology predominant in a society and the level of residential and logistical mobility. Chapman (1977), Koldehoff (1987), Parry and Kelly (1987), and Kelly and Todd (1988) asserted that there is a correlation between the vise of a standardized core technology and a highly mobile lifestyle. Using a range of North American examples primarily from permanent village sites, Parry and Kelly (1987) demonstrated a correlation between a decrease in residential mobility and a reduction in the use of standardized cores. They concluded that, as groups become more sedentary, emphasis in lithic technological organization changed to the expedient production of unretouched flakes, or to the use of unstandardized cores. Parry and Kelly (1987) suggested that the primary advantages of standardized core technology relate to its portability and conservation of raw material. Once there was no longer a need for these advantages in the technological organization, then the extra time spent manufacturing standardized cores was no longer justifiable, and groups stockpiled raw material to use in unstandardized core technology. However, Parry and Kelly (1987) and others restricted their investigation to the base camps of semi-sedentary groups and did not investigate the nature of the core technology used in special-purpose sites. Also, they were confusing the production of standardized tools with curation and the production of nonstandardized tools with expediency. Parry and Kelly (1987) also suggest that compared to expedient tool technology, standardized tool technology is costly to use, manufacture, and maintain. Once a need for a mobile technology decreased, then expedient core technology may have 29 replaced prepared core technology. Although this model refers primarily to biface core technology and the production of formed tools, i t offers support for my e a r l i e r hypothesis relating the use of microcore technology to high mobility i n upland areas (Greaves 1986). According to Nelson (1989), the key principle of a transportable design i s that the tool k i t be made i n one location and carried to the location of use. I f mobility affects tool k i t design, then certain material implications may occur i n archaeological assemblages. To meet the constraints of mobility, a transportable tool k i t should contain only a few items, and be lightweight (Gould 1968; Ebert 1979; Lee 1979). In order to maintain a small toolkit, a group may conserve and maintain those items. I f a toolkit has few items, some of these tools may be either versatile or flex i b l e , as discussed above. Although Nelson (1989) only considered v e r s a t i l i t y and f l e x i b i l i t y i n relation to composite tools, i t i s easy to apply these concepts to simple tools which assume multiple functions throughout their lifespan. Again, a versatile tool class w i l l display several different types of use-wear indicating the number of tasks i n which i t i s used (Shott 1986). F l e x i b i l i t y i s defined as the range of tasks i n which a tool i s used, and a f l e x i b l e tool class can be identified by the number of unrelated tasks i n which i t i s used (Shott 1986). While Nelson (1989) suggested that large-blade technology i s not suited to transport, other standardized core technologies are, e.g. b i f a c i a l or microcore. 4. Conservation of Raw Material This i s the classic explanation for the use of standardized core technology: the use of standardized tools may allow highly mobile groups to transport sufficient raw material from the source to the location of use so 30 that both anticipated and unanticipated needs can be met (Johnson 1987a). Standardized cores used to produce multipurpose tools may be common in the lithic technology of residentially mobile populations because of the need to transport raw materials from source to source. In addition, infrequent or new occupants of an area may continue to retain conservative technologies, even though raw material may be abundant, because of unf ami liarity with the location of the resource (Kelly 1988). Hayden (1989) has also suggested that blade tools were adopted for woodworking and other tasks because edges could be resharpened repeatedly, resulting in a reduction of time spent in procurement of lithic resources, which may have been scarce or seasonally unavailable. Finally, use of standardized core technology may be confined to one particular raw material type, which has restricted access for a variety of reasons. Using ethnographic and archaeological data, some archaeologists proposed that the availability of suitable raw material is the primary influence on tool kit design (Gould and Sagger 1985), and that variability in core form, reduction technique, tool maintenance and recycling are responses to shortages of stone (Bamforth 1986). Custer (1987) suggested that the low incidence of standardized cores at a temporary logistical camp occupied by semi-sedentary hunter-gatherers is due to the abundance of locally available raw material. However, he failed to note that the majority of tools at this camp are the product of standardized bifacial technology. Parry and Kelly (1987) and Clark (1987) also proposed that the use of bifacial tools is a response to a shortage of raw material by highly mobile groups. Bamforth (1986) argued that maintenance and recycling are more closely related to availability of stone material than to settlement organization or time 31 limits on the activities for which the tools are used. However/ Nelson (1989) pointed out that technology is part of an economic system, and tool use and recycling solves problems of adaptation that involve a variety of environmental and social constraints. People may carry/ maintain and reuse tools because the resource is highly mobile or requires immediate processing, or because considerable energy has already been expended in the manufacture of the tool. Although Nelson suggests that lithic material is not a resource in the same sense that edible resources go through a natural cycle, its availability is s t i l l constrained by seasonal factors, such as snow coverage or high elevations. Lithic resources can also be circumscribed to particular localities, or restricted in access by some groups. If the shortage of lithic raw material affects tool kit design, we would expect to find smaller core sizes, a higher percentage of bipolar cores, flake tools more aggressively utilized prior to discard, and a larger number of tools that were maintained and reused regularly (Bamforth 1986). In addition, when the majority of debitage derives from later reduction stages, then the resharpening of tools to extend their use life is indicated (Bamforth 1986). Distribution of Standardized Pore Technology Another approach, although not explanatory in the same sense as the approaches outlined above, is the investigation of the distribution of the various steps in the manufacturing trajectory, and use and discard patterns of the particular core technology under study. A salient feature of standardized core technology is the staging required: several different knapping procedures need to be carefully followed in a predetermined 32 sequence, in order to transform a chunk of lithic raw material into a suitable core and then into flakes or blades. Each stage can be considered a separate task or production episode (Kelly 1984; Clark 1987). All sequential tasks can be, but are not always, completed at one location. The potential for segmenting the production process either temporally or spatially has important implications for the organization of standardized core technologies. Regional studies of technological organization have investigated the distributional implications of technological planning in terms of classes of sites: base camps, residential camps, field camps, and stations (processing locations, extractive sites, and lookouts) (Binford 1977, 1978b, 1979, 1980; Raab, Cande and Stahl 1979; Custer 1987). Residential sites (base camps and residential camps) should provide the time to manufacture and repair tools. In a technological strategy based on curation, transportable tools should be made at base camps as well as residential camps (Binford 1979; Ebert 1986; Kelly 1988). Binford (1979), Torrence (1983) and Ebert (1986) also argued that transportable tools should be returned to the base camps to be maintained and recycled. However, Hayden (1987) suggested that only complex tools, such as hafted bifaces, will be returned for maintenance. In addition, standardized cores should be prepared for future reduction at residences. Expedient tool manufacture should also occur at residential camps i f the raw material is available (Keeley 1982; Parry and Kelly 1987; Kelly 1988). At special-purpose sites (field camps, processing locations, extractive sites, and lookouts) very little tool manufacture or maintenance may occur (Camilli 1983). Activities are probably focused on those tasks that can only be done efficiently or effectively at that particular location. Assemblages 33 at special-purpose sites may be indicative of a smaller range of activities than would be found at residences (Camilli 1983; Raab, Cande and Stahl 1979). Technological planning should focus on curated, transported tool kits. Nelson (1989) also suggested that when raw materials and time are available, and the location is regularly reused, some expedient tool manufacture may occur. Debitage at special-purpose sites should derive primarily from resharpening and rejuvenation of worn and dull tools (Ebert 1986) and from segments of composite tools that break and are not salvaged (Raab, Cande and Stahl 1979; Hayden 1987). If the transported tool kit is designed with sequential flexibility, the module that breaks may be different from one special activity site to the next (Nelson 1989). Thus the forms of tool fragments deposited in special activity sites may be highly variable among those sites as a class (Camilli 1983). Discussion The results of recent research into the organizati n of standardized core technology have indicated that the traditional explanation of raw material conservation is only one of several factors which may operate in favour of selecting this type of tool production. Other potential factors are which have yet to be adequately investigated include: a high level of residential mobility (Koldehoff 1987; Parry and Kelly 1987; Johnson 1987a, 1987b), a high level of time stress (Torrence 1983), a limited resource base (Torrence 1983; Johnson 1987a), and a specialized function (McNemey 1987). In addition, the role of any particular standardized core technology may change through time and/or space (Kelly and Todd 1988; Greaves 1986, 1987). Johnson 34 (1987a:11) has formulated a general statement about the use of standardized (prepared) and unstandardized (amorphous) core technologies, which can be considered as a model for further testing and refinement. Amorphous core technologies conserve tool production time and maintain a maximum number of trajectory options at the expense of raw material to derive a broad array of cutting edges and angles. They are ideal for subsistence systems based on diversified resources in an area where raw material is abundant. In areas where lithic resources are not common, more conservative technologies, usually bifacial, were used. However, the distinction between resource zone and non-resource zone is nullified at the transition from mobile to non-mobile settlement strategies. Sedentary societies were able to procure and store enough raw material to allow the use of amorphous core technologies. Sedentism and prepared core technologies are also coincident but the one major exception, the Paleoindian assemblages, points to the common underlying factor in these technologies. They are part of a technological system which is focused on one major activity with specific tool requirements. So the results of previous research appear to indicate that standardized core technologies are associated with a shortage of raw material and a specialized use in both sedentary and residentially mobile societies. However, the research on which this general statement is based has the following limitations. Investigation of archaeological data was confined, in most cases, to either a single site or a set of base camps within the same area (Custer 1987; Parry and Kelly 1987). in addition, analysis of assemblages was restricted to either the tools, or selected classes of debitage (Parry and Kelly 1987; Arnold 1987; Koldehoff 1987; Johnson 1987b). As well, very limited use-wear analysis was attempted to associate the actual tasks in which the products of standardized core technologies were used with other parameters of the design and distribution of these tools. Although there are significant differences among the various types of standardized core technologies, Johnson's (1987a) model also assumes that the same organizational role was fulfilled by a l l of them in hunter-gatherer 35 societies. Finally, with few exceptions (cf. Custer 1987; Kelly 1988) there has been very little research directed toward explicating the role of standardized core technologies in residentially and logistically mobile hunting-gathering societies. To conclude, this model is too simplistic because i t ignores the potential variability in the use of standardized core technologies through space as well as through time. The following section will present a model of a particular type of standardized core technology to serve as a test case for the assumptions discussed above. Model of the Organization of Microcore Technology Definition of Microcore Technology Microcore technology is a specialized subset of standardized core technology that comprises the production and use of microblades for tools. A microcore is a standardized core which displays evidence of a striking platform or of core edge preparation, as well as preparation of the fluted face, or ridges where microblades will be later detached. According to Sanger (1970b), the most useful criteria for the recognition of a microcore are evidence of successive removal of blades and evidence of deliberate core platform preparation. Microblades, as defined by Sanger (1970b), are small, long narrow parallel-sided flakes, with at least two linear dorsal ridges, and evidence of a prepared platform. Within a microcore technology, blades are detached from a carefully prepared core and are used to perform tasks with a minimum of further modification. Several types of microcores exist in northwestern North America (Sanger 36 1968), including wedge-shaped (Fladmark 1986a; Sanger 1970b), blocky with a wide keel (Sanger 1968), blocky with little preparation of sides and keel (Sanger 1968), cube-like (Sanger 1968), and bifacial (Fladmark 1985). This research deals only with the wedge-shaped core with a narrow keel, originally defined by Sanger (1970b), and found throughout the Interior Plateau of North America. The wedge-shaped microcore is not a flexible design because i t produces only one type of flake and cannot be used as a tool without sustaining damage which makes its further use as a core impossible, unless completely rejuvenated. The core itself has little, i f any, potential for creating other tool forms, except perhaps as a form of crude heavy scraper or as a source of a few generalized flakes. Goals of the Model The traditional models of microcore technology relate its use to: movements of Athapaskan speakers; the production of a specialized hafted tool; and a foraging subsistence-settlement system. These models have never been explicitly formulated and tested. As well, they do not account for the widespread geographical and chronological distribution of microcore technology, nor its apparent association with late prehistoric sites occupied by semi-sedentary hunter-gatherers utilizing a combination of foraging and collecting resource procurement strategies. A viable and testable model of the organization of microcore technology should account for its utility within a l l pertinent adaptive contexts. The primary goal of the model presented below is to explicate the role of microcore technology as an adaptive response used by residentially and logistically mobile hunter-gatherers. The model will predict how microcore 37 technology is organized in prehistoric foraging and collecting hunting-gathering cultures, and will be multi-facetted in order that variability in adaptive technological responses may be incorporated and tested. The model will outline what microcore technology is designed to do within the system in terms of the technological goals discussed in the previous section, and will describe the predicted distribution of the manufacturing trajectory within the regional settlement system. In order to establish the significance of microcore technology, i t is not enough to simply demonstrate the actual use or uses of microblades because the same tasks can be carried out successfully by other tools. The model must provide an explanation for why people selected this method for producing artifacts of this particular type. Salient Features of the Model The model, which is elaborated on below, predicts that microcore technology is an adaptive response to high residential mobility, a shortage of lithic raw material, and a need for a flexible and/or versatile tool assemblage. In addition, the model predicts that microblades, which are the end product of microcore technology, are both manufacture and used at residential camps, associated with both foraging and collecting resource procurement strategies. Thus, microcore technology can be understood as a useful technological strategy for societies which are primarily collectors as well as those which are primarily foragers. In addition, microcore technology is not considered to be restricted to any one ethnic group. 38 Technological Organization Goals This section describes the predictions of the model in terms of what microcore technology is designed for, in relation to the technological organizational goals of the society using i t . The model posits that microcore technology was a method for producing and preserving a maintainable tool kit. Maintainable tools function under diverse conditions, and are most adaptive in a context of continuous need, unpredictable scheduling, and low risk (Bleed 1986). Maintainable designs can either be flexible or versatile (Bleed 1986). A flexible tool is one which can be used in a wide range of tasks (Shott 1986). A versatile tool is one which is multi-purpose, that is, used in more than one type of tasks. According to the model, microcore technology is designed to enhance maintainability by producing versatile and/or flexible tools and tool assemblages. The model also predicts that transportability is a second major advantage of microcore technology. According to Nelson (1989), the key principle of a transportable design is that the tool kit be made in one location and carried to the location of use. While Nelson (1989) suggested that large-blade technology is not suited to transport, microcore technology is because specialized production tools are not required and the core is small and light (Kelly 1984). In addition, only the core itself should be portable; the blades are expediently produced, used, and discarded. Microblades themselves are not practically portable except in the form of a core or in a haft. Detached microblades are extremely fragile, and highly susceptible to breakage and edge damage. Microblades may have been hafted into a composite tool and carried to other sites in the haft (Flenniken 39 1981). The third design goal derived from the model associates microcore technology with the conservation of lithic resources of a particular quality or type, by groups facing new territories or seasonally restricted access. Microcore technology represents an extremely efficient, systematic use of raw material. Microcores conserve raw material because they contain a large quantity of tool edge in relation to amount of raw material. Once the core is adequately shaped, each flake removed produces a usable blade; approximately 50 blades per core can be produced. In addition, each blade has a high ratio of useable edge to total amount of material (McNemey 1987; Hofman 1987; Sheets 1978; Morrow 1987; Clark 1987; Parry and Kelly 1987). Thus, microcore technology is particularly adaptive in an environment where high quality lithic resources are absent, at least seasonally, or scarce. In addition, infrequent or new occupants of an area may continue to retain conservative technologies, even though raw material may be abundant, because of unfamiliarity with the location of the resource (Kelly 1988). Finally, use of standardized technology may be confined to one particular aw material type, which has restricted access for a variety of reasons. Distribution of Microcore Technology This section describes the predicted distribution of the manufacturing trajectory of microcore technology within the regional settlement pattern. A salient feature of blade technology is the staging required - several different knapping procedures need to be carefully followed, and in their proper sequence, in order to transform a lump of raw material into a suitable core and then into blades. Each stage can be considered a separate 40 task or production episode. All sequential tasks can be produced at one location but need not be. However, i t is remember to realize that microblades are best produced in batches, as core manufacture and blade removal both require a significant amount of preparation time and concentration on the task. It would have been disadvantageous to spend time setting up a core for microblade removal i f only one blade were removed. Thus, the possibility of dividing up the production process temporally or spatially has important implications for the potential organization of microlithic industries. According to the model, microcore technology in northwestern North America occurs as an organizational response to the high residential mobility associated with two types of resource procurement strategies: (1) primarily foraging, and (2) semi-sedentary, logistically organized collecting. Microcore technology associated with time periods or regions where foraging strategies predominated will be found in sites which can be identified as residential base camps, according to Binford's (1980) model. Thus, forager-occupied residential camps will be the locus of microcore preparation, reduction and discard, and also the locus of microblade production, use and discard. Microcore technology associated with semi-sedentary collecting groups will be a strategy utilized during seasons when the groups were s t i l l residentially mobile, collecting and processing bulk resources. Thus, collector-occupied residential camps will also be the locus of most manufacturing, utilization, and discard activities associated with microcore technology. In addition, the model also predicts that microcores may have been manufactured at seasonally-occupied base camps, for further reduction into blades during other seasons at residential camps. 41 The model associates the production and use of microblades with residential camps because the manufacture of microblades required time and concentration which was available at residential camps. Microcore technology should be associated with sites occupied by small family groups rather than special task groups where time is in short supply. At short-term residential camps, blade production has economic value because raw material may not be stockpiled, especially i f the location of the camp is changed annually or is not predictable. In addition, i t is proposed that microblades were a multi-purpose tool which would have been most useful in a residential camp where a larger variety of tasks were performed. The model postulates that microblades were used for a variety of tasks, related more to food processing, and tool and container manufacture than food procurement. These are tasks which were probably performed more commonly at residential camps than at field camps. In the case of semi -sedentary logistically-organized collectors, i t is suggested that preparation of microcores, but not production of microblades, may have taken place at base camps, to take advantage of time available when procurement and processing of resources was not a major activity. At seasonally-occupied base camps, the probability that lithic resources were stockpiled in anticipation of time available for tool manufacture and repair very likely precludes the necessity for using a time-consuming technique like microcore technology to produce tools for immediate use. The model predicts that microcores were prepared in advance at residential sites, cached there or taken to the next residential camp. Microlithic debitage deposited in residential camps should consist of primarily nonviable microcores, plus microcore preparation and rejuvenation 42 flakes. The majority of microblades should have been produced and used at short-term residential camps. Microblades were probably produced in batches, and there may be only segments of the manufacturing trajectory present. A few worn microblades may be deposited at field camps or processing sites, i f they were originally fastened to a haft and discarded when no longer useful. Application of the Model to the Study Areas In order to test the model of the organization of microlithic technology presented above, a region was selected which is known to have been inhabited protohistorically by semi-sedentary, logistically organized hunter-gatherers. Within this region, the southern Interior Plateau of British Columbia, two study areas were chosen which contain large numbers of microlithic and non-microlithic sites. Between the areas, the edible resource base differs in variability and productivity, while lithic raw material varies in availability. Previous archaeological research in both areas indicates seasonal occupation, and a resource procurement strategy based on high residential mobility and collection of resources for winter storage. It is unclear at the present time whether the microlithic sites pre-date or are contemporaneous with the non-microlithic sites. However, the environment and resource base are assumed to be similar. A major hypothesis derived from the model is that microlithic sites in the study areas, two upland valleys, are associated with the inferred ethnographic subsistence-settlement pattern. The model predicts that microcore technology was not an essential feature of the technological organization at winter base camps in the major river valleys. The major advantages of this technology relate to its portability and conservation of 43 material/ which would not have been advantages in a sedentary camp. An exception to this prediction may have been the preparation of microcores during the winter months for further reduction at upland camps during the spring to fa l l months. Microlithic sites will be residential camps where families stayed for periods up to several weeks, while procuring seasonally available resources. Non-microlithic sites will be limited activity sites, either field camps where task groups stayed for a few days while procuring a specific resource/ or stations/ where a single task was carried out. Research Hypotheses The research hypotheses are derived directly from the model and relate specifically to expectations regarding the characteristics of the lithic assemblages associated with microlithic and non-microlithic sites in the study areas, that is, in upland valleys. The hypotheses are designed to determine which factors relating to technological goals are operative in the subsistence-settlement system associated with microlithic assemblages in the study areas. Test implications derived from the hypotheses were reduced to be non-redundant. HYPOTHESIS 1; Microcore technology is associated with the inferred  ethnographic subsistence-settlement pattern in the study areas. The inferred ethnographic subsistence-settlement pattern in the study areas can be described as a esidentially mobile collecting strategy. In Chapter III, ethnographic and archaeological data will be examined in order to provide a model of the way in which upland valleys were utilized during the ethnographic period. This model will include a description of the types 44 of settlements occupied and their location within the biogeoclimatic zones present in each valley. Hypothesis 1 will be tested by comparing the settlement types and associated biogeoclimatic zones of microlithic and non-microlithic sites with those predicted by the ethnographic model. Test Implications: 1. Microlithic sites will be residential camps, while non-microlithic sites will be special-purpose sites, either field camps or stations. 2. Microlithic sites will be located in the appropriate biogeoclimatic zones. HYPOTHESIS 2: Micropore technology is designed to contribute toward a  maintainable tool assemblage. Maintainable designs can be achieved by the production of either flexible or versatile tools and/or assemblages (Bleed 1986). A flexible tool is one which can be used in a wide range of tasks (Shott 1986), while a more flexible assemblage is one which contains the same or a smaller number of morphological tool types, but a larger number of actual uses, as a less flexible assemblage. A versatile tool is one which is multi-purpose, that is, used in more than one type of tasks. The range of functions for both individual tool types and tool assemblages will be determined by use-wear analysis. A comparison of individual tool types and tool assemblages from microlithic and non-microlithic sites will be carried out in order to determine whether or not microcore technology produces a more maintainable tool and/or assemblage. Test Implications: 1. Microblades will be a more flexible tool than other tool types. 45 2. Tool assemblages at microlithic sites will be more flexible than those at non-microlithic sites. 3. Microblades wil l be a more versatile tool than other tool types. 4. Tool assemblages at microlithic sites w i l l be more versatile than those at non-microlithic sites. HYPOTHESIS 3; Microcore technology is designed to contribute toward a  transportable tool assemblage. To meet the constraints of mobility, a transportable tool kit must contain only a few items, and be lightweight (Gould 1968; Ebert 1979; Lee 1979). In order to maintain a small toolkit, a group must conserve and maintain those items. If a toolkit has few items, some of these tools must be either versatile or flexible, as discussed above. Another important implication of an organizational strategy oriented toward transport of cores as well as tools is that the majority of tasks will be carried out by the use of expedient flake and cobble tools. A comparison of microlithic and non-microlithic assemblages will be carried in order to determine whether or not microcore technology is associated with the production of transportable assemblages. Test Implications: 1. Complete formed tools in microlithic sites w i l l be lighter and smaller than complete formed tools in non-microlithic sites. 2. Evidence of microcore manufacture and rejuvenation will not be found at a l l sites containing microblades. 3. Some microlithic sites will contain microblades belonging to different stages of the production sequence. 46 4. Microcores found i n the study sites w i l l not be viable. 5. Microlithic sites w i l l have a very high micxoblade/microcore ratio. HYPOTHESIS 4: Micropore technology i s associated with the conservation of  l i t h i c raw material. If the shortage of l i t h i c raw material i s a major design of microcore technology, microlithic sites should contain smaller unstandardized cores, flake tools more aggressively u t i l i z e d prior to discard, and the use of other conservative technologies, such as b i f a c i a l core technology. In addition, when the majority of debitage derives from later reduction stages, then the resharpening of tools to extend their use l i f e i s indicated (Bamforth 1986). Finally, i f microcore technology i s a method for conserving specific l i t h i c materials, then microlithic artifacts w i l l tend to manufactured from those types. Microlithic and non-microlithic assemblages w i l l be compared i n order to determine Whether or microcore technology i s a strategy for conserving stone. Test Implications; 1. Microlithic sites w i l l have a proportionally greater amount of late stage debitage from resharpening than non-microlithic sites. 2. Unstandardized cores i n microlithic sites w i l l be smaller than unstandardized cores i n non-microlithic sites. 3. Microlithic sites w i l l contain a proportionally greater number of flake tools which display evidence of multiple uses than non-microlithic sites. 4. Microlithic sites w i l l contain a proportionally greater number of b i faces, biface fragments and b i f a c i a l thinning flakes than non-47 microlithic sites. 5. Microcores and microblades will be manufactured from raw materials which are more restricted in distribution in the study areas. The goal of the research to be presented in the following chapters is to describe and explain the sources of assemblage variability between microlithic and non-microlithic sites. The following chapter describes the geographical characteristics of the region and places the study areas in archaeological and ethnographic context. 48 CHAPTER III THE STUDY AREAS Biophysical Environment Geographic Setting Upper Hat Creek Valley is situated east of the Fraser River on the transition zone between the Fraser and Thompson Plateaus (Figure 1) (Holland 1964). Oriented in a north-south direction, Upper Hat Creek Valley is approximately 24 kilometres long. The western boundary consists of the rounded peaks of the Clear Range, rising to a maximum of 2280 metres. The steep western slopes of the Clear Range drop down to the Fraser River, while the eastern slopes incline gradually down to the Upper Hat Creek Valley floor. High points in the Clear Range are Blustry Mountain at 2315 metres, and Cairn Peak at 2318 metres. The eastern margin of Upper Hat Creek Valley includes the western slopes of the Cornwall and Trachyte Hills, which mark the western margin of the Thompson Plateau. These hills are relatively low, up to approximately 2000 metres above sea level, and of more moderate relief. The valley floor slopes from an elevation of 840 metres in the north to approximately 1200 metres in the south. Highland Valley is situated in the Thompson Plateau (Figure 1) (Holland 1964). The valley trends northwest-southeast and is approximately 22 kilometres long. In contrast to Upper Hat Creek Valley, Highland Valley is narrower, with steeper valley walls. The Pimainus Hills mark the southern boundary at an elevation of 1800 metres. The highest peak in the mountain range to the north is South Forge Mountain at 1890 metres. Valley walls, Figure 1. Location of study areas: I. Upper Hat Creek Valley; II. Highland Valley. 50 which reach a height of approximately 1900 metres, slope quite sharply to the floor which is a gently rolling upland of low relief between 1200 and 1500 metres above sea level. Paleoenvironmental Setting This section presents the most recent evidence relating to the final major glaciation and subsequent deglaciation in southern British Columbia. In addition, recent syntheses of Late Pleistocene and Holocene climatic and vegetational sequences are presented to indicate how the resource base in the study areas may have been affected. Fraser Glaciation and Subsequent Deglaciation The final glacial period, the Fraser Glaciation, began approximately 25,000 B.P. as ice tongues proceeding east from the Coast Mountains coalesced into the Cordilleran ice sheet (Clague 1981; Armstrong et al. 1965). Radiocarbon dates from localities in the Fraser valley lowlands and southeastern British Columbia indicate that much of southern British Columbia remained ice-free until after about 19,000 to 20,000 B.P. (Clague et al. 1980; Fulton and Smith 1978). Subsequent buildup in area and depth of ice was probably very rapid, until the glacial climax slightly after 15,000 B.P. (Clague 1985). Ice covered a l l land surfaces east of the Fraser Valley, except perhaps the highest peaks of the Clear Range (Ryder 1976). There is no detailed regional scheme of deglaciation, although i t is thought to have started in the Columbia Mountains and proceeded in a northwesterly direction (Fulton 1975). Retreat first occurred along the southern, eastern and western margins of the Cordilleran Ice Sheet, 51 approximately 13,000 B.P., and proceeded rapidly. The ice sheet thinned and receded through downwasting, with uplands appearing f i r s t through the ice cover and dividing the ice sheet into a series of valley tongues i n response to local conditions (Clague 1981). The oldest reli a b l e postglacial radiocarbon date i n the Interior Plateau i s 11,000+180 B.P., from the Arrow Lakes (Clague 1981). Bog-bottom radiocarbon dates indicate that, by 9500 to 10,000 B.P., the plateaus and valleys of the Interior Plateau were completely deglaciated, and a l l g l a c i a l lakes were drained (Clague 1981; Fulton 1969). During and immediately following deglaciation, and prior to the establishment of vegetation, rapid aggradation occurred i n river valleys and on lowlands (Clague 1981; Ryder 1971a, 1971b). Throughout the interior, freshly deglaciated d r i f t was eroded from the uplands and valley walls, and re-deposited at lower elevations i n fans, deltas, and floodplains (Church and Ryder 1972). Probably the majority of postglacial deposits, except i n lake basins and at the mouths of major rivers, were l a i d down rapidly within a few hundred years of deglaciation (Clague 1981). In addition, unvegetated surfaces were subject to eolian activity, and substantial amounts of loess were deposited i n protected areas (Clague 1981). As vegetation became established and slopes stabilized, sediment supply to streams and rivers decreased. The number of existing landforms completed 6000 years ago i n the southern Interior Plateau indicate that the rates of erosion and deposition have since been very small (Church and Ryder 1972). Modern drainage patterns were established i n the study areas during deglaciation. Hebda (1983) estimated that Upper Hat Creek valley was ice-free at approximately 13,000 B.P., although this date may be too old due to 52 contamination by "old carbon" from local coal deposits. During the final deglaciation/ the ice front retreated rapidly from Upper Hat Creek Valley, and stagnant ice deposits were rare. Meltwater may have originally drained south and east through Oregon Jack Creek or northward over the ice (Aylsworth 1975). In addition, Upper Hat Creek Valley has been subject to periodic landslides, in the form of slumps, debris flows and earthflows (Clague 1981). Hebda (1983) estimated that Highland Valley was probably ice-free at approximately the same time, or perhaps a thousand years later, as Upper Hat Creek Valley because both valleys are situated at a similar elevation. During the late phase of deglaciation, a tongue of ice apparently extended into the valley from the Thompson River Valley. As this ice retreated, ice-contact glaciofluvial deposits and at least one block of stagnating ice were left behind (Hebda 1983). A brief examination of the basal sediments of drained McNaughton Lake provided a series of radiocarbon dates, with the earliest date of 9600+70 B.P. indicating that the lake was probably established shortly after deglaciation of the valley, between 10,000 and 11,000 B.P. (Clague 1988). However, determination of a subsequent chronology of sediment deposition within the lake may be impossible due to the possibility of at least some of the radiocarbon dates being affected by the ingestion of old carbon by living organisms (Clague 1988). In addition, some of the dated material is wood transported from elsewhere in the valley and which is very likely older that the enclosing lake sediments (Clague 1988). An analysis of sedimentary sequences from Big Divide Lake indicated that this lake was probably inundated during the entire postglacial period, and that this area of Highland Valley may have been ice-free for over 12,000 53 years (Hebda 1983). Since deglaciation/ the physical appearance of Highland valley has probably changed very little, with the exception of alluvial deposits from streams and some mass-wasting along slopes (Hebda 1983). Holocene Climatic and Vegetational Sequence Recent syntheses of Holocene climatic and vegetational changes in southern British Columbia are based primarily on data from the coast, Okanagan Valley, Lillooet area, Marion Lake, Yale and the lower Fraser Canyon (Alley 1976; Mathewes and Heusser 1981; Mathewes and Rouse 1975; King 1980; Mathews 1979; Desloges and Ryder 1990; Ryder and Thomson 1986). These are supplemented by data from paleoecological studies of cores from lakes in both Upper Hat Creek Valley and Highland Valley (Hebda 1979; Clague 1988). During late deglaciation, approximately 13,000 to 12,000 B.P., the climate was probably cool and continental, with pioneering aboreal species (Clague 1981; Mathewes 1973, 1985). Hebda (1982, 1983) interpreted upland vegetation during this time period as pioneering grassland, with grasses, sedges (Cyperaceae), and sagebrush (Artemesia). Following this initial cold period, there was a rapid transition to a cooler, moist climate, which as ted from approximately 12,000 to 10,500 B.P. with a corresponding increase in mountain hemlock, balsam, spruce and western hemlock. The bulk of data relating to this time period derives from the coast, but investigation of several interior sites also indicates a corresponding cooler, moister climate there (Hebda 1982). Pollen diagrams from Finney Lake in Upper Hat Creek Valley show abundant aspen (Populus  tremuloides) and lodgepole pine (Pinus contorta) or western white pine 54 (Pinus monticttla). Hebda (1982) proposed that aspen may have formed a parkland with grasses and sage, or even closed forest stands in moister areas, and that pine forests probably occupied the upper slopes. Mathewes (1985) noted that there was a dramatic change in pollen diagrams from the period between 10,500 to 10,000 B.P., corresponding to a rise in summer temperatures associated with a decline in precipitation. The climate in most areas of southern British Columbia became as warm as, or warmer than, the modern climate. The dating of this xerothermic interval varies slightly between regions. Evidence collected from Marion Lake, indicating maximum temperatures and minimi mi precipitation between about 10,000 and 7500 B.P. (Mathewes and Heusser 1981) is supported by data from lakes in the Lillooet area, Fraser Canyon area, and Yale area (King 1980; Mathewes and Rouse 1975; Mathewes and Heusser 1981). Pollen samples taken from Kelowna Bog indicate a xerothermic interval in that area between approximately 8400 and 6600 B.P. (Alley 1976). Basing his interpretation on pollen diagrams from Finney Lake, in Upper Hat Creek Valley, and other interior sites, Hebda (1982) limited the dating of the xerothermic interval to between 10,000 and 8000 B.P. In addition, Hebda (1982, 1983) suggested that much of the southern Interior Plateau, below 1300 metres, was covered by sage-grasslands, while at higher elevations or in north facing, moister areas, forests were dominant. The following period, from 8000 to 4500 B.P. was moister, although s t i l l warmer than the present, and may have been characterized by forest expansion into intermediate and unstable grasslands (Hebda 1982, 1983). Hebda (1983) interpreted the vegetation as almost continuous grassland cover at lower to 55 mid-elevations in the southern Interior Plateau, with a transition to forest approximately 100 metres above the modern transition zone. However, this emphasis on such widespread grassland is derived from a limited data base, and is not widely supported in current interpretations (June Ryder, personal communication 1990). Lakes within the study areas may have been situated within grassland and perhaps rose to nearly modern levels, while development of poorly drained wetlands in these upland valleys during the latter half of the inteval is also implied (Hebda 1982, 1983). After approximately 6000 B.P., the climate became cooler and moister throughout southern British Columbia, with subsequent fluctuations corresponding to the three Neoglacial advances (Clague 1981; Alley 1976; Mathewes 1985). Three minor readvances occurred between approximately 6000 and 5000 B.P., between 3500 and 2000 B.P., and after 900 B.P., with a glacial maximum during the last few centuries (Desloges and Ryder 1990; Ryder and Thomson 1986). Although Hebda (1982) suggested that modem climatic conditions appeared approximately 4500 B.P. in the southern interior, Mathewes (1985) proposed that Hebda's data from Upper Hat Creek valley indicates a shift to moister conditions approximately 8000 B.P., with a second moisture increase about 4500 B.P. King's (1980) detailed work at two lakes near Lillooet as well as Hebda's work in Upper Hat Creek Valley demonstrate the marked variability in the timing of climatic changes in the British Columbia. Mathewes (1985) concluded that this mid-Holocene period can be viewed as transitional, both climatically and vegetatively, and clarification of intra-regional variation depends on further analysis of well-dated pollen and fossil sequences. 56 Hebda (1983) suggested that from 4500 B.P. to the present, the climate was perhaps initially cooler and moister, but gradually established a pattern similar to the modern dim te. From 4500 to 3000 B.P., the grassland areas shrunk, and forest areas descended to the valley floor. Between 3000 and 2000 B.P., extant vegetation boundaries and types became established, and lakes became established at modern levels. In addition, data from fossil pollen assemblages in the Okanagan Valley suggest that, during this latter phase, three moister intervals, tentatively correlated with the dated Neoglacial advances, were characterized by increases in birch (Betula), alder (Alnus) and hazel (Corylus) (Alley 1976). Although a tentative climatic sequence has been established for the study area, there has not been a study of the effects of climatic change on landform development or terrestrial sedimentation in the southern Interior Plateau. The probable response to glacial advance was an increase in fluvial discharge but other responses were also possible (J. Ryder, personal coanunication 1990). Any changes in the intensity and frequency of storms, and in moisture and temperature patterns may have also affected the intensity of surface erosion and resultant deposition of sediment (Ryder 1982). In addition, little is known about the paleohydrology of the Fraser-Thompson basin following early fluvial degradation, while the magnitude of change in the fluvial regime and its effects on salmon-carrying capacity are unknown. Finally, in the semi-arid areas of the southern Interior Plateau, a minor decrease in precipitation or an increase in temperaturewould have produced marked changes in vegetation (Clague 1981). These changes would most likely have manifested themselves as altitudinal movements in vegetation boundaries, with grassland communities occurring at higher 57 elevations during warmer, drier periods (Clague 1981). Although there are no data relating directly to the sequence of change in species, and grazing or breeding patterns of fauna in either valley, there must have been significant change accompanying these climatic and vegetation shifts. Hebda (1983:251) suggested that the following may have occurred in the study area: 1. Now-extinct Late Pleistocene megafauna were present shortly after deglaciation, although there is no evidence of these species. 2. Fauna adapted to open arid sage-grassland, and grassland (elk, bighorn sheep, antelope, and perhaps bison) were dominant in the early to mid-Holocene. These were replaced by modern fauna after 4500 B.P.. 3. Around 4500 B.P., more abundant waterfowl appeared in the marshy sloughs of the early to mid-Holocene. Mathews (1979) proposed that salmon would have established themselves in interior streams as soon as the ice had retreated, i.e. by 9500 to 10,000 B.P. throughout the Interior Plateau. In addition, hypothesized changes in vegetation, climate and hydrological systems ca. 5000-4000 B.P. may have reduced ungulate populations in the Canadian Plateau, wh le increasing the size and reliability of salmon runs (Kujit 1988). However, the impact of paleoenvironmental changes on the size and reliability of specific salmon runs is, as yet, unknown. The most recent archaeological data from both the Fraser and Columbia Plateaus indicate that those areas supporting large ungulate populations in the historic period were utilized more intensively after 4500 B.P. (Kujit 1988; Richards and Rousseau 1987). 58 Modern Environmental Setting  Climate The most distinctive climatic feature of this region i s the low annual precipitation due to the plateau's position within the "rainshadow" of the Coast Mountains. The continental climate i s characterized by marked seasonal variation: summers are usually warm and dry/ while winters can be severe (Farley 1979). In addition/ alti t u d i n a l v a r i a b i l i t y i s readily observable/ with higher elevations receiving more precipitation. The local weather patterns of Upper Hat Creek Valley and Highland Valley r e f l e c t this feature/ having cooler wetter summers and longer winters with considerably more snowfall than the nearby major river valleys. In addition/ Upper Hat Creek Valley i s situated i n a zone with a s l i g h t l y higher mean annual precipitation/ and a higher mean daily winter temperature than Highland Valley (Farley 1979). Drainage The major drainage i n Upper Hat Creek Valley i s provided by Hat Creek, which flows for approximately 4 0 kilometres down the valley floor before entering the Bonaparte River/ a tributary of the Thompson River. A series of seasonal and permanent tributaries flow from both east and west valley walls into Hat Creek (Figure 2). Although tributary flows have not yet been studied i n de t a i l , the prehistoric location of these creeks was similar to the modern situation (Pokotylo 1978). Creek levels may vary considerably during the year, with a peak during May and June (Pokotylo 1978). Alternate water sources/ although alkaline/ are provided by a series of small lakes and sloughs on the western slopes south of Ambusten Creek/ and along the Figure 2 . Hat Creek Valley drainage system. 60 western benchlands. Springs are present in the same general area, although the exact distribution is unknown (Pokotylo 1978). In contrast, the major drainage in Highland valley is provided through two lake systems (Figure 3). Prior to mine construction in the early 1970's, Big Divide Lake and Twentyfour Mile Lake drained west through Pukaist Creek into he Thompson River (Areas Associates 1983). Quiltanton Lake and Little Divide Lake drain east through Witches Brook and Guichon Creek to the Nicola River, a tributary of the Thompson River. Other water sources are not known at this time. A large wetland area, immediately downstream from Twentyf our Mile Lake, has been covered by recent mine construction (Brolly 1981). With the exception mentioned above, the drainage patterns have probably not altered within the last 8,900 years (Ryder 1971a). Floral Resources According to the British Columbia biogeoclimatic zone distribution map (Krajina 1965), Upper Hat Creek valley spans four major biogeoclimatic zones: Ponderosa Pine-Bunchgrass, Interior Douglas Fir, Engelmann Spruce-Subalpine Fir and Alpine Tundra. Highland Valley lies entirely within the Interior Douglas Fir zone (Areas Associates 1986). Table 1 provides salient climatic features of each zone. The following brief description of vegetational characteristics (Krajina 1965; Mathewes 1978; Mitchell and Green 1981; Alexander 1989) emphasizes potential food and non-food resources (Turner 1978, 1979). Appendix I provides Latin and common names for floral resources mentioned in the text. 62 Table l . Characteristics of biogeoclimatic zones in the study areas. Ponderosa Pine-Bunchgrass Interior Douglas Fir Engelmann Spruce-Sub-Alpine Fir Alpine Tundra Elevation 275-915 m 300-1525 m 1225-2290 m >1830 m Temperature Mean annual (in Celsius): 6 to 10 4 to 9 1 to 4 -4 to -1 January mean -8 to -3 -12 to -3 -18 to -7 -18 to -1 July mean 18 to 22 17 to 21 12 to 16 7 to 11 Maximum 38 to 44 36 to 43 32 to 37 21 to 28 Minimum -41 to -21 --46 to -32 -56 to -34.5 -45 to -20 No. of frost-free days 100-200 75-200 50-100 <25 Precipitation (in centimetres): Annual total 19 to 36 36 to 56 41 to 183 70 to 280 Annual snowfall 50 to 152 76 to 178 175 to 1016 531 to 1955 snowfall as % of annual total 19 to 42 21 to 35 43 to 72 72 to 74 Driest month 0.7 to 1.5 1.3 to 2.8 1.5 to 6.6 2.3 to 12.2 Wettest month 2.9 to 5.1 5.1 to 8.9 6.4 to 25.4 7.6 to 36.2 Climate: continental: semi-arid to microthermal subhumid microthermal continental subhumid to humid microthermal subalpine continental cold humid alpine tundra (pseudo-arctic) Sources: Krajina 1965; TERA Environmental Resource Analyst Limited 1978 63 1. Ponderosa Pine-Bunchqrass Zone This is the driest and warmest zone in British Columbia/ favouring the development of steppe-like ocmmunities dominated by grasses/ xerophytic herbs and shrubs. Ponderosa pine is the climax tree species. Climax shrubs include Saskatoon berry/ sagebrush/ snowbrush, penstemon, squaw currant and wild rose. Climax herbs include balsamroot/ mariposa l i l y , wild thistle, spring beauty/ toadflax, yellowbell, alum root, long-flowered stoneseed, stoneseed, biscuit root and prickly-pear cactus. On the dry soils of wooded terraces and slopes, shrubs and associated herbs include kinnikinnick, snowbrush, common juniper, Rocky Mountain juniper, penstemon, soapberry, wild onion, toadflax and wild strawberry. Along water courses, vegetation includes western white birch, black hawthorn, silverberry, balsam poplar, trembling aspen, wild rose, willow, rye-grass, Oregon grape, oceanspray and Saskatoon berry. The shoots and stems of the various herbs are ready for collecting between March and May. The most abundant food resources include spring beauty and balsamroot, ready for harvesting in late May to early June. Most berry plants ripen during June, July and early August. 2. Interior Douglas Fir Zone Increased precipitation in this zone favours forest development and the formation of small lakes. The climax tree is Douglas fi r , while Ponderosa pine also grows well here. In the drier pinegrass subzone, shrubs and herbs include Rocky Mountain juniper, balsamroot, chocolate l i l y , stoneseed, wild carrot, Indian celery, penstemon, pipsissewa, rough-fruited fairy bells and tiger l i l y . In the wetter forest subzone, around lakes and along creeks, shrubs and herbs include common juniper, soapberry, dwarf blueberry, 64 grouseberry, Saskatoon berry/ kinnikinnick/ snowbrush/ red twinberry/ Oregon grape, penstemon/ sticky currant/ wild rose, wild onion, toadflax, wild strawberry/ alum-root/ twinflower, black hawthorn, swamp gooseberry, northern black currant, dwarf wild rose, thimbleberry, silverweed, false Solomon's seal, star-flowered Solomon's seal and stinging nettle. In the moister forest subzone, trees include Rocky mountain maple, western white birch, balsam poplar, trembling aspen, western red cedar, subalpine f i r , Engelmann spruce, white spruce and willow. In the grassland subzone, the more important resources include black lichen and lodgepole pine bark, ready for collecting i n late A p r i l and mid-May. Other major shoots and roots (spring beauty and balsamroot) are available between late March and June, while the majority of berries ripen between June and August. In the forest subzone, the more important resources are available between late A p r i l and mid-May. Plants on the valley bottom include C o t t o n w o o d bark, silverweed and cow parsnip, while those on the drier slopes include false Solomon's seal, mariposa l i l y , nodding onion, balsamroot, Indian celery, biscuitroot and blackcap. Berries are especially abundant i n June and early July, and ca t t a i l s and tule are available i n August. In September and October, silverweed and Cottonwood mushrooms are ready for collecting around the lakes. 3. Engelmann Spruce-Subalpine F i r Zone At elevations above 1200 metres, there i s a gradual transition to this forest zone. The climax trees are subalpine f i r and Engelmann spruce. Other trees occurring uncommonly include western white birch, white pine, western red cedar, western hemlock and trembling aspen. Associated climax shrubs are 65 Sitka mountain ash, twin blueberry, oval-leafed blueberry and grouseberry. Climax herbs include bunchberry, twinflower and raspberry. In the drier areas, lodgepole pine and Engelmann spruce occur along with raspberry, mountain bilberry, and grouseberry. In the moister areas, western red cedar is accompanied by swamp gooseberry, thimbleberry, willow, western skunk cabbage, false Solomon's seal, star-flowered Solomon's seal, Labrador tea, mountain valerian, timbergrass, red twinberry and sticky currant. One of the more abundant food plants is spring beauty, ready in early to mid-June. Later in mid-August, the roots of avalanche l i l y , spring beauty and tiger l i l y are s t i l l available, primarily from rodent caches. Although berries are also available at this time, they are neither abundant nor reliable. In September, whitebark pine seeds and black lichen are ready. 4. Alpine Tundra Zone The Alpine Tundra zone occurs at elevations above 1800 metres, on several mountain peaks along the western margin of Upper Hat Creek Valley. Vegetation is dominated by an abundance of low shrubs, grasses and sedge species. In addition, a few species of stunted trees appear, including white bark pine, alpine f i r , lodgepole pine and Engelmannn spruce. Shrubs include crowberries, Labrador tea, red twinberry, willow, Sitka mountain ash, twin blueberry, oval-leafed blueberry and grouseberry. Herbs include spring beauty, wild strawberry, yellow avalanche l i l y and mountain valerian. The earliest floral resource to appear in this zone is spring beauty available in early June. Other major resources include avalanche l i l y , which is ready in July and August, and blueberries, which ripen in August. 66 Faunal Resources Upper Hat Creek valley and Highland valley both lie within the Dry Forest biotic area of mammal distribution (Cowan and Guiget 1965) but local variation is marked. In addition, substantial changes to both the species present and size of populations have occurred within the last 200 years. Large populations of elk, bighorn sheep and mountain goat replaced deer along the lower Thompson River in the late 1700 's, but had disappeared by 1850, when the deer population again increased (Teit 1900). Upper Hat Creek Valley appears to provide a suitable habitat for elk (R.G. Matson, personal communication 1990). In addition, moose did not move into southern British Columbia until after 1920 (Cowan and Guiget 1965). Although mountain goat and bighorn sheep may have inhabited the study areas prior to contact, suitable habitats for these species are limited (Environment Canada 1974) and their populations were presumably small. The modern distribution of mammals in Upper Hat Creek Valley and Highland Valley is probably representative of only the last 4500 years. Environment Canada (1974) identified several areas within Upper Hat Creek Valley with a potentially high capability for sustaining wintering populations of mule deer. Highland Valley has a moderate capability for the support of mule deer and moose, with south-facing slopes providing small areas of high capability (Areas Associates 1983). According to historic sources (Areas Associates 1983), an area near Little Divide Lake provided a wintering range for deer. Appendix I provides Latin and common names of potential mammal resources in both valleys, based on land capability studies and the location of contemporary breeding and migration routes (Banfield 1974; Cowan and Guiget 1965). 67 Contemporary studies of breeding and migration ranges which overlap the study area have identified seventeen freshwater fish species (Carle et al. 1973). Only nine are of potential economic importance as subsistence resources: Dolly varden (Salvelinus malma); Rainbow trout (Salmo gairdneri); Mountain whitef ish (Posopium williamsoni); Cutthroat trout (Salmo clarki); longnose sucker (Catastomus catastomus); northern squawfish (Ptychocheilus  oreqonensis); peamouth chub (Mylocheilus caurinum); burbot or ling (Lota  lota) and perhaps Bridgelip sucker (Catostomus oolumbianus). Hat Creek is the only stream in Upper Hat Creek valley that may have supported an anadromous fish population; other creeks are seasonal or subject to winter freeze-up. The larger lakes do not presently contain fish, and the capability of Upper Hat Creek valley to support anadromous fish species is uncertain (Pokotylo 1978). Spawning populations of pink, coho and Chinook salmon were observed in the Bonaparte River in the early 1970's, before dam construction reduced the size and extent of the run. Data on fish species available in Highland Valley are also ambiguous. Rainbow trout (Salmo gairdneri) are now present in the major lakes as well as in Pukaist Creek and Witches Brook (Areas ssociates 1983). Rainbow and steelhead trout, coho salmon (Oncorhynchus kisutch). and chinook salmon (O.  tshawvtscha), spawn in the lower reaches of Guichon Creek and the Nicola River, while whitef ish, squawf ish, and ling are present in Mamit Lake and Guichon Creek. Brolly (1981) noted that these species may have been introduced by European ranchers, and asserted that native fish species were not available for aboriginal exploitation prior to 1900. However, Teit (1900:348) stated that the Thompson Indians stocked lakes themselves by transporting live trout between lakes, in addition, although the lack of 68 f i s h remains i n archaeological sites may indicate an absence of f i s h (Areas Associates 1986), f i s h remains may have been deliberately removed from prehistoric sites after preparation of the flesh (Albright 1984). Information collected from range maps (Godfrey 1986) and descriptions of habitat characteristics (Guiget 1955, 1958) indicates that approximately 200 species of birds have breeding ranges and/or migration routes which include Upper Hat Creek Valley and Highland Valley. A series of small seasonal ponds along the western slopes of Hat Creek Valley i n the Ponderosa Pine-Bunchgrass zone provides areas of high capability for supporting waterfowl populations (Canada Department of Regional Economic Expansion 1970). There are also similar areas around Kamloops and near Nicola Lake (Areas Associates 1983, 1986). Although sources show the same number of bird species potentially available i n Highland Valley, i n actual fact, there appear to be fewer. The lakes, creeks and wet meadows of Highland Valley support a moderate population of mallards, widgeon, bufflehead, common goldeneye, common loons and scaup, while upland areas support ruffed and blue grouse (Areas Associates 1983). However, birds may have been more abundant i n the past, as a large area of wetlands was eliminated by the construction of a tailings pond (Brolly 1981). Appendix I l i s t s Latin and common names of birds with potential economic value as subsistence resources (Godfrey 1986; Guiget 1955, 1958). The following description of the seasonal a v a i l a b i l i t y of faunal resources i s organized by biogeoclimatic zones (Alexander 1989; Mathewes 1978). 1. Ponderosa Pine-Bunchgrass Zone The majority of resource species are concentrated along water courses and 69 forest margins: snowsnoe hare, beaver, squirrel, bear, deer, moose and elk. Migrating animals, such as deer, bighorn sheep, elk and moose, will be most numerous in the spring and late fall as they move between the river terraces and the higher Engelmann Spruce-Subalpine Fir Zone. Freshwater fish species, particularly trout and whitef ish, are probably available in Hat Creek. Bird species include ruffed grouse, available a l l year, and blue grouse, available during spring and summer. In addition, the wetland subzone in Upper Hat Creek Valley provides a stopping place during spring and fall migration for numerous wetland birds. 2. Interior Douglas Fir Zone Resource species present in this zone include: deer, bighorn sheep, mountain goat, grizzly bear, black bear, yellow-bellied marmot, wolf, coyote, wolverine, weasel, snowsnoe hare, porcupine, red squirrel, northern flying squirrel, cougar, lynx, bobcat, red fox, marten, mink, fisher and short-tailed weasel. During the winter months, ungulates, such as deer, bighorn sheep and elk, and small game mammals, especially snowsnoe hare, are available along the treeline. In late March and early April, ungulates and small mammals are s t i l l present, although in poor condition. Later, in late April and mid-May, deer and other ungulates, along with grouse and small game mammals, are available at higher elevations in this zone. Ungulates and other mammals are most abundant and in prime condition in the fall months, as they migrate down the mountains from this zone to the Ponderosa Pine-Bunchgrass zone in Upper Hat Creek Valley and to the major river terraces. In Highland Valley, the lakes environment provides wetland habitats for various species of birds. In addition, although available evidence is inconclusive, trout may have spawned in the inlet and outlet streams of the 70 lakes in mid-May. 3. Engelmann Spruce-Subalpine Fir Zone The resource species located in the Interior Douglas Fir zone are also present in this zone. In particular, areas of parkland attract deer for summer grazing. While mountain goats and most of the smaller mammals winter in this zone, the following mammals migrate to the lower-lying grasslands in Upper Hat Creek Valley and along the river terraces: deer, bighorn sheep, elk, moose, black bear, grizzly bear, wolf, coyote, wolverine, cougar, bobcat, lynx and marten (Environment Canada 1974). 4. Alpine Tundra Resource species found here include: deer, grizzly bear, black bear, wolf, coyote, wolverine and long-tailed weasel. In addition, small numbers of bighorn sheep, mountain goat and elk may have utilized the tundra in the past. With the exception of mountain goat, these species were probably present only during the summer months, because of the severe winters. Lithic Resources Preliminary studies of lithic resources in Upper Hat Creek Valley have indicated that several types of lithic materials are present. The earliest data on lithic sources comes from Dawson (1894:212b), who noted a limestone conglomerate formation containing "pebbes of chert", ranging from the eastern end of Marble Canyon southward into the Trachyte Hills. Stryd (1973:189-190) mapped a "chert quarry" located near the confluence of Medicine Creek and Hat Creek, but did not describe its characteristics or distribution. During archaeological survey, Pokotylo (1978) collected more detailed data, and noted that secondary deposits of eroded and redeposited 71 nodules of chert and, to a lesser extent, basalt are sporadically distributed i n g l a c i a l d r i f t throughout the survey areas. In particular, dense concentrations of basalt occur along the eastern slopes above Medicine Creek, while chert i s more frequent to the south. The basalt category includes both coarse-grained and vitreous types, while chert and chalcedony are present i n a range of colours, often within the same nodule. The only known source of basalt i n the northern portion of the Clear Range i s i n forested country at the headwaters of Maiden Creek, at the north end of the valley above Marble Canyon (Alexander 1989). According to informants, knowledge of the location was restricted because of the importance of basalt as a trade item and as a raw material for tool manufacture. Informants did not state whether the basalt was obtained during resource procurement trips or whether special excursions were necessary. Hayden et a l . (1987) suggested that access to l i t h i c sources i n Upper Hat Valley was d i f f e r e n t i a l l y restricted for several thousand years by residential groups l i v i n g along the Fraser River, but the evidence for this hypothesis i s sparse. Family groups collecting and hunting i n the uplands may have procured localized l i t h i c resources as an embedded act i v i t y . Information on l i t h i c sources i n Highland Valley i s not as detailed. The a v a i l a b i l i t y of l i t h i c resources i s unknown, but good quality vitreous l i t h i c resources are not present i n the valley (Areas Associates 1983). Some of the less vitreous basalt and quartzite may derive from g l a c i a l deposits and valley stream beds (Areas Associates 1986). Sources for the cryptocrystalline material have been identified near Ashcroft, approximately 30 kilometres away, on the Thompson River (Areas Associates 1986). Cultural Environment 72 Synthesis of Prehistoric Subsistence-Settlement System Regional Cultural-Historical Sequence Sanger (1970a) based the first regional cultural-historical sequence on an assumption that the ethnographic pattern of dependence on stored salmon during the winter months extended back at least 7000 years. However, recent research indicates an apparent shift, approximately 4000 to 4500 B.P., from a highly mobile, primarily foraging system based on mammal hunting to a semi-sedentary, logistically organized collecting system dependent on stored resources, primarily salmon (Fladmark 1982; Richards and Rousseau 1987; Stryd 1971). The following synthesis describes the regional cultural-historical sequence, initially proposed by Fladmark (1982) and expanded by Richards and Rousseau (1987) to incorporate evidence of local variability. 1. Early Prehistoric Period (pre-8000 B.P.) The only radiocarbon-dated site from this period in the southern Interior Plateau is Gore Creek near Kamloops. The site contains the post-cranial remains of an adult male dated at approximately 8250 B.P, and lacks associated artifacts or features. Results of stable carbon is tope analysis of the bone collagen indicate that the protein source of this individual was predaninantly terrestrial (Chisholm 1987). Other evidence of Early Prehistoric occupation is limited to undated surface finds of projectile points which are typologically similar to early points in other areas (Stryd and Rousseau 1988). The data are presently too sparse to reconstruct the Early Prehistoric subsistence-settlement system. 73 2. Middle Prehistoric Period (8000 - ca. 4000/3500 B.P.) The Middle Prehistoric Period i s well-represented at a series of excavated sites along the Thompson and Fraser River valleys and i n Highland Valley. The latter portion of the Middle Prehistoric has been subdivided into the Lehman Phase (6000/5500 - 4400 B.P.) and the Lochnore Phase (5500 -4000/3500 B.P.) (Areas Associates 1986; Stryd and Rousseau 1988). Preliminary analyses indicated that the differences between these two phases relate primarily to the l i t h i c technological industry (Areas Associates 1986); whether the ultimate cause i s functional, chronological or ethnic, as Stryd and Rousseau (1988) have proposed, i s s t i l l unknown. Archaeological data from test excavations indicated that the occupations were short-term, with no evidence for semi-permanent dwellings or food storage. Richards and Rousseau (1988) speculated that the absence of pithouses and storage p i t s constitutes evidence for a foraging strategy during the Lehman Phase. On the other hand, Hayden et a l . (1987) have suggested that a pithouse excavated at Keatley Creek may date to the Lehman Phase, implying that the basic social organization of the winter village was already established during the Middle Prehistoric period. As well, winter settlements may have consisted of either brush houses or mat lodges, as their use i s documented, although infrequently, during the ethnohistoric period (Teit 1909). As well, food storage was not limited to the locus of the winter village. Teit (1900, 1909), and Dawson (1891) noted that storage caches consisted of p i t s or above-ground platforms, often located at the processing s i t e or some distance from the winter village. In addition, the Mount Currie L i l l o o e t and Chase Shuswap used elevated box caches for long-term storage at fishing stations, and underground caches for foods which 74 were susceptible to frost damage, consumed later in the year, or reserved for a food shortage. Therefore, the absence of semi-subterranean houses and storage pits does not, in itself, constitute clear evidence for a foraging strategy. The majority of the stone tools were manufactured from fine-grained basalt (Stryd and Rousseau 1988). The subsistence-settlement pattern is characterized by a reliance on portable dwellings, and a fairly generalized subsistence economy, with salmon slowly beconiing more important during the latter half of this period (Fladmark 1982). Analyses of faunal material from one site demonstrated that the major food resources were small mammals, ungulates and fresh-water shellfish (Areas Associates 1983, 1986; Richards and Rousseau 1988). However, stable carbon isotope analysis of human burials dated to the latter part of Middle Prehistoric indicated that, while terrestrial mammals provided the bulk of dietary protein, consumption of marine resources, i.e. salmon, increased to approximately 40% (Chisolm 1987). Although the results from these two sets of analyses appear to be contradictory, they may be indicative of local variability in the consumption of salmon. Salmon utilization declined away from the major river valleys (Chisholm 1987). As well, taphonomic processes may have resulted in the differential preservation of terrestrial and marine bone, or the under-or over-representation of subsistence species. The majority of Middle Prehistoric sites are located within the major river valleys. This characteristic may be due to more intensive surveys within these areas, or a lack of chronologically diagnostic artifacts in upland sites. Alternately, utilization of upland areas may not have been widespread until intensive storage practices associated with population 75 expansion necessitated the more extensive exploitation of additional resource zones (Greaves 1986). The prevailing view of the chronological significance of microcore technology places i t within this period (Fladmark 1986; Hayden et al 1987). Drynoch is the oldest radio-carbon dated Middle Period site in the interior of British Columbia. It was occupied approximately 7500 years ago, and contains one microblade. Lehman is one of the better known Middle Period sites, and contains a large collection of microblades and cores. Unfortunately/ the stratigraphy is complex and the dating has been questioned by several researchers (Pokotylo 1978; Fladmark 1986). According to Richards and Rousseau (1988:40), the Lehman Phase lacks microcore technology, based on excavations at Oregon Jack Creek and Rattlesnake Hill sites in the Thompson River Valley, and the Lehman site in the Fraser River Valley. This lack of agreement with previously published reports is puzzling, as Sanger (1966) states that in Zone II at the Tifthman site, with a single date of about 6600 B.P., microblades constitute nearly 50% of the assemblage. Securely-dated single component sites from this period are rare and microblades are commonly found in association with large projectile points, assumed to be at least 4500 years old. However, a pre-pithouse component, associated with Lehman Phase projectile points, containing microblades has been located recently at Keatley Creek (Hayden et al. 1987) 3. Late Prehistoric Period (ca. 4000/ 500 - 200 B.P.) The ethnographic pattern of semi-sedentism and reliance upon winter storage of salmon and other resources appears to derive from the beginning of the Late Prehistoric period. This period is characterized by the extensive occupation of winter pithouses, associated with storage pits, use 76 of a wide range of l i t h i c resources, and a l o g i s t i c a l subsistence system focused on salmon. Nonetheless, preliminary archaeological data indicate that variation i n subsistence-settlement practices occurred within the Late period. Pokotylo and Froese (1983) suggested that intensive root processing began approximately 2250 B.P. and became less common after 1200 B.P. As well, the results of Langemann's (1987) study of faunal material from the L i l l o o e t area indicated that salmon use increased significantly from approximately 1800 to 1200 B.P., and then decreased. Other studies (Fladmark 1982; Hayden et a l . 1986; Richards and Rousseau 1987) indicated that there may have been a population increase between 1800 and 1200 B.P. i n the mid-Fraser region. There i s also a corresponding increase i n the number of Late Prehistoric sites i n upland areas and along the Thompson River Valley (Pokotylo and Froese 1983; Areas Associates 1983, 1986; Richards and Rousseau 1987; Alexander and Matson 1987). L i t h i c technology i n the Late Prehistoric also appears to d i f f e r considerably from that of the Middle Prehistoric (Areas Associates 1986; Stryd and Rousseau 1988). A wider range of l i t h i c raw materials was u t i l i z e d , with assemblages composed of up to 20% exotic materials. Richards and Rousseau (1987) proposed that the Late Prehistoric period should be subdivided into three cultural horizons, based on changes i n projectile point morphology, bone and antler technology, architectural features, burial customs, and regional exchange networks. The Shuswap Horizon (4000/3500 B.P. - 2400 B.P.) i s characterized by large housepits, lanceolate or triangular spear and a t l a t l points, relatively simple l i t h i c technology based on the use of locally available raw materials, a well-developed, primarily u t i l i t a r i a n , bone and antler industry, and burials 77 within habitations. Although the relative importance of individual species is not yet known, subsistence activities were apparently focused on the hunting of large and small terrestrial mammals and birds, the collecting of fresh water mussels, and the fishing of salmon and other fresh water species. There is no direct archaeological evidence for plant collecting although i t undoubtedly occurred. Shuswap components have not yet been indisputably identified in upland localities, and subsistence activities may have been focused on the exploitation of resources in major river valleys, close to main seasonal base amps (Richards and Rousseau 1987). Although Hayden et al. (1987:22) suggested that sometime during the Shuswap or the following Plateau Horizon, large socioeconomic coresidential corporate groups established ownership of and controlled access to edible and lithic resources, there is, as yet, very little archaeological evidence to support this hypothesis. The Plateau Horizon (2400 - 1200 B.P.) is characterized by smaller housepits, bilaterally barbed, corner-notched or basally-notched atlatl and arrow points, an increase in the quality of chipped stone tool technology with a reliance on high quality, often exotic, raw material, an elaboration of the bone and antler industry, an increase in trade with the coast, and frequent cremation (Richards and Rousseau 1987). The most significant change in subsistence practices is a focus on the intensive exploitation of upland root resources (Pokotylo and Froese 1983). The Kamloops Horizon (1200 - 200 B.P.) is characterized by housepits which are highly variable in size, Kamloops side-notched and multi-notched arrow points, an increase in the quality and importance of both ground stone tool technology and bone and antler technology, flexed burials in shallow 78 pits, occasionally acxxinpanied by ornamental objects, and an increase in inter-regional trade. Subsistence focused on salmon and a continued reliance on upland root resources. Until recently, microcore technology was considered to be absent from Late Period sites. However, the majority of identified sites are located in major river valleys, representing only a portion of the subsistence-settlement pattern. In addition, archaeologists have frequently considered microlithic artifacts associated with housepits and Late period projectile points to be indicative of a mixed assemblage (Donahue 1975; Sanger 1967; Fladmark 1976). Until further research clarifies the chronological extent of microcore technology, its presence in Late Period sites should not be interpreted a priori as intrusive. Local Prehistoric Subsistence-Settlement Pattern 1. Upper Hat Creek Valley A series of four surveys located 223 sites on the valley floor, lower slopes and one headwaters area, interpreted as indicative of substantial occupation throughout the last 7000 years (Pokotylo 1978; Beirne and Pokotylo 1979). However, Pokotylo's tentative estimate of site antiquity was based entirely on comparative artifact morphology. In particular, microcore technology was inferred to be diagnostic of only Middle Prehistoric sites. As well, a detailed comparison of projectile points with well-dated points from other contexts has not yet been done. Although recent surveys in the Cornwall Hills and along the upper reaches of Oregon Jack have located sites containing evidence of microcore technology, these sites have not yet been dated by independent means (Rousseau and Gargett 1987; Rousseau 1989). 79 Table 2 l i s t s those sites which are dated by radiocarbon or by projectile point typology. A l l are attributed to the Late Prehistoric period. This interpretation f i t s with the paleoenvironmental data, suggesting that Upper Hat Creek Valley was not significantly exploited u n t i l the modern environment was well established. On the other hand, we may be unable to accurately identify sites dating to the Middle or Early Prehistoric periods because of s i t e v i s i b i l i t y or a lack of chronologically diagnostic ar t i f a c t s . Table 2. Cultural-historical a f f i l i a t i o n of Upper Hat Creek Valley sites. Middle Prehistoric Late Prehistoric Lehman Lochnore Shuswap Plateau EeRj71* EeRj46* EeRj55a* EeRjlOl* EeRJ93* EeRk42* EeRk43* EeRj20 EeRJc52 EeRjlOO EGRJ42 *dated by radiocarbon Kamloops EeRj64 EeRj8 EeRj55d* EeRjl* EeRj53* Pokotylo's (1978) detailed analysis of forty-four sites confirmed his hypothesis that Upper Hat Creek Valley was an important part of the regional subsistence-settlement system. Relative to the intensity of occupation at f a l l salmon procurement sites and winter housepit sites, a l l sites i n Upper Hat Creek Valley appear to represent short term occupations. Two s i t e classifications, one based on a manufacturing stage typology of debitage and one based on a technological tool typology, were produced by cluster 80 analysis and multidimensional scaling of 44 surface assemblages. Each assemblage was assigned to one of five debitage groups and one of f i v e tool groups. Pokotylo(1978) concluded that, although membership of the f i v e debitage groups differed from that of the five tool groups, s i t e occupation and use patterns tended to coincide. The following s i t e typology was constructed u t i l i z i n g a combination of assemblage and environmental characteristics (Pokotylo 1978:327-329): (1) hunting camps (limited a c t i v i t y sites) where hunting and butchering tasks) were carried out: optimal overview; debitage derived from late stage tool finishing and possible tool maintenance; tools consisting of fragments of projectile points and bifaces representing hunting and butchering a c t i v i t i e s . (2) hunting camps (local base camps for extractive activities) where various other maintenance and extractive tasks were carried out: optimal overview; debitage derived from the entire range of reduction steps, resulting from the manufacture and rejuvenation of hunting and butchering tools; tools consisting of fragments of projectile points and bifaces. (3) small plant-gathering/processing sites (limited a c t i v i t y s i t e s ) : low frequency and diversity of expediently made tools; cultural depressions interpreted as root roasting ovens; small surface area. (4) plant-gathering/processing sites (local base camps for extractive activities) where other maintenance and extractive tasks were also carried out: debitage derived from a wide range of manufacturing step; high frequency and variety of tool types; cultural depressions interpreted as root roasting ovens; large surface area. Although direct evidence of si t e seasonality i s absent, Pokotylo used 81 modern environmental data to suggest that the valley was occupied during the late spring, summer and f a l l , for the procurement and processing of roots for storage, and the hunting and butchering of ungulates. A later study (Pokotylo and Froese 1983) of cultural depressions inferred to be cooking p i t s provided information on the nature of prehistoric root processing a c t i v i t i e s i n Upper Hat Creek valley. The morphology of prehistoric cooking p i t s appears to be similar to those described for the ethnohistoric period, except for two notable differences: apparent reuse and a much greater basin size i n some p i t s . Three types of base processing camps were distinguished: (l) sites with only cultural depressions; (2) sites with small cultural depressions, interpreted as single-use cooking p i t s , associated with l i t h i c scatters reflecting extractive a c t i v i t i e s ; and (3) sites with larger cultural depressions, interpreted as multiple-use cooking p i t s , associated with l i t h i c scatters produced by maintenance a c t i v i t i e s . Radiocarbon dates suggest a correlation between p i t size, incidence of reuse and chronology. The largest p i t s with multiple basins date between 2250 and 1000 B.P., a period previously hypothesized as being characterized by population growth, and intensification of salmon exploitation and storage, i n major winter village sites i n the Fraser River valley (Stryd 1971, 1974). The apparent contemporaneity of the intensified collection and storage of both root resources and salmon indicate that prehistoric subsistence-settlement patterns i n Upper Hat Creek valley may have been substantially different from the ethnohistoric period (Pokotylo and Froese 1983). 82 2. Hicrhland Valley Two surveys located 54 sites in the Lake Zone area, immediately adjacent to the valley-bottom lakes (Brolly 1981; Areas Associates 1983, 1986). Table 3 lists those sites dated by radiocarbon or projectile point morphology. Again, absence of Early Prehistoric period sites may be attributed to lack of site visibility. However, agreement of these data with those from Upper Hat Creek Valley suggest that small population size made the early regular exploitation of upland areas unnecessary. Table 3. Cultural-historical affiliation of Highland Valley sites. Middle Prehistoric Late Prehistoric Lehman Lochnore Shuswap Plateau Kamloops EdRg2* EcRglB* EcRg4B EcRglB* EcRgiB* EcRg4C EdRgiB* EdRglC EcRg2BB EcRg2-II EcRg4H EcRg4F EcRglA-II EcRglO EdRgS EdRg2-III* EdRgS EdRg2 EcRg3-I EcRg6 *dated by radiocarbon Highland Valley was apparently intermittently occupied throughout the last 6000 years for the hunting and butchering of ungulates and waterfowl, and for fishing (Areas Associates 1986). However, restriction of the survey to the lake-shore area, as well as excavation subject to time and budgetary constraints, may mean that additional sites remain to be discovered. Therefore, the extent of prehistoric use of the valley may well be significantly under-estimated. Arnoud Stryd (personal communication 1990) commented that the resource base in Highland Valley during the Middle Prehistoric was probably more 83 diverse than in previous times, with reference specifically to ungulate populations. The lakeshore location of Middle period Lehman Phase sites indicates a reliance on upland lake fishing, although waterfowl and aquatic mammals may also have been hunted (Stryd and Rousseau 1988). Lithic and faunal remains at the largest site, EdRg2, are indicative of food processing and camp maintenance activities (Areas Associates 1986). Investigators applied an intuitive site typology, adapted from Binford (1980) which incorporates the following site types (Areas Associates 1983:26-27): (1) residential camp: a base settlement from which work parties forage or leave on longer procurement trips; also the location where most processing, manufacturing and maintenance activities occurred; will contain archaeological evidence for food preparation and consumption, as well as tool production and/or rejuvenation. (2) field camp: a temporary operational centre where a task group sleeps, eats and maintains itself while away from the residential camp; will consist of a variety of types (fishing, hunting, etc.). (3) station: combines Binford's (1980) location where extractive tasks are carried out, station where special-purpose groups gather to collect information, and cache where resources are stored. Interpretation of site assemblages was based on the following assumptions: archaeological assemblages from residential camps will contain a greater quantity and variety of debris because of the larger number and variety of activities performed there; and archeological assemblages from field camps and stations will be smaller, more homogeneous, more task specific, and seldom multi-component (except large fishing stations) (Areas 84 Associates 1983:27). Assignment of each assemblage to a s i t e type was based on the number, variety, and types of a c t i v i t i e s represented; the size and composition of the inferred residential or task group; and the inferred length of occupation. These data were derived from s i t e location; s i t e size; and the quantity, variety and distribution of faunal remains, tools, debitage, charcoal, fire-altered rock, and features (Areas Associates 1983:27). During the Lochnore Phase, several small residential base camps, f i e l d camps, and a station were located i n Highland valley. The largest, EcRglB, was the locus of a series of occupations during which resources obtained by hunting and fowling were processed and l i t h i c tools were manufactured (Areas Associates 1986). As already discussed above, the major difference between the Lehman and Lochnore Phases appears to be technological and primarily related to the morphology of l i t h i c tools. Further research i s required to investigate the cultural significance of this technological v a r i a b i l i t y , particularly because the Lochnore Phase may be directly ancestral to the Plateau Pithouse Tradition (Stryd and Rousseau 1988). The majority of sites located i n Highland valley have been assigned to a new archaeological construct, the Quiltanton Complex, formulated to account for the presence of microlithic assemblages i n Highland Valley, and other upland areas (Areas Associates 1986; Lawhead and Stryd 1985). On the basis of a r t i f a c t cross-dating, geological dating and three radiocarbon dates, the Quiltanton Complex was provisionally dated at approximately 2,100 to 1,000 B.P., but may extend back to 4,500 B.P. or forward into the next millennium (Areas Associates 1986). However, only two sites (EcRg2AA and EdRglB) have been dated by radiocarbon to the Late Prehistoric period. I f the majority of 85 these sites date to the Middle Prehistoric period, then a significant decrease i n u t i l i z a t i o n of the valley i s indicated. Diagnostic t r a i t s consist of a well-developed microlithic industry, use of poor quality basalts and l i t h i c raw material other than basalt, a high frequency of tools used as gravers, and a high incidence of multiple tools. The t h i r t y sites assigned to the Quiltanton Complex represent the most intensive prehistoric use of the Highland Valley. Site types include four residential camps, twelve f i e l d camps, nine l i t h i c reduction and/or tool production stations, and f i v e of unknown function (Areas Associates 1986). Limited faunal remains and residue analysis indicated that subsistence a c t i v i t i e s included the procurement and processing of ungulates and plants (Areas Associates 1986). The Quiltanton Complex sites were probably inhabited by single family groups on a seasonal basis. As i n Upper Hat Creek Valley, there are no sites clearly attributable to the Shuswap Phase. Five of the sites attributed to the Plateau Horizon are characterized by small assemblages, and identified as f i e l d camps, or limited a c t i v i t y sites. The sixth i s interpreted as a residential camp, but several occupations may have added to the size and complexity of the l i t h i c scatter. During the Plateau Horizon, Highland Valley was probably occupied by smal hunting parties from one of the major river valleys (Areas Associates 1983, 1986). Prehistoric u t i l i z a t i o n of Highland Valley continued into the Kamloops Horizon. The six sites are small and identified as hunting camps and f i e l d camps. Again, the valley was probably used by small hunting parties from the major ri v e r valleys (Areas Associates 1983, 1986). 86 Micropore Technology i n the Southern Interior Plateau  Introduction The following comprises an overview of recent research into the cultural significance of microlithic assemblages i n the southern Interior Plateau. I n i t i a l interpretations focused on their importance as chronological indicators, that i s , as technological markers of the earliest cultures i n the area (Borden 1952; Sanger 1967). Associated with this emphasis was an attempt to trace the technology involved from i t s origins, presumably i n Asia, where microcore technology i s dated several millenia ea r l i e r (Borden 1952; MacNeish 1963). The second major interpretation focused on an association between microcore technology and ethnic group a f f i l i a t i o n , or more specifically, the southern spread of Athapaskan-speaking groups (Sanger 1966, 1967, 1968, 1969, 1970a, 1970b; Donahue 1975; Areas Associates 1983, 1986). The third major interpretation proposed that microcore technology was functionally specific and that microlithic sites were used for a unique task or set of related tasks (Pokotylo 1978; Fladmark 1986a, 1986b; Sanger 1968). Chronological and Geographic Distribution 1. The Lochnore-Nesikep Locality Sanger considered the Plateau Microblade Tradition, tentatively radiocabon-dated between 5,600 and 6,600 B.P. at the Lochnore-Nesikep loc a l i t y , to be at the peak of popularity between 7,000 and 3,500 B.P., i n components interpreted as pre-pit house village occupations. Sanger suggested that, after 3,500 B.P. microblades became increasingly scarce, and were absent from the tool inventory by 2,000 B.P., when house p i t villages appeared. 87 Sanger (1968:114) originally defined the Plateau Microblade Tradition as being characterized by the following: 1. Microblade cores u t i l i z i n g a weathered surface for a striking platform which i s usually modified only at the core edge. Multiple blow striking platform preparation i s scarce, and core rejuvenation tablets are not known. 2. Microblades are usually removed from only one end of the core. 3. Core rotation, resulting i n more than one striking platform, i s very unusual. 4. Fluted surfaces commonly contract to a wedge-shaped keel. 5. The technique of preparing the fluted surfaces i s currently unknown, but the apparent absence of ridge flakes may be very important i n this respect. Later descriptions (Sanger 1970b) did not include the weathered striking platform. In addition, other researchers (cf. Ludowicz 1983) noted that they did not observe this attribute during recent examinations of the Lochnore-Nesikep microcores. Sanger (1967, 1969, 1970a) depended heavily on the presence or absence of microcore technology to define successive stages i n the f i r s t cultural-h i s t o r i c a l model of Interior Plateau prehistory. The data base derived from extensive excavations at seven major sites, mostly house p i t villages, i n the Lochnore-Nesikep l o c a l i t y of the middle Fraser River valley. Interpretation of the stratigraphically complex sites was d i f f i c u l t , and Sanger based his chronological scheme on geological context, stratigraphic position, judgemental estimates of the significance of formal variation i n adjacent assemblages, amount of patination on l i t h i c tools, a r t i f a c t typology and radiocarbon dates selected to agree with other estimates of age. 2. Chronology of Microcore Technology i n Other Areas of the Plateau Other researchers continued to use Sanger's original chronological model to cross-date assemblages from other regions. In several cases. 88 archaeologists ignored Sanger's assertion that microcore technology persists u n t i l 2,000 B.P. i n favour of an interpretation that excluded microcore technology from housepit sites altogether (Donahue 1977, Fladmark 1976), i.e. after 4,000 to 4,500 B.P. In addition, researchers often used Sanger's model to substantiate their own claims of mixed components at sites where microlithic components are associated with housepits or late radiocarbon dates (e.g. Donahue 1975). On the basis of excavations conducted at housepit sites i n the L i l l o o e t area on the Fraser River, stryd (1972, 1973) suggested that microcore technology disappeared i n this area a l i t t l e earlier, approximately 3,000 B.P. However, the Lil l o o e t area deposits were also extensively disturbed. There are forty-one used and retouched microblades i n three sites i n the sample, but the prepared core platforms which would provide conclusive evidence of microcore technology are absent (Stryd 1972, 1973). Fladmark (1982) suggested that an acceptable date for the termination of microcore technology i n the Interior Plateau i s approximately 4,000 to 5,000 B.P., and i s compatible with a similar date for other regions. This interpretation i s supported by recent data from the Keatley Creek si t e , near the Fraser River, where microblades were located i n association with Lehman Phase projectile points i n pre-pithouse deposits (Hayden et a l . 1987). In addition, several sites dated to the last 5,000 years do not contain microlithic a r t i f a c t s : these include Punchaw Lake i n the Central Interior (Fladmark 1976, 1982); T e z l i i n the northern Interior (Donahue 1977); Deer Park Phase sites along the Arrow Lakes (Turnbull 1977) and Rattlesnake H i l l i n the Thompson River valley (Lawhead and Stryd 1985). However, these are either pithouse villages or sites with evidence of long-term occupation. 89 Cn the other hand/ there are a number of sites containing microlithic components with more recent dates and dissimilar context. In some cases, microlithic components appear to be clearly associated with radiocarbon dated p i t house deposits. At NatalXuz Lake i n the Central Interior/ microblades and a microcore are associated with a small hearth/ dated at about 2/000 B.P. (Borden 1952; Donahue 1975). At T e z l i , a microlithic component i s associated with housepits and three radiocarbon dates ranging from approximately 190 to 3850 B.P. (Donahue 1975). At the uikatcho si t e , microblades appear to be associated with a proto-historic p i t house assemblage (Donahue 1973). Microblades at the Punchaw Lake s i t e are apparently dispersed throughout several occupations/ and were i n i t i a l l y associated with radiocarbon dates of approximately 600 and 2500 B.P. (Fladmark 1976; He "liner 1977). However/ after recent re-examination of the assemblage/ Fladmark (1986a) concluded that there i s no microlithic component at Punchaw Lake. At Anahim Lake to the south/ several microlithic assemblages are associated with radiocarbon dates between approximately 1,800 and 120 B.P. from housepit assemblages (Wilmeth 1978). Wilmeth (1978) suggested that the majority of associations are probably the result of post-depositional disturbance, but did accept a date of 1,600 to 2,000 B.P. for microcore technology i n this area. Fladmark (1986a) suggested that a late persistence of microcore technology around the Anahim Lake area, and perhaps as far north as Natalkuz seemed reasonable i n view of the evidence, and noted that i t may be associated with the use of deposits of a high quality raw material i n this area. In addition, radiocarbon dates ranging from 1,250 B.P. to 1,180 B.P. are 90 associated with a "pure" microblade component at the Dantikto s i t e on the nearby Dean River (Wilmeth 1971:2). At two housepit sites near Williams Lake, radiocarbon dates between 1/800 and 1,100 B.P. are associated with several microblades (Whitlam 1976). Finally, at Marron Lake i n the Okanagan valley, Grabert (1974) located a substantial microlithic assemblage dated at 2,500 B.P. i n a housepit at a s i t e interpreted as an upland hunting and stone working station. 3. Microcore Technology i n Upland Areas As discussed above, the presence of microcore technology i n sites along the major ri v e r valleys i s often used, with varying degrees of r e l i a b i l i t y , as a chronological indicator. But the frequent absence of microcore technology at housepit sites does not a p r i o r i rule out i t s presence at Late Period sites i n upland valleys where the organization of resource procurement strategies and the a c t i v i t i e s carried out differed from those i n the major ri v e r valleys. Now there i s evidence that microcore technology may be later, at least i n the study areas. A l l sites, including microlithic sites, located to date i n the two valleys are shallow, and appear to be f a i r l y recently deposited. Twenty-two sites i n the Highland valley contain microblades i n association with l i t h i c scatters and a possible short-term dwelling, at EcRg2AA. This s i t e i s also well-dated by radiocarbon at 1120 to 1900 B.P. (Areas Associates 1983, 1986). In addition, i n Upper Hat Creek valley, at least sixteen sites are known to contain mi c r o l i t i c assemblages. Microblades are associated with Kamloops points i n several l i t h i c scatters and may be associated with a radiocarbon date of approximately 2000 B.P. from a multi-component roasting p i t (David Pokotylo, personal communication 1984). 91 Non-microlithic sites i n these upland valleys may not always be later or earlie r i n time than microlithic sites; they may simply represent functional variation i n a synchronic subsistence-settlement system. At the present time, we cannot assume that microlithic assemblages i n upland area have the same chronological significance as those i n other areas of the plateau, particularly the major ri v e r valleys. Ethnic A f f i l i a t i o n and Microcore Technology Research into the ethnic a f f i l i a t i o n s of microcore technology has been ongoing for several decades. Borden (1952) was the f i r s t to link early nucrolithic-bearing cultures at the Natalkuz Lake s i t e with repeated migrations by Athapaskans. Both Donahue (1975) and Dumond (1969) rejected Borden's hypothesis on lex i c o s t a t i s t i c a l and chronological grounds. Later, Wilmeth (1971, 1978) attempted to identify the archaeological remains of Athapaskan speakers at Anahim Lake, but was unable to clearly isolate a prehistoric Athapaskan component. Following Borden's lead, Donahue's (1975) research at the Ulkatcho and T e z l i sites was an attempt to investigate the prehistory of Athapaskan-speaking people i n the northern interior plateau. Microblades and microcores occur at both sites, but have been mixed with f a l l e n roof material. Donahue (1975) f a i l e d to detect Athapaskan-speakers archaeologically, but continued to defend the hypothesis that the southern migration of Athapaskans was a progressive move that occurred over several millennia i n the northern interior plateau. Although he maintained that the use of microcore technology did not persist past 4,500 B.P., Donahue (1975) continued to imply a relationship between microlithic assemblages and Athapaskans. 92 Further progress i n this direction was made by Matson (Matson 1985; Matson et a l . 1980; Magne and Matson 1980, 1985) at Eagle Lake. Matson developed and applied the pa r a l l e l direct h i s t o r i c approach i n two areas with similar environments but occupied by different ethnic groups. The histor i c Lulua Phase, dated at approximately 1877 A.D., and the prehistoric Eagle Lake Phase exhibit some of the v a r i a b i l i t y predicted i n dwelling shape and projectile point types for Athapaskan sites (Matson 1985). Magne (in Magne and Matson 1985) also suggested that the large number of microlithic sites i n upland areas adjacent to the Thompson River Valley may be attributable to Athapaskan speakers. Stryd and Lawhead (in Areas Associates 1986; Lawhead and Stryd 1985) formulated a new archaeological construct to account for the presence of microlithic assemblages i n Highland Valley, and other upland areas, the Quiltanton Complex, defined earlier. Stryd and Lawhead suggested that the most parsimonious explanation i s ethnic group a f f i l i a t i o n , because s i t e assemblages assigned to the Quiltanton Complex appear to be contemporaneous with, but different from the Kamloops and Highland Phases. That i s , the Quiltanton Complex i s the archaeological manifestation of Athapaska-speakers adapted to a montane-forest-small lake environment provided by Highland Valley, Upper Hat Creek Valley and similar areas near Kelowna and Kamloops. However, sites clearly a f f i l i a t e d with Athapaskan speakers i n other areas do not have microblades (e.g. Wyatt 1972). In addition, there i s no evidence of other Athapaskan t r a i t s , such as those discovered by Matson (Matson et a l 1980) i n the Chilootin, or Workman (1978) i n the Yukon. Furthermore, the nature of the other t r a i t s associated with the Quiltanton complex suggests 93 that factors such as raw material a v a i l a b i l i t y and conservation, and mobility may be significant. Before we accept Stryd's ethnic explanation, we should formulate and test other plausible hypotheses to account for the presence of microcore technology i n upland areas. Subsistence-Settlement Pattern and Microcore Technology Pokotylo (1978) analyzed forty-four surface assemblages from Upper Hat Creek valley, sixteen of which contain microblades. Separate cluster analyses performed on a manufacturing stage typology of debitage and a technological tool typology produced two s i t e classifications. The s i t e c l a s s i f i c a t i o n based on tool types included microlithic assemblages i n three clusters: (1) an uninterpretable microlithic group; (2) a group interpreted as intensive occupations with wide ranging a c t i v i t i e s ; and (3) a group interpreted as brief task specific occupations. The s i t e c l a s s i f i c a t i o n based on debitage types also included microlithic assemblages i n three clusters: (1) a group interpreted as a wide range of manufacturing steps, with emphasis on the early and late stages; (2) a group interpreted as a wide range of manufacturing steps, with emphasis on the middle stages; and (3) a group interpreted as the early stages of manufacturing. The membership of the tool groups differed from that of the debitage groups. None of the microlithic sites exhibited any interpretable patterning with biophysical variables, although this result may relate to a data collection problem rather than random patterning (David Pokotylo, personal orarmmication 1984). Pokotylo (1978) noted that there i s a mutually exclusive distribution of sites with microblades and sites with formed unifaces, and suggested a dichotomy between cutting and scraping tool use. Pokotylo's preliminary 94 results do indicate that microlithic sites i n the Upper Hat Creek Valley may have served an exclusive purpose but additional data and analyses are needed to test this hypothesis. An analysis of the technological attributes of tools and debitage i n Highland Valley sites produced the following intuitive s i t e typology: residential camps; f i e l d camps; tool production stations; l i t h i c reduction stations; undifferentiated stations; drop/discards; and unknown (Areas Associates 1983, 1986). Microblades appear i n a l l s i t e types except undifferentiated stations and drop/discards. This study did not indicate a different purpose for microlithic sites; however, other characteristics common to these sites include the use of poor quality basalt and l i t h i c materials other than basalt, the presence of numerous graving implements, and a higher percentage of "multiple tools", presumably identified by the presence of different types of use-wear. Further analysis may demonstrate that these i n i t i a l differences represent not only variation i n s i t e use but also i n the level of mobility of the s i t e occupants (Greaves 1987). Ludowicz's (1983) research was the f i r s t attempt to quantify and explain the v a r i a b i l i t y between microlithic and non-microlithic assemblages i n a formal model. The research method was a comparison of v a r i a b i l i t y associated with microlithic assemblages from a major riverine area, the Lochnore-Nesikep l o c a l i t y , and an upland area, the Upper Hat Creek Valley. Applying Binford's (1979) model of the technological organization of tools, Ludowicz derived a settlement strategy which was then compared to that of the his t o r i c Interior Salish. Only the tools were available for analysis, and variables were chosen which reflected the relative amount of time involved i n tool manufacture and use, or the intensity of s i t e occupation. Riverine 95 microlithic assemblages are characterized by a higher percentage of formed tools, which represent a greater i n i t i a l expenditure of energy, and upland assemblages contain a higher percentage of expediently-produced, or flake, tools (Ludowicz 1983). In addition, microcore preparation and rejuvenation flakes constitute a larger percentage of upland assemblages, and may indicate that microblade manufacture and microcore rejuvenation were more common a c t i v i t i e s i n upland valleys than i n riverine sites (Ludowicz 1983:158). However, the sample from Lochnore-Nesikep may not have been representative. Finally, Ludowicz (1983) suggested that microcore technology, because i t i s present at sites identified as base camps, may represent features of a collector resource procurement strategy i n place before the inferred s h i f t to a semi-sedentary collecting system. Although the dating of microlithic sites i s not secure, this interpretation seems reasonable, i n view of the probable need for at least limited food storage as a response to the marked seasonal climate, already discussed above. The Use of Microblades The actual use of microblades i n relation to subsistence practices has been primarily the object of informed speculation rather than methodological research. Sanger (1968) suggested that microblades were used as hafted engraving tools. Fladmark (1986) proposed that they were side-slotted as barbs i n projectile shafts, and were used primarily i n hunting a c t i v i t i e s . In addition, Purvis (1971) suggested that microblades were used to cut vegetal materials used i n the manufacture of baskets. Arnoud Stryd (personal cqrimmication 1984) also proposed that microblades may have been used for the procurement and/or processing of waterfowl. Finally, Pokotylo (1978) 96 suggested that ndcroblades served as a cutting tool of some type. However, to date, these hypotheses have not been tested, nor i s there any clear evidence for hafted microblades i n the southern Interior Plateau. However, at least one, and possibly two, end hafted quartz microblades have been located at the Hoko River s i t e i n Washington, i n association with sixteen unhafted quartz microblades but no microcores (Croes 1989, personal communication 1990). Although a systematic study of use-wear patterns on microblades from a regional sample has not yet been published, experimental studies using blade tools (Semenov 1964; Yerkes 1983) indicate the s u i t a b i l i t y of blade tools for the manufacture of bone, antler and shell items. Examination of experimentally-produced microliths used to s l i c e f i s h , and archaeological microliths (not microblades) from the Hoko River s i t e i n Washington revealed no v i s i b l e use-wear patterns on either group (Flenniken 1981). Thus, Flenniken (1981) proposed that the archaeological microliths were hafted individually into wooden handles and used as f i s h knives. Since women now process a l l f i s h and, i n the recent past, collected and processed materials used to create hafts and bindings, Flenniken suggested that, prehistorically, women manufactured and curated these composite knives. This hypothesis i s plausible i n view of the manufacture and curation of hide scraping tools reported by Tahltan women (Albright 1984). In both cases, the tool i s a composite, technologically simple tool, used i n a single location for a special purpose. In order to extend the analogy to the manufacture and use of microblades, we must f i r s t invesigate their function and distribution within the archaeological setting. We should also remember that microcore technology i s a more complex technology and may have been the work 97 of partially-specialized individuals. As well, Pokotylo and Hanks (1989) have recorded the exclusive manufacture of composite, curated "women's" tools by males among the Mountain Dene of the Mackenzie Paver Valley. The major attempt to determine the function or functions of microblades i n the Interior Plateau derived from Stryd and Lawhead's work i n the Highland Valley (Loy 1983, 1986). The f i r s t analysis was conducted on residue present on a single microblade; Loy (1983:258) concluded that " i t derived from scraping wet leather, perhaps i n the f i r s t preparative stages of tanning". My intuitive impression of this result i s that microblades are too small i n working area to be used as a scraping tool (also see Albright (1984) for a description of a hide scraper). Results of the second residue analysis indicated that microblades were used for working both f l o r a l and faunal material (Loy 1986). Loy analyzed twenty-two microblades: six display no residue and are cl a s s i f i e d as unused, while the remaining seventeen display residues identified as either plant or mammal. Of the blades associated with mammals, eight were used as knives, three as bone scrapers, and one as a bone engraver. According to Loy (1986), a l l used microblades analyzed (16) i n the second study were either side- or end-hafted, but a discussion of the evidence used to infer hafting was not provided. However, Ley's published work on residue analysis i n this region lacks sufficient methodological d e t a i l for successful replication by others, and should be cited with caution. Further residue analysis applying the beter documented methods now available (e.g. Gurfinkel and Franklin 1988) would be useful addition to studies of microblade function. In addition, Stryd and Lawhead (Areas Associates 1983) suggested that preliminary examination of wear patterns on some microblades from Highland 98 Valley supported Sanger's (1968) hypothesis that microblades were used primarily for working organic materials such as bone, wood, or antler. Again, there i s no discussion of the wear patterns or their implications for the interpretation of the uses of microblades. Discussion The preceding discussion of recent research into microcore technology the Southern Interior Plateau indicates that the Plateau Microblade Tradition did disappear by 4,000 B.P. i n most areas of the southern Interior Plateau. However, well-dated components, particularly from the upland valleys but also from other areas, signify that microcore technology may have persisted well into the pithouse period of prehistoric occupation. The chronological range and geographic distribution of microcore technology are too great to be identified with any one ethnic group. As well, their occurrence i n large numbers i n upland assemblages and results of residue analysis may indicate that they were not always used as fish-knives (Flenniken 1981; Loy 1986). In the major ri v e r valleys, microblades appear to be either absent from p i t house v i l l a g e sites, or earlier i n time (Hayden et a l . 1987; Sanger 1970; Stryd 73). In upland valleys, microblades appear to be associated with both intensively occupied and limited a c t i v i t y settlements. Pokotylo (1978) has pointed out an apparent dichotomy between the occurrence of microcore technology and tool used as scrapers, as well as the apparent lack of association between microcore technology and biophysical environmental patterning. In Highland Valley at least, microcore technology appears to be associated with archaeological indicators of high residential and/or l o g i s t i c a l mobility, and conservation of raw material: poor quality raw 99 material, and multiple use tools (Greaves 1986, 1987). Researchers working i n the Interior Plateau have proposed an association between microcore technology and the forager resource procurement strategy (Donahue 1977; Ludowicz 1983; Hayden et a l . 1987), a v a i l a b i l i t y of high quality raw material (Fladmark 1985), functional s p e c i f i c i t y (Pokotylo 1978) and ethnic group a f f i l i a t i o n (Donahue 1977; Areas Associates 1983, 1986). Additional explanations for the use of microcore technology include a high level of residential and l o g i s t i c a l mobility (Greaves 1986) and a shortage of high quality raw material (Greaves 1987). However, to date, none of the potential models presented above have been e x p l i c i t l y formulated and tested locally, or at a regional level. In order to explore further the possible organizational role of microcore technology i n upland valleys of the southern Interior Plateau, the following section examines the ethnographic evidence for use of upland valleys. The goal of this section i s to demonstrate a need for, and an opportunity to use microcore technology during the portion of the regional subsistence-settlement pattern located i n upland valleys. Synthesis of Ethnographic Subsistence-Settlement System  Introduction Upper Hat Creek valley was part of the territory of two bands: the Spences Bridge band of the upper Thompson Indians claimed most of the valley while the Bonaparte band of the Shuswap Indians claimed the lower part of Hat Creek, passing through Marble Canyon to Pavilion, at the northern end (Teit 1900, 1909). Teit's (1900:170) statement on the Spences Bridge band i s the most e x p l i c i t description of the use of Upper Hat Creek Valley by 100 aboriginal groups: "Their hunting grounds extend back for t h i r t y or forty miles on each side of the Thompson River, and include the upper half of Hat Creek". The Shuswap "t r i b e " claimed the entire Hat Creek Valley (Dawson 1891:5). In addition to these two groups, Upper Hat Creek Valley may have been u t i l i z e d by groups with adjoining t e r r i t o r i e s : the Idllooet-speaking L i l l o o e t band, the Shuswap-speaking Bonaparte band, and the Thompson-speaking Upper Fraser band (Teit 1900). Information on his t o r i c aboriginal occupation of Highland Valley i s less clear. In a map that demarcates Shuswap territory, Highland Valley f a l l s within Athapaskan-speaking Nicola territory (Teit 1909). However, the Nicola band of the Upper Thompson Indians claimed hunting grounds extending 50 to 70 kilometres back from the Nicola River between Spences Bridge and Nicola Lake; t h i s area included Highland Valley (Teit 1900). In addition, the Spences Bridge band of the Upper Thompson Indians resided along the Thompson River from Spences Bridge to Ashcroft, and used hunting grounds again extending 50 to 70 kilometres on either side of the river, and including Highland Valley (Teit 1900). Informants from present-day villages along the Thompson River above Spences Bridge s t i l l u t i l i z e food resources from the valley on an annual basis, and claim more extensive use i n the past before the disruption caused by mining a c t i v i t i e s (Sylvia Albright, personal cxxnnunication 1989). Finally, Dawson (1891) provided several Shuswap place names for l o c a l i t i e s within Highland Valley, indicating that some Shuswap bands were familiar with this area. Stryd and Lawhead (Areas Associates 1983) suggested that native groups infrequently v i s i t e d Highland Valley, as there i s no archival record of aboriginal a c t i v i t i e s such as hunting, fishing, plant collecting or fowling 101 (Kennedy 1983). But i n 1886, the Tnornpson-speaking Piminos of the Cook's Ferry band petitioned for reserve lands i n the valley, stating that they were already l i v i n g i n the valley during the summer months, while the livestock grazed i n the meadows. After 1889, the Piminos harvested swamp hay i n the summer for horses and cattle, and resided on four reserves i n the valley a l l winter. In addition, elders from present-day native oomrmrnities along the east bank of the Thompson River, north of Spences Bridge, stated that they r e c a l l annual moves to Highland Valley for hunting, berry collecting and ranching; this pattern apparently continues to a limited extent today, despite the disruption caused by mining a c t i v i t i e s (Sylvia Albright, personal communication 1989). Although root resources are evidently not available i n Highland Valley at the present time, variation i n climate and vegetation probably produced a more diverse resource base i n the past (Arnoud Stryd, personal cxsriminication 1990). Taking into account these data, Highland Valley may have been a more integral part of the annual subsistence-settlement system i n the prehistoric past, for both spring and f a l l resource procurement a c t i v i t i e s . Apart from the now-extinct Nicola Indians (Boas 1895), the Thompson and Shuswap Indians were the most probable protohistoric occupants of both Upper Hat Creek Valley and Highland Valley (Figure 4). The following generalized account of subsistence and settlement patterns i n the study area treats the two groups together. In addition, information relating to the neighbouring Salish-speaking L i l l o o e t and the Athapaskan-speaking Chilcotin i s included because these groups exploited a similar biophysical environment i n an analogous manner. 102 Figure 4. Language groups i n study areas. 103 Ethnographic sources provide limited information on the nature of subsistence patterns during the spring and summer, the primary season of occupation of upland areas. Most of the "tribe lived i n the mountains during the greater part of the year, moving about from one root-digging or deer-hunting ground to another, according to the harvest-time of certain roots and berries, or as the deer changed their feeding-grounds during the seasons" (Teit 1900:230). Although large population aggregates are known for unusually r i c h resource areas such as Botanie valley (Teit 1900), the extended family group, or several such groups, was probably the typical work group (Dawson 1891). An examination of early ethnographic sources (Boas 1890; Dawson 1891; Hill-Tout 1899, 1905, 1978; Teit 1900, 1906, 1909), which are based on direct informant observation of subsistence-settlement practices, provides limited information on aboriginal use of upland areas. These descriptions are supplemented by reconstructions of early practices and observations of modern resource collection and processing methods i n the Interior Plateau (Turner 1978, 1979; Kennedy and Bouchard 1978; Alexander 1989; Burnard 1987; Bouchard and Kennedy 1975a, 1975b, 1979). Subsistence-Settlement System i n the Study Areas Several sources (Boas 1890; Dawson 1891; Hill-Tout 1899; Palmer 1975a, 1975b; Teit 1900, 1909) emphasized the dependence of the Thompson and Shuswap people on a diverse resource base. As outlined above, the dominant characteristic of the biophysical environment i n the study area i s habitat zonation or division of the environment into a l t i t u d i n a l l y placed zones defined on the basis of f l o r a l and faunal resources, available at different times of the year. The cultural response to this zonation was the formation 104 of work parties and resident groups of varying sizes and composition, and the use of both l o g i s t i c a l and residential mobility i n order to collect, store and consume resources at the optimum time. This strategy also ensured that i f one resource fa i l e d , other resources could be added as needed or exploited more intensively. In November, hunting parties moved into the Interior Douglas Forest zone to procure deer, the most important faunal resource, as they migrated during the rutting season from the higher mountains to wintering grounds i n the lower h i l l s . Bow and arrows were the principal hunting method, although deer-fences, corrals, surrounds, large nets and snares were also very common (Teit 1900). Hunting grounds were used by families from several villages, although the deer fences were owned and maintained by individuals (Ray 1942). Methods for capturing other mammals included spring-traps, snares, spears, nets, and dogs, wherever there was deep snow i n the mountains, men on snowshoes used dogs to run down deer and elk, which were either shot or clubbed to death. Generally, hunting parties consisted of several men, although larger groups might cooperate for a deer or elk drive. The presence of women at hunting camps i s indicated by the erection of temporary hut for pubescent g i r l s near the hunting lodge (Teit 1900). Hunting lodges of heavy poles covered with sticks and bark spread with fir-branches were constructed i n sheltered valleys, close to good hunting areas. Occasionally, these lodges would be permanent, with earth banked up to one metre high around the perimeter. Often, lodges would be re-occupied, especially i f located near deer fences or topography suitable for deer drives. Another type of hunting lodge was the '•brush-house'', constructed of a ligh t pole framework covered with branches, and used only once i n winter 105 or early spring. In addition to the extractive tasks associated with the procurement and processing of deer meat/ a c t i v i t i e s at hunting camps also included the i n i t i a l steps of dressing skins for clothing/ which would be completed later at the winter camp (Teit 1909). The storage location of f a l l hunting products i s not known; i t i s assumed that a l l skins and meat were transported to the s i t e of the winter village. In December/ the people moved to winter pithouses, clustered i n villages located along sheltered/ south facing terraces near sources of fresh water and wood, i n major river valleys. During this period, stored subsistence resources/ primarily salmon, roots and berries, provided the major food supply. In addition, to supplement the stored food, groups of males continued to hunt deer and elk i n wintering grounds which included Upper Hat Creek Valley (Teit 1900; Alexander 1989) and perhaps Highland Valley. Ice fishing for whitef ish may also have been an important winter a c t i v i t y i n Highland Valley. By late March or A p r i l , families had l e f t the winter houses and were spread throughout the lower elevations i n the uplands, i n the Interior Douglas F i r zone and the Ponderosa Pine-Bunchgrass zone. Fishing i n lakes and rivers at various elevations was the major subsistence activity, supplemented by hunting and plant gathering (Teit 1900). Camps constructed at fishing areas varied i n size from a few to as many as 100 dwellings, each containing a family (Teit 1900). Act i v i t i e s conducted at fishing camps concentrated on the procurement, processing, and consumption of fresh f i s h , although a few plant resources may also have been available at this time. The growing season began i n late A p r i l , and much later at higher elevations. Collection and processing of various plant resources continued 106 through summer u n t i l the f i r s t snowfall (Turner 1978). The f i r s t plants available grew i n the Ponderosa Pine-Bunchgrass zone and were dug as early as late A p r i l . Groups of women using digging sticks and baskets (Teit 1900) procured roots, which were threaded on bark or grass strings and either hung up to dry or baked i n earth ovens for winter storage (Teit 1900). Slightly later, various types of berries became available i n the Ponderosa Pine-Bunchgrass zone and the Engelmann Spruce-Subalpine F i r zone (Turner 1978). In June, the earliest varieties of service berries were ready, and by July, various species were ripening at higher elevations. Finally, during the f a l l months, bog-cranberries, high-bush cranberries, wild rose hips, and crowberries were collected. In addition, black tree lichen was collected throughout the summer months. Berries were eaten fresh, but at least half of the harvest was processed and dried for winter consumption (Turner 1978). Service-berries, soapberries, wild cherries, huckleberries, raspberries, brambleberries, and rose-pips were dried by being thinly spread upon mats exposed to the sun, or baked i n cakes, boiled i n baskets, and f i n a l l y spread onto a layer of fresh pine-needles or gras , to dry i n the sun (Teit 1909, 1906). Cambium was another important plant resource gathered during May and June from upland areas i n the Ponderosa Pine-Bunchgrass zone, the Interior Douglas F i r zone and the Engelmann Spruce-Subalpine F i r zone. Two species were commonly collected by interior groups: lodgepole pine and ponderosa pine (Palmer 1975b). Pine cambium was at i t s prime for harvesting for only a few weeks, from mid-May to mid-June. The bark was removed f i r s t , and the cambium was scraped off i n long strips, with antler or bone scrapers. Pine cambium was usually eaten fresh, but some groups also dried i t for future 107 use (Turner 1978). Alexander (1989) suggested that bird populations i n the upland areas were never very large, and there are few references to birds i n the ethnographic literature. Species were collected as much for their feathers as for the meat (Kennedy and Bouchard 1978). various species of birds were available i n every zone, at different times of the year, and i t i s probable that these species were taken on an encounter basis during ungulate hunts or searches for plants. Both valleys provide wetland habitats where limited numbers of wetland birds were available during spring and f a l l migrations. Birds were snared or chased with bow and arrow (Teit 1900, 1909). Slow-moving species l i k e ptarmigan could have been simply grabbed by hand. Although individuals fished during the f i r s t run of spring salmon i n July, other a c t i v i t i e s (i.e., plant collection and hunting) took precedence at t his time, and intensive fishing did not take place u n t i l August or early September (Bouchard and Kennedy 1975a). From mid-August to mid-September, everyone lived at the salmon fishing stations along the major rivers. Dried salmon and other f i s h were taken to winter village sites, where they were usually placed into an elevated box cache (Kennedy and Bouchard 1978), or into a lined underground cache (Teit 1906). Another subsistence ac t i v i t y carried out i n late summer or early f a l l was the collection and storage of nuts. I t i s not known how the problem of the simultaneous a v a i l a b i l i t y of salmon and nut resources was solved, because the involvement of women was crucial. The preferred method of gathering nuts was to extract them from the caches of squirrels, and this could have been accomplished after the salmon processing was completed (Alexander 1989). The Shuswap and Thompson preferred ponderosa pine (Pinus ponderosa) or the more 108 important white-bark pine (Pinus albicaulus) nutlets, the latte r of which were available at higher elevations at this time. Groups of women, camping for several days (Dawson 1891), collected, cooked or roasted the nutlets, mixed them with berries and stored them. The women also secured roots and seeds i n the f a l l by robbing the nests of sguirrelsand mice (Teit 1900). During October, the people cached salmon caught during the previous months and went to the mountains to hunt (Teit 1909, 1900). Several hunters and their families moved into the uplands where they established temporary hunting camps. Hunting continued u n t i l sufficient meat was prepared for winter storage (Bouchard and Kennedy 1979). In addition, women remaining at the winter vi l l a g e may have continued to collect lower elevation foods (Alexander 1989). The hunting and trapping of game - primarily deer, marten, mink, fisher, beaver, fox, and lynx - continued into November. Several types of caches were used to store food. The most common type among the Thompson bands was the circular c e l l a r for f i s h and other food (Teit 1909). The box-cache was also used extensively, especially i n wooded areas where i t was placed either i n trees or on platforms. According to Teit (1900), the Thompson stored salmon i n such elevated caches for several years, always ensuring an emergency supply. The Li l l o o e t (Bouchard and Kennedy 1975a), and probably the other groups as well, constructed two types of underground caches: one kind, made very carefully and lined with bark, was intended to store food undisturbed u n t i l the following spring; the other was less carefully made, and situated close to the house for foods used throughout the winter (Teit 1906; Kennedy and Bouchard 1978). The location of caches i n upland areas appears to depend on the distance 109 from the resource procurement location to the winter village. For example, Alexander and Matson (1987) have located over 300 cache pits in the parkland zone of the Potato Mountain Range in the central Interior Plateau. In Upper Hat Creek Valley and Highland Valley, no cache pits have yet been located (Pokotylo 1978; Areas Associates 1983, 1986). Resources collected in these two valleys were probably stored at the winter villages, located within a day's travel along the Fraser or Thompson Rivers. Temporary summer dwellings were located at hunting and fishing areas, and temporary camps were also associated with root-baking ovens near root-gathering grounds (Dawson 1891). These camps consisted of wood-frame lodges covered with mats, branches, bark or skins (Dawson 1891; Teit 1900, 1906). At locations where large numbers of people gathered annually for a short time, such as fishing camps and extensive root-gathering grounds, more substantial dwellings with log foundations were constructed (Teit 1900, 1909). Remains of this type of lodge have been seen at high elevations in the Pavilion and Fountain Valleys, and along the terraces of the east bank of the Fraser River (Kennedy and Bouchard 1978). Men hunted and trapped, while women were responsible for the gathering and preparation of roots, berries, and other foods (Teit 1900; Dawson 1891). Although females processed al l plant resources, males helped to con truct the earth ovens used to bake roots for winter storage (Dawson 1891). In addition to these processing activities, the manufacture and repair of tools and utensils probably occurred at summer and fall camps. The Shuswap "used at least 37 species of plants for technological purposes, including the making of tools, weapons, dwellings, canoes, fibres and cords, dyes, baskets, fire, smoke, and sounds" (Palmer 1975b:35). There 110 would have been an optimal time for the procurement and processing of each plant species required, although preparation of the f i n a l product may have been delayed u n t i l the winter months. For example, from September to October, the Mount Currie Li l l o o e t gathered the Indian hemp plant (Apocynun cannabinum) from which fishing lines and nets are made (Bouchard and Kennedy 1975b). Sedge (Scirpus acutus), an important mat-maJcLng material, was gathered i n late summer and early f a l l , while bark from the western white birch (Betula papvrifera), used for making baskets and canoes, was collected most easily i n late spring and early summer (Turner 1979). Both males and females participated i n these a c t i v i t i e s , depending on the amount of work involved and who would use the f i n a l product. Turner (1979:34) states that "as a general rule, the men were involved with the harvesting and construction of wood, as well as of larger sheets of bark for canoes. The women were usually responsible for collecting and preparing fibrous plant materials.... and roots for making baskets, mats, and clothing. However, i f the fibres were to be used as fishing l i n e and net, or i n some way involved with fishing, hunting, or woodworking, the men might gather and process i t as well." Subsistence technology among the three Interior Salish groups was practically identical (Teit 1900), with most of the tools and utensils being made of stone, bone, wood, bark, skins, matting or basketry (Table 4). Although iron was introduced i n the mid-1700's, i t was rare u n t i l 1810, when i t gradually began to replace a l l other materials (Teit 1909). Iron flakers, scrapers, knives, digging sticks, and projectile points were i n common usage when Teit made his ethnographic observations; however, antler was s t i l l preferred for working stone, and bone awls and needles were s t i l l i n use (Teit 1900, 1909). A wide variety of l i t h i c materials was used for projectile points, clubs, axe-heads, chisels, adzes, knives, pestles, hammers, arrow-smoothers, whetstones, f i l e s , mortars, anvils, and skin-Table 4. Interior Salish subsistence technology. i l l Item Material Use (Activity and Worked Material) Knife gritstone cut & work jade and serpentine beaverteeth cut & work jade and serpentine cut arrowshaft smoothers cut and carve wood carve,incise wood & stone chipped stone cut and carve wood cut,carve antler & bone cut roots cut bark cut grass o cut fat from meat o m s l i c e flesh for drying •> s l i c e salmon for drying slate cut f i s h Pipe polisher equisetum polish steatite pipes Scraper stone scrape skins sharp stone scrape fat off flesh bone scrape skin for moccasins bone,horn scrape cambium from bark Hand-hammer pebble s p l i t boulders trim edges of s p l i t boulder remove flakes from edge ground stone drive i n wedge, chisel, stake Flaker antler prepare f i n a l shape of boulder Wedge elk-antler, hardwood, bone cut down trees Mallet wood drive i n light stake,peg Club,axe-head stone drive i n light stake, peg Adze jade,serpentine hafted to wood construct canoes D r i l l stone d r i l l holes i n wood Incisor bone decorate wood Sharpener gritstone,sand sharpen and polish bone equisetum sharpen and polish bone Pounder sharp stick beat skin u n t i l soft stone beat skin u n t i l soft Chisel antler cut down trees bone,horn,stone scrape hair o f f skin Shovel wood remove snow Frame bent sticks support skin over f i r e to smoke poles tied together stretch skins for drying Needle wood,bone,horn sew skins Awl bone sew skins s p l i t roots for baskets sew baskets strengthen basket rims Table 4, continued. Item Material Use (Activity and Worked Material) Case Bag Pin Thread Root-digger o *> Board o Netting stick Scraper Digging stick Scoop 1 1 Separator Pestle Basket Basket,lid Basket,small Basket,conical Basket, f l a t antler skin thorn willow bark,deer sinew, buckskin wood 1 •> • wood • wood wood hard wood wood horn, wood wood mat, bark, roots mat,bark,roots mat,bark,roots mat,bark,roots mat,bark,roots Basket,oval Basket Basket Pot bark bark bark stone Grinding stone stone Vessel Cup Spoon Sti r r e r Tongs Fire d r i l l Long stick Slowmatch Bag Bag Bag String stone bark bark wood, horn, bone wood,antler wood wood wood bark deer paunch bladder salmon skin bark,grass store needles,awls store sewing materials sew skins sew skins dig roots cut bark strings cut roods and bulrushes shred and clean bark fibre cut skin for bags weaving bark thread nets dig hole for winter house dig hole for winter house collect blueberries make wooden spoons make horn,bone spoons make wooden stirrers,tongs separate bark from tree mash berries for drying cook food, heat l i q u i d store food,clothing store sewing tools transport store tobacco store bait store fishing tackle store food cooking berries soaking skins store paint,ochre grind meat,berries grind tobacco store o i l hold fish,meat,roots drink eat food s t i r l i q u i d food l i f t hot stone produce f i r e gather firewood carry f i r e store animal f a t store animal marrow store salmon o i l thread roots 113 Table 4, continued. Item Material Use (Activity and Worked Material) BOW wood,sinew, bark hunt Arrow wood hunt Arrowhead stone (basalt) hunt Arrowshaft smoother sandstone prepare arrows Quiver tanned hide, sagebrush carry arrows Spear wood hunt, f i s h Spear point stone hunt Bagnet bark, wood and horn catch f i s h Club wood,stone k i l l f i s h Hook & line bone,bark catch f i s h Sinker stone catch salmon Trap wood catch trout canoe bark,wood transportation,fishing Sources: Bay 1942; Teit 1900, 1906, 1909; Dawson 1891; Turner 1979 114 scrapers. Elk, caribou and deer antler was used to make wood chisels, stone flakers, wedges, tool handles, and scrapers. Bone, wood, and horn needles and awls were used for sewing skin clothing with bark, sinew or skin thread. Birch-bark and spruce-bark were used to make baskets for storage and plant collection, while cedar and spruce roots were woven into baskets for cooking, storage and plant collection. Mats for floor and wall coverings, bags and place mats were woven of tule, bulrushes, and bark. Netting for hunting was constructed from bark or hemp, using wooden netting sticks. Cedar, yellow pine, and Cottonwood were used for dugout or bark canoes, although several bands did not have access to suitable material and traded for canoes (Teit 1900). Finally, snowshoes were constructed of mountain maple, yew and deerhide, while tumplines used for carrying family possessions, meat, and baskets of roots and berries were constructed of buckskin (Teit 1900). Males were responsible for working i n stone, bone, and wood, while females prepared skins, matting, and basketry (Teit 1900). The early ethnographers did not record either the manufacture or the use of microblades. We should note that Teit, Boas and others restricted most of their observations to l i f e i n the winter pithouse villages, where microcore technology may have disappeared by 3,500 B.P. In addition, most of the stone tools had already been replaced by metal equivalents. Microblades would have been suitable as analogues for several of the tools l i s t e d i n Table 7: knife, scraper, d r i l l , incisor or awl, used on wood,antler, bone, roots, bark, grasses, fat, flesh and skins. Discussion 115 The subsistence-settlement system described above has been compiled from observed and reconstructed data from the late nineteenth century and therefore i t s antiquity i s unknown, i n addition, both Upper Hat Creek Valley and Highland Valley were probably u t i l i z e d during only a portion of the annual subsistence-settlement cycle throughout the prehistoric period. As well, the subsistence-settlement pattern varied considerably throughout the centuries. For example, there i s evidence i n both valleys of a hiatus i n occupation during the Shuswap Horizon. In addition, occupation during the Middle Prehistoric period was substantially more widespread i n Highland Valley, while exploitation of Upper Hat Creek Valley was more intensive during the Plateau Horizon. Finally, the use of microcore technology was not observed by early ethnographers and there i s no direct analogue for i t s replacement within or before the historic period. Paleoenvironmental data from the southern Interior Plateau indicate that extensive grasslands may have already boon i n place i n upland areas by approximately 10,000 B.P. Archaeological data on the Early Prehistoric period i n other areas of the Interior Plateau suggest that subsistence was based on the hunting of large ungulates (Fladmark 1982). Although there i s , as yet, no archaeological evidence of prehistoric occupation of upland areas during the Early Prehistoric period, small groups of hunters may have u t i l i z e d these areas. Ethnographic information presented above indicates that hunting i n the uplands was carried out i n the f a l l by either individuals or groups of males, accompanied by a few women to process the game. Hunting technology at this time included spears, probably traps, 116 snares, and nets, and perhaps the use of group f a c i l i t i e s such as surrounds, fences, and corrals.Sites would have included small residential or f i e l d camps, k i l l and processing sites and lookout stations. There i s some archaeological evidence for the increasing (or beginning) use of upland areas during the Middle Prehistoric period, as warmer temperatures raised alpine tree-lines and reduced the amount of permanent snow. In Highland Valley, at least, the resource base was probably more diverse at this time, since a larger area of open grassland would have produced a corresponding increase i n the ungulate population (Arnoud Stryd, personal onwrnmication 1990). Upper Hat Creek Valley contains no sites clearly attributed to this period, and i t may be that edible resources, particularly roots, were not well-established. If Upper Hat Creek Valley were exploited for the same resources as during the ethnohistoric period, small, mobile family groups would have u t i l i z e d the valley from summer to f a l l , as ungulates moved into summer grazing areas, and roots ripened. Ethnographic information on the use of upland areas pertains only to exploitation of resources by specialized task groups on a large scale for winter storage. However, taking into account the highly seasonal climate, i t s probable that small scale storage was necessary for a successful adaptation to this region. Although i n the ethnohistoric period, roots were collected and prepared for storage during spring and summer months, R. G. Matson (personal communication 1989) suggested that, prior to the large-scale use of salmon which commenced sometime after 4500 B.P., roots may have been exploited primarily during the f a l l . Although ethnographic information relating to the exploitation of root resources i s sparse, sites would have remained small and consisted of small residential 117 base camps, small resource procurement sites and lookout stations for hunting. The latt e r two would probably be archaeologically in v i s i b l e . Small hunting parties would have consisted primarily of males, and would have transported the game back to the residential base camp for processing. In addition, females would have collected and processed the bulk of the f l o r a l resources, either at the small residential camps, or near the procurement sites. Highland Valley contains at least eight sites clearly attributed to the Middle Prehistoric period. As discussed above, archaeological data from the Lochnore Phase sites indicate an occasional, seasonal use of the valley for the hunting of ungulates, rabbit and beaver. As well, the lakeshore location of these sites may indicate the harvesting of waterfowl and freshwater molluscs. This probably occurred during the spring and f a l l months, by mobile family groups occupying small seasonal base camps, i n association with hunting camps and lookout stations. The number and size of well-dated Late Prehistoric sites i n both valleys indicate that the most extensive use of these upland areas occurred during this period. I t seems l i k e l y that the Late Prehistoric pattern was, for the most part, similar to that reconstructed for the ethnographic period. Paleoenvironmental data indicate that the environment during the Late Prehistoric assumed modern characteristics, with a few short, colder, wetter periods. The presence of pithouses and storage p i t s i n the major river valleys clearly indicate a semi-sedentary settlement pattern and a collecting subsistence system. After the intensification of salmon resources, then perhaps i t was also necessary to intensively collect and process more upland resources for storage i n order to last out the winter. 118 Thus, there should be larger and, possibly, more sites i n the uplands i n the Late Prehistoric period. According to currently available data, prehistoric occupation of Upper Hat Creek valley was underway by at least 2400 B.P., and continued throughout the Late Prehistoric period. Pokotylo (1978) inferred that the major use of the valley was for spring plant gathering a c t i v i t i e s , and for f a l l hunting a c t i v i t i e s . However, the presence of a housepit s i t e at the northern end of the valley indicates a winter occupation for at least one season. Waterfowl would have been available also i n the f a l l , and l i t h i c material was available whenever there was no snow cover. Archaeological data indicate that intensification of root resources began approximately 2000 B.P., and continued to the historic period although apparently i n smaller quantities. Although ethnographic sources (Teit 1900, 1906, 1909; Dawson 1891) described root gathering as being primarily conducted by small parties of women, i t i s more probable that the intensive collection and processing implied by the presence of large cooking p i t s dated to the Plateau Horizon was carried out by large family groups, occupying f a i r l y large seasonal base camps, associated with processing sites. At the same time, small hunting camps were occupied primarily by males. During the ensuing Kamloops Period, processing of root resources for storage decreased i n intensity; therefore, the base camps and processing sites should be smaller. Hunting probably continued i n the same manner. In Highland valley, the major resources were deer, f i s h and waterfowl, indicating that the major use of the valley was during the f a l l months. Highland Valley sites represent primarily f a l l occupations, with a few spring, summer and winter sites (Areas Associates 1983). Unlike the 119 situation i n Upper Hat Creek valley, there i s no archaeological evidence to indicate that intensification of edible resources occurred during the Late Prehistoric period, nor any evidence of a winter residence. However, i t i s s t i l l probable that groups using the Highland Valley during the Late Prehistoric period were primarily collectors, focusing on resources to be processed, transported and stored close to winter pithouse villages within the major ri v e r valleys. Thus, we can predict that seasonal residential base camps w i l l be larger, and perhaps more numerous, and that hunting camps and other special purpose sites w i l l be both more common and show evidence of sequential use. Table 5 demonstrates the inferred ethnographic subsistence-settlement pattern i n these two upland valleys throughout the year. The most noticeable characteristic i s the high level of both residential and l o g i s t i c a l mobility selected for the e f f i c i e n t exploitation of altitudinally-zoned and seasonally restricted resources. In addition, a wide range of resources were procured, and consumed or processed for winter storage. Therefore, a need for either a wide range of functionally specific tools or a smaller number of multi-purpose tools i s demonstrated. Finally, high quality l i t h i c raw material i s limited i n a v a i l a b i l i t y only i n Highland Valley, although access to certain cherts or highly vitreous basalts may have been restricted i n Upper Hat Creek Valley. Chapter two presented contemporary archaeological theory dealing with the organizational role of l i t h i c technology and derived a testable model for the organizational role of microcore technology i n the subsistence-settlement system of semi-sedentary hunter-gatherers. The present chapter re-examined archaeological and ethnographic data from the Southern Interior Table 5. Inferred ethnographic subsistence-settlement pattern i n uplands. Month Zone Resource Group Site A c t i v i t i e s October- IDF ungulates males, FC maintenance, November- hide preparation families P butchering IDF ungulates males P butchering >«!KKI«' ungulates males, FC maintenance families P butchering AT ungulates males P butchering December- IDF ungulates,fish males P butchering, early March small mammals fishing IDF ungulates males P butchering Late March- IDF plants, females, FC maintenance A p r i l ungulates families P butchering IDF plants,ungulates females, P plant processing, small mammals males butchering FFB6 plants, females, P plant processing, ungulates families butchering Mid-May- IDF f ish,ungulates families FC maintenance early June plants,waterfowl IDF plants, females, PR plant processing, ungulates males P butchering ESSF plants, families RC maintenance ungulates P butchering, PR plant processing AT plants, females, P plant processing, ungulates males PR butchering June-mid IDF berries,spruce females P plant processing July roots,plants PR plant processing PPBG berries, families RC maintenance plants P plant processing ESSF roots families FC maintenance P plant processing AT plants, males, P butchering, ungulates females PR plant processing Mid- ESSF plants, families RC maintenance August- ungulates P butchering, September PR plant processing IDF plants females P plant processing AT ungulates families P butchering AT: Alpine Tundra FC: f i e l d camp ESSF: Engelmann Spruce-Subalpine F i r P: procurement s i t e IDF: Interior Douglas F i r PR: processing s i t e PPBG: Ponderosa Pine-Bunchgrass RC: residential camp 121 Plateau i n order to provide a model of the way i n which people manufacturing and using microlithic tools may have organized their annual subsistence-settlement system i n upland valleys. The assumption i s made here that the inferred ethnographic subsistence-settlement pattern presented i n Table 5 i s similar to that followed by prehistoric peoples using upland valleys during the Kamloops and Plateau Horizons. The following chapters w i l l test the research proposition advanced here that the cultural significance of microcore technology i s functional, that i s , related not only to the purpose of the tool i t s e l f , but also to i t s organizational role within the subsistence-settlement system. The following chapter describes the data base used to test the model of the organizational role of microcore technology presented i n Chapter I I . 122 CHAPTER XV THE ARCHAEOLOGICAL DATA BASE This chapter describes the archaeological data base vised i n the following analytical chapter: the l i t h i c a r t i f a c t c l a s s i f i c a t i o n scheme, and the study sites. The nature of this particular data base provides information on both technological behaviour, represented by the material by-products of tool use, manufacture and discard, and settlement behaviour, as indicated by the environmental characteristics of the microlithic and non-microlithic sites selected for investigation. A p i l o t study was completed i n order to investigate and select the most ef f i c i e n t and useful methods of data measurement and analysis. Following the analysis and interpretation of preliminary results, the number of attributes and some of the values recorded were changed. During the following discussion of the descriptive data, a brief reference i s made, when appropriate, to the methods and results of the p i l o t study, to j u s t i f y any changes. The f i r s t section provides definitions of the a r t i f a c t c l a s s i f i c a t i o n used, while the second section describes the s i t e selection process and provides pertinent biophysical and archaeological characteristics. Art i f a c t Descriptions Introduction According to the theoretical orientation of this research, stone tool technology i s viewed as a strategy selected by a population to f u l f i l l needs deriving from the procurement, transport, processing and storage of 123 resources. The particular subsistence strategy chosen w i l l be a primary determinant of settlement types and locations (Jochim 1981). I f the settlement strategy comprises the di f f e r e n t i a l use of locations for different a c t i v i t i e s , then the entire range of behaviour involving stone tools w i l l probably not be represented at a l l s i t e locations within the regional settlement round. The variation i n a c t i v i t i e s conducted at each s i t e w i l l be reflected i n v a r i a b i l i t y i n the kind and distribution of the debris produced across s i t e locations. Thus, i f technological behaviour i s responsive to the kind of subsistence strategy selected, then v a r i a b i l i t y i n the manufacture, use, maintenance and discard of tools, as evidenced i n the archaeological record, w i l l r e f l e c t the overall subsistence-settlement strategy of the population under study. The data base i n this study was chosen for i t s potential to r e f l e c t v a r i a b i l i t y i n both the manufacture and maintenance of tools, and i n the use and discard of tools. The data base consists of the flaked l i t h i c artifacts which constitute the dominant, and often the only, class of material remains found i n sites from the study area. The morphological a r t i f a c t typology i s a modification of one developed and used successfully by David Pokotylo (1978). The goals associated with this particular typology are: (1) to identify and document the processes of manufacturing, maintenance and rejuvenation of tools; (2) to identify v a r i a b i l i t y i n tool manufacturing and core preparation strategies; and (3) to identify and document v a r i a b i l i t y i n tool use. The artifacts are divided into two categories: (1) debitage, or by-products of tool manufacture, resharpening, and rejuvenation; and (2) tools, or any a r t i f a c t exhibiting evidence of use by intentional retouch or use wear. 124 L i t h i c Material Types The type of stone selected i s an important part of the technological strategy of any group and can provide information on d i f f e r e n t i a l selection and use of l i t h i c raw material i n relation to their s u i t a b i l i t y for particular tools and tool manufacturing processes. The p i l o t study used a l i s t of l i t h i c material types developed from the Upper Hat Creek Archaeological Project: vitreous basalt, non-vitreous basalt, chert, chalcedony, obsidian, cherty a r g i l l i t e , guartzite, lowgrade volcanic, steatite, quartz crystal, and cobble of undetermined l i t h i c type. I t was d i f f i c u l t to differentiate between chert and chalcedony, which often occur together i n the same ar t i f a c t . In addition, the majority of arti f a c t s were either basalt or chert, with only a few or no examles of the other raw material types. Accordingly, for the remainder of the study, the following l i t h i c material type categories were used: vitreous basalt, non-vitreous basalt, chert and/or chalcedony, and other. Debitage Classification L i t h i c debitage i s defined as the unmodified (by use or intentional retouch) waste products of tool manufacture, resharpening, and rejuvenation. Debitage includes nodules of raw material, cores and core fragments, b i f a c i a l preforms, flakes and flake shatter, block shatter, and microblades. I t i s often assumed that l i t h i c debitage, as an immediate byproduct, i s deposited at the locus of production, and not removed from the s i t e for further use (Pokotylo 1978). One of the goals of this research i s to test this assumption, particularly with respect to microcore technology. The major goal i n adopting this debitage typology i s to provide mutually exclusive categories which w i l l provide information on nianufacrturing, resharpening and rejuvenation processes which may vary between sites. Each a r t i f a c t type i s described below and illustrated i n Figures 5, 6 and 7. Non-microlithic Debitage (Figures 5 and 6) 1. Nodule refers to a chunk of l i t h i c raw material, with cortex s t i l l intact, and no evidence of flake removal (Figure 5a). 2. Core refers to a piece of l i t h i c raw material from which some flakes have been removed. Some cortex may remain and negative flake scars are present (Figure 5b). 3. Platform-remnant bearing flake i s defined by the presence of a portion of an intact striking platform (Figure 6a). 4. B i f a c i a l thinning flake has an extensively facetted, steeply angled, lipped striking platform (Crabtree 1972) and a curved side view (Figure 6b). 5. Flake shatter exhibits at least one flake margin, and a v i s i b l e differentiation between the dorsal and ventral surfaces (Figure 6c). 6. Block shatter lacks an obvious flake margin, or dorsal and ventral surfaces (Figure 6d). 7. Bipolar flake displays two striking platforms at opposing ends of the flake's long axis (Figure 6e). 8. Preform i s a b i f a c i a l l y flaked blank exhibiting a regular outline, but not finished into a recognizable tool class (Figure 6f). Microlithic Debitage (Figure 7) 1. Microcore has been prepared for the purpose of microblade removal. Defining attributes are the presence of a series of p a r a l l e l flake scars 126 Figure 5. Non-micxolithic debitage: a-nodule; b-core. Figure 6. Non-micxolithic debitage: a-platform-rannant bearing flake; b-bifacial tiiinning flake; c-flake shatter; d-block shatter; e-bipolar flake; f-preform. along the d i s t a l end, and a prepared striking platform perpendicular to the flake scars. Microcores can be placed i n one of several stages, depending on whether or not the core has been prepared, u t i l i z e d or discarded immediately prior to i t s f i n a l handling (Figure 7a). 2. Microcore fragment i s the d i s t a l portion of a microcore, or the f l u t i n g surface, defined by the presence of p a r a l l e l flake scars and a platform perpendicular to the scars (Figure 7b). 3. Microcore preparation flake i s defined by the presence of a series (two or more) of p a r a l l e l flake scars and the remnants of a perpendicular striking platform (F gure 7c). 4. Microcore rejuvenation flake exhibits portions of a series of par a l l e l flake scars and a deeply lipped striking platform (Figure 7d). 5. Microblade i s a narrow flake with long p a r a l l e l flake margins, a prepared platform perpendicular to the bulb of percussion, and a triangular or prismatic cross-section (Figure 7e). 6. Microblade proximal fragment i s the proximal portion of a microblade with the striking platform (Figure 7f). 7. Microblade medial fragment i s the medial portion of a microblade lacking both striking platform and d i s t a l termination (Figure 7g). 8. Microblade d i s t a l fragment i s the d i s t a l portion of a microblade lacking the striking platform (Figure 7h). Tool Classification A tool i s defined as a l i t h i c a r t i f a c t which has been modified by intentional retouch or by use-wear. I t i s assumed that formed, or precisely-shaped, tools required a substantial effort i n terms of time and s k i l l , and 129 F i g u r e 7. M i c r o l i t h i c d e b i t a g e : a-mic rooore ; b-micxocore f ragment ; c-microcore p r e p a r a t i o n f l a k e ; d-microcore r e j u v e n a t i o n f l a k e ; e-microb lade ; f -m i c rob l ade p r o x i m a l f ragment ; g-microb lade med ia l f ragment ; h-microblade d i s t a l f ragment . possibly specific raw materials as well. In addition, formed tools, because of the control exerted over outline and size, are more l i k e l y to have been hafted. Thus, formed tools were probably curated more often than expediently-manufactured flake tools (Pokotylo 1978; Hayden 1981). The following tool classes r e f l e c t the technological features of both formed and expediently-manufactured tools. Each tool type was cl a s s i f i e d on the basis of two observable attributes: basic shape, and/or the presence of retouch or u t i l i z a t i o n scars. The criterion for classifying each formed tool was i t s basic outline shape, while the criterion selected a r b i t r a r i l y for classifying an expedient tool was the presence of at least one modified section with continuous scarring for a minimum distance of 2 mm, or two contiguous negative flake scars. Assignment to a basic shape category implicitly recognizes the amount of energy expended during manufacture. For example, on the surface of a biface, the modification i s extensive, entailing a larger amount of energy expenditure than on a modified flake, where the modification i s marginal. Although several of the tool categories have traditional labels that imply function, i n this study they refer only to the basic outline shape. A l l a r t i f a c t s were examined under low magnification (X10) i n order to determine the presence or absence of modification or retouch scars. No attempt was made to differentiate between scar patterns resulting from manufacturing retouch and use retouch because, i n many cases, i t i s not possible without the use of high-power magnification (Knudson 1983; Keeley 1980). The presence of either type of modification was considered sufficient to c l a s s i f y an a r t i f a c t as a tool. Tool types are described below, and illustrated i n Figures 8, 9 and 10. 131 Formed Tools (Figure 8) 1. Uniface was i n i t i a l l y defined by Sanger (1970a: 76) as having "well-defined outlines, reflecting, presumably, deliberate shaping". The unifacial modification extends over at least one-third of the edge (Figure 8a). 2. Graver i s characterized by a unifacially or b i f a c i a l l y retouched spur or projection, although the rest of the a r t i f a c t may be variable i n outline (Figure 8b). 3. Biface i s a symmetrical tool b i f a c i a l l y flaked over the entire surface. Tips are either pointed or rounded, and there i s no evidence of hafting (Figure 8c). 4. Biface end fragment i s a proximal or d i s t a l portion of a biface including the t i p (Figure 8d). 5. Biface medial fragment i s the body portion of a biface and lacks a t i p (Figure 8e). 6. Projectile point i s symmetrical and b i f a c i a l l y flaked over the entire surface, with a sharply-pointed d i s t a l t i p and a base modified for hafting by notching, stemming or thinning (Figure 8f). 7. Projectile point fragment i s a portion of a projectile point and has a portion of the base, and perhaps the body (Figure 8g). Expedient Tools (Figures 9 and 10) 1. Modified cobble exhibits at least one area of continuous modification, usually a crushed surface (Figure 9). 2. Modified flake i s a flake (of any type defined above) having one or more modified areas exhibiting continuous retouch for more than 2 mm, or more than two contiguous negative flake scars. Modification i s restricted to 132 Figure 8. Formed tools: a-uniface; b-graver; c-biface; d-biface end fragment; e-biface medial fragment; f-projectile point; g-projectile point fragment. Figure 9. Expedient tools: modified cobble. within one third of the flake margins (Figure 10a). 134 Microlithic Tools (Figure 10) 1. Modified microblade exhibits at least one area of continuous modification (Figure 10b). 2. Modified microblade proximal fragment exhibits at least one area of continuous modification (Figure 10c). 3. Modified microblade medial fragment exhibits at least one area of continuous modification (Figure lOd). 4. Modified microlade d i s t a l fragment exhibits at least one area of continuous modification (Figure lOe). 5. Modified microcore preparation flake exhibits at least one area of continuous modification (Figure lOf). 6. Modified microcore rejuvenation flake exhibits at least one area of continuous modification (Figure lOg). The next section reports on the si t e selection process and describes the biophysical characteristics/ sampling strategies/ a r t i f a c t inventories, and interpretations of the data by si t e investigators. Site Descriptions Introduction This section describes the twenty-four sites selected for study from the Upper Hat Creek and Highland valleys. A l l sites were c l a s s i f i e d as single component by the original investigators (Areas Associates 1983, 1986; 135 Figure 10. Expedient and microlithic tools: a-modified flake; b-modified microblade; c-modified microblade proximal fragment; d-modified microblade medial fragment; e-modified microblade distal fragment;f-modified microcore preparation flake; g-modified microcore rejuvenation flake. Pokotylo and Beirne 1978; Beime and Pokotylo 1979; Beirne 1979a, 1979b; Pokotylo et a l . 1983). i n order to test the model proposed i n chapter II, an equal number (6) of microlithic and non-microlithic sites from each upland valley are selected i n order to reduce bias related to variable sample size. I t i s assumed that both the microlithic and non-microlithic assemblages precede the introduction of metal to the study area, and that the non-microlithic sites were occupied by semi-sedentary hunter-gatherers using a subsistence-settlement strategy similar to that described ethnographically. The entire flaked l i t h i c collection available for each s i t e i s analyzed, except i n the cases of two very large microlithic sites, EeRilO and EeRj55, which are sampled by selection of those excavation units which contain the highest percentage of microlithic artif a c t s . The sampling strategy i s judgmental, i n that the following s i t e selection c r i t e r i a were used wherever possible: a range of biogeoclimatic zones; a range of s i t e areas and a r t i f a c t densities; independent dating by radiocarbon; and s i t e excavation as well as surface collection. A r t i f a c t densities for p a r t i a l l y collected Highland valley sites are taken from estimates calculated by area investigators. Each microlithic s i t e contains a minimum of two microlithic artifacts, and each non-microlithic s i t e (except EcRg4A) i s dated by the presence of projectile points to either the Kamloops Horizon or the Plateau Horizon. Table 6 provides details on s i t e selection c r i t e r i a . Table 7 provides summary data for the study sites: s i t e type, inferred s i t e function, and a r t i f a c t frequencies, broken down into tools and debitage. The majority of Upper Hat Creek sites selected for this study were not previously c l a s s i f i e d into functional s i t e types; the exceptions are 137 Table 6. Site selection c r i t e r i a . Site Biogeo- Radio- Cultural- Site A r t i f a c t Micro-climatic carbon Historical Area Density l i t h i c Zone Date B.P. Association M2 M2 Artifacts Upper Hat Creek valley Microlithic Sites EeRilO+* iDF(gr) 352 3.99 99 EeRj49+* iDF(gr) 84 5.71 92 EeRJ55+* PPBG 1220+70 Plateau 104 40.70 58 EeRj56+ iDF(gr) 892 3.62 8 EeRJ60+ PPBG 1448 2.25 15 EeRj62+ iDF(gr) 416 1.94 16 Upper Hat Creek Valley Non-microlithic Sites EeRj8+ XDF(gr) Kamloops 88 5.52 EeRj20+ iDP(gr) Plateau 244 1.14 EeRj42+ iDF(gr) Plateau 2242 2.62 EeRj64+ iDF(gr) Kamloops 372 0.92 EeRjlOO+ IDF(gr) Plateau 160 1.02 EeRk52+ ESSF Plateau 160 0.90 Highland Valley Microlithic Sites EcRg2AA+* IDF(gr) 1490+150 Quiltanton 125 73.49 840 1920+210 1120+170 EcRg2CC* IDF(gr) Quiltanton 850 7.94 3 EcRg4C+ IDF(gr) Quiltanton 150 3.62 29 EcRg4 J IDF (gr) Quiltanton 220 6.00 24 EdRglA* IDF(gr) Quiltanton 190 15.15 45 EdRglB+* iDF(gr) 140+80 Quiltanton 60 34.50 57 Highland Valley Non-microlithic Sites EcRg4A+* rDF(gr) 90 2.91 EcRg4B+ TDF(gr) Plateau 150 8.26 EcRg4D iDF(gr) Kamloops 200 2.67 EcRg4E IDF(gr) Plateau 120 3.00 EdRgS* TDF(gr) Kamloops 2000 0.06 EdRg6+* iDF(gr) Kamloops 150 45.00 +: complete or block surface collection *: excavation ESSF: Engelmann Spruce-Subalpine F i r IDF(gr): Interior Douglas F i r (grassland) PPBG: Ponderosa Pine-Bunchgrass Table 7. Summary data for the study sites. Site Site Previous Debitage Tool Ar t i f a c t Type Classification Total Total Total Upper Hat Creek Valley Microlithic Sites EeRilO LS camp s i t e 94 1292 1386 EeRj49 LS 33 447 480 EeRj55 LSR camp s i t e 44 4164 4208 EeRj56 LS intensive occupation 26 3206 3232 EeRj60 LS intensive occupation 51 3252 3303 EeRj62 LS intensive occupation 49 756 805 Upper Hat Creek Valley Non-microlithic Sites EeRj8 LS limited a c t i v i t y 9 477 486 EeRj20 LS limited a c t i v i t y 8 272 280 EeRj42 LS intensive occupation 105 5759 5864 EeRj64 LS intensive occupation 22 321 343 EeRjlOO LS 14 149 163 EeRk52 LSR limited a c t i v i t y 14 130 144 Highland Valley Microlithic Sites EcRg2AA LSP residential camp 576 3632 4208 EcRg2CC LS residential camp 18 273 291 EcRg4C LS residential camp 37 476 513 EcRg4J LS f i e l d camp 18 51 69 EdRglA LS f i e l d camp 27 170 197 EdRglB LS f i e l d camp 73 755 828 Highland Valley Non-microlithic sites EcRg4A LS f i e l d camp 33 197 230 EcRg4B LSH residential camp 23 210 233 EcRg4D LS f i e l d camp/station? 3 32 35 EcRg4E LS station 1 18 19 EdRgS LSH residential camp 8 105 113 EdRg6 LS f i e l d camp 14 887 901 LS: l i t h i c scatter LSP: l i t h i c scatter with post moulds LSR: l i t h i c scatter with roasting p i t LSH: l i t h i c scatter with hearth 139 EeRilO and EeRj55, both interpreted as camp sites (Beirne 1979a, 1979b). However, some Upper Hat Creek Valley s i t e assemblages were previously interpreted as resulting from either intensive occupation, or limited ac t i v i t y , based on tool and debitage analysis, and s i t e locational characteristics. Specific details on analytical methods and interpretations were provided i n Chapter III. These interpretations are provided i n Table 7, and elaborated on i n the following s i t e descriptions. Two sites, EeRj49 and EeRjlOO, not previously analyzed are indicated by a dashed l i n e . Terms used to designate s i t e function for Highland Valley sites are those adapted from Binford (1980) by Stryd and Lawhead (Areas Associates 1983), and w i l l be b r i e f l y re-defined below. A residential camp i s the base where most manufacturing, processing, and maintenance a c t i v i t i e s occur, and where the group resides during prolonged collecting t r i p s . A f i e l d camp i s the temporary residence and work location of a smaller l o g i s t i c a l group. A station i s a place where resources are procured and perhaps processed, a place where task groups stay while gathering information on resource a v a i l a b i l i t y , or a place where resources are temporarily stored. Site descriptions are compiled from project reports (Pokotylo and Beirne 1978; Beirne and Pokotylo 1979; Beirne 1979a; Beirne 1979b; Areas Associates 1983, 1986; Pokotylo et a l . 1983), graduate theses (Pokotylo 1978; Ludowicz 1983; Magne 1985), and B r i t i s h Columbia Provincial Archaeological Site Report Forms. Upper Hat Creek Valley The s i t e assemblages analyzed i n this study were collected during two projects: the Upper Hat Creek Archaeological Project, run from 1976 to 1979, and the Hat Creek Archaeological Project, 1982. The f i r s t project was an assessment of the cultural heritage resource base i n the valley, prior to the proposed construction of a thermal e l e c t r i c development. Only the grassland valley bottom and forested lower slopes were included i n a s t r a t i f i e d random sampling scheme, because of budget and time constraints, and to comply with the project terms of reference. Sampling units consisted of 400 m quadrats, individually assigned to the grassland or forest strata. The sampling fraction was 7.78 %. Sites were a r b i t r a r i l y defined as a concentration of l i t h i c artifacts equal to or more than six items i n a 2 m by 2 m area. A l l sites encountered were completely surface-collected i n a 2 m grid or sampled by transects. Only those sites which were completely surface-collected were selected for this study. Other data collected included the contemporary plant coninunity, the topography, drainage pattern-water source, extent of overview of the surrounding area from the si t e , and ease of access to the s i t e . During the second phase of this project, selected sites were excavated i n a judgmental sampling scheme. The second project was intended to provide additional data on the Middle Prehistoric period i n upland valleys through the excavation of selected microlithic sites. A secondary objective was to provide i n i t i a l information on the nature and distribution of archaeological sites i n the alpine and subalpine zones by a survey conducted along the banks of an unnamed tributary of Anderson Creek, from the headwaters at the divide, to a location upstream of i t s junction with Anderson Creek. A l l f l a t and gently sloping terrain within 50 to 200 metres of the stream was surveyed, and a l l sites located were completely surface collected with a 2 m grid system. Cultural depressions within sites were test-pitted, and any organic material was collected for radiocarbon dating. Data on the contemporary plant cormninity, the topography/ drainage pattern-water source/ overview and s i t e access were also collected. The following sections provide descriptions of the biophysical characteristics/ archaeological assemblages/ and previous interpretations of te prehistoric significance of the study sites. Site locations are provided i n Figure 11. Tabulations of artifacts and raw material types are provided i n Tables 8, 9 and 10. Microlithic Sites 1. EeRilO This i s a large s i t e with a low-density surface scatter, situated on the north bank of an unnamed tributary to Medicine Creek. Site access i s easy i n a l l directions, and a good overview spans 180 degrees/ from north-west to south. In 1977/ a surface collection located one hundred and seventy ar t i f a c t s . In 1978, ten 2 m units were randomly selected for excavation/ and yielded 2402 art i f a c t s . No features or f l o r a l remains were found/ and faunal remains are limited to 5 fragments of land-snail from one excavation unit. Diagnostic arti f a c t s include a lanceolate projectile point and formed gravers thought to represent the Middle Prehistoric period. However, similar projectile points have also been found i n more recent assemblages, and the gravers may be characteristic of the Late Prehistoric period. Beirne (1979a) interpreted the s i t e as a camp sit e , with a c t i v i t y areas for tool manufacture and waste disposal. In addition to the surface collection, units 1 and 5, containing 88% of the microlithic artifacts, were selected for this study. The study sample includes 94 tools and 1292 pieces of debitage. Figure 11. Location of Upper Hat Creek Valley study s i t e s . >r*r^r* o Q r. r> n r^  f» ra ra r* t* r* ra r* ra ra n r» = S X » X 3? 39 X 33 X S 3 3 3 ? 3 X X X X — — 1°°° S o S t i i ; • s o t a, » »- o n Sf t o g wodUMao U O I I S B »nj » in — o* or . Ul I* O — K) C* *• ^4 — ~^  — ro «n <» A ro — — f\J ro lo LD  o > W *» «• *> — co, ro —• ro • ro —' —" cn *0 • < • * • u ro (r* • 0 t/» CT» 0> W —. I > l/i l/> ~> Q> lo m j*. fo » J at ro — o IO ~ CJ u> ro SITE NODULE COSE PLATFORM-REMNANT BEARING FLAKE BIFACIAL THINKING FLAKE FLAKE SHATTER BLOCK SHATTER BIPOLAR FLAKE PREFORM MICROCORE MICROCORE FRAGMENT MICROCDRE PREPARATION FLAKE MICROCORE REJUVENATION FLAKE MICROBLADE MICROBLADE PROXIMAL FRAGMENT MICROBLADE MEDIAL FRAGMENT MICROBLADE DISTAL FRAGMENT UNIFACE GRAVER BIFACE BIFACE END FRAGMENT BIFACE MEDIAL FRAGMENT PROJECTILE POINT PROJECTILE POINT FRAGMENT MODIFIED FLAKE MODIFIED COBBLE MODIFIED MICROBLADE MODIFIED MICROBLADE PROXIMAL FRAGMENT MODIFIED MICROBLADE MEDIAL FRAGMENT MODIFIED MICROBLADE DISTAL FRAGMENT MODIFIED MICROCORE PREPARATION FLAKE MODIFIED MICROCORE REJUVENATION FLAKE TOTAL •SS4TS Aptr+s sscaoB s^urico Abtrarfeai -pBjx+jtf '8 STQ^l Table 9. A r t i f a c t percentage counts across study s i t e s . TVIQl 3xvn HOiiVNjAnna 3ua:ou3iw aiiiiaoM ixni N O i i r m i i M 3HO:OU3IH aimoOH iHjworuj tvisia 3on«o«iw amman uuwarm -mow 3aniau3iw asuioow ixiNsmi IKWIXOUJ 3on.oMtw aiumow 3aniaii3iM tnuiaoH 311103 0 JHIODW n r u o i u i o a w uticmiiijiroiu ixiNnvnnaiH m i n l 3 . n i «mo 13»ii«n tx)nOT»i m a m lo tno . Jm U H N 3 T K I UHiiow. l o ruonN* l o n i a i i i w l i n i «oiii«unn» i»o::in:-l i r t ixo iumi iM I«O30>::M UUnTtDi 3W030H3IM fl«0il<4 j r n i t roj i K . U T M $ « : O U M U 1 H 1 i m i n m nmiMiHiiTmiii m n :NI»TH IKT««I» n .onnj 1103 l i n a c K l H i l l I I I I H l l U l i l i l i l l 1 3 3 3 33332 2 33333 23.1323 SS32§2 3 3 3 3 32 3 5 5 2 5S 3 3 33 55 I I I 5 323 2332 3 55 3 323 333 55 23 3 235 33 I 5 3333 3 = 3 3 =33333 2-=-322 2 2 2 3 333 3 23 2 3 2 33 33 3 5 5 23 3i32335 233232 55 1 3 3 5 33 3 5 n 3 5 5 o a' d o 33 3 3 3 3 12=11= 331133 321333 l i s s i mm 33223 35 52 35 §3 5 2 333535 S5II3S- mm 532323 3 255 3 3 2 3 2 2 23 3 S1IIII null III!!! Illlll Table 10. Frequency and percentage counts of l i t h i c raw material type. Site Vitreous Non-Vitreous Chert and/or Other Total Basalt Basalt Chalcedony Fr. Pet. Fr. Pet. Fr. Pet. Fr. Pet. Fr. Pet. EeRilO 808 58.3 177 12.8 395 28.5 6 0.4 1386 100.0 EeRj49 127 26.5 41 8.5 283 58.9 29 6.1 480 100.0 EeRj55 201 4.7 2948 70.1 1055 25.1 4 0.1 4208 100.0 EeRj56 165 5.1 2983 92.3 81 2.5 3 0.1 3232 100.0 EeRj60 302 9.1 2057 62.3 937 28.4 7 2.1 3303 100.0 EeRj62 195 24.2 541 67.2 68 8.4 1 0.2 805 100.0 EeRj8 70 14.4 396 81.5 0 0.0 20 4.1 486 100.0 EeRj20 41 14.7 195 69.6 42 15.0 2 7.0 280 100.0 EeRj42 1152 19.6 2883 49.2 1820 31.0 9 2.0 5864 100.0 EeRj64 52 15.2 67 19.5 223 65.0 1 0.3 343 100.0 EeRjlOO 39 23.9 110 67.5 14 8.6 0 0.0 163 100.0 EeRk52 66 45.8 75 52.1 3 2.1 0 0.0 144 100.0 EcRg2AA 2188 51.9 1477 35.1 285 6.8 258 6.2 4208 100.0 EcRg2CC 46 15.8 233 80.1 6 2.1 6 2.0 291 100.0 EcRg4C 21 4.1 483 94.2 8 1.6 1 0.1 513 100.0 EcRg4J 29 42.0 36 52.2 2 2.9 2 2.9 69 100.0 EdRglA 146 74.1 38 19.3 12 6.1 1 0.5 197 100.0 EdRglB 712 85.9 105 12.7 10 1.2 1 0.2 828 100.0 EcRg4A 87 37.8 130 56.5 3 1.3 10 4.4 230 100.0 EcRg4B 43 18.4 179 76.8 7 3.0 4 1.8 233 100.0 EcRg4D 1 2.8 34 97.2 0 0.0 0 0.0 35 100.0 EcRg4E 6 31.6 12 63.1 0 0.0 1 5.3 19 100.0 EdRg5 0 0.0 111 98.2 2 1.8 0 0.0 113 100.0 EdRg6 567 62.9 332 34.9 1 1.1 1 1.1 901 100.0 146 2. EeRi49 This small l i t h i c scatter i s located on a small saddle and ridge, 6 metres from an unnamed creek, s i t e access i s easy i n a l l directions, and the overview spans 360 degrees. In 1977, two 2 m units were excavated i n arbitrary 10 cm levels. A l l microblades were located i n a s i l t layer, indicating deposit on by runoff from above the s i t e . In 1982, an additional two 2 m units were excavated, i n arbitrary 5 cm levels, and again, the microlithic ar t i f a c t s were located i n s i l t . The tool assemblage was uninterpretable, and the debitage derived from the early steps i n the reduction sequence (Pokotylo 1978). The l i t h i c assemblage under study here originates from the 4 excavation units and the surface collection, and consists of 33 tools and 447 pieces of debitage. 3. EeRi55 This s i t e i s a l i t h i c scatter associated with a cultural depression identified as the remains of a multi-component root-roasting oven (Beirne 1979b:49). I t i s located on a gentle slope i n Houth Meadow, at the north end of the valley, with easy access from the east. Overview from the s i t e i s good from northwest to southeast. Fresh water i s located i n a tributary to Hat Creek, 150 metres to the south. A surface collection located one hundred and sixty-eight a r t i f a c t s . The s i t e was excavated i n 1977 and 1978, i n three areas designated on the basis of surface features. Area B, where the surface collected artifacts were located, contained over 98% of the t o t a l number (8947) of art i f a c t s collected. The root-roasting ovens are dated to approximately 1220 years B.P. Two projectile points located i n Area B are chronologically diagnostic of the Plateau Horizon. Beirne (1979b) interpreted Area B as a camp-site, probably associated with the repeated use of the earth-ovens for roasting roots and meat, and delineates a c t i v i t y areas used for waste disposal and resource processing on the basis of a r t i f a c t densities and ethnographic analogy. Analysis of the surface collection indicated that the tools were primarily expedient, representative of a limited term occupation, and that the debitage derived from a broad range of reduction a c t i v i t i e s encompassing the majority of basic manufacturing steps (Pokotylo 1978). Unit 3B, which contains 95% of the microlithic artifacts, and twelve fragments of unidentifiable bone, was selected as the sample for this study, i n addition to the surface collection. The study sample includes 44 tools and 4164 pieces of debitage. 4. EeRi56 This large l i t h i c scatter i s located on a l i g h t l y forested r o l l i n g plain i n Houth Meadow, with easy access from a l l directions. Overview from the si t e i s clear only to the southeast. Distance to fresh water i s 100 metres. Earlier analysis indicated that the assemblage was characterized by a high average tool frequency and diversity indicative of a more intensive occupation and a wide range of a c t i v i t i e s , and the majority of manufacturing steps were represented (Pokotylo 1978). The a r t i f a c t assemblage from the surface collection includes 26 tools and 3206 pieces of debitage. 5. EeRi60 This s i t e i s a large surface l i t h i c scatter located i n a high open area, near Lloyd Creek i n Houth Meadow. Access to the s i t e i s easy from a l l directions, and the overview from the s i t e spans 360 degrees. Distance to fresh water i s less than 100 metres. Analysis of tools and debitage indicated again an intensive occupation with a wide range of a c t i v i t i e s (Pokotylo 1978). The assemblage studied here comprises the surface 148 collection and contains 51 tools and 3252 pieces of debitage. 6. EeRi62 This s i t e i s a smaller l i t h i c surface scatter, located on a r o l l i n g plain and ridge i n Houth Meadow, adjacent to an unnamed creek. Again, access to the s i t e i s easy i n a l l directions, but the overview i s restricted to the south. Distance to fresh water i s approximately 20 metres. Pokotylo's (1978) analysis indicated an intensive occupation, with a wide range of l i t h i c manufacturing stages performed at the s i t e . The materials studied here include 49 tools and 756 pieces of debitage from the surface collection. Non-microlithic Sites 1. EeRj8 This l i t h i c scatter i s located i n a more heavily treed area, on a gentle slope 250 metres west of Hat Creek. Access to the s i t e i s easy i n a l l directions, and the overview spans 360 degrees. The s i t e assemblage included tools suggestive of a brief s i t e occupation and debitage from a wide range of manufacturing steps (Pokotylo 1978). The surface collection studied here includes 9 tools and 477 pieces of debitage. 2. EeRi20 This l i t h i c scatter i s located on a r o l l i n g plain, adjacent to Finney Creek to the east. Site access i s easy i n a l l directions, but overview i s restricted to the south. Previous analysis indicated that the s i t e assemblage included expediently-mamifactured tools from a limited term occupation, and debitage from a wide range of reduction a c t i v i t i e s (Pokotylo 1978). The surface collection includes 8 tools and 272 pieces of debitage. 149 3. EeRi42 This i s a very large l i t h i c surface scatter, located on a bench to the north of Anderson Creek. Access to the s i t e i s easy from the west, and overview spans 240 degrees, form east to west. Distance to fresh water i s 25 metres. According to earlier analysis, the assemblage contained tools from an intensive occupation with a wide range of a c t i v i t i e s , and debitage from a wide range of manufacturing steps (Pokotylo 1978). The assemblage includes 105 tools and 5759 pieces of debitage from the surface collection. 4. EeRi64 This l i t h i c scatter i s situated on an open grassland bench immediately west of Hat Creek. Site access i s good from the north and south, with a 120 degree overview i n the same directions. The s i t e assemblage contained tools used i n a wide range of intensive a c t i v i t i e s , and debitage indicative of a wide range of manufacturing steps with emphasis on the later reduction stages (Pokotylo 1978). The assemblage studied here includes 22 tools and 321 pieces of debitage from the surface collection. 5. EeRilOO This s i t e i s located on gently r o l l i n g open grassland, adjacent to an unnamed creek. Access to the s i t e and overview from the s i t e are very easy i n a l l directions. Distance to fresh water i s less than 10 metres. There i s no previous detailed analysis of the s i t e assemblage, which includes 14 tools and 149 pieces of debitage from a surface collection. 6. EeRk52 This s i t e i s located i n the sub-alpine zone i n a low saddle i n the headwaters area of Anderson Creek, i n the Clear Range. Site access i s easy i n a l l directions, and overview from the s i t e spans 180 degrees, from west to east. Distance to fresh water i s more than 100 metres. The s i t e i s a l i t h i c surface scatter, associated with a cultural depression identified as a root-roasting oven. Shovel testing of four lm units revealed a very shallow subsurface deposit of 5 to 15cm i n depth. Preliminary analysis suggested that the s i t e assemblage, l i k e a series of others i n the alpine zone, i s characterized by an expedient technology, and probably dates to the Late Prehistoric period (Pokotylo et a l . 1983). The small area and limited variety i n the assemblage are indicative of a limited a c t i v i t y / occupation s i t e used on a seasonal basis. The assemblage studied here includes 14 tools and 130 pieces of debitage from both the surface collection and two test p i t s . Highland Valley The s i t e assemblages analyzed i n this study were collected by Areas Associates i n 1982 and 1985, as part of a contract with Ocminco Ltd. to record and investigate cultural heritage resources before proposed expansion of the copper mine. Archaeological survey was restricted to the Lake Zone adjacent to the two major lake systems, within the grassland portion of Interior Douglas F i r biogeoclimatic zone. A l l sites located are situated along the northwest margins of the three major lakes. An i n i t i a l reconnaissance of the area located twenty-one prehistoric sites, which were separated by s t e r i l e areas of at least 100 metres i n width and extended, i n some cases, over 10,000 sq metres (Brolly 1981). After reviewing the local ethnographic and archaeological literature, and contemporary hunter-gatherer theory, Stryd and Lawhead (Areas Associates 1983) proposed that prehistoric sites i n Highland valley were more l i k e l y to be small, oval or circular 151 clusters of artif a c t s reflecting short-term use by small groups of people. Therefore, during the major investigation (Areas Associates 1983, 1986), the cri t e r i o n for establishing s i t e boundaries was changed to an a r t i f a c t density of 2 per square metre, or 4 per square metre i f clusters were contiguous or overlapped. Sites were re-examined, re-mapped as fi f t y - f o u r smaller sites, and tested i n the following manner. An arbitrary 1 x 1 m grid was placed over the s i t e surface, and the 1 m units were systematically sampled according to the following scheme: large sites were sampled at a rate of 1% or every 10th unit, and small sites were sampled at a rate of 4% or every 5th unit. The ground surface of each unit selected was cleared by removal of the surface vegetation and l i t t e r mat. Each 1 m unit selected for clearance i s defined as a Surface Exposure Unit (SEU). In addition, a single test p i t was excavated into the centre of every f i f t h SEU, or every 50 m. Selected sites were investigated more intensively by either block surface exposure or excavation, and a r t i f a c t collection. Other data recorded included modem vegetation cover and drainage pattern. Artifacts analyzed i n this study include the to t a l sample for each si t e , from surface collections, SEU program, test p i t s , and block excavations. Figure 12 provides locations of the study sites. Tabulations of artifacts and raw material types are provided i n Tables 11, 12 and 13. Microlithic Sites 1. EcRg2AA This l i t h i c scatter i s located 105 metres south of Quiltanton Lake, on a large terrace. Artifacts were located i n thirtee SEU's cleared at 5 m intervals, and f i f t y - s i x units of an extensive block excavation program. 0 L 0.5 1.0km _J I Features located during excavation include twenty-six post moulds that are interpreted as the remains of a spring-fall habitation stnxcsture, a possible rock alignment which may be a wall, a p i t , and a fire-altered rock scatter. Several pieces of fire-altered rock and three deer phalanges were also located. Three radiocarbon dates obtained from post moulds indicate that the s i t e was re-occupied several times within the last 800 years; however, other data may indicate a single occupation. The investigators (Areas Associates 1983) interpreted the s i t e as a residential camp where a variety of maintenance tasks took place, including l i t h i c tool production, hide working, and food preparation. The assemblage studied here includes 576 tools and 3632 pieces of debitage. 2. ECPXT2QC This very large s i t e i s situated 2 metres above Quiltanton Lake on a large lake-edge f l a t largely denuded of vegetation by recent a c t i v i t i e s . Eight cultural SEU's and two excavation units, located at 10 m intervals, contained a r t i f a c t s . According to s i t e investigators (Areas Associates 1983), the s i t e probably consists of two small a c t i v i t y areas, a tool production area and a medium-sized residential camp. The assemblage included i n this study consists of 18 tools and 273 pieces of debitage. 3. EcPxr4C This s i t e i s a large l i t h i c scatter located approximately 50 metres from Quiltanton Lake i n a saddle between two knolls. Investigation included a complete surface collection of one hundred and one units, and the clearance of four SEU's placed at 10 m intervals. The s i t e was interpreted as a residential camp, with a wide range of tools and debitage representing a variety of maintenance a c t i v i t i e s (Areas Associates 1983). The assemblage 154 included i n this study contains 37 tools and 476 pieces of debitage. 4. EcRa4J This l i t h i c scatter i s situated on a large, poorly-drained f l a t along the shore of Quiltanton Lake. Artifacts were located i n thirteen SEU's placed at 5 m intervals. In addition to l i t h i c a r t ifacts, a fragment of unidentifiable calcined bone was located. The investigators suggested that the s i t e was a tool production station and possible f i e l d camp, where a c t i v i t i e s such as the manufacture of wood, antler, bone and l i t h i c tools occurred (Areas Associates 1983). The assemblage under study here includes 18 tools and 51 pieces of debitage. 5. EdPxrlA This l i t h i c scatter i s located between Twenty-four-Mile Lake and Big Divide Lake, 25 metres from an unnamed creek connecting the two lakes. Artifacts were located i n six SEU's placed at 10 m intervals, i n one 50 cm by 50 cm test p i t within the most productive SEU and i n seven 1 m excavation units placed within a visually defined a r t i f a c t cluster. The s i t e was interpreted as a moderate-sized campsite where a variety of a c t i v i t i e s occurred, including plant processing, implement tooling, organic tool manufacture and l i t h i c tool production (Areas Associates 1986). The assemblage studied here includes 27 tools and 170 pieces of debitage. 6. EdRglB This small l i t h i c scater i s situated approximately 50 metres north of Big Divide Lake, on an open grassy f l a t . During the i n i t i a l investigation, two SEU's located at 10 m intervals were exposed, and one 50 by 50 cm test p i t i n the most productive SEU, and four 1 by 1 m test p i t s i n the v i c i n i t y of the same SEU were excavated. Later, twenty-four 1 m excavation units were 155 judgmentally placed within the same area. The investigators interpreted the s i t e as a f i e l d camp where a variety of a c t i v i t i e s occurred, including food processing and a l l stages of l i t h i c tool production (Areas Associates 1986). The assemblage studied here includes 73 tools and 755 pieces of debitage. Non-microlithic Sites 1. EcRj4A This small l i t h i c scatter i s situated approximately 75 metres from the shore of Lake Quiltanton. Investigation included a systematic SEU test and a block surface exposure of a portion of the s i t e with the highest a r t i f a c t density. One fragment of deer skull (possibly recent) was located along with four hundred and ninety-five pieces of fire-altered rock i n a dispersed fashion with no discernible hearth. The s i t e assemblage includes primarily expedient tools representing a wide range of maintenance and manufacturing a c t i v i t i e s , and debitage from the later stages of tool manufacture and finishing. The investigators interpreted the s i t e as a small b r i e f l y occupied f i e l d camp Where l i t h i c tools were finished and organic materials were worked (Areas Associates 1983). The assemblage consists of 33 tools and 197 pieces of debitage from forty-two SEU's. 2. ECPXT4B This l i t h i c scatter i s situated on a large f l a t approximately 80 metres from Quiltanton Lake. The investigation consisted of an exposure of eight SEU's at 5 m intervals, and a continuous block exposure i n the area with the highest a r t i f a c t density. A small microlithic component was located i n the southwest corner of the s i t e ; however, the investigators considered i t to be the result of a second, later occupation of the s i t e and i t i s not included 156 i n the present study. Exposed artifacts were clustered around two hearths, and a quantity of fire-altered rock was noted and weighed. The s i t e assemblage includes tools inferred to represent a variety of maintenance a c t i v i t i e s and debitage from the f i n a l stages of tool production and finishing. The s i t e was interpreted as a small residential camp (Areas Associates 1983). The assemblage includes 23 tools and 210 pieces of debitage from thirty-two SEU's. 3. EcRgJD This large l i t h i c scatter i s located at the base of a moderate slope, 60 metres northwest of Quiltanton Lake. Investigation consisted of the exposure of twelve SEU's placed at 5 m intervals. Low a r t i f a c t density indicates a brief occupation. The investigators interpreted the s i t e as a small transitory camp or task station where l i t h i c , and possibly organic, tools were finished (Areas Associates 1983). The assemblage under study here consists of 3 tools and 32 pieces of debitage from ten SEU's. 4. EcPxr4E This very small l i t h i c scatter i s located approximately 35 metres from Quiltanton Lake. Investigation consisted of the exposure of six SEU's located 5 m apart. A small unidentifiable mammal bone fragment was located near the centre of the a r t i f a c t cluster. The s i t e assemblage includes expedient tools and debitage from the f i n a l stages of tool manufacture. The si t e was interpreted as a small briefly-occupied campsite or task station of similar age to EcRg4D. EcRg4D and EcRg4E may be small clusters of the same s i t e (Areas Associates 1983). The assemblage includes 1 tool and 18 pieces of debitage. 157 5. EdRoS This large l i t h i c scatter i s located on a lacustrine terrace 60 metres from the northeast end of Twenty-four Mile Lake. Because the s i t e i s so large (2000 square metres), the investigation consisted of a systematic inspection of the entire s i t e i n order to locate a l l material/ and a limited SEU and test program. Within an area of 250 m2, eighteen SEU's were placed at 5 m intervals/ and a single test p i t was excavated within each SEU containing cultural material (five). Artifacts diagnostic of the Kamloops Phase include the basal fragment of a Kamloops projectile point/ and a small triangular projectile point with a possible unilateral side notch. The s i t e assemblage contains tools u t i l i z e d i n a variety of tool production and maintenance tasks and debitage indicative of the intermediate and f i n a l stages of tool manufacture. One small unidentifiable fragment of calcined bone and two pieces of fire-altered rock were recovered/ and two small hearths were also noted. The s i t e was interpreted as a b r i e f l y occupied residential camp (Areas Associates 1983). The assemblage consists of 8 tools and 105 pieces of debitage. 6. EdRcr6 This small dense l i t h i c scatter i s located on the same terrace as EdRgS at the northeast end of Twenty-four Mile Lake. Site investigation consisted of the excavation of twenty-nine SEU's placed at 5 m intervals, and the excavation of twenty-five contiguous 1 m units within the area of highest a r t i f a c t density. Two small fragments of bone were located: one burnt fragment of an unidentifiable species and one f i s h bone fragment/ possibly recent. A small rock hearth was located along the southern edge of the excavated cluster/ along with fragments of fire-altered rock i n smaller 158 clusters. The s i t e assemblage i s characterized by primarily expedient tools probably used i n tool production and maintenance tasks, and debitage resulting from the intermediate and f i n a l stages of tool reduction. The investigators interpreted the s i t e as a small l o g i s t i c a l camp and stone tool production station (Areas Associates 1983). The assemblage includes 14 tools and 887 pieces of debitage from 26 SEU's. Summary of Site Data Base The preceding section provides a brief description of the biophysical characteristics, archaeological assemblages, and interpretation of the prehistoric significance of the sites selected for the study. A variety of biogeoclimatic zones are included i n the Upper Hat Creek valley sample, while only one zone was available for study i n Highland Valley. Sites from Upper Hat Creek Valley are located i n the Interior Douglas F i r , Ponderosa Pine-Bunchgrass and Engelmann Spruce-Subalpine F i r zones, while those from Highland Valley are a l l situated i n the Interior Douglas F i r zone, around the three major lakes on the valley floor. Major cultural-historical associations include the Plateau Horizon (2400 to 1200 B.P.), and the Kamloops Horizon (1200 to 200 B.P.); at this time, the majority of microlithic sites are not clearly associated with either of these constructs. Only one microlithic s i t e i n the Upper Hat Creek Valley sample i s dated by radiocarbon; this i s EeRjSS, attributed to the Plateau Horizon. Two microlithic sites i n the Highland Valley sample are radiocarbon-dated: EcRg2AA, dated to both the Plateau and Kamloops horizons, but attributed by the original researchers to the Quiltanton Complex, and EdRglB, dated to the Kamloops Horizon, but also attributed to the Quiltanton Complex (Areas Associates 1983, 1986). Due to the s i t e selection process, the majority of non-microlithic sites are associated with either the Plateau Horizon or Kamloops Horizon. A t o t a l of twenty-four l i t h i c scatter sites were selected, with an equal number located i n each valley. Microlithic and non-microlithic sites f a l l into three s i t e function categories: residential camps, f i e l d camps and stations. The t o t a l number of artifacts ranges from 69 to 4208 for the microlithic sites, and from 19 to 5864 for the non-microlithic site s . Estimated s i t e areas range from 60 to 1448 square metres for the microlithic sites, and from 88 to 2242 square metres for the non-microlithic site s . A r t i f a c t densities range from 1.94 to 73.49 per square metre for microlithic sites, and from 0.06 to 45.00 per square metre for non-microlithic sites. The minimum and maximum numbers of microlithic a r t i f a c t s are 3 and 840, respectively. Discussion The l i t h i c analysis was completed entirely by the author, and results are internally consistent. Some discrepancies between a r t i f a c t counts provided i n this study and those provided by the original investigators are to be expected. Only chipped stone artifacts were included; ground stone, bone and antler a r t i f a c t s , and historic artifacts were not. According to a r t i f a c t forms provided by the Royal B r i t i s h Columbia Museum, art i f a c t s are missing from several of the Highland Valley sites. In addition, a l l assemblages were re-classified, and the c r i t e r i a used for this study may not agree with those of previous researchers. For example, the a r t i f a c t typology used for the 160 Highland Valley assemblages (Areas Associates 1983, 1986) includes fiv e types of flakes with striking platforms, while this study contains only three types of flakes with platforms; none of these are dire c t l y equivalent, according to definitions provided. Several problems arose during the s i t e selection process: the sampling of biogeoclimatic zones i s incomplete for both valleys; only a very small number of sites are radiocarbon-dated; the majority of sites were surface collected and not excavated; and the surface collection of several of the sites from Highland Valley was systematically sampled, based on a small sampling fraction. In addition, the c r i t e r i a used to define s i t e boundaries created, as the researchers predicted (Areas Associates 1983), clusters of small sites which may be a c t i v i t y areas within a single settlement. The mean area of the Highland Valley s i t e sample, 358.8 sq m, i s smaller than that of the Upper Hat Creek Valley s i t e sample, 546.8 sq m. However, at this early stage of archaeological research, i t i s impossible to determine whether the s i t e area i s a consequence of the aboriginal settlement pattern or of the archaeological research method. Therefore, the representativeness of the data, particularly from Highland Valley sites, i s not known at this time. In spite of the shortcomings mentioned above, this i s the f i r s t investigation of microcore technology to focus on a comparison of both tools and debitage from a series of sites i n two separate geographic areas within the Canadian Southern Interior Plateau. The strength of the study l i e s i n the large sample available for analysis: twenty-four sites, from two upland valleys, with a t o t a l a r t i f a c t sample of 28,331 tools and debitage. In addition, the data from both valleys were collected i n a compatible manner, i.e. complete biophysical information, and similar a r t i f a c t collection and recording techniques. The follcrwing chapter describes the development and application of analytical methods used on the data base. 162 CHAPTER V ANALYTICAL METHODS AND RESULTS This chapter describes the development and application of the three analytical methods used i n this study to investigate the nature and distribution of a c t i v i t i e s associated with microlithic technology: debitage analysis, tool analysis, and a c t i v i t y area analysis. These analyses were carried out i n order to re-classify the function of the study sites using equivalent c r i t e r i a , and i n order to provide the data required for testing the research hypotheses presented i n Chapter II. To j u s t i f y any changes made, the following discussion of analytical methods refers b r i e f l y , when appropriate, to the methods and results of the p i l o t study. Debitage Analysis Non-Microlithic Debitage  Introduction The manufacture of chipped stone tools i s a subtractive technology and therefore, the resultant debitage displays oombinations of attributes which constitute evidence of the entire manufacturing, resharpening and rejuvenation sequence. The manufacture and maintenance of stone tools generally produces a large amount of waste material, or debitage, and usually occurs adjacent to or close by l i v i n g areas (Carr 1984). In addition, l i t h i c debitage i s deposited and either l e f t at the place of tool manufacture or maintenance, or discarded within a metre or two of the workplace (Spurling and Hayden 1984). Therefore, an analysis of l i t h i c 163 debitage can provide an accurate interpretation of the nature of a c t i v i t i e s involving the manufacture and rejuvenation of l i t h i c tools at the s i t e , even though the tools themselves may no longer be present. In addition, ethnographic and ethnoarchaeological studies indicate that the entire range of a c t i v i t i e s involving tool manufacture, use and rejuvenation may not be present at every s i t e occupied during the annual subsistence-settlement cycle (Binford 1980). Residential sites, either base camps or residential camps, should contain debitage from a l l stages of manufacture, resharpening, and rejuvenation, while special purpose sites ( f i e l d camps and stations) should contain debitage from only one or two of these stages (Binford 1980; Camilli 1983; Pokotylo 1978; Raab et a l . 1979). Therefore, differentiation among these various stages involving chipped stone tools i s integral to distinguishing among the various types of sites occupied by any one prehistoric aboriginal group. The goal of manufacturing stage analysis i n this study i s to calculate the relative importance of sequential manufacturing stages i n each s i t e . The conclusions are relevant only for the sites analyzed i n this study because the method i s primarily comparative. Non-inicrolithic debitage i s defined as the waste or non-utilized products that result from the manufacture, rejuvenation, and resharpening of flake blanks and formed tools. Although the majority of current experimental research i s confined to the quantification of biface reduction stages (Magne 1985; Magne and Pokotylo 1981; Newcomer and Sieveking 1980; Raab et a l . 1979; Morrow 1984; McAnany 1989), the non-microlithic debitage i n this study probably also results from the manufacture of other formed tools such as unifaces and gravers. In addition, although the stages of rejuvenation and resharpening are included i n the definition, the parameters for these stages 164 have not been quantified by experimental research. Pokotylo 1s (1978:250) results suggested that "the technological processes involved i n the manufacture of chert stone tools are quite different from and/or more subtle than those surrounding the production of basalt tools". However, Magne's (1985) study found no substantial differences based on the attributes selected for analysis. Therefore, u n t i l further research c l a r i f i e s this problem, this study treats basalt and chert debitage as a single sample for the purpose of deriving manufacturing stages. Selection of Attributes Quantitative attributes have been used to indicate progression i n manufacturing stage on platform-remnant bearing flakes (Pokotylo 1978; Magne and Pokotylo 1981; Magne 1985; Raab et a l . 1979; McAnany 1989). Pokotylo (1978) used multivariate s t a t i s t i c a l techniques on archaeological data to reduce a l i s t of 19 variables to 5 independent variables reflecting reduction stages: weight, ventral flaking angle, dorsal scar count, striking platform width, and bulb of applied force. In a later study of experimentally produced core-reduction and blank-reduction debitage, Magne and Pokotylo (1981) concluded that four attributes contribute the greatest amount of non-redundant information on reduction stages for platform-remnant bearing flakes: weight, dorsal scar count, platform scar count, and cortex cover. Magne (1985) later determined that weight does not contribute significantly to the identification of reduction stages, but excluded flakes less than 5 mm i n length, as well as flakes considered suitable for further reduction as blanks from the analysis. Several studies have found that, as 165 manufacturing progresses, the size, measured as weight and/or maximum dimension, of flaxes produced decreases (Pokotylo 1978; Magne 1985; Raab et a l . 1979; Newcomer and Sieveking 1980). Consequently, this study uses maximum dimension as a measure of size, because this measurement does not require specialized equipment and involves less data collection time per ar t i f a c t than recording weight. The presence of b i f a c i a l thinning flakes constitutes r e l i a b l e evidence of the manufacture or resharpening of bifaces, although a strategy for calculating sequential manufacturing stages for b i f a c i a l thinning flakes has not yet been developed (Magne 1985). Magne and Pokotylo (1981) suggested that b i f a c i a l thinning flakes derive from the later stages of biface manufacture, after core reduction, and during the later stages of blank reduction. In addition, Magne (1985) found that bipolar flakes derive from the early stages of core reduction. Although the set of attributes measured on platform-remnant bearing flakes i s also measured on b i f a c i a l thinning flakes and bipolar flakes, manufacturing stages are not calculated for these two a r t i f a c t classes. Flake shatter and block shatter, by definition, lack striking platforms. However, Magne and Pokotylo (1981) found that both the size, measured by weight and maximum dimension, and the amount of cortex present decrease throughout the manufacturing sequence. The f i n a l l i s t of attributes recorded for non-microlithic debitage i s given below, with a definition of each attribute and procedures for i t s measurement. Figure 13 provides a diagram of attributes selected for platform-remnant bearing flakes, flake shatter and block shatter. A. Maximum dimension i s measured i n millimetres by comparing each flake 166 0. STRIKING PLATFORM E. STRIKING PLATFORM C. DORSAL FLAKE SURFACE WIDTH SCARS VENTRAL SIDE LONGITUDINAL DORSAL SIDE CROSS-SECTION Figure 13. Attributes measured on platform-remnant bearing flakes and shatter. 167 with a set of c i r c l e s pre-measured by intervals of 1.0 millimetres and taking the corresponding measurement from the smallest c i r c l e which would contain the entire flake. B. Cortex cover i s the amount of cortex, or natural raw material surface, remaining on the flake's dorsal surface, measured i n 25% increments, from 0% to 100%. The amount of cortex cover decreases throughout the reduction process. C. Dorsal scar count i s the t o t a l number of scars on the flake's dorsal surface, except those attributed to platform preparation. A larger number of dorsal flake scars indicates increased preparation of the core face, a later step i n the manufacturing sequence. D. Striking platform preparation refers to the type of platform preparation for each flake, particularly relating to evidence for the number of previous flake removals indicating advanced reduction stages. Values for this attribute are, i n random order: - p a r t i a l l y removed, preparation unidentifiable -single facet platform -multiple facet platform -cortex-covered platform -cortex and single-facet platform -cortex and multiple-facet platform -ground/polished platform E. Striking platform width i s the maximum dorsal-ventral measurement i n millimetres, perpendicular to the striking platform width. Striking platform width should decrease throughout the manufacturing sequence. 168 Definition of Mairufacturinq Stages Mthough some archaeologists (e.g. Camilli 1983; Pokotylo 1978; Sullivan 1987) have examined attribute patterning at the s i t e level i n order to determine the nature of manufacturing and rejuvenation processes, another method also used successfully assigns each piece of debitage to a specific manufacturing stage (Magne and Pokotylo 1981; Magne 1985; Hayden and Hutchings 1989). During analysis of the p i l o t study data, both methods were used, and better differentiation among the sites was achieved with the second method. Even though a certain percentage of flakes and shatter w i l l undoubtedly be assigned to an incorrect category, the main goal to be achieved i s the determination of the relative importance of sequential manufacturing stages, as demonstrated by the debitage deposited at the s i t e . This study uses the definition of manufacturing stages provided i n Magne and Pokotylo (1981). Three major a c t i v i t i e s are involved i n the production of a biface: l . core reduction to produce flake blanks; 2. flake blank reduction to produ e the f i n a l biface, and 3. resharpening and use breakage. Core reduction produces early stage flake and block shatter and early stage platform-remnant bearing flakes. Late core and early blank reduction produces middle and late stage platform-remnant bearing flakes. Late blank reduction produces late stage flake and block shatter, and b i f a c i a l thinning flakes. Following Magne (1985) and Pokotylo (Magne and Pokotylo 1981), v a r i a b i l i t y i n the attributes defined above i s used to assign platform-remnant bearing flakes and shatter to sequential manufacturing stages. Magne and Pokotylo (1981) determined that three stages of platform-remnant bearing flake reduction can be cl a s s i f i e d with an accuracy rate of 74.32% using a 169 combination of four attributes: weight, dorsal scar count, platform scar count, and cortex cover. Only one of these attributes has mutually exclusive values for each reduction stage: weight. Magne's (1985) later study determined that either of two attributes, dorsal scar count or platform scar count, can be used to differentiate platform-remnant bearing flakes into manufacturing stages. This study uses dorsal scar count because the p i l o t study indicated that approximately 38% of the sample of platform remnant bearing flakes are missing a portion of the striking platform and i t i s not possible to determine the number of platform scars. During analysis of the p i l o t study data, an attempt was made to determine the corresponding values of maximum dimension for each manufacturing stage. Boxplots of the maximum dimension of platform-remnant bearing flakes arranged by the number of dorsal scars were produced separately for each s i t e (Figures 14 and 15). Flake sizes are plotted individually when the sample size f a l l s below four. There i s a weakly defined relationship between these two attributes, with maximum dimension often showing larger median and maximum values when the number of dorsal scars equals or exceeds three. This result supports Magne's (1985) conclusion that size, measured either by weight or maximum dimension, i s not a useful discriminating variable. Magne and Pokotylo (1981) also found that early and late shatter can be differentiated on the basis of size, measured here by maximm dimension, and the amount of cortex present. Two stages of flake and block shatter reduction were cl a s s i f i e d with an accuracy rate of 97.5% using two attributes: weight and cortex cover. Magne's (1985) later study indicated that the presence of cortex on the dorsal face of flake and block shatter i s associated with the early stage of core reduction. Therefore, this study 170 EeRj8 o CO co EeRj49 g 2 o or Ld co 2 23 Z I 20 40 60 MAXIMUM DIMENSION (mm) 8 0 | 100 to o CO - - c o CO EeRj56 | 2 cr LU m 2 Z - -CD-20 40 60 80 MAXIMUM DIMENSION (mm) | 100 >3 cn cr < u to -I < CO cc o a o cr LU cn 5 z to o to - - L _ - -DO-20 40 60 80 MAXIMUM DIMENSION (mm) 100 < to EeR]64 § 2 O cr LU CD 1 1 20 40 60 80 MAXIMUM 0IMENSION (mm) | 100 Figure 14. Boxplot of mavimiwi dimension by number of dorsal scars in Upper Hat Creek Valley pilot study sites. CO o co - - o EcRg2CC § 2 u . o cr LU co cn o CO < CO cr - -CD-— • • 20 40 60 MAXIMUM DIMENSION (mm) 80 _l 100 - -co-EcRg4C § 2 o cr co — •• • 20 40 60 MAXIMUM DIMENSION (mm) 80 _l 100 EcRg4D EdRg5 CO CJ CO cr o o u. o cr LU CO 2 ZJ 20 40 60 80 MAXIMUM 0IMENSI0N (mm) 100 co cr < CJ CO _) < CO cr o o u. o cr LU GO 2 Z >3 - -CL> 20 40 60 MAXIMUM 0IMENSI0N (mm) 80 _! 100 Figure 15. Boxplot of maximum dimension by number of dorsal Highland Valley pilot study sites. scars 172 uses the presence of cortex as the discriminating variable between the early and late stages of shatter production. Again, the r e l i a b i l i t y of maximum dimension as a stage discriminator was explored during the p i l o t study using boxplots of maximum dimension on flake and block shatter arranged by the presence/absence of cortex, for each s i t e (Figures 16 and 17). Although the majority of pieces of debitage i n a l l size categories do not display any cortex, the relationship between these two attributes appears to be stronger than between maximum dimension and dorsal scarring. In a l l sites except those lacking shatter without cortex, the median and maximum values of maximum dimension are larger for cases with cortex than for those without cortex. Although the size of the overlap i n this sample indicates i t s lack of r e l i a b i l i t y at this time, i t appears as though maximum dimension may be a more useful discriminator for shatter than for platform-remnant bearing flakes. This p o s s i b i l i t y should be investigated i n further experimental work. Finally, Magne (1985) ascertained that bipolar flakes result from the early stages of core reduction, and that b i f a c i a l thinning flakes result rem the later stages of blank reduction. Therefore, these two debitage types are counted and grouped with the appropriate stages without further discrimination. Tables 11 and 12 provide the discriminating values for three stages of platform-remnant bearing flake reduction and two stages of flake and block shatter reduction. Each a r t i f a c t i n these three classes i s assigned, according to the discriminating values, to one manufacturing stage. The frequencies of debitage assigned to sequential manufacturing stages by the methods described above are given i n Table 13. These counts also include 173 EeRj49 WITH CORTEX WITHOUT CORTEX - — [ _L 20 40 60 80 MAXIMUM DIMENSION (mm) EeRj 56 WITH CORTEX WITHOUT CORTEX 20 40 60 80 MAXIMUM DIMENSION (mm) EeRj 8 WITH CORTEX WITHOUT CORTEX - - r n — \ ---CD-J _ 20 40 60 80 MAXIMUM DIMENSION (mm) EeRj 64 WITH CORTEX WITHOUT CORTEX T 3 .** *** • 100 _J 100 100 20 40 60 MAXIMUM DIMENSION (mm) 80 100 Figure 16. Boxplot of m a y " * ™ ™ dimension by amount of cortex in Upper Hat Creek Valley pilot study sites. 174 WITH CORTEX EcRg2CC WITHOUT CORTEX --LD-20 40 60 MAXIMUM DIMENSION (mm) 80 100 EcRg4C WITH CORTEX WITHOUT CORTEX - - C D -- - 0 J 20 40 60 MAXIMUM DIMENSION (mm) 80 100 EcRg4D WITH CORTEX WITHOUT CORTEX --o 20 40 MAXIMUM 60 DIMENSION (mm) 80 100 EdRg5 WITH CORTEX WITHOUT CORTEX 20 40 MAXIMUM 60 DIMENSION (mm) 80 _ l 100 Figure 17. Boxplot of maximum dimension by amount of cortex i n Highland Valley p i l o t study sites . Table 11. Discriminating values for debitage classes on platform-remnant bearing flakes. Attribute Manufacturing stage Early Middle Late Dorsal Scar Count 0 - 1 2 > 3 Table 12. Discriminating values for debitage classes on shatter. Attribute Manufacturing Stage Early Late Cortex present absent Table 13. Frequencies of debitage assigned to manufacturing stages. Late Core/ Early Core Early Blank Late Blank Reduction Stage Reduction Stage Reduction Stage Site Early Shatter Early PRB Middle PRB Late FRB BTF Late Shatter Total EeRilO 18 7 111 295 0 782 1213 EeRj49 23 3 2 76 22 235 361 EeRj55 177 57 148 412 12 3299 4105 EeRj56 115 15 51 759 4 2254 3198 EeRj60 185 3 45 245 6 2730 3214 EeRj62 31 3 13 129 2 563 741 EeRj 8 43 3 7 30 0 393 476 EeRj20 3 1 8 32 3 224 271 EeRj42 139 17 138 785 14 4656 5749 EeRj64 29 2 4 54 0 227 316 EeRj100 5 1 3 23 1 113 146 EeRk52 4 1 6 29 5 85 130 EcRg2AA 175 15 124 603 3 2103 3023 EcRg2CC 10 2 8 70 10 169 269 EcRg4C 9 3 3 81 0 366 462 EcRg4J 2 0 4 13 0 21 40 EdRglA 4 1 9 45 2 70 131 EdRglB 5 3 33 140 2 522 705 EcRg4A 10 1 15 56 0 115 197 EcRg4B 1 3 23 48 4 129 208 EcRg4D 0 0 0 8 0 24 32 EcRg4E 1 1 5 3 0 8 18 EdRg5 0 0 0 26 4 75 105 EdRg6 50 2 24 147 0 662 885 PRB: Platform-remnant bearing flake BTF: B i f a c i a l thinning flake 177 bipolar flakes and b i f a c i a l thinning flakes. Results In order to enable direct comparisons among sites, and to provide a suitable data base for later multivariate techniques, the frequency counts of debitage classes are converted to percentage data and re-arranged according to the s i t e function c l a s s i f i c a t i o n made by the original s i t e investigators, presented i n Chapter IV (Table 14). With specific reference to assemblages from Upper Hat Creek Valley described as representative of an intensive occupation, Pokotylo (1978) also referred to them as local base camps for extractive a c t i v i t i e s . For comparative purposes therefore, these sites, plus the two sites previously c l a s s i f i e d as camp sites, are grouped with Highland Valley residential camps. The second group of sites contains Upper Hat Creek Valley sites previously interpreted as limited a c t i v i t y sites, either for hunting or plant procurement, grouped with what should be a similar set of sites previously interpreted as f i e l d camps from Highland Valley (Pokotylo 1978; Areas Associates 1983, 1986). The f i r s t group of intensive occupations/residential camps contains debitage from a broad range of manufacturing steps, with the exception of EcRg4D and EdRgS. These two sites were sampled and s i t e interpretation i s based p a r t i a l l y on estimated s i t e size. In addition, this group of sites contains, on the average, a higher proportion of debitage from the earliest stages of core reduction than do the second group of sites, limited a c t i v i t y / f i e l d camps. The one s i t e identified as a station, EcRg4E, contains the highest combined percentage of early and middle debitage, indicating a technological strategy focused on core reduction and early blank reduction. 178 Table 14. Percentages of debitage assigned to manufacturing stages. Late Core/ Early Core Early Blank Late Blank Reduction Stage Reduction stage Reduction Stage Site Early Early Middle Late Late Shatter PRB PRB PRB BTF Shatter Total intensive occupation/residential camps EeRj42 2.4 .3 2.4 13.6 .2 81.1 100.0 EeRj56 3.6 .5 1.6 23.7 .1 70.5 100.0 EeRj60 5.8 .1 .4 7.6 .2 85.9 100.0 EeRj62 4.2 .4 1.8 17.4 .3 75.9 100.0 EeRj64 9.2 .6 1.3 17.1 .0 71.8 100.0 EcRgZAA 5.8 .5 4.1 19.9 .1 69.6 100.0 EcRg2CC 3.7 .7 2.9 26.0 3.7 63.0 100.0 EcRg4C 21.4 .6 .6 17.5 .0 59.9 100.0 EcRg4B .5 1.4 11.1 23.1 1.9 62.0 100.0 EdRgS .0 .0 .0 24.8 3.8 71.4 100.0 limited a c t i v i t y / f i e l d camp EeRilO 1.5 .6 9.2 24.3 .0 64.4 100.0 EeRj 8 9.0 .6 1.5 6.3 .0 82.6 100.0 EeRj20 1.1 .3 2.9 11.8 1.1 82.8 100.0 EeRj55 4.3 1.4 3.6 10.0 .3 80.4 100.0 EeRk52 3.1 .8 4.6 22.3 3.8 65.4 100.0 EcRg4J 5.0 .0 10.0 32.5 .0 52.5 100.0 EdRglA 3.1 .9 6.9 34.4 1.5 53.2 100.0 EdRglB .7 .4 4.7 19.9 .3 74.0 100.0 EcRg4A 5.0 .5 7.6 28.4 .0 58.5 100.0 EcRg4D .0 .0 .0 33.3 .0 66.7 100.0 EdRg6 5.6 .2 2.7 16.6 .0 74.9 100.0 station EcRg4E 5.5 5.5 27.8 16.7 .0 44.5 100.0 no previous c l a s s i f i c a t i o n EeRj49 6.4 .8 .6 21.1 6.1 71.1 100.0 EeRj100 3.4 .7 2.1 15.8 .7 77.3 100.0 PRB: Platform-remnant bearing flake BTF: B i f a c i a l thinning flake 179 The only s i t e with previously unanalyzed debitage, EeRjlOO, i s very similar i n both frequency and percentage of stage-classified debitage to EeRk52, cl a s s i f i e d as a limited a c t i v i t y s i t e . Mthough the majority of sites i n Upper Hat Creek Valley do not have a previously assigned s i t e type, the interpretation of the relative importance of manufacturing stages present agrees with those presented by other researchers (Chapter TV). The single exception i s EeRj 49, earl i e r presented as a s i t e with emphasis on the early manufacturing stages; this analysis includes artifacts not available to the original researcher. In order to determine whether the percentage of debitage classes assigned to manufacturing can be used to differentiate among the functional s i t e classes, a series of Mann-Whitney tests was applied to the percentage data presented i n Table 14, after stations (1 site) and previously unanalyzed sites (2 sites) were deleted from the sample. None of the debitage classes discriminates among the s i t e classes at the pre-selected s t a t i s t i c a l level of 0.05, although middle stage platform-remnant bearing flakes produce a Mann-Whitney s t a t i s t i c which i s nearly significant (Table 15). The best discriminators i n terms of debitage classes are: early stage shatter, middle stage platform-remnant bearing flakes, and b i f a c i a l thinning flakes; each of these derives from a different manufacturing stage. The worst discriminators are: early stage platform-remnant bearing flakes, late stage platform-remnant bearing flakes, and late stage shatter, again each from a different manufacturing stage. A second series of Mann-Whitney tests was run on the three generalized manufacturing stages themselves, after combining the debitage classes assigned to each stage i n Table 14. Again, the results are not s t a t i s t i c a l l y 180 Table 15. Mann-Whitney two-sample tests on debitage class percentages grouped by previous s i t e c l a s s i f i c a t i o n . Debitage Class Rank Sum Mann-Whitney Probability S i t e l Site2 S t a t i s t i c Early shatter 120.5 110.5 65.5 0.459 Early PRB 109.5 121.5 54.5 0.972 Middle PRB 84.0 147.0 29.0 0.067 Late PRB 101.5 129.5 46.5 0.549 BTF 122.5 108.5 67.5 0.364 Late Shatter 116.0 115.0 61.0 0.673 PRB: Platform-remnant bearing flake BTF: B i f a c i a l thinning flake S i t e l : Residential camp/intensive occupation Site2: F i e l d camp/limited a c t i v i t y Table 16. Mann-Whitney two-sample tests on manufacturing stages grouped by previous s i t e classification. Manufacturing Rank Sum Mann-Whitney Probability Stage S i t e l Site2 S t a t i s t i c Early core reduction 115.5 115.5 60.5 0.698 Late core/early blank reduction 94.0 137.0 39.0 0.260 Late blank reduction 122.0 109.0 67.0 0.398 S i t e l : Residential camp/intensive occupation Site2: F i e l d camp/limited a c t i v i t y 181 significant, although the probability levels associated with the respective Mann-Whitney s t a t i s t i c s are much lower, indicating a better discrimination between settlement types when debitage i s placed into three previously manufacturing stages rather than six debitage classes (Table 16). In relation to manufacturing stages, the probability that the sites represent samples drawn from different populations i s substantially lower than the probability associated with the worst discriminators i n debitage classes. In fact, the probability levels associated with the three manufacturing stages approximates the median values between probability levels for the best and worst discriminators. In view of this result, and i n order to incorporate a l l debitage data, further analyses using debitage partitioned into manufacturing stages w i l l use the three generalized manufacturing stages devised by Magne (1985). Generally, these results support Binford's (1980) model that residential camps w i l l contain debitage from a wide range of manufacturing stages, while stations w i l l contain debitage from a more limited range of manufacturing stages. In addition, these results support Chatter's (1987) model which predicts differences i n emphasis on reduction stages between residential camps and f i e l d camps. However, there are some exceptions: for example, EcRg4B, c l a s s i f i e d as a residential camp, contains less than 2% early stage debitage, and EdRgS, also a residential camp contains no early stage debitage. The latter result may be due to s i t e sampling, as already discussed. According to Binford's (1980) model, f i e l d camps and stations should contain debitage from only one or two stages of tool manufacture, rejuvenation and resharpening. The results provided i n Table 14 indicate that this prediction i s not supported. In fact, the v a r i a b i l i t y i n 182 proportions of individual manufacturing stages between settlement types i s considerably more subtle than Binford's (1980) and Chatters' (1987) models of assemblage structure predict. Finally, the variation i n emphasis on reduction stages, as evidenced i n this study, relates primarily to the late core/early blank reduction stage and to the late blank reduction stage. Sites interpreted as intensive occupations or residential camps contain a higher percentage of debitage from the late blank reduction stage, while sites interpreted as limited a c t i v i t y or f i e l d camps contain a higher percentage of debris from the late core/early blank reduction stage. This result may indicate a greater emphasis i n residential camps on f i n a l tool finishing, tool maintenance and rejuvenation. Microlithic Debitage  Introduction A salient feature of microcore technology i s the staging reguired: several different flintknapping procedures must be followed i n a predetermined order. Each stage of microcore preparation and reduction can considered a separate task, to be completed i n one location or i n several. Thus, the way i n which the various stages are divided up spatially and temporally has important implications for the potential organization of microcore technology. One important feature of the model being tested i n this research associates the preparation of microcores with an emphasis on the use of curated technology. Under these conditions, i t i s l i k e l y that microcores were prepared at spring and summer residence camps. Archaeological indications of microcore preparation should of unsuccessfully prepared 183 micxocores and microcore fragments, microblades from the earliest stages of ridge preparation, and microcore preparation flakes. A second feature of the model associates the production, use and discard of microblades with spring and summer residence camps. Archaeological evidence of this type of technological organization would consist of the presence of microcore rejuvenation flakes, and both used and unused microblades from a complete range of production stages i n residential sites. There i s , as yet, no conclusive evidence for hafted microblades i n the southern Interior Plateau. Although Loy (1986) interpreted several of the Highland valley microblades as being hafted, he did not discuss the evidence for this interpretation, and the microblades i n question do not display the same type of wear patterns. I f some microblades were selected for hafting into a curated, composite tool, then a less complete microblade production sequence would be present at the locus of tool manufacture, consisting of only those blades unsuitable for hafting because of irregularities of size and shape (Hofrnan 1987). In addition, broken blades removed from the composite tool might be located at the s i t e where la s t used. Current research (Kelly 1984) indicates that microblade manufacture can only be differentiated from biface manufacture at the f i n a l product stage. That i s , i f microblades, microcore preparation flakes, microcore rejuvenation flakes or microcores are not present, i t i s not possible to determine from an examination of flakes and shatter whether or not a microcore was prepared or microblades manufactured at an archaeological s i t e . Therefore, the analyses which follow are confined entirely to an examination of microcores and microblades from recent experimental replication and archaeological sites. 184 Implications of Previous Research The only available report (Kelly 1984) of experimental inicroblade manufacture describes an attempt to discriminate between intentionally and fortuitously produced microblades, and the resultant debitage. Kelly's study has important implications for this research because the technique of microblade manufacture used i s similar to that suggested for the archaeological specimens for the B r i t i s h Columbia Interior Plateau (Sanger 1970b), i n that platform preparation i s minimal and flakes or nodules are used as blanks. Kelly (1984) reported on the experimental replication of both microblades and bifaces beginning with selection of raw material, heat treatment, blank reduction and ending with core exhaustion or tool breakage. Raw material was a l o c a l l y available variegated chalcedony, and tools consisted of basalt hammerstones for percussion flaking, and copper tipped pressure flakers. Indirect percussion with a vise and anvil was not as successful as the pressure or direct free-hand percussion techniques, because of the small size of the core. Debitage was not collected u n t i l the experiment was finished. According to Kelly (1984), the most suitable core blanks were blocky flakes with one semi-flat, s l i g h t l y concave surface for the platform. In addition, the core had a relatively f l a t surface perpendicular to that selected for the platform and was at least 3cm long. Minimal platform preparation ensured a platform angle of s l i g h t l y less than 90-degrees which f a c i l i t a t e s the contact between the flaker and core. On flakes without this natural platform, platform preparation consisted of the removal of small flakes transverse to the edge where microblades were subsequently detached, or l a t e r a l l y across the platform surface. The platform edge was ground to 185 strengthen i t and to remove any overhang. The next step i n microcore manufacture was the preparation of a face by the removal of "primary" microblades which l e f t longitudinal ridges to guide later microblade detachment. Kelly (1984:43) described the sequence of microblade removal and microcore rejuvenation: The microblades were detached by either direct free-hand percussion or pressure. For each technique/ the core was held i n a leather-covered hand and the percussor or pressure tool was held i n the opposite hand. The point of contact was at the juncture of the a r r i s and the platform/ or between the two arrises. The former position produced microblades that are triangular i n cross-section whereas the latt e r position produced microblades with a trapezoidal cross-section. Six percussion and six pressure microblade cores were experimentally manufactured resulting i n 85-90 microblades from each... .After a series of microblades was removed from one face, the edge was abraded and microblade production continued. In some cases, i t was necessary to prepare a clean platform surface by removing multiple small flakes, as previously described. I f the knapper encountered too many flaws on the face or platform/ a new face or platform was prepared wherever a clean surface was available. This process continued u n t i l either the core was too small or too flawed to allow further reduction. Both the microcores and microblades produced by the pressure technique are indistinguishable from those produced by direct free-hand percussion. In addition, the replication of both microcore and b i f a c i a l core technology produced b i f a c i a l thinning flakes and primary blades. Although Kelly's research objective was not the identification of manufacturing stages within microcore technology/ she observed that primary blades, which are detached after platform preparation i n order to provide a fluted face for intentional microblade removal, are linear flakes with some cortex cover and a single longitudinal dorsal a r r i s . The early stage of microblade removal apparently produces shorter, narrower blades with shorter, more numerous dorsal scars than the later stages of microblade removal. 186 Experimental Replication The following section describes the results of an additional attempt to replicate microblade production by knapper Alexander Mackie, who was coached by Professor David Pokotylo at the Laboratory of Archaeology, University of Br i t i s h Columbia, i n 1987. The raw material was obsidian from Glass Buttes, Oregon, and the tools consisted of a cobble hammer stone, an antler hammer, a wooden vise, and an antler punch. Three microcores were prepared but only one successfully produced microblades. The knapper's comments were recorded, and a l l debitage, including microblades, was collected i n bags labelled with the corresponding number and stage: 1. Selection of Core Blank The knapper selected a round or ovoid cortex-covered cobble which could be s p l i t into two similarly-sized pieces, each with an interior cortex-free surface to serve as a platform. 2. Platform Preparation The knapper used a hard hammer to roughen the surface of the platform with abrasion. 3. Preparation of Ridges The desired result of this step was the production of several long, straight, p a r a l l e l ridges on the face of the core which would guide future microblade removal. This produced several blade-like flakes which are the equivalent of Kelly's (1984) primary blades, discussed above. This step appears to be the most d i f f i c u l t to complete successfully and on Core #2, discarded after this step, the knapper f a i l e d to produce the necessary number and quality of ridges. The knapper switched to the antler hammer i n order to detach several primary blades. Between blows, the hammer was used 187 to roughen the platform edge, and remove any overhang. The knapper oonmented that the deal platform i s situated at a 90-degree angle to the fluted face, and that a b i f a c i a l technique was used to remove primary blades. 4. Microblade Removal For t h i s step, the knapper placed the prepared microcore into a vise, and switched to use of an antler punch with hammer stone. Platform roughening was continued between punches. The knapper placed the punch direc t l y over a ridge and attempted to remove a blade with a direct right-angle push. On Core #1, the knapper switched to the indirect punch method after removal of the f i r s t blade created an undesirable set of ridges; however, the punch was placed too far from the platform edge and the flake detached was too wide and removed most of the guiding ridges. Core #3, Which had three guiding ridges, slipped i n the vice and a hinge flake was detached. Core rejuvenation, described below, was carried out on both cores before further microblade removal was attempted. The knapper successfully removed nineteen microblades, i n four separate steps, from Core #3 after the f i r s t rejuvenation. After each group of microblades was removed, the platform edge was roughened with the cobble hammer stone. 5. Microcore Rejuvenation The knapper attempted rejuvenation on Core #1 after removal of one microblae, and on Core #3 after an unsuccessful attempt to remove the f i r s t microblade, discussed above, and after the f i r s t nineteen blades were removed. The goal of rejuvenation i n both cases was to re-establish the pa r a l l e l guiding ridges which are essential for successful microblade production. Frequently, the ridges were too far apart to produce microblades, and the knapper attempted to create additional ridges. Another 188 frequent problem was the appearance of an overhang on the platform edge after the removal of either primary blades or microblades. Removal of this overhang often resulted i n the removal of one or more guiding ridges. Finally, the knapper noted that i f the original platform was uneven, consistent removal of evenly-sized blades was d i f f i c u l t . Rejuvenation of Core #1 ended when removal of a large hinge flake reduced the platform to an unworkable length. Rejuvenation of Core #3 produced a workable fluted face, and two microblades were removed. Again, the guiding ridges were too far apart, and the platform collapsed. Once more, the knapper rejuvenated the ridges and detached two more microblades. The ridges were again removed. Further rejuvenation was required, and one additional microblade was produced. According to the knapper, the f i n a l attempt to rejuvenate the ridges f a i l e d because either too much force was used, the punch was placed too far back on the platform, or the platform angle was not quite 90-degrees. The twenty-four successfully produced microblades from Core #3, and the exhausted core are illustrated i n Figure 18. Summary of Experimental Microblade Production Microcore preparation consists of a series of procedures requiring a high level of knapping s k i l l , a variety of tools, and a high level of concentration. Errors are not easy to correct, and fa i l u r e to produce the c r i t i c a l number of ridge scars of the desired length and distance apart results i n a basically unusable piece of raw material. The most important features on a cobble to be reduced into a microcore are: a smooth interior surface free of impurities, and a minimum overall size of 5 to 6 cm. The c r i t i c a l features produced by knapping on a successful microcore are: at 189 Figure 18. Experimentally produced microblades and exhausted core. 190 least four or fi v e straight, p a r a l l e l ridges running almost the entire length of the d i s t a l face, and a platform striking angle of approximately 90 degrees. Consistent problems included too great a distance between the guiding ridges, too much force placed on the punch, too much distance between the punch t i p and the platform edge, and a fail u r e to maintain the platform striking angle at 90 degrees. A useful addition to this set of experimental replications would be to vary the raw material types and to vary the blade production techniques. Only indirect percussion was used by Alexander Mackie,. whereas Kelly's (1984) most successful methods were direct pressure and free-hand percussion. Selection of Attributes Kelly (1984) described two generalized manufacturing stages i n microblade production, separated by platform rejuvenation. Primary microblades, or those detached during ridge preparation, display remnants of cortex and a single dorsal scar. In addition, as mentioned above, Kelly (1984) suggested that early stage microblades are shorter, and narrower, with more dorsal or ridge scars than late stage microblades. Sanger (1968) proposed that microblade length w i l l not change, but that width w i l l decrease throughout the reduction sequence. As well, he suggested that the ventral striking platform angle w i l l probably not alter, and that a l l striking platforms w i l l be ground. Other studies of microcore technology indicate that striking platform width i s greater i n microblades produced during the early stages of core reduction (Hofman 1987). Based on the studies of microcore technology mentioned above, the 191 following attributes were selected for measurement: blade length (A), blae width (B), blade thickness (C), number of ridge scars (D), ventral striking platform angle (E), amount of cortex (F), number of platform preparation scars (6), and striking platform width (H) (see Figure 19). The blade length, blade width, blade thickness and striking platform width were measured i n millimetres. The striking platform was examined for evidence of grinding and the number of scars counted. The number of ridge scars were counted on each microblade. ventral striking platform angle was measured i n five degree intervals, i n the same way that this angle was measured on platform-remnant bearing flakes. The amount of cortex present was measured i n 25% increments from 0% to 100%. The experimental replication described above produced twenty-four complete microblades, i n two stages, early and late, separated by platform rejuvenation. The early stage i s defined as the production of a l l microblades up to the f i r s t attempt at microcore rejuvenation; the late stage i s defined as the production of a l l microblades after the rejuvenation. Table 17 provides measurements for a l l attributes on each microblade. There i s no cortex present i n either the experimental sample or the archaeological sample, and this attribute was deleted from further analysis. In addition, no evidence of grinding on the striking platfor was noted and only the number of platform scars was retained as a measure of this attribute. In order to determine i f the values for these seven attributes d i f f e r , i n the predicted manner, between the two stages, measures of central tendency and dispersion were calculated separately for each attribute: minimum, maximum, median, mean and standard deviation. Tables 18 and 19 Figure 19. Attributes measured on experimentally produced microblades. 193 Table 17. Attribute values of experimentally produced microblades i n order of detachment from microcore. Stage Order Length Width Thick- Number Number Platform Platform (Mm) (Mm) ness of of Angle Width (Mm) Ridge Scars Platform Scars (Degrees) (Mm) 1 1 21.5 4.2 1.0 2 1 90 0.9 1 2 28.8 10.1 2.5 3 1 90 1.0 1 3 30.0 9.0 3.0 3 1 90 0.8 1 4 31.6 8.9 2.9 4 1 75 1.6 1 5 35.5 9.7 3.0 6 2 85 1.2 1 6 29.0 8.8 1.6 3 1 85 1.3 1 7 23.5 6.0 1.5 3 1 85 1.7 1 8 23.8 5.4 2.0 2 1 85 1.0 1 9 34.9 8.8 2.1 4 1 90 1.6 l 10 35.0 9.9 3.0 3 2 85 2.3 1 11 31.0 8.8 3.1 1 1 85 1.8 1 12 36.6 9.6 4.4 4 2 80 2.2 l 13 29.5 11.5 3.5 1 1 85 1.6 l 14 35.0 9.6 2.3 3 2 85 1.5 1 15 32.7 10.4 2.2 3 2 85 2.0 l 16 36.2 7.9 2.5 4 1 80 1.0 l 17 30.0 7.4 2.0 3 2 85 1.0 1 18 34.1 10.5 2.0 2 2 85 2.0 1 19 31.5 6.4 1.6 2 1 90 1.2 2 20 34.0 9.9 3.0 4 3 90 1.7 2 21 36.0 10.9 2.6 6 1 90 1.4 2 22 36.5 12.3 4.0 4 1 80 2.4 2 23 36.4 13.4 3.0 3 2 85 1.6 2 24 27.8 6.6 2.0 2 1 85 2.0 194 Table 18. Measures of central tendency and dispersion for early stage of microblade production (N=19). Attribute Minimum Maximum Median Mean Standard Deviation Length (Mm) 21.5 36.6 31.5 31.1 4.4 Width (Mm) 4.2 11.5 8.9 8.6 1.9 Thickness (Mm) 1.0 4.4 2.3 2.4 0.8 Number of Ridge Scars 1.0 6.0 3.0 2.9 1.2 Number of Platform Scars 1.0 2.0 1.0 1.4 0.5 Platform Angle (Degrees) 75.0 90.0 85.0 Platform Width (Mm) 0.8 2.3 1.5 1.5 0.5 Table 19. Measures of central tendency and dispersion for late stage of microblade production (N=5). Attribute Minimum Maximum Median Mean Standard Deviation Length (Mm) 27.8 36.5 36.0 34.1 3.7 Width (Mm) 6.6 13.4 10.9 10.6 2.6 Thickness (Mm) 2.0 4.0 3.0 2.9 0.7 N mber of Ridge Scars 2.0 6.0 4.0 3.8 1.5 Number of Platform Scars 1.0 3.0 1.0 1.6 0.9 Platform Angle (Degrees) 80.0 90.0 85.0 Platform Width (Mm) 1.4 2.4 1.7 1.8 0.4 195 provide these s t a t i s t i c s for the early and late stages of microblade production. Examination of these tables indicates that microblade length, width, and thickness are greater i n the late stage of production, as Kelly (1984), Hofrnan (1987) and Arnold (1987) predict. In addition, the striking platform angle does not d i f f e r greatly, as predicted by Sanger (1968). However, two attributes, number of ridge scars and platform width, do not vary i n the manner predicted by Kelly (1984) and Hofrnan (1987); both display greater values i n the later stage. Thus, an examination of these descriptive s t a t i s t i c s indicates that there are between-stage differences i n only some of the selected attributes. In order to determine i f the microblades i n the two production stages are derived from similar or different populations, the empirical v a l i d i t y of the two groups (stages) i s tested by applying a series of Mann-Whitney two-sample tests on the raw data matrix provided i n Table 17. This nonparametric test examines two samples to determine i f their respective populations have different population distributions, by ranking the scores on each attribute i n a composite distribution including both groups. Then, the U-statistic i s calculated, based on the rank sum which has a corresponding probability, for each attribute (Thomas 1976; Downie and Heath 1974). A prior significance level of 0.05 i s selected. Table 20 provides the ranked sums, U-statistic and probability level for each attribute i n the experimental sample. None of the attributes are significant at the 0.05 level of probability, although the value for blade width i s 0.055, very close to the pre-selected significance level. Professor R.G. Matson (personal ccmnunication 1990) suggested that this result be checked by hand because the SYSTAT program used to analyze the data uses approximations, instead of tables, which rely 196 Table 20. Mann-Whitney two-sample tests on attributes measured on experimental microblades grouped by stage. Attribute Rank Sum Mann-Whitney Probability Early Late S t a t i s t i c Length 217.00 83.00 27.00 0.145 Width 210.50 89.50 20.50 0.055 Thickness 220.50 79.50 30.50 0.224 Number of Ridge Scars 220.50 79.50 30.50 0.208 Number of Platform Scars 232.50 67.50 42.50 0.675 Platform Angle 233.00 67.00 43.00 0.723 Platform Width 216.00 84.00 26.00 0.124 Table 21. Experimental microblade attribute principal components analysis factor loadings. Attribute Factor 1 Factor 2 Factor 3 Factor 4 Length (0.884) 0.243 -0.078 0.048 Width (0.838) 0.087 0.103 0.410 Thickness (0.819) -0.120 -0.120 0.264 Number of Ridge Scars 0.486 (0.567) -0.551 -0.243 Number of Platform Scars 0.520 0.372 (0.562) 0.459 Platform Angle -0.437 (0.656) 0.398 0.396 Platform Width (0.640) -0.479 0.358 -0.113 Percent of t o t a l variance 46.586 17.504 13.390 9.707 explained NOTE: Bracketed () values indicate significant loadings discussed i n text. 197 on re l a t i v e l y equal sample sizes. The hand-calculated U-statistic i s 20.0, which i s significant at a probability level of 0.05. Thus, the results of this test indicate that blade width i s the only attribute measured which differentiate among the two stages of microblade production. Although the Mann-Whitney tests indicate that only one of the attributes, d i f f e r s significantly between the two stages, there i s s t i l l a p o s s i b i l i t y that a combination of attributes may distinguish between the stages. Thus a series of multivariate tests i s conducted on the experimental microblade data: principal components analysis, multiple discriminant analysis and cluster analysis. The f i r s t step i n the multivariate analysis i s the selection of an attribute set which provides the greatest amount of technological information about the microblade sample and contains the least amount of redundant information. Principal components analysis i s used i n archaeological data analysis to determine i f artifacts, attributes or s i t e characteristics covary, and the extent of their correlation with one another (Sherman 1988). Principal components analysis defines underlying patterns of variation common to the group of variables being considered (Shennan 1988). In this study, principal components analysis (Wilkinson 1988) i s applied to the data matrix given i n Table 17, to determine the nature of the relationships among the attributes selected to measure microblade manufacturing stages, and to reduce the set of attributes to a smaller number for further clustering and multiple discriminant analysis. The analysis of seven microblade attributes results i n a 4-factor solution (Table 21). Factor 1, which accounts for 46.586% of the sample variance, has four highly-loading attributes: blade length, width, thickness and striking 198 platform width. These attributes a l l increase with production stage, and relate to the overall dimensions of the microblades. The second factor accounts for 17.504% of variance and i s dctninated by number of ridge scars and striking platform angle. These two attributes are c r i t i c a l to the successful production of microblades, and were, therefore, under tight control with l i t t l e v a r i a b i l i t y being desirable. Factor 3 i s a specific factor primarily defined by a single attribute - number of platform scars -and contributing to 13.39% of the variance. Again, this attribute was controlled by the knapper as he continually roughened the platform. Factor 4 contains no highly loading attributes, and contributes to 9.707% of the variance. The results of the principal components analysis provide the c r i t e r i a for selecting those attributes to be used i n further multivariate analysis: length, width, thickness and platform depth, the highly loading attributes from Factor 1. The second step i n this analysis i s a cluster analysis of the attributes recorded for the experimental sample. Cluster analysis has been used most successfully i n archaeological data analysis to group either cases (Q-mode analysis) or attributes (R-mode analysis) together based on s i m i l a r i t i e s among the multivariate data sets (Matson and True 1974; Pokotylo 1978; Greaves 1982: Magne 1985). The clustering, or group formation, i s performed on a matrix of s i m i l a r i t i e s or distances between the objects being studied. This study uses the Euclidean distance measure, the one most commonly used with interval or ratio level data. Ward's error sum of squares method i s used on a standardized data matrix of the four attributes previously selected after factor analysis (length, width, thickness and platform depth) to produce the groups (Matson and True 1974). The best solution i s a two-199 cluster grouping at the f a i r l y high similarity level of 2.864 (Figure 20), shown along the right margin. Numbers along the l e f t margin of the diagram indicate the stage of production (l=early, 2=late), and the order of detachment from the core, i n brackets. Each cluster contains microblades from both stages (Table 22). Cluster 1 contains 7 microblades from the f i r s t half of the early stage, 2 microblades from near the end of the second half of the early stage, and the f i n a l microblade produced i n the late stage. Two of these latt e r three are the f i r s t and l a s t microblades to be produced after micxocore rejuvenation. Cluster 2 contains 3 microblades from the f i r s t half of the early stage, 8 microblades from the second half of the f i r s t stage, and 4 microblades from the middle part of the late stage. Table 23 provides measures of central tendency and dispersion for the four microblade attributes used i n the cluster analysis. Cluster 2 microblades are longer, wider, thicker, and have a wider striking platform than cluster 1 microblades. Mann-Whitney two-sample tests are conducted on the four attributes to see i f the differences between the two clusters are significant. Table 24 provides the results, and a l l attributes have a significance level of less than 0.05, indicating s t a t i s t i c a l l y significant differences between the clusters. I f the experimental sample size, both microcores and microblades, had been larger, these figures could have used as discriminating values from determination of production stage on archaeological assemblages. Examination of the composition of the clusters also provides some preliminary indications about how microcore rejuvenation may affect microblade attributes, particularly those relating to size. The microblades produced i n the f i r s t half of the early stage are smaller than those 200 1(1) 1(8) 1(7) 2(24) 1(19) 1(17) 1(6) 1(2) 1(3) 1(16) 1(5) 2(21) 1(14) 1(9) 2(23) 1(18) 1(15) 1(10) 1(11) 1(4) 2(20) 1(13) 2(22) 1(12) 0.0 | 10.000 Figure 20. Ward's cluster analysis of experimental microblades. 201 Table 22. Classification of experimental microblades by clusters (N=24). Stage Cluster 1 Cluster 2 Early 8 11 (42.11%) (57.89%) Late 1 4 (20.0%) (80.0%) Table 23. Measures of central tendency and dispersion for two clusters of experimental microblades (N=24). Attribute Cluster Minimum Maximum Mean Standard Deviation Length (Mm) 1 21.5 31.5 27.3 3.5 2 29.5 36.6 34.3 2.2 Width (Mm) 1 4.2 10.1 7.1 1.9 2 7.9 13.4 10.1 1.4 Thickness (Mm) 1 1.0 3.0 1.9 0.6 2 2.0 4.4 2.9 0.7 Platform Width (Mm) 1 0.8 2.0 1.2 0.4 2 1.0 2.4 1.7 0.4 202 Table 24. Mann-Whitney two-sample tests on attributes measured on experimental microblades grouped by cluster. Attribute Rank Sum Cluster 1 Cluster 2 U S t a t i s t i c Probability Length 49.00 251.00 4.0 o . o o Width 60.00 240.00 15.00 0.02 Thickness 60.00 240.00 15.00 0.02 Platform Width 69.50 230.50 24.50 0.01 Table 25. Classification probabilities for experimental microblade sample (N=24). Stage Stage Early Late Early 73.7% 26.3% Late 20.0% 80.0% 203 produced i n the second half of the early stage. However, the l a s t microblade to be produced before microcore rejuvenation and the f i r s t microblade to be produced after microcore rejuvenation are small, and resemble those produced at the beginning of microcore reduction. The other three blades produced after successful microcore rejuvenation are large, and resemble those produced before rejuvenation. The f i n a l stage i n the multivariate analyses i s multiple discriminant analysis, used to cla s s i f y cases into groups (dependent categorical variables) on the base of v a r i a b i l i t y i n selected attributes (independent metric variables) (Wilkinson 1988). Discriminant analysis derives a linear combination of the metric measurements for two or more variables that best discriminates among the previously defined groups. This study applies the simultaneous method for computing the discriminant function, i n which a l l the independent variables are considered concurrently (Wilkinson 1988). The raw data matrix specifies a grouping variable, stage, and the four discriminating variables selected by factor analysis: length, thickness, width, and platform width. Table 25 provides the results. Seventy-three per cent of the early stage microblades and eighty percent of the late stage microblades are c l a s s i f i e d correctly, figures well above the f i f t y percent prior probability of accurate classification. Results Although these results indicate that microblades from the f i r s t half of the early stage are differentiated from those i n the second half of the early stage, there i s no clear clustering of early stage microblades and late stage microblades. This preliminary experimental research does indicate 204 seme differentiation among blades resulting from early and late stages of production. The attributes that appear to vary i n the most predictable manner are: length, width, thickness and platform width, a l l related to overall size. However, as discussed above, microcore platform rejuvenation appears to have substantial effects on the v a r i a b i l i t y i n these attributes. Microblades produced immediately before and after are the same size as those produced i n the f i r s t half of the early stage of production. As well, these results do not conform to the usual expectations for l i t h i c reduction, where the products of late stage reduction tend to be smaller than those from the early stage of reduction. This i s probably due to the unique outline shape of the wedge-shaped microcore, which i s deeper i n the middle than at the d i s t a l end where blades would be removed i n the early stage. Although number of ridge scars, number of platform scars and ventral platform flaking angle also vary, the v a r i a b i l i t y i s not direct and may also be related to platform and f l u t i n g face rejuvenation. As the knapper commented, control over number of ridge scars and platform angle i s crucial for the successful production of microblades. The knapper attempted to maintain a flaking angle of 90-degrees, and sets of two p a r a l l e l ridge scars. In addition, the striking platform was continually roughened between blows, ensuring a minimal number of platform scars on each successful blade. These preliminary results are promising i n that i t may be possible to discriminate between stages of microblade removal before and after platform and f l u t i n g face rejuvenation. As well, the high level of covariation of length, width and thickness of microblades indicates that i t may be possible to use only one of these attributes, i n addition to striking platform width, as successful stage discriminators. However, these results are based on 205 analysis of the products of the reduction of a single microcore by a r e l a t i v e l y inexperienced knapper. In addition, the effects of raw material type and the technique of microblade removal on v a r i a b i l i t y of the attributes are unknown. The former i s a particularly important factor for this study, because the archaeological samples are composed of basalt and chert, and the experimental sample i s obsidian. As well, other technological goals, such as production of microblades for hafting (Kelly 1984, Flenniken 1981) production for time efficiency (Arnold 1987), production for conservation of raw material (Arnold 1987), and production for trade (Hofrnan 1987) may have an effect on microblade attributes. Although the researchers cited above have hypothesized about the material implications of variation i n these technological goals, to date no experimental replication has been carried out to evaluate these hypotheses. Additional replicative experiments, u t i l i z i n g the several manufacturing techniques described above and carried out on various raw material types, and with different technological goals, are essential to the future construction of a manufacturing stage sequence for microlithic technology. Although the intent of this section was to develop a manufacturing stage typology for microblades, the experimental replicative data are insufficient for the determination of a reliable typology. Therefore, the archaeological sample of microblades i s not partitioned into manufacturing stages for the s i t e function portion of the data analysis. In addition, as already discussed, debitage from microcore preparation i s indistinguishable from debitage resulting from b i f a c i a l core preparation. However, some flakes contain remnants of the flut i n g face, and these can be c l a s s i f i e d as either microcore rejuvenation flakes or microcore preparation flakes (Areas 206 Associates 1983). Therefore, the c l a s s i f i c a t i o n of microlithic debitage for th i s study i s confined to the following: unused microblades, unused microcore preparation flakes, and unused microcore rejuvenation flakes. In addition, the experimental data are considered insufficient for the determination of the presence or absence of a complete manufacturing sequence. Manufacturing Typology for Microcores This typology i s primarily judgmental and i s included here as a f i r s t step i n the experimental reproduction of a microcore reduction trajectory. The c l a s s i f i c a t i o n i s based on consideration of several attributes of microblades, discussed above, and comments by the knapper. Mthough the experimental replication produced three microcores, the f i r s t two were shattered by attempts at ridge scar preparation and were not analyzed. The thi r d i s i l l u s t r a t e d i n Figure 17. In order to produce microblades successfully, the striking platform angle must approximate 90-degrees, the ridge scars must be straight, p a r a l l e l , and extend to the base of the core, and the top of the core must be at least 3 cm long and 2 cm wide (Kelly 1984). A schematic diagram of a successfully prepared microcore i s il l u s t r a t e d i n Figure 21. During analysis of the archaeological sample, each microcore and microcore fragment was measured, drawn to scale, and described i n terms of platform angle, number of ridge scars, and maximum dimensions. In addition, the condition of the basal edge was noted, i n order to assess whether the core had been placed i n a vise or other type of holding device. The following stage typology was produced (see Figure 22): PROXIMAL END STRIK PLATF ANGLf DISTAL END Figure 21. Microcore attributes. a b Figure 22. Microcore typology. 209 1. Blank (Figure 22-a) This i s the f i r s t step of microcore preparation. The basic outline i s wedge-shaped and the striking platform i s present, although the striking platform angle i s not 90-degrees, and there are no ridge scars. Cortex remains on the d i s t a l end, to be removed during f l u t i n g face preparation. There are no retouch scars on the basal edge, which may be cortex. 2. Reject (Figure 22-b) This stage refers to a microcore that has not been prepared successfully for microblade production. The basic outline s t i l l i s wedge-shaped. The d i s t a l end lacks ridge scars, and i s often marred by large hinge scars. Although the platform may s t i l l be a sufficient size, removal of the hinge scar and successful establishment of guiding ridges would reduce the platform too much. The basal edge has a few small step or feather scars. 3. Viable (Figure 22-c) A viable microcore retains a striking platform angle of approximately 90-degrees, a suitable set of guiding ridge scars, and a striking platform of suitable dimensions. The entire basal edge i s covered with step scars. 4. Exhausted (Figure 22-d) An exhausted microcore exhibits at least the remnants of successfully placed guiding ridges, a striking platform angle that i s considerably less than 90-degrees, and a striking platform that would be too small i f the striking platform angle were corrected. The basal edge i s severely scarred, polished or crushed. Tool Analysis 210 Introduction A major factor i n determining the type of archaeological s i t e i s the actual function of the tools used at the s i t e , and the range of uses i n relation to the number of morphological types. Mthough some morphological tool types display a strong correlation with a particular use or combination of uses, the majority of types were probably used i n a variety of tasks (Vaughan 1985). Therefore, a more complete interpretation of the a c t i v i t i e s carried out at each s i t e should be based on the results of use-wear analysis, i n conjunction with other types of archaeological evidence. Binford's (1980) model predicted that residential camps w i l l be the locus of the greatest number of a c t i v i t i e s involving tools; f i e l d camps w i l l be used for a smaller number of a c t i v i t i e s ; and stations w i l l contain the most limited range of a c t i v i t i e s . Chatters (1987) predicted that residential camps w i l l contain a more diverse assemblage than f i e l d camps, that i s , a wider variety of morphological tool types f u l f i l l i n g a large number of tasks. F i e l d camps should contain a smaller number of morphological tool types used for a variety of tasks because of the constraints placed on tool k i t size by high mobility. In order to incorporate these predictions into the differentiation among the various s i t e types, which i s the ultimate goal of this chapter, i t i s necessary to determine for each s i t e the range and relative amounts of a c t i v i t i e s performed with tools and the proportion of tools which are specialized or multi-purpose. The goal of the tool analysis i n this study i s determination of the range of uses associated with the various tool types found i n the study sites. In 211 order to accomplish this, this study uses the Employable unit (ED) concept as developed by Knudson (1983) i n which each separate modified portion of a tool i s coded individually. Knudson (1983:10) defined the ED as that implement segment or portion (continuous edge or projection) deemed appropriate for use i n performing a specific task, e.g. cutting, scraping, perforating, d r i l l i n g , chopping. The unit i s identified by deliberate retouch and/or post-production u t i l i z a t i o n modification, and i t s boundaries are defined subject to the analyst's own conception of "habitual use". This approach enables the determination of the f u l l range of uses for each tool, and the identification of those tools which were used for more than one purpose. No differentiation was made between intentional retouch and u t i l i z a t i o n modification. In order to be included, modification must be continuous for a minimum distance of 2 mm or two negative flake scars. Figure 23 i l l u s t r a t e s an Employable Unit on a microblade. Attribute Selection and Recording Vaues for each attribute are derived from experimental replication (Odell 1981; Odell and Odell-Vereecken 1980) which has particular relevance for this study because local Cache Creek basalt was used to make the tools. In addition, these studies reported a success rate of approximately 70% for a c t i v i t y and 61% to 68% for worked material, using low-power magnification (X10). Attributes were selected for their usefulness i n indicating the motion used (e.g. scraping, cutting) and the nature of the worked material (e.g. soft, medium, hard). Table 26 provides the attributes selected and values recorded for modification patterns which might be expected to occur on the tools analyzed. In addition to the attributes l i s t e d , two other attributes were i n i t i a l l y recorded: extent of modification and thickness of the EU measured 2 mm Figure 23. Employable unit on a microblade. Table 26. Attributes and values recorded on Employable Units. 213 Attribute Value Position of EU Surface Tip/distal end Edge Base/proximal end Notch Placement of modification Uhifacial B i f a c i a l Around t i p Direction of modification Oblique Perpendicular to edge Parallel to edge Parallel to long axis Perpendicular to long axis Multidirectional Unknown Type of modification Feather scars Hinge and step scars Striations Polish Crushing/pitting Snapping Polish and feather scars Feather scars on one side, hinge and step scars on opposite side Size of scars Small (visible with hand lens X 0) Large (visible to naked eye) Small and large Not applicable Angle of EU Acute (less than 45 degrees) Obtuse (more than 90 degrees) Steep (between 45 and 90 degrees) Not applicable 214 from the tool edge. During the analysis of the p i l o t study data, these attributes were found to be redundant for determining the function of the ED, and were deleted from the remainder of the study. The surfaces and edges of a l l microblades, microblade fragments, and formed tools were examined under low-power magnification (X10). A l l debitage which displayed some indication of post-production modification, v i s i b l e without magnification, was examined under low-power magnification. Each modified segment of the a r t i f a c t was coded separately. For example, a flake with two modified edges and a modified t i p would be coded as having 3 ED's. Modification was coded separately for each ED even i f the wear patterns were identical. Once each modified a r t i f a c t was coded, the ED's were assigned to a category including motion and worked material, depending on the values of the relevant attributes. Basically, this decision-making process followed the same sequence of steps for each ED. F i r s t , the type of motion used i n the a c t i v i t y was determined. This was done by considering: position of ED, placement of modification, direction of modification, and angle of ED. Table 27 provides a correlation of attribute values with motion (Odell 1981; Odell and Odell-Vereecken 1980). As an example, i f an ED on a flake was coded as: position of ED edge placement of modification b i f a c i a l direction of modification oblique angle of ED acute then the interpretation of the motion used would be cutting. Second, the type of material worked by the a r t i f a c t was determined. The hardness of the worked material i s a major factor i n analysis of use-wear (Vaughan 1985). Experimental tests involving stone tools have demonstrated that variation i n hardness of the different worked materials i s correlated 215 Table 27. Correlation of attribute values with notion. Attribute Value Motion Position of ED Surface Hammer Tip/distal end Grave/bore/wedge Edge Cut/scrape Base/proximal end Cut/scrape/wedge Notch Cut/scrape Placement of Dnifacial Scrape/hammer modification B i f a c i a l Cut Around t i p Grave/bore/wedge Direction of Oblique Cut modification Perpendicular to edge Scrape/cut/hammer Parallel to edge Cut Parallel to long axis Grave/bore Perpendicular to long axis Grave/bore Multidirectional Grave/bore Unknown Angle of ED Acute Cut/grave/bore Obtuse Scrape/wedge/chop Steep Scrape Not applicable Hammer Table 28. Correlation of attribute values with worked material. Attribute Value Worked Material Type of modification Size of scars Type of modification Size of scars Type of modification Size of scars Type of modification Size of scars Feather scars/striations/polish Soft Small/small/not applicable Feather scars Soft-medium Large Hinge and step scars/striations Hard-medium Small/large Hinge and step scars Hard Large 216 with variation i n the size and shape of the resultant scars (Tringham et a l . 1974; Odell 1981; Odell and Odell-VereecXen 1980). The following categories of hardness are based on George Odell 1s classifications (Odell 1981; Odell and Odell-Vereecken 1980:101): soft: animal products such as meat, skin and f a t ; soft vegetal substances such as tubers, rhizomes, stalks and leaves; soft-medium: soft woods, particularly coniferous, and firm pliable substances such as fresh stalks; hard-medium: hard woods, soaked antler and fresh bone; hard: bone i n any state, dry antler, dry wood, carcass. Attributes considered for the interpretation of the hardness of the worked material include type of modification and size of scars. Table 28 provides a correlation of attribute values with worked material. The coded values for these attributes for every EO were examined and interpreted according to Tables 30 and 31, and the EU was assigned to the appropriate category. For example, an EU coded as: type of modification feather scars size of scars large was interpreted as having been used on soft-medium worked material. Employable Unit Typology The EU types were established by combining the two categories described above, motion and worked material. There are sixteen mutually exclusive types present i n the t o t a l study sample (Table 29). This typology i s independent of tool morphology. In order to calculate the number of single purpose and multi-purpose tools, the functional interpretation for each EU i s determined. Those tools with more than one function, that i s , more than one EU type, are designated as multi-purpose tools. A tool with more than one EU, but a l l of the same type, i s c l a s s i f i e d as a single-purpose tool. 217 Table 29. Employable Unit types. EU Number Motion Worked Material 1 Grave/bore Soft-medium 2 Grave/bore Soft 3 Grave/bore Hard-medium 4 Scrape Soft-medium 5 Cut Soft-medium 6 Cut Hard 7 Cut Hard-medium 8 Cut Soft 9 Scrape soft 10 Scrape Hard-medium 11 Scrape Hard 12 Wedge Hard 13 Hammer Hard 14 Grave/bore Hard 15 Chop Hard 16 Chop Hard-medium 218 Table 30 provides frequency counts of the number of each ED type present i n the study sites . These figures w i l l not match exactly the numbers of tools present i n each s i t e because, i n many cases, each tool w i l l have more than one ED. This table does not indicate how many u t i l i z e d a r t i f a c t types are present but rather the relative proportion of each a c t i v i t y i n the s i t e . Results Table 31 demonstrates the number of Employable Unit types present and the worked material types present i n each s i t e . In Upper Hat Creek, the sites previously interpreted as camp sites, and those sites with tools previously interpreted as representative of an intensive occupation contain the highest numbers of EU types. In addition, these EU's were used on a l l four kinds of worked material. The four sites previously interpreted as limited a c t i v i t y tend to contain fewer EU types, and a more limited range of worked materials. Overall, those sites c l a s s i f i e d as residential camps, i n Highland valley, display the highest number of EU's present, and f i e l d camps display the second highest number of EU's present. The majority of residential camps have EU's which have been used to work a l l four types of material, while f i e l d camps contain EU's which have been used on three to four types of material. The single station contains only one EU type, used on one type of material. In general, these data support Binford's (1980) and Chatters' (1987) predictions about the range of a c t i v i t i e s occurring at various s i t e types. That i s , residential camps were the locus of the greatest variety of a c t i v i t i e s , f i e l d camps were used for a smaller range of a c t i v i t i e s , and stations were used for the most limited range of a c t i v i t i e s . Table 30. Frequency counts of Employable Unit types. 219 Site Employable Unit Types 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 EeRilO 2 1 8 3 1 5 19 66 8 1 EeRj49 1 1 12 10 3 9 7 2 4 EeRj55 3 1 5 10 10 22 1 1 1 EeRj56 1 1 9 4 6 2 3 1 1 EeRj60 2 12 11 10 12 16 1 5 1 EeRj62 2 3 26 7 9 7 19 1 3 EeRj 8 1 5 7 3 1 2 EeRj20 1 3 1 2 1 1 1 EeRj42 2 4 3 44 9 15 25 22 41 4 3 1 1 2 EeRj64 1 13 1 6 2 1 5 EeRj100 3 4 5 2 4 1 EeRk52 4 2 2 4 1 1 EcRg2AA 2 48 3 15 15 5 23 217 438 20 11 1 6 15 1 EcRg2CC 1 6 14 1 3 5 1 1 2 2 EcRg4C 2 3 14 3 9 15 10 3 1 1 3 EcRg4J 2 2 3 3 18 1 EdRglA 1 7 8 3 18 1 EdRglB 7 4 18 6 42 3 1 EcRg4A 3 7 2 4 4 10 17 1 2 EcRg4B 3 5 9 3 6 2 3 EcRg4D 1 2 3 EcRg4E 3 EdRg5 1 1 2 1 2 1 1 1 EdRg6 3 1 9 1 1 Total 10 70 27 213 59 101 154 337 726 45 54 5 9 28 4 6 220 Table 31. Frequency counts of Employable Unit types, worked material types, and tool:ED ratio. Site Previous Number of Number of Ratio of Classification ED Worked Tool Type: Types Material Types ED Type EeRilO camp s i t e 9 4 1:2.25 EeRj49 9 4 1:3.00 EeRj55 camp s i t e 9 4 1:3.00 EeRj56 intensive occupation 7 4 1:1.75 EeRj60 intensive occupation 7 4 1:2.33 EeRj62 intensive occupation 9 4 1:2.25 EeRj 8 limited a c t i v i t y 6 3 1:3.00 EeRj20 limited a c t i v i t y 6 3 1:2.00 EeRj42 intensive occupation 13 4 1:4.33 EeRj64 intensive occupation 5 3 1:1.66 EeRj100 8 4 1:2.66 EeRk52 limited a c t i v i t y 8 4 1:4.00 EcRg2AA residential camp 14 4 1:2.80 EcRg2CC residential camp 10 4 1:3.33 EcRg4 residential camp 11 4 1:3.66 EcRg4J f i e l d camp 6 4 1:3.00 EdRglA f i e l d camp 6 3 1:3.00 EdRglB f i e l d camp 9 3 1:3.00 EcRg4A f i e l d camp 8 4 1:4.00 EcRg4B residential camp 8 4 1:2.66 EcRg4D station 3 3 1:3.00 EcRg4E f i e l d camp/station? 1 1 1:1.00 EdRgS residential camp 8 3 1:8.00 EdRg6 f i e l d camp 5 4 1:5.00 221 Table 31 also demonstrates the relationship between the number of morphological tool types present i n each s i t e , and the number of ED types present i n each s i t e . The number of morphological types i s used here as a measure of tool diversity, and the number of ED types i s used as a measure of the range of tasks involving these tools. The ra t i o of morphological tool type to ED type i s provided i n the right-hand column of the table. This figure i s a measure of how specialized the tool assemblage i s : the higher the ratio, the more generalized the tools are. Those Highland valley sites c l a s s i f i e d as residential camps contain, on the average, a less specialized tool assemblage with fewer uses per tool type than sites c l a s s i f i e d as f i e l d camps. In addition, those Upper Hat Creek sites previously interpreted as being intensively occupied with a wide range of a c t i v i t i e s (EeRj 56, ErRj60, EeRj62, EeRj42, ErRj64) have a more specialized assemblage with fewer uses per tool than those sites characterized as having expedient tool assemblages (EeRj55, EeRj8, EeRj20, EeRk52). This result corroborates Chatters 1 (1987) original prediction which links the wide range of a c t i v i t i e s performed at residential camps with a need for a more diverse tool assemblage. Activity Area Analysis Introduction Previous researchers (Pokotylo 1978; Magne 1985) who have successfully used manufacturing stage debitage analysis i n order to produce a s i t e typology have treated the s i t e assemblage as one unit. However, Binford (1978a) pointed out that s i t e structure, or the patterning of a r t i f a c t location within a si t e , reflects technological organization, a c t i v i t y 222 structure and disposal mode. Camilli (1983) also proposed that a s i t e should be considered as a spatial composite of a c t i v i t y areas which w i l l vary i n a predictable manner according to the occupational history of the s i t e . Significant intrasite v a r i a b i l i t y and patterning related s p e c i f i c a l l y to the organization of microcore technology may be more readily apparent i f sites are divided into a c t i v i t y areas. In addition, the majority of sites i n the study area consist of l i t h i c scatters without features, and a c t i v i t y area analysis provides a potentially useful method for determining the relative complexity and diversity of such sites. Identification of areas within a s i t e where l i t h i c tool manufacture has taken place does not preclude the occurrence of other a c t i v i t i e s involving tools or materials which do not preserve as well, or a c t i v i t i e s which do not produce a large amount of debris. However, a c t i v i t y area identification and analysis w i l l enable a determination of the complexity and arrangement of at least some of the a c t i v i t i e s at sites, for comparative purposes. Ethnographic, ethnoarchaeological, and experimental approaches to the study of tool use and manufacture have indicated that certain kinds of tools do tend to be used together, constituting tool k i t s (Carr 1984). However, these tool k i t s and the areas where they are deposited tend to be polythetic, that i s the a r t i f a c t types (morphological) are similar but not identical and no single a r t i f a c t type i s essential to the definition of the ac t i v i t y (Carr 1984). In addition, tool k i t s and a c t i v i t y areas tend to overlap, that i s similar a r t i f a c t types w i l l occur i n more than one k i t or area (Carr 1984). Activity areas, as the term i s used here, are defined as locations within a s i t e where tools and debris cluster together because of localized use, manufacture, disposal and/or storage. 223 Recent ethnoarchaeological research (Murray 1980; Spur l i n g and Hayden 1984; Binford 1978a, 1983; Yellen 1977; Hayden 1987; Lee 1979; Gould 1980) has substantiated the assumption that a c t i v i t y areas can be defined on the basis of spatial patterning of l i t h i c tools and debitage. Using ethnoarchaeologically derived data from two recently occupied sites i n Western Australia, Spurling and Hayden (1984) noted that l i t h i c debris i s always discarded within 1 to 2 metres of the workplace. As well, tool manufacture for individual consumption (as opposed to manufacture for trade i n complex societies) generally takes place adjacent to the l i v i n g areas within a s i t e , usually next to individual family hearths (Yellen 1977; Yellen and Lee 1976; Spurling and Hayden 1984; 0'Connell 1977). Murray's (1980) ethnographic research indicated that semi-sedentary groups who reside at a s i t e for at least one season w i l l discard elements away from the area of use only i f this area i s enclosed. The same population at short-term sites w i l l discard elements away from the area of use i f i t i s a temporary shelter or at the location of use i f i t i s outside. Yellen and Lee (1976) noted that !Rung huts serve as storage places, sources of shade and markers of family ground; very few a c t i v i t i e s take place within them. Ethnographic and archaeological data from the study areas indicate that they were occupied during spring, summer and f a l l months by highly mobile residential and l o g i s t i c a l groups. Shelters, i f any were used, consisted of temporary mat or brush lodges. Thus the implications of Murray's (1980) and Yellen and Lee's (1976) findings are that most a c t i v i t i e s involving l i t h i c tools and debris would have taken place outside, and that discarded elements would have been deposited at the location of use. I t seems unlikely that f lintknapping would have been carried out inside a dwelling (notwithstanding 224 the interpretation of the mat house at EcRg2AA by Areas Associates 1983), because of v i s i b i l i t y problems, and the potential discomfort of sleeping or stepping on sharp debris. Recent ethnoarchaeological research (Binford 1983) has suggested that basic intra-site patterns are largely the result of the mechanics of the human body. These patterns of space use are redundant, and consist of the use of different areas of the s i t e for different purposes: sleeping, seating, hearths, beds, and debris accumulation. A family unit occupies a similar amount of space, depending on number of days of occupation. For example, the !Kung occupy areas ranging i n size from 7.29 to 16.43 m2 during stays lasting 2 to 10 days, and areas ranging i n size from 5.08 to 35.87 m2, during stays lasting 11 to 40 days. Given that l i t h i c tool manufacture produces predictable spatial forms of debris scatter, sites which are composites of functionally similar a c t i v i t i e s should exhibit the same regularities (Camilli 1983). Residential camps should have a larger number of ac t i v i t y areas than f i e l d camps and stations, since spatial arrangement of a c t i v i t i e s w i l l r e f l e c t the number of social units and/or families, at the s i t e (Price 1978). Residential camps should also contain a number of a c t i v i t y areas similar i n size and function, again due to the number of family units occupying the s i t e (Camilli 1983; Yellen 1977). In addition, increased occupation spans and a greater number of occupants, may result i n more specialized use of areas due to the need for increased spatial separation of a c t i v i t i e s due to such considerations as space required for bulk processing, the amount of debris produced, and scheduling concerns (Binford 1978a; rami H i 1983). Thus, residential camps are expected to contain more specialized and discrete a c t i v i t y areas than 225 f i e l d camps. Fi e l d camps should require less spatial separation of a c t i v i t i e s and more generalized a c t i v i t y areas. Stations w i l l have the least complex spatial organization. The objectives of the a c t i v i t y area analysis i n this study are to define the spatial limits of each area, and to define the a c t i v i t i e s which took place within that area, as evidenced by l i t h i c debitage and tools. To establish a relative index of s i t e size i n terms of level of a c t i v i t i e s which occurred, measures of area size, density, and s p e c i f i c i t y of l i t h i c contents are established. Activity areas are identified as specialized or generalized, and discrete or blended. Method The method used to determine ac t i v i t y areas was developed and refined during the p i l o t study. L i t h i c tool manufacture generates such a large amount of debris within a relatively small area (cf. Newcomer and Sieveking 1980) that s t a t i s t i c a l techniques are not required to identify the resultant scatters (Carr 1984). Visual inspection of s i t e maps provides an adequate picture of interpretable structure, both i n the spatial patterning of individual items, and i n assemblage v a r i a b i l i t y over the s i t e area (Price 1978). A l l study sites are l i t h i c scatters containing large quantities of debitage and very few features. In addition, previous research i n the study area (Pokotylo 1978; Areas Associates 1983, 1986) indicates that these sites are the remains of mobile hunter-gatherers. Results of ethnographic research suggests that, under these circumstances, stone tools, both expediently-produce and curated, w i l l be discarded within .5 m of the workplace, along with manufacturing, rejuvenation, and resharpening debris (Hayden 1976; 226 Spurling and Hayden 1984; Gould 1978; Murray 1980). Therefore, the analytical method used here i s visual inspection, based on the following working assumptions: 1. Discard of l i t h i c debitage occurs at the locus of manufacture, or resharpening or rejuvenation; 2. Discard of formed l i t h i c tools occurs at or near the locus of use or f i n a l handling; 3. Discard of expedient l i t h i c tools occurs at or near the locus of manufacture and use; 4. Differential distribution of l i t h i c a r t i f a c t s i s due to separate u t i l i z a t i o n at dist i n c t locations within the s i t e . The presence/absence of artifacts i s used to define a c t i v i t y areas. Proportions of manufacturing debris and Employable Units are used to define a c t i v i t i e s . V a r i a b i l i t y i n proportions w i l l r e f l e c t differences i n a c t i v i t i e s . Previous determinations of a c t i v i t y area patterning i n Highland Valley sites were incorporated into this study because they included f i e l d data and impressions not available to the author. In addition, s i t e EeRj49, i n Upper Hat Creek Valley was interpreted as one a c t i v i t y area because geomorphological evidence indicates that artifacts were re-deposited by runoff from above the s i t e . The following method was used to define a c t i v i t y areas for each s i t e : 1. Prepare a two-dimensional unit map of each s i t e with the number of art i f a c t s , debitage and tools, i n each collection unit. 2. Select units with higher density deposition of arti f a c t s (debitage and tools) surrounded by lower-density units, and a r b i t r a r i l y group these contiguous units into spatially discrete a c t i v i t y areas. "Higher density" i s 227 defined only in relation to surrounding units and not by an arbitrary number of a r t i f a c t s . "Activity areas" are separated by units which contain the lowest numbers of ar t i f a c t s . I f two low-density units are adjacent, then the a c t i v i t y area boundary i s drawn between them; otherwise, the boundary i s drawn a r b i t r a r i l y on one side or the other of the low-density unit. Ethnographic and experimental research has indicated that l i t h i c debitage from a worker w i l l scatter i n a radius approximately 2 to 4 metres wide, depending on whether the worker i s i n a seated or standing position (Camilli 1983). In order to account for both these p o s s i b i l i t i e s , a single unit containing one or more artifacts i s considered to be a separate a c t i v i t y area i f separated by at least 5 m from other units containing a r t i f a c t s . However, i n Highland valley sites which were sampled, the a c t i v i t y areas are separated by an arbitrary distance of more than 10 m. 3. Prepare a table l i s t i n g each a r t i f a c t type i n the a c t i v i t y areas. 4. Compile descriptive data for each a c t i v i t y area: type of features, area size i n square metres, a r t i f a c t t o t a l and density of arti f a c t s per square metre. 5. Determine the emphasis placed on manufacturing stages and tool use i n each a c t i v i t y area, by the following: (a) Determine the Employable Unit cl a s s i f i c a t i o n nd corresponding task for each tool (see above under Tool Analysis). Assign the task to the appropriate a c t i v i t y area, and calculate the percentage of each type of act i v i t y . (b) Calculate the percentage of debitage assigned to each of the three generalized manufacturing stages, for each a c t i v i t y area, according to the method described above. 228 (c) Determine the presence/absence of microlithic a r t i f a c t s for each a c t i v i t y area. Description of A c t i v i t y Areas This section provides descriptive data for a c t i v i t y areas i n each s i t e . Table 32 provides the descriptive data for a c t i v i t y areas i n the study sites: type of feature, size of area, a r t i f a c t t o t a l and a r t i f a c t density. Appendix II contains maps of the a c t i v i t y areas showing a r t i f a c t counts for each collection unit, and locations of tools, microlithic ar t i f a c t s and features. Numbers assigned to individual a c t i v i t y areas within sites are underlined, eg. 1. Act i v i t y Area Clusters In order to determine the major characteristics of each a c t i v i t y area, and to compare a c t i v i t y areas containing microlithic a r t i f a c t s with those containing only non-microlithic artifacts, cluster analysis i s applied to percentage counts of three manufacturing stages derived from analy i s of flaxes and shatter, and percentage counts for Employable Units i n each a c t i v i t y area. Employable Units 15 and 16 are deleted from analysis because of their low occurrence i n the study sample. Percentage data are selected for analysis i n order to provide an indication of the relative emphasis placed on the various manufacturing stages and tool usage monitored by Employable Units. Microlithic debitage i s not included i n this analysis because the result would be a two-cluster grouping into microlithic and non-microlithic a c t i v i t y areas. The clustering method i s Wards, discussed above, and the distance measure i s the standardized Euclidean distance coefficient. Table 32. Activity area descriptive data. Site Area Feature Size A r t i f a c t A r t i f a c t (M2) Total Density/M 2 EeRilO 1 400 17 0.04 2 340 36 0.11 3 216 18 0.08 4 672 581 0.86 5 160 734 4.59 EeRj49 1 112 480 4.28 EeRj55 1 24 30 1.25 2 RP 160 4142 25.89 3 80 36 0.45 EeRj56 1 140 57 0.41 2 280 169 0.60 3 256 429 1.68 4 88 27 0.31 5 1320 2407 1.82 6 224 143 0.64 EeRj60 1 144 151 1.05 2 264 267 1.01 3 420 65 0.15 4 360 891 2.48 5 320 65 0.20 6 1196 1864 1.56 EeRj 62 1 100 80 0.80 2 72 109 1.51 3 432 488 1.13 4 72 103 1.46 5 144 23 0.16 EeRj 8 1 240 484 2.02 2 4 1 0.25 3 4 1 0.25 EeRj20 1 160 80 0.50 2 224 149 0.67 3 60 51 0.85 EeRj42 1 4 1 0.25 2 1080 2333 2.16 3 392 884 2.26 4 36 3 0.08 5 900 1201 1.33 6 704 597 0.85 7 780 588 0.75 8 240 197 0.82 9 220 60 0.27 RP: Roasting p i t Table 32, continued. Site Area Feature Size A r t i f a c t A r t i f a c t (M2) Total Density/M 2 EeRj64 1 24 13 0.54 2 168 102 0.61 3 64 54 0.84 4 32 71 2.22 5 224 103 0.46 EeRj 100 1 64 38 0.59 2 72 48 0.67 3 80 38 0.48 4 112 36 0.32 5 24 3 0.13 EeRk52 1 RP 140 38 0.27 2 180 103 0.57 3 24 3 0.13 EcRc/2AA 1 10 10 1.00 2 66 61 0.92 3 PM(26) 168 4118 24.52 4 1 8 8.00 5 1 11 11.00 EcRg2CC 1 24 128 5.33 2 48 123 2.56 3 264 40 0.15 EcRg4C 1 20 20 1.00 2 — — 30 60 2.00 3 12 26 2.17 4 121 407 3.36 EcRg4J 1 6 2 0.33 2 11 4 0.36 3 176 63 0.36 EdRglA 1 190 197 1.04 EdRglB 1 PM(3) 60 828 13.80 EcRg4A 1 20 34 1.70 2 6 38 6.33 3 9 63 7.00 4 40 95 2.38 EcRg4B 1 H(2) 150 233 1.55 EcRg4D 1 272 35 0.13 EcRg4E 1 73 19 0.26 EdRgS 1 H 130 35 0.27 2 1 17 17.00 3 1 34 34.00 4 1 27 27.00 EdRg6 1 H{2) 38 900 23.68 2 1 1 1.00 H: Hearth PM: Post mould RP: Roasting p i t 231 Figure 24 reproduces the cluster deridrogram, stewing 9 interpretable groups. Tables 33 and 34 provide input data for debitage and tools, respectively, arranged by cluster. Those a c t i v i t y areas containing microlithic debitage are marked with an asterisk, i n order to provide an indication of other a c t i v i t i e s occurring i n association with microblade production. On the basis of these groups, the mean proportion of debitage and EU types i n each a c t i v i t y area i s re-calculated for each cluster-group (see Table 35). In addition, the mean and range of size, i n square metres, and a r t i f a c t density, per square metre, are calculated for each cluster (see Table 36). These measures are used as the basis of the following interpretation. In order to provide an i n i t i a l indication of the attributes which best differentiate the clusters, a series of Kruskal-Wallace analysis of variance tests i s run on the original raw data set grouped according to the Ward's cluster analysis. Table 37 presents the results. Nine of the 18 a c t i v i t y area attributes are significant at the 0.05 level of probability, and an additional two at 0.056 and 0.057, representing samples drawn from at least two different populations. Those attributes Which do not significantly differentiate among the clusters are those which occur i n comparatively few a c t i v i t y areas, i n very low percentages: cores, and Employable Units 1 and 2, 10, and 12 to 14. Cluster 1 contains 11 microlithic and 7 non-microlithic a c t i v i t y areas. These areas are large, more than 200 square metres, with a high a r t i f a c t density. A l l stages of core reduction and manufacturing are represented, with emphasis on the later stages. In regard to tool usage, a l l EU's are represented i n this cluster, although emphasis i s placed on EU9, which i s 0 . 0 ttttfucm Eillll(J) EiljCH ttlf2U( Edtgid) Ell|4JIJ) EtljHCl Efl](l((la J-1 E<lf4A<]) ItlqlUC] ElIIINJI Ellill(4)l EMtlllllt I t t I I I I I | I I 200,0 4)1 • i 232 5h. E<t,2U(J)l El(4JC)l UI,1»U>« E<l«4>(l) Eil|4E(l) EtljHCl Eil<4CI2) EU»J(I> Eiljlllll) EHtf(J) til]!••<]) Ecl|2CC(l)l Ell|4l(t) El|4t(4) bt|4]IT) _ , ( i i j t i m i 4—' E>l|4]|]l J Etl|tl(]| Ell|4C(ll E>l|«)(]) l(M<CCI El|41(l)l ll|41(l) t.llJKl) tiiiiin:) (l4«Cl Itl|4>l4l El|l(4l« f<»l<l(ll llljlllll liMKIUll—«i (ilium _r uioui — MlH I i.iiuiii _h IlHKI _ \ tiiiiiuit _r hi II I! _J lil|H(«l _ Ciiiiui _ h ttiiiiiji _ j (iiiiaii) Jr Eil|JSI3l __. mom —i " iinsii]) -,h liiiHd) s\ Eiit*i(u r EllllKJ) _ _ [lf.tKI fiiiliilli (ll,Ilil3). llliill!) Etii<mi (tli(4(]l W|53(I) lil|3tllll UliHI llllJlll) blllllt) (ill)HI Eil Eil EiljllO) tll|lllj)l LiljHUJ ltl|lll(31 E>I|I3] Cfl|l(» [|I(<JIIM — . Eii|4i(iii r ] f •|li(3l — . • |3tll) - i t — liill) Jl •iiini _ r IJU 1 4(11 _J-^  11(31 ^ Figure 24. Ward's cluster analysis of a c t i v i t y area attributes. Table 33. Debitage data for a c t i v i t y area cluster analysis grouped by cluster. Cluster Site Area C E M L Total EcRg2AA 4* 100.0 100.0 EeRilO 3 12.5 6.2 81.3 100.0 EeRj60 1* 8.2 6.1 85.7 100.0 EcRg2AA 1* 100.0 100.0 EdRg6 1 0.1 5.9 19.2 74.8 100.0 EeRj42 5 0.1 2.7 15.1 82.1 100.0 EeRj62 2 3.1 19.2 77.7 100.0 EeRj55 2* 0.1 5.5 13.6 80.8 100.0 EeRj60 6* 0.3 5.8 10.4 83.5 100.0 EcRg4A 3 3.4 37.9 58.7 100.0 EcRg2AA 3* 1.6 6.8 23.2 68.4 100.0 EeRilO 5* 0.6 2.0 31.7 65.7 100.0 EeRilO 4* 0.9 2.2 35.7 61.2 100.0 EcRg2AA 2* 24.4 75.6 100.0 EcRg4J 3* 2.7 5.4 43.2 48.7 100.0 EdRglA 1* 3.0 41.2 55.8 100.0 EcRg4A 1 11.1 51.9 37.0 100.0 EcRg4E 1 23.5 23.5 53.0 100.0 EeRj56 3 7.5 26.3 66.2 100.0 EcRg4C 2* 3.6 14.5 81.9 100.0 EdRg5 1 30.3 69.7 100.0 3 EeRj100 4 3.0 3.1 24.2 69.7 100.0 EdRg6 2 100.0 100.0 EeRj100 2 4.4 8.9 4.4 82.3 100.0 *: Contains microlithic artifacts C: Cores E: Early core reduction stage M: Late core/early blank reduction stage L: Late blank reduction stage Table 33, continued. Cluster Site Area C E M L Total EcRg2CC 1* 5.1 36.4 58.5 100.0 EcRg4B 1 0.5 1.9 33.4 63.6 100.0 EcRg4A 4 5.0 31.2 63.8 100.0 EeRj42 7 0.5 3.6 12.5 83.4 100.0 EeRj62 3* 0.7 5.3 17.8 76.2 100.0 EeRj42 3 0.1 2.2 15.8 81.9 100.0 EeRj60 2 0.4 4.6 9.2 85.8 100.0 EcRg4C 4* 1.6 19.7 78.7 100.0 EeRj42 2 0.1 3.0 16.7 80.2 100.0 EcRg4C 3* 13.0 4.3 82.7 100.0 EeRj49 1* 0.5 8.8 20.4 70.3 100.0 EcRg4A 2 6.2 31.2 62.6 100.0 EeRk52 1 3.2 38.7 58.1 100.0 EeRj100 3 2.9 17.1 80.0 100.0 EcRg4C 1* 5.3 15.8 78.9 100.0 EeRj42 4 100.0 100.0 EeRj62 4* 3.8 15.2 81.0 100.0 EeRj42 8 4.7 16.3 79.0 100.0 EeRj42 6 0.3 1.5 17.0 81.2 100.0 EcRg2CC 2* 0.9 5.3 25.4 68.4 100.0 EeRj64 5 3.2 11.8 11.8 73.2 100.0 EdRgS 2 21.4 78.6 100.0 EeRj62 1* EeRk52 2 EeRj64 4 EeRj56 5* EeRj20 2 EeRj60 4 EeRj8 1 EeRj20 3 EcRg2CC 3 EeRj55 3 EdRg5 3 EeRk52 3 EeRj64 2 EcRg4D 1 EeRilO 2 1.4 2.8 25.3 2.1 22.9 1.5 10.6 18.2 0.1 3.4 23.8 2.1 16.0 1.4 4.5 5.7 0.2 9.7 7.6 2.1 2.1 12.5 15.8 2.9 5.7 14.3 3.2 29.0 33.3 10.4 21.9 25.0 5.8 35.3 70.5 100.0 75.0 100.0 69.7 100.0 72.7 100.0 81.9 100.0 88.4 100.0 82.5 100.0 83.3 100.0 84.2 100.0 77.1 100.0 67.8 100.0 66.7 100.0 67.7 100.0 75.0 100.0 58.9 100.0 *: Contains microlithic artifacts C: Cores E: Early core reduction stage M: Late core/early blank reduction stage L: Late blank reduction stage 235 Table 33, continued. Cluster Site Area C E M L Total 6 EeRj56 4 50.0 50.0 100.0 EeRj56 2* 3.6 40.7 55.7 100.0 EcRg2AA 5* 33.3 66.7 100.0 EcRg4J 2 25.0 75.0 100.0 EeRj42 9 1.7 20.0 78.3 100.0 EeRj64 3 1.8 3.7 20.4 74.1 100.0 EeRj55 1 10.0 20.0 70.0 100.0 EeRj56 6* 0.7 10.5 18.9 69.9 100.0 EdRgS 4 14.8 85.2 100.0 EeRj20 1 13.7 86.3 100.0 EeRilO 1 11.8 88.2 100.0 7 EeRj62 5 40.0 60.0 100.0 EeRj56 1 8.9 26.8 64.3 100.0 EeRj100 1 5.6 19.4 75.0 100.0 EeRj60 3 1.6 15.9 7.9 74.6 100.0 EeRj60 5* 48.2 7.1 44.7 100.0 EeRj64 1 27.3 9.1 63.6 100.0 8 EeRj100 5 EeRj 8 3 EeRj 8 2 9 EcRg4J 1* EeRj42 1 *: Contains microlithic artifacts C: Cores E: Early core reduction stage M: Late core/early blank reduction stage L: Late blank reduction stage 236 Table 34. Tool data f o r act i v i t y area cluster analysis grouped by cluster. Cluster Site Area EU1 EU2 EU3 EU4 EUS EUS EU7 EU8 EU9 EU1B EU11 EUI2 EU13 EU14 Total I EcRg2AA 4* £6.7 33.3 18B.B 188.0 EeRilB 3 20.8 2B.B 6B.B EsftjSB U 2B.B 88.B 188.8 EcRg2AA It 1BB.B 1BB.B EdRgS 1 23.1 7.7 61.5 7.7 188.8 EeRj42 S 17.1 2.9 11.4 2B.B 2.9 42.9 2.9 1BB.B EeRj62 2 8.3 25. B 16.7 8.3 41.7 1BB.B EeRjSS 2< 5.8 1.8 7.7 19.2 19.2 42.3 1.9 1.8 1BB.B EeRj6a 6* 6.9 2B.7 13.8 24.1 31.B 3.4 1BB.B EcRg4A 3 12.5 37.5 5B.B 1BB.B EcRg2AA 31 B.3 5.9 B.4 1.8 1.8 B.6 2.6 26.5 53.4 2.5 1.4 B.l 8.7 1.9 1BB.B EeRilB 5t 1.6 4.7 4.7 2B.3 64.1 3.1 1.6 IBB.8 EeRilB 4» 1B.B 5.8 5.8 12.5 55.B 12.5 1BB.B EdRglB 1* 8.6 4.9 22.2 7.4 51.9 3.7 1.2 1BB.B EcRg2AA 2t 8.3 16.7 16.7 58.3 188.8 EcRg4J 3t 3.8 7.7 11.5 3.8 69.2 3.8 1BB.8 EdRglA It 2.6 18.4 21.8 7.9 47.4 2.6 180.8 EcRg4A 1 2B.B 1B.B 18.B 1B.8 1B.8 4B.B 1BB.B 2 EcRg4E 1 1BB.B 180.8 EeRj56 3 5B.B 5B.B 18B.B EcRg4C 2t 11.1 11.1 11.1 55.6 11.1 18B.B EdRgS 1 33.3 33.3 33.3 188.8 3 EeRjIBB 4 33.3 33.3 33.3 lee.e EdRg6 2 5B.B 5B.B 188.8 EeRjIBB 2 28.8 68.8 28.B 188.8 »: Contains aicrolithic artifacts Table 34, continued. Cluster Site Area EU1 EU2 EU3 EU4 EUS EUS EU7 EU9 EUS EU10 EUU EU12 EU13 EU14 Total EcRg2CC U 13.3 33.3 13.3 26.7 6.7 6.7 1BB.B EcRg4B 1 9.7 16.1 29.B 9.7 19.4 6.5 9.7 100.B EcRg4A 4 4.0 16.8 12.B 8.8 2B.B 28.B 4.8 8.B 180.0 EeRj42 7 5.3 10.5 21.1 5.3 26.3 5.3 26.3 1BB.B EeRj62 3i 2.6 2.6 31.6 5.3 23.7 2.6 26.3 S.3 100.8 EeRj42 3 27.3 9.1 36.4 18.2 9.1 10B.0 EeRj60 2 14.3 14.3 14.3 28.6 28.6 188.0 EcRg4C 4* 2.4 2.4 29.3 2.4 9.8 24.4 19.5 7.3 2.4 1BB.8 EeRj42 2 1.5 3.0 17.9 6.8 11.9 7.5 23.9 17.9 3.8 3.8 l.S 1.5 1.5 100.8 EcRg4C 3* 2B.B 2B.0 28.8 4B.B 188.8 EeRj49 It 2.0 2.0 24.5 2B.4 6.1 18.4 14.3 4.1 8.2 taa.a EcRg4A 2 28.6 14.3 28.6 28.6 108.8 EeRkS2 1 28.6 28.6 28.6 14.3 180.8 EeRjlBB 3 25.8 75.8 IBB.8 EcRq4C It 58.8 58.8 108.8 EeRj42 4 33.3 66.7 188.0 EeRj62 4t 20.0 48.B 28.8 20. a 180.0 EeRj42 8 48. B 18.8 1B.B 2B.B 1B.B 1B.8 100.8 EeRj42 6 3.3 3.3 43.3 18.B 3.3 18.8 3.3 28.8 3.3 180.0 EcRg2CC 2t 23.5 47.1 5.9 5.9 5.9 11.8 188.0 EeRj64 5 10.8 68.8 10.a 20.0 188.0 EdRg5 2 58.8 SB.B 188.8 EeRj62 It 36.4 27.3 9.1 18.2 9.1 EeRk52 2 28.6 28.6 28.6 14.3 EeRj64 4 37.5 12.5 12.5 12.5 12.5 12.5 EeRj56 5t 4.3 4.3 34.8 17.4 26.1 4.3 8.7 EeRj28 2 14.3 28.6 14.3 14.3 14.3 14.3 EeRj60 4 3.7 18.5 14.8 14.8 7.4 11.1 3.7 18.5 EeRjB 1 6.7 20.0 48. B 13.3 EeRj20 3 33.3 33.3 33.3 Ecfig2CC 3 25.B 25.0 25. B 25.8 EeRj55 3 58.8 SB.B EdRg5 3 2B.B 4B.B 28. B EeRk52 3 25. B SB.B 25.8 EeRj64 2 25.8 SB.B 25.8 EcRg4D 1 12.5 25.8 62.5 EeRilB 2 SB.B SB.B 188.8 108.0 100.0 180.0 180.0 7.4 180.0 6.7 13.3 188.8 108.0 108.8 100.0 2B.B 100.B 108.8 1B8.B 188.8 100.0 t: Contains l i c r o l i t h i c artifacts 238 Table 34, continued. Cluster S i t e Area EU1 EU2 EU3 EU4 EUS EU6 EU7 EU8 EUS EU10 EU11 EU12 EU13 EU14 Total EeRj56 4 EeRj56 2 i EcRg2AA 5» EcRg4J 2 EeRj42 9 EeRj64 3 EeRjSS 1 EeRj56 S* EdRgS 4 EeRj2B 1 EeRilB 1 7 EeRjS2 S 108.8 100.0 EeRjSS 1 100.8 108.8 EeRjl00 1 188.8 180.0 EeRj£B 3 100.0 100.0 EeRj60 S» 100.0 100.0 EeRj64 1 66.7 33.3 100.0 8 EeRjlBB 5 83.3 1S.7 180.0 EeR j8 3 33.3 33.3 33.3 100.0 EeRj8 2 33.3 33.3 33.3 100.0 S EcRg4J U 33.3 SS.7 1BB.0 EeRj42 1 100.0 180.0 »: Contains • i c r o l i t b i c a r t i f a c t s Table 35. Mean proportion of a c t i v i t y area attributes grouped by Ward's cluster analysis. Cluster C E M L EU1 EU2 EU3 EU4 EUS EUS EU7 EU8 EUS EU1B EU 11 EU12 EU13 EU14 1 0.4 4.4 22.4 72.8 0.5 2.7 1.2 7.1 1.6 3.7 9.3 16.3 54.6 1.9 8.2 8.1 B.l 2 8.7 23.7 67.7 2.8 11.1 8.3 2.8 59.7 15.3 3 2.5 4.8 9.6 84.0 6.7 11.1 27.8 47.8 4 0.3 4.7 19.4 75.6 3.0 3.2 4.3 33.0 11.0 4.8 11.2 12.5 18.7 1.4 3.7 B.l B.l S 0.6 4.2 20.4 74.8 1.9 2.8 23.8 2.9 35.8 3.1 3.8 1.9 4.2 17.3 8.4 6 0.2 2.7 24.4 72.7 7 8.3 17.6 18.4 63.7 94.4 5.5 8 22.2 49.9 5.6 11.1 11.1 9 16.7 83.3 C: Cores E: Early core reduction stage H: late core/early blank reduction stage L: Late blank reduction stage Table 36. Mean size and a r t i f a c t density of a c t i v i t y area clusters. Cluster Size (M2) Ar t i f a c t Density/M 2 240 1 236.5 6.6 2 122.2 1.0 3 61.7 0.7 4 226.3 2.6 5 244.3 3.1 6 133.9 3.9 7 185.3 0.3 8 10.7 0.2 9 5.0 0.7 Table 37. Kruskal-Wallis tests on a c t i v i t y area attributes grouped by Ward's cluster analysis. Attribute H (X2) Degrees of Probability Freedom Cores 11.642 8 0.168 Early* 17.488 8 0.025 Middle* 16.301 8 0.038 Late* 20.527 8 0.009 EUl 11.909 8 0.155 EU2 11.779 8 0.161 EU3 15.108 8 0.057 EU4* 59.411 8 0.000 EU5 15.159 8 0.056 EU6* 52.625 8 0.000 EU7* 37.862 8 0.000 EU8* 41.395 8 0.000 EU9* 62.076 8 0.000 EU10 11.389 8 0.181 EU11* 41.257 8 0.000 EU12 1.753 8 0.988 EU13 9.328 8 0.315 EU14 7.772 8 0.456 *: Significant at 0.05 level of probability Early: Early core reduction stage Middle: Late core/early blank reduction stage Late: Late blank reduction stage 241 inferred to represent the scraping of soft materials, EU8, inferred to represent cutting soft materials, and EU7, inferred to represent cutting hard-medium materials. Cluster 1 a c t i v i t y areas are interpreted as locations of generalized tasks, including food preparation and consumption, secondary animal processing, and tool and container manufacture and maintenance. Cluster 2 contains 4 a c t i v i t y areas, l microlithic and 3 non-microlithic. They are medium-sized, between 100 and 200 square metres, with a f a i r l y low a r t i f a c t density. There i s representation from a l l stages of core and blank reduction, with particular emphasis on the early stages. Tool use i s low, with only six types of EU's being present. EU7, which i s inferred to represent cutting hard-medium materials, EUll, inferred to represent scraping hard materials, and EU3, inferred to represent graving/boring hard-medium materials, are dominant. Cluster 2 a c t i v i t y areas are interpreted as places of specialized use where tool and container manufacture and maintenance occurred. Cluster 3 contains 3 non-microlithic a c t i v i t y areas. These areas are re l a t i v e l y small, less than 100 square metres, with a f a i r l y low a r t i f a c t density. The debitage derives from a l l stages of core reduction and tool manufacture, with emphasis on the later stages. Again, tool usage i s quite limited, with representation from five EU's. The most numerous types are EUll, scraping hard materials, EU9, scraping soft materials, and EU5, cutting soft-medium materials. Cluster 3 i s interpreted as representing specialized areas, where a f a i r l y limited range of maintenance a c t i v i t i e s occurred. Cluster 4 contains 7 microlithic and 15 non-microlithic a c t i v i t y areas. These areas are large, with relatively high a r t i f a c t densities. Again, 242 debitage derives from a l l core reduction and manufacturing stages, with emphasis on the late stages. Tool usage as indicated by EU's i s highly varied, with f a i r l y even representation from a l l fourteen EU's. Emphasis i s placed on EU4, scraping soft-medium materials, EUS, cutting soft materials, EU7, cutting hard-medium materials, and EU5, cutting soft-medium materials. Cluster 4 i s interpreted as representing areas of generalized use, including a wide range of maintenance a c t i v i t i e s . Cluster 5 contains 2 microlithic and 13 non-microlithic a c t i v i t y areas. These areas are very large, with high a r t i f a c t densities. The debitage derives from a l l stages of core reduction and tool manufacture, with emphasis on the late stages. Again, tool usage i s highly varied, with f a i r l y equitable representation from twelve EU's. The most common types are EU6, cutting hard materials, EU4, scraping soft-medium materials, and EUll, scraping hard materials. Cluster 5 i s interpreted as representing generalized areas, where a wide range of maintenance a c t i v i t i e s took place. Cluster 6 contains 3 microlithic and 8 non-microlithic a c t i v i t y areas. These a c t i v i t y areas are medium-sized, with a f a i r l y high a r t i f a c t density. A l l manufacturing stages are represented, with emphasis on the middle and late stages. There i s no indication of tool usage. Cluster 6 i s interpreted as specialized a c t i v i t y areas focusing on l i t h i c tool manufacture and rejuvenation. Cluster 7 contains 1 microlithic and 5 non-microlithic a c t i v i t y areas. Activity areas are medium-sized, with a very low a r t i f a c t density. A l l stages of core reduction and tool manufacture are represented, with a high percentage deriving from the early stages. Tool usage i s limited to EU4, scraping soft-medium materials, and EU6, cutting hard materials. Cluster 7 243 i s interpreted as the use of specialized locations for the secondary processing of animal products, including skins, with some l i t h i c core reduction, and tool manufacture and maintenance. Cluster 8 contains 3 non-microlithic a c t i v i t y areas. Again, these areas are small, with a very low a r t i f a c t density. No core reduction or l i t h i c tool manufacturing debris i s present. Tool usage focuses on EU6, cutting hard-medium materials, EU4, scraping soft-medium materials, EU13, hammering hard materials, EUll, scraping hard materials, and EU9, scraping soft materials. Cluster 8 i s interpreted as representing the specialized use of locations where a few tasks occurred, focusing on non-lithic tool and container manufacture and/or the processing of animal and vegetal foods. Cluster 9 contains 2 microlithic a c t i v i t y areas. These a c t i v i t y areas are very small with a very low a r t i f a c t density. No core reduction or l i t h i c tool manufacturing debris i s present, and tool usage i s confined to EUS, which represents cutting soft materials, and EU2, which represents graving/boring soft materials. Cluster 9 i s interpreted as representing locations where specialized tasks occurred, the cutting and graving/boring of either soft animal products or vegetal substances. Discussion During the introduction above, i t was proposed that an analysis of the number, type and discreteness of a c t i v i t y areas would provide a method for differentiating among residential camps and f i e l d camps of mobile hunter-gatherers. Table 38 provides the data base for the following determination of how successful these measures are. Residential camps should contain a larger number of a c t i v i t y areas Table 38. Activity area measures used to discriminate among sites. 244 Site Previous Number of Percentage of Percentage of Classification Activity Specialized Discrete Areas Areas Areas EeRilO camp s i t e 5 20.0 100.0 EeRj49 1 0.0 0.0 EeRj55 camp s i t e 3 33.3 33.3 EeRj56 intensive occupation 6 83.3 100.0 EeRj60 intensive occupation 6 33.3 16.7 EeRj62 intensive occupation 5 20.0 20.0 EeRj 8 limited a c t i v i t y 3 66.7 100.0 EeRj20 limited a c t i v i t y 3 33.3 0.0 EeRj42 intensive occupation 9 22.2 33.3 EeRj64 intensive occupation 5 40.0 20.0 EeRj 100 5 80.0 100.0 EeRk52 limited a c t i v i t y 3 0.0 33.3 EcRg2AA residential camp 5 20.0 100.0 EcRg2CC residential camp 3 0.0 EcRg4C residential camp 4 25.0 0.0 EcRg4J f i e l d camp 3 66.7 EdRglA f i e l d camp 1 0.0 o EdRglB f i e l d camp 1 0.0 •> • EcRg4A f i e l d camp 4 0.0 0.0 EcRg4B residential camp 1 0.0 100.0 EcRg4D f i e l d camp 1 0.0 o EcRg4E station 1 100.0 o EdRg5 residential camp 4 50.0 EdRg6 f i e l d camp 2 50.0 *> • 245 (Price 1978; Camilli 1983). This prediction i s supported by the data provided i n Table 38. In Upper Hat Creek Valley, the mean number of ac t i v i t y areas (5.6) i n those sites identified as representative of an intensive occupation i s higher than the mean number of a c t i v i t y areas (3.0) i n those sites identified as special a c t i v i t y with expedient tool assemblages. The mean number (3.4) of ac t i v i t y areas i n those sites identified as residential camps i n Highland Valley i s higher than the mean number (2.0) of a c t i v i t y areas i n those sites identified as f i e l d camps. Residential camps should contain specialized a c t i v i t y areas (Camilli 1983). The data, as presented i n Table 38, p a r t i a l l y support this prediction. In Upper Hat Creek Valley, a l l sites contain specialized a c t i v i t y areas, except EeRj49 which has already been interpreted as a secondary deposition. In addition, those sites previously identified as representative of an intensive occupation with a wide range of a c t i v i t i e s (EeRj56, EeRj60, EeRj62, EeRj64, EeRj42) and camp sites (EeRilO) have a sl i g h t l y higher percentage of specialized a c t i v i t y areas than those sites identified as limited a c t i v i t y with expedient tool assemblages (EeRj55, EeRj8, EeRj20, EeRk52). In Highland Valley, half of the sites previously identified as residential camps contain a f a i r l y low percentage of specialized a c t i v i t y areas, while one third previously identified as f i e l d camps contain a high percentage of specialized a c t i v i t y areas. On the other hand, the short term use of sites i n the study areas, as recorded i n the ethnographic literature, w i l l undoubtedly influence the nature of a c t i v i t y area deposition. Residential camps may not have been occupied long enough for specialized a c t i v i t y areas to have been formed. In addition, the remains deposited at f i e l d camps may resemble those expected 246 i n a specialized a c t i v i t y area, because of the limited numbers and types of a c t i v i t i e s which occurred there. As well, the size of a c t i v i t y areas, particularly those from Highland Valley sites, exceeds the size of ethnographically documented a c t i v i t y areas. In these cases, the a c t i v i t y areas as defined i n this study may actually be more accurately interpreted as sites, or composites of ac t i v i t y areas. Finally, the a c t i v i t y areas i n residential camps should be spatially more discrete than those i n f i e l d camps (Camilli 1983; Binford 1978a). The data presented i n Table 38 appear to support this prediction. In Upper Hat Creek Valley, a l l sites previously c l a s s i f i e d as either representative of an intensive occupation or camp sites have some spatially discrete a c t i v i t y areas; i n contrast, only two of the three sites previously identified as limited a c t i v i t y sites have spatially discrete a c t i v i t y areas. In addition, the mean percentage of discrete a c t i v i t y areas i s s l i g h t l y higher i n intensively occupied sites (46.2%) than the mean percentage of discrete a c t i v i t y areas i n limited a c t i v i t y sites (44.4). However, the high number of Highland Valley sites that were systematically sampled present d i f f i c u l t i e s i n determining whether or not ac t i v i t y areas are discrete or blended. Those sites for which a determination i s impossible are marked with a "?". Two of the three sites identified as residential camps contain a c t i v i t y areas which are a l l spatially discrete. Only two of the eight sites identified as f i e l d camps have spatially discrete a c t i v i t y areas, but as mentioned above, the greatest proportion of these sites are sampled, making determination impossible. Settlement Types 247 Method This section focuses on the definition and c l a s s i f i c a t i o n of settlement types. In order to produce clusters based on s i t e function data, cluster analysis i s applied to presence/absence data on a c t i v i t y area types present i n the study sites (Table 39). The complete linkage clustering method was used on a matrix of Jaccard's coefficients calculated from the dichotomous data matrix. Activity area clusters are based on percentages of generalized manufacturing stages and Employable Unit types, and indicate the range of a c t i v i t i e s that occurred i n sites, as well as the relative number of redundant clusters. Again, microlithic debitage i s not included because i t i s necessary for the s i t e c l a s s i f i c a t i o n scheme to be independent of the presence or absence of microcore technology. Results The analysis produces six interpretable clusters, and one outlier, described below (see Figure 25). Microlithic sites are indicated with an asterisk (*). The input data for the cluster analysis are re-ordered according to cluster membership, and presented i n Table 39. Additional data used i n the following interpretation are presented i n Table 40: s i t e area, a r t i f a c t density, and a c t i v i t y area characteristics. Table 41 presents settlement type frequencies according to individual study areas. Cluster 1 sites, 2 microlithic and 2 non-microlithic, are located i n Highland Valley. A l l contain a c t i v i t y area Cluster 1, large, high-density areas representing a wide range of tasks. Two sites also contain a c t i v i t y 248 Table 39. Ac t i v i t y area cluster membership re-ordered by settlement cluster membership. Settlement Site Activity Area Clusters Clusters 1 2 3 4 5 6 7 8 9 1 EdRg6 + + *EdRgiB + •EdRglA + EcRg4A + + 2 EeRj42 + + + *EcRg4J + + *EcRg2AA + + 3 *EeRj55 + + + *EeRilO + + + EeRj20 + + EdRg + + + + *EeRj56 + + + + EeRj 64 + + + + 4 *EeRj62 + + + + *EeRj60 + + + EeRk52 + + *EcRg2CC + + 5 EcRg4D + EeRj 8 + 6 EeRj100 + + + EcRg4B + *EeRj49 + *EcRg4C + + o u t l i e r EcRg4E + *: indicates microlithic s i t e -1.000 I i 1.000 249 EdRg6 EdRglB* BdRglA* EcRg4A EeRj42 EcRg4J* EcRg2AA* EeRj55* EeRil0* EeRj20 . EdRg5 EeRj56* EeRj64 EeRj62* EeRj60* EeRk52 EcRg2CC* EcRg4D EeRj 8 EeRj100 EcRg4B EeRj49* EcRg4C* EcRg4E Figure 25. Complete linkage cluster analysis of s i t e attributes. 250 Table 40. Site characteristics re-ordered by settlement cluster membership. Settlement Site Site A r t i f a c t Number of Percentage of Clusters Area Density A c t i v i t y Specialized M2 /Mg Areas Areas 1 EdRg6 150 45.00 2 50.0 *EdRglB 60 34.50 1 0.0 *EdRglA 190 15.15 1 0.0 EcRg4A 90 2.91 4 0.0 2 EeRj42 2242 2.62 9 22.2 *EcRg4J 220 6.00 3 66.7 *EcRg2AA 125 73.49 5 20.0 3 *EeRj55 104 40.70 3 33.3 *EeRilO 352 3.99 5 20.0 EeRJ20 244 1.14 3 33.3 EdRgS 2000 0.06 4 50.0 * EeRj 56 892 3.62 6 83.3 EeRj64 372 0.92 5 40.0 4 * EeRj62 416 1.94 5 20.0 *EeRj60 1448 2.25 6 33.3 EeRk52 160 0.90 3 0.0 *EcRg2CC 850 7.94 3 0.0 5 EcRg4D 200 2.67 1 0.0 EeRj 8 88 5.52 3 66.7 6 EeRj100 160 1.02 5 80.0 EcRg4B 150 8.26 1 0.0 *EeRJ49 84 5.71 1 0.0 *EcRg4C 150 3.62 4 25.0 outlier EcRg4E 120 3.00 1 100.0 *: indicates microlithic s i t e Table 41. Settlement type frequencies i n study sample. Study Area Settlement Type 1 2 3 4 5 6 7 Upper Hat Creek Valley 0 1 5 3 1 2 0 Highland Valley 4 2 1 1 1 2 1 Type 1: F i e l d camp Type 2: Small residential camp Type 3: Large residential camp Type 4: Small residential camp Type 5: Fi e l d camp Type 6: Small residential camp Type 7: Station 252 area cluster 3, specialized locations where a limited range of maintenance a c t i v i t i e s occurred, and cluster 4, also large, generalized use areas. The low freguency of a c t i v i t y areas and the low percentage of specialized areas indicates that these sites are small f i e l d camps, where maintenance a c t i v i t i e s occurred and the processing of animal products, i n particular, took place. Cluster 2. contains 2 microlithic sites from Highland valley and 1 non-microlithic s i t e from Upper Hat Creek valley. A l l sites contain a c t i v i t y area cluster 1, as described above, and a c t i v i t y area cluster 6, specialized areas focusing on l i t h i c tool manufacture and maintenance. EeRj42 also contains two other types of a c t i v i t y area clusters: cluster 4, which i s again the locus of generalized tasks, and cluster 9, a specialized area where soft animal or vegetal products were worked. Thus, this s i t e may be a composite of two or more occupaions. Settlement cluster 2 i s interpreted as small residential camps, because of the f a i r l y high number of a c t i v i t y areas present, and the high percentage of specialized areas. Cluster 3 consists of 3 microlithic sites and 2 non-microlithic sites from Upper Hat Creek Valley, and 1 non-microlithic s i t e from Highland Valley. A c t i v i t y area cluster 5 i s present i n a l l sites; this cluster represents areas of generalized use where a variety of maintenance a c t i v i t i e s occurred. In addition, a c t i v i t y area cluster 6 i s present i n a l l sites; this cluster i s interpreted as specialized a c t i v i t y areas focusing on the manufacture and rejuvenation of l i t h i c tools. A l l sites i n settlement cluster 3 are large, with high numbers of a c t i v i t y areas, and are interpreted as large residential camps, where a wide variety of maintenance a c t i v i t i e s occurred, as well as an emphasis on tool manufacture. 253 Cluster 4 contains 2 microlithic and 1 non-microlithic s i t e from Upper Hat Creek Valley and 1 microlithic s i t e from Highland Valley. Generalized a c t i v i t y area clusters 4 and 5 are represented i n a l l sites; the microlithic sites i n Upper Hat Creek Valley also contain a low percentage of a c t i v i t y area cluster 7, interpreted as specialized locations for the secondary processing of animal products, core reduction, and tool manufacture and maintenance. Settlement cluster 4 i s also interpreted as small residential camps, similar to settlement cluster 2, but with less emphasis on the use of specialized areas. Cluster 5 comprises only 2 sites, both non-microlithic, one from each valley. The generalized a c t i v i t y area cluster 5 i s the only one present i n the Highland Valley s i t e ; a c t i v i t y area cluster 8, indicative of a few tasks focusing on non-lithic tool manufacture and resource processing i s also present i n the Upper Hat Creek Valley s i t e . These sites are also small and contain few a c t i v i t y areas. Settlement cluster 5 represents small f i e l d camps Where primarily maintenance a c t i v i t i e s occurred. Cluster 6 consists of two sites from each valley, one microlithic and one non-microlithic. These sites are similar to settlement cluster 2, i n that a generalized a c t i v i t y area cluster i s present i n a l l sites, with a few occurrences of several specialized clusters: cluster 2 where tool manufacture occurred, cluster 3 where a few maintenance a c t i v i t i e s took place, cluster 7 where resource processing and tool manufacture occurred, and cluster 8, where organic tools were manufactured. Settlement cluster 6 i s interpreted as small residential camps, because of the number of a c t i v i t y areas present, and the degree of specialization. The outlier i n this cluster analysis, EcRg4E, contains only a c t i v i t y area 254 cluster 2, a specialized location where tool and/or container manufacture and maintenance occurred. Thus, this s i t e i s interpreted as a station where tool manufacturing was carried out, but not the maintenance a c t i v i t i e s normally associated with a residential s i t e . This interpretation coincides with the original investigators' assessment of s i t e function (Areas Associates 1983). Discussion Camilli (1983) suggested that assemblages which result exclusively from a c t i v i t i e s performed during residential occupations may be uncemmon at the seasonal residential camps of mobile, logistically-organized hunter-gatherers. That i s , re-occupation of residential sites for l o g i s t i c a l tasks i s probably very common, and i t may be impossible to identify this kind of depositional history by analyzing l i t h i c assemblages. However, the use of more extensive analysis of ac t i v i t y area patterning, along with detailed information on tool manufacturing and maintenance and tool usage strategies i s used i n this study to produce a usable typology. The original goal of the settlement analysis was to establish a s i t e c l a s s i f i c a t i o n compatible with that predicted by the ethnographic model presented i n Chapter III. That scheme recognizes three basic s i t e types: residential camps, f i e l d camps or temporary camps, and stations or procurement and processing sites. The cluster analysis performed i n this study on a c t i v i t y area clusters produces six groups and one outlier. Clusters 2, 3, 4, and 6 are interpreted as residential camps; clusters 1 and 5 are interpreted as f i e l d camps; and the outlier i s interpreted as a station. 255 V a r i a b i l i t y among the kinds of residential camps derives from not only size, i n terms of the occupation area and the number of art i f a c t s deposited, but also from d i f f e r e n t i a l emphasis on tool use relating to the processing and consumption of various resources which are seasonally restricted i n av a i l a b i l i t y . Table 42 provides a cross-tabulation of settlement cluster types by biogeoclimatic zone. Cluster 3 sites, more predominant i n Upper Hat Creek Valley, are located i n the Interior Douglas F i r biogeoclimatic zone, with the exception of EeRj55, which i s located i n the Ponderosa Pine-Bunchgrass zone. According to the inferred ethnographic subsistence-settlement pattern presented i n Table 5 (Chapter III), the only residential sites expected i n the Interior Douglas F i r zone are f i e l d camps, occupied during early spring and summer while plants, ungulates, f i s h and waterfowl are procured and processed. Therefore, either cluster 3 sites are not representative of the ethnographic pattern, or, as Alexander (1989) suggests, larger sites are the result of re-occupation of favoured l o c a l i t i e s . Residential camps are expected i n the Ponderosa Pine-Bunchgrass zone, during summer months, while f l o r a l resources were gathered. Cluster 2 sites are located i n the Interior Douglas F i r zone. These sites are smaller than cluster 3 sites, and may more closely resemble the f i e l d camps inferred for the ethnographic period. In addition, cluster 6 sites are a l l located i n the same biogeoclimatic zone, and are interpreted as small residential camps. Again, the interpretation of cluster 2 sites as residential camps may indicate some deviation from the ethnographic subsistence-settlement pattern. Cluster 4 sites are located i n three biogeoclimatic zones: Interior Douglas F i r , Ponderosa Pine-Bunchgrass, and Engelmann Spruce-Subalpine F i r . Table 42. Summary of settlement types by biogeoclimatic zones. Biogeoclimatic Settlement Type Zone 1 2 3 4 5 6 7 Interior Douglas F i r 4 Ponderosa Pine-Bunchgrass 0 Engelmann Spruce-Subalpine F i r 0 3 5 2 1 4 1 0 1 1 0 0 0 0 0 1 0 0 0 Type 1: Fi e l d camp Type 2: Small residential camp Type 3: Large residential camp Type 4: Small residential camp Type 5: F i e l d camp Type 6: Small residential camp Type 7: Station 257 According to the inferred ethnographic subsistence-settlement pattern, residential camps are expected i n the latter two zones. Families occupied the Ponderosa Pine-Bunchgrass zone during mid-summer, to procure f l o r a l resources. The Engelmann Spruoe-Siibalpine F i r zone was exploited during early summer and early f a l l , for plants and ungulates. Those sites interpreted as f i e l d camps should also exhibit v a r i a b i l i t y , according to the season of occupation. Clusters 1 and 5, a l l located i n the Interior Douglas F i r biogeoclimatic zone, represent f i e l d camps, the inferred s i t e type for this zone during the f a l l , early spring, and early summer months. The major resources collected and processed at these times were respectively: ungulates; ungulates and early plants; and f i s h , ungulates, plants and waterfowl. Cluster 1 sites are a l l located i n Highland Valley and the processing of animal products i s indicated. Cluster 5 sites are located i n both valleys, and a r t i f a c t analysis indicates a variety of materials, including edible resources, were processed. The analyses of debitage, tools, a c t i v i t y areas and the resultant settlement typology, presented above, have provided the pertinent data base which w i l l be used to test the model of the organization of microcore technology. The research hypotheses associated with the model as i t pertains to the study area w i l l be tested i n the following chapter. 258 CHAPTER VT TEST OF MODEL This chapter constitutes a test of the model of the technological organization of microcore technology presented i n Chapter II. The model i s based on Binford's (1979) suggestion that the reduction strategies of tool manufacture and the form of the tools themselves w i l l vary according to the subsistence-settlement strategy selected at that particular time, for that particular area. The model predicts that microcore technology i s a technological strategy associated with the inferred ethnographic subsistence-settlement pattern i n the study areas; a maintainable tool assemblage; a transportable tool assemblage; and the conservation of l i t h i c raw material. The following section i s a test of the four research hypotheses, and their test implications. Evaluation of Research Hypotheses HYPOTHESIS 1: Microcore technology i s associated with the inferred  ethnographic subsistence-settlement pattern i n the study areas. Test Implications; 1. Microlithic sites w i l l be residential camps, while non-microlithic sites w i l l be special-purpose sites, either f i e l d camps or stations. In order to test this implication, the settlement types are collapsed into three types (residential camps, f i e l d camps, stations) i n order to examine the v a r i a b i l i t y between microlithic and non-microlithic sites further. According to the figures i n Table 43, over 80% of the microlithic 259 Table 43. Frequency and percentage counts of residential camps, f i e l d camps and stations i n study sample. Sites Residential Camps Fie l d Camps Stations Microlithic 10 (83.3%) 2 (16.7%) 0 Non-microlithic 7 (58.3%) 4 (33.3%) 1 (8.4%) Table 44. Frequency counts of settlement types i n Upper Hat Creek Valley. Sites 2 Settlement Type 3 4 5 6 Microlithic Non-microlithic 0 1 3 2 3 1 0 1 1 0 Type 2: Small residential camp Type 3: Large residential camp Type 4: Small residential camp Type 5: F i e l d camp Type 6: Small residential camp Table 45. Frequency counts of settlement types i n Highland Valley. Sites 1 Settlement Type 2 3 4 5 6 7 Microlithic Non-microlithic CM CM 2 0 0 1 1 0 0 1 1 1 0 1 Type l Type 2 Type 3 Type 4 Type 5 Type 6 Type 7 Fie l d camp Small residential camp Large residential camp Small residential camp Field camp Small residential camp Station 260 sites are c l a s s i f i e d as residential camps, while only 58% of the non-microlithic sites are c l a s s i f i e d as residential camps. In order to test the s t a t i s t i c a l significance of the association between microcore technology and settlement type, the single station was removed from the sample, and a Fisher Exact Test i s applied to the following n u l l hypothesis: Ho: There i s no association between technological s i t e type (microlithic and non-microlithic) and settlement type. Fisher's Exact Test i s selected because the sample size i s between 20 and 40, and the expected value for one or more c e l l s i s less than 5 (Thomas 1976; Wilkinson 1988). The result i s a probability level of 0.28, indicating that the relationship i s l i k e l y due to chance. In addition, a Phi-square s t a t i s t i c of 0.05 indicating a very weak relationship supports this result. The next step i s an investigation of how the association between microcore technology and settlement type may vary between valleys. Tables 44 and 45 provide the frequencies of sites c l a s s i f i e d into each of the seven settlement types i n Upper Hat Creek valley and Highland Valley. Although, again, the expected c e l l values are too small for a significance test, there are some differences i n the distribution. Microlithic sites i n Upper Hat Valley f a l l into only three types: large residential camps, and two varieties of small residential camps. Fiel d camps and stations are not present i n the study sample. As a contrast, microlithic sites i n Highland Valley f a l l into four types: f i e l d camps, and three types of small residential camps. Thus, microcore technology i n Upper Hat Creek Valley sites does f i t the predicted pattern of being associated with residential camps rather than f i e l d camps or stations. But i n Highland Valley sites, microcore technology does not f i t the predicted pattern of being associated 261 only with residential camps. A corollary to the prediction relating to the settlement type associated with microcore technology i s that non-microlithic sites w i l l be either f i e l d camps or stations, that i s , functional variants i n a synchronic subsistence-settlement pattern. Examination of Tables 46 and 47 demonstrates that this prediction i s not true for either valley. Non-microlithic sites i n Upper Hat Creek Valley are c l a s s i f i e d as both residential camps and i e l d camps; i n addition, the majority of non-microlithic sites are residential camps. In order to determine i f the settlement type membership of the microlithic groups i s significantly different from that of the non-microlithic group i n Upper Hat Creek Valley, the n u l l hypothesis i s tested on data i n Table 46: Ho: There i s no association between technological s i t e type (microlithic and non-microlithic) and settlement type. Fisher's Exact test produces a probability level of 1.00, and indicates that there i s no significant association between microcore technology and settlement type. The Phi-squared s t a t i s t i c (Wilkinson 1987), which has maximum value of 1.00, i s 0.09 indicating a very weak association between microcore technology and settlement type. The situation for non-microlithic sites i n Highland Valley i s s l i g h t l y different, with the majority being c l a s s i f i e d as f i e l d camps and stations rather than residential camps. However, contrary to the prediction, non-microlithic sites f a l l into a l l three settlement types. 2. Microlithic sites w i l l be located i n the appropriate biogeoclimatic zones. 262 Table 46. Frequency counts of residential camps and f i e l d camps i n Upper Hat Creek Valley. Sites Residential Camps Fie l d Camps Microlithic 6 0 Non-microlithic 5 1 11 1 Fisher Exact Test P = 1.00 Phi-square = 0.09 Table 47. Frequency counts of residential camps and f i e l d camps i n Highland Valley. Sites Residential Camps Field Camps Microlithic 4 2 Non-microlithic 2 3 263 Table 48 compares the archaeolccricail subsistence-settlement pattern, asinterpreted from results i n Chapter V, with the inferred ethnographic subsistence-settlement pattern, presented ear l i e r i n Table 5, Chapter III. Four Upper Hat Creek Valley residential microlithic sites (EeRilO, ErRj56, EeRj 49, ErRj62) are situated i n the Interior Douglas F i r zone, which should only contain f i e l d camps and stations. The latter two are small residential sites and could conceivably be mis identified f i e l d camps. In addition, the other two Upper Hat Creek Valley microlithic sites, which are also interpreted as residential, are located i n the Ponderosa Pine-Bunchgrass Zone, where residential camps and stations were occupied during the ethnographic period. Although no stations i n the study sample are located i n this biogeoclimatic zone, the limited sample size may be a factor as well as sampling bias toward selection of sites containing projectile points. I f the microlithic sites i n Upper Hat Creek Valley a l l date to the Plateau horizon, as may EeRj55, then the subsistence-settlement pattern represented by these sites d i f f e r s from that of the ethnographic period. One plausible interpretation i s a more intensive use of the Interior Douglas F i r zone during the Plateau Period indicated by large residential camps, associated with intensified collection and processing of resources. This supports the hypothesis put forward by Pokotylo and Froese (1983) after intensive investigation of a series of root roasting ovens. If the microlithic sites are earlier chronologically than the non-microlithic sites, then again the subsistence-settlement pattern represented by microlithic sites d i f f e r s from that of the ethnographic period. But i n this case, the existence of groups of residential camps without corresponding f i e l d camps and stations indicates a subsistence procurement 264 Table 48. Comparison of archaeological subsistence-settlement pattern with inferred ethnographic subsistence-settlement pattern. Site Biogeo- Archaeological Cultural- Inferred climatic Settlement Historical Ethnographic Zone Type Association Settlement Types Upper Hat Creek valley Microlithic Sites EeRilO IDF large RC FC, S EeRj49 IDF small RC FC, S EeRj55 PPBG large RC Plateau RC, S EeRj56 IDF large RC FC, S EeRj60 PPBG small RC RC, S EeRj62 IDF small RC PC, S Upper Hat Creek Valley Non-microlithic Sites EeRj 8 IDF FC Kamloops FC, S EeRj20 IDF large RC Plateau FC, S EeRj42 IDF small RC Plateau PC, S EeRj64 IDF large RC Kamloops PC, S EeRj100 IDF large RC Plateau PC, S EeRk52 ESSF small RC Plateau RC, FC, S Highland Valley Microlithic Sites EcRg2AA IDF small RC Quiltanton RC?, FC, S EcRg2CC IDF small RC Quiltanton RC?, PC, S EcRg4C IDF small RC Quiltanton RC?, FC, S EcRg4J IDF small RC Quiltanton RC?, FC, S EdRglA IDF FC Quiltanton RC?, FC, S EdRglB IDF FC Plateau RC?, FC, S Highland Valley Non-microlithic sites EcRg4A IDF PC RC?, FC, S EcRg4B IDF small RC Plateau RC?, PC, S EcRg4D IDF PC Kamloops RC?, FC, S EcRg4E IDF S Plateau RC?, PC, S EdRgS IDF large RC Kamloops RC?, FC, S EdRg6 IDF FC Kamloops RC?, FC, S RC: Residential camp FC: Fiel d camp S: Station IDF: Interior Douglas F i r PPBG: Ponderosa Pine-Bunchgrass ESSF: Engelmann Spruce-Subalpine F i r 265 strategy which i s primarily foraging rather than collecting. I t has been suggested (Kujit 1988; Fladmark 1986) that, prior to the large-scale procurement of salmon resources, subsistence was based on the hunting of ungulates, as well as other small mammals and f l o r a l resources. Prior to 4500 B.P., Upper Hat Creek Valley may have been occupied by more mobile hunter-gatherers, relying on abundant ungulate resources i n the more extensive grasslands. Another p o s s i b i l i t y that must be considered i s that vegetation boundaries have most l i k e l y shifted several times since 4500 B.P.. Although paleobotanical data, discussed i n Chapter III, indicate that modern vegetation patterns were established i n Upper Hat Creek Valley by 4500 B.P., i n i t i a l l y , the climate may have been moister and cooler. In addition, two neoglacial advances which occurred between 3500 and 2000 B.P., and after 900 B.P. probably had the same effect, while the climate during the interval between the advances was presumably warmer and drier. During cooler, wetter periods, the boundary between the Ponderosa Pine-Bunchgrass zone and the Interior Douglas F i r zone probably shifted to below the modern transitional zone. Thus, the residential camps currently situated i n the Interior Douglas F i r zone may very well have been located prehistorically i n the Ponderosa Pine-Bunchgrass zone, where residential camps are predicted, according to the inferred ethnographic subsistence-settlement pattern. Therefore, evaluation of this test implication must await more detailed paleoenvironmental data, and a more secure dating of microlithic sites. The situation i n Highland Valley i s less clear due to equivocal data on the scale of ethnographic exploitation of the valley. Archival records indicate very l i t t l e aboriginal interest i n this area, while more recent 266 interviews suggest the valley may have been a more integral part of the regional subsistence-settlement pattern i n the prehistoric past. According to Alexander (1989), the lake zone within the Interior Douglas F i r zone would have been the locus of intensive f i s h and waterfowl procurement, from residential base camps re-occupied annually. While the archaeological evidence from the areas surveyed to date i n Highland Valley do not indicate substantial re-occupation of sites, Areas Associates (1983, 1986) do predict the use of residential camps, f i e l d camps and stations i n the Interior Douglas F i r zone. Table 48 indicates that a l l three settlement types are found i n the Highland Valley lake zone. However, the only station i s a non-microlithic s i t e . Even i f the dates assigned by Areas Associates (1986) to the Quiltanton complex, which may co-exist with the Plateau horizon, are accurate, the subsistence-settlement pattern represented by the microlithic sites i n Highland Valley does resemble that inferred for the ethnographic present. That i s , the resource procurement strategy i s primarily l o g i s t i c a l . In addition, the subsistence-settlement pattern represented by the non-microlithic sites i n Highland Valley also resembles that of the ethnographic period. Therefore, the situation i n Highland Valley i s unlike that i n Upper Hat Creek Valley where there i s evidence of a different strategy involving either intensification of the Interior Douglas F i r zone resources, or a primarily foraging resource procurement strategy. Finally, the p o s s i b i l i t y must also be considered that biogeoclimatic zones have also shifted i n Highland Valley. Again, during the two f i n a l neoglacial advances, cooler, moister conditions probably increased the amount of forest cover, and may affected the vegetation of the Interior 267 Douglas F i r zone around the valley bottom lakes to the extent that at least lower sections became Ponderosa Pine Bunchgrass instead. The study sites are located i n such close proximity that a change i n vegetation boundary would have affected the entire sample. However, lack of ethnographic data on predicted settlement types i n either biogeoclimatic zone of this valley means that consideration of movement of the contemporary grassland/forest border does not actually aid i n the evaluation of the test implication. HYPOTHESIS 2; Microcore technology i s designed to contribute toward a  maintainable tool assemblage. 1. Microblades are a more flexible tool than other tool types. Shott's (1986) definition of f l e x i b i l i t y refers to a tool class that i s used i n a wide range of tasks. In order to measure this tool attribute, the type of EU's recorded as present on each morphological tool class i s presented i n Tables 49, 50, and 51, for the t o t a l s i t e sample, microlithic sites, and non-microlithic sites respectively. Microlithic tools are a very fl e x i b l e tool, displaying a wide range of Employable Unit types (see Table 49). Over the t o t a l s i t e sample, microblades display 12 Employable Unit types, which i s fewer than only two tools: modified flakes and bifaces. An important difference between the two valleys i s also evident: microlithic tools i n Upper Hat Creek Valley display fewer EU types than do the same tools i n Highland Valley. Thus, the microlithic tools i n Highland Valley are interpreted as more flexible than i n Upper Hat Creek Valley. However, the most fle x i b l e tool i n both microlithic and non-microlithic assemblages i s the modified flake. Although microblades do display a large number of Employable Units, Table 49. Presence-absence of EU types on morphological tool types. 268 Tool Type Employable Unit Type 1 2 3 4 5 6 7 8 9 10 11 1 2 1 3 14 15 16 Microblade + + + + + + + + + + + + Flake + + + + + + + + + + + + + + + + Uhiface + + + + + + + + Biface + + + + + + + + + + + + + + Graver + + + + + + Cobble + + + + Preform + + Table 50. Preserice-absence of EU types on morphological tool types i n microlithic sites. Tool Type Employable Unit Type 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Microblade + + + + + + + + + + + + Flake + + + + + + + + + + + + Uhif ace + + + + + + + Biface + + + + + + + + + + + + Graver + + + + + + Cobble + + + Preform + + Table 51. Presence-absence of EU types on morphological tool types i n non-microlithic sites. Tool Type Employable Unit Type 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Flake + + + + + + + + + + + + + + Uhif ace + + + + + Biface + + + + + + + + + + + Graver Cobble + Preform 269 Table 52 demonstrates that nearly 50% of microblades c l a s s i f i e d as modified display Eu9, inferred to represent the scraping of soft materials. On the basis of the entire range of ED types which appear on microblades, they can be considered a f l e x i b l e tool. However, microblades appear to be closely associated with a specific function, particularly i n Upper Hat Creek Valley. In addition, the wear traces resulting from the original use may not be v i s i b l e except under high-power magnification (Michael Blake, personal communication 1990). To test this hypothesis, three microblades from Upper Hat Creek Valley study sites were re-examined under high-power magnification by Dr. Michael Blake and the author, at the Laboratory of Archaeology, University of B r i t i s h Columbia. The following artifacts were examined and findings recorded; interpretations of use are those given by Dr. Blake. 1. EeRj49:266 unused microblade X40 - surface "glossy" X80 - edge "rounded?",slightly dulled X160 - p a r a l l e l striations on top of flake scars interpreted as used to cut meat 2. EeRilO:85 modified microblade proximal fragment X160 - unmodified edge uneven, lumpy, no gloss, no rounding X40 - modified edge has flakes off but not dulled X320 - modified edge - same as X40 interpreted as used i n scraping 3. EeRj49:24 unused microblade proximal fragment X80 - rounded edge? X160 - edge dulled X320 - edge polished interpreted as non-abrasive use-wear The implication of the results of this extremely brief use-wear analysis i s that probably at least some of the microblades c l a s s i f i e d as unused i n this study were, i n fact, used. Microblades with v i s i b l e use-wear interpreted as scraping soft materials may be those which have been very heavily used i n this a c t i v i t y . As well, some of the microblades c l a s s i f i e d i n this study as Table 52. Percentage counts of EU types on morphological t o o l types. Tool Type • Employable Unit Type 1 2 3 4 5 6 7 8 9 IB 11 12 13 14 15 16 Mi c r o l i t h i c Sites Microblade 8.6 6.2 B.4 2.4 B.8 8.2 7.1 22.2 58.8 B.4 B.2 B.a Flake B.4 3.9 1.1 12.5 3.B 2.7 8.1 2B.3 41.6 1.8 2.1 B.l 1.6 Uniface 11.1 3.7 3.7 59.3 3.7 14.8 3.7 Biface 3.5 4.7 9.4 4.7 21.2 21.2 17.6 4.7 2.4 4.7 1.2 1.2 3.5 Graver 15.4 7.7 7.7 3B.8 38.8 7.7 Cobble 22.2 11.1 11.155.5 Prefort 58.B 5B.B Non- t i c r o l i t h i c S i t e s Flake B.9 8.9 3.6 25.9 5.3 8.2 18.2 11.5 22.7 2.3 5.3 8.3 8.3 1.6 8.3 8.3 Uniface 38.B 1B.B 38.8 1B.B 2B.B Biface 1.6 9.5 6.3 3B.2 23.8 7.9 6.3 9.5 1.6 1.6 1.6 Graver Cobble 1BB.B Pref o r i 271 as single purpose, with only type of ED, may be multi-purpose, with the finer work, e.g. meat or f i s h s l i c i n g , being finished f i r s t , and then the coarser work, e.g. scraping plants or graving bone, being completed with an already s l i g h t l y dulled edge. The actual extent of this type of use-wear should be investigated on a large sample of microblades, from a variety of locations, under high power magnification. Until that time, the results of the tests carried out above to determine the relative f l e x i b i l i t y should be qualified to indicate that microblades do have the potential to be flexible. 2. Tool assemblages at microlithic sites are more fle x i b l e than those at non-microlithic sites. Considering the assemblages as a whole, microlithic tool assemblages, which contain a larger number of morphological tool types than do non-microlithic tool assemblages, were used for the same range of task applications types (see Tables 50 and 51). Thus, on the basis of this investigation, microlithic assemblages are interpreted as less flexible than non-microlithic assemblages. 3. Microblades are a more versatile tool than other tool types. According to shott's (1986) definition, versatile tools are those which have the greatest range of task applications. This can be measured by defining the number and type of ED's present on each tool, to determine the percentage of tools that were used i n more than one task (multi-purpose tools). A l l tools were cl a s s i f i e d as either single purpose or multi-purpose i n Chapter V, based on the number and type of Employable Unit present. 272 Table 53 provides percentages of those morphological tool types that are c l a s s i f i e d as versatile tools. Microblades constitute the least versatile class. In Upper Hat Creek valley microlithic sites, 82.5% of the microlithic tools are single purpose, while i n Highland valley, 70.5% of microlithic tools are single purpose. In microlithic sites, gravers, unifaces and bifaces are the most versatile classes, while i n non-microlithic sites, bifaces and unifaces are the most versatile classes. One explanation for why microblades do not appear to be a versatile, that i s , multi-purpose tool class, i s that the wear traces for the original intended use are not v i s i b l e except under high-power magnification, as already discussed above. Until further research c l a r i f i e s this issue, the results of the tests carried out above to determine the relative v e r s a t i l i t y should be qualified to indicate that microblades do have the potential to be versatile. 4. Tool assemblages at microlithic sites are more versatile than those at non-microlithic sites. Table 54 provides the frequency and percentage counts of single and multi-purpose tools for microlithic and non-microlithic sites i n the two valleys. In order to determine i f there is a significant association, a n u l l hypothesis was formulated and tested with the Chi-square test of independence: Ho: there i s no association between technological s i t e type and tool type (single and multi -purpose). Results are shown i n Table 54. The Chi-square s t a t i s t i c i s significant at a 0.05 level of probability, therefore the n u l l hypothesis i s rejected. 273 Table 53. Percentage counts of morphological tool types that are versatile. Site Microblade Flake Cobble Uniface Graver Biface Preform Microlithic 23.5 22.4 0.0 40.0 85.7 37.2 0.0 Non-microlithic 27.2 0.0 50.0 51.6 Table 54. Chi-square test for significance of association between technological type and versatile tools. Technological Tool Type Site Type Single Purpose Multi-purpose Microlithic 809 251 1060 (76.3%) (23.7%) Non-microlithic 176 83 259 (67.9%) (32.1%) 985 334 1319 Chi-square = 7.706 P = significant at 0.05 Phi-square = 0.006 274 However, the Phi-square s t a t i s t i c indicates a very weak association, so the significance of the chi-square test may be due to the large sample size. Although the distribution i s s t a t i s t i c a l l y significant, i t i s exactly opposite to that predicted by the model. Microlithic sites contain a smaller percentage of versatile tools than non-microlithic sites (see Table 54). One plausible explanation for this result i s that microblades, which do not appear to be multi-purpose but do display several different types of Employable Units, constitute a suitable replacement for multi-purpose tools. Although a wide range of tasks i s indicated for microblade use (see Table 52), further exploration of the association between microcore technology and the dominant tasks, interpreted from tool analysis, carried out at the sites i s warranted. Pokotylo (1978) suggested that the mutually exclusive distribution of microblades and formed unifaces i n his sample indicates a dichotomy between cutting and scraping tool use. Earlier analyses (Table 8, Chapter XV) do not support the observation that unifaces are not found i n microlithic sit e s . However, there i s an interesting dichotomy between cutting and scraping tool use that i s further c l a r i f i e d by EU analysis. Table 55 presents the percentage values of the most frequent EU types present i n both microlithic and non-microlithic sites. The dominant tasks i n microlithic sites are scraping and cutting soft materials; i n contrast, the dominant task i n non-microlithic sites i s scraping soft-medium materials, although scraping soft materials, and cutting hard-medium and hard materials are also prominent. Thus the difference between the two types of sites relates not to the actual motions involved but to the nature of the worked materials. This result could be due to seasonal v a r i a b i l i t y i n sites occupations, or d i f f e r e n t i a l emphasis on resources, and therefore tasks Table 55. Percentage counts of the most frequent EU types. 275 EU Motion Worked Material Microlithic Non-microlithic Sites Sites 3 graving/boring soft 4.5 1.0 9 scraping soft 43.4 18.7 4 scraping soft-medium 8.6 21.9 11 scraping hard 2.8 6.3 8 cutting soft 20.3 10.6 5 cutting soft-medium 2.5 6.3 7 cutting hard-medium 7.3 13.1 6 cutting hard 3.3 13.1 10 scraping hard-medium 2.9 2.7 2 graving/boring soft 4.4 1.0 Table 56. Frequency counts of ac t i v i t y area types. Technological Activity Area Clusters Type 1 2 3 4 5 6 7 8 9 Microlithic 12 1 0 8 2 3 1 0 1 Non-microlithic 6 3 3 14 13 8 5 3 1 Table 57. Chi-square test for significance of association between technological type and ac t i v i t y area type. Activity Area Type Technological Generalized Specialized Site Type Microlithic 22 6 28 Non-microlithic 33 23 56 55 29 84 Chi-square = 3.186 P = significant at 0.074 Phi-square = 0.03 276 performed at the sites. As further evidence, Table 56 provides a frequency count of microlithic and non-microlithic a c t i v i t y areas present i n each a c t i v i t y area cluster. Cluster 1 i s the most numerous among microlithic a c t i v i t y areas; this cluster contains a wide variety of EU's, but the dominant ones are EU7, EU8 and EU9 (scraping soft materials, and cutting soft and hard-medium materials. Clusters 4 and 5 are the most numerous among non-microlithic a c t i v i t y areas: cluster 4 contains a wide variety of EU types, but emphasis i s placed on EU's 4, 5, 7 and 8 (scraping soft-medium materials, cutting soft-medium materials, cutting hard materials and cutting soft materials); while cluster 5 contains a higher proportion of EU's 4, 6 and 11 (scraping soft-medium materials, cutting hard materials and scraping hard materials. Again, the differences between the two types of a c t i v i t y areas relate to the type of materials being worked, with soft and soft-medium materials being dominant i n microlithic a c t i v i t y areas and harder materials being dominant i n non-microlithic a c t i v i t y areas. The expected c e l l values i n this table are too low for a significance test. In order to produce a table with high enough expected c e l l values for significance testing, the a c t i v i t y area clusters are collapsed into two types, generalized and specialized. This analysis i s based on the assumption that assemblages designed for task v e r s a t i l i t y w i l l also exhibit a larger number of EU types, that i s , w i l l be generalized rather than specialized. Table 57 provides results of a Chi-square test carried out on a frequency count of the number of microlithic and non-microlithic a c t i v i t y areas present i n each of these two types. Approximately 78% of the microlithic a c t i v i t y areas are generalized, while only 58% of the non-microlithic 277 a c t i v i t y areas are generalized. The Chi-square s t a t i s t i c i s significant at a probability level of 0.074, indicating that this distribution may not be due to chance. However, the Phi-square s t a t i s t i c i s 0.03, indicating a very weak association between microoore technology and generalized a c t i v i t y areas. HYPOTHESIS 3; Microcore technology i s designed to contribute toward a  transportable tool assemblage. 1. Complete formed tools i n microlithic sites w i l l be lighter and smaller than complete formed tools at non-microlithic sites. In order to determine i f this test proposition i s true, the weight, measured i n grams, and size (length, width, and thickness), measured i n millimetres, were recorded for a l l complete formed tools. Table 58 provides the median for each of these measures. Complete formed tools from microlithic sites are, on the average, lighter and smaller than those from non-microlithic sites. In order to determine i f these measurements discriminate significantly between the two samples of tools from microlithic and non-microlithic sites, a Mann-Whitney two-sample test i s carried out a data f i l e that contains weight and a size index, calculated by multiplying length x width x thickness/100. The results of the test are provided i n Tables 59 and 60. The Mann-Whitney s t a t i s t i c , i n both cases, i s not significant at the 0.05 level. Thus, these two measures of size do not discriminate between complete tools from microlithic and non-microlithic sites. 2. Evidence of microcore manufacture w i l l not be found at a l l sites containing microblades. 278 Table 58. Median of weight and size of complete formed tools (N=53). Site Type Weight (gm) Length (cm) Width (cm) Thickness(cm) Microlithic 4.55 32.30 23.90 6.50 Non-microlithic 18.78 42.77 32.05 10.73 Table 59. Mann-Whitney two-sample tests on weight. Attribute Rank Sum U Probability Microlithic Non-microlithic S t a t i s t i c Weight 737.0 694.0 302.0 0.411 Table 60. Mann-Whitney two-sample tests on size. Attribute Rank Sum U Probability Microlithic Non-microlithic S t a t i s t i c Size 725.0 706.0 290.0 0.300 279 Table 61 provides frequency counts of microlithic ar t i f a c t s i n the study sites. A l l sites contain complete microblades and/or microblade fragments. However, only seven of the twelve sites contain microcores, indicating that the microblades were manufactured from a core which was later removed from the s i t e . Other microlithic debitage which clearly indicates the preparation or rejuvenation of microcores includes microcore preparation flakes and microcore rejuvenation flakes (Kelly 1984). Only two of the sites contain microcore preparation flakes, supporting the test implication that microcore preparation did not occur at a l l sites where microblade production occurred. Both sites are located i n Highland Valley, indicating a difference i n the organization of microcore technology between the two valleys. In addition, both sites where microcores were prepared are small residential camps (EcRg2AA, EcRg4J). One f i e l d camp (EdRglA) contains a microcore rejuvenation flake, but no microcores or preparation flakes, indicating that one microcore was rejuvenated there, presumably for the further production of microblades. The archaeological pattern evident i n Highland Valley does support the model, which predicts that microcores w i l l be manufactured only at residential camps. No sites i n Upper Hat Creek Valley contain microcore preparation flakes, suggesting that microcores were prepared elsewhere i n the regional settlement pattern. Again, this pattern does f i t the model, since microcores may have been prepared at the main winter base camps i n the major ri v e r valleys, as suggested i n the model i n Chapter II. 3. Some microlithic sites w i l l contain microblades belonging to different stages of the production sequence. Table 61. Frequency counts of microlithic a r t i f a c t s . Site Microcores and Rejuvenation Preparation Microblades Fragments Flakes Flakes and Fragments Upper Hat Creek Valley sites EeRilO 2 103 EeRj49 1 3 88 EeRj55 58 EeRj56 1 7 EeRj60 1 7 EeRj62 2 2 12 Highland Valley sites EcRg2AA 20 22 57 718 EcRg2CC 1 2 EcRg4C 23 EcRg4J 1 1 22 EdRglA 1 44 EdRglB 57 281 A manufacturing stage typology for microblade production was not constructed i n this study, as planned, due to the limited nature of the experimental data base. Therefore, this implication remains to be tested i n future work. 4. Microcores found i n the study sites w i l l not be viable. Table 62 provides frequency and percentage data for types of microcores present i n study sites. Only one s i t e , i n Highland Valley, contains a potentially viable microcore. Two other sites, one i n each valley, contain blanks with enough platform surface to be fashioned into productive microcores. The two dominant categories of microcores, exhausted and rejects, occur i n both valleys, at six sites. These results indicate that microcores are l e f t at the locus of use primarily because they are no longer, or never were, viable. That i s , exhausted microcores were discarded because further rejuvenation was impossible, and rejected microcores were discarded because of unsuccessful preparation. 5. Microlithic sites w i l l have a very high micax>blade/microcore ratio. The next step i s an estimate of whether or not the itdcrooores l e f t at the sites were used to produce the microblades also deposited there. The frequency of complete microblades and microblade proximal fragments, both mo