?Copyright 2012 Amanda K. Taylor University of Washington Abstract Creating and Transcending Territorial Boundaries in Late Holocene Pacific Coast Communities Amanda K. Taylor Chair of Supervisory Committee: Dr. Julie K. Stein Anthropology In this research, I investigate precontact territorial behavior in the San Juan Islands, Washington and San Nicolas Island, California. Drawing on economic defensibility models, I generate hypotheses for change over time in boundary defense and permeability in the context of Late Holocene climate and settlement pattern change. Defensive characteristics of archaeological sites and lithic procurement patterns should reflect increased boundary defense and smaller territories when resources are adequate to the needs of the community. Extra-local materials should increase in abundance during times of resource scarcity. For the San Juan Islands case study, data on visibility, elevation, and distance to lookouts do not indicate significant changes through time in site location consistent with changes in boundary defense. Dissimilarities between artifacts from the Watmough Bay site (45-SJ-280) on Lopez Island and and nearby beach cobble toolstone suggests lithic procurement beyond the local beach at 1600-1000 cal BP, consistent with predictions for minimal boundary defense due to resource scarcity at that time. I do not find increased use of nearby beach cobbles or slate at 600 cal BP-Contact during a period of predicted increased boundary defense. Data on the spatial and temporal distribution of extra-local materials are insufficient to evaluate hypotheses regarding boundary permeability. For the San Nicolas Island case study, data on elevation and distance to lookouts do not indicate significant changes through time in site location consistent with changes in boundary defense. For the lithic procurement study using data from Tule Creek Village (CA-SNI-25) Mound B and CA-SNI-106, I found few changes through time in material type or artifact dimensions at either site that would indicate shifts in cobble procurement locations. Increases in abundance of chert are correlated with increases in sample size. For both study areas, results indicate potential sample size issues that must be resolved to further investigate my predictions. People may have engaged in boundary defense at a larger scale than the village; community boundaries may have always been permeable to interactions between kin and friends. This study has implications for social complexity studies and research on human adaptations to resource abundance and scarcity. Creating and Transcending Territorial Boundaries in Late Holocene Pacific Coast Communities Amanda K. Taylor A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy University of Washington 2012 Reading Committee: Julie K. Stein, Chair J. Benjamin Fitzhugh Angela E. Close Ren? L. Vellanoweth Program Authorized to Offer Degree: Anthropology i TABLE OF CONTENTS List of Figures ................................................................................................................. iv List of Tables ................................................................................................................ viii Chapter 1: Introduction .....................................................................................................1 Territoraility in Coastal Settings .....................................................................3 Ethnographic Research on Territoriality in the San Juan Islands ....................5 Ethnographic Research on Territoriality in the Southern Channel Islands ....10 Theoretical Perspectives on Territoriality .................................................... 12 Human Behavioral Ecology Studies of Boundary Defense ...........................14 Boundary Permeability ...................................................................................18 Hypotheses for Precontact Territoriality on the Pacific Coast of North America .........................................................................................................23 Predictions ......................................................................................................27 Dissertation Organization ...............................................................................28 Chapter 2: Pacific Coast Paleoenvironments and Settlement Patterns ...........................30 The San Juan Islands Study Area ..................................................................31 Gulf of Georgia Paleoenvironment ...............................................................33 Marine Paleoenvironment in the Gulf of Georgia .........................................34 Terrestrial Paleoenvironment in the Gulf of Gerogia ....................................35 Gulf of Georgia Marine Resources ...............................................................37 Gulf of Georgia Terrestrial Resources ...........................................................39 San Juan Islands Settlement Patterns .............................................................41 Territoriality Predictions for the San Juan Islands .........................................48 San Juan Islands Defensive Sites ...................................................................50 The Southern Channel Islands Study Area and San Nicolas Island ...............54 Southern Channel Islands Paleoenvironment .................................................57 Marine Paleoclimate in the Channel Islands .................................................57 Marine Resources and Climate Change in the Channel Islands ....................60 Terrestrial Resources and Climate Change in the Channel Islands ................61 Settlement Pattern Data for San Nicolas Island .............................................62 Territoriality Predictions for San Nicolas Island ............................................70 Defensive Sites on San Nicolas Island ...........................................................71 Conclusions ...................................................................................................75 Chapter 3: Stratigraphic Context and Dating of Lithic Assemblages ............................ 77 Excavations at Watmough Bay .....................................................................77 Stratigraphy and dating at Watmough Bay ...................................................80 Temporal Analytic Units for Watmough Bay ...............................................88 The San Nicolas Island Sites ..........................................................................92 Excavations at Tule Creek Village, Mound B ................................................93 Stratigraphy and Dating at Mound B . .......................................................... 94 Temporal Analytic Units for Mound B ....................................................... 108 Excavation, Stratigraphy and Dating at CA-SNI-106 ..................................110 Chapter Summary .........................................................................................112 Chapter 4: Results of Toolstone Surveys ..................................................................... 113 The Lithic Landscape of the San Juan Islands ............................................114 ii Cobble Surveys in the San Juan Islands ......................................................125 San Juan Islands Toolstone Summary Statistics ........................................ 133 The Lithic Landscape of San Nicolas Island ................................................138 Cobble Surveys on San Nicolas Island ........................................................146 San Nicolas Island Toolstone Summary Statistics ..................................... 155 Quarry Atractiveness ....................................................................................163 Calculations and Conclusions ......................................................................170 Chapter Summary .........................................................................................171 Chapter 5: Flaked Stone Technology on the Pacific Coast ..........................................173 General Analytic Methods ...........................................................................173 The Watmough Bay Lithic Assemblage .....................................................176 San Nicolas Island Lithic Analysis .............................................................203 Calculations and Comparisons .....................................................................223 Chapter 6: Toolstone Procurement Predictions and Lithic Analysis ........................... 225 Toolstone Procurement and Processing Predictions ...................................226 Analytic Methods: Reduction Sequence Analysis .......................................230 Testing Model Predictions: Toolstone Conservation ................................. 231 Testing Model Predictions: Exchange ..........................................................233 Procurement Predictions for the Watmough Bay Site ..................................234 Testing Procurement Predictions at Watmough Bay: Material Type ...........236 Testing Procurement Predictions at Watmough Bay: Size, Shape and Cortex Appearance .......................................................................................242 Testing Procurement Predictions at Watmough Bay: Processing ................249 Toolstone Conservation at Watmough Bay ..................................................252 Exchange at Watmough Bay ........................................................................253 Summary of Toolstone Procurement Results for Watmough Bay ...............255 Evaluating an Alternative Hypothesis for Lithic Procurment at Watmough Bay ................................................................................................................256 Procurement Predictions for San Nicolas Island ..........................................258 Testing Procurement Predictions on San Nicolas Island: Material Type .....261 Testing Procurement Predictions on San Nicolas Island: Cobble Shape .....270 Testing Procurement Predictions on San Nicolas Island: Processing Prior to Transport ..................................................................................................276 Toolstone Conservation on San Nicolas Island ............................................280 Exchange on San Nicolas Island ..................................................................285 San Nicolas Island Lithic Procurement Summary .......................................288 An Alternative Hypothesis for Lithic Procurement on San Nicolas Island .290 Chapter Summary .........................................................................................291 Chapter 7: Conclusions .................................................................................................294 Summary of Results for the San Juan Islands .............................................296 Summary of Results for San Nicolas Island ................................................303 Territoriality and Scale ............................................................................... 310 Climate Change and Resource Access ........................................................312 Modeling Territorial Behavior . ...................................................................313 Data Sensitivity to Territorial Behavior ...................................................... 314 Utility of a Comparative Approach ..............................................................315 iii Implications for Social Complexity Studies .................................................317 Relevance to Contemporary Issues ..............................................................317 References ................................................................................................................... 319 Appendix A Defensive characteristics of sites on the San Juan Islands, WA .............351 Appendix B Defensive characteristics of sites on San Nicolas Island, CA ................. 353 Appendix C Zirconium and strontium concentrations (ppm) for archaeological and geological samples of fine-grained volcanic rock from the San Juan Islands, WA...............................................................................................355 iv LIST OF FIGURES Figure Number Page 1.1 Map of the San Juan Islands ..................................................................................2 1.2 Map of San Nicolas Island .....................................................................................3 1.3 Dyson-Hudson and Smith (1978) economic defendability model ....................... 15 1.4 Smith?s (1988) risk reduction model ....................................................................22 1.5 Fitzhugh et al.?s (2011) social network model .....................................................23 2.1 Paleoclimate in the Late Holocene Gulf of Georgia .............................................37 2.2 Map of perennial freshwater streams in the San Juan Islands ..............................40 2.3 Settlement pattern map for the San Juan Islands at 4000-3500 cal BP ................43 2.4 Settlement pattern map for the San Juan Islands at 3500-3000 cal BP ................44 2.5 Settlement pattern map for the San Juan Islands at 3000-2500 cal BP ................44 2.6 Settlement pattern map for the San Juan Islands at 2500-2000 cal BP ................45 2.7 Settlement pattern map for the San Juan Islands at 2000-1500 cal BP ................45 2.8 Settlement pattern map for the San Juan Islands at 1500-1000 cal BP ................46 2.9 Settlement pattern map for the San Juan Islands at 1000-500 cal BP ..................46 2.10 Settlement pattern map for the San Juan Islands at 500 cal BP-Contact ..............47 2.11 Summed probability plot for San Juan Islands dates. ...........................................48 2.12 Comparison of Gulf of Georgia summed probability plots ..................................48 2.13 Paleoclimate in the Late Holocene on San Nicolas Island ...................................59 2.14 Site distribution on San Nicolas Island prior to 5000 cal BP ...............................65 2.15 Site distribution on San Nicolas Island at 5000-4500 cal BP ...............................65 2.16 Site distribution on San Nicolas Island at 4500-4000 cal BP ...............................66 2.17 Site distribution on San Nicolas Island at 4000-3500 cal BP ...............................66 2.18 Site distribution on San Nicolas Island at 3500-3000 cal BP ...............................67 2.19 Site distribution on San Nicolas Island at 3000-2500 cal BP ...............................67 2.20 Site distribution on San Nicolas Island at 2500-2000 cal BP ...............................68 2.21 Site distribution on San Nicolas Island at 2000-1500 cal BP ...............................68 2.22 Site distribution on San Nicolas Island at 1500-1000 cal BP ...............................69 2.23 Site distribution on San Nicolas Island at 1000-500 cal BP .................................69 2.24 Site distribution on San Nicolas Island at 500 cal BP-Contact .............................70 2.25 Summed probability plot for San Nicolas Islands sites ........................................70 3.1 Map of the San Juan Islands and the Watmough Bay site ....................................78 3.2 View from above of Watmough Bay ....................................................................78 3.3 Plan view map of Watmough Bay excavations ....................................................80 3.4 Stratigraphic profile at 0N18W, Watmough Bay .................................................82 3.5 Stratigraphic profile at 0N9W, Watmough Bay ...................................................82 3.6 Stratigraphic profile at EXU1, Watmough Bay ....................................................83 3.7 Hearth feature at EXU1, Watmough Bay .............................................................84 3.8 Watmough Bay dates ............................................................................................88 3.9 Dates and temporal analytic units at Watmough Bay ...........................................90 3.10 Spatial analytic units at Watmough Bay ...............................................................91 3.11 Map of San Nicolas Island with Tule Creek Village and SNI-106 .......................92 3.12 View to the east of Mound B excavation ..............................................................94 3.13 Plan view map of Mound B excavation ................................................................96 v 3.14 Mound B dates .....................................................................................................99 3.15 North wall profiles at Mound B .........................................................................100 3.16 Plan view and north wall of Unit 43, Mound B .................................................102 3.17 North walls of units 13 and 32, Mound B ..........................................................102 3.18 North and south wall profiles of units 32 and 13, Mound B .............................103 3.19 Percent fines in trench units at Mound B ...........................................................105 3.20 Percent organics and carbonates in trench units at Mound B ............................107 3.21 Temporal analytic units at Mound B .................................................................110 3.22 Stratigraphy and dating of an index unit at CA-SNI-106 ..................................111 4.1 Fine-grained volcanic (FGV) artifacts from San Juan Island ............................115 4.2 Map of the Watts Point volcanic center .............................................................117 4.3 Biplot showing geochemistry of Howe Sound and Mt. Baker FGV .................122 4.4 Biplot showing geochemistry of FGV artifacts .................................................123 4.5 Biplot showing Watts Point FGV artifacts ........................................................123 4.6 Biplot showing geochemistry of geological samples of FGV ...........................124 4.7 Biplot showing Watts Point FGV geological samples .......................................124 4.8 Map of cobble survey areas on San Juan Island ................................................126 4.9 Map of cobble survey areas on Lopez Island .....................................................127 4.10 Cobble survey area at Watmough Bay beach ....................................................128 4.11 Cobble survey area at Aleck Bay beach ............................................................129 4.12 Cobble survey area at Agate Beach ...................................................................130 4.13 Cobble survey area at American Camp .............................................................131 4.14 Cobble survey area at Eagle Cove .....................................................................131 4.15 Cobble survey area at False Bay ........................................................................132 4.16 Cobble survey Snug Harbor ...............................................................................133 4.17 San Nicolas Island cobbles eroding out of conglomerate ..................................139 4.18 Metavolcanic cobble ..........................................................................................141 4.19 Porphyrtic metavolcanic artifacts from Mound B .............................................142 4.20 Metavolcanic porphyry cobble from San Nicolas Island ...................................142 4.21 Quartzite artifacts from Mound B ......................................................................144 4.22 Quartz artifacts from Mound B ..........................................................................144 4.23 Cico chert biface from Mound B .......................................................................145 4.24 Monterey banded chert biface from Mound B ...................................................145 4.25 Map of extra-local toolstone sources for San Nicolas Island ............................146 4.26 Map of cobble survey areas on San Nicolas Island ...........................................147 4.27 Survey area 1 at Corral Harbor ..........................................................................148 4.28 Survey area 2 at Corral Harbor ..........................................................................149 4.29 Survey area near Tule Creek Village .................................................................150 4.30 Survey area 1 at Bollywood Beach ....................................................................151 4.31 Survey area 2 at Hollywood Beach ....................................................................151 4.32 Survey area at NAVFAC ...................................................................................152 4.33 Survey area at Thousand Springs .......................................................................153 4.34 Survey area 1 at Radar Row...............................................................................154 4.35 Survey area 2 at Radar Row...............................................................................154 4.36 Least cost paths between Tule Creek Village and cobble areas ........................168 4.37 Least cost paths between CA-SNI-106 and cobble areas ..................................169 vi 5.1 FGV flake from an angular cobble, Watmough Bay, obverse ............................182 5.2 FGV flake from an angular cobble, Watmough Bay, reverse.............................182 5.3 Manufacturing process for angular FGV cobbles ...............................................185 5.4 Multi-platform unpatterned FGV core from Watmough Bay, obverse ..............186 5.5 Multi-platform unpatterned FGV core from Watmough Bay, reverse ...............186 5.6 Plan-view schematic of multi-platform unpatterned FGV core ..........................186 5.7 Multi-platform 90 degree FGV core from Watmough Bay, obverse ..................187 5.8 Multi-platform 90 degree FGV core from Watmough Bay, reverse ..................187 5.9 Plan-view schematic of a multi-platform 90 degree FGV core ..........................187 5.10 Exhausted multi-platform FGV core from Watmough Bay, obverse .................188 5.11 Exhausted multi-platform FGV core from Watmough Bay, reverse ..................188 5.12 Plan-view schematic of exhausted multi-platform FGV core .............................188 5.13 FGV rotated core from Watmough Bay, oberse .................................................189 5.14 FGV rotated core from Watmough Bay, reverse ................................................189 5.15 Plan-view schematic of FGV rotated core from Watmough Bay .......................189 5.16 FGV rotated core from Watmough Bay, obverse ...............................................190 5.17 FGV rotated core from Watmough Bay, reverse ................................................190 5.18 Plan-view schematic of FGV rotated core from Watmough Bay .......................190 5.19 Unidirectional FGV core from Watmough Bay, obverse ...................................191 5.20 Unidirectional FGV core from Watmough Bay, reverse ....................................191 5.21 Opposed platform core on a round FGV cobble, obverse ..................................192 5.22 Opposed platform core on a round FGV cobble, reverse ...................................192 5.23 Chopper core on a round FGV cobble, obverse ..................................................192 5.24 Chopper core on a round FGV cobble, reverse ...................................................192 5.25 FGV scaled piece from Watmough Bay, obverse ...............................................195 5.26 FGV scaled piece from Watmough Bay, reverse ................................................195 5.27 FGV scaled piece from Watmough Bay, obverse ...............................................195 5.28 FGV scaled piece from Watmough Bay, reverse ................................................195 5.29 FGV scraper from Watmough Bay, obverse .......................................................196 5.30 FGV scraper from Watmough Bay, reverse .......................................................196 5.31 FGV flake with edge damage from Watmough Bay, obverse ............................196 5.32 FGV flake with edge damage from Watmough Bay, reverse .............................196 5.33 FGV bifaces from Watmough Bay, obverse .......................................................199 5.34 FGV bifaces from Watmough Bay, reverse ........................................................199 5.35 FGV willow leaf-shaped point and stemmed point, obverse ..............................200 5.36 FGV willow leaf-shaped point and stemmed point, reverse ...............................200 5.37 FGV triangular points from Watmough Bay, obverse ........................................201 5.38 FGV triangular points from Watmough Bay, reverse .........................................201 5.39 Slate knife from Watmough Bay, obverse ..........................................................203 5.40 Slate knife from Watmough Bay, reverse ...........................................................203 5.41 Split cobble reduction sequence on San Nicolas Island......................................209 5.42 Schematic of split cobble core from Mound B ...................................................210 5.43 Split cobble cores from Mound B, obverse ........................................................211 5.44 Split cobble cores from Mound B, reverse .........................................................212 5.45 Diagonal core reduction sequence ......................................................................214 5.46 Diagonal cobble core from Mound B, obverse ...................................................215 vii 5.47 Diagonal cobble core from Mound B, reverse ....................................................215 5.48 Plan-view schematic of diagonal cobble core from Mound B, obverse .............215 5.49 Decapitate reduction technique ...........................................................................216 5.50 Decapitated core from Mound B, obverse ..........................................................216 5.51 Decapitated core from Mound B, reverse ...........................................................216 5.52 Unpatterned multidirectional cores from Mound B, obverse .............................217 5.53 Unpatterned multidirectional cores from Mound B, reverse ..............................217 5.54 Plan-view schematic of an unpatterned multidirectional core from Mound B ...217 5.55 Flake from Mound B with damaged margin, obverse ........................................218 5.56 Flake from Mound B with damaged margin, reverse .........................................218 5.57 Drills from Mound B, obverse ............................................................................219 5.58 Drills from Mound B, reverse .............................................................................219 5.59 Retouched drill from Mound B, obverse ............................................................220 5.60 Retouched drill from Mound B, reverse .............................................................220 5.61 Scrapers from Mound B, obverse .......................................................................220 5.62 Scrapers from Mound B, reverse ........................................................................220 5.63 Bifaces from Mound B, obverse .........................................................................221 5.64 Bifaces from Mound B, reverse ..........................................................................221 5.65 Manufacture of a sandstone saw .........................................................................222 6.1 Central place foraging model applied to cobble reduction .................................228 6.2 Proposed processing strategies for sites with high and low boundary defense ..229 6.3 Predictions for lithic procurement at Watmough Bay ........................................235 6.4 Proportions of FGV and slate by weight, Stein/Phillips excavation ...................240 6.5 Proportions of FGV and slate by weight for the Munsell excavation.................240 6.6 Distribution of extra-local materials at Watmough Bay .....................................255 6.7 Procurement predictions for Mound B ...............................................................260 6.8 Proportions of material types by weight at Mound B .........................................264 6.9 Proportions of material type proportions by weights at CA-SNI-106 ................264 6.10 Flake with a round platform, obverse .................................................................273 6.11 Flake with a round platform, reverse ..................................................................273 6.12 Decapitation flake, obverse .................................................................................273 6.13 Decapitation flake, reverse..................................................................................273 viii LIST OF TABLES Table Number Page 1.1 Predictions for boundary defense and permeability for the San Juan Islands ......25 1.2 Boundary defense and permeability in a coastal environment .............................26 2.1 Primary freshwater sources in the San Juan Islands .............................................41 2.2 Boundary defense and permeability in the San Juan Islands ................................49 2.3 Results of ANOVA comparing mean defensive measures for San Juan Islands sites, sites assigned to the first period to which they date ....................................52 2.4 Results of ANOVA comparing mean defensive measures for San Juans Islands sites, sites assigned to every period to which they date ........................................52 2.5 Results of t-tests comparing means for defensive measures for sites inhabited before and after 600 cal BP. ..................................................................................53 2.6 Results of t-tests comparing mean defensive measures for big sites and small sites for all time periods and for sites inhabited after 600 cal BP.........................53 2.7 Predictions for boundary defense and permeability for San Nicolas Island .........71 2.8 Results of an ANOVA comparing mean defensive measures for San Nicolas Island sites, sites assigned to the first period to which they date ..........................73 2.9 Results of an ANOVA comparing mean defensive measures for San Nicolas Island sites, sites assigned to every time periods to which they date ...................74 2.10 Site location counts for each time period, sites assigned to the first period to which they date .....................................................................................................75 2.11 Site location counts for each time period, sites assigned to each period to which they date ................................................................................................................75 3.1 Basic field descriptions of stratigraphy at Watmough Bay ..................................82 3.2 Radiocarbon dates for the Watmough Bay site, 1968 excavation ........................85 3.3 Radiocarbon dates for the Watmough Bay site, 2004 excavation ........................87 3.4 Basic field descriptions of strata at Mound B .......................................................95 3.5 Radiocarbon dates for Tule Creek Village, Mound B ..........................................97 3.6 Results of a grain size analysis at Mound B .......................................................106 3.7 Results of a loss-on-ignition analysis for bulk samples from Mound B .............106 3.8 Radiocarbon dates for CA-SNI-106 ....................................................................111 4.1 Descriptive statistics for six FGV cobble areas in the San Juan Islands ............136 4.2 Results of an ANOVA for cobble size/shape data for the San Juan Islands.......136 4.3 Results of a Bonferonni post-hoc analysis for cobble length for cobble areas on the San Juan Islands .......................................................................................137 4.4 Results of ?2 tests for ordinal values for cobble areas in the San Juan Islands ...138 4.5 Proportions of material types at each cobble area on San Nicolas Island ..........155 4.6 Descriptive statistics for metavolcanic cobbles at six cobble areas on San Nicolas Island......................................................................................................157 4.7 Results of an ANOVA comparing mean size and shape measurements for metavolcanic cobbles on San Nicolas Island ......................................................158 4.8 Results of a Bonferonni post-hoc analysis for cobble length for cobble areas on San Nicolas Island...............................................................................................159 4.9 Results of a Bonferonni post-hoc analysis for circumference ratio for cobble areas on San Nicolas Island ................................................................................160 ix 4.10 Results of a Bonferonni post-hoc analysis for flatness for cobble areas on San Nicolas Island...............................................................................................161 4.11 Results of a Bonferonni post-hoc analysis for elongation for cobble areas on San Nicolas Island...............................................................................................162 4.12 A ?2 test comparing proportions of flat and round metavolcanic cobbles ..........162 4.13 Wilson?s Attractive Index values for San Juan Islands cobble areas..................171 4.14 Wilson?s Attractive Index values for San Nicolas Island quarry ares for Tule Creek Village ......................................................................................................172 5.1 Measurement descriptions for lithic analysis of flaked stone tools ....................175 5.2 Frequences of flakes and cores from the Watmough Bay site ............................177 5.3 Basic statistics on tools and groundstone from the Watmough Bay site ............178 5.4 Results of an ANOVA test comparing dimensions of flake types at Watmough Bay, FGV only ....................................................................................................182 5.5 Results of an ANOVA and post-hoc analysis comparing intact cobble Dimensions of Watmough Bay FGV cores and beach cobbles ..........................193 5.6 Location of cortex and use-wear or retouch by flake-tool type .........................197 5.7 Results of a ?2 test comparing proportions of a round and flat platforms for FGV flakes and scaled pieces at Watmough Bay ...............................................198 5.8 Results of t-tests comparing mean dimensions for unmodified FGV flakes, scrapers, and scaled pieces at Watmough Bay ....................................................202 5.9 Descriptive data for the Mound B and SNI-106 flakes and cores ......................205 5.10 Descriptive data for the Mound B and SNI-106 formal tools and miscellaneous stone artifacts ......................................................................................................206 6.1 Predictions for the Watmough Bay lithic assemblage ........................................235 6.2 Material type counts and weights for the Watmough Bay lithic assemblage .....239 6.3 Results of ?2 tests comparing proportions of FGV and slate for the Stein/ Phillips excavation, Watmough Bay ...................................................................241 6.4 Results of ?2 tests comparing proportions of FGV and slate for the Munsell excavation, Watmough Bay ................................................................................242 6.5 Basic statistics for FGV cores from Watmough Bay that have original cobble dimensions intact and cobbles from beaches on the San Juan Islands ...............244 6.6 Results of ?2 tests comparing proportions of cores (1600-1000 cal BP) made on angular and round cobbles from Watmough Bay with cobble areas on the San Juan Islands ..................................................................................................245 6.7 Results of ?2 tests comparing proportions of smooth and rough cortex FGV flakes from Watmough Bay (1600-1000 cal BP) with cobble areas on the San Juan Islands ..................................................................................................247 6.8 Results of ?2 tests comparing proportions of smooth and rough cortex FGV cores from Watmough Bay (1600-1000 cal BP) with smooth and rough cortex cobbles from each cobble area ............................................................................248 6.9 Summary of results for shape, size, and appearance for artifacts from Watmough Bay ...................................................................................................249 6.10 FGV flake types at Watmough Bay and ?2 results comparing first flakes and later flakes between 1600-1000 cal BP and 600 cal BP-Contact time periods ...251 6.11 Results of ?2 tests comparing number of dorsal flake scars between time periods at Watmough Bay ................................................................................................251 x 6.12 Results of a ?2 test comparing proportions of slate and FGV between separate spatial areas at Watmough Bay ...........................................................................257 6.13 Summary of predictions for lithic analysis at Mound B and SNI-106 ...............261 6.14 Material weights for Mound B and SNI-106 based on debitage ........................263 6.15 Results of ?2 tests comparing proportions of metavolcanic and quartzite flakes between time periods at CA-SNI-106. ................................................................266 6.16 Data on proportions of metavolcanic, porphyrtic metavolcanic, and metavolcanic porphyrtic toolstone at Mound B and SNI-106 ............................268 6.17 Results of a ?2 test comparing proportions of metavolcanic, porphyrtic metavolcanic, and metavolcanic porphyry flakes at Mound B ...........................268 6.18 Results of ?2 tests comparing proportions of metavolcanic, porphyrtic metavolcanic, and metavolcanic porphyry flakes at SNI-106 ............................269 6.19 Early and late stage flakes associated with different San Nicolas Island core reduction strategies .............................................................................................273 6.20 Flake type counts at Mound B and SNI-106 .......................................................275 6.21 Results of a ?2 test comparing proportions of round and flake reduction Sequence debitage at Mound B and SNI-106 at 1500-500 cal BP .....................276 6.22 Results of ANOVA tests comparing mean weight and size for primary, secondary, and tertiary flakes .............................................................................278 6.23 Attributes related to reduction for Mound B and SNI-106 .................................278 6.24 Results of ?2 tests comparing proportions of flakes and flake attributes associated with early and later stage reduction at Mound B and SNI-106 .........279 6.25 Results of a ?2 test comparing proportions of flake types at 1500-500 cal BP and 500 cal BP-Contact at Mound B, all material types included ......................280 6.26 Results of a ?2 test comparing proportions of flake types at 1500-500 cal BP and 500 cal BP-Contact at Mound B, only metavolcanic rock included ............280 6.27 Results of ?2 tests comparing proportions of exhausted core flakes between time periods at SNI-106 ......................................................................................282 6.28 Results of of ?2 tests comparing proportions of non-cortical shatter and other flakes at Mound B ...............................................................................................283 6.29 Results of ?2 tests comparing proportions of non-cortical shatter and other flakes at SNI-106 ................................................................................................284 6.30 Results of ?2 tests comparing proportions of chert and quartz to metavolcanic flakes at Mound B ...............................................................................................287 6.31 Results of a ?2 test comparing proportions of quartz to metavolcanic flakes at SNI-106 ...............................................................................................................288 6.32 Results of ?2 tests comparing proportions of material types for two spatial areas at Mound B ................................................................................................291 7.1 Predictions and results for change over time in defensive sites in the San Juan Islands .................................................................................................................297 7.2 Predictions and results for change over time in lithic procurement at Watmough Bay ...................................................................................................300 7.3 Predictions and results for change over time in conservation and processing at Watmough Bay ...............................................................................................301 7.4 Predictions and results for change over time in boundary permeability at Watmough Bay ...................................................................................................302 xi 7.5 Predictions and results for defensive characteristics of sites on San Nicolas Island ...................................................................................................................304 7.6 Results of lithic analyses for Mound B ...............................................................306 7.7 Results of lithic analyses for SNI-106 ................................................................307 7.8 Results of boundary permeability analyses at Mound B and SNI-106 ...............309 xii ACKNOWLEDGEMENTS This research has benefited greatly from the guidance of my supervisory committee. Ren? Vellanoweth shared his passion for the archaeology of the southern Channel Islands and generously allowed me to analyze the Mound B lithic assemblage and participate in fieldwork on San Nicolas Island. Angela Close?s fascinating perspective on the flake tool technology of the San Juan Islands inspired my investigation of toolstone procurement and manufacture at Watmough Bay. Ben Fitzhugh challenged and greatly clarified my thinking on the theoretical framework of this project. Olivier Bachmann served as a supportive Graduate School Representative. My advisor Julie Stein provided invaluable enthusiasm, support, and wisdom over the years regarding both research and teaching. She introduced me to archaeology in the San Juan Islands and improved my work through her innate understanding of human and natural landscapes. I am indebted to the organization, efficiency, and encouragement of Graduate Program Advisor Catherine Zeigler and Fiscal Specialist John Cady. I would also like to express my appreciation for excellent mentorship and a solid background in archaeological method and theory from my undergraduate archaeology professors at Hamilton College, Tom Jones, Charlotte Beck, and Mike Cannon. For access to the Watmough Bay assemblage and for a research fellowship that supported my lithic analysis and sourcing work, I thank the archaeology department at the Burke Museum: Steve Denton, Peter Lape, Kelly Meyers, Megon Noble, and Laura Phillips. My research was also funded by a National Science Foundation Dissertation Improvement Grant (BSC1043916). Thanks also to Rich Bailey, District Archaeologist for the Spokane District Bureau of Land Management, for access to the Watmough collections. I thank the many members of the San Juan Islands community who participated in settlement pattern research by welcoming us onto xiii their properties, providing information about the archaeological record, and for the lemonade on hot days and coffee on cold days. I thank Lola Deane, the Mendez family, San Juan County Parks, Friday Harbor Labs, and Snug Harbor Resort for access to their beaches and fine-grained volcanic rocks for my beach cobble study. Craig Skinner at Northwest Research Obsidian Studies Laboratory conducted geochemical analyses for this study and I thank him for his efficiency and input. I am also grateful for the insights of the cultural resources department at the Samish Indian Nation, the Swinomish Indian Tribal Community, and the Lummi Nation, especially Lena Tso at the Lummi Tribal Historic Preservation Office whose concerns and ideas helped structure the field methods and goals of the San Juan Islands Archaeology Project. Several graduate students at the University of Washington provided field assistance in the San Juan Islands. I thank Erik Gjesfjeld, Emily Peterson, Colby Phillips, Alecia Spooner, and Catherine West for their hard work and friendship. Stephanie Jolivette was a wonderful partner in sourcing adventures, fieldwork, research, and public outreach. For the San Nicolas Island field research, I thank Steve Schwartz and Lisa Thomas-Barnett of the Naval Air Station Facility, Point Mugu for facilitating my research on the island. I also thank the California State University, Los Angeles archaeology graduate and undergraduate students for their assistance with cobble surveys and for sharing their knowledge and excitement about San Nicolas Island archaeology. Amira Ainis, Richard Guttenberg, Bill Kendig, and Johanna Marty provided vital research collaboration, hospitality, and friendship. Without the intellectual support and camaraderie of my friends at the University of Washington Department of Anthropology, I would still be stuck in the basement of Denny Hall and I never would have had so much fun along the way. It was a privilege to work with such a brilliant, dynamic, and warm group of people. Special gratitude to cohort mates Jack Johnson, xiv Becky Kessler, Megan Luce, Emily Peterson, and Haiying Zhang, and to Shelby Anderson, Jacob Fisher, Adam Freeburg, Christina Giovas, Bob Kopperl, Lisbeth Louderback, Molly Odell, Lauren Rhodes, and Natasha Slobodina for advice, input, encouragement, laughter, and time well-spent at Finn MacCool?s. I thank my Archy 205, Archy 299, and Coastal Archaeology students for constantly renewing my enthusiasm for archaeology. I also thank my many ?civilian? friends outside the department for their encouragement and fun distractions from my work, especially Gabi Rizzuto for her love of science. Thanks to Ira Glass, KEXP, and Ballard Mandarin for getting me through long evenings of analysis and writing and The Dray for nights off. Finally, my parents, sisters, in-laws, and out-laws not only put up with me during this long process, but had my back the whole way through. A special thanks to my grandmother, Ruth Nathanson, for inspiring me to go to great lengths to pursue intellectual curiosity and a fulfilling career. I cannot possibly adequately thank my husband Matt Saunders for keeping our lives on track as I finished my degree, for providing a logical and insightful sounding board for thoughts and ideas, and for his unconditional love and support. 1 Chapter 1: Introduction When Europeans arrived on the Pacific Coast of North America in the 1500s, they found large villages connected through exchange, intermarriage, kinships, friendships, and rivalries. My dissertation research investigates the development of territoriality in the context of complex and interconnected coastal communities and Late Holocene climate change. I focus in particular on peoples of the San Juan Islands, Washington and San Nicolas Island in the southern Channel Islands, California (Figure 1.1, 1.2). The San Juan Islands and southern Channel Islands study areas are well-suited to an investigation of the emergence of territoriality because territorial behaviors are recorded in the ethnographic literature, paleoenvironmental reconstructions are detailed, and settlement patterns are thoroughly documented. A comparison of the Northwest and California coasts facilitates archaeological investigations of the unique behaviors of coastal peoples (e.g., Ames 2005b; Coupland 2004; Erlandson 2001; Fitzhugh and Kennett 2010; Lightfoot 1993) and more rigorously tests the predictions of territoriality models. Territorial behavior is a critical dimension of the emergence of sociopolitical complexity among complex hunter-gatherers, defined here as horizontal differentiation into groups such as households or domestic units, and/or vertical division into ranks associated with different levels of prestige or access to resources (Fitzhugh 2003). This research does not focus on evaluating social complexity models but provides new insights that will be useful in addressing how and why social and political systems in both study areas changed over time. To test hypotheses and predictions regarding the emergence of territoriality in the San Juan Islands and southern Channel Islands, I investigate the relationships between boundary defense, boundary permeability, and resource procurement. I examine settlement patterns and temporal and spatial patterns in the acquisition and conservation of local and extra-local toolstone throughout the Late Holocene. This research project required excavation, settlement 2 pattern analysis, toolstone surveys, and lithic analysis. Lithic assemblages analyzed include the Watmough Bay site (45-SJ-280), Lopez Island, San Juan Islands, Washington and Tule Creek Village (CA-SNI-25) Mound B and CA-SNI-106, San Nicolas Island, southern Channel Islands, California. Where possible I have attempted to parallel the analyses for each study area; however, different components of the research took different trajectories and required more or less investigation and explanation. Figure 1.1. Map of the San Juan Islands and Gulf of Georgia. 3 Figure 1.2. Map of San Nicolas Island and southern California. Territoriality in Coastal Settings Territory size, shape, degree of boundary defense, and degree of boundary permeability are all elements of human territoriality, a complex set of behaviors that restrict access and resource use within a defined spatial area (Cashdan 1983; Dyson-Hudson and Smith 1978). I define a territory as an area that a group defends and identifies as a community-owned space. Land ownership systems range from exclusive control of land and resources to common pool systems where land and resources are accessible to community members based on rules and traditions (e.g., Ostrom 1990) or expectation of future reciprocation (e.g., Eerkens 1999:298). 4 Like other aspects of human adaptation, people change their territorial strategies in response to increases and decreases in resources, population density, and other social and environmental variables. The boundaries that demarcate territories can be built out of wood, earth, or stone; they can also be natural physical boundaries or social boundaries that leave no physical trace (Tringham 1972). Territoriality is a topic of particular interest on parts of the Pacific Coast where resource density and economic modes of production supported high population aggregations and sedentism. Yesner (1980:739) notes that maritime foraging of densely packed and patchy resources often works best from a single location. Use of boats facilitates resource gathering from geographically distant resource patches, thus villages are often located at a point on the landscape where people can easily access a few different resources. These settlements tend to be near resources that are easy to collect and process such as shellfish, seabird colonies, marine mammal rookeries, and fishing grounds. People use storage to maintain resource surpluses at the habitation site (Renouf 1984:21). In such contexts, staying in one spot allows communities to maintain exclusive access to nearby abundant and predictable resources, and conversely, sedentism is impossible if resources are insufficient to support the group. Rosenberg (1998) proposes that sedentism is desirable in contexts of high population pressure because it allows more exclusive access to the most productive parts of a territory. In ?cheating at musical chairs?, some groups of people stop moving on their seasonal round and instead establish a more permanent settlement in the most resource-productive area they can find. For both case studies, the communities that I investigate may not have been fully sedentary; ethnographic examples indicate that people occupied different parts of the landscape during different seasons. Several researchers describe this type of semi-sedentism on the 5 continuum of between sedentism and mobility (Kelly 1991, 1992; Rocek and Bar-Yosef 1998). Here, I use Binford?s (1980:13) definition of semi-sedentary groups as those who ?shift from one to another fixed settlement at different seasons or who occupy more or less permanently a single settlement from which a substantial proportion of the population departs seasonally to occupy shifting camps.? Since the 1970s, several ethnographic studies have focused on complex semi-sedentary hunter-gatherer-fishers in marine environments (e.g., Acheson 1988, 2003; Acheson and Gardner 2004; Ackerman and Ackerman 1973; Angelbeck 2009; Aswani 2002; Barber 2003; Begossi 1995; Durrenberger and P?lsson 1987; McCay 1978; Richardson 1982; Schaepe 2006, 2009). These studies suggest modifications to traditional territoriality models for complex hunter- gatherer fishers. For example, habitation sites may be at a distance from the most productive resources on the landscape, yet people still establish exclusive access to these areas through social of physical boundary defense. Establishing and maintaining boundaries on the open water involves different costs and benefits than maintaining boundaries on land. As well, with high population density, rapid boat transport, and complex social structures, communities are highly inter-connected within and across territorial boundaries. Ethnographic analysis of territoriality and community in both study areas provides further context for this research. Ethnographic Research on Territoriality in the San Juan Islands The ethnographic record of the central Northwest Coast provides a rich resource on the ways that people established, maintained, and crossed group boundaries. I focus specifically on Coast Salish communities, defined by Suttles (1960b[1987:29]) as the groups that lived in the area around the Georgia Strait, Strait of Juan de Fuca, Puget Sound, and on the open coast 6 between the Olympic Peninsula and Willipa Bay. The native peoples of the San Juan Islands are identified by ethnographers as Central Coast Salish and Straits Salish speakers (Carlson 2001:23; Suttles 1990). The land tenure system of the Coast Salish in the San Juan Islands is characterized by communities that establish and maintain ownership to territories on land and sea based on descent from an ancestor who fell from the sky in the distant past in the time of the Transformers (Suttles 1987:9, 1960a; Thom 2005:292). Families trace their line and rights to property, knowledge, and resources to the First Ancestors (Barnett 1955:141, 291; Thom 2005:84-85). Evidence of boundary defense among the Coast Salish comes from a variety of sources. European colonizers painted images and composed letters and ship logs that attested to violent interactions between tribes and Spanish and English explorers (Angelbeck 2009:1, 69-129), and conflict between groups. Ethnographers wrote of skirmishes, land seizing, and slave-raiding between the Coast Salish and southern Kwakwaka?wakw communities in the 1700s that ended with a marriage alliance in the 1800s (Thom 2005:361). Weapons included slate or metal knives, stone, wood, or bone clubs, bow and arrows, spears, slings, and buckskin armor (Angelbeck (2009:102-108). Angelbeck notes that warfare and construction of defensive sites among the Coast Salish increased after European colonization began, perhaps due to the effects of social disruptions caused by a smallpox epidemic, uneven distribution of firearms, and conflicts between Coast Salish groups, traders, and colonists. The St?:l? of the lower Fraser Valley and other groups have place names and stories associated with conflict and defense (Schaepe 2006, 2009). Defensive sites recorded by ethnographers and archaeologists also attest to boundary defense. The Coast Salish built palisaded forts on bluff-tops, trenches, embankments, rock-wall fortifications, refuges in defensive areas, lookouts in high areas, and skulls on posts and other 7 symbolic displays (Angelbeck 2009:108; Schaepe 2006, 2009). Many rock wall fortification systems are still visible today. Some built and some natural, these structures were designed by shape and location to operate over a large area within a system of social cooperation of kin groups under a central political leadership. The walls protected people and communicated power (Carlson 2010; Schaepe 2006). In general, river travel and resource access was more restricted than ocean travel and marine resource access (Carlson 2010:53). Based on ethnographic accounts, different territorial strategies operate simultaneously at different scales of community for the Coast Salish. Suttles 1964[1987:209] discusses six community tiers: (1) families that live together in a section of a plank house and participate in a common domestic economy; (2) several related families in a plank house who co-host ceremonies; (3) groups of houses on a beach that share a common group name and ceremonial rights; (4) several groups of houses on a long stretch of shoreline that share a common language, subsistence methods, and ritual; (5) kin groups that hold rights to resource patches; and (6) all of the people who participate in subsistence activities and ceremonies that may or may not parallel residential units (see also Thom 2005:280; Robinson 1963:27). Both residence and descent groups were dynamic social structures that varied through time and space (Thom 2005:285). There have been particularly pronounced changes since the beginning of the 20 th century when smaller groups were pushed into larger villages. Groups split, reunited, and relocated based on resource availability. They maintained salmon camps in the spring, traveled to high-elevation berry patches in the summer, held potlatch ceremonies and hunting trips in the fall, and fished for sturgeon and held large gatherings in the winter (Carlson 2001:62; Elmendorf 1971:357; Suttles 1960a[1987:23). 8 The concept of community most commonly discussed by ethnographers in narratives of territoriality is the village, which consists of multiple groups of houses on a beach who united for group defense (Angelbeck 2009:226; Suttles 1951:277). Outsiders were actively excluded unless they had permission to trespass (Richardson 1982; Suttles 1960a[1987:16]). Angelbeck (2009:257) describes a decentralized cooperative strategy with different ?nodes? of defense that communicated using scouts, lookouts, messengers, and signal stations. Among some groups such as the St?:l?, multiple villages might band together for defense against an outside threat by coordinating defensive structures and aggressive strategies (Schaepe 2006, 2009). Boundary defense also took place at the level of the kinship group but centered on productive resources rather than village territories. During the postcontact period, productive resource areas were controlled by the more powerful extended families (Carlson 2010; Richardson 1982; Robinson 1963; Suttles 1960a,c[1987]); Thom 2005:283). Standing in a family group demonstrated through public gifting during feasts and potlatches secured access to family resources (Thom 2005). A host family or village invited members of other communities to acquire food, blankets, shell ornaments, baskets, clothing, stone tools, canoes, and slaves (Suttles 1960a[1987:19-21]). Suttles (1951:68-69; 1960a[1987:20]) records family ownership of camas, fern, wapato, and shellfish collecting areas as well as reef net locations. Stern (1934:47) records clam gardens created and maintained by the Lummi in the 1920s on descent group-owned beaches. The same was true for waterfowl, deer, mountain goat, and seal hunting areas, and sturgeon traps (Barnett 1955:251; Robinson 1963). Salmon weirs and traps were sometimes controlled by a village rather than a family (Thom 2005:320). Resources that were extremely abundant, such as salal berries and shellfish, were rarely restricted (Richardson 1982; Carlson 2010:47). Knowledge about exploiting resources was passed down from generation to generation 9 (Suttles 1987:8). Trespass by non-kin could result in violent confrontation or an obligation to provide reparations (Thom 2005:377). Village boundaries were sometimes aggressively defended, but they were also permeable to certain inter-group interactions between kin or families allied through marriage or friendship (Elmendorf 1971:358; Suttles 1963). Kinship was bilateral with lineages reckoned through male lines. Residence was usually patrilocal (Suttles 1960a[1987:9]). Kin and marriage alliances enabled resource sharing, which balanced out temporal and spatial fluctuations in subsistence resources (Carlson 2010; Miller 1989; Suttles 1960a,b[1987]). Elmendorf (1971:359) notes frequent ?reciprocal visiting? and exchange of food resources between kin in different Coast Salish communities. Carlson (2001:27; 2010:49) records visits from kin from the Gulf Islands and Vancouver Island to the Fraser Valley to harvest resources. High status families retained intensive knowledge of their geneology and low status families or ?worthless people? were those who had forgotten who they were and where they came from. If an inter-village relationship was established through marriage, the two families could exchange goods and property as long as the marriage lasted. This involved a series of exchanges that in some cases became competitive. If one family had a surplus of herring they could bring it to parents-in-law, receive a blanket in exchange, and later use that blanket to thank the parents- in-law for gifts of camus bulbs and dried sturgeon (Suttles 1960a[1987:17]). Traditions of intermarriage between certain communities led to alliances and reciprocal access between descent or residence groups (Onat 1984; Thom 2005:298). 10 Ethnographic Research on Territoriality in the Southern Channel Islands The people who lived on the southern Channel Islands during the post-Contact period are known as the Island Gabrielino or the Tongva. They are Uto-Aztecan speakers whose traditional tribal territory includes Santa Catalina Island, San Clemente Island, San Nicolas Island, and Santa Barbara Island and the central coast and interior south of the Chumash territory to Newport Bay and east from the coast to the San Gabriel and Santa Ana Mountain (McCawley 1996, 2002). Little is known specifically of the Nicole?o because they were removed to missions on the mainland by 1835 prior to any visits by ethnographers. One woman was left behind on San Nicolas Island and died soon after she was brought to the mainland in 1853 (Meighan 1954). Texts from early Spanish explorers beginning the 1500s suggest that the Tongva approached the newcomers with friendly behavior and eager participation in exchange of clothing, beads, and other items (Bean and Smith 1978:547; McCawley 1996:4-9, 25,90-91). However, when attempts were made to colonize Tongva territory on the mainland in the late 1700s, they were met with violent and organized resistance under the principal chief of Porciuncula (Castillo 1999:48). Similar to the Coast Salish, the Tongva had complex communities. For mainland groups, the settlement round was characterized by dispersed family units at certain times during the year and larger settlements at other times depending on resource distribution. Primary settlements controlled political and spiritual life, secondary settlements served as trade loci, and small special purpose sites were used for hunting, fishing, shellfish harvesting, and processing of various materials (McCawly 1996:27-29). Villages were organized into one or more lineage groups (several related families). Each village was headed by a tomyaar, an inherited position of political, spiritual, and economic power usually served by men. In some cases, tomyaars had 11 authority over several communities. Lineage groups included related families descended from an ancestor through the paternal line who shared access to hunting areas, acorn and seed gathering locales, and shellfish beds. Families contributed a share of their food to a reserve managed by a tomyaar and hoarding had consequences (McCawley 1996:111). Each lineage group had an elite class that held ceremonial rights and political sway over the chief, a middle class of craftspeople and skilled laborers, poorer people, and slaves (McCawley 1996:104-105; 2002). The ethnographic record indicates violent conflict between Tongva groups due to trespassing, particularly between the coast and inland; however, food and other resources could be shared across group boundaries during formal gatherings. Lineages were grouped into wildcat and coyote moieties that owned specific ritual knowledge and objects. During inter-group gatherings, families representing both moieties were needed to complete ceremonies through ritual and paraphernalia. Groups in different village would take turns hosting ceremonies (Harrington 1942; McCawley 1996:89). Goods exchanged at these meetings included plant foods, seal, sea otter skins, red ochre, shell beads, plant foods, steatite, soapstone bowls and pipes, and other manufactured items and raw materials (McCawly 1996:112-113). Resource access beyond the village territory could also be obtained through inter- community marriage alliances. Tomyaars created intra and inter-community alliances through marriages, sometimes even outside of the language boundaries of the Tongva territory with the Cahuilla, Chumash, Serrano, and Luise?o (Bean and Smith 1978:547). The elite class tended to favor marriage partners from other communities for political and economic gain (McCawley 1996:104). If the leadership declared war, allies through marriage ties from other communities might be given gifts in exchange for assistance (McCawly 1996:106-107). 12 Boundary permeability was also negotiated through professional guilds that cross-cut lineage groups such as shaman guilds, plank canoe guilds and artisan guilds (McCawley 1996:10). One way that the Tongva restricted movement across the region was through access to the te?aat, a plank canoe. Only those individuals who were part of a guild had access to knowledge about how to build and use watercraft. Gamble (2008) notes that among the neighboring Chumash, canoes were only owned by families that had inherited wealth and power (Arnold 1992, 1993, 1995a; Gamble 2008). Theoretical Perspectives on Territoriality In my research, I endeavor to synthesize ethnographic data, ideas about territoriality from anthropological theory, and data from both study areas to understand how and why shifts in territorial strategies occurred in both the Gulf of Georgia and the southern Channel Islands. The ethnographic record provides an important place to start in understanding how territorial behavior?both boundary defense (Dillian 2003) and boundary permeability?manifested itself among complex hunter-gatherer-fishers of the Pacific Coast. Along with long-term continuities between the ethnographic record and the deeper past, there are also significant discontinuities (Moss 2011:24). The ethnographic record thus provides inspiration for developing hypotheses that can be integrated with anthropological theories and tested using the archaeological record. Below I review theoretical perspectives on the relationships between human territoriality and the natural and social environment. I outline my approach to investigating this behavior among coastal hunter-gatherer fishers by drawing on models from human behavioral ecology, archaeological studies, and ethnographic research. Finally, I create a set of basic predictions to 13 evaluate my approach that focus on lithic procurement strategies on the Pacific Coast of North America. A variety of theoretical perspectives are successfully used to explore human territoriality and conflict. Anthropologists apply concepts from boundary theory (Kooyman 2006), Marxist- based modes of power, Bourdieu?s (1977, 1990) practice theory, anarchy theory (Angelbeck 2009), materialist theories (McCauley 1990; Haas 1990), theories about innate human capacity for aggression (Maschner and Reedy-Maschner 1998), and evolutionary or human behavioral ecology (e.g., Krebs and Davies 1997; Smith and Winterhalder 1992; Winterhalder and Smith 2000). To develop hypotheses on territoriality, I draw mainly on human behavioral ecology, an approach that is grounded on the concept that humans evolved through natural selection to be flexible in our ability to adapt to the environment. We are capable of adjusting our extended phenotype to achieve an optimal adaptation to survive and reproduce, thus our fitness-enhancing behavior is predictable (Boone and Smith 1998: 144; Stephens and Krebs 1986). The goal of human behavioral ecology is to explore ?the differential persistence of variability in behavior over time? (Kelly 2000:64). Several human behavioral ecologists investigate the relationships between humans and their environment through the lens of human territoriality (Axelrod 1997; Boone 1983, 1992; Cashdan 1983, 1990, 1992; Durham 1976; Dyson-Hudson and Smith 1978; Smith 1981, 1983; Smith and Winterhalder 1992). The simple, abstract, and deductive models provided by human behavior ecology are useful tools for disentangling relationships between environment and behavior (Broughton and O?Connell 1999; Kennett 2005:12; Smith 1988:225). Expectations derived from human behavioral ecology models can be articulated with one another and used to generate testable predictions for each study area (Fitzhugh 2003). 14 The goals of this research are to better understand the emergence and nature of territoriality in the San Juan Islands and southern Channel Islands and to use those case studies to investigate the processes of emergence of territoriality in complex hunter-gatherer societies. The use of two study areas helps to more rigorously test the same hypotheses in more than one environmental and social context. Determining how the same hypotheses play out in more than one study areas provides more information about the strengths and weaknesses of those hypotheses, but I do not intend to use the study areas to prove the validity of abstract human behavioral ecology models. Rather, the models that I draw on are conceptual tools for developing hypotheses and articulating assumptions in a formal way. Human Behavioral Ecology Studies of Boundary Defense Several recent archaeology studies use concepts from human behavioral ecology to explore territorial behavior in coastal and island settings. Many of these studies draw on an economic defendability model originated by Brown (1964) that assumes that organisms defend boundaries if the benefits of controlling access to resources outweigh the costs of actively defending boundary. Dyson-Hudson and Smith (1978:26) build on this concept and present a model that predicts different territorial strategies given different degrees of resource density and predictability (Figure 1.3). If resource supply is far greater than demand, costs of defending territorial boundaries?energy spent on boundary maintenance and defense, risk of injury, confinement to a smaller area for resource access?outweigh the benefits of exclusive access to land and resources. The benefits are low because everyone on the landscape has more than enough food and water. If resources are scarce and unpredictable, territorial behavior is costly because people must move within a larger area to obtain sufficient resources and larger areas are 15 more difficult to defend. In situations of resource superabundance or scarcity/unpredictability, Dyson-Hudson and Smith (1978) predict that people will not create geographically stable territorial systems. They suggest that where resources are abundant and predictable but just adequate to meet a group?s needs, the benefits associated with defending a discrete area will likely outweigh the costs. Because resources are abundant, people can remain in a small area that is easy to defend. If they let others access their food and water, they would not have enough for the community to subsist on. Although Dyson-Hudson and Smith do not explicitly state that this scenario assumes that a population is at the carrying capacity of the local environment, their model assumes a threat of intruders who do not have enough resources of their own due to spatial and temporal fluctuations in productivity levels. This would occur to certain communities if the larger population of the region was at or near carrying capacity. Figure 1.3. Predictions of the economic defendability model reproduced from Dyson- Hudson and Smith (1978:26). 16 Some recent human behavioral ecology studies on territoriality in coastal settings also draw on social science models of competition (e.g., Carneiro 1970; Cohen 1977; Haas 1990) triggered by subsistence stress due to increased population density or decreased resource abundance (Bawden and Reycraft 2000; Boone 1992; Ember and Ember 1992; Jones et al. 1999; Lambert and Walker 1991). These studies emphasize that people also engage in alliances and other types of cooperative behaviors during periods of conflict (Gumerman 1986), which I will discuss in further detail below. Kennett and Kennett (2000) propose that on the Northern Channel Islands, the origins of a variety of competitive and cooperative responses such as violence and production of trade goods occurred at 1500-650 cal BP during a period of climatic instability and associated increased risk of resource shortfall. Their review of settlement pattern analysis indicates sites with more defensive characteristics during this period, and osteological data shows an increase in violence. Kennett and Kennet propose that increased fishing may be consistent with more intra-group cooperation. Evidence for exchange of food, beads, and microdrills also increases at this time. Kennett and his colleagues do not use the economic defendability model, but they use other human behavioral ecology models to further explore territorial behavior in the Channel Islands (e.g.,Kennett 2005; Kennett and Clifford 2004; Kennett et al 2009). Kennett and Clifford (2004) propose that during the Middle Holocene, Northern Channel Islands peoples moved between large semi-permanent coastal villages (prime locations for accessing shellfish on the rocky coast and freshwater) and inland hilltop villages (prime location for accessing plant resources and defensive sites) based in part on the presence or absence of other groups. They may have followed a bourgeois game theory strategy where players fight for resources that they 17 currently own (play the hawk) but do not challenge ownership of resources that they do not own (play the dove) if other options are available. Kennet (2005) traces the shift from an ideal free to an ideal despotic distribution in the Northern Channel Islands (see also Winterhalder et al. 2010). He suggests that people select locations for central places to achieve an optimal balance between accessing high-ranked resources and minimizing costs of travel and transport of resources. In an ideal free distribution where people are free to select any resource patch to exploit, they should aggregate near patches with the highest return rates, in this case areas on the coast near the mouths of large drainages. When the per capita yield of the patch decreases due to an increase in population density, at least some of the group should move to the next best patch near smaller drainages. Territorial behavior occurs when some groups protect the best resource patches and leave secondary resource patches to groups that are less competitive. This results in an ideal despotic distribution. Based on an increase in primary villages in less desirable places (from a central place foraging standpoint) at 1500 BP, Kennett suggests that there is a shift from an ideal free to ideal despotic distribution on the Northern Channel Islands during the Late Holocene due to increased population density resulting in territorial behavior. In her dissertation research and subsequent articles on land tenure on the Fiji Islands, Field (2003, 2004, 2005; Field and Lape 2010) integrates an economic defendability model with a social scientific approach to competition (e.g., Kennett and Kennett 2000; Lambert 1993; Lambert and Walker 1991; Raab and Larson 1997). She hypothesizes that people use competitive strategies when resources are abundant and predictable (Dyson-Hudson and Smith 1978) but both competition and conflict increase rather than decrease when resources are temporally unpredictable. She hypothesizes further that cooperative behaviors also increase at 18 times of resource unpredictability and scarcity. In Fiji, Field?s GIS-based spatial analysis of human activities and resource distribution indicates that climatic fluctuations associated with the El Ni?o Southern Oscillation (ENSO) caused drops in agricultural productivity. She tests her hypotheses using the distribution of fortified and unfortified sites, the distribution of buildings with religious architecture, and the distribution of objects used in exchange such as pottery. Field (2003) finds that in the area of the Sigatoka Valley that was most vulnerable to ENSO fluctuations due to a severe dry season, competitive behavior was most intense but cooperation was also present. In areas of the upper Sigatoka Valley where the environment was more stable, there is less evidence for violent conflict. Religious structures indicate inter-group cooperation. Boundary Permeability The research described above focuses on boundary defense suggests that both boundary defense and cooperative strategies that allow people to transcend boundaries are related components of a territorial adaptation. In contexts in which resources tend to be scarce or unpredictable, groups buffer against resource variance in many ways (see review in Gamble 2005:100-101). One important strategy is to share information and exchange resources with other individuals or groups (Boone 1992; Cashdan 1990; Dyson-Hudson and Smith 1978; Gumerman 1986; Halstead and O?Shea 1989). I define boundary permeability as the ability of individuals or families to strategically transcend territorial boundaries, often based on the strength of inter-village relationships between kin or friends. People transcend boundaries through reciprocal access agreements, ceremonial gatherings, marriage ties, and exchange (Fitzhugh et al. 2011:87; Malinowski 1922; McCoy et al. 2010; Rautman 1993; Smith 1988:241; Wiessner 1982). Just as boundary defense varies based on the costs and benefits of investing in 19 boundary maintenance, the degree and nature of boundary permeability can be investigated using a cost/benefit analysis. Boundary permeability and reciprocity arrangements occur in many contexts including kin, marriage, and friendship relationships and political alliances. Studies from anthropology (Wiessner 1977), biology (Hamilton 1964), sociobiology (Palmer 1991), evolutionary psychology (Dunbar and Kenyatta 2008; Dunbar et al. 1995; Kruger 2003), and human behavioral ecology (Boone 1992) all present evidence for frequent and mutually beneficial interactions between kin. On the Pacific Coast, dense populations and frequent interactions between groups facilitated by boat travel likely encouraged strong ties between groups. Marriage ties and friendships may have been as important or even more important than genetic ties. Marriage ties often involve the transfer of resources from one family to another through the initial transfer of bridewealth from the new husband and his family to the family of the new wife. This may lead to lasting relationship and exchange of resources between the two families (Apostolou 2008). The ethnographic literature for the Gulf of Georgia and the Channel Islands attests to the strength and complexity of these relationships. Hill et al. (2009:188) note ?extraordinary cooperation? involving economic transactions, political alliances, religious ceremonies, and other interactions between non-kin in many communities. They note than in many experiments, people choose to cooperate rather than compete despite minimal evidence of immediate rewards or benefits to their behavior. Thus boundary permeability is an essential part of territoriality models because it both structures and is structured by choices about boundary defense. The primary costs of boundary permeability include travel (both on land and over water), conflict with other groups, loss of resources through gifting or exchange, and potential for failure 20 of future reciprocation (Aldenderfer 1998; Cashdan 1990, 1992; Halstead and O?Shea 1982; Fitzhugh et al. 2011; Kennett and Kennett 2000; Smith 1988). Travel costs are determined not only by the difficulty of the landscape but also by the degree to which boundaries that must be crossed are defended. If boundary defense is low it should not be difficult for families who belong to different groups to meet and even form temporary communities. If boundary defense is high, the arrangements that allow people to transcend boundaries must become more formalized within the sociopolitical framework of the group. In this context, people will not be able to cross a boundary without giving something back in return. They must spend time and energy in maintaining the inter-group relationship. Thus, the relative costs of boundary permeability should be higher at times when boundary defense is more intense. The benefits are also lower at this time because resource availability is adequate. The benefits of boundary permeability include information exchange, marriage partners, acquisition of objects that may elevate individual and family status in the community, access to rare or essential resources beyond territorial boundaries, and ?buffering? against potential resource shortfall (Aldenderfer 1998; Arnold 1990a; Cashdan 1990, 1992; Fitzhugh et al. 2011; Halstead and O?Shea 1982; Janetski 2002; Kennett and Kennett 2000). Sharing between groups is particularly important because more information decreases uncertainty about resource distribution and abundance. Studies of reduction in unpredictable shortfall (sometimes referred to as risk reduction or risk minimization) indicate that people often make choices to minimize the possibility that they will fail at procuring a desired amount of resources (Halstead and O?Shea 1989; Hiscock 1994; Wiessner 1982). The chance that two groups will experience the same resource fluctuations decreases with distance (Mackie 2001; Smith 1988; Wiessner 1982). In choosing alliance partners, people must weigh the costs of transport and travel against the 21 possibility that an alliance partner may also be experiencing resource shortfall due to climate change or natural disaster (Cashdan 1985; Kennett and Kennett 2000; Rautman 1993; Rensink et al. 1991; Whallon 1989, 2006). Based on these considerations, boundary permeability should increase in contexts of resource scarcity or uncertainty when people perceive a greater benefit to exchange, sharing resources, and sharing information. The benefits of inter-group relationships have been addressed using a variety of models. In human behavioral ecology models used to predict pooling or sharing behavior, the relationship between resource abundance (harvest size) and value of the resource to the forager is expressed as a sigmoid-shaped utility function. The value of the resource increases quickly when resources are scarce but eventually that value levels off and increases more slowly (diminishing marginal utility) as harvest size increases. Thus, during times when resource availability far exceeds the needs of the community, extra food provides less value to the group so that the benefits (insurance against potential future shortfall, relationship-building, marriage partners) provided by sharing are greater than the costs incurred by losing the resource. This finding parallels the tolerated theft model (Blurton Jones 1984). During times of resource scarcity, resources have a relatively higher value to the group so the benefits of sharing are less likely to outweigh the loss of the resource (Figure 1.4; Smith 1988; Kohler and Van West 1996; Winterhalder, Lu and Tucker 1999). Based on this model, groups with extremely abundant resources should be willing to share or allow raccess to groups that have less, particularly if they know it could be their turn for shortfall in the near future. Groups with adequate resource may be less likely to share. This model does not distinguish between adequate and extremely scarce or unpredictable resources. Given low boundary defense and higher benefits of sharing information and subsistence 22 resources (Fitzhugh et al. 2011), groups with scarce or unpredictable resources may be more likely to engage in inter-group relationships. Figure 1.4. Reproduced from Smith?s (1988:235) risk reduction model of resource sharing. Individuals who share resources gain higher value from their resources [V(x )] than individuals who do not share resource [V (? + ?/2)]. A cost-benefit analysis of social networking by Fitzhugh et al. (2011) explores the issue of inter-group relationships in times of scarcity using a broad view of boundary permeability that incorporates not only environmental variability also interaction cost variability. They suggest that when environmental productivity/predictability and networking costs are low, people should pursue more connections at various scales from local to supraregional. If networking costs are high and resources are unproductive/unpredictable, people should rely more on individual traders rather than full-scale group interactions. When networking costs are high and resources are productive/predictable and, people should become more insular. If networking costs are low and resources are productive/predictable, people should interact in a competitive way. The most tightly interconnected communities are expected when resources are unproductive/unpredictable and networking costs are low (Figure 1.5) Fitzhugh et al. (2011) apply this model using ceramic 23 and lithic artifacts from the Kuril Islands between Japan and Siberia (see also Phillips 2011). I follow their basic model in integrating boundary defense and boundary permeability predictions to create a set of hypotheses for territoriality on the Pacific Coast of North America; however, I differentiate between formal and informal interactions across boundaries. Formal interactions are defined as community-sanctioned events, meetings, gifting, and trader-led exchange. Informal interactions encompass reciprocal access of common use areas, temporary groupings of family groups, and unplanned encounters. Figure 1.5. Reproduced from Fitzhugh et al. (2011:97), preditions for social networks given different levels of environmental predictability and networking costs. Hypotheses for Precontact Territoriality on the Pacific Coast of North America My hypotheses integrate an economic defendability model with expectations for boundary permeability based on the costs and benefits of inter-village relationships on the Pacific Coast. In constructing hypotheses for economic defendability, I consider resource 24 adequacy relative to population levels. If population density is low, food and water resources are more likely to exceed the needs of the population. I also design these hypotheses with the assumption that the distribution of terrestrial, marine, and freshwater resources in coastal areas is patchy. Shellfish occur on certain beaches, fish are abundant in kelp beds or along certain routes, and plant resources are found only in certain environments (Kennett 2005:29). Patchniness encourages territorial behavior because aggregated resources are less costly to defend (Cashden 1983). In general, if food and water resources adjacent to a community far exceed its needs either due to low population density or extremely abundant and spatially and temporally predictable resources, I predict minimal boundary defense. Since the value of the resource is higher to outside groups than to the group adjacent to the resource, they should tolerate ?theft? (e.g., Blurton Jones 1984). At times when resources are extremely abundant in one patch, they may also be abundant in multiple patches on the landscape and other groups may not need to trespass unless population grows rapidly. Although boundaries should be permeable, benefits of interactions should be low and therefore infrequent. There may be other less resource-driven reasons for interactions across boundaries such as obtaining marriage partners and prestige items. Where resources are scarce and/or unpredictable, territories are large and poorly defined and boundary defense should be minimal. If a community does not have a valuable patch to defend, they will not spend time and energy protecting their resources against outsiders. Unhindered travel and greater value of information in buffering against shortfall will encourage more frequent and informal interactions across boundaries. Finally, when villages are located near abundant and predictable resources that adequately satisfy the requirements of the community, people should engage in aggressive boundary defense. The 25 benefits of defending the resource will outweigh the costs. Aggressive boundary defense is associated with smaller territories because they are easier to defend given sufficient resources (Cashdan 1992). If population density is at or near carrying capacity, temporal and spatial perturbations in food and water resources will cause some people or groups to attempt to trespass on others? territories, resulting in aggressive boundary defense. In this context, boundary permeability should be low due to more restricted movement across the landscape. Interactions should be formal (Table 1.1). Table 1.1. General predictions for boundary defense and permeability for a community. This basic model must be further adjusted for coastal groups because marine and terrestrial resource productivity often shift independently of one another, which alters cost/benefit calculations for boundary defense and permeability. I generate the following scenarios for territoriality strategies in a coastal context (Table 1.2): In environments with long-term marine resource abundance adequate to the needs of the group but poor or unpredictable terrestrial resources, people should move more freely on land and focus on boundary defense at villages on the coast adjacent to fishing or shellfishing areas. Inter-group interaction on land should be frequent and informal. Over water or on the coast, interactions should be limited and formal. In environments with long-term terrestrial resource abundance adequate to the needs of the group but poor or unpredictable marine resources, boundary defense should focus on inland resources and freshwater sources. Interactions on land should be limited and formal. Interactions on water or on beaches should be more frequent and more informal. Resources Defense Permeability Unproductive/Unpredictable Low High, informal Adequate High Low, formal Highly productive/Predictable Low Low, formal Boundary 26 Table 1.2. Expectations for boundary defense and permeability in a coastal environment. Finally, the alternative hypothesis to the above is that although some degree of defense took place at the level of the village, kin and marriage relationships may have rendered boundaries permeable to certain people or families regardless of the marine or terrestrial productivity. Cashdan (1983) suggests that even when people do not engage in active boundary defense, they tend to participate in systems of reciprocal altruism in which they control access to social groups. This is particularly effective with sparse and unpredictable resources in which information exchange is vital to increasing foraging efficiency. In other words, the benefits of frequent interactions may always have outweighed the costs. If this is the case, people should defend areas beyond a single village and they should access resources outside that area as well. As Kennett (2005) notes, complex group formations often occur in coastal areas because many foraging expeditions depends on cooperation between group members. This alternative hypothesis is more consistent with descriptions of interconnected communities in the ethnographic record for both study areas. Marine Terrestrial Water Land Water Land - - Low Low High, informal High, informal + - High Low Low, formal High, informal - + Low High High, informal Low, formal + + High High Low, formal Low, formal Boundary Defense PermeabilityRes urc s 27 Predictions To test the hypotheses presented above, I analyze settlement pattern data and lithic procurement. To determine if high boundary defense occurs when marine and terrestrial resources are adequately productive and predictable, I test whether there are significant differences in defensive characteristics of sites occupied during different climatic at demographic regimes. The bulk of the analysis focuses on toolstone procurement patterns. Unlike fish, birds, and other subsistence resources, rocks do not shift with changing environmental conditions. If the area over which people could access resources decreased during periods productive/predictable resources due to increased boundary defense, the lithic assemblage should reflect the toolstone that was available nearby. If the area over whch people accessed resources increased during period of unproductive/unpredictable resources due to decreased boundary defense, the lithic assemblage should reflect a variety of toolstone sources and/or the most attractive toolstone areas on the landscape. Alternatively, if boundaries always remained permeable to kin or friends, different areas of the site should show significant differences in toolstone sources corresponding with different resource access that is based on inter-group family connections rather than village connections. In both study areas, most toolstone comes in the form of local beach cobbles. I use cobble size, shape, and cortex appearance to match lithic artifacts at habitation sites to sources on the landscape. I also consider degree of processing and conservation as an indicator of easier or more difficult access to high quality toolstone. None of these lines of evidence are sufficient on their own to demonstrate change over time in toolstone access, but separate lines of evidence can be ?triangulated? to establish convergence in support of a hypothesis (Wylie 2002). 28 I examine boundary permeability mainly through analysis of the abundance and distribution of items that would have marked exchange relationships between groups such as extra-local toolstone, beads, and other ornamental objects. Although food and everyday materials were also likely exchanged between groups, more unique objects and materials can be recognized as status markers. They serve as a visible means to maintain a relationship. Frequent and informal intergroup interaction should result in a more dispersed pattern of exchange items while infrequent and formal intergroup interaction should results of concentrations of rare or ornamental objects in certain households. Dissertation Organization In Chapter 2, I discuss current research on Late Holocene climate change in both study areas, current settlement pattern research, and other relevant background information on precontact lifeways and previous research in each study area. I present a quantitative analysis of the defensive characteristics of sites for both the San Juan Islands and San Nicolas Island. In Chapter 3, I provide detailed information on the archaeological sites from which I analyze lithic assemblages to test my predictions. I describe spatial and temporal analytic units based on dating and stratigraphy at each site. For the Watmough Bay site, excavations were completed in 1968 and 2004. Extensive dating efforts by previous researchers made creating analytic units relatively straightforward. For the Tule Creek Village Mound B site, sediment analysis and detailed stratigraphic analysis were required to create temporal units. Chapters 4-6 focus on the lithic analysis used to test the hypotheses presented in Chapter 1. In Chapter 4, I discuss the results of field surveys in both study areas to determine the distribution, extent, and accessibility of high quality toolstone. For the San Juan Islands study 29 area, this required geochemical analyses to determine the source provenance of the toolstone. For San Nicolas Island, toolstone surveys were sufficient since the provenance of the material was identified in previous studies. Chapter 5 provides background information on the flake tool technology, formal tool technology, and reduction sequences for each site based on the lithic analysis. In Chapter 6, I discuss my predictions for the lithic assemblages at each site and the results of analyses designed to test these predictions to investigate changes in lithic procurement. In Chapter 7, I summarize the results of the research in the context of larger theoretical, methodological, and topical issues for each study area. I also identify areas for future research. 30 Chapter 2: Pacific Coast Paleoenvironments and Settlement Patterns General similarities in the San Juan Islands and southern Channel Islands, such as an island setting, reliance on marine resources, boat transport, and large multi-house villages, allows the application of the same hypotheses in both places. Differences between the study areas in distance from the mainland, distance between islands, paleoenvironment, and settlement pattern facilitates a more rigorous test of hypotheses about territoriality in coastal hunter-gatherer-fisher communities. Exploring questions from a broader coastal archaeology framework rather than from one study area also helps me to challenge some of the assumptions associated with the culture histories and research traditions of the Gulf of Georgia and Southern California Bight. In this chapter, I review information about the culture histories, environments, and settlement patterns of each study area. Since the hypotheses discussed in the previous chapter consider territoriality in the context of resource availability, detailed examination of paleoenvironmental data are essential to predicting shifts in boundary defense and boundary permeability strategies in the Late Holocene. Settlement pattern data are an essential component of an investigation into territoriality for several reasons. Demographic data are important because population density affects relative resource abundance. Calculating actual population levels using archaeological data is beyond the scope of this research, but site size and abundance provides insights on the size of precontact communities. The location of sites on the landscape are also important to understanding which resources people chose to settle near and defend. Finally, I also examine the distribution and characteristics of defensive sites in both regions to test predictions about active defense of village sites. 31 The San Juan Islands Study Area Less than ten kilometers offshore, the San Juan Islands are part of an archipelago of 450 islands north of the Puget Sound between Vancouver Island and the Washington and British Columbia coasts. Along with the Gulf Islands, the Lower Fraser River, the Strait of Georgia, northern Puget Sound, and southeastern Vancouver Island, they form part of the Gulf of Georgia culture area (Stein 2000; Suttles 1990). Most of the San Juan Islands are held by private landowners (approximately 85%) with the exception of San Juan Island National Historical Park, Washington State Parks, San Juan County Parks, areas protected by federal agencies, land trusts and other organizations, and land held by the University of Washington. The coastal landscape is characterized by rocky and sandy beaches. Inland areas are characterized by open prairies, farmland, and mixed coniferous forest. Today, several Coast Salish speaking native communities include the San Juan Islands within their traditional territories. These include the Lummi Nation, the Samish Indian Nation, the Swinomish Indian Tribal Community, the Songhees Nation, and the Saanich Nation The coastal archaeological record is dominated by shell midden sites. Inland sites are scarce due to poor preservation of organic material in acidic soil and limited research. These sites are characterized by isolated finds of lithic artifacts. Kenady et al. (2008) also report a surface lithic scatter on a high bedrock outcrop on San Juan Island. The culture history of the Gulf of Georgia area presents a narrative of a terrestrial- oriented initial colonization (Cascade Phase, 9000-4500 cal BP), followed by increased use of marine resources and related fishing technology. Shell middens appear in the Locarno Beach Phase at 4500-2500 cal BP, although shell-bearing deposits are found earlier elsewhere on the Northwest Coast (Stein 2000). Gulf of Georgia researchers propose an increase in population 32 density and sedentism at ca. 2500-1500 cal BP during the Marpole phase (Croes and Hackenberger 1988; Matson 1983, 1985; Matson and Coupland 1995). The appearance of larger houses and villages during the Marpole is thought to reflect increased social stratification and changes in the ways that resources were controlled (Ames 1994, 1995; Ames and Maschner 1999; Grier 2003). Researchers also cite evidence for intensified salmon fishing and associated tools (Burley 1980; Croes and Hackenberger 1988; Matson 1983, 1992; Mitchel 1971), ornamental and artistic objects, wood and bone craft specialization (Burley 1980; Matson and Coupland 1995), large houses, multihouse settlements (Ames 1994; Grier 2001, 2003), and increased regional exchange of obsidian and other materials (Carlson 1994; Grier 2003). The roots of sociopolitical structures observed in the historic period are thought by some to lie in the Marpole phase (Burley 1980; Carlson 1960; Matson and Coupland 1995; Mitchell 1971). Lepofsky et al. (2005) suggest that some sociopolitical developments are associated with a warmer and drier Gulf of Georgia during the Marpole that led to resource scarcity that was most acute outside the Fraser River region. This may have encouraged exchange relationships between the Fraser Valley and other Gulf of Georgia communities. People with kin relationships in the Fraser would have developed higher status. Similarities in artifact assemblages, artistic and status-marking items such as labrets and other decorative objects, and the distribution of materials considered trade items throughout the Gulf of Georgia region provides evidence for increased or stabilized social and economic relationships throughout the region (Burley 1981; Grier 2003). The period following the Marpole, known as the ?Late Phase? or ?San Juan Phase? (1500 cal BP-Contact), is associated with a continuation of many of the cultural trends that appeared during the Marpole but a decrease in abundance and variety of ?elaborate goods?, possibly due to 33 a change in burial practices and/or a shift in social organization (Lepofsky et al. 2005:273), a decrease in chipped stone, and an increase in bone and antler tools. In general, the San Juan Phase is poorly studied and this research will contribute to a greater understanding of this time period. Gulf of Georgia Paleoenvironment Today, the Northwest Coast climate is classified as ?mild maritime? with cool summers and wet and mild winters. Due to the rain shadow effect of the Olympic Mountain Range, San Juan Islands summers are drier than mainland summers. Dominant tree species include douglas fir (Pseudostuga mensiesii), Pacific madrone (Arbutus menziesii), white oak (Quercus garryana) big leaf maple (Acer macrophyllum), grand fir (Abies grandis), lodge pole pine (Pinus contorta), Pacific yew (Taxus brevifolia), western hemlock (Tsuga heterophylla), western red cedar (Thuja plicata) and red alder (Alnus rubra). Understory species include ocean spray (Holodiscus discolors), snowberry (Symphoricarpus albus), nootka rose (Rosa nootkanensis), and salal (Gaultheria shallon) (Mitchell 1971; Wessen 1986). Due to mixing of cold and highly saline ocean waters with brackish surface waters, the diverse marine environment of the Gulf of Georgia is characterized by rich kelp forests and eelgrass beds. Freshwater is more limited on the San Juan Islands than on the adjacent mainland. Below I review both marine and terrestrial climate change in the Salish Sea from the beginning of the late Holocene at 4500 cal BP. Since change over time in sea surface temperature and upwelling both have important implications for the productivity of marine flora and fauna (Daniels 2009; Kozloff 1990, 1993), both are considered in this review of marine 34 climate change. My discussion of terrestrial climate change focuses on temperature and precipitation. Marine Paleoenvironment in the Gulf of Georgia Following a relatively stable and productive Middle Holocene, the Late Holocene (4500 cal BP-Contact) Gulf of Georgia marine environment was characterized by changes in upwelling and sea surface temperature, both of which affect the productivity of the marine resources upon which native communities relied (Figure 2.1). Climate researchers use increases and decreases in fauna and oxygen isotopes to pinpoint shifts in upwelling and sea surface temperature, which do not always change in tandem in the Gulf of Georgia (Daniels 2009). The period from 4000- 2800 cal BP is characterized by cooler sea surface temperature based on analysis of diatoms, silicoflagellates and biogenic silica in Effingham Inlet, Vancouver Island (Hay et al. 2007). According to Daniels? (2009) sea surface temperature study using stable isotope data from Protothaca staminea (littleneck clam) from the English Camp site on San Juan Island, ocean waters warmed at ca. 700-1400 cal BP, cooled at ca. 1400-1000 cal BP, and warmed again at ca. 1000-300 cal BP. Preliminary data suggest possible warming at 300-100 cal BP, but Daniels (2009:80-82) was unable to rule out sampling issues. Regarding upwelling, the period from 4000-2800 cal BP was characterized by decreased upwelling based on analysis of diatoms, silicoflagellates and biogenic silica in Effingham Inlet, Vancouver Island (Hay et al. 2007). Changing proportions of fish taxa in the same area indicate increased upwelling afterwards at 2800-1500 cal BP (Wright et al. 2005). Daniels establishes upwelling history by building on a local marine reservoir chronology created by Deo et al. (2004) using paired shell and charcoal dates. Since increased local upwelling causes an increase 35 in C 14 depleted carbon in surface waters, changes in marine reservoir between local and global surface water (?R) can be used as a proxy for increased upwelling (Daniels 2009:51). Daniels also considers stable carbon and oxygen isotope data from shell samples from the English Camp and Watmough Bay sites on the San Juan Islands. Like Wright (see also Tunnicliffe 2001), she finds an increase in upwelling at 3000-1400 cal BP. She also notes a decrease in upwelling at 1000-600 cal BP followed by an increase in upwelling 600 cal BP to modern times (Daniels 2009:70). Terrestrial Paleoenvironment in the Gulf of Georgia On land, a decrease in temperature and increase in precipitation at 3000-2400 BP are consistent with a brief neoglacial advance (Brown and Hebda 2003; Hallett et al. 2003; Hebda 1995; Long et al. 1998; Pellatt et al. 2001). At about 3000 years ago, increased precipitation and decreased temperature led to closing of the canopy in western hemlock and cypress-family forests, an increase in wetlands (Brown and Hebda 2003), and more cedar, western helmlock, spruce, and douglas fir (Pellatt et al. 2001). Glacial advances are documented throughout British Columbia at this time (Koch et al. 2007; Luckman 1994). At 2400-1200 cal BP, the Gulf of Georgia region experienced a warmer and drier climate associated with anomalies in the mid- troposphere caused by climate forcing (Hallett 2001; Lepofsky et al. 2005). Glacial retreats (Koch et al. 2007) and increases in soil charcoal in upper Fraser lake sediments associated with increased fire caused by drier wood and increased electrical storms (Hallett 2001; Hallett et al. 2003; Lertzman et al. 2002) establish the Fraser Valley Fire Period in southwestern British Columbia. 36 Paleoecologists in the San Juan Islands (Fujikawa 2002; Sugimura et al. 2008) identify a similar climatic shift in the San Juan Islands. The findings are based on both an increase in charcoal and an increase in fire-adapted alder (Alnus) pollen in sediment cores from bogs on Mt. Constitution on Orcas Island. Brown and Hebda (2003) also propose an increase in fires on Vancouver Island after 1940 cal BP based on charcoal records from lake sediments. It is possible that anthropogenic burning played a role in increased fire frequency in the Gulf of Georgia region, but the weight of the evidence and contemporaneous droughts in western North America (Gavin et al. 2003; Hallett et al. 2003; Meyer and Pierce 2003) support a warmer and drier climate at 2400-1200 cal BP (Fujikawa 2002; Hallett et al. 2003; Lepofsky et al. 2005:277-278). Paleoecological research in the Olympic Mountains (Gavin and Brubaker 1999) and southwestern British Columbia (Hallett et al. 2003) suggests that after a brief period of increased precipitation after the Fraser Valley Fire Period, the climate returned to warmer and drier conditions in the islands at 1050-600 cal BP (Gavin and Brubaker 1999). At 300-100 BP, the climate cooled again. Evidence for this cooling event has been found in pollen data on western Vancouver Island (Gavin and Brubaker 1999) and the Olympic Peninsula (Greenwald and Brubaker 2001), tree ring data from Mount Rainier (Graumlich and Brubaker 1986), glacial sediment in the British Columbia Coast Mountains (Koch et al. 2004; 2007; Larocque and Smith 2003; Ryder 1987, 1989; Smith and Desloges 2000; Walker and Pellat 2003), and increased charcoal in Upper Fraser lake sediment (Hallett et al. 2003). 37 Figure 2.1. A summary of terrestrial climate, upwelling, and sea surface temperature for the Late Holocene Gulf of Georgia. Gulf of Georgia Marine Resources To address hypotheses regarding shifts in human territoriality in the context of climate shifts, I consider the impact of paleoenvironmental changes on the resources upon which precontact Gulf of Georgia communities depended. Shell middens are typically dominated by littleneck clam (Protothaca staminea), butter clam (Saxidomus giganteus), and blue mussel (Mytilus trossulus). People also gathered cockle (Cardiidae), horse clam (Tresus spp.), snail (Gastropoda), limpet (Littorina spp.), dogwinkle (Nucella spp.), periwinkle (Lottidae), barnacle (Balanus spp.), bentnose clam (Macoma spp.) and sea urchin (Strongylocentrotus spp.) (Wessen 1986). The San Juan Islands offer a fishing advantage in that salmon (Oncorhynchus spp.) returning to the Fraser River to spawn must pass through narrow passages between the islands (Mitchell 1971). People also fished for steelhead (Salmo gairdneri), eulachon (Thaleichthys pacificus), smelt (Spirinchus spp. and other genera), sturgeon (Acipenser spp.), Pacific lamprey (Lampetra tridentata), halibut (Hippoglossus stenolepis), Pacific herring (Clupea harengus), cod (various genera), dogfish (Squalus acanthias), perch (various genera), sculpins (Cottidae), and flatfish (Pleuronectidae) (Suttles 1990). Native communities exploited Waterfowl such as 38 cormorants (Phalacrocorax auritus, penicillatus, and pelagicus), sea ducks (Merigini) and pochard ducks (Aythyini) (Bovy 2005). A variety of marine mammals were available for hunting including northern elephant seal (Mirounga angustirostris), California sea lion (Zalophus californianus), northern sea lion (Eumetopia jubatus), Dall?s porpoise (Phocoenoides dalli), and harbor seal (Phoca vitulina). The Gulf of Georgia environment is characterized by rich marine resources but because productivity varies over both space and time based on precipitation, altitude, salinity, and shoreline (Suttles 1960b [1987:26-27]), depictions of this area as a ?Garden of Eden? (Erlandson 1994) are misleading. Higher sea surface temperature diminishes microscopic marine plants and kelp forests, resulting in diminished fish and shellfish populations. Warmer sea surface temperatures are also associated with the algal species that cause paralytic shellfish poisoning (Horner et al. 1997). Late summer droughts cause fires and increase erosion which increases silt and streams and diminishes spawning. In places like the lower Fraser River where all five species of Pacific salmon thrive, people may not have been adversely impacted (Lepofsky et al. 2005). On the San Juan Islands, however, a decline in fish populations could have caused a shortfall in subsistence resources. Bovy (2005, 2007) notes that warm water events like El Ni?o can diminish some seabird populations on the Northwest Coast. Thus, during periods of warmer sea surface temperature and decreased upwelling, many marine resources would have become less abundant and/or more spatially unpredictable. Based on the hypotheses outlined in the previous chapter, I predict that this would have resulted in more boundary defense surrounding more abundant and predictable marine resource patches (unless they far exceeded the needs of the group), and decreased defense of boundaries surrounding less abundant/predictable resource patches. I also predict increased permeability to informal inter-group interactions. 39 Gulf of Georgia Terrestrial Resources Regarding the terrestrial environment, people gathered plants such as camas (Camassia quamash), wapato (Sagittaria latifolia), and a variety of berries (Barnett 1955; Stein 2000; Suttles 1990; Wessen 1986). They hunted terrestrial mmamals including river otter (Lontra canadensis), Columbia black-tailed deer (Odocoileus hemionus), bear (Ursus spp.), and elk (Cervus elaphas) (Wessen 1986; Jim Kenagy pers. comm). During the drier Fraser Valley Fire Period and Medieval Warm Period, the islands would have been more adversely impacted than the neighboring mainland for several reasons, First, they are less ecologically diverse with only 2 of 13 Gulf of Georgia biogeoclimatic variants (Lepofsky et al. 2005). Lepofsky et al. (2005:278) note that some mammals such as mule deer, elk, marmot, and snowshoe hare would have flourished in more open forests following large fires, but other taxa would have declined. Increased open meadows would have negatively affected some plant resources although berries and camas may have thrived in a more open environment. Despite an increase in some subsistence resources, a shortage of freshwater on the San Juan Islands may have been problematic for people and for certain animals during times of drought. The islands are drier than the surrounding area due to the rain shadow effect of the Olympic Mountains; there are only six perennial streams to choose from. Most streams on the islands have no flow between June and November (Dietrich 1975:68; Wixom and Snow 2004). The higher elevation areas on eastern Orcas Island receive more rainfall than other areas with approximately 76-114 centimeters per year while southern San Juan Island and Lopez Island are drier, receiving 51-64 inches per year (Detrich 1975:60) (Figure 2.2; Table 2.1). 40 Figure 2.2. Map showing streams that were perennial freshwater sources (arrows with numbers refer to Table 2.1 below) on the San Juan Islands modified from Taylor et al. 2011. 41 Table 2.1. Descriptions of primary freshwater sources in the San Juan Islands based on Dietrich 1975; Wixom and Snow 2004. Reproduced from Taylor et al. (2011). Map # Freshwater Source Area Description 1 Cascade Bay, Orcas Island Cascade Bay provides access to an unnamed creek 400 m SW of Cascade Lake (volume 4,600 acre-ft.). This is a high precipitation area surrounding Mt. Constitution with a large spring that feeds Cold Creek, a high-flow perennial stream that runs into Cascade Lake. 2 Buck Bay, Orcas Island Buck Bay provides access to the mouth of Cascade Creek, a high discharge stream fed by Mountain Lake (volume 8,800 acre-ft.) located in the high precipitation area surrounding Mt. Constitution. 3 Unnamed Bay, Blakely Island The large bay on western Blakely Island provides access to an unnamed creek and is 200 m from Spencer Lake (volume 5,400 acre-ft.). 4 Swifts Bay, Lopez Island The Swifts Bay watershed is fed by Hummel Lake (volume 272 acre-ft.). An unnamed stream runs from the lake to the bay. 5 False Bay, San Juan Island The False Bay watershed is fed by streams running from Trout Lake (volume 1,400 acre-ft.) on Mt. Dallas and Zylstra Lake (volume 350 acre-ft.). San Juan Valley Creek begins at Trout Lake and runs year round. 6 Garrison Bay, San Juan Island The source of freshwater to Garrison Bay is a year- round creek with its head on the north side of Mt. Cady, a high precipitation area on northern San Juan Island. San Juan Islands Settlement Patterns Prior to this research, settlement pattern data on the San Juan Islands consisted of dates from only a handful of the largest shell midden sites. The goal of the San Juan Islands Archaeological Project (2005-2009) was to date additional ?big? shell midden sites (larger than 3,000 square meters) and ?small? shell midden sites (smaller than 3,000 square meters) (Taylor et al. 2011). Definitions of site size for different time periods for multi-component sites was based on augering and spatial data from previous excavations. I conducted this project with Dr. Julie Stein (Principal Investigator), and Stephanie Jolivette (University of Washington Graduate 42 Student). We sampled shell middens throughout the islands by augering and collecting shell from the eroding bank. The results of this project are reported in scholarly articles and unpublished reports at the Burke Museum and the Washington State Department of Archaeology and Historic Preservation (Taylor and Stein 2006, 2007; Taylor et al. 2009a, 2011). From 2005-2009 the SJIAP obtained a total of 84 dates from 41 sites (Taylor et al. 2011:293-296) and combined these data with 145 previously published dates (Bovy 2005; Daniels 2009; Deo et al. 2004; Stein et al. 2003; Walker 2003) from a total of 50 sites in the San Juan Islands. Based on work by Deo et al. (2004) and refined by Daniels (2009), the regional marine reservoir correction value (?R) is 400 years at 0-600 cal BP and 1000-3000 cal BP and 0 years at 600-1000 cal BP due to decreased upwelling. The data show small numbers of sites on the islands beginning at 4000 cal BP, an increase in sites at approximately 2500 cal BP, and the highest frequency of dates at 650-300 cal BP (Figure 2.3-2.11). The scarcity of older sites is likely due to sea level change and erosion. Sea level history in this region is complex due to influences of isostatic depression and rebound, eustatic sea level change, and tectonic processes (Clague and James 2002; Fedje et al. 2009; Hutchinson 1992; James et al. 2005; Mosher and Hewitt 2004) . In the Cascadia region, as the glaciers melted and isostatic rebound began, relative sea level fell from 75 meters above sea level to modern levels by 11,700 + 170 RCYBP (James et al. 2005; Mosher and Hewitt 2004:25). The land continued to emerge and sea level may have been over 10 meters below its present level at 9000 RCYBP (see also Clague 1981; Dethier et al. 1996). At 5000 RCYBP, relative sea level rose to within two meters of modern sea level because eustatic sea level change outpaced tectonic uplift (Clague and James 2002; Mazzoti et al. 2008). Sea level rose to within a meter of its present position within the last 2000 years (Fedje et al. 2009; Grier et al. 2009; Whittaker and 43 Stein 1992). Underwater excavation for early sites has been attempted in the Gulf Islands (Easton and Moore 1991) and may hold potential in the San Juan Islands. Megafauna finds (Kenady et al. 2007, 2011; Wilson et al. 2009) suggest a potential Terminal Pleistocene/Early Holocene human occupation on the islands. Efforts to find inland lithic manufacture, hunting, and habitation sites are ongoing. Figure 2.3. Settlement pattern map for the San Juan Islands at 4000-3500 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites. 44 Figure 2.4. Settlement pattern map for the San Juan Islands at 3500-3000 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites. Figure 2.5. Settlement pattern map for the San Juan Islands at 3000-2500 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites. 45 Figure 2.6. Settlement pattern map for the San Juan Islands at 2500-2000 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites and white squares indicate small sites. Figure 2.7. Settlement pattern map for the San Juan Islands at 2000-1500 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites and white squares indicate small sites. 46 Figure 2.8. Settlement pattern map for the San Juan Islands at 1500-1000 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites and white squares indicate small sites. Figure 2.9. Settlement pattern map for the San Juan Islands at 1000-500 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites and white squares indicate small sites. 47 Figure 2.10. Settlement pattern map for the San Juan Islands at 500-0 cal BP. Numbers indicate site numbers recorded in the Department of Archeology and Historic Preservation site database. Black circles indicate big sites and white squares indicate small sites. The slight increase in sites on the San Juan Islands at 2500 cal BP is similar to findings for the Canadian Gulf of Georgia (Lepofsky et al. 2005); however, a more dramatic increase in sites at 650-300 cal BP only occurs in the San Juan Islands (Figure 2.11, 2.12). Site size differs significantly before and after 650 cal BP (?2 = 16.931, 1 df, p <0.001), with more big sites prior to 650 cal BP and more small sites afterwards. One possible explanation for the discrepancy between the two records concerns climate change. The Fraser Valley area provided more resources than the Gulf Islands and San Juan Islands at 2400-1200 cal BP during the Fraser Valley Fire Period. Lepofsky et al. (2005) note that the Fraser region had large volume of freshwater, ecological diversity, a large salmon run, and high waterfowl populations. Their summed probability plot that shows an increase in sites and/or a dispersion of people across the landscape supports the hypothesis that people clustered in a productive area of the Gulf of Georgia. Perhaps after the climate became cooler and wetter after the Medieval Warm Period at 600 cal BP, part of the larger Fraser Valley population relocated to the San Juan Islands. 48 Research on site vulnerability to erosion based on fetch, bathymetry, vegetation, and landform by Taylor et al. (2011) indicates that erosion is also probably a factor in the smaller number of sites at 2500-15000 cal BP but is not enough to explain this pattern. Figure 2.11. Summed probability plot for San Juan Islands dates reproduced from Taylor et al. 2011. Following the same protocol as Lepofsky et al. (2005), this plot incorporates only one date per site per 200-year interval to ensure that the better-dated large sites do not bias the chronology. Figure 2.12. Summed probability plot for San Juan Islands based on SJIAP data and for the Canadian Gulf of Georgia reproduced from Lepofsky et al. (2005). Both plots incorporate only one date per site per 200-year interval. Territoriality Predictions for the San Juan Islands To create more specific expectations for boundary defense and permeability in the Late Holocene San Juan Islands I apply the boundary defense/permeability concepts outlined in the previous chapter to the specific environment and settlement patterns of the San Juan Islands record. The arbitrary time periods chosen for both settlement pattern and lithic analysis are 49 based on both paleoenvironmental shifts and the dating of the assemblages from the Watmough Bay site. During the period from 3500-2500 cal BP, a cool wet terrestrial climate and a shift towards increased upwelling would have provided a resource supply that far exceeded the demands of a relatively low population. I predict minimal boundary defense and moderate boundary permeability reflecting longer-distance travel and the need to maintain large social networks to find marriage partners. Population may have risen slightly during the 2500-1600 cal BP period when terrestrial productivity was low and marine productivity was high. I propose that marine resources were still so abundant that people would not have defended boundaries around their villages. I predict minimal boundary defense and moderate boundary permeability reflecting longer-distance travel and the need to maintain large social networks to find marriage partners. During the 1600-1000 cal BP period, low terrestrial and marine productivity should be associated with minimal boundary defense on most of the landscape and high permeability on both land and water due to resource unpredictability. During the 600 cal BP-Contact period, with both high terrestrial and marine productivity and a higher population, communities located near adequately productive resource patches should defend boundaries against intruders more intensely and participate in more formal but less frequent boundary crossings (Table 2.2). Table 2.2. Expectations for Late Holocene boundary defense and permeability for the San Juan Islands based on paleoenvironmental and demographic shifts. San Juan Islands Population Marine Terrestrial Water Land Water Land 600 cal BP-Contact High + + High High Low, formal Low, formal 1600-1000 cal BP Medium - - Low Low High, informal High, informal 2500-1600 cal BP Medium + - Low Low Moderate, informal Moderate, informal 3500-2500 cal BP Low - + Low Low Moderate, informal Moderate, informal Resources Boundary Defense Permeability 50 San Juan Islands Defensive Sites Quantifying the abundance and distribution of defensive sites after 650 cal BP in the San Juan Islands provides information on the amount of effort people put towards the active defense of boundaries. Previous research on central Northwest Coast settlement patterns suggests an upsurge in defensive sites at approximately 900-1000 cal BP (Keddie 1984, 1996; Moss and Erlandson 1992). Angelbeck?s (2009) archaeological, ethnographic, and ethnohistoric dissertation research on the Coast Salish indicates an increase in fortified defensive sites at 1600- 500 BP and again during Euroamerican colonization. Radiocarbon dating on trench embankment sites for the Coast Salish shows an increase in sites at approximately 400 cal BP (Angelbeck 2009:262). Schaepe (2006, 2009) describes multi-village corporate family group defensive systems using built and natural structures on the Fraser River. Considering the San Juan Islands record alone, however, indicates minimal evidence of built defensive sites. Suttles (1949:70) notes defensive structures built by the Lummi and Anglebeck reports four trench-embankment fortifications on the southern islands. He also reports a rock-wall fortification on Hunter Bay on southern Lopez Island. Angelbeck (2009:224-225) notes that although some archaeologists have assumed that the Coast Salish were defending their territory against people who lived to the North, there is no concentration of defensive sites of any type at that border, rather they are relatively evenly dispersed throughout the region. To determine if people in the San Juan Islands aggressively defended their boundaries during certain time periods, I look beyond built earthworks that can be affected by post- depositional processes and modern disturbances. I rely on natural site location and context to investigate the defensive potential of a site by using Martindale and Supernant?s (2009) defensibility measures. Martindale and Supernant propose simple quantitative measures of 51 visibility, site elevation, accessibility, and area that can be calculated for almost any site based on site forms and reports. For this study, I focus on visibility and elevation measures because these can be calculated most accurately and they are more appropriate to the San Juan Islands sites that often have completely accessible approaches. Visibility is calculated as the degrees of visibility beyond 100 meters divided by the total degrees of approach around the site. Higher visibility is assumed to make a site more defendable than lower visibility because people would have more time to prepare for a potential attack. Elevation is measured as the slope measured from a point inside the site to a point just outside the site. A higher slope gives an advantage to defenders because it is more difficult for attackers to access the site and forces the attackers to congregate at an area near the defenders as they attempt to gain access, making them more vulnerable to a counter-attack (Martindale and Supernant 2009:195). I measure elevation change towards the water assuming that enemies are coming from water rather than land. The elevation measure is expressed as the ARCTAN of the elevation difference from the inside to the outside of the site divided by the site radius. Radians are converted into degrees and divided by 90 to provide a ratio between 0-1. I also calculate distance from site to a lookout point (point on the landscape with greater than 200 degrees of visibility over water) with the expectation that increased defensiveness should correspond with decreased distance to a lookout (Appendix A). In comparing defensive characteristics of sites, I consider two sets of data. For the first dataset, I assign each site to the earliest time period to which it dates. For the second dataset, all multi-component sites are assigned to every time period to which they date. The first dataset emphasizes the period when the site location was first chosen since people may pick a site due to its defensive capabilities but continue to use it due to community tradition or for other less 52 strategic reasons. The second dataset emphasizes each set of sites on the landscape used during each time period. If boundary defense in the San Juan Islands was more intense at 600 cal BP-Contact, I expect a significantly higher value for visibility and elevation measures and lower measures for distance to lookout during this time period. Although only certain sites would be located near resource patches that would be worth the effort of defending, overall, higher values for these sites should increase the overall mean for sites dating to the 600 cal BP-Contact time period. I use one-way ANOVA tests to determine if differences in mean visibility, elevation, and distance to lookout measures for each time period are significantly greater than differences that might be produced by random sampling. The results of this analysis indicate no significant difference in defensibility measures between time periods for either the first or second dataset (Table 2.3, 2.4). Table 2.3. Results of an ANOVA for visibility, elevation, and distance to lookout measures for sites in the San Juan Islands. Sites are assigned to the first time period in which they appear. Table 2.4. Results of an ANOVA for visibility, elevation, and distance to lookout measures for sites in the San Juan Islands. Sites are assigned to every time period in which they are inhabited. Time Per. (cal BP) n x Visibility F Sig. x Elevation F Sig. x Distance to lookout (m) F Sig. 600-Contact 32 0.44 0.191 0.477 0.24 0.477 0.7 358.13 0.68 0.569 1600-1000 8 0.49 0.19 400 2500-1600 7 0.44 0.22 407.14 3500-2500 4 0.41 0.18 750 Time Per. (cal BP) n x Visibility F Sig. x Elevation F Sig. x Distance to lookout (m)F Sig. 600-Contact 43 0.46 0.147 0.931 0.22 0.306 0.82 387.44 0.806 0.495 1600-1000 13 0.44 0.19 476.92 2500-1600 10 0.43 0.2 485 3500-2500 4 0.41 0.18 750 53 I also used Student t-tests to compared mean defensibility values before 600 cal BP and after 600 cal BP, combining time periods prior to 600 cal BP when boundary defense is predicted to be least intense, and comparing that sample to sites dating to after 600 cal BP when boundary defense is expected to be least intense. Results of this analysis indicate no significant difference in mean visibility, elevation, or distance to lookout before and after 600 cal BP for either the first or second dataset (Table 2.5, 2.6). Potentially, the large sites that date to the 600 cal BP-Contact period could be considered sites near high-value resource patches that would be more likely to be defended. Considering all time periods, there is also no significant difference in mean visibility, elevation, or distance to lookout between big sites and small sites (Table 2.5). Considering only sites that date to 600 cal BP-Contact (using the dataset in which sites are included in every time period in which they appear), there is also no significant difference in mean visibility, elevation, or distance to lookout between big sites and small sites (Table 2.6). Table 2.5. Results of t-tests comparing means for defensive measures for sites before and after 600 cal BP, equal variances not assumed. Dataset 1 = Sites are assigned to the period in which they first appear. Datset 2 = Sites are assigned to every time period in which they are inhabited. Table 2.6. Results of t-tests comparing means for defensive measures for big sites and small sites dating to all time periods. Results of t-tests comparing means for defensive measures for big sites and small sites dating to after 600 cal BP, equal variances not assumed. Dataset 1 n x VisibilityF t p (2-tailed) x Elev. F t p (2-tailed) x Distance to lookout (m) F t p (2-tailed) Before 600 cal BP 19 0.45 1.83 0.22 0.83 0.2 0.89 -1.1 0.28 476.32 0.33 0.79 0.43 After 60 cal BP 32 . 4 0.24 358.13 Dataset 2 Before 600 cal BP 27 0.43 1.09 -0.614 0.541 0.19 2.21 -0.954 0.344 520.37 1.03 1.3 0.199 After 60 cal BP 43 . 6 .22 387.44 All time period n x VisibilityF t p (2-tailed) x Elev. F t p (2-tailed) x Distance to lookout (m) F t p (2-tailed) Big 14 0.42 6.87 -0.57 0.57 0.16 6.15 -1.87 0.067 567.86 0.486 1.44 0.157 Small 37 0.46 0.25 339.46 600 cal BP-Contact Dataset 2 Big 17 0.43 0.85 -0.92 0.36 0.2 0.01 -0.67 0.51 417.65 1.8 0.34 0.74 Small 26 0.48 0.23 367.69 54 Overall, quantitative analysis comparing defensibility sites between time periods is not consistent with the prediction that people more actively defended their boundaries at 600 cal BP- Contact when resources at sites near productive resource areas were adequate to communities? needs. The dearth of evidence for trenches, fortifications, and other human-made modifications based on Angelbeck?s (2009) research also suggests that violent conflict may not have presented a threat to people on the San Juan Islands during the Late Holocene. Given the lack of change through time in defensive characteristics of sites, I also consider the sensitivity of the analyses to changes in territorial behavior. It is possible that the elevation and visibility indices used do not fully capture the ways that people defended sites against outsiders. Locations on the landscape that people considered acceptable for establishing communities may also have happened to have comparable defensive characteristics. Finally, it is also possible that the scale of active boundary defense against outside threats was greater than the village or even the San Juan Islands. If villages interconnected by kinship, marriage, and alliance defended themselves against outside threats from the north or the south, sites with higher defensive indices might exist at those boundaries. The Southern Channel Island Study Area and San Nicolas Island The Channel Islands of the Southern California Bight stretch from Point Conception to the border between the United States and Mexico. The northern group of islands in the Santa Barbara Channel region include Anacapa, Santa Cruz, Santa Rosa and San Miguel. They are associated with the Island Chumash band of the Chumash Nation. Channel Islands to the south include Santa Catalina, Santa Barbara, San Clemente and San Nicolas Island. These islands are within the traditional territory of the Tongva people (Johnston 1962). That the Tongva are Uto- 55 Aztecan speakers like the Western Shoshone to the east while the Chumash are Hokan speakers suggests a possible incursion of peoples from the east into the southern Channel Islands during the Late Holocene. Researchers use evidence from linguistic data (Kroeber 1976), bead styles (Howard and Raab 1993; Vellanoweth 2001a), and human osteology (Reinman and Townsend 1960) to determine when a possible incursion occurred but different evidence points towards different dates. The communities that lived on San Nicolas Island are referred to as the Nicole?o. The most isolated of the Channel Islands, San Nicolas Island is located 98 kilometers from the mainland and is approximately 5.6 kilometers long and 13 kilometers wide (Vellanoweth et al. 2002). The land is now owned and managed by the United State Navy. The modern coastal landscape is characterized by both rocky and sandy beaches and the interior is a large plateau covered in sand dunes. The archaeological record in both coastal and inland areas is dominated by shell-bearing sites, many of which are located on sand dunes (Afifi 2000). The southern Channel Islands culture history begins with short term seasonal occupations during the Terminal Pleistocene (13,000-10,000 cal BP) and Early Holocene (10,000 cal BP- 7500 cal BP). Based on assemblages from early sites, particularly the well-documented Eel Point site (SCLI-43) on San Clemente Island, people relied on marine resources from the first human settlement of the islands (Erlandson 1994; Erlandson et al. 2009, 2011; Rick et al. 2005a). Early sites on the northern Channel Islands indicate a marine-oriented adaptation, emphasis on rocky shore shellfish, and secondary reliance on fish and birds (Rick et al. 2001). Formal chipped stone tools are rare, but leaf-shaped bifaces, crescents, and small contracting stem points have been found. The first appearance of Olivella biplicata shell beads associated with exchange networks occur at several sites. Ornaments are rare (Cannon 2006; Rick et al. 2005a). 56 During the Middle Holocene (7500-4500 cal BP), more people moved to the Channel Islands and lived there year-round. They continued to focus on marine resources, particularly kelp bed environments, and incorporated a greater variety of fish and shellfish in their diet. An increase in the number of inland site and artifacts associated with plant resources such as mortars and pestles (Basgall 1987) suggests an increase in use of inland areas. Technology remains similar to the Early Holocene except for the appearance of side-notched and contracting stemmed points and increased diversity in stone, bone, and shell tools (Cannon 2006; Rick et al. 2005a). Throughout the Channel Islands, evidence of social complexity?increased sedentism, permanent houses, exchange, and social hierarchy?first appears during the Middle Holocene and intensifies during the late Holocene after approximately 3500 years ago (Glassow 2004; Rick et al. 2005a). Based on AMS dating of ornamental shell beads, long distance exchange networks between the southern Channel Islands and the mainland date back at least 5000 years (Vellanoweth 2001a) and there is an increase in variety of bead styles throughout the Middle Holocene (Rick et al. 2005a). By the Late Holocene (4500 cal BP-Contact), there is evidence on both the northern and the southern Channel Islands for increased complexity with larger and more sedentary villages, more exchange, sophisticated craftwork, evidence of social stratification, and complex ritual practices (e.g., Arnold 2001, 2004; Bartelle et al. 2010; Vellanoweth et al. 2008). Lambert (1993, 1994) proposes that analyses of human remains from the northern Channel Islands provide evidence of increased population density, circumscription, and inter-group violence. During this period, new tools appear including the single-piece fishhook, toggling harpoon, and small projectile points. On the northern Channel Islands, chert microblade technology flourished. Stone tool technology is mainly characterized by expedient stone and bone tools and small projectile 57 points. People focused on marine resources but pursued more deep water fish, decreased their reliance on shellfish, and increased the variety of shellfish that they gathered (Rick et al. 2005a). Southern Channel Islands Paleoenvironment Today, the southern Channel Islands climate is characterized as a ?Mediterranean? climate with mild temperatures during both the dry summers and wet winters. Dominant plant taxa include the giant coreopsis (Coreopsis gigantea), rattlesnake weed (Daucus pusillus), silver beach weed (Ambrosia chamissonis), and coyote brush (Baccharis pilularis). A large percentage of the annual precipitation comes in the form of fog; the island receives only 16.5 centimeters of rainfall annually (Vellanoweth et al. 2002:83). Freshwater is mainly located in the northwest part of the island where there are 12 perennial springs and seeps (Burnham et al. 1963; Vellanoweth et al. 2002:83). San Nicolas Island lacks richness and diversity in terrestrial flora and fauna but marine upwelling and mixing creates a productive marine environment (Rick et al. 2005a). To use the San Nicolas Island archaeological dataset to test a set of hypotheses on the relationship between territorial strategies and resource distribution requires a review of the paleoclimate record. It also requires a consideration of the impact of climate change on the abundance, predictability, and distribution of marine and terrestrial subsistence resources and freshwater. Marine Paleoclimate in the Channel Islands Below, I review paleoclimatic reconstructions of ocean upwelling, sea surface temperature, and terrestrial temperature and precipitation to establish a chronology of resource availability for San Nicolas Island. Sea surface temperature fluctuated significantly in the 58 Channel Islands during the Late Holocene, affecting the productivity of marine resources. Marine paleoclimate research has centered mainly on the Northern Channel Islands. A 25-year resolution oxygen isotope analysis of sediment cores from the Santa Barbara Basin indicate cool sea surface temperatures between approximately 3800-2900 BP, warm sea surface temperatures between 2900-1500 BP, coolest sea surface temperatures between 1500-500 BP, and a return to warm sea surface temperatures between 500-Contact (Kennett and Kennett 2000; Kennett 2005). Oxygen isotope data on a large sample of California mussel shells from the Northern Channel Islands also suggests colder sea surface temperature at 1400-400 cal BP and possibly at 250 cal BP (Kennett and Kennett 2000). These findings contradict previous work by Pisias (1978, 1979) who analyzed changes in radiolarian assemblages in a Santa Barbara Basin sediment core, potentially due to problems with the stratigraphic integrity of the core (Kennett and Kennett 2000:382). Future work may further clarify change over time in sea surface temperature in the Santa Barbara Basin, but for the purposes of this study I follow Kennet and Kennett?s interpretation. Sea surface temperature alone is not enough to determine marine productivity in the Southern California Bight. Researchers use a marine productivity index based on temperature differences between surface ocean water and deep ocean water. This index demonstrates degree of upwelling and other features of ocean circulation. Temperatures at the surface and in deeper water are determined using oxygen isotope data on planktonic foraminifera that live in those environments (Kennett 2005:67; Kennett and Kennet 2000:383-384; Kennett et al. 1995, 2007:352-353; Pak et al. 1997; Pak and Kennett 2002). Marine productivity index results demonstrate high productivity at 3800-2800 cal BP and 1000-500 cal BP. Despite cold 59 temperatures at 1500-1000 cal BP, upwelling did not increase in intensity (Kennett 2005; Kennett and Kennett 2000; Kennett and Ingram 2005). Regarding the terrestrial record, Kennett and Kennett (2000) suggest that during the Late Holocene, terrestrial drought correlates relatively well with times of cold sea surface temperature. Larson and Michaelson?s (1989) tree-ring data from the southern California Bight and Stine?s (1994) lake level data from the Sierra Nevadas indicate droughts at 1450-1150 and 970-700 cal BP. Gamble (2005:96) argues that these data are not convincing because Michaelson and Larson?s data has not been published and Stine?s data is from outside the region. However, pollen data from the Twin Rivers Marsh, San Nicolas Island indicates that between 1375-1250 and 920-420 BP, conditions were warmer and drier with cooler and wetter periods in between (Davis et al. 2003) (Figure 2.5). People also experienced substantial environmental variability at a decadal scale due to El Ni?o/Southern Oscillations (ENSO) events (Gamble 2005:98; see also Sandweiss et al. 1996 for analysis of ENSO events in Peru). Figure 2.13. Climate change expressed through precipitation, upwelling, and sea surface temperature for the Middle and Late Holocene, San Nicolas Island. 60 Marine Resources and Climate Change in the Channel Islands The paleoclimate record tells part of the story of human interaction with the environment, but it also important to consider more specifically how the resources that people depended on were affected by climate change. Regarding marine resources, people mainly focused on the intertidal and nearshore areas. Shellfish were a consistent staple of Channel Islanders? diet (Glassow 1980; Kennett 1998). Common taxa include five species of abalone (Haliotis spp.), California mussel (Mytilus californianus), wavy topshell (Astraea undosa), limpet (Lottia gigantean, Megathura cenulata, Fissurella volcano), chiton (Mopalia ciliate, Cryptochiton stelleri), norris top shell (Norrisia norrisi), and turbans (Tegula spp.) (Vellanoweth et al. 2002:84; Cannon 2006). People also caught fish from a variety of trophic levels including cabezon (Scorpaenichthys marmoratus), California sheephead (Semicosyphus pulcher), lingcod (Ophiodon lingatus), rockfish (Sebastes spp.), surfperch (Embiotocidae), swordfish and marlin (Xiphiida), and tunas (Scombridae) (Erlandson et al. 2009:715). They hunted sea mammals including sea otters (Enhydra lutris), California sea lions (Zalophus californianus), northern elephant seals (Mirounga angustirostris), and Pacific harbor seals (Phoca vitulina). The archaeological record also demonstrates use of marine birds such as cormorants (Phalacorcorax spp.) and gulls (Laridae) (Erlandson et al. 2009:715; Vellanoweth et al. 2002:84-85). The main impact of marine climate change on the Nicole?o?s marine food source would have been declines in fish and shellfish communities caused by damage to plant communities of the nearshore and intertidal zone during times of warmer sea surface temperature. Kelp (Macrocystis spp.) forests that grow along a submarine shelf extending four kilometers offshore (Cannon 2006:46; Engle 1994:18) decline when sea surface temperature is high. They are particularly vulnerable if nutrient availability is low, as is often the case in southern California 61 (Dayton et al. 1999; Edwards 2004; Tegner et al. 2001; Gerard 1997; Steneck et al. 2002). Long term sea surface temperature above 20 degrees celcius destroys kelp forests and reduces marine productivity. This is observed during ENSO events every several years (Kennett and Kennett 2000:381). Taxa that rely on the kelp forest ecosystems of southern California include sea urchins, abalone (Haliotis spp.), sheephead fish (Semicossyphus pulcher), spiny lobster (Panulirus interuptus), sea snail (Tegula spp.), halfmoon fish (Medialuna californiesis), greenfish (Girella nigricanus), and sea otter (Enhydra lutris), (Steneck et al. 2001:440). A recent study on abalones indicates that they rely on drift kelp to feed, and that warm temperatures also hinder their larval dispersal (Tegner et al. 2001). Abalones were a major source of food and tools for native communities and were also a food source for sea otters (Enhydra lutris). Terrestrial Resource and Climate Change in the Channel Islands Important terrestrial resources in the region include silver lupine (Lupinus albifrons), malva rosa (Lavatera assurgentiflora), Morman tea (Ephedra sp.), prickly pear cacti (Opuntia littoralis), and sage (Penstemom speciosus) (Thomas 1995; Yatsko and Raab 2009). At SNI-35, soil samples were analyzed for macrobotanical remains and included wild cucumber (Marah sp.), red maids (Calandrinia sp.), legume family (Fabaceae), manzanita (Arctostaphylos sp.), and blueberry/huckleberry (Vaccinium sp.) Pollen found on ground stone indicates processing of cheno-am, and sagebrush and sea grass may have been used to make containers (Thomas 1995:27). Fuel wood like silver lupine may have been depleted on the island by the Late Holocene (Vellanoweth 2001b:205). Domestic dog (Canis familiaris) and island fox (Urocyon littoralis) were the largest native land animals. Smaller animals such as the island night lizard 62 (Xantusia riversiana), side-blotched lizard (Uta stansburiana), white-footed deer mouse (Peromyscus maniculatus), and land snail (Micrarionta spp.) would not have been a significant food source due to their small size (Vellanoweth 1998; Vellanoweth et al. 2002). Regarding the terrestrial environment of San Nicolas Island, droughts would have been a significant problem because surface water is limited in the summer months (Rick et al. 2005a: 173). Yatsko (2000a, 2003) tests the hypothesis that human settlement on San Clemente Island was significantly affected by drought conditions during the Medieval Warm Period and finds that water would have been limited under the best conditions. People relied to a greater degree on plentiful marine food resources (Vellanoweth et al. 2002:84), but plants in particular were likely an important source of carbohydrates and nutrients. Settlement Pattern Data for San Nicolas Island Settlement pattern data for San Nicolas Island provides important insights into demographic data, distribution of communities relative to resources, and active defense of village boundaries. Much of the settlement pattern research on San Nicolas Island was conducted by California State University Los Angeles (CSULA) archaeologists in the 1990s who surveyed sites and excavated index units from a variety of environmental zones including the inland plateau, the slope, and the coastal plain (Martz and Rosenthal 2001; Martz 2002, 2008). The CSULA project recorded a total of 535 sites and obtained 68 radiocarbon dates from 41 sites. Site sampling has also been conducted as part of cultural resources management projects by Petra Resources and Statistical Research (Vellanoweth et al 2002) and through other archaeological investigations bringing the total number of dated sites on San Nicolas Island to 61 sites. Many sites have components dating to more than one period. Afifi?s (2000) research on 63 site location preference suggests that people may have preferred dune sites for habitation site locations and non-dune areas for special purpose sites. Village sites also tend to be closer to the ocean that special purpose sites, although preferences for locations of special purpose sites changed through time. Site destruction caused by modern activities presents a challenge to settlement pattern analysis on San Nicolas Island. The landscape has undergone dramatic changes in the last hundred years due to erosion, which has been accelerated by denudation of the landscape caused by sheep grazing from 1857 through the 1950s (Martz et al. 2005:67). Modern disturbance caused by construction of roads, buildings and other Navy activities has also impacted some sites. Lastly, much of the Terminal Pleistocene/Early Holocene record of the Southern California Coast has been inundated by ocean waters that lay 90 meters below present levels at 14,000 BP and rose to 30 meters below present by 10,000 BP. Relative sea level continued to rise gradually through the Holocene. At 8000 RCYBP the shoreline was 10 meters below its present level, at 6000 RCYBP it was 5 meters below present, and by 5000 RCYBP it reached modern levels (Inman 1983; Masters and Aiello 2007). Lack of commercial and private development has preserved the archaeological record on San Nicolas Island, as has the lack ground squirrels and other burrowing animals. The oldest site on San Nicolas Island is CA-SNI-339. It is located on the southeast coast and dates to nearly 8000 cal BP (Schwartz and Martz 1992). The number of sites on the island (and possibly population density) increases in the Middle Holocene after about 6500 cal BP, especially along the northwest coast of San Nicolas Island. A total of 22 sites date to this period. Site types include habitation sites, lithic sites, and shellfish processing sites (Martz 2005). During the Late Holocene, an increase in the number of sites to 42 and an increase in dispersion of sites 64 across the island parallels an increase in population and multifamily sedentary villages in the southern Channel Islands and Los Angeles Basin (Koerper and Drover 1983; Rick et al. 2005a; Vellanoweth et al. 2002). By plotting the sites on a map in 500 year intervals, it is possible to see finer-grained patterns in the distribution of site types (Figure 2.14-2.24). Dates are calibrated using CALIB 6.0 (Suiver et al. 2010). For marine shell, I applied a Marine09 calibration (Reimer et al. 2004) and ?R value of 225 ? 35 (Kennett et al. 1997; Prior et al. 1999; Rick et al. 2005b; Vellanoweth 2001a). Prior to 5000 cal BP until 3000 cal BP, most sites are residential and are concentrated on the northwest coast of the island. A few camps and lithic sites are also apparent at 5000-4000 cal BP. At 3000-2500 cal BP there is an increase in sites on the plateau. There is an increase in non- habitation sites at 2500-2000 cal BP. Site frequency appears to decrease at 2000-1500 cal BP but increases afterwards. A summed probability plot using 159 radiocarbon dates from 61 sites across the island indicates possible peaks in population just after 3000 cal BP and later at 1000- 500 cal BP (Figure 2.25). Ethnographic research on the proto-historic period suggests that the pre-contact population may have been between 600-1,200 individuals on San Nicolas Island. The total number of Tongva on the islands and the mainland is estimated at approximately 5000 (Bean and Smith 1978:539-540; McCawley 1996). 65 Figure 2.14. Site distribution on San Nicolas Island prior to 5000 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. Figure 2.15. Site distribution on San Nicolas Island at 5000-4500 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. 66 Figure 2.16. Site distribution on San Nicolas Island at 4500-4000 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. Figure 2.17. Site distribution on San Nicolas Island at 4000-3500 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. 67 Figure 2.18. Site distribution on San Nicolas Island at 3500-3000 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. Figure 2.19. Site distribution on San Nicolas Island at 3000-2500 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. 68 Figure 2.20. Site distribution on San Nicolas Island at 2500-2000 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. Figure 2.21. Site distribution on San Nicolas Island at 2000-1500 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. 69 Figure 2.22. Site distribution on San Nicolas Island at 1500-1000 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. Figure 2.23. Site distribution on San Nicolas Island at 1000-500 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. 70 Figure 2.24. Site distribution on San Nicolas Island after 500 cal BP. Numbers indicate site numbers recorded in the California Office of Historic Preservation site database. Figure 2.25. Summed probability plot for San Nicolas Island (n = 159 radiocarbon dates from 61 sites). Following the same protocol as Lepofsky et al. (2005), this plot incorporates only one date per site per 200-year interval to ensure that the better-dated sites do not bias the chronology. Territoriality Predictions for San Nicolas Island Based on paleoenvironmental shifts and settlement pattern data for San Nicolas Island, I propose a chronology for boundary defense and permeability. The arbitrary time periods used for both settlement pattern and lithic analysis are based on paleoenvironmental shifts and dating 71 for the lithic assemblage from Tule Creek Village, Mound B. At 5000-4000 cal BP, I predict minimal boundary defense due to a low population. Costs of inter-group interactions may outweigh the benefits, but some informal interaction should occur to preserve social networks for marriage partners and exchange of prestige objects. At 3000-1500 cal BP and 500 cal BP- Contact, a cool and productive terrestrial climate should provide adequate food and water resources but due to a higher population, boundary defense should be focused on inland areas near highly productive resource patches. Less productive marine areas should be shared by a number of villages. Boundaries in inland areas should be less permeable than those in marine areas due to higher costs and lower benefits of inter-group interactions. For the 1500-500 cal BP period, adequate marine resources and impoverished terrestrial resource should encourage more boundary defense in highly productive marine areas. Due to lower costs and higher benefits of exchange and reciprocal access , informal inter-group interactions in marine areas should increase (Table 2.7). Table 2.7. Prediction for boundary defense and permeability for San Nicolas Island based on paleoenvironmental and settlement pattern shifts during the Late Holocene. Defensive Sites on San Nicolas Island No sites with trenches, ditches, walls, or other evidence of defensive earthworks have been reported on San Nicolas Island. Lambert (1997) notes that on the narrow coastal plains of the Santa Barbara Channel, defensive sites are rare and poorly preserved. Kennett (2005:180- 183) suggests that in the Late Holocene/Early Contact Period, the even distribution of San Nicolas Island Population Marine Terrestrial Water Land Water Land 500 cal BP-Contact High - + Low High High, informal Low, formal 1500-500 cal BP High + - High Low Low, formal High, informal 3000-1500 cal BP High - + Low High High, informal Low, formal 5000-4000 cal BP Low - + Low Low Moderate, informal Moderate, informal Resources Defense Permeability 72 communities on the Northern Channel Islands and lack of intervisibility between them is indicative of the partitioning of island areas into small territories owned by lineal descent groups. These groups were connected through marriage but in competition over resources. He suggests that the large village sites on Santa Rosa Island and San Miguel Island from this period are located near perennial water sources and strategically positioned to provide a view over the ocean and coast, suggesting defense against potential threats from the mainland or nearby villages. He also proposes that some villages had secondary habitation sites in the interiors to maintain access to interior plant resources and keep watch over the interior. Cemeteries may have served as territorial markers. To test my predictions for boundary defense, I calculate Martindale and Supernant?s (2009) measure for elevation, the ARCTAN of the elevation difference from the inside to the outside of the size divided by the site radius. I also calculate distance from site to a lookout point (greater than 200 degrees of visibility over water) (Appendix B). Site location is classified based on plateau, slope (steep escarpment slope from the plateau to the coastal plain), and coastal plain. As for the San Juan Islands case study I use two datasets to test the predictions below, one where sites are assigned to the earliest time period to which they date and one where sites are assigned to all time periods to which they date. Based on the predictions for territorial behavior discussed above, if boundary defense is low at 5000-4000 cal BP, I predict significantly lower values for elevation and higher values for distance to lookout during this time period. In other words, site location should be chosen without regard to defensive qualities. The 3000-1500 cal BP and 500 cal BP periods should have higher values for inland sites and lower values for marine sites where boundary defense is predicted to be low. The reverse should be true for the 1500-500 cal BP period. Although 73 boundary defense should only be higher at certain sites near productive resource areas, the higher defensive values for those sites should increase the overall mean elevation and distance to lookout values. The types of sites during these time periods should also differ in location with more residential sites located near important and productive resource patches. Results of an ANOVA for both datasets indicate no significant differences between time periods in mean elevation or distance to lookout for all sites. This is the case if all sites are considered, if only marine sites are considered, or if only terrestrial sites are considered (Table 2.8, 2.9). Results of a ?2 test reveals no statistically significant difference in site location between time periods whether considering all sites (?2 = 4.22, 6 df, p = 0.646) or habitation sites only (?2 = 2.35, 6 df, Fisher?s exact p = 0.88) (Table 2.10, 2.11). Table 2.8. Results of an ANOVA comparing mean elevation and distance to lookout for >5000-3000 cal BP, 1500-500 cal BP, 3000-1500 cal BP, and 500-Contact. Sites are assigned to the first period to which they date. Sites Time Per. (cal BP) n x Elevation F Sig. x Distance to lookout (km) F Sig. All > 5000-3000 26 0.19 0.046 0.99 1.22 0.69 0.56 3000-1500 12 0.17 1.15 1500-500 14 0.18 1.07 500-0 8 0.17 0.75 Coastal > 5000-3000 9 0.25 1.52 0.23 0.64 0.45 0.72 3000-1500 6 0.11 0.86 1500-500 6 0.17 0.63 500-0 6 0.2 0.55 Plateau/Slope > 5000-3000 17 0.15 0.9 0.46 1.52 0.31 0.82 3000-1500 6 0.24 1.45 1500-500 5 0.22 1.88 500-0 2 0.12 1.35 74 Table 2.9. Results of an ANOVA comparing mean elevation and distance to lookout for >5000-3000 cal BP, 1500-500 cal BP, 3000-1500 cal BP, and 500 cal BP-Contact. Sites are assigned to all time periods to which they date. Table 2.10. Site location counts for each time period. All sites are assigned to the first time period to which they date. Table 2.11. Site location counts for each time period. All sites assigned to all time periods to which they date. These results are not consistent with shifts in defensive site characteristics on San Nicolas Island that mirror environmental shifts hypothesized to be associated with changes in active boundary defense. I also cannot rule out the possibility that the lack of statistically significant differences in measures of defensiveness between time periods is the small sample size. It is possible that site locations were chosen for reasons other than their defensive capabilities. Sites Time Per. (cal BP) n x Elevation F Sig. x Distance to lookout (km) F Sig. All > 5000-3000 26 0.19 0.227 0.877 1.22 0.182 0.91 3000-1500 21 0.17 1.34 1500-500 23 0.18 1.2 500-0 16 0.21 1.13 Coastal > 5000-3000 9 0.25 0.99 0.409 0.64 0.317 0.81 3000-1500 8 0.16 0.76 1500-500 12 0.17 0.59 500-0 8 0.21 0.54 Plateau/Slope > 5000-3000 17 0.16 0.381 0.767 1.52 0.29 0.83 3000-1500 13 0.19 1.7 1500-500 11 0.19 1.85 500-0 8 0.21 1.71 >5000-3000 3000-1500 1500-500 500-Contact Coastal Plain 9 6 9 6 Slope 12 3 1 1 Plateau 5 3 4 1 Time Period (cal BP) >5000-3000 3000-1500 1500-500 500-Contact Coastal Plain 9 8 12 8 Slope 12 7 5 4 Plateau 5 6 6 4 Time Period (cal BP) 75 Alternatively, the lack of change through time in defensive characteristics of sites mayindicate a problem with elevation and distance to lookout measures in investigating defense characteristics of sites. Perhaps boundary defense was achieved using built defensive features that did not preserve well over time or were destroyed by the post-depositional processes that have greatly altered the San Nicolas Island landscape. It is also possible that locations on the landscape that people considered acceptable for establishing communities due to protection from the elements and access to food and water also had similar defensive characteristics. Conclusions Many Pacific Coast archaeology studies attempt to link cultural changes and climate changes (e.g., Arnold 1992; Fladmark 1975, Glassow 1996; Prentiss and Chatters 2003; Johnson 2004; Kennett 1998; Kennett and Kennett 2000; Lambert 1997; Lepofsky et al. 2005; Mitchell 1971; Morgan 2009; Raab et al. 1995; Raab and Larson 1997). The challenge of these studies is to provide a strong chronology for environmental change, determine how specific aspects of that environmental change affected people, and to determine if a chronology of cultural change corresponds with the environmental changes. In this chapter, I have established a chronology of environmental change for both study areas, discussed the specific impact of environmental changes on subsistence resources, and used settlement pattern to determine when precontact population on the San Juan Islands and San Nicolas might have been high enough that subsistence resource supply was at or near the level of demand. To begin to test the territoriality hypothesis discussed in Chapter 1, I used data on the defensive characteristics of archaeological sites to determine if increases in active defense correspond to time periods when resources were abundant. For both study areas, visibility, elevation, and distance to lookout did not show statistically significant differences between time 76 periods associated with shifts in resource availability. This may indicate that people in the San Juan Islands and San Nicolas Island did not defend boundaries around productive resource areas. It is also possible that their territorial behavior did not emphasize aggressive attacks on one another at the level of the village, but rather groups of villages joined together to face a common threat. This supports the alternative hypothesis for this study that is based on descriptions of the complex and interconnected communities from the ethnographic record in both study areas. Kin and marriage relationships may have rendered boundaries permeable to certain people or families regardless of the marine or terrestrial productivity. Alternatively, territorial behavior may not have centered on violent conflict but may instead have emphasized resource procurement. In the chapters that follow, I investigate whether toolstone procurement patterns indicate shifts in resource access that correspond with changes in resource availability during the Late Holocene. 77 Chapter 3: Stratigraphic Context and Dating of Lithic Assemblages To investigate changes in territorial behavior on the Pacific Coast, I test predictions about lithic procurement patterns by analyzing toolstone availability and toolstone use at habitation sites on the San Juan Islands and southern Channel Islands. These sites include the Watmough Bay site (45-SJ-280) on Lopez Island, San Juan Islands, Washington and Tule Creek Village (CA-SNI-25) Mound B and CA-SNI-106 on San Nicolas Island, Channel Islands, California. Since predictions for change over time in boundary defense and permeability are tied to changes in resource abundance, it is necessary to compare precisely dated lithic assemblages from time periods associated with different climate regimes. In this chapter, I discuss stratigraphy and chronology at Watmough Bay, Tule Creek Village, and CA-SNI-106 to establish the temporal and spatial analytic units used in this study. Excavations at Watmough Bay The Watmough Bay site is an approximately 9,500 meter 2 shell midden located on southern Lopez Island on a northeast-facing bay. It is on land owned by the Washington State Bureau of Land Management and the San Juan County Land Bank (Figure 3.1). The site is on a beach bar between a sandy beach and an extensive freshwater marsh. The first investigations of the site were conducted by David Munsell in 1968 as a University of Washington field school. In 2004, Julie Stein and Laura Phillips directed a small excavation as part of a site stabilization project (Figure 3.2). Artifacts and sediment samples from Watmough Bay are housed at the Burke Museum , Seattle. 78 Figure 3.1. Map of the San Juan Islands showing the location of the Watmough Bay site. Figure 3.2. View from above to the east of Watmough Bay showing the 1968 and 2004 excavation areas and surrounding landforms. 79 The 1968 excavation at Watmough Bay covered a 36 meter 2 area. Most units were 2 x 1 meters although some were expanded to investigate features. Three 1 x 1 meter balks between the units were excavated at the end of the project. The site was excavated by trowel and screened through ?-inch mesh. Students usually excavated in 20-centimeter (cm) arbitrary levels, but in some cases, features were screened separately and assigned separate bag numbers (Bovy 2005:24-25; Bovy et al. 2007; Field Notebooks on file at the Burke Museum, Accn. 1996-121). In August of 2004, Julie Stein and Laura Phillips directed the excavation of two 1 x 1 meter units on the east side of the Watmough Bay beach in association with a site stabilization project conducted by the Bureau of Land Management to minimize erosion of the shell midden (Figure 3.3). The project was a collaborative effort between the Burke Museum, the Bureau of Land Management, and the Samish Indian Nation. Based on methods used by Stein at the English Camp and Burton Acres shell midden excavations (Parr et al. 2002, 2011), the site was excavated by trowel in 10-cm arbitrary layers within larger natural layers. If excavators encountered <10-cm lenses of sediment that were different in color and texture than surrounding sediment, these lenses were excavated and screened separately. All sediment was screened through 1-inch, ?-inch, ?-inch and ?-inch inch mesh on site. All cultural material was saved from 1-inch and ?-inch screens. Initially only material from every fourth bucket was saved from the ?-inch screen and from every eigth bucket for the ?-inch screen. For layers below 2F in EXU1 and all of EXU2, 100% of the material recovered in the ? and ?- inch screen was saved (Bovy et al. 2007). 80 Figure 3.3 Plan view map of Watmough Bay showing the 1968 excavation. Modified from Bovy 2005 Figure 2.4). The Kubota trench was an area excavated as part of the bank stabilization project. Stratigraphy and Dating at Watmough Bay The typical stratigraphic profile at Wamough Bay is characterized by an upper plow zone approximately 30-50 cm thick composed of gray brown, brown, or dark brown sediment with small amounts of fragmented shell and historic artifacts. During the 1968 excavation, the plow zone was designated Stratum I and typically removed as one level. The plow zone is underlain 81 by a 50-100 cm thick shell midden layer (Stratum II and III) with dark silty sediment, abundant artifacts, lenses of charcoal, ash, and sea urchin spines. Below the shell midden layer lies a dark layer (Stratum IV) that contains little to no shell but relatively abundant cultural material, and a layer of natural beach sands and gravels, designated Stratum V (D. Croes, G. Jenkins, A. Richardson, R. Schalk field notebooks; Table 3.1; Figure 3.4). In other Northwest Coast shell middens, such as English Camp on San Juan Islands, geoarchaeological analysis indicates that post-depositional leaching of carbonate caused by inundation of sea water decreases the amount of shell present in the lowest strata of shell middens (Stein 1992). This process almost certainly accounts for the appearance of Stratum IV at Watmough Bay. Dating results indicate that the age of Stratum IV varies across the units (Figure 3.4, 3.5). Excavators in 1968 and in 2004 observe spatial variation in stratigraphy across the site. Bovy (2005) notes that based on her review of field notebooks, the closest unit to the bay, 9N3W, contains more sparsely distributed shell than units to the southwest. It also contains layers of sand with abundant fish and bird bone below 136 cm below the surface (cmbs) and gravel layers. At EXU1 and EXU2 (2004 excavation), the stratigraphy follows the same basic description as units to the northwest; however, the plow zone is less than 10 cm thick and historic artifacts and modern objects were not encountered below 10 cm below the surface (cmbs) (Figure 3.6). The shell midden layer (Stratum II and III) was approximately 40 cm thick, which is thinner than the 1968 units. Additionally, in EXU1 at approximately 80-90 cmbs, excavators discovered a stone slab hearth feature (Figure 3.7). 82 Table 3.1. Basic field description of the stratigraphy at Watmough Bay based on descriptions in G. Jenkins? notebook. Stratum Color Texture Compaction Cultural Materials I Brown Silty soil Moderate II Gray- brown Fine silty soil Many shell fragments III Brown Sandy clay, gritty with gravel Compact Fine shell fragments, FMR IV Black Sandy silt with pebbles, cobbles Compact Abundant charcoal, FCR, shell is scarce V Brown Beach sand and gravel None Figure 3.4 A typical stratigraphic profile at 0N18 W, modified from a profile drawing by Richardson and Jenkins. Figure 3.5. Stratigraphic profile at 0N9 W, modified from a profile drawing by Jeffrey and Kaschko. 83 Figure 3.6. Profile drawing for west wall EXU1 based on field profile drawn by J.K. Stein and S. Johnson, August 2004. 84 Figure 3.7. Hearth feature at EXU1, Watmough Bay. Both stratigraphy and dating indicate that beneath the plow zone, the site has been minimally disturbed by modern activities or erosion. A total of 51 radiocarbon dates have been obtained for the site (Bovy 2005; Bovy et al. 2007; Daniels 2009; Deo et al. 2004; Stein et al. 2003), including eight for this project (Table 3.2, 3.3). Dating results indicate that the site was occupied from as early as 3000 cal BP until the historic period. A majority of the dates occur at 1600-1200 cal BP, but there are also clusters of dates at 2800-2400 cal BP and 1000-400 cal BP indicating periods of more permanent occupation. In all units but 1N/9W and EXU1 there are no 85 chronological reversals within the stratigraphic profile (Figure 3.6). In 1N9W, dates at 80-100 and 120-140 indicate a chronological reversal that may be due to erosion or post-depositional disturbance (Figure 3.8). In EXU1, dates at 10-20 cmbs and 22 cmbs are younger than an additional date at 10-20 cmbs. This may be due to a calibration issue since the older date is on shell and the younger dates are on wood. It is also possible that the younger wood date at 22 cmbs is attributable to a contaminated wood sample, or a sample that is from the inner part of a tree. Chronology at the Watmough Bay site is relatively straightforward; however, there are some potential dating issues that I consider in creating analytic units. For example, in Unit 9N/3W, a shell sample (Acme mitra) and bird bone sample (Uria cf. aalge) from 120-140 cmbs date to 1250-1370 cal BP and 3580-3830 cal BP, respectively. The bird bone date is much older than dates found in strata below. Although it is possible that the >3000 cal BP date resulted from post-depositional disturbance, it seems more likely that the bone date is in error due to processing or calibration problems. In Unit EXU1, a date on a hardwood branch at 10-20 cmbs at 520-640 cal BP does not match a shell dated to 1300-1470 BP cal. from the same layer. If one date is from the top of the layer and the other is from the bottom, it is possible that the discrepancy does not indicate disturbance but rather very rapid accumulation. It is also possible that the upper layers have been disturbed by modern activities or wave action. Finally, in excavation units in the trench at 0N, >2000 cal BP deposits in lower levels of 0N9W indicate a potential older deposit, but there are too few dates in adjacent units to determine if older deposits are present at that depth across the site. 86 Table 3.2. Radiocarbon dates for the Watmough Bay site, 1968 excavation. Sample # Unit Depth (cm)a 14C Age BP Cal. Age Range, 2 Sigma BP b Material Reported OS89919 0N0E 0-40 2150 ? 30 1240-1370 Polyplacophora Taylor 2012 OS89920 0N0E 160-180 2280 ? 25 1350-1510 Balanus spp. Taylor 2012 OS45737 0N3W 40-60 2360 ? 30 1410-1600 P. staminea Bovy 2005 OS88418 Blk.C 120-130 2290 ? 25 1360-1510 Bivalve Taylor 2012 OS89921 0N6W 40-60 2340 ? 30 1390-1575 Polyplacophora Taylor 2012 OS84405 0N9W 30-50 1930 ? 25 1000-1170 P. staminea Taylor 2012 Beta119323 0N9W 169 2110 ? 50 1949-2180, 2240- 2303 Charcoal Stein et al. 2005 Beta119320 1N/9W 60 120 ? 50 0-0, 7-150, 170-280 Charcoal Stein et al. 2005 OS42278 1N9W 60-80 2330 ? 30 1380-1560 Mollusca Bovy 2005 OS42279 1N9W 80-100 2450 ? 35 1520-1710 P. staminea Bovy 2005 CAMS56453 1N9W 120-140 2240 ? 50 1290-1500 Stronglyocentrotus sp. Deo et al. 2004 Beta119321 1N9W 140 2200 ? 50 2065-2083, 2110- 2340 Pseudotsuga menziesii Deo et al. 2004 Beta119322 1N9W 153 2090 ? 40 1949-2152, 2280- 2290 Charcoal Stein et al. 2005 OS42360 Blk.A 60-80 2340 ? 30 1390-1580 P. staminea Bovy 2005 OS42361 Blk.A 60-80 2320 ? 30 1380-1550 Polyplacophora Bovy 2005 OS42362 Blk.A 60-80 2310 ? 30 1370-1540 Tresus sp. Bovy 2005 OS42363 Blk.A 100-120 2360 ? 30 1410-1600 Polyplacophora Bovy 2005 OS45738 0N18W 40-60 2340 ? 30 1390-1580 P. staminea Bovy 2005 OS42364 0N18W 60-80 2330 ? 30 1380-1560 Polyplacophora Bovy 2005 OS42365 0N18W 80-100 2380 ? 35 1420-1650 P. staminea Bovy 2005 OS42366 0N18W 100-120 2400 ? 30 1480-1680 Bivalve Bovy 2005 OS42735 0N18W 100-120 2300 ? 30 1360-1560 Bivalve Bovy 2005 OS42736 0N18W 100-120 2280 ? 25 1350-1510 Bivalve Bovy 2005 Beta119317 0N24W 60-80 1350 ? 100 1010-1030, 1050- 1420, 1470-1510 Charcoal Stein et al. 2005 Beta119318 0N24W 80-100 1560 ? 50 1350-1550 P. menziesii branch Deo et al. 2004 CAMS56454 0N24W 80-100 2330 ? 50 1350-1610 Bivalve Deo et al. 2004 CAMS56455 0N24W 100-120 2170?50 1220-1440 Balanus sp. Deo et al. 2004 Beta119319 0N24W 100-120 1580?50 1350-1560 Conifer branch Deo et al. 2004 OS89922 3S0E 0-40 2140 ? 35 1220 - 1370 Nucella spp. Taylor 2012 OS84404 3S0E 40-60 2090?30 1170-1300 Bivalve Taylor 2012 OS89923 3S0E 80-100 2200 ? 30 1270 - 1410 P. staminea Taylor 2012 Beta119316 12S0E 61 2360 ? 50 2190-2190, 2210- 2230, 2310-2520, 2530-2540, 2590- 2620, 2640-2700 Thuja/Tsuga branch Deo et al. 2004 CAMS56451 12S0E 60-80 3150 ? 40 2340-2610, 2630- 2650 P. staminea Deo et al. 2004 OS45739 9N3W 40-60 2130 ? 30 1220-1350 Mytilus californianus Bovy 2005 OS42367 9N3W 120-140 2160 ? 30 1250-1370 Acme mitra Bovy 2005 Beta193785 9N3W 120-140 3430 ? 40 3580-3740, 3740- 3780, 3790-3830 Uria cf. aalge humerus Stein et al. 2005 Beta119324 9N3W 165 2640 ? 40 2720-2840 Conifer branch Deo et al. 2004 CAMS56452 9N3W 160-180 3320 ? 50 2520-2810 Gastropoda Deo et al. 2004 aDepths are cm below surface. bThe marine reservoir correction follows values established by Deo et al. (2004) and Daniels (2009). At 0-600 cal BP and 1000-3000 cal BP, ?R = 400 years. At 600-1000 cal BP ?R = 0 years. 87 Table 3.3 Radiocarbon dates for the Watmough Bay site, 2004 excavation. Sample # Unit Depth (cm)a 14C Age BP Cal. Age Range, 2 Sigma BP b Material Reported OS66822 EXU1 10-20 550 ? 30 520-560, 590-640 Hardwood branch Daniels 2009 OS68460 EXU1 10-20 1840 ? 25 1300-1470 P. staminea Daniels 2009 Beta203751 EXU1 22 970 ? 40 790-960 P.menziesii branch Bovy et al. 2006 OS66820 EXU1 41-50 1110 ? 25 960-1060 Softwood twig Daniels 2009 Beta203752 EXU1 78 1530 ? 40 1340-1520 Unidentified conifer Bovy et al. 2006 Beta203753 EXU1 103 2550 ? 40 2490-2640, 2670- 2750 Unidentified conifer Bovy et al. 2006 Beta203756 EXU1 108 2640 ? 40 2720-2840 Charred plant Bovy et al. 2006 Beta203754 EXU1 116 2700 ? 40 2750-2870 Populus/Salix sp. Bovy et al. 2006 Beta203756 EXU1 119 2540 ? 40 2490-2640, 2650- 2750 Alnus sp. small branch Bovy et al. 2006 Beta203757 EXU1 130 2690 ? 40 2750-2860 Conifer Bovy et al. 2006 OS68461 EXU2 10-20 1160 ? 25 320-470 P. staminea Daniels 2009 OS66823 EXU2 10-20 665 ? 30 560-600, 630-670 Alnus sp. or Betula sp. Daniels 2009 OS66847 EXU2 50-60 1080 ? 30 930-1020, 1020- 1060 Hardwood cf. Sambucus Daniels 2009 aDepths reported as cm below surface. bThe marine reservoir correction used for the shell follows values established by Deo et al. (2004) and refined by Daniels (2009). At 0-600 cal BP and 1000-3000 cal BP, the regional correction value (?R) was 400 years. At 600-1000 cal BP ?R was 0 years. 88 Figure 3.8. Watmough Bay dates (2-sigma calibrated). Temporal Analytic Units for Watmough Bay After dating and analyzing the stratigraphy at Watmough Bay, I divided the lithic assemblage into four temporal analytic units based on date clustering the timing of environmental shifts (Figure 3.9). The earliest period, 3500-2500 cal BP, corresponds to a period 89 of cool and wet terrestrial conditions and productive marine conditions. The next period, 2500- 1600 cal BP, is the beginning of the Fraser Valley Fire period when terrestrial conditions were warm and dry and marine conditions were productive. The 1600-1000 cal BP period at the end of the Fraser Valley Fire Period saw a continued warm dry terrestrial environment and a decline in marine productivity. Most artifacts from the Watmough Bay site and the English Camp site date to this period. The latest period at 600 cal BP-Contact corresponds to increased terrestrial moisture and increased in marine productivity. Most strata/level are assigned to a time period based on a radiocarbon date for that level. In some cases, such as Unit 0N15W, units had no radiocarbon dates but I assigned time periods based on dating from two adjacent units or an adjacent unit and a level above or below. In other cases, such as 33S0E where no nearby units were dated, it was not possible to assign the artifacts to a time period. Artifacts that were not assigned to a time period were not included in some of the lithic analyses depending on the research question. Some of the lithic analyses require a comparison of two spatial areas with the sites. For Watmough Bay, I chose two areas that date to the 1600-1000 cal BP time period due to the larger sample size. I choose units with larger sample sizes that might potentially correspond to different household within the village because they are far away from one another. These include 0N24W and 0N18W on the northwest side of the site and 0N0E, 1.5N0E, 0N3W and Balk C on the southeast side of the site (Figure 3.10). 90 Figure 3.9. Dates and temporal analytic units at Watmough Bay, not to scale. 91 Figure 3.10. Spatial analytic units at Watmough Bay. Unit 1: 0N24W, 40-160 cmbs. Unit 2: 0N0E 0-200 cmbs, 1.5N0E 40-160 cmbs, 0N3W 60-180 cmbs, Balk C 0-160. 92 The San Nicolas Island Sites To examine change over time in lithic procurement on San Nicolas Island, I created temporal analytic units for the lithic assemblages for Tule Creek Village (CA-SNI-25) Mound B and SNI-106. Temporal units were based on the stratigraphy and chronology of these sites and correspond to time periods associated with climate shifts. Lithic analysis focuses on Mound B, but also includes an assemblage from CA-SNI-106, an inland habitation site located 2.5 km southwest of Mound B (Figure 3.11). Figure 3.11. Map of San Nicolas Island showing the locations of Tule Creek Village and CA-SNI-106. White shaded areas indicate recorded site locations. The map was generated using ARCGIS and LIDAR data provided by Steve Schwartz and Richard Guttenberg. 93 Excavations at Tule Creek Village, Mound B Tule Creek Village is an approximately 145,000 meter 2 shell-bearing site on the upper plateau of the northern coast of San Nicolas Island, located within a kilometer of the shoreline (Martz 2008). The site overlooks Corral Harbor, a protected inlet where marine resources and toolstone would are abundant. Inhabited as early as 5000 cal BP through the historic period, Tule Creek Village includes several discrete areas of dense artifacts and features. This study focuses on the area known as Mound B, an approximately 100 meter 2 area considered primarily a residential locality due to the diversity and abundance of artifact types (Figure 3.11, 3.12). Site information was first recorded by Reinman and Associates in 1984 and updated by California State University Los Angeles (CSULA) researchers in 1997 (Martz 2002, 2008; Reinman and Lauter 1984). Site testing was conducted by CSULA and Humboldt State University students under Dr. Patricia Martz in 1996 (Martz 2008). The most recent excavations at Mound B began in 2001 under the direction of Ren? Vellanoweth. Excavations were conducted during Humboldt State University field school through 2007 and CSULA field schools through 2009. All 1x1 meter units were excavated in arbitrary 10-cm levels within strata using trowels, brushes, and scoops. Excavated sediment and cultural materials were collected in 10-liter graduated buckets and sediment was screened through ?-inch mesh. Larger artifacts, shells, bones and stones were collected in the field. Material that remained in the screen and bulk samples were collected for further analysis in the lab. Cultural materials are housed at both at the CSULA laboratory and at the San Nicolas Island Archaeology Laboratory on San Nicolas Island. Lithic artifacts included in this study were analyzed at the CSULA laboratory and at the University of Washington Department of Anthropology. 94 Figure 3.12. View to the east of Tule Creek Village with Mound B excavation in the foreground. Photo courtesy of the Humboldt State University San Nicolas Island Field School. Stratigraphy and Dating at Mound B Excavations at Tule Creek Village (CA-SNI-25) exposed approximately 100 meter 2 . Geoarchaeological investigations focused on a trench that bisected the mound from northwest to southeast and included units 12, 13, 32, 43, and 48. Lithic analysis was conducted on artifacts from only those units where stratigraphy and dating were well understood, including units 11, 13, 32, 43, 48, 52, and 58 (Figure 3.13). The stratigraphy and dating of these units is discussed below. The basic stratigraphy at Mound B was characterized by a surface deposit of modern dune sand (Stratum I) underlain by a darker colored sandy layer with abundant artifacts (Stratum 95 II). In several units, excavators encountered a lighter colored compact surface that alternated with Stratum II. During the 2001-2008 field season, this stratum was designated Stratum IIA. During the 2009 field season, alternating light and compact layers and dark layers were given additional numerical designations. Stratum II was the first dark layer encountered and Stratum IIA was the first light layer encountered. Any subsequent light layers were numbered IIA1, IIA2, etc. Any subsequent dark layers were numbered IIB1, IIB2, etc. Darker and lighter cultural strata were underlain by a transitional sandy layer with fewer artifacts (Stratum IIB), and a non-cultural sandy layer (Stratum III) (Table 3.4). Table 3.4. Basic field descriptions of strata at Mound B. Stratum Color Texture Compaction Cultural Materials I 10YR 3/2 Very dark grayish brown Sand Very loose Moderate II 10 YR 3/1 Very dark gray Silty sand Loose Abundant IIA 10YR 4/2 Dark yellowish brown Silty sand Compact Abundant IIB 10YR 3/2 Very dark grayish brown Silty sand Loose Scarce III 10YR 5/3 Brown Fine sand Very loose Scarce or None 96 Figure 3.13. Map of CA-SNI-25 Mound B showing excavation units included in dating, geoarchaeological, and lithic analyses for this study. Modified from Cannon 2006 Figure 7.3). Lithic artifacts analyzed are from Units 11, 13, 32, 43, 48, 52 and 58. Three sets of radiocarbon dates were submitted for the Mound B locality of Tule Creek Village, one reported by Cannon (2006) as part of her MA thesis research at Humboldt State University, a second submitted by Vellanoweth in 2008, and a third submitted as part of this dissertation research. All radiocarbon dating was conducted at the National Ocean Sciences 97 Accelerator Mass Spectometry (NOSAMS) Facility at Woods Hole, Massachusetts. Dates were calibrating using CALIB 6.0 (Stuiver et al. 2010). All but two of the radiocarbon samples were marine shell, and corrections for both global (Reimer et al. 2004) and regional variation in radiocarbon age of sea surface water were applied (Kennett et al. 1997; Prior et al. 1999; Rick et al. 2005b; Vellanoweth 2001). Radiocarbon dating suggests that Mound B was inhabited from as early as 5000 cal BP to as late as 250 cal BP. The occupation may have been continuous, but clusters of dates at 5000- 4400 cal BP, 2000-1500 cal BP, and 600-250 cal BP indicate potential for more intensive occupations separated by periods of use by smaller communities or perhaps abandonment (Table 3.5, Figure 3.14). In the four excavation units where Stratum IIA is absent, post-depositional disturbance is minimal. In each, the lower layers of Stratum I and the upper layers of Stratum II represent a ca. 250-600 cal BP occupation. Although no noticeable differences in color and texture of sediment were noted between upper and lower layers of Stratum II in the field, they accumulated during two different time periods. The middle and lower layers represent a ca. 1500-2000 cal BP occupation. In some units, the lowest levels of Stratum II and upper layers of Stratum IIB represent occupations older than 4000 cal BP (Table 3.5, Figure 3.13). In Unit 48, dates on features indicate that the inhabitants of Mound B sometimes altered the chronology of the stratigraphic profile. Here, a shell sample from a pit feature at 120 cm below datum (cmbd) dates to 280-460 cal BP, much later than the deposit surrounding it but consistent with use of the pit during the terminal occupation of the site (Figure 3.15). 98 Table 3.5. Radiocarbon dates for Tule Creek Village, Mound B Sample # Unit/Stratum/ Feature Depth (cm) a Uncorrected 14 C Age BP Cal. Age Range, 2 Sigma BP b Material Reported OS-54354 11 SII/L3 46 880 ? 30 150-160,190-210, 220-430 H. cracherodii Cannon 2006 OS-69733 12 SII/L3 40 2160 ? 20 1370-1590 H. cracherodii CSULA 2008 OS-69745 13 SIIA/L2 52 905 ? 20 260-430 H. cracherodii CSULA 2008 0S-69739 13 SII/L4 64 2220 ? 30 1420-1680 H. cracherodii CSULA 2008 OS-69744 13 SII/L4 87.4 1010 ? 20 320-490 H. cracherodii CSULA 2008 OS-84406 32 SII/L1 36 3270 ? 30 2720-2930 H. cracherodii Taylor 2012 OS-84407 32 SII/L5 47 860 ? 35 140-410 H. cracherodii Taylor 2012 OS-69740 32 SII/L7 84.5 2270 ? 35 1490-1760 M. californianus CSULA 2008 OS-69747 32 SIIB/L1 109 3220 ? 25 2690-2860 H. cracherodii CSULA 2008 OS-84409 43 SII/L1 53 875 ? 30 150-160, 190-420 Tegula spp. Taylor 2012 OS-69743 43 SIIB/L2 69 915 ? 20 260-430 H. cracherodii CSULA 2008 OS-69742 43 SIIB/L2 86 4730 ? 30 4560-4820 H. rufescens CSULA 2008 OS-69746 43 Pit 1 93 4720 ? 25 4550-4810 H. cracherodii CSULA 2008 OS-55025 47 Fox Feature 77 1540 ? 35 1360-1520 U. littoralis Cannon 2006 OS-54356 47 Fox Feature 79 2330 ? 30 2210-2220, 2310- 2370, 2390-2400, 2410-2460 U. littoralis Cannon 2006 OS-84410 48 SI/L2 57 1040 ? 30 330-520 H. cracherodii Taylor 2012 OS-69737 48 SII/L1 70 1630 ? 30 830-1060 H. cracherodii CSULA 2008 OS-69738 48 SII/L4 Feature 1 120 950 ? 20 280-460 H. cracherodii CSULA 2008 OS-54413 52 Feature 63 225 ? 35 520-460c H. cracherodii Cannon 2006 OS-84408 52 SII/L1 70 2380 ? 25 1620-1860 H. cracherodii Taylor 2012 OS-54357 55 Pit 1 85 4890 ? 35 4800-5030 H. rufescens Cannon 2006 OS-54358 57 Hearth 94 4800 ? 30 4620-4900 H. cracherodii Cannon 2006 OS-84411 58 SII/L1 51 935 ? 25 270-450 Haliotis spp. Taylor 2012 OS-54359 58 Bird Feature 74 2450 ? 30 1700-1940 H. cracherodii Cannon 2006 OS-84412 58 SII/L3 81 3230 ? 25 2700-2880 H. cracherodii Taylor 2012 OS-54360 58 Pit 1 93 4750 ? 35 4570-4840 H. cracherodii Cannon 2006 OS-54362 60 Compact 49 1670 ? 25 900-1100 H. cracherodii Cannon 2006 OS-54361 61 Compact 50 1070 ? 45 330-550 L. gigantea Cannon 2006 OS- 69734 64 S3/L1 100 4590 ? 25 4390-4650, 4670- 4670 H. cracherodii CSULA 2008 a Depths are reported as cm below datum as originally recorded, but different datums were used for different units. b Dates calibrated using CALIB 6.0 (Stuiver et al. 2010). For marine shell, a Marine09 calibration (Reimer et al. 2004) and ?R value of 225 ? 35 were applied (Kennett et al. 1997; Prior et al. 1999; Rick et al. 2005b; Vellanoweth 2001a). c 1 Sigma Cal BP date reported by Cannon 2006. CALIB 6.0 indicates that this age range cannot be calibrated. 99 Figure 3.14. Tule Creek Village, Mound B dates (2-sigma calibrated). Dates reported by Cannon (2006), CSULA (2008), and as part of the current study. 100 Figure 3.15. North wall profiles showing the locations of the samples dated and the corrected ages obtained for the samples. The profiles are the units used in the lithic analysis at Mound B. The horizontal dimension is at ? scale. 101 In units where Stratum IIA is present, Mound B stratigraphy is complex. The compact and lighter-colored Stratum IIA was first encountered in units 59, 60, and 61 and also appeared in units 11, 12, 13, 15, and 32 (Figure 3.16, 5.17). Transitions between Stratum II and Stratum IIA are typically gradual and mottling is minimal. In 2008, the excavation team hypothesized that the stratigraphy represents post-depositional disturbance associated with the construction of a 1950s-era road adjacent to the site. The road is still present near the site, but little is known about the nature of its construction. Lighter-colored compact Stratum IIA could have resulted from the transport and redeposition of Stratum I by road construction, workers? shovels, or mechanical excavation equipment. Alternatively, lighter (Stratum IIA) strata might have resulted when clay particles washed downslope from the road, onto the site surface, downward through the profile. Another possibility is that Stratum IIA represents periods of intensive aeolian deposition of carbonate-rich silt. In these scenarios, darker layers might result from an increase in organic particles introduced by activities such as cooking, processing, or discard of refuse (Marty et al. 2010; Taylor et al. 2009b). Radiocarbon dating and sediment analysis were used to investigate these hypotheses. Radiocarbon dating results support the hypothesis that the alternating light and dark layers were caused post-depositional disturbance caused by road construction. In units 11, 13, and 32, the profiles where Stratum IIA is present show chronological reversals (Figure 3.14). For example, in Unit 32, a date of 2720-2930 cal BP in the uppermost level of Stratum II is underlain by a date of 140-410 cal BP in Stratum II/Level 5 (Figure 3.17). In Unit 13, dates for the upper layers of Stratum IIA and Stratum II appear in chronological order; however, Stratum II appears to dip deep into Stratum IIB in the SE corner of the unit, and a date of 320-490 cal BP at 87 102 cmbd is younger than those above it although this could be due to a pit feature as in Unit 48 (Figure 3.18). Figure 3.16. Plan view and north wall of Unit 43, Mound B, showing darker Stratum II overlying lighter-colored Stratum IIA. Figure 3.17. North walls of Units 13 and 32, Mound B, showing the lighter-colored Stratum IIA within Stratum II. 103 Figure 3.18. North and south wall profiles of units 32 and 13 showing stratigraphy and dating results. The profile drawing of Unit 32 is based on profiles drawn in the field and on level forms. The profile drawing of Unit 13 was based primarily on level forms. 104 Sediment analysis of bulk samples from units 11, 12, 13, 15, 32, and 43 also supports the hypothesis that alternating light and dark layers within Stratum II were caused by mechanical road construction disturbance. If the light, compact Stratum IIA had originated from downward movement of road clay or exotic wind-blown silt, the grain-size distribution of Stratum IIA samples would be different from the grain-size distribution from Stratum II samples. However, a grain-size analysis conducted with CSULA graduate students Johanna Marty and Nicholas Poister (152H type hydrometer and sieve shaker) shows no statistically significant difference in abundance of silt and clay-sized particles in light layers and dark layers (Marty et al. 2010; Taylor et al. 2009b). An independent samples t-test indicates that proportion of fines to sand is not significantly greater in light Stratum IIA layers (x = 6.20, n = 8) than in darker Straum II layers (x = 5.70, n = 8); t(df = 14) = 0.8325, p = 0.73 (Table 3.6, Figure 3.19). These results suggest that Stratum IIA more likely resulted from excavation, erosion, and deposition of Stratum I and II than from an increase in wind-blown silts and clays. 105 Figure 3.19. Percent fines (silt and clay) in units 43, 32, 13, 15, 11, and 12, Mound B based on work by Marty et al. (2010) and Taylor et al. (2009b). A loss on ignition (LOI) analysis conducted by University of Washington undergraduate Jordan Martinez (Martinez 2010) suggests that compaction of Stratum IIA is likely attributable to higher carbonate content. An independent samples t-test indicates that percent carbonate was significantly higher in lighter layers (x = 13.62, n = 6) than in darker layers (x = 10.56, n = 7); t(df = 11) = 5.56; p = 0.00 (Table 3.7, Figure 3.20). Since the LOI analysis did not include a sample from the road or from Stratum I, it is difficult to determine why Stratum IIA had higher carbonate levels but presumably the road was made using a carbonate-rich material and it was deposited on the site during construction. The percentage of organic matter showed no significant difference in lighter (x = 2.31, n = 6) and darker strata (x = 2.04, n = 7); t(11) = 0.9344; p = 0.3702, which suggests that differences in cultural activities did not contribute significantly to color changes. 106 Table 3.6. Results of a grain size analysis at Mound B. Unit Stratum Depth below datum (cm) Field Description Weight Sand Weight Fines % Fines 43 II 61-75 Dark 80.67 7.48 8.48% 43 IIB 80-85 Transition to non-cultural 85.25 4.57 5.09% 43 III 90-100 Sterile 82.04 3.32 3.89% 32 II 30-45 Dark 82.13 11.60 12.37% 32 IIA 52-61 Light 78.29 10.10 11.43% 32 IIA 66-78 Light 82.42 6.94 7.77% 32 IIB 88-98 Transition to non-cultural 85.11 4.55 5.07% 32 III 120-125 Sterile 89.76 6.72 6.96% 13 IIA1 7-12 Light 88.47 5.02 5.37% 13 IIA 9-14 Light/Disturbed 87.58 5.17 5.58% 13 IIB1 19-27 Dark 62.30 2.01 3.13% 13 IIA2 27-34 Light 87.52 5.04 5.44% 13 IIB2 37-47 Dark 86.79 6.06 6.52% 13 IIB 46-59 Transition to non-cultural 95.34 2.01 2.07% 15 IIB2 26-34 Dark 90.72 4.03 4.26% 15 IIB 40-48 Transition to non-cultural 94.54 2.01 2.08% 11 IIA1 60-63 Light 87.88 6.03 6.42% 12 Surface Road 76.07 9.11 10.69% 12 II 11-19 Dark 89.67 3.02 3.25% 12 IIA1 24-28 Light 89.66 4.03 4.30% 12 IIB1 30-38 Dark 88.45 4.04 4.37% 12 IIA 40-45 Light 89.25 3.02 3.27% 12 IIB2 47-54 Dark 92.77 3.01 3.20% 12 IIB 60-68 Transition to non-cultural 92.12 1.00 1.08% Table 3.7. Results of a loss-on-ignition analysis for bulk samples from units 43, 32, 13, and 15 at Mound B based on work by Martinez (2010). Deposit % Organic % Carbonate Field Description Unit 43 SII.61-75 2.39 12.05 Dark Unit 43 SIII.90-100 0.01 9.35 Sterile Unit 32 SII.30-45 2.85 14.46 Dark Unit 32 SIIA.52-61 2.27 14.93 Light Unit 32 SIIA.66-78 1.77 10.65 Light Unit 32 SIIB.88-98 1.08 10.32 Transition to non-cultural Unit 32 SIII.120-125 1.04 12.66 Sterile Unit 32 SIIA.9-14 2.87 12.76 Light Unit 32 SII.48-59 1.22 9.56 Dark Unit 32 SIIA.60-63 2.45 13.09 Light Unit 32 SIIB.80-85 1.37 9.64 Transition to non-cultural Unit 13 SIIA1.7-12 2.06 14.53 Light Unit 13 SIIB1.19-27 2.03 10.66 Dark Unit 13 SIIA2.27-34 2.31 11.97 Light Unit 15 SIIA1.9-17 2.41 10.31 Light Unit 15 SIIB2.26-34 2.48 10.78 Dark Unit 15 SIIB2.40-48 1.24 9.92 Dark 107 Figure 3.20. Percent organics and carbonates in units 43, 32, 13, and 15, Mound B based on work by Martinez (2010). 108 Temporal Analytic Units for Mound B After dating and analyzing the stratigraphy at Mound B, I divided the lithic assemblage into temporal analytic units (Figure 3.21). Based on environmental shifts discussed in the previous chapter and the clustering of dates at Mound B, I break the chronology into four periods: At 5000-4000 cal BP both marine and terrestrial resources are abundant. For the 3000- 1500 cal BP and 500 cal BP-Contact periods, terrestrial climate is cool and productive while marine environment is less productive. For the 1500-500 cal BP period, marine resources should be abundant and terrestrial resources should be less productive. Because of post-depositional disturbance at Mound B in Units 11, 13, and 32, I excluded almost all of the disturbed upper layers of these units from the analysis and only included the lower levels below the area that appeared disturbed in the stratigraphic profile for which I had radiocarbon dates for that specific 10 cm level or which were between two dated levels. For Unit 11, I included 117 flakes from Stratum II levels 2 and 3 (500 cal BP-Contact) based on a date for Stratum III Level 3 and the location of both of these levels below the areas with the lighter Stratum IIA. For Unit 13 I included 69 flakes from a discrete lens of Stratum IIA level 2 that dated to 500 cal BP-Contact and which may represent a redeposited Stratum I. I also included 196 flakes from Stratum II Level 3 and 4 (1500-500 cal BP) based on their location below the disturbed area and a date for the upper part of the level. A younger date for the lower part of the level is problematic but because most of the artifacts came from the upper part of Stratum level 4 I chose to include them in the 1500-500 cal BP sub-assemblage. The very young data in Unit 13 Stratum IIB (lowest cultural stratum) is also problematic and may be caused by disturbance, but based on the appearance of the stratigraphy, only the two upper strata were affected by road construction and the younger date in Unit 13 Stratum IIB is more likely caused by a pit feature or 109 another post-depositional event. For Unit 32, I included artifacts from below the areas with Stratum II Level 4 where dates and stratigraphy indicate that there was no additional disturbance. These included 32 flakes from Stratum II Level 6 (1500-500 cal BP) and 161 flakes from Stratum II Level 4 and 5 (500 cal BP-Contact). By including some artifacts from these disturbed units, I may have created a limited amount of mixing between time periods, but the dating is consistent enough in the majority of the assemblage that these effects should be swamped by the other artifacts. Thus, spatial units for comparing lithic artifacts across the site within the same time period (500 cal BP-Contact) consist of Spatial Unit 1: Unit 58, Stratum I, Stratum II levels 1-2 and Spatial Unit 2: Unit 43, Stratum I and II. These areas of the site were chosen as spatial analytic units because they are far enough away from one another that they are probably not associated with the same household and because they are well-dated. 110 Figure 3.21. Temporal analytic units at Mound B, not to scale. Those strata that are not enclosed in a time period box were not included in some analyses because dating was uncertain. Spatial units (500 cal BP-Contact) consist of Unit 58, Stratum I, Stratum II levels 1-2 and Unit 43, Stratum I and II. Excavation, Stratigraphy and Dating at CA-SNI-106 In addition to the lithic assemblage from Mound B, I also analyzed a small assemblage of lithic artifacts from CA-SNI-106. This approximately 20,000 m 2 habitation site on a sand dune was first investigated by Reinman and his students in 1984 (Martz 2002). The lithic assemblage comes from a 1.5 m x 1.5 m unit excavated in arbitrary levels during a CSULA field school directed by Martz in 1994. All sediment was screened through ?-inch mesh. The stratigraphy at this site is characterized by a thick (approximately 60 cm) shell-bearing stratum with lenses of 111 silt and ash underlain by sand with no cultural materials present (Figure 3.22). Most artifacts were found at 10-20 cmbs, and a total of 480 lithic artifacts were recovered. This assemblage is located at the archaeology laboratory on San Nicolas Island. Figure 3.22. Stratigraphy and dating of an index unit at CA-SNI-106. Table 3.8. Radiocarbon dates for CA-SNI-106. Sample # Unit/Stratum/Feature Depth (cm) Uncorrected 14 C Age BP Calibrated Age Range, 2 Sigma BP a Material Reported Beta 96686 Index Unit 10-18 2530 ? 60 1750-2110 charcol Martz 2008 Beta 243467 Index Unit 20-30 3030 ? 40 2370-2700 Olivella sp. bead Martz 2008 Beta 96687 Index Unit 30-40 2490 ? 70 1700-2960 charcol Martz 2008 Beta 243466 Index Unit 50-60 3910 ? 40 3450-3750 Mytilus sp. bead Martz 2008 Beta 96688 Index Unit 80-90 2740 ? 60 2000-2330 charcol Martz 2008 a Dates calibrating using CALIB 6.0 (Stuiver et al. 2010) ), for shell dates Marine09 calibration (Reimer et al. 2004) and ?R value of 225 ? 35 were applied (Kennett et al. 1997; Prior et al. 1999; Rick et al. 2005b; Vellanoweth 2001a). 112 Dates reported by Martz (2008) suggest that the upper levels of the unit at SNI-106 are approximately 2000 years old with an occupation that ranges from 2000-3000 cal BP below 20 cmbs (Figure 3.20, Table 3.8). The charcoal dates are in stratigraphic order and the shell dates appear slightly older than the charcoal dates, which suggests a potential discrepancy with the shell calibration. The small shell beads could also have moved vertically within the profile. Artifacts from 10-18 cmbs were placed within a 1000-2000 cal BP analytic unit, and artifacts from 20-50 and 60-90 cmbs were placed within a 2000-3000 BP analytic unit. Those from 50-60 cmbs were excluded from the analysis due to the dating uncertainty. Chapter Summary In this chapter, I provided data on radiocarbon dating and stratigraphy at the Watmough Bay site on Lopez Island, Tule Creek Village on San Nicolas Island, and CA-SNI-106 on San Nicolas Island. I integrated these data with Late Holocene climate change information to divide the assemblages into sub-assemblages dating to approximately 1000-year time intervals. These temporal units will be used to determine if lithic procurement shifts parallel changes in resource abundance and predictability. I also created spatial units to compare between households within the Watmough Bay and Tule Creek Village sites to test hypotheses about differences in resource access based on kin affiliation. 113 Chapter 4: Results of Toolstone Surveys Settlement pattern research provides an essential framework for investigating territory size and boundary defense. Lithic procurement patterns bridge landscape and site-scale analysis and provide further insight on circumscription, inter-village relationships, and resource access. In studies of highly mobile hunter-gatherers, source provenance analyses are often used to establish the presence and extent of lithic conveyance zones and mobility patterns within these areas (e.g., Bamforth 1990; Bettinger 1982; Blades 1999; Brantingham 2006; Dillian 2003; Eerkens 1999, 2010; Eerkens et al. 2008; Holdaway et al. 2010; Jones et al. 2003; Seeman 1994; Skinner et al. 2004; Smith 2010; Tankersley 1990; Walsh 1998). The sources of the toolstone are typically geochemically distinct and located in geographically limited areas. Travel is usually pedestrian and distance is roughly equivalent to travel cost. In many of these studies, exchange?transfer of objects or materials owned by one social group to another social group?is thought to be rare due to low population density. On the Pacific Coast of North America, establishing the ?lithic landscape? (Wilson 2007:391) is complicated by toolstone distribution, boat travel, and frequent exchange. Due to erosion by wave action, toolstone is often found in the form of rounded beach cobbles. In both the San Juan Islands and the southern Channel Islands, the geological deposits from which toolstone was collected stretched over hundreds of kilometers. People could have transported bulk loads of toolstone over long distances using both pedestrian transport and large boats. They also likely engaged in exchange of both everyday and extralocal or rare lithic raw material such as chert and obsidian. Despite these complexities, stone artifacts at archaeological sites in both study areas represent the resource acquisition activities of the people who lived there. Interpreting the lithic procurement record requires a creative and speculative approach that 114 begins with a comprehensive understanding of the nature and extent of toolstone resources on the landscape. I focus mainly on flaked stone technology but also incorporate data from ground stone technology. In this chapter, I present the results of field and laboratory research on the lithic landscape of both study areas as an essential background to testing predictions regarding territorial strategies. I consider not only the characteristics of toolstone collection areas but also explore the question of desirability of different source areas to native inhabitants of the San Juan Islands and San Nicolas Island. The Lithic Landscape of the San Juan Islands The main source of toolstone for precontact communities of the San Juan Islands was fine-grained volcanic rock (FGV) (Figure 4.1). In Gulf of Georgia literature, this material is also referred to by archaeologists as basalt (Carlson 1960), dacite (Bakewell 1996, 2005), and crystalline volcanic rock or CVR (Close 2006, 2011) (see Bakewell 2005:1-8 for a detailed description of archaeological nomenclature for San Juan Islands toolstone). It is dark gray to black in color, fine-grained, and sometimes contains small (1 mm) white phenocrysts. Relative to chert and obsidian, the toolstone is difficult to work; however, it creates durable tools. Other types of toolstone present in small amounts in San Juan Islands assemblages include slate, schist, quartzite, quartz, metasedimentary rock, chert, and obsidian. Slate, schist, and quartz were all available at many local outcrops including a large slate outcrop at Watmough Bay. Quartzite was available in the form of local beach cobbles eroding out of glacial deposits. 115 Figure 4.1. Artifacts made from FGV found on a San Juan Island beach. Unlike FGV, chert and obsidian are quite rare in most assemblages on the San Juan Islands. They are unavailable locally and would have been acquired during travel or through exchange. Chert deposits have been found in the Skagit drainage basin (Waitt 1977) and the Fraser River Valley (Lian and Hiscock 2001). There are numerous source areas for obsidian in Washington, Oregon, and British Columbia mapped by Craig Skinner on the Northwest Research Obsidian Studies Laboratory website (Skinner 2011). Carlson (1994) uses XRF data on artifacts from throughout British Columbia to analyze changes in exchange patterns in the Pacific Northwest. He finds a decrease in long-distance exchange and an increase in local Garibaldi obsidian after approximately 2500 cal BP. Reimer (2000:207) also surveys the distribution of Mt. Garibaldi obsidian in the Gulf of Georgia. Because the source area is small and difficult to 116 find, he suggests that access may have been restricted to a small family groups. Large amounts of Garibaldi obsidian are found in assemblages in the Squamish River Drainage. Smaller amounts of this material are found as far as 100 kilometers away. Ground stone artifacts at Watmough Bay are made from basalt, sandstone, shale, argillite, nephrite, and steatite. Nephrite and steatite may not be available locally but no formal survey has been conducted to determine if this is the case. Fraser Canyon nephrite was used throughout the Gulf of Georgia by 2,500 BP (Grier 2003). Nephrite occurs as alluvial cobbles along the Fraser River, and among British Columbia Plateau groups. Nephrite celts were likely considered status markers and were part of a trade system between the Fraser Canyon to the Shuswap Lakes and Nicola Valley (Darwent 1996). Previous lithic analysis demonstrates that FGV toolstone was obtained in the form of small water-worn cobbles and pebbles (Bakewell 1996, 2005; Close 2006:159; Kornbacher 1992). In his dissertation research, Bakewell (1996, 2005) proposes that precontact native peoples of the San Juan Islands traveled by boat to southern British Columbia where they gathered cobbles of FGV. His geochemical analysis of San Juan Islands artifacts indicates that the ultimate source of the toolstone was an eruption from the Watts Point volcanic center, the southernmost volcanic flow of the Mt. Garibaldi volcanic belt on Howe Sound in southern British Columbia and approximately 0.02 km 3 (Bye et al. 2000:1) (Figure 4.2). 117 Figure 4.2. Map showin the Watts Point volcanic center. Beaches on Howe Sound where FGV cobbles were abundant are marked with black circles. Bakewell (2005:85-86) suggests that Watts Point FGV was carried from the outcrop to the beaches of Howe Sound by longshore currents; it was collected and transported by boat to the San Juan Islands. He suggests that people did not collect FGV from beaches in the vicinity of English camp because (1) he found few natural FGV cobbles during a surface survey of the beaches near English camp, and (2) the smooth glassy appearance of the material suggests a post-glacial eruption (after 10,000 BP) in which case Watts Point FGV could not have been transported by glaciers to the San Juan Islands. Bakewell also based his assumption of a post- glacial Garibaldi eruption on potassium-argon dating conducted by Green (1989; Green et al. 118 1988); however, a review of Green?s dating research indicates that the Watts Point flow dates to between 90,000 + 30,000 BP to 130,000 + 30,000 BP. He notes potential for an additional Holocene dacite flow but does not provide a date. Recent geological investigation of the Watts Point outcrop by Bye and her colleagues suggests that the eruption occurred in a subglacial or englacial environment, rather than a post-glacial environment, because the FGV is found beneath glacial till (Bye et al. 2000:1). Many Northwest Coast researchers suggest that contrary to Bakewell?s conclusions, precontact peoples in the San Juan Islands and throughout western Washington acquired cobbles of toolstone from local beaches rather than from Howe Sound (Conca 2000; Herbel et al. 2001; Kornbacher 1992; Morgan et al. 1999; Wessen 1993). Given a Middle or Late Pleistocene origin, Watts Point FGV could have been transported by ice to the San Juan Islands during the Fraser glaciation at 18,000-13,000 BP (Armstrong et al. 1965; Booth 1987; Dethier et al. 1995; Easterbrook 1969, 1986, 1992; Thorson 1980). As waves eroded glacial marine drift deposits on the San Juan Islands after the glacial ice melted, cobbles of Watts Point FGV and other stone would have eroded out of finer sediment and blanketed beaches where glacial marine drift was abundant in moraines, outwash, wave-cut banks, and creeks. Prior to Kwarsick?s (2008, 2010) thesis work and this project, only limited geochemical testing and survey had been conducted to address this hypothesis (e.g., Morgan et al. 1999:C4). Kwarsick collected cobbles from beaches and rivers on the northern Olympic Peninsula and surrounding areas and submitted 51 cobbles to the Northwest Research Obsidian Studies Laboratory for energy dispersive X-Ray Fluorescence (EDXRF) analysis. Of these, over 60% matched the Watts Point source. Of a sample of 111 artifacts from 54 archaeological sites on the Olympic Peninsula and surrounding region, 100 artifacts matched the Watts Point source. 119 Kwarsick?s work supports glacial transport of Watts Point FGV throughout Western Washington. It is also demonstrates that people differentially selected this material as toolstone. During the summers of 2007-2009, I conducted geological fieldwork and geochemical testing to resolve the source provenance of FGV toolstone for precontact peoples of the San Juan Islands. Geochemical analyses to characterize trace element composition of toolstone sources and artifacts were conducted by Craig Skinner at Northwest Research Obsidian Studies Laboratory using EDXRF with a Spectrace 5000 energy dispersive X-ray fluorescence spectrometer with a resolution of eV FHWM for 5.9 KeV X-ray (1000 counts per second) in an 30 mm 2 area. This method is relatively inexpensive, requires minimal sample preparation, is non-destructive, accurate, and has been used successfully for numerous FGV source provenance studies (e.g., Johnson 2010; Jones et al. 1997; Mills et al. 2008, 2010; Page 2008). For geological samples, I removed flakes from FGV beach cobbles to provide flat, clean surfaces for irradiation. For the XRF process, samples are subjected to radiation that moves electrons from higher to lower energy shells. Other electrons replace the one that have moved, releasing fluorescent X- rays. The energy released creates a signal specific to the chemical element present in the sample (Pollard et al. 2007). For the FGV samples submitted for this study, elements identified and quantified include zinc (Zn), lead (Pb), rubidium (Rb), strontium (Sr), yttrium (Y), zirconium (Zr), niobium (Nb), titanium (Ti), manganese (Mn), barium (Ba), sodium (Na), potassium (K), and iron oxide (Fe 2 O 3 ). Elemental concentrations for the sample are calculated in parts per million by converting energy spectra peaks into measurable quantities using a regression calibration to known standards (Glascock et al. 1998). If diagnostic element values for geological samples and artifacts fall within approximately two standard deviations of upper and lower limits 120 of chemical variability for the source material, the tested sample or artifact is identified to the source. The Northwest Reseach Obsidian Studies Laboratory provides additional information on their methodology on their website (http://www.obsidianlab.com/info_xrf.html). Below, I use biplots of Sr (ppm) and Zr (ppm) concentrations to visually display differentiation of Watts Point FGV and ?Unknown FGV? source material. Geological samples of Watts Point FGV were collected from a modern stone quarry at an exposure at the side of Route 99 (n = 4) and from nearby Murrin Provincial Park where FGV pebbles are located on the ground surface (n = 6). Survey of Howe Sound beaches indicates abundant FGV cobbles (Figure 4.2). Geochemical testing of these samples confirms that they are Watts Point FGV. Two samples from upper Boulder Creek near Mt. Baker, Washington displayed different geochemical signatures from San Juan Islands lithic artifacts and geological samples from Watts Point. Geological samples from beaches in the Puget Sound area at Double Bluff Beach on Whidbey Island, Quartermaster Harbor on Vashon Island, and the Dungeness Spit on the Olympic Peninsula also displayed an unknown FGV (Figure 4.3; Appendix C). To determine if Watts Point FGV was transported by glaciers, I collected FGV pebbles and cobbles from beaches and glacial deposits in the San Juan Islands. I submitted rocks collected from Schoen Beach (n = 2), Agate Beach (n = 29) Deadman?s Cove (n = 3), False Bay (n = 36), Watmough Bay (n = 11), and from Blind Bay glacial deposits (n = 9). Of the 72 cobbles visually identified as FGV from surface deposits, 64 were identified geochemically as Watts Point FGV (Figure 4.4, 4.5). One possible explanation for the presence of Watt Point FGV on San Juan Islands beaches is that people brought the cobbles to archaeological sites and materials from these sites are eroding out onto the beaches at Schoen Beach, Blind Bay, and Watmough Bay. This argument is not supported by the presence of Watts Point FGV in glacial 121 deposits and the presence of Watts Point FGV at Deadman?s Cove, False Bay, and Agate Beach where no archaeological sites have been recorded. The most parsimonious explanation is that the cobbles are eroding out of the glacial deposits. Many of the small pebbles collected from pits excavated into glacial deposits at Blind Bay (n = 8 of 9) and False Bay (n = 8 of 10) were identified as an unknown FGV source. This suggests that the larger FGV cobbles found on the beach are size sorted by wave action, and many of the cobbles that are an appropriate size for use as cores are Watts Point FGV. To assess the ubiquity of Watts Point FGV in San Juan Islands tool assemblages, I sent FGV artifacts from the Deane site (45-SJ-150; n = 20), the Watmough Bay site (45-SJ-280; n = 50), and the English Camp site (45-SJ-24; n = 30) to Northwest Research Obsidian Studies Laboratory for geochemical analysis. I chose artifacts that had at least one non-cortical flat surface. I also chose samples from different time periods and/or different areas of the site to ensure that the samples represent the spatial and temporal diversity of toolstone at each site. By far the majority of the artifacts (n = 93 of 100) were identified as Watts Point FGV (Figure 4.4, 4.5; Appendix C). These data support the hypothesis that native communities carefully selected a particular type of cobble and that they primarily procured toolstone from local beaches rather than traveling to Howe Sound, although they may have done so occasionally. 122 Figure 4.3. Biplot of zirconium and strontium concentrations (ppm) from EDXRF data for Howe Sound and Mt. Baker samples. Data points within the ellipse are identified as Watts Point FGV. 123 Figure 4.4. Biplot of zirconium and strontium concentrations (ppm) from EDXRF data for stone artifacts from the Deane site, English Camp, and Watmough Bay. Figure 4.5. Biplot of zirconium and strontium concentrations (ppm) from EDXRF data for stone artifacts from the Deane site, English Camp, and Watmough Bay showing Watts Point and unknown source FGV distribution. 124 Figure 4.6. Biplot of zirconium and strontium concentrations (ppm) from EDXRF data for geological samples from San Juan Islands cobble areas. Figure 4.7. Biplot of zirconium and strontium concentrations (ppm) from EDXRF data for geological samples from San Juan Islands cobble areas showing Watts Point and unknown source FGV distribution. 125 Cobble Surveys in the San Juan Islands It may not be possible to use a traditional geochemical source provenance analysis to determine where past communities in the San Juan Islands acquired their toolstone; however, information about the range of variation in size and shape of the cobbles on local beaches provides important clues about procurement behavior. I systematically surveyed six cobble beaches on the San Juan Islands during the summer of 2010 (Figure 4.8, 4.9). I chose beaches for survey that were within a boat ride from the Watmough Bay site and had high potential for exposed glacial deposits. On all six beaches, quaternary glacial deposits are present and high wave energy deflates large cobbles out of glacial deposits and erodes away sand and pebble- sized rocks. Recent GIS work by Finlayson (2006) identified moderate to high fetch beaches. FGV cobbles are visually distinct from other material types due to their dark color and reddish- brown to black cortex appearance. I conducted cobble surveys in 100 meter 2 areas at American Camp, Watmough Bay, False Bay, Snug Harbor, Agate Beach, and Aleck Bay. Survey areas were located in the sections of the beaches where cobbles appeared most abundant. I surveyed at one-meter intervals within the survey areas and recorded cortex appearance and dimensions of FGV cobbles for all cobbles over 2 cm in maximum dimension. Below I briefly describe the characteristics of each cobble survey area. 126 Figure 4.8 Cobble survey areas on San Juan Island. Dark gray areas indicate quaternary glacial deposits. 127 Figure 4.9 Cobble survey areas on Lopez Island. Dark gray areas indicate quaternary glacial deposits. Watmough Bay beach is owned by the Bureau of Land Management and the San Juan County Land Bank. It is approximately 6980 meter 2 , bounded to the southeast and northwest by large bedrock outcrops. To the southwest lies the archaeological site and a large wetland. Due to relatively low wave energy and/or the source sediment, the beach is predominantly composed of sand and many FGV cobbles there are relatively small (< 3 cm in maximum dimension). I noted a total of 64 FGV cobbles within the 100 meter 2 survey area (Figure 4.10). 128 Figure 4.10. View to the north of the 10 x 10 meter survey area at Watmough Bay. The 100 meter 2 survey area on Aleck Bay beach, Lopez Island, is located on private land owned by the Mendez family. Based on the Department of Archaeology and Historic Preservation (DAHP) database, there are no recorded archaeological sites sites on Aleck Bay. The Mendez property includes a section of the cobble beach that extends around the bay. The total area of this beach is approximately 8818 meter 2 . The abundance of large driftwood logs suggests that there are periods of high wave energy on Aleck Bay beach. I noted 30 cobbles in the 100 meter 2 survey area; however, approximately 20% of the area was obscured by seaweed and driftwood (Figure 4.11). 129 Figure 4.11 View to the northeast of the 10 x 10 meter 2 survey area at Aleck Bay beach. Agate Beach on Lopez Island is located on land owned by San Juan County Parks. No archaeological sites have been recorded on or near Agate Beach. Rounded and waterworn FGV cobbles are abundant on this beach intermixed with other rock types. The total size of this beach is approximately 4900 meter 2 . Because the pebbles on the beach decrease in size towards the intertidal zone, I created a 5 x 20 meter survey area closer to the high-tide line. Approximately 20% of the survey area was obscured by driftwood. I noted 69 FGV cobbles within the 100 meter 2 survey area (Figure 4.12). 130 Figure 4.12. View to the southeast of the survey area at Agate Beach. The beach at American Camp on San Juan Island is located on land owned by San Juan Island National Historic Park. The western end of this beach, Eagle Cove, is on land owned by San Juan County Parks. The entire southern end of this island is a long glacial moraine and the cobble beach is composed of waterworn pebbles deflated from this moraine and outwash. Because this landform is so large (approximately 81,538 meter 2 ), I surveyed within two 10 x 10 m 2 areas to better estimate the range of variability in cobble abundance, size and shape. The first survey was located in the middle section of the landform. Most rocks were too small to be used as toolstone and a variety of rock types were present. A total of 10 FGV cobbles were found in this area (Figure 4.13). At the western end of the landform at Eagle Cove, 60% of the survey area was obscured by driftwood, but I identified a total of 26 FGV cobbles within the 10 x 10 meter 2 131 survey area (Figure 4.14). No archaeological sites have been recorded in the vicinity of American Camp beach although a shell midden has been recorded at Eagle Cove. Figure 4.13. View to the southeast at the the survey area at American Camp. Figure 4.14. View to the east of the survey area at Eagle Cove. 132 False Bay Beach on San Juan Island is located on land owned by the University of Washington. The beach is approximately 33,743 square meters and it is backed by a high cliff composed of glacial marine drift. Due to the rapidly rising tide, I surveyed within a 5 x 20 meter area at the high tide line. Approximately 40% of the survey area was obscured by seaweed, driftwood, and silt and clay but I found a total of 94 FGV cobbles, both rounded and angular. There are no recorded sites at False Bay (Figure 4.15). Figure 4.15. View to the northwest of the survey area at False Bay. Snug Harbor beach is located on land owned by Snug Harbor Resort and Marina on Mitchell Bay, San Juan Island. Today, the cobble beach is approximately 33,743 meter 2 ; however, extensive modern disturbance associated with the construction of the marina, resort, 133 parking lot, and sea wall have changed the size of the original cobble collection area. In the 10 x 10 meter survey area I found both round and angular cobbles (Figure 4.16). A shell midden is eroding from behind the seawall adjacent to the beach, and I noted FGV artifacts, tested cobbles (one flake removed), and cores in the survey area. Figure 4.16. View to the southeast at Snug Harbor. San Juan Islands Toolstone Summary Statistics Differences in the characteristics of toolstone available at Watmough Bay and at other nearby beaches are essential to evaluating change over time in land use by precontact peoples in the context of shifting territorial behavior. When people live in smaller and more highly defended territories, they should be constrained in their choice of toolstone to the cobbles on 134 Watmough Bay beach. When territories are larger and more poorly defended, people should procure cobbles from other cobble beaches. To distinguish Watmough cobbles from other cobble beaches, I investigate the cobble assemblages to determine if there are significant differences in size and shape. In the final section of this chapter, I use Wilson?s (2007a,b; 2011; Browne and Wilson 2011) quarry attractiveness calculation to evaluate which cobble areas on the island would have been most desirable to native inhabitants. To investigate whether differences in the composition of glacial marine drift and beach morphology affect the size and shape of cobbles on different beaches on the San Juan Islands, I use length as a proxy for size and measure and describe several dimensions of shape. Cobble flatness is measured as a ratio of length + width/2 x thickness (Wentworth 1922) where higher values represent flatter cobbles. Cobble elongation is measured as a ratio of width x 100/length (Sames 1966) where elongated cobbles have lower values. In the field, I classified cobbles into general categories of ?round? or ?angular?, described cortex as black or gray, and described texture as unpolished (rough) or polished (smooth and waterworn). Results of an ANOVA test to compare sample means indicate a statistically significant difference in mean length (used as a proxy for size) between Watmough Bay cobbles and mean length from all other cobble areas except American Camp. Watmough Bay cobbles are significantly smaller than cobbles from the other beaches. Most cobble areas are significantly different from one another in mean length (Table 4.1, 4.2, 4.3). Results of ANOVA also suggest a statistically significant difference in mean elongation between cobble areas; a Bonferroni post- hoc analysis indicates that only Watmough Bay and False Bay are significantly different in mean elongation (p = 0.05) whereas all other cobble areas are statistically similar in mean elongation (Watmough Bay and Agate Beach p = 0.08, all other comparisons p = 1.0) therefore this measure 135 is not helpful in distinguishing between cobble areas. Roundness and flatness measures show no significant differences between cobble beaches on the San Juan Islands. Results of ?2 tests indicate that ?round? versus ?angular? categories may be more helpful in distinguishing between cobble areas. Although Watmough Bay is not significantly different in proportion of round and angular cobbles from other beaches, Aleck Bay and American Camp have significantly higher proportions of angular cobbles and False Bay has significantly higher proportions of round cobbles (Table 4.4). Results of ?2 tests also indicate that Watmough Bay and Agate Beach have significantly higher proportions of smooth cobbles and False Bay has significantly higher proportions of rough cobbles (Table 4.4). Results of ANOVA tests and ?2 tests confirm that size, shape category, and cortex texture are all useful for tracing change over time in lithic procurement in the San Juan Islands. In Chapter 6, I use this information to determine when people procured more toolstone from Watmough Bay beach and when they procured toolstone from elsewhere on the San Juan Islands. 136 Table 4.1 Descriptive statistics for six FGV cobble areas in the San Juan Islands. Table 4.2. Results of an ANOVA for cobble size/shape data for San Juan Islands beaches. df x sq. F Sig. Max. Length Between Groups 5 12284.99 51.74 0.00 Within Groups 418 237.458 Total 423 Flatness Between Groups 5 0.223 0.211 0.96 Within Groups 418 1.056 Total 423 Elongation Between Groups 5 638.032 2.263 0.05 Within Groups 418 281.985 Total 423 Watmough Bay min max x min max x min max x Quarry Area (m 2 ): 6980 20 70 37.13 1 4.36 1.65 33.33 109.09 71.17 Survey Area (m 2 ): 100 ? var. ? var. ? var. Total FGV cobbles >2 cm: 64 11.68 136.46 0.51 0.26 17.64 311.22 Aleck Bay min max x min max x min max x Quarry Area (m 2 ): 8820 46 137 76.77 11.16 2.53 1.71 6.25 105.88 75.62 Survey Area (m 2 ): 100 ? var. ? var. ? var. Total FGV cobbles >2 cm: 30 23.74 563.5 0.32 0.1 17.5 306.35 Agate Beach min max x min max x min max x Quarry Area (m 2 ): 4910 34 108 58.39 1.08 3.23 1.73 28.41 122.64 79.33 Survey Area (m 2 ): 100 ? var. ? var. ? var. Total FGV cobbles >2 cm:70 15.98 255.46 0.45 0.2 15.13 228.95 American Camp/Eagle Cove min max x min max x min max x Quarry Area (m 2 ): 81540 23 85 44.47 1.03 3.17 1.61 43.18 129.41 76.35 Survey Area (m 2 ): 100 ? var. ? var. ? var. Total FGV cobbles >2 cm: 36 13.49 182.03 0.39 0.16 18.65 347.66 False Bay min max x min max x min max x Quarry Area (m 2 ): 33740 31 141 69.73 0.86 2.94 1.59 6.17 142.31 79.23 Survey Area (m 2 ): 100 ? var. ? var. ? var. Total FGV cobbles >2 cm: 94 20.26 410.35 0.32 0.11 17.78 316.19 Snug Harbor min max x min max x min max x Quarry Area (m 2 ):760 32 79 51.34 1.04 21.17 1.69 41.18 131.7 75.69 Survey Area (m 2 ): 100 ? var. ? var. ? var. Total FGV cobbles >2 cm: 130 9.71 94.26 1.75 3.05 15.75 248.12 Max. length (mm) Flatness (mm) Elongation (mm) 137 Table 4.3. Results of a Bonferonni post-hoc analysis for cobble length for cobble areas on the San Juan Islands. Source Area Source Area Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval Lower Bound Upper Bound Watmough Agate -21.26 2.67 0.00 -29.13 -13.39 Aleck -39.64 3.41 0.00 -49.71 -29.58 American -7.35 3.21 0.34 -16.82 2.13 False Bay -32.61 2.50 0.00 -39.98 -25.24 Snug -14.21 2.35 0.00 -21.16 -7.27 Agate Watmough 21.26 2.67 0.00 13.39 29.13 Aleck -18.38 3.36 0.00 -28.31 -8.45 American 13.91 3.16 0.00 4.58 23.24 False Bay -11.35 2.43 0.00 -18.53 -4.17 Snug 7.05 2.28 0.03 0.30 13.79 Aleck Watmough 39.64 3.41 0.00 29.58 49.71 Agate 18.38 3.36 0.00 8.45 28.31 American 32.29 3.81 0.00 21.05 43.54 False Bay 7.03 3.23 0.00 -2.51 16.57 Snug 25.43 3.12 0.45 16.21 34.64 American Watmough 7.35 3.21 0.34 -2.13 16.82 Agate -13.91 3.16 0.00 -23.24 -4.58 Aleck -32.29 3.81 0.00 -43.54 -21.05 False Bay -25.26 3.02 0.00 -34.18 -16.35 Snug -6.87 2.90 0.28 -15.43 1.70 False Bay Watmough 32.61 2.50 0.00 25.24 39.98 Agate 11.35 2.43 0.00 4.17 18.53 Aleck -7.03 3.23 0.45 -16.57 2.51 American 25.26 3.02 0.00 16.35 34.18 Snug 18.40 2.09 0.00 12.24 24.55 Snug Watmough 14.21 2.35 0.00 7.27 21.16 Agate -7.05 2.28 0.03 -13.79 -0.30 Aleck -25.43 3.12 0.00 -34.64 -16.21 American 6.87 2.90 0.28 -1.70 15.43 False Bay -18.40 2.09 0.00 -24.55 -12.24 138 Table 4.4. Results of ?2 tests for ordinal values recorded for cobble area data for the San Juan Islands. The Lithic Landscape of San Nicolas Island Toolstone on San Nicolas Island is available primarily in the form of metavolcanic and metasedimentary cobbles from an Eocene-era cobble conglomerate strata (Vedder and Norris 1963; Figure 4.17). This deposit is the only major source of toolstone on the island (Clevenger 1982). Cobbles eroding out of the conglomerate are easily accessible and densely distributed on beaches. They are available but more sparsely distributed in canyon drainages and inland blowouts (Clevenger 1982). Clevenger (1982) located nine potential quarry locations on San Nicolas Island. In the 1980s, Reinman and his students mapped approximately 40 cobble Round Expected Angular Expected Total ? 2 df p Adjusted Residuals Watmough Bay 39 36.38 25 27.62 64 41.97 5 <0.01 0.72 Agate Beach 36 39.79 34 30.21 70 -1 Aleck Bay 11 17.05 19 12.95 30 -2.31 American Camp 11 20.46 25 15.54 36 -3.33 False Bay 77 53.43 17 40.57 94 5.56 Snug Harbor 67 73.89 63 56.12 130 -1.47 Black Expected Gray Expected Total ? 2 df p Adjusted Residuals Watmough Bay 46 45.43 18 18.57 64 30.08 5 <0.01 0.17 Agate Beach 41 49.69 29 20.31 70 -2.51 Aleck Bay 20 21.30 10 8.70 30 -0.54 American Camp 17 25.56 19 10.44 36 -3.29 False Bay 65 66.73 29 27.27 94 -0.45 Snug Harbor 112 92.29 18 37.71 130 4.58 Unpolished Expected Polished Expected Total ? 2 df p Adjusted Residuals Watm ugh B y 7 29.11 48.00 25.89 55 125.6 5 <0.01 -4.1 Agate Beach 4 27.52 48.00 24.48 52 -4.48 Aleck Bay 20 15.88 10.00 14.12 30 1.03 American Camp 14 14.29 13.00 12.71 27 -0.08 False Bay 77 47.10 12.00 41.90 89 4.36 Snug Harbor 77 65.10 46.00 57.90 123 1.48 139 locations throughout the island. These data were included in a GIS database for island cultural and natural resources. Of these previously recorded cobble areas, I chose six areas to survey based on proximity to Tule Creek Village and SNI-106. I included both inland and beach cobble areas to investigate variability in material type, size, and shape based on setting. Pinpointing the exact provenance of a cobble is not possible based on material type or geochemistry because they all come from the same geological deposit (Clevenger 1982:29). Differences in geomorphology and geology of cobble areas produce differences in cobble size, shape, and material type, thus I use data from cobble surveys to investigate change through time in procurement strategies. Erlandson et al. (2008) note that in the Channel Islands, cobbles with internal fractures and those made of softer material are destroyed by waves on beaches but remain intact in inland outcrops. Sandstone outcrops tend to be accessible on beaches due to the surface geology of the island and erosion of other sediment by waves. Figure 4.17. Cobbles eroding out of an Eocene-era cobble conglomerate stratum. 140 In choosing inland cobble areas to survey, I considered the degree to which modern erosion altered the modern distribution of cobbles relative to the precontact distribution of cobbles on San Nicolas Island. During the 1850s, sheep ranching began on the island and continued through the 1940s. Island vegetation was destroyed and large sand dunes formed (Schwartz and Rossbach 1993). The dunes move gradually across the landscape and cover up both archaeological sites and cobble collection areas. In other areas, dune erosion creates a palimpsest where artifacts that date to several time periods are redeposited onto the same surface. Because postcontact inland blowouts and streambeds are not in the same place as precontact inland cobble areas, my data from cobble surveys in those areas may not represent the exact cobble areas used by native communities but are representative of the kinds of inland outcrops that were used. In beach cobble areas, wave energy is the most important erosional force along the coast. Cobble beaches exist where the cobble conglomerate layers are present at the surface. Due to the unique geology of the island, lithic assemblages on San Nicolas Island are primarily composed of metamorphic rocks that erode out of a softer sedimentary matrix. These cobbles are hard and some are coarse-grained and difficult to work. Metasedimentary cobbles and finer-grained metavolcanic cobbles fracture more easily than metavolcanic cobbles that are highly porphyritic with large feldspar crystals (Rosenthal 1996). Rosenthal (1996:304) notes that the geology of San Nicolas Island provides ?few materials suitable for flaked stone tool manufacture? other than the ?less desirable? metamorphic cobbles; however, another perspective is that for the Nicole?o who were accustomed to this material type, adequate toolstone was ubiquitous on the landscape. It may have been more difficult to work than chert or obsidian, but given the abundance of lithic artifacts made from local toolstone on sites across the island, native 141 communities found metavolcanic and metasedimentary rock well-suited to their technological needs. Ta?kiran (2001:55-60) describes in detail the rock types available on the southern Channel Islands and notes that the metamorphism of the San Nicolas Island cobbles can be classified as ?low grade? because parent rocks can be identified by mineralogical composition. These include metamorphosed equigranular and aphanitic latite, basalt, and rhyolite. Since the mineralogical composition of the rock was difficult to determine macroscopically with a high degree of precision, I used broader categories for metavolcanic rocks following the California State University Los Angeles archaeology laboratory protocol. (1) Metavolcanic rocks have few to no phenocrysts (larger crystals within a finer groundmass) (Figure 4.18); (2) Porphyritic metavolcanic rocks have phenocrysts (5-50%) but they do not dominate the groundmass (Figure 4.19); (3) Metavolcanic porphyry rocks have abundant (50%) phenocrysts. These rocks are typically coarse-grained, although in some cases, the phenocrysts fuse together and the rock is fine-grained (Figure 4.20). This classification system for metavolcanic rocks is not based on geological terminology but on a macroscopically consistent archaeological typology relevant to stone tool manufacture. 142 Figure 4.18. A metavolcanic cobble split in the lab. Figure 4.19. Porphyrtic metavolcanic artifacts from Tule Creek Village. Photo courtesy of Jay Flaming?s archaeological photography students. Figure 4.20. A metavolcanic porphyrtic cobble still embedded in the matrix. 143 Quartzite artifacts are identifiable based on their crystalline and sugary appearance characteristic of recrystallized sandstone (Figure 4.21). Small amounts of quartz (Figure 4.22) and fine-grained metasedimentary rock known as ?island chert? are also found in assemblages on San Nicolas Island. These materials were available in small quantities throughout the islands although exact source locations for these materials are unknown. Along with local toolstone, native peoples of San Nicolas Island also used chert that came from other islands and the mainland. Chert artifacts are present in the Tule Creek Village assemblage and range in appearance from banded, solid colored, and translucent. No chert has been found on San Nicolas Island but some have been located on the Northern Channel Islands. Santa Cruz Island chert ranges from white to dark brown to gray and was used to create microblades by craft specialists as part of a shell bead industry during the Late Holocene (Arnold 1987, 1990b, 1992). Cico Chert (4.23) is found on northeastern San Miguel Island. It is translucent and ranges in color from white to gray to brown with some reddish brown and purple rocks and parallel layers or homogenous veins (Erlandson et al. 1997) . Erlandson et al. (2008) also report an outcrop of ?Tuquan? Monterey Chert on eastern San Miguel Island. They note that there may also be a source of toolstone-quality Monterey banded chert and siliceous shales on Santa Rosa Island and elsewhere on San Miguel Island. Monterey chert (Figure 4.24) is widespread along the mainland coast from the Camp Pendleton area to San Francisco. Franciscan chert, which comes in a variety of colors but is not banded, is widespread along the mainland coast from north of Santa Barbara to Oregon. (Cannon 2006:81). Most obsidian comes from the Coso volcanic field inland (Rick et al. 2001a; Figure 4.25). Regarding ground stone, the Nicole?o crafted vessels, shaft straighteners, beads, and ceremonial objects from sandstone, steatite, and serpentine. Tabular sandstone slabs were also 144 used to make saws likely used in the production of shell tools and ornaments (Kendig et al. 2010). While sandstone was available on San Nicolas Island, steatite and serpentine were not. It would have been acquired directly or through exchange from Santa Catalina Island or from the mainland (Cannon 2006:82; Hudson and Blackburn 1987:35). Following previous work in the Channel Islands (e.g., Cannon 2006), I assume that most exotic stone was brought to San Nicolas through exchange in return for sandstone objects, ceremonial toshaawt stones, food resources, and Olivella beads (Cannon 2006; Vellanoweth et al. 2002). The cost of travel to procure these materials would have been great due to the distance between the island and the other sources (Figure 4.25). If people had quarried and collected the material directly, it probably would have been present at sites in larger quantities. Exchange is a more parsimonious explanation, but the possibility of direct procurement cannot be ruled out. Figure 4.21. Quartzite artifacts from Figure 4.22. Quartz artifacts from Tule Tule Creek Village, San Nicolas Island. Creek Village, San Nicolas Island. Photo courtesy of Jay Flaming?s Photo courtesy of Jay Flaming?s archaeological photography students. archaeological photography students. 145 Figure 4.23. Cico chert biface Figure 4.24. Monterey banded chert biface from Tule Creek Village, Mound B. from Tule Creek Village, Mound B 146 Figure 4.25. Map of extra-local toolstone sources for San Nicolas Island, California. The dashed line shows the distribution of Franciscan chert on the mainland. The dotted line shows the distribution of Monterey chert on the mainland and on Santa Cruz Island. The triangle shows the location of Cico chert on San Miguel Island. The circle shows the location of Coso obsidian. Cobble Surveys on San Nicolas Island During the summer of 2009 I worked with field school students at California State University, Los Angeles (CSULA) to record characteristics of cobble deposits on northwest San Nicolas Island in the vicinity of Tule Creek Village, Mound B and CA-SNI-106. Surveys were conducted in 50-200 meter 2 areas at 3 beaches and 3 inland areas (Figure 4.26). All cobbles over 2 cm in maximum dimension were recorded along transects spaced 1 meter apart. During cobble 147 surveys, we could distinguish between metavolcanic, quartzite, and sandstone but quartzite is likely underrepresented because weathered cobbles were more difficult to identify. Figure 4.26. Map showing the locations of cobble survey areas on San Nicolas Island. 1 ? Corral Harbor 2 ? Tule Creek Cobble Area 3 ? Hollywood Beach 4 ? NAVFAC Cobble Area 5 ? Thousand Springs 6 ? Radar Row. 148 The beach cobble area nearest to Tule Creek Village is at Corral Harbor, a cove protected by a sandstone outcrop that would also have served as an optimal location for fishing, hunting sea mammals, and gathering shellfish. At this approximately 5040 meter 2 cobble area, I set up a 10 x 10 meter survey area on the east side of the beach and a 5 x 20 meter survey area on the west side of the beach. This beach is characterized by both abundant metamorphic cobbles and abundant sandstone due to the erosion of a large sandstone outcrop. Within both survey areas I noted a total of 1075 cobbles of which 497 were identified as metavolcanic (Figure 4.27, 4.28). Figure 4.27. Survey Area 1 at Corral Harbor, view to the southeast. Sandstone outcrop in the background. 149 Figure 4.28. Survey Area 2 at Corral Harbor, view to the northeast. Protected harbor in the background. The inland cobble area nearest to Tule Creek Village is an extensive blowout adjacent to the site. It measures 1545 meter 2 . I could not be sure that this blowout would have been present at the time that the site was occupied, but since there was a creek nearby, but since there is a creek nearby, cobbles were likely present in the creek bed. The presence of flakes and cores also indicates procurement from this cobble area. Within a 10 x 10 meter survey area I noted a total of 129 cobbles of which 74 were metavolcanic. Many of the cobbles were identified as broken or ?low quality? indicating that they were made of soft and crumbling metasedimentary rock (Figure 4.29). 150 Figure 4.29. Survey area near Tule Creek Village. Another beach cobble area southeast of Corral Harbor designated ?Hollywood Beach? would have been readily accessible to the occupants of CA-SNI-25. Sandstone outcrops surround the beach. At this approximately 28,010 meter 2 quarry area, I set up two 5 x 5 meter survey areas, one on the east side of the beach and the other on the west side of the beach. Both metamorphic cobbles and sandstone slabs are abundant at this quarry area. I recorded a total of 663 cobbles of which 223 were identified as metavolcanic (Figure 4.30, 4.31). 151 Figure 4.30. Survey Area 1 at Hollywood beach, view to the southeast. Sandstone outcrop and eroding bank in the background. Figure 4.31. Survey Area 2 at Hollywood beach, view to the northwest. Protected cove in the background. 152 Near the NAVFAC building on an inland beach terrace approximately 100 meters inland is large blowout with abundant cobbles in some locations. The area where cobbles were present on the surface was approximately 3750 meter 2 . Within a 10 x 10 meter survey area I counted 101 cobbles of which 79 were identified as metavolcanic. Sandstone was not abundant in this area (Figure 4.32). Figure 4.32. NAVFAC survey area, view the northeast. On the northwestern tip of the island, the Thousand Springs cobble area is located on a high eroding cobble conglomerate cliff. Vegetation is sparse and the top of the cliff is an erosional surface due to high wind. There may be modern disturbance associated with road construction. Cobbles are also abundant on the beach below. In the approximately 5740 meter 2 area where cobbles were abundant on top of the cliff, we surveyed a 10 x 10 meter area. I noted a total of 1612 cobbles of which 1452 were metavolcanic. Sandstone was not abundant in this area. (Figure 4.33). 153 Figure 4.33. Cobble area at Thousand Springs, view to the north. At the southern edge of the high interior plateau on San Nicolas Island, blowouts with exposures of cobbles are abundant because vegetation in scarce. One area of dense cobbles is located off of Radar Row Road. In an approximately 10,130 meter 2 area, I surveyed one 10 x 10 meter area on the south side of the road (Area 1) and another 10 x 10 meter area on the north side of the road (Area 2). I recorded a total of 146 cobbles of which 90 were metavolcanic. Many of the cobbles were identified as broken or ?low quality? indicating that they were made of soft and crumbling metasedimentary material. Sandstone in this area was scarce (Figure 4.34, 4.35). 154 Figure 4.34. Radar Row Survey Area 1 on the south side of the road. Figure 4.35. Radar Row Survey Area 2 on the north side of the road. 155 San Nicolas Island Toolstone Summary Statistics Differences in the characteristics of toolstone available near Tule Creek Village and CA- SNI-106 and other cobble area are used to investigate territorial circumscription during the Late Holocene. To determine if and when people defended smaller territories, I compare toolstone available near Tule Creek Village and CA-SNI-106 and toolstone available elsewhere on the landscape. Proportions of material types (metavolcanic, quartzite, and sandstone) differ substantially between cobble areas with higher amounts of quartzite at Corral Harbor and Thousand Springs and higher amounts of sandstone at Corral Harbor and Hollywood Beach (Table 4.5). In comparisons of cobble shape and size, I focus on metavolcanic rock cobbles because the lithic assemblage is composed predominantly of that material type thus sample size is large enough conduct statistical analyses comparing lithic assemblage to cobble assemblage characteristics. Table 4.5. Proportions of material types at each cobble area on San Nicolas Island. Cobble Area Metavolcanic Sandstone Quartzite Corral Harbor 497 510 48 Tule Creek 74 1 6 Hollywood Beach 223 291 14 NAVFAC 79 3 3 Thousand Springs 1452 40 98 Radar Row 90 0 6 To investigate whether differences in the composition of glacial marine drift and beach morphology affect the size and shape of cobbles on different beaches on the San Nicolas Island, I use length as a proxy for size and measure several dimensions of shape. Unlike the San Juan Islands cobbles, all metamorphosed cobbles on San Nicolas Island are relatively round. To distinguish between cobbles with different degrees of roundness, I quantify this variable as both a ratio of maximum to minimum circumference to determine the regularity of the shape (higher 156 numbers are more regular) and as an average of the ratio between each dimension (width/length + thickness/length + thickness/width)/3 (Armstrong et al. 2003). Based on this measure, I define ?flat? cobbles as cobbles with roundness values below 0.5 or above 1.5 to and round cobbles as those with roundness values between 0.5-1.5. Cobble flatness is a ratio of length + width/2 x thickness (Wentworth 1922) where higher values represent flatter cobbles. Cobble elongation is a ratio of width x 100/length (Sames 1966). Elongated cobbles have lower values. Results of an ANOVA test comparing mean length, circumference ratio, flatness, and elongation measures for metavolcanic rocks indicate significant differences in cobble size and shape between almost all of cobble areas (Table 4.6, 4.7). Bonferonni post-hoc analyses show which areas are significantly different from one another in mean size and shape measurements (Table 4.8-4.11). Mean cobble size quantified as mean length for Hollywood Beach and Tule Creek (approximately 8-9 cm) are significantly lower than for Thousand Springs and Corral Harbor (approximately 6 cm). For circumference ratio, the major difference is between Thousand Spring which has a higher mean ratio (x = 1.39) and Corral Harbor (x = 1.29), therefore this measure is less useful unless it necessary to distinguish between those two specific cobble areas. For flatness, Corral Harbor has a significantly lower mean flatness value (x = 1.41) and Hollywood Beach has significantly higher mean flatness value (x = 2.61), suggesting a higher abundance of flatter cobbles. This result is also reflected in a Bonferonni post-hoc analysis for elongation. Mean elongation value for Corral Harbor cobbles (x = 96.54) is significantly greater than all other cobble areas. Thus, for metavolcanic cobbles, size varies between cobble areas and shape is different for Corral Harbor, Hollywood Beach, and Thousand Springs cobble areas. 157 Table 4.6. Descriptive statistics for metavolcanic cobbles at six cobble areas on San Nicolas Island. Corral Harbor min max x min max x min max x min max x min max x Quarry Area (m 2 ): 6980 20 300 60.22 0.99 3 1.29 0.29 2.98 0.94 0.27 18 1.41 33.33 300 96.54 Survey Area (m 2 ): 200 ? var. ? var. ? var. ? var. ? var. Total MV cobbles >2 cm:497 2.9 8.4 0.26 0.07 0.28 0.08 1.13 1.28 31.33 981.29 Tule Creek Cobbles min max x min max x min max x min max x min max x Quarry Area (m 2 ): 1545 40 160 86.55 1.05 4.75 1.39 0.42 3.1 0.63 0.24 4 2.17 38.89 112.5 76.76 Survey Area (m 2 ): 100 ? var. ? var. ? var. ? var. ? var. Total MV cobbles >2 cm: 74 2.59 6.68 0.43 0.19 0.31 0.09 0.7 0.49 15.24 232.3 Hollywood Beach min max x min max x min max x min max x min max x Quarry Area (m 2 ): 2810 50 220 96.03 1 3 1.34 0.31 1.1 0.56 0.8 6.33 2.61 37.5 150 75.53 Survey Area (m 2 ): 100 ? var. ? var. ? var. ? var. ? var. Total MV cobbles >2 cm:223 2.74 7.5 0.22 0.05 0.12 0.02 1.02 1.04 15.35 235.68 NAVFAC Cobbles min max x min max x min max x min max x min max x Quarry Area (m 2 ): 3750 50 130 74.56 1 7.5 1.4 0.33 0.82 0.59 1.25 5.75 2.18 33.33 116.67 71.28 Survey Area (m 2 ): 100 ? var. ? var. ? var. ? var. ? var. Total MV cobbles >2 cm: 79 2.01 4.06 0.73 0.54 0.11 0.01 0.77 0.59 15.56 242.06 Thousand Springs min max x min max x min max x min max x min max x Quarry Area (m 2 ): 5740 30 210 62.86 0.82 21 1.39 0.29 4.85 0.64 0.15 11 2.01 14.29 375 77.46 Survey Area (m 2 ): 100 ? var. ? var. ? var. ? var. ? var. Total MV cobbles >2 cm: 1452 2.04 4.18 0.83 0.68 0.2 0.04 0.73 0.53 21.99 483.55 Radar Row min max x min max x min max x min max x min max x Quarry Area (m 2 ):10,130 50 230 89.61 1 1.94 1.39 0.39 2.87 0.61 0.27 3.67 2.17 50 112.5 74.2 Survey Area (m 2 ):50 ? var. ? var. ? var. ? var. ? var. Total MV cobbles >2 cm: 90 2.54 6.43 0.18 0.03 0.26 0.07 0.61 0.38 13.67 186.95 Max. length (mm) Circumference Ratio (mm) Roundness (mm) Flatness (mm) Elongation (mm) 158 Table 4.7. Results of an ANOVA for metavolcanic cobbles on San Nicolas Island comparing means for size and shape measurements. df Mean Square F Sig. Max. Length Between Groups 5.00 2369.62 204.89 0.00 Within Groups 3720.00 11.57 Total 3725.00 Circum.Ratio Between Groups 5.00 0.85 2.43 0.03 Within Groups 3720.00 0.35 Total 3725.00 Flatness Between Groups 5.00 196.30 177.11 0.00 Within Groups 3720.00 1.11 Total 3725.00 Elongation Between Groups 5.00 53710.91 91.27 0.00 Within Groups 3720.00 588.51 Total 3725.00 159 Table 4.8. A Bonferroni post-hoc analysis of ANOVA results comparing mean length for metavolcanic cobbles from six cobble areas on San Nicolas Island. Maximum Length x difference Std. Error Sig. 95% Cofidence Lower Bound Upper Bound Tule Creek Corral Harbor 1.44 0.32 0.00 0.51 2.37 Thousand Springs 2.53 0.31 0.00 1.62 3.44 NAVFAC 1.31 0.45 0.06 -0.02 2.63 Radar Row 0.12 0.41 1.00 -1.09 1.33 Hollywood Beach -2.40 0.33 0.00 -3.36 -1.44 Corral Harbor Tule Creek -1.44 0.32 0.00 -2.37 -0.51 Thousand Springs 1.09 0.13 0.00 0.70 1.49 NAVFAC -0.13 0.35 1.00 -1.17 0.91 Radar Row -1.32 0.30 0.00 -2.20 -0.44 Hollywood Beach -3.84 0.17 0.00 -4.33 -3.35 Thousand Springs Tule Creek -2.53 0.31 0.00 -3.44 -1.62 Corral Harbor -1.09 0.13 0.00 -1.49 -0.70 NAVFAC -1.22 0.35 0.01 -2.25 -0.20 Radar Row -2.41 0.29 0.00 -3.27 -1.55 Hollywood Beach -4.93 0.16 0.00 -5.39 -4.47 NAVFAC Tule Creek -1.31 0.45 0.06 -2.63 0.02 Corral Harbor 0.13 0.35 1.00 -0.91 1.17 Thousand Springs 1.22 0.35 0.01 0.20 2.25 Radar Row -1.19 0.44 0.10 -2.48 0.10 Hollywood Beach -3.71 0.36 0.00 -4.78 -2.64 Radar Row Tule Creek -0.12 0.41 1.00 -1.33 1.09 Corral Harbor 1.32 0.30 0.00 0.44 2.20 Thousand Springs 2.41 0.29 0.00 1.55 3.27 NAVFAC 1.19 0.44 0.10 -0.10 2.48 Hollywood Beach -2.52 0.31 0.00 -3.43 -1.61 Hollywood Beach Tule Creek 2.40 0.33 0.00 1.44 3.36 Corral Harbor 3.84 0.17 0.00 3.35 4.33 Thousand Springs 4.93 0.16 0.00 4.47 5.39 NAVFAC 3.71 0.36 0.00 2.64 4.78 Radar Row 2.52 0.31 0.00 1.61 3.43 160 Table 4.9. A Bonferroni post-hoc analysis of ANOVA results comparing mean circumference ratio for for metavolcanic cobbles from six cobble areas on San Nicolas Island. Circum. Ratio x difference Std. Error Sig. 95% Cofidence Lower Bound Upper Bound Tule Creek Corral Harbor 0.03 0.06 1.00 -0.13 0.19 Thousand Springs -0.04 0.05 1.00 -0.20 0.12 NAVFAC -0.04 0.08 1.00 -0.27 0.19 Radar Row -0.03 0.07 1.00 -0.24 0.18 Hollywood Beach -0.06 0.06 1.00 -0.23 0.11 Corral Harbor Tule Creek -0.03 0.06 1.00 -0.19 0.13 Thousand Springs -0.07 0.02 0.05 -0.14 0.00 NAVFAC -0.06 0.06 1.00 -0.25 0.12 Radar Row -0.06 0.05 1.00 -0.21 0.10 Hollywood Beach -0.09 0.03 0.05 -0.17 0.00 Thousand Springs Tule Creek 0.04 0.05 1.00 -0.12 0.20 Corral Harbor 0.07 0.02 0.05 0.00 0.14 NAVFAC 0.00 0.06 1.00 -0.17 0.18 Radar Row 0.01 0.05 1.00 -0.14 0.16 Hollywood Beach -0.02 0.03 1.00 -0.10 0.06 NAVFAC Tule Creek 0.04 0.08 1.00 -0.19 0.27 Corral Harbor 0.06 0.06 1.00 -0.12 0.25 Thousand Springs 0.00 0.06 1.00 -0.18 0.17 Radar Row 0.01 0.08 1.00 -0.22 0.23 Hollywood Beach -0.02 0.06 1.00 -0.21 0.16 Radar Row Tule Creek 0.03 0.07 1.00 -0.18 0.24 Corral Harbor 0.06 0.05 1.00 -0.10 0.21 Thousand Springs -0.01 0.05 1.00 -0.16 0.14 NAVFAC -0.01 0.08 1.00 -0.23 0.22 Hollywood Beach -0.03 0.05 1.00 -0.19 0.13 Hollywood Beach Tule Creek 0.06 0.06 1.00 -0.11 0.23 Corral Harbor 0.09 0.03 0.05 0.00 0.17 Thousand Springs 0.02 0.03 1.00 -0.06 0.10 NAVFAC 0.02 0.06 1.00 -0.16 0.21 Radar Row 0.03 0.05 1.00 -0.13 0.19 161 Table 4.10. A Bonferroni post-hoc analysis of ANOVA results comparing mean flatness for for metavolcanic cobbles from six cobble areas on San Nicolas Island. Flatness x difference Std. Error Sig. 95% Cofidence Lower Bound Upper Bound Tule Creek Corral Harbor 0.62 0.10 0.00 0.33 0.91 Thousand Springs 0.07 0.10 1.00 -0.21 0.35 NAVFAC -0.12 0.14 1.00 -0.53 0.30 Radar Row -0.03 0.13 1.00 -0.40 0.35 Hollywood Beach -0.92 0.10 0.00 -1.22 -0.63 Corral Harbor Tule Creek -0.62 0.10 0.00 -0.91 -0.33 Thousand Springs -0.55 0.04 0.00 -0.67 -0.43 NAVFAC -0.74 0.11 0.00 -1.06 -0.41 Radar Row -0.65 0.09 0.00 -0.92 -0.37 Hollywood Beach -1.54 0.05 0.00 -1.70 -1.39 Thousand Springs Tule Creek -0.07 0.10 1.00 -0.35 0.21 Corral Harbor 0.55 0.04 0.00 0.43 0.67 NAVFAC -0.18 0.11 1.00 -0.50 0.13 Radar Row -0.10 0.09 1.00 -0.36 0.17 Hollywood Beach -0.99 0.05 0.00 -1.14 -0.85 NAVFAC Tule Creek 0.12 0.14 1.00 -0.30 0.53 Corral Harbor 0.74 0.11 0.00 0.41 1.06 Thousand Springs 0.18 0.11 1.00 -0.13 0.50 Radar Row 0.09 0.14 1.00 -0.31 0.49 Hollywood Beach -0.81 0.11 0.00 -1.14 -0.48 Radar Row Tule Creek 0.03 0.13 1.00 -0.35 0.40 Corral Harbor 0.65 0.09 0.00 0.37 0.92 Thousand Springs 0.10 0.09 1.00 -0.17 0.36 NAVFAC -0.09 0.14 1.00 -0.49 0.31 Hollywood Beach -0.90 0.10 0.00 -1.18 -0.61 Hollywood Beach Tule Creek 0.92 0.10 0.00 0.63 1.22 Corral Harbor 1.54 0.05 0.00 1.39 1.70 Thousand Springs 0.99 0.05 0.00 0.85 1.14 NAVFAC 0.81 0.11 0.00 0.48 1.14 Radar Row 0.90 0.10 0.00 0.61 1.18 162 Table 4.11. A Bonferroni post-hoc analysis of ANOVA results comparing mean elongation for metavolcanic cobbles from six cobble areas on San Nicolas Island. Table 4.12. A ?2 test comparing proportions of ?flat? metavolcanic cobbles (roundness values < 0.5 or > 1.5 to ?round? metavolcanic cobbles (roundness value 0.5-1.5) at San Nicolas Island cobble areas. Elongation x difference Std. Error Sig. 95% Cofidence Lower Bound Upper Bound Tule Creek Corral Harbor -18.19 2.26 0.00 -24.83 -11.55 Thousand Springs -1.92 2.22 1.00 -8.43 4.60 NAVFAC 5.60 3.22 1.00 -3.86 15.07 Radar Row 2.32 2.93 1.00 -6.29 10.93 Hollywood Beach 3.16 2.33 1.00 -3.70 10.01 Corral Harbor Tule Creek 18.19 2.26 0.00 11.55 24.83 Thousand Springs 16.27 0.96 0.00 13.47 19.08 NAVFAC 23.79 2.52 0.00 16.38 31.21 Radar Row 20.51 2.14 0.00 14.22 26.79 Hollywood Beach 21.34 1.20 0.00 17.83 24.86 Thousand Springs Tule Creek 1.92 2.22 1.00 -4.60 8.43 Corral Harbor -16.27 0.96 0.00 -19.08 -13.47 NAVFAC 7.52 2.49 0.04 0.21 14.83 Radar Row 4.24 2.10 0.65 -1.92 10.39 Hollywood Beach 5.07 1.12 1.78 8.36 NAVFAC Tule Creek -5.60 3.22 1.00 -15.07 3.86 Corral Harbor -23.79 2.52 0.00 -31.21 -16.38 Thousand Springs -7.52 2.49 0.04 -14.83 -0.21 Radar Row -3.28 3.14 1.00 -12.51 5.94 Hollywood Beach -2.45 2.59 1.00 -10.06 5.16 Radar Row Tule Creek -2.32 2.93 1.00 -10.93 6.29 Corral Harbor -20.51 2.14 0.00 -26.79 -14.22 Thousand Springs -4.24 2.10 0.65 -10.39 1.92 Thousand Springs 3.28 3.14 1.00 -5.94 12.51 Hollywood Beach 0.84 2.22 1.00 -5.68 7.35 Hollywood Beach Tule Creek -3.16 2.33 1.00 -10.01 3.70 Corral Harbor -21.34 1.20 0.00 -24.86 -17.83 Thousand Springs -5.07 1.12 0.00 -8.36 -1.78 NAVFAC 2.45 2.59 1.00 -5.16 10.06 Radar Row -0.84 2.22 1.00 -7.35 5.68 Flat Expected Round Expected Total ? 2 df p Adjusted Residuals Corral Harbor 42 82.52 455.00 414.48 497.00 156.75 5 <0.001 -5.48 Tule Creek 18 12.29 56.00 61.71 74.00 1.81 Hollywood Beach 97 37.03 126.00 185.97 223.00 11.33 NAVFAC 20 13.12 59.00 65.88 79.00 2.11 Thousand Springs 204 241.10 1248.00 1210.90 1452.00 -4.14 Radar Row 20 14.94 70.00 75.06 90.00 1.46 163 Quarry Attractiveness Surveys of local cobble beaches and outcrops on both the San Juan Islands and San Nicolas Island reveal that in both places, people could have chosen from a number of toolstone collection areas. To examine how territorial circumscription may or may not have influenced procurement choices, I consider the relative value of each quarry in terms of desired toolstone shape and size, collection costs, and transport costs. To facilitate a standardized evaluation of quarry ?attractiveness,? I employ a toolstone attractiveness index (AI) developed by Wilson (2007a,b, Wilson 2011; Browne and Wilson 2011). Wilson?s AI is based on economic research on the attributes that draw people to places or products. Wilson (2007) notes that gravity models can be useful in understanding the sources that ?should? have been used based on geographic and geologic factors, with the assumption that deviations from those predicted patterns should be useful in elucidating the ?human factors? that also influence toolstone choice. Her measure is expressed as a ratio of the quality of the source, extent of the source, and size of the average cobble to the difficulty of the terrain, cost of extraction, and scarcity of the material. Below I define each variable in the formula and discuss how I measure it. I define toolstone quality for a cobble area based mainly on the variety of material available. Sites for which more than one rock type was abundant were ranked a ?3? because this would have allowed people to choose between different tool types to create tools for different purposes. Those where only one rock type was abundant was ranked a ?2?. Cobble areas on San Nicolas Island where soft and friable rocks were abundant I ranked a ?1? because people would have had AI = (toolstone quality) (extent of source) (100) x size (difficulty of terrain) (extraction cost) scarcity 164 to sort through lower quality material to ensure that the toolstone that they selected was hard enough to be flaked. I determined extent of source through GPS mapping in the field and ranked each area from 1-4 based on the range and average size of smaller and larger sites in both study areas ( 1 = 0 = 100 meters 2 ; 2 = 100-2500 meters 2 ; 3 = 2500 ? 10,000 meter2; 4 > 10,000 m2). This variable is important in terms of toolstone reliability. A source is more attractive if people are certain that toolstone will be abundant and easy to find. The cost of extraction variable was given a ?1? in all cases because almost all cobbles can be found loose on beaches or in blowouts and quarrying would have been minimal. Because this variable is the same in all cases, the value is not meaningful. For Wilson?s formula, size and scarcity are considered as a ratio to provide an overall measure of the size of cobbles at a site that is most determined by the most abundant size classes but also incorporates rarer size classes. Size is defined as the maximum dimension (cm) of the modal size. Scarcity is determined by the amount surface area covered by cobbles of that size (1 = > 50% ; 2 = 25-50%; 3 = 5-25%, 4 < 5%). As a baseline, I considered 100% surface area coverage to be one cobble per every square meter of the survey area. First, the modal size/scarcity ratio is calculated. Additional size/scarcity ratios are added in order of decreasing abundance of the size class, but with each ratio, the denominator is raised to a higher power (squared, then cubed, etc.) so that the less abundant size classes are given less weight. To provide an example, the most abundant cobble size class at a cobble area is 6 cm; the second most abundant is 5 cm, and the third most abundant is 4 cm. The 6 cm cobbles are given a ?1? scarcity ranking because they all cover >50% of the surface area of the site. The 5 cm cobbles are given a ?2? scarcity ranking because they cover 25-50% of the surface area of the site. The 4 cm cobbles 165 are given a ?4? scarcity ranking because they cover less than 5% of the surface area of the cobble area. The size/scarcity value is calculated as 6/1 + 5/2 2 + 4/4 3 . To calculate a difficulty of terrain variable, I modify Wilson?s (2007a, b) protocol to estimate travel cost in terms of kcal/hour to account for boat travel and for least cost path distance rather than straight-line distance from site to source for pedestrian travel (Browne and Wilson 2011). Boats change the basics of transport costs associated with lithic procurement in coastal areas by allowing transport of heavy or bulky loads and by facilitating faster and farther travel between islands and the mainland and into the interior (Ames 2002; Arnold and Bernard 2005; Blair 2010; Durham 1060; Fitzhugh and Kennett 2010; Renouf 1988). Ease of boat access to a shellfish bed, fishing stream, or lithic source is limited by beach morphology or the distance a boat can travel up a river or creek. No direct archaeological evidence of watercraft has yet been discovered in the San Juan Islands, but people likely used dugout canoes made from western red cedar logs similar to those recorded historically (Holm 1994; Neel 1995). The earliest palynological evidence for red cedar in the Gulf of Georgia occurs at 6000-7000 BP (Moss et al. 2007) followed by widespread distribution of adzes, wedges, and other woodworking tools after 5000 BP (Hebda and Mathewes 1984). During historic times, Northwest Coast dugout canoes could carry as much as five tons and allowed a foraging radius of 30 km per day (Ames 2002). If toolstone was not directly collected at a site, the most efficient way to transport the material on the islands would be to use canoe. I calculate calories per kilometer in paddling a canoe based an average rate of 400 calories per hour assuming 6.5-8 kilometers an hour (Dillon and Oyen 2008). Canoe travel is more efficient relative to pedestrian travel in the San Juan Islands than the Channel Islands because the wave energy is much lower and vegetation in the inland areas is denser. For the 166 purposes of this study, I assume that whenever possible, people traveled by canoe rather than walking. That people reached the Channel Islands during the Terminal Pleistocene indicates sophisticated watercraft technology from the earliest settlement of the island (Erlandson et al. 2009; Rick et al. 2001b). The emergence of sewn plank canoe technology observed during the historic period has not yet been resolved (Arnold 1995b; Cassidy et al. 2004; Des Lauriers 2005; Fagan 2004; Gamble 2002; Heizer 1938, 1940; Jones and Klar 2005). Arnold (1992, 1995b) describes specialized knowledge required to chop, plane, sew, and caulk boards of redwood or pine that drifted to the islands from northern California, a process requiring over 500 person- hours of labor. In historic times, plank canoes carried up to two tons of cargo at 20 kilometers per hour (see also Hudson et al. 1978:137). Although people likely transported abundant material, including toolstone, by boat both between islands and between different points on the island, I presume that given the high energy wave environment and the long distances between inland sites and the coastline, pedestrian transport was the more efficient way to carry cobbles from inland cobble areas to inland sites and from beach cobble areas to beach sites. To determine travel cost based on pedestrian travel on San Nicolas Island, I create cost surfaces based on a raster digital elevation model using Path Distance and Cost Path tools in ArcGIS ESRI TM with Spatial Analyst extension. Following Wilson?s example, I based round- trip calorie expenditure between cobble areas and archaeological sites based on estimates created by Jones and Madsen (1989). After generating the cost paths, I break the lines into segments and calculate the number of calories burned per segment based on the gradient for both the trip to the source and the return trip. Because denudation of plants and associated erosion substantially changed the topography of the island, I consider the cost paths to represent general estimates of 167 the cost of pedestrian travel between source and site rather than a representation of a path that the Nicole?o might have used (Figure 4.36, 4.37). For CA-SNI-106, I calculate the cost path between the site and the Radar Row cobble area as a representation of the probable distance between the site and an inland cobble area, though not necessary that particular area. I calculate the cost path to Vizcaino point as the nearest beach cobble area even though I was not able to access those cobbles at the time that the fieldwork was conducted. 168 Figure 4.36. Least cost paths between Tule Creek Village and surrounding cobble areas. 169 Figure 4.37. Least cost paths between CA-SNI-106 and surrounding cobble areas. 170 Calculations and Conclusions Based on all of the variables discussed above, by far the most attractive cobble area for precontact inhabitants of the Watmough Bay site would have been the Watmough Bay cobbles (Table 4.13). Despite the small size of the cobbles, the low transport cost should encourage people to take cobbles from the beach adjacent to the site if enough were available. Given the low density of adequate FGV cobbles on the beach, it is possible that families may have exhausted the supply of the toolstone on the beach faster than erosion could have replenished it. If that was the case, the next most desirable location based on these calculations is Aleck Bay followed by Agate Beach. If people at Watmough Bay did not choose Watmough cobbles, it may be because they ran out, or because they desired cobbles of a different size or shape to make their tools. This calculation demonstrates, however, that if during times of increased boundary defense people were constrained in their toolstone procurement opportunities to Watmough Bay beach, this should be a relatively attractive option for them. Transport costs are minimal and slate is available to provide a secondary toolstone source. The other cobble areas are all relatively similar in their AI value. On San Nicolas Island, the most attractive area overall would have been Hollywood beach due to the abundance and variety of material available, the size of the area, and the relatively low cost of travel down a more gentle slope (Table 4.14). If only metavolcanic cobbles are considered, Thousand Springs is the most attractive area due to the density of high quality cobbles. These data suggest that although there are cobbles near Tule Creek Village, if people wanted to procure a larger amount of high quality toolstone, despite the travel costs, it might be more efficient to travel to another point on the landscape to procure cobbles. If people were unable to travel freely, they may have had to rely on cobble areas closer to home. 171 Chapter Summary Data on the material types, size, shape, and attractiveness of the cobble areas on both the San Juan Islands and on San Nicolas Islands provides an essential background for generating predictions about procurement activities in the context of proposed changes in territorial behavior during the Late Holocene on the Pacific Coast. In both study areas, toolstone is ubiquitous locally; however, subtle differences in the characteristics of cobble areas defined peoples? procurement choices and leave a signature of their ?lithic landscape? to determine where they did and did not travel to collect cobbles. Understanding the characteristics of cobble areas provides an insight into how people could have collected cobbles if they could move freely across the landscape and if they were not constrained in their toolstone choices based on their technology. My research explores scenarios for procurement given constraints on movement and the organization of flake-tool technologies. Table 4.13. Wilson?s Attractive Index values for San Juan Islands cobble areas for Watmough Bay. Cobble Area Size/ Scarcity Extent Value Quality Extracti on Cost Distance by Water from Watmough (m) Rate of Calorie Burn Terrain Difficulty (Cal/km) AI tmough Bay 2.09 3 3 1 0.01 0.5 0.5 3766.7 Aleck Bay 2.72 3 2 1 6.42 321 321 5.08 Agate Beach 2.43 3 2 1 10.83 541.5 541.5 2.7 American Camp 1.35 4 2 1 18.07 903.5 903.5 1.2 False Bay 3.06 4 2 1 26.29 1314.5 1314.5 1.86 Snug Harbor 3.54 2 2 1 38.68 1934 1934 0.73 172 Table 4.14. Wilson?s Attractive Index values for San Nicolas Island quarry areas for Tule Creek Village. Cobble Area Siz /Scarcity Size/Scarcity (MV) Extent Quality Quality (MV) Extraction Terrain Difficulty (Cal/km) AI AI (MV) Corral 18.95 4.76 3 3 2 1 481.77 35.4 5.93 Tule Creek 4.26 3.69 2 1 2 1 52.1 16.35 28.33 Hollywood 105.01 19.15 3 2 2 1 537.69 117.18 21.37 NAVFAC 4.12 4.12 3 2 2 1 744.12 3.32 3.32 Thousand 35.22 39.09 3 3 2 1 571.81 55.43 41.02 Radar 7.77 7.32 4 1 2 1 552.54 5.62 10.6 173 Chapter 5: Flaked Stone Technology on the Pacific Coast To test hypotheses regarding change over time in lithic procurement associated with shifts in territoriality on the Pacific Coast, I develop and test predictions for lithic assemblages from archaeological sites on the San Juan Islands and the southern Channel Islands. The predictions center on procurement, processing, reduction strategies, toolstone conservation, and exchange. To develop these predictions, I first establish the steps of the manufacturing process for the lithic technologies at each site. In both study areas, precontact peoples relied primarily on flake tools. This type of technology is often referred to as expedient technology, characterized by minimal core preparation and retouch (Andrefsky 2009; Bleed 1986; Kelly 1988; Nelson 1991; Odell 1998; Parry and Kelly 1987; Teltser 1991). My focus in this chapter is on flaked stone tools, but I also provide data on ground stone tools from each assemblage. I draw on previous lithic studies in each study area and empirical data from the Watmough Bay, Tule Creek Village Mound B, and CA-SNI-106 assemblages to investigate the technological traditions for each site. General Analytic Methods During analysis, each artifact was given a unique number linking the artifact to a site, unit, stratum, level, and/or depth, and artifact type (?f? for flake, ?c? for core, or ?t? for tool). Where more than one artifact type for that unit/level/stratum was present, artifacts were numbered sequentially and bagged together. For example, the first flake analyzed from Watmough Bay, Unit 0N9W from the 150-160 cmbs level was identified as 280.0N9W.150- 160.F.1. The first flake analyzed from Mound B Unit 13 Stratum II Level 2 was identified as 25B.13.2.2.F.1. Where more than one flake, core, or tool is analyzed from a single site/unit/stratum/level/depth, they are distinguishable for future analysis by weight. Details of 174 measurements and identification of attributes for flakes, cores, and tools are noted in Table 5.1. Individual attribute measurements and descriptions were chosen based on lithic studies that demonstrate their utility in identifying different stages of reduction or manufacturing techniques such as bipolar and biface reduction, particularly in combination with one another (e.g., Andrefsky 2009; Close 2006; Dibble 2005; Inizan et al. 1999; Odell 2004; Root 2004; Steffen et al. 1998). Regarding statistical analyses in this chapter and the following chapter, I use ?2 tests to evaluate differences in the distribution of nominal and ordinal variables (e.g., platform type, number of dorsal scars, cortex location) between two populations, in this case sub-samples of an assemblage from two or more time periods. A significant p-value (< 0.05) indicates that the proportions of each class are not randomly distributed given the magnitude of the differences and the sample size of both populations. Adjusted residuals (AR) are used to determine which classes are driving significant differences (Everitt 1977; Grayson and Delpech 2003). To analyze differences in means of ratio variables for assemblages (e.g., length, width, thickness), I use Student?s t-tests (two populations) and ANOVA tests with Bonferroni post-hoc analysis (more than two populations) if sample sizes are similar. Statistically significant differences are demonstrated if p < 0.05. Statistical analyses were completed using SPSS ? 11.5. 175 Table 5.1. Measurement descriptions for lithic analysis of flaked stone tools. All Artifacts Measurement Description Material Type, color, coarse/smooth, glossy/dull Cortex Appearance (rough, smooth), location, approximate % of total possible surface area that could be cortical given artifact type. Size Class Measured for all artifacts using a chart with concentric circles at 1 cm intervals. Artifacts were assigned a size class based on the smallest circle by which they could be circumscribed. Original Cobble Describe shape of original cobble if possible and note which dimensions, if any, indicate the size of the original cobble. Use Extent, location, appearance of chipping or dulled edges Weight Measured for all artifacts (grams). Flakes Measurement Description Platform Cortical, lisse , dihedral, faceted Platform Shape Round vertical (parallel to the direction of force), Round horizontal (perpendicular to the direction of force), flat. Platform lip Present/Absent Platform Angle Interior angle between the platform and ventral face Bulb of Percussion Pronounced/Diffuse Termination Feathered/Hinged/Step/Overpassed Concavity Ventral face concave, convex, or straight Transverse Cross-Section Oval/triangle/semicircle/irregular Dorsal scars Number and orientation (parallel, bidirectional, or mixed). Maximim Length Parallel to the direction of force (if unbroken) (mm). Maximim Width Perpendicular to the direction of force (if unbroken) (mm). Maximim Thickness Perpendicular to maximim length and width (mm). Breakage Orientation of breakage relative to direction of force. Cores Measurement Description Length Parallel to the main axis of flaking (mm). Width Perpendicular to the main axis of flaking (mm). Thickness Perpendicular to length and width (mm). Platform Number, type (lisse, dihedral, multifaceted, cortical, crushed), angle to flaking surface. Removals Number, orientation (unidirectional, bidirectional, multidirectional). Rotation Rotated/Semi-rotated Retouched tools Measurement Description Length Parallel to the retouched, sharp, or pointed end (mm). Width Perpendicular to length. (mm) Thickness Perpendicular to length and width (mm). Origin Tool made on a flake/core. Flake features Features of the platform, dorsal face, termination that are apparent despite retouch/use are recorded and/or measured. Core feature Features of the platform(s), removals that are apparent despite retouch/use are Retouch features Morphology of retouch (scaled/parallel), angle of retouch (abrupt, semi-abrupt, low), direction (obverse/inverse/alternative/crossed/bifacial), extent (short/long/invasive/covering), location and distribution (entire/partial), delineation (straight, convex, concave, sinuous, shoulder). 176 The Watmough Bay Lithic Assemblage I analyzed a total of 2367 flakes (including slate fragments), 242 cores (including fragments), 219 ground stone tools, and 282 formal tools (Table 5.2, 5.3). A majority of the artifacts (63% of the flakes and 56% of the cores) come from the 1600-1000 cal BP time period with smaller samples from the 3500-2500 cal BP, 2500-1600 cal BP, and 600 cal BP-Contact time periods. 177 Table 5.2. Frequencies of flakes and cores from the Watmough Bay site. Site Watmough Bay Watmough Bay Watmough Bay Watmough Bay Watmough Bay Time Period 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Undated Flakes Coarse-grained volcanic 2 4 17 0 7 Chert 3 0 10 0 3 FGV 78 92 753 50 370 Schist 10 3 39 0 25 Slate 68 7 582 41 72 Argillite 0 1 0 0 2 Quartz 0 2 70 3 10 Quartzite 0 1 7 0 2 Metasedimentary 0 0 2 0 1 Metavolcanic 0 0 2 0 1 Nephrite 0 0 2 0 0 Sandstone 0 0 1 0 0 Unknown 4 0 11 0 4 Total 179 110 1496 94 496 Cores (FGV) Unworked 3 0 0 0 2 Unidirectional 1 0 6 0 4 Split/Tested 4 0 6 1 4 Exhausted 1 1 3 0 2 Bipolar 1 5 6 0 3 Multidirection Unpatterned 1 1 15 1 5 Flaked flake 0 1 3 2 0 90 Degree 0 1 7 1 7 Fragment 0 8 45 2 11 Chopper 0 0 4 1 0 Rotated 0 0 5 0 6 Opposed Platform 0 1 6 0 1 Unidentifable 0 0 3 0 2 Total 11 18 109 8 47 Cores (Chert) Multidirection Unpatterned 0 0 1 0 0 Fragment 1 2 0 0 0 Cores (Quartz) Unworked 0 0 1 0 0 Fragment 0 0 17 0 1 Cores (Coarse-grained volcanic) Split/Tested 0 0 1 0 1 Bipolar 0 0 1 0 0 Fragment 0 0 1 0 0 Chopper 0 0 1 0 0 Cores (Nephrite) Unworked 0 0 1 0 0 Cores (Quartzite) Unidirectional 0 0 1 0 0 Fragment 0 0 0 2 0 Cores (Slate) Fragment 0 0 0 0 1 178 Table 5.3. Basic statistics on tools and groundstone from the Watmough Bay site. Material types is FGV unless noted otherwise. A comparison of the Watmough Bay assemblage to published analyses of the English Camp lithic artifacts (Close 2006, 2011; Kornbacher 1989, 1992) and an intra-site comparison of material excavated in 1968 by Munsell and Stein and Phillips in 2004 indicates that the Munsell Site Watmough Bay Watmough Bay Watmough Bay Watmough Bay Watmough Bay Time Period 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Undated Tools (FGV unless noted) Scrapers 1 6 (1 chert) 18 (2 chert) 0 12 (1 chert) Retouched Flakes 1 1 22 (1 chert, 1 shale, 1 quartzite, 1 slate) 0 6 Choppers 0 1 5 0 3 (1 quartzite) Knives 0 2 (slate) 12 (6 slate, 5 schist) 0 14 (7 slate, 4 schist, 3 metasedimentar y) Pointed Tools 0 1 10 0 2 Hammerstones 0 0 2 0 0 Bifaces 1 (CGV) 2 (1 undet.) 26 (2 undet., 3 schist, 2 slate, 2 CGV) 2 4 (1 slate) Leaf-Shaped Points 0 0 1 0 0 Projectile Points 0 0 6 (1 undet.) 1 2 (1 slate) Stemmed Points 0 0 4 (1 slate) 0 1 Corner-notched poin 0 0 0 0 1 (chert) Triangles 0 0 7 1 2 Notches 0 0 2 0 3 (1 slate, 1 CGV) Scaled Pieces 3 12 (1 undet.) 46 (1 slate) 0 37 (2 CGV, 3 slate, 1 schist, 1 quartzite) Total 6 25 161 4 87 Bead/Ornament 1 0 2 0 5 Net weight 0 1 15 0 11 Incised shale/other 0 1 31 1 7 Unmodified shale/other 0 6 70 0 22 Misc. Ground stone 0 0 0 0 3 Abrader 0 0 3 0 1 Adze/Axe 0 0 3 0 1 Labret 0 0 0 0 1 Flaked slate/shale/other 0 0 2 0 0 FMR 0 2 11 0 6 Grinding stone 0 0 1 0 0 Pecked Stone 0 0 2 0 2 Point 0 0 2 0 2 Vessel Fragment 0 0 1 0 0 Hammerstone 0 0 0 0 0 179 excavators did not consistently collect small flakes (< 10 mm in maximum dimension) and flakes without platforms. Considering only FGV, the Stein/Phillips assemblage contains a higher proportion of small flakes (n = 23,16%) than the Munsell excavation (n = 4, 0.3% ). The Stein/Phillips excavation contains a higher proportion of FGV flakes without platforms (n = 117, 81.3%) compared to the Munsell excavation (n = 686, 57.2%). The ratio of FGV flakes to cores also differs substantially between the two Watmough assemblages and the English Camp OpD (45-SJ-24) assemblage. At OpD, there is a ratio of 5387 flakes to 83 cores including unworked cobbles (64.9 flakes per core). The Munsell assemblage contains 1200 FGV flakes and 169 cores (7.10 flakes per core), and the Stein excavation contains 144 flakes and 10 cores (4.4 flakes per core). Differences in core/flake ratios between OpD and Watmough Bay could be the result of differences in the manufacturing process between the sites; however, I suspect that flakes are under-represented at Watmough Bay. The field notes on the Watmough Bay excavation do not describe a process by which flakes were selected or discarded in the field by the excavation team. Difference in the amounts of slate and schist from the two Watmough excavations also suggests that the Munsell excavation team collected less of the slate they encountered than the Stein/Phillips excavation team. The Munsell excavation contains 333 slate fragments of 1650 total flakes (2%). The Stein/Phillips excavation contains 517 slate fragments of 717 total flakes (72%). The Stein/Phillips excavation is closer to the slate outcrop and many of these fragments were deposited naturally into the shell midden, but different collection strategies likely contributed to such a large difference in the proportion of slate between the two areas of the site. I take these factors into consideration in my analysis. 180 In analyzing the debitage, I was guided by two published analyses of San Juan Islands lithic assemblages from the English Camp shell midden site. Kornbacher (1989,1992) provides an analysis of lithic artifacts from the Operation A (OpA) assemblage that date to ca. 1700 cal BP ? Contact (Stein et al. 2003). Kornbacher concludes that expedient flake tool technology is the primary goal of lithic manufacture at the site. She tests the hypothesis that a decrease in abundance of chipped stone artifacts after 1000 cal BP is associated with a shift in lithic technology rather than just a decrease in the use of flaked stone tools. Despite a decrease in microblades and an increase in tool classes, similarities in flake attributes before and after 1000 cal BP do not support an overall change in manufacturing practices. Kornbacher (1992) proposes a hypothetical model of cobble reduction where a first flake is removed from the edge of a cobble and subsequent removals widen the flaking surface so that the remainder of the cobble can be flakes with minimal cortex on the dorsal face and margins of the flake. Close (2006, 2011) presents a cha?ne op?ratoire approach to technological organization at English Camp Operation D (OpD) by investigating the social context of the entire life history of the artifacts from raw material procurement, production (creation of blanks, selection and shaping of blanks for retouched tools, selection of unshaped blanks for scaled pieces), and management (use, maintenance, discard). The site dates to ca. 2000-1000 cal BP with most dates clustering at 1550-1250 cal BP (Stein et al. 2003). Close finds that a major emphasis of the lithic technology at this site was to create multipurpose flake tools with cortical backs, some that show splintering from use (scaled pieces) and others that do not (naturally backed flakes). Close (2006:18) distinguishes scaled pieces from cores by removals that are too small to be used as tools. She distinguishes them from retouched flakes based on the presence of removals that were not intended to shape the flake. People at OpD made triangular points in a spatially separate area 181 of the site from the flaked tools. The two different technological trajectories may indicate differences in manufacturing and tool use activities based on gender. Based on Close?s analysis, lithic procurement at OpD centered on small irregularly- shaped beach cobbles (Close refers to these as pebbles). The manufacturing process began with the ?decapitation? of a corner of cobble using a natural edge to channel the force of a hard hammer percussion blow to create a long flaking surface. First flakes are thin, cortical flakes with a relatively high maximum length/maximum width ratio (Close 2006:161). Subsequent flakes were struck from a cortical platform adjacent to the first flake scar. The high frequency of secondary and tertiary flakes with wide cortical platforms and/or cortical margins, many of which are opposite non-cortical margins that show extensive wear, suggest that knappers intentionally set up cores to create flakes that retain cortex on one edge (Figure 5.1, 5.2). When one platform was exhausted, the core could be rotated 180 degrees to exploit an additional cortical platform and repeat the process described above. Close also suggests that people made triangular projectile points on non-cortical flakes with feathered terminations. The thicker proximal end was shaped to a point by retouch but additional retouch shaping on the rest of the flake was usually minimal. The bases of most triangles were often thin and asymmetrical with dulled corners. Close describes, explores, and verifies this manufacturing process using extensive description and quantitative analysis. 182 Figure 5.1. FGV flake from an angular cobble from Watmough Bay, obverse (280.1N9W.40-60.F.2). Figure 5.2. FGV flake from an angular cobble from Watmough Bay, reverse (280.1N9W.40-60.F.2). The manufacturing process at Watmough Bay appears to be similar to English Camp as described by Close rather than Kornbacher. Based on the uniform high quality of the FGV and the presence of 17 tested (one flake removed) cobbles, raw material was sometimes tested at the site. Given the low number of first flakes (n = 32) per the total number of cores (n = 242), some cobbles may have been tested where they were collected. Cobble surveys indicate that both round and angular cobbles were available on the beaches of the San Juan Islands, but people 183 more often chose angular ones. A total of 11 of the cores at Watmough Bay are made on smooth round waterworn cobbles while 48 are identified as angular; the reminder could not be matched to a cobble type. That the mean length/width ratio is significantly higher for flakes with cortical platforms and no dorsal flake scars (x 1.26) and flakes with cortical platforms and 4+ dorsal flake scars (x 0.94) suggests that people made an effort to create a longer than average flakes to begin the reduction sequence (Table 5.4). Some of the cores are made on waterworn cores from earlier deposits. Table 5.4. Results of an ANOVA test comparing length, width, thickness, length/width ratio by flake type from the Watmough Bay assemblage, FGV only. Flake type is determined by platform type (1 = lisse, 4 = cortical) and number of dorsal flake scars. After the decapitation flake was removed from the cobble, flintknappers at Watmough Bay removed additional flakes in a variety of ways based on the shape, size, and workability of the cobble as well as the desired end product (Figure 5.3). The most common strategy was to flake the core from multiple platforms in an unpatterned manner (Table 5.2, Figure 5.4-5.9). There are a total of 28 cores of this type of which 4 are exhausted (Figure 5.10-5.12). There are also 14 multiple-platform cores where flakes were removed mainly at 90 degree angles from one another. Secondary and tertiary flakes are removed from flat cortical platforms adjacent to the initial flaking surface. There are 18 cases of cortical platforms and 18 cases of single facet P atform Single Single Single Single Cortical Cortical Cortical Cortical F Sig. Bonferonni Sig. Do sal Scars 0 1 2,3 4+ 0 1 2,3 4 Flake Type 1.0 1.1 1.2 1.4 4.0 4.1 4.2 4.4 Length n 12.00 48.00 73.00 54.00 27.00 64.00 80.00 43.00 3.10 0.00 1.2 and 1.4 - 0.04 Length x 39.01 32.06 30.70 36.38 38.04 33.06 32.83 37.85 Width 10.00 40.00 66.00 53.00 27.00 65.00 72.00 46.00 4.06 0.00 1.2 and 4.2 - 0.00; Width x 30.79 30.40 27.30 32.32 31.79 33.61 35.18 37.35 1.2 and 4.4 - 0.00 Thickness n 13.00 50.00 79.00 61.00 30.00 72.00 87.00 50.00 2.15 0.04 Thickness x 11.26 8.83 8.73 10.31 10.60 9.42 9.98 10.91 Length/Width n 10.00 38.00 63.00 47.00 24.00 60.00 66.00 39.00 2.59 0.01 4.2 and 4.0 - 0.04 Length/Width x 1.31 1.21 1.18 1.17 1.32 1.06 0.99 1.09 184 platforms on the cores. The small number of flakes with a single facet platform and 100% dorsal cortex (n = 15) suggests that it was more common to use a non-cortical flaking surface later on in the reduction sequence. At some point, the core was rotated or partially rotated, the core was decapitated again, and the flintknapper began flaking from the new flaking surface maintaining a cortical margin. Another common flaking strategy at Watmough Bay was to decapitate or split a cobble and remove flakes from a cortical platform all around the flaking surface, producing a rotated multi-directional core (n = 12). This technique creates flakes with wide cortical platforms (Figure 5.13-5.18). Unidirectional cores (n = 12) are similarly designed but flakes are removed from a flat cortical surface or flaked surface around the outside of a cobble along the same axis (Figure 5.19-5.20). 185 Figure 5.3. Proposed manufacturing process for angular FGV cobbles at Watmough Bay. The white circular object attached to the rock is a barnacle. 186 Figure 5.4. Photograph of a multi-platform unpatterned FGV core from Watmough Bay, obverse (280.BalkF.40.C.1). Figure 5.5. Photograph of a multi-platform unpatterned FGV core from Watmough Bay, reverse (280.BalkF.40.C.1). Figure 5.6. Plan-view schematic of a multi-platform unpatterned FGV core from Watmough Bay (280.BalkF.40.C.1). 187 Figure 5.7. A multi-platform 90 degree FGV core from Watmough Bay, obverse (280.0N9W.50-70.C.2). Figure 5.8. A multi-platform 90 degree FGV core from Watmough Bay from Watmough Bay, reverse (280.0N9W.50-70.C.2). Figure 5.9. Plan-view schematic of a multi-platform 90 degree FGV core from Watmough Bay (280.0N9W.50-70.C.2). 188 Figure 5.10. An exhausted multiple platform FGV core from Watmough Bay, obverse (280.0N24W.40-60.C.1). Figure 5.11. An exhausted multiple platform FGV core from Watmough Bay, reverse (280.0N24W.40-60.C.1). Figure 5.12. Plan-view schematic of a an exhausted multiple platform FGV core from Watmough Bay (280.0N24W.40-60.C.1) 189 Figure 5.13. A FGV rotated core from Watmough Bay, obverse (280.0N15W.130.C.1). Figure 5.14. A FGV rotated core from Watmough Bay, reverse (280.0N15W.130.C.1). Figure 5.15. Plan-view schematic of a FGV rotated core from Watmough Bay (280.0N15W.130.C.1). 190 Figure 5.16. A FGV rotated core from Watmough Bay, obverse (280.0N0E.Oversize.280- 300.C.1). Figure 5.17. A FGV rotated core from Watmough Bay, reverse (280.0N0E.Oversize.280- 300.C.1) Figure 5.18. A plan-view schematic of a FGV rotated core from Watmough Bay (280.0N0E.Oversize.280-300.C.1). 191 Figure 5.19. Photograph of a unidirectional FGV core from Watmough Bay, obverse (280.3S0E.180-200.C.2). Figure 5.20. Photograph of a unidirectional FGV core from Watmough Bay, reverse (280.3S0E.180-200.C.2). Round cobbles were flaked using opposed platform or chopper core strategies. For a bipolar opposed platform technique, the rock was placed on an anvil and force was applied from above, often splitting the cobble and/or flaking it on two faces (Figure 5.21, 5.22). Battering is typically present on both ends of these cobbles on their longer axis. Some of the flatter angular cores were also reduced using this technique. For the ?chopper? strategy, the cobble was decapitated on one end and then that flake scar was used as a platform for a second flake creating an acute edge (Figure 5.23, 5.24). 192 Figure 5.21. Opposed platform Figure 5.22. Opposed platform core on a round core on a round waterworn cobble, waterworn cobble, reverse. obverse (280.12S0E.50.C.1). Figure 5.23. Chopper core on a round Figure 5.24. Chopper core on a round waterworn FGV cobble (280.0N0E.30. waterworn FGV cobble. C.1). 193 Along with FGV cores at Watmough Bay, there are also a small number of quartz cores (n =18). All are unpatterned chunks or fragments therefore some may just be shatter. There are also quartzite cores (two fragments and one unidirectional core), and chert cores (two fragments, 1 exhausted, and 1 multi-directional unpatterned core). There is also one unworked nephrite cobble at the site. A size comparison of Watmough Bay cores and cobbles from Watmough Bay beach and the other cobble areas that I surveyed on the San Juan Islands indicates that people at Watmough Bay preferred larger cobbles than those found on the beach adjacent to the site. For FGV cores with two opposite cortical surfaces, it is possible to measure a dimension of the original cobble (intact cobble dimension or ICD). The ICD is equal to or smaller than the maximum dimension of the original cobble, therefore mean ICD for an assemblage should underestimate mean cobble size for the cobbles from which the cores were made . Results of an ANOVA comparing mean ICD for Watmough cores and cobble lengths for surveyed beaches on the San Juan Islands indicates a significantly larger ICD for Watmough cores compared to cobble lengths for Watmough Bay beach (Table 5.5). This suggests that either people picked the largest cobbles on the beach or that they often went elsewhere to collect toolstone. Table 5.5. Results of an ANOVA and Bonferroni post-hoc analysis comparing ICD for Watmough Bay FGV cores and length dimensions for FGV cobbles from six cobble beaches on the San Juan Islands (F 39.42 Sig. 0.00). For 48 cores that had an I x 50.17. n x Mean Difference Std. Error Sig. Watmough ICD American Camp 36 44.47 5.70 3.59 1.00 Watmough 64 37.13 13.05 3.11 0.00 False Bay 94 69.73 -19.56 2.89 0.00 Snug Harbor 130 51.34 -1.17 2.75 1.00 Agate Beach 70 58.39 -8.21 3.05 0.15 Aleck Bay 30 76.77 -26.60 3.79 0.00 194 As at OpD, it appears that the primary goal of lithic manufacture at Watmough Bay was to create flake tools and scaled pieces which have invasive splintering and crushing caused by use (Figure 5.25-5.28). People also made scrapers with unifacial retouch designed to create a steep edge (Figure 5.29, 5.30), and other intentionally retouched flakes such as pointed and notched tools. Some scaled pieces were later retouched and some flakes were retouched, used as scrapers, and also used in a manner that caused bifacial splintering and crushing. Because of the overlap between these tool categories, I considered all scaled pieces and scrapers within a single framework of ?flake tool? and recorded the nature and degree of use and retouch (e.g., unifacial or bifacial, one or more margin, continuous or discontinuous, short or invasive). A total of 232 flakes show macroscopic evidence of damage on at least one margin (Figure 5.31, 5.32), but I cannot determine without microscopic analysis and/or residue analysis if that damage is caused by use, manufacture, or post-depositional processes. Because all FGV flake tools are at least 40 mm in maximum dimension, I did not include flakes smaller than 40 mm in maximum dimension in ?2 tests comparing flake tools and flakes to determine preferences for creating flake tools. People at Watmough Bay did not consider FGV flakes smaller than approximately 40 mm in maximum dimension to be useful for creating tools. 195 Figure 5.25 FGV scaled piece from Figure 5.26. FGV scaled piece from Watmough Bay, obverse (280.3S0E.80.T.1). Watmough Bay, reverse. Figure 5.27. FGV Scaled piece from Figure 5.28. FGV Scaled piece from Watmough Watmough Bay, obverse Bay, reverse. (280.0N15W.34.T1). 196 Figure 5.29. FGV scraper from Watmough Figure 5.30. FGV scraper from Watmough Bay, obverse (280.9N3W.32.T.1). Bay, reverse. Figure 5.31. FGV flake showing edge Figure 5.32. FGV flake showing edge damage from Watmough Bay, obverse damage from Watmough Bay, reverse. (280.0N9W.146.T.1). 197 The relationship between cortical margins and use observed by Close at OpD is present at Watmough Bay but is not statistically meaningful. For FGV scaled pieces, 44 of 83 scaled pieces have a cortical margin either opposite from or adjacent to the scaled edge. In most cases, the cortical edges are either adjacent or adjacent and opposite to the cutting or scraping end of the tools. In other cases the cortical margins are opposite from the cutting or scraping end of the tool (Table 5.6). A ?2 test comparing proportions of FGV flakes > 40 mm in maximum dimension with cortical margins with FGV scaled pieces with cortical margins does not show a statistically significant difference (?2 = 3.37, 1 df, p = 0.07) in cortical and non-cortical margins. This does not support the hypothesis that people specifically chose flakes with cortical margins to create scaled pieces. For FGV scrapers, 19 of 17 have a cortical margin either opposite from or adjacent to the retouched edge (Table 5.6). A ?2 test comparing proportions of FGV flakes > 40 mm in maximum dimension with cortical margins with FGV scrapers with cortical margins does not show a statistically significant difference (?2 = 1.21, 1 df, p = 0.27) in cortical versus non-cortical margins, which does not support the hypothesis that people specifically chose flakes with cortical margins to create scrapers. Scaled pieces and scrapers are statistically similar in incidence of cortical margins (?2 = 2.58, 1 df, p = 0.12), and similar in incidence of opposite and adjacent cortical margins (?2 = 0.20, 1 df, p = 0.65). Table 5.6. Location of cortex and location of use-wear or retouch for FGV flake tools from Watmough Bay. Ajacent Opposite Same Undet. Adjacent/ Opposite Adjacent/ Same Opposite/ Same All Margins Total Scaled Pieces 24 15 1 1 2 0 1 0 44 Retouched Flakes 10 4 1 0 3 0 0 0 18 Scrapers 11 2 1 0 5 0 0 0 19 Cortex Location Relative to Use/Retouch 198 Scaled pieces made from FGV do not have significantly more dorsal flake scars than > 40 mm FGV flakes (?2 = 2.94, 1 df, p = 0.09). There were no significant differences in platform facet between non-used, used flakes and scaled pieces (?2 = 3.52, 1 df, p = 0.06). Flakes with both cortical and non-cortical platforms were used to make these tools. Only a handful of flakes with dihedral or multi-faceted platforms were found in the assemblage. Scaled pieces had significantly more round vertical platforms (platforms that are curved longitudinally rather than flat) than expected given the ratio for non-used flakes although sample size is small (?2 = 36.15, 1 df, p = 0.00 (Table 5.7). Table 5.7. Results of a ?2 test comparing proportions of round and flat platforms for FGV flakes and scaled pieces at Watmough Bay. Results of a several t-tests indicate that the people at Watmough Bay selected sturdier flakes for scaled pieces and wider flakes for minimal retouch. Scaled pieces have significantly greater mean lengths and thicknesses than > 40 mm unmodified FGV flakes. Retouched flakes (excluding perforators and notched tools) have significantly greater mean widths than unmodified flakes. Scrapers are not significantly different from the unmodified flakes in any dimension. For length and width, this is potentially because extensive retouch decreases the length and width of the original flake, although thickness should not be affected (Table 5.8) Artifact Type (FGV) Stats Round Vertical Flat ?? DF p Sca ed Pieces Count 18 14 36.15 1 <0.001 Expected 5.57 26.43 AR 6.01 -6.01 Flakes > 40 mm Count 61 361 Expected 73.43 348.57 AR -6.01 6.01 Platform 199 Along with the flake tools, the Watmough assemblage also contains bifaces (n = 34), shapes flakes or cores that are flaked on both faces from parallel opposing axes (Kelly 1988:718) (Figure 5.33, 5.34) and projectile points (n = 26). Along with unidentifiable tips and fragments (n = 9) points types include one willow leaf-shaped point, one corner-notched point, stemmed points (n = 5, Figure 5.35, 5.36) and triangles (n = 10, Figure 5.37, 5.38). Because the projectile points have been heavily worked, I had difficulty determining the kinds of flakes that were preferred to make these artifacts. FGV is the most common material type with three points/bifaces made from coarser grained volcanic rock, 1 made from chert, and 11 made from slate, schist, or other metasedimentary material that do not facture conchoidally. Cortex is present on only a few bifaces. The FGV bifaces are significantly wider than the average unworked > 40 mm FGV flake, suggesting that people picked wider flakes for use (Table 5.8). Length and thickness were not significantly different for FGV flakes and bifaces, but this may be because people chose thicker and longer-than-average flakes that were thinned and reduced in size through bifacial thinning and shaping. Figure 5.33. Two FGV bifaces from Figure 5.34. Two FGV bifaces from Watmough Bay, obverse (left - 280.0N0E. Watmough Bay, reverse. 25.T.1, right ? 280.0N0E.45.T1). 200 Figure 5.35. FGVwillow leaf-shaped point (left -280.0N15W.106.T.1) and stemmed point (right - 280.9N24W.70.T.1 ) from Watmough Bay, obverse. Figure 5.36. FGV willow leaf-shaped point (left -280.0N15W.106.T.1) and stemmed point (right - 280.9N24W.70.T.1 ) from Watmough Bay, reverse. 201 Figure 5.37. FGV triangular points from Watmough Bay, obverse (left ? 280.0N0E.100.T.1, middle ? 280.2S0E.70.T.1, right ? 280.0N18W.36.T.1 ) Figure 5.38. FGV triangular points from Watmough Bay, reverse (left ? 280.0N0E.100.T.1, middle ? 280.2S0E.70.T.1, right ? 280.0N18W.36.T.1 ) 202 Table 5.8. Results of independent samples t-tests (equal variances not assumed) comparing mean length, width and thickness for > 40 mm unmodified FGV flakes, FGV scrapers, and FGV scaled pieces at Watmough Bay. Along with FGV flaked stone manufacture, the Watmough Bay site contains evidence of flaked and ground slate technology. Large pieces of slate could be removed from an outcrop on the southeast end of the beach or at the base of the cliff. Slate bifaces (n = 5), chipped slate unifaces (n = 14) (Figure 5.39, 5.40), and ground slate tools such as net weights, adzes, and points are present at the site. There are also a number of rough unifaces (n = 9) and bifaces (n = 2) made from schist, a material that is coarser-grained and more durable than slate and also Dimension Flakes Scrapers F Sig. t df Sig. (2-tailed) Length n 444.00 33.00 0.06 0.51 -1.75 37.2 0.09 Length x 36.33 39.63 Width n 500.00 33.00 0.13 0.72 0.291 37.6 0.77 Width x 33.32 32.80 Thickness n 908.00 33.00 2.31 0.13 -1.97 33.66 0.06 Thickness x 9.69 11.39 Dimension Flakes Scaled Pieces F Sig. t df Sig. (2-tailed) Length n 444.00 87.00 7.77 0.01 -2.21 105.66 0.03 Length x 36.33 39.92 Width n 500.00 87.00 0.57 0.05 -2.18 117.21 0.03 Width x 33.32 36.18 Thickness n 908.00 87.00 21.26 0.00 -4.12 94.48 0.00 Thickness x 9.69 12.29 Dimension Flakes Retouched flakes F Sig. t df Sig. (2-tailed) Length n 444.00 25.00 4.07 0.04 -1.03 25.47 0.31 Length x 36.33 39.41 Width n 500.00 25.00 3.31 0.07 -2.56 25.51 0.02 Width x 33.32 40.96 Thickness n 908.00 25 5.46 0.02 -0.74 24.71 0.47 Thickness x 9.69 10.53 Dimension Flakes Bifaces F Sig. t df Sig. (2-tailed) Length 444.00 22 0.249 0.62 -0.91 23.88 0.37 Length x 36.33 38.19 Width n 500.00 22 2.34 0.13 2.28 25.29 0.03 Width x 33.32 33.32 Thickness n 908.00 22 4.23 0.04 0.67 23.4 0.05 Thickness x 9.69 9.69 203 probably available locally. On the Northwest Coast, unifacially or bifacially chipped and/or ground slate fragments are often called knives. They are associated with fish processing due to their thin edges and the ease of resharpening (Hayden 1989; Graesch 2007). Byproducts of slate knife production are small chips that would slide through ?? mesh (Graesch 2007:583), therefore evidence of slate processing at the Munsell excavation was likely only partially recovered. Graesch?s (2007) reduction sequence for slate provides a useful baseline for the artifacts at Watmough Bay. Slate fragments could be removed from the bedrock using hard hammers or they could be collected from the beach. Further thinning could be accomplished using bipolar or hard hammer percussion. These fragments can be used without further chipping or they can be chipped along the margins using soft-hammer percussion or pressure flaking to thin and shape the fragile edge. Some slate knives from Northwest Coast assemblages were ground on their edges using sandstone abraders because a more regular edge would break and chip less easily during use. Among the Coast Salish during the post-contact period, slate knives were sometimes hafted (Barnett 1955). Figure 5.39. Slate knife, obverse Figure 5.40. Slate knife, reverse. (280.1.5N0E.175-180.T.1). 204 San Nicolas Island Lithic Analysis For this research I analyzed a total of 4597 flakes (including ground sandstone fragments), 114 cores, and 168 retouched and groundstone tools from the Tule Creek Village site (CA-SNI-25, Mound B). At CA-SNI-106, I analyzed a total of 474 flakes, 1 retouched tool, and 6 cores (Table 5.9). Similar to the lithic technology of the San Juan Islands, San Nicolas Island lithic technology centers primarily on expedient tools. Large metavolcanic and metasedimentary cobbles are split or decapitated and flaked to create flake tools, some of which show damage to their margin (n = 217 at Mound B, n = 6 at SNI-106). Many flakes from Mound B were probably used but do not show evidence of use damage because of the hardness and coarse grain of the material. The Nicole?o also made retouched scrapers, choppers, drills, reamers, and bifaces from local toolstone and from off-island chert. In addition to flaked stone, the Mound B and SNI-106 assemblages contain ground stone tools and debris such as sandstone fragments from vessels, saws, pestles, and other objects, and steatite and serpentine beads and ornaments. Other lithic materials include angular, discolored, and potlidded fire-affected rock (FAR), tarring pebbles that were used to apply asphaltum to baskets and other items, iron concretions, calcite, waterworn pebbles, and a variety of other miscellaneous rocks. All stone material at Mound B and SNI-106 was brought there by people; no rock occurs naturally on the sandy dunes where the sites are located. This research focuses primarily on the flaked stone assemblage but I also present data on the other lithic materials for Mound B (Table 5.10). 205 Table 5.9. Descriptive data for the Mound B and SNI-106 flakes and cores. Site Tule Creek Tule Creek Tule Creek Tule Creek Tule Creek CA-SNI- 106 CA-SNI- 106 CA-SNI- 106 Time Period 5000-4000 BP 3000-1500 BP 1500-500 BP 500 BP- Contact Undated 3000-1500 BP 1500-500 BP Undated Flakes Metavolcanic 81 144 452 971 1685 113 225 35 Metasedimentary 4 11 18 56 63 0 0 0 Quartzite 14 34 21 Quartzite 14 18 85 171 216 34 21 3 Sandstone 7 2 34 51 173 0 1 0 Exotic Chert 9 8 15 40 63 3 2 0 Obsidian 0 0 0 0 1 0 0 1 Island Chert 1 0 2 1 5 0 1 0 Quartz 1 12 10 47 87 20 12 2 Total 131 195 616 1337 2293 204 283 41 Cores (Metavolcanic unless noted) Unworked Cobble 1 (quartzite) Tested Cobble 1 1 0 0 0 0 0 Split Rotated 0 0 0 7 (1 quartzite) 9 (2 quartzite) 0 0 Diagonal Cores 0 1 1 1 (sandstone) 7 (1 quartzite) 0 0 Decaptation Cores 0 0 1 1 2 0 0 Exhauste Cores 1 0 4 (1 quartzite) 8 (2 sandstone) 12 (1 quartzite, 1 metasedimen tary, 1 quartz) 1 0 Cores on Flakes 0 0 1 1 6 (1 sandstone) 0 0 Core Fragments 0 0 15 (1 sandstone, 1 quartzite, 1 quartz) 9 (5 sandstone, 1 chert, 1 quartzite) 21 (4 quartzite, 4 sandstone) 4 (1 quartzite, 1 sandstone) 1 206 Table 5.10. Descriptive data for the Mound B and SNI-106 formal tools and other miscellaneous stone artifacts. Site Tule Creek Tule Creek Tule Creek Tule Creek Tule Creek CA-SNI- 106 CA-SNI- 106 CA-SNI- 106 Time Period 5000-4000 BP 3000-1500 BP 1500-500 BP 500 BP- Contact Undated 3000-1500 BP 1500-500 BP Undated Tools (Metavolcanic unless noted) Damaged Flakes 4 3 33 59 118 1 5 Retouched Flakes 2 4 0 4 (1 chert, 1 quartzite) 6 (1 chert) 0 0 Drill 2 1 15 (1 metased., 1 chert, 3 quartz, 2 quartzite, 1 sandstone) 12 (2 chert, 2 quartzite) 15 (3 chert, 1 silicious shale, 2 quartzite, 2 quartz) 0 0 Biface/Point 0 2 (1 quartzite, 1 chert) 2 (1 chert) 2 (1 chert, 1 quartz) 11 (2 quartz, 2 chert) 0 0 Scraper/Knife 0 1 11 (1 quartzite) 14 (3 metasediment ary, 1 quartzite) 6 (4 chert, 1 quartzite) 0 1 Reamer 0 0 1 0 5 (4 sandstone, 1 quartzite) 0 0 Chopper 0 0 0 0 5 (1 quartzite) 0 0 Ground Stone 1 0 0 2 (1 sandstone, 1 serpentine) 3 0 0 Bead/Ornament 0 1 (quartz crystal) 3 (steatite) 3 (2 steatite) 11 (6 serpentine, 5 steatite) 0 0 Saw Fragment 0 0 1 (sandstone) 0 8 (sandstone) 0 0 Other tool type 0 1 2 0 6 0 0 FAR 8 46 39 242 287 Tarring Pebbles 6 2 8 95 166 207 In contrast to lithic research on the Northern Channel Islands that investigates microblades (Arnold 1990b; Arnold et al. 2001) and bifaces (Cassidy 2004; Pletka 2001), previous research on flaked stone tools on San Nicolas Island explores expedient flake tool technology (but see Rosenthal 1996). Clevenger?s (1982) thesis research on the lithic assemblage at CA-SNI-11 investigates reduction strategies and cobble sources on the island. She describes a split round cobble reduction strategy with a continuum of flexible actions performed to create flake tools used without further modification. Ta?kiran (2001) builds on Clevenger?s research and investigates the ways that availability, quality, and characteristics of available toolstone affects technological organization on San Nicolas Island and San Clemente Island. Based on research at CA-SNI-39 she suggests that tabular and oval cobbles were often split diagonally rather than lengthwise , which results in several different potential round and flat cobble reduction trajectories. Ta?kiran?s results suggest that because cobbles were abundant on San Nicolas Island, metavolcanic and metasedimentary rock was not conserved. Core preparation was designed to facilitate reduction of desired flake types rather than reduction to exhaustion. Rosenthal?s (2008) study of stone artifacts from 20 units from sites across San Nicolas Island adds a broader perspective to previous site-focused studies. She uses data from the index units to learn about artifact production, use, and discard. Rosenthal observes fewer cores than expected given the volume of flakes. She suggests that core reduction occurred at the cobble source; however, some exhausted cobbles may also have been smashed to create flakes. Rosenthal also notes that index units have unequal proportions of material types indicating differential access to source materials, selection of certain material types, or both. 208 My analysis of core and flake types at Mound B supports Ta?kiran?s assertion that people on San Nicolas Island used different core reduction strategies depending on the original cobble shape. At SNI-106, all cores are either exhausted cores or core fragments but I assume that similar technological processes took place there. At Mound B, a common reduction strategy for round and ovoid cobbles is to place a cobble on an anvil stone and strike perpendicular to the anvil to split the cobble in half. The next step is to remove flakes from the rounded cortical surface towards the center of the split face (Clevenger 1982; Ta?kiran 2001; Figure 5.41). This centripetal reduction results in a ?tortoise shell? appearance for a rotated split cobble core (Rondeau 1995). In some cases, after the core is worked to the point that the platform angle becomes too obtuse to flake effectively, it is rotated 90 degrees and additional flakes are removed using the non-cortical flaked split face as a platform. Ta?kiran describes that in a variation of the split cobble core, the cobble is split and then flakes re removed using the non- cortical split face as a platform. The core is then rotated. Because I did not identify any cores that fit this description and noted few first flakes with lisse platforms and 100% dorsal cortex, I cannot demonstrate that this technique was used at Mound B or SNI-106. I identified seven split rotated cores in the Mound B assemblage (Figure 5.42-5.44). 209 Figure 5.41. Split cobble reduction sequence. 210 Figure 5.42. Schematic showing plan view of obverse and reverse of split cobble core from Mound B (25B.32.2.4.C.3) 211 Figure 5.43. Split cobble cores from Mound B, obverse. Top row (left - 25B.52.I.3.shell- lithic feature.C.1, right - 25B.32.II.3.C.1). Middle row (left - 25B.11.II.3.C.2, right - 25B.48.I.2.C.1. Bottom row (left - 25B.11.1.2.C.5, right - 25B.52.I.3. shell-lithic feature. C.2.). 212 Figure 5.44. Split cobble cores from Mound B, reverse. Top row (left - 25B.52.I.3.shell- lithic feature.C.1, right - 25B.32.II.3.C.1). Middle row (left - 25B.11.II.3.C.2, right - 25B.48.I.2.C.1. Bottom row (left - 25B.11.1.2.C.5, right - 25B.52.I.3. shell-lithic feature. C.2.). 213 For oblate and tabular cobbles, people favored diagonal and decapitate core reduction sequences described by Ta?kiran (2001). In the diagonal reduction sequence, the cobble is placed on an anvil and split diagonally down the long axis. Flakes are then removed from the split face, using the round cortical surface as a platform (Figure 5.45). The core can be rotated or semi- rotated to remove additional flakes. I identified 4 cores matching this description at Mound B (Figure 5.46-5.48). For the decapitate reduction sequence, a flake is removed from one end of the cobble using a flat cortical surface as a platform. Additional flakes are struck using the same platform, the force of the blow following the ridge made by the first flake scar (Figure 5.49). I identified 4 cores matching this description at Mound B (Figure 5.50, 5.51). Ta?kiran describes a similar core reduction sequence where a tabular cobble is decapitated and then the flake scar is used as a platform for the next flake, but I did not identify any cores matching this description. These reduction sequence descriptions produce flakes with considerable variation at the beginning of the reduction sequence, but later in the reduction sequence as the cobble core is flaked and rotated, the flakes that result become more similar to one another. Cores at this stage are multidirectional and unpatterned or exhausted (Figure 5.52-5.54). 214 Figure 5.45. Diagonal core reduction sequence. 215 Figure 5.46. Diagonal cobble core Figure 5.47. Diagonal cobble core from Mound B, obverse (25B.52.2.1.C.2). from Mound B, reverse. Figure 5.48. Plan-view schematic of diagonal cobble core from Mound B, obverse (25B.52.2.1.C.2). 216 Figure 5.49. Decapitate reduction technique. Figure 5.50. Decapitated core from Mound Figure 5.51. Decapitated core from Mound B, obverse (25B.11.II.3.C.1). B, reverse. 217 Figure 5.52. Unpatterned multidirectional Figure 5.53. Unpatterned multi- cores from Mound B, obverse (Top - directional cores from Mound B, 25B.52.I.3.Shell-lithic.C.3, Bottom - reverse. 25B.32.II.1b.C.2). Figure 5.54. Plan-view schematic of an unpatterned multidirectional cores from Mound B (25B.32.II.1b.C.2). For the Mound B assemblage, a majority of the cores were fragments or exhausted and substantially smaller than cobbles at the cobble areas therefore it was not possible to calculate an estimate for original cobble size. Based on macroscopic analysis of flake margins, 141 flakes 218 show chipping and damage that may be caused by use, manufacture, or post-depositional processes (Figure 5.55, 5.56). Most flakes were likely used without retouch or other modification and do not show evidence of wear from scraping or chopping. Given the abundance of chipped stone tools at the site and the scarcity of formal tools made from metavolcanic rocks, the goal of most lithic manufacture was to create flake tools to be used immediately for tasks at Mound B or potentially elsewhere on the landscape. Based on the core reduction strategies, people at Mound B did not attempt to remove cortex from the flakes prior to use, but rather, used the cortex as a natural back. Figure 5.55. Flake from Mound B with Figure 5.56. Flake with Mound B with damage on the margin, obverse damage on the margin, reverse. (25B.43.2.3.F.3-40, 42-43). A total of 10 of the 168 retouched tools from Mound B were made on cores, with the remainder of the retouched tools made on flakes. Many of the formal tools are made from exotic cherts and rarer fine-grained island materials such as quartz and siliceous shale. Focusing on the two largest retouched tool assemblages from the 1500-500 cal BP and 500 cal BP-Contact assemblages from Mound B, the most common tools are drills (Figure 5.57-5.60) and scrapers (Figure 5.61-62) with a handful of bifaces/projectile points (Figure 5.63, 5.64), retouched flake 219 tools, choppers, and reamers. Drills are made from a variety of different materials including chert, metavolcanic rock, quartzite, quartz, and sandstone. There are several different manufacturing methods (Preziosi 2001).Some are made by minimally shaping a small flake or bade or by taking advantage of a break. Other drills are made by bisecting a flake with an axial break and have a triangular cross-section. In some cases, small flakes are removed from the distal end of a flake at an oblique angle, creating a trapezoidal cross section. If a drill bit becomes dull it can be resharpened through retouch (Figure 5.59, 5.60). The scrapers and bifaces at Mound B also come in a variety of forms. Some show extensive retouch on one or more margins; others show minimal retouch on a single margin. For 1500-500 cal BP and 500 cal BP- Contact sub-assemblages, material type is relatively evenly distributed. Most scrapers are made using metavolcanic rock or quartzite. Finally, another important tool used at Mound B is the sandstone saw. These tools are made on first flakes from sandstone cores. The proximal ends are thinned through secondary retouch and then the tool is used without further modification (Figure 5.65) (Kendig et al. 2008, 2010:200). Figure 5.57. Drills from Mound B, obverse Figure 5.58. Drills from Mound B, reverse. (25B.11.2.2.T.1-3). 220 Figure 5.59. Retouched drill from Mound B, Figure 5.60. Retouched drill from Mound B, obverse (25B.32.2.1.T.1). reverse. Figure 5.61. Scrapers from Mound B, Figure 5.62. Scrapers from Mound B, obverse (25B.11.1.1.T.1-2). reverse. 221 Figure 5.63. Bifaces from Mound B, from left 25B.44.II.1.T.2, 25B.31.I.1.T.1; 25B.34. I.4.T.1, and 25B.49.I.2.T.2, obverse. Figure 5.64. Bifaces from Mound B, from left 25B.44.II.1.T.2, 25B.31.I.1.T.1; 25B.34. I.4.T.1, and 25B.49.I.2.T.2, reverse. 222 Figure 5.65. Manufacture of a sandstone saw. Photos courtesey of Bill Kendig, CSULA. 223 Conclusions and Comparisons The cobble-core expedient flake tool technologies of the San Juan Islands and southern Channel Islands provided people with an important toolkit for tasks such as cleaning fish, processing meat, processing plant foods, and a range of other activities. The study of two assemblages characterized by similar technologies, material types, and cobble availability provides important insights on flaked beach cobble technology. First, analyses of cores indicate that people selected different shaped cobbles for different purposes. They also approached reduction of angular, round, oblate, and flat cobbles in different ways. People at Watmough Bay specifically chose angular cobbles and used the natural ridges to channel the force of the blow to decapitate the cobble and remove the first few flakes. People at Mound B used round, oval, and tabular cobbles and they typically split their cobbles rather than decapitating them. This is at least partly attributable to the fact that the metamorphic toolstone on San Nicolas Island was harder and more difficult to work than the FGV on the San Juan Islands. In both study areas, however, people had their choice of a number of different core reduction strategies that were based on cobble shape and desired flake morphology. Unpatterned cores are more common than patterned cores in both study areas. An interesting similarity in the flake tools created by people on both the San Juan Islands and on San Nicolas Island is the use of cortex as a natural back as described by Close (2006) for the OpD site on San Juan Island. Cores were set up to create flakes with cortex either adjacent to or opposite from the sharp edge perhaps to serve as a handle to push or pull the tool with greater force. In both study areas, flakes chosen for use were often thicker more robust flakes than those that were not, and they often had round platforms and multiple dorsal flake scars. That sturdier flakes are used for the cutting and scraping tasks that un-modified flakes is not surprising?they 224 would have been easier to hold, broken less frequently, and performed tasks more efficiently. The prevalence of used and modified flakes with multiple dorsal flake scars may indicate that at times it took a few tries to set up a core in such a way that the appropriate flake could be removed. This demonstrates the planned and purposeful nature of flake tool technology despite a lack of formal or standardized end product. 225 Chapter 6: Toolstone Procurement Predictions and Lithic Analysis Archaeologists who study highly mobile hunter-gatherers often use toolstone procurement patterns to investigate changes in mobility and territory size (e.g., Beck et al. 2002; Blades 1999; F?blot-Augustins 1993, 2009; Gamble 1999; Jones et al. 2003; Kuhn 1995, 2004). Assuming that ?raw material is embedded in basic subsistence schedules? (Binford 1979:259), researchers imagine a small group of people moving between habitation sites and subsistence opportunities, encountering toolstone outcrops as they travel. They discard exhausted formal tools, collect and work toolstone, and continue across the landscape with cores and blanks. In contrast, archaeologists? concept of lithic procurement among sedentary or semi-sedentary people is a small group traveling from a village to gather a bulk load of unprocessed toolstone. They bring it home, make several informal flake tools, and store the rest for later use (e.g., Blair 2010; Close 2006; Odell 1998, 2000). Procurement concepts from mobility studies can be applied to more sedentary coastal peoples but must be fine-tuned to account for flake tool technology, logistical procurement, boat transport, exchange, use of beach cobbles, and other characteristics of the coastal record. In this context, procurement patterns reflect the limits of landscape access due to social and physical boundaries. In Chapter 1, I outlined hypotheses for lithic procurement behaviors in different territorial contexts based on economic defensibility and social networking models. I suggested that boundary defense should center on parts of the landscape where resources are adequate to the needs of the population?not so abundant that theft or reciprocal access would be tolerated and not so scarce that the costs of defending the resource would have outweighed the benefits. Boundary permeability should be facilitated by increased benefits and decreased costs of inter- group interaction associated with an unstable or impoverished resource base. In coastal settings, 226 separate terrestrial and marine environmental shifts should foster complex territorial strategies that, in turn, affect lithic procurement, processing, and conservation. In this analysis, I also address an alternative hypothesis. If boundary defense did not occur at the level of the village, access to resource areas might instead be determined by kin groups that transcend village boundaries. If this is the case, different families within a village have access to different parts of the landscape including toolstone collection areas. This territorial organization should be reflected in differences in material type within different parts of the site. For each study area in this study, I focus first on testing the primary hypotheses that are based on economic defendability models. I then briefly explore how the data fit with the alternative hypothesis. This party of the study is less detailed because more spatial data and analysis would be required to more fully explore access to resources on the landscape determined by kin affiliation. Toolstone Procurement and Processing Predictions Based on economic defensibility models, if a habitation site is located in an area considered , increased boundary defense and associated use of a smaller territory should encourage toolstone procurement near a habitation site. Decreased boundary defense associated with scarce or unpredictable resources allows toolstone procurement from the most attractive sources on the landscape because people move over larger spaces. Change over time in access to certain material types, cobble size, and cobble shape should reflect change over time in access to different parts of the landscape. Shifts in resource access should also be reflected in degree of processing at a cobble area prior to transport. Using a central place foraging model framework, degree of processing at a 227 quarry depends on travel costs. Toolmakers seek to minimize costs of transport and field processing (material testing and early stage core reduction) and maximize the amount of high utility toolstone delivered to a habitation site from a source (Bettinger et al. 1997; Metcalfe and Barlow 1992). Utility is defined as the amount of material from a core that is usable as a tool (Bettinger et al. 1997). This model has been applied to quarry behavior in several instances (e.g., Beck et al. 2002; Elston 1990; Kelly 2001; but see Brantingham 2003 for an alternative perspective). These studies indicate that when a quarry site is far from a habitation site, people are likely to increase load utility by testing cobble quality, a practice that has been documented ethnographically and archaeologically at quarry sites (Binford and O?Connell 1984; Ericson 1984; Ross et al. 2003; Torrence 1986). They may also begin early stage reduction to open a flaking surface and remove cortex. Particularly for coastal groups with access to efficient boat travel, distance may be secondary to boundary defense in determining travel cost on a crowded and sociopolitically complicated landscape. People traversing contested areas may incur social and economic costs, face a threat to their well-being, or risk retribution. They should increase the utility of the load prior to transport to ensure that it carries a high enough value to be worth the dangerous trip (Figure 6.1). Another possible interpretation of lithic procurement strategies is that people who are taking toolstone from another groups? territory without permission would work quickly to load up a boat or basket with stones without attracting the attention of their neighbors, therefore toolstone collected from a distance from a habitation site in a territorial regime would not be processed, contrary to the above logic. I propose that this strategy may have been used at times, but given probable inter-connectedness between groups, it would have been more effective to communicate with the other group and use the opportunity to engage in exchange or ceremonial 228 activities. If inter-village relationships were so tense that they could not engage in such relationships, they could have made due with more local material, shell, or bone until the relationship improved. Thus, for villages in high boundary defense/low permeability contexts located near toolstone areas, I expect minimal testing prior to transport and toolstone conservation if local resources are limited. If toolstone areas are located outside of territorial boundaries, increased processing at the source reflects increased social or economic cost in obtaining the resource. Low boundary permeability should be reflected in more formal exchange relationships. Figure 6.1. Graphical representation of the central place foraging model applied to cobble reduction strategies when toolstone (modified from Beck et al. 2002 Figure 5). In contrast, for sites in low defense/high permeability areas, people should acquire toolstone from the most attractive source areas on the landscape. If those areas are far away and travel costs are high, testing offsets a higher transport cost. If travel costs are low, all processing 229 should occur at the habitation site. Toolstone conservation should be minimal unless travel costs are extremely high. Higher boundary permeability and greater benefits of inter-group relationships should lead to more intensive and less formal exchange relationships (Figure 6.2). Figure 6.2. Proposed processing strategies for sites with high and low boundary defense strategies. To test hypotheses about change over time in territorial behavior on the Pacific Coast, I use these general predictions to guide the development of predictions more specific to the 230 archaeological record of the San Juan Islands and southern Channel Islands. I take into account the lithic technology, paleoenvironmental shifts, distribution of toolstone, and transport strategies for each study area. To determine the validity of the predictions I use data collected on local toolstone abundance and morphology and individual attribute analysis of stone artifact assemblages from the Watmough Bay site on Lopez Island, and Tule Creek Village Mound B and CA-SNI-106 on San Nicolas Island. Analytic Methods: Reduction Sequence Analysis On both the San Juan Islands and on San Nicolas Island, source areas cannot be identified using geochemical methods, but material type, cobble size, cobble shape, quarry processing decisions, and toolstone conservation efforts provide insights on where and how people acquired toolstone. Investigating these behaviors requires reduction sequence analysis, defined as the culturally and physically patterned way that people in the past reduced pieces of stone into useful tools (Bleed 2001). Reduction sequence analysis, originated by Holmes (1894), breaks up a continuous and flexible stone tool manufacturing process into abstract stages (Bleed 2001; Johnson 1989; Pecora 2001). Analyses conducted by researchers who employ a similar cha?ne op?ratoire approach (e.g., Andouze 1999; Bodu 1996; Boeda 1995; Close 2006; Inizan et al. 1999; Sellet 1993), consider the entire life history of a tool from procurement to discard in the context of the social and cultural traditions of the toolmakers. For this research, I focus on procurement and processing decisions prior to transport to a habitation site in the larger social context of boundary defense and permeability. For each assemblage, I define early, middle, and late stage reduction based on core types, debitage analysis, and other information on the manufacturing process for each site. Platform 231 facets, percent cortex, dorsal flake scars, and flake termination characteristics for debitage associated with early, middle, and late-stage reduction have been well-established for biface technology (e.g. Andrefsky 2005; Dibble et al. 2005; Kessler et al. 2009; Phillips 2011). These concepts must be adjusted to the core-and-flake technologies used on the San Juan Islands and southern Channel Islands. Debitage attributes associated with reduction stages for each assemblage are defined based on the reduction sequence for each site, a diacritical approach (Sellett 1993). Although I also considered a cortical surface analysis to assess degree of reduction prior to transport (e.g., Douglass and Holdoway 2011; Douglass 2008), I found that this analysis was not sensitive to reduction behavior in either study area because toolmakers attempted to retain cortex on flake tools. To examine processing decisions, I use individual attribute analysis of cores and flakes to determine the degree to which the earliest stages of reduction are represented at the habitation sites. Testing Model Predictions: Toolstone Conservation Traditional models of technological organization equate highly mobile people with curated technology?formal tool forms designed for portability in anticipation of flexible use far from a toolstone source (Bamforth 1986; Binford 1973, 1979; Bleed 1986; Odell 1996). Sedentary people who can better predict time and place of tool use often rely on quickly made but ?wasteful? expedient technology characterized by minimal core preparation and retouch (Andrefsky 2009; Bleed 1986; Kelly 1988; Nelson 1991; Odell 1998; Parry and Kelly 1987; Teltser 1991). Although the curation concept emphasizes the role of mobility in determining technological organization, many researchers have demonstrated that raw material quality, shape and availability are equally if not more important than mobility (Andrefsky (1994 a,b; Bamforth 232 1991; Bradbury and Franklin 2000; Kuhn 1991; MacDonald 2008; Wallace and Shea 2006). Low quality or highly abundant material tends to be associated with informal tools that are relatively effective for most tasks (Andrefsky 2009; Frison 1979). Likewise, increased sedentism may restrict access to high quality toolstone, Thus, when toolstone is more difficult to acquire due to increased boundary defense, people will make more of an effort to conserve the resource by minimizing waste (e.g., Shott 1989; Dibble 1995) and maximizing cutting edge relative to toolstone volume (e.g., Bradbury and Franklin 2000; Brantingham and Kuhn 2005; Braun 2005; Kuhn 1991; Prasciunas 2007). In this study I consider use of cores to exhaustion to assess evidence of conservation. In general, when people choose a more conservative manufacturing strategy, they favor more formal prepared core designs to maximize the ratio of perimeter to volume of raw material (Andrefsky 1987; Brantingham et al. 2000; Brantingham and Kuhn 2001; Goodyear 1979:4-6; Morrow 1997). They also may use cores to exhaustion and smash exhausted cores to produce additional flakes. Evidence for this technique includes exhausted cores, core fragments, ?chunks?, and non-cortical bipolar flakes with multiple dorsal flake scars. Bipolar flakes are difficult to identify for coarser-grained materials. They are characterized by evidence of force on opposite ends such as two bulbs of percussion, negative flake scars on the ventral face, crushing on both ends, axial breakage, and opposing concentric rings or ripples (Close 2006:12-15; Odell 1996:70). Intensive bipolar flaking of exhausted cores should also be evident in increased non- cortical shatter, defined as pieces smaller than < 30 mm that do not show any flake features. I also consider retouch intensity to determine if people are taking steps to make their formal tools last longer based on the degree of invasiveness of retouch flakes (e.g., Clarkson 2002). 233 Testing Model Predictions: Exchange Exchange was an important means of lithic procurement on the Pacific Coast. If direct access to toolstone sources was restricted, exchange relationships may have provided an alternative way to acquire toolstone (Morrow and Jeffries 1989; Whallon 2006). Archaeologists have difficulty separating direct procurement from reciprocal access and exchange using stone artifacts alone. It is equally difficult to distinguish different types of exchange such as middleman and down-the-line methods (Earle 2010; Gould and Saggars 1985; Renfrew 1977). Renfrew and his colleagues (Renfrew et al. 1968, Renfrew 1975, 1977) propose that different kinds of access, including exchange, result in different ?fall-off curves? when percentage of lithic raw material is plotted against distance from the source (and see McCoy et al. 2002, 2010 for more recent examples). Torrence (1986) suggests that there is a distance beyond which people directly acquire material (dependent on transport technology, the value of the material, and other variables). As noted by McCoy et al. (2010:2553) and Earle (2010), use of gravity models often leads to problems of equifinality and other methodological and theoretical difficulties. To explore exchange on the Pacific Coast, I use chert and other extra-local materials as a potential marker of exchange. People could have traveled long distances to directly procure chert, and perhaps sometimes they did, but it also would have been passed between kin, friends, and allies to mark relationships. Chert has a markedly different color and texture than the local material in both study areas. It also tends to be made into more formal tools, perhaps indicating that people considered it to be useful for different kinds of social and technological traditions than FGV on the San Juan Island or metavolcanic rock on the Channel Islands. Other materials that people may have considered to be fundamentally different from local toolstone are rare local 234 toolstones with different flaking properties such as quartz and nephrite. At times when exchange was infrequent and formal due to higher boundary defense, small amounts of extra-local or rare material should be concentrated in discrete areas of the site representing formal relationships between certain individuals or families with other communities. At times when exchange was more frequent and informal due to higher boundary permeability, many people in the village could engage in exchange and other inter-village relationships. Exotic tools and flakes should be spread more evenly across the site. Procurement Predictions for the Watmough Bay Site Based on the territoriality models discussed in Chapter 2, I predict that lithic procurement patterns at Watmough Bay at 3500-2500 and 2500-1600 cal BP should reflect minimal boundary defense due to low population density and food and water resource abundance well beyond the demands of the communities on the San Juan Islands. People should gather toolstone from the most attractive sources on the islands. Inter-community interactions should be informal but moderately frequent. With fewer people on the landscape, people would have interacted freely to obtain resources, information, marriage partners, and exchange partners. I expect increased boundary defense and decreased permeability at 600 cal BP-Contact when population density increased and both terrestrial and marine environments were productive. Subsistence resources would have been just adequate for some groups near productive resource areas like Watmough Bay and inadequate for others. If people at Watmough Bay defended the area around the site and, as a result, focused more resource procurement nearby, this should have several ramifications for lithic procurement. People should increase their use of FGV from Watmough Bay beach, decrease processing prior to transport because 235 transport costs were so low, and increase use of local slate. Since Watmough Bay had a relatively sparse FGV cobble density, I expect intensive toolstone conservation. Exchange between sites on the islands should be limited and formal. At 1600-1000 cal BP, impoverished terrestrial and marine resources should encourage minimal boundary defense. At the Watmough Bay site, people should acquire toolstone from the most attractive sources on this island, which include Agate Beach and Aleck Bay. Boat transport would have kept transport costs relatively low, but more core testing should occur when toolstone was acquired from beaches beyond Watmough Bay than when it was acquired from the beach adjacent to the site. Toolstone conservation should be minimal due to easy access to more toolstone. Increased boundary permeability should be apparent based on higher amounts of chert distributed more evenly across the site (Figure 6.3; Table 6.1). Figure 6.3. Predictions for lithic procurement at Watmough Bay during the Late Holocene. The gray/green shaded area represents territorial boundaries around a village site. 236 Table 6.1. Predictions for the Watmough Bay lithic assemblage. Testing Procurement Predictions at Watmough Bay: Material Type To test the prediction that the inhabitants of Watmough Bay were more geographically restricted in their lithic procurement at 600 cal BP-Contact, I investigate whether people increased their reliance on the slate outcrops adjacent to the site during that time period. Based on ethnographic examples, Northwest Coast archaeologists have proposed that change over time in slate use correlates with increased salmon processing (Graesch 2007; Mitchell 1971; Morin 2004; Lepofsky et al. 2000). Another potential explanation is that for people at Watmough Bay, slate supplemented their raw material needs when FGV was more difficult to access. Despite differences in toolstone characteristics, if people needed a sharp edge for cutting and scraping and FGV was scarce due to territorial circumscription, they may have increased their use of slate. Even if some of the slate in the Watmough assemblage was deposited in the shell midden through natural processes, if there is a dramatic increase in slate tool manufacture, I expect a parallel increase in slate debris created during the manufacturing process. 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP- Contact Territoriality Minimal defense/moderate permeability Minimal defense/moderate permeability Minimal defense/high permeability High defense/low permeability FGV Source Aleck Bay, Agate Beach, Watmough Aleck Bay, Agate Beach, Watmough Aleck Bay, Agate Beach, Watmough Watmough Slate ? ? ? ? Exchange Moderate, informal Moderate, informal Frequent, informal Limited, formal Testing/Early Stage Reduction ? ? ? ? Conservation ? ? ? ? 237 To test this prediction, I consider the Munsell and Stein/Phillips excavations separately to control for different collection strategies for unmodified slate fragments. For the Munsell excavation, I assume that only the bigger fragments were collected but that they were collected consistently. The Stein/Phillips team collected most or all of the slate they encountered, therefore results for that sub-assemblage are more reliable. For both excavations, comparison of slate and FGV by percent weight is not consistent with my predictions. For the Stein/Phillips excavation, percent slate is much lower for the 600 cal BP-Contact assemblage (48.49%) than the 1600-1000 cal BP assemblage (71.75%). In the Munsell assemblage, there is a lower percentage of slate for the 600 cal BP-Contact assemblage (3.22%) than the 2500-1600 cal BP (20.72%) or the 1600- 1000 cal BP assemblage (38.44%). These results may be affected by small sample sizes in all but the 1600-1000 cal BP assemblages (Table 6.2; Figure 6.4, 6.5). Results of ?2 tests comparing counts of FGV and slate also fail to support the prediction that slate use intensified during times of hypothesized territorial circumscription. To ensure that differential breakage in different units did not bias the results, I divided the assemblage into > 3 cm maximum dimension and < 3 cm maximum dimension flakes. I chose this size category as the point to divide the assemblage because this is the size class for the smallest flaked tools (of any material type) in the assemblage and therefore smaller flakes might be considered a byproduct rather than an end product. For the Stein/Phillips excavation there is significantly more slate than expected in the 3500-2500 cal BP assemblage both for all flakes and for < 3 cm maximum dimension flakes (Table 6.3). If only the 1600-1000 cal BP and 600 cal BP-Contact are considered, difference in slate and FGV counts are not significant (?2 = 1.104, 1 df, p = 0.293). 238 For the Munsell excavation, there is significantly more slate than expected in the 3500- 2500 cal BP assemblages and less than expected in the three later assemblage for all flakes. Results are similar for flakes > 3 cm and < 3 cm in maximum dimension (Table 6.4). One possible explanation for this result is that the slate the people used to make tools at Watmough Bay was not the same as that available from the outcrop near the site. Some of the finished tools are made out of material that looks similar?light gray, fragile, and slightly metamorphosed. Other tools appear to be made of a denser material. In future research I plan to determine the variability of slate quality near the site and at other outcrops both on the San Juan Islands and on the mainland. A preliminary result of this analysis, however, does not indicate that people used more slate at times of hypothesized increased boundary defense and may indicate the opposite, that people used were more restricted to local raw material at 3500-1500 cal BP. 239 Table 6.2. Material type counts and weights for flake assemblages from the Stein/Phillips and Munsell excavations from the Watmough Bay site, Lopez Island. Excavation Stein/PhillipsStein/Phillips Stein/Phillips Munsell Munsell Munsell Munsell Time Period (cal BP) 3500-2500 1600-1000 600 - Contact 3500-2500 2500-1600 1600-1000 600 - Contact FGV 11 117 7 67 92 637 43 Metavolcanic/Basalt 0 2 0 2 4 17 0 Argillite 0 0 0 0 1 0 0 Chert 1 2 0 2 0 8 0 Schist/Metasedimentary 0 0 0 10 3 43 0 Nephrite 0 1 0 0 0 1 0 Quartz 0 32 3 0 2 38 0 Quartzite 0 4 0 0 1 3 0 Sandstone 0 0 0 0 0 1 0 Slate 11 463 38 57 7 120 3 Undetermined 0 6 0 0 0 5 0 Total 23 627 48 138 110 873 46 Frequency % FGV 47.83% 18.66% 14.58% 48.55% 83.64% 72.97% 93.48% MV/Basalt 0.00% 0.32% 0.00% 1.45% 3.64% 1.95% 0.00% Argillite 0.00% 0.00% 0.00% 0.00% 0.91% 0.00% 0.00% Chert 4.35% 0.32% 0.00% 1.45% 0.00% 0.92% 0.00% Schist/Metasedimentary 0.00% 0.00% 0.00% 7.25% 2.73% 4.93% 0.00% Nephrite 0.00% 0.16% 0.00% 0.00% 0.00% 0.11% 0.00% Quartz 0.00% 5.10% 6.25% 0.00% 1.82% 4.35% 0.00% Quartzite 0.00% 0.64% 0.00% 0.00% 0.91% 0.34% 0.00% Sandstone 0.00% 0.00% 0.00% 0.00% 0.00% 0.11% 0.00% Slate 47.83% 73.84% 79.17% 41.30% 6.36% 13.75% 6.52% Undetermined 0.00% 0.96% 0.00% 0.00% 0.00% 0.57% 0.00% Weight (g, sum) FGV 42.8 417.3 21.6 419.7 531 2089.1 412.3 Slate 17.2 1076 20.9 323 146.7 1738.5 13.7 Schist/Metasedimentary 0 0 0 78.7 20.9 437.7 0 Quartz 0 6.2 0.6 0 9.4 202.7 0 Chert 0.001 0.1 0 1.5 0 54.3 0 Total 60.001 1499.6 43.1 822.9 708 4522.3 426 Weight % FGV 71.33% 27.83% 50.12% 51.00% 75.00% 46.20% 96.78% Slate 28.67% 71.75% 48.49% 39.25% 20.72% 38.44% 3.22% Schist/Metasedimentary 0.00% 0.00% 0.00% 9.56% 2.95% 9.68% 0.00% Quartz 0.00% 0.41% 1.39% 0.00% 1.33% 4.48% 0.00% Chert 0.00% 0.01% 0.00% 0.18% 0.00% 1.20% 0.00% 240 Figure 6.4. Proportions of FGV and slate by weight (g) for the Stein/Phillips excavation, Watmough Bay, Lopez Island. Figure 6.5. Proportions of FGV and slate by weight (g) for the Munsell excavation, Watmough Bay, Lopez Island. 241 Table 6.3. Results of ?2 tests comparing proportions of FGV and slate by count for the Stein/Phillips excavation, Watmough Bay, Lopez Island. Time Per. (cal BP) Flake Size Stats FGV Slate ?? DF p 3500-2500 All Count 11.00 11.00 12.92 2 0.002 Expected 4.56 17.44 AR 3.44 -3.44 1600-1000 Count 117.00 463.00 Expected 120.31 459.69 AR -1.06 1.06 600-0 Count 6.00 38.00 Expected 9.13 34.87 AR -1.20 1.20 3500-2500 < 3 cm Count 9.00 9.00 12.84 2 0.002 Expected 3.55 14.45 AR 3.28 -3.28 1 00-1000 Count 89.00 370.00 Expected 90.56 368.44 AR -0.55 0.55 600-0 Count 4.00 36.00 Expected 7.89 32.11 AR -1.61 1.61 242 Table 6.4. Results of ?2 tests comparing proportions of FGV and slate by count for the Munsell excavation, Watmough Bay, Lopez Island. Time Per. (cal BP) Flake Size Stats FGV Slate ?? DF p 3500-2500 All Count 67.00 57.00 79.39 3 <0.001 Expected 101.49 22.60 AR -8.53 8.53 2500-1600 Count 92.00 7.00 Expected 80.96 18.04 AR 3.02 -3.02 1600-1000 Count 637.00 120.00 Expected 619.03 137.97 AR 3.30 -3.30 600-0 Count 43.00 3.00 Expected 37.62 8.38 AR 2.10 -2.10 Time Per. (cal BP) Flake Size Stats FGV Slate ?? DF p 3500-2500 > 3 Count 34.00 42.00 73.32 3 <0.001 Expected 61.28 14.72 AR -8.33 8.33 2500-1600 Count 62.00 6.00 Expected 54.83 13.17 AR 2.30 -2.30 1600-1000 Count 510.00 103.00 Expected 494.26 118.74 AR 3.36 -3.36 600-0 Count 35.00 3.00 Expected 30.64 7.36 AR 1.83 -1.83 Time Per. (cal BP) Flake Size Stats FGV Slate ?? DF p 3500-2500 < 3 cm Count 33.00 15.00 11.24 1 0.001 Expected 40.34 7.66 AR -3.35 3.35 1600-1000 Count 125.00 15.00 Expected 117.66 22.34 AR 3.35 -3.35 243 Testing Procurement Predictions at Watmough Bay: Size, Shape, and Cortex Appearance If people were restricted to Watmough Bay for lithic procurement at 2500-1600 cal BP and 500 cal BP-Contact, their cores and debitage should reflect the size, shape, and cortex appearance of the cobbles on that beach. As demonstrated in Chapter 5, Watmough Bay beach cobbles are significantly smaller than those at nearby Aleck Bay and Agate Beach. In fact, they are smaller than cobbles from all of the other beaches included in the survey. Although the proportion of round cobbles at Watmough Bay beach is not significantly greater than at the other beaches on the island, given the apparent preference for angular cobbles and the relative scarcity of cobbles at Watmough, restriction to Watmough beach would result in increased use of round cobbles. Watmough Bay also has significantly more polished (highly waterworn) cobbles than most other beaches in the survey except nearby Agate Beach. Thus, when people were collecting from Watmough Bay, core and flake assemblages should indicate use of smaller, rounder, and smoother cobbles than during periods when they were collecting cobbles from elsewhere on the islands. Beginning with cobble size, only the 1600-1000 cal BP assemblage contains enough cores with original cobble dimensions intact to characterize original cobble size. I predicted that during this period people were not restricted to Watmough Bay beach. Results of an ANOVA comparing mean length for cores and cobbles indicates that mean core size is significantly larger than average cobble size for Watmough Bay. It is also significantly smaller than mean cobble size at Agate Beach and Aleck Bay. The Watmough Bay cores are most similar in size to cobbles found at American Camp and Snug Harbor (Table 6.5). If people were selecting the larger cobbles from the southern Lopez Island beaches, I would expect a lower coefficient of variation for the cores than for the cobble areas. However, the core COV is actually higher than for the 244 cobble areas. A consideration of the distribution of size classes for the Watmough Bay cobbles and the cores with intact cobble dimensions from the 1600-1000 cal BP cobble assemblage shows a slightly different different distribution of cobble sizes. For the Watmough Bay beach FGV cobbles (n = 66), a majority of the cobbles fall into the 20-30 mm (37.9%), 30-40 mm (28.8%) and 40-50% (21.2%). Comparing Watmough FGV cores with intact dimensions (n = 31) indicates fewer 20-30 mm cobbles (9.7%), but similar numbers of 30-40 mm cobbles (29.0%) and 40-50 mm cobbles (25%), and greater number of 50-60 mm cobbles (16.1%). Sample sizes are small enough that these results can only be considered prelimary. I also cannot rule out the possibility that people simply preferred and chose the slightly larger cobbles available on the beach. Table 6.5. Basic statistics for FGV cores from Watmough Bay that have original cobble dimensions intact and cobbles from beaches on the San Juan Islands. Results of an ANOVA comparing mean cobble dimensions for cores from the 1600-1000 cal BP assemblage with cobbles from beaches on the San Juan Islands. To test the hypotheses about cobble shape I use ?2 tests to compare proportions of cores and flakes made from angular and round cobbles from the 1600-1000 cal BP assemblage with cobbles from the natural beaches. The data indicate a significantly higher proportion of angular cobble cores at the Watmough Bay site than at Watmough Bay beach (p < 0.001), Snug Harbor (p < 0.001), and Agate Beach (p < 0.001). The core assemblage from Watmough Bay is similar in proportion of angular and round cobbles to American Camp (p = 0.146), False Bay (p = 0.916) n Mean SD COV F Sig. Bonferroni Sig. Wa ough Bay Cores 31 48.24 15.90 32.95% 44.04 0.00 American Camp 36 44.47 13.49 30.34% 1.00 Watmough Bay Beach 64 37.13 11.68 31.47% 0.02 False Bay 94 69.73 20.26 29.05% 0.00 Snug Harbor 130 51.34 9.71 18.91% 1.00 Agate Beach 70 58.39 15.98 27.37% 0.05 Aleck Bay 30 76.77 23.74 30.92% 0.00 245 and Aleck Bay (p =0.069; Table 6.6). These results indicate that at 1600-1000 cal BP, people were accessing cobble areas other than Watmough Bay beach. Table 6.6. Results of ?2 tests comparing proportions of cores (1600-1000 cal BP) made on angular and round cobbles from Watmough Bay with cobble areas on the San Juan Islands. Type Stats Angular Round ?? DF p 1600-1000 cal BP Cores Count 47 11 2.111 1 0.146 Expected 44.13 13.87 AR 1.45 -1.45 American Camp Cobbles Count 23 11 Expected 25.87 8.13 AR -1.45 1.45 Type Stats Angular Round ?? DF p 1600-1000 cal BP Cores Count 47 11 21.632 1 <0.001 Expected 34.9 24 AR 4.65 -4.65 Watmough Bay Cobbles Count 21 34 Expected 33.1 21.9 AR -4.65 4.65 Type Stats Angular Round ?? DF p 1600-1000 cal BP Cores Count 47 11 0.011 1 0.916 Expected 47.25 10.76 AR -0.11 0.11 False Bay Cobbles Count 76 17 Expected 75.76 17.25 AR 10.76 -0.11 Type Stats Angular Round ?? DF p 1600-1000 cal BP Cores Count 47 11 17.529 1 <0.001 Expected 33.94 24.06 AR 4.19 -4.19 Snug Harbor Cobbles Count 63 67 Expected 2.24 3.16 AR -4.19 4.19 Type Stats Angular Round ?? DF p 1600-1000 cal BP Cores Count 47 11 14.39 1 <0.001 Expected 36.7 21.3 AR 3.8 -3.8 Agate Beach Cobbles Count 34 36 Expected 44.3 25.7 AR -3.8 3.8 Type Stats Angular Round ?? DF p 1 00-1000 cal BP Cores C unt 47 11 3.304 1 0.069 Expected 43.5 14.5 AR 1.82 -1.82 Aleck Bay Cobbles Count 19 11 Expected 22.5 7.5 AR -1.82 1.82 246 To test predictions about cortex appearance, I compare FGV flakes from the 2500-1600 cal BP, 1600-1000 cal BP, and 600-0 cal BP periods with the cobble assemblages. I expected a higher number of smooth and polished cobbles at 2500-1600 cal BP and 600 cal BP-Contact. Proportions of smooth and polished flakes for these three assemblages were not significantly different (?2 = 0.56, 2 df, p = 0.76). A comparison of the 1600-1000 cal BP FGV flake assemblage with all of the cobble areas shows no significant differences between in proportions of smooth and rough cortex (Table 6.7). There are too few cores to compare cortex appearance between time periods, but I instead compare the large 1600-1000 cal BP core and flake assemblages with the cobble beaches. There are no significant differences in proportions of smooth and rough cortex cores and cobble areas (Table 6.8). This measurement is effectively neutral with respect to my predictions. 247 Table 6.7. Results of a ?2 test comparing proportions of smooth and rough cortex FGV flakes at Watmough Bay (1600-1000 cal BP) with smooth and rough cortex cobbles at each cobble area. Aleck Bay and Watmough Bay are not included due to a small sample size for rough cobbles. Type Stats Smooth Rough ?? DF p 1600-1000 cal BP flakes Count 374.00 59.00 0.43 1.00 0.51 Expected 372.73 60.27 AR 0.65 -0.65 American Camp Cobbles Count 28.00 6.00 Expected 29.27 4.73 AR -0.65 1.06 Type Stats Smooth Rough ?? DF p 1600-1000 cal BP flakes Count 374.00 59.00 0.01 0.00 0.94 Expected 373.76 59.24 AR 0.08 -0.08 False Bay Cobbles Count 74.00 12.00 Expected 74.24 11.76 AR -0.08 0.08 Type Stats Smooth Rough ?? DF p 1600-1000 cal BP flakes Count 374.00 59.00 0.03 1.00 0.87 Expected 374.55 58.45 AR -0.16 0.16 Snug Harbor Count 113.00 17.00 Expected 112.45 17.55 AR 0.16 -0.16 Type Stats Smooth Rough ?? DF p 1600- 0 cal BP flakes Count 374.00 59.00 2.28 1.00 0.13 Expected 377.91 55.09 AR -1.51 1.51 Agate Beach Count 65.00 5.00 Expected 61.09 8.91 AR 1.51 -1.51 248 Table 6.8. Results of ?2 tests comparing proportions of smooth and rough cortex FGV cores from Watmough Bay (1600-1000 cal BP) with smooth and rough cortex cobbles at each cobble area. Type Stats Smooth Rough ?? DF p 1600-1000 cal BP Cores Count 374.00 59.00 0.43 1.00 0.51 Expected 372.73 60.27 AR 0.65 -0.65 American Camp Cobbles Count 28.00 6.00 Expected 29.27 4.73 AR -0.65 0.65 Type Stats Smooth Rough* ?? DF p 1600-1000 cal BP Cores Count 374.00 59.00 Expected AR Watmough Bay Cobbles Count 55.00 0.00 Expected AR *Small sample size Type Stats Smooth Rough ?? DF p 1600-1000 cal BP Cores Count 80.00 13.00 0.00 1.00 1.00 Expected 80.01 12.99 AR -0.05 0.03 False Bay Cobbles Count 74.00 12.00 Expected 73.99 12.01 AR 0.05 -0.03 Type Stats Smooth Rough ?? DF p 1600-1000 cal BP Cores Count 80.00 13.00 0.04 1.00 0.85 Expected 80.49 12.51 AR 0.00 0.00 Snug Harbor Cobbles Count 112.51 17.49 Expected 77.88 17.57 AR 0.00 0.00 Type Stats Smooth Rough* ?? DF p 1600-1000 cal BP Cores Count 80.00 13.00 1.90 1.00 0.17 Expected 82.73 10.27 AR -1.38 1.38 Agate Beach Cobbles Count 65.00 5.00 Expected 62.27 7.73 AR 1.38 -1.38 Type Stats Smooth Rough ?? DF p 1600-1000 cal BP Cores Count 80.00 13.00 Expected 62.18 AR 2.03 Aleck Bay Cobbles Count 29.00 1.00 Expected 22.82 AR -2.03 *Small sample size 249 Overall, results for cobble size and shape provide several clues regarding the provenance of cobbles collected by people from the Watmough Bay site during the 1600-1000 cal BP period (Table 6.9). Based on an analysis of cores and flakes, the cobbles that people used were on average larger and more angular than those found on Watmough Bay beach. Results for cortex appearance do not indicate significant differences between the 1600-1000 cal BP assemblage and Watmough Bay beach or any other beach, suggesting that people did not have a preference for a particular cortex texture or that they collected cobbles from a variety of beaches. These data suggest that at 1600-1000 cal BP, people collected toolstone from nearby Aleck Bay, Agate beach, or other beaches throughout the islands. Due to smaller sample sizes for the other time periods, I was not able to determine if the data are consistent with toolstone procurement shifts through time. Table 6.9. Summary of results for shape, size, and appearance of artifacts from Watmough Bay. Cells left blank indicate that sample size was too low to reach meaningful conclusions. 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Size Predictions ? ? ? ? Results Cores from cobbles larger than those found at Watmough Bay beach. Shape Pr ic ion More Angular More Angular More Angular More Round Shape Re ults Cores/flakes indicate more angular than round cobbles. Cort x Predictions Different proportions f smooth/rough from Watmough Beach. Different proportions of smooth/rough from Watmough Beach. Different proportions of smooth/rough from Watmough Beach. Similar to Watmough Beach Results No significant differences between flake assemblages. No significant differences between cores and cobble areas, flakes and cobble areas, or flake assemblages. No significant differences between flake assemblages. 250 Testing Procurement Predictions at Watmough Bay: Processing To investigate procurement strategies on the San Juan Islands, I also consider processing decisions using reduction sequence analysis and other observations of the assemblage. Based on the reduction sequences described in Chapter 5, cobble test and the first steps of the reduction process should be characterized by the removal of a fully cortical decapitation flakes and perhaps a second flake that follows the ridge of the first flake scar to open a flaking surface and determine the quality of the material. Second flakes removed from the core have cortical platforms, a single dorsal flake scar, and often have a cortical margin. In contrast, later stage flakes have cortical platforms and multiple dorsal flake scars or single facet platforms and multiple flake scars. I expect that if people acquired cobbles primarily from Watmough Bay beach at 600 cal BP-Contact, lack of processing prior to transport should be reflected in larger number of first and second flakes relative to later stage flakes found at the site during those time periods. Results indicate no first flakes in the two earlier assemblages. The sample size of flakes identified as ?first? or ?later? is too small in the 3500-2500 cal BP period (n = 6) to be useful for analysis. That there are no first flakes in the 2500-1600 cal BP assemblage and 19 later stage flakes may indicate that first flakes may be removed elsewhere. A ?2 test comparing first flakes and later stage flakes for the later two assemblages reveals no significant difference in flake types between periods, but sample size for the 600 cal BP-Contact period is too small to reach a definitive conclusion (Table 6.10). A ?2 test comparing numbers of dorsal flake scars between time periods indicates significantly more flakes with 0 or 1 flake scars during the 600 cal BP- Contact time period than during the other time periods (Table 6.11). 251 Table 6.10. FGV flake types at Watmough Bay and ?2 results comparing first flakes and later flakes between 1600-1000 cal BP and 600 cal BP-Contact time periods. Table 6.11. Results of ?2 tests comparing number of dorsal flake scars between time periods at Watmough Bay. Time Per. (Cal BP) Stats First flake Later flake 3500-2500 * Count 0 6 2500-1600 * Count 0 19 *Not included in ?? test, small sample size. ?? DF Fisher's exact p 1600- 00 Count 21 100 0.987 1 0.390 Expected 22.17 98.83 AR -0.99 0.99 600-0 Count 3 7 Expected 1.83 8.16 AR 0.99 -0.99 Per. (Cal. BP) Stats 0,1 2,3 4+ ?? DF p 35 0-2500 Count 11.00 12.00 8.00 18.49 6 0.005 Expected 12.27 11.88 6.85 AR -0.48 0.04 0.51 2500-1600 Count 23.00 33.00 13.00 Expected 27.32 26.45 15.24 AR -1.13 1.72 -0.69 1600-1000 Count 198.00 197.00 113.00 Expected 201.12 194.71 112.18 AR -0.63 0.47 0.20 600-0 Count 19.00 1.00 6.00 Expected 10.29 9.97 5.74 AR 3.57 -3.69 0.12 252 Toolstone Conservation at Watmough Bay If high quality lithic raw material was difficult to acquire during times of increased territorial circumscription at 600 cal BP-Contact, people would have used toolstone conservation strategies to make their supply last longer. One way to look for conservation is to determine the degree to which cores were flaked to exhaustion rather than just discarded when they became more difficult to flake due to small size or obtuse platform angles. Regarding cores, for the 2500-1600 cal BP time period, there are 11 core fragments/exhausted cores and 9 cores of other types. For the 1600 -1000 cal BP period, there are 67 core fragments and exhausted cores and 57 cores of other types. For both periods, this suggests some degree of conservation and potential bipolar smashing of exhausted cores, but there is no discernible difference between time periods. Sample sizes for the 3500-2500 cal BP time period and 600 cal BP-Contact time period are too small to analyze quantitatively. A comparison of non-cortical shatter between time periods is not useful for this case study because the Munsell excavation collection strategy was biased against smaller flakes and the Stein- Phillips excavation sample is too small. I also consider retouch intensity to determine if people are taking steps to make their formal tools last longer based on the degree of invasiveness of retouch flakes (e.g., Clarkson 2002). A qualitative assessment of retouch intensity suggests that the later two assemblages are more heavily retouched than the earlier ones. In the 3500-2500 cal BP assemblage, there are 3 scaled pieces, a scraper with minimal retouch, a more extensively retouched flake, and a rough biface. In the 2500-1600 cal BP assemblages, there are 12 scaled pieces, 2 rough bifaces, 1 unretouched chopper, 1 retouched pointed tool, 1 minimally retouched flake, and 6 scrapers. Of these, three are minimally retouched with short retouch flake scars that cover only part of one 253 margin. In contrast, two others have short flake scars that cover the entirety of one or margin. A chert scraper shows extensive retouch with invasive flake scars on all margins. In the 1600-1000 cal BP assemblage, there are 46 scaled pieces and 26 bifaces (most of which are early stage and assymetrical with large flake scars) but also a large number of retouched tools. The leaf-shaped (n =11), stemmed (n = 4) triangular (n = 7), and projectile point tips and fragments (n =6) show extensive retouch, mainly long or invasive flake scars on the entirety of both lateral margins. For the scrapers (n = 17), the three with long retouch flake scars have retouch covering the entire margins. For those with short flake scars, about half are partial and the other half are entire. For the 600 cal BP-Contact assemblage, there are no scaled pieces and two extensively retouched projectile points. The only assemblage that contained flakes identified as biface thinning flakes was the 1600-1000 cal BP assemblage, likely because of the larger sample size. Overall, data are not sufficient to test whether conservation efforts at Watmough Bay increased at 2500-1600 cal BP and 600 cal BP-Contact. The presence of both exhausted cores and other core types during the 1600-1000 cal BP and 600 cal BP-Contact as well as a substantial number of extensively retouched formal tools during these periods are consistent with some degree of toolstone conservation, but the high ratio of cores to flakes in general suggest that conservation did not guide manufacturing decisions. Exchange at Watmough Bay The amount of exotic raw material found during any time period at Watmough Bay is almost insignificant (Table 6.2), which makes it difficult to test predictions for boundary permeability. I predicted increased boundary permeability during times of food scarcity at 1600- 254 1000 cal BP and decreased permeability with increased subsistence resource abundance and increased boundary defense during the 600 cal BP period. There are three chert flakes and one core in the 3500-2500 cal BP assemblage, 1 chert scraper and 2 cores in the 2500-1600 cal BP assemblage, 10 chert flakes, 1 core, and 3 retouched tools, in the 1600-1000 cal BP assemblage, and no chert in the 600 cal BP- Contact assemblage. The higher amount of chert in the 1600- 1000 cal BP assemblage and the lack of chert in the 600 cal BP-Contact assemblage are consistent with predictions, but the larger sample size would have the same effect. Sample size for the other time periods that the results for these are tentative at best. Only the 1600-1000 cal BP period has a large enough sample size to analyze the distribution of chert and other rare materials across the site. My prediction for this time period was that frequent and informal exchange should be reflected in an even distrution of extra-local and rare toolstone across households at Watmough Bay. On the basis of visual inspection, I found that chert, nephrite, steatite, serpentine, and quartz tools and flakes were dispersed relatively evenly, although they were not abundant (Figure 6.6). I also considered the distribution of incised shale?small tabular stones with lines, dashes, and patterns?as potential markers of household identify or status. Site chronology is not precise enough to determine if these objects were deposited contemporaneously since these units were deposited over hundreds of years, but I assume that household identity and location remained constant over several generations. There are clusters where three or more extra-local/rare objects occur in 0N24W, EXU1, 9N3W, and 3S0E, but overall, the distribution of these material is relatively even across the site. This distribution is consistent with a more informal exchange pattern during this time period. 255 Figure 6.6. Distribution of extra-local materials, beads, ornaments, and incised shale at Watmough Bay. Summary of Toolstone Procurement Results for Watmough Bay The Watmough Bay dataset provides mixed results regarding the prediction that the 600 cal BP-Contact sub assemblages should indicate decreased use of local FGV beach cobbles and nearby slate outcrops. The small datasets for several of the time periods made it difficult to fully employ a diachronic perspective, I was able to assess whether the 1600-1000 cal BP assemblage reflected procurement, processing, and conservation behaviors associated with minimal territorial circumscription. Based on results for slate use, in most cases weights and counts of this local material do not increase at 600 cal BP-Contact despite predictions that people should rely more heavily on toolstone available at Watmough Bay. Consistent with my prediction that lithic procurement did not take place at Watmough Bay at 1600-1000 cal BP, cobble size and cobble shape for the lithic assemblage dating to that period indicates the use of larger and more angular cobbles than would 256 be the case if people relied exclusively on Watmough Bay beach. Regardless of where people were getting toolstone, ratios of first flakes to later flakes and cortical surface analysis indicates little or no processing of toolstone prior to transport to Watmough Bay. Some effort was made to conserve toolstone, but based on the numbers of exhausted cores, core fragments, and extensively retouched tools, conservation does not define the technology during any time period. Regarding boundary permeability, the distribution of chert, quartzite, and steatite in the 1600- 1000 cal BP assemblage is consistent with infrequent but informal interactions between groups. Other types of stone, foodstuffs, or perishables might be better indicators of exchange, or exchange may not play a large role in daily life at Watmough Bay. Evaluating an Alternative Hypothesis for Lithic Procurement at Watmough Bay As an alternative hypothesis, I propose that territoriality and resource access that operated at the family or household level rather than at the level of the community. To address this alternative hypothesis, I compare the two spatial areas that date to 1600-1000 cal BP: Area 1 (0N24W and 0N 18W, n = 98 flakes) on the northwest side of the site and Area 2 (0N0E, 1.5N0E, 0N3W and Balk C, n = 204 flakes) on the southeast side of the site. If different families had access to different toolstone as a result of kinship ties beyond the village, there should be a significant difference in the ratio of FGV to slate flakes. A preliminary exploration of this hypothesis indicates that there is a significant difference in the amount of slate > 3 cm in maximum dimension between the two spatial areas. There is a higher proportion of slate in Area 1, which is farther away from the bedrock outcrop (Table 6.12). 257 Table 6.12. Results of a ?2 test comparing counts of > 3 cm slate and FGV in Area 1 and Area 2 at Watmough Bay. Sample size for the < 3 cm fraction was too small for quantitative analysis Access to different beaches by different families should also be apparent in differences in cobble size, angularity, cortex appearance, and processing strategy. Results of ?2 tests comparing proportions of smooth and rough cortex flakes do not show significant differences between areas (?2 = 0.91, 1 df, p = 0.34). Sample sizes for other cobble variables were too small for quantitative comparisons. There is no significant difference between early and later stage flakes (?2 = 0.83, 1 df, p = 0.18). This preliminary analysis is not consistent with dramatic differences in the spatial distribution of toolstone at Watmough Bay that would be associated with differential lithic procurement by households. Sp tial Uni s St t Slate FGV ?? DF p Ar 1 Count 26 39 26.437 1 <0.001 Expected 12.6 52.4 AR 5.14 -5.14 Area 2 Count 12 119 Expected 25.4 105.6 AR -5.14 5.14 258 Procurement Predictions for San Nicolas Island To address hypotheses on territorial behavior using the San Nicolas Island record, I test predictions for lithic procurement patterns using assemblages from Tule Creek Village Mound B and CA-SNI-106. I compare assemblages from these sites dating to 5000-4000 cal BP, 3000- 1500 cal BP, 1500-500 cal BP and 500 cal BP-Contact. The assemblage from CA-SNI-106 dates to the middle two time periods. I make predictions for the San Nicolas Island lithic assemblages based on the environmental history, demographic information, toolstone availability, and lithic technology specific to this study area (Table 6.19). Because both Tule Creek Village and CA- SNI-106 are inland sites, I assume that most transport of local toolstone is pedestrian, therefore travel distance is a rough proxy for travel cost. For the 5000-4000 cal BP assemblage I predict minimal evidence for boundary defense due to a lower population and food and water resources that would have far exceeded the needs of the communities on the islands. Unconstrained movement on the landscape should be apparent in procurement from the most attractive cobble area at Thousand Springs where all types of toolstone are abundant, densely distributed, and accessible. The lithic assemblage should indicate testing (removal of a first and second flake) of all toolstone prior to transport due to a relatively high travel cost over land. Since toolstone was abundant and accessible, toolstone conservation should be minimal. Because boundaries would be poorly defended and the small groups of people would have interacted as part of social networks to gain marriage partners, information, and prestige items, exchange should be moderate. Chert and obsidian artifacts should be relatively few but evenly distributed across the site, consistent with limited informal exchange. At 3000-1500 cal BP and 500 cal BP-Contact, population was higher, the terrestrial climate was cool and wet, and the marine environment was unproductive. I predict increased 259 evidence for inland-focused territoriality for groups near productive resource patches (adequate but not exceeding the needs of the community), including Mound B because it was located adjacent to a freshwater creek. If people defended a smaller territory at this time, this should be reflected in a decrease in quartzite and sandstone that would have been scarce near the site. Short travel distance should be apparent in minimal testing and working of metavolcanic rock at the nearby cobble area prior to transport to the site. Metavolcanic rock should be conserved at Mound B due to a limited supply of high quality material nearby. This should not be the case at SNI-106 which was near a more secondary resource area without ready access to freshwater and associated resources. The costs of protecting the resources near that site should outweigh the benefits therefore people should acquire metavolcanic rock from a larger area. During special purpose trips to acquire coastal resources including quartzite and sandstone, inhabitants of both sites would have had to negotiate access through contested inland areas. Once at the beach cobble areas, they should take time to test quartzite prior to transport to offset high costs of travel and access by increasing the utility of the load. I also predict conservation of quartzite and sandstone to minimize trips to marine areas. Due to low boundary permeability in terrestrial areas, chert and quartz should be rare at the site but the flakes and tools that are present should be clustered in certain parts of the site reflecting formal relationships between kin or friends with other groups. At 1500-500 cal BP, marine resources were abundant and predictable while terrestrial resources were relatively impoverished. Since territoriality should be focused on the island shorelines, people occupying Tule Creek Village and SNI-106 during these times would have had their choice of inland quarry areas. Due to longer travel distances overall, people should test metavolanic rock prior to transport. Efforts to conserve metavolcanic toolstone should be 260 minimal. Access to marine resources, including quartzite and sandstone, would have been more difficult than the periods before and after because it would have required a relationship with a group in the marine zone. As during the 3000-1500 cal BP and 500 cal BP-Contact periods, I predict early stage reduction of quartzite prior to transport because access to that resource incurs both travels costs and negotiation costs. For the same reasons, I also predict toolstone conservation to minimize trips to resupply. Chert and quartz should increase due to decreased inland boundary defense and increased need to share terrestrial resources. It should be spread more evenly across the site reflecting more informal exchange (Figure 6.7; Table 6.13). Figure 6.7. Graphical depiction of procurement predictions for Mound B. Gray/green shaded area indicates territorial boundaries around a village. 261 Table 6.13. Table summarizing predictions for lithic analysis at Mound B and SNI-106. Testing Procurement Predictions on San Nicolas Island: Material Type Since sandstone and quartzite are significantly more abundant in shoreline cobble areas than inland cobble areas, changes in abundance of these materials is one indicator of shifts in access to coastal areas on San Nicolas Island. If people had complete access to any toolstone area at 5000-4000 cal BP, this should be evident in a more even distribution of material types. At 1500-500 cal BP, if the most attractive cobble areas near the shoreline were controlled by nearby villages during a marine territoriality regime, inland groups like those as Tule Creek Village and CA-SNI-106 would have relied more heavily on nearby inland cobble sources where sandstone and quartzite were rare. The 3000-1500 cal BP and 500 cal BP periods of terrestrial-centered boundary defense should fall somewhere in between because travel to the shoreline may have been limited by other inland communities but coastal areas themselves should be open. Mound B 5000-4000 Mound B 3000- 1500 Mound B 1500-500 Mound B 500 - Contact SNI-106 3000- 1500 SNI-106 1500-500 Territoriality Defense low, permeability high. Terrestrial Marine Terrestrial Terrestrial Marine Metavolcanic Source Thousand Springs Tule Creek Inland quarries Tule Creek Multiple inland quarries Multiple inland quarries Quartzite/San dstone Source Thousand Springs Corral Harbor Most attractive Corral Harbor Beach quarries Beach quarries Exchange Informal Formal Formal Formal Formal Formal Testing/Early Stage Reduction at Source All toolstone Quartzite and sandstone All material All toolstone All toolstone All toolstone Conservation Minimal All toolstone Quartzite and sandstone All toolstone Quartzite and sandstone Quartzite and sandstone Site and Date (Cal BP) 262 To test these predictions, I compare proportions of material type by weight and count for the debitage assemblage to determine both the overall amount of material at the site and the amount and size of debitage. I include both groundstone and sandstone flakes because they can be hard to distinguish from each other, and because grinding and flaking were often part of the same technological processes to make sandstone saws, bowls, and other objects. A comparison of proportions of material type by weight (Table 6.14; Figure 6.8) indicates that in accord with the predictions there is relatively more quartzite in the 5000-4000 cal BP Mound B assemblage than during any other period. The high percentage of sandstone in the 3500-1500 cal BP period is attributable to one very large sandstone bowl fragment. That the second largest amount of sandstone is in the 5000-4000 cal BP assemblage is in accord with the predictions. That the amounts of sandstone and quartzite are lower in the three later periods at Mound B is also consistent with the predictions, but the fact that these percentages are not lowest during the 1500- 500 cal BP period is less consistent with my predictions. At SNI-106, there is almost no sandstone present in the assemblage (1 sandstone fragment was found in the 1500-500 cal BP assemblage and none were found in the 500-0 cal BP assemblage), which is consistent with difficult access at inland sites and/or preference for other material types. Percent quartzite indicates a slightly smaller amount at 1500-500 cal BP than at 500 cal BP-Contact, which is consistent with the predictions (Table 6.14, Figure 6.9). 263 Table 6.14. Material weights for Mound B and SNI-106 based on debitage only. Site Mound B Mound B Mound B Mound B SNI-106 SNI-106 Time Period 5000-4000 BP 3000-1500 BP 1500-500 BP 500 BP-Contact 3000-1500 BP 1500-500 BP Count Metavolcanic 81 144 452 971 113 225 Metasedimentary 4 11 19 56 0 0 Quartzite 14 18 85 171 34 21 Sandstone 7 2 34 51 0 1 Exotic Chert 9 8 13 40 3 3 Quartz 1 12 10 47 20 12 Total 116 195 613 1336 170 262 Frequency % Metavolcanic 69.80% 73.85% 73.74% 72.68% 66.50% 85.90% Metasedimentary 3.50% 5.64% 3.10% 4.19% 0.00% 0.00% Quartzite 12.10% 9.23% 13.87% 12.80% 20.00% 8.00% Sandstone 6.00% 1.03% 5.55% 3.82% 0.00% 0.40% Chert 7.80% 4.10% 2.12% 2.99% 1.80% 1.20% Quartz 0.90% 6.15% 1.63% 3.52% 11.80% 4.60% Weight (g, sum) Metavolcanic 337.6 797.6 3142.4 5391.7 510.4 1066.6 Quartzite 137.5 53.9 496.7 802.9 109.9 164.1 Sandstone 109.6 593.6 384.3 712.8 0 0 Chert 13.7 15.9 38 61.2 11.1 6.1 Quartz 0 16.4 24.1 96.4 45.5 164.1 Total 598.4 1477.4 4085.5 7065 676.9 1400.9 Weight % Metavolcanic 56.40% 53.99% 76.92% 76.32% 68.90% 76.10% Quartzite 23.00% 3.65% 12.16% 11.36% 14.80% 11.70% Sandstone 18.30% 40.18% 9.41% 10.09% Chert 2.30% 1.08% 0.93% 0.87% 1.50% 0.40% Quartz 0.00% 1.11% 0.59% 1.36% 6.10% 11.70% 264 Figure 6.8. A comparison of material type proportions by weight for flakes at Mound B. Figure 6.9. A comparison of material type proportions by weight for flakes at SNI-106. Results of ?? tests comparing counts of metavolcanic, quartzite, and sandstone flakes indicate no significant differences in proportions of material type between time periods at Mound B. Significant differences in proportions of metavolanic and quartzite flakes at SNI-106 are in 265 accord with the predictions. For these analyses, I compare three sub-groups of flakes: all flakes, > 3 cm maximum dimension flakes, and < 3 cm maximum dimension flakes. I chose this size class to divide the assemblage because with the exception of one broken drill, all tools are 3 cm in maximum dimension or greater. Separating the smaller fraction reveals bias caused by large amounts of small shatter. A ?? test comparing proportions of metavolcanic and quartzite flakes for all flake sizes for the Mound B assemblages indicates no significant differences between the four assemblages (?? 2.21, 3 df, p = 0.53). There are also no significant differences for > 3 cm (?? 4.30, 3 df, p = 0.23) and < 3 cm (?? 3.20, 3 df, p = 0.36) sub-samples. In comparing only the two larger assemblages, 1500-500 cal BP and 500 cal BP-Contact, there are no significant differences in proportions of metavolcanic and quartzite flakes whether considering all flakes (?? = 0.02, 1 df, p 0.65), > 3cm flakes (?? 0.25, 1 df, p = 0.62), or < 3 cm flakes (?? 0.05, 1 df, p = 0.83). Proportions of metavolcanic flakes and sandstone flakes/fragments for the 1500-500 cal BP and 500 cal BP- Contact assemblages are not significantly different for all flakes (?? 2.5, 1 df, p 0.11), > 3 cm flakes (?? 1.4, 1 df, p = 0.23) or < 3 cm flakes (?? 1.07, 1 df, p = 0.30). For CA-SNI-106, ?? tests indicate significantly greater proportions of quartzite flakes than expected in the 3000-1500 cal BP time period for all flakes and for < 3 cm flakes. The 0.08 p-value for > 3 cm flakes is nearly significant (Table 6.15). This result and the lack of sandstone at this site meets the prediction that people would have had the most difficulty accessing quartzite-rich marine cobble areas during the 1500-500 cal BP time period. 266 Table 6.15. Results of ?? tests comparing proportions of metavolcanic and quartzite flakes between time periods at CA-SNI-106, significant values in bold. Another material difference that I considered in investigating differences in territorial circumscription between time periods is sub-type of metavolcanic rock: metavolcanic (MV, few or no phenocrysts), porphyritic metavolcanic (PMV, many phenocrysts), and metavolcanic porphyry (MVP, groundmass dominated by phenocrysts). MV rock is often finer grained and MVP is almost always coarser grained, although in some cases, the MVP has been subject to so much heat and pressure that the phenocrysts and groundmass have completely fused. Different raw materials have strengths and weaknesses for different tasks because they fracture differently. Materials that are easy to work tend to be less durable (Bradbury et al. 2008; Frison 1991; Terry et al. 2008). The MV toolstone would have been easier to work but PMV and MVP may have been preferred for cutting or sawing activities. Change over time in proportions of metavolcanic Time Per. (cal BP) Flake Size Stats Metavolcanic Quartzite ?? DF p 3000-1500 All Count 113 34 17.466 1 <.001 Expected 126.8 20.2 AR -4.2 4.2 1500-500 Count 226 20 Expected 212.2 33.8 AR 4.2 -4.2 Time Per. (cal BP) Flake Size Stats Metavolcanic Quartzite ?? DF p 3000-1500 > 3cm Count 39 10 3.176 1 0.08 Expected 42.3 6.7 AR -1.8 1.8 1500-500 Count 68 7 Expected 64.7 10.3 AR 1.8 -1.8 Time Per. (cal BP) Flake Size Stats Metavolcanic Quartzite ?? DF p 3000-1500 < 3cm Count 74 24 14.977 1 <.001 Expected 84.5 13.5 AR -3.9 3.9 1500-500 Count 158 13 Expected 147.5 23.5 AR 3.9 -3.9 267 rock types could reflect natural differences in the cobble areas used, changes in cobble collection areas, or changing preference in tool type. When people at Mound B had more options available in collecting MV toolstone (5000-4000 cal BP and 1500-500 cal BP) they may have selected more MV rock. When they were more limited in their collection opportunities due to inland- centered territoriality (3000-1500 cal BP and 500 cal BP-Contact), MV should decrease and PMV and MVP should increase. At SNI-106, proportions of MV should stay the same through time if their boundary defense strategies did not change. Results of ?? test comparing proportions of MV, PMV, and MVP flakes for all size classes at Mound B and for > 3 cm and < 3cm flakes show mixed results (Table 6.16). Considering all size classes, the proportion of PMV flakes is significantly higher during the 3000-1500 cal BP period and significantly lower during the 1500-500 cal BP period. Contrary to the predictions, however, the proportion of MVP flakes is significantly higher during the 1500- 500 cal BP period (Table 6.17). Results of ?? tests for the > 3cm (?? 9.72, 6 df, p = 0.137) and < 3 cm flakes (?? 7.41, 6 df, p = 0.28) do not show statistically significant differences between time periods. At SNI-106, consistent with predictions, there is significantly more MV during the 1500-500 cal BP period than the 3000-1500 cal BP for all flakes, > 3 cm flakes, and <3 cm flakes (Table 6.18). 268 Table 6.16. Data on proportions (count and weight) of metavolcanic, porphyritic metavolcanic, and metavolcanic porphyrtic toolstone at Mound B and SNI-106. Table 6.17. Results of a ?2 test comparing proportions of metavolcanic, porphyritic metavolcanic, and metavolcanic porphyry flakes at Mound B, all flakes included. Site Mound B Mound B Mound B Mound B SNI-106 SNI-106 Time Period 5000-4000 BP 3000-1500 BP 1500-500 BP 500 BP-Contact 3000-1500 BP 1500-500 BP Count MV 22 24 111 215 22 97 PMV 54 101 262 536 81 105 MVP 5 16 77 113 10 23 Total 81 141 450 864 113 225 Frequency % MV 27.50% 17.02% 24.67% 30.94% 19.50% 43.10% PMV 67.50% 71.63% 58.22% 77.12% 71.70% 46.70% MVP 6.25% 11.35% 17.11% 16.26% 8.80% 10.20% Time Per.(cal BP) Size Stats MV PMV MVP ?? DF p 5000-4000 All Count 22.00 54.00 5.00 14.89 6.00 0.02 Expected 19.63 50.24 11.13 AR 0.63 0.89 -2.03 3000-1500 Count 24.00 1 1.00 16.00 Expected 34.17 87.45 19.38 AR -2.10 2.47 -0.87 1500-500 Count 111.00 262.00 77.00 Expected 109.06 279.09 61.86 AR 0.25 -1.97 2.47 500-0 Count 215.00 535.00 113.00 Expected 209.14 535.23 118.63 AR 0.70 -0.02 -0.84 269 Table 6.18. Results of ?2 tests comparing proportions of metavolcanic, porphyritic metavolcanic, and metavolcanic porphyry flakes at SNI-106. To summarize material type comparisons, although the weight results indicate the highest proportion of quartzite at Mound B during the 5000-4000 cal BP period when boundary defense is predicted to be minimal, flake count results do not show significant differences between time periods. For SNI-106, however, a significantly higher proportion of quartzite in the 3000-1500 cal BP assemblage is consistent with less boundary defense around marine areas where quartzite was more abundant. The scarcity of sandstone at SNI-106 may reflect differences in manufacturing activities compared to Mound B, or more difficult access due to the greater distance between SNI-106 and sandstone sources. Results comparing MV, PMV, and MVP suggest increases in coarser grained metavolcanic rock in the 3000-1500 cal BP assemblage at Mound B when inland territoriality may have hindered raw material choice. There is also an Time Per. (cal BP) Size Stats MV PMV MVP ?? DF p 3000-1500 All Count 22 81 10 20.641 2 <.001 Expected 39.8 62.2 11 AR -4.3 4.4 -0.4 1500-500 Count 97 105 23 Expected 79.2 123.8 22 AR 4.3 -4.4 0.4 Time Per.(cal BP) Size Stats MV PMV MVP ?? DF p 3000-1500 > 3cm Count 9 28 2 9.958 2 <.001 Expected 14.3 20.2 4.4 AR -2.2 3.1 -1.5 1500-500 Count 30 27 10 Expected 24.7 34.8 7.6 AR 2.2 -3.1 1.5 Time Per. (cal BP) Flake Size Stats MV PMV MVP ?? DF p 3000-1500 < 3cm Count 13 53 8 13.808 2 0.001 Expected 25.5 41.8 6.7 AR -3.7 3.2 0.6 1500-500 Count 67 78 13 Expected 54.5 89.2 14.3 AR 3.7 -3.2 -0.6 270 increase in finer-grained metavolcanic rock at SNI-106 during the 1500-500 cal BP period when decreased inland boundary defense would have allowed more open access to any inland quarry areas and allowed people more choice in their raw material procurement. Testing Procurement Predictions on San Nicolas Island: Cobble Shape Data on original cobble shape provides additional information on change over time in procurement patterns. Based on field observations and cobble measurements from beaches and outcrops on San Nicolas Island, most areas had a mixture of round cobbles that people considered suited to their split cobble reduction sequence and flatter cobbles suited to diagonal, decapitation, and chopper reduction strategies. For this analysis, I assume that people sought both cobble types to create different flake tool forms, although one or the other may have been more efficient or favored for other reasons. Tabular metavolcanic cobbles (those with roundness values below 0.5 or above 1.5) are much rarer than round metavolcanic cobbles at all of the quarries. If people were more limited in their cobble access at 3000-1500 cal BP and 500 cal BP- Contact time periods due to inland boundary defense at Mound B, I therefore expect fewer flakes associated with diagonal and decapitation strategies. I expect that flakes associated with round and flat reduction sequences at SNI-106 should remain similar through time. Core sample sizes are too small to provide meaningful data to address this prediction. Debitage associated with different reduction sequences can be used to distinguish flat and round reduction strategies in assemblages. Each reduction strategy creates a trajectory of flake types identified by unique combinations of platform shapes, platform facets, dorsal flake scars, and cortex location. Combinations of flake attributes can also be used to categorize flakes as primary, secondary, tertiary, or middle stage flakes in the reduction sequence (Table 6.19). For 271 round cobble reduction sequences, primary (defined as 1.1 in Table 6.19) flakes associated with the Split Cobble 1 reduction sequence have round cortical platforms (6.10, 6.11) with no dorsal cortex because they are removed from the round cortical perimeter towards the center of the split face. The second set of flakes (1.2) might have a few dorsal flake scars because the toolmaker has already gone around the perimeter once. The next group of flakes (1.3) are struck from that same cortical perimeter towards the center of the split face but because multiple flakes have been struck from that surface, there are multiple dorsal flake scars. Flakes defined as 1.4 are struck from a split rotated core that is rotated 90 degrees so that the original flaking surface becomes the platform. Flakes have dihedral or multi-faceted platforms, dorsal cortex, and few dorsal flake scars. Flakes defined as 1.5 are removed in the same manner but have more dorsal flake scars as the knapper continues to work the increasingly exhausted core. For Split 2 cores, which are also associated with round cobbles, the cobble is split and the first flakes are removed using the split face as a platform around the perimeter of the cobble. First flakes (2.1) have a flat lisse platform and 100% dorsal cortex. The next set of flakes (2.2) are similar but have more dorsal flake scars because the toomaker has gone around the perimeter once. When the core is rotated and the original flaking surface is used as a platform, the next set of flakes (2.3) have lisse, dihedral, or mutli-faceted platforms and no dorsal cortex. As more flakes are removed from this surface, flakes have more dorsal flake scars (2.4). The flat cobble reduction sequences produce different flake from the split cobble sequences, although there is some overlap between flakes created using different core reduction strategies. The first flakes removed during the diagonal reduction sequence should have round cortical platforms with one dorsal flake scar and a step termination (3.1). As the toolmaker continues to work from the cortical platform using the ridge of previous flake scars to direct the 272 force of the blow, the next set of flakes (3.2) will be similar to the first flakes but will have more dorsal flake scars. If the core is rotated, the platform may come from the flatter side of the cobble (3.3). First flakes removed the decapitate or chopper reduction sequences have round cortical platforms and dorsal cortex with no flakes scars (Figure 6.12, 6.13; 4.1, 5.1). For the decapitate sequence, the next set of flakes removed from the cortical flat platform adjacent to the initial flaking surface (4.2) follow the original flake scar and therefore have one flake scar. The next sets of flakes (4.3, 4.4) have more flake scars. The second set of flakes (5.2) removed during the chopper reduction sequence are struck the original flake scar (platform) and therefore have a flat lisse platform and dorsal and lateral cortex. The third set of flakes have less dorsal cortex because they are flaked onto a previously flaked surface surface (5.3) and the next set of flakes have more dorsal flake scars as the toolmaker continues to remove flakes, alternating sides of the cobble (5.4). I do not expect that these flake types are absolutely diagnostic of each reduction sequence due to overlap and idiosyncrasies of the knapping process. In some cases, flake attributes associated with two different reduction techniques overlap. Split Cobble 1 tertiary flakes have round cortical platforms with cortex located on the platform and a lateral margin, and have multiple flake scars. These attributes also define Diagonal reduction sequence tertiary flakes. Characteristics of secondary, tertiary, and later stage flakes for each of these reduction sequences are described in the table below. The table describing the flake types for round and split cobble reduction sequences provides a general overview of flakes more likely to be associated with rounder and flatter reduction sequences, and a place to start in assessing change through time in use of cobbles of different shapes. 273 Figure 6.10. Flake with a round platform, Figure 6.11. Flake with a round platform obverse (25B.13.2A.3.F.4). reverse (25B.13.2A.3.F.4). Figure 6.12. Decapitation flake from Figure 6.13. Decapitation flake from decapitate or chopper reduction sequence, decapitate or chopper reduction sequence, obverse (25B.58.IIB.1.F.5). reverse. 274 Table. 6.19. Primary, secondary, tertiary, and later stage flakes associated with different San Nicolas Island core reduction strategies. 1 2 3 4 5 Split 1 Platform Shape Round Round Round Flat Flat Type 1 Platform Facet Cortical Cortical Cortical Dihedral, Multi Multi Cortex Location Platform Platform/Lateral Platform/Lateral Dorsal Dorsal, Margins, or None Dorsal Flake Scars 1 2, 3 4+ 0-2 3+ Termination Feather Feather/Hinge Overlap 3.3 2.4 Split 2 Platform Shape Flat Flat Flat Flat Type 2 Platform Facet Lisse Lisse Lisse, Dihedral, Multi Dihedral, Multi Cortex Location Dorsal Dorsal None None Dorsal Flake Scars 0 1+ 1 2+ Overlap 5.2 1.5 Diagonal Platform Shape Round Round Flat/Round Type 3 Platform Facet Cortical Cortical Cortical Cortex Location Platform/L ateral Platform/Lateral Platform/Lateral Dorsal Flake Scars 1 2,3 4+ Termination Step Step Overlap 4.4, 1.3 Decapitate Platform Shape Round Round/Flat Flat Flat Type 4 Platform Facet Cortical Cortical Cortical Cortical Cortex Location Platform, dorsal, lateral Platform and Lateral/Distal Platform/Lateral/ Distal Platform/L ateral/Dist al Dorsal Flake Scars 0 1 2 3+ Termination Feather Overlap 5.1 3.3 Chopper Platform Shape Round Flat Flat Flat Type 5 Platform Facet Cortical Lisse Lisse Lisse Cortex Location Platform and dorsal, lateral Dorsal/Lateral Lateral Lateral Dorsal Flake Scars 0 0 1 2+ Overlap 4.1 2.1 275 Overall, a total of 423 flakes from dated strata were identifiable to a particular reduction sequence at Mound B with 23 of those from overlapping categories. A total of 47 flakes were identifiable at SNI-106 with 3 of those from overlapping categories (Table 6.20). Table 6.20. Flake type counts at Mound B and SNI-106. Results of ?? tests comparing proportions of round (Split Cobble) and flat (Diagonal, Decapitate, and Chopper) reduction sequence debitage in the 1500-500 cal BP and 500 cal BP- Contact assemblages showed no significant differences through time (?? 1.04, 1 df, p = 0.31). This was also the case for SNI-106 for 3000-1500 and 1500-500 cal BP assemblages (?? .029, 1 df, p = 0.864). These results are not consistent with the prediction that there should be fewer flat reduction sequence flakes at 3000-1500 cal BP and 500 cal BP - Contact. Flake Type 5000-4000 3000-1500 1500-500 500-0 3000-1500 1500-500 1.1 1 1 5 11 2 3 1.2 3 3 10 22 2 5 1.3 3 7 17 1 5 1.4 1 2.2 4 7 2.3 3 5 2.4 1 1 1 7 5 13 3.1 2 3.2 1 1 2 4.2 5 2 22 48 4.3 4 8 18 64 4.4 3 8 52 53 5.1 1 5.3 1 1.3/3.3 2 1.3/3.4 2 1 1.5/2.4 1 2 2.1/5.2 1 1 5 3.3/4.4 4 5 4.1/5.1 1 1 Time Period (Cal BP) Mound B SNI-106 276 A comparison between Mound B and SNI-106 showed significantly different proportions of round and flat reduction flakes during the 1500-500 cal BP period. There were significantly more flat cobble reduction sequence flakes at Mound B and significantly more round cobble reduction sequence flakes at SNI-106 (Table 6.21). This suggests that the inhabitants at SNI-106 had access to cobble areas with more flat cobbles, which is consistent with my prediction. I cannot rule out the possibility that they preferred flat cobbles to create their tools or had different manufacturing processes at SNI-106 than elsewhere on the island. Table. 6.21. Results of a ?? test comparing proportions of round and flat reduction sequence debitage at Mound B and SNI-106 at 1500-500 cal BP. Toolstone Procurement on San Nicolas Island: Processing Prior to Transport Change over time in toolstone processing decisions provides an important source of data to further investigate procurement behaviors on San Nicolas Island. I use proportions of first flakes compared to later stage flakes as an indicator of processing prior to transport. Based on my predictions, the analysis should show minimal testing of metavolcanic toolstone at Mound B at 3000-1500 cal BP and 500 cal BP - Contact when increased territorial circumscription encouraged procurement near the habitation site. For SNI-106, all toolstone should show similar moderate processing prior to transport. If people were never constrained to procuring metavolcanic rock from near the site, people should always test and process prior to transport to ensure that the costs of transport were outweighed by the utility of the stone. For both sites, since Tim Per. (cal BP) Stats Round Flat ?? DF p Mound Count 34 109 31.92 1 <.001 Expected 50.16 92.84 AR -5.65 5.65 SNI-106 Count 33 15 Expected 16.84 31.16 AR 5.65 -5.65 277 quartzite cobbles were more abundant at a distance from the habitation sites and would have required negotiation with other groups to access the source, I expect to see testing and the earliest stages of processing prior to transport during all time periods, but particularly at 1500- 500 cal BP when marine territoriality would have made coastal resources more difficult to access for inland groups. Bringing back a higher-utility toolstone load would ensure less frequent visits into unfriendly territory. Early stage reduction of metavolcanic and quartzite cobbles prior to transport should be evidenced by a small number of first and second flakes relative to later stage flakes and a higher expected to actual cortical surface ratio at the habitation site. To examine the representation of reduction sequence at Mound B and SNI-106, I use the flake types for each reduction sequence described in Table 6.24. I group the primary flakes from each reduction sequence (first flakes struck from a core), the secondary flakes from each reduction sequence (the second or third set of flakes struck from a core, and the tertiary flakes from each reduction sequence (the next identifiable set of flakes struck from a core). Results of an ANOVA test for the Mound B assemblage demonstrates significant differences in both mean weight and size class (flake maximum dimension grouped by 1 cm intervals) for secondary and tertiary flakes. The mean for tertiary flakes is significantly higher than the mean secondary flakes for both measurements (Table 6.22). This result is consistent with San Nicolas Island lithic reduction sequences that do not progress from large to small flakes. Along with primary, secondary, and tertiary flakes, I also analyze the dorsal flake scars and percent cortex because these variables are strongly associated with reduction stage (Table 6.23). 278 Table 6.22. Results of ANOVA tests comparing mean weight and size for primary, secondary, and tertiary flakes. Table 6.23. Attributes related to reduction for Mound B and SNI-106. Results of reduction sequence analysis at Mound B and SNI-106 are consistent with some but not all of the predictions. Proportions of tertiary flakes are significantly higher during the Dependent Variable Reduction Stage n Mean F Sig. x Difference Std. Error Sig. Weight 1 23.00 7.84 7.85 0.00 1 2 0.97 2.78 1.00 3 -4.36 2.87 0.39 2 252.00 6.87 2 1 -0.97 2.78 1.00 3 -5.33417* 1.35 0.00 3 138.00 12.2 3 1 4.36 2.87 0.39 2 5.33417* 1.35 0.00 Siz Class 1 23.00 4 7.95 0.00 1 2 0.09 0.32 1.00 3 -0.52 0.33 0.34 2 252.00 3.91 2 1 -0.09 0.32 1.00 3 -.61301* 0.15 0.00 3 138.00 4.52 3 1 0.52 0.33 0.34 2 .61301* 0.15 0.00 Reduction Stage Comparison Bonferroni Post-Hoc Analysis 1 2 3 0 1 2 3 4+ Cortical Lisse Dihedral Multi Crushed 0% 1-25% 25-50% 50-100% Mound B (Cal BP) 5000-4000 cal BP Metavol. 2 9 3 0 12 11 11 11 19 10 0 0 1 10 15 3 3 Quartzite 0 3 0 0 3 6 1 2 1 0 1 0 4 2 1 0 3000-1500 Metavol. 1 12 10 3 18 16 14 20 35 11 1 3 0 12 25 10 2 Quartzite 0 2 0 1 0 4 1 1 2 2 0 0 0 2 1 1 0 1500-500 Metavol. 1 2 2 14 61 71 62 75 104 49 2 1 2 43 83 23 2 Quartzite 2 9 10 2 12 12 16 11 26 8 0 0 1 6 16 12 1 500-0 Metavol. 1 125 55 26 157 171 95 111 209 81 6 2 10 74 157 63 8 Quartzite 1 25 7 8 19 35 20 8 35 7 1 0 0 9 25 13 4 SNI-106 (Cal BP) 3000-1500 Metavol. 1 5 7 0 10 15 9 15 15 7 1 0 0 7 9 6 1 Quartzite 1 0 3 0 2 4 4 6 5 1 1 0 0 2 5 0 0 1500-500 Metavol. 2 19 23 0 17 31 23 20 31 29 2 2 0 31 15 18 0 Quartzite 1 3 1 0 3 3 0 3 6 2 0 0 0 2 1 5 0 % Cortex 1 Platform FacetDorsal ScarsReduction 1 Includes only flakes with platforms because % cortex is estimated based on total possible cortical area on platform and dorsal surface. 279 1500-500 cal BP compared to the 500 cal BP-Contact periods at Mound B both if all flakes are considered and if only metavolcanic flakes are considered. More later stage reduction at 1500- 500 cal BP is consistent with procurement from a broader geographical area and processing prior to transport. That there are no significant differences in flake types for SNI-106 is consistent with the prediction that procurement of metavolcanic rock should change little over time at that site. Results of ?2 tests comparing attribute states associated with reduction stage do not indicate statistically significant differences at Mound B or at SNI-106. Sample size for quartzite flakes is too small to test most of the predictions (Table 6.24, 6.25, 6.26). The small number first flakes at the site relative to other kinds of debitage is unsurprising because there is only one first flake per core. Table 6.24. Results of ?2 tests comparing proportions of flake types and attributes associated with earlier and later stage reduction at Mound B and SNI-106. Site Variable Material Time Periods (Cal BP) ? 2 df p MoundB Flake Typ All 1500-500, 500-0 13.16 2 <0.001 Mou dB Metavolcanic 1500-500, 500-0 9.42 2 0.01 MoundB Quartzite 1500-500, 500-0 *NA SNI-106 Metavolcanic 3000-1500, 1500-500 0.25 2 0.89 MoundB Dorsal Scars All All 14.36 9 0.11 Moun B Metavolcanic All 13.76 9 0.13 MoundB Quartzite 1500-500, 500-0 3.86 3 0.28 SNI-106 Metavolcanic 3000-1500, 1500-500 1.78 2 0.62 MoundB % Cortex All 1500-500, 500-0 1.51 3 0.68 MoundB Metavolcanic 1500-500, 500-0 *NA MoundB Quartzite 1500-500, 500-0 *NA SNI-106 Metavolcanic 3000-1500, 1500-500 2.8 2 0.25 *NA = sample size too small 280 Figure 6.25. Results of a ?2 test comparing proportions of flake types at 1500-500 cal BP and 500 cal BP-Contact at Mound B, all material included. Figure 6.26. Results of a ?2 test comparing proportions of flake types at 1500-500 cal BP and 500 cal BP-Contact at Mound B, only metavolcanic rock included. Toolstone Conservation on San Nicolas Island The predictions of the territoriality hypotheses for San Nicolas Island indicate that increased terrestrial territorial circumscription at 3000-1500 cal BP and 500 cal BP-Contact should encourage increased conservation of toolstone at Mound B. At 5000-4000 cal BP and 1500-500 cal BP, free access to resources across the landscape would require minimal conservation. Marine territoriality at 1500-500 cal BP would require conservation of quartzite and sandstone because they would have been more difficult to obtain. At SNI-106 there should be minimal conservation of metavolcanic rock at 3000-1500 cal BP or 1500-500 cal BP due to minimal boundary defense. As at Mound B, quartzite should be conserved, particularly at 1500- 500 cal BP. Time Per. (cal BP) Material Stats 1 2 3 ?? DF p 1500-500 All Count 5.00 62.00 57.00 13.2 2 0.001 Expected 6.81 75.63 41.56 AR -0.88 -3.09 3.62 500-0 Count 15.00 160.00 65.00 Expected 13.19 146.37 80.44 AR 0.88 3.09 -3.62 Time Per. (cal BP) Material Stats 1 2 3 ?? DF p 1500-500 Metavolcanic Count 3.00 51.00 45.00 13.2 2 0.001 Expected 5.74 59.47 33.79 AR -1.45 -2.14 2.92 500-0 Count 14.00 125.00 55.00 Expected 11.26 116.53 66.21 AR 1.45 2.14 -2.92 281 Evidence to test these predictions comes from core, flake, and formal tool assemblages. Given the small number of cores per the amount of debitage at the site, many cores were probably smashed into shatter during bipolar reduction. In all of the assemblages, exhausted cores and fragments by far outnumber patterned cores. Because toolstone is so difficult to work on San Nicolas Island, people likely used bipolar reduction throughout the reduction sequence and particularly in the early stages of reduction to split open the cores and at the end of the reduction sequence to create useful flakes from angular and exhausted cores. In the 1500-500 cal BP assemblage, 19 of 22 cores are fragments or exhausted. In the 500 cal BP-Contact assemblage, 18 of 27 cores are fragments or exhausted. There are too few quartzite cores (n = 2 in the 1500-500 cal BP assemblage and n = 3 in the 500 cal BP-Contact assemblage) to compare them to the metavolcanic cores. The small number of cores relative to flakes for all of the assemblages may indicate that all were worked to exhaustion, consistent with toolstone conservation during all time periods. To further investigate change over time in bipolar reduction of exhausted cores, I test the flake assemblage for change over time in debitage associated with bipolar manufacture. One potential indicator of flake production from exhausted or later-stage cores are non-cortical multi- faceted platform flakes with multiple dorsal flake scars; however, there were fewer than 10 of these flakes in each assemblage at Mound B. For SNI-106, a significantly higher proportion of these flakes were identified in the 3000-1500 cal BP assemblage than the 1500-500 cal BP assemblage if all material types are considered (Table 6.27). 282 Table 6.27. Results of ?2 tests comparing proportions of exhausted core flakes between time periods at SNI-106. If people were using cores to exhaustion and then smashing them on an anvil to create additional usable flakes, this should be indicated by an increase in non-cortical flakes with bipolar features and non-cortical shatter. Bipolar features include a crushed platform, a crushed distal end, breakage parallel to the direction of force, and flaking on the ventral face. This analysis must be considered tentative because none of these features alone guarantee that a flake was created through bipolar reduction, and there are only eight cases where flakes have two bipolar attributes. Almost half of the flakes with bipolar attributes have only breakage parallel to the direction of force. At Mound B, there is no significant differences in flakes with bipolar attributes between time periods (?? 5.40, 3 df, p = 0.145). Sample size is too small for quantitative comparison for SNI-106. Non-cortical shatter < 3 cm in maximum dimension and lacking an identifiable dorsal or ventral face is another potential indicator of bipolar reduction of exhausted cores. At Mound B, a ?2 test comparing proportions of shatter and debitage indicates a significantly low proportion of shatter in the 1500-500 cal BP assemblage both if all toolstone is considered and if only metavolcanic toolstone is considered. For metavolcanic flakes, there is also a significantly higher Material Time Per. (cal BP) Stats Exhausted core flakes Other ?? DF p All 3000-1500 Count 27.00 144.00 5.70 1.00 0.02 Expected 19.31 151.69 AR 2.39 -2.39 1500-500 Count 22.00 241.00 Expected 29.69 233.31 AR -2.39 2.39 Materi l Time Per. (cal BP) Stats Exhausted core flakes Other ?? DF p Metavolcanic 3000-1500 Count 18.00 95.00 3.21 1.00 0.07 Expected 13.04 99.96 AR 1.79 -1.79 1500-500 Count 21.00 204.00 Expected 25.96 199.04 AR -1.79 1.79 283 proportion of shatter in the 500 cal BP-Contact assemblage (Table 6.28). There is no significant difference in non-cortical shatter between 1500-500 cal BP and 500 cal BP-Contact time periods for quartzite (?2 = 0.11, 1 df, p = 0.74). Table 6.28. Results of ?2 tests comparing proportions of non-cortical shatter and other flakes at Mound B. In contrast to Mound B, at SNI-106 there are no significant differences in proportions of non-cortical shatter whether all material types are included or if only metavolcanic flakes are included (Table 6.29). Comparing quartzite and metavolcanic shatter at SNI-106 and Mound B, there is no significant difference between the two material types for the 1500-500 cal BP time period (?2 = 0.092, 1 df, p = 0.761) or during the 500 cal BP-Contact time period (?2 = 1.58, 1 df, p = 0.21). Results for non-cortical shatter for Mound B are consistent with decreased non-cortical shatter and potentially toolstone conservation at 1500-500 cal BP. At 500 cal BP, non-cortical Time Per. (cal BP) Material Stats Shatter Other ?? DF p 5000-4000 All Count 30.00 86.00 11.44 3.00 0.01 Expected 27.09 88.91 AR 0.66 -0.66 3000-1500 Count 47.00 148.00 Expected 45.54 149.46 AR 0.26 -0.26 1500-500 Count 114.00 503.00 Expected 144.10 472.90 AR -3.36 3.36 500-0 Count 338.00 999.00 Expected 312.26 1024.74 AR 2.60 -2.60 Time Per. (cal BP) Material Stats Shatter Other ?? DF p 5000-4000 Metavolcanic Count 20.00 61.00 10.96 3.00 0.01 Expected 17.25 63.75 AR 0.76 -0.76 3000-1500 Count 27.00 117.00 Expected 30.67 113.33 AR -0.78 0.78 1500-500 Count 74.00 378.00 Expected 96.27 355.73 AR -3.00 3.00 500-0 Count 230.00 741.00 Expected 206.81 764.19 AR 2.84 -2.84 284 shatter increases. This may fit with the prediction that that people had to conserve toolstone at 500 cal BP due to territorial circumscription. For SNI-106 there are no significant differences in proportions of shatter over time (the p value is significant but adjusted residuals do not show significant differences between time periods), which is consistent with a hypothesis of no change in boundary defense strategies. An increases in flakes from exhausted cores at 3000-1500 cal BP is not consistent with that prediction, but this may result from a small sample size. Table 6.29. Results of ?2 tests comparing proportions of non-cortical shatter and other flakes at SNI-106. Another index of conservation to consider is intensity of retouch of formal tools. I predicted that intensity of retouch of formal tools should be higher at 500 cal BP-Contact, but the small sample sizes make it difficult to test that prediction. For drills, 2 of 15 tools have retouch flaking at 1500-500 cal BP and 1 of 12 has retouch flaking at 500 cal BP-Contact. For scrapers, most show minimal retouch limited to part of one margin with short flake scars. There are a few instances where the flake scars are present on an entire margin and are long, invasive, or covering. Approximately half of the scrapers show evidence of use. In the 1500-500 cal BP period, all scrapers have retouch on only one margin and only 2 of 10 have more than partial retouch on a margin. In the 500 cal BP-Contact period, 2 scrapers of 14 have retouch on more Material Time Per. (cal BP) Stats Non-cortical shatter Other ?? DF p All 3000-1500 Count 58.00 113.00 2.73 1.00 0.10 Expected 66.19 104.81 AR -1.65 1.65 1500-500 Count 110.00 153.00 Expected 0.66 0.42 AR 1.65 -1.65 Materi l Time Per. (cal BP) Stats Non-cortical shatter Other ?? DF p Metavolcanic 3000-1500 Count 34.00 79.00 3.18 1.00 0.07 Expected 41.46 71.54 AR -1.78 1.78 1500-500 Count 90.00 135.00 Expected 82.54 142.46 AR 1.78 -1.78 285 than one margin, and 4 have retouch on entire margins rather than part of a margin. These patterns indicate a potential increase in retouch during the later period at Mound B. To summarize results for conservation, I predicted increased conservation at 3000-1500 cal BP and 500 cal BP-Contact as a result of inland territoriality at Mound B. I also predicted an increase in conservation of quartzite at 1500-500 cal BP at both Mound B and SNI-106 as a result of more difficult access to coastal cobble areas. Tests of these predictions using exhausted cores, bipolar reduction of exhausted cores, and retouched tools generally did not support this prediction. The low abundance of cores and high proportion of exhausted cores and core fragments for assemblages for all time periods from both sites indicates either some degree of conservation at all times, or that smashing cores was part of the technology. There was no significant difference in non-cortical flakes associated with exhausted cores from Mound B. The significantly lower proportion of non-cortical shatter from Mound B at 1500-500 cal BP is also indicates decreased conservation efforts at that time. The lack of increase in non-cortical shatter at SNI-106 is consistent with unchanged boundary defense strategies that would have maintained similar lithic procurement strategies. Quartzite samples were too low to test most of the predictions. There was no significant difference in proportion of metavolcanic and quartzite non- cortical shatter at 1500-500 cal BP and 500 cal BP-Contact. Regarding formal tools, data on scraper retouch indicate a potential increase in retouch during the 500 cal BP-Contact period, but drill retouch does not appear to change over time. Exchange on San Nicolas Island Toolstone procurement, processing, and conservation are associated with boundary defense. Boundary permeability is an equally important component of a territoriality strategy that 286 I investigate using evidence for informal and formal exchange on San Nicolas Island. Based on a social network model hypotheses, I suggest that at 5000-4000 cal BP, exchange should be infrequent among resource-secure and sparsely distributed communities. At 3000-1500 cal BP and 500 cal BP to contact, increased boundary defense in the San Nicolas Island interior should be associated with lower boundary permeability. Chert and quartz should be rare at Mound B but the flakes and tools that are present should be distributed in clusters reflecting formal relationships between households and kin in other groups. For the 1500-500 cal BP assemblage, increased boundary permeability associated with a more resource poor inland environment should be associated with an increase in chert and quartz across the site and a more even distribution of these materials. I test predictions about exchange using extra-local chert, mainly Franciscan and Monterey banded chert that would have traded between islands and within groups on the island. It is fine-grained and highly visible. As well as being useful for creating retouched tools it was also likely a toolstone that carried symbolic weight. The lack of cores made from non-local material suggests that chert was imported as tools or blanks (Ta?kiran 2001). I also consider quartz as a possible inter-island trade material. Available in small veins and outcrops on the island, it may have been accessible on a variety of beaches but is also a highly visible and rare material. Quartz crystals were used for ceremonial purposes (Bartelle et al. 2010). Because there were few chert and quartz flakes > 3 cm at either site, I was not concerned that a large number of small flakes might be skewing the counts so I did not separate the larger and smaller size fractions. Consistent with expectations for small groups that interacted widely across territorial boundaries to maintain ties with friends and marriage partners, results of ?2 tests results show a 287 significantly higher proportion chert flakes compared to metavolcanic flakes during the 5000- 4000 cal BP period at Mound B. With a total of 81 metavolcanic flakes, 9 chert flakes makes up a larger proportion than for any of the other sub-assemblages even though the sample size itself is not larger (Table 6.30). If that earliest assemblage is removed, differences in proportion of chert flakes between time period are not significantly different (?2 = 2.259, 2 df, p = 0.32). Also at Mound B there is a higher amount of quartz in the 3000-1500 cal BP assemblage (Table 6.36). At SNI-106, there are no significant differences in chert between time periods (?2 = 0.73, 1 df, Fisher?s p = 0.41) but there is a significantly higher proportion of quartz at 3000-1500 cal BP (Table 6.31). Results showing more extralocal and rare material at 3000-1500 cal BP are not in accord with predictions for decreased boundary permeability during this time period. Table 6.30. Results of ?2 tests comparing proportions of chert and quartz to metavolcanic flakes at Mound B . Time Per. (cal BP) Stats Chert Metavolcanic ?? DF p 5000-4000 Count 9.00 81.00 10.62 3.00 0.01 Expected 3.67 86.33 AR 2.92 -2.92 3000-1500 Count 8.00 144.00 Expected 6.19 145.81 AR 0.78 -0.78 1500-500 Count 13.00 452.00 Expected 18.95 446.05 AR -1.63 1.63 500-Contact Count 40.00 971.00 Expected 41.19 969.81 AR -0.30 0.30 Time Per. (cal BP) Stats Quartz Metavolcanic ?? DF p 3000-1500 Count 12.00 144.00 9.88 2.00 0.01 Expected 6.58 149.42 AR 2.27 -2.27 15 0-500 Cou t 10.00 452.00 Expected 19.49 442.51 AR -2.59 2.59 500-Contact Count 47.00 971.00 Expected 42.94 975.06 AR 1.03 -1.03 288 Table 6.31. Results of a ?2 test comparing proportions of quartz to metavolcanic flakes at SNI-106. Regarding the spatial distribution of chert and quartz I found too many unknowns in controlling for spatial and temporal issues to draw firm conclusions. There are only two units that date to the 1500-500 cal BP period, thus spatial patterning cannot be assessed. For the 500 cal BP-Contact period, chert and quartz artifacts are relatively evenly spread throughout the units with slightly of these artifacts more found in Unit 43 (n = 30) than in the other units. San Nicolas Island Lithic Procurement Summary To examine lithic procurement patterns to investigate change over time in territoriality, I considered access to different material types, access to different shaped cobbles, processing prior to transport, conservation, and exchange. I predicted minimal boundary defense and high boundary permeability at 5000-4000 cal BP, inland boundary defense at 3000-1500 cal BP and 500 cal BP-Contact, and marine boundary defense at 1500-500 cal BP. During times of inland boundary defense, people at Mound B and SNI-106 would have been restricted in their procurement opportunities to the area around the site where quartzite, sandstone, and flatter cobbles are rarer. In general, their toolstone options would have been more limited. During the 1500-500 cal BP period they could have procured cobbles from a variety of inland sources, but Time Per. (cal BP) S ats Quartz Metavolcanic ?? DF p 3 00-1500 Count 20.00 113.00 10.73 1.00 <0.001 Expected 11.5 121.50 AR 3.28 -3.28 1500-500 Count 12.00 225.00 Expected 20.50 216.50 AR -3.28 3.28 289 access to quartzite and sandstone would still have been limited, perhaps even more so, by boundary defense by people who lived in marine areas. In testing these scenarios, I found that small sample sizes for the earliest assemblage at 5000-4000 cal BP and for quartzite and sandstone artifacts made it difficult to fully address all of my predictions. The rarity of quartzite and sandstone at both Mound B and Tule Creek Village is consistent with either restricted ability to access that material or use of nearby toolstone simply because rocks are heavy and convenience was more important than material type. If shifts in territoriality occurred during the time periods when I predicted, and these shifts truly determined resource access, at least most of the indices that I used to distinguish between procurement strategies should have shown some kind of significant difference, even if it was not in the predicted direction. Instead, most of the indices do not show significant differences between time periods. For the 1500-500 cal BP period, more of the indices of procurement and manufacture support the predictions than during the other periods. At Mound B, there is a decrease in PMV and at at SNI 106, there is an increase in MV. This may reflect less restriction on resource access and increased testing of cobbles from inland sources. There are also more tertiary flakes at Mound B, which would be consistent with more processing prior to transport. The decrease in non-cortical shatter indicates decreased conservation efforts. For the other time periods, however, the data largely show no significant differences or are not in accord with the predictions. There are also few significant differences in chert and quartz that would support changes in boundary permeability during the time periods predicted. The incongruence in results for Mound B and SNI-106 provides additional evidence that my predictions are not supported by the data. For example, during the 1500-500 cal BP period, there is an increase in finer-grained MV toolstone at SNI-106 and a decrease in PMV toolstone at 290 Mound B. Flake types associated with flat cobbles increase at Mound B and those associated with round cobbles increase at SNI 106. The cortex ratio is low at Mound B and high at 106. Non-cortical shatter increases at Mound B and decrease at SNI-106. All of these differences may simply reflect different manufacturing activities at the two sites or sampling error; however, greater similarities for the two sites would support a more important role for lithic procurement in determining those manufacturing traditions. An Alternative Hypothesis for Lithic Procurement on San Nicolas Island To further address my research questions, I evaluate an alternative hypothesis that boundary defense operated beyond the scale of the village. As a preliminary test, I compare lithic artifacts from strata/levels that date to the 500 cal BP-Contact time period in two separate spatial areas that may correspond to separate households: (1) Unit 58, Stratum I, Stratum II levels 1-2, and (2) Unit 43, Stratum I and II. Results of ?? tests comparing proportions of metavolcanic flakes and sandstone, chert, and quartz indicate significantly higher proportions of sandstone, chert, and quartz in Spatial Unit 1 and quartzite in Spatial Unit 2 (Table 6.32). A comparison of tools from the two units also reveals significant differences, but toolstone distribution does not mirror the flake results. In Spatial Unit 1 there are 12 retouched tools, none of which are chert. The tools consist of scrapers, drills, and retouched flakes. There are also two ground stone beads. In Spatial Unit 2 there are also 12 tools but material type includes not just metavolcanic toolstone but also quartz and chert. There are no beads in this unit, and there is one projectile point. Further research is required to distinguish between manufacturing activity differences and resource procurement differences. 291 Table 6.32. Results of ?2 tests comparing proportions of material types for two spatial areas at Mound B. Chapter Summary The results of lithic analysis for Watmough Bay, Mound B, and SNI-106 demonstrate both the potential and the challenges of using a lithic procurement approach to examine territorial behavior in coastal settings. In both study areas, the question became, were there time periods when people used more toolstone from near the site than from elsewhere on the landscape? If they relied on toolstone collected near the site, is this because there was territorial circumscription associated with increased boundary defense? For Watmough Bay, at least during the 1600-1000 cal BP period, the data indicate that people did not primarily procure toolstone from the beach near the site. The size and shape of Spatial Unit Stats Metavolcanic Quartzite ?? DF p 1 Count 181 21 14.56 1 <0.001 Expected 164.17 37.83 AR 3.82 -3.82 2 Count 266 82 Expected 282.83 65.17 AR -3.82 3.82 Spatial Unit Stats Metavolcanic Sandstone ?? DF p 1 Count 181 16 8.94 1 0.003 Expected 187.76 9.24 AR -2.99 2.99 2 Count 266 6 Expected 259.24 12.76 AR 2.99 -2.99 Spatial Unit Stats Metavolcanic Chert/Quartz ?? DF p 1.00 Count 181.00 14.00 1.26 1.00 0.26 Expected 177.53 17.47 AR 1.12 -1.12 2.00 Count 266.00 30.00 Expected 269.47 26.53 AR -1.12 1.12 292 cores and cortex appearance for cores and flakes is more similar to beaches on southern San Juan Island than on southern Lopez Island, a relatively easy boat ride away. There are several different ways to explain procurement in this context. People from Watmough Bay could have visited southern San Juan Island to collect subsistence or water resources or meet with kin. If Watmough was used seasonally and families lived in other areas during summer or spring months as indicated in the ethnographic record, they could have collected larger and more angular cobbles in these areas to bring them back to Watmough Bay in the winter, knowing that this resource would be limited. The distribution of chert, quartz, and incised shale suggests little change through time, minimal use of these materials, and a relatively even distribution across the site. My ability to investigate change through time in boundary permeability at Watmough Bay was limited due to the small sample of extra-local and rare toolstone. Other material types must be considered to further address this aspect of my research question. For the San Nicolas Island case study, the lithic landscape is more complex and so are the predictions. The lithic record does not indicate that people were limited in their procurement opportunities to Tule Creek at 3000-1500 and 500 cal BP-Contact. There is slightly more evidence for procurement from a variety of toolstone sources and decreased toolstone conservation at Mound B at 1500-500 cal BP, but the record for SNI-106 shows the opposite pattern. It would have been interesting to compare in more detail differences between sandstone, quartzite, and metavolcanic procurement practices, but the smaller sample of quartzite and sandstone made that difficult. The weight of the evidence is against boundary defense playing a significant role in toolstone procurement on San Nicolas Island. Similarities in the distribution of chert and quartz over time do support change over time in boundary permeability expressed through changes in exchange patterns. In combination with the settlement pattern data, the lithic 293 data do not support boundary defense or permeability at the level of the village on San Nicolas Island during the Late Holocene. My preliminary analysis evaluating differences in toolstone access at different parts of the site at 500 cal BP-Contact reveals different proportions of material types, but flake toolstone and retouched tool toolstone are dissimilar at different parts of the site. Further research on lithics in both areas would be required to disentangle manufacturing and procurement. Analysis of the stratigraphy and spatial distribution of other artifacts would be required to assess potential separation of households at Mound B. 294 Chapter 7: Conclusions In this research, I explored territorial behavior in two study areas on the Pacific Coast, the San Juan Islands and southern Channel Islands. By testing hypotheses drawn from human behavioral ecology models using settlement pattern and lithic procurement data, I investigated whether territorial behavior occurred at the scale of the village in response to shifts in marine and terrestrial resources. A comparative perspective facilitated rigorous hypothesis testing and contributes to a more detailed understanding of coastal settlement patterns, lithic procurement strategies, and technologies. Investigating change over time in territorial behavior on the Pacific Coast is fundamental to larger questions about the development of social, political, and economic inequalities in semi-sedentary communities, human-environment interactions, and human response to climate change. This study centered on boundary defense and permeability strategies as an adaptation to resource abundance and predictability. Based on economic defensibility models, I hypothesized that when subsistence resources far exceed or fall far short of a community?s needs, the costs of defending a territory will outweigh the benefits. When subsistence resources are abundant and predictable enough to just adequately fulfill resource needs, people should defend small territories around villages that are located near productive resource patches. Due to temporal and spatial fluctuations in the environment, some people will always have less and will therefore attempt to encroach on others? territory. As a result, resource procurement near productive patches?including procurement of toolstone?is restricted to smaller areas. During times of resource scarcity, people should invest more effort toward maintaining inter-village ties to buffer against resource shortfall; during times of adequate abundance they should invest less effort 295 toward exchange and other inter-group activities due to higher costs of crossing aggressively defended boundaries and lower benefits of risk management. When I used this basic hypothesis to create scenarios for the development of territorial behavior during the Late Holocene in the San Juan Islands and southern Channel Islands, I considered shifts in marine and terrestrial environments, the locations of habitation sites relative to those resources, and settlement pattern information. As an alternative hypothesis, I suggested that village boundaries may have always been permeable to resource acquisition by multiple groups due to strong kin relationships, marriage, and friendship ties between villages. Because of the spatially and temporally unpredictable nature of coastal resource distribution and/or social and cultural phenomena that are not affected by subsistence strategies, families within villages may have maintained access to specific resource areas and shared access or resources?including toolstone?with other related families. I tested predictions specific to the archaeological record of both study areas using data on the defensive characteristics of archaeological sites and lithic procurement patterns. Landscape and site scale data provided two different perspectives on potential shifts in boundary defense and permeability. For a landscape perspective on boundary defense around villages, I analyzed defensive properties of sites from across both study areas. Built earthworks are almost non- existent in the Gulf of Georgia and southern Channel Islands, but choice of site location also reveals strategies for defense against unwanted visitors. I measured site defensiveness based on visibility, elevation, and distance to lookouts for all dated sites on the San Juan Islands and measured elevation and distance to lookouts for all dated sites on San Nicolas Island. I then compared defensive properties of sites between time periods associated with different environmental regimes. 296 Site-scale data on toolstone procurement provides an additional perspective on territorial behavior. The majority of my dissertation analysis focuses on toolstone access at Watmough Bay on Lopez Island and Tule Creek Village Mound B and SNI-106 on San Nicolas Island. Because toolstone sources cannot be pinpointed to specific geographic locations in either study area, I compared proportions of material types, cobble shapes, cortex appearance, and other features of the lithic assemblages with cobbles from potential toolstone collection areas both adjacent to the sites and elsewhere on the landscape. I investigated whether the lithic assemblages matched the toolstone available in the vicinity of the site during periods when I predicted smaller and more aggressively defended territories. In exploring toolstone procurement, processing, and conservation at each site, I considered boat and pedestrian transport, reduction sequences for cobble-centered technologies, and lithic traditions characterized mainly by unmodified flake tools. Summary of Results for the San Juan Islands For the San Juan Islands case study, neither defensive site data nor lithic procurement analysis were consistent with predicted changes through time in boundary defense and permeability at the scale of the village. I predicted that at 600 cal BP-Contact, adequately productive marine resources and higher populations would have encouraged communities near productive resource patches to defend the area surrounding the village to maintain an adequate resource supply. Boundary defense should be lower during periods of lower population (3500- 2500 cal BP and 2500-1600 cal BP) when resources would have been extremely abundant relative to the population and at 1600-1000 cal BP when a drop in marine productivity would have made resources too scarce to be worth defending in most areas. Boundary permeability 297 should be highest at times of resource scarcity. Further analysis is required to determine whether the incongruence between the data and the predictions is attributable to stability in territorial behavior through time, were not consistent with the predictions because territorial behavior did not change over time, because the data or model were inadequate, or because territoriality did not occur at the scale of the village. Results of comparisons of visibility, elevation, and distance to lookout values before and after 600 cal BP indicated no significant difference between time periods (Table 7.1). Perhaps coastal site locations were more strongly determined by access to fish, shellfish, sea birds, or other resources. People may also have chosen habitation sites based on their ability to pull a canoe onshore or fish offshore. If the reason that the defensive characteristics did not register change through time is because multiple communities banded together against groups to the north or the south, investigating those strategies would require an analysis of different parts of the islands for sites that could be used as lookouts, areas from which to attack intruders, or refuges where people could hide from invading groups. Table 7.1. Predictions and results for change over time in defensive sites on the San Juan Islands. Results of lithic analyses to investigate changes through time in lithic procurement centered on determining if the lithic assemblage from Watmough Bay that dates to 600 cal BP- 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Supports hypothesis? Territo iality Minimal defense/moderate permeability Minimal defense/moderate permeability Minimal defense/high permeability Increased defense/low permeability Defensive site predictions ? ? ? ? Results No sig. differences in visibility, elevation, distance to lookout. No sig. differences in visibility, elevation, distance to lookout. No sig. differences in visibility, elevation, distance to lookout. No sig. differences in visibility, elevation, distance to lookout. No 298 Contact indicates lithic procurement from a smaller (actively defended) area in the vicinity of Watmough Bay while assemblages from the previous time periods reflect a population that procured resources from a larger area. Small sample sizes for all but the 1600-1000 cal BP assemblage precluded a full diachronic analysis of lithic artifacts from the Watmough Bay site, but most measures did not support an increase in boundary defense at 600 cal BP-Contact (Table 7.2). I predicted that people should increase their use of the slate outcrop near the Watmough Bay site at 600 cal BP-Contact, but comparisons (by weight and by count) of slate during different time periods were not consistent with this trend. I also predicted an increase in flakes with smooth cortex at 600 cal BP-Contact similar to those found on Watmough Bay beach, but I found no significant differences between the four time periods in cortex appearance for flakes. Analyses designed to explore changes in processing and conservation strategies linked to hypotheses about procurement also indicated few changes through time (Table 7.3). I predicted that procurement adjacent to Watmough Bay at 600 cal BP ? Contact should be associated with decreased processing prior to transport due to minimal travel costs. I found no significant differences in proportions of first flakes between the assemblages although there were significantly more flakes with 0 or 1 flake scars rather than multiple flake scars for the 600 cal BP-Contact assemblage. I also predicted increased conservation of FGV at 600 cal BP-Contact due to the scarcity of cobbles of appropriate size and shape on Watmough Bay beach. I found no significant differences in the numbers of exhausted cores between the 2500-1600 cal BP and 1600-1000 cal BP time periods for which sample size was large enough to compare between assemblages. Retouch intensity was intensive during the 1600-1000 cal BP and 600 cal BP- Contact time periods during the two earlier periods, but due to small samples sizes and the subjectivity of this analysis, this result is inconclusive. 299 Finally, some analyses tentatively confirmed procurement of FGV from beyond Watmough Bay beach during the 1600-1000 cal BP time period in which the lithic assemblage was largest (Table 7.2). Cores (with original cobble dimensions intact) from Watmough Bay beach dating to 1600-1000 cal BP were typically larger than the cobbles from the beach. An analysis of cores and flakes from this time period also indicates the use of more angular cores than would have been available at Watmough Bay beach. People could have chosen the larger or more angular cobbles from this beach, but based on my field observations, the supply of large angular cobbles on the beach would have been insufficient to sustain communities over the hundreds of years that Watmough Bay was occupied. The substantially higher number of cores and flakes with rough cortex rather than the smooth waterworn cortex found on cobbles at Watmough Bay beach also indicates the use of other beaches during this time period. These findings are consistent with my hypothesis that people procured toolstone from a relatively large area at 1600-1000 cal BP when marine resources were scarcer; however, I cannot determine whether this behavior represents a shift in territorial behavior or a strategy for lithic procurement that remained unchanged through time. 300 Table 7.2. Predictions and results for change over in time lithic procurement at Watmough Bay based on lithic assemblage analyses. 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Supports hypothesis? Territoriality Minimal defense/moderate permeability Minimal defense/moderate permeability Minimal defense/high permeability Increased defense/low permeability FGV Source Aleck Bay, Agate Beach, Watmough Aleck Bay, Agate Beach, Watmough Aleck Bay, Agate Beach, Watmough Watmough Size Predictions ? ? ? ? Results Cores from cobbles larger than those found at Watmough Bay beach. Yes Slate predictions ? ? ? ? Slate Weights Munsell ? Munsell, Stein/Phillips ? Munsell, Stein/Phillips ? No Slate Counts Munsell, Stein/Phillips ? Munsell, Stein/Phillips ? Munsell, Stein/Phillips ? Munsell, Stein/Phillips ? No Shape Predictions More Angular More Angular More Angular More Round Results Cores/flakes indicate more angular than round cobbles. Yes Cortex Predictions Different proportions of smooth/rough from Watmough Beach. Different proportions of smooth/rough from Watmough Beach. Different proportions of smooth/rough from Watmough Beach. Similar to Watmough Beach Results No sig. differences between flake assemblages. No sig. differences between flake assemblages. Sig. more rough cortex cores/flakes than at Watmough beach. No sig. differences between flake assemblages. Inconclusive 301 Table 7.3. Predictions and results for change over time in conservation and processing at Watmough Bay based on lithic assemblage analyses. Along with testing hypotheses about boundary defense, I also considered change over time in boundary permeability. I hypothesized that during times of resource scarcity, higher benefits of cooperating and exchanging information and lower costs of transgressing boundaries would encourage increased inter-group interaction at 1600-1000 cal BP. Inter-group relationships should be lower at 600 cal BP-Contact when resources were adequate and boundary defense was high. During the earlier time periods from 2500-1600, small populations should encourage moderate interaction for information exchange and marriage partners. I tested those predictions using extra-local and rare toolstone at Watmough Bay. As predicted, I noted an increase in chert at 1600-1000 cal BP and an absence of chert at 600 cal BP-Contact; however, 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Supports hypothesis? Territoriality Minimal defense/moderate permeability Minimal defense/moderate permeability Minimal defense/high permeability Increased defense/low permeability FGV Source Aleck Bay, Agate Beach, Watmough Aleck Bay, Agate Beach, Watmough Aleck Bay, Agate Beach, Watmough Watmough Conservation predictions ? ? ? ? Exhausted cores No sig. difference No sig. difference Inconclusive Resouch intensity Less retouch More intensive retouch More intensive retouch Inconclusive Proce si g predictions ? ? ? ? Results No first flakes No sig. difference in proportion of first flakes. No sig. difference in proportion of first flakes. Sig. more flakes with 0 or 1 flake scar. Inconclusive 302 these results are likely a function of the larger sample at 1600-1000 cal BP and small sample at 600 cal BP-Contact. I also predicted that during times of high boundary defense, inter-group interaction should be more formal and involve fewer families. This should be evidenced through an uneven distribution of extra-local and rare lithic raw material across the site. The only assemblage for which there was enough extra-local or rare raw material to test this prediction was the 1600-1000 cal BP assemblage. The even distribution of raw material across the site is consistent with informal inter-group interactions predicted during this time period (Figure 7.4). Without the ability to compare the distribution of chert, quartz, and nephrite between time periods, the distribution of these materials does not provide conclusive answers about group interaction systems in the San Juan Islands during the Late Holocene. Table 7.4. Predictions and results for change over time in boundary permeability at Watmough Bay based on analyses of extralocal and raw materials. Overall, lithic procurement data are not consistent with the hypothesis that there was increased boundary defense and decreased permeability at 600 cal BP-Contact in the San Juan Islands. These results may, in fact, signify that territorial behavior did not develop until the Contact period, but there are other important possibilities to consider regarding sample size, the scale of boundary defense, the impact of climate change on resource distribution, potential 3500-2500 cal BP 2500-1600 cal BP 1600-1000 cal BP 600 cal BP-Contact Supports hypothesis? Territoriality Minimal defense/moderate permeability Minimal defense/moderate permeability Minimal defense/high permeability Increased defense/low permeability Exchange predictioncs Moderate, informal Moderate, informal High, informal Low, formal Results Small amount of chert. Small amount of chert. More chert, evenly distributed across the site. No chert. Inconclusive 303 problems with the model in predicting territorial behavior, and the sensitivity of the data to territoriality. I discuss these possibilities after summarizing results for the San Nicolas Island analyses. Summary of Results and Directions for San Nicolas Island Similar to the San Juan Islands case study, defensive site and lithic procurement data for the San Nicolas Island study area did not support the hypothesis that shifts in territorial behavior correspond to shifts in subsistence resource productivity on the landscape. Further research must be conducted to determine whether those results indicate that the Nicole?o were never territorial or whether there was a problem with the research approach or analysis. Regarding defensive sites, I predicted that at 5000-4000 cal BP, subsistence resources should far exceed the needs of the small population on the island therefore boundary defense should be minimal. At 2500-1500 cal BP and 500 cal BP-Contact, boundary defense should center on terrestrial resources. Sites near productive patches should exhibit defensive characteristics consistent with aggressive defense of adequate resources against intruders. At 500 cal BP-Contact, an increase in marine productivity should be associated with increased defensive characteristics of sites located near productive marine resources. Results of an ANOVA comparing mean elevation and distance to lookout values failed to show significant differences in defensive characteristics of sites between time periods for all sites, coastal sites, inland sites, or habitation sites (Table 7.5). This may indicate minimal boundary defense during any time period on San Nicolas Island, but it is probably more indicative of a lack of sensitivity of elevation and distance to lookout measurements to shifts in territorial behavior. People may have chosen particular areas for habitation sites due to proximity to resources, protection from 304 the elements, access to streams, community traditions, or social values. If all of the communities on the island considered themselves to be part of a larger San Nicolas Island community, those habitation sites with the best visibility to watercraft approaching from the mainland or from other islands should have the most prominent lookouts. Sites that were often the first point of attack should have the highest elevation values because sites with higher elevation values are more difficult to access. Villages could communicate with one another about potential threats by signaling or messengers. Further research on potential water routes between islands might provide an avenue for research on the defensive characteristics of San Nicolas Island as an island community. Table 7.5. Predictions and results for defensive characteristics of sites on San Nicolas Island. Defensive characteristics are based on site elevation measures and distance to lookout. Regarding lithic procurement on San Nicolas Island, I predicted two different trajectories for Tule Creek Village Mound B and SNI-106 because Mound B was more likely a productive terrestrial resource area and SNI-106 was more likely a secondary resource area. If the area around Mound B was defended more intensely at 3500-1500 cal BP and 500 cal BP-Contact due 5000-4000 cal BP 3500-1500 cal BP 1500-500 cal BP 500 cal BP-Contact Supports hypothesis? Territoriality Minimal defense/high permeability Inland defense/Marine permeability Marine defense/Inland permeability Inland defense/Marine permeability Predictions for d fen ive sit s Minimal Terrestrial Marine Terrestrial All sites No sig. differences No sig. differences No sig. differences No sig. differences No Coastal sites No sig. differences No sig. differences No sig. differences No sig. differences No Inland sites No sig. differences No sig. differences No sig. differences No sig. differences No 305 to more abundant and predictable terrestrial resources, lithic procurement should take place within a smaller area surrounding Tule Creek Village. This would result in decreased access to the finest-grained metavolcanic rock (MV) and increased use of the round cobbles that are more common at Tule Creek than at coastal sites. Processing prior to transport should decrease since the distance from the cobble area to the habitation site is short. People should conserve metavolcanic rock because high-quality material would have been relatively limited. At SNI- 106, none of these factors should show changes through time because with scarcer and less predictable resources, the people at that habitation site would be less likely to engage in costly boundary defense activities. My results indicate few changes through time in lithic procurement at either site (Table 7.6, 7.7). At Mound B there were no significant differences in MV toolstone through time. Based on proportions of flakes associated with round and flat cobble reduction sequences, there was also minimal change through time in access to round and flat cobbles. One analysis that may be consistent with increased conservation of metavolcanic rock at 500 cal BP-Contact is a decrease in tertiary flakes relative to other flake types during that time period, suggesting more primary reduction at the site. Since many other manufacturing processes and technological factors could contribute to that shift, this positive result alone is inconclusive. Finally, based on proportions of exhausted cores, flakes from exhausted cores, flakes with bipolar attributes, and non-cortical shatter, the only potential change through time in toolstone conservation is a significant increase in metavolcanic non-cortical shatter at 500 cal BP-Contact. Because other technological and manufacturing strategies could cause that shift and none of the other indices of conservation showed positive results, my overall impression is that conservation strategies changed little through time. 306 Table 7.6. Results of lithic analyses for Mound B. At SNI-106, some results point towards change over time in procurement strategies and others are more consistent unchanged strategies. The significant decrease in MV toolstone at 3500-15000 cal BP and the increase in this material at 1500-500 cal BP indicates reduced procurement opportunities during the later time period; however, a significant decrease in flakes from exhausted cores at that time supports the opposite conclusion. The lack of significant 5000-4000 cal BP 3500-1500 cal BP 1500-500 cal BP 500 cal BP-Contact Supports hypothesis? Sources Most attractive Tule Creek/Beaches Inland areas/Beaches Tule Creek/Beaches Material Type Predictions More quartzite, sandstone, finer-grained metavolcanic rock Some quartzite, sandstone, coarser- grained metavolcanic rock Least quartzite, sandstone, fine- grained metavolcanic rock Some quartzite, sandstone, coarser- grained metavolcanic rock Quartzite/Sandst one Weight Quartzite ?, Sandstone ? Quartzite lowest, Sandstone ? Quarzite ?, Sandstone ? Quarzite ?, Sandstone ? Yes Quartzite/Sandst one Count No sig. differences No sig. differences No sig. differences No sig. differences No Metavolcanic results No sig. differences PMV ? PMV ? MVP ? No sig. differences No Cobble Shape Predictions Flat/Round Flat/Round Fewer Flat Flat/Round Flake Types No sig. differences No sig. differences No sig. differences No Processing prior to transport Predictions Metavolcanic ? Quartzite ? Metavolcanic ? Quartzite ? Metavolcanic ? Quartzite ? Metavolcanic ? Quartzite ? Results No sig. differences Metavolcanic tertiary flakes ? Metavolcanic tertiary flakes? Yes Conservation Predictions None Conserve MV, Q Conserve Q Conserve MV, Q Exhausted cores Many Many Neutral Flakes from exhausted cores Few Few Few Few No Flakes with ip l r attributes No sig. differences No sig. differences No sig. differences No sig. differences No Non-cortical shatter MV ?, Quartzite no sig. difference MV ?, Quartzite no sig. difference Yes Retouch Potential increase in retouch intensity. Yes 307 differences in flakes associated with flat and round reduction sequences and unchanged proportion of non-cortical shatter indicates unchanged boundary defense strategies. (Table 7.7). Overall, these results tentatively support an unchanged boundary defense strategy and the need for further research linking lithic procurement activities and territorial behavior. Table 7.7. Results of lithic analyses for SNI-106. Along with analyzing metavolcanic rock, I also consider shifts in access to quartzite and sandstone in the context of proposed shifts in territorial strategies. At both Mound B and SNI- 3500-1500 cal BP 1500-500 cal BP Supports hypothesis? Territoriality Inland defense/Marine permeability Marine defense/Inland permeability Sources Inland areas/Beaches Inland areas/Beaches Material Type Predictions Some quartzite, sandstone, metavolcanic rock similar Least quartzite, sandstone, metavolcanic rock similar Quartzite/Sandst one Weights Sandstone rare, quarzite ? Sandstone rare, quartzite ? Yes Quartzite/Sandst one Counts Quartzite ? Quartzite ? Yes Metavolcanic results MV ? MV ? No Cobble Shape Predictions Flat/Round Flat/Round Flake Types No sig. differences between time period at 106, but more flat reduction sequence flakes than at Mound B. No sig. differences Yes Processing prior to transport Predictions All toolstone All toolstone Results No sig. differences No sig. differences Yes Conservation Predictions Conserve Q but not metavolcanic Conserve Q but not metavolcanic Flakes from exhausted cores ? ? No Non-cortical shatter No sig. differences No sig. differences Yes 308 106, I predicted that access to these materials (which are more abundant in coastal cobble areas) would be most difficult at 1500-500 cal BP when villages near productive marine resource patches should aggressively defend their territories. Access to quartzite, sandstone, and other marine resources would have been negotiated with shoreline villages and therefore have exacted a higher cost than during periods when marine resources were more accessible. In considering material proportions by weight for Mound B, I found less quartzite and sandstone at 1500-500 cal BP than at 500 cal BP-Contact, although quartzite percentages were lowest during the preceding period at 3500-1500 cal BP. When considering material type by count, there were no significant differences between time periods (Table 7.6). At SNI-106, sandstone was rare both at 3500-1500 cal BP and 1500-500 cal BP. Proportions of quartzite were lower at 1500-500 cal BP both considering count and weight (Table 7.7). Thus at SNI-106, if preference or manufacturing strategies are not a factor, access to quartzite may have been more difficult when predicted during a time of marine territoriality. Detailed analysis of other marine resources at this site would be beneficial in isolating changes in the ways that people accessed coastal areas at 1500- 500 cal BP. Likewise, lithic resources are important to separating environmental and social reasons for shifts in procurement because rocks are not affected by upwelling or sea surface temperature. Finally, another aspect of my lithic analysis for San Nicolas Island was to determine whether shifts in boundary permeability through time at Mound B (sample size of extra-local material at SNI-106 was too small for this analysis) reflected my prediction for increased inter- group interactions at 1500-500 cal BP when terrestrial resources were scarce and marine resource were less accessible. During this period, communication and reciprocal access between inland groups would have yielded greater benefits than during times of resource abundance. 309 Movement across the inland landscape would also have been less difficult during this period. Contrary to my predictions, I found no significant differences in chert or other rare lithic materials such as quartz during the later two time periods at Mound B (Table 7.8). There was a significantly higher proportion of chert at 5000-4000 cal BP when I predicted only moderate inter-group relationships to maintain marriage partners and information exchange. This could be a function of a small sample size. The proportion of quartz was highest at 3500-1500 cal BP when I predicted infrequent exchange. Due to small sample sizes and limited spatial information from Mound B, I was not able to reach definite conclusions about the intensity of formal of exchange on the island. Future research may focus on increasing the sample size of chert and quartz from throughout this site to determine the distribution of the material and change over time in proportions of rare toolstone through time. Table 7.8. Results of boundary permeability analyses at Mound B and SNI-106. The lithic procurement data for Mound B and SNI-106 are not consistent with my hypotheses for shifts in territorial behavior primarily determined by shifts in marine and terrestrial resource access. These results may indicate that communities on San Nicolas Island shared resource ownership and allowed full access throughout the Late Holocene. It is also possible that territoriality did not take place at the scale of the village, that millennial-scale climate change did not affect resource distribution on the ground, that the model may not be 5000-4000 cal BP 3500-1500 cal BP 1500-500 cal BP 500 cal BP-Contact Supports hypothesis? T rritori lity Minimal defense/high permeability Inland defense/Marine permeability Marine defense/Inland permeability Inland defense/Marine permeability Exchange Predictions Moderate, informal Infrequent, formal Frequent, Informal Infrequent, formal Results Mound B Chert ? Quartz ? No sig. diff. No sig diff. No Result SNI-106 Quartz ?, chert no sig. diff. Chert no sig. diff. No 310 useful for accurately predicting territorial behavior, or that the data are not sensitive to shifts in territoriality. Territoriality and Scale One possibility in interpreting my primarily negative results is that boundary defense and resource ownership took place at a scale larger than the village. In discussing the defensive characteristics of archaeological sites above, I suggested several potential ways to test whether communities on the San Juan Islands or San Nicolas Island banded together to defend their land and resources against groups from other islands or the mainland. In testing my proposed alternative hypothesis, I also examined this issue of scale by examining whether resources belonged to kin who shared goods and resources across permanently permeable boundaries between family groups in different villages. I suggested that if different families shared access between groups, different spatial areas of archaeological sites should show differences in toolstone access. My preliminary results for the Watmough Bay site indicated that for the 1600- 1000 cal BP period there may have been differences in use of FGV and slate in Spatial Unit 1 and 2 but there were no differences in processing or conservation activities. At Mound B, results of ?? tests comparing proportions of metavolcanic flakes and sandstone, chert, and quartz for the 500 cal BP-Contact assemblage indicate significantly higher proportions of sandstone, chert, and quartz in Spatial Unit 1 and quartzite in Spatial Unit 2 (Table 6.39). A comparison of tools from the two units also reveals significant differences, but toolstone distribution does not mirror the flake results. To more fully explore household access to toolstone, it will be necessary to conduct more detailed analysis of households and household boundaries (e.g., Ames 1996, 2006; Ames et al. 311 1992; Arnold 2006; Arnold et al. 1997; Graesch 2000; Hayden 1997, 1998; Hayden et al. 1996; Lepofsky et al. 2009; Prentiss et al. 2008; Rick 2007). For both Watmough Bay and Mound B, it is possible that the two defined spatial areas are close enough together that they were actually part of the same household. Analysis of the stratigraphy of contemporaneous deposits dating to 1600-1000 cal BP at Watmough Bay does not indicate distinct and separate house features. The hearth found in the 2004 excavation unit EXU1 is likely far enough away that it represents a separate household or activity area from the Munsell units, but differences in collection strategies between the two excavations makes it difficult to compare between them. At Mound B, the areas of the site that are well-dated are relatively close to one another. A comparison of the Mound B artifacts with those from a contemporaneous residential area of Tule Creek at Mound A may be more appropriate for determining differences in access to toolstone based on household. At Mound B, households are difficult to identify due to an absence of distinct floor, wall, hearth, and post hole features. Further complicating the issue, for both sites families may extend beyond the confines of a house structure and members of different households may share resources with one another. Another challenge of testing a permanent permeability hypothesis is that differences in material type ratios, original size and shape of cobbles, processing, and conservation may be attributable to differences in manufacturing activities, technological preferences, tool use, or other differences in household activities across the sites. To control for this possibility, it will be necessary to address the broader picture of resources procurement at the site. For Watmough Bay, shellfish and bird assemblages have been analyzed (Bovy 2005; 2007; Daniels 2009) but fish, mammal, and macrobotanical analyses have not been conducted. For Mound B, these analyses are in progress by the California State University, Los Angeles Archaeology 312 Laboratory. Because marine resources come from a distance from Tule Creek Village, differences in fish and shellfish assemblages would be particularly useful in determining if different families fished and collected shellfish on different beaches. To more rigorously test whether territoriality took place at a more regional scale, it may also be helpful to further investigate connections between sites by examining the distribution of extra-local toolstone at multiple sites on the landscape. Climate Change and Resource Access Another avenue for investigating the lack of congruence between my predictions and data is that the millennial-scale marine and terrestrial climate change data that I discuss in Chapter 2 may not have affected people at a local scale in the manner that I predicted. In other words, shifts in upwelling in the Gulf of Georgia may bring about a slight decrease in salmon and herring, but not enough to affect peoples? perceptions of subsistence resources and risk. A warmer and drier terrestrial climate on San Nicolas Island may decrease availability of freshwater but increases availability or distribution of other plant resources such as sagebrush and grasses used to make containers or scrub used for firewood (Thomas 1995). Another possibility is that during times of scarcity, resources were abundant in certain pockets of the landscape?where the salmon run remained adequate or where freshwater was always available, and these may have always been loci of boundary defense for the groups who lived nearby. Perhaps resources were always so patchy and unpredictable in these coastal settings that it made sense for kin and friends to ensure reciprocal access to nearby resource areas. Alternatively, fish, shellfish and small game may have always been so abundant that people did not feel the need to defend particular spots. 313 To address resource availability and distribution at a more local scale, I plan to synthesize published research from archaeological studies and site reports from both study areas to determine whether, when, and where marine and terrestrial resource were actually scarce on the ground. In particular, studies that provide data on resource depression of high-ranked prey (e.g., Daniels 2009) provide information on the degree to which people were affected by resource stress. Data that indicates active efforts to maintain a stable and sustainable resource base must also be part of the conversation about resource scarcity (Campbell and Butler 2010). As discussed above, ongoing research on faunal research at Mound B will provide invaluable information on subsistence strategies at that site. For both study areas, improved resolution of the paleoenvironmental record and the distribution of resources on the landscape will help in testing models that relate resource access to human behavior. Further study of ethnographic examples of boundary defense and permeability would also be helpful in redefining the expectations of the model. Modeling Territorial Behavior This study demonstrates both the utility and the shortcomings of a human behavioral ecology approach to articulate expectations for the behaviors of complex hunter-gatherer fishers groups that considers both the natural and cultural environment. Although human behavioral ecology models are more commonly applied to small egalitarian hunter-gatherers, they are also is increasingly applied to hunter-gatherers identified as ?complex? in that they exhibit differences in economic, social, and political rank both within and between groups (e.g., Fitzhugh 2003; Kohler and Van West 1996; Neiman 1997; Prentiss and Kuijt 2004). Using human behavioral ecology models to create hypotheses about human territoriality allowed me to construct explicit 314 and testable predictions for the archaeological record. One way to interpret my results is that people in both study areas maintained relatively permeable boundaries because they evaluated the costs and benefits of defending village boundaries in a manner that was not primarily defined by shifts in resource access but rather by considerations more purely in the social realm. A human behavioral ecology model provides a good place to start in generating an explanation for the development of territoriality but may be inadequate for capturing the many unpredictable nuances of human decision making about who to exclude from a community, where to access resources, and with whom to share or exchange resources. Hill et al. (2009:188) note that humans show a capacity for cooperative and altruistic behavior with both kin and nonkin. In this light, it may be difficult to predict territoriality in the context of environmental change. A more inductive approach that incorporates and discusses possibilities for both competition and cooperation may allow me to examine patterns in the data on stone artifacts that I had not considered in generating predictions that focus on boundary defense. This approach would allow me to consider several possibilities for territorial behavior simultaneously, rather than emphasizing one primary hypothesis to test. Data Sensitivity to Territorial Behavior A final consideration in interpreting my results concerns the sensitivity of my data to shifts in boundary defense and permeability. Small sample sizes of extra-local materials for both the Watmough Bay and Mound B assemblages make addressing boundary permeability particularly problematic. There is also the problem of small sample sizes, particularly for the two earlier assemblages at Watmough Bay and Mound B and for the entire assemblage at SNI-106. Increasing the samples through more dating at Mound B to increase the 500 cal BP-Contact 315 assemblage would be particularly useful in re-testing the hypotheses discussed here. For Watmough Bay, increasing the sample would be more difficult because all lithic artifacts were analyzed and most deposits were dated. Increasing the sample from the 600 cal BP-Contact period would require more excavation in the area of the 2003 excavation units where both older and younger deposits were found. Another important consideration regarding data sensitivity is that although people may have maintained and defended boundaries surrounding food resources such as fishing areas or shellfish beds, they may not have done so with toolstone because it was so abundant both in the San Juan Islands and on San Nicolas Island. People may have tolerated the ?theft? of toolstone when other groups came in their canoes or on foot to procure a load of cobbles. It is also possibly that although theft was not tolerated, people who lived near good sources of lithic raw material participated in exchange for other resources with groups. For these reasons, toolstone may not serve as a good proxy for territorial behaviors. Utility of a Comparative Approach A comparison of two study areas reveals intriguing similarities and differences in the ways that people procured cobbles and where and how they flaked the cobbles to create tools. In the San Juan Islands case study, people chose angular cobbles of a specific size to make their tools. The most likely scenario is that they piled cobbles in their watercraft, perhaps along with other supplies, and brought them to the site with minimal testing or early stage reduction prior to transport. They do not appear to have emphasized toolstone conservation given the high ratio of cores to flakes at the site and the scarcity of bipolar shatter. Cores were reduced with minimal 316 preparation and in many cases, the desired tool was created by removing a large flake from the core and then using it for a scraping or cutting task. On San Nicolas Island, people would have had to carry toolstone from cobble areas on beaches and blowouts across the island back to the site. The low proportion of primary flakes compared to later stage flakes may reflect testing and early stage reduction prior to transport reflecting higher transport costs than on the San Juan Islands. Similar to the other case study, most cores were prepared minimally although patterned cores are somewhat more common for the Mound B assemblage. Many flakes struck from the cores were considered waste products but many others were used without further modification. Unlike the scaled pieces from the Watmough Bay assemblage, the used flakes from Mound B show less visible wear. Some flakes show signs of edge damage but it is difficult to link that damage to use. The metavolcanic toolstone from Mound B is slightly harder than the FGV at Watmough Bay. In both cases, the skill with which the toolmakers reduced cores and created flake tools demonstrates that the ?expedient? technology that they relied on involved not only physical strength, but also foresight and planning. During collection, there would have been more uncertainty regarding toolstone quality on San Nicolas Island than on the San Juan Islands because the San Nicolas Island cobbles vary more in hardness and in grainsize. It would have been difficult to determine type of toolstone for the cortex alone, although experienced collectors were probably able to use subtle visual clues to determine quality. In both study areas, it would have been easier to split the cobbles on the beach because of availability of natural rock anvils and hammers that would have been heavy to bring up to the sites, and because of the strenuous activity and disturbance created by these activities. 317 Implications for Social Complexity Studies A study of territorial adaptations on the Pacific Coast that indicates potential for high levels of boundary permeability through time could provide new insights on complex hunter- gatherers and explanations for the emergences of social complexity on both the Northwest Coast and in the Channel Islands. Further research is required for me to draw that conclusion for either study areas; however, the lack of clear indication of village boundary defense during time periods when other signs of social complexity occur may provide insights for cooperation- focused models. In contrast to Pacific Coast complexity models that emphasize competition (e.g., Matson 1983, 1985; Matson and Coupland 1995; Raab and Larson (1997), some models explore the ramifications of different forms of cooperation within and between villages. For example, Lepofsky et al. (2005) provide a cooperation-focused model in which regional differences in resource availability due to the warmer and drier Fraser Valley Fire Period spurred social interactions between Gulf of Georgia villages to maintain access to the relatively rich Fraser Valley during a time of scarcity. Families with access to Fraser Valley kin saw a rise in social and economic status. Similarly, for the Channel Islands, Kennett and Kennet?s (2000) research on both competitive and cooperative responses to climate change in the Northern Channel Islands indicates more cooperative strategies after 650 cal BP as marine productivity increased and terrestrial food and water availability decreased. Relevance to Contemporary Issues Research on boundary defense and permeability has broader implications for modern studies about conflict in the context of resource scarcity and human response to climate change in general. With more development, my research may support arguments by political ecologists 318 that resource scarcity does not necessarily lead to conflict. Although earlier political ecology research assumed scarcity and conflict to be linked, especially with regard to water (Huth 1996; Vasquez 1993), researchers increasingly find that people use ingenuity and inter-group cooperation to deal with difficulties in accessing water (Toset et al. 2000). They may also move away, attempt to use other resources, begin new forms of cooperation (Ostrom 1990; O?Lear 2005; Ostrom et al. 1994), and management of resources (Thompson and Price 2002). Facing increased resource unpredictability due to global climate change, political ecologists may be able to draw on archaeological examples of sustained cooperative relationships through periods of environmental upheaval to understand when, how, and why people share resources across boundaries. 319 REFERENCES Acheson, J. M. 2003. Capturing the Commons: Devising Institutions to Manage the Maine Lobster Industry. University Press of New England, Lebanon, NH. Acheson, J. M. 1988. The Lobster Gangs of Maine. University Press of New England, Hanover, NH. Acheson, J. M. and R. J. Gardner. 2004. Strategies, Conflict, and the Emergence of Territoriality: The Case of the Maine Lobster Industry. American Anthropologist 106:296-307. Ackerman, R. E. and L. A. Ackerman. 1973. Ethnoarchaeological Interpretations of Territoriality and Land Use in Southwestern Alaska. Ethnohistory 20:315-334. Afifi, A. 2000. Prehistoric Settlement Pattern on San Nicolas Island, California: A GIS Application. Unpublished M.A. thesis. Department of Anthropology, California State University, Los Angeles. Aldenderfer, M. 1998. Montane Foragers: Asana and the South-Central Andean Archaic. University of Iowa Press, Iowa City. Ames, K. M. 2006. Thinking about Household Archaeology on the Northwest Coast.. Pages 16- 38 in E. A. Sobel, D. A. Trieu Gahr, and K.M. Ames, editors. Household Archaeology on the Northwest Coast. International Monographs in Prehistory, Ann Arbor, MI. Ames, K. M. 2005. Tempo and Scale in the Evolution of Social Complexity in Western North America: Four Case Studies. Pages 56-78 in T.R. Pauketat and D.D. Loren, editors. North American Prehistory. Blackwell Press, London. Ames, K. M. 2002. Going by boat: The forager-collector continuum at sea. Pages 19-52 in B. Fitzhugh and J. Habu, editors. Beyond Foraging and Collecting: Evolutionary Change in Hunter-Gatherer Settlement Systems. Kluwer/Plenum, New York. Ames, K. M. 1996. Life in the Big House: Household Labor and Dwelling Size on the Northwest Coast. Pages 178-200 in G. Coupland and E. B. Banning, editors. People Who Lived in Big Houses: Archaeological Perspectives on Large Domestic Structures. Prehistory Press, Madison, WI. Ames, K. M. 1995. Chiefly Power and Household Production on the Northwest Coast. Pages 155-187 in T. D. Price and G. M. Feinman, editors. Foundations of Inequality. Plenum Press, New York. Ames, K. M. 1994. The Northwest Coast: Complex Hunter-Gatherers, Ecology and Social Evolution. Annual Review of Anthropology 23:209-229. Ames, K. M. 1992. Household archaeology of a southern Northwest coast plank house. Journal of Field Archaeology 19(3): 275-290. Ames, K. M. and H.G.D. Maschner. 1999. Peoples of the Northwest Coast: Their Archaeology and Prehistory. Thames and Hudson, London. 320 Andouze, F. 1999. New Advances in French Prehistory. Antiquity 71:167-175. Andrefsky, W. 2009. The Analysis of Stone Tool Procurement, Production, and Maintenance. Journal of Archaeological Research 17:65-103. Andrefsky, W. 2005. Lithics: Macroscopic Approaches to Analysis, Second Edition. Cambridge University Press, Cambridge. Andrefsky, W. 1994a. Raw-Material Availability and the Organization of Technology. American Antiquity 59:21-34. Andrefsky, W. 1994b. The geological occurrence of lithic material and stone tool production strategies. Geoarchaeology 9:375-391. Andrefsky, W., Jr. 1987. Diffusion and Innovation from the Perspective of Wedge Shaped Cores in Alaska and Japan. Pages 13-44 in J.K. Johnson and C.A. Morrow, editors. The Organization of Core Technology. Westview Press, Boulder, CO. Angelbeck, W. P. 2009. ?They Recognize No Superior Chief? Power Practice, Anarchism and Warfare in the Coast Salish Past. Unpublished Ph.D. dissertation. Department of Anthropology, University of British Columbia, Vancouver. Apostolou, M. 2008. Bridewealth and Brideservice as Instruments of Parental Choice. Journal of Social, Evolutionary, and Cultural Psychology 2(3):89-102. Armstrong, B. G., M. Weiss, R. I. Grieg, J. Dines, J. Gooden, and G. P. Aldred. 2003. Determining screening fractions and kernel roundness with digital image analysis. Proceedings of the 53rd Australian Cereal Chemistry Conference, Adelaide, Australia. Armstrong, J. E., D.R. Crandell, D. J. Easterbrook, and J.B. Noble. 1965, Late Pleistocene stratigraphy and chronology in southwestern British Columbia and northwestern Washington. Geological Society of America Bulletin 76: 321-330. Arnold, J. E. 2006. Households and Production on the Pacific Coast: The Northwest Coast and California in Comparative Perspective. Pages 270-285 in . A. Soebel, D. A. Trieu Garr and K. M. Ames, editors. Household Archaeology on the Northwest Coast. International Monographs in Prehistory, Ann Arbor, MI. Arnold, J. E. 2001. The Origins of a Pacific Coast Chiefdom: the Chumash of the Channel Islands. University of Utah Press, Salt Lake City. Arnold, J. E. 1995a. Social Inequality, Marginalization, and Economic Process. Pages 87-103 in T. D. Price and G.M. Feinman, editors. Foundations of Social Inequality. Plenum Press, New York. Arnold, J. E. 1995b. Transportation, Innovation, and Social Complexity Among Maritime Hunter-Gatherer Societies. American Anthropologist 97:733-747. Arnold, J. E. 1992. Complex Hunter-Gatherer-Fishers of Prehistoric California ? Chiefs, Specialists, and Maritime Adaptations of the Channel Islands. American Antiquity 57:60-84. 321 Arnold, J. E. 1990a. An Archaeological Perspective on the Historic Settlement Pattern on Santa Cruz Island. Journal of California and Great Basin Anthropology 12:112-127. Arnold, J. E. 1990b. Lithic Resource Control and Economic Change in the Santa Barbara Channel Region. Journal of California and Great Basin Anthropology 12:158-172. Arnold, J. E. 1987. Craft Specialization in the Prehistoric Channel Islands, California. University of California Press, Berkeley. Arnold, J. E. and J. Bernard. 2005. Negotiating the coasts: status and the evolution of boat technology in California. World Archaeology 37:109-131. Arnold, J. E., E.L. Ambos, and D.O. Larson, 1997. Geophysical Surveys of Stratigraphically Complex Island California Sites: New Implications for Household Archaeology. Antiquity 71: 157-168. Arnold, J. E., A.M. Preziosi, and P. Shattuck. 2001. Flaked Stone Craft Production and Exchange in Island Chumash Territory. Pages 113-131 in J. E. Arnold, editor. The Origins of a Pacific Coast Chiefdom: The Chumash of the Channel Islands. University of Utah Press, Salt Lake City. Aswani, S. 2002. Assessing the Effects of Changing Demographics and Consumption Patterns on Sea Tenure Regimes in the Roviana Lagoon, Solomon Islands. Ambio 31:272-284. Axelrod, R. 1997. The Complexity of Cooperation: Agent Based Models of Competition and Cooperation. Princeton University Press, Princeton, NJ. Bakewell, E. F. 2005. The Archaeopetrology of Vitrophyric Toolstones, with Applications to Archaeology in the Pacific Northwest. Unpublished Ph.D. dissertation. University of Washington, Department of Anthropology, Seattle, WA. Bakewell, E. F. 1996. Petrographic and geochemical source-modeling of volcanic lithics from archaeological contexts: A case study from British Camp, San Juan Island, Washington. Geoarchaeology 11:119-140. Bamforth, D. B. 1991. Technological organization and hunter-gatherer land use: A California example. American Antiquity 56:216-234. Bamforth, D. B. 1990. Settlement, raw material, and lithic procurement in the central Mojave Desert. Journal of Anthropological Archaeology 9:70-104. Bamforth, D. B. 1986. Technological efficiency and tool curation. American Antiquity 51:38-50. Barber, I. 2003. Sea, land and fish: spatial relationships and the archaeology of South Island Maori fishing. World Archaeology 35:434-448. Barnett, H. 1955. The Coast Salish of British Columbia. University of Oregon Press, Eugene, OR. 322 Bartelle, B. G., R. L. Vellanoweth, E. S. Netherton, N. W. Poister, W. E. Kendig, A. F. Ainis, R. J. Glenn, J. V. Marty, L. Thomas-Barnett, and S. J. Schwartz. 2010. Trauma and pathology of a buried dog from San Nicolas Island, California, U.S.A. Journal of Archaeological Science 37(11):2721-2734. Basgall, M. E. 1987. Resource Intensification among Hunter-Gathers: Acorn Economies in Prehistoric California. Research in Economic Anthropology 9:21-52. Bawden, C. and R. M. Reycraft. 2000. Environmental disaster and the archaeology of human response. Maxwell Museum of Anthropology, Albuquerque. Bean, L. J. and C. R. Smith. 1978. Gabrielino. Pages 538-549 in R. F. Heizer, editor. Handbook of North American Indians. Smithsonian Institution, Washington, D.C. Beck, C., A. K. Taylor, G. T. Jones, C. M. Fadem, C. R. Cook, and S. A. Millward. 2002. Rocks are heavy: transport costs and Paleoarchaic quarry behavior in the Great Basin. Journal of Anthropological Archaeology 21:481-507. Begossi, A. 1995. Fishing Spots and Sea Tenure: Incipient Forms of Local Management in Atlantic Forest Communities. Human Ecology 23:387-407. Bettinger, R. L., R. Malhi, and H. McCarthy. 1997. Central place models of acorn and mussel processing. Journal of Archaeological Science 24:887-899. Bettinger, R. L. 1982. Aboriginal Exchange and Territoriality in Owens Valley, California. Pages 103-127 in J. E. Ericson and T.K. Earle, editors. Contexts for Prehistoric Exchange. Academic Press, New York. Binford, L. R. 1980. Willow smoke and dogs' tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45:4-20. Binford, L. R. 1979. Organization and Formation Processes: Looking at Curated Technologies. Journal of Anthropological Research 35:255-273. Binford, L. R. 1973. Interassemblage variability: The Mousterian and the "functional" argument. Pages 227-254 in C. Renfrew, editor. The Explanation of Culture Change: Models in Prehistory. Duckworth, London. Binford, L. R. and J. F. O'Connell. 1984. An Alyawara day: The stone quarry. Journal of Anthropological Research 40:406-432. Blades, B. S. 1999. Aurignacian lithic economy and early modern human mobility: new perspective from classic sites in the V?z?re valley, France. Journal of Human Evolution 37:91- 120. Blair, S. E. 2010. Missing the boat in lithic procurement: Watercraft and the bulk procurement of tool-stone on the Maritime Peninsula. Journal of Anthropological Archaeology 29:33-46. 323 Bleed, P. 2001. Trees or Chains, Links or Branches: Conceptual Alternative for Consideration of Stone Tool Production and Other Sequential Activities. Journal of Archaeological Method and Theory 8:101-127. Bleed, P. 1986. The optimal design of hunting weapons: Maintainability or reliability. American Antiquity 51:737-747. Blurton Jones, N. G. 1984. A selfish origin for human food sharing: Tolerated theft. Ethology and Sociobiology 5(1):1-3. Bodu, P. 1996. Magdalenian hunters of Pincevent. Lithic Technology 21:66-70. Boeda, E. 1995. Levallois: A volumetric construction, methods, a technique. Pages 41-68 in H. L. Dibble and O. Bar-Yosef, editors. The Definition and Interpretation of Levallois Technology. University of Pennsylvania Press, Philadelphia. Boone, J. L. 1992. Competition, Conflict, and the Development of Social Hierarchies. Pages 301-337 in E. A. Smith and B. Winterhalder, editors. Evolutionary Ecology and Human Behavior. Aldine de Gruyter, New York. Boone, J. L. 1983. Noble Family Structure and Expansionist Warfare in the Late Middle Ages: A Socioecological Approach. Pages 79-96 in M. A. Little and R. Dyson-Hudson, editors. Rethinking Human Adaptation: Biological and Cultural Models. Westview Press, Boulder, CO. Boone, J. L. and E. A. Smith 1998. Is It Evolution Yet? A Critique of Evolutionary Archaeology. Current Anthropology 39 SUPP/1: S141-S174. Booth, D. B. 1987. Timing and processes of deglaciation along the southern margin of the Cordilleran ice sheet. Pages 71?90 in W. F. Ruddiman, H.E. Wright, Jr., editors. North America and Adjacent Oceans During the Last Deglaciation. Geological Society of America, The Geology of North America, Boulder, CO. Bourdieu, P. 1977. The Outline for a Theory of Practice. Cambridge University Press, New York. Bourdieu, P. 1990. The Logic of Practice. Cambridge University Press, New York. Bovy, K. 2007. Global human impacts or climate change?: explaining the Sooty Shearwater decline at the Minard site, Washington State, USA. Journal of Archaeological Science 34:1087- 1097. Bovy, K. M. 2005. Effects of Human Hunting, Climate Change and Tectonic Events on Waterbirds along the Pacific Northwest Coast during the Late Holocene. Unpublished Ph.D. dissertation. Department of Anthropology, University of Washington, Seattle. Bovy, K. M., Phillips, Laura S. and Stein, Julie K. 2007. Watmough Bay Site Stabilitization Project, 45-SJ-280. Descriptive Preliminary Report submitted to the Bureau of Land Management. Manuscript on file at the Burke Museum of Natural History and Culture, Seattle. 324 Bradbury, A. P. and J. D. Franklin. 2000. Material variability, package size and mass analysis. Lithic Technology 25:42-58. Bradbury, A. P., P. J. Carr, and D. R. Cooper. 2008. Raw material and retouched flakes. Pages 223-256 in J. Andrefsky, W., editor. Lithic Technology: Measures of Production, Use, and Curation. Cambridge University Press, Cambridge. Brantingham, P. J. 2006. Measuring Forager Mobility. Current Anthropology 47:435-459. Brantingham, P.J. 2003. A Neutral Model of Stone Raw Material Procurement. American Antiquity 68(3):487-509. Brantingham, P. J. and S. L. Kuhn. 2001. Constraints on Levallois Core Technology: A Mathematical Model. Journal of Archaeological Science 28:747-761. Brantingham, P. J., J. Olsen, J. Rech, and A. Krivoshapkin. 2000. Raw Material Quality and Prepared Core Technologies in Northeast Asia. Journal of Archaeological Science 27:255-271. Braun, D. R. 2005. Examining flake production strategies: Examples from the Middle Paleolithic ofsouthwest Asia. Lithic Technology 30:107-125. Broughton, J. M. and J. F. O'Connell. 1999. On Evolutionary Ecology, Selectionist Archaeology, and Behavioral Archaeology. American Antiquity 64:153-165. Brown, J. L. 1964. The Evolution of Diversity in Avian Territorial Systems. Wilson Bulletin 76:160-169. Brown, K. J. and Hebda, R.J. 2003. Coastal Rainforest Connections Disclosed Through a Late Quaternary Vegetation, Climate, and Fire History Investigation from the Mountain Hemlock Zone on Southern Vancouver Island, British Columbia, Canada. Review of Palaeobotany and Palynology 123:247-269. Browne, C. L. and L. Wilson. 2011. Resource selection of lithic raw materials in the Middle Palaeolithic in southern France. Journal of Human Evolution 61:597-608. Burley, D. M. 1980. Marpole: Anthropological Reconstructions of a Prehistoric Northwest Coast Culture Type. Simon Fraser University, Burnaby, British Columbia. Burnham, W. L., F. Kunkel, W. Hofmann, and W. C. Peterson. 1963. Hydrologic reconnaissance of San Nicolas Island, California. Pages 1539-1540 in Contributions to the hydrology of the United States, 1960. US Geological Survey, Washington, D.C. Bye, A., B.R. Edwards, and C.J. Hickson, 2000. Preliminary field, petrographic, and geochemical analysis of possible subglacial, dacitic volcanism at the Watts Point volcanic centre, southwestern British Columbia. Geological Survey of Canada, Current Research 2000-A-20, pp. 1-9, http://www.nrcan.gc.ca/gsc/bookstore. Campbell, S. K. and V. L. Butler. 2010. Archaeological Evidence for Resilience of Pacific Northwest Salmon Populations and the Socioecological System over the last 7,500 years. Ecology & Society 15(1):1-17. 325 Cannon, A. 2006. Giving Voice to Juana Maria's People: The organization of shell and exotic stone artifact production and trade at a late Holocene village on San Nicolas Island, California. Unpublished M.A. thesis. Department of Social Science, Environment and Community Interdisciplinary Program, Humboldt State University, Arcata, CA. Carlson, K.T. 2010. The power of place, the problem of time: Aboriginal identity and historical consciousness in the cauldron of colonialism. University of Toronto Press, Toronto, Ontario. Carlson, K. T., and A. J. McHalsie. 2001. A St?:lo-Coast Salish Historical Atlas. Douglas & McIntyre ; University of Washington Press, Seattle, WA; St?:lo Heritage Trust, Chilliwack, BC. Carlson, R. L. 1994. Trade and Exchange in Prehistoric British Columbia. Pages 307-361 in T. Baugh and J. Ericson, editors. Prehistoric Exchange Systems in North America. Plenum Press, New York. Carlson, R. L. 1960. Chronology and Culture Change in the San Juan Islands, Washington. American Antiquity 25:562-586. Carneiro, R.L. 1970. A Theory of the Origin of the State. Science 169:733-738. Cashdan, E., editor. 1990. Risk and Uncertainty in Tribal and Peasant Economies. Westview Press, Boulder, CO. Cashdan, E. 1992. Spatial Organization and Habitat use. Pages 237-266 in E. A. Smith and B. Winterhalder, editors. Ecology, Evolution, and Human Behavior. Aldine de Gruyter, New York. Cashdan, E. 1983. Territoriality among human foragers: ecological models and an application to four bushman groups. Current Anthropology 24:47-66. Cassidy, J., L. M. Raab, and N. A. Kononenko. 2004. Boats, Bones and Biface Bias: The Early Holocene Mariners of Eel Point, San Clemente Island. American Antiquity 69:109-130. Castillo, E. D. 1999. Blood Came from Their Mouths: Tongva and Chumash Responses to the Pandemic of 1801. American Indian Culture and Research Journal 23:47-61. Clague, J. J. 1981. Late Quaternary Geology and Geochronology of British Columbia, Part 2: Summary and Discussion of Radiocarbon-dated Quaternary History. Survey of Canada Paper 80-35. Clague, J. J. and T. S. James. 2002. History and isostatic effects of the last ice sheet in southern British Columbia. Quaternary Science Reviews 21:71-87. Clarkson, C. 2002. An index of invasiveness and Archaeological verification. Journal of Archaeological Science 61: 65-75. Clevenger, J. M. 1982. The Split Cobble Reduction Technology on San Nicolas Island, CA. California State University, Long Beach, Long Beach, California. Close, A. E. 2011. A Cha?ne Operatoire Analysis of the Flaking of Stone. Pages 65-87 in A. K. Taylor and J. K. Stein, editors. Is it a House? Archaeological Excavations at English Camp, San Juan Island, Washington. University of Washington Press, Seattle. 326 Close, A. E. 2006. Finding the People who Flaked the Stone at English Camp (San Juan Island). University of Utah Press, Salt Lake City. Cohen, N.M. 1977. Food Crises in Prehistory. Yale University Press, New Haven. Conca, D. 2000. Archaeological Investigations at Site 45-CA-432: Re-evaluating Mid-Holocene Land Use on the Olympic Peninsula, Washington. Unpublished M.A. thesis. Department of Anthropolology, Western Washington University, Bellingham, WA. Coupland, G. 2004. Complex Hunter-Gatherers of the Southern California Coast: A View from One Thousand Miles North. Pages 172-184 in J.E. Arnold, editor. Foundations of Chumash Complexity: Perspectives in California Archaeology. Cotsen Institute of Archaeology, University of California Los Angeles. Croes, D. R. and S. Hackenberger. 1988. Hoko River Archaeological Complex: Modeling Prehistoric Northwest Coast Economic Evolution. Pages 19-85 in B. Issac, editor. Prehistoric Economies of the Pacific Northwest Coast, Suppl. 3. JAI Press, Greenwich, Connecticut. Daniels, P. S. 2009. A Gendered Model of Prehistoric Resource Depression: A Case Study on the Northwest Coast of North America. Unpublished Ph.D. dissertation. Department of Anthropology, University of Washington, Seattle. Darwent, J. 1996. The Prehistoric Use of Nephrite on the British Columbia Plateau. Unpublished M.A. thesis. Department of Archaeology. Simon Fraser University, Burnaby, BC. Davis, O. K. W., Stephen L., and Palacios-Fest, Manuel R. 2003. Palynological Holocene Environments of Twin Rivers Marsh, San Nicolas Island, California. Technical Report 97-18. Statistical Research, Tucson. Dayton, P. K., Tegner, M.J., Edwards, P.B., and Riser, K.L. 1999. Temporal and spatial scales of kelp demography: the role of oceanography climate. Ecological Monographs 69:219-250. Deo, J. N., J. O. Stone, and J. K. Stein. 2004. Building Confidence in Shell: Variations in the Marine Radiocarbon Reservoir Correction for the Northwest Coast Over the Past 3,000 Years. American Antiquity. 69:771. Des Lauriers, M.M. 2005. The Watercraft of Isla Cedros, Baja California: Variability and Capabilities of Indigenous Seafaring Technology along the Pacific Coast of North America. American Antiquity 70(2):342-360. Dethier, D. P., Pessl Jr., F., Keuler, R.F., Balzarini, M.A., Pevear, D.R. 1995. Late Wisconsinan glaciomarine deposition and isostatic rebound, northern Puget Lowland, Washington. Geological Society of America Bulletin 107:1288?1303. Dethier, D. P., W. H. Crane, M. A. Hirshfeld, J. K. Holbert, W. S. Morgan, and R. E. Sylvester. 1996. Morphology and seismic stratigraphy of marine banks, northern Puget Lowland, Washington. Geological Society of America Abstracts with Programs 28:A-435. Dibble, H. L. 1995. Middle Paleolithic Scraper Reduction: Background, Clarification, and Review of the Evidence to Date. Journal of Archaeological Method and Theory 2:299-368. 327 Dibble, H. L., U. A. Schurmans, R. P. Iovita, and M. V. McLaughlin. 2005. The Measurement and Interpretation of Cortex in Lithic Assemblages. American Antiquity 70:545-560. Dietrich, W. 1975. Surface water resources of San Juan County.Pages 59-126 in R. H. Russell, editor. Geology and Water Resources of the San Juan Islands, San Juan County, Washington. Washington State Department of Ecology, Olympia, WA. Dillian, C. D. K. 2003. An Archaeological Approach to Territoriality and Boundary Defense among Northern California Hunter-Gatherers. Pages 123-140 in J. Habu, Savelle, James M., Koyama, S., and Hitomi, Hongo editors. Hunter-Gatherers of the North Pacific Rim. National Museum of Ethnology, Osaka, Japan. Dillon, P. S. and J. Oyen, editors. 2008. Canoeing. Human Kinetics, Inc., Champaign, IL. Douglass, M. J. and S. Holdoway. 2011. Changing perspectives in Australian archaeology. Part IV. Quantifying stone raw material size distributions: investigating cortex proportions in lithic assemblages from western New South Wales. Technical Reports of the Australian Museum, Online 23:45-57. Douglass, M. J., S. J. Holdaway, P. C. Fanning, and J. I. Shiner. 2008. An Assessment and Archaeological Application of Cortex Measurement in Lithic Assemblages. American Antiquity. 73:513-526. Dunbar, R. I. M. and C. Kenyatta. 2008. Kinship, Risk and the Willingness to Invest. Journal of Evolutionary Psychology 6:111-128. Dunbar, R. I. M., A. Clark, and N. Hurst. 1995. Conflict and co-operation among the Vikings: Contingent behavioral decisions. Ethology and Sociobiology 16:233-246. Durham, B. 1960. Indian Canoes of the Northwest Coast. Copper Canoe Press, Seattle. Durham, W. H. 1976. Resource Competition and Human Aggression, Part I: A Review of Primitive War. The Quarterly Review of Biology 51:385-414. Durrenberger, E. P. and G. P?lsson. 1987. Ownership at sea: fishing territories and access to sea resources. American Ethnologist 14:508-522. Dyson-Hudson, R. and E. A. Smith. 1978. Human Territoriality: An Ecological Reassessment. American Anthropologist 81:21-41. Earle, T. 2010. Exchange Systems in Prehistory. Pages 205-217 in C. Dillian and C. White, editors. Trade and Exchange: Archaeological Studies from History and Prehistory. Springer Verlag, New York. Easterbrook, D. J. 1969. Pleistocene Chronology of the Puget Lowland and San Juan Islands, Washington. Geological Society of America Bulletin 80:2273-2286. 328 Easterbrook, D. J. 1986. Stratigraphy and chronology of Quaternary deposits of the Puget Lowland and Olympic Mountains of Washington and the Cascade Mountains of Washington and Oregon. Pages 145?159 in V. ?ibrava, D.Q. Bowen, and G. Richmond, G. editors. Quaternary Glaciation in the Northern Hemisphere. Quaternary Science Reviews. Easterbrook, D. J. 1992. Advance and Retreat of Cordilleran Ice Sheets in Washington, U.S.A. G?ographie Physique et Quaternaire 46:51-68. Easton, N. A. and C. D. Moore. 1991. Test excavations of subtidal deposits at Montague Harbour, British Columbia, Canada-1989. International Journal of Nautical Archaeology 20:269-280. Edwards, M. S. 2004. Estimating scale-dependency in disturbance impacts: El Ninos and giant kelp forests in the northeast Pacific. Oceologia 138:436-447. Eerkens, J. W. 1999. Common Pool Resources, Buffer Zones, and Jointly Owned Territories: Hunter-Gatherer Land and Resource Tenure in Fort Irwin, Southeastern California. Human Ecology 27:297-318. Eerkens, J. W., K. J. Vaughn, M. Linares-Grados, C. A. Conlee, K. Schreiber, M. D. Glascock, and N. Tripcevich. 2010. Spatio-temporal patterns in obsidian consumption in the Southern Nasca Region, Peru. Journal of Archaeological Science 37:825-832. Eerkens, J. W., A. M. Spurling, and M. A. Gras. 2008. Measuring prehistoric mobility strategies based on obsidian geochemical and technological signatures in the Owens Valley, California. Journal of Archaeological Science 35:668-680. Elmendorf, W. W. 1971. Coast Salish Status Ranking and Intergroup Ties. Southwestern Journal of Anthropology 27:353-380. Elston, R. G. 1990. A Cost-Benefit Model of Lithic Assemblage Variability. Pages 153-163 in R. Elston and E. Budy, editors. The Archaeology of James Creek Shelter. University of Utah Anthropological Papers no. 115, Salt Lake City. Ember, M. and C.R. Ember. 1992. Resource Unpredictability, Mistrust, and War: A Cross- Cultural Study. Journal of Conflict Resolution 36:246-262. Engle, J. M. 1994. Perspectives on the Structure and Dynamics of Nearshore Marine Assemblages of the California Channel Islands. Pages 13-26 in W. L. Halborsen and G. J. Maender, editors. The Fourth California Islands Symposium: Update on the Status of Resources. Santa Barbara Museum of Natural History, Santa Barbara. Ericson, J. 1984. Toward the analysis of lithic production systems. Pages 1-9 in J. Ericson and B. Purdy, editors. Prehistoric Quarries and Lithic Production. Cambridge University Press, Cambridge. Erlandson, J. D. 2001. The Archaeology of Aquatic Adaptation: Paradigms for a New Millennium. Journal of Archaeological Research 9:287-350. Erlandson, J. M. 1994. Early Hunter-Gatherers of the California Coast. Plenum, New York. 329 Erlandson, J. M., Rick, T.C., Braje, T.J., Casperson, M., Culleton, B., Fulfrost, B., Garcia, T., Guthrie, D.A., Jew, N., Kennett, D.J., Moss, M.L., Reeder, L., Skinner, C., Watts, J., Willis, L. 2011. Paleoindian seafaring, maritime technologies, and coastal foraging on California's channel islands. Science 331(6021): 1181-1185. Erlandson, J. D., T. C. Rick, and T. J. Braje. 2009. Fishing Up the Food Web?: 12,000 Years of Maritime Subsistence and Adaptive Adjustments on California's Channel Islands. Pacific Science 63:711-724. Erlandson, J. M., T. J. Braje, and T. C. Rick. 2008. Tuqan Chert: A "Mainland" Monterey Chert Source on San Miguel Island, CA. Pacific Coast Archaeological Society Quarterly 40:23-34. Erlandson, J. M., D. J. Kennett, R. J. Behl, and I. Hough. 1997. The Cico Chert Source on San Miguel Island, California. Journal of California and Great Basin Anthropology 19:124-130. Everitt, B. S. 1977. The Analysis of Contingency Tables. Chapman & Hall, London. Fagan, B. 2004. The House of the Sea: An Essay on the Antiquity of Planked Canoes in Southern California. American Antiquity 69(1):7-16. F?blot-Augustins, J. 1993. Mobility strategies in the Late Middle Palaeolithic of Central Europe and Western Europe: Elements of stability and variability. Journal of Anthropological Archaeology 12:211-265. F?blot-Augustins, J. 2009. Revisiting European Upper Paleolithic Raw Material Transfers: The Demise of the Cultural Ecological Paradigm? Pages 25-46 in B. Adams and B. S. Blades, editors. Lithic Materials and Paleolithic Societies. Wiley-Blackwell, Oxford, UK. Fedje, D. W., I. D. Sumpter, and J.R. Southon. 2009. Sea-levels and Archaeology in the Gulf Islands National Park Reserve. Canadian Journal of Archaeology 33(2): 234-253. Field, J. S. 2005. Land tenure, competition and ecology in Fijian prehistory. Antiquity 79:586- 600. Field, J. 2004. Environmental and climatic considerations: a hypothesis for conflict and the emergence of social complexity in Fijian prehistory. Journal of Anthropological Archaeology 23: 79-99. Field, J. S. 2003. The Evolution of Competition and Cooperation in Fijian Prehistory: Archaeological Research in the Sigatoka Valley, Fiji. Unpublished Ph.D. dissertation. Department of Anthropology, University of Hawai'i, Manoa. Field, J. S. and Lape, P.V. 2010. Paleoclimates and the emergence of fortifications in the tropical Pacific islands. Journal of Anthropological Archaeology 29:113-124. Finlayson, D. P. 2006. The Geomorphology of Puget Sound Beaches. Unpublished Ph.D. dissertation. Department of Earth and Space Sciences, University of Washington, Seattle Fitzhugh, B. 2003. The Evolution of Complex Hunter-Gatherers: Archaeological Evidence from the North Pacific. Kluwer-Plenum, New York. 330 Fitzhugh, B., Phillips, S. Colby, and Gjesfjeld, Erik. 2011. Modeling Variability in Hunter- Gatherer Information Neworks: An Archaeological Case Study from the Kuril Islands. Pages 85- 115 in R. Whalon, Lovis, W., and Hitchcock, R., editor. Information and its Role in Hunter- Gatherer Bands. UCLA Cotson Institute for Archaeology, Los Angeles. Fitzhugh, B. and D. J. Kennett. 2010. Modeling Seafaring Intensity and Island-Mainland Exchange Along the Pacific Coast of North America. Pages 69-80 in A. Anderson, J. Barrett, and K. Boyle, editors. The Global Origins of Seafaring. University of Cambridge Press, Cambridge. Fladmark, K. 1975. A Paleoecological Model for Northwest Coast Prehistory. National Museum of Man, Ottawa, Ontario. Frison, G. C. 1979. Observations on the Use of Stone Tools: Dulling of Working Edges on Some Chipped Stone Tools in Bison Butchering. Pages 259-268 in B. Hayden, editor. Lithic Use-Wear Analysis. Academic Press, New York. Fujikawa, W. Y. 2002. A 7600-year vegetation and fire history of Mt. Constitution, Orcas Island, Washington, U.S.A. Unpublished M.S. thesis. Department of Earth and Space Sciences, University of Washington, Seattle. Gamble, C. 1999. The Palaeolithic Societies of Europe. Cambridge University Press, Cambridge. Gamble, L. H. 2008. The Chumash World at European Contact: Power, Trade, and Feasting among Complex Hunter-Gatherers. University of California Press, Berkeley. Gamble, L. H. 2005. Culture and Climate: Reconsidering the Effect of Palaeoclimatic Variability Among Southern California Hunter-Gatherer Societies. World Archaeology 37:92-108. Gamble, Lynn H. 2002. Archaeological Evidence for the Origin of the Plank Canoe in North America. American Antiquity 67(2):301-315 Gavin, D. and L.B. Brubaker. 1999. A 6000-Year Soil Pollen Record of Subalpine Meadow Vegetation in the Olympic Mountains, Washington, USA. Journal of Ecology 87:106-122. Gavin, D., L.B. Brubaker, L. B., and K.P. Lertzman. 2003. Holocene Fire History of a Coastal Temperate Rain Forest Based on Soil Charcoal Radiocarbon Dates. Ecology 84:186-201. Gerard, V. A. 1997. The role of nitrogen nutrition in high-temperature tolerance of kelp, Laminaria saccharina. Journal of Phycology 33:800-810. Glascock, M. D., G.E. Braswell, and R.H. Cobean. 1998. Systematic Approach to Obsidian Source Characterization. Pages 15-65 in M. S. Shackley, editor. Archaeological Obsidian Studies: Method and Theory. Society for Archaeological Science. Glassow, M. A. 2004. Identifying Complexity during the Early Prehistory of Santa Cruz Island Pages 7-17 in J.E. Arnold, editor. Foundations of Chumash Complexity. Perspectives in California Archaeology. Cotsen Intitute of Archaeology, University of California, Los Angeles. Glassow, M.A. 1996. Purisme?o Chumash Prehistory: Maritime adaptations along the Southern California coast. Harcourt Brace College Publishers, Fort Worth. 331 Goodyear, A. C. 1979. A Hypothesis for the Use of Cryptocrystalline Raw Material Among Paleo-Indian Groups of North America. Institute of Archaeology and Anthropology, University of South Carolina, Columbia, SC. Gould, R. A. and S. Saggers. 1985. Lithic Procurement in Central Australia: A Closer Look at Binford's Idea of Embeddedness in Archaeology. American Antiquity 50:117-136. Graesch, A. P. 2007. Modeling ground slate knife production and implications for the study of household labor contributions to salmon fishing on the Pacific Northwest Coast. Journal of Anthropological Archaeology 26:576-606. Graesch, A. P. 2000. Chumash Houses, Households, and Economy: Post-Contact Production and Exchange on Santa Cruz Island. Unpublished M.A. thesis. Department of Anthropology. University of California, Los Angeles. Graumlich, L. J., and Brubaker, L. B. 1986. Reconstruction of annual temperature (1590-1979) for Longmire, Washington, derived from tree rings. Quaternary Research 25:223-234. Grayson, D. K. and Delpeche, F. 2003. Ungulates and the Middle-to-Upper Paleolithic transition at Grotte XVI (Dordogne, France). Journal of Archaeological Science 30:1633-1648. Green, N. L. 1989. Geology and Petrology of the Mount Garibaldi Volcanic Field, Garibaldi Volcanic Belt, Southwestern British Columbia, Canada. Continental Magmetism Abstracts New Mexico Bureau of Mines and Mineral Resources:114. Green, N. L., R. L. Armstrong, J. E. Harakal, J. G. Souther, and P. B. Read. 1988. Eruptive history and K-Ar geochronology of the late Cenozoic Garibaldi volcanic belt, southwestern British Columbia. Geological Society of America Bulletin 100:563-579. Greenwald, D. N. and Brubaker, L.B. 2001. A 5000-year record of disturbance and vegetation change in riparian forests of the Queets River, Washington, USA. Canadian Journal of Forest Research 31:1375-1385. Grier, C. 2003. Dimensions of Regional Interaction in the Prehistoric Gulf of Georgia. Pages 170-187 in R. G. Matson, Coupland, Gary, and Mackie, Quentin, editor. Emerging from the Mist. Studies in Northwest Coast Culture History. University of British Columbia Press, Vancouver, British Columbia. Grier, C. 2001. The Social Economy of a Prehistoric Northwest Coast Plankhouse. Unpublished Ph.D. dissertation. Department of Anthropology, Arizona State University, Tempe, Arizona. Grier, C., Dolan, P., Derr, K., and McLay, E. 2009. Assessing Sea Level Changes in the Southern Gulf Islands of British Columbia Using Archaeological Data from Coastal Spit Locations. Canadian Journal of Archaeology 33(2):254-280. Gumerman, G. J. 1986. The Role of Competition and Cooperation in the Evolution of Island Societies. Pages 29-42 in P. V. Kirch, editor. Island Societies: Archaeological Approaches to Evolution and Transformation. Cambridge University Press, New York. 332 Haas, J. 1990. Warfare and the evolution of tribal polities in the prehistoric southwest. Pages 171-189 in J. Haas, editor. The Anthropology of War. Cambridge University Press, Cambridge. Hallett, D. 2001. Holocene Fire History and Climate Change in Southern British Columbia, Based on High Resolution Analyses of Sedimentary Charcoal. Unpublished Ph.D. dissertation. Simon Fraser University, Burnaby, British Columbia. Hallett, D. J., Lepofsky, D.S., Mathewes, R.W., and Lertzman, K.P. 2003. 11000 Years of Fire History and Climate in the Mountain Hemlock Rain Forests of Southwestern Brtish Columbia Based on Sedimentary Charcoal. Canadian Journal of Forest Resources 33:292-3112. Halstead, P. and J.O. O?Shea. 1989. Bad Year Economics: Cultural Responses to Risk and Uncertainty. Cambridge University Press, New York. Hamilton, W. D. 1964. The genetic evolution of social behavior. Journal of Theoretical Biology 7:1-52. Harrington, J. P. 1942. Culture Element Distributions: XIX California Coast. University of California Anthropological Records 7:1-46. Hay, M. B., A. Dallimore, R.E. Thomson, S.E. Clavert, and R. Pienitz. 2007. Siliceous Microfossil Record of Late Holocene Oceanography and Climate along the West Coast of Vancouver Island, British Columbia (Canada). Quaternary Research 67: 33-49. Hayden, B. 1998. Observations on the Prehistoric Social and Economic Structure of the North American Plateau. World Archaeology 29: 242-261. Hayden, B. 1989. From chopper to celt: The evolution of resharpening techniques. Pages 7-16 in R. Torrence, editor. Time, Energy and Stone Tools. Cambridge University Press, Cambridge. Hayden, B., Reinhardt, G.A., Holberg, D., Crellin, D. 1996. Pages 151-163 in G.Coupland and E.B. Banning, editors. Space Per Capita and the Optimal Size of Housepits. Monographs in World Prehistory, No. 27. Prehistory Press, Madison, WI. Hebda, R. J. 1995. British Columbia Vegetation and Climate History with Focus on 6 ka BP. G?ographie physique et Quaternaire 49:55-79. Hebda, R. J. and R. W. Mathewes. 1984. Holocene History of cedar and Native Indian Cultures of the North American Pacific Coast. Science 225:711-713. Heizer, R. F. 1940. The Frameless Plank Canoe of the California Coast. Primitive Man 13:80-89. Heizer, R. F. 1938. The Plank Canoe of the Santa Barbara Channel Region, California. Ethnological Studies 7:193-227. Herbel, B. C., Olson, D.L., and Schwarzmiller, R. 2001. Artifact Descriptions. In R. Schalk, Kenady, Steven, Herbel, B., Schwarzmiller, R., Kopperl, R.E., Olson, D.L., and Breidenthal, M., editor. Archaeological Testing at Cama Beach State Park. Cascadia Archaeology, Report Submitted on behalf of Washington State Parks and Recreation Commission, on file at the Department of Archaeology and Historic Preservation Olympia, WA. 333 Hill, K., M. Barton, M., and A.M. Hurtado. 2009. The emergence of human uniqueness: Characters underlying behavioral modernity. Evolutionary Anthropology 18(5):187-200. Hiscock, P. 1994. Technological responses to risk in Holocene Australia. Journal of World Prehistory 8:267-292. Holdaway, S., W. Wendrich, and R. Phillipps. 2010. Identifying low-level food producers: detecting mobility from lithics. Antiquity 84:185-194. Holm, B. 1994. Canoes of the Northwest Coast. Pages 259-264 in W. Fitzhugh and V. Chaussonnet, editors. Anthropology of the North Pacific Rim. Smithsonian Institution Press, Washington, D.C. Holmes, W. H. 1894. Natural history of flaked stone implements. Pages 120-139 in C. S. Wake, editor. Memoirs of the International Congress of Anthropology. Schulte, Chicago. Horner, R. A., D. L. Garrison, and F. G. Plumley. 1997. Harmful Algal Blooms and Red Tide Problems on the U.S. West Coast. Limnology and Oceanography 42:1076-1088. Howard, W. J. and L. M. Raab. 1993. Olivella Grooved Rectangle Beads as Evidence of Early- Period Southern Channel Islands Interaction Sphere. Pacific Coast Archaeological Society Quarterly 29(3):1-11. Hudson, D. T., and T. C. Blackburn. 1987. The Material Culture of the Chumash Interaction Sphere, Vol. V: Manufacturing Processes, Metrology, and Trade. Ballena Press Anthropological Papers No. 31. Ballena Press, Menlo Park, CA. Hudson, T., Timbrook, J., and Rempe, M. 1978. Tomol: Chumash Watercraft as Described in the Ethnographic Notes of John P. Harrington. Ballena Press Anthropological Papers No. 9, Soccorro. Hutchinson, I. 1992. Holocene Sea Level Change in the Pacific Northwest: A Catalogue of Radiocarbon Dates and an Atlas of Regional Sea level Curves. Institute for Quaternary Research, Discussion Paper, No. 1, Simon Fraser University, BC. Huth, P. K. 1996. Standing Your Ground: Territorial Disputes and International Conflict. University of Michigan Press, Ann Arbor. Inizan, M. -L., Reduron-Ballinger, M., Roche, H., and Tixier, J. 1999. Technology and Terminology of Knapped Stone. Pr?histoire de la Pierre Tail?e, 5. Cercle de Recherches et d'Etudes Pr?historiques, Nanterre. Inman, D. 1983. Application of coastal dynamics to the reconsruction of paleocoastlines in the vicinity of La Jolla, California. Pages 1-49 in P. Masters and N. Flemming, editors. Quaternary Coastlines and Marine Archaeology. Academic Press, New York. James, T. S., I. Hutchinson,J.V. Barrie, K.W. Conway, and D. Mathews. 2005. Relative sea-level change in the northern Strait of Georgia, British Columbia. G?ographie physique et Quaternaire 59(2/3): 113-128. 334 Janetski, J. C. 2002. Trade in Fremont Society: contexts and contrasts. Journal of Anthropological Archaeology 21:344-370. Johnson, J. K. and C.A. Morrow, 1989. The utility of production trajectory modeling as a framework for regional analysis. Pages 119-138 in D. O.Henry and G.H. Odell, editors. Alternative Approaches to Lithic Analysis. American Anthropological Association, Washington, D.C. Johnson, J. R. 2004. Social Responses to Climate Change Among the Chumash Indians of South- Central California. Pages 149-171 in Jones, L.M. and Jones, T.L, editors. Prehistoric California: Archaeology and the Myth of Paradise. The University of Utah Press, Salt Lake City, UT. Johnson, P. R. 2010. Elemental Analysis of Fine-Grained Basalt Sources from the Samoan Island of Tutuila: Applications of Energy-Dispersive X-Ray Fluorescence (EDXRF) and Instrumental Neutron Activation Analysis (INAA) Toward an Intra-Island Provenance Study. Pages 143-160 in M. S. Shackley, editor. X-Ray Fluorescence Spectrometry (XRF) in Geoarchaeology. Springer, New York. Johnston, B. E. 1962. California's Gabrielino Islands. Southwest Museum, Los Angeles, California. Jones, G. T., C. Beck, E. E. Jones, and R. E. Hughes. 2003. Lithic Source Use and Paleoarchaic Foraging Territories in the Great Basin. American Antiquity 68:5-38. Jones, G. T., D. G. Bailey, and Charlotte Beck. 1997. Source Provenance of Andesite Artifacts Using Non-Destructive Analysis. Journal of Archaeological Science 2:929-943. Jones, K. J. and D. B. Madsen. 1989. Calculating the Cost of Resource Transportation: A Great Basin Example. Current Anthropology 30:529-534. Jones, T. L., Brown, Gary.M., Raab, L. Mark, McVickar, Janet L., Spaulding, W. Geoffrey, Kennett, Douglas J., York, Andrew, and Walker, Phillip L. 1999. Environmental Imperatives Reconsidered: Demographic Crises in Western North America During the Medieval Climatic Anomaly. Current Anthropology 40:137-170. Jones, T. L., and K. A. Klar. 2005. Diffusionism reconsidered: linguistic and archaeological evidence for prehistoric Polynesian contact with southern California. American Antiquity 70: 457?484. Keddie, G. 1996. Aboriginal Defensive Sites. Royal British Columbia Museum, Victoria, B.C. www.royalbcmuseum.bc.ca/Content_Files/Files/. Keddie, G. 1984. Fortified Defensive Sites and Burial Cairns of the Songhees Indians. The Midden 16:7-9. Kelly, R. L. 2001. Prehistory of the Carson Desert and Stillwater Mountains. Environment, Mobility, and Subsistence in a Great Basin Wetland. University of Utah Press, Salt Lake City. 335 Kelly, R. L. 2000. Elements of a Behavioral Ecological Paradigm for the Study of Prehistoric Hunter-Gatherers. Pages 63-78 in M. B. Schiffer, editor. Social Theory in Archaeology. University of Utah Press, Salt Lake City, UT. Kelly, R. L. 1992. Mobility/Sedentism:Concepts, Archaeological Measures, and Effects. Annual Review of Anthropology 21:43-66. Kelly, R. L. 1991. Sedentism, sociopolitical inequality, and resource fluctuations. Pages 135-150 in S. Gregg, editor. Between Bands and States: Sedentism, Subsistence and Interaction in Small Scale Societies. Southern Illinois University Press, Carbondale, IL. Kelly, R. L. 1988. The Three Sides of a Biface. American Antiquity 53(4): 717-734. Kenady, S. M., M. C. Wilson, R. F. Schalk, and R. R. Mierendorf. 2011. Late Pleistocene butchered Bison antiquus from Ayer Pond, Orcas Island, Pacific Northwest: Age confirmation and taphonomy. Quaternary International Quaternary International 233:130-141. Kenady, S. M., R. F. Schalk, M. Wolverton, M. C. Wilson, and R. R. Mierendorf. 2008. A new perspective on the DeStaffany Site, an early lithic site in the San Juan Islands Washington. Current Research in the Pleistocene 25:105-108. Kenady, S. M., M. C. Wilson, and R. F. Schalk. 2007. Indications of butchering on a late- Pleistocene Bison antiquus from the maritime Pacific Northwest. Current Research in the Pleistocene 24:167-170. Kendig, W. E., K. N. Smith, and R. L. Vellanoweth. 2008. Sandstone Saws of Tule Creek Village (CA-SNI-25). 7th California Island Symposium, Oxnard, CA. Kendig, W. E., K. N. Smith, R. L. Vellanweth, J. A. Allen, C. M. Smith, and A. M. Points. 2010. The Use of Replicative Studies in Understanding the Function of Expedient Tools: The Sandstone Saws of San Nicolas Island, California. Journal of California and Great Basin Anthropology 30:193-210. Kennett, D.J. 2005. The Island Chumash: Behavioral Ecology of a Maritime Society. University of California Press, Berkeley. Kennett, D. J. and J. P. Kennett. 2000. Competitive and cooperative responses to climatic instability in coastal southern California. American Antiquity 65:379-395. Kennett, D. J. and R.A. Clifford. 2004. Flexible Strategies for Resource Defense on the Northern Channel Islands of California: An Agent Based Model. Pages 21-50 in S. M. Fitzpatrick, editor. Voyages of Discovery: The Archaeology of Islands. Praeger, Westport, Connecticut and London. Kennett, J. P. and B.L. Ingram. 1995. A 20,000-year record of ocean circulation and climate change from the Santa Barbara Basin. Nature 377:510-514. Kennett, D. J., B. Winterhalder, J. Bartruff, and J. M. Erlandson. 2009. An ecological model for the emergence of institutionalized social hierarchies on California's Northern Channel Islands. Pages 297-314 in S. Shennan, editor. Pattern and Process in Cultural Evolution. University of California Press, Berkeley. 336 Kennett, D. J., Kennett, James P., Erlandson, Jon M., and Cannariato, Kevin G. 2007. Human responses to Middle Holocene climate change on California's Channel Islands. Quaternary Science Reviews 26:351-367. Kennett, D., B. Ingram, J. Erlandson, and P. Walker. 1997. Evidence for Temporal Fluctuations in Marine Radiocarbon Reservoir Ages in the Santa Barbara Channel, Southern California. Journal of Archaeological Science 24:1051-1059. Kessler, R. A., C. Beck, and G. T. Jones. 2009. Trash: The Structure of Great Basin Paleoarchaeic Debitage Assemblages in Western North America. Pages 144-159 in B. Adams and B. S. Blades, editors. Lithic Materials and Paleolithic Societies. Wiley-Blackwell, Oxford. Koch, J., Osborn, G.D., and Clague, J.J. 2007. Pre-'Little Ice Age' Glacier Fluctuations in Garibaldi Provincial Park, Coast Mountains, British Columbia, Canada. The Holocene 17:1069- 1078. Koch, J., Menounos, B., Claue, J., and Osborn, G. 2004. Environmental Change in Garibaldi Provincial Park, Southern Coast Mountains, British Columbia. Geoscience Canada 31:127-135. Koerper, H.C. and Drover, C.E. 1983. Chronology Building for Coastal Orange County: The Case from CA-ORA-119-A. Pacific Coast Archaeological Society Quarterly 19(2):1?34 Kohler, T. A. and Van West, C.R. 1996. The Calculus of Self-Interest in the Development of Cooperation: Sociopolitical Development and Risk among the Northern Anasazi. Pages 169-196 in J. A. Tainter and B.B. Tainter, editors. Evolving Complexity and Environmental Risk in the Prehistoric Southwest. Addison-Wesley, Reading, MA. Kooyman, B. 2006. Boundary theory as a means to understanding social space in archaeological sites. Journal of Anthropological Archaeology 25:424-435. Kornbacher, K. D. 1992. Shell Midden Lithic Technology: Analysis of Stone Artifacts from British Camp. Pages 163-192 in J. K. Stein, editor. Deciphering a Shell Midden. Academic Press, San Diego. Kornbacher, K. D. 1989. Shell midden lithic technology: An investigation of change at British Camp (45SJ24), San Juan Island. Unpublished M.A. thesis. Department of Anthropology and Sociology, University of British Columbia, Vancouver, British Columbia. Kozloff, E. N. 1993. Seashore Life of the Northern Pacific Coast: An Illustrated Guide to Northern California, Oregon, Washington, and British Columbia. University of Washington Press, Seattle. Kozloff, E. N. 1990. Invertebrates. Saunders College Publications, Philadelphia. Krebs, J. R., and Davies, N. B. 1997. Behavioural Ecology: an Evolutionary Approach (Fourth Edition). Blackwell Science, Malden, Massachusetts. Kroeber, A. L. 1976 (reprinted from 1926). Handbook of the Indians of California. Smithsonian Institution, Washington, D.C. 337 Kruger, D. 2003. Evolution and altruism: Combining psychological mediators with naturally selected tendencies. Evolution and Human Behavior 24:118-125. Kuhn, S. L. 2004. Upper Paleolithic raw material economies at Ucagizli Cave, Turkey. Journal of Anthropological Archaeology 23:43-48. Kuhn, S. L. 1995. Mousterian lithic technology: An ecological perspective. Princeton University Press, Princeton. Kuhn, S. L. 1991. "Unpacking" reduction: Lithic raw material economy in the Mousterian of west-central Italy. Journal of Anthropological Archaeology 10:76-106. Kwarsick, K. C. 2010. Lithic Raw Material Procurement and the Technological Organization of Olympic Peninsula Peoples. Unpublished M.A.thesis. Department of Anthropology, Washington State University, Pullman, Washington. Kwarsick, K. 2008. A Stone's Throw Away: Local Sources for Olympic Peninsula Dacites (Poster). 73rd Annual Meeting of the Society for American Archaeology, Vancouver, British Columbia. Lambert, P. and Walker, P.L. 1991. Physical Anthropological Evidence for the Evolution of Social Complexity in Coastal Southern California. Antiquity 65:963-973. Lambert, P. 1997. Patterns of Violence in Prehistoric Hunter-gatherer Societies of Coastal Southern California. Pages 77-109 in D. L. Martin and D. W. Grayer, editors. Troubled time: violence and warfare in the past. Overseas Publishers Association, Amsterdam. Lambert, P. 1993. Health in Prehistoric Populations of the Santa Barbara Channel Islands. American Antiquity 58:509-522. Larocque, S. J. and Smith, D.J. 2003. Little Ice Age Glacial Activity in the Mt. Waddington Area, British Columbia Coast Mountains, Canada. Canadian Journal of Earth Sciences 40:1413- 1436. Lepofsky, D., D. M. Schaepe, A. P. Graesch, M. Lenert, P. Ormerod, K. T. Carlson, J. E. Arnold, M. Blake, P. Moore, and J. J. Clague. 2009. Exploring Sto:Lo-Coast Salish Interaction and Identity in Ancient Houses and Settlements in the Fraser Valley, British Columbia. American Antiquity 74:595-626. Lepofsky, D., K. Lertzman, and D. Hallett. 2005. Climate Change and Culture Change on the Southern Coast of British Columbia 2400-1200 Cal. B.P.: An Hypothesis. American Antiquity 70:267-293. Lertzman, K., Gabin, Daniel, Hallett, Douglas, Brubaker, Linda, Lepofsky, Dana, and Mathewes, Rolf. 2002. Long-Term Fire Regime Estimated from Soil Charcoal in Coastal Temperate Rainforests. Conservation Ecology 6: Electronic document, http://www.consecol.org/vol6/iss2/art5. 338 Lian, O. B. and S. R. Hicock. 2001. Lithostratigraphy and limiting optical ages of the Pleistocene fill in Fraser River valley near Clinton, south-central British Columbia. Canadian Journal of Earth Sciences 38:839-850. Lepofsky, D., M. Blake, D. Brown, S. Morrison, N. Oakes, and N.T. Lyons. 2000. The archaeology of the Scowlitz site, Southwestern British Columbia. Journal of Field Archaeology 27: 391-416. Lightfoot, K. G. 1993. Long-term developments in complex hunter-gatherer societies: Recent perspectives from the pacific coast of North America. Journal of Archaeological Research 1(3): 167-201. Long, C. J., Whitlock, C., Bartlein, P.J., Millspaugh, S.H. 1998. A 9000-Year Fire History from the Oregon Coast Range, Based on a High-Resolution Charcoal Study. Canadian Journal of Fire Resources 28:774-787. Luckman, B. H. 1994. Evidence for Climatic Conditions between ca. 900-1300 A.D. in the Southern Canadian Rockies. Climate Change 26:171-182. MacDonald, D. H. 2008. The role of lithic raw material availability and quality in determining tool kit size, tool function, and degree of retouch: A Case study from Skink Rockshelter (46NI445), West Virginia. Pages 216-232 in J. Andrefsky, W., editor. Lithic Technology: Measures of Production, Use and Curation. Cambridge University Press, Cambridge. Mackie, Q. 2001. Settlement Archaeology in a Fjordland Archipelago: Network Analysis, Social Practice and the Built Environment of Western Vancouver Island, British Columbia, Canada, Since 2,000 BP. Archaeopress, Oxford. Malinowski, B. 1922. Argonauts of the Western Pacific. Waveland Press, Long Grove, IL. Martindale, A. and K. Supernant. 2009. Quantifying the defensiveness of defended sites on the Northwest Coast of North America. Journal of Anthropological Archaeology 28:191-204. Martinez, J. 2010. Loss on Ignition to Determine Organic and Carbonate Content on San Nicolas Island, California, Poster. Poster Presented at the 13th Annual University of Washington Undergraduate Research Symposium Seattle, WA. Marty, J., Taylor, Amanda K., Poister, Nicholas, Martinez, Jordan, and Vellanoweth, Rene. 2010. Anatomy of a Mound: Sediment Analysis at Tule Creek Village, San Nicolas Island, CA. Poster Presented at the Society for American Archaeology Meetings St. Louis, MO. Martz, P. C. 2008. 4000 years on GHALAS-AT: Part One of the San Nicolas Island Index Unit Analysis. Report prepared for Naval Air Weapons Station China Lake, CA through a Cooperative Agreement with California State University Los Angeles N68711-00-LT-0041. Martz, P. 2005. Prehistoric subsistence and settlement on San Nicolas Island. Pages 65-82 in D. Garcelon and C. Schwemm, editors. Proceedings of the Sixth California Islands Symposium. National Park Service Technical Publication SHIS-05-01, Institute for Wildlife Sutdies, Arcata, CA. 339 Martz, P. C. 2002. San Nicolas Island Prehistoric Archaeological Sites Mapping and Recordation Project. On file, South Central Coastal Archaeological Information Center, California State University, Fullerton, CA. Martz, P. and Rosenthal, E.J. 2001. The Maritime Hunter-Gatherers of San Nicolas Island, California. Pages 59-70 in C.Stevenson, G.Lee, and F.J. Morin, editors. Pacific 2000 Proceedings of the Fifth International Conference on Easter Island and the Pacific. Easter Island Foundation. Maschner, H. D. G. and K. L. Reedy-Maschner. 1998. Raid, Retreat, Defend (Repeat): The Archaeology and Ethnohistory of Warfare on the North Pacific Rim. Journal of Anthropological Archaeology 17:19-51. Masters, P. M. and I. W. Aiello. 2007. Chapter 3. Postglacial Evolution of Coast Environments. Pages 35-52 in T. L. Jones and K. A. Klar, editors. California Prehistory: Colonization, Culture, and Complexity. AltaMira Press, Lanham, MA. Matson, R. G. 1992. The Evolution of Northwest Coast Subsistence. Research in Economic Anthropology Supplement 6:367-430. Matson, R. G. 1985. The relationship between sedentism and status inequalities among hunter- gatherers. Pages 245-252 in M. Thompson, Garcia, M.T., Kense, F.J., editors. Status, Structure and Stratification. Archaeological Association of the University of Calgary, Calgary, Alberta. Matson, R. G. 1983. Intensification and the development of cultural complexity: the Northwest versus the Northeast Coast. Pages 124-148 in R. J. Nash, editor. The Evolution of Maritime cultures on the Northeast and Northwest Coasts of North America. Simon Fraser University, Burnaby, BC. Matson, R. G. and G. Coupland 1995. Prehistory of the Northwest Coast. Academic Press, San Diego. Mazzoti, S., C. Jones, et al. 2008. Relative and absolute sea level rise in western Canada and northwestern United States from combined tide gauge-GPS analysis. Journal of Geophysical Research 113(C11019): 1-19. McCauley, C. 1990. Conference overview. Pages 1-25 in J. Haas, editor. The Anthropology of War. Cambridge University Press, Cambridge. McCawley, W. 2002. A Tale of Two Cultures: The Chumash and the Gabrielino. Pages 41-65 in J. H. Altschul and D. R. Grenda, editors. Islanders and Mainlanders: Prehistoric Context for the Southern California Bight. SRI Press, Tucson. McCawley, W. 1996. The First Angelinos: The Gabrielino Indians of Los Angeles. Malki Museum Press, Banning, CA. McCay, B. J. 1978. Systems Ecology, People Ecology, and the Anthropology of Fishing Communities. Human Ecology 6:397-422. 340 McCoy, M. D. and T. N. Ladefoged 2010. Reconstructing lithic supply zones and procurement areas: an example from the Bay of Islands, northland, New Zealand. Journal of Pacific Archaeology 1(2): 174-183. Meighan, C. 1954. The Nicoleno. Pacific Discovery 7:22-27. Metcalfe, D. and K. R. Barlow. 1992. A model for exploring the optimal trade-off between field processing and transport. American Anthropologist 94:340-356. Meyer, G. A. and Pierce, J.L. 2003. Climate Controls on Fire-Induced Sediment Pulses in Yellowstone National Park and Central Idaho: A Long-Term Perspective. Forest Ecology and Management 178:89-104. Miller, B. G. 1989. Centrality and Measures of Regional Structure in Aboriginal Western Washington. Ethnology 28:265-276. Mills, P. R., S. P. Lundblad, J. S. Field, A. B. Carpenter, W. K. McElroy, and P. Rossi. 2010. Geochemical sourcing of basalt artifacts from Kaua'i, Hawaiian Islands. Journal of Archaeological Science 37:3385-3393. Mills, P. R., S. P. Lundblad, J. G. Smith, P. C. McCoy, and S. P. Naleimaile. 2008. Science and Sensitivity: A Geochemical Characterization of the Mauna Kea Adze Quarry Complex, Hawaii Island, Hawaii.American Antiquity 73:743-758. Mitchell, D. H. 1971 Archaeology of the Gulf of Georgia Culture Area: A Natural Region and Its Culture Types, Syesis 4(Supplement No.1), Victoria, British Columbia. Morgan, V. E. 1999. The SR-101 Sequim Bypass Archaeological Project: Mid- to Late-Hoocene Occupations on the Northern Olympic Peninsula, Clallam County, Washington. Eastern Washington University Reports in Archaeology and History 100-108. Archaeological and Historical Services, Eastern Washington University, Cheney, Washington. Morin, J. 2004. Cutting Edges and Salmon Skin: An Investigation of Variation in Salmon Processing Technology on the Northwest Coast. Canadian Journal of Archaeology 28:281-318. Morrow, C. A. and R. W. Jeffries. 1989. Trade or embedded procurement? A test case from southern Illinois. Pages 27-33 in R. Torrence, editor. Time, Energy and Stone Tools. Cambridge University Press, Cambridge. Morrow, T. A. 1997. A chip off the old block: alternative approaches to debitage analysis. Lithic Technology 22:51-69. Mosher, D. C. and A. T. Hewitt. 2004. Late Quaternary deglaciation and sea-level history of eastern Juan de Fuca Strait, Cascadia. Quaternary International 121:23-29. Moss, M. L. 2011. Northwest Coast: Archaeology as Deep History. The SAA Press, Washington, D.C. Moss, M. L., Peteet, D.M., and Whitlock, C. 2007. Mid-Holocene Culture and Climate on the Northwest Coast of North America. Pages 491-529 in D. G. Anderson, Maasch, K.A., and 341 Sanweiss, D.H., editors. Climate Change and Cultural Dynamics: A Global Perspective on Mid- Holocene Transitions. Elsevier, New York. Moss, M. L. and J. M. Erlandson. 1992. Forts, Refuge Rocks, and Defensive Sites: the Antiquity of Warfare along the North Pacific Coast of North America. In: Maritime Cultures of Southern Alaska: Papers in Honor of Richard H. Jordan. Arctic Anthropology 29:73-90. Neel, D. 1995. The Great Canoes: Reviving a Northwest Coast Tradition. University of Washington Press, Seattle. Neiman, F. 1997. Conspicuous Consumption as Wasteful Advertising: A Darwinian Perspective on Spatial Patterns in Classic Maya Terminal Monument Dates. Pages 267-290 in C.M. Barton and G.A. Clark, editors. Rediscovering Darwing: Evolutionary Theory in Archaeological Explanation. Archaeological Papers of the American Anthropological Association No. 7. American Anthropological Association, Arlington, VA. Nelson, M. C. 1991. The study of technological change. Pages 57-100 in M. B. Schiffer, editor. Archaeological Method and Theory. University of Arizona Press, Tucson. Odell, G. H. 2004. Lithic Analysis. Kluwer, New York. Odell, G. H. 2000. Stone Tool Research at the End of the Millennium: Procurement and Technology. Journal of Archaeological Research 8:269-331. Odell, G. H. 1998. Investigating Correlates of Sedentism and Domestication in Prehistoric North America. American Antiquity 63:553-571. Odell, G. H. 1996. Economizing behavior and the concept of "curation.". Pages 51-80 in G. H. Odell, editor. Stone Tools: Theoretical Insights into Human Prehistory. Plenum, New York. O'Lear, S. 2005. Resource concerns for territorial conflict. GeoJournal 64:297-306. Onat, A. B. 1984. The Interaction of Kin, Class, Marriage and Property Ownership and Residence with Respect to Resource Locations Among the Coast Salish of the Puget Lowlands. Northwest Anthropological Research Notes 18:86-96. Ostrom, E., R. Gardner, and J. Walker. 1994. Rules, Games, and Common-pool Resources. Cambridge University Press, New York. Ostrom, E. 1990. Governing the Commons. Cambridge University Press, New York. Page, D. 2008. Fine-Grained Volcanic Toolstone Sources and Early Use in the Bonneville Basin of Western Utah and Eastern Nevada. Unpublished M.A. thesis. Department of Anthropology, University of Nevada, Reno, Reno, NV. Pak, D. K. and J. P. Kennett. 2002. A foraminiferal isotopic proxy for upper water mass stratification. Journal of Foraminiferal Research. 32:319-327. Pak, D., Kennett, J.P., and Kashgarian, M. 1997. Panktonic Foraminiferal Proxies of Hydrographic Conditions in the Santa Barbara Basin. EOS 78:359. 342 Palmer, C. T. 1991. Kin-Selection, Reciprocal Altruism, and Information Sharing Among Maine Lobstermen. Ethology and Sociobiology 12:221-235. Parr, M., Stein, Julie K. and Phillips, Laura S. 2002. Chapter 4: Field and Laboratory Methods and Procedures. Pages 33-45 in J. K. a. P. Stein, Laura S., editor. Vashon Island Archaeology: A View from Burton Acres Shell Midden. University of Washington Press, Seattle. Parr, M., L. S. Phillips, and J. K. Stein. 2011. Field Methods. Pages 21-28 in A. K. Taylor and J. K. Stein, editors. Is it a House? Archaeological Excavations at English Camp, San Juan Island, Washington. University of Washington Press, Seattle. Parry, W. J. and Kelly, R.L. 1987. Expedient core technology and sedentism. Pages 285-304 in J. K. Johnson and C.A. Morrow, editors. The Organization of Core Technology. Westview Press, Boulder, CO. Pecora, A. M. 2001. Chipped stone tool production strategies and lithic debris patterns. Pages 173-191 in J. Andrefsky, W., editor. Lithic Debitage: Context, Form, Meaning. University of Utah Press, Salt Lake City. Pellatt, M., Hebda, R.J., and Mathewes, R.W. 2001. High Resolution Holocene Vegetation History and Climate from Hole 1034B, ODP Leg 169S, Saanich Inlet Canada. Marine Geology 174:211-226. Phillips, S. C. 2001. Networked Glass: Lithic Raw Material Consumption and Social Networks in the Kuril Islands, Far Eastern Russia. Unpublished Ph.D. dissertation. Department of Anthropology, University of Washington, Seattle. Pisias, N. G. 1979. Model for paleoceanographic reconstructions of the California current for the last 8,000 years. Quaternary Research 11:373-386. Pisias, N. G. 1978. Paleoceanography of the Santa Barbara Basin During the Last 8000 Years. Quaternary Research 10:366-384. Pletka, S. 2001. Bifaces and the Institutionalization of Exchange Relationships. Pages 133-163 in J. E. Arnold, editor. The Origins of a Pacific Coast Chiefdom: The Chumash of the Channel Islands. University of Utah Press, Salt Lake City. Pollard, M. C., C. Batt, B. Stern, and S.M.M. Young. 2007. Analytical Chemistry in Archaeology. Cambridge University Press, Cambridge. Prasciunas, M. M. 2007. Bifacial cores and flake production efficiency: An experimental test oftechnological assumptions. American Antiquity 72:334-348. Prentiss, A. M., G. Cross, T.A. Foor, D. Markle, M. Hogan, and D.S. Clarke. 2008. Evolution of a Late Prehistoric Winter Village on the Interior Plateau of British Columbia: Geophysical Investigations, Radiocarbon Dating, and Spatial Analysis of the Bridge River Site. American Antiquity 73: 59-82. Prentiss, W. C. and I. Kujit 2004. Evolution of collector systems on the Canadian Plateau. Pages 35-45 in W.C. Prentiss and I. Kujit, editors. Complex Hunter-Gatherers: Evolution and 343 Organization of Prehistoric Communities on the Plateau of Northwestern North America. W. C. Prentiss and I. Kujit. University of Utah Press, Salt Lake City, UT. Preziosi, A. M. 2001. Standardization and Specialization: The Island Chumash Microdrill Industry. Pages 151-164 in J.E. Arnold, editor. The Origins of a Pacific Coast Chiefdom: The Chumash of the Channel Islands. University of Utah Press, Salt Lake City, UT. Prior, C., Gray, D.W., and Sander, J. 1999. Radiocarbon Analysis of the Upwelling Effect on Marine Shell from San Nicolas Island, Ventura County, California. Statistical Research Technical Report 97?20, Redlands, CA. Raab, L. M. and D. O. Larson. 1997. Medieval climatic anomaly and punctuated cultural evolution in coastal Southern California. American Antiquity 62(2):319-336. Raab, L.M., Bradford, K., Porcasi, J.F., and Howard, W.J. 1995. Return to Little Harbor, Santa Catalina Island, California: A Critique of the Marine Paleotemperature Model. American Antiquity 60(2):287-308. Rautman, A. E. 1993. Resource Variability, Risk, and the Structure of Social Networks: An Example from the Prehistoric Southwest. American Antiquity 58(3): 403-424. Reimer P.J., B., M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C. Blackwell, P.G., Buck, C.E., Burr, G., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, T.P., Guilderson, K.A., Hughen, K.A., Kromer, B., McCormac, F.G., Manning, S., Bronk Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talama, S., Taylor, F.W., van der Plicht, J., Weyhenmeyer, C.E. . 2004. Radiocarbon Comparison from 26-50 cal kyr BP - NotCa104 -- Comparison/Calibration 14C Records 26-50 cal kyr BP. Radiocarbon 46:1029-1058. Reimer, R. 2000. Extreme Archaeology: The Results of Investigations at High Elevation Regions in the Northwest. Unpublished M.A. thesis. Department of Archaeology, Simon Fraser University, Burnaby, BC. Reinman, F. M. and Lauter, G.A. 1984. San Nicolas Island Cultural Resource Survey Report. F.M. Reinman and Associates for Pacific Missile Test Center, Point Mugu. Reinman, F. M. and Townsend, S.J. 1960. Six Burial Sites on San Nicolas Island. University of California Archaeology Survey Annual Report 2:1-134. Renfrew, C. 1977. Alternative models for exchange and spatial distribution. Pages 71-90 in J. Ericson and T. Earle, editors. Contexts for Prehistoric Exchange. Academic Press, New York. Renfrew, C. 1975. Trade as action at a distance. Pages 3-59 in J. Sabloff and C. C. Lamber- Karlovsky, editors. Ancient Civilization and Trade. University of New Mexico Press, Albuquerque. Renfrew, C., J. E. Dixon, and J. R. Cann. 1968. Obsidian and the origins of trade. Scientific American 218:38-46. 344 Renouf, M.A.P. 1988.Sedentary coastal hunter-fishers: an example from the Younger Stone Age of northern Norway. Pages 102-115 in Bailey, G. and Parkington, J., editors. The Archaeology of Prehistoric Coastlines. Cambridge University Press, Cambridge. Renouf, M. A. P. 1984. Northern Coast Hunter-Fishers: An Archaeological Model. World Archaeology 16:18-27. Rensink, E., J. Kolen, and A. Spieksma. 1991. Patterns of Raw Material Distribution in the Upper Pleistocene of Northwestern and Central Europe. Pages 142-159 in A. Montet-White and S. Holen, editors. Raw Material Economies among Prehistoric Hunter-Gatherers. University of Kansas, Lawrence. Richardson, A. 1982. The Control of Productive Resources on the Northwest Coast of North America. Pages 93-112 in N. Williams and E. Hunn, editors. The Control of Productive Resources on the Northwest Coast of North America. Westview Press, Boulder. Rick, T. C. 2007. Household and Community Archaeology at the Chumash Village of Niaqla, Santa Rosa Island, California. Journal of Field Archaeology 32: 243-263. Rick, T., J. Erlandson, R. Vellanoweth, and T. Braje. 2005a. From Pleistocene Mariners to Complex Hunter-Gatherers: The Archaeology of the California Channel Islands. Journal of World Prehistory 19:169-228. Rick, T. C., Vellanoweth, Rene L., and Erlandson, Jon. 2005b. Radiocarbon dating and the ?old shell? problem: direct dating of artifacts and cultural chronologies in coastal and other aquatic regions. Journal of Archaeological Science 32:1641-1648. Rick, T. C., C. E. Skinner, J. M. Erlandson, and R. L. Vellanoweth. 2001a. Obsidian Source Characterization and Human Exchange Systems on California?s Channel Islands. Pacific Coast Archaeological Society Quarterly 37:27-44. Rick, T. C., J. M. Erlandson, and R. L. Vellanoweth. 2001b. Paleocoastal Marine Fishing on the Pacific Coast of the Americas: Perspectives from Daisy Cave, California. American Antiquity 66:595-613. Robinson, S. 1963. Spirit Dancing Among the Salish Indians, Vancouver Island, British Columbia. Unpublished Ph.D. dissertation. Department of Anthropology, University of Chicago, Chicago, IL. Roceck, T. R. and O. Bar-Yosef, editors. 1998. Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites.Peabody Museum Bulletin 6, Harvard University, Cambridge, MA. Rondeau, M. F. 1995. Cobble Core Types of California. Proceedings of the Society for California Archaeology 8:33-45. Root, M. J. 2004. Technological Analysis of Flake Debris and the Limitations of Size-grade Techniques. Pages 65-94 in C. T. Hall and M. L. Larson, editors. Aggregate Analysis in Chipped Stone. The University of Utah Press, Salt Lake City. 345 Rosenthal, E. J. 1996. San Nicolas Island Bifaces: A Distinctive Stone Tool Manufacturing Technique. Journal of California and Great Basin Anthropology 18:303-314. Rosenthal, E. J. 2008. A Comparison of Index Unit Stone Artifacts from San Nicolas Island. Pages 122-209 in P. Martz, editor. 4000 Years on Ghalas-at: Part One of the San Nicolas Island Index Unit Analysis Program. Prepared for Naval Air Weapons Station China Lake, CA Through a Cooperative Agreement with California State University, Los Angeles N68711-00-LT-0041, Los Angeles. Ross, A., B. Anderson, and C. Campbell. 2003. Gunumbah: Archaeological and Aboriginal meanings at a quarry site on Moreton Island, southeast Queensland. Australian Archaeology 57:75-81. Ryder, J. M. 1987. Neoglacial History of the Stikine-Iskut Area, Northern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences 24:1294-1301. Ryder, J. M. 1989. Holocene Glacier Fluctuations (Canadian Cordillera).in R. J. Fulton, editor. Quaternary Geology of Canada and Greenland. Geological Survey of Canada, Ottawa, Ontario. Sames, C. W. 1966. Morphometric data of some recent pebble associations and their application to ancient deposits. Journal of Sedimentary Petrology 36:126-142. Sandweiss, D. H., Richardson, J.B., Reitz, E.J., Rollins, H.B., and Maasch, K.A. 1996. Geoarchaeological evidence from Peru for a 5,000 years B.P. onset of El Nino. Science 273:1531-1533. Schaepe, D. M. 2006. Rock Fortifications: Archaeological Insights into Pre-Contact Warfare and Sociopolitical Organization Among the Sto:lo of the Lower Fraser River Canyon, B.C. American Antiquity 71:671-705. Schaepe, D. M. 2009. Pre-Colonial Sto:lo-Coast Salish Community Organization: An Archaeological Study. Unpublished Ph.D. dissertation. University of British Columbia, Vancouver, BC. Schwartz, S.J. and P. Martz. 1992. An Overview of the Archaeology of San Nicolas Island, Southern California. Pacific Coast Archaeological Society Quarterly 28(4):46-75. Schwartz, S. J. and K. A. Rossbach. 1993. A Preliminary Survey of Historic Sites on San Nicolas Island. Proceedings of the Society for California Archaeology 6:189-198. Seeman, M. F. 1994. Intercluster Lithic patterning at Nobles Pond: a case for disembedded procurement among Early Paleoindian societies. American Antiquity 59:273-288. Sellet, F. 1993. Cha?ne op?ratoire: The concept and its application. Lithic Technology 18:106- 112. Shott, M. J. 1989. On Tool-Class Use Lives and the Formation of Archaeological Assemblages. American Antiquity 54:9-30. 346 Skinner, C. E. 2011. United States and Canada Obsidian Source Catalog. Northwest Research Obsidian Studies Laboratory http://www.sourcecatalog.com/index.html. Skinner, C. E., Thatcher, Jennifer E., Jenkins, Dennis L., and Oetting, Albert C. 2004. X-Rays, Artifacts, and Procurement Ranges: A Mid-Project Snapshot of Prehistoric Procurement Patterns in the Fort Rock Basin of Oregon. Pages 221-232 in D. L. Jenkins, Connolly, Thomas J., and Aikens, C. Melvin, editor. Early and Middle Holocene Archaeology of the Northern Great Basin. University of Oregon Anthropological Papers, Eugene, OR. Smith, D. J., and Desloges, J.R. 2000. Little Ice Age History of Tzeetsaytsul Glacier, Tweedsmuir Provincial Provincial Park, British Columbia. G?ographie physique et Quaternaire 54:135-141. Smith, E. A. 1988. Risk and Uncertainty in the 'Original Affluent Society': Evolutionary Ecology of Resource Sharing and Land Tenure. Pages 222-251 in T. Ingold, D. Riches, and J. Woodburn, editors. Hunters and Gatherers I: History, Evolution, and Social Change. Berg, New York. Smith, E. A. 1983. Anthropology, Evolutionary Ecology, and the Explanatory Limitations of the Ecosytem Concept. Pages 51-85 in E. F. Moran, editor. The Ecosystem Concept in Anthropology. Westview Press, Bould, CO. Smith, E. A. 1981. The Application of Optimal Foraging Theory to the Analysis of Hunter- Gatherer Group Size. Pages 36-65 in B. Winterhalder, and Smith, E.A., editors. Hunter-Gatherer Foraging Strategies. University of Chicago Press, Chicago. Smith, E. A., and Winterhalder, B. 1992. Evolutionary Ecology and Human Behavior. Aldine de Gruyter, New York. Smith, G. M. 2010. Footprints across the Black Rock: Teporal Variability in Prehistoric Foraging Territories and Toolstone Procurement Strategies in the Western Great Basin. American Antiquity 75:865-885. Steffen, A., A. E. Skinner, and P. W. Ainsworth. 1998. A View to a Core: Technological Units and Debitage Analysis. Pages 131-146 in A. F. Ramenofsky and A. Steffen, editors. Unit Issues in Archaeology. University of Utah Press, Salt Lake City. Stein, J. K. 2000. Exploring Coast Salish Prehistory: The Archaeology of San Juan Island. University of Washington Press, Seattle. Stein, J. K., editor. 1992. Deciphering a Shell Midden. Academic Press, San Diego. Stein, J. K., J. N. Deo, and L. S. Phillips. 2003. Big Sites Short Time: Accumulation Rates in Archaeological Sites. Journal of Archaeological Science 30:297-316. Steneck, R. S., M. H. Graham, B. J. Bourque, D. Corbett, J. M. Erlandson, J. A. Estes, and M. J. Tegner. 2002. Kelp forest ecosystems: biodiversity, stability, resilience and future. Environmental Conservation 29:436-459. Stephens, D. W. and Krebs, J. R. 1986. Foraging Theory. Princeton University Press, Princeton, NJ. 347 Stern, B. 1934. The Lummi Indians of Northwest Washington. Columbia University Press, New York. Stuiver, M., Reimer, P.J. and Reimer, R.W. 2010. CALIB 6.0 [WWW program and documentation]. Sugimura, W., D. Sprugel, L. Brubaker, and P. Higuera. 2008. Millennial-scale changes in local vegetation and fire regimes on Mount Constitution, Orcas Island, Washington, USA, using small hollow sediments. Canadian Journal of Forest Research 38:539-552. Suttles, W. 1990. Central Coast Salish. Pages 199-208 in W. Suttles, editor. Handbook of North American Indians. Smithsonian Institution, Washington, D.C. Suttles, W. 1987. Coast Salish Essays. University of Washington Press, Seattle. Suttles, W. 1963. The Persistence of Intervillage Ties Among the Coast Salish. Ethnology 2:512- 524. Suttles, W. 1960a. Affinal Ties, Subsistence, and Prestige Among the Coast Salish. American Anthropologist 62:296-305. Suttles, W. 1960b. Variation in Habitat and Culture on the Northwest Coast. Verlag, Vienna. Suttles, W. 1951. The Economic Life of the Coast Salish of Haro and Rosario Straits. Unpublished Ph.D. dissertation. Department of Anthropology, University of Washington Seattle, Washington. Suttles, W. 1949. Field notebooks 4 to 7. Wayne Suttles Papers, 1946-86. Pacific Northwest Special Collections, University of Washington, Seattle. Tankersley, K. B. 1990. Late Pleistocene Lithic Exploitation in the Midwest and Midsouth: Indiana, Ohio, and Kentucky. Pages 259-299 in Tankersley, K.B. and Isaac, B.L., editors. K. B. Early Paleoindian Economies of Eastern North America. JAI Press, Greenwich, Connecticut. Ta?kiran A. N. 2001. Flaked-Stone Technology on the Channel Islands, CA. Unpublished Ph.D. dissertation. Department of Anthropology, University of California, Riverside, Riverside, CA. Taylor, A. K. and J. K. Stein. 2007. The San Juan Islands Archaeological Project, 2006. Report on file at the Department of Archaeology and Historic Preservation, Olympia, WA. Taylor, A. K. and J. K. Stein. 2006. A Report on Archaeological Investigations on Shaw Island and San Juan Island, Washington. Report on file at the Department of Archaeology and Historic Preservation, Olympia, WA. Taylor, A. K., J. K. Stein, and S. A. E. Jolivette. 2011. Big Sites, Small Sites, and Coastal Settlement Patterns in the San Juan Islands, Washington, USA. Journal of Island and Coastal Archaeology 6:287-313. Taylor, A. K., S. A. E. Jolivette, and J. K. Stein. 2009a. The San Juan Islands Archaeological Project, 2008. Report on file at the Department of Archaeology and Historic Preservation, Olympia, WA 348 Taylor, A. K., Marty, Johanna., Poister, Nicholas, and Vellanoweth, Rene L. 2009b. Anatomy of a Mound: Site Formation Processes at Tule Creek Village Mound B. Paper Presented the Society for California Archaeology Meetings, Modesto, CA. Tegner, M. J., Haaker, Pete L., Riser, Kristin L. and Vilchis, L. Ignacio. 2001. Climate Variability, Kelp Forests, and the Southern California Red Abalone Fishery. Journal of Shellfish Research 20:755-763. Teltser, P. A. 1991. Generalize Core Technology and Tool Use: A Mississippian Example. Journal of Field Archaeology 18:363-375. Thom, B. D. 2005. Coast Salish Senses of Place: Dwelling, Meaning, Power, Property and Territory in the Coast Salish World. McGill University, Montreal. Thomas, L. 1995. Archaeobotanical Research on San Nicolas Island: Current Directions. Pacific Coast Archaeological Society Quarterly 31:23-32. Thompson, M. and M. F. Price 2002. In praise of clumsiness: understand man and nature as a single but complex system. IHDP [International Human Dimensions Program] Newsletter 14? 16. Available at http://www.ihdp.uni-bonn.de/html/publications/ update/IHDPUpdate02_01.html. Thorson, R. M. 1980. Ice-sheet glaciation of the Puget Lowland, Washington, during the Vashon stade (Late Pleistocene). Quaternary Research 13:303-321. Torrence, R. 1986. Production and Exchange of Stone Tools: Prehistoric Obsidian in the Aegean. Cambridge University Press, Cambridge. Toset, H. P. W., N. P. Gleditsch, and H.Hegre. 2000. Shared rivers and interstate conflict. Political Geography Political Geography 19(8): 971-996. Tringham, R. 1972. Territorial demarcation of prehistoric settlements.in P. J. Ucko, R. Tringham, and G.W. Dimbleby editors. Man, Settlement and Urbanism. Schenkman Publishing Company, Cambridge, MA. Tunnicliffe, V., O'Connell, J.M. and McQuoid, M.R. 2001. A Holocene Record of Marine Fish Remains from Northeastern Pacific. Marine Geology 174:197-210. Vasquez, J. A. 1993. The War Puzzle. Cambridge University Press, New York. Vedder, J. G. and R. M. Norris. 1963. Geology of San Nicolas Island, California. Geological Survey Professional Paper 369. Prepared in cooperation with the U.S. Dept. of the Navy, Office of Naval Petroleum and Oil Share Reserves. U.S. Government Printing Office, Washington, D.C. Vellanoweth, R. L. 2001a. AMS radiocarbon dating and shell bead chronologies: Middle Holocene exchange and interaction in western North America. Journal of Archaeological Science 28:941-950. Vellanoweth, R.L. 2001b. Coastal Archaeology of Southern California: Accounts from the Holocene. Unpublished Ph.D. Dissertation, University of Oregon, Eugene, OR. 349 Vellanoweth, R.L. 1998. Earliest Island Fox Remains on the Southern Channel Islands: Evidence from San Nicolas Island, California. Journal of California and Great Basin Anthropology 20(1):100-108. Vellanoweth, R.L., Martz, P.C., and Schwartz, S.J. 2002. Complexity and the Late Holocene Archaeology of San Nicolas Island, California. Pages 91-99 in J.M. Erlandson and T.L. Jones, editors. Catalyst to Complexity: The Late Holocene Archaeology of the California Coast. UCLA Cotsen Institute of Archaeology, Los Angeles, CA. Waitt, J., RB. 1977. Evolution of glaciated topography of upper Skagit drainage basin, Washington. Arctic and Alpine Research 9:183-192. Walker, I. R., and Pellatt, M.G. 2003. Climate Change in Coastal British Columbia. Canadian Water Resources Journal 28:531-560. Wallace, I. J. and J. J. Shea. 2006. Mobility patterns and core technologies in the middle Paleolithic of the Levant. Journal of Archaeological Science 33:1293-1309. Walsh, M. R. 1998. Lines in the sand: competition and stone selection on the Pajarito Plateau, New Mexico. American Antiquity 63:573-593. Wentworth, C. K. 1922. The shapes of beach pebbles. Professional Papers of the U.S. Geological Survey 131C:75-83. Wessen, G. C. 1986. Prehistoric Cultural Resources of San Juan Couny, Washington. A Report Prepared for the Washington State Office of Archaeology and Historic Preservation. Wessen and Associates, Kirkland, WA. Wessen, G. C. 1993. An Overview of Archaeological Activities Conducted by Western Heritage, Inc. in the Lake Cushman Project Area, 1988-1991. Report prepared for Tacoma Public Utilities. Wessen and Associates, Seattle, Tacoma, Washington. Whallon, R. 1989. Elements of Cultural Change in the Later Paleolithic. Pages 433-454 in P. Mellars and C. Stringer, editors. The Human Revolution, Behavioral and Biological Perspectives on the Origins of Modern Humans. Princeton University Press, Princeton. Whallon, R. 2006. Social networks and information: Non-"utilitarian" mobility among hunter- gatherers. Journal of Anthropological Archaeology 25:259-270. Whittaker, F. H. and J. K. Stein. 1992. Shell Midden Boundaries in Relation to Past and Present Shorelines. Pages 25-42 in J. K. Stein, editor. Deciphering a Shell Midden. Academic Press, Inc., San Diego. Wiessner, P. 1977. Hxaro: A Regional System of Reciprocity for Reducing Risk Among the !Kung San. Unpublished Ph.D. dissertion. Department of Anthropology, University of Michigan, Ann Arbor. Wiessner, P. 1982. Risk, reciprocity and social influences on !Kung San economics. Pages 61-84 in E. Leacock and R. Lee, editors. Politics and History in Band societies. Cambridge University Press, Cambridge. 350 Wilson, L. 2011. Raw material economics in their environmental context: an example from the Middle Palaeolithic of southern France. Geological Society, London, Special Publication 352:163-180. Wilson, L. 2007a. Terrain difficulty as a factor in raw material procurement in the Middle Palaeolithic of France. Journal of Field Archaeology 32:315-324. Wilson, L. 2007b. Understanding Prehistoric Lithic Raw Material Selection: Application of a Gravity Model. Journal of Archaeological Method and Theory 14:388-411. Wilson, M. C., S. M. Kenady, and R. F. Schalk. 2009. Late Pleistocene Bison antiquus from Orcas Island, Washington, and the biogeographic importance of an early postglacial land mammal dispersal corridor from the mainland to Vancouver Island. Quaternary Research 71:49- 61. Winterhalder, B. and E.A. Smith. 2000. Analyzing adaptative strategies: human behavioral ecology at twenty-five. Evolutionary Anthropology 9:51-72. Winterhalder, B., D. J. Kennett, M. N. Grote, and J. Bartruff. 2010. Ideal free settlement of California's Northern Channel Islands. Journal of Anthropological Archaeology 29:469-490. Winterhalder, B., Lu, Flora, and Tucker, Bram. 1999. Risk-sensitive adaptive tactics: Models and evidence from subistence studies in biology and anthropology. Journal of Archaeological Research 7:301-348. Wixom, T. and Snow, B. 2004. San Juan County Water Resources Management Plan WRIA 2. San Juan County Board of County Commissioners, Friday Harbor, WA. Wright, C. A., Dallimore, A., Thomson, R.E., Patterson, R.T., and Ware, D.M. 2005. Late Holocene Paleofish Populations in Effingham Inlet, British Columbia, Canada. Paleogeography, Palaeoclimatology, Paleoecology 224:367-384. Yatsko, A. and Raab, M.L. 2009. San Clemente Island. Pages 23-39 in M. L. Raab, Cassidy, Jim, Yatsko, Andrew, and Howard, William J., editor. California Maritime Archaeology. Altamira Press, New York. Yesner, D. R. 1980. Maritime Hunter-Gatherers: Ecology and Prehistory. Current Anthropology 21(6):727-750. 351 Appendix A. Defensive characteristics for sites on the San Juan Islands, Washington. Site 45SJ# Landform Size Time range occupied (cal BP) Degrees visibility Degrees visibility/ degrees approach Distance to lookout (m) (vis. >180 degrees) Elev. difference (site interior to exterior) (m) Site radius (m) arctan radians Elevation (arctan degrees/90) 1.00 Sloping bluff Big 2500-1000 70 0.19 300 12 50 0.24 0.15 2.00 Lagoon and spit Big 2500-1600 153 0.43 700 2 40 0.05 0.03 3A Lagoon and spit Small 2500-1600 170 0.65 500 2 10 0.20 0.13 3C Lagoon and spit Small 600-0 180 0.72 300 1 5 0.20 0.13 6.00 Lagoon and spit Small 600-0 200 0.63 200 3 10 0.29 0.19 9.00 Beach Big 600-0 130 0.36 400 5 20 0.24 0.16 23.00 Beach Small 1600-1000 80 0.22 200 2 5 0.38 0.24 24A Beach Big 1600-0 130 0.36 200 5 10 0.46 0.30 24D Inland area Big 2500-1000 180 0.50 300 12 30 0.38 0.24 26.00 Beach Small 600-0 130 0.36 300 2 2 0.79 0.50 27.00 Beach Small 600-0 160 0.44 600 3 5 0.54 0.34 47.00 Beach Small 600-0 70 0.19 150 1 10 0.10 0.06 60.00 Beach Small 600-0 60 0.17 250 3 2 0.98 0.63 61.00 Bluff over beach Small 600-0 90 0.25 250 6 15 0.38 0.24 70.00 Bluff over beach Small 2500-1600, 600-0 200 0.56 0 6 40 0.15 0.09 71.00 Beach Small 600-0 200 0.56 0 1 10 0.10 0.06 72.00 Beach Small 600-0 220 0.61 0 2 5 0.38 0.24 89.00 Beach Small 600-0 180 0.50 3000 2 10 0.20 0.13 95.00 Beach Small 600-0 160 0.44 800 5 10 0.46 0.30 105.00 Isthmus Big 2500-1600, 600-0 150 0.42 400 4 10 0.38 0.24 120.00 Beach Small 600-0 220 0.61 0 6 25 0.24 0.15 124.00 Tombolo Small 600-0 200 1.00 0 2 25 0.08 0.05 147.00 Bedrock Point Small 600-0 220 0.61 0 2 15 0.13 0.08 150.00 Beach Small 600-0 120 0.33 500 3 10 0.29 0.19 165.00 Beach Big 3500-2500, 1600-1000 160 0.44 1000 12 30 0.38 0.24 169.00 Tombolo and lagoon Big 2500-1600, 600-0 180 0.50 600 10 40 0.24 0.16 201.00 Beach Small 600-0 240 0.67 0 3 5 0.54 0.34 202.00 Beach Small 600-0 270 0.75 0 6 50 0.12 0.08 225.00 Point Small 600-0 170 0.47 450 6 20 0.29 0.19 352 Appendix A, continued. Defensive characteristics for sites on the San Juan Islands, Washington. Site 45SJ# Landform Size Time range occupied (cal BP) Degrees visibility Degrees visibility/ degrees approach Distance to lookout (m) (vis. >180 degrees) Elev. difference (site interior to exterior) (m) Site radius (m) arctan radians Elevation (arctan degrees/90) 239.00 Beach Small 1600-1000 130 0.36 200 3 5 0.54 0.34 251.00 Beach Small 600-0 240 0.63 0 6 10 0.54 0.34 254.00 Beach Big 1600-0 180 0.50 700 6 100 0.06 0.04 274.00 Spit Big 1600-0 150 0.57 600 3 10 0.29 0.19 277.00 Beach Small 600-0 170 0.47 500 2 10 0.20 0.13 278.00 Beach Big 3500-0 180 0.50 800 3 10 0.29 0.19 279.00 Beach Big 1600-0 180 0.50 1200 3 20 0.15 0.09 280.00 Beach Big 3500-0 50 0.19 600 2 10 0.20 0.13 282.00 Point Small 1600-0 260 0.72 0 6 40 0.15 0.09 307.00 Beach Small 600-0 150 0.42 300 12 30 0.38 0.24 324.00 Beach Small 600-0 50 0.14 150 5 5 0.79 0.50 364.00 Beach Small 600-0 130 0.36 1800 2 10 0.20 0.13 407.00 Lagoon and spit Small 2500-1600 130 0.36 650 5 3 1.03 0.66 450.00 Tombolo Small 1600-0 100 0.67 100 3 10 0.29 0.19 451.00 Beach Small 600-0 80 0.22 100 2 5 0.38 0.24 453.00 Bluff over beach Small 600-0 180 0.50 600 6 5 0.88 0.56 460.00 Beach Small 600-0 120 0.33 50 5 5 0.79 0.50 461.00 Beach Small 600-0 90 0.25 100 3 10 0.29 0.19 481.00 Beach Small 600-0 70 0.24 60 3 10 0.29 0.19 483.00 Beach/ inland area Big 600-0 170 0.47 150 12 50 0.24 0.15 507.00 Beach Small 600-0 60 0.17 250 3 5 0.54 0.34 509.00 Beach Small 600-0 100 0.28 200 2 5 0.38 0.24 353 Appendix B. Defensive characteristics for sites on San Nicolas Island, California. Site CASNI# Landform Type Elev. Change Distance to water Lookout (> 180 degrees over water) (m) Time range (cal BP) Elevation difference (interior to exterior of site) (m) Site radius (m) arctan (radians) Elevation (arctan degrees/90) 5 Coastal Plain Residential 50 0.5 1500-500 50 100 0.46 0.30 6 Coastal Plain Camp 5 0.6 1500-500 5 50 0.10 0.06 8 Coastal Plain Residential 25 0.1 5000-3000 25 100 0.24 0.16 11 Coastal Plain Residential 50 0 5000-3000, 3000-500, 500-0 50 100 0.46 0.30 14 Coastal Plain Residential 25 0.9 5000-3000 25 100 0.24 0.16 15 Coastal Plain Residential 50 0.3 >5000 50 50 0.79 0.50 16 Coastal Plain Residential 75 0.3 5000-3000 75 250 0.29 0.19 18 Coastal Plain Residential 75 0.9 500-0 75 150 0.46 0.30 21 Slope Residential 5 1.8 5000-3000, 1500-500 5 50 0.10 0.06 25 Plateau Residential 75 1.9 5000-3000, 3000-500, 500-0 75 150 0.46 0.30 38 Coastal Plain Residential 5 0.7 1500-500 5 25 0.20 0.13 39 Coastal Plain Residential 50 1 3000-0 50 150 0.32 0.20 40 Slope Residential 25 1.5 5000-3000 25 50 0.46 0.30 41 Coastal Plain Residential 50 0.5 >5000 50 75 0.59 0.37 43 Slope Residential 25 3.5 5000-3000, 3000-500, 500-0 25 100 0.24 0.16 51 Slope Residential 125 0.5 3000-1500 125 175 0.62 0.39 56 Coastal Plain Residential 75 2.3 5000-3000 75 200 0.36 0.23 72 Coastal Plain Camp 5 0.5 3000-1500 5 50 0.10 0.06 73 Coastal Plain Residential 5 0.4 1500-500 5 75 0.07 0.04 74 Coastal Plain Residential 5 0.15 3000-1500 5 50 0.10 0.06 76 Coastal Plain Fishing, small 5 0.7 500-0 5 25 0.20 0.13 79 Plateau Residential 5 2 1500-500 5 50 0.10 0.06 84 Plateau Stone artifact manufacture/s hellfish processing 5 3 1500-500 5 12.5 0.38 0.24 102 Plateau Camp 5 2.1 3000-1500 5 50 0.10 0.06 105 Plateau Camp 25 2.5 5000-3000, 3000-1500 25 150 0.17 0.11 106 Plateau Residential 25 2.8 3000-1500 25 100 0.24 0.16 117 Slope Shellfish processing 75 0.7 3000-1500 75 100 0.64 0.41 129 Coastal Plain Multi-use 25 0.6 1500-500 25 50 0.46 0.30 130 Coastal Plain Camp 5 0.9 3000-1500 5 50 0.10 0.06 131 Coastal Plain Camp 5 0.4 5000-3000, 1500-500 5 50 0.10 0.06 354 Appendix B, continued. Defensive characteristics for sites on San Nicolas Island, California. Site CASNI# Landform Type Elev. Change Distance to water Lookout (> 180 degrees over water) (m) Time range (cal BP) Elevation difference (interior to exterior of site) (m) Site radius (m) arctan (radians) Elevation (arctan degrees/90) 147 Coastal Plain Fishing camp 5 1 3000-1500 5 25 0.20 0.13 157 Slope Residential 25 1.3 5000-3000 25 150 0.17 0.11 160 Coastal Plain Residential 50 1.4 1500-500 50 250 0.20 0.13 161 Coastal Plain Residential 25 0.95 5000-3000, 3000-1500 25 50 0.46 0.30 162 Slope Small shell midden 5 0.5 1500-500 5 50 0.10 0.06 163 Coastal Plain Fishing camp 5 0.6 1500-500 5 50 0.10 0.06 164 Slope Shell midden 25 0.9 5000-3000 25 50 0.46 0.30 165 Slope Residential 25 0.7 5000-3000 25 25 0.79 0.50 168 Slope Residential 50 0.7 5000-3000, 3000-500, 500-0 50 200 0.24 0.16 169 Slope Camp 5 0.9 5000-3000, 3000-1500 5 12.5 0.38 0.24 170 Slope Stone artifact manufacture and shellfish processing 5 0.9 5000-3000 5 50 0.10 0.06 171 Slope Residential 5 1.3 5000-3000, 3000-1500 5 50 0.10 0.06 184 Coastal Plain Residential 5 0.5 500-0 5 50 0.10 0.06 204 Slope Camp 75 1.8 3000-1500 75 150 0.46 0.30 214 Plateau Residential 5 2.2 500-0 5 500 0.01 0.01 238 Coastal Plain Fishing camp 5 1.6 3000-1500 5 25 0.20 0.13 284 Slope Multi-use 5 2.1 5000-3000 5 25 0.20 0.13 290 Plateau Multi-use 5 2.1 1500-500 5 12.5 0.38 0.24 315 Slope Shellfish processing 5 0.5 500-0 5 12.5 0.38 0.24 316 Plateau Multi-use processing 5 1.1 5000-3000 5 25 0.20 0.13 323 Coastal Plain Residential 25 0.4 500-0 25 25 0.79 0.50 328 Coastal Plain Residential 5 0.3 1500-500 5 6.25 0.67 0.43 329 Coastal Plain Multi-use 5 0.4 500-0 5 25 0.20 0.13 339 Slope Residential 25 0.5 >5000, 1500-500, 500-0 25 50 0.46 0.30 340 Coastal Plain Camp 5 0.4 500-0 5 50 0.10 0.06 342 Coastal Plain Multi-use 5 0.6 1500-500 5 50 0.10 0.06 346 Plateau Shellfish processing 25 1.8 1500-0 25 25 0.79 0.50 347 Plateau Shellfish processing 5 1.7 5000-3000 5 12.5 0.38 0.24 351 Plateau Residential 5 2.6 >5000, 3000-0 5 150 0.03 0.02 361 Plateau Multi-use 5 0.8 3000-1500 5 25 0.20 0.13 355 Appendix C. Zirconium and strontium concentrations (ppm) for archaeological and geological samples of fine-grained volcanic rock from the San Juan Islands, WA. Site No. Catalog No. Specimen Type Zr (ppm) + Sr (ppm) + Source Watts Point Quarry 50 1621-1 Geological 121 8 773 12 Watts Point Watts Point Quarry 51 1621-2 Geological 121 8 753 12 Watts Point Watts Point Quarry 52 1621-3 Geological 120 8 719 12 Watts Point Watts Point Quarry 53 1621-4 Geological 123 8 753 11 Watts Point Porteau Cove 32 1616-1 Geological 149 8 824 12 Watts Point Porteau Cove 33 1616-2 Geological 152 8 855 12 Watts Point Porteau Cove 34 1616-3 Geological 153 8 1030 12 Unknown Lighthouse Beach 35 1617-1 Geological 140 8 834 12 Watts Point Lighthouse Beach 36 1617-2 Geological 148 8 918 12 Watts Point Lighthouse Beach 38 1617-3 Geological 144 8 786 12 Watts Point Horseshoe Bay Ferry Beach 39 1618-1 Geological 126 8 745 12 Watts Point Horseshoe Bay Ferry Beach 40 1618-2 Geological 126 8 787 11 Watts Point Whytecliffe Park 41 1619-1 Geological 129 8 767 12 Watts Point Whytecliffe Park 42 1619-2 Geological 121 8 754 11 Watts Point Whytecliffe Park 43 1619-3 Geological 125 8 758 12 Watts Point Murrin Park 44 1620-1 Geological 140 8 860 12 Watts Point Murrin Park 45 1620-2 Geological 129 8 795 12 Watts Point Murrin Park 46 1620-3 Geological 126 8 728 12 Watts Point Murrin Park 47 1620-4 Geological 130 8 819 12 Watts Point Murrin Park 48 1620-5 Geological 116 8 708 12 Watts Point Murrin Park 49 1620-6 Geological 123 8 752 12 Watts Point Boulder Creek 55 1623-1 Geological 50 8 424 11 Unknown Boulder Creek 56 1623-2 Geological 58 8 1050 12 Unknown Watmough Bay 23 1726-1 Geological 127 7 621 9 Watts Point Watmough Bay 24 1726-2 Geological 137 7 668 9 Watts Point Watmough Bay 25 1726-3 Geological 124 7 698 10 Watts Point Watmough Bay 26 1726-4 Geological 128 7 660 10 Watts Point Watmough Bay 27 1726-5 Geological 127 7 681 10 Watts Point Watmough Bay 28 1726-6 Geological 132 7 732 9 Watts Point Watmough Bay 29 1726-7 Geological 126 7 681 10 Watts Point Watmough Bay 30 1726-8 Geological 157 7 426 9 Unknown Watmough Bay 31 1726-9 Geological 126 7 656 9 Watts Point Watmough Bay 32 1726-10 Geological 133 7 742 10 Watts Point Watmough Bay 33 1727-1 Geological 135 7 701 9 Watts Point Agate Beach 34 1727-2 Geological 125 7 663 10 Watts Point Agate Beach 35 1727-3 Geological 128 7 737 10 Watts Point Agate Beach 36 1727-4 Geological 129 7 664 9 Watts Point Agate Beach 37 1727-5 Geological 121 7 751 10 Watts Point Agate B ch 38 1727-6 Geological 128 7 681 10 Watts Point Agate Beach 39 1727-7 Geological 122 7 698 10 Watts Point Agate Beach 40 1727-8 Geological 123 7 639 9 Watts Point Agate Beach 41 1727-9 Geological 125 7 666 9 Watts Point Agate Beach 42 1727-10 Geological 124 7 778 10 Watts Point Agate Beach 43 1727-11 Geological 126 7 711 10 Watts Point Agate Beach 44 1727-12 Geological 126 7 746 10 Watts Point 356 Appendix C, continued. Zirconium and strontium concentrations (ppm) for archaeological and geological samples of fine-grained volcanic rock from the San Juan Islands, WA. Site No. Catalog No. Specimen Type Zr (ppm) + Sr (ppm) + Source Agate Beach 45 1727-13 Geological 138 7 730 10 Watts Point Agate Beach 46 1727-14 Geological 162 7 766 10 Watts Point Agate Beach 47 1727-15 Geological 135 7 723 10 Watts Point Agate Beach 48 1727-16 Geological 132 7 688 10 Watts Point Agate Beach 49 1727-17 Geological 126 7 648 9 Watts Point Agate Beach 50 1727-18 Geological 129 7 676 9 Watts Point Agate Beach 51 1727-19 Geological 121 7 695 10 Watts Point Agate Beach 52 1727-20 Geological 131 7 645 9 Watts Point Agate Beach 53 1727-21 Geological 152 7 60 9 Unknown Agate Beach 54 1727-22 Geological 124 7 660 9 Watts Point Agate Beach 55 1727-23 Geological 124 7 705 10 Watts Point Agate Beach 56 1727-24 Geological 125 7 666 9 Watts Point Agate Beach 57 1727-25 Geological 131 7 685 9 Watts Point Agate Beach 58 1727-26 Geological 97 7 626 9 Unknown Agate Beach 59 1727-27 Geological 111 7 874 10 Unknown Agate Beach 60 1727-28 Geological 125 7 645 9 Watts Point Agate Beach 61 1727-29 Geological 125 7 693 9 Watts Point Agate Beach 62 1727-30 Geological 121 7 629 9 Watts Point False Bay 13 1725-1 Geological 108 7 149 9 Unknown False Bay 14 1725-2 Geological 123 7 357 9 Unknown False Bay 15 1725-3 Geological 142 7 854 10 Watts Point False Bay 16 1725-4 Geological 138 7 683 9 Watts Point False Bay 17 1725-5 Geological 123 7 645 9 Watts Point False Bay 18 1725-6 Geological 136 7 702 9 Watts Point False Bay 19 1725-7 Geological 135 7 660 9 Watts Point False Bay 20 1725-8 Geological 135 7 644 10 Watts Point False Bay 21 1725-9 Geological 138 7 767 10 Watts Point False Bay 22 1725-10 Geological 135 7 687 10 Watts Point False Bay 31 FalseBay.s.18 Geological 117 7 222 9 Unknown False Bay 32 FalseBay.s.19 Geological 141 7 705 9 Watts Point False Bay 33 FalseBay.s.20 Geological 136 7 684 9 Watts Point False Bay 34 FalseBay.1.20-30.1 Geological 92 7 655 9 Unknown False Bay 35 FalseBay.1.20-30.2 Geological 43 7 38 9 Unknown False Bay 36 FalseBay.1.30-85.1 Geological 109 7 359 9 Unknown False Bay 37 FalseBay.1.85-140.1 Geological 87 7 60 9 Unknown False Bay 38 FalseBay.2.0-20.1 Geological 95 7 410 9 Unknown False Bay 39 FalseBay.2.0-20.2 Geological 70 7 92 9 Unknown False Bay 40 FalseBay.2.30-50.1 Geological 117 7 355 9 Unknown False Bay 16 FalseBay.s.3 Geological 130 7 702 9 Watts Point False Bay 17 FalseBay.s.4 Geological 124 7 695 10 Watts Point False Bay 18 FalseBay.s.5 Geological 136 7 721 9 Watts Point False Bay 19 FalseBay.s.6 Geological 131 7 682 9 Watts Point False Bay 20 FalseBay.s.7 Geological 133 7 692 9 Watts Point False Bay 21 FalseBay.s.8 Geological 136 7 674 9 Watts Point False Bay 22 FalseBay.s.9 Geological 134 7 702 9 Watts Point 357 Appendix C, continued. Zirconium and strontium concentrations (ppm) for archaeological and geological samples of fine-grained volcanic rock from the San Juan Islands, WA. Site No. Catalog No. Specimen Type Zr (ppm) + Sr (ppm) + Source False Bay 23 FalseBay.s.10 Geological 130 7 650 9 Watts Point False Bay 24 FalseBay.s.11 Geological 119 7 680 10 Watts Point False Bay 25 FalseBay.s.12 Geological 131 7 660 9 Watts Point False Bay 26 FalseBay.s.13 Geological 129 7 664 9 Watts Point False Bay 27 FalseBay.s.14 Geological 124 7 677 9 Watts Point False Bay 28 FalseBay.s.15 Geological 122 7 644 9 Watts Point False Bay 29 FalseBay.s.16 Geological 129 7 644 10 Watts Point False Bay 30 FalseBay.s.17 Geological 123 7 681 10 Watts Point False Bay 14 FalseBay.s.1 Geological 118 7 623 10 Watts Point False Bay 15 FalseBay.s.2 Geological 127 7 750 10 Watts Point Blind Bay 1 Deane.1.20-35.1 Geological 130 7 753 9 Watts Point Blind Bay 2 Deane.1.35-50.1 Geological 71 7 150 9 Unknown Blind Bay 3 Deane.1.50-60.1 Geological 43 7 37 9 Unknown Blind Bay 4 Deane.1.50-60.2 Geological 124 7 165 9 Unknown Blind Bay 5 Deane.1.50-60.3 Geological 66 7 130 9 Unknown Blind Bay 6 Deane.2.0-20.1 Geological 133 7 209 9 Unknown Blind Bay 7 Deane.2.60-70.1 Geological 112 7 240 9 Unknown Blind Bay 8 Deane.2.60-70.2 Geological 105 7 229 9 Unknown Blind Bay 9 Deane.2.60-70.3 Geological 178 7 162 9 Unknown Deadman's Cove 10 Deadman.s.1 Geological 128 7 750 10 Watts Point Deadman's Cove 11 Deadman.s.2 Geological 140 7 899 10 Unknown Deadman's Cove 12 Deadman.s.3 Geological 113 7 1242 10 Unknown Schoen Beach 57 1624-1 Geological 126 8 674 11 Watts Point Schoen Beach 58 1624-2 Geological 128 8 768 11 Watts Point Double Bluff 59 1625-1 Geological 213 8 522 11 Unknown Quartermaster Harbor 54 1622-1 Geological 149 8 259 11 Unknown Dungeness Spit 13 Dungeness.s.1 Geological 49 7 270 9 Unknown English Camp (45-SJ-124) 1 SAJH 4231 Artifact 135 7 687 10 Watts Point English Camp (45-SJ-124) 2 SAJH 42951 Artifact 136 7 664 9 Watts Point English Camp (45-SJ-124) 3 SAJH 43001 Artifact 134 7 663 9 Watts Point English Camp (45-SJ-124) 4 SAJH 43103 Artifact 132 7 657 9 Watts Point English Camp (45-SJ-124) 5 SAJH 43126 Artifact 135 7 654 9 Watts Point English Camp (45-SJ-124) 6 SAJH 43138 Artifact 125 7 648 9 Watts Point English Camp (45-SJ-124) 7 SAJH 104523 Artifact 126 7 711 9 Watts Point English Camp (45-SJ-124) 8 SAJH 104654 Artifact 123 7 618 9 Watts Point English Camp (45-SJ-124) 9 SAJH 104708 Artifact 128 7 659 9 Watts Point English Camp (45-SJ-124) 10 SAJH 104709 Artifact 129 7 655 9 Watts Point English Camp (45-SJ-124) 11 SAJH 104776 Artifact 127 7 663 9 Watts Point English Ca p (45-SJ-124) 12 SAJH 104782 Artifact 129 7 779 10 Watts Point English Camp (45-SJ-124) 13 SAJH 104783 Artifact 128 7 692 10 Watts Point English Camp (45-SJ-124) 14 SAJH 104805 Artifact 116 7 677 10 Watts Point English Camp (45-SJ-124) 15 SAJH 104812 Artifact 123 7 648 9 Watts Point English Camp (45-SJ-124) 16 SAJH 104869 Artifact 109 7 1138 10 Unknown English Camp (45-SJ-124) 17 SAJH 104870 Artifact 143 7 946 10 Unknown English Camp (45-SJ-124) 18 SAJH 104871 Artifact 132 7 693 9 Watts Point 358 Appendix C, continued. Zirconium and strontium concentrations (ppm) for archaeological and geological samples of fine-grained volcanic rock from the San Juan Islands, WA. Site No. Catalog No. Specimen Type Zr (ppm) + Sr (ppm) + Source English Camp (45-SJ-124) 19 SAJH 104895 Artifact 128 7 707 9 Watts Point English Camp (45-SJ-124) 20 SAJH 104905 Artifact 130 7 671 9 Watts Point English Camp (45-SJ-124) 21 SAJH 104906 Artifact 130 7 790 10 Watts Point English Camp (45-SJ-124) 22 SAJH 104909 Artifact 136 7 727 9 Watts Point English Camp (45-SJ-124) 23 SAJH 104910 Artifact 133 7 664 10 Watts Point English Camp (45-SJ-124) 24 SAJH 104938 Artifact 129 7 677 9 Watts Point English Camp (45-SJ-124) 25 SAJH 104975 Artifact 124 7 659 9 Watts Point English Camp (45-SJ-124) 26 SAJH 125816 Artifact 125 7 771 9 Watts Point English Camp (45-SJ-124) 27 SAJH 125901 Artifact 87 7 413 9 Unknown English Camp (45-SJ-124) 28 SAJH 125902a Artifact 131 7 692 9 Watts Point English Camp (45-SJ-124) 29 SAJH 125902b Artifact 128 7 672 9 Watts Point English Camp (45-SJ-124) 30 SAJH 125928 Artifact 123 7 641 9 Watts Point Deane Site (45-SJ-150) 1 3 Artifact 134 7 694 9 Watts Point Deane Site (45-SJ-150) 2 16 Artifact 127 7 267 9 Unknown Deane Site (45-SJ-150) 3 19 Artifact 127 7 746 10 Watts Point Deane Site (45-SJ-150) 4 34 Artifact 122 7 673 10 Watts Point Deane Site (45-SJ-150) 5 45 Artifact 121 7 662 10 Watts Point Deane Site (45-SJ-150) 6 51 Artifact 124 7 704 9 Watts Point Deane Site (45-SJ-150) 7 69 Artifact 127 7 699 10 Watts Point Deane Site (45-SJ-150) 8 75 Artifact 125 7 686 9 Watts Point Deane Site (45-SJ-150) 9 77 Artifact 137 7 749 9 Watts Point Deane Site (45-SJ-150) 10 86 Artifact 130 7 658 9 Watts Point Deane Site (45-SJ-150) 11 101 Artifact 127 7 655 9 Watts Point Deane Site (45-SJ-150) 12 120 Artifact 127 7 655 9 Watts Point Deane Site (45-SJ-150) 13 136 Artifact 129 7 677 9 Watts Point Deane Site (45-SJ-150) 51 SJ150.32 Artifact 144 7 724 10 Watts Point Deane Site (45-SJ-150) 52 SJ150.27 Artifact 129 7 667 9 Watts Point Deane Site (45-SJ-150) 53 SJ150.33 Artifact 129 7 649 9 Watts Point Deane Site (45-SJ-150) 54 SJ150.134 Artifact 134 7 707 9 Watts Point Deane Site (45-SJ-150) 55 SJ150.105 Artifact 131 7 694 9 Watts Point Deane Site (45-SJ-150) 56 SJ150.105 Artifact 127 7 722 10 Watts Point Deane Site (45-SJ-150) 57 SJ150.92 Artifact 133 7 711 10 Watts Point Watmough Bay (45-SJ-280) 1 1 Artifact 150 8 742 11 Watts Point Watmough Bay (45-SJ-280) 2 2 Artifact 145 8 671 12 Watts Point Watmough Bay (45-SJ-280) 3 3 Artifact 151 8 836 12 Watts Point Watmough Bay (45-SJ-280) 4 4 Artifact 144 8 733 11 Watts Point Watmough Bay (45-SJ-280) 5 5 Artifact 144 8 722 11 Watts Point Watmough Bay (45-SJ-280) 6 6 Artifact 141 8 742 11 Watts Point Watmough Bay (45-SJ-280) 7 7 Artifact 138 8 679 11 Watts Point Watmough Bay (45-SJ-280) 8 8 Artifact 140 8 696 11 Watts Point Watmough Bay (45-SJ-280) 9 9 Artifact 118 8 731 11 Watts Point Watmough Bay (45-SJ-280) 10 10 Artifact 132 8 803 11 Watts Point Watmough Bay (45-SJ-280) 11 11 Artifact 135 8 661 11 Watts Point Watmough Bay (45-SJ-280) 12 12 Artifact 143 8 744 11 Watts Point Watmough Bay (45-SJ-280) 13 13 Artifact 144 8 736 11 Watts Point 359 Appendix C, continued. Zirconium and strontium concentrations (ppm) for archaeological and geological samples of fine-grained volcanic rock from the San Juan Islands, WA. Site No. Catalog No. Specimen Type Zr (ppm) + Sr (ppm) + Source Watmough Bay (45-SJ-280) 42 LB88/3 Artifact 122 7 665 9 Watts Point Watmough Bay (45-SJ-280) 43 LB88/4 Artifact 114 7 629 9 Watts Point Watmough Bay (45-SJ-280) 44 LB88/5 Artifact 123 7 655 9 Watts Point Watmough Bay (45-SJ-280) 45 LB88/6 Artifact 125 7 667 9 Watts Point Watmough Bay (45-SJ-280) 46 LB88/7 Artifact 73 7 183 9 Unknown Watmough Bay (45-SJ-280) 47 LB88/8 Artifact 123 7 673 10 Watts Point Watmough Bay (45-SJ-280) 48 LB88/9 Artifact 123 7 720 10 Watts Point Watmough Bay (45-SJ-280) 49 LB88/10 Artifact 122 7 638 9 Watts Point Watmough Bay (45-SJ-280) 50 SJ150.80 Artifact 123 7 764 9 Watts Point Watmough Bay (45-SJ-280) 14 14 Artifact 143 8 725 11 Watts Point Watmough Bay (45-SJ-280) 15 15 Artifact 139 8 714 11 Watts Point Watmough Bay (45-SJ-280) 16 16 Artifact 129 8 763 11 Watts Point Watmough Bay (45-SJ-280) 17 17 Artifact 144 8 745 11 Watts Point Watmough Bay (45-SJ-280) 18 18 Artifact 135 8 672 12 Watts Point Watmough Bay (45-SJ-280) 19 19 Artifact 139 8 673 11 Watts Point Watmough Bay (45-SJ-280) 20 20 Artifact 266 8 176 11 Unknown Watmough Bay (45-SJ-280) 21 21 Artifact 149 8 733 11 Watts Point Watmough Bay (45-SJ-280) 22 22 Artifact 127 8 795 12 Watts Point Watmough Bay (45-SJ-280) 23 23 Artifact 125 8 719 11 Watts Point Watmough Bay (45-SJ-280) 24 24 Artifact 142 8 725 11 Watts Point Watmough Bay (45-SJ-280) 25 25 Artifact 136 8 692 11 Watts Point Watmough Bay (45-SJ-280) 26 26 Artifact 120 8 699 11 Watts Point Watmough Bay (45-SJ-280) 27 27 Artifact 138 8 697 11 Watts Point Watmough Bay (45-SJ-280) 28 28 Artifact 147 8 878 12 Watts Point Watmough Bay (45-SJ-280) 29 29 Artifact 137 8 740 11 Watts Point Watmough Bay (45-SJ-280) 30 30 Artifact 122 8 751 12 Watts Point Watmough Bay (45-SJ-280) 31 LB94/4 Artifact 123 7 642 9 Watts Point Watmough Bay (45-SJ-280) 32 LB94/5 Artifact 141 7 900 10 Unknown Watmough Bay (45-SJ-280) 33 LB94/6 Artifact 130 7 739 10 Watts Point Watmough Bay (45-SJ-280) 34 LB94/7 Artifact 130 7 684 9 Watts Point Watmough B y (45-SJ-280) 35 LB94/8 Artifact 127 7 678 9 Watts Point Watmough Bay (45-SJ-280) 36 LB94/9 Artifact 128 7 679 9 Watts Point Watmough Bay (45-SJ-280) 37 LB94/10 Artifact 121 7 696 10 Watts Point Watmough Bay (45-SJ-280) 38 LB92/2 Artifact 126 7 731 10 Watts Point Watmough Bay (45-SJ-280) 39 LB92/3 Artifact 116 7 755 10 Watts Point Watmough Bay (45-SJ-280) 40 LB92/4 Artifact 122 7 608 10 Watts Point Watmough Bay (45-SJ-280) 41 LB88/2 Artifact 120 7 668 9 Watts Point 360 VITA Amanda Taylor was born in New York City and raised in Croton-on-Hudson, New York. She earned a Bachelor of Arts degree in Archaeology and American Studies at Hamilton College in Clinton, New York in 2002. At the University of Washington, Seattle, Amanda earned a Master of Arts in Anthropology in 2006 and a Doctor of Philosophy in Anthropology in 2012. She has conducted archaeological fieldwork in the Great Basin, New York State, the Pacific Northwest, Alaska, and the California Channel Islands. She currently lives in Seattle, Washington and will begin working as a Visiting Assistant Professor of Anthropology at Pacific Lutheran University, Tacoma, in Autumn of 2012.