UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Inter -and intra-site heterogeneity as sources for faunal assemblage variability : an analysis of fish… Johnson, Raini Abigale 2019

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2019_september_johnson_raini.pdf [ 1.51MB ]
Metadata
JSON: 24-1.0380481.json
JSON-LD: 24-1.0380481-ld.json
RDF/XML (Pretty): 24-1.0380481-rdf.xml
RDF/JSON: 24-1.0380481-rdf.json
Turtle: 24-1.0380481-turtle.txt
N-Triples: 24-1.0380481-rdf-ntriples.txt
Original Record: 24-1.0380481-source.json
Full Text
24-1.0380481-fulltext.txt
Citation
24-1.0380481.ris

Full Text

  INTER- AND INTRA-SITE HETEROGENEITY AS SOURCES FOR FAUNAL ASSEMBLAGE VARIABILITY: AN ANALYSIS OF FISH TAXA FROM NORTHERN TSIMSHIAN ARCHAEOLOGICAL SITES   by RAINI ABIGALE JOHNSON B.A., University of British Columbia, 2016  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS  in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  (Anthropology)  THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)   August 2019   Raini Abigale Johnson, 2019 ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the thesis entitled: Inter- and intra-site heterogeneity as sources for faunal assemblage variability: an analysis of fish taxa from Northern Tsimshian archaeological sites   submitted by Raini Abigale Johnson in partial fulfillment of the requirements for the degree of Master of Arts  in Anthropology  Examining Committee: Andrew Martindale, Anthropology Supervisor  Bruce Granville Miller, Anthropology Supervisory Committee Member  David Pokotylo, Anthropology External Examiner   iii  Abstract This thesis examines archaeological faunal assemblages of fish taxa from Northern Tsimshian village sites in the Prince Rupert Harbour (PRH) and surrounding area on the northern coast of British Columbia. Previously, the PRH has shown evidence for salmon dominated faunal assemblages which has led researchers to deem the region as an area of “extreme salmon specialization” (Coupland et al., 2010: 189). This thesis asks how representative this trend is within the study region by exploring the relationship between the three most prevalent fish species: salmon, herring, and smelt/other fish. When 45 faunal samples from 34 sites within the region were examined, 11 samples from nine archaeological sites are not dominated by salmon. Species variability within and between faunal assemblages was examined through the use of relative abundance and density calculations. Patterning due to variability in sampling methodology (column versus auger samples) and in the historic (i.e., representative of what is ‘in the ground’) spatial and temporal variables of inter-site location, re-occupation status, sample depth, intra-site sampling location, and site type were investigated through exploratory data analysis and statistical tests. This analysis found that all non-salmon dominated faunal assemblages came from two distinct contexts: sites from the Dundas Island Group and the Chatham Sound region. The density of fish was significantly larger in back midden samples than shell terrace and house floor samples, and camp sites had lower densities of faunal material than village sites. This research concludes with a call for action to validate subsistence patterns with substantial data before inferring regional patterns of subsistence. Archaeological investigations in the study region should take into consideration the effects of spatial and temporal heterogeneity on both vertical and horizontal planes when excavating, examining, and interpreting fauna. iv  Lay Summary This thesis examines archaeological fish assemblages from Northern Tsimshian village sites in the Prince Rupert Harbour and surrounding area on the northern coast of British Columbia. Previous work within the region focused on salmon as the primary food resource and proposed the region as a location of ‘extreme salmon specialization.’ This thesis examines the variability in abundance of three prominent fish species, salmon, herring, and smelt/other fish. Using relative abundances and densities, the patterning in non-salmon dominated samples are investigated via visual and statistical methods in terms of diversity in sampling methodology, site location, re-occupation status, sample depth, sampling location, and site type. Fish assemblages differ between camps and village sites, geographic regions, and sample locations from within sites. This research calls for a more data driven interpretation of subsistence and suggests that regional archaeological investigations in consider the effects of space and time when examining and comparing faunal assemblages.    v  Preface This thesis is an original, unpublished, intellectual product of the author, Raini Abigale Johnson.    vi  Table of Contents Abstract .......................................................................................................................................... iii Lay Summary ................................................................................................................................. iv Preface............................................................................................................................................. v Table of Contents ........................................................................................................................... vi List of Figures ................................................................................................................................ ix List of Tables .................................................................................................................................. x Acknowledgements ........................................................................................................................ xi Dedication ..................................................................................................................................... xii Introduction ..................................................................................................................................... 1 Background: The cultural history of the Northern Tsimshian ........................................................ 4 The occupational history of the Northern Tsimshian ................................................................. 5 Regional ecology and Northern Tsimshian seasonal subsistence patterns ................................. 7 The history of archaeology in Northern Tsimshian territory ...................................................... 9 Utilizing faunal analysis to understand Northern Tsimshian subsistence patterns ..................... 9 Salmonopia and the demonstrability of exploited fish taxa on the Northwest Coast ........... 10 Systemic problems which affect fish species representation ................................................ 12 Materials and Methods: Studying fish taxa from Northern Tsimshian sites ................................ 14 Faunal sample overview ........................................................................................................... 14 Temporal bracketing ............................................................................................................. 16 Comparing column and auger samples ................................................................................. 17 Methodology ................................................................................................................................. 19 vii  Faunal NISP counts............................................................................................................... 19 Examining variability between and within faunal samples .................................................. 22 Variable Descriptions............................................................................................................ 23 Visual Assessment of Data Trends ....................................................................................... 24 Statistical Analysis of Data Trends ....................................................................................... 25 Results: On the variability of fish species..................................................................................... 26 The consequences of sampling method .................................................................................... 26 The consequences of inter-site location .................................................................................... 29 The consequences of sample depth ........................................................................................... 33 The consequences of re-occupation status ................................................................................ 34 The consequences of intra-site sampling location .................................................................... 36 The consequences of site type ................................................................................................... 38 Summary of results ............................................................................................................... 41 Analysis and Discussion: Taxonomic patterning through historic variables ................................ 42 Diversity in relative abundance as a result of inter-site location .............................................. 42 Diversity in density as a result of intra-site sampling location ................................................. 44 Diversity in density as a result of site type ............................................................................... 46 Conclusions: Considering methodological issues when retrieving and analyzing fauna ............. 47 Future prospects on studying Northern Tsimshian subsistence ................................................ 47 Bibliography ................................................................................................................................. 49 Appendices .................................................................................................................................... 54 viii  Appendix 1: The Dataset .......................................................................................................... 54 Appendix 2: Statistical Results ................................................................................................. 56 Relative Abundance of Fish Taxa by Sampling Method ...................................................... 56 Density of Fish Taxa by Sampling Method .......................................................................... 57 Relative Abundance of Fish Taxa by Inter-Site Location .................................................... 58 Relative Abundance of Fish Taxa by Re-occupation Status ................................................. 62 Density of Fish Taxa by Intra-Site Location......................................................................... 64 Relative Abundance of Fish Taxa by Site Type ................................................................... 69 Density of Fish Taxa by Site Type ....................................................................................... 71    ix  List of Figures Figure 1. Map of study region, showcasing important Northern Tsimshian Rivers, Islands, and Localities. ........................................................................................................................................ 4 Figure 2. Map of analyzed sites by sampling method from the Dundas Island Group and three PRH regions (The Inner Harbour, Metlakatla Pass, and Chatham Sound). .......................................... 14 Figure 3. Relative abundance of fish taxa by sampling method. .................................................. 26 Figure 4. Density of fish taxa by sampling method. ..................................................................... 28 Figure 5. Relative abundance of fish taxa by inter-site location. .................................................. 29 Figure 6. NISP of fish taxa by relative depth below surface (DBS) from select auger samples. . 34 Figure 7. Relative abundance of fish taxa by re-occupation status. .............................................. 35 Figure 8. Density of fish taxa by intra-site location...................................................................... 36 Figure 9. Relative abundance of fish taxa by site type. ................................................................ 39 Figure 10. Density of fish taxa by site type. ................................................................................. 40 Figure 11. Diversity of fauna from village sites in the Chatham Sound with pie charts. ............. 43    x  List of Tables Table 1. Ethnographic seasonally harvested foods (adapted from the Port Simpson Curriculum Committee, n.d.: 5). ........................................................................................................................ 8 Table 2. Procurement and processing overview of faunal samples. ............................................. 14 Table 3. Regional overview of basal and terminal occupation dates for sampled sites. ............... 17 Table 4. Rank order abundance and ubiquity of identified fish taxa by sample. .......................... 20 Table 5. Statistically different salmon samples (x) in terms of inter-site location. ...................... 30 Table 6. Statistically different smelt/other fish samples (x) in terms of inter-site location. ......... 33 Table 7. Statistically different salmon samples (x) in terms of intra-site location. ...................... 37    xi  Acknowledgements I would like to first thank my thesis supervisor, Andrew Martindale, who has supported my work for the last five years and encouraged me to pursue a master’s degree. Andrew, I look forward to being encouraged to pursue a doctorate degree in the future.  This thesis would also not have been possible without the support of my whole community, my committee member Bruce Granville Miller, my external reviewer David Pokotylo, the support and encouragement of my fellow grad students, and the funding I received through the Social Sciences and Humanities Research Council of Canada (SSHRC) Canadian Graduate Scholarship-Master’s Award. Special thanks to TJ Brown, Jonathan Duelks, and Eric Simons for your camaraderie and research collaborations. Thank you to my family for your love and support. And thank you to Kenneth M. Ames for the extraordinary archaeological work you accomplished on the NWC. Your work has been an inspiration to many young archaeologists including myself and will continue to be for many generations to come. May you rest in peace.  Finally, thank you to the Qualicum, Cowichan, ʷməθkʷəy̓əm (Musqueam), Səl̓ílwətaʔ (Tsleil-Watuth), Stó:lō, Sts’ailes, Songhees, Esquimalt and WSÁNEĆ Coast Salish Nations and the Metlakatla and Lax Kw’alaams Northern Tsimshian Nations  for allowing me to grow up, learn, and work on your traditional ancestral unceded territories. Thank you to all the First Nation individuals who have taught me so much over the years and who I hope I can serve through my work in the future.    xii  Dedication To my gran,  for inspiring me and being inspired by me,  for teaching me to appreciate the beauty of small things,  and for always pushing me to learn deeper.  1  Introduction Pre-contact fisheries were fundamental modes of subsistence procurement for ancestral Indigenous groups on the Northwest Coast (NWC) and continue to be important for Indigenous people today. Pre-contact fisheries were interconnected with regional forms of socio-political complexity, ownership, and trade. Based on ethnographic and archaeological evidence, ancestral Northern Tsimshian people from the contemporary Prince Rupert Harbour (PRH) and surrounding region of northern British Columbia retained diets that were heavily focused on fish. In this northern region of the NWC, the three most ubiquitous fish species utilized were from the salmon (Oncorhyncus spp.), herring (Clupeidae spp.), and smelt (Osmeridae spp.) families.  The importance of salmon to ancestral diets has long been discussed on the NWC and many researchers have thought salmon and its ability to be captured and stored for winter was a fundamental driver for the complex Developed Northwest Coast Pattern (DNWCP) seen amongst coastal people in the late Holocene. However, other archaeologists have suggested that the faunal data, and therefore ancestral diets, from many regions of the NWC are more variable and diverse. Recent zooarchaeological investigations have found that salmon is not the most abundant fish species found in many NWC contexts (e.g., McKechnie and Moss, 2016). To confirm subsistence patterns and the role salmon played in the DNWCP, more data-driven studies must be completed to confirm that ethnographic trends extend back through the Holocene.  Although more variability in fish assemblages is found in many NWC contexts, investigations in Northern Tsimshian territory have continued to show that salmon dominates faunal assemblages, which led a group of researchers to deem village sites in Northern Tsimshian territory as home to evidence of “extreme salmon specialization” (Coupland et al., 2010: 189). This thesis examines if the trend of ‘extreme salmon specialization’ (i.e., salmon as the most 2  abundant fish taxa) is retained within the study region when using a larger dataset of fine-screened faunal samples (n=45) from 34 Northern Tsimshian archaeological sites, many of which have never before been examined for faunal remains. The rank abundance of the three most ubiquitous fish taxa, salmon, herring, and smelt, was explored using a count of number of identified specimens (NISP). This thesis finds that salmon is the most abundant fish species in 76% of the sampled faunal assemblages, with herring dominating 13% of samples, and smelt 11%. With 24% of the examined samples not dominated by salmon, the goal of this research became to understand if and what variables were patterning the dataset by investigating the relative abundance and density of fish taxa.  The effect of diversity in sampling methodology was the first variable investigated in this study. The faunal samples utilized in this analysis came from two distinct sampling methods, column sampling and auger sampling. The utility of augers for collecting representative faunal material is contested within the discipline yet has been shown to retrieve representative material, especially when taxonomic trends are distinct enough and have low heterogeneity (Cannon, 2000). After examining the effects of sampling, this research began investigating historic variables, defined as historically accurate and representative of the faunal signatures ‘in the ground,’ to determine if patterns in faunal abundance and density could be explained by diversity in space and time. Fish taxa variability was examined along the lines of inter-site location (investigating four distinct geographic regions - the Inner Harbour, Metlakatla Pass, Chatham Sound, and Dundas Island Group), re-occupation status (comparing sites re-occupied after the regionally defined hiatus period to sites not re-occupied), sample depth (examining changes in taxonomic abundance within samples), intra-site sampling location (exploring the variability in density of fish taxa from back middens, house floors, and shell terraces), and site type (comparing camp sites to villages).  3  The variables were examined for statistically significant differences in sampling methods. The results of 2-sample comparison tests show no significant difference in relative abundance and density between the two sampling methods. There is a significant difference between inter-site locations in the relative abundance of salmon. The Chatham Sound has a distinct salmon signature from the Inner Harbour and Metlakatla Pass. and the Metlakatla Pass and Dundas Island Group also differ significantly in smelt/other fish. There is no significant difference in herring between inter-site location. There was no statistically significant difference in relative abundance between re-occupied sites and sites not re-occupied, however, when examining the relationship between NISP and sample depth there is evidence for change over time. There is a significant difference between intra-site locations in the density of salmon with the back midden and both house terraces and shell terrace samples being significantly different from each other. There is no significant difference in herring or smelt/other fish between intra-site locations. There is no statistically significant difference in relative abundance between site types, however, Dundas Island Group village sites and camp sites differ statistically in density. Villages and camps are significantly different in terms of both salmon and smelt/other fish.   This thesis determines that Northern Tsimshian faunal assemblages, thus likely and ancestral diets, from the study region are, firstly, far more variable in relative abundance and density of fish taxa than previously proposed; secondly, that ‘extreme salmon specialization’ occurred at many sites but not all sites; and thirdly, that variability in fish taxa is spatially and temporally dependent on both vertical and horizontal patterns between sites and within them. These conclusions bring forward important methodological concerns about the importance of sampling and comparing similar kinds of data. Statements about subsistence should be confined to the spatial and temporal brackets from which the data originate, not prescribed to an entire 4  region without due consideration for comparability and sampling. Archaeologists who aim to utilize column and auger sampling methods to gain information on faunal assemblages and subsistence in the study region, and likely elsewhere on the NWC, should standardize the methods of sampling, the locations within sites from which samples are taken, and maintain temporal control. Without standardization, sampling for fauna in the study region is limited to inferences on the locations sampled - not the entire site or region, however, it can be stated that there is evidence for great diversity of fish species in taxonomic trends, abundance, and density.  Background: The cultural history of the Northern Tsimshian  Northern Tsimshian territory is situated on the northern Northwest Coast (NWC) of British Columbia and has been home to the Northern Tsimshian1 people since time immemorial (Figure 1). Northern Tsimshian territory traditionally included the coast and coastal islands, counting the                                                  1 Also called the Nine Tribes by Martindale and Marsden (2003), or the Coast Tsimshian or Metlakatla Tsimshian in other publications (e.g. Coupland et al., 2010).  Figure 1. Map of study region, showcasing important Northern Tsimshian Rivers, Islands, and Localities. 5  archipelago known as the Dundas Island Group (Dundas, Dunira, Melville, and smaller islands), located between the mouths of the two great northern rivers, the Skeena and the Nass (Figure 1) (Marsden and Galois, 1995; Martindale and Marsden, 2011). The cultural history of the region has been well documented through use of the adawx (Tsimshian oral histories) (Marsden, 2002; Martindale, 2006; Martindale and Marsden, 2003; Martindale et al., 2017a) and the archaeological record (Ames, 2005; Martindale et al., 2017a, 2017b). An ancient alliance saw 10 and later nine different Northern Tsimshian tribes inhabiting the study region (Martindale et al., 2017a). All nine tribes, the Gitwilgyoots, Ginax’angiik, Gitnadoiks, Gitzaxiaai, Giluts’aaw, Gits’iis, Gispaxlo’ots, Gitlaan, and Gitando had winter villages located in the protected areas through the Metlakatla Pass and around what is today the town of Prince Rupert (Figure 1) (Martindale et al., 2017a). These nine Northern Tsimshian tribes are today represented by the Lax Kw’alaams and Metlakatla First Nations (Figure 1).  There are two other linguistic groups within the Tsimshian nation, the Interior Tsimshian (the Gits’ilaasu and Gits’mgeelm) and Southern Tsimshian (the Gitk’a’ata and Gitkxaala) (Martindale et al., 2017a) whose territory is beyond the scope of this study. The occupational history of the Northern Tsimshian  A distinguishable characteristic of the study region and other NWC archaeological sites is the presence of large terraced shell-bearing features (in the past, these features were classified under the term shell midden) (Ames, 2005; Archer, 2001; Martindale et al., 2017a). The largest known terraced shell-bearing feature in the region is located in the PRH and extends more than 200,000 m2 and is over 8m deep (Letham et al., 2017). NWC shell-bearing sites are classified as villages if surface depressions of wooden plank houses (often in straight or curved rows) are present (Archer, 2001) or as non-villages or camp sites if shell terraces show no evidence for surface or subsurface structural depressions (Brewster and Martindale, 2011; Cannon, 2002).  6  At present, the earliest radiocarbon dates from the PRH are two camp sites which date between 8000 and 9000 calibrated years before present (cal BP) along with over 9000 cal BP dates from the Dundas Island Group and Stephens Island (Letham et al., 2016; 2018). Currently 66 village sites within the study region have been identified (Martindale et al., 2017a). The earliest dated houses are recorded at 6500 cal BP (Ames, 2005; Ames and Martindale, 2014; Martindale et al., 2017a, 2017b). These houses were arranged in small village formations, defined as villages of small size with small houses (Martindale et al., 2017a). By approximately 4500-3000 cal BP, villages within the study region grew in both size and number, with sites encompassing greater areas and having larger volumes of shell and an increased number of houses (Martindale et al., 2017a). These characteristically large villages contained houses of both small and large sizes with each village having the capacity to house hundreds of people (Letham et al., 2017; Martindale et al., 2017a, 2017b). The spread of these large villages is associated with a period of increased population size, density, and social complexity (Martindale and Marsden, 2003; Letham et al., 2017; Martindale et al., 2017a, 2017b). A hiatus in occupation occurred from approximately 1200 to 1000 cal BP where the entire region was depopulated (Martindale et al., 2017a, 2017b). It is thought that the population migrated inland to defensive sites in interior valleys in response to increased warfare with other nations (Marsden, 2001; Martindale et al., 2017a, 2017b). A smaller number of sites in the PRH region were repopulated after 1000 cal BP and remained into the contact period (Martindale et al., 2017a, 2017b). European contact in the region dates to 1787, although foreigners had a presence on the northern coast since 1741 (Martindale et al., 2017b). The subsistence economy in the region is difficult to date, but evidence from the Dundas Island Group camp site GcTr-6 suggests a complex fishing economy existed by 9000 years ago (McLaren et al., 2011).  Storage and surplus economics likely date from at least the establishment of large 7  village populations (4500 years ago) and undoubtedly were associated with small villages (by 6500 years ago) (Martindale et al., 2017a). Cannon (1998) defines the beginning of storage economics from faunal data on the central coast by at least 6000 years ago. Regional ecology and Northern Tsimshian seasonal subsistence patterns The region has a rich local ecology. Located in the Coastal Western Hemlock Zone (Pojar et al., 1991). The local vegetation consists of western hemlock, western red cedar, yellow cedar, lodgepole pine, Sitka spruce, and Douglas fir (Ames, 2005; Eldridge et al., 2015), and understory of salal, blueberry, bunchberry, huckleberry, deer fern, skunk cabbage, and a variety of mosses (Ames, 2005; Eldridge et al., 2015).  Common invertebrates include blue mussel, clams, barnacles, amphipods, limpets, and hermit crabs (Ames, 2005; Eldridge et al., 2015). Marine mammals local to the region include orcas, harbour and Dall’s porpoises, northern Steller sea lions, harbour seals, northern fur seals, and sea otters (Ames, 2005; Eldridge et al., 2015). Terrestrial mammals include Sitka, blacktail, and mule deer, caribou, mountain goats, mountain sheep, wapiti, black and grizzly bears, wolves, lynx, dogs, mink, hoary marmots, martens, muskrats, river otters, beavers, striped skunks, porcupines, and later Norway rats in the post-contact era (Ames, 2005; Eldridge et al., 2014). Canada geese, mallard ducks, and common mergansers are the only waterfowl species that regularly breed in the region, but many other species inhabit it seasonally (Ames, 2005). Bald eagles, ospreys, peregrine falcons, and ravens are present throughout the area and auklets are common on the smaller islands (Ames, 2005; Eldridge et al., 2015). Many anadromous and marine fish species are found in the region’s waterscape. Five species of salmon (sockeye, pink, coho, chum, and chinook) are present (Ames, 2005; Eldridge et al., 2015). Other species of anadromous fish include steelhead and four species of smelt (including eulachon) (Ames 2005). Herring, halibut, dogfish, and cod species are common marine fish (Ames, 2005; Eldridge et al., 2015).  8  Table 1. Ethnographic seasonally harvested foods (adapted from the Port Simpson Curriculum Committee, n.d.: 5).  Pre-contact ancestral Northern Tsimshian people were a semi-sedentary population with a socially stratified hunter-gatherer-fisher delayed return and surplus economy (Martindale et al., 2017a, 2017b). Subsistence focused on a diverse diet of seasonally available plants, invertebrates, and vertebrates (Table 1) with resources being stored and stockpiled in the spring, summer, and early fall for use during the winter months (Boas, 1916; Garfield, 1939, 1951; Halpin and Seguin, 1990; Miller, 1997; Port Simpson Curriculum Committee, n.d). Information on these seasonal rounds has been mentioned in fragments within a multitude of ethnographic works (Boas, 1916; Garfield, 1939, 1951; Halpin and Seguin, 1990; Miller, 1997) and summarized in two mid-2000s publications (Ames, 2005; MacDonald, 2006). The autumn and winter seasons were spent in winter villages where shellfish gathering occurred. Social activities such as ceremonies and potlaches were the focus of the Nine Tribes during this time. In late winter, the population moved to fishing villages on the lower Nass river to fish eulachon, then to outer islands in late spring for deep sea fishing and seaweed extraction. By midsummer the population moved again, this time to the Skeena River for salmon fishing, berry picking, and plant gathering.  Winter clams, cockles, china hats, mussels, abalone, crabs, sea urchins, sea prunes, sea cucumbers, octopus, winter spring salmon, grey cod, ling cod, black cod, red snapper, halibut, deer, mountain goat, mink, marten, otter, fox, muskrat Spring seaweed, berry sprouts, cow parsnip, licorice fern root, devil’s club, hemlock bark, rhubarb, jack pine sap, pine needles, abalone, crabs, sea urchin, sea prunes, sea cucumber, herring spawn, eulachon, spring salmon, halibut, sea lion, seal, seagull eggs Summer Salmonberries, raspberries, gooseberries, elderberries, huckleberries, dwarf blueberries, red and black currents, bunchberries, fireweed, soapberries, devil’s club, licorice fern root, abalone, crabs, sea urchins, sea prunes, sea cucumbers, octopus, sockeye salmon, coho salmon, pink salmon, halibut, bear Fall Wild crabapple, rose hips, bog cranberries, salalberries, licorice fern root, abalone, crabs, sea urchin, sea prunes, sea cucumber, octopus, dog salmon, spring salmon, halibut, moose, mountain goat, goose, duck 9  The history of archaeology in Northern Tsimshian territory  The archaeology of the Northern Tsimshian people has been heavily studied for the last century with a primary focus on sites located in the PRH. The region’s ‘shell heaps’ were first analyzed by Harlan I. Smith in the early 1900’s during the Jessup North Pacific Expedition (MacDonald and Inglis, 1980; Ames, 2005). Excavations by the Smithsonian’s Philip Drucker in 1938, and University of British Columbia professor Charles E. Borden along with local high school student James Baldwin in 1954 followed (MacDonald and Inglis, 1980; Ames, 2005). Starting in the mid 1960’s, the PRH was host to 10 large-scale excavations as part of the Canadian Museum of Man’s (now Canadian Museum of History) North Coast Prehistory Project (NCPP) (Ames, 2005). The NCPP ran from 1966 to 1980 and was led by George MacDonald and Richard Inglis (Ames, 2005). Kenneth Ames from Portland State University reanalyzed the NCPP and published a summary of the excavations and an analysis of the artifacts in 2005 (Ames, 2005). Large excavation and survey projects led by archaeologists Gary Coupland (University of Toronto), David Archer (North West Community College), Andrew Martindale (University of British Columbia), and Morley Eldridge (Millennia Ltd.) soon followed. In 2005 Andrew Martindale began conducting archaeological investigations within the greater Northern Tsimshian region on the Dundas Island Group and Stephens Island. All these projects collected faunal evidence to some degree and proposed historical interpretations based on these data. Utilizing faunal analysis to understand Northern Tsimshian subsistence patterns “Undoubtedly the main raison d'être for zooarchaeology has been to reconstruct diet (what people ate) and subsistence (how people obtained food)” Jonathan Driver (1993: 84).  Hunter-gatherer-fisher societies like the Northern Tsimshian can be said to have a subsistence-based economy (Ames, 2005). The study of subsistence-based economics through zooarchaeological faunal analysis explores how and what invertebrate and vertebrate resources 10  people harvested, processed, stored, consumed, and disposed of (Lyman, 2008). Faunal analysis is the practice of analyzing assemblages of fauna from archaeological contexts (Lyman, 2008). A faunal assemblage can be defined as a collection of animal remains from a distinct horizontally and vertically constrained space (Lyman, 2008). While the examination of faunal remains can tell us a considerable amount about past environments, bio-geography, and ancestral cultural values (Driver, 1993, Lyman, 2008), this thesis focuses on the use of faunal remains as evidence of past human subsistence activities to reconstruct subsistence patterns in Northern Tsimshian territory.  Salmonopia and the demonstrability of exploited fish taxa on the Northwest Coast Many ancestral people across much of the NWC relied heavily on fish resources for subsistence (e.g., Butler and Campbell, 2004; McKechnie and Moss, 2016). A recent study of fauna by Coupland et al. (2010) of five village sites in the PRH which focused on vertebrate faunal material found that approximately 80 to 99% of faunal assemblages were made up of fish remains, with the remainder encompassing terrestrial mammal, sea mammal, and bird remains. A similar isotopic marine to terrestrial ratio was observed in ancestral Tsimshian diets (Chisholm et al., 1982) and, ethnographically, fish played and continue to play a vital role in Northern Tsimshian subsistence (Boas, 1916; Garfield, 1951; Port Simpson Curriculum Committee, n.d.). Since the early 1900’s archaeologists, anthropologists, and ethnographers have examined the role of salmon in Indigenous NWC diets, with some authors likening the dietary importance of salmon on the coast to bison on the plains (Wissler, 1938). The storability of salmon, evident ethnographically, arguably became a fundamental driver for the progress of the Developed Northwest Coast Pattern (DNWCP) (Wissler, 1938; Drucker, 1965; Fladmark, 1975; Mitchell and Donald, 1988; Matson and Coupland, 1994) with the NWC defined as a “coastal strip … known for its wet but relatively mild climate, thick forests, numerous coastal resources, and abundant salmon” (Matson and Coupland, 1994: 2). The DNWCP causally linked the ability to maintain 11  surpluses of salmon to degrees of cultural and economic development and complexity which rivals that of past agricultural societies (Wissler, 1938; Drucker, 1965; Fladmark, 1975; Mitchell and Donald, 1988; Matson and Coupland, 1994). Complexity in this case was attributed to the prevalence of large semi-sedentary village populations with culturally dictated laws of ownership as well as class structures arising across much of the NWC between 3500 - 2500 years ago (Matson and Coupland, 1994). Today, it is not the complexity the DNWCP suggests that is contested but the role salmon played in forming and maintaining the cultural complexity of ancestral NWC people. A theoretical problem that the DNWCP and subsequent researchers faced was the ability to test the assumption of salmon as a developmental driver, a dominant subsistence resource, archaeologically. Monks (1987: 119) coined the term “salmonopia” to describes NWC researchers’ inability to see past the importance of salmon as a food resource and examine other fish species, and other vertebrate and invertebrate species, as well as plants, all which were fundamental to NWC Indigenous subsistence. Recent archaeological work has showcased that other fish species besides salmon have been heavily utilized on the NWC with some regions showing subsistence trends not dominated by salmon yet maintaining the degrees of complexity defined in the DNWCP (Butler and Campbell, 2004; Cannon, 2001; Monks, 1987, 2006; Orchard, 2007; McKechnie and Moss, 2016; Patton et al, 2019). Specifically, there is clear evidence for the importance and dominance of herring at many sites (Monks, 1987, 2006; McKechnie et al., 2014; Moss et al., 2015; McKechnie and Moss, 2016). A recent meta-analysis of fish assemblages from over 200 sites located between Oregon and Alaska found that herring was more ubiquitous (98.2%) than salmon (95.9%) (McKechnie and Moss, 2016: 4). This study also found that sites located close together contained similar taxonomic composition and taxon abundances and suggested that faunal 12  assemblages were “strongly associated with place and appear less subject to variations over time” (McKechnie and Moss, 2016: 13). On the Northern BC coast (a region which encompasses the Northern Tsimshian territory), salmon ranked first in abundance in 74.0% of samples, while herring ranked first in 9.7% (McKechnie and Moss, 2016). When ubiquity (frequency of occurrence) of fish taxa from the Northern BC coast region is examined, the three most ubiquitous are salmon (100.0%), herring (97.0%), and smelts (55.0%) (McKechnie and Moss, 2016). As salmon (Oncorhyncus spp.), herring (Clupea pallasii), and smelts (Osmeridae spp.) which include the often-indistinguishable subspecies of eulachon (Thaleichthys pacificus), surf smelts (Hypomesus pretiosus), and capelin (Mallotus villosus), are the three most ubiquitous species in the Northern BC region (McKechnie and Moss, 2016) these are the fish taxa that this analysis will focus on.  Published literature concerning the Northern Tsimshian region has continued to support salmon as the most ubiquitous and abundant taxa (Coupland et al., 2010). In a large-scale excavation project of five PRH (Inner Harbour) sites, Gary Coupland and his team found that salmon species contributed to approximately 90.0-95.0% of the fine-screened fish assemblage, with herring ranging from 2.5 to 5.5% and other identified fish species ranging from 1.0 to 4.5% (Coupland et al., 2010). This led Coupland et al. (2010:189) to deem the PRH and surrounding Northern Tsimshian territory as a region of “extreme salmon specialization.” Understanding and further substantiating or disproving this taxonomic trend, and the labeling of the study area as a region of ‘extreme salmon specialization” was a prime motive of this thesis.  Systemic problems which affect fish species representation  Beyond an unsubstantiated, ethnographically based, assumption of salmon domination, NWC archaeological evidence may have overrepresented salmon due to lack of sufficiently fine screening methodology (3.2-2.0 mm) (Cannon, 2000; McKechnie, 2005; Stewart and Wigen, 13  2003). Fine-screened samples allow for the capture of bones of smaller species (such as herring and smelts), small bones of larger species, and bone fragments. Consequently, faunal assemblages which do not utilize fine screen size skew faunal data towards larger species and unfragmented remains (Casteel, 1976; Stewart and Wigen, 2003). It is now accepted that for adequate retrieval of remains, mesh should be less than 3.0 mm but larger than 1.4 mm in size (Cannon, 2000; Casteel, 1976; McKechnie, 2005; Stewart at el., 2003; Stewart and Wigen, 2003). Stewart et al. (2003) found that 1.4 mm screens did not warrant the time and labour costs of identifying remains.  In this analysis the use of less than 3.0 mm screen should adequately capture salmon and herring bones yet smelt species may still be underrepresented (McKechnie and Moss, 2016). In the study region, vertebrate faunal assemblages have been identified in excavations since the 1960’s (Allaire, 1979; Inglis, 1977; MacDonald, 1969). However, excavations only began systematically focusing on faunal remains and utilizing screens with small mesh size in the late 1970’s (May, 1979; Coupland et al., 1993, 2002, 2006, 2010; Stewart et al., 2003, 2009; McKechnie, 2013; Patton, 2011; Patton et al., 2019). In this thesis, only fauna from samples screened through 2 mm mesh or less will be utilized so as to not skew data towards large species.  Further systemic problems which affect the taxonomic representation in fish assemblages include differential taphonomic effects and total number of elements, (Butler and Chatters, 1994; Driver, 1993; Grayson, 1984; Wheeler, 1978; Wigen and Stucki, 1988) as well as issues with identification (Cannon, 2000; Orchard, 2007; Wigen and Stucki, 1988). Salmon in general is more identifiable than herring or smelt, and salmon vertebrae are identifiable even when fragmented which can overrepresent the abundance of salmon. Differences in consumption and butchery practices may also overrepresent certain species, for example, smelts ethnographically were commonly consumed bone-in or rendered into an oil, leading to underrepresentation in the 14  archaeological record (McKechnie and Moss, 2016; Patton et al., 2019). Differences in shell accumulation may also affect preservation and abundance of species over time (Stein, 1992).  Materials and Methods: Studying fish taxa from Northern Tsimshian sites Faunal sample overview This analysis compiles published (Brewster and Martindale, 2011; Coupland et al., 2010; Patton et al., 2019) and unpublished (Adams, 2016) faunal data from 34 archaeological sites in Northern Tsimshian territory, specifically from the Dundas Island Group and the PRH (Figure 1 and Figure 2). Forty-five samples were examined in this analysis. Table 2 gives an overview of the procurement and processing methods utilized in the analyzed faunal samples.  Table 2. Procurement and processing overview of faunal samples. Collected By No. of sites Borden Designation  Sampling method Screen size (mm) Volume (l) Wet/dry screening Katherine Patton, Gary Coupland 5 GbTo-28 GbTo-31 GbTo-46 GbTo-77 GcTo-06 column 6.3, 2.0, 1.4 1.0 dry Figure 2. Map of analyzed sites by sampling method from the Dundas Island Group and three PRH regions (The Inner Harbour, Metlakatla Pass, and Chatham Sound).  15  Andrew Martindale 3 GbTo-23 GbTo-24 GbTo-34 column 6.3, 2.0 0.5-1.0 wet and dry Andrew Martindale 10 GbTo-04 GbTo-24 GbTo-34 GcTo-01 GcTo-06 GcTo-27 GcTo-28 GcTo-39 GcTo-51 GcTo-52 auger 6.3, 2.0 0.25-1.25 wet  Natalie Brewster, Andrew Martindale 18 GcTq-01 GcTq-04 GcTq-05 GcTq-06 GcTq-07 GcTq-08 GcTq-09 GcTq-10 GcTq-11 GcTq-12 GcTq-13 GcTr-05 GcTr-06 GcTr-08 GcTr-09 GcTr-10 GdTq-01 GdTq-03 auger 6.3, 2.0 ~0.75 wet   These samples were wet and dry screened (GbTo-23 only wet screened) using 6.3- and 2.0-mm mesh and identified using NISP at the University of British Columbia by Iain McKechnie (now at the University of Victoria), Mariko Adams, and Raini Johnson. Martindale also collected auger samples at 5-10 cm increments with volumes ranging from 250-1250 ml. These samples were wet screened in the lab using 6.3- and 2.0-mm mesh and identified using NISP by Aubrey Cannon (McMaster University). The Dundas Island Group auger samples were collected by Natalie Brewster (McMaster) and Martindale in 5-10 cm increments with volumes of approximately 770 ml in volume. These samples were wet screened in the lab using 6.3- and 2.0-mm mesh and identified using NISP by Natalie Brewster. Bags of material from the beginning and end of column and auger samples which contained no faunal elements were cut from this analysis under the assumption that no fauna equates to non-cultural material from pre- or post-occupation periods (Appendix 1). Although there is diversity in sampling and processing methodology, I argue below that these methods are comparable in both screen size and volume. The samples utilized in this analysis came from 34 sites from four distinct regions based on regional geography, the Inner Harbour (also known as the PRH proper), the Metlakatla Pass, the Chatham Sound (coastal sites), and the Dundas Island Group, also called the Dundas Archipelago (see Appendix 1: The Dataset). These samples came from village sites and camp sites 16  (Appendix 1), and were retrieved from three distinct locations within sites, the shell terrace itself, house floors, and back middens (Appendix 1). Shell terraces can be defined as the shell-bearing features on which houses were constructed. The samples taken from shell terraces came from areas in front of and beside house depressions. House floor samples were taken from inside house depressions. The back midden of a site is the area located behind the houses furthest from the shoreline. Back middens are areas often associated with refuse disposal. Unfortunately, the location of samples from the Dundas Island Group sites are not yet published (Appendix 1). The faunal material in this analysis came from samples which ranged in size between 1 - 41 bags of material and had total sample matrix volumes between 1.00 and 80.6 liters (Appendix 1).  Temporal bracketing Unfortunately, all the faunal samples in this analysis lack corresponding radiocarbon dates, and therefore, temporal control. As a result, it is outside of the scope of this thesis to examine relationships between trends in fauna and time. While basal and terminal dates for the majority of the sites (22/34) analyzed here do exist (Table 3), there is no consistent relationship between relative depth and time. Letham et al. (2017) showed that shell accumulation rates were not consistent through time, with the relationship between depth and time varying dramatically within sites. Consequently, the only relationships between temporal dates that can be examined are through the overall temporal brackets (basal, hiatus, and terminal dates). For example, sites re-occupied after the hiatus period (~1200-1000 cal BP) can be compared to sites which were not re-occupied (Table 3). It is impossible, however, to reliably associate any of these faunal samples more specifically than within these rough temporal brackets. While this analysis is unable to discuss time as a continuous or precise variable, relative depth will be utilized to discuss some changes in faunal patterns through time. Samples from greater depths are from earlier periods than those from shallow depths in undisturbed contexts.  17  Table 3. Regional overview of basal and terminal occupation dates for sampled sites. Region Site  14C dates Earliest date (basal) Latest date (terminal)     (n) (Cal BP) (Cal BP) Inner Harbour  GbTo-23 1 27 6657 341  GbTo-28 7 2842 1178  GbTo-311 49 4787 396  GbTo-46 8 2622 1743  GcTo-061 19 5889 735 Metlakatla Pass GbTo-041 8 3180 240  GbTo-341 65 5721 319  GcTo-011 4 3436 482  GcTo-39 6 3470 1574 Chatham Sound GbTo-24 8 3155 1269  GbTo-77 7 3239 1943  GcTo-27  7 4160 1509  GcTo-28 5 3079 1587  GcTo-51 8 2846 1354  GcTo-52 6 2719 1749 Dundas Island Group GcTq-01 5 2787 1711  GcTq-05 12 9195 1355  GcTr-05 4 2544 1414  GcTr-08 11 7257 1840  GdTq-01 9 6338 1758 1 Sites which were re-occupied after the hiatus period (~1200 -1000 cal BP).  Comparing column and auger samples As stated above, these faunal samples were retrieved using column and auger sampling techniques and their material processed and examined for the presence of vertebrate faunal remains. It has been generally accepted that small samples of material are as sufficient as large-scale excavations in determining information on the focus and intensity of NWC fisheries (e.g., Casteel, 1970, 1976; Meighan et al., 1958). In fact, field sorted material from excavation units tends to be more time consuming, biased against small species and elements, and less representative of taxonomic trends when compared to lab processed column samples (Casteel, 1976). Both column and augering are common archaeological sampling methods used in NWC archaeological investigations and data from both methods are often combined to create larger datasets (e.g., Patton et al., 2019). Unfortunately, neither method is standardized, making it common to have archaeological samples, such as in this study, that differ in both volume and mass (Cannon, 2000; Driver, 1993; Gray, 2008; McKechnie, 2005; Moss 1989, 2007; Moss et al., 2017; Patton, 2011; Stein, 1992). Column samples are commonly utilized for collecting fauna and are considered to retrieve a representative faunal collection of small elements such as fish, accurately 18  capturing faunal signatures if combined with adequate processing methods. Column samples are taken by bagging material either during an excavation or by sampling from an eroding portion of a site (often called a slump face). Auger samples are taken using a hand-held bucket auger that drills into the earth and retrieves columnar sections of material as it penetrates down. Auger samples maintain depth precision, in that material is recovered in specific increments (often between 5 to 10 cm) but does not account for changes in stratigraphic layers. Augers also only collect material smaller than their bit diameter. The auger samples in this analysis were retrieved using a bucket auger with a 7 cm bit, therefore, animal remains larger than ~4-6 cm in diameter would likely not be retrieved and not likely representative of mammalian and avian elements. Augers have been utilized primarily as a tool to establish the presence of cultural material, but recent studies have turned to augers as a minimally invasive tool to collect faunal material in multi-component sites (Cannon, 2000; Patton et al., 2019). Column sampling erosion faces, and auger sampling are common methods utilized due to their efficiency in time and cost compared to excavations (Cannon, 2000). There is debate within the archaeological community of the ability of augers to capture faunal material representative of what is actually occurring within the wider horizontal and vertical space of a site, yet an archaeologically sound study at the NWC Namu site and surrounding region found that auger samples did collect representative faunal material for fish (Cannon, 2000). This study took auger samples along with corresponding core samples, which take continuous sections of sediment and retain stratigraphy, from areas located next to previous full-scale excavation units, matched stratigraphy and depth, dated transitions, and determined that the density of fish taxa from the augers was comparable to the original excavation units and could sufficiently be used to determine resource focus and intensity as well as changes over time (Cannon, 2000). As a simple statistical sampling pattern, auger and column samples are more 19  likely to be representative of patterns in space and time if taxonomic trends are clearly patterned and have limited modality, i.e., if there are patterns that are consistent, visible, and involve few variable species. Only fish taxa will be discussed in this analysis because the size of the utilized auger bit is too small to capture a representative collection of avian and mammalian species. Since the overall ratio of fish (~80-100%) to other vertebrate species in the PRH (Coupland et al., 2010) indicates that fish is the dominant vertebrate subsistence resource for Northern Tsimshian people, augering and column sampling will capture important subsistence trends.  Methodology Faunal NISP counts The faunal data from the samples utilized in this analysis were counted using an NISP measure which is the “most fundamental unit by which faunal remains are tallied” (Lyman, 2008: 27) and is the most commonly used counting method on the NWC (Appendix 1). The NISP calculation adds up the total number of identifiable faunal skeletal elements (e.g., bones, teeth, antler, fragments of each) to the most specific taxon (class, order, family, genus, or species) possible (Lyman, 2008). Taxonomical identifications also tend to be indicative of skeletal element (Lyman, 2008), for example a salmon bone can also be identified as a salmon vertebra. 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑖𝑖𝑖𝑖𝑛𝑛𝑛𝑛𝑖𝑖𝑖𝑖𝑜𝑜𝑖𝑖𝑛𝑛𝑖𝑖 𝑠𝑠𝑠𝑠𝑛𝑛𝑠𝑠𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑠𝑠 1 𝑖𝑖𝑖𝑖𝑛𝑛𝑛𝑛𝑖𝑖𝑖𝑖𝑜𝑜𝑖𝑖𝑖𝑖𝑛𝑛𝑖𝑖𝑛𝑛 𝑛𝑛𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑖𝑖 (𝑛𝑛𝑜𝑜𝑛𝑛𝑛𝑛, 𝑖𝑖𝑜𝑜𝑜𝑜𝑖𝑖ℎ,𝑜𝑜𝑛𝑛𝑖𝑖𝑓𝑓𝑛𝑛𝑛𝑛𝑛𝑛𝑖𝑖) =  1 𝑠𝑠𝑠𝑠𝑛𝑛𝑠𝑠𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛 Minimum number of individuals (MNI) and meat weight calculations are not utilized in this analysis for two reasons. First, augers take samples vertically and are not likely to retrieve whole specimens disposed of in horizontal space, thus they are less likely to be representative of numbers of individual animals than elements. Secondly, PRH fish assemblages in general are 20  dominated by vertebral fish remains, not complete skeletons, possibly due to past human behavior. A relative lack of cranial elements could be attributed to processing practices. Some authors (e.g., Calvert, 1973; Chatters, 1984) have suggested that when collecting salmon backbone-in transportation occurred, whereby fish heads were disposed of at the catching location and only the flesh with the backbone in were transported back to winter villages for storing. If de-heading occurs at logistic locations like camp sites, it is likely that camps and villages would show different fish element signatures (Chatters, 1984). An alternative hypothesis is that non-vertebral remains such as cranial bones tend to be thinner and less dense than vertebrae allowing for quicker taphonomic affects and fragmentation (Butler and Chatters, 1994; Driver, 1993; Wheeler, 1978; Wigen and Stucki, 1988). Unlike the high degree of vertebral fish remains from these sites, mammalian and avian remains when found tend to be more complete or more diverse in element representation (Coupland et al., 2010). For these reasons stated above, fish species may be underrepresented in these NISP calculations.  In this analysis, all three major fish taxa, salmon, herring, and smelt, were counted using the NISP count (Appendix 1). A fourth category, a count of all other identified fish species, was utilized to show further diversity within the samples. Unfortunately, eight samples used in this analysis did not publish the individual results of smelt NISP and instead lumped smelt NISP into the all other identified fish category (Appendix 1). It can be assumed that smelt was present in many if not all of these samples (see Patton et al., 2019) yet for the purpose of statistical analysis the NISP of smelt will be combined with the NISP of other fish creating an NISP smelt/other fish count. Table 4. Rank order abundance and ubiquity of identified fish taxa by sample.   Rank   Site No. of samples  1st 2nd 3rd GbTo-04 1 Salmon (397) Herring (32) Smelt (9) GbTo-23 1 Salmon (72) Herring (41) Smelt (0) 21  GbTo-24 2 Smelt (8) Salmon (5) Herring (0)   Smelt (47) Salmon (26) Herring (3) GbTo-28 2 Salmon (65) Herring (21) Smelt not ID’d   Salmon (503) Herring (136) Smelt not ID’d GbTo-31 4 Salmon (106 Herring (8) Smelt not ID’d   Salmon (71) Herring (0) Smelt not ID’d   Salmon (179) Herring (4) Smelt not ID’d   Salmon (302) Smelt (12) Herring (9) GbTo-34 2 Salmon (161) Herring (6) Smelt (1)   Salmon (106) Herring (63) Smelt (0) GbTo-46 3 Salmon (53) Herring (8) Smelt not ID’d   Salmon (302) Herring (38) Smelt not ID’d   Salmon (106) Smelt (53) Herring (30) GbTo-77 4 Salmon (5) Herring (2) Smelt (1)   Smelt (55) Herring (45) Salmon (4)   Herring (107) Smelt (89) Salmon (72)   Salmon (28) Smelt (25) Herring (14) GcTo-01 1 Salmon (114) Herring (14) Smelt (6) GcTo-06 2 Salmon (127) Herring (17) Smelt (10)   Salmon (204) Herring (3) Smelt not ID’d GcTo-27 1 Herring (69) Salmon (30) Smelt (0) GcTo-28 1 Herring (142) Salmon (29) Smelt (2) GcTo-39 1 Salmon (200) Herring (43) Smelt (3) GcTo-51 1 Salmon (82) Herring (19) Smelt (1) GcTo-52 1 Herring (45) Salmon (19) Smelt (3) GcTq-01 1 Salmon (187) Smelt (123) Herring (51) GcTq-04 1 Salmon (36) Herring (23) Smelt (0) GcTq-05 1 Salmon (703) Herring (103) Smelt (82) GcTq-06 1 Smelt (54) Herring (13) Salmon (12) GcTq-07 1 Salmon (143) Smelt (43) Herring (24) GcTq-08 1 Salmon (39) Smelt (5) Herring (4) GcTq-09 1 Herring (22) Salmon (7) Smelt (4) GcTq-10 1 Salmon (18) Smelt (5) Herring (5) GcTq-11 1 Herring (8) Salmon (0) Smelt (0) GcTq-12 1 Salmon (23) Smelt (4) Herring (0) GcTq-13 1 Smelt (17) Salmon (16) Herring (2) GcTr-05 1 Salmon (552) Smelt (72) Herring (46) GcTr-06 1 Salmon (7) Herring (4) Smelt (0) GcTr-08 1 Salmon (188) Herring (89) Smelt (53) GcTr-09 1 Salmon (59) Herring (14) Smelt (2) GcTr-10 1 Salmon (415) Smelt (59) Herring (25) GdTq-01 1 Salmon (332) Smelt (29) Herring (17) GdTq-03 1 Salmon (8) Herring (7) Smelt (0)  NISP counts can be utilized to analyze the ubiquity and rank abundance of fish taxa (Table 4). Ubiquity is a measure of abundance and examines the presence absence of material between contexts. Rank abundance looks at the rank of species by frequency within single contexts. For this analysis the first, second, and third most abundant taxa are noted (Table 4). When examining the ubiquity of fish taxa between the faunal samples from the study region, salmon is present in 44 of 45 samples (97.7%), herring in 43 samples (95.5%) and smelt is found in 30 out of the 37 samples in which smelt was aimed to be identified (81.0%) (Table 4). When ranking the species, out of the 45 samples, salmon ranks first in 34 samples (75.6%), second in eight samples (17.8%), 22  and third in three samples (6.7%) (Table 4). This is one of the first instances where salmon is not the most abundant taxa from all PRH and Northern Tsimshian faunal samples. Herring ranks first in six samples (13.3%), second in 25 samples (55.6%), and third in 14 samples (31.1%) (Table 4). Smelt ranks first in five samples (11.1%), second in 12 samples (26.7%), and third in 28 samples (62.2%) (Table 4). Salmon and herring can both be considered ubiquitous taxa within the study region. I would also argue that smelt (cf. Patton et al., 2019) is also a ubiquitous taxon that is underrepresented in this analysis due to smelt specific sampling biases and differential use.  Examining variability between and within faunal samples Analysis of fish taxa from these faunal samples shows that within the study area, salmon is not always the most abundant species, nor is salmon any more ubiquitous than herring (Table 4). With 11 samples from nine archaeological sites not dominated by salmon, questions arise as to what the drivers of this pattern could be. Is the diversity in species abundance and rank caused by variability in sampling, therefore misrepresenting faunal data, or are there historic ‘in the ground’ variables affecting the faunal assemblages, such as differences in use, space, or time?  This analysis employs standardized measurements of relative abundance (%) and density (fish/litre) to examine patterns within and between the sampled faunal assemblages (Appendix 1). Relative abundance is defined as the percent composition of a species relative to the total number of identified specimens from a single context, in this case a single faunal sample.  𝑅𝑅𝑛𝑛𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑅𝑅𝑛𝑛 𝐴𝐴𝑛𝑛𝑛𝑛𝑛𝑛𝑖𝑖𝑖𝑖𝑛𝑛𝑠𝑠𝑛𝑛 = �𝑇𝑇𝑖𝑖𝑇𝑇𝑜𝑜𝑛𝑛𝑇𝑇𝑜𝑜𝑖𝑖𝑖𝑖𝑖𝑖𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁� 𝑇𝑇100 Relative abundance was calculated by dividing the individual NISP of the three major fish taxa: salmon, herring, and smelt as well as the NISP of all other identified fish specimens by the total NISP and then multiplied by 100 to create the relative percentage of a sample for which a fish taxon takes up. Relative abundance allows fish taxa from different samples to be compared to each 23  other even when samples differ in total NISP. 𝐷𝐷𝑛𝑛𝑛𝑛𝑠𝑠𝑖𝑖𝑖𝑖𝐷𝐷 = 𝑇𝑇𝑖𝑖𝑇𝑇𝑜𝑜𝑛𝑛 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑉𝑉𝑜𝑜𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛 (𝑖𝑖)  The density of fish taxa is calculated by dividing the NISP of the individual fish taxa by the total matrix volume (l) of faunal samples. A density calculation gives the NISP of an individual taxon found in one litre of material. Density allows for fish taxa from different samples to be compared to each other even when samples vary in volume.  Variable Descriptions Primary categories of investigated variables included sampling method (variability in technology), regional site location (inter-site variability), re-occupation status and sample depth (variability in time), sampling location (intra-site variability), and site type (variability in use).  As stated above, augers have been said to capture representative enough faunal data for fish from NWC shell-bearing sites (Cannon, 2000), and faunal data from column and auger samples are often combined in publications (e.g. Patton et al., 2019). In this analysis, the relative abundance and density of fish taxa from village sites were compared for patterning between the two sampling methods. Then, the historical ‘representative’ variables of inter-site location, re-occupation status, sample depth, intra-site sampling location, and site type were examined for patterning.  First, to examine variability in site location (i.e., inter-site regional heterogeneity), the village sites were sorted by the four distinct geographic regions within Northern Tsimshian territory: The Inner Harbour, Metlakatla Pass, Chatham Sound, and the Dundas Island Group. Patterns in inter-site location were then examined by the relative abundance of fish taxa.  As discussed above, the lack of temporal control impedes subsistance investigations within the study region, however, two ordinal temporal variables can be analyzed: re-occupation status 24  and sample depth. Six of the village sites discussed in this analysis were reoccupied after the hiatus period (~1200 to 1000 cal BP) (Edinborough et al., 2017; Martindale et al., 2017a): three sites from the Inner Harbour (GbTo-23, GbTo-31, Gcto-06) and three sites from the Metlakatla Pass (GbTo-04, GbTo-34, GcTo-01). No sites from the Chatham Sound region or the Dundas Island Group were re-occupied post hiatus. Re-occupied and not re-occupied sites from the Inner Harbour were compared and examined by the relative abundance of fish taxa. To inspect the effects of depth, a subset of eight auger samples were examined comparing the relationship between NISP and depth below surface (DBS) from which the bags of material were retrieved.  To examine variability in sampling location (i.e., intra-site heterogeneity) four Inner Harbour village sites were selected (GbTo-28, GbTo-31, GbTo-46, and GcTo-06) and sorted by sample location (shell terraces, house floors, back middens). The relative abundance and density of fish taxa from each intra-site location were compared and examined for patterning.  Finally, variability in site type (i.e., village or camp site) was examined and the relative abundance and density of fish taxa from two site types compared. It is not known what effect that site type, primarily village versus camp, has on faunal assemblages. The sites from the Inner Harbour, Metlakatla Pass, and Chatham Sound examined in this analysis are all classified as villages, however, the Dundas Island Group sites consist of both villages and camps. Site type trends were examined using only the faunal assemblages from the Dundas Island Group sites. Visual Assessment of Data Trends  The relationship between variables was visually portrayed using bar graphs and box and whisker charts. The box and whisker chart can be read as; the circles as outliers (more than 2/3 times the 1st quartile), the top whisker is the greatest value (excluding the outliers), the top of the box/1st quartile (25% of the data is greater than this value),  the median line where 50% of the data is greater than this value, the bottom of the box/3rd quartile (25% of the data is less than this value),  25  and the bottom whisker which is the lowest value. The x indicates the mean or average of the data.  Statistical Analysis of Data Trends Variability in sampling method, inter-site location, re-occupation status, intra-site location, and site type were analyzed via bivariate 2-sample comparison statistics using the PAST (PAleontological STatistics) version 3.25 software. The full results of the statistical analysis can be found in Appendix 2: Statistical Results. The standard 2-sample comparison hypothesis is stated below:  𝑇𝑇ℎ𝑛𝑛 𝑁𝑁𝑛𝑛𝑖𝑖𝑖𝑖 𝐻𝐻𝐷𝐷𝑠𝑠𝑜𝑜𝑖𝑖ℎ𝑛𝑛𝑠𝑠𝑖𝑖𝑠𝑠 (𝐻𝐻0) = 𝑇𝑇𝑇𝑇𝑜𝑜 𝑓𝑓𝑛𝑛𝑜𝑜𝑛𝑛𝑠𝑠𝑠𝑠 𝑖𝑖𝑛𝑛𝑛𝑛 𝑖𝑖𝑛𝑛𝑖𝑖𝑇𝑇𝑛𝑛 𝑜𝑜𝑛𝑛𝑜𝑜𝑛𝑛 𝑖𝑖ℎ𝑛𝑛 𝑠𝑠𝑖𝑖𝑛𝑛𝑛𝑛 𝑠𝑠𝑜𝑜𝑠𝑠𝑛𝑛𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑜𝑜𝑛𝑛 𝑜𝑜𝑜𝑜 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑇𝑇ℎ𝑛𝑛 𝐴𝐴𝑖𝑖𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑖𝑖𝑖𝑖𝑖𝑖𝑅𝑅𝑛𝑛 𝐻𝐻𝐷𝐷𝑠𝑠𝑜𝑜𝑖𝑖ℎ𝑛𝑛𝑠𝑠𝑖𝑖𝑠𝑠 (𝐻𝐻1) = 𝑇𝑇𝑇𝑇𝑜𝑜 𝑓𝑓𝑛𝑛𝑜𝑜𝑛𝑛𝑠𝑠𝑠𝑠 𝑖𝑖𝑛𝑛𝑛𝑛 𝑜𝑜𝑛𝑛𝑜𝑜𝑛𝑛 𝑖𝑖𝑖𝑖𝑜𝑜𝑜𝑜𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑖𝑖 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑠𝑠𝑜𝑜𝑠𝑠𝑛𝑛𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑜𝑜𝑛𝑛𝑠𝑠 Due to small sample sizes (n<10) in some groups, only 2-sample comparisons will be utilized in this analysis instead of n-sample ANOVA (one-way analysis of variance) tests. These data will be examined using Exploratory Data Analysis (EDA) and Shapiro-Wilks normality tests. Normally distributed samples will be compared using a t-test for equal means after equal variance is tested for. Non-normally distributed samples will be examined using a Mann-Whitney U test. For tests between two variables, the significance level (p-value) of α=0.05 will be utilized, meaning that samples with p-values >0.05 will fail to reject the null hypothesis.  For tests with more than two variables a Bonferroni correction to the significance level (p-value) will be implemented to avoid false positive significance results. The Bonferroni correction is calculated as original p-value divided by the number of comparison tests run. For example, 0.5/6 = 0.0083, meaning that the p-value of these tests must be <0.0083 to be from statistically different populations.  26  Results: On the variability of fish species The consequences of sampling method Relative abundance values were used to analyze the fish taxa from the two different sampling methods (Figure 3). Of the 38 valid samples, 21 from augers and 17 from column samples (Figure 3).   Figure 3. Relative abundance of fish taxa by sampling method. When examining the relative abundance of salmon in auger samples, the data have a non-normal distribution (Shapiro-Wilk W=0.8781, p=0.01345). The relative abundance of salmon in column samples have a normal distribution (W=0.9132, p=0.1134). A Mann-Whitney U test conducted between auger (median = 70.1%, n=21) and column samples (median = 64.36%, n=17) does not indicate a statistically significant difference in the relative abundance of salmon between the two sample groups (U=146, p=0.3475).   When examining the relative abundance of herring, the auger samples have a non-normal distribution (W=0.7765, p=0.000294), while column samples have a normal distribution (W=0.9044, p=0.08048). A Mann-Whitney U test conducted between the augers (median = Relative Abundance of Fish Taxa by Sampling MethodGbTo-04GbTo-24GbTo-34GcTo-01GcTo-06GcTo-27GcTo-28GcTo-39GcTo-51GcTo-52GcTq-01GcTq-04GcTq-05GcTq-06GcTq-07GcTq-10GcTq-11GcTr-05GcTr-08GcTr-10GdTq-01Auger0102030405060708090100Relative Abundance (%)Salmon Herring Smelt / Other FishGbTo-23GbTo-24GbTo-28(A)GbTo-28(B)GbTo-31(A)GbTo-31(B)GbTo-31(C)GbTo-31(D)GbTo-34GbTo-46(A)GbTo-46(B)GbTo-46(C)GbTo-77(A)GbTo-77(B)GbTo-77(C)GbTo-77(D)ColumnAuger0102030405060708090100Column27  13.89%, n=21) and columns (median = 15.79%, n=17) does not indicate a statistically significant difference in the relative abundance of herring between auger and column samples (U=156.5, p=0.52789).   Data for relative abundance of smelt/other fish in auger samples have a non-normal distribution (W=0.7221, p=0.0000526), while the relative abundance of smelt/other fish in column samples have a non-normal distribution (W=0.815, p=0.003354). A Mann-Whitney U test conducted between the auger (median = 10%, n=21) and column (median = 12.5%, n=17) samples shows no statistically significant difference in density of smelt/other fish between auger and column samples (U=157, p=0.53744).   In all three cases, there is no statistically significant difference in the relative abundance of fish taxa from auger and column samples between the two sampling methods.  Density values were used to analyze the fish taxa from the two different sampling methods (Figure 4). Of the 38 valid samples, 21 from augers and 17 from column samples (Figure 4).  Salmon density values in auger samples have a non-normal distribution (W=0.8773, p=0.01304). The density of salmon in column samples also have a non-normal distribution (W=0.6948, p=0.0001009). A Mann-Whitney U test conducted between the augers (median = 6.96, n=21) and columns (median = 10.5, n=17) shows no statistically significant difference in the density of salmon between auger and column samples (U=115.5, p=0.066496).   28   Figure 4. Density of fish taxa by sampling method. With respect to herring density, both auger sample and column sample data have non-normal distributions (auger samples W=0.7647, p=0.0001987; column samples W=0.7217, p=0.0002056). A Mann-Whitney U test was conducted between the augers (median = 1.05 fish/litre, n=21) and columns (median = 2 fish/litre, n=17) does not indicate a statistically significant difference in the density of herring between auger and column samples (U=155, p=0.49941).  The density of smelt/other fish in auger samples the data have a non-normal distribution (W=0.8173, p=0.00122). The density of smelt/other fish in column samples also have a non-normal distribution (W=0.6561, p=0.000000105)A Mann-Whitney U test was conducted between the augers (median = 0.97, n=21) and columns (median = 1, n=17), indicates no statistically significant difference in the relative abundance of herring between auger and column samples (U=165, p=0.70266).   In all three cases, there is no statistical difference between the density of fish taxa from auger and column samples.  Density of Fish Taxa by Sampling MethodGbTo-04GbTo-24GbTo-34GcTo-01GcTo-06GcTo-27GcTo-28GcTo-39GcTo-51GcTo-52GcTq-01GcTq-04GcTq-05GcTq-06GcTq-07GcTq-10GcTq-11GcTr-05GcTr-08GcTr-10GdTq-01Auger0102030405060708090DensitySalmon Herring Smelt / Other FishGbTo-23GbTo-24GbTo-28(A)GbTo-28(B)GbTo-31(A)GbTo-31(B)GbTo-31(C)GbTo-31(D)GbTo-34GbTo-46(A)GbTo-46(B)GbTo-46(C)GbTo-77(A)GbTo-77(B)GbTo-77(C)GbTo-77(D)GcTo-06ColumnAuger010203040506070Column29  The consequences of inter-site location  Figure 5. Relative abundance of fish taxa by inter-site location. Relative abundance values were used to analyze the fish taxa from the four geographic regions within Northern Tsimshian territory (Figure 5). Of the 38 valid samples, 12 are from the Inner Harbour, 5 from the Metlakatla Pass, 10 from the Chatham Sound, and 11 from the Dundas Island Group. Given the Metlakatla Pass has a small sample size (n=5), comparisons were made using t-tests not n-sample tests. As six pairwise comparisons will be made the Bonferroni corrected significance value will be used, which in this case is 0.05/6 = 0.00833. When examining the relative abundance of salmon, all four regions: Inner Harbour (W=0.9217, p=0.3006), Metlakatla Pass (W=0.9528, p=0.7571), Chatham Sound (W=0.899, p=0.2137), and Dundas Island Group (W=0.8872, p=0.1283) have normal distributions.  A t-test comparing the Inner Harbour samples (mean = 80.18%, n=12) and the Metlakatla Pass (mean = 77.34%, n=5) shows equal variance (F=1.753, p=0.62067) between sample groups, and a result of 0.40154 (p=0.69368), indicating no significant difference between the two groups. A t-test of samples from the Inner Harbour (mean = 80.18%, n=12) and the Chatham Sound (mean Relative Abundance of Fish Taxa by Re-occupation StatusGbTo-23GbTo-28(A)GbTo-28(B)GbTo-31(A)GbTo-31(B)GbTo-31(C)GbTo-31(D)GbTo-46(A)GbTo-46(B)GbTo-46(C)GcTo-06(A)GcTo-06(B)Inner Harbour0102030405060708090100Relative Abundance (%)Salmon Herring Smelt / Other FishGbTo-04GbTo-34(A)GbTo -34(B)GcTo-01GcTo-39Metlakatla PassGbTo-24(A)GbTo-24(B)GbTo-77(A)GbTo-77(B)GbTo-77(C)GbTo-77(D)GcTo-27GcTo-28GcTo-51GcTo-52Chatham SoundGcTq-01GcTq-04GcTq-05GcTq-06GcTq-07GcTq-10GcTq-11GcTr-05GcTr-08GcTr-10GdTq-01Dundas Island GroupInner Harbour0102030405060708090100Metlakatla Pass Chatham SoundDundas Island Group30  = 32.30%, n=10), with equal variance (F=2.5056, p=0.15323) between samples, has a value of 6.1022 (p=0.00000579), indicating the two regional groups are significantly different. A t-test was conducted between samples from the Inner Harbour (mean = 80.18%, n=12) and the Dundas Island Group (mean = 56.67%, n=11). There is unequal variance (F=4.0263, p=0.031246) between samples and the unequal variance t-test has a value of 2.4786 (p=0.026138), indicating no significance between the two groups. A t-test was conducted between samples from the Metlakatla Pass (mean = 77.34%, n=5) and the Chatham Sound (mean = 32.30%, n=10). There is equal variance (F=4.3922, p=0.1679) between samples and the t-test has a value of 4.2047 (p=0.001031), indicating the two regional groups are significantly different. A t-test was conducted between samples from the Metlakatla Pass (mean = 77.34%, n=5) and the Dundas Island Group (mean = 56.67%, n=11). There is equal variance (F=7.058, p=0.074678) between samples and the t-test has a value of 1.5532 with a (p=0.14268), indicating no significant difference between the two groups. A t-test was conducted between samples from the Chatham Sound (mean = 32.30%, n=10) and the Dundas Island Group (mean = 56.67%, n=11). There is equal variance (F= 1.6069, p=0.48804) between samples and the t-test has a value of 2.1682 (p=0.043044), indicating no significant difference between the two groups. In sum, salmon from the Chatham Sound region shows statistically different results than the samples from the Inner Harbour and Metlakatla Pass (Table 5). Table 5. Statistically different salmon samples (x) in terms of inter-site location.   When examining the relative abundance of herring, the Inner Harbour (W=0.8986, p=0.1523), Metlakatla Pass (W=0.9333, p=0.6187), and Chatham Sound (W=0.9349, p=0.4977) Inner Harbour (salmon) Metlakatla Pass (salmon) Chatham Sound (salmon) Dundas Island Group (salmon)Inner Harbour (salmon) xMetlakatla Pass (salmon) xChatham Sound (salmon) x xDundas Island Group (salmon)31  have normal distributions while the Dundas Island Group (W=0.7257, p=0.001004) has a non-normal distribution.  A t-test comparing the Inner Harbour samples (mean = 11.61%, n=12) and the Metlakatla Pass (mean = 20.18%, n=5) shows equal variance (F=1.4271, p=0.57782) between sample groups, and a result of 1.4385 (p=0.17082), indicating no significant difference between the two groups. A t-test of samples from the Inner Harbour (mean = 11.61%, n=12) and the Chatham Sound (mean = 36.12%, n=10) has unequal variance (F= 6.8671, p=0.0041191) between samples and the unequal variance t-test has a value of 2.6334 (p=0.022985), indicating no significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Inner Harbour samples (median = 10.51%, n=12) and the Dundas Island Group median = 13.35%, n=11). The Mann-Whitney U test has a value of 45 (p=0.20706), indicating no significant difference between the two groups. A t-test was conducted between samples from the Metlakatla Pass (mean = 20.18%, n=5) and the Chatham Sound (mean = 36.12%, n=10) had equal variance (F= 4.8119, p=0.14471) between samples and the t-test has a value of 1.2039 (p=0.25009), indicating no significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Metlakatla Pass (median = 17.34, n=5).and the Dundas Island Group (median = 13.35%, n=11). The Mann-Whitney U test has a value of 22 (p=0.57109), indicating no significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Chatham Sound (median = 31.82%, n=10) and the Dundas Island Group (median = 13.35%, n=11). The Mann-Whitney U test has a value of 35 (p=0.16971), indicating no significant difference between the two groups. In all six cases there is no significant difference in the relative abundance of herring from 32  the four regions.  When examining the relative abundance of smelt/other fish, the Metlakatla Pass (W=0.7945, p=0.7146), and Chatham Sound (W=0.8536, p=0.06417) have normal distributions while the Inner Harbour (W=0.8352, p=0.02423) and Dundas Island Group (W=0.8178, p=0.01613) have non-normal distributions.  A non-parametric Mann-Whitney U test was conducted between samples from the Inner Harbour (median = 3.23%, n=12) and the Metlakatla Pass (median = 2.02%, n=5). The Mann-Whitney U test has a value of 22 (p=0.42806), indicating no significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Inner Harbour (median = 3.23%, n=12) and the Chatham Sound (median = 23.94%, n=10). The Mann-Whitney U test has a value of 25 (p=0.022876), indicating no significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Inner Harbour (median = 3.23%, n=12) and the Dundas Island Group (median = 18.14%, n=11). The Mann-Whitney U test has a value of 26 (p=0.01503), indicating no significant difference between the two groups. A t-test was conducted between samples from the Metlakatla Pass (mean= 2.47%, n=5) and the Chatham Sound (mean = 31.59%, n=10). There is unequal variance (F= 172.2, p=0.00016305) between samples and the unequal variance t-test has a value of 2.9903 (p=0.009202), although close to the Bonferroni correction value (0.0083) the p-value still indicates no significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Metlakatla Pass (median = 2.02, n=5) and the Dundas Island Group (median = 18.14%, n=11). The Mann-Whitney U test has a value of 0 (p=0.002222), indicating a significant difference between the two groups. A non-parametric Mann-Whitney U test was conducted between samples from the Chatham Sound (median = 23.94%, n=10) and the 33  Dundas Island Group (median = 18.14%, n=11). The Mann-Whitney U test has a value of 55 (p= 0.97191), indicating no significant difference between the two groups.  In sum, the smelt/other fish from the Metlakatla Pass differ significantly from the Dundas Island Group sample (Table 6).  Table 6. Statistically different smelt/other fish samples (x) in terms of inter-site location.  The consequences of sample depth Patterns are visible within the subset of auger samples chosen to examine the relationship between fish species NISP and the depth below surface (DBS) from which the sample matrix was retrieved. Without corresponding dates, it cannot be confirmed that trends in depth occurred at the same time in different samples, however, it does suggest that over time changes happened. Between approximately 150 and 250 cm DBS herring spikes in five samples (GbTo-34, GcTo-27, GcTo-28, GcTo-51, and GcTo-52) and salmon spikes in four samples (GbTo-04, GbTo-34, GcTo-01, and GcTo-51) (Figure 6). Double spikes of salmon are also present in four samples (GbTo-04, GcTo-01, GcTo-39, GcTo-51) (Figure 6). Inner Harbour (smelt/other fi sh )Metlakatla  Pass  (smelt/other fi sh )Chatham Sound (smelt/other fi sh )Dundas  Is land Group (smelt/other fi sh )Inner Harbour (smelt/other fi sh )Metlakatla  Pass  (smelt/other fi sh ) xChatham Sound (smelt/other fi sh )Dundas  Is land Group (smelt/other fi sh ) x34   The consequences of re-occupation status Relative abundance values were used to analyze fish taxa from 12 samples from Inner Harbour sites which were re-occupied after the hiatus period (n=7) compared to sites which were not re-occupied (n=5) (Figure 7). When examining the relative abundance of salmon from re-occupied sites (W=0.8525, p=0.1295) and not re-occupied sites (W=0.9851, p=0.9599) both have normal distributions. A 2-05101520250 50 100 150 200 250 300NISPDBS (cm)G b T o - 3 40246810121416180 50 100 150 200 250 300NISPDBS (cm)G c T o - 0 105101520250 50 100 150 200 250 300 350 400NISPDBS (cm)G c T o - 2 701020304050600 50 100 150 200 250 300NISPDBS (cm)G b T o - 0 4010203040500 50 100 150 200 250 300 350 400NISPDBS (cm)G c T o - 2 801020304050600 50 100 150 200NISPDBS (cm)G c T o - 3 90246810120 50 100 150 200 250 300NISPDBS (cm)G c T o - 5 1024681012140 50 100 150 200 250 300NISPDBS (cm)G c T o - 5 2Figure 6. NISP of fish taxa by relative depth below surface (DBS) from select auger samples. 35  sample comparison t-test was conducted between the re-occupied (mean = 86.64%, n=7) and not re-occupied (mean = 71.15%, n=5). There is equal variance (F=1.0979, p=0.87309) between samples and the t-test has a value of 2.1585(p=0.056251), indicating no significant difference between the two groups.  Figure 7. Relative abundance of fish taxa by re-occupation status. When examining the relative abundance of herring, re-occupied sites (W=0.6999, p=0.003696) have a non-normal distribution while not re-occupied sites (W=0.8615, p= 0.2336) have a normal distribution. A Mann-Whitney U test was conducted between the re-occupied sites (median = 13.89%, n=7) and not re-occupied sites (median = 15.97%, n=5). The Mann-Whitney U test has a value of 6 (p=0.074035), indicating no significant difference between the two groups. When examining the relative abundance of smelt/other fish from re-occupied sites (W=0.8383, p=0.09574) and not re-occupied sites (W=0.8995, p=0.4074) both have normal distributions. A 2-sample comparison t-test was conducted between the re-occupied (mean = 4.77%, n=7) and not re-occupied (mean = 12.99%, n=5). There is equal variance (F=4.9771, p=0.082104) between samples and the t-test has a value of 1.5662 (p=0.14836), indicating no significant difference between the two groups. Relative Abundance of Fish Taxa by Re-occupation StatusRe-occupied0102030405060708090100Relative Abundance (%)Salmon Herring Smelt / Other FishNot Re-occupiedRe-occupied0102030405060708090100Not Re-occupied36  In all three cases, there is no statistically significant difference between the relative abundance of fish taxa from re-occupied and not re-occupied Inner Harbour sites.   The consequences of intra-site sampling location  Figure 8. Density of fish taxa by intra-site location. Density values were used to analyze the fish taxa from the four geographic regions within Northern Tsimshian territory (Figure 8). Of the 12 valid samples, 3 are from back middens, 3 are from house floors, and 6 are from shell terraces. Given the small sample sizes comparisons were made using t-tests not n-sample tests. As four pairwise comparisons will be made the Bonferroni corrected significance value will be used, which in this case is 0.05/3 = 0.0167. When examining the density of salmon, the back midden (W=0.9057, p=0.4037), house floor (W=0.7811, p=0.07022), and shell terrace (W=0.9726, p=0.9093) samples all have normal distributions.  A t-test comparing the back midden (mean = 53.98 fish/litre, n=3) and house floor (mean = 8.91 fish/litre, n=3) samples shows equal variance (F=33.512, p=0.057951) between sample groups, and a result of 4.2472 (p=0.013188), indicating that there is a significant difference Density of Fish Taxa by Site TypeBack Midden01020304050607080DensitySalmon Herring Smelt / Other FishHouse Floor Shell TerraceBack Midden010203040506070House Floor Shell Terrace37  between the two groups. A t-test comparing the back midden (mean = 53.98 fish/litre, n=3) and shell terrace (mean = 10.53 fish/litre, n=6) samples shows equal variance (F=7.8278, p=0.057658) between sample groups, and a result of 5.5255 (p=0.000882), indicating that there is a significant difference between the two groups. A t-test comparing the house floor (mean = 8.91 fish/litre, n=3) and shell terrace (mean = 10.53 fish/litre, n=6) samples shows equal variance (F=4.2811, p=0.40026) between sample groups, and a result of 0.4009 (p=0.70045), indicating there is no significant difference between the two groups.  In sum, there is a statistical difference in the density of salmon between the back midden and the house floor and shell terrace samples (Table 7).  Table 7. Statistically different salmon samples (x) in terms of intra-site location.  When examining the density of herring, the back midden (W=0.8813, p=0.3282), house floor (W=0.8833, p=0.3341), and shell terrace (W=0.9619, p=0.8345) samples all have normal distributions.  A t-test comparing the back midden (mean = 5.89 fish/litre, n=3) and house floor (mean = 2.43 fish/litre, n=3) samples shows equal variance (F=8.9536, p=0.20093) between sample groups, and a result of 1.3211 (p=0.25697), indicating no significant difference between the two groups. A t-test comparing the back midden (mean = 5.89 fish/litre, n=3) and shell terrace (mean = 0.63 fish/litre, n=6) samples shows equal variance (F=112.99, p=0.00013789) between sample groups, and a result of 2.1163 (p=0.16744), indicating no significant difference between the two groups. A t-test comparing the house floor (mean = 2.43 fish/litre, n=3) and shell terrace (mean = 0.63 fish/litre, n=6) samples shows equal variance (F=12.619, p=0.022236) between sample groups, Back Midden (salmon) House Floor (salmon) Shell Terrace (salmon)Back Midden (salmon) x xHouse Floor (salmon) xShell Terrace (salmon) x38  and a result of 2.1366 (p=0.1566), indicating no significant difference between the two groups.   In sum, there is no statistical difference in the density of herring between the three intra-site sample locations. When examining the density of smelt/other fish, the back midden (W=0.7668, p=0.03758), sample has a non-normal distribution while house floor (W=0.9155, p=0.4366) and shell terrace (W=0.9241, p=0.5355) samples have normal distributions.  A Mann-Whitney U test was conducted between back midden (median = 0.40 fish/litre, n=3) and house floor (median = 2.50 fish/litre, n=3) samples. The Mann-Whitney U test has a value of 3 (p=0.66252), indicating no significant difference between the two groups. A Mann-Whitney U test was conducted between back midden (median = 0.40 fish/litre, n=3) and shell terraces (median = 0.33 fish/litre, n=6) samples. The Mann-Whitney U test has a value of 7 (p=0.69733), indicating no significant difference between the two groups. A t-test comparing the house floor (mean = 3.17 fish/litre, n=3) and shell terrace (mean = 0.36 fish/litre, n=6) has unequal variance (F=31.495, p=0.0029332) between samples and an unequal variance t-test has a value of 2.4267 (p=0.13214), indicating no significant difference between the two groups. In sum, there is no statistical difference in the density of smelt/other fish between the three intra-site sample locations. The consequences of site type Relative abundance values were used to analyze fish taxa from 18 samples from Dundas Island Group region 7 of which are categorized as camps and 11 of which are villages (Figure 9) When examining the relative abundance of salmon, the village site (W=0.8872, p=0.1283) and camp site (W=0.8992, p=0.326) samples have normal distributions. A t-test comparing villages (mean = 56.67%, n=11) and camps (mean = 57.83%, n=7) has equal variance (F=1.4294, 39  p=0.68618) between samples with a t-test has a value of 0.089339 (p=0.92992), indicating no significant difference between the two groups.  Figure 9. Relative abundance of fish taxa by site type. When examining the relative abundance of herring, the village site (W=0.7257, p=0.001004) sample has a non-normal distribution while the camp site (W=0.9123, p=0.4123) samples has a normal distribution. A Mann-Whitney U test was conducted between villages (median = 13.35%, n=11) and camps (median = 18.67%, n=7). The Mann-Whitney U test has a value of 35 (p=0.78585), indicating no significant difference between the two groups. When examining the relative abundance of smelt/other fish, the village site (W=0.8178, p=0.01613) sample has a non-normal distribution while the camp site (W=0.815, p=0.05748) samples has a normal distribution. A Mann-Whitney U test was conducted between villages (median = 18.14%, n=11) and camps (median = 14.81%, n=7). The Mann-Whitney U test has a value of 36 (p=0.85626), indicating no significant difference between the two groups. In sum, there is no statistically significant difference in the relative abundance of fish taxa between site types from the Dundas Island Group. Density values were used to analyze fish taxa from 18 samples from Dundas Island Group Relative Abundance of Fish Taxa by Site TypeGcTq-01GcTq-04GcTq-05GcTq-06GcTq-07GcTq-10GcTq-11GcTr-05GcTr-08GcTr-10GdTq-01Villages0102030405060708090100Relative Abundance (%)Salmon Herring Smelt / Other FishGcTq-08GcTq-09GcTq-12GcTq-13GcTr-06GcTr-09GdTq-03CampsVillages0102030405060708090100Camps40  region 7 of which are categorized as camps and 11 of which are villages (Figure 10).   Figure 10. Density of fish taxa by site type. When examining the density of salmon, the village site (W=0.9097, p=0.2417) and camp site (W=0.8833, p=0.2414) samples have normal distributions. A t-test comparing villages (mean = 5.78 fish/litre, n=11) and camps (mean = 2.23 fish/litre, n=7) has unequal variance (F=6.9318, p=0.027704) between samples with an unequal valiance t-test has a value of 2.2194 (p=0.043664), indicating a significance between the two groups.  When examining the relative abundance of herring, the village site (W=0.9109, p=0.2501) sample has a non-normal distribution while the camp site (W=0.7707, p=0.02076) samples has a normal distribution. A Mann-Whitney U test was conducted between villages (median = 0.96 fish/litre, n=11) and camps (median = 0.47 fish/litre, n=7). The Mann-Whitney U test has a value of 17.5 (p=0.063229), indicating no significant difference between the two groups. When examining the density of smelt/other fish, the village site (W=0.8988, p=0.1787) sample has a normal distribution while the camp site (W=0.5847, p=0.0001795) sample has a non-normal distribution. A Mann-Whitney U test was conducted between villages (median = 1.23 fish/litre, n=11) and camps (median = 0.38 fish/litre, n=7). The Mann-Whitney U test has a value Density of Fish Taxa by Site TypeVillages0246810121416DensitySalmon Herring Smelt / Other FishCampsVillages0246810121416Camps41  of 16 (p=0.04632), indicating a significant difference between the two groups. In sum, there is a statistically significant difference in the density of salmon and the density of smelt/other fish between camps and villages. Summary of results These results indicate that, inter-site location, sample depth, intra-site location, and site type are variables in which statistical and visible patterns in fish taxa in either relative abundance or density are present and have the potential to influence Northern Tsimshian faunal samples while sampling method and time (in a re-occupation sense) do not.  The results of the statistical analysis show that there was no significant difference in relative abundance and density between the two sampling methods. There is a statistical difference between inter-site locations in the relative abundance of salmon. The Chatham Sound has a distinct salmon signature from the Inner Harbour and the Metlakatla Pass. The Metlakatla Pass also differs from the Dundas Island Group in smelt/other fish. There is no significant difference in herring between regions. There was no significant difference in relative abundance between re-occupied sites and sites not re-occupied however when examining the relationship between NISP and sample depth there is evidence for change over time. There is a significant difference between intra-site locations in the density of salmon with the back middens differing significantly from both the house floor and the shell terrace samples. There is no significant difference in herring or smelt/other fish between intra-site locations. There is no significant difference in relative abundance between site types however Dundas Island Group village sites and camp sites differ statistically in density. Villages and camps are significantly different in terms of both salmon and smelt/other fish.  42  Analysis and Discussion: Taxonomic patterning through historic variables  Subsistence studies, as discussed in this thesis, are paramount for understanding ancestral human activities that pertain specifically to diet as well as to larger cultural practices. The results of this thesis showcase the historic effects of inter- and intra-site variability and site type on faunal patterns with relative abundance of fish taxa differing by region, density differing by sample location within sites and by site type. These results allow us to begin theorizing on the cultural roots of this variability of fish taxa in Northern Tsimshian territory. Diversity in relative abundance as a result of inter-site location All faunal samples not dominated by salmon came from two distinct regional contexts, sites from the Dundas Island Group and sites from the Chatham Sound. Two sites from the Dundas Island Group dominated by herring and two by smelt while four sites from the Chatham Sound are dominant in herring and three in smelt. The Chatham Sound is significantly different from the Inner Harbour and Metlakatla Pass regions in terms of their relevant abundance of salmon and the Dundas Island Group is significantly different from the Metlakatla Pass in smelt/other fish.  Greater diversity in species use in these regions is likely attributed to differences in subsistence practices between occupants of the coastal sites and sites located in more protected areas such as the Inner Harbour and Metlakatla Pass. For one, the local ecology of these sites differs due to their location. When examining the habitats in which the species examined in this work thrive, the largest salmon-bearing rivers in the region are the Skeena and the Nass, while herring and smelt are often found on the outer coast and islands. The results of this thesis could suggest that local resource extraction occurred more often on the outer coast than in the Inner Harbour. Local resource extraction is the act of using resources from a very local environment compared to logistical resource extraction whereby resources are retrieved from extra-local camp 43  sites and transported or traded in (Binford, 1980). The Inner Harbour is known to be a low diversity area where very little local resources are local to the large village sites. The results of this thesis may suggest that the diversity in inter-site location could be caused by a difference in local versus logistical resource extraction.  Figure 11. Diversity of fauna from village sites in the Chatham Sound with pie charts.  Note that sites with multiple samples (GbTo-24 [n=2], GbTo-77 [n=4]) have been combined for this figure.  This diversity within the Chatham Sound can further expand on this point. When the data is mapped onto the landscape, which shows a relationship between site location and faunal assemblage signature (Figure 11), further intra-regional patterns are visible. Three (GcTo-27, GcTo-28, GcTo-52) of the four sites located on the west facing side of the Tsimshian Peninsula are dominated by herring (Figure 11). The two samples from GbTo-24 on Tugwell Island are dominated by other fish, notably dogfish and smelt (Figure 11). The four samples from GbTo-77 located on the west side of Digby Island have samples high in herring and smelt (Figure 11). These results further suggest difference in local habitat or site use.  It is also possible that a change in site use over time from camps to villages occurred in the 44  Chatham Sound region. When the relationship between fish NISP and relative depth below surface (DBS) was analyzed to examine changes in faunal signatures over time through vertical space, spikes in the abundance of herring at approximately 200cm below surface are evident in in all four samples from the Tsimshian Peninsula (GcTo-27, GcTo-28, GcTo-51, and GcTo-52) (Figure 6). It cannot be confirmed that 200 cm DBS occurs at the same time at all four sites without corresponding dates, but the trend of spiked herring suggests similar changes to the use or abundance of resource extraction over time. This potential transition of site use over time may correspond with known historical settlement patterns and demographics. A known population boom occurred in the study region at approximately 2500 cal BP which saw new village sites being inhabited in less ideal locations (Martindale et al., 2017a, 2017b). It would make sense that ancestral Northern Tsimshian people would populate and construct villages on top of camp sites they already used, and which had shell-bearing features already built. Unfortunately, without proper temporal data this cannot yet be tested, however, corresponding core samples were taken along with many of the auger samples allowing for a future project to further examine and date these samples. Diversity in density as a result of intra-site sampling location  Similar to variability and potential changes in site use over time seen through vertical examinations of faunal trends, variability in horizontal space is patterned. When examining the diversity in faunal density by intra-site location a significant pattern differentiates the back midden samples from those from the house floors and shell terraces in terms of the density of salmon. Back middens have a mean density of 53.98 fish/litre while house floors have a mean density of 8.91 fish/litre and shell terraces have mean density of 10.53 fish/litre. Examining these means, back middens therefore have approximately six times more salmon than house floors, and five times 45  more salmon than shell terraces.  These results could suggest that the horizontal spatial variability in faunal density evident in this analysis is the result of differential material accumulation and area use within a site. Diversity within shell-bearing terraced sites within the study region has been mentioned in recent archaeological studies, yet has remained mostly untested (Letham et al., 2017; Martindale et al., 2017b; Patton et al., 2019) and the effects of intra-site heterogeneity on faunal remains is are currently unknown. Literature (Blukis Onat, 1985; Classen, 1991; Marquardt, 2010) has begun to examine shell-matrix archaeological sites as locations of constructed “planned placement of prepared shell” (Blukis Onat, 1985; 201) rather than viewing shell-bearing sites largely as food debris (i.e., shell middens). A recent study by Letham et al. (2017) on shell accumulation rates in the PRH found areas of high accumulation and areas of low accumulation during the same time periods within individual sites, suggesting that PRH shell-bearing sites are not homogenous entities. At one site, radiocarbon dating showed directionality of material deposition from north to south (Letham et al., 2017: 28), which could suggest planned construction over time as a result of a need to increase the size of the habitation platform. Although not yet systematically tested, it is likely that there are fundamental differences in shell and faunal material accumulation rates between back midden locations and shell terraces. Furthermore, if portions of sites are different in how they were built, they are likely also different in what they were used for, and what kinds of material, in this case subsistence remains, are found within them. Based on the diversity in density of faunal remains seen in this analysis back middens are likely areas where culturally dictated refuse disposal occurred, while shell terraces are the result of purpose-built foundation construction. Shell terraces could have portions of quickly accumulated material while back middens may be more likely to be made up of slowly accumulated material.  46  If samples were taken or matched with corresponding dates, shell accumulation rates could be confirmed as was done in the study by Letham et al. (2017). If faunal density is tested from these three intra-site locations throughout the greater Northern Tsimshian region, and house floor, shell terrace, and back midden locations are fundamentally different in shell accumulation and faunal density, then our archaeological methods of retrieving faunal material in the future must speak to this and standardization in intra-site sample location may be required. This could play a fundamental role in legal debates interested in past subsistence if back midden locations were excluded from faunal testing endeavors. Diversity in density as a result of site type When examining the diversity in faunal density by site type using the Dundas Island Group samples a strong statistical pattern differentiates the back midden, house floors, and shell terraces in terms of the density of salmon and of smelt/other fish. Village sites have a mean salmon density of 5.78 fish/litre while camps have a mean salmon density of 2.23 fish/litre. Village sites have a mean smelt/other fish density of 1.23 fish/litre while camps have a mean smelt/other fish density of 0.38 fish/litre. Comparing the mean values, villages therefore have approximately two and a half times more salmon and three times more smelt/other fish than camps. This pattern holds well with what we know about logistic resource extraction sites and collector hunter-gatherer-fisher populations (Binford, 1980). Camp sites are likely locations where resources are retrieved from local environments and often processed before being brought back to villages to be consumed and stored (Binford, 1980). Camp sites are more likely to be located nearest to specific resource catchments, with locations less determined by other habitation factors, unlike village sites which are often, but not always (Supernant and Cookson, 2014) in highly accessible and sheltered locations. Therefore, the faunal remains disposed of in camps are more 47  likely products of catch and consume subsistence and pre-storage processing. Village sites would more likely be catchments for resources retrieved from a multitude of camp site locations. Archaeologically, it would be expected that camp sites show more localized resources in less dense amounts. Camp sites in the study region not located near salmon-bearing rivers such as the Nass or the Skeena would likely show low amounts of salmon. Conclusions: Considering methodological issues when retrieving and analyzing fauna  The results of this analysis identify important methodological issues that need to be assessed when faunal analysis is a goal of archaeological investigations. These consideratrions are currently specific to projects in the study area but are likely also pertinent to other regions on the NWC. Firstly, variability in fish taxa is spatially and temporally dependent on both vertical and horizontal planes between sites and within them. Secondly, sampling methodologies and technologies should be standardized within the NWC archaeological community so larger regional studies can be conducted using easily coalesced data. Lastly, and the major conclusion of this thesis, is that salmon was not the most important subsistence resource in all locations within Northern Tsimshian territory, which signifies that claims of regional subistance practices, like the Northern Tsimshian region as an area of “extreme salmon specialization” need to be substantially investigated in data driven studies and should be confined to the spatial and temporal brackets from which the data originate, not prescribed to an entire region without due consideration for comparability and sampling. Future prospects on studying Northern Tsimshian subsistence Although the lack of temporal control negates an ability for regional conclusions to be drawn beyond the increased taxonomic variability shown within this thesis, and spatial and temporal patterning to be fully understood, it is a hope that this thesis can be used as a pilot study 48  to structure future archaeological investigations within Northern Tsimshian territory and throughout the greater NWC. Future research prospects will be to continue examining new Northern Tsimshian sites for faunal signatures, to systematically test the trend of highest faunal density in back middens, and to cross-examine and date corresponding core samples to the auger samples utilized in this thesis to create temporal control within the samples.  Finally, as Wayne Suttles (1968: 67) so aptly put it “I am afraid I have raised more questions about the Northwest Coast [NWC subsistence, and archaeological practices] than I have been able to answer”. Yet, it is of the upmost importance for archaeologists to consider and standardize the method of sampling, the locations within sites from which samples are taken, and to maintain temporal control. These factors contain important implications for future legal and scholarly debates through both the effects of testing location on faunal signatures and the predominant focus on specific taxa, in this case salmon.   49  Bibliography Adams, M. 2016. Exploring the Efficacy of Faunal Recovery Techniques in the Northwest Coast  of North America. Honours thesis, University of British Columbia, Vancouver. Allaire, L. 1979. The cultural sequence at Gitaus: A case of prehistoric acculturation. In (Ed.) Inglis, R., MacDonald, G.M., Skeena River Prehistory, Archaeological Survey of Canada, Mercury Series, paper 87, pp. 18-52.  Ames, K.M. 2005. The North Coast Prehistory Project Excavations in Prince Rupert Harbour, British Columbia: The Artifacts. Site report. Portland State University. Ames, K.M., Martindale, A. 2014. Rope Bridges and Cables: A Synthesis of Prince Rupert  Harbour Archaeology. Canadian Journal of Archaeology 38(1): 140–178. Archer, D.J.W. 2001. Village patterns and the emergence of ranked society in the Prince Rupert  area. In: Cybulski, J.S. (Ed.), Perspectives on Northern Northwest Coast Prehistory. Archaeological Survey of Canada Paper No. 160, Canadian Museum of Civilization, Hull, pp. 203–223. Binford, L. 1980. Willow smoke and dog's tails: hunter-gatherer settlement systems and  archaeological site formation. American Antiquity. 45(1): 4–20. Blukis Onat, A.R. 1985. The multifunctional use of shellfish remains: From garbage to  community engineering. Northwest Anthropological Research Notes 19(2): 201–207. Brewster, N., Martindale, A. 2011. An archaeological history of Holocene fish use in the Dundas  Island Group, British Columbia. In: Moss, M., Cannon, A. (Eds.), The Archaeology of North Pacific Fisheries. University of Alaska Press, Fairbanks, pp. 247–264. Boas, F. 1916. Tsimshian mythology. In 31st Annual Report of the Bureau of American Ethnology for the Years (1909-1910), Washington, pp. 29-1037. Butler, V.L., Campbell, S. 2004. Resource Intensification and Resource Depression in the Pacific  Northwest of North America: A Zooarchaeology Review. Journal of World Prehistory 18(4): 327–405.  Butler, V.L., Chatters, J.C. 1994. The role of bone density in structuring prehistoric salmon bone  assemblages. Journal of Archaeological Science 21: 413-424. Cannon, A. 1998. Contingency and agency in the growth of Northwest Coast maritime  economies. Artic Archaeology 35(1): 57-67.              2000. Assessing Variability in Northwest Coast Salmon and Herring Fisheries: Bucket- Auger Sampling of Shell Midden Sites on the Central Coast of British Columbia. Journal of Archaeological Science 27: 725-737.             2002. Sacred power and seasonal settlement on the central Northwest Coast. In:  Fitzhugh, B., Habu, J. (Eds.), Beyond Foraging and Collecting: Evolutionary Change in Hunter-Gatherer Settlement Systems. Kluwer Academic/Plenum Publishers, New York, pp. 311–338 Casteel, R.W. 1970. Core and column sampling. American Antiquity 35(4): 465-467.             1976. Comparison of column and whole unit samples for recovering fish remains. World  Archaeology 8(2): 192-196. Chatters, J.S. 1984. Dimensions of site structure: the archaeological record of two sites in  Okanagan County, Washington. U.S. Army Corps of Engineers, Seattle District, Seattle, WA.  Chisholm, B.S., Nelson, E.D., Schwarcz, H.P. 1982. Stable-Carbon Isotope Ratios as a Measure  of Marine versus Terrestrial Protein in Ancient Diets. Science, 216(4550): 131–1132. 50  Classen, C. 1991. Normative Thinking and Shell Bearing Sites. Journal of Archaeological  Methods and Theory 3: 249–298. Coupland, G., Bissell, C., King, S. 1993. Prehistoric subsistence and seasonality at Prince  Rupert Harbour: evidence from the McNichol Creek site. Canadian Journal of Archaeology 17: 59–73. Coupland, G., Stewart, K., Hall, T.J., 2002. Archaeological Investigations in the Prince Rupert  Area in 1999 and 2000: The Differentiated Regional Economy at Prehistoric Prince Rupert. Archaeological Branch, Ministry of Sustainable Resource Development, Victoria. Coupland, G., Stewart, K., Patton, A.K., Banahan, J., 2006. Archaeological Investigations in the Prince Rupert Area, 2002–2004: Permit 202–208, Final Report. Archaeology Branch, British Columbia Ministry of Sustainable Resource Development, Victoria. Coupland, G., Stewart, K., Patton, K. 2010. Do You Ever Get Tired of Salmon? Evidence  for Extreme Salmon Specialization at Prince Rupert Harbour, British Columbia. Journal of Anthropological Archaeology 29:189–207. Driver, J. Zooarchaeology in British Columbia. BC Studies 99: 77-105.  Drucker, P. 1965. Cultures of the North Pacific Coast. Harper and Row, New York.  Edinborough, K., Porcic, M., Martindale, A., Brown, T.J., Supernant, K., Ames, K.M.  Radiocarbon test foe demographic events in written and oral history. PNAS 114(47): 12436-12441. Eldridge, M., Parker, A., Mueller, C., Crockford, S. 2015. Archaeological Investigations at Ya asqalu'i? Kaien Siding, Prince Rupert Harbour. Report Prepared for CN Environmental Assessment. Manuscript on file, Victoria. Fladmark, K.R.1975. A Peleoecological model of Northwest Coast prehistory. National Museum  of Man Mercury Series Archaeological Survey of Canada Paper 43, Ottawa.  Garfield, V. 1939. Tsimshian clan and society. University of Washington Publications in  Anthropology, No. 7. Seattle.              1951. The Tsimshian and their neighbours. In: Garfield, V., Wingert, P. (Eds.), The  Tsimshian Indians and Their Arts, Douglas and McIntyre. Vancouver, pp. 3-70. Gray, B.S. 2008. Sampling methods in Northwest Coast household archaeology: a simulation  approach using faunal data from the Ozette site. Master’s Thesis, University of Victoria, Victoria.  Grayson, D.K. 1984. Quantitative Zooarchaeology. Academic Press, New York.  Halpin, M., Seguin, M. 1990. Tsimshian people: Southern Tsimshian, Coast Tsimshian, Nishga,  and Gitksan. In: Suttles, W. (Ed.), Handbook of North American Indians, Volume 7: Northwest Coast, Smithsonian Institute, Washington, pp. 267-284.  Inglis, R. 1977. The Coast Tsimshian seasonal subsistence pattern: the ethnographic and  archaeological record. Paper presented at the 30th Northwest Anthropological Conference, Victoria.  Letham, B., Martindale, A., McLaren, D., Brown, T., Ames, K.M., Archer, D.J.W., Marsden, S. 2015. Holocene settlement history of the Dundas Islands archipelago, northern British Columbia. BC Studies 187: 51–86. Letham, B., Martindale, A., Macdonald, R., Guiry, E., Jones, J., and Ames, K.M. 2016.  Postglacial Relative Sea-level History of the Prince Rupert Area, British Columbia, Canada. Quaternary Science Reviews 153: 156–191.  Letham, B., Martindale, A., Supernant, K., Brown, T.J., Cybulski, J., Ames, K. M. 2017.  Assessing the Scale and Place of Large Shell-Bearing Site Occupation in the Prince 51  Rupert Harbour Area, British Columbia. The Journal of Island and Coastal Archaeology 0: 1–35.  Letham, B., Martindale, A., Waber, N., Ames, K.M. 2018. Archaeological Survey of Dynamic  Coastal Landscapes and Paleoshorelines: Locating Early Holocene Sites in the Prince Rupert Harbour Area, British Columbia, Canada. Journal of Field Archaeology 43(3):181-199. Lyman, L.R. 2008. Quantitative Paleozoology. Cambridge Manuals in Archaeology. Cambridge. MacDonald, G.F. 1969. Preliminary culture sequence from the Coast Tsimshian area, British Columbia. Northwest Anthropological Research Notes 3: 240-254.              2006. Coast Tsimshian Pre-Contact Economics and Trade: An Archaeological an Ethno-Historical Reconstruction. Report submitted to Ratliff & Co.  MacDonald, G. F., Inglis, R. I. 1980. An Overview of the North Coast Prehistory Project (1966-1980). BC Studies 48: 37–63.  Marquardt, W.H. 2010. Shell Mounds in the Southeast: Middens, monuments, temple mounds,  rings or works? American Antiquity 75(3): 551–570. Marsden, S. 2001. Defending the mouth of the Skeena: Perspectives on Tsimshian Tlingit  relations. In: Cybulski, J., (Ed.), Perspectives on Northern Northwest Coast Prehistory. Mercury Series, Archaeology Survey of Canada Paper 160. Ottawa: Canadian Museum of Civilization, pp. 61–106.             2002. Adawx, Spanaxnox, and the Geopolitics of the Tsimshian. BC Studies 35: 101-135.  Marsden, S., Galois, R. 1995. The Tsimshian, the Hudson’s Bay Company, and the geopolitics of  the northwest coast fur trade, 1787–1840. Canadian Geographer 39(2): 169–183. Martindale, A., 2006. Methodological issues in the use of Tsimshian oral traditions (adawx) in  archaeology. Canadian Journal of Archaeology 30(2): 158–192.             2009. Entanglement and tinkering: structural history in the archaeology of the Northern  Tsimshian. Journal of Social Archaeology 9(1): 59–91.             2014. Archaeology taken to court: unraveling the epistemology of cultural tradition in the  context of Aboriginal title cases. In: Ferris, N., Harrison, R. (Eds.), From the Margins: The Archaeology of the Colonized and Its Contribution to Global Archaeological Theory. Oxford University Press, Oxford, pp. 397–422. Martindale, A., Marsden, S. 2003. Defining the Middle Period (3500 BP to 1500 BP) in Tsimshian History through a Comparison of Archaeological and Oral Records. BC Studies 138: 13-50.              2011. Analysis of Territorial Claims Made in “Lax Kw'alaams: Review of Historical and Ethnographic Sources” and “Metlakatla: Review of Historical and Ethnographic Sources”, Documents Prepared by the Aboriginal Research Division of the British Columbia Ministry of Attorney General Legal Services Branch. Martindale, A., Marsden, S., Patton, K., Ruggles, A., Letham, B., Supernant, K., Archer, D., McLaren, D., Ames, K.M. 2017a. The role of small villages in Tsimshian territory through the mid-late Holocene. Journal of Social Archaeology 17(3): 285–325. Martindale, A., Letham, B., Supernant, K., Brown, T.J., Edinborough, K., Duelks, J., Ames, K.M. 2017b. Urbanism in Northern Tsimshian archaeology. Hunter Gatherer Research 3(1): 133–163. Matson, R.G., Coupland, G., 1994. Prehistory of the Northwest Coast. Left Coast Press, Walnut Creek. 52  May, J., 1979. Archaeological Investigations at GbTn-19, Ridley Island, A Shell Midden in the Prince Rupert Area. Manuscript No. 1530. Archaeological Survey of Canada Archives, Ottawa. McKechnie, I., 2005. Column sampling and the archaeology of small fish in Ts'ishaa. In:  McMillan, A.D., St. Clair, D.E. (Eds.), Ts'ishaa: Archaeology and Ethnography of a Nuu-chah-nulth Origin Site in Barkley Sound. Archaeology Press, Burnaby, pp. 206–223.             2013. An Archaeology of Food and Settlement on the Northwest Coast. Doctorate  Dissertation, University of British Columbia, Vancouver.  McKechnie, I., Moss, M. 2016 Meta-analysis in Zooarchaeology Expands Perspectives on  Indigenous Fisheries of the Northwest Coast of North America. Journal of Archaeological Science: Reports 8: 470-485. McKechnie, I., Lepofsky, D., Moss, M.L., Butler, V.L., Orchard, T.J., Coupland, G., Foster, F., Caldwell, M., Lertzman, K., 2014. Archaeological data provide alternative hypotheses on Pacific herring (Clupea pallasii) distribution, abundance, and variability. PNAS 111(9): E807–E816. McLaren, D., Martindale, A., Fedje, D., Mackie, Q. 2011. Relict shorelines and shell middens of  the Dundas Island Archipelago. Canadian Journal of Archaeology 35: 86-116.  Meighan, C.W., Pendergast, D.M., Swartz, B.K., Wissler, M.D. 1958. Ecological interpretation  in archaeology: part 1. American Antiquity 24(1): 1-23.  Miller, J. 1997. Tsimshian Culture: A Light Through the Ages. University of Nebraska Press,  Lincoln. Mitchell, D.H., Donald, L. 1988. Archaeology and the study of Northwest Coast economies. In:  Isaac, B.L. (Ed.), Prehistoric Economies of the Northwest Coast, Research in Economic Anthropology, Suppl. 3. JAI Press, Greenwich, pp. 293–351. Monks, G.G. 1987. Prey as bait: The Deep Bay example. Canadian Journal of Archaeology  11(1): 119–142.             2006. The fauna from Ma’acoah (DfSi-5), Vancouver Island, British Columbia: an  interpretive summary. Canadian Journal of Archaeology 30: 272–305. Moss, M.L. 1989. Analysis of the vertebrate assemblage. In (Ed.) Davis, S.D., The Hidden Falls  Site, Baranof Island, Alaska. Alaska Anthropological Association Monograph Series, Brockport, pp. 93–130.             2007. The Killisnoo Picnic Ground (49-SIT-124) revisited: assessing archaeological    recovery of vertebrate faunal remains from Northwest Coast shell middens. Journal of Northwest Anthropology 41: 1–17. 2012. Understanding variability in Northwest Coast faunal assemblages: beyond economic intensification and cultural complexity. Journal of Island and Coastal Archaeology 7: 1–22.             2016. The nutritional value of pacific herring: an ancient cultural keystone species on the  Northwest Coast of North America. Journal of Archaeological Science: Reports 5: 649–655. Moss, M.L., Rodrigues, A., Speller, C.F., Yang, D.Y., 2015. The historical ecology of Pacific  herring: tracing Alaska native use of a forage fish. Journal of Archaeological Science: Reports 8: 504-512. Moss, M.L., Minor, R., Page-Botelho, K., 2017. Native American fisheries of the southern  53  Oregon coast: fine fraction needed to find forage fish. Journal of Californian Great basin Archaeology 37(2), 169–182. Orchard T.J. 2007. Otters and Urchins: Continuity and Change in Haida Economy during the  Late Holocene and Maritime Fur Trade Periods. Doctorate Dissertation, University of Toronto, Toronto. Patton, K. 2011. Reconstructing Houses: Early Village Social Organization in Prince Rupert  Harbour, British Columbia. Doctorate dissertation, University of Toronto, Toronto. Patton, K., Martindale, A., Orchard, T.J., Vanier, S., Coupland, G. 2019. Finding eulachon: The  use and cultural importance of Thaleichthys pacificus on the norther Northwest Coast of North America. Journal of Archaeological Science: Reports 23: 687-699. Pojar, J., Meidinger, D., Klinka, K. 1991. Concepts. In Ecosystems of British Columbia. Special  Report Series of BC Ministry of Forestry 6: 9-38. Port Simpson Curriculum Committee. n.d. Port Simpson Foods: A Curriculum Development  Project. The People of Port Simpson and School District No. 52 (Prince Rupert).  Stein, J. 1992. Deciphering a shell midden. Academic Press, San Diego. Stewart, K.M., Wigen, R.J. 2003. Screen size and the need for reinterpretation: a case study  from the Northwest Coast. Florida Museum of Natural History Bulletin 44(1): 27–34. Stewart, K., Coupland, G., Naughton, D. 2003. Effects of screen size on fish element recovery in  northern Northwest Coast middens. In: Stewart, K., Stewart, F. (Eds.), Transitions in Zooarchaeology. Canadian Museum of Nature, Ottawa, pp. 53–63. Stewart, K.M., Stewart, F.L., Coupland, G. 2009. Boardwalk, northern Northwest Coast, Canada:  a new face to an old site. Canadian Journal of Archaeology 33: 205–233.  Supernant, K., Cookson, C. 2014. Mapping social cohesion in Prince Rupert Harbour, BC: A  social application of GIS to the archaeology of the Northwest Coast. Canadian Journal of Archaeology 38: 179-210.  Wheeler, A. 1978. Problems with identification and interpretation of archaeological fish remains.  In: Brothwell, D.R., Thomas, K.D., Clutton-Brock, J. (Eds.), Research Problems in zooarchaeology. London: Institute of Archaeology, Occasional Publication No. 3, pp. 69-75.  Wigen, R.J., Stucki, B.R. 1988. Taphonomy and stratigraphy in the interpretation of economic  patterns at Hoko River Rockshelter. Research in Economic Anthropology, Supplement 3 87-146. Wissler, F.C., 1938 [1917]. The Relation of Nature to Man in Aboriginal America. McMurtrie,  New York.    54  Appendices Appendix 1: The Dataset  *Samples at start and end with no fauna cut from volume assuming non-cultural material.  TotalSalmonHerringSmeltOther FishSalmonHerringSmelt/Other FishSalmonHerringSmelt/Other FishGbTo-23 Village Yes shell terrace column 41 36.67 113 72 41 0 0 63.72 36.28 0 1.96 1.12 0GbTo-28 Village No back midden column 15 15 641 503 136 Not ID’d 2 78.47 21.22 0.31 33.53 9.07 0.13GbTo-28 Village No house floor column 6 6 101 65 21 Not ID’d 15 64.36 20.79 14.85 10.83 3.5 2.5GbTo-31 Village Yes shell terrace column 10 10 114 106 8 Not ID’d 0 92.98 7.02 0 10.6 0.8 0GbTo-31 Village Yes shell terrace column 9 9 73 71 0 Not ID’d 2 97.26 0 2.74 7.89 0 0.22GbTo-31 Village Yes shell terrace column 9 9 187 179 4 Not ID’d 4 95.72 2.14 2.14 19.89 0.44 0.44GbTo-31 Village Yes shell terrace column 19 19 323 302 9 12 0 93.5 2.79 3.72 15.89 0.47 0.63GbTo-46 Village No back midden column 5 5 342 302 38 Not ID’d 2 88.3 11.11 0.58 60.4 7.6 0.4GbTo-46 Village No house floor column 10 10 77 53 8 Not ID’d 16 68.83 10.39 20.78 5.3 0.8 1.6GbTo-46 Village No house floor column 10 10 190 106 30 53 1 55.79 15.79 28.42 10.6 3 5.4GcTo-06 Village Yes back midden column 3 3 243 204 3 Not ID’d 36 83.95 1.23 14.81 68 1 12GcTo-06 Village Yes shell terrace auger 48* 18.25 160 127 17 10 6 79.38 10.63 10 6.96 0.93 0.88GbTo-04 Village Yes shell terrace auger 19 16.38 445 397 32 9 7 89.21 7.19 3.6 24.24 1.95 0.98GbTo-34 Village Yes shell terrace auger 18* 13.8 229 161 66 1 1 70.31 28.82 0.87 11.67 4.78 0.14GbTo-34 Village Yes shell terrace column 10 10.1 169 106 63 0 0 62.72 37.28 0 10.5 6.24 0GcTo-01 Village Yes shell terrace auger 19* 15.85 136 114 14 6 2 83.82 10.29 5.88 7.19 0.88 0.5GcTo-39 Village No shell terrace auger 12* 10.55 248 200 43 3 2 80.65 17.34 2.02 18.96 4.08 0.47GbTo-24 Village No shell terrace auger 20* 7.375 31 5 0 8 18 16.13 0 83.87 0.68 0 3.53GbTo-24 Village No shell terrace column 30 13.75 116 26 3 47 40 22.41 2.59 75 1.89 0.22 6.33GbTo-77 Village No shell terrace column 1 1 8 5 2 1 0 62.5 25 12.5 5 2 1GbTo-77 Village No house Floor column 2 2 104 4 45 55 0 3.85 43.27 52.88 2 22.5 27.5GbTo-77 Village No shell terrace column 9 9 277 72 107 89 9 25.99 38.63 35.38 8 11.89 10.89GbTo-77 Village No back midden column 7 7 68 28 14 25 1 41.18 20.59 38.24 4 2 3.71GcTo-27 Village No shell terrace auger 24* 17.38 102 30 69 0 3 29.41 67.65 2.94 1.73 3.97 0.17GcTo-28 Village No shell terrace auger 21* 16.25 179 29 142 2 6 16.2 79.33 4.47 1.78 8.74 0.49GcTo-51 Village No shell terrace auger 15* 11.53 106 82 19 1 4 77.36 17.92 4.72 7.11 1.65 0.43GcTo-52 Village No shell terrace auger 13* 9.7 68 19 45 3 1 27.94 66.18 5.88 1.96 4.64 0.41Relative Abundance (%) DensityTotal matrix volume (l)No. of bagsNISPInner HarbourMetlakatla PassChatham SoundRegionBorden no.Site typeRe-occupied?Sampling locationSample method55        TotalSalmonHerringSmeltOther FishSalmonHerringSmelt/Other FishSalmonHerringSmelt/Other FishGcTq-01 Village No unknown auger 2 48.8 382 187 51 126 18 48.95 13.35 37.7 3.83 1.05 2.95GcTq-04 Village No unknown auger 2 23.55 64 36 23 0 5 56.25 35.94 7.81 1.53 0.98 0.21GcTq-05 Village No unknown auger 4 80.6 896 703 103 82 8 78.46 11.5 10.04 8.72 1.28 1.12GcTq-06 Camp No unknown auger 2 13.5 82 12 13 54 3 14.63 15.85 69.51 0.89 0.96 4.22GcTq-07 Village No unknown auger 1 13.05 204 143 24 33 4 70.1 11.76 18.14 10.96 1.84 2.84GcTq-08 Camp No unknown auger 1 8.25 49 39 4 5 1 79.59 8.16 12.24 4.73 0.48 0.73GcTq-09 Camp No unknown auger 1 9.8 33 7 22 4 0 21.21 66.67 12.12 0.71 2.24 0.41GcTq-10 Village No unknown auger 1 13.35 36 18 5 5 8 50 13.89 36.11 1.35 0.37 0.97GcTq-11 Village No unknown auger 1 9.05 11 0 8 0 3 0 72.73 27.27 0 0.88 0.33GcTq-12 Camp No unknown auger 1 10.65 27 23 0 4 0 85.19 0 14.81 2.16 0 0.38GcTq-13 Camp No unknown auger 1 4.25 35 16 2 17 0 45.71 5.71 48.57 3.76 0.47 4GcTr-05 Village No unknown auger 4 50.8 671 552 46 72 1 82.27 6.86 10.88 10.87 0.91 1.44GcTr-06 Camp No unknown auger 2 17 14 7 4 0 3 50 28.57 21.43 0.41 0.24 0.18GcTr-08 Village No unknown auger 2 44.7 347 181 89 53 24 52.16 25.65 22.19 4.05 1.99 1.72GcTr-09 Camp No unknown auger 1 16.5 75 59 14 2 0 78.67 18.67 2.67 3.58 0.85 0.12GcTr-10 Village No unknown auger 4 53.55 500 415 25 59 1 83 5 12 7.75 0.47 1.12GdTq-01 Village No unknown auger 3 24.45 379 332 17 29 1 87.6 4.49 7.92 13.58 0.7 1.23GdTq-03 Camp No unknown auger 1 25.5 18 8 7 0 3 44.44 38.89 16.67 0.31 0.27 0.12No. of bagsTotal matrix volume (l)NISP Relative Abundance (%) DensityDundas Island GroupRegionBorden no.Site typeRe-occupied?Sampling locationSample method56  Appendix 2: Statistical Results Relative Abundance of Fish Taxa by Sampling Method  Relative Abundance Salmon Herring Smelt / Other Fish   Auger Column  Auger Column Auger Column Descriptive Statistics N 21 17 21 17 21 17 Min 0 3.85 0 0 0.8733624 0 Max 89.21 97.26 79.33 43.27 83.87097 75 Sum 1193.83 1101.53 522.37 296.12 383.8199 302.3563 Mean 56.84905 64.79588 24.87476 17.41882 18.27714 17.78566 Std. error 6.251469 6.756716 5.390271 3.532714 4.846728 5.292872 Variance 820.6981 776.1046 610.1555 212.1612 493.3063 476.2464 Stand. dev 28.64783 27.85865 24.70133 14.56575 22.2105 21.82307 Median 70.1 64.36 13.89 15.79 10 12.5 25 prcntil 28.675 48.485 8.74 2.69 4.593127 0.4484039 75 prcntil 81.46 90.64 32.38 30.64 24.73146 31.90006 Skewness -0.6442441 -0.7942182 1.331536 0.474726 2.023508 1.419185 Kurtosis -1.020227 -0.1161344 0.3772693 -1.107758 3.727431 1.622654 Geom. mean 0 53.8396 0 0 9.980348 0 Coeff. var 50.3928 42.99448 99.30276 83.62076 121.5207 122.7003 Normality               N 21 17 21 17 21 17 Shapiro-Wilk W 0.8781 0.9132 0.7765 0.9044 0.7221 0.8157   p(normal) 0.01345 0.1134 0.0002936 0.08048 0.00005262 0.003354               Tests for equal medians               N 21 17 21 17 21 17 Mean rank 9.9211 9.5789 11.355 8.1447 11.342 8.1579 Mann-Whitn U 146 156.5 157 p (same med.) 0.3475 0.52789 0.53744 Z 0.93945 0.63123 0.61668 Monte Carlo permutation (p) 0.3613 0.5252 0.5256                 57  Density of Fish Taxa by Sampling Method  Density Salmon Herring Smelt / Other Fish   Auger Column  Auger Column Auger Column Descriptive Statistics N 21 17 21 17 21 17 Min 0 1.89 0 0 0.1449275 0 Max 24.24 68 8.74 22.5 4.222222 27.5 Sum 145.81 276.28 43.05 72.65 26.16356 72.76203 Mean 6.943333 16.25176 2.05 4.273529 1.245884 4.280119 Std. error 1.403074 4.779514 0.4564449 1.421164 0.255429 1.71607 Variance 41.34092 388.3439 4.37518 34.335 1.370124 50.06323 Stand. dev 6.429691 19.70644 2.091693 5.859607 1.170523 7.075537 Median 6.96 10.5 1.05 2 0.9737828 1 25 prcntil 1.63 4.5 0.88 0.635 0.4231053 0.1777778 75 prcntil 10.915 17.89 2.98 6.92 1.579801 5.863636 Skewness 1.215408 1.986468 1.948559 2.18961 1.390556 2.535761 Kurtosis 1.335607 3.137203 4.18525 5.302637 1.057956 7.191027 Geom. mean 0 9.270041 0 0 0.8178153 0 Coeff. var 92.60236 121.2572 102.0338 137.114 93.95121 165.3117 Normality               N 21 17 21 17 21 17 Shapiro-Wilk W 0.8773 0.6948 0.7647 0.7217 0.8173 0.6561   p(normal) 0.01304 0.0001009 0.0001987 0.0002056 0.00122 0.00003834               Tests for equal medians               N 21 17 21 17 21 17 Mean rank 9.1184 10.382 10.158 9.3421 10.421 9.0789 Mann-Whitn U 115.5 155 165 p (same med.) 0.066496 0.49941 0.70266 Z 1.8351 0.67542 0.38174 Monte Carlo permutation (p) 0.0646 0.4951 0.7101                 58  Relative Abundance of Fish Taxa by Inter-Site Location  Relative Abundance Salmon Herring Smelt / Other Fish   Inner Harbour Metlakatla Pass Chatham Sound Dundas Island G.  Inner Harbour Metlakatla Pass Chatham Sound Dundas Island G.  Inner Harbour Metlakatla Pass Chatham Sound Dundas Island G.  Descriptive Statistics N 12 5 10 11 12 5 10 11 12 5 10 11 Min 55.79 62.72 3.85 0 0 7.19 0 4.49 0 0 2.941176 7.8125 Max 97.26 89.21 77.36 87.6 36.28 37.28 79.33 72.73 28.42105 5.882353 83.87097 69.5122 Sum 962.26 386.71 322.97 623.42 139.39 100.92 361.16 217.02 98.35731 12.36735 315.8797 259.5718 Mean 80.18833 77.342 32.297 56.67455 11.61583 20.184 36.116 19.72909 8.196443 2.47347 31.58797 23.59744 Std. error 4.085349 4.780213 7.083917 8.562029 3.060603 5.664262 8.785886 5.986953 2.752025 1.04324 9.680143 5.627734 Variance 200.2809 114.2522 501.8188 806.3917 112.4075 160.4193 771.918 394.2796 90.88372 5.441751 937.0516 348.3853 Stand. dev 14.15206 10.68888 22.40131 28.39704 10.60224 12.66568 27.78341 19.85648 9.533295 2.332756 30.6113 18.66508 Median 81.665 80.65 26.965 56.25 10.51 17.34 31.815 13.35 3.227448 2.016129 23.93953 18.13725 25 prcntil 65.4775 66.515 16.1825 48.95 2.3025 8.74 14.0875 6.86 0.3802082 0.4366812 4.655054 10.04464 75 prcntil 93.37 86.515 46.51 82.27 19.54 33.05 66.5475 25.65 14.84232 4.738929 58.41346 36.11111 Skewness -0.4105031 -0.5098371 1.075144 -0.9600126 1.135843 0.5035064 0.2828751 2.230977 1.051109 0.7122684 0.716335 1.645498 Kurtosis -1.231766 -1.308702 0.6562716 0.2117139 1.333787 -1.727146 -1.24655 5.387501 0.07118157 -0.3744475 -0.9745696 2.947836 Geom. mean 78.97325 76.72868 25.01303 0 0 16.89945 0 14.01709 0 0 17.0298 18.44803 Coeff. var 17.64853 13.82028 69.36035 50.10545 91.27402 62.75107 76.92826 100.6457 116.3102 94.31108 96.90808 79.09792 Normality                           N 12 5 10 11 12 5 10 11 12 5 10 11 Shapiro-Wilk W 0.9217 0.9528 0.899 0.8872 0.8986 0.9333 0.9349 0.7257 0.8352 0.9581 0.8536 0.8178   p(normal) 0.3006 0.7571 0.2137 0.1283 0.1523 0.6187 0.4977 0.001004 0.02423 0.7945 0.06417 0.01613                              59  Relative Abundance Salmon   Inner Harbour Metlakatla Pass Inner Harbour Chatham Sound Inner Harbour Dundas Island G.  Metlakatla Pass Chatham Sound Metlakatla Pass Dundas Island G.  Chatham Sound Dundas Island G.    N 12 5 12 10 12 11 5 10 5 11 10 11 Tests for equal variance                           Variance 200.28 114.25 200.28 501.82 200.28 806.39 114.25 501.82 114.25 806.39 501.82 806.39 F 1.753 2.5056 4.0263 4.3922 7.058 1.6069 p (same var.) 0.62067 0.15323 0.031246 0.1679 0.074678 0.48804 Critical F value (p=0.05): 8.7935 3.5879 3.5257 8.9047 8.8439 3.9639 Monte Carlo permutation (p) 0.3657 0.0251 0.1796 0.0199 0.0953 0.2543 Exact permutation (p) 0.36846 0.025278 0.17771 0.018648 0.093636 0.24526                           Tests for equal means Mean 80.188 77.342 80.188 32.297 80.188 56.675 77.342 32.297 77.342 56.675 32.297 56.675 95% conf. max 89.18 90.614 89.18 48.322 89.18 75.752 90.614 48.322 90.614 75.752 48.322 75.752 95% conf. min 71.197 64.07 71.197 16.272 71.197 37.597 64.07 16.272 64.07 37.597 16.272 37.597 Difference between means 2.8463 47.891 23.514 45.045 20.667 24.378 95% conf. interval (parametric): (-12.262 17.955) (31.52 64.262) (4.3195 42.708) (21.901 68.189) (-7.8717 49.207) (0.8457 47.909) 95% conf. interval (bootstrap): (-8.6343 13.983) (33.473 64.051) (5.0222 40.423) (30.242 61.474) (2.1533 38.07) (4.6393 45.895) t 0.40154 6.1022 2.5476 4.2047 1.5532 2.1682 p (same mean) 0.69368 0.000005792 0.018741 0.0010306 0.14268 0.043044 Critical t value (p=0.05) 2.1314 2.086 2.0796 2.1604 2.1448 2.093 Uneq. var. t  0.45265 5.8565 2.4786 5.271 2.1076 2.1937 p (same mean) 0.66043 0.000034676 0.026138 0.00015134 0.053789 0.041137 Monte Carlo permutation (p) 0.696 0.0002 0.0157 0.0013 0.1338 0.0429 Exact permutation (p) 0.68988 0.000029382 0.01457 0.001665 0.13484 0.044852    60  Relative Abundance Herring   Inner Harbour Metlakatla Pass Inner Harbour Chatham Sound Inner Harbour Dundas Island G.  Metlakatla Pass Chatham Sound Metlakatla Pass Dundas Island G.  Chatham Sound Dundas Island G.    N 12 5 12 10 12 11 5 10 5 11 10 11 Tests for equal medians                           Mean rank         5.3478 6.6522     3 5.5 6.1905 4.8095 Mann-Whitn U     45   22 35 p (same med.)     0.20706   0.57109 0.16971 Z     1.2617   0.56644 1.3731 Monte Carlo permutation (p)     0.2185   0.5828 0.1697 Exact permutation     0.2115   0.58333 0.17335                           Tests for equal variance                           Variance 112.41 160.42 112.41 771.92     160.42 771.92         F 1.4271 6.8671   4.8119     p (same var.) 0.57782 0.0041191   0.14471     Critical F value (p=0.05): 4.2751 3.5879   8.9047     Monte Carlo permutation (p) 0.7337 0.0248   0.0938     Exact permutation (p) 0.73756 0.024893   0.092241                               Tests for equal means Mean 11.616 20.184 11.616 36.116     20.184 36.116     36.116   95% conf. max 18.352 35.911 18.352 55.991     35.911 55.991     55.991   95% conf. min 4.8795 4.4575 4.8795 16.241     4.4575 16.241     16.241   Difference between means 8.5682 24.5   15.932     95% conf. interval (parametric): (-4.1272 21.264) (6.4331 42.567)   (-12.658 44.522)     95% conf. interval (bootstrap): (-2.8867 19.771) (7.1558 41.738)   (-3.374 34.715)     t 1.4385 2.8287   1.2039     p (same mean) 0.17082 0.010376   0.25009     Critical t value (p=0.05) 2.1314 2.086   2.1604     Uneq. var. t  1.3308 2.6334   1.5241     p (same mean) 0.22819 0.022985   0.15146     Monte Carlo permutation (p) 0.1651 0.0095   0.2492     Exact permutation (p) 0.1713 0.0092029   0.25341      61  Relative Abundance Smelt / Other Fish   Inner Harbour Metlakatla Pass Inner Harbour Chatham Sound Inner Harbour Dundas Island G.  Metlakatla Pass Chatham Sound Metlakatla Pass Dundas Island G.  Chatham Sound Dundas Island G.    N 12 5 12 10 12 11 5 10 5 11 10 11 Tests for equal medians                           Mean rank 6.8235 2.1765 4.6818 6.8182 4.5217 7.4783     0.9375 7.5625 5.2381 5.7619 Mann-Whitn U 22 25 26   0 55 p (same med.) 0.42806 0.022876 0.01503   0.0022224 0.97191 Z 0.79251 2.2755 2.4317   3.0588 0.035209 Monte Carlo permutation (p) 0.4435 0.0206 0.0113   0.0007 1 Exact permutation 0.44037 0.020104 0.012623   0.00045788 1                           Tests for equal variance                           Variance             5.4418 937.05         F       172.2     p (same var.)       0.00016305     Critical F value (p=0.05):       8.9047     Monte Carlo permutation (p)       0.031     Exact permutation (p)       0.031635                               Tests for equal means Mean             2.4735 31.588         95% conf. max             5.37 53.486         95% conf. min             -0.42303 9.6901         Difference between means       29.115     95% conf. interval (parametric):       (-1.0627 59.292)     95% conf. interval (bootstrap):       (10.531 46.151)     t       2.0843     p (same mean)       0.057426     Critical t value (p=0.05)       2.1604     Uneq. var. t        2.9903     p (same mean)       0.01482     Monte Carlo permutation (p)       0.0445     Exact permutation (p)       0.041958      62  Relative Abundance of Fish Taxa by Re-occupation Status Relative Abundance Salmon Herring Smelt / Other Fish   Re-occupied Not re-occupied  Re-occupied Not re-occupied  Re-occupied Not re-occupied  Descriptive Statistics N 7 5 7 5 7 5 Min 63.72 55.79 0 10.39 0 0.31 Max 97.26 88.3 36.28 21.22 14.81 28.42 Sum 606.51 355.75 60.09 79.3 33.41 64.95 Mean 86.64 71.15 8.58 15.86 4.77 12.99 Std. error 4.55 5.64 4.82 2.3 2.1 5.55 Variance 144.63 158.78 162.76 26.38 30.99 154.22 Stand. dev 12.03 12.6 12.76 5.14 5.57 12.42 Median 92.98 68.83 2.79 15.79 2.74 14.85 25 prcntil 79.38 60.08 1.23 10.75 0 0.45 75 prcntil 95.72 83.39 10.63 21.01 10 24.6 Skewness -1.33 0.31 2.22 0.01 1.23 0.05 Kurtosis 1.34 -0.7 5.16 -2.93 0.44 -2.22 Geom. mean 85.85 70.26 0 15.17 0 4.37 Coeff. var 13.88 17.71 148.62 32.38 116.63 95.6 Normality               N 7 5 7 5 7 5 Shapiro-Wilk W 0.8525 0.9851 0.6999 0.8615 0.8383 0.8995   p(normal) 0.1295 0.9599 0.003696 0.0279 0.09574 0.4074               Tests for equal medians               N     7 5     Mean rank     2.8333 3.6667     Mann-Whitn U   6   p (same med.)   0.074035   Z   1.7864   Monte Carlo permutation (p)   0.0724   Exact permutation     0.073232                       63  Relative Abundance Salmon Smelt / Other Fish   Re-occupied Not re-occupied  Re-occupied Not re-occupied    N 7 5 7 5 Tests for equal variance           Variance 144.63 158.78 30.986 154.22 F 1.0979 4.9771 p (same var.) 0.87309 0.082104 Critical F value (p=0.05): 6.2272 6.2272 Monte Carlo permutation (p) 0.8949 0.0986 Exact permutation (p) 0.8952 0.09596           Tests for equal means Mean 86.644 71.15 4.7727 12.99 95% conf. max 97.767 86.796 9.9208 28.409 95% conf. min 75.522 55.504 -0.37545 -2.4299 Difference between means 15.494 8.217 95% conf. interval (parametric): (-0.49979 31.488) (-3.4725 19.907) 95% conf. interval (bootstrap): (3.2766 28.59) (-1.9155 18.692) t 2.1585 1.5662 p (same mean) 0.056251 0.14836 Critical t value (p=0.05) 2.2281 2.2281 Uneq. var. t  2.1401 1.3836 p (same mean) 0.062776 0.22335 Monte Carlo permutation (p) 0.0575 0.1462 Exact permutation (p) 0.058081 0.14268                64  Density of Fish Taxa by Intra-Site Location  Density Salmon Herring Smelt / Other Fish   Back Midden House Floor Shell Terrace Back Midden House Floor Shell Terrace Back Midden House Floor Shell Terrace Descriptive Statistics N 3 3 6 3 3 6 3 3 6 Min 33.53 5.3 1.96 1 0.8 0 0.1333333 1.6 0 Max 68 10.83 19.89 9.07 3.5 1.12 12 5.4 0.8767123 Sum 161.93 26.73 63.19 17.67 7.3 3.76 12.53333 9.5 2.174958 Mean 53.97667 8.91 10.53167 5.89 2.433333 0.6266667 4.177778 3.166667 0.362493 Std. error 10.45609 1.806221 2.642622 2.481552 0.8293237 0.1650791 3.911869 1.146492 0.1444557 Variance 327.9896 9.7873 41.9007 18.4743 2.063333 0.1635067 45.90815 3.943333 0.1252046 Stand. dev 18.11048 3.128466 6.473075 4.298174 1.436431 0.4043596 6.775555 1.985783 0.3538426 Median 60.4 10.6 9.25 7.6 3 0.64 0.40 2.50 0.33 25 prcntil 33.53 5.3 5.71 1 0.8 0.33 0.1333333 1.6 0 75 prcntil 68 10.83 16.89 9.07 3.5 0.9775 12 5.4 0.6928623 Skewness -1.395265 -1.721525 0.3093145 -1.506928 -1.498959 -0.4685639 1.729033 1.340467 0.3926228 Kurtosis -2.333333 -2.333333 -0.6265646 -2.333333 -2.333333 -0.3681455 -2.333333 -2.333333 -1.332452 Geom. mean 51.64082 8.473641 8.436607 4.100218 2.032793 0 0.8617739 2.784953 0 Coeff. var 33.55243 35.11185 61.46297 72.97409 59.0314 64.52546 162.1808 62.70893 97.61365 Normality                     N 3 3 6 3 3 6 3 3 6 Shapiro-Wilk W 0.9057 0.7811 0.9726 0.8813 0.8833 0.9619 0.7668 0.9155 0.9241   p(normal) 0.4037 0.07022 0.9093 0.3282 0.3341 0.8345 0.03758 0.4366 0.5355                      65   Density Salmon   Back Midden House Floor Back Midden Shell Terrace House Floor Shell Terrace   N 3 3 3 6 3 6 Tests for equal variance               Variance 327.99 9.7873 327.99 41.901 9.7873 41.901 F 33.512 7.8278 4.2811 p (same var.) 0.057951 0.057658 0.40026 Critical F value (p=0.05): 39 8.4336 39.298 Monte Carlo permutation (p) 0.0963 0.2844 0.3253 Exact permutation (p) 0.1 0.29762 0.33333               Tests for equal means Mean 53.977 8.91 53.977 10.532 8.91 10.532 95% conf. max 98.966 16.682 98.966 17.325 16.682 17.325 95% conf. min 8.9877 1.1385 8.9877 3.7386 1.1385 3.7386 Difference between means 45.067 43.445 1.6217 95% conf. interval (parametric): (15.606 74.527) (24.853 62.037) (-7.9432 11.187) 95% conf. interval (bootstrap): (31.043 65.437) (28.812 62.558) (-3.955 6.9067) t 4.2472 5.5255 0.4009 p (same mean) 0.013188 0.00088238 0.70045 Critical t value (p=0.05) 2.7764 2.3646 2.3646 Uneq. var. t  4.2472 4.0283 0.50663 p (same mean) 0.04622 0.045869 0.62806 Monte Carlo permutation (p) 0.0664 0.0107 0.7125 Exact permutation (p) 0.05 0.011905 0.71429           66  Density Herring   Back Midden House Floor Back Midden Shell Terrace House Floor Shell Terrace   N 3 3 3 6 3 6 Tests for equal variance               Variance 18.474 2.0633 18.474 0.16351 2.0633 0.16351 F 8.9536 112.99 12.619 p (same var.) 0.20093 0.00013789 0.022236 Critical F value (p=0.05): 39 8.4336 8.4336 Monte Carlo permutation (p) 0.2027 0.2891 0.3079 Exact permutation (p) 0.2 0.28571 0.32143               Tests for equal means Mean 5.89 2.4333 5.89 0.62667 2.4333 0.62667 95% conf. max 16.567 6.0016 16.567 1.051 6.0016 1.051 95% conf. min -4.7873 -1.135 -4.7873 0.20232 -1.135 0.20232 Difference between means 3.4567 5.2633 1.8067 95% conf. interval (parametric): (-3.8078 10.721) (1.3797 9.147) (0.40147 3.2119) 95% conf. interval (bootstrap): (-0.13333 8.18) (2.15 10.075) (0.69167 3.39) t 1.3211 3.2046 3.0401 p (same mean) 0.25697 0.01497 0.018843 Critical t value (p=0.05) 2.7764 2.3646 2.3646 Uneq. var. t  1.3211 2.1163 2.1366 p (same mean) 0.29665 0.16744 0.1566 Monte Carlo permutation (p) 0.3019 0.0236 0.05 Exact permutation (p) 0.25 0.02381 0.047619            67   Density Herring   Back Midden House Floor Back Midden Shell Terrace House Floor Shell Terrace   N 3 3 3 6 3 6 Tests for equal variance               Variance 18.474 2.0633 18.474 0.16351 2.0633 0.16351 F 8.9536 112.99 12.619 p (same var.) 0.20093 0.00013789 0.022236 Critical F value (p=0.05): 39 8.4336 8.4336 Monte Carlo permutation (p) 0.2027 0.2891 0.3079 Exact permutation (p) 0.2 0.28571 0.32143               Tests for equal means Mean 5.89 2.4333 5.89 0.62667 2.4333 0.62667 95% conf. max 16.567 6.0016 16.567 1.051 6.0016 1.051 95% conf. min -4.7873 -1.135 -4.7873 0.20232 -1.135 0.20232 Difference between means 3.4567 5.2633 1.8067 95% conf. interval (parametric): (-3.8078 10.721) (1.3797 9.147) (0.40147 3.2119) 95% conf. interval (bootstrap): (-0.13333 8.18) (2.15 10.075) (0.69167 3.39) t 1.3211 3.2046 3.0401 p (same mean) 0.25697 0.01497 0.018843 Critical t value (p=0.05) 2.7764 2.3646 2.3646 Uneq. var. t  1.3211 2.1163 2.1366 p (same mean) 0.29665 0.16744 0.1566 Monte Carlo permutation (p) 0.3019 0.0236 0.05 Exact permutation (p) 0.25 0.02381 0.047619            68   Density Smelt / Other Fish   Back Midden House Floor Back Midden Shell Terrace House Floor Shell Terrace   N 3 3 3 6 3 6 Tests for equal medians               Mean rank 1.5 2 1.8889 3.1111     Mann-Whitn U 3 7     p (same med.) 0.66252 0.69733     Z 0.43644 0.38892     Monte Carlo permutation (p) 0.6981 0.6596     Exact permutation 0.7 0.65476                   Tests for equal variance               Variance         3.9433 0.1252 F     31.495 p (same var.)     0.0029332 Critical F value (p=0.05):     8.4336 Monte Carlo permutation (p)     0.151 Exact permutation (p)     0.15476               Tests for equal means Mean         3.1667 0.36249 95% conf. max         8.0996 0.73383 95% conf. min         -1.7663 -0.00884 Difference between means     2.8042 95% conf. interval (parametric):     (0.96034 4.648) 95% conf. interval (bootstrap):     (0.6411 4.33) t     3.5961 p (same mean)     0.008785 Critical t value (p=0.05)     2.3646 Uneq. var. t      2.4267 p (same mean)     0.13214 Monte Carlo permutation (p)     0.0109 Exact permutation (p)     0.011905    69  Relative Abundance of Fish Taxa by Site Type  Relative Abundance Salmon Herring Smelt / Other Fish   Villages Camps Villages Camps Villages Camps Descriptive Statistics N 11 7 11 7 11 7 Min 0 21.21 4.49 0 7.8125 2.666667 Max 87.6 85.19 72.73 66.67 69.5122 48.57143 Sum 623.42 404.81 217.02 166.67 259.5718 128.5143 Mean 56.67455 57.83 19.72909 23.81 23.59744 18.35918 Std. error 8.562029 8.977343 5.986953 8.800584 5.627734 5.477617 Variance 806.3917 564.1488 394.2796 542.152 348.3853 210.03 Stand. dev 28.39704 23.75182 19.85648 23.28416 18.66508 14.49241 Median 56.25 50 13.35 18.67 18.14 14.81 25 prcntil 48.95 44.44 6.86 5.71 10.04464 12.12121 75 prcntil 82.27 79.59 25.65 38.89 36.11111 21.42857 Skewness -0.9600126 -0.2426458 2.230977 1.088211 1.645498 1.780065 Kurtosis 0.2117139 -1.251859 5.387501 0.8175834 2.947836 4.089092 Geom. mean 0 52.83327 14.01709 0 18.44803 13.92877 Coeff. var 50.10545 41.07179 100.6457 97.7915 79.09792 78.93823 Normality               N 11 7 11 7 11 7 Shapiro-Wilk W 0.8872 0.8992 0.7257 0.9123 0.8178 0.815   p(normal) 0.1283 0.326 0.001004 0.4123 0.01613 0.05748               Tests for equal medians               N     11 7 11 7 Mean rank     5.6111 3.8889 5.9444 3.5556 Mann-Whitn U   35 36 p (same med.)   0.78585 0.85626 Z   0.2717 0.18113 Monte Carlo permutation (p)   0.7805 0.8632 Exact permutation     0.79142 0.86011                  70  Relative Abundance Salmon   Villages Camps   N 11 7 Tests for equal variance       Variance 806.39 564.15 F 1.4294 p (same med.) 0.68618 Critical F value (p=0.05): 5.4613 Monte Carlo permutation (p) 0.6486 Exact permutation (p) 0.65334       Tests for equal means Mean 56.675 57.83 95% conf. max 75.752 79.797 95% conf. min 37.597 35.863 Difference between means 1.1555 95% conf. interval (parametric): (-26.262 28.573) 95% conf. interval (bootstrap): (-21.472 22.713) t 0.089339 p (same mean) 0.92992 Critical t value (p=0.05) 2.1199 Uneq. var. t  0.093139 p (same mean) 0.92706 Monte Carlo permutation (p) 0.9281 Exact permutation (p) 0.93056       71  Density of Fish Taxa by Site Type Density Salmon Herring Smelt / Other Fish   Villages Camps Villages Camps Villages Camps Descriptive Statistics N 11 7 11 7 11 7 Min 0 0.31 0.37 0 0.2123142 0.1176471 Max 13.58 4.73 1.99 2.24 4.222222 4 Sum 63.53 15.66 11.43 4.55 18.14955 5.926353 Mean 5.775455 2.237143 1.039091 0.65 1.649959 0.8466218 Std. error 1.439407 0.6853422 0.1520195 0.2830278 0.3664378 0.5318182 Variance 22.79081 3.287857 0.2542091 0.5607333 1.477043 1.979814 Stand. dev 4.773972 1.813245 0.5041915 0.7488213 1.215337 1.407059 Median 4.05 2.16 0.96 0.47 1.23 0.38 25 prcntil 1.35 0.41 0.7 0.24 0.9737828 0.1212121 75 prcntil 10.87 3.76 1.28 0.85 2.835249 0.7272727 Skewness 0.3316704 0.1810411 0.8180426 2.010818 1.016426 2.520859 Kurtosis -1.480471 -2.048932 0.2362443 4.428987 0.5654824 6.473212 Geom. mean 0 1.433031 0.9302291 0 1.223533 0.3789534 Coeff. var 82.65967 81.05182 48.52237 115.2033 73.65859 166.1968 Normality               N 11 7 11 7 11 7 Shapiro-Wilk W 0.9097 0.8833 0.9109 0.7707 0.8988 0.5847   p(normal) 0.2417 0.2414 0.2501 0.02076 0.1787 0.0001795               Tests for equal medians               N     11 7 11 7 Mean rank     6.9722 2.5278 7.0556 2.4444 Mann-Whitn U   17.5 16 p (same med.)   0.063229 0.04632 Z   1.8576 1.9925 Monte Carlo permutation (p)   0.0551 0.0449 Exact permutation     0.058384 0.044118                  72  Density Salmon   Villages Camps   N 11 7 Tests for equal variance       Variance 22.791 3.2879 F 6.9318 p (same var.) 0.027704 Critical F value (p=0.05): 5.4613 Monte Carlo permutation (p) 0.0385 Exact permutation 0.041384       Tests for equal means Mean 5.7755 2.2371 95% conf. max 8.9826 3.9141 95% conf. min 2.5683 0.56017 Difference between means 3.5383 95% conf. interval (parametric): (-0.49399 7.5706) 95% conf. interval (bootstrap): (0.5839 6.5336) t 1.8602 p (same mean) 0.081341 Critical t value (p=0.05) 2.1199 Uneq. var. t  2.2194 p (same mean) 0.043664 Monte Carlo permutation (p) 0.0812 Exact permutation (p) 0.082673      

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0380481/manifest

Comment

Related Items