AN INVESTIGATION OF ANASAZI TURKEY PRODUCTION IN SOUTHWESTERN COLORADO. Natalie D. Munro

Transcription

1 AN INVESTIGATION OF ANASAZI TURKEY PRODUCTION IN SOUTHWESTERN COLORADO Natalie D. Munro B.Sc., Southern Methodist University, 1991 THESIS SUBMITTED IN PARTIAL'FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of Archaeology 63 Natalie D. Munro 1994 SIMON FRASER UNIVERSITY May 1994 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other means, without permission of the author.

2 APPROVAL Name: Degree: Title of thesis: Natalie D. Munro Master of Arts (Archaeology) An Investigation of Anasazi Turkey Production in Southwestern Colorado Examining Committee: Chair: Dr. Roy Carlson Jonathan C. Driver ior Supervisor Dr. Mark Skinner Associate Professor Department of Archaeology ~ r ~Thael. Blake External Examiner Department of Anthropology and Sociology University of British Colombia Date Approved: May 18, 1994

3 PARTIAL COPYRIGHT LICENSE I hereby grant to Simon Fraser University the right to lend my thesis or dissertation (the title of which is shown below) to users of the Simon Fraser University Library, and to make partial or single copies only for such users or in response to a request from the library of any other university, or other educational institution, on its own behalf or for one of its users. I further agree that permission for multiple copying of this thesis for scholarly purposes may be granted by me or the Dean of Graduate Studies. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Title of ThesisIDissertation: I AN INVESTIGATION OF ANASAZI TURKEY PRODUCTION IN SOUTHWESTERN COLORADO Author: Signature Name May 18, 1994 Date

4 ABSTRACT The current study is an investigation of turkey production and its potential as an indicator of intensification, resource access and domestication in southwestern Colorado during the latter part of the Anasazi occupation (ca. A.D ). The archaeological remains of turkeys are utilised to investigate four primary questions. The first is whether the turkeys in the sample were domesticated or wild. A reexamination of current divisions of southwestern archaeological turkey species and breeds is performed to address this problem. Second, the changing function of the turkey in Anasazi society is evaluated. Third, intensification in turkey production and its relationship to human population aggregation is examined by tracing temporal change in the proportion of turkeys in Mesa Verde region faunal assemblages. Finally, spatial variation in the intensity of turkey production is considered within the context of resource access, land tenure and community organisation during Pueblo I11 in the Sand Canyon Locality. The osteological examination of turkey bone did not reveal whether the turkeys in the Sand Canyon sample were domesticated or wild. Many questions are raised as to the validity of current subdivisions of Meleagris gallopavo into subspecies or distinct breeds. It is proposed that variation may instead be a result of environmental factors, isolation or natural variation within a population. However, numerous contextual indicators suggest that the turkeys in the sample are domesticated. The function of the turkey in Anasazi society was evaluated by searching for indicators of its value as a ritual or utilitarian resource and evidence for food or feather use in the archaeological record. The data support the utilisation of the turkey for all of these purposes, though its primary function varied over time.

5 Both temporal and spatial variation in the intensity of turkey production are recognised in the study area. Proportions of turkey in Mesa Verde region faunal assemblages increased through time, most dramatically between Pueblo I1 and 111. This is correlated with the initiation of small-scale population aggregation and the specific conditions present at the time. These include increased population densities, restricted mobility and agricultural intensification, which led to the reduction of 'big game' in the Mesa Verde region. As a result I propose that the Anasazi intensified turkey production to ensure access to a reliable meat source. In the Sand Canyon Locality during Pueblo III cliff/talus/bench sites exhibited significantly higher proportions of turkey than other locations. It is conjectured that this resulted from the presence of a land tenure system which allowed differential access to land. The occupants of the canyon sites were limited to marginal agricultural land and were thus required to increase domestic turkey production to meet their requirements for animal protein.

6 ACKNOWLEDGMENTS I would like to begin with a special thank you to my advisory committee. I am most indebted to Jon Driver, my supervisor, for providing invaluable guidance, support and financial assistance which ensured the completion of this project. Mark Skinner read and commented on various drafts and provided many insightful ideas. Thanks also go to Michael Blake for taking the time to serve on my committee as my external advisor. Secondly, I would like to extend a sincere thank you to the Crow Canyon Archaeological Center, particularly for providing financial assistance and a friendly and supportive atmosphere while I was undertaking the analysis for this project. Special thanks go to Karen Adams, Mark Hovesak, Bill Lipe and Melita Swain for their time and assistance with many technical aspects of this research, including making the faunal collections available to me. Thanks also to Bruce Bradley and Charmion McKusick for donating turkey skeletons for my identifications. Most importantly I would like to thank Mark Varien for tirelessly answering my many questions, and for his continual encouragement. I am grateful for the stimulating discussion, suggestions and support of my fellow graduate students. Thanks to John Darwent for sharing his first hand experiences with raising turkeys. Rick Schulting's input was particularly valuable, especially when it related to anything statistical. Otherwise his discussion was thought provoking and enabled me to straighten out many of my own ideas in my head. Finally, thanks to Hartley Odwak, for clearing up seemingly endless glitches on the computer and being an unrelenting wall of support throughout the duration of this project. Not to be forgotten is my family: Neil, Jennifer, Shane, Camille, Kara, Krista, Pauline, Evan, Grace and Mimi. Thank you for your infinite faith!

7 TABLE OF CONTENTS APPROVAL... n... ABSTRACT... m ACKNOWLEDGMENTS... v LIST OF TABLES... ix LIST OF FIGURES Introduction... 1 Objectives... 1 Geographic Location... 2 Environmental Setting... 4 Cultural History - The Mesa Verde Region... 5 Research History of The Crow Canyon Archaeological Center... 8 Sampling Design - The Sand Canyon Locality Sand Canyon Pueblo The Duckfoot Site The Green Lizard Site The Site Testing Program Cougar Cub Alcove Sampling Design - The Mesa Verde Region Methods Provenience Taxon Element, Part, Side, Breakage and Length Modifications Sexing Aging CHAPTER I INTRODUCTION 1 Determination of Breed 23 Detection of Temporal and Spatial Variation in Turkey Production Quantitative Procedures Organisation of Research CHAPTER IT THE TURKEY: BACKGROUND AND UTILIS ATION Introduction Habitat, Subsistence and Reproduction Origin of the Southwestern Turkey Entrance of the Turkey into Southwestern Lifeways Southwestern Species and Breeds The Function of Southwestern Archaeological Turkeys 43 Historic Documentation ; Ethnographic Documentation Archaeological Evidence... 47

8 Previous Archaeological Interpretation Introduction Domestication Intensification Population Aggregation Southwestern Models of Aggregation Environmental Deterioration Defense Leadership Exchange Competition Reduction CHAPTER III THEORETICAL BACKGROUND 54 Population Pressure -65 The Aggregation Process on the Colorado Plateau The Processes Behind Aggregation in the Sand Canyon Locality Population Increases Restricted Mobility Resource Scarcity Agricultural Intensification Aggregation Synthesis...: Aggregation in the Dolores Valley Resource Access and Land Tenure Expectations for the Assemblage Intensification Intersite Variation in Turkey Production in the Sand Canyon Locality CHAPTER IV RESULTS Introduction Evidence for turkey domeshcatlon Osteological data Contextual Evidence The Turkey's Function in Anasazi Society Evidence for Feather Utilisation ) The Mesa Verde Region ) The Sand Canyon Locality Evidence for Food Utilisation ) The Mesa Verde Region ) The Sand Canyon Locality Evidence for Bone Utilisation ) The Mesa Verde Region ) The Sand Canyon Locality Intensity of Turkey Production: Temporal Change Evidence from the Mesa Verde Region Evidence from the Sand Canyon Locality Intersite Variation Within the Sand Canyon Locality vii

9 ... Introduction Is There a Turkey in the House? Osteological Evidence Conclusion of osteological research Conclusion of contextual evidence The Role of the Turkey in Anasazi Society CHAPTER V DISCUSSION A ND CONCLUSIONS 136 Conclusion 149 Temporal Change: Intensification in Turkey Production Conclusion Intersite Variation in Turkey Production Within the Sand Canyon Locality Conclusion Problems Encountered Suggestions for Further Research REFERENCES APPENDIX A: Elemental Distribution of Turkey by Site APPENDIX B: Measurements of Tarsometatarsi APPENDIX C: Statistics Testing Spatial Variation in Turkey Production

10 LIST OF TABLES Table 1.1 Table 4.1 Table 4.2 Table 4.3 Table 4.4 Table 4.5 Table 4.6 Table 4.7 Sampled sites from the Sand Canyon Locality and their dates of occupation Ranges of means of greatest lengths of tarsometarsi by breed and sex Means of Greatest Length of tarsi from sites in the Mesa Verde Region Ratio of eggshell to bone by site in the Sand Canyon Locality Age structure of the turkey population in the Sand Canyon Locality by site Indices of preservation, the ratio of the distal ends of Sylvilagus sp. tibiae and humeri to their proximal ends at sites in the Sand Canyon Locality Distribution of cutrnarks by element and the percentage of each element with cuts The location of cutmarks on turkey elements in the Sand Canyon Locality Table 4.8 Table 4.9 Part representation of humeri, tibiotarsi and radii from Pueblo 111contexts in the Sand Canyon Locality Sex structure of the turkey population by site in the Sand... Canyon Locality Table 4.10 Artifacts divided by type from Pueblo I11 contexts in the Sand Canyon Locality Table 4.11 Artifacts divided by element from Pueblo I11 contexts in the Sand Canyon Locality Table 4.12 Percentage of turkey per site in the regional sample from the... Mesa Verde region -123 Table 4.13 Turkey proportions divided by time period from sites in the... Sand Canyon Locality -128 Table 4.14 Site Sample for Spatial analysis within the Sand Canyon Locality... ix -130

11 Table 4.15 Percentage of turkey bone in Pueblo 111 sites in the Sand Canyon Locality by topographic location Table 4.16 Statistics comparing the cliff/talus/bench sites and the other Pueblo I11 sampled sites in the Sand Canyon Locality... combined 134

12 LIST OF FIGURES Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4 Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 The Southwest United States... 3 Sampled sites from the Sand Canyon Locality Sampled sites from the Mesa Verde region Measurements taken from the turkey tarsometatarsus Distribution of the breadth of the distal end of tarsometarsi in the Sand Canyon Locality Simple regression of tarsometatarsi measurements: distal breadth versus greatest length Greatest length distribution of tarsometatarsi from the Sand Canyon Locality Skeletal depiction of Meleagris gallopavo indicating the locations of common cutrnarks on turkey skeletons in the Sand Canyon Locality sample Bar graph representing the proportion of turkey versus. lagomorphs and artiodactyls at Pueblo I11 sites grouped by topographic location from the Sand Canyon Locality Average percent and ranges of turkey at Pueblo I11 sites in the Sand Canyon Locality, grouped by topographic... location..i33

13 CHAPTER I INTRODUCTION Introduction In this study the remains of domestic animals are used as an alternative to traditional data sets (i.e. architecture, agricultural productivity) to study changes in the Anasazi community which occupied the Sand Canyon Locality, Colorado. The study examines the shifting role of the turkey (Meleagris gallopavo) as the community underwent population aggregation. This will provide insight into the function the turkey played in Anasazi prehistory and will supplement current research addressing broader questions concerning community organisation and intensification in the Sand Canyon Locality. The temporal parameters under investigation range from ca. A.D. 500, when the turkey first appeared as a domesticate in the northern Southwest, to A.D when regional abandonment occurred. The primary spatial focus will be on the Sand Canyon Locality, currently under investigation by the Crow Canyon Archaeological Center. Data from the Mesa Verde region will also be considered to provide greater temporal depth. Objectives This project is designed with the dual goals of clarifying the turkey's role in Anasazi prehistory and addressing broader research problems related to intensification and resource access. The study investigates Anasazi turkey production since its inception in the Mesa Verde region, particularly within the Sand Canyon Locality. Four primary questions will be addressed by utilising archaeological turkey assemblages as a data base. The first two relate specifically to the turkey and its role within pueblo society. They include: 1) reexamining

14 the assignment of distinct species and breeds to archaeological turkeys in the Southwest, particularly the differentiation between domestic and wild birds; and 2) determining the function of the turkey within Anasazi society and how it changed through time. The final two objectives have been designed to investigate more general problems which have not traditionally been addressed using faunal data sets. These are intended to fit within the broader research goals of the Crow Canyon Archaeological Center. They involve: 1) examining variation in the intensity of turkey production through time and its relationship to large-scale aggregation in southwestern Colorado, and 2) investigating intersite variation in turkey production within the context of resource access systems and sociocultural organisation in the Sand Canyon Locality during Pueblo 111. It is hoped that the results of this study will aid overall research by substantiating or refuting current hypotheses not only in the study area, but in the greater Southwest as well. Geographic Location The Sand Canyon Locality has been selected as the primary research domain. It is situated 18 km from modem day Cortez in southwestern Colorado, and in close proximity to other major Anasazi centres, including Mesa Verde National Park, the Dolores Drainage and Hovenweep National Monument. Each of these experienced similar culture histories, though temporal variation existed (Breternitz et al. 1986, Rohn 1977, Varien et al. 1992). The locality occupies an area of 225 sq. km in the McElmo Drainage which flows into the San Juan River. Boundaries are designated by McElmo Creek to the south and Yellow Jacket Creek to the north. The eastlwest borders are arbitrarily assigned by an arc of 7.5 kilometres which is drawn to the east of Goodman Point Ruin and to the west of Sand Canyon Pueblo. These two sites are the largest within the locality and are

15 Figure 1.1: The Southwest United States.

16 surrounded by numerous small sites (Lipe 1992a). Between ca. A.D. 100 and A.D. 1300, the study area was occupied by the Anasazi after which a large scale rapid abandonment of the region occurred (Cordell 1984, Lipe 1992a). A regional sample has also been selected to address aspects of the study which require greater temporal depth (intensification and function). The area known as the Mesa Verde region is defined as the territory bounded to the north by the Dolores Valley, to the south by the New Mexico border, and ranges east to west from Mesa Verde to the Utah border. This region has been chosen since it surrounds the Sand Canyon Locality thus minimizing geographic and culture variability (see Figure 1.3). In addition, some more specific comparisons will be made with data collected by the Dolores Archaeological Program which undertook research in the Dolores Valley 13 kilometres north of Cortez. Environmental Setting Topographically, the Sand Canyon Locality is situated on a circular, structural uplift known as the McElmo Dome. The area is laced with a complex network of mesas and canyons, created by the erosion of sandstone formations (Kuckelman et al. 1991). Mesas occupy the flat land between the canyons and are carpeted with arable soil. In contrast, canyon walls are steep and rocky, hosting benches and talus slopes which were frequently used for Anasazi habitation, though they accommodated less productive acreage. Elevations range between 1,700-2,100m (5,600-7,000 feet) (Kuckelman et al. 1991). The southern portion of the Sand Canyon Locality is dominated by the Lower McElmo Drainage. This area exhibits a distinct microenvironment due to its lower elevation and sparse vegetation. The climate is predominantly arid, though not desert-like. Sufficient precipitation for dry farming falls in most seasons, though the climate is well known for its dramatic and unpredictable fluctuations. Local

17 temperatures range from an average -3 degrees Celsius in January to 22 degrees Celsius in July (Walker 1990a). This allows for approximately frost free days (depending on elevation), which is an adequate growing season for maize cultivation. The flora and fauna which inhabit the Sand Canyon Locality today are much the same as they were during the Anasazi occupation. Vegetation on the mesa tops is dominated by drought resistant species such as sage brush (Artemisia tridentata), pinyon pine (Pinus edulis) and juniper (Juniperus osteosperma). The latter two also thrive on the canyon rims and talus slopes. The canyon bottoms supply enough moisture to support riparian species such as cottonwood (Populus sp.) and willow (Salix sp.)(adams 1992). The lower McElmo provisions sparser vegetation including yucca and cactus (Varien et al. 1992). Common faunal species include mule deer (Odocoileus hemionus), coyotes (Canis latrans), foxes (Vulpes sp.), cottontails (Sylvilagus sp.) and jackrabbits (Lepus sp.). Several species of rodents are also abundant. Birds are represented by eagles, hawks, turkey vultures, jays, flickers, swallows, swifts, magpies, hummingbirds and numerous songbirds (Kuckelman et al. 1991). Cultural History - The Mesa Verde Region The first human residents in southwestern Colorado were the Paleoindians. Remains from their occupation are rare, and have been attributed to temporary, seasonal habitation in the area (Cordell 1984, Lightfoot 1991). The Archaic tradition replaced the Paleoindian Period ca. 5,000 B. C. and was characterised by increased selectivity of habitation sites and the adoption of maize agriculture at an equivocal date of approximately 1,000 B.C. (Vierra 1990) (see Wills 1988 and Minnis 1985 for alternative arguments). By A.D. 100 the human population inhabiting the northern Southwest is

18 known as the Anasazi who can be recognised by a series of diagnostic traits. Their chronology as formulated by Kidder in the late 1920s, commences with Basketmaker I1 (100 B.C.-A.D. 500) (Cordell 1984). During this period local population densities were low, though gradual population growth took place throughout the occupation (Orcutt et a ). During Basketmaker I1 the Anasazi inhabited shallow subterranean pithouse structures, associated with storage cysts. Settlements appeared either as loose clusters or dispersed hamlets. The local subsistence strategy focused on a mixed economy of hunting, gathering and limited agriculture emphasising maize and squash production. Typical artifacts include basketry, twined bags, atlatls, dart weapons, rabbit fur blankets and two handed manos with troughed metates (Cordell 1984). Basketmaker I11 commenced ca. A.D. 500 and persisted until A.D Population densities remained low throughout Baskermaker 111, though they were higher than preceding periods (Orcutt et al. 1990). The subterranean pithouse continued as the primary habitation structure but boasted new architectural features such as antechambers and sipapus. The subterranean depth of the structures also increased (Cordell 1984). The appearance of small villages and the occasional Great Kiva, are also noted, though hamlets remained popular for habitation (Wills and Windes 1989). Fertile agricultural soil became the primary determinant of site location and increasing emphasis was placed on agriculture, particularly following the introduction of beans ca A.D. 400 (Ford 1981) which are an excellent nutritional balance for maize and squash (Minnis 1985). Innovations in material culture include the appearance of gray-ware ceramics and the bow and arrow (Cordell 1984). Pueblo I was initiated ca. A.D. 750 and terminated by A.D This period was characterised by episodes of aggregation, increased sedentism and the expansion of agriculture. Local populations were higher than ever and

19 settlements were less evenly distributed over the landscape as a result of localised aggregations (Adler and Varien 1991). Architectural changes were manifested in the pithouse which underwent structural transformations perhaps as the result of a functional transition to a ritual entity (Varien and Lightfoot 1989, Wilshusen 1989). The resulting structures are known as kivas. Habitations evolved from the above ground storage rooms, forming small contiguous units (Prudden 1903). Great Kivas were more common during Pueblo I and are believed to have functioned as integrative facilities for increasing populations. Innovations in material culture include the introduction of cotton into the agricultural complex for use as a textile and food source. Ceramics were more refined and cradleboards which produce a distinctive flattening of infants skulls were introduced (Fiedel 1987). Pueblo I1 was initiated ca. A.D. 900 and endured until A.D During this period settlements remained small and consisted of a series of wattle and daub or masonry surface rooms and kivas. By A.D the Anasazi occupied their maximum extent, leaving few environments uninhabited (Adler and Varien 1991). Changes in ceramics are noted in the appearance of corrugated wares and an increase in decorated pottery (Cordell 1984). Also, turkey feather blankets largely replaced those previously manufactured from rabbit fur. It was during Pueblo I1 that the development of a large-scale Anasazi community known as the Chaco Phenomenon took place in the San Juan Basin in New Mexico. This preplanned society was linked to an extended community throughout the Four Comers region by a complex road system. Chaco Canyon was the centre of influence in the region while at its cultural peak, as reflected by the appearance of Great Houses referred to as Chacoan outliers in numerous locations outside of the San Juan Basin. Several outliers occupied the Colorado Plateau though the Chacoan community collapsed during the mid A.D. 1100's (Lekson 1991).

20 Pueblo I11 represents the culmination of the Anasazi occupation in the Mesa Verde region. This period endured from ca. A.D to 1300 and terminated with the large-scale rapid abandonment of the region. Settlement patterns were characterised by dispersed communities composed of small scattered habitation sites and aggregated villages (Varien et al. 1994). Population density reached its maximum, while habitation sites were located in closer proximity to one another. Sites were also located in association with less arable soil than earlier occupations, as the majority of productive soil was already under cultivation (Adler 1990a). Changes in material culture include an increase in decorated ceramics and the appearance of flat metates which were set in bins. By the late A.D. 1200's, construction terminated in most areas of the Mesa Verde region and abandonment followed soon after (by A.D. 1300) (Lightfoot 1991). Many hypotheses have been developed to explain the mysterious disappearance of the Anasazi. The most widely accepted proposes that the Anasazi migrated south and became integrated into communities in these areas (Lipe 1992b). This also explains the increase in population in the Mogollon Highlands, Upper Rio Grande, Hopi Mesas, Acoma and Zuni. Several suggestions have been advocated to explain why the Anasazi headed south. These include environmental degradation, high population densities and intersite hostility which may have driven the Anasazi away from the Mesa Verde region. The attractiveness of religion and organisation in the south are suggested as proponents which may have attracted the Anasazi to these communities. As yet the exact cause of abandonment is unknown though it is probable that it resulted from a combination of the above factors. Research History of The Crow Canyon Archaeological Center The Crow Canyon Archaeological Center is a not-for-profit research

21 facility located in Cortez, Colorado. It was established in 1983 with the dual goals of combining research objectives with education (Lipe 1992a). The Sand Canyon Locality was proposed as a useful region for study as an alternative to the extensive research concentrated on the exceptional settlements at Mesa Verde National Park. Little other research concerning the Pueblo I11 period has been undertaken elsewhere in the Mesa Verde region, particularly on large aggregated settlements and communities. In addition, the area houses a high density of settlements, and was inhabited throughout Anasazi prehistory. The locality includes Sand Canyon Pueblo, an aggregated Pueblo I11 site which sits at the head of Sand Canyon and is ideal for the investigation of large Pueblo I11 aggregated sites and the processes of aggregation. The Sand Canyon Locality was therefore selected as a domain with high potential for the examination of Pueblo I11 community organisation (Lipe 1992a). The Sand Canyon Locality covers an area of 225 sq. km. Dimensions of the study locale were chosen to represent the areal extent researchers believe would be required to support a Pueblo I11 Anasazi community in terms of subsistence and habitation area (Lipe 1992a). The size determined was based on Willey and Phillips definition of a locality, which states that it is larger than a settlement, yet smaller than a region. The community is defined as "...the muxima1 group of people who reside close to one another and who interact regularly on a face-to-face basis" (Lightfoot 1991 : 1). The Crow Canyon Archaeological Center has four primary research objectives. The researchers wish to: 1) identify the sociocultural organisation and define the communities inhabiting the Sand Canyon Locality during Pueblo 111; 2) examine cultural, social and environmental changes during this time;

22 3) compare their data on a larger scale to increase understanding of the changing patterns within the greater Southwest; 4) build the foundations for the above problems by performing a number of instrumental studies such as chronology building and the reconstruction of past environments. In an attempt to achieve the proposed objectives, several major projects have been initiated over the past ten years. These include archaeological excavations at Sand Canyon Pueblo, the Duckfoot Site, the Green Lizard Site and the Site Testing Program which sampled thirteen sites (Lipe 1992a). The completion of these projects will conclude the research planned for the period. In addition, instrumental studies have been conducted by the environmental archaeology program, research associates and students. The current project has been undertaken with the dual objective of complementing research currently underway at the Crow Canyon Archaeological Center and increasing the body of knowledge regarding the turkey in southwestern archaeology. To achieve these goals the applicable methods and goals as defined by Crow Canyon have been incorporated into the research design. Prior analyses of fauna excavated by the Crow Canyon Archaeological Center have been undertaken by various associated researchers. Neusius (1985) performed a preliminary analysis of the Sand Canyon fauna and later analyses were undertaken by Walker (1990a ) and Brand (1991). Walker also examined the remains from the Duckfoot Site (1989) and the Green Lizard Site (1990b). Fauna from four sites excavated during the small site testing program were analysed in Brand (1991) and were synthesised with data from the remaining nine sites by Driver et al. (1995). In addition, a study by Bullock (1992) concerning

23 the bone tools from selected sites from the Sand Canyon Locality has recently been completed. Sampling Design - The Sand Canyon LocaIity In this study, the sample was chosen to conform to research already completed or in progress in the Sand Canyon Locality. The majority of fauna recovered from past excavations is included in the sample. Four major projects have produced animal remains from 17 sites. All turkey remains from each site were examined with the exception of the assemblage from Sand Canyon Pueblo which was subsampled. 1 5MT3868 Duckfoot Site A.D Kenzie Dawn Hamlet A.D A.D A.D G and G Hamlet A.D A.D Shorlene's Site A.D A.D Cougar Cub Alcove A.D Roy's Ruin A.D Lillian's Site A.D Sand Canyon Pueblo A.D Green Lizard Site A.D Troy's Tower A.D Mad Dog Tower A.D Catherine's Site A.D Saddlehorn Hamlet A.D Stanton's Site A.D Castle Rock Pueblo A.D Lookout House A.D Lester's Site A.D Table 1.1: Sampled sites from the Sand Canyon Locality and their dates of occupation.

24

25 (See Table 1.1 for a list of sampled sites from the Sand Canyon Locality and Figure 1.2 for a map). Sampling procedures are therefore identical to those used during each of the four projects. Faunal remains which were excavated by the Crow Canyon Archaeological Center have been catalogued and stored using standard laboratory procedures (Schwab et al. 1989). Materials were easily accessed for this study, thanks to Crow Canyon's generous cooperation. The sampling procedures for all sites in the current sample are discussed below. Sand Canyon Pueblo Sand Canyon Pueblo (5MT765) is a large Pueblo I11 site located at the head of Sand Canyon. The site underwent a short, continuous occupation between ca. A.D and 1300 when abandonment occurred (Bradley 1992). Excavations at the site have been directed by Bruce Bradley since Prior to this, in the first year of excavation (1984) Charles Adams shared the directorship. Due to the large size of the site only a small percentage is examined each year. To date less than 10% of the site has been excavated. The sampling strategy implemented for excavations at Sand Canyon Pueblo was devised by Adams and Bradley (Bradley 1992). It is designed to sample diversity in two types of structural groups referred to as kiva suites and architectural units. Kiva suites consist of a kiva and all areas with which it is deemed to be in association. Architectural units may incorporate one or more kiva suites and are defined by structural continuity. Spaces between structures and other architectural features designate their boundaries. Three types of architectural units have been identified: the kiva dominated roomblock (less than 5 rooms per kiva); the standard block (5-16 rooms per kiva); and room dominated blocks

26 (more than 16 rooms per kiva). Sand Canyon Pueblo is composed of a total of fourteen architectural units (Bradley 1992). The sampling strategy was designed to emphasise the variation in the types of architectural units therefore, judgmental sampling was utilised to select architectural units within which kiva suites were excavated in their entirety. At the conclusion of the 1989 season a total of six kiva structural groups had been excavated at Sand Canyon Pueblo. Due to the large quantity of faunal remains recovered, it was necessary to subsample the material to create a manageable data base for this project. The sample utilised here represents approximately half of the total turkey bone available from the excavations completed by the end of the 1989 season. An attempt was made to maintain the original sampling goals of Bradley and Adarns in devising the subsample. Three architectural blocks have thus been chosen, one from each group of roomblocks previously mentioned. These include architectural block 100 (kiva dominated), architectural block 1000 (standard) and architectural block 300 (room dominated). In addition, the remains from non-structures 209,210 and structure 1204 were included in the sample. These units were chosen for potential comparisons with studies performed by other researchers affiliated with the Crow Canyon Archaeological Center (Hegmon 1991, Chao 1991). Fauna was collected and catalogued uniformly for all Sand Canyon excavations and identifications were performed by Walker (1990a) and Brand (1991). The latter was supervised by Jon Driver at Simon Fraser University. The Duckfoot Site Between 1983 and 1987 Ricky Lightfoot directed excavations at the Duckfoot Site (5MT3868). The site is a small Pueblo I habitation occupied between A.D and 880. Site layout is according to Prudden's (1903)

27 definition of the unit pueblo and consists of 19 rooms, 4 pithouses, a plaza and a midden. The sampling design called for the excavation of the site in its entirety and resulted in the excavation of all rooms and pithouses, the entire plaza and 90% of the midden. Walker (1989), in his faunal analysis of Duckfoot views the sample as representative of the entire local fauna at the site. Duckfoot is the only site excavated in the Sand Canyon Locality which lacks a Pueblo I11 component. As such, it is ideal for comparative purposes and for the detection of temporal change in the Sand Canyon Locality. The Green Lizard Site In 1987 excavation was initiated at The Green Lizard Site (5MT3901) under the direction of Ed Huber (Huber and Lipe 1992). The site was chosen due to its suitability for comparison with Sand Canyon Pueblo since although they were contemporaneous, they differ greatly in overall size (Green Lizaid is much smaller). The area chosen for excavation includes one kiva and all associated rooms (a kiva suite) in the western portion of the site. This judgmental selection allows comparisons with the kiva suites excavated at Sand Canyon Pueblo. One test trench was also judgmentally excavated in the eastern half of the site to search for a second kiva suite which was eventually located. In addition to these excavations, a stratified random sample was instigated using lxlm test pits. The random pits were designed to test six sampling strata identified in the midden and the peripheral areas of the site (Huber and Lipe 1992). All faunal identifications and analysis for Green Lizard were performed by Walker (1990b). The Site Testing Program In 1988 the Site Testing Program was initiated under the direction of Mark Varien. Thirteen sites believed to be contemporaneous with Sand Canyon Pueblo

28 were selected for excavation in an attempt to examine organisation within the Pueblo I11 community (Varien et al. 1992). The same excavation and recording strategies were implemented at each tested site creating an excellent body of data from which to draw comparisons. Variation among Pueblo I11 sites is high in the Sand Canyon Locality. Site selection was "designed to sample this diversity and to obtain data useful in evaluating and more fully understanding the variation that is present" (Varien et al. 1992: 49). Sources of variation include: a) geographic location: the upper and lower Sand Canyon, b) physiographic setting: the lower canyon, the talus slopes, canyon benches and mesa tops, c) site size: small and medium sites (no large sites are available for study other than Sand Canyon Pueblo, currently under investigation), d) site layouts, e) occupation dates within Pueblo 111: prior to or contemporaneous with Sand Canyon Pueblo. An attempt was made to choose at least one site from each of the categories within the five criteria. As a result thirteen sites were selected and excavated. Sites tested in the program were sampled using a stratified random approach, although occasional judgmental samples were taken. One by one metre test pits were used for both strategies. Strata were defined following the mapping of each site. Sampling variability between sites was reduced due to the similarity in site layouts and components which created like strata. Random samples were excavated from all layers identified at each site, though the number of sampling units varied between sites, according to size (Varien 1992). Finally, the faunal assemblage was analyzed by students at Simon Fraser University using standard procedures devised by Driver (1991). Cougar Cub Alcove Fauna from one additional site was included in the sample. Cougar Cub

29 Alcove is a small cliff face/talus slope habitation similar to those tested in the Site Testing program (Kuckelman et al. 1991). It was sampled by the crew working on the Site Testing Program in Only two test pits were excavated in an attempt to obtain tree-ring samples for dating. The site had been extensively damaged by looters, thus the samples were taken before further injury could result. Once again, fauna was identified by students at Simon Fraser University under the guidance of Jon Driver (1991). Sampling Design - The Mesa Verde Region The Mesa Verde sample consists of the sites from the Sand Canyon Locality sample as reviewed above, as well as numerous other sites from the surrounding region. Sites ranging in time from Basketmaker I11 to Pueblo I11 were examined. Sites were selected for inclusion in the Mesa Verde sample based on the availability of site reports which contained quantified faunal data. Because the sites were excavated and analyzed by a variety of researchers using incommensurate strategies for recording and quantification there are serious compatibility problems amongst the data. Limited analysis will be performed keeping these factors in mind, thus the resulting trends in the faunal record through time and across space should be interpreted with caution. Sites included within the regional sample are depicted on the regional map in Figure 1.3. Methods All turkey and large bird bone recovered from units in the sample as described above, were selected for analysis. Each bone was studied individually and the following information was recorded: provenience, taxon, element, part of element, side, breakage and length. When present, modifications were noted and specimens were sexed, aged and/or measured when possible.

30

31 Provenience Provenience information was recorded according to the system employed by the Crow Canyon Archaeological Center (Schwab et al. 1989). Each bone received both a provenience designation and field specimen number following excavation. These numbers refer the bones to the specific context from which they were removed. They also make it possible to determine the matrix in which the specimen was found. Taxon Taxon for each specimen was recorded as either turkey or large bird. Elements were assigned to the large bird category if they were larger than a mallard duck, but could not be identified as turkey with certainty (Driver 1991). This occurred when diagnostic features were missing or obscured or when complete elements from large bird species exhibited no obvious morphological differences. For example, distal phalanges are virtually identical in many large bird species, therefore they cannot be assigned to a species with certainty. For the purpose of the current analysis large bird is incorporated within the turkey category since turkeys are by far the most ubiquitous large bird in the assemblage. Other large birds were rare in the Sand Canyon Locality assemblages, thus it is unlikely that the high quantity of large bird bone belonged to a species which was not even identified in the assemblage. Element, Part, Side, Breakage and Length The element, part, side, breakage and length categories were recorded according to Driver's manual (1991). This manual has guided all analyses of Crow Canyon fauna since its implementation in 1991, and allows for easy comparison between data sets. Element was recorded for all identifiable specimens before

32 they were assigned to taxon. If the element was not identifiable the taxon was determined to be unidentifiable as well. Part refers to the portion of the element which was preserved. Each part is defined in the manual and assigned a particular numerical code. Paired elements were sided if sufficient diagnostic features were preserved. Breakage refers to the condition of the articular ends of an element and when applicable its anterior and posterior ends. Letter codes were utilised to indicate intact, culturally or naturally modified ends and various breakage patterns. Length was recorded by measuring maximum length to the nearest millimetre using sliding calipers. Modifications Modifications were also recorded according to Driver (1991). The presence of both natural and cultural modifications including cutmarks, tool manufacture, polish, abrasion, rodent gnawing, carnivore chewing, pathologies and burning were noted. In addition, cutmarks, artifacts and localised burning were sketched on plates illustrated by Harvey et al. (1968). This enabled later reference to the exact locations of cutmarks and burning and to the types of artifacts so that frequencies could be compiled and grouped during analysis. Sexing Specimens were sexed when at least one of three criteria was present. These include size, presence of spurs and presence of medullary bone. a) Size. There is notable size sexual dimorphism within southwestern turkey populations (Meleagris gallopavo). Schorger (1966) claims that the adult turkey hen reaches only one half the weight of the male. It has also been argued that the largest female and the smallest males exhibit no overlap in body size regardless of breed or species, once they have reached adulthood (McKusick

33 1986). McKusick (1986) argues that the smallest turkey in the Southwest prehistorically was the Small Indian Domestic (Meleagris gallopavo tularosa). Therefore, individuals with elements smaller than the male SID should be female, while those which are larger should be male. To supplement her work, McKusick created illustrations of male Small Indian Domestic elements to use as aids during sexing. The plates include sketches of the innominate, sternum, coracoid, humerus, ulna, radius, carpometacarpus, femur, tibiotarsus and tarsometatarsus. During the analysis for this project, adult bones with at least one complete end were compared to McKusick's illustrations. Bones which were larger than the sketches were designated male, while those which were smaller were designated female. Increasing familiarity with the study assemblage made the distinction more obvious. b) Presence of Spurs. Discounting the rare exception spurs develop only on the tarsometatarsi of male turkeys (Schorger 1966, McKusick 1986). In this project a bone was identified as male if a spur was present. Conversely, the absence of spurs on the tarsi was not used to indicate femaleness. The spur, is easily susceptible to separation from the bone due to post-depositional processes and does not form until the bird is approximately seven or eight months of age (McKusick 1986), thus its absence does not exclude it from being a male specimen. c) Presence of Medullary Bone. Medullary bone develops in female birds during periods of egg-laying. It serves as a temporary storage area for the minerals necessary in eggshell formation and is most visible in the marrow cavity of long bones starting a few weeks prior to laying (Driver 1982, Rick 1975). Throughout the laying season medullary bone is depleted and eventually disappears. It can be recognised as a light porous bone which outwardly resembles sponge toffee. Elements in which it was found were recorded as

34 females, those specimens in which it was not exhibited were left unsexed if no other sexing criteria were present. The absence of medullary bone does not indicate that an element is male, as it can only be detected in females during part of the year, otherwise male and female bone is similar. Aging.. Elements were assigned to one of four growth stages if they displayed at least one intact end. In this project the stages were defined by altering categories devised by McKusick (1986) who had in turn, modified age stages created by Hargrave (1965). Hargrave established the following four divisions: Juvenile (hatching to 5 months); Immature ( 6-11 months); Young Adult (1-2 years) and Adult (2 years and older). McKusick made modifications to these stages by subdividing the juvenile stage into small (hatching to one month), medium (2-3 months) and large (3-5 months) classes. She also added an Old Adult category to incorporate birds four years or older. The changes were based on her familiarity with turkey remains and her resulting ability to distinguish between birds which were relatively close in age. For the purpose of this research further changes were made. McKusick's small juvenile (SJ) stage was not altered and includes birds between hatching and one month of age. The medium juvenile group is also left intact representing birds between 2-3 months. The immature group however, have been combined with the large juveniles and young adults to make up the present immature category, which refers to birds between 3 months and two years. It was necessary to condense these groups as size sexual dimorphism becomes apparent at this age yet there are no criteria to distinguish between the sexes prior to adulthood. This makes it extremely difficult to differentiate between an immature male and a juvenile female, for example. The final age

35 group is the adult category and includes all buds over two years of age. Old adults are not distinguished from adults in this research, as they are also difficult to differentiate. Bones were aged according to two criteria: size and degree of ossification. Size was utilised to identify members of the first two categories, whereas ossification was used to recognise the latter two. Ossification is a process which bird bones undergo as they age. At hatching bones are composed of cartilage and as the individual ages the cartilage is gradually replaced by bone (it ossifies). When the process is completed the bird is considered an adult. Sub-adult birds may be recognised by the texture of their bone which is not yet fully ossified and appears more porous and spongy than adult bone. As birds approach adulthood the spongy texture can only be recognised at the articular ends of long bones. The size limits of the young age groups have been illustrated by McKusick (1986: 27-31). These plates are particularly useful for aging bones from the small and medium juvenile categories. During these stages both males and females are similar in size and sexual dimorphism is not yet a factor. Bones were assigned to the immature category if they were larger than McKusick's plates for medium juveniles, but were not yet fully ossified. Bones exhibiting complete ossification were classified as adults. Determination of Breed Two osteological methods were used in an attempt to determine if the turkeys in the sample were domestic or wild, and whether they could be assigned to a specific breed or subspecies. These are described as follows. 1) Character identification. McKusick (1986) states that not only is it possible to differentiate between wild and domestic turkeys on the basis of the presence of diagnostic characteristics, but also to determine their breed. The

36 specific characteristics for three turkey groups (wild form of Meleagris gallopavo merriami, domestic form of Meleagris gallopavo merriami and Meleagris gallopavo tularosa) are described in detail by McKusick (1986: 35-53). In this study applicable elements were grouped and laid out for examination. An attempt was made to identify features which McKusick has attributed to the three turkey groups mentioned above. The specimens were also compared with each other and similarities and differences were noted. Figure 1.4: Measurements taken from the turkey tarsometatarsus. 24

37 2) Measurements. Two measurements were selected from von den Driesch (1976) with the intent of isolating the different turkey subspecies or breeds by size. They are as follows (See Figure 1.4): a) Greatest length (GL) of the tarsometatarsus. This measurement was only taken from complete tarsi. b) Greatest breadth (Bd) of the distal end of the tarsometatarsus. According to von den Driesch (1976: 129) Bd is "measured in projection at right angles to the longitudinal axis of the bone". This measurement was only taken if the distal end of the tarsometatarsus was complete, Measurements were subjected to statistical analysis in an attempt to determine how many turkey populations were present and to which group they belonged. Of particular interest was whether the population(s) were domestic or wild. Detection of Temporal and Spatial Variation in Turkey Production: Quantitative Procedures To detect temporal and spatial variation in turkey production within the Mesa Verde region, the data from three faunal groups considered to have major economic importance were selected for comparison. These include turkey (turkey and large bird), artiodactyls (deer, elk, mountain sheep, etc.) and lagomorphs (cottontails and jackrabbits). The three groups were selected as they are frequently cited as the most valued sources of animal protein in the Anasazi diet (Bertram 1991, Brand 1991, Neusius 1986, Nickens 1981, Shelley 1993, Walker 1990a). It is likely that rodents were also utilised however they are not considered as important as the three other groups. Their exclusion also prevents problems raised when attempting to determine which animals are remnants of human activity and which are intrusive. Rodents are notorious for burrowing in

38 archaeological sites following abandonment and subsequently dying in situ (S haffer 1992). By dividing the frequency of turkeys by the total frequency of the three groups deemed to be of major economic importance [turkey/ (turkey + artiodactyls + lagomorphs)] proportions were produced (these may also be multiplied by 100% for easier comparison). These are useful for examining variation in the importance of turkey in relation to the other major economic species. It is hoped that the comparison of proportions will allow the detection of change in the intensity of turkey consumption and production and the relative importance of the above faunal groups in the diet of the local inhabitants. Proportions have been chosen to explore the problem of turkey intensification because they eliminate the need to utilise absolute numbers. They are also useful for intersite comparisons which use different quantitative strategies. Faunal reports with quantified data from sites in the study area are relatively rare and unfortunately those that do exist are often incomparable due to the use of incommensurate methods of quantification (MNI of NISP). MNI and NISP cannot be directly compared, though MNI is in fact derived from NISP. By reducing absolute data to proportions, comparisons are facilitated. This method has its share of problems, largely relating to the differences between and within the two methods of quantification (Grayson 1979, 1984). Proportions also facilitate the comparison of assemblages with varying sample sizes. Proportions, however, are not without their problems, particularly when attempting to measure intensification. In this study intensification is defined as an increase in the amount of energy invested into turkey production. Energy refers to the amount of labour expended for production which is measured using calories. Difficulties are created when attempting to utilise faunal remains to estimate labour investment since accurate estimates require knowledge of the

39 exact amount of turkeys produced, the size of the human population and the work required to construct and maintain pens. Herding and food production are also two labour consuming activities which are necessary for turkey husbandry. It is impossible to determine accurate estimates of such practices using archaeological data. Instead, it is argued that increases in the amount of turkey bone in faunal assemblages per unit capita, not only indicate an increase in consumption, but also an increase in energy investment. This is based on the assumption that the production of more turkeys requires the input of a larger quantity of energy to perform the activities discussed above. As a result increases in turkey production in this study are argued to be representative of intensification. The problems with this argument lie primarily in identifying increases in the quantity of turkeys produced by utilising proportions. Proportions are a measure which depict the amount of a group in relation to the total of the groups being compared. The total proportions of all groups being evaluated must equal one. In this study the number of turkey bones in a given faunal assemblage was divided by the total number of artiodactyl, lagomorph and turkey bones to determine the proportion of turkey. This allows for easy comparison of the intensity of turkey production between sites and over time. Problems arose because proportions are relative measures. An increase in the frequency of one component does not necessarily represent an absolute rise in that component but may indicate a decrease in a different component. Because the total proportions in each assemblage must add up to one, if one component increases an associated decrease must occur in one or more of the other components. As a result is may be difficult to determine what changes are absolute increases and thus indicative of intensification and which are simply changes in resource mix.

40 In defense of the use of proportions, there are few viable alternatives. An alternate method advocates determining the absolute quantity of turkey produced at a site. In most cases, however, this is impossible to quantify as the type of data required is not available in the archaeological record. Faunal assemblages should not be considered equivalent to the animal resources originally utilised at a site as the death assemblage is highly distorted due to taphonomic the effects of taphonornic processes. This method also requires site occupation and population estimates in order to draw comparisons between assemblages. A second alternative is to use an external artifact class such as ceramic pot sherds or debitage as a control against which to compare the data class under examination (in this case turkey bone). This approach assumes that the external data classes accumulate at a constant rate. Population and site occupation times must be estimated to determine accumulation rates. When tested these estimates have proven to be fairly reliable (i.e. Blinman 1986), however several new variables are introduced into the analysis increasing the potential for error. These are related to the reliability of population or site occupation length estimates, and the assumption that individuals produce a standard quantity of pot sherds over a certain length of time. In addition though this data is available from sites in the Sand Canyon Locality, it is not recorded from most other Mesa Verde region sites. In this study ceramic and lithic data from Pueblo I11 sites in the Sand Canyon Locality were investigated to determine their reliability as a control to measure changes in turkey production. Preliminary examination indicated that comparing turkey bone to lithic (debitage) and ceramic data created inconsistent results in the quantities of turkey produced. This suggests that additional variables affected the lithic and ceramic assemb1ages;but not in the same way and thus this method was not utilised for this research.

41 A final alternative involves determining the density of turkey bones per square unit of excavation. This method is practical because it eliminates the need to compare the frequencies of turkey bone against other species. However, once again confounding variables come into play. These include variation in sedimentation rates and the location of excavation units (i.e. in middens or roomblocks etcetera). In addition, the data for the volume of units excavated which are essential for this comparison, are available for the Sand Canyon Locality, but not from the majority of sites in the Mesa Verde region, making this type of analysis impossible. The above discussion of the problems associated with the definition of intensification and the use of proportions should be taken into account when considering the interpretation of the study results. Organisation of Research The four problems addressed in this research may be examined as separate entities, however they are also contingent on one another to some degree. Initially it is necessary to establish whether the turkey population in the sample is domestic or wild. This will guide later interpretations concerning the turkey's function. Likewise it is necessary to establish function before viable conclusions explaining variation in the intensity of the turkey's use through time and across space can be drawn. To address the primary objectives in this study results obtained by implementing the above methods are discussed in subsequent chapters. Chapter I1 provides background on past research concerning southwestern archaeological turkeys, particularly the debate surrounding its changing function in Anasazi culture. In Chapter I11 the theoretical aspects of this research are introduced, specifically domestication, aggregation and intensification theory. The chapter concludes with the presentation of several hypotheses evoked by the theoretical

42 discussion. The data collected for this project and its analysis are presented in Chapter IV which is divided into four sections each addressing a major objective of this research. The data regarding the turkey's domestication is analysed, followed by the presentation of contextual data indicating the role the turkey played in the day to day operation of Anasazi society. The final two sections trace the changing proportions of the turkey through time and across space in the faunal assemblages from the Mesa Verde region. Chapter V, the concluding chapter, is a forum for the discussion and synthesis of the data within the context of the theoretical framework presented earlier. An attempt is made to resolve and integrate the original problems addressed in this research and to provide suggestions for further research.

43 CHAPTER I1 THE TURKEY: BACKGROUND AND UTILISATION Introduction The primary purpose of this study is to examine the dynamic role of the turkey within the changing community in the Sand Canyon Locality, Colorado. The intensification of components of the cultural system have been examined by the Crow Canyon Archaeological Center by using several forms of data including architecture, geographic information systems and agricultural productivity. The faunal record has not yet been considered. Domestic animals have the potential to yield valuable insight into intensification and the social and political systems which operate within human society as they are directly controlled by human groups. The turkey has been selected largely because it was the only domestic animal present in the study locality, with the exception of occasional dog specimens. Turkeys are also one of the most common species found in the faunal assemblages from the study area and therefore are assumed to have played an important role in the local economic sphere. Finally, the turkey allows for the study of more specific problems currently under debate in southwestern zooarchaeology. These include determining whether the turkeys in the study area were wild or domestic, whether they may be assigned to specific breeds and their function within Anasazi society. The current chapter will set the stage for further discussion by reviewing the literature on southwestern turkeys. It begins with an overview of turkey habitat, subsistence and reproduction. This is followed by background on the turkey's phylogeny which includes paleontological discussion and a review of the hypotheses regarding archaeological races and species. The remainder of the chapter includes a synthesis of material addressing the role that the turkey played

44 within Anasazi society. This discussion focuses on the debate over the turkey's function as a food or feather source and how these arguments are substantiated by the archaeological record. Habitat, Subsistence and Reproduction The turkey is the largest member of the Galliforme family and is native to North America. Their habitat is located in intermontane regions with elevations ranging between 6,000-10,000 feet. In the Southwest this area is forested by ponderosa pine, pinyon pine and Garnbel's oak. The forest provides shelter and roosting areas for the birds, while small shrubby plants on the forest floor create an understory which offers protection and subsistence (Schorger 1966). Water is the turkey's primary limiting resource. The quantity ingested in food or temporarily pooled as a result of rainfall is insufficient to meet the turkey's requirement of 500 rnl to 1 litre (112-1 quart) a day. Turkeys therefore require year round access to a dependable water supply and rarely forage further than one mile from a reliable source (Schorger 1966). The intermontane environment is able to supply the turkey with abundant water and food. Turkeys forage from a broad range of flora and fauna, acquiring the majority of their nutrients from plant sources. Their diet shifts according to food availability in their habitat. In the Southwest wild turkeys subsist mainly on acorns, pine nuts, bemes, buds and insects. They have also been known to ingest small reptiles (Schorger 1966). Turkeys generally feed twice a day, in the morning and afternoon by scratching and overturning the undergrowth. The amount of food they consume depends on age, sex, size and season, but the average adult has been estimated to eat 250g (half a pound) of food per day (Schorger 1966). On average, female wild turkeys nest once a year and lay two clutches of eggs during each nesting period. Each clutch averages between eggs.

45 Laying begins in April or May and lasts for three to three and a half months (Schorger 1966). Eggs are generally incubated for 28 days and the hatching rate is between four to eight poults per hen (Windes 1987). Modern domestic turkeys have been selected to lay year round, but it is unlikely that prehistoric southwestern species were manipulated to reproduce at this rate. Poults require warm conditions to survive. They are fragile and susceptible to death by chilling and wetting, predation or drought (Schorger 1966). Human protection could potentially reduce these dangers, increasing the survival of young and thus alter a flock's population profile. Origin of the Southwestern Turkey There has been much controversy over the origin and descent of southwestern archaeological turkeys. To shed light on this problem an attempt will be made to trace the evolution of the turkey from the appearance of its first ancestors in North America to M.g. merriami, the species which inhabited the Southwest prehistorically and today. Data concerning fossil remains and the available literature will also be reviewed to interpret the turkey's phylogeny. The fossil record provides the earliest evidence for the presence of turkeys in the New World. Unfortunately most fossil remains are fragmentary and represent few elements. As a result, determining the relationship of the identified remains and differentiating between species is difficult. Steadman (1980) hypothesised that the subfamily Meleagridinae, to which the turkey belongs, evolved from a Phasianine origin. The two share several characteristics in common on the furcula, scapula and tarsometatarsus. It is probable that the Meleagridinae branched from the Phasianine after the latter migrated to the New World via the Bering Strait land bridge. This hypothesis is supported by the lack

46 of Meleagridinae fossils in the Old World and the presence of Phasianine characteristics in some fossils on the turkey lineage (Steadman 1980). Steadman (1980) revised the taxonomy of the turkey subfamily Meleagridinae by recognising only three genera though previously five had been proposed. The additional two genera are now encompassed within Steadman's three which include Rhegminoris, Proagriocharis and Meleagris. Rhegminoris and Proagriocharis represent ancient forms found only in Miocene and Pliocene contexts. Rhegminoris colobates is the earliest known fossil on the turkey lineage and was recovered from a Miocene deposit in Florida. It expresses traits from both the Meleagridinae and Phasianine. The earliest Meleagridinae specimen was recovered from Upper Pliocene deposits in Nebraska and is designated as Proagriocharis kimballensis. It is not known if this specimen is on the direct lineage to Meleagris gallopavo (Steadman 1980). Fossils recovered from later contexts belong only to the genus Meleagris to which several species have been identified in a variety of geographic locations. Confusion arises when attempting to create a phylogeny for the Meleagridinae subfamily. The only specimens recovered from Pleistocene deposits in the Southwest are the extinct Meleagris crassipes and M. gallopavo (Rea 1980). M. crassipes went extinct between 3,300 and 6,600 B.P., before the appearance of the first M. gallopavo remains in the area, thus ruling it out as a possible progenitor. M. crassipes is also the most distant of all modem and fossilised turkey species recovered from North America according to Breitburg's (1988) shape analysis, which evaluated four measurements on turkey tarsometatarsi. In light of this evidence M. crassipes is discounted as a possible ancestor of M. gallopavo. Breitburg (1988) emphasises that there is no evidence other than that presented above for the presence of Pleistocene turkeys in the Southwest or

47 Mesoamerica. Some fossils, however have been recovered from eastern North America (M. leopoldi and Meleagris spp.) and what is now California (M. californica). The two eastern species were recovered from Irvingtonian deposits in Florida. The progenitors of these specimens are unknown, thus their relationship to M. g. merriami cannot be established. M. californica is thought to have evolved from M. gallopavo populations which became isolated in California due to arid conditions in Arizona and southeast California. Alternatively, it is possible that the two species evolved convergently due to selection by similar environmental conditions (Steadman 1980). Breitburg (1988) also performed shape analysis on measurements of the tarsometatarsi of M. g. merriami and M.g. silvestris. M.g.merriami is the subspecies found in southwestern archaeological contexts and currently inhabits highland forested areas in New Mexico. M. gallopavo silvestris is a modem subspecies from eastern North America. The results of his analysis suggest that the two subspecies evolved from the same parent population, though the common ancestor is unknown. Five subspecies of M.gallopavo have been identified in Mesoamerica and/or the Southwest. These include the two subspecies which have been recovered from southwestern archaeological contexts; M. g. merriami and M.g. tularosa. All five subspecies continue to inhabit the region today with the exception of M.g. tularosa which is extinct. It should be noted that the identification of M.g. tularosa as an independent subspecies has been questioned by some researchers (Breitburg 1988, Senior and Pierce 1989). The remaining subspecies inhabit ranges at least partially within Mesoamerica. M.g. intermedia (Rio Grande turkey) is a small bird which occupies the region between eastern Kansas and central Mexico, including the Yucatan Peninsula (Lang and Harris 1984, Schorger 1966, Steadman 1980). This subspecies appeared in the record

48 following the termination of the Pleistocene. It may have been separated from M. ocellata which occupied the coastal region of north Veracruz and Tarnaulapis, by a wet barrier which arose following the Pleistocene (Steadman 1980). M.g. mexicana (Gould's Wild Turkey) inhabits the southeast corner of New Mexico. Its range also extends southward along the Sierra Madre Occidental to northern Jalisco, Mexico. The final subspecies, M.g. gallopavo (Mexican turkey) inhabits central Mexico south of the ranges of M.g. intermedia and M.g. mexicana and is the species locally exploited by central Mesoamerican cultures (Lang and Hams 1984). The origins of the five subspecies of M. gallopavo have yet to be determined. There is no osteological evidence to indicate that the latter three are related to M. g. merriami or M.g. tularosa. McKusick (1980a) stresses that she can find no characters which relate them to those found at southwestern archaeological sites. As illustrated by the data presented above the details of the turkey's phylogeny are still unclear. In addition there is a lack of archaeological material documenting the domestication process. It is necessary to recover additional fossil evidence indicating the turkey's evolution if a comprehensive phylogeny is to be created. Entrance of the Turkey into Southwestern Lifeways The motivation behind the domestication of animals has been a subject of intense discussion in the archaeological literature. Numerous models have been proposed to explain why it occurred and often focus on stressed populations. In these models adoption of domesticates is seen as a buffer to prevent resource stress. Numerous models citing driving factors such as population pressure, climatic change, sedentism and fluke occurrences have also been advocated. Recently, Hayden (1990, 1992) has proposed an alternate model arguing that

49 stressed conditions were not prime movers. Instead he claims that complex hunterlgatherers with a stable resource base, social inequality, and competitive individuals, domesticated animals as prestige items to be served up during competitive feasts. He supports this model by arguing that many early domestic plants and animals were not adopted as major dietary staples (i.e. bottle gourds and dogs) and others did not become so until they had already been domesticated for long periods. Keeping the alternatives in mind, current models which seek to explain the entrance of turkeys into Anasazi society are reviewed below. No attempt will be made to explain why and where turkeys were adopted by the Anasazi in this research. Still, this issue is directly applicable to the upcoming discussion regarding the turkey's domestication. Presently two major hypotheses exist to explain the pathway the turkey followed to become a domesticate in Anasazi society, yet they differ significantly. The first hypothesis suggests that the turkey was not indigenous to the Southwest, but to Mesoamerica where it was domesticated. It is proposed that it later diffused into the Southwest via various trade routes (Hargrave 1970, McKusick 1980a, 1986). Recently, McKusick has argued that the domesticated turkey (Meleagris gallopavo merriami) moved north as a component of the Upper Sonoran Agricultural Complex (USAC). The USAC included beans, squash, bottle gourd and maize and diffused in waves to the Southwest beginning ca. A.D. 1 (Ford 1981). Trade began with the importation of turkey feathers, which lead to the eventual movement of entire birds. This argument is substantiated by evidence citing the appearance of domestic turkeys in the Southwest ca. A.D years, at the same time that the USAC was diffusing. The hypothesis, however is not supported by archaeological or paleontological data in Mesoamerica as no remains of M. g. merriami's ancestors

50 have been recovered (Breitburg 1988). This, may be a result of the paucity of archaeological data which has been retrieved from the region to date. The alternate hypothesis proposes that M.g. merriami was indigenous to the Southwest and was therefore domesticated locally. Turkeys are believed to have been available prehistorically in the mountain ranges of Colorado, Arizona and New Mexico (Breitburg 1988, Olsen 1968, Schorger 1966). Modern wild populations of M.g. merriami inhabit the montane regions of New Mexico today (Lang and Harris 1984) though this is partially due to human intervention. Breitburg (1988) proposes that the turkey was domesticated by the Anasazi, who initially hunted and eventually raised turkeys from local wild populations. He argues that specimens recovered from Basketmaker I11 contexts were not domesticated, but were wild buds obtained by hunting. It would be interesting to test this hypothesis using stable isotope analysis. It is likely that domestic turkeys were subsisting at least partially on maize, which has a C4 photosynthetic pathway, unlike their wild counterparts. Contributions to the diet by Cq plants can be recognised by examining the 12~:13~ isotopic ratios in archaeological turkey bone (Van der Mewre 1982). Unfortunately this hypothesis has not been tested in this research as no turkey bones from definite Basketmaker I11 contexts were available in the sample. Breitburg (1988) substantiates his argument by drawing on the negative evidence for southwestern turkey ancestors in Mesoamerica and on the existence of ideal environmental conditions for turkey habitation in the territory occupied by the Anasazi. This claim can be proven no more than McKusick's argument for domestication in Mesoamerica. In neither area does archaeological or paleontological data provide evidence for an ancestor of M. gallopavo merriami. The presence of modem populations of M.g. merriami in Anasazi territory lends some support for Breitburg's argument, but it has been postulated that these may

51 be the descendants of domesticated birds which went feral in prehistoric times, or following the abandonment of the Anasazi region ca. A.D (McKusick 1980a, 1986). In light of the data presented above, it is impossible to be sure of the origins of turkey domestication in the Southwest. The recovery of evidence indicating the presence of ancestral Meleagris gallopavo populations may aid in solving the debate, but until then it remains unresolved. Southwestern Species and Breeds Turkeys recovered from southwestern archaeological contexts are assigned to the subspecies Meleagris gallopavo merriami, which inhabits the region today (Schorger 1966). In 1961 Schorger identified a second subspecies following the examination of a mummified turkey from Tularosa Cave, New Mexico (Schorger 1961). This specimen possessed several diagnostic features which differentiated it from M.g. merriami. Unfortunately the majority of these features are perishable (i.e. feather colouration and scales on the tarsometatarsus) and thus only survive under unusual conditions. One major osteological difference, however, was recognised: the length of the tarsometarsus was noticeably shorter in the Tularosa specimen. Approximately six years later Schorger located a second mummified turkey with identical characteristics at Canyon del Muerto, in Northeastern Arizona (Schorger 1970). He approached McKusick and Matthews, two researchers studying similar specimens at the Southwest Archaeological Center who confirmed the presence of the diagnostic features. After establishing the differences Schorger proposed a new extinct subspecies, M.g. tularosa (Tularosa's turkey). McKusick (1980a, 1986) expanded on Schorger's work by developing her own classification for the southwestern turkey groups. She identified three

52 groups as follows: a) the Small Indian Domestic Turkey (M.g. tularosa.), b) the Large Indian Domestic Turkey (M.g. merriami ) and c) Merriarn's Wild Turkey (also M.g. merriami). By examining osteological characteristics on numerous turkey skeletons, from all regions in the Southwest, McKusick concluded that it is possible to differentiate between the two subspecies mentioned above and also between the wild and domestic forms of M.g. merriami. She claims that the three groups may be identified by the recognition of specific characters on the mandible, pelvis, coracoid, scapula, humerus, carpometacarpus, femur, tibiotarsus and tarsometatarsus. A few metric measurements are also reported to substantiate the osteological characters. The following prehistory of the three turkey races is proposed by McKusick (l980a, 1986), but is not agreed upon by all researchers (i.e. Breitburg 1988, Senior and Pierce 1989). The Small Indian Domestic (SID), the subspecies which Schorger identified at Tularosa Cave, may be recognised by its 'small stature and short tarsi. It fust appeared in the Southwest in west-central New Mexico between 300 and 150 B.C. The establishment of the SID was concurrent with the initiation of agriculture, thus it is believed that it may have diffused along the same trade routes as the agricultural complex. The SID predominantly occupied the peripheral areas of the Southwest until its extinction when the Pueblos fell to the Spaniards ca. A.D McKusick contends that it functioned primarily as a ceremonial species, not as a dietary item as SIDs do not appear as food refuse in the archaeological record (1980a). The Large Indian Domestic (LID) is the most frequently recovered turkey subspecies from southwestern archaeological sites, though it was adopted much later than the SID. McKusick proposed that it first appeared in Anasazi Basketmaker sites ca. A.D After its introduction it rapidly increased in popularity and soon outnumbered the SID's, though the SID continued to be

53 raised in peripheral areas. McKusick suggests that LIDs required a lower energy investment than LIDs and thus once introduced it was adopted by much of the population. Still, the groups who had originally invested energy into the SID continued to raise them. The LID served a predominantly economic function as a dietary component, and as a raw material source for tools and jewelry. The last record of LID flocks appeared in A.D. 1723, shortly before the population went extinct. Merriam's Wild Turkey (MWT) is the wild form of M.g. merriami. It is believed to be a descendant of LID populations which went feral shortly after their arrival in the Southwest (McKusick 1980a, 1986). The presence of the large, robust MWT in the archaeological record has been interpreted as the result of hunting, while the domestic forms SID and LID are believed to have been maintained by human populations, hence the osteological differences. Other researchers have explained the presence of MWT populations as the result of turkeys returning to the wild at the time of the Spanish Conquest in the sixteenth century (Hayes and Lancaster 1984, Rea 1980). The majority of faunal reports which delved into the question of turkey domestication following McKusick's reclassification of southwestern archaeological turkeys cite her work, yet many analysts have had difficulty recognising the three groups (Bertram 1991, Breitburg 1988, Senior and Pierce 1989). McKusick (1986) admits that not all researchers will be able to perceive the characteristic morphological traits which differentiate the groups. As a result the method is highly subjective. Often researchers prefer to rely on measurements as they are less subjective, however few have been produced to distinguish the three groups. In sum, though some researchers claim to have met with success using McKusick's methods (Lang and Harris'1984), many are beginning to question their validity ( Bertram 1991, Breitburg 1988, Senior and Pierce 1989).

54 Senior and Pierce (1989) have recently disputed McKusick's divisions based on their analysis of the Homol'ovi I11 faunal assemblage. An attempt was made to fit three adult turkey skeletons into McKusick's scheme by identifying the differentiating characteristics, yet their examination met with mixed results. The majority of their measurements fit into the ranges for the LID and most elements in the skeletons exhibited LID characteristics, however certain elements (i.e. coracoid and carpometacarpus) from the same skeletons shared traits with the SID. As an alternative they postulate that the "variation in turkey morphology may be due to phenotypic plasticity in response to the environment rather than the result of genetic changes" (1989: 245). Thus, they recognise the presence of patterned variation in many of the turkey's osteological characters, but they do not attribute it to speciation. It must be taken into account that this research was based upon a small sample size, and the environmental effects on turkey bone have not yet been tested. Still the idea is an interesting one and should not be discounted. A similar view has been presented by Breitburg (1988) who performed shape, size and principal component analysis on turkeys identified by McKusick (LIDS from Casas Grandes and SIDs from Gran Quivira). His results indicate that it is impossible to discriminate statistically between the SID, LID and MWT. Instead he hypothesises that the size reduction in the Gran Quivira population may have occurred as a result of isolation which prevented gene flow between the sample and parent wild or domestic populations. Size reduction may have been further emphasised due to environmental or nutritional pressures (Breitburg 1988). Perhaps the smaller turkeys were inhabiting areas with a limited water supply and adopted a smaller body size accordingly. Alternately, they may have been nutritionally stressed as a result of their captive lifestyle. Food supply would have been controlled by their captors and may have been insufficient to

55 promote full growth. Like Senior and Pierce (1989), Breitburg does not attribute osteological variation within the southwestern turkeys to different subspecies, but to variation within a single species. The Function of Southwestern Archaeological Turkeys The function of the turkey in Pueblo society has long been debated by archaeologists. Two sources of contention figure prominently in this debate. Did the turkey play a primarily ritual or economic role and was it predominantly used as a feather or food source? It has been argued that in some areas the turkey served only a ceremonial function and was domesticated solely for the utility of its feathers for ritual purposes and as a sacrificial animal (Akins 1986, Lange 1951). Alternatively, turkeys may have been raised to serve an economic function, primarily as a meat source and secondarily for their bones and feathers. Turkey bone is an excellent raw material for the manufacture of implements or jewelry while feathers may be utilised in the creation of blankets and clothing. The most probable is a combined approach, suggesting that the turkey was exploited for both ritual and utilitarian purposes (Breitburg 1988, Hargrave 1965, Lang and Harris 1984, McKusick 1980b). McKusick (1980a, 1986) promotes a variation on this hypothesis by suggesting that there were two breeds of domestic turkeys in the Southwest, which served different functions. She proposes that the Small Indian Domestic (SID) was a ritual bird whereas the Large Indian Domestic (LID) played a utilitarian role. An attempt will be made to address the confusion surrounding turkey utilisation by reviewing historic and ethnographic documentation, the archaeological literature and the contextual data from the archaeological record in the Mesa Verde region.

56 Historic Documentation Documents recorded by members of the Spanish expeditions, provide valuable insight into the turkey's role in prehistoric Pueblo culture. It should be noted that the Anasazi were no longer a recognisable cultural group at the time of the Conquest, but had abandoned the Colorado Plateau. It is generally accepted that they were assimilated into the Pueblo cultures to the south (i.e. Hopi, Zuni, Acoma) (Lipe l992b). The domestic turkey continued to play an important role in Pueblo society following this merger, thus the examination of historic and ethnographic records may provide suggestions about its earlier function in the Mesa Verde region. Modern and historic data are not considered direct analogues to what occurred in the prehistoric past. Instead they are viewed as valuable tools for guiding the interpretation of archaeological information. It is also essential to consider the biases of the chroniclers when drawing interpretations from historic documents. It is possible that some researchers selectively included particular details while others were excluded. The turkey may have also served functions which were not recorded by the Spaniards. Many Spanish documents from the sixteenth and seventeenth centuries remark on the presence of turkeys at several pueblos such as Tiwa, Tanos, Zuni, Acoma and the Keresan pueblos, though Hopi is noted as an exception (Reed 1951, Schroeder 1968, Winship 1896). There are references for the exclusive use of the turkey for feathers to manufacture blankets and robes, while others document utilisation for meat products and still others report that the birds were kept for both purposes (Reed 1951, Schroeder 1968). The interpretation varies according to the researcher and the cultural group. In 1540 during the first Spanish expedition, de Coronado reported that the 'Indians' raised turkeys only for their feathers. He also mentions that the Tiwa utilised the feathers in the manufacture of blankets and cloaks as a replacement

57 for cotton which they did not produce (as stated in Reed 1951). The use of turkey as a dietary item is recorded in the chronicles of the second Spanish expedition in the sixteenth century. Apparently, the Piros corralled hundreds of birds for later consumption, though they also utilised their feathers to create blankets (Reed 1951, Schroeder 1968). In 1626 Fray Geronimo de Zarate- Salmeron discussed the multiple role of the wild turkey in Zuni society as a meat and feather source. Feathers were utilised to manufacture cloaks and were presented as offerings which were placed in anthills for good luck while traveling (Reed 1951). Luxan the chronicler of the Espejo expedition reported that the Piros manufactured feather blankets, but made no reference to the turkey's utilisation for food. This exemplifies the presence of incongruous information within the Spanish documents. In a 1598 document don Juan de Onate discussed prayer sticks, which he described as small multi-coloured sticks attached to turkey feathers. He used the word 'idols' to describe the sticks, implying that they had ritual value (Gnabasik 1981). The above evidence indicates that historically turkey feathers had ritual as well as utilitarian significance. According to historic documentation the turkey served multiple functions in both the utilitarian and spiritual spheres of Puebloan life, though the type of utilisation varied extensively between cultural groups. Ethnographic Documentation As with historic data, ethnographic information should be used only to provide suggestions for the interpretation of past behaviour, not as a direct analogue. Ethnographic data are valued particularly for the insight they provide into spiritual significance, a realm which may only be speculated upon when interpreting archaeological information. Several modern Pueblo groups raise M.g.

58 merriami today as a source of food and/or feathers for the manufacture of utilitarian artifacts and ritual objects. The consumption of turkey meat by Pueblo groups has frequently been cited in ethnographic literature from the twentieth century. Turkeys have been listed as food items at several pueblos including Taos (Parsons 1970), Isleta, Laguna, Santa Ana and Cochiti. In some cases wild birds are hunted, though domesticated birds are also maintained at the pueblos (Gnabasik 1981). The Hopi are an exception and do not eat turkeys due to a taboo forbidding their consumption. The bird still figures prominently in numerous ritual events. References to the ceremonial utilisation of turkeys, particularly their feathers is extensive in the ethnographic literature (Gnabasik 1981, Hawley-Ellis 1968, Lange 1950, Parsons 1970, Reed 1951, Schroeder 1968). Turkey feathers are frequently incorporated into katsina ceremonies. They may be attached to masks, various body parts (i.e. nostrils, ears, hands) or props during dances. Following the dances feathers are returned to their respective owners and caches to be stored for future ceremonies (Gnabasik 1981). Turkey feathers are also commonly used as prayerfeathers and as a component of prayersticks. These are manufactured for a variety of occasions and ceremonies and have been documented at Zia, San Felipe, Jemez, Hopi, Zuni and Santo Dorningo (Gnabasik 1981). Prayersticks and prayerfeathers are offerings which are commonly presented at shrines outside of the pueblos (Parsons 1970). Their meaning varies according to occasion, cultural group and location. Unfortunately prayersticks and prayerfeathers are rarely recovered from the archaeological record as they are extremely fragile and are usually 'offered' in exposed areas (Gnabasik 1981). Featherstrings may be fabricated from turkey feathers and are utilised to bind prayersticks during manufacture. Ethnographic sources also cite offerings of individual feathers. For example, Cochiti warriors

59 have been known to make offerings of turkey feathers and corn meal so that they will be more courageous than Apache and Navajo warriors. At Taos Pueblo loose feathers are deposited in irrigation ditches after cleaning to ensure the happiness of the water spirits (Parsons 1970). The spiritual importance of turkey feathers is further illustrated by their inclusion in human graves at the time of burial. Individual turkey feathers or prayersticks are often presented as offerings in graves at both Jemez and Taos Pueblo (Gnabasik 1981, Parsons 1970). Archaeological Evidence Faunal assemblages provide several lines of evidence which may clarify why and how a species was exploited by humans. Indicators in the archaeological record with potential to distinguish between faunal populations utilised for economic purposes and those raised for their ritual value are discussed below. The following indicators provide evidence for the butchery and/or the subsequent consumption of animals: (1) Cutmarks. Cutmarks are distinctive parallel lacerations on bone which frequently result from processing animals. Cutrnarks are not expected to appear on all butchered bone, instead it has been suggested that they may be the result of sloppy or inexperienced work (Lyman 1979). Therefore negative evidence is not necessarily regarded as support for the argument against butchery. Cutmarks in close proximity to the articular ends of long bones suggest that tendons were severed to disarticulate an animal for cooking or eating (Hayes and Lancaster 1984). Elements at important junctions in the disarticulation process tend to receive cuts more frequently. Lang and Harris (1984) include a thorough discussion of the locations on

60 Anasazi turkey skeletons which frequently display cutmarks and their probable functions. These include: a) cuts on the proximal end of the ulna and the radius which result from wing disarticulation, b) cuts on ulna shafts which result from the removal of the secondary feathers, c) cuts on the cervical vertebra which result from removal of the neck, d) cuts on the distal tibiotarsus and proximal tarsometatarsus which result from the removal of the lower leg which does not contain any flesh, e) cuts on the proximal tibiotarsus and distal femur which result from the disarticulation of the drumstick, f) cuts on the proximal scapula and proximal humerus which result from sectioning of the wing and g) cuts on the distal radius which result from the removal of the wing tip which is attached to the primary feathers. (2) Burning. Bones have the potential to char if cooked, discarded in hearths or deposited within structures which are later burned. Because numerous methods may potentially result in burning, the assumption that burned bone indicates food use, should not be made. Localised burning (when only portions of the bone are charred), is a more accurate indicator of food use. This pattern results from roasting articulated flesh over a flame. The ends of the bone are exposed to the heat while the shaft is sheltered by the flesh, thus only the ends become charred (Lyman 1979). Bones may also appear somewhat porous, similar to immature bone, as a result of boiling (Bertram 1991). Accidental boiling is unlikely so it is probable that bones with this appearance resulted from food preparation. (3) Breakage and cancellous bone extraction. Bird long bones may be recovered with their ends broken or cut in an attempt to extract the cancellous bone in their interior. The proximal and distal ends of long bones and the keel of the sternum contain a high concentration of 'cancellous bone and are frequently damaged in archaeological contexts. The exposed spongy bone was

61 undoubtedly valued as a raw fat and energy source or perhaps as a stock for soup (Hargrave 1965). (4) Population Structure. When domesticated animals are raised for subsistence purposes they tend to be culled at a younger age than those utilised for secondary products. It is most energy efficient to butcher an animal as it reaches maximum size. When maintained beyond this point domestic animals require additional energy input, yet gain no further weight (Greenfield 1991, Hesse 1982, Higgs and Jarman 1972, Senior and Pierce 1989). Turkeys reach maximum size just prior to adulthood, with the exception of breeding females and are expected to be culled at this point. It is assumed that some females of reproductive age will be maintained as they are necessary to propagate the flock. Only one male turkey of at least two years of age is needed to ensure reproduction in a flock of hens, thus it is expected that there will be a lower proportion of adult males. (5) Disarticulation and dispersion of bone. The presence of mixed, disarticulated bone representing individuals of a variety of sizes and ages is cited for evidence of meat consumption. Animal skeletons are frequently disarticulated to facilitate meat removal prior to preparation for eating (Hargrave 1965, Lang and Harris 1984, Senior and Pierce 1989). Scattered configurations imply that bones were discarded randomly as waste products after the meat was utilised. This contrasts with the recovery of complete skeletons from deliberately buried birds. (6) Ossified tendons separated from limb bones. The tendons associated with the tibiotarsus and tarsometatarsus in the turkey's lower leg calcify as the bird ages, creating ossified splints (Schorger 1966: 92). These are frequently recovered from the archaeological record. If meat is separated from the

62 lower leg prior to burial the tendons will be removed with it and they will not be recovered in association with the leg bones. Negative or contrary evidence for the features discussed above may imply that birds were raised for feather utilisation. If this is the case evidence for butchery, localised burning and cancellous bone extraction are not expected. Instead the presence of the following features support arguments for feather utilisation: (1) Age and Sex Profiles. Archaeological assemblages which are products of feather utilisation, should be structured with equal numbers of male and female adults. Birds retained past adulthood will continue to produce feathers until death, thus it is efficient to support them past adulthood (Greenfield 1991, Hesse 1982, Higgs and Jarman 1972, Senior and Pierce 1989). Feathers from male and female turkeys are similar, with the exception of the tips of breast feathers which are black in males and brown in females (McKusick 1986, Schorger 1966). It is unknown whether there was a preference for brown or black feathers, which would result in the preponderance of a particular sex in the archaeological record, thus confounding potential interpretation. Later discussion on this issue is based on the assumption that feathers from males and females were utilised equally. (2) Articulated Burials. It is anticipated that the majority of turkeys which were raised as feather producers will be buried intact after death (Hargrave 1965). Still, the possibility that disarticulated turkey skeletons represent birds exploited for their feathers should not be ignored. For example, it is likely that taphonomic factors will affect complete skeletons and may potentially provoke disarticulation. Some elements may be transported out of situ by agents such as rodents or alluvial movement (Medlock 1975, Micozzi 1991, Nash and Petraglia 1987). Skeletons may be disarticulated and scattered prior to burial if they are

63 used for secondary products such as bone for tool and jewelry manufacture. It is also possible that articulated sections of turkeys may be separated for ritual purposes, for example, intact wings or tarsi and feet. The range of possibilities mentioned above may potentially cloud interpretations of dispersal patterns, thus caution should be exercised. (3) Association of ossified tendons with limb bones. In keeping with the assumption presented in the previous section, it is expected that ossified tendons will be recovered in situ with tibiotarsi and tarsometatarsi. This is based on the premise that an intact turkey carcass will decay in situ so the splints will remain associated with the bone. Once again this may be confounded by taphonornic processes. (4) Feather Artifacts. Though feathers are extremely fragile and subject to rapid disintegration, it is expected that feather artifacts will be recovered from contexts with excellent preservation. For example, feather blankets, cloaks and cordage may potentially withstand decay in desiccated locations. Fortunately there are many sites which exhibit these conditions in the Southwest due to the aridity. The possibility that the Anasazi were utilising turkeys for meat and feathers for both economic and ritual purposes should not be ignored. If this is the case, indicators of both are expected to be recovered from the archaeological record. The features discussed above have been cited by many researchers in an attempt to determine the function of turkeys in a particular assemblage (Bertram 1991, Hargrave 1965, Lang and Harris 1984, Senior and Pierce 1988). The prevailing views and interpretations of archaeological data relating to the function of the turkey are presented below. k 8

64 Previous Archaeological Interpretation The majority of interpretations of the turkey's function in the archaeological literature advocate that it was initially exploited as a feathers source, though this evolved to incorporate food utilisation (Breitburg 1988, Hargrave 1965, Lange 1950, McKusick 1980a, 1986, Reed 1951, Schorger 1961, 1966, 1970, Windes 1987). Evidence cited in support of this view includes the paucity of bones recovered from Basketmaker I1 and I11 contexts, though feather artifacts have frequently been recovered from early components (Hayes and Lancaster 1975, McKusick 1986, Morris 1939, Rohn 1977). The transition to food use is believed to have taken place ca. A.D. 900 when turkey bones appear in the record in increasing abundance. The presence of cutmarks, burning and disarticulation convinced many researchers that turkeys were being consumed by this time, though their feathers continued to be a valuable raw material (Hargrave 1965, Lange and Harris 1984, McKusick 1980a, 1986, Schorger 1966). Other research featuring southwestern archaeological turkeys has been performed by Capone and Schoeninger (1991) to determine whether cooking can be detected in turkey bone by utilising the amino acid racemization of aspartic acid as an indicator. Racemization is the conversion of one structure of an amino acid (L-enantiomers) into its mirror image (D-enantiomers), until there is a stable, equal mixture of each. The racemization process is affected by temperature and unfortunately also by post-depositional conditions. Capone and Shoeninger propose that quantifying the proportion of D-enantiomers to L-enantiomers may enable the researcher to determine whether a bone has been heated. Heating bone results in higher proportions of D-enantiomers than unheated bone from the same context. Control samples (non-turkey) and turkey samples from four sites in the Southwest were tested to determine whether heating could be detected in archaeological bone. It was concluded that the duration of cooking in prehistoric

65 times was probably too short to detect by using racernization, as one and a half hours of boiling or baking were required to produce notable differences in amino acid proportions. Unfortunately the length of time for which turkeys were cooked in prehistory can not be determined but it is unlikely that they were cooked for such long periods. Hopefully further exploration will be performed in this area and lead to the discovery of a more effective method. Data supporting or contradicting the presence of domestic turkeys and their function at. Anasazi archaeological sites will be presented in Chapter IV. The following Chapter will concentrate on the theoretical background of the other three problems concerning domestication, intensification and resource access.

66 CHAFTER m THEORETICAL BACKGROUND The theoretical background for domestication, aggregation, intensification and resource access are presented in this chapter. These concepts provide the theoretical framework for later interpretation of the data relating to domestication and spatial and temporal change in turkey production. This chapter also serves as a forum for the presentation of the expectations for the major problems addressed in this study. It commences with a discussion on domestication and the expectations for the study assemblage. Next is a review of recent explanatory models of the aggregation process in the Southwest followed by an outline of settlement change in the Mesa Verde region and a synthesis of the factors which may have promoted aggregation. A final model explaining aggregation provides a hypothesis for settlement change in the Sand Canyon Locality. The chapter concludes with a series of expectations for the structure and changes in the turkey assemblage in relation to this model. These expectations are evaluated against the data from the Sand Canyon Project, Dolores Archaeological Program and Mesa Verde region in the following chapter. The turkey was chosen for study due to its widespread use as a prehistoric domesticated animal in the Southwest. As mentioned previously domestic animals have high utility for archaeological research as they provide a rare opportunity to examine direct remnants of purposeful human behaviour. It is essential to establish that the turkey is in fact'a domestic animal to guide

67 interpretations in this research, rather than making this assumption based on previous studies. Past research has been performed in an attempt to differentiate osseous turkey remains into subspecies and to make wild and domestic divisions. It has been argued that wild and domestic forms of a species may be separated according to diagnostic osteological characters or by analyzing standard measurements (Armitage 1986, Bokonyi 1969, Boessneck and von den Driesch 1978). Much debate has ensued over this claim and many questions require clarification before it may be accepted. For example, what constitutes a domestic animal? Can bones from wild and domestic animals really be differentiated osteologically? If they can be, what processes have led to the changes and are they caused by natural or cultural agents? In order to tackle these questions it is necessary to establish a definition for domestication. The much cited classic definition was created by Bokonyi (1969: 219) and states that "the essence of domestication (is): the capture and taming by man of animals of a species with particular behavioural characteristics, their removal from their natural living area and breeding community, and their maintenance under controlled breeding conditions for profit". This definition has been subject to frequent attack primarily due to its emphasis on the dichotomy of an animal's wild or domestic state. Ducos (1978: 54) counters with a definition of his own which states that "domestication can be said to exist when living animals are integrated as objects into the socioeconomic organization of the human group, in the sense that, while living, those animals are objects for ownership, inheritance, exchange, trade etc., as are the other objects (or persons) with which human groups have something to do." This can be taken farther to a position that states that domestication should be viewed as a continuum along which animals partake in a range of economic relationships with 55

68 humans (Higgs and Jarman 1972, Jarman and Wilkinson 1972, Meadow 1989, Rindos 1984). According to this definition domestic animals may be virtually indistinguishable from wild ones depending on where they fall along the continuum. It should therefore be noted that though animals may have been maintained and controlled by humans they may be identical to wild fauna osteologically. Osteological variation in bone has frequently been used to differentiate species or domestic and wild forms of the same species. Bone preserves better than most other animal byproducts which may be used to separate species (i.e. feathers, fur or voice). Changes in bone are expected as a result of domestication due to selective pressures from humans which are proposed to differ significantly enough from natural selection to be detectable (Higgs and Jarman 1972). Changes in bone resulting from domestication have also been attributed to living conditions and isolation from the wild gene pool. Because bone is an adaptable and flexible substance, changes in body form often leave their signature on the skeleton. Size diminution is frequently cited as a mark of domestication resulting from poor nutrition and/or living conditions (Annitage 1986, Higgs and Jarman 1972, Meadow 1989). Small animals may be the product of domestication since the diversity of food which is available to them decreases and they are often forced to forage in overgrazed areas (Meadow 1989). Owning more animals with smaller volumes also benefits the herder as the animals can disperse over a greater foraging area to reduce overgrazing (Jarman and Wilkinson 1972). It should be noted that size reduction may also occur as a result of natural causes such as warming climates, thus diminution does not necessarily imply domestication (Jarman and Wilkinson 1972, Meadow 1989; Olsen 1979). Osteological change may also occur as a result of isolation which limits 56

69 gene flow within a small group. Populations become isolated when they are prevented from coming into contact with wild members of their species due to human intervention. The resulting reduction in gene flow will separate domestic animals from wild ones since subsequent mutations and genetic change will not be shared, causing the two groups to diverge (Breitburg 1988, Jarman and Wilkinson 1972). Once again isolation may occur naturally so it should be apparent that diminution can be seen neither as a definite nor necessary indicator of domestication. Osteology is not always a reliable indicator of domestication, therefore it is necessary to examine other types of archaeological data to determine an animals' wild or domestic state. In the case of the turkey, alternate sources of data include eggshell and gizzard stones. Eggshell may have been introduced to prehistoric sites through the collection of wild turkey eggs or if an adult population was breeding in captivity. Gizzard stones are ingested by turkeys to facilitate food digestion in the craw (Schorger 1966). They may be identified by their rounded edges, high polish and small surface peck marks. Evidence for the ingestion of debitage and ceramic sherds for use in turkey gizzards implies that turkeys were inhabiting human sites (Bertram 1989, Breitburg 1988, Rohn 1971, Windes 1986). Only by foraging at sites either while they were occupied or following occupation by humans could the birds ingest these discarded cultural objects. The population structure of a faunal assemblage may also indicate whether an animal was domesticated by providing information about the ages and sexes of the turkeys which were selected for culling (Greenfield 1991, Hesse 1982, Higgs and Jarman 1972, Senior and Pierce 1986). The recovery of juvenile or immature bones has similar implications as eggshell since it is unlikely that poults were hunted in the wild for their economic value. It is probable instead that they were hatched under confinement. Architectural evidence such as pens or other 57

70 retaining structures clearly indicate that turkeys were maintained at a site. Pens are enclosures coated in layers of consolidated turkey droppings, and are often located in abandoned rooms. It is unknown if turkeys were permanently enclosed within the pens or if they were able to forage during the day and were restrained at night to prevent their escape (Cattanach 1980, Rohn 1971, Schorger 1966). It cannot be determined archaeologically whether the Anasazi clipped their turkeys wings to prevent flight. Clipping results in the removal of one or more of the primary flight feathers which articulate with the ulna. There is no need to cut the feathers during removal, but instead they can be pulled out without maiming the skin or bone. If clipping was performed it would be difficult if not impossible to recognize osteologically. In this research it is hypothesised that the turkey sample in the Sand Canyon Locality represents a domestic population, though this may not be apparent by using osteological criteria. As discussed in Chapter 11, most previous research concerning southwestern archaeological turkeys advocates that the distinction between breeds and wild and domestic groups may be recognised by identifying certain diagnostic characters on bones. This claim will be reexamined in this study. The utility of measurements will also be examined to determine their effectiveness for separating breeds. Guidelines for the separation of species were drawn from McKusick's research (1980a, 1986) and measurements as defined in Chapter I have been taken to test this method. In addition evidence for alternative indicators of domesticity drawn from contextual data will be recorded from sites in the Mesa Verde region and the Sand Canyon Locality. This information will be provided in an attempt to identify domestic populations if they cannot be recognised by using osteological data. The results will be presented in the following chapter. Once it has been determined whether the turkeys in the sample were 58

71 domestic or wild it will be possible to evaluate broader social and economic problems. The remainder of this chapter is dedicated to the presentation of the background which will be necessary to interpret temporal and spatial variation in the intensity of turkey production within the faunal assemblages. This section begins with a review of the concept of intensification and the aggregation process in the Mesa Verde region. The effect of settlement changes on the use of local turkey populations will be examined within the context of intensity (or intensification). Intensity is "...the amounts of population, material, information or energy use per unit area or per capita" (Lipe 1992a: 5) as defined for the research performed at the Crow Canyon Archaeological. Intensification is often used as a tool to measure sociocultural development and cultural complexity (Adler 1990a). Generally, as the intensity of the components within a sociocultural system increase the system becomes more complex. This may create a greater demand for resources within the socioeconomic system and lead to higher costs for their procurement. When demands for additional resources are increased, human groups will adopt various strategies as risk prevention mechanisms, such as diversifying the subsistence base or intensifying modes of production already in use (Minnis 1985). These strategies generally require greater inputs of labour for production and processing and are often employed on less productive resources (Lipe et al. 1991). As previously discussed, (see Chapter I) intensification in this project is viewed as an increase in the energy invested into turkey production per capita. Energy investment takes place in the form of labour. Maintaining domestic animals requires a greater labour investment in exchange for a reliable food source (Earle 1980). Documenting the changes in reliance on domestic animals is 59

72 therefore a means of estimating the degree of intensification throughout the sociocultural system. In this research intensification of turkey production will be determined by examining changes in the proportions of turkeys in the sampled faunal assemblages. This is based on earlier discussions which indicate that evidence for increased consumption reflects intensification. Numerous changes in southwestern Colorado between A.D. 200 and 1300 influenced the turkey's role in Anasazi society. Among these, shifts in community organisation may have had a prominent effect on the intensity of turkey exploitation. Aggregation occurs when a settlement pattern undergoes a transition from small dispersed sites to large, concentrated groups (Adler 1990a). Increasing population, decreased mobility and intensification of production are believed to have promoted aggregation (Kohler 1992). These factors'were also affected by fluctuating local climatic conditions (Dean et al. 1985, Schlanger 1988). In southwestern Colorado the Anasazi underwent gradual change from a mobile hunter gatherer lifestyle to the habitation of dispersed sedentary settlements to complex aggregated communities based on mixed agricultural economies. The process commenced uniformly throughout the region, though variation existed between localities as it progressed (Cordell 1984, Varien et al. 1994). Aggregation occurred twice, once in the late A.D. 800's and again in the A.D to 1280 period. Each episode was preceded by an interval of dispersal and followed by a regional depopulation or abandonment (Varien et al. 1994). The aggregation which characterised some areas in the Mesa Verde region in the late A.D. 800's (e.g. the Dolores Valley) is absent in the Sand Canyon Locality. This discussion will focus on the changes in settlements in the Mesa Verde region, 60

73 and will provide a local example of aggregation from each of the two episodes described above. Before the impact of the aggregation process on turkey production may be evaluated, background information on aggregation in the Southwest and its causes will be reviewed. Southwestern Models of Aggregation A plethora of hypotheses have been offered as explanations for Anasazi aggregation. These range from models which claim that aggregation results from only one factor, to those that propose multivariate explanations. Unicausal models include those which attribute the cause of aggregation to climatic stress, defense or leadership. More integrative models argue that aggregation was the product of environmental depletion due to high population densities and/or as a result of competition reduction strategies. A brief summary of these models is presented below. Environmental Deterioration Models citing climatic fluctuations as the impetus for aggregation suggest that aggregation transpired as a result of subsistence stress due to environmental deterioration. Poor climatic conditions caused periods of drought and forced populations to amalgamate around the few remaining water sources and to intensify production as a buffer against depleted conditions (Adler 1990a, Hill and Trierweiler 1986). Dean et al. (1985) and Schlanger (1988) cite climatic shifts as the trigger for human adaptive mechanisms which attempt to offset subsistence stress. By shifting population locations and intensifying agriculture during times of environmental stress the Anasazi compensated for decreases in productivity caused by periods of low rainfall. Agricultural intensification was often associated with aggregated communities as the concentrated population 61

74 provided a greater labour pool and allowed for the localised exploitation of optimum land (Hill and Trierweiler 1986). Defense Arguments for aggregation as a mechanism for defense have long been proposed in the Southwest. It was originally suggested that the Anasazi were forced into aggregated communities for protection against enemy peoples (Kidder 1924 in Wilcox and Haas 1991). 'The enemy' were identified as Numic speaking or Athabaskan hunter-gatherers who raided the more permanent Anasazi villages. This view was later replaced by one which promoted conflict between neighbouring Anasazi communities as the cause for warfare (Haas and Creamer 1990 and others). Recent research by Wilcox and Haas (1991) suggests that the formation of aggregated villages in the northern Southwest was largely influenced by the presence of warfare. Populations are believed to have amalgamated as a defensive mechanism which operated by providing protection through 'strength in numbers'. Environmental deterioration is cited as the primary cause for warfare which was initiated as a result of competition over depleted resources, due to environmental deterioration. The scarcity of food led Anasazi villages to raid their neighbours to ensure adequate supplies, this in turn led to the aggregation of populations for protection against raiding. Evidence cited as indicative of warfare includes 1) architecture such as stockades, forts and towers, 2) artifacts, particularly shields, 3) osteological remains, for example mutilated or crushed human bone or their direct association with artifacts such as projectile points and axes and 4) rock art. Leblanc (1978) has put forth a similar argument to explain settlement change in the El Mono Valley in New Mexico. In approximately A.D the basic settlement pattern in the valley changed from one of tight groups of small 62

75 villages located in areas with high visibility to one with sites surrounded by high unbroken outer walls, which Leblanc suggests were designed for defense. He believes the change in settlement pattern stemmed from the presence of warfare in the valley which developed following a population influx due to the paucity of arable land in neighbouring areas. The resulting dense population led to conflict over resources and raiding which forced the inhabitants into defensive sites ca. A.D Kintigh (1985) furthered Leblanc's research by confirming the location of some but not all sites in easily defensible locations. He suggests that local groups may have responded differently to warfare, thus some chose new site locations in defensible areas while others did not. Similarly sites in highly inaccessible locations in the Kayenta region have been designated defensive and used as evidence to promote arguments for warfare (Haas and Creamer 1990). An alternative claim by Cordell et al. (1994) contends that the evidence for warfare on the Colorado Plateau has been ambiguous at best, and it has been suggested that the location of so-called 'defensive sites' may have been more of a hindrance than a help during attacks (Rohn 1989). Leadership It has been argued that the existence of an elite was essential to the maintenance of aggregated communities and the development of intensification (Upham 1982, Lightfoot 1984). Advocates of this model suggest that elites were necessary to control social systems which were essential for the operation of large aggregates. Elites had preferential access to resources and directed redistribution to the remaining population. The presence of elites is therefore expected to draw the population into aggregated sites to facilitate their inclusion within the system of redistribution. Difficulties arise when attempting to prove this argument 63

76 archaeologically. There has been much debate over the presence of an elite in Anasazi society. Braun and Plog (1982) contend that societies in the northern Southwest had an essentially 'non-hierarchical' social organisation whereas Lightfoot and Upham (1989) and Johnson (1989) claim that Anasazi society was stratified. The leadership model assumes that Anasazi society was stratified, though this has not been established by the archaeological community. Exchange Increasingly complex exchange networks are closely linked to leadership models of aggregation. Factors such as increasing population density, agricultural intensification and reduced mobility are believed to trigger the development of exchange networks as a buffer strategy. Trade alliances gave local populations access to peripheral resources which were previously unavailable (Plog 1986). The development of more complex trade networks is hypothesised to be directly associated with aggregation and the presence of elites (Braun and Plog 1982). Elites are considered responsible for the formation and maintenance of exchange alliances which connected the aggregated centres and the smaller settlements with larger villages. As a result the presence of elites had the potential to attract the surrounding population to the aggregated centres. Competition Reduction Hunter-Anderson (1979) proposes that aggregation occurred in the Cochiti Reservoir area, New Mexico to prevent competition over valuable resources. She claims that aggregation was associated with climatic events, though it was not necessarily caused by them. Population increases due to migration from the north resulted in the intensification of agriculture and the decreased availability of productive land. This created the potential for

77 competition over limited resources. Hunter-Anderson claims that it was more economical for local inhabitants to withdraw from the area rather than engage in warfare, as the expense of intensifying production was less than that of conflict. As a result, this settlement strategy resulted in the development of buffer zones which surrounded aggregated communities and could be used to obtain resources without interfering with neighbouring groups. Population Pressure Kohler (1989, 1992) promotes a similar argument to Hunter-Anderson's by claiming that population growth led to the risk of competition. Population growth is cited as the instigator of agricultural intensification as it created a scarcity of productive land. Agricultural intensification required a higher energy investment and created greater risk in the case of crop failure, as energy was predominantly invested in the production of fewer resources. subsistence risk led to competition and, therefore, aggregation was necessary to control access to limited resources. Like Hunter-Anderson's model, aggregation was a strategy to eliminate competition as the community restricted the use of resources to its members. The Aggregation Process on the Colorado Plateau Prior to A.D. 1 during the Archaic Period, the Colorado Plateau was inhabited by mobile hunter-gatherers. Population densities were low and dispersed over the productive regions of the landscape. The introduction of maize agriculture at an equivocal date of 1,000 B.C (Vierra 1990) is argued to have promoted the adoption of an increasingly sedentary lifestyle (see Minnis 1985 and Wills 1988 for alternate argument). Following A.D. 1 there was an increase in pithouse structures which functioned as domiciles, though the 65

78 inhabitants probably continued to be seasonally mobile (Gilman 1987). Small, dispersed pithouse settlements consisting of two or three structures developed (Matson 1991) and campsites established to exploit seasonal resources were frequently utilised. The pithouse dwellers emphasised agriculture as their primary mode of subsistence though they maintained hunting and gathering as dietary supplements (Decker and Tieszen 1989, Minnis 1989). In the late A.D. 500's the Basketmaker I11 period commenced. The settlement pattern changed as subterranean pithouses became more substantial and non-contiguous surface structures appeared (Bullard 1966). In some areas in the Mesa Verde region small-scale aggregation was initiated, creating a settlement pattern of loose, dispersed clusters which were perhaps the first villages (Adler 1992a, Wills and Windes 1989). These villages contained greater numbers of pithouses and larger populations than hamlets, which are small settlements composed of a few households. Integrative features in the form of Great Kivas also made their appearance, though they were not yet common. Rises in population correlated with increasing sedentism which may have resulted from an increasing dependence on agriculture. Because of the importance of agriculture there is a strong relationship between deep, arable soils and the location of habitation sites in the Mesa Verde region (Adler 1989). The best soils were situated on the flat mesa tops where the majority of settlements were also located (Adler l992a). By A.D. 750 the Pueblo I period was initiated. During Pueblo I (ca. A.D ) an architectural transition occurred within the pithouses, which may be related to function (Varien and Lightfoot 1989, Wilshusen 1989). Pithouses underwent architectural transformations into structures known as kivas. Simultaneously, the Anasazi adopted above.ground buildings for living areas and storage. The surface rooms formed small contiguous roomblocks to the north of 66

79 each kiva, composing unit pueblos characteristic of this time period (Hayes and Lancaster 1976, Prudden 1903). It was also common for two or more units to coalesce and form small villages. The number of great kivas increased to integrate the larger population. Villages were the predominant settlement type, though the occupation of smaller hamlets continued. In the Sand Canyon Locality there was a paucity of sites due to increasing aggregation to the north and east in the Dolores area, which attracted much of the local population. As a result the settlement pattern remained dispersed with small-scale clustering. Sites in the Sand Canyon Locality at this time had the highest degree of association with arable soil during the entire Anasazi occupation (Adler 1992a). Pueblo I1 was initiated ca. A.D. 900 and persisted until A.D The Anasazi were at their widest geographical extent, occupying all habitable topographic areas (Lekson 1992). Pueblo I1 may be subdivided into Early and Late periods, with Early Pueblo I1 dating from A.D. 900 to 1050 and Late Pueblo I1 continuing until A.D Little is known about the 900s in the Mesa Verde region. There was a massive depopulation of the area, though it is unclear what happened to the inhabitants. In the early 1000's population was on the rise again as evidenced by the appearance of sites such as Gnatsville and the sites in the South Canal area (Kent 1991, Kuckelman and Morris 1988). Late Pueblo I1 (A.D ) is characterised by the appearance of Great Houses, believed to be Chacoan outliers, which were situated in areas previously occupied by dispersed communities (Hallasi 1979, Varien et al. 1994). The Great Houses exerted influence until ca A.D when the Chacoan system declined. Peak population attainment and subsequent abandonment of several communities peripheral to the Sand Canyon Locality occurred during this period. Population within the Sand Canyon Locality however, continued to increase and large-scale aggregation began. In early Pueblo I1 (A.D ), settlements were 67

80 dispersed, but a marked increase in population led to clustering, larger sites and an increasingly aggregated population. Site groupings were evident in two locations in the Sand Canyon Locality and became more apparent throughout the period. These include the Goodman Point Community and the Sand Canyon Community. The association of sites with the best arable soils declined somewhat as aggregation progressed (Varien et al. 1994). Pueblo 111, which commenced ca A.D. 1150, was the culmination of the Anasazi habitation in the Mesa Verde region. It terminated with the abandonment of the area by ca A.D In the Sand Canyon Locality, population increased, as did site size, while agriculture was further intensified. The beginning of large-scale aggregation was reflected in the appearance of multiroomblock sites, and the close proximity of sites to one another (Adler 1989). Between approximately A.D a major shift in settlement location took place. Several mesa top sites were abandoned and the population relocated to the canyon rims, talus slopes and canyon benches. The movement correlated with the founding of Sand Canyon Pueblo, a large aggregated site at the head of Sand Canyon. The shift in population created a further decrease in the association between site location and arable soil (Adler 1994). This may have been an adaptation to land stress (the move freed up productive soils on the mesa tops), or an attempt to access run-off water from the mesas (Schlanger 1988). Abandonment of the Sand Canyon Locality occurred at the end of the thirteenth century. The Processes Behind Aggregation in the Sand Canyon Locality An intensive study of the aggregation process in the Sand Canyon Locality has been undertaken by Adler (1990a). He creates a comprehensive hypothesis explaining why and how the process occurred by emphasising the role of the 68

81 community and the development of social control over access to valuable resources (land tenure systems). The community is the spatial unit within which aggregation is examined. It encompasses the sites and clusters of sites which are integrated by public architectural features such as great kivas or great houses (Adler and Varien 1991). Using this definition, two communities have been identified within the Sand Canyon Locality: the Goodman Point Community and the Sand Canyon Community. These are recognised as independent entities due to the clustering of sites surrounding integrative facilities and their separation by hinterlands with low population density. The communities are named after the two large sites, Sand Canyon Pueblo and Goodman Point Ruin, which were constructed at the culmination of aggregation in the Sand Canyon Locality. The community examined in this project is the Sand Canyon Community. Adler and Varien (1991) argue that it existed prior to the development of Sand Canyon Pueblo in the mid-thirteenth century. During its early stages (ca A.D. 1000) it was composed of dispersed habitation sites, which later aggregated into large, densely populated sites. Though households within the community were relatively mobile, Adler and Varien (1991) argue that the community may be recognised archaeologically by continuity in the location of public architectural integrative features such as kivas and plazas, and the clustering of the settlements during the A.D period. Factors which played an influential role in promoting aggregation include population increases, restricted mobility, resource scarcity and agricultural intensification (Adler 1990a). Each of these will be discussed and supplemented with additional research from the area.

82 Population Increases Increasing populations require the procurement of additional resources to ensure the survival of the group. A 1992 survey of the Upper Sand Canyon provided ample evidence for local population increases. Adler (1992b) developed population estimates by assessing room counts for each habitation site and multiplying them by an expected number of inhabitants. Temporally, he creates four sub-periods within Pueblo I1 and Pueblo I11 (A.D , , , ). The results, representing average, momentary populations, illustrate definite increases in population density through time. Studies undertaken by Dean et al. (1985) and Schlanger (1988) corroborate Adler's reconstruction of population growth. By examining the issue from a regional perspective both studies conclude that there is a general trend toward increasing populations from A.D. 1 to 1100 in the Mesa Verde region. Following A.D it is more difficult to characterise population patterns as there is much local variation. Growth persisted in some areas though population began to decline in others leading to eventual abandonment. The Sand Canyon Locality was one area which continued to prosper and grow until ca. A.D when the area was rapidly abandoned (Adler 1992b). Restricted Mobility A viable solution to resource stress is movement to areas of higher productivity. The utilisation of shifting agriculture in the Mesa Verde region resulted in the depletion of plots of land of their nutrients, thus forcing the inhabitants to move to new areas to give the worn land an opportunity to replenish itself (Kohler and Matthews 1988, Lekson 1992, Schlanger 1988). Mobility allows for the procurement of adequate resources without intensifying production (Kohler and Matthews 1988). It may, however, have ceased to be an 70

83 option in the Mesa Verde region at some point (Dean et al. 1985). Mobility restrictions result from factors such as environmental, geographical or social circumscription, increases in population or a combination thereof. Environmental circumscription may be caused by climatic shifts which decrease the productivity of an area. For example, in the Mesa Verde Region climatic change had the potential to reduce the growing season at high elevations, thus promoting migration to lower elevations with longer growing seasons (Schlanger 1988). Geographic circumscription occurs when surrounding lands are not suited for agriculture. As well, decreases in rainfall have the potential to cause crop failure (Adler 1990a). Lastly, social circumscription may result if land ownership is recognised. In the Sand Canyon Locality mobility became restricted due to increasing populations and the farming of peripheral agricultural land. The constriction of the farming belt between ca. A.D and 1275, due to deteriorating climatic conditions further limited the population from moving to adjacent areas (Schlanger 1985). Restricted mobility therefore often forces solutions which will increase yields from resources which are already being exploited, most notably by agricultural intensification. Resource Scarcity It has been argued that aggregation is linked to the degree of resource scarcity (Adler 1990b). Resources utilised by the Anasazi which were subject to scarcity include arable land, wild flora and fauna and water. Constant population growth in the Sand Canyon Locality required that additional resources be extracted from the environment to support higher populations. As a result the depletion of valuable resources such as land' occurred. Resources fluctuated in response to the dramatic and unpredictable nature of the climate on the Colorado 71

84 plateau. (Dean et al. 1985, Schlanger 1988). It was necessary to adjust the subsistence system to ensure sufficient animal and plant sources in the diet in response to population pressure and environmental perturbations. Scarcity of resources thus often led to the intensification of available food sources (Minnis 1985). One of the primary resources available to the Anasazi was arable land. As populations increased and agriculture was intensified, more productive land was utilised and became increasingly limited. Farmers were therefore forced to cultivate less arable soil. Resource depletion in the Sand Canyon Community is evidenced by the decreasing association of habitations with desirable land. The association of sites with the most productive soil peaked in Pueblo I and then decreased through Pueblo I1 and I11 (Adler 1990a). This indicates that the premium land was already under cultivation so people were forced to farm less productive land. Van West (1990) has performed a study which models climatic shifts and soil productivity to determine whether the land was capable of meeting the subsistence needs of estimated Anasazi populations by agriculture. She concludes that the population did not surpass the carrying capacity of the land. This assumes unrestricted mobility and an effective exchange system to send products from high to low areas of productivity. Currently, there is insufficient data to determine patterns of regional food exchange. It should be noted that the amount of acreage per person was reduced as time progressed and this may have led to stressful conditions. It is likely that stress resulted at times when populations increased and mobility was restricted. While the stress was not sufficient to wipe out Anasazi populations, it limited the range of mobility of the inhabitants and created the potential for social tension concerning ownership of valuable acreage.

85 The production of domestic animals is expected to intensify in response to resource scarcity. The depletion of wild fauna can lead to an increased emphasis on a local, reliable faunal source to ensure a continued supply of meat in the diet. Agricultural Intensification Agricultural intensification results when the amount of labour utilised to boost productivity per unit area of arable land is increased. It has been suggested that the process is initiated when mobility becomes restricted and groups are unable to offset resource scarcity by relocating to lands of higher productivity (Adler 1990a, Dean et al. 1985, Decker and Tieszen 1989, Schlanger 1988). As a result the inhabitants are forced to further exploit the local environment to derive required yields, thus intensifying production (Boserup 1965). This may be performed by introducing water control features, irrigation, fallow cycles etc. At first agriculture was practiced on the optimum available land in the Sand Canyon Locality. As time progressed this association peaked as most of the best agricultural soil continued to be used and it was necessary for newcomers to occupy less desirable land. This is apparent by Pueblo I1 when the average plot of land under cultivation was less fertile than in previous times indicating a shortage of prime soil for cultivation (Adler 1992a). As a result more energy was expended to derive a successful harvest. The association of agriculturally based habitations with less arable land than the preceding periods can be viewed as indicative of intensification (Adler 1990a). The shift in the location of habitation sites from the mesa tops to the canyon rims in the late twelfth and early thirteenth centuries has also been cited as a strategy to intensify production by exploiting the irrigation potential of the canyons which were situated close to springs and flood waters (Schlanger 1988). Further evidence for intensification is indicated in the archaeological 73

86 record by the presence of certain features. These include water collection areas (i.e. reservoirs and terraces), storage and habitation structures of greater permanence and an increase in the diversity of land utilised for agricultural activity (Kohler and Matthews 1988, Schlanger 1988). Adler (1992a) noted that these features show up primarily post A.D in the Sand Canyon Locality, and cites them as evidence for agricultural intensification. Dean et al. (1985) support Adler's view in their regional analyses of agricultural intensification. They cite the appearance of reservoirs, irrigation facilities and terraces by A.D and suggest that they were widely distributed across the Colorado Plateau by A.D The production of domestic animals is another means by which the subsistence system can be intensified. An increase in reliance on faunal domesticates has similar implications to agricultural intensification as production of both may potentially be controlled. Aggregation Synthesis Adler (1990a) synthesises the factors discussed above to create a comprehensive explanation for large-scale aggregation. As previously discussed the model was developed for the Sand Canyon Community, during Pueblo I1 and m. In the Sand Canyon Locality, continuous growth raised populations to sufficient levels to play a restricting role over the mobility of the inhabitants. Between A.D the area of fertile land in the Mesa Verde region decreased, further lowering mobility potential (Schlanger 1985). Rising populations and the increasing utilisation of productive land resulted in the scarcity of arable land. To ensure the procurement of an adequate food supply the Anasazi intensified their food production. This is evidenced by a population 74

87 shift from sites on the mesa tops to sites occupying talus slopes and canyon benches. The move placed new settlements in close proximity to essential, reliable water supplies such as springs and the run-off from mesas (Adler 1990a, Schlanger 1988). Smaller groups were attracted to localities with aggregated populations partially due to their proximity to the few reliable water sources. Archaeological evidence for the aggregation of the Sand Canyon Community is abundant in the locality. The early stages of aggregation were marked by the appearance of large, multi-roomblock pueblos with many more architectural components than previous settlements. Site size also increased from an average of 6 to 9 room roomblocks in Pueblo I1 to more than thirteen rooms during Pueblo I11 (Adler and Varien 1991). Prior to A.D only 20% of the population lived in sites with three or more households, whereas 40% of the population resided in sites with more than fifteen rooms by A.D (Adler 1990a). Sites in late Pueblo I1 also tended to be associated with reliable water sources. In addition, site density became higher as is expected in aggregated situations. Ca many small settlements amalgamated to form Sand Canyon Pueblo which was also surrounded by cluster of small sites (Lip 1992b). Aggregation in the Dolores Valley The Anasazi occupation of the Dolores River Valley was intensively studied between 1978 and 1985 as part of a salvage effort prior to the damming of the Dolores River. Human occupation in the Dolores Valley was rare before A.D. 600, though there is evidence for a few temporary campsites which were occupied seasonally (Kane 1986). By A.D. 600 groups were beginning to migrate into the Dolores Valley and local populations slowly increased. At this time settlements were small and dispersed and site location was selected according to the presence of arable soil. By A.D. 840 population growth 75

88 exploded and large-scale aggregation took place. Much of the population accumulated in large multi-roomblock villages. Shortly after the population reached its peak the Dolores area was gradually depopulated and nearly entirely abandoned (Kane 1986). ource Acc- Resource access is an additional factor which is closely tied to the environmental, economic and social factors of population aggregation. According to Adler (1990a: 18) an intensification in the utilisation of a resource is generally associated with "an increase in the social definition of the means of access to (the) resource". Access refers to who has the right to utilise a particular resource. The primary resource available to the Anasazi was arable land, thus this discussion will focus on the concept of land tenure systems. Land tenure is defined as "the system or rights utilized against other people with respect to resource access and use" (Adler 1992: 1). Land tenure systems form to minimise potential conflict which results from increasing competition over a resource. It has been argued that competition develops due to the presence of the same factors which promote aggregation, that is increasing population densities, reduced mobility and resource scarcity, particularly agricultural land (Adler 1990a, 1992% Kohler 1992). Land tenure acts to formalise access to land within the community to reduce conflict and to protect it from competition from outsiders. By utilising cross-cultural data Adler (1990a, 1992a) argues that the Sand Canyon community had the authority to make decisions over and above the economic units from which it was composed, in effect empowering the community to mediate conflicts over its resources. Land tenure systems with access below the community level may also be viewed as enablers for large-scale aggregation. The creation of a 76

89 'socially negotiated landscape' allows a community to aggregate without conflict because economic units are guaranteed access to productive land. If there was no guarantee it is likely that people would avoid amalgamating in large groups and reside on their plots of land to protect their access. The presence of high population densities in aggregated villages forced agriculturists to travel greater distances between habitation sites and agricultural fields. It was therefore necessary to indicate land ownership to ensure continued access. Field houses are isolated masonry structures composed of one or two rooms with a proposed seasonal occupation. It has been argued that field houses functioned as symbols of land ownership (Kohler 1992). In the northern Southwest their appearance is contemporaneous with the development of soil and water control features which have been interpreted as indicators of agricultural intensification. The development of land tenure systems is viewed as a possible initiator of "differential economic power" which may eventually lead to stratification. It has been argued further that land tenure is a "system of controlled passage of land from one generation to the next" (Kohler 1992: 631). Adler (1990a) suggests that this results from increasing energy investments into individual land plots. As a result, those economic units with access to the most productive agricultural soils may potentially become more wealthy over succeeding generations. In the Dolores Valley, though it is believed that formalised land tenure systems were in effect, it is argued that residential mobility and subsequent abandonment of the area prevented economic stratification from occurring (Kohler 1992). Adler (1990a, 1991, 1992a) proposes that there was a 'general trend' toward a formalised land tenure system in the Sand Canyon Locality during Anasazi prehistory. Using cross-cultural data he argues that agricultural intensification within the locality was the initial impetus for the development of a 77

90 land tenure system in which multi-households were the primary access units. This ensured the protection of the owner's energy investment and reduced conflict over arable land. It is manifested in the archaeological record by the appearance of larger habitations which housed more than one family and are believed to represent primary access groups. Later, ca. A.D multi-roomblock sites blossomed on the Anasazi landscape, housing increasingly large numbers of families and integrating the inhabitants from numerous smaller sites. Adler (1992a) concludes that the Anasazi utilised a variety of agricultural strategies which were most likely associated with distinct access systems. Thus, there is potential for the existence of a variety of access strategies for a single resource. The data utilised to support models of aggregation and investigate intensity in past studies have been largely architectural, environmental or the product of settlement studies and surveys. In this research an attempt will be made to establish the relationship between the aggregation process in the Sand Canyon Locality and domestic turkey production. This will be performed within the terms of the model presented above and the context of intensification. As previously mentioned domestic animals are of great utility for this type of study as they are controlled by humans and thus have potential to provide insight into past human behaviour. It follows that changes in turkey production may also mirror changes within the broader sociocultural system. Intensification Restricted mobility, population increases and resource scarcity create the need for intensified food production as populations no longer have the option of migrating to neighbouring areas with more ubiquitous resources. An 78

91 intensification in domestic animal production often results when there are insufficient quantities of local wild game to meet the animal protein requirements of a given area. Though Van West and Lipe (1992) have argued that there was sufficient arable land to meet the caloric demands of Anasazi populations during all periods of their occupation in the Sand Canyon Locality, this assumes a system of redistribution was in place to transfer resources from areas of high to low productivity. It is argued here that meat was also a valued resource for both nutritional and social reasons. It is suggested that animal protein was an important dietary component regardless of the productive capacity of agricultural land. Corn dominates most archaeological botanical assemblages in the Southwest, yet it is deficient in two amino acids (tryptophan and lysine) and niacin (Nickens 1981). Though beans are able to compensate for these deficiencies by providing the missing components, animal protein is a more complete source of protein than any single plant resource, which must be combined to meet dietary requirements. Several paleopathalogical studies have concluded that the presence of porotic hyperostosis in prehistoric southwestern skeletons was caused by an overdependence on maize and a low meat diet (El-Najaar 1976, Palkovich 1980). Porotic hyperostosis is created by increased blood production promoted by anaemia. It occurs mainly in thin bones and causes the expansion of the marrow cavity. This results in the thinning of the outer layer of bone giving it a porous appearance (Martin et al. 1985). Porotic hyperostosis is believed to be caused by nutritional factors rather than hereditary conditions. These may be influenced by diet, parasites or an unsanitary environment. An overdependence on maize in the diet is suggested as a primary cause of porotic hyperostosis in southwestern archaeological sites. Maize inhibits iron absorption in the body which may potentially promote iron deficiency anaemia. 79

92 Supporters of this view substantiate their argument by linking the increase in porotic hyperostosis with the simultaneous increase in maize dependency between Basketmaker I11 and Pueblo I11 (El-Najaar 1976). Recent research however has revealed that maize production was probably intensified to feed a growing population not to contribute to a larger portion of the diet. The amount of maize consumed per individual during Basketmaker I11 is believed to be similar to quantities consumed in Pueblo 111. This has been supported by isotopic and coprolite studies on the Mesa Verde and Black Mesa (Decker and Tieszen 1989, Matson and Chisolm 1991, Minnis 1989, Stiger 1979). Kent (1986) offers an alternative explanation for porotic hyperostosis by arguing against maize overdependence and attributing poor iron absorption to the conditions created by increasing aggregation and sedentism, which also intensified between Basketmaker I11 and Pueblo 111. She maintains that meat was an important contributor to the diet and should not be dismissed due to the presence of porotic hyperostosis. Kent's argument is based on the premise that aggregation and sedentism promoted unsanitary living conditions which prompted the infestation of parasites and bacteria which also inhibit iron absorption. Anaemia is still argued to be the cause of porotic hyperostosis but it is believed to have been promoted by different conditions. An additional argument can be made for the importance of meat in the social sphere. Kent (1989: 9) asks why "meat and hunting (are) consistently and ubiquitously valued more than plants and gathering or farming" amongst hunter-gatherer and horticultural groups. By performing a cross-cultural examination of the perceptions surrounding the value placed on hunting and animals Kent concludes that they are attributed greater importance due to the presence of an 'intellect'. Wild animals are frequently grouped on a similar intellectual level with humans whereas plants are not. As a result many societies

93 have come to believe that hunting is more dangerous and interesting and requires more skill than gathering, thus attributing it with higher status. Though, Kent suggests that domesticated animals have less value in many societies because they can be controlled and are therefore more likely to be viewed as objects (there are exceptions, especially in pastoralist societies), the value of meat is apparent in many cultural groups today. In light of the above discussion, the assumption is made that meat played an integral role in Anasazi subsistence strategies, though it is unlikely that it equaled plants in dietary importance. If there are demands for resource intensification it is therefore expected that changes in the faunal procurement strategy will result. Speth and Scott (1985,1989) argue that the number of large to small species of wild game often increases in archaeological faunal assemblages as sites become 1) more stable as sedentary communities 2) larger and more aggregated and 3) more dependent on horticulture. Agricultural intensification requires an increased time and labour investment which imposes constraints upon the time available for hunting activities. Agricultural activity is also expected to degrade adjacent environments requiring hunters to travel further to procure meat. Finally, hunts are proposed to occur less often during the growing season due to agricultural commitments. The game sought is therefore expected to be larger and of the highest quality protein. Population aggregation also benefits communal big game hunts as a larger pool from which to draw hunters is available. It has further been suggested that agricultural intensification may promote a reduction in wild game resources due to increased harvesting in areas surrounding habitation sites (Earle 1980, Minnis 1989, Shelley 1993). This takes place when horticultural communities lower the productivity of their environment due to prolonged agricultural activity causing the depletion of game surrounding 81

94 the site (Speth and Scott 1989). Local small game may be overexploited due to high demands for animal protein in heavily populated areas. This often forces hunters to travel greater distances to procure large animal resources. Scott and Speth (1985, 1989) state that an emphasis on big game hunting will occur only if large animals are plentiful in the buffer zones surrounding sites. In this study it is proposed that increases in population density at the regional level created a reduction in hinterland resources due to predation pressure in much of the Mesa Verde Region during Pueblo 111. Hunting pressures from numerous surrounding communities would undoubtedly reduce the availability of big game, thus restricting its potential as a reliable meat source. The conditions cited above (restricted mobility, agricultural intensification, resource scarcity and population increases) which promoted aggregation in the Sand Canyon community are also proposed to have caused a reduction in the availability of large animals therefore reducing big game hunting and forcing reliance on an alternate meat source. Domestic turkeys are suggested to be the ideal resource for intensification in this scenario. Increased input into raising domestic animals may be detected in the archaeological record by examining architectural evidence, the exploitation of secondary products and increased proportions of domestic animals in relation to wild ones within faunal assemblages. Intensification in turkey production is expected to appear architecturally in the form of pens which require a greater energy investment to construct and maintain. It may also result in the exploitation or intensification of secondary products. For example, if the turkey is being exploited primarily as a meat source, intensification may lead to the utilisation of its other economically valued components, such as feathers (Sherratt 1983). Finally, the frequency of turkeys in relation to other commonly exploited wild fauna such as artiodactyls and lagomorphs is expected to rise in relation to 82

95 less intensified times. This is a result of the increased availability and convenience of the domesticated food source and also because of its increase in numbers due to intensification. In sum, the fust expectation relating to temporal change is that the proportion of turkey in the faunal assemblage will increase as the community becomes more aggregated due to the presence of the conditions promoted in Adler's model. Most importantly, regional population density must have reached a high enough level to cause a reduction in hinterland faunal resources. Change and how it was influenced by the prevailing conditions at the time will be monitored by examining the relative proportion of turkey within faunal assemblages throughout the Anasazi occupation in the Mesa Verde region. Two major episodes of aggregation took place in southwestern Colorado during prehistory. An example of each will be examined to determine the relationship between resource intensification and aggregation. The fust occurred 'in the late 800's and is exemplified by settlements in the Dolores Valley, and the second took place in the mid-thirteenth century as depicted by the Sand Canyon Locality. Data from earlier time periods in each location are also considered to detect change between periods of aggregated versus non-aggregated situations, Results are presented in Chapter IV. Intersite Variation in Turkey Production in the Sand Canyon Locality The second expectation is related to resource access and intersite variation in the intensity of turkey production within the Sand Canyon Locality. It is proposed that increasingly defined systems of land tenure may result in differential turkey production between sites. This is based on the premise that though there was enough land for everybody in the Sand Canyon Locality (Van West 1990), it was not equally productive. As a result some groups had access to 83

96 better agricultural land than others. It is expected that sites with access to marginal lands will exhibit higher percentages of turkey as they require a dietary supplement to meet the potential shortfall in agricultural production. This should be reflected in the archaeological record in the form of differential proportions of turkeys between site assemblages. It is difficult to predict which sites will have preferred access to agricultural land as there are no obvious indicators in the archaeological record. If differential access does exist it is proposed that Sand Canyon Pueblo will have access to superior agricultural land assuming that it was a major integrative center. Sites in the Lower Sand Canyon are also expected to have access to high quality land because they are located far enough from Sand Canyon Pueblo (6-7 krn) to be out of the range of its agricultural territory. In the following chapter faunal data will be divided by topographic location in the Sand Canyon Locality to detect intersite variation in turkey production. Data relating to the intensification of turkey production, the turkey's role in Anasazi society and domestication will be reviewed in the following chapter as well.

97 CHAPTER IV RESULTS Introduction To address the hypotheses presented in Chapters I1 and 111, it is necessary to evaluate the applicable data from the archaeological record. Data from the Sand Canyon Locality were collected from the study sample according to the methods reviewed in Chapter I. Information from sites in the surrounding Mesa Verde region was drawn only from the archaeological literature, predominantly site and faunal reports. This chapter opens with a review of measurement and contextual data designed to determine whether the sample contains wild and/or domestic turkeys. Second, the debate over the turkey's function in Anasazi society is considered by examining age and sex profiles, artifacts and other culturally modified bone (i.e. cutmarks, localised burning). Finally, variations in faunal proportions through time and across space are examined to address the questions regarding intensification and social access to resources. Evidence for turkey domestication Before proceeding, it is essential to establish whether turkeys were in fact raised by the prehistoric inhabitants in the Sand Canyon Locality and the Mesa Verde region. Several arguments presented here are based on the assumption that the bones in the sample belonged to domestic animals. In this research domestication has been defined to include all animals which were maintained by humans at archaeological sites. If it can be established that turkeys were inhabiting human sites it can be concluded that at least some of the turkeys examined in this study were domesticated. In an attempt to determine if the turkeys in the sample were wild or.

98 domesticated, applicable elements were examined and measured to detect osteological differences. Additional evidence was sought from the non-osseous remains in the sampled sites. As well, methods utilised in previous research for sexing and differentiating subspecies based on variation in osteological features were reexamined. Osteological data As discussed in previous chapters, McKusick (1980a, 1986) has advocated that the presence of certain features on the turkey skeleton are diagnostic of particular subspecies or breeds. The three groups which can potentially be identified in the study sample include the Small Indian Domestic (Meleagris gallopavo tularosa ), the Large Indian Domestic (Meleagris gallopavo merriami) and Merriam's Wild Turkey (also Meleagris gallopavo merriami). It has been argued that Merriam's Wild Turkey is a large robust turkey, the Large Indian Domestic is similar to Merriam's Wild Turkey though smaller and less robust and the Small Indian Domestic can be distinguished by its short, scaly tarsi and distinct feather colouration (Hargrave 1970, McKusick 1980a, 1986). This study was initiated with an attempt to identify turkey subspecies in the sample by searching for morphological differences between individuals. Complete or near complete specimens of the same element were laid out and intensively searched for variation. The examination did not result in the recognition of any obvious differences. The only noted variation was minimal and inconsistent between individuals. Subsequently, an attempt was made to identify characteristics which were claimed to be diagnostic of a particular breed. In this case, instead of simply searching for differences between specimens, specific, defined characteristics were sought: McKusick's (1986: 35-53) descriptions for the characteristic features were utilised. Upon examination I was

99 unable to recognise any of the characters identified as being diagnostic of a specific breed by McKusick on any Sand Canyon Locality specimen. These observations do not support the validity of McKusick's methods for identifying subspecies and/or breeds, but neither do they dismiss them. It is possible that only one breed is represented in the sample and thus no significant differences could be noted. Variation in this case would be a result of natural fluctuations within a population rather than the differences between two or more populations. Mckusick (personal communication 1992) suggests that the breed represented in the sample is the Large Indian Domestic turkey (the domestic Meleagris gallopavo merriami), which she claims was the most common breed of turkey in the study area prehistorically. I, however could not identify the individuals in my sample as LIDS by using her descriptions of diagnostic morphological characters. As an alternative, measurements were applied in an attempt to determine if more than one subspecies or breed was represented. The possibility that the sample is composed of only one turkey breed was tested by examining measurements of the tarsometarsus. This element was chosen for four reasons. 1) It is the only element which can be reliably sexed due to the occasional presence of spurs on males (Schorger 1966). 2) It is flagged by McKusick (1986) as being noticeably variable in length between breeds. 3) If osteological changes occurred as a result of domestication, the tarsometatarsus has high potential to be affected since domesticated turkeys are expected to adapt to a terrestrial lifestyle, and increase the use of their legs and feet (Breitburg 1988). 4) After performing principal components analysis, shape, size and generalised distance coefficients and multivariate dicrirninant analysis, Breitburg (1988) designated the greatest length measurement of the tarsometatarus the most informative for detecting differences between wild and domestic turkey populations.

100 Before attempting to differentiate turkey subspecies based on size it must be determined if sexual dimorphism exists in the sampled population. This is designed to prevent confusion when attempting to separate breeds on the basis of size, as the different sexes of one breed could be mistaken for members of two different breeds. When graphing a population with distinct sexual dimorphism a bimodal shape should result, with each mode representing one sex. Sex can be assigned using the following criteria. 1) Medullary bone. Its presence definitely indicates that the specimen is female, whereas its absence provides no information (Driver 1982, Rick 1975). 2) Spur on the tarsometatarsus. Its presence definitely indicates that the specimen is male, whereas its absence provides no information (Schorger 1966). 3) Size. According to McKusick (1986) all female prehistoric Southwestern turkeys were smaller than the male Small Indian Domestic turkey. It is therefore likely that any Sand Canyon Locality turkey which was larger than the male SID illustrated by McKusick (1986: 22-33) is male whereas smaller specimens could be male or female. If dimorphism exists in the sample it is expected to be apparent in bar graphs depicting the distribution of measurement data for the tarsometatarsus. The measurements recorded in this project include the greatest length and the breadth of the distal end of the tarsometarsus for all specimens on which either or both measurements could be taken. Most of the bones in the sample were incomplete, therefore more measurements of distal breadth were taken than greatest length (See figure 4.1 for the distribution of distal breadth measurement.). Greatest length measurements of tarsometatarsi are ideal for comparison because most previous research concerning osteological differentiation in turkeys

101 Individuals sexed by size Spurred males Breadth of Distal End Figure 4.1: Distribution of the breadth of the distal end of tarsometarsi in the Sand Canyon Locality. have emphasised this measurement as being the most important (Breitburg 1988, McKusick 1986). Data for greatest length measurements of tarsometatarsi are available to test and compare with the measurements in this research. Greatest lengths are the most difficult measurements to take as the element must be complete and intact specimens are rarely recovered from the archaeological record. To compensate for this problem the breadths of the distal ends were used to estimate the length of elements which were incomplete (see Appendix B). This was performed by plotting a scatter graph of the eight specimens which had both greatest length and distal breadth measurements (Figure 4.2). A simple regression line with a high RA2 value of was plotted on the scatter graph. The resulting equation (y= x) was used to calculate the greatest lengths of incomplete tarsometatarsi, where y=greatest length and

102 x=breadth of distal end. The actual and estimated greatest length data were plotted on a bar graph for statistical examination (Figure 4.3) , y = ~ R Y = C1 m Q) I I I I I Breadth of Distal End Figure 4.2: Simple Regression of Tarsometatarsi Measurements: Distal breadth versus Greatest Length The sample was sexed according to the criteria previously reviewed. No tarsometarsus in the sample exhibited evidence for medullary bone. The spurred (male) tarsometatarsi are indicated in Figure 4.3. The specimens with a greatest length exceeding 132 mm are also probably male following McKusick's data for the male SID turkey (length of illustration). It can be seen from Figures 4.1 and 4.3 that the definite male specimens fall on the right side of the bimodal distribution. It can also be seen that the smallest definite males in the sample are larger than the minimum size proposed for male SIDs as illustrated by McKusick (1986), since they are larger than 132 rnrn. At this point it is concluded that the sample is composed of one population of sexually dimorphic turkeys. It is possible that the odd individual representing a different species is also present. This however, cannot be recognised by using

103 measurement data alone because if other breeds are represented in the sample they fit well into the range of the dominant breed. - Individuals sexed by size Greatest Length in mm Figure 4.3: Greatest length distribution of tarsometatarsi from the Sand Canyon Locality. A second question may now be addressed. Can the breed of the population in the sample be identified? In Table 4.1 ranges of the means of the greatest length of tarsometatarsi in four prehistoric populations separated by sex and identified by McKusick are presented. These data is problematic as McKusick fails to include pertinent information about her sample. For example, sample size and specific ranges of greatest length measurements (only means are given) for each population are not provided. If absolute ranges were given for each turkey group rather than their means we would undoubtedly be able to detect overlap in the ranges between the different breeds. This point challenges the validity of identifying subspecies by using measurement data alone. In order

104 to utilise McKusick's data for comparison it is necessary to divide the Sand Canyon sample into male and female groups and average the greatest length measures from each. These may then be fitted into her ranges of means to determine the breed of the sample population. In addition, sample sizes for each group were not provided, thus they have not been included here. Breed Small Indian Domestic Large Indian Domestic (152.29) (123.04) Merriam's Wild Turkey ( ) (131.02) Gould's Turkey (172.85) (136.80) Table 4.1: Ranges of means of greatest lengths of tarsometarsi by breed and sex. Data adapted from McKusick (198654). Note only one figure is provided for the Small Indian Domestic as only one population was measured. As previously mentioned a greatest length of 132mrn was selected as the dividing line between the sexes as this is the length of the male SID in McKusick's (1986) illustration. It was demonstrated above that the distribution of measurements in the sample correlate well with this division. All definite males (with spurs) fall on the right side of the graph, indicating that the male population occupies the mode with greater lengths. The mean greatest length of the female tarsometatarsus in the sample is 123 rnm while the mean for males is 156 rnrn. Both fall well within the range of means presented by McKusick (1986) for the Large Indian Domestic turkey, the breed which she argues was dominant in the Mesa Verde region. Additional measurement data from Breitburg (1988) were examined to determine how other turkey populations fit irito McKusick's scheme. Breitburg

105 analysed data from three prehistoric Anasazi sites in the Mesa Verde Region; Long House, Mug House and MVS820, Mesa Verde (See Table 4.2). Site Mean of Greatest Mean of Greatest Long House (10) (20) Mug House (5) (13) MVS8320, Mesa Verde (2) (8) (n) = sample size Table 4.2: Means of Greatest Length of tarsi from sites in the Mesa Verde Region. Adapted from Breitburg (1988: ). Breitburg's data from the six samples presented above also fits into McKusick's range of means (Table 4.1) for the Large Indian Domestic turkey. Though the greatest lengths from the sample in this study and others. within the Mesa Verde region (Breitburg 1988) fit within the range for the LID (McKusick 1986), the differentiation of the turkey population into the proposed subspecies is questioned. Means of greatest length measurements of tarsometatarsi taken by Breitburg from prehistoric turkey populations breeds, resulted in a range of different lengths depending on the degree of mixing and the subspecies mixed. Identification of potential breeds by using osteological measurements would be virtually impossible in this case. Instead, following Breitburg (1988) and Senior and Pierce (1989) it is suggested that variation may be the result of environmental factors or isolation rather than genetic differentiation. See Chapter V for further discussion and conclusions. Contextual Evidence The remaining evidence includes non-osseous data, which have the

106 potential to provide information regarding the wildness or domesticity of turkeys. Applicable archaeological evidence was discussed in Chapter 111. Briefly this includes the presence of a) eggshell, b) gizzard stones, c) accumulations of turkey droppings, d) birds of all ages and e) turkey pens or retaining structures. The site reports from the Mesa Verde region and the Sand Canyon Locality were examined for the presence of these items and the findings are reported below. 1) Eggshell. Turkey eggshell has been reported from sites in the Mesa Verde region as early as Pueblo I in Mancos Canyon (Emslie 1977). It was also prolific in the Pueblo I1 excavations from the Hovenweep Laterals Project (Bertram 1991, Moms 1991), the South Canal sites (Kuckelman and Morris 1988) and Gnatsville (Kent 1991). Changes in the quantity of eggshell from the Laterals Project is directly correlated with the frequency of turkey bone in the assemblage (Bertram 1991). Reports of large quantities of shell also come from the Mesa Verde Pueblo I11 sites of Long House (McKusick 1980b), Mug House (Rohn 1971), and others in Johnson Canyon (Nickens 1981). In sum, eggshell appears to be a common item recovered at Pueblo period sites in the region, though it was not found at all sites, such as those excavated by the Dolores Archaeological Program (Neusius 1986) and others with early dates. Large quantities of eggshell were recovered from all sites from the Sand Canyon Locality sample with the exception of 5MT1690 which had a particularly small bone assemblage. Even the earliest excavated site 5MT3868 (Duckfoot) exhibited notable quantities of eggshell, though its turkey assemblage was substantially smaller than all others. The proportion of turkey eggshell to turkey bone was determined for each site within the Sand Canyon Community sample (Table 4.3). Interestingly, the Duckfoot Site (5MT3868), which dates to Pueblo I exhibited the highest eggshell to bone ratio (1.6 fragments of eggshell per bone) with the exception of Mad Dog Tower (5MT181), which had an

107 extremely small sample size (n=3). Perhaps this is due to the minute quantity of turkey bone in an otherwise large faunal assemblage, and/or high fragmentation of a few eggs. Twenty-nine eggshell pieces were recovered and each of these had lengths of less than 2 cm. It is possible that together these fragments represent only one egg. The other sites which date to Pueblo I1 and Pueblo 111 contexts had lower ratios with an average ratio of.84 eggshell fragments to turkey bones. - Site Total Table 4.3: Ratio of eggshell to bone by site in the Sand Canyon Locality. 2) Gizzard Stones. At Pueblo I1 sites excavated during the Hovenweep Laterals Project rounded, pock-marked debitage flakes were interpreted as turkey gizzard stones (Bertram 1991). They were also noted in late Pueblo I1 and early Pueblo I11 contexts at the Mustoe site (Gould 1982) and at Mug House (Rohn 1971). Gizzard stones from the latter site were mainly eroded debitage and were found in concentrated numbers in Room 46A, which has been interpreted as a

108 turkey pen. They were also distributed in random scatters throughout the site (Rohn 1971). Gizzard stones recovered from the sites in the Sand Canyon Locality exhibited a pattern similar to eggshell. They were present in every site in the sample, with the exception of Cougar Cub Alcove (5MT1690) for which there was no data available. They were also surprisingly abundant at the Pueblo I period Duckfoot Site (5MT3868) in comparison to the low proportion of turkey bone. Exact numbers of gizzard stones from each site were not recorded, though the quantity of provenience designation and field specimen numbers listed suggest that they were common (there is at least one stone for each provenience number). 3) Droppings. Scattered droppings have been found at Pueblo I and Pueblo I1 sites in Mancos Canyon, and in Pueblo 111 deposits at Long House, Lion House, Hoy House and Mug House (Emslie 1977, McKusick 1980b, Nickens 1981, Rohn 1977). The distribution of these coprolites indicate that the turkeys were wandering about the site while foraging for food. This is also suggested by the presence of insects in the feces at Site 1676 on Wetherill Mesa (Cattanach 1980). Insects are an important component in the diets of wild turkeys (Schorger 1966) and it follows that birds in captivity would consume a similar diet if given the freedom to forage. The sites in the Sand Canyon Locality do not provide clear evidence for the presence of turkey droppings. This may be a result of poor preservation conditions, caused partially by exposure as the site is located in a fairly open environment. In Structure 1008 (the biwalled, D-shaped structure) at Sand Canyon Pueblo (5MT765) "a thin layer of powdery, yellow-ochre-colored deposit, perhaps the remains of decomposed organic materials" (Bradley 1992: 89) was identified. A fairly large quantity of turkey gizzard stones were

109 associated with this layer. This description is similar to those given for accumulations of turkey dung in other reports (Breitburg 1988, Lang and Harris 1984). Bradley however, makes no further interpretation about the layer than to suggest that it contains organic material. It is possible that this room may have been inhabited by turkeys for a short time following its initial occupation by humans prior to the abandonment of the site. 4) Population Structure. Sub-adult birds have been reported at three sites dating to the Pueblo periods. In Mancos Canyon young juveniles less than four weeks old were found at Pueblo I sites (Emslie 1977). As well, numerous immature turkey bones were found in contexts which postdate AD 900, such as the immature and juvenile turkeys recorded at Big Juniper House, a late Pueblo I1 and early Pueblo I11 habitation (Swannack 1969). Birds of all ages are reported from the Pueblo 111 Johnson Canyon sites (Nickens 1981). Unfortunately, the demographics of turkey populations were not reported from other sites. Elements within the Sand Canyon Locality sample which displayed intact articular surfaces were aged according to the methods described in Chapter I. A total of 67% of the elements were aged and the resulting population structure is depicted in Table 4.4. The turkey assemblages from most sites in the Sand Canyon Locality had adult populations which made up over 90% of their totals. The percentages of adult elements ranged between 86 and 100% of the aged specimens. Only sites with sample sizes less than 15 had 100% adult populations, and may be attributed to inadequate sample size. High adult populations are expected in flocks raised for either meat or feathers, since optimal products of both are produced by adults birds. For example, populations raised for meat are expected to be culled when the birds reach maximum size to take full advantage of the highest yield of meat. A small number of sub-adults are also expected in the death assemblage as a result of

110 natural deaths from disease or chills. This number is expected to be lower than in natural populations, where mortality is high (Schorger 1966) since birds in captivity are protected by humans. Subadult bones may also be underrepresented as they are more subject to taphonomic processes due to their fragility. The majority of the sites in the Sand Canyon Locality are 'open air' sites increasing their susceptibility to destruction. Discounting preservation bias the above distribution is suggestive of a domesticated population, though the possibility of a hunted wild population can not be ignored. If turkeys were hunted it can also be expected that adults would be selected for their large size and therefore adult birds would also be chosen. Site 5MT MT MT MT MT1690 5MT181 5MT1825 5MT262 5MT3868 5MT3901 5MT3918 5MT3930 5MT3936 5MT3951 5MT3967 5MT5152 5MT765 Total Table 4.4: Age structure of the turkey population in the Sand Canyon Locality by site. 5) Pens. Many Pueblo I11 sites on both Wetherill Mesa and in Johnson 98

111 Canyon on the Mesa Verde provide architectural evidence for turkey pens (Rohn 1971, Swannack 1969). A deposit of droppings.6 to 2 metres deep and 13 by 40 metres in length and width was excavated alongside the structures at Long House (Schorger 1966). An accumulation of this magnitude could only be the result of the long-term residence of a fairly large group of turkeys. Site 5MTUMR2614 in Johnson Canyon also provides clear evidence of a pen in the form of an unroofed enclosure which contained a deep layer of turkey dung (Nickens 1981). A third pen existed at Mug House in Room 46A, where a thick layer of compressed dung was found to be mixed with turkey feathers and gizzard stones (Rohn 1977). No evidence, for turkey pens, other than the thin powdery layer of organic material in Structure 1008 at Sand Canyon Pueblo (5MT765 ) as described earlier, was apparent in the Sand Canyon Locality. It is possible that further excavation at Sand Canyon Pueblo may reveal pens or retaining structures, but though a variety of different areas and types of structures have been sampled less than 10% of the site had been excavated by the end of the 1989 field season (Bradley 1992). The evidence provided above will be discussed in the following chapter. The conclusions will affect subsequent interpretations regarding the turkey's function and its potential for intensification. In the next section, data which may potentially reveal how the turkey functioned in Anasazi society is presented. The Turkey's Function in Anasazi Society As reviewed in Chapter 11, there has been much discrepancy over the turkey's role in Anasazi society. Many debates have been waged to determine whether the birds were raised for food or feathers. Researchers advocating the importance of feathers discount the use of meat at some sites and emphasise the

112 ritual and utilitarian role of feathers and/or eggs in ceremonies and as offerings (Akins 1986, Lange 1950). The argument focusing on the importance of meat, does not ignore the importance of feathers or other turkey byproducts but suggests that they played a secondary role to subsistence. It also stresses that turkeys were initially raised as a feather source, though this changed as production intensified and meat became the primary resource (Breitburg 1988, Hargrave 1965,1970, Lang and Harris 1984, McKusick 1980a, 1986, Windes 1987). Evidence for both feather and food utilisation will be presented from sites in the Mesa Verde region and the Sand Canyon Locality. Evidence for Feather Utilisation The researchers who advocate that the domesticated turkey's role in southwestern society was to provide a feather source, support a dual use for this commodity. It is argued that feathers served both as a raw material for manufacturing utilitarian objects and as a valued component in ritual and ceremonial events. Its role as each is considered below. 1) The Mesa Verde Region a) Utilitarian Function. Turkey feather utilisation had begun on the Colorado Plateau by Basketmaker I1 when the turkey first appeared in the regional archaeological record. Evidence for textiles manufactured from feathers exist at sites dated as early as Basketmaker 11 near Durango, Colorado and Basketmaker I11 at present day Mesa Verde National Park (Moms 1939, Rohn 1971). Due to excellent preservation conditions feather artifacts survived at the Mesa Verde sites. There are no feathers reported from other sites in the study area regardless of time period. This is probably attributable to poor preservation rather than the absence of feather items in other regional sites. Reports from the Mustoe

113 site and Mancos Canyon refer to the lack of feather artifacts in their assemblages, but both cite preservation as the cause (Emslie 1977, Gould 1982). Though feather artifacts have been recovered from Basketmaker sites the majority were retrieved from those dating to the Pueblo.periods. Feather artifacts have been recorded at Long House, Mug House and Step House on the Mesa Verde (McKusick 1980b, Rohn 1977), and at Lion House and Hoy House which are located in Johnson Canyon to Mesa Verde's south (Nickens 1981). Fortunately, these artifacts have retained their original form thanks to exceptional preservation. The Anasazi frequently utilised turkey feathers in the manufacture of cordage, which was created by twisting vegetal fibres (i.e. yucca) with turkey down or flight feathers (McKusick 1980b, 1986, Moms 1954, Rohn 1971). Subsequently, it was woven into items such as blankets and robes. Feather socks made from knotless netted cordage, interwoven with turkey down were recovered from Long House (Cattanach 1980). Turkey feathers also served as fletching for arrows. McKusick (l980b: 393) noted that an arrow from Long House was "fletched with sections of turkey vane supported by the split rachis which extended beyond the vane at each end so that they might be bound to the shaft with sinew". The transition from the spear and atlatl to the bow and arrow as the primary weapon occurred during Basketmaker I11 and corresponds to the adoption of turkeys by the Anasazi (Cordell 1984). By keeping turkeys captive the Anasazi would have had a reliable feather source available for fletching arrows. Mug House and Long House have each yielded isolated finds of corncobs which have turkey feathers inserted into their base (Cattanach 1980, Rohn 1971). These have been interpreted as potential shuttlecocks as the Hopi currently manufacture objects of a similar description which are used to play games, though the method is not explained (Nickens 1981).

114 b) Ritual function. The presence of articulated turkey burials in the archaeological record imply that the turkey had ritual significance and was not eaten. Alternately, skeletons are expected to be dispersed in the archaeological record following meat consumption as a result of processing methods (Hargrave 1965, Senior and Pierce 1989). In the Mesa Verde region turkey burials have been recovered from contexts dating as early as Basketmaker I11 (Morris 1991, Schorger 1966). Most were fully articulated and no evidence for food use was reported. Two turkey burials from the Knobby Knee site have been dated to the Basketmaker I11 period, though this date was disputed by the principal investigator of the site and the faunal analyst who advocated a Pueblo I11 date (Moms 1991, Bertram 1991). The primary argument made against the early date, was based on the premise that Basketmaker I11 turkey burials were highly unusual (Bertram 1991). This, however is disputable as they have been recovered from other parts of the region. A Basketmaker I11 burial was excavated from the Twin Trees Site at Mesa Verde and dates to A.D. 800 (Schorger 1966). Turkey burials were also mentioned in reports from the excavations at Badger House (Hayes and Lancaster 1975). Considering the paucity of Basketmaker I11 sites with available faunal data and the low frequencies of total turkey bone in the study area at this time, turkey burials are relatively common in the region. Turkey burials have also been reported at Pueblo I1 and I11 sites in the region. These burials appear to have ritual significance due to the location of their interment. Most skeletons were found in association with structures assumed to have ritual significance (i.e. kivas) or human burials. In Mancos Canyon, Emslie (1977) excavated the dismembered skeletons of two adult hens which were scattered over the benches of a Late Pueblo I1 kiva. Other turkey burials recovered from kivas include a young hen excavated from beneath the

115 floor of Kiva D at Mug House (Rohn 1971). In Mancos Canyon, a human burial in the kiva fill was associated with a complete turkey egg and human Burial 14 was interred alongside a juvenile turkey less than four weeks old (Emslie 1977). In addition, burial 30 at Long House was recovered in direct association with an adult turkey hen though no other inclusions were recovered from the grave (Cattanach 1980). The context of these data suggests that the turkeys were purposely placed in these locations for ritual purposes, though the reasoning behind these actions cannot be discerned. In addition, ethnographic evidence exists for the ritual burial of intact turkey skeletons by the Zuni whose offerings of turkey carcasses often result in burial (Gnabasik 1981). Other data which suggest that turkeys were attributed ritual value involves the utilisation of feathers in ceremonial events. Several human burials in the Mesa Verde area were wrapped in turkey feather blankets prior to interment (Nickens 1981, Rohn 1977). Data recording the use of turkey feathers in ceremonies as isolated items, prayer sticks or in association with regalia is also documented in both historical and ethnographic reports concerning the Anasazi and their descendants (see Chapter 11). Cattanach (1980: 358) refers to the presence of turkey feather 'aspergills' which he describes as a group of feathers which are arranged in rows and sewn together with yucca or split feather vane twine. Though he does not mention their function he includes them in a section describing perishable ceremonial items. In addition twigs wrapped with feathers have been interpreted to have ritual function as well, possibly as offerings. Unfortunately, due to their fragility, few feather artifacts have been recovered from archaeological contexts outside of Mesa Verde National Park. 2) The Sand Canyon Locality Unfortunately the conditions of preservation at the sites in the Sand

116 Canyon Locality were not conducive for the protection of fragile remains such as feathers. There is no evidence for feathers at any of the sites within the sample. Negative evidence however, should not be considered proof of non-existence. Preservation conditions at sites in the Sand Canyon Locality have been examined by Driver et al. (1995) who analysed variation in the quality of preservation between sites by examining aspects of the faunal assemblages. It was concluded that the sites were not differentially affected by preservation following the analysis of six indices which measured the degree of preservation in relative terms. One of these indices (proximal:distal ends of long bones) is also an effective indicator of the severity of the affects of preservation conditions on the faunal assemblages. Long bones often exhibit differential preservation between their distal and proximal ends due to variation in shape and density (Brain 1967). Driver et al. (1995) chose the combined distal ends versus the combined proximal ends of both the humerus and tibia of Silvilagus sp. as an indicator of preservation. Sylvilagus sp. was chosen as it is the most common mammal in the majority of sampled sites. In a site with optimal preservation it is expected that there will be an equal number of distal and proximal ends of humeri and tibiae. Past studies, however have indicated that when affected by taphonomic processes the distal ends of both elements exhibit the best preservation (Brain 1967). The ratio of distal to proximal ends is therefore expected to increase as the quality of preservation decreases. Six of the largest sites in the Sand Canyon Locality were selected for analysis as they have large enough sample sizes to produce useful data. The results, reproduced below (Table 4.5) suggest that preservation is much less than optimal. Distal ends of humeri and tibiae outnumber the proximal ends substantially in each case. In light of this evidence

117 it may be concluded that it is unlikely that remains as fragile as feathers could survive to the present time. site Saddlehorn Hamlet (5MT262) 14:O Castle Rock Pueblo (5MT1825) 8: 1 Lillian's Site (5MT3936) 1.8: 1 Catherine's Site (5MT3967) 6:l Kenzie Dawn Hamlet (5MT5 152) 2.5: 1 Stanton's Site (5MT10508) 5: 1 Proximal Table 4.5: Indices of preservation, the ratio of the distal ends of Sylvilagus sp. tibiae and humeri to their proximal ends at sites in the Sand Canyon Locality. Adapted from Driver et al. (1995). Evidence for turkey burials is equally scant in the Sand Canyon Locality. Perhaps, no burials existed or the area sampled by excavation was unusually sparse in articulated interments. Evidence for Food Utilisation The following discussion focuses on the recovery of evidence indicating the utilisation of turkeys as a subsistence item. Bone artifacts are also discussed as they are likely byproducts of food use, since bone becomes available as a raw material after the flesh is removed. As previously mentioned, indicators in the faunal record which imply utilisation of turkeys for food include: a) cutmarks; b) bone mutilation for the extraction of cancellous areas; c) disarticulated skeletons and dispersed bone; d) localized burning; e) the dissociation of tibiotarsi and tarsometatarsi with ossified tendons and f) age and sex profiles with predominantly young adult males and

118 adult females. Evidence from the Mesa Verde Region and Sand Canyon Locality are presented below. 1) The Mesa Verde Region. A limited number of Basketmaker sites were examined within the Mesa Verde Region. Each produced small assemblages of turkey bones in relation to other species in the sample. Because the sample is small it is not surprising that evidence for food use was not detected in Basketmaker contexts in this study. It is possible that turkeys were not being used as a food source at this time or that the sample is too small to detect the expected evidence. In either case turkeys do not appear to be an economically valuable meat source. The remaining discussion focuses on the Pueblo periods. a) Cutmarks. Strong evidence for food utilisation appears during Pueblo I in the Mesa Verde Region. Cutmarks, though rare, are found on distal tibiotarsi at sites in Mancos Canyon (Emslie 1977). It is probable that these resulted from cutting the tendons at the junction of the tibiotarsus and tarsometatarsus which enables the removal of the lower leg and foot which contain no meat (Lang and Harris 1984). Cuts are found most frequently on bones during the Pueblo I11 period at Lion House, Hoy House (Nickens 1981), Long House (McKusick 1980b), Mug House (Rohn 1971) and other sites on Wetherill Mesa (Hargrave 1970). Once again, the distal tibiotarsus displays cutmarks the most frequently. Unfortunately, details of butchery (i.e. location of cuts) are not recorded in most reports. b) Breakage and Cancellous bone extraction. The breakage of cancellous areas of turkey bone was not mentioned in any of the reports examined, with the exception of the Pueblo I11 sites on the Mesa Verde. It is probable that either this data was not considered for most reports or the

119 exploitation of cancellous bone did not begin until Pueblo I11 when turkey utilisation intensified. Elements which exhibited frequent cancellous damage include the proximal tibiotarsus, the proximal and distal humerus and the anterior sternum (McKusick 1980b, Rohn 1971, Hargrave 1970). These areas were undoubtedly selected for their large quantities of spongy bone which could be accessed for its high fat content and nutritional value. c) Disarticulated and dispersed bone. The turkey bones at all sites, with the exception of the articulated burials previously mentioned, were disarticulated and dispersed. McKusick (l986b) notes that this pattern was not apparent at Long House until after A.D Previously, a high proportion of the few turkey bones recovered from sites in the Mesa Verde region belonged to articulated burials. This date is therefore frequently cited as the initiation of meat use on Wetherill Mesa (McKusick 1986). Turkey bones from the Dolores sites were also disarticulated prior to burial, though other indicators of food use were scarce (Neusius 1986). d) Burned Bone. Charred bone is reported from the Pueblo I11 sites in Johnson Canyon (Nickens 1981) and at Mug House (Rohn 1971). Data regarding the specifics of the burning is not mentioned, so the intensity and extent is not known. Because examined only superficially in the regional reports, this data cannot be included in the present study. Bertram's (1991) analysis of the Hovenweep Laterals fauna is an exception. He states that turkey bones which were burned and "probably cooked" appeared in all contexts which contained "rich turkey assemblages" (1991: 1006). e) Ossified Tendons. Reports from sites on Wetherill Mesa mention that tibiotarsi and tarsometatarsi were not found in association with ossified tendons during excavation (Hargrave 1977, Rohn 1971). This implies that the bones were discarded following removal of the flesh when the bony tendons were still

120 imbedded in the leg muscles. If the legs decayed while encased in flesh, the tendons which form in the turkey's lower leg, would remain in association with the applicable leg bone, unless disturbed by taphonomic processes. f) Population Structure. Information relating to the population structure in the regional reports was not examined due to insufficient reporting and lack of compatibility between sites. A final piece of evidence for the consumption of turkey in the Mesa Verde region is the presence of a small turkey bone in a human coprolite found in Johnson Canyon (Nickens 1981). It is probable that it was mistakenly consumed in a meal containing turkey meat. 2) The Sand Canyon Locality Much evidence for the utilisation of the turkey as a food source exists in the Sand Canyon Locality. The earliest site, Duckfoot (5MT3868) which dates to Pueblo I, has a small proportion of turkey bones. Of these only one is burned. No other modifications were noted in the turkey assemblage. This may be a product of the small quantity of turkey bones from this site or it may indicate that turkeys were not used as a food source at this time. The remaining sites in the Sand Canyon Locality date predominantly to the Pueblo I11 period and contain more extensive evidence of food exploitation. a) Cutmarks. Cutmarks were recorded during analysis regardless of their association with other cultural modifications. All cutrnarks were illustrated for future reference, to determine whether they were created as a result of processing methods or tool making. In the following analysis all cutmarks which were judged to be independent of artifact manufacture are considered. Artifacts which were modified at one end and had cutmarks on the opposing end's articular surface (i.e. tibiotarsi awls with proximal tips, often have independent cutmarks

121 on their distal condyles) are considered to have been created independently of artifact manufacture for the purpose of this analysis as well. Turkey bones with cutmarks were identified at all Pueblo I11 sites in the Sand Canyon Locality sample with the exception of Mad Dog Tower (5MT181) and Cougar Cub Alcove (5MT1690). These sites have turkey sample sizes of three and four respectively and are dismissed on account of sample size. - i Number with % of Total Turkey Element Cutmarks Elements Cut Cervical 4 3% Fibula 1 2% Foot Phalanx 3 1% Humerus 8 5% Patella 1 100% Terminal Foot Phalanx 2 3% Radius 3 3% Tarsometatarsus 10 3% Tibiotarsus 34 12% Ulna 3 3% Total 69 2% Table 4.6: Distribution of cutmarks by element and the percentage of each element with cuts. Table 4.6 indicates the distribution of cutrnarks on turkey bone in the Sand Canyon Locality by element and site. Of the total turkey bones in the sample 2% displayed cutmarks. The elements which were cut the most frequently include the tibiotarsus (12% of total tibiortarsi), the humerus (5% of total humeri) and the tarsometatarsus (3% of total tarsi). Elements which were cut with less frequency include the cervical vertebra, the foot phalanx the terminal foot phalanx, the ulna and the radius. In addition, one fibula and one patella bore cutmarks. A skeletal depiction of the turkey is presented in Figure 4.4 to indicate the locations of commonly recorded cutmarks in the Sand Canyon Locality turkey assemblage. The patterning in cutmark distribution may be interpreted by

122 - Cervical Vertebrae Humeral Shaft 4 DM Humerus Foot Phalanx jb 8at P - Distal Tamometatam~s 'ce denuind Foot Phalanx Figure 4.4: Skeletal depiction of Meleagris gallopavo indicating the locations of common cutmarks on turkey elements in the Sand Canyon Locality sample. Adapted with permission from Gilbert et al. (1985).

123 utilising the suggestions discussed in Chapter I1 from Lang and Harris (1984) (also see Table 4.7 which depicts the regions on each element which were frequently cut). # of Turkey % of Total Portion of Element with Cuts Turkey Cuts Cervical Vertebra 4 6% Proximal Humerus 1 1% Shaft of Humerus 4 6% Distal Humerus 3 4% Shaft of Ulna 1 1% Distal Ulnae 2 3% Shaft of Radius 2 3% Distal Radius 1 1% Shaft of Tibiotarsi 2 3% Distal Tibiotarsi 32 46% Shaft of Fibula 1 1% Patella 1 1% Proximal Tarsometatarsus 5 7% Shaft of Tarsometatarus 2 3% Distal Tarsometatarsus 3 4% Foot Phalanx Terminal Foot Phalanx 3 2 4% 3% Total % Table 4.7: The location of cutmarks on turkey elements in the Sand Canyon Locality. The junction which exhibited the highest proportion of cuts falls between the distal tibiotarsus and proximal tarsometatarus. Cuts in this area were probably created when removing the turkey's lower leg which is covered in a scaly skin and contains no meat. Cuts on proximal tibiotarsi, the patella which rests between the femur and tibiotarsus and the fibula which articulates with the proximal tibiotarsus are proposed to have resulted from the disarticulation of the drumstick. Cutting the ligament at the junction of the femur and tibiotarsus could potentially result in cutmarks on any of these elements. Cutmarks on the ulna shaft may be produced during removal of the secondary feathers. This may be

124 indicative of both feather and meat utilisation. It is proposed that cutmarks on the cervical vertebrae were created by the disarticulation of the neck from the body. This hypothesis is supported by the presence of cuts on cervical vertebrae and their absence on all other vertebrae. Cutmarks are discussed further in Chapter V. b) Breakage and cancellous bone extraction. During analysis each specimen was assigned a code which indicates which portion of the element is represented. Long bones, were given individual codes for complete bones, fragments with complete or partial proximal ends, fragments with complete or partial distal ends, articular ends and shaft fragments. The codes for the humerus and tibiotarsus, which have large spongy areas on their proximal ends were assessed to determine which parts were frequently destroyed. Data for the radius was examined as a control because both ends possess little cancellous bone. The proximal and distal ends are also similar in size, shape and density (See Table 4.8) Humerus Tibiotarsus Radius Part Represented F % F % F % Complete 5 3% 1 0% 5 4% Proximal End Intact 11 7% 2 1% 13 11% Partial proximal end intact 22 15% 7 3% 5 4% Distal End Intact Partial Distal End Intact % 7% 84 30% 52 19% % 6% Shaft Fragment 91 61% % 77 63% Total I Table 4.8: Part representation of humeri, tibiotarsi and radii from Pueblo I11 contexts in the Sand Canyon Locality. These data must be interpreted with caution because although spongy areas are frequently broken to access their interiors they are also more susceptible to taphonomic processes. Spongy bone has more potential to be damaged postdepositionally due to its delicate nature. Data for the control element (radius)

125 does not show differential preservation between the proximal and distal ends. Approximately the same numbers of intact proximal (13) and distal (15) ends were recovered. Interestingly, the same pattern was apparent for the humerus. There were 11 of both intact proximal and distal ends. The tibiotarsus however, showed dramatic differentiation in the representation of proximal and distal ends. Only two proximal tibiotarsi were intact (1% of total), while 84 (30% of total) distal ends were intact. This is the expected distribution for an element which had its spongy area mutilated to extract additional nutrients. This distribution is also expected for a bone which has one dense and one spongy end. Still, the high preservation of the proximal humerus implies that the patterning can not be attributed to preservation alone. c) Skeletal disarticulation and dispersed elements. As previously mentioned, there were no articulated turkey burials recovered from Pueblo I11 sites in the Sand Canyon Locality. All specimens in the sample were. disarticulated and recovered from dispersed contexts. It is likely that taphonomic forces played a role in creating this pattern of disposal, yet its pervasiveness suggests that it can be attributed to the disarticulation of skeletons prior to deposition. Disarticulation often results from the processing and consumption of meat and dispersal is a common means for the disposal of refuse. d) Burning. Though many bones in the Sand Canyon Locality assemblage exhibit a form of burning, they are not considered indicative of food use unless the burning is localised. The interpretation that bone which is entirely burned results from cooking is avoided for fear that bones which were burned by other means (i.e. disposal in hearths or rooms which were later burned) will distort the assemblage. Localised burning provides the most reliable evidence for cooking and is thus used as the minimum figure for burned bone resulting from cooking. Undoubtedly, eliminating the majority of burned bones as indicators of

126 cooking would result in the under representation of cooked bone in the assemblage. Therefore figures for the total burned bone are also presented. The actual number of bones considered to be indicative of cooking is expected to fall somewhere between these two figures. A total of 40 (1% of the total assemblage) specimens in the Pueblo 111, Sand Canyon sample expressed localised burning, whereas 269 (9%) exhibited burning of any type (burned black, brown, white or localised). As a final note, it is probable that turkeys were also cooked by boiling which may not leave any indicators on bone. e) Ossified Tendons in Situ. The context of ossified tendons at the time of excavation was impossible to assess since analysis took place a few years later and this information was not recorded. As a result the contextual placement and the associations of each specimen are unknown. A total of 59 ossified tendons were identified in the faunal assemblage, indicating that preservation was conducive to their survival. f) Age and Sex Profiles. In this study sub-adult birds (with the exception of some immature males) were not sexed due to complications created by sexual dimorphism, thus a comprehensive profile of the population structure could not be presented. As an alternative, age and sex charts have been produced separately to provide some insight into the structure of the local turkey population. The age profile for Sand Canyon Locality turkeys was presented in a previous section (Table 4.4). A brief overview is presented below. A high proportion of specimens in the sample were aged during analysis (67% of total sample). Of these, adults composed the majority of the individuals, representing 94% of the aged sample, This average is also representative of the sites on an individual basis. The percentage of adults ranged between 91 and 100% of the aged population when sites were examined as distinct entities. The four sites which have 100% adult populations each have sample size of less than

127 15, thus the high proportions are probably a result of small sample size. Subadults were uncommon and increased in abundance with increasing age. Five percent of the aged sample were immature, 0.6% were medium juveniles and 0.2% were small juveniles. It should be noted that bones from subadults are more susceptible to destruction by taphonomic processes due to their fragility. Fragility decreases with age, and as a result young individuals may be underrepresented in the sample. Twenty-eight percent of the total specimens in the Sand Canyon Locality sample were sexed (see Table 4.9). The resulting sexed population displayed a relatively even distribution of male and female specimens. The females were represented by 44% of the sexed adult population while 56% were male. Site 5MT181 5MT262 5MT765 5MT1690 5MT1825 5MT3868 5MT3901 5MT3918 5MT3930 5MT3936 5MT3951 5MT3967 5MT5152 5MT MT MT MT11338 Total Table 4.9: Sex structure of the turkey population by site in the Sand Canyon Locality.

128 Evidence for Bone Utilisation 1) The Mesa Verde Region Little evidence exists for the modification of turkey bone during Basketmaker 111. Utilitarian artifacts were primarily constructed from the more robust large mammal and lagomorph bones. The few turkey bone artifacts that have been recorded are primarily hollow tubes, cut most frequently from the shaft of the ulna and radius. It is probable that these served a decorative function as evidence of light polishing has been found in the interior of some tubes. This may be a result of stringing the tubes on a piece of rope as beads (Bertram 1991). Turkey bone artifacts increased in frequency during the Pueblo periods. Emslie (1977) reports that they were often recovered from Pueblo I and I1 sites in Mancos Canyon. It appears that as turkeys became more common at archaeological sites, associated increases in turkey artifacts resulted. This is further illustrated by a shift in the species utilised for bone tool production. Prior to Pueblo I1 the majority of utilitarian artifacts were manufactured from mammal bone, predominantly lagomorphs and artiodactyls. By A.D. 900, however there was a shift to the predominant use of turkey bone as a raw material for these artifact types (Cattanach 1980, Swannack 1969). Prior to this only two turkey bone tools were reported from Wetherill Mesa (McKusick 1980b). This pattern can also be detected in the data collected from other sites in the region. At the Escalante site which dates to early Pueblo I1 when the turkey was beginning to gain importance in faunal assemblages, turkey bone tools were only the third most important material for bone implement construction (Hallasi 1979). The Hovenweep Laterals sites had virtually no turkey bone tools during Basketmaker 111 and few by Pueblo 11, yet the Pueblo I11 assemblages were "dominated overwhelmingly by (turkey) bone tools and beads" (Bertram 1991: 1104). At other sites with Pueblo I11 components such as Knobby Knee Stockade

129 (Kuckelman and Moms 1988), Lowry Ruin (Martin 1939) and Mug House (Rohn 197 I), turkey artifacts also figured prominently. Artifact types frequently represented by turkey bone in the regional sites included needles, awls, awls with grooves, tubes, beads, scapula tools, scrapers and whistles (Rohn 1971, Martin 1939, Bertram 1991, Swannack 1969). 2) The Sand Canyon Locality During analysis of the Sand Canyon Locality fauna, artifacts were recorded and illustrated. The majority of artifacts were classified according to Schwab and Bradley (1987), though a few additions and specifications have been made to allow further distinction between types in this study. The type categories utilised here include awls, awls with grooves, tubes, tubes with holes, beads, scapula tools and other modified bone. Other modified bone refers to miscellaneous specimens, including abraded or polished bone, discarded ends, and tools of unknown function. Only one artifact was recovered from Pueblo I deposits in the Sand Canyon Locality. The artifact is a tibiotarsus classified as other modified bone. The tibiotarsus has ground distal condyles but displays no evidence for function. The proximal end of the specimen was broken during excavation, thus it is possible that an awl tip may have been removed though it was not recovered. The distal condyles and trochlea of awls from tibiotarsi and tarsometatarsi are frequently cut or ground in the Pueblo I11 assemblage, increasing the likelihood of this conjecture. If sample size is considered the lone artifact appears more significant as it represents 6% of the turkey bone assemblage, which is comparable to proportions of turkey artifacts in later assemblages. The turkey artifact also represents 2% of the total artifact assemblage from Duc kfoot, which is notable but not as high as in other assemblages.

130 % of Total Artifact Type Frequency Turkey Artifacts Awls % Awls with Grooves 9 4% Beads 1 0.4% Discarded Ends 9 4% Ground 7 3% Polished 11 5% Scapula Tools 14 6% Tubes 42 18% Tubes with Holes 3 1% Unknown Function 7 3% Total 238 Table 4.10: Artifacts divided by type from Pueblo I11 contexts in the Sand Canyon Locality. There is an abundance of turkey artifacts in Pueblo I11 assemblages from the Sand Canyon Locality. A total of 238 were recovered from the 16 sites combined, composing 8% of the total turkey assemblage. Their distribution by type ( Table 4.10) and element (Table 4.1 1) are depicted here. i % of Element % of Turkey Modified into Element Artifacts Artifacts Rib 1 6% Sternum Scapula Humerus Ulna Radius Carpometacarpus Femur Tibiotarsus Tarsometatarsus Total I 238 8% Table 4.11: Artifacts divided by element from Pueblo I11 contexts in the Sand Canyon Locality.

131 The primary tool type identified is the awl (56% of tools) which was manufactured from a variety of long bones including the radius, ulna, carpometacarpus, tibiotarsus and tarsometatarsus, though the latter two dominated. In the turkey these elements generally have narrow shafts and more robust cortical bone than other elements. This provides the necessary support for the weaving and puncturing activities for which the awls were intended. Awls with grooves have been grouped separately to emphasise the presence of the groove which has been attributed to the friction of fibres during weaving (Bullock 1992). There are nine awls with grooves in the assemblage (4% of total artifacts), which when combined with the original awl category brings the awl total to 59% of the total turkey artifacts. Beads and tubes are separated arbitrarily in the Crow Canyon Archaeological Center guidelines (Scwab and Bradley 1987). No quantitative boundary was provided for the division but it has been estimated to fall at approximately three centimetres (Bullock 1992). Beads are less than three centimetres in length while tubes are larger. When the data were collected for this study, the boundary between the two was set at a shorter length. Beads are deemed to be less than approximately 1.5 centimetres in length while tubes are longer. Both beads and tubes were manufactured from long bone shafts which were cut at both ends, usually by the groove and snap method. They are considered to have similar functions, predominantly as decorative objects such as jewelry. Possibly they were strung and worn as necklaces. Most tubes had brightly polished exteriors, though few revealed interior polish which would indicate friction from stringing. Tubes compose 18% of the total turkey artifacts and were manufactured largely from humerus, radius and ulna shafts, though the odd tibiotarsus and tarsometatarsus were also utilised. Three tubes with holes (1% of total artifacts) were also identified in the sample and once again were

132 grouped separately to emphasise the presence of their holes which serve an unknown function. One source (Schorger 1966) postulates their function as whistles which were used as wild turkey callers. Their inclusion in the bone tube category brings its total to 19% of the total artifacts. The lone bead in the sample (0.4% of total artifacts) was cut from a radius shaft. Artifacts manufactured from scapulae compose 6% of the artifact assemblage. The scapulae are predominantly intact proximally, but the distal portion of the blade has been removed by an angled cut. The function of the scapulae artifacts is unknown and a literature search revealed that their presence was only referenced in two reports (Bertram 1991, Bullock 1992), one which deals with much of the same material discussed here. It has been suggested that these tools were utilised as scrapers for soft vegetal material or pottery (Bertram 1991), but the lack of any indication of usewear on the proposed scraping surface has resulted in the rejection of this hypothesis (Bullock 1992). It is also possible that the diagonal cuts may be a byproduct of butchering, possibly resulting from the removal of the wing. Still, there is no evidence for cutmarks on any of these artifacts, and because the scapula does not articulate with any bones at its distal end, it is seemingly pointless to remove its lower end to aid in disarticulation. Potentially residue analysis could reveal additional information which may clarify the function of the scapulae tools. Bullock (1992) refers to their presence as anomalous as they have rarely been found in prehistoric Anasazi sites outside of the region or prior to Pueblo 111. She suggests that they may have utility as markers of late prehistoric contexts. Possibly artifacts manufactured from scapulae were created only in southwestern Colorado and did not diffuse elsewhere. The final artifact category is other modified bone. This group is most frequently represented by polished elements which are not otherwise modified.

133 There are 11 polished bones in the sample which account for 5% of the total artifacts. The category 'elements with ground condyles' refers exclusively to tibiotarsi and tarsometatarsi which have heavy abrasion on at least one of their distal condyles or trochlea. The function of this modification is unknown. It is possible that the elements may be the remnants of broken awls as some awls in the assemblage display the same alteration on their distal ends. There are 7 artifacts with ground condyles, representing 3% of the turkey artifact assemblage. Artifacts which have been identified as discarded ends or blanks from tube manufacture compose 4% of the total artifacts. These generally take the form of distal or proximal ends of wing long bones, though the occasional tibiotarsus and tarsometatarsus display similar alterations. Each specimen has a straight cut on the shaft. These often appear to be formed by groove and snap and are not believed to serve a function except as debitage of bone tube manufacture. Seven tools (3% of turkey artifacts) of unknown function which do not fit into any of the above categories were also identified. Each of these were manufactured from tarsometatarsi or tibiotarsi and are similar to awls with the exception of their tips. The modified edge tends to be gently rounded and polished yet relatively flat rather than pointed like an awl. The tips therefore cannot be used to pierce as an awl can. Possibly they were utilised for weaving in cases where sharp tips were not required. A morphological and functional analysis of artifacts from six sites in the Sand Canyon Locality (also included within the sample in this project) has been performed by Bullock (1992). Her research brings to light several noteworthy observations which relate to this research. Avian bone was the dominant raw material selected for bone tool manufacture in all of her sampled sites, though the dominance was not as strong at Sand Canyon Pueblo. Of these, the elements

134 modified most frequently are the tibiotarsus and tarsometatarsus. Awls are noted as the most abundant artifact type followed by bone tubes and beads. Intensity of Turkey Production: Temporal Change Faunal reports from Anasazi sites in the Mesa Verde region were consulted to examine change through time in turkey proportions. Data from the faunal assemblages from sites in the Sand Canyon Locality sample are also included in this regional sample. The sites in the sample were divided into time periods according to chronological information within the site reports. The data are grouped by time period and presented in Table There are numerous problems with the regional analysis. These result from incompatibility amongst the data due to variables such as taphonornic factors, excavation strategies, sample size, identification criteria and quantification methods. There is no standardised method for faunal analysis in the study region, thus each researcher utilised hisher own strategies. This has resulted in the creation of data that can not be directly compared without creating numerous errors and generalisations. These problems are recognised in this research and the reader is urged to consider them when interpreting the following analysis. The regional analysis is included only to provide a general interpretation of the change in turkey populations throughout time in the Mesa Verde region. The sampling strategy for the Mesa Verde Region was reviewed in Chapter I. Although an attempt was made to select data quantified using NISP, at times this was not available and MNI figures are recorded as an alternative (the quantification strategy is indicated on Table 4.12). It is believed that it is possible to identify general trends in the data by using proportions though quantification strategies are incompatible. These trends however must be interpreted with caution.

135 ~ --- Basketmaker m A.D A.D A.D A.D. 685 Basketmaker 111 Basketmaker EI Total Site Knobby Knee (5MT2525) Dolores Modeling Period I Shallow House (5MT8822) Dos Bobs Hamlet (5MT8837) % Reference 5% Bertram (1991) 1% Neusisus (1986) 33% Kuckelman & Moms (1988 2% Kuckelman & Morris (1988 2% NISP M N T I Pueblo I A.D A.D A.D. 780 A.D A.D A.D Pueblo I Total NISP Pueblo I Total MNI Dolores Modeling Periods II-IV Little Cahone House (5MT8838) 5~~2347 5MT2559 Duckfoot Site (5MT3868) 2% Neusius (1986) 5% Kuckelman & Morris ( % Emslie (1977) 40% Emslie (1977) 0% Walker (1989) 2% 28% A.D A.D A.D A.D A.D A.D A.D A.D A.D Pueblo 11 Pueblo 11 Pueblo II - Total ~ NISP Pueblo II Total MNI '~indweed House (5MT8834)' 4%'~uckelman & Morris (1988' 1 27 Chameleon House (5MT8836) 13% Kuckelman & Morris (1988: 1 8 Casa Bisecada 15MT8829) 50% Kuckelman & Morris (1988' 2 4 Norton House (5~8839) 1 % Kuckelman & Morris (1988' 1 89 Dripping Springs (5MT2527) 10% Bertram (1991) 3 30 Escalante (5MT2149) 8% Hallasi (1979) Dominguez Ruin 19% Reed (1979) 5 26 '~oundtree Pueblo (5MT2544) ' 25%'~ertr& (1991) ' Wallace Ruin (5MT6969) 15% Shelley (1993) Hanson Pueblo (5MT3879) 16% Rood (1991) Gnatsville 20% Kent (1986) % % I Pueblo JI A.D A.D Hov House (5MT2150) 39% Nickens (1981) A.D ; / ~ioi ~ouse (5~~2156) / 26%j~ickens i1981 j A.D l Mustoe Site I 39%IGould (1982) A.D ]G and G Hamlet (5MTll338) 42%/Driver kt al. (1995) A.D IKenzie Dawn Hamlet (5MT %1Driver et al. (1995) A.D A.D A.D '~nobb~ Knee (5MT2525) ' 70%'~ertram (1991) Roundtree Pueblo (5MT2544) 81% Bertram (1991) MU^ HOUW (5~~1228) 54% Rohn (1971) Table 4.12: Percentage of turkey per site in the regional sample from the Mesa Verde region. The percentage sign in the third column refers to the percentage of turkeys and the n in column sever indicates the total NISP when NISP values are recorded and the total MNI when MNI values are recorded.

136 Table 4.12 continued: Percentage of turkey per site in the regional sample from the Mesa Verde region. The percentage sign in the third column refers to the percentage of turkeys and the n in column sever indicates the total NISP when NISP values are recorded and the total MNI when MNI values are recorded. The total percentages for each time period were calculated only with data from sites using NISP. MNI figures are not included in totals due to the incompatibility of the data as discussed above. The totals were calculated by combining the frequencies of turkey, lagomorphs and artiodactyls from the sites which utilised NISP and performing the calculation for proportions as described in Chapter I. The averages are included in an attempt to balance the range of sample sizes for the various sites and to provide a broader overview of change throughout the region, by eliminating local variation. The problem of intensification can also be addressed more comprehensively by examining general trends. Evidence from the Mesa Verde Region Evidence for turkey utilisation is rare in the Mesa Verde Region prior to Basketmaker 111. Most remains dating to early time periods which have been identified as Meleagris gallopavo have been discounted in recent years as misidentifications (McKusick 1986). Currently, the oldest turkey remains from

137 Southwestern Colorado were recovered from Basketmaker I1 sites near present day Durango in the form of feather blankets (Morris 1939). These sites, however are outside of the study area. This does not necessarily indicate that turkeys were inhabiting the area at the time. It is possible that feathers were being traded from other regions for their utility in the manufacture of warm clothing. Alternatively, the absence of earlier remains may be influenced by the paucity of archaeological data from Basketmaker I1 and the Archaic Period in the region. Faunal data was not available for examination from sites occupied prior to Basketmaker I11 and the following discussion begins ca. A.D a) Basketmaker 111. Reports from three sites and the Dolores Archaeological Program were consulted for faunal data from the Basketmaker I11 period (See Table 4.12). Following their initial documentation, turkey remains appear in Basketmaker I11 contexts in low percentages indicating that they played a limited role in economic activities. They compose only 2% of the Basketmaker I11 assemblage of major economic species. Lagomorphs and artiodactyls were the major sources of animal protein and secondary animal products. In addition, very few turkey bone tools and no cutmarks were reported from this time period. b) Pueblo I. Average turkey percentages in the Mesa Verde Region during Pueblo I are virtually indistinguishable from those in Basketmaker 111 deposits. Faunal data were drawn from four sites and the Dolores Archaeological Program. Turkey percentages increased insignificantly and comprised only 2% of the species deemed to be of high economic value. Individually the three sites quantified by NISP correlate well with this figure. Alternatively, two Pueblo I sites from Mancos Canyon (5MT2347 and 5MT2559) were quantified using MNI, and therefore are not included in the regional average. The turkey percentages from these sites with values of 25% and 40% respectively are

138 notably higher than the other Pueblo I sites. This may be a result of sample size, the utilisation of MNI as a quantitative method rather than NISP or local variability. Unfortunately there is little comparative data available from the nearby Mesa Verde or Johnson Canyon sites during Pueblo I which could be used to test the latter option. c) Pueblo 11. A large sample of sites was examined from the Pueblo I1 period, including fourteen sites with data quantified using NISP and four that utilised MNI. Turkey percentages increased notably at this time and comprised 12% of the economic species from sites quantified by NISP. A substantial range of variability is apparent when the sites are examined independently, though the anomalous values tend to be associated with smaller sample sizes. The Dolores area has particularly low percentages as it was largely abandoned by this time, explaining the associated decrease in turkey abundance. Turkey production was never intensified during occupation of the Dolores area, regardless of time period. Percentages calculated from the four sites quantified by MNI each have high turkey values ranging from 19 to 33%. In addition to increasing percentages, quantities of turkey bone tools and evidence for subsistence use (i.e. cutmarks and burning) also increase during Pueblo I1 (Cattanach 1980, McKusick 1980b) d) Pueblo 111. Turkey percentages increase dramatically in Pueblo I11 assemblages, and represent the most abundant fauna in the combined regional assemblage. The compilation of data from 20 sites with a total sample size of 12,547 indicates that turkey compose 49% of major economic species in the Pueblo I11 assemblages. Individual assemblages in the sample retain high turkey percentages, the lowest value being 26%. Percentages at other sites range as high as 81%, confing the significant increase in this period. Only one site in the Pueblo I11 sample was calculated using MNI. This site displayed a turkey percentage of 39% which is consistent with the NISP quantified sites in the

139 sample. The abundance of turkey in the assemblage is accompanied by similar increases in its use as a raw material for bone tools, and a proliferation of retaining structures which have been interpreted as turkey pens at various sites (Cattanach 1980, Rohn 1970, 1977, Swannack 1969). Evidence from the Sand Canyon Locality The sample from the Sand Canyon Locality is not well suited for the examination of temporal change due to the paucity of sites from early periods. The Duckfoot Site (5MT3868) is the only site in the sample which was entirely occupied prior to Pueblo 111. At some other sites both Basketmaker I11 and Pueblo I and I1 components were excavated, but the sampled units were dominated by Pueblo 111 deposits, thus the data is attributed to this time period. As a result the sample is composed of one Pueblo I site and thirteen Pueblo I11 sites. Analysis of the fauna from the Sand Canyon Locality was performed by Walker (1989, 1990a). Walker utilised different criteria for his identifications than the analysts of the material from the Site Testing Program and the more recent material from Sand Canyon Pueblo. He frequently identified specimens to taxon when they could not be assigned to a specific element. Later identifications (Brand 1991, Driver et al. 1995) under the direction of Driver (1991) required that a specimen be designated as a specific element before it could be assigned to taxon, otherwise the bone was considered unidentifiable. In order to eliminate the discrepancy between the two methods, the data from Duckfoot was requantified using Driver's criteria. This involved recounting the NISP's of lagomorphs, artiodactyls and turkey/large bird from the raw data from the site (Walker 1989, 1990b). Entries were not counted unless they were identified as specific elements (i.e. femur or humerus, not long bone).

140 A comparison is drawn between the data from the Duckfoot site which dates to Pueblo I and the accumulated data from Pueblo I11 in an attempt to detect temporal change in turkey production in the Sand Canyon Locality. There is a notable increase in the percentage of turkey bone in comparison to lagomorphs and artiodactyls between the Pueblo I and I11 periods in the Sand Canyon Locality (See table 4.13). Pueblo III Castle Rock Pueblo Saddlehorn Hamlet Shorlene's Site Roy's Ruin Lillian's Site Troy's Tower Kenzie Dawn Hamlet G and G Hamlet Catherine's Site Lester's Site Lookout House Stanton's Site Sand Canyon Pueblo Total Table 4.13: Turkey proportions divided by time period from sites in the Sand Canyon Locality. The percentage of turkey in the Duckfoot assemblage is 2%, whereas in Pueblo I11 sites within the Sand Canyon Locality it averages 46% of the major economic species. There is an extensive range in the percentages of turkeys in the Pueblo I11 sites (between 32 and 70%) but it is clear even from the sites with the lowest percentages that turkeys were far more important than they were in

141 Pueblo I. Though the temporal depth in the Sand Canyon Locality sample is less than optimum, it matches the pattern described for the Mesa Verde region. These results indicate that turkey production was intensified throughout the Sand Canyon Locality, as was previously demonstrated by the regional sample. Intersite Variation Within the Sand Canyon Locality The sites examined within the Sand Canyon Locality were predominantly contemporaneous. The majority contain a prominent Pueblo I11 component occupied during the thirteenth century. These sites present an ideal opportunity to examine variation across space within what is postulated to be one community (Lipe 1992a). Not only are the sites largely contemporaneous but because they were tested as part of a series of projects by the Crow Canyon Archaeological Center, excavation and sampling strategies were similar allowing easy comparison between sites. The Sand Canyon Locality sample has been reduced for the following spatial analysis. The sites selected include twelve of the thirteen sites sampled for the Site Testing Program (Varien et al. 1992) and portions of the data from Sand Canyon Pueblo. One site, Mad Dog Tower (5MT181) was tested as part of the Site Testing Program but was excluded from this analysis due to its small sample size (Driver et al. 1995). The other two sites which were eliminated are the Duckfoot site, as it has a much earlier occupation date and the Green Lizard Site due to the use of incompatible methods for recording the faunal data. The assemblage from Sand Canyon Pueblo was subsampled to incorporate only the fauna analysed according to the guidelines set up in Driver's (1991) manual to retain compatibility within the sample. Walker's (1989) data from an earlier analysis were therefore excluded. To detect spatial variation in the faunal assemblage the sites in the sample

142 were grouped according to topographic location. The sites are located either on the mesa tops, the cliff/talus/bench slopes or in the Lower McElmo drainage (see Table 4.14) and were depicted earlier on a map of the Sand Canyon Locality (Figure 1.2). S i t e p h i c Number Site Name with SCP Location 5MT5152 Kenzie Dawn Hamlet No Mesa Top 5MT11338 G and G Hamlet No Mesa Top 5 MT3 951 Troy's Tower Nones Mesa Top 5MT3918 Shorlene's Site No Mesa Top 5MT3930 Roy's Ruin No Mesa Top 5MT3936 Lillian's Site No Mesa Top 5MT765 Sand Canyon Pueblo CliffITalus 5MT10508 Stanton's Site Yes Cliff/ralus 5MT10246 Lester's Site Yes Cliff/ralus 5MT10459 Lookout House Yes Cliff/Talus 5MT3967 Catherine's Site Nones Canyon Bench 5MT262 Saddlehorn Hamlet Yes Lower Canyon 5MT1825 Castle Rock Pueblo Yes Lower Canyon Table 4.14: Site Sample for Spatial analysis within the Sand Canyon Locality. 1) Mesa Top Sites. Mesas are steep-sided, flat plateaus which are intersected by canyons. They are located on the raised land above Sand Canyon Pueblo, in the Upper Canyon. This area is occupied by six sites in the sample (Shorlene's Site, Roy's Ruin, Lillian's Site, G and G Hamlet, Troy's Tower and Kenzie Dawn Hamlet )(See figure 1.2). The majority of these sites were occupied prior to A.D. 1250, though a few are also partially contemporaneous with Sand Canyon Pueblo. Each of the six sites are within 2 km of Sand Canyon Pueblo. 2) Cliff/ Talus1 Bench Sites. The mesa top where the above sites are located is intersected by Sand Canyon, which boasts steep slopes and numerous sites along its walls. Within the current sample their are four cliff/talus/bench sites (Catherine's Site, Lester's Site, Lookout House and Stanton's Site). All are

143 located in the upper Sand Canyon and are contemporaneous with Sand Canyon Pueblo (A.D ). These sites are also located within 2 krn of Sand Canyon Pueblo. 3) The Lower McElmo Drainage. The lower McElmo drainage is located at the southern end of Sand Canyon. Elevation in the lower canyon is at least 300m lower than the upper end creating a distinct microenvironment. The two sites (Saddlehorn Hamlet and Castle Rock Pueblo) in this area are the furthest from Sand Canyon Pueblo averaging about 7 krn down canyon. They are also contemporaneous with Sand Canyon Pueblo. Location Lower Canyon Saddlehorn Hamlet 44% 215 Castle Rock Pueblo 45 % 748 Average 45 % 963 Mesa Tops Shorlene's Site 32% 75 Roy's Ruin 49% 39 Lillian's Site 41% 153 Troy's Tower 56% 34 Kenzie Dawn Hamlet 33% 461 G and G Hamlet 42% 36 Average 36% 798 Cliff/Talus/Bench Catherine's Site 67% 284 Lester's Site 70% 89 Lookout House 66% 128 Stanton's Site 70% 757 Average 69 % 1258 Sand Canyon Pueblo 37 % 2099 Table 4.15: Percentage of turkey bone in Pueblo I11 sites in the Sand Canyon Locality by topographic locatian.

144 Lower Sand Canyon Mesa Tops Cliff/TaluslBench Figure 4.5: Bar graph representing the percentages of turkey versus lagomorphs and artiodactyls at Pueblo I11 sites grouped by topographic location from the Sand Canyon Locality.

145 4) Sand Canyon Pueblo. Sand Canyon Pueblo is considered a separate entity, for the purposes of comparison. Though it is technically a cliff/talus/'bench site due to its location on the canyon rim at the head of Sand Canyon, it is assigned its own category because of its extreme size in relation to all other sites in the sample. I I I Sand Canyon Pueblo Mesa Top Lower Canyon Cliff/Bench/Talus Site Figure 4.6: Average percent and ranges of turkey at Pueblo I11 sites in the Sand Canyon Locality, grouped by topographic location. To detect variation between the Pueblo I11 sites, the percentage of turkey at each site was calculated using the same methodology as described in the previous section for the analysis of temporal change. The percentage of turkey out of lagomorphs, artiodactyls and turkeys combined ranges between 32 and 70% at the individual Pueblo I11 sites in the Sand Canyon Locality. The majority of the sites fall between 32 and 49%. Four sites however, have percentages

146 between 66 and 70%. Interestingly, when the sites are grouped by topographic location as described above, all sites in the latter group are cliffftalusfbench sites. The turkey percentages and the ranges for the highest and lowest turkey percentage for each topographic location are depicted in Figure 4.6. Grouping the sites enabled a comparison between groups with larger sample sizes, potentially reducing error associated with sample size. An interesting pattern resulted. When the major economic faunal species were averaged the cliff/talus/bench site assemblage exhibited a turkey percentage of 69%. The averages of the other three groups are much lower, though they are similar to one another (37% at Sand Canyon Pueblo, 36% on the mesa tops and 45% in the lower canyon). The difference between the cliffftalusfbench sites and the three other groups combined has been tested to determine if it is statistically significant (see Appendix C and Table 4.16). Ninety-five percent confidence intervals were established for each of the two groups as follows:.385f for the combined group and.688f.022 for the cliff/talus/bench sites. The difference between the two groups is statistically significant as it is apparent that there is no overlap at two and even three standard errors (greater than 95% confidence interval). The low variation within the cliff/talusfbench group is also noteworthy, though the sample size is small. - i 3 l i f f P T a l u s l B e n c h r Sampled Sites Average Turkey Proportion Variance Standard Deviation Standard Error % Confidence Interval Range 95% Confidence Interval 0.665, , Table 4.16: Statistics comparing the cliff/talus/bench sites and the other Pueblo I11 sampled sites in the Sand Canyon Locality combined. 134

147 The trend depicted above is particularly interesting considering the proximity of the sites to one another. The sites on the mesa tops and on the cliff/talus/bench areas are all within 2 lun of Sand Canyon Pueblo (See Figure 1.2). As a result the variation seems to be related to the topographic environment of the site, which may in turn be associated with other factors (i.e. social or economic). These possibilities will be discussed in the following chapter. Discussion and conclusions integrating the results of the data relating to the four problems reviewed above will also be provided in Chapter V.

148 CHAPTER V DISCUSSION AND CONCLUSIONS Introduction This research has been designed to increase our awareness of the utility of biocultural data sets to substantiate studies which go beyond environmental reconstruction and subsistence strategies (i.e. interpretations of social and political structure). Faunal data from archaeological contexts in the Mesa Verde region have been utilised to address questions relating specifically to domestic animals and more generally to community and social organisation. In this final chapter the data which were presented in Chapter IV will be consulted with reference to the four problems addressed in previous chapters. The chapter commences by addressing the question of domestication and reexamining previous divisions of turkey breeds and subspecies. This is followed by an interpretation of the turkey's function in Anasazi society. Next, discussion centers on the final two problems which are concerned with temporal and spatial variation in the intensity of turkey production in the study area. Models are proposed to explain intensification in turkey production through time and across space. Finally the chapter concludes with a review of problems encountered during the study and some suggestions for future research. Is There a Turkey in the House? An attempt was made to determine if the turkey population in the sample was wild or domestic by examining both osteological characters and supporting evidence such as eggshell, gizzard stones, droppings, age structures and pens. Methods used in previous research to separate prehistoric southwestern turkeys

149 into breeds were also examined to test their validity. The conclusions are as follows. Osteological Evidence Previous research concerning southwestern turkeys has focused on the identification of subspecies based partially on osteological variation (Lang and Harris 1984, McKusick 1980a, 1986, Schorger 1961,1966). This was performed by recognising diagnostic characteristics and analysing measurements. In this research an effort has been made to reexamine the validity of these methodologies, while attempting to identify the species composition of the sample. Data presented in Chapter IV, reveal that characters on bone which have been used to separate subspecies of turkeys in the past could not be identified in this research. Based on this and evidence presented below the argument is made that the existence of breeds and domestic populations cannot be established by utilising osteological comparisons and measurements. McKusick (1986) herself states that not all individuals will be able to recognise the differences between breeds. She became aware of this after attempting to point out features to a naturalist who was unable to identify similarities or differences. The failure to recognise diagnostic characteristics in this research may have occurred if only one breed is present in the sample. For example, if the assemblage is composed exclusively of Large Indian Domestic turkeys, no obvious differences should be detectable amongst the specimens, as was the case. It was however, impossible to identify any of the specimens specifically as LIDS based on the existing descriptive criteria of diagnostic features. These differences should still be apparent regardless of whether there are one or more breeds represented in the sample. The failure to identify subspecies may have also been a product of the

150 experience of the researcher. The individuals involved in the pioneering research concerning prehistoric southwestern turkeys (McKusick, Hargrave, Schorger) were well seasoned in the field of avian zooarchaeology and have performed hands-on research with a range of turkeys from all comers of the Southwest. Certainly this has provided them with an eye for noting variation within turkey populations. However, if only researchers with a specialisation in turkey identifications are capable of recognising the distinctions, they are undoubtedly subtle and subjective and should be interpreted as representing different subspecies with extreme caution. Subtle differences may simply represent natural variation within a population composed of one species. It is indisputable that osteological variation exists within the species Meleagris gallopavo, but splitting the population into distinct subspecies or breeds based on this variation may be unwarranted. As previously mentioned few researchers working on faunal assemblages in various locations in the Southwest have met with success using McKusick's methods. In fact, the majority of the specimens which have been assigned to breed were identified by McKusick herself or with her consultation (Lang and Harris 1984, McKusick 1974, 1980b, 1981,1982). Others have questioned the validity of these identifications based on her failure to consider the effect of environmental variability on osseous phenotype (Breitburg 1988, Senior and Pierce 1989). The results of multivariate statistical analysis by Breitburg (1988) also failed to identify any significant differences between the proposed breeds. The results of multivariate discriminant analysis, size, shape and generalised distance coefficients and principal components analysis were unable to discriminate between any of the southwestern turkeys included in Breitburg's study, though members of each of the proposed groups were included in the sample. In addition

151 southwestern turkeys proved to be statistically indistinguishable from the southeast and northeast breeds of Meleagris gallopavo silvestris, although they could be discriminated from archaeological Mexican turkey specimens. Breitburg did recognise a reduction in tarsi and humeral length in some archaeological turkey samples (i.e. Gran Quivira) and attributes this to isolation from wild parent populations. Environmental and nutritional effects are cited as other possible contributing factors. Still, he also claims that "southwestern turkey populations and potential parent populations cannot be distinguished with confidence" (1988: 89). Unfortunately in the current study the sample was derived from a fairly homogeneous environment, so environmental effects could not be tested. This research into turkey domestication has revealed some potentially serious flaws in the methodology used to separate breeds and distinguish domestic from wild turkeys based on osteological characteristics. Fuither difficulties arose when attempting to determine the identity of the sample population by using metric data. It has been proposed above that not only the methodology is at fault but the division of the species Meleagris gallopavo merriami into subspecies is as well. A primary problem lies in the ambiguity of the definition of the proposed breeds in the first place. There are no modem parameters which define the extent of variation within each group since most of the breeds are extinct today. Instead the parameters are defined by the metric data which McKusick provides for the different groups (Table 4.1). These represent the ranges of means of populations which McKusick herself has identified as belonging to the different breeds using highly subjective criteria as previously discussed. It was shown above using Breitburg's research, that individuals identified to these proposed breeds cannot be distinguished statistically. Because the expected ranges of natural variability

152 for each breed are undefinable and the estimates that we do have are indistinguishable statistically, it is impossible to accurately assign an unknown individual to a subspecies based on metric data. Furthermore, individuals potentially fit within the given parameters of more than one of the possible breeds. The problem discussed above is better illustrated when considering data from the present study. The measurement data examined in this research exhibit a range of 4.5 cm (up to 25% of the total length of the element) in the greatest length of male tarsi. These differences are large enough to place individuals within McKusick's ranges of sample means for each of the four different breeds, though the mean of the tarsi in the current sample falls within the LID category. For example, when attempting to assign a breed to one tarsus with an arbitrary length of 160mm, the specimen appears to fit between the LID and MWT categories in McKusick's scheme. However, it also fits in the SID or the Gould's turkey range as a more extreme variant in the population. This distribution depicts that though the mean of the group falls within one breed the ranges for the different breeds are so close together that any given individual potentially fits into several categories. As a result it is impossible to assign individuals to a single group with certainty. This is a problem which corresponds only to the use of measurements. It must be pointed out that McKusick does not strongly advocate this method to differentiate various breeds, but suggests that character identifications are better suited (1986). Finally, by using means to identify the breeds represented within a single faunal assemblage the presence of more than one turkey subspecies or breed may be overlooked. If two groups were present in an assemblage averaging their greatest lengths would make it appear that individuals which belong to the less common breed belong to the more ubiquitous breed in the assemblage. Bimodal

153 distributions which often mark the presence of two populations only cause further confusion in this case due to sexual dimorphism and the methodological problems with sexing. Perishable evidence such as feather colouration and short, scaly tarsi has been used to distinguish SIDs from the other proposed breeds (McKusick 1986, Schorger 1961, 1966). Unfortunately these types of data have only been preserved under rare circumstances. Few mummified turkeys with these characteristics (identified as SIDs) have been recovered and were identified as SID's based on the anomalous feather colouration and tarsi. The osteological characteristic; short tarsi, was therefore also assumed to be associated with the SID. Short tarsi which are frequently recovered from the archaeological should not be automatically associated with SIDs as the sample size of mummified SIDs is too small to be reliable. Again, assigning subspecies based on this criteria is risky at best. Conclusion of osteological research In this research, it could not be determined whether the turkeys were wild or domestic based on osteology. This however, does not imply that they were not domesticated. Osteological change often occurs as a result of alterations in lifestyle particularly changes in diet, environment or activity. It is possible that though turkeys were under the control of humans, they maintained the freedom to forage for food and were only restricted to pens at night when they could not be observed(breitburg 1988, Cattanach 1980, Rohn 1971, Schorger 1961). This possibility has been supported by the presence of scattered droppings around some sites in the regional sample, and the recovery of high quantities of pine pollen and insects in the analysis of turkey feces. The presence of the latter two suggests that the turkeys were not feeding entirely in confinement but also had

154 some freedom to forage. This would prevent any severe lifestyle changes which may have resulted in genetic osteological change. A second possibility is that the captive turkey population was interbred with wild populations, thus preventing isolation from the wild gene pool (Schorger 1966). The presence of these factors would make it difficult if not impossible to separate breeds based entirely on osteological comparisons and measurements. It is further concluded that the division of prehistoric southwestern archaeological turkeys into three proposed breeds, the Large Indian Domestic (Meleagris gallopavo merriami), the Small Indian Domestic (Meleagris gallopavo tularosa) and Merriam's Wild Turkey (Meleagris gallopavo merriami) is unreliable and can not be statistically proven. It is unquestionable that patterned variation existed within the southwestern turkey population but the assignment of distinct breeds and/or subspecies is as yet unfounded. Much confusion has resulted from attempts to use proposed methods for recognising turkey subspecies since the affinity of groups has not been proven statistically. The breeds have also been defined by data which is ambiguous at best. As suggested by Senior and Pierce (1989) additional research into the effects of environmental factors on phenotypic plasticity is required to address the variation within the turkey population. Conclusion of contextual evidence Abundant evidence for the maintenance of turkeys at Anasazi sites was drawn from both the Sand Canyon Locality and the Mesa Verde region (see Chapter IV). Data which indicates that domesticated turkeys were present at archaeological sites includes eggshell, gizzard stones, juvenile turkey bones, turkey droppings and retaining structures such as pens. This evidence suggests that turkeys were born, raised and killed at human sites and thus were under

155 human control. In Chapter I11 it was determined that domestication refers to a relationship between humans and animals which may take a variety of forms along a continuum leading from limited to complex human control. The removal of animals from their natural environment and the control, maintenance and breeding of these animals therefore constitutes domestication. As a result, the turkeys within the sample in this project are assumed to be domestic animals. This does not exclude the possibility that turkeys were occasionally introduced into the archaeological record as a result of hunting or capture from the wild for interbreeding with the domestic population. The assumption that turkeys were domesticated is essential to the remainder of this research as patterning in the turkey assemblages is viewed as a reflection of human behaviour since the turkeys were under their control. This leads us to ask for what purpose were the Anasazi raising their turkeys? The Role of the Turkey in Anasazi Society The debate over the turkey's function in Anasazi society may be addressed by asking two primary questions: did the turkey play a predominantly ritual or economic role, and was it exploited primarily as a source of food or feathers? The responses to these questions are interrelated and by integrating them a comprehensive interpretation may be developed. Evidence from the Mesa Verde region provides a strong argument for the utility of turkey feathers as both ritual and economic objects. Feathers were not only an important resource at the time of the turkey's adoption by the Anasazi, but continued to be so throughout the occupation of the Colorado Plateau. A review of the feather artifacts recovered from several sites in the Mesa Verde region, predominantly those which now occupy Mesa Verde National Park, was presented in Chapter IV. The artifacts reflect both utilitarian and ceremonial

156 function. Feather robes, blankets and arrow fletching are believed to represent the latter, whereas prayer sticks, turkey burials and feather aspergills are believed to have spiritual value. Unfortunately, reports from sites outside of Mesa Verde National Park including those for the Sand Canyon Locality, failed to present evidence for the presence of feathers at their sites, largely because none existed. The abundance of feather artifacts at some sites and their nonexistence at others probably results from differential preservation. The presence of feather artifacts at Mesa Verde National Park is attributed to excellent preservation due to the arid climate and the protection of many sites from exposure as they are located in cliff overhangs. The quality of preservation at these sites is apparent due to the presence of other fragile remains such as vegetal fibres (basketry and sandals), leather, wood and archaeobotanical remains (Cattanach 1980, Hayes and Lancaster 1975, Rohn 1971, 1977, Swannack 1969). It is postulated that feather artifacts at sites outside of what is now Mesa Verde National Park were utilised during the site's occupation, but have since been lost due to their fragile nature. This is supported by the low representation of other types of fragile remains at these sites. It is proposed that feather artifacts were utilised at sites in the study region which were external to present day Mesa Verde National Park based on the material recovered from sites with optimal preservation, the conditions of preservation at the sites and the fragile nature of the remains. In light of the inventory of feather artifacts at the Mesa Verde sites it is argued that they were utilised for both ritual and utilitarian purposes. Evidence for ritual in the archaeological record is often ambiguous at best. Though artifacts may appear to have ceremonial function it is extremely difficult to evaluate the ideology with which they were treated. In this case, ethnographic and historic data have been utilised speculatively as analogy to aid in the

157 reconstruction of the past. Information provided by burials (human and turkey) and artifacts are also considered valuable tools for interpreting ritual function. Numerous ethnographic and historic references citing the importance of turkeys in Pueblo rituals were reviewed in Chapter 11. Feathers figured prominently in dances and katsina rituals and various turkey components were presented as offerings and were occasionally interred with human burials. Additional artifacts such as prayer sticks, prayer feathers, masks and costumes, which played roles in Pueblo ceremony were also often partially constructed from turkey components. Archaeological evidence which implies the ritual utilisation of turkeys is also available. This appears mainly in the form of articulated turkey burials and their frequent association with human graves and architecture argued to have ritual significance (i.e. kivas). Burials and artifacts (i.e. feather aspergills, prayer sticks, feather wrapped twigs) with potential ritual significance have also been recovered from all time periods after and including Basketmaker I11 in the Mesa Verde region, though they were not retrieved from sites in the Sand Canyon Locality. In light of the evidence, it is surmised that turkeys were an important source of raw material which played multiple roles in the Anasazi ceremonial sphere. Criteria proposed to be indicative of food utilisation are present in a variety of forms in the turkey assemblages from the Sand Canyon Locality and the Mesa Verde region. Applicable data in the form of cutmarks, breakage patterns, burning, ossified tendons, dispersion of bone and age and sex profiles was reviewed in the previous chapter and will now be discussed. Cutmarks are mentioned in the majority of reports from the Mesa Verde region and are found predominantly on the distal condyles of tibiotarsi. In the Sand Canyon Locality cutmark locations cotrespond well with those proposed by Lang and Harris (1984) to be the result of skinning and disarticulation

158 activities. It has also been suggested that an experienced individual would rarely leave evidence of butchering on the skeleton, but cutmarks may instead be the product of a sloppy job (Lyman 1979). If this is the case we would expect that a much greater proportion of the bones were butchered than the cutmarks suggest. Evidence for burning is inconclusive in terms of food utilisation. It is difficult to determine whether bones were burned during cooking or as a result of disposal in hearths or structures which were later burned. It is also possible that distinctive patterns which were created by roasting were later obscured by additional burning following disposal. Localised burning is present but is not ubiquitous enough to provide substantial evidence for cooking. Though it is probable that some of the burning in the assemblage resulted from cooking, this can not be stated with certainty. In addition, it is likely that methods other than roasting were utilised for cooking meat, therefore cooked bone need not be burned. Boiling is a likely cooking alternative which leaves no visible effects on bone. Capone and Shoeninger's (1991) study attempted to detect boiling by measuring amino acid racemization in turkey bones. Unfortunately, this method was proven ineffective unless bones were boiled for longer than 1.5 hours and were unaffected taphonomic processes. It is therefore concluded that localised burning is the only reliable evidence of cooking in the assemblage. It was identified in the Sand Canyon Locality assemblages but in low frequencies. Breakage patterns on the humerus, tibiotarsus and radius were examined to provide insight into the use of cancellous bone as a dietary supplement. It was demonstrated that the spongy proximal end of the tibiotarsus was virtually absent in the assemblage though the more compact distal end was frequently represented. The humerus which was expected to express a similar pattern, was instead more similar to the control sample, the radii. The radius has minimal cancellous bone and articulated ends of approximately equal size. Both the

159 humerus and radius in the Sand Canyon Locality assemblages had approximately 50% representation of both proximal and distal ends. The proximal end of the humerus however was expected to be preserved less frequently as a result of cancellous bone extraction. It is possible that the distal end of the humerus was also large enough to warrant utilisation for cancellous bone extraction unlike the distal tibitotarsus and proximal and distal radius which are dense and smaller in volume. The primary limitation for this interpretation involves the effects of taphonomic factors which are more likely to be destructive to bones with large spongy sections because they have larger surface areas and are less dense than cortical bone. The humerus, however was not affected to the same degree as the tibiotarsus as would be expected if both were altered by taphonornic factors. Articulated burials were occasionally recovered from sites in the Mesa Verde region (Bertram 1989, Cattanach 1980, Rohn 1971, Schorger 1966), however all turkey assemblages were dominated by dispersed, disarticulated bone. All bones in the Sand Canyon assemblage were also disarticulated and recovered from dispersed contexts. In addition, turkey lower leg bones were not found in situ with ossified tendons, which implies that the bones were discarded after the flesh was removed. Because the Sand Canyon Locality fauna was not examined in situ during excavation, contextual information regarding ossified tendons is sketchy at best. Still, the contextual information that is available suggests that turkeys were taken apart and defleshed intentionally, presumably for consumption. Finally the age and sex profiles from the Sand Canyon Locality were evaluated. Data from other locations in the Mesa Verde Region were not examined due to the of lack of information and incompatibility between the different data bases. The distribution of the sexes in the Sand Canyon Locality is virtually equal which suggests that one sex was not preferred over the other. In

160 terms of age the population is dominated by adult individuals (94%). This scenario is most supportive of a population raised for feathers since both sexes are equally able to produce them. Feathers from male and female turkeys are similar except for some variation in colouration. High proportions of adults are also expected to represent a population raised for feather utilisation because feathers are continually produced throughout a birds lifetime, thus, it is beneficial to keep birds until they die naturally as feathers can be harvested multiple times. It is expected that a profile of a turkey assemblage raised for meat would indicate high proportions of young adult males as they would be culled when they reach their maximum size to minimise energy expenditure. Females are expected to survive longer as they are required not only for reproduction, but potentially also as egg producers which may have served other functions as well (i.e. food, glaze on painted masks [McKusick 19821). In this study, young adults were not distinguished from old adults making it impossible to discern such a pattern. Interpretations of the population structure are made with caution as it is necessary to assign value to the different turkey resources to draw conclusions. It is difficult to determine, for example, if the colouration of male feathers was preferred over female feathers, or whether female birds were valued as a more delectable meat source. Preferences of past peoples are ambiguous at best but had the potential to greatly influence the age and sex structure of the population. The high proportion of turkey artifacts in the faunal assemblage indicates the value of turkey bone as a source of raw material. It is probable that it became more popular due to its increased availability, which may have led to the recognition of its attractive properties for tool manufacture. As frequencies in raw turkey bone rose so did the number of artifacts which were created from it. This may explain why turkey bone artifacts are rare at early sites from

161 Basketmaker I11 and Pueblo I, as turkey bone was low in availability as a raw material. As a result, it appears that as a resource bone was secondary to meat. Conclusion In light of the data presented in Chapter IV and the interpretation presented above it is concluded that the turkey was a multi-purpose resource which played an important role in both the ritual and economic spheres of Anasazi society. It is suggested that its function was both utilitarian and ceremonial and that it served as a source for feathers, meat, bone and eggs. It is further proposed that the turkey served multiple functions since its inception in Anasazi culture. The utilisation of feathers predominated during the Basketmaker periods and Early Pueblo I, as reflected by the paucity of turkey bone and indicators of food utilisation during these early periods. It is unclear whether the turkeys were being raised at Anasazi sites at this point. Before this can be determined it is essential to establish with certainty, the prehistoric range of Meleagris gallopavo merriami. If the turkey was not present in the Mesa Verde region prehistorically it would not be possible to obtain specimens by local hunting. There is evidence for eggshell and gizzard stones at sites at this time, however no pens or turkey droppings have been identified. It is possible that low intensity turkey production which left a light archaeological signature in comparison to the more intensified production of later times was taking place. Alternatively, wild turkeys may have been available to the Anasazi prehistorically though they were not domesticated until Pueblo I. At this time (Pueblo I) more turkey bone appears in the archaeological record and pens and turkey dung make their appearance at numerous sites in the Mesa Verde region. Increased availability also led to the increased utilisation of turkey bone as a raw material for tool manufacture.

162 Temporal Change: Intensification in Turkey Production In previous discussion regarding quantification (see Chapter I) it was determined that increases in turkey proportions potentially indicate intensification, though the possibility that they may also represent shifts in the faunal mix can not be ignored. The argument for an increasing emphasis on turkey production in the Mesa Verde region during the Anasazi occupation, is supported not only by higher proportions of turkey in assemblages, but by increasing frequencies of turkey elements in faunal assemblages. In early assemblages (Basketmaker I11 and Pueblo I) turkeys were represented by very few elements regardless of sample size. This results in lower turkey proportions and also suggests that turkey utilisation was minimal regardless of the proportions of other species. In most later assemblages the frequencies of turkeys is much higher than earlier times regardless of sample size, therefore indicating that the trend toward increased production of turkeys is real and not simply a reflection of changes in the use of other species. Artiodactyls do appear to be decreasing in some assemblages by Pueblo 111, however this still does not fully account for the dramatic increases in turkey proportions. In addition, the number of turkey pens excavated in Mesa Verde sites increases as Pueblo I11 approaches, further reflecting an increase in energy investment. In light of the above discussion, it is concluded that the dramatic increases in the proportion of turkeys in regional faunal assemblages between Basketmaker I11 and Pueblo 111 (See Chapter IV) is indicative of the intensification of domestic turkey production in the Mesa Verde region. The increase is also apparent on a local level in the turkey population in the Sand Canyon Locality between Pueblo I and 111. Initially it appears that intensification and aggregation are correlated as they both arose during Pueblo 111. Instead the specific conditions surrounding

163 aggregation and intensification must be considered. In this research two episodes of aggregation from differing time periods were compared to detect whether aggregation was the causal factor behind the intensification of domestic turkey production. Two notable episodes of aggregation took place in the Mesa Verde region during Anasazi prehistory. The first occurred during Pueblo I in the late A.D. 800's and was exemplified by changes in settlement patterns in the Dolores region (Varien et al. 1992). The second episode took place on a large-scale during Pueblo 111, as illustrated by changes in the Sand Canyon Community (Adler 1990a, 1990b). An examination of the faunal data from the Dolores Archaeological Program indicates that the proportion of turkey in the faunal assemblage was relatively constant throughout the process of aggregation, exhibiting only nominal increases (Neusius 1986). Turkey proportions ranged from 1% in Modeling Period 1 (A.D ) to 6% in Modeling period 5 (A.D ). Thus, it cannot be convincingly argued that large scale intensification in turkey production took place during this episode of aggregation. Data presented in Chapter IV however, indicate large increases in the proportion of turkey in the Pueblo 111 faunal assemblages in the Sand Canyon Locality. This pattern is also apparent in the larger regional sample. What conditions were present during the second aggregation episode but not the first, to induce intensification? On closer examination it appears that the intensification of turkey production in the Sand Canyon Locality took place prior to large-scale aggregation into Sand Canyon Pueblo (ca. A.D. 1250). Data from mesa top sites which were occupied in the early thirteenth century indicate that turkey production was already intensified in relation to previous times. On average 36% of the mesa top faunal assemblages were composed of turkey,

164 whereas Pueblo I assemblages contained only 2% turkey. By this time aggregation had begun on a small-scale, population density in the Sand Canyon Locality was on the rise, mobility was more restricted and agricultural strategies were being intensified. These factors, however, did not trigger large-scale aggregation until ca. A.D If intensification in turkey production was not induced by pressures created by large aggregated villages as originally expected, what was the cause? It is maintained that aggregation and intensification are correlated due to causation by similar conditions. It is suggested, however, that the two are triggered when the causal factors reach differing levels. By examining aggregated situations in both the Dolores Valley and the Sand Canyon Locality it appears that population density was a crucial differentiating factor, when acting under conditions of restricted mobility, reduced resources and agricultural intensification. It appears that regional population density at the time of aggregation in the two communities differed. Local population densities prior to both aggregations were high as a result of increases over the preceding periods due to migration and natural growth. Density across the region however was lower during the former event (Varien et al. 1994). The settlement pattern in the Mesa Verde region during Pueblo I (A.D ) when the Dolores aggregation took place was marked by clustering. In some areas site densities were high (i.e. the Dolores Valley), however these were separated by sparsely populated hinterlands (Varien et al. 1994). It has been argued that the sequence of rapid population aggregation, short-use life and subsequent abandonment characteristic of the Dolores region reflects a strategy of high mobility. Groups migrated in and by way of low intensity agriculture depleted the environment then moved on in search of more productive areas (Kohler and Matthews 1988, Kohler 1989). This

165 was a viable strategy during Pueblo I when mobility was a reasonable option and regional population densities were still low. The inhabitants of the Dolores Valley were able to offset higher demands for resources created by increasing populations by relocation and mobility. It is argued that when agricultural intensification did begin it was due to restrictions by climatic and geographic circumscription rather than population circumscription from surrounding communities (Schlanger 1988). During the late 800's the territory surrounding the area was not heavily utilised thus it could be exploited for hunting with little competition from other communities. As a result it is argued that wild game reduction did not occur and big game hunting remained a viable option to obtain animal protein. This is supported by the high quantities of artiodactyl remains in the Dolores faunal assemblage. Artiodactyls represent between 21 and 57% of the major economic faunal groups over the seven periods of occupation in the Dolores Valley (Neusius 1986). At the time of aggregation artiodactyls were well represented with proportions ranging between 41 and 47% of the assemblage. Overall, this pattern indicates that large game was available for procurement throughout Pueblo I at least in the Dolores River Valley. The Pueblo I11 inhabitants of the Sand Canyon Locality were also exposed to increasing pressure from climatic and geographic circumscription as the farmbelt narrowed due to climatic effects (Schlanger 1988). Regional populations on the Colorado Plateau also rose dramatically during this period (Varien et al. 1994). By Pueblo 111, sudden regional population increases and restricted mobility on the Colorado Plateau potentially led to wild game scarcity in the Sand Canyon Locality, largely due to the increased numbers of people hunting in a limited area. In addition, pressure from populations in neighbouring localities may have led to hunting from several directions, thus further reducing the large game

166 population and lowering its potential as a reliable meat source. Increased utilisation of land for agricultural activity has also been known to drive large animals from an area and may have contributed to the reduction. A reduction in the hunting of big game is reflected in the faunal assemblages in the Sand Canyon Locality (Brand 1991, Driver 1993, Driver et al. 1995). Artiodactyls are represented in extremely low frequencies in the smaller sites in the Sand Canyon Locality (total approximately 2%). Instead highly valued large game resources are concentrated at Sand Canyon Pueblo (15%) which is interpreted as the major integrative and ceremonial center in the community. This implies that big game was being conserved, possibly for ritual events or social feasting. It seems likely that only under conditions of scarcity would the inhabitants of the smaller villages stop consuming artiodactyls since they were highly valued. Overall big game represent a minimal proportion of the faunal assemblages in the Sand Canyon Locality in comparison to their proportions in the Dolores assemblages (Neusius 1986). Recent isotopic and coprolite studies have concluded that the Anasazi depended primarily on plant resources and exploited animal protein as a secondary resource (Decker and Tieszen 1989, Minnis 1989, Stiger 1979). It has also been argued convincingly that all of the Anasazi's nutritional requirements could have potentially been met by cultigens alone in the Sand Canyon Locality (Van West and Lipe 1992). The diet of domestic animal populations was probably largely composed of cultigens as well. This is depicted in the Sand Canyon Locality where isotopic analysis of turkey bones from Pueblo I11 components had delta 1 3 values ~ of -10.8, -9.5 and -9.9 respectively. This indicates a diet high in C4 plants, notably maize in the study area (M.A. Katzenberg personal communication to J. Driver 1993). Despite the emphasis on plants, evidence for the use of animals as a secondary subsistence source is

167 abundant in the archaeological record. It is argued here that meat was a valued resource, and an important dietary component even if it was not nutritionally essential. Animals provide a more complete protein than plants (Nickens 198 l), and their procurement is often associated with status (Kent 1989). Meat and fat are also highly desired resources in virtually all ethnographic societies (Abrarns 1987, Hayden 1981, Speth and Spielmann 1983). In addition, animals were not only valued as a food source but for secondary products as well. Fur, leather and feathers are essential for the manufacture of warm clothing to enable survival during cold winters. Animal products also figure prominently in the ceremonial sphere of past and present societies. Finally, bone is a strong but pliable substance which is well suited for the manufacture of tools of various functions. In sum, animal resources were undoubtedly highly valued for both utilitarian and ceremonial functions. It is therefore expected that if an important animal source is reduced, it will be compensated for by exploiting a more available fauna. Population increases are expected to have a similar effect on animals as they do on plants. If wild resources are unable to meet the needs of a growing population and mobility is not an option the intensification of domesticates is expected to result. Because wild resources can not be intensified in response to increasing demands, increases in population density are expected to promote mobility, reduce wild resource availability and/or lead to the intensification of domestic resources. Domestic resources have the potential to be controlled and manipulated by human groups to achieve greater yields (to a point) and thus support higher populations. In the case of animals, turkeys were the primary domestic animal in the region (though dogs were also important), thus turkey production is expected to intensify if wild game was unable to meet the needs of the existing population. Domestic animals aquire a higher investment of energy since they require regular maintenance and must be fed (Earle 1980), it is

168 therefore likely that the intensification of turkeys took place at a time when wild game was in low supply. Conclusion As proposed above, it is unlikely that wild game was scarce during the Dolores aggregation thus the intensification of turkey production was unnecessary to meet demands for animal protein and did not occur. Alternatively, the later aggregation in the Sand Canyon Locality took place when regional populations were higher, mobility was restricted and more land was under cultivation. Wild game populations were therefore reduced as a result of overhunting and infringement by humans on their territory. It is suggested that the intensification in domestic turkey production was undertaken as a suitable alternative. It is further proposed that aggregation did not encourage the intensification of domestic animals in the study area, but the two were promoted by similar conditions. Adler's model (1990a) of aggregation for the Sand Canyon Locality argues that decreased mobility, increased population densities, agricultural intensification and resource scarcity were causal factors. These are in effect the same causal factors which appear to have led to the intensification of turkeys in the Sand Canyon Locality. Though aggregation and intensification are caused by similar circumstances their thresholds differ. The intensification of turkey production took place during early Pueblo 111, as reflected in the mesa top site assemblages (A.D ) when aggregation had begun on a small scale (average site size doubled from 6 to 13 rooms) but prior to the coalescence of Sand Canyon Pueblo. It is argued that big game populations were reduced by this time, though conditions had not yet reached the point where large-scale aggregation was a necessary strategy.

169 The difference between the Dolores and Sand Canyon Locality situations seems to be a matter of degree, namely in population density. A higher regional population density in the Mesa Verde region during Pueblo I11 led to restrictions in mobility which prevented movement to exploit neighbouring resources. Higher populations also created increased demands on wild fauna and flora and agricultural production. These factors induced the intensification of domestic plants and animals to ensure sustenance of the population. Alternatively, in the Dolores area, mobility was prevented by climatic and geographic but not by population circumscription. Arable land was limited by climatic and geographic factors which led to the intensification of agriculture but not animals. In the surrounding hinterlands low human population densities persisted as most people were accumulating in proximity to arable land. Population density in surrounding areas were also lower than they were in Pueblo I11 allowing populations of large animals to thrive and avoid overexploitation. Big game remained accessible as is reflected by the high quantities of large mammal bones in Dolores faunal assemblages (Neusius 1986). Intersite Variation in Turkey Production Within the Sand Canyon Locality A distinct pattern in the proportion of turkey in Sand Canyon Locality faunal assemblages emerges when the Pueblo I11 sites are divided by topographic location. Between A.D , prior to the aggregation of Sand Canyon Pueblo, a dispersed settlement pattern of small mesa top sites was characteristic of the Sand Canyon Locality. Ca. A.D these were abandoned and the community aggregated into Sand Canyon Pueblo, a large multi-component site located on the canyon rim at the head of Sand Canyon. Several small settlements occupying the cliffs, taluses and benches along the canyon walls were also founded at this time. All of these sites are located within 2 krn of Sand Canyon

170 Pueblo and are believed to be integrated within the same community (Adler and Varien 1991, Varien et al. 1994). The lower canyon (McElmo drainage) was also occupied by numerous small settlements during Pueblo 111. Castle Rock Pueblo is the largest site in this area, though it is not of the same magnitude as Sand Canyon Pueblo. In Chapter IV it was demonstrated that cliff/talus/bench sites located in Sand Canyon had significantly greater proportions of turkey (69%) in comparison to lagomorphs and artiodactyls than the other topographic areas. The mesa tops, lower canyon and Sand Canyon Pueblo had percentages of 36%, 45% and 37% respectively. It was also proven that the difference between the cliff/talus/bench sites and the remaining sites in the Pueblo I11 sample combined is statistically significant. Driver et al. (1995, Driver 1993) consider both environmental/economic and social models to explain the spatial patterning of turkey production within the Sand Canyon Locality faunal assemblages. Environmental/econornic models have largely been discounted as ineffective explanations. For example, one environmental model proposes that the faunal assemblages differ as they are products of their topographic location. Differentiation between faunal assemblages is proposed to result from hunting animals which inhabit the environment surrounding the site. Deer, for example are not expected to be numerous in the canyon sites as they inhabit open environments such as mesa tops. This model may be discounted as each of the canyon sites are located within several hundred metres of the mesa top sites and are therefore expected to have overlapping catchment areas and access to similar species. Sand Canyon Pueblo also occupies a canyon location, yet its assemblage expresses distinctively different faunal proportions.

171 Driver et al. (1995, Driver 1993) also consider social models as alternate explanations for the patterning in the faunal assemblages. In these, inhabitants of sites which were contemporaneous with Sand Canyon Pueblo are considered to be either a) integrated within the community or b) excluded from and intimidated by the larger community. The first two hypotheses operate under the premise that the cliff/talus/bench sites were united as one community which was integrated by Sand Canyon Pueblo. The first model suggests that Sand Canyon Pueblo served primarily as a ceremonial center, and a focal point for feasting. Its faunal assemblage is therefore expected to be dominated by species which were ritually valued. Ethnographically, artiodactyls are credited with high prestige due to the ritual importance of hunting. To the contrary, the acquisition of turkey meat is attributed no significant ritual importance, though turkey feathers were valued (Gnabasik 1981). At Sand Canyon Pueblo there is evidence for higher percentages of turkeys in room dominated architectural blocks, whereas artiodactyls and lagomorphs appear in greater frequencies in kiva dominated blocks. Kivas are commonly believed to be associated with ritual and ceremonial activity, thus the faunal distribution supports the model. Alternatively, the distribution of fauna in the canyon sites (higher proportions of turkey, lower proportions of artiodactyls and lagomorphs) is viewed as representative of domestic activity. It is proposed that families raised and maintained turkeys near their homes and lagomorphs were hunted in the canyons. Artiodactyls are uncommon in these assemblages as they were publicly hunted for ritual activities which centered around Sand Canyon Pueblo. A more intensive examination of the Sand Canyon fauna must be undertaken before this hypothesis can be more fully evaluated.

172 The second social model suggests that the cliff/talus/bench sites were production centers for the community centered at Sand Canyon Pueblo. It is proposed that the inhabitants of the cliff/talus/bench sites were producing turkeys to trade with Sand Canyon Pueblo, though it has not been determined what benefits or products they received in exchange. This model seems improbable as it is unlikely that the cliff/talus/bench residents would be willing to give up the important activity of big game hunting to focus entirely on domestic animal production (Gnabasik 1981). Until evidence for trade between the sites is recovered this model remains highly speculative. A third hypothesis operates under the condition that the canyon sites were not integrated within the community centered at Sand Canyon Pueblo but were in competition with it. The inhabitants of the cliff/talus/bench sites are viewed as 'socially/politically disadvantaged' due to the presence of differential access to hunting grounds. Using intimidation and sheer numbers Sand Canyon Pueblo was able to dominate access to big game hunting on the mesa tops, restricting the cliff/talus/bench groups to hunting in the canyon. The large percentages of turkey in the canyon site assemblages are viewed as attempts to compensate for the minimal returns from hunted meat. As an alternate meat source turkeys are ideal as they could be raised at the sites to ensure access to adequate animal resources. In this research an additional hypothesis is proposed to explain intersite variation in turkey proportions in the Sand Canyon Locality. This model includes components of the social hypotheses reviewed above. The following discussion is presented within the framework of resource access as discussed in Chapter 111. Following Adler and Varien (1991) it is maintained that the sites in the Upper Sand Canyon are incorporated within one community (the Sand Canyon Community). The Lower Canyon sites may be an exception, alternatively

173 belonging to a separate, smaller community for which Castle Rock Pueblo was the integrative center. In this case ties with the Sand Canyon Community would have existed primarily through trade (Gleichman and Gleichman 1992). The sites in the Upper Sand Canyon are believed to be integrated by Sand Canyon Pueblo, rather than in competition with it as described in the previous model. It is argued that a land tenure system providing access to multi-household groups was in effect in the Sand Canyon Community during Pueblo 111. The rules governing access to land were mediated by the community. The land tenure system enabled the resolution of conflict brought on by competition over valued resources and ensured that the energy investment of the primary economic group was protected. It should also be noted that land tenure systems which operate below the communal level potentially result in differential access to quality soil. During Pueblo I11 the presence of numerous field houses within the Sand Canyon Locality and the contemporaneity of their development with intensified agricultural practices, suggests that formalised land tenure systems below the communal level were in existence. Primary access groups for some agricultural activities were smaller than the community, probably at the multi-household level (Adler 1992b). Field houses are believed to be indicators which demarcate land ownership (Kohler 1992). It is therefore suggested that ownership of land plots was recognised by the community and passed on to future generations, even after aggregation into Sand Canyon Pueblo. It has been demonstrated that the association between habitations and optimum arable land was at its lowest during Pueblo I11 (Adler 1992b). This is largely the result of the shift in site location to the canyon walls where soil was patchier. In addition the best arable land was probably already under cultivation. Migrants to the Sand Canyon Locality during Pueblo I11 when populations rapidly increased, would have no choice but to farm less arable land as nothing

174 else was available. As a result, some multi-household groups are believed to have had access to more marginal land than others, creating the potential for differential wealth accumulation. The community is believed to have collapsed before economic differentiation became marked. Though Van West (1990) has demonstrated that there was sufficient arable land to adequately maintain the Sand Canyon Community, this assumes that there was movement of agricultural products from areas of high to areas of low productivity. Productivity in the Sand Canyon Locality was not necessarily evenly distributed. It is proposed that some landowners had access to land capable of producing a surplus whereas others may have had to supplement their diet using alternative resources. The most viable alternative under the latter conditions was to intensify the production of domestic animals. Conclusion It is concluded that the individuals who inhabited the mesa top sites during the early thirteenth century later coalesced to form Sand Canyon Pueblo. By demarcating their land with symbols of ownership such as field houses they were able to retain access to their land and therefore obtained sufficient agricultural yields for their sustenance. Formalised tenure systems were in place to provide the community with the jurisdiction to ensure that the rules of ownership were upheld and to resolve conflict. The occupants of the cliff/talus/bench sites are proposed to have migrated into the Sand Canyon Locality at a later date and therefore received marginal land, as the optimum arable land was already claimed by the original inhabitants. Turkey production is argued to have intensified to serve as a dietary supplement. This is reflected in the high percentage of turkey bone in the cliff/talus/bench sites during late Pueblo 111.

175 Problems Encountered The primary problem encountered during this research concerns the incompatibility of faunal data within the regional sample. An attempt was made to elucidate regional trends in domestic turkey production, however zooarchaeological information was rarely reported consistently. Variation existed in several aspects of the faunal reports including quantitative strategies, identification procedures and excavation and sampling methods. Inconsistencies between quantification methods were particularly problematic. Though NISP was the most common strategy used, MNI was a popular alternative. The two methods were rarely presented together which created difficulties, as absolute values quantified using NISP cannot be directly compared with those using MNI. Many researchers also neglected to explain the methods they used to calculate MNI, though alternate methods may result in significantly different results (1979). In this study, proportions were used in an attempt to eliminate quantitative units, however this introduced additional confounding variables (see below). The use of dissimilar criteria to guide identifications also creates discrepancies within the data. An example was provided earlier (see Chapter IV) when attempting to compare fauna analysed by Walker with that identified according to Driver (1991) in the Sand Canyon Locality. It was necessary to requantify Walker's data using Driver's criteria to enable comparison with the other sites in the sample. Fortunately this was possible as Walker provided an appendix of his raw data which included clear descriptions of each element and species. Driver (1992) makes a call for the standardisation of faunal identifications, though unfortunately standard criteria has not yet been adopted and probably will not be in the near future. The methods currently in use are diverse as are the arguments of support advocated by the researchers who use

176 them. To establish a system on which we all agree will be a laborious if not impossible task. A second problem concerns the use of proportions and/or percentages to indicate spatial and temporal variation. Proportions are a measure which depict the amount of a group in relation to the total of the groups being compared. The total proportions of all groups being evaluated must equal one. In this study the number of turkey bones in a given faunal assemblage was divided by the total number of artiodactyl, lagomorph and turkey bones to determine the proportion of turkey. This allows for easy comparison of the intensity of turkey production between sites and over time. Problems arose because proportions are relative measures. An increase in the frequency of one component does not necessarily represent an absolute rise in that component but may indicate a decrease in a different component. Because the total proportions in each assemblage must add up to one, if one component increases an associated decrease must occur in one or more of the other components. As a result is extremely difficult to determine what changes are absolute changes and which are simply responding to other change. In defense of this method there are few viable alternatives. The use of absolute values in faunal analysis introduces an infinite number of confounding variables making their utilisation unworthwhile for this research. A second alternative is to use an external artifact class such as ceramic cooking pot sherds or debitage as a control against which to compare the data class under examination (in this case turkey bone). This approach assumes that the external data classes accumulate at a constant rate. Population and site occupation times have been estimated to determine accumulation rates. When tested these estimates have proven to be fairly reliable (i.e. Blinman 1986), however several new variables are introduced into the analysis increasing the potential for error.

177 These are related to the reliability of population or site occupation length estimates, and the assumption that individuals produce a standard quantity of pot sherds over a certain length of time. The potential of this method is recognised and it is recommended as a possible venue for further research. It is essential to consider the problems discussed above when attempting to identify regional trends such as the one depicting turkey intensification in this study. Suggestions for Further Research Additional research which would benefit the current study involves the examination of environment and its effect on bone remodeling in turkeys. This type of research has potential to clarify the confusion surrounding the presence of turkey breeds and subspecies. Several researchers have questioned the reality of the races claimed to be present in southwestern archaeological assemblages and have proposed that osteological variation may instead have been created by environmental and/or nutritional factors. These may have induced osseous change in the turkey skeleton but do not necessarily predicate the naming of new subspecies. Studies which examine the bone of modem turkeys living under controlled environmental and/or nutritional conditions would be useful for determining their effects on the shaping of turkey bone. To date no research of this nature has been performed. Nitrogen isotope analysis of human bone is also suggested as a promising venue for further research. Nitrogen isotopic ratios (15~:14N) preserve in bone collagen and may be used to estimate the importance of plant and animal products in the human diet (Ambrose and DeNiro 1986). The 1 5 isotope ~ concentration in bone increases concurrently with trophic level. In this research the technique would be useful for comparing the diet of individuals who inhabited the cliff/talus/bench sites versus the inhabitants of the remaining Sand

178 Canyon Locality sites during Pueblo 111. The model presented in this study proposes that the inhabitants of the cliff/talus/bench sites were supplementing their diet with meat to compensate for lower agricultural yields as they were farming marginal soils. As a result they are expected to exhibit higher 1 5 values ~ than the other groups, which would indicate that they were consuming a larger proportion of meat. Some isotopic studies on Sand Canyon Locality human bone are currently being undertaken by Annie Katzenberg at the University of Calgary. In this study an attempt was made to promote the utilisation of alternate data sets such as faunal or archaeobotanical remains to address problems dealing with social and community organisation. Traditionally this type of research has been performed using architectural remains, settlement data or 'high status' artifact classes. Future archaeological research will be greatly benefited if it is recognised that ecological data sets traditionally associated with economic and environmental research have high potential to substantiate research outside of these arenas as well. Overall, hypotheses which can be corroborated with multiple data sets provide much stronger arguments than those based on one type of data. Therefore, it is hoped that the results of this study can be used to support research performed by the Crow Canyon Archaeological Center and within the greater Southwest.

179 REFERENCES Abrams, H. L. Jr The Preference for Animal Protein and Fat: A Cross-Cultural Survey, in Food and Evolution Toward a Theory of Human Food Habits, pp Temple University Press, Philadelphia, PA. Adams, K.R The Environmental Archaeology Program, in The Sand Canyon Archaeological Project A Progress Report, edited by W.D. Lipe, pp , Occasional Paper No. 2, Crow Canyon Archaeological Center, Cortez, CO. Adler, M. 1990a Communities of Soil and Stone: An Archaeological Investigation of Population Aggregation Among the Mesa Verde Region Anasazi; A.D Ph.D. Dissertation, Department of Anthropology, University of Michigan, Ann Arbor, MI. 1990b Community, Aggregation and Land Tenure: The Search for Common Ground. Paper presented at the Annual Meetings for the Society of American Archaeology, Las Vegas, NV. 1992a Land Tenure Systems and Prehistoric Land Use: Perspectives from the American Southwest. Paper presented at the 57th Annual Meetings of the Society for American Archaeology, Philadelphia, PA. 1992b The Upland Survey. In The Sand Canyon Archaeological Project A Progress Report, edited by W. D. Lipe, pp Crow Canyon Archaeological Center Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO Population Aggregation and the Anasazi Social Landscape: The View from the Four Comers, in The Ancient Southwestern Community: Models and Methods for the Study of Prehistoric Social Organization, edited by W. H. Wills and R.D. Leonard, pp , University of New Mexico Press, Albuquerque, NM. Adler, M., and M. Varien 1991 The Changing Face of the Community in the Mesa Verde Region A.D ; Paper presented at the Second Anasazi Symposium, Mesa Verde National Park, CO. - Akins, N. J Prehistoric Faunal Utilization in Chaco Canyon Basketmaker I11 through Pueblo 111. In Environment and Subsistence of Chaco Canyon, New Mexico, edited by F. J. Mathien. National Parks Service, Albuquerque, NM. Ambrose, S.H. and M.J. DeNiro

180 1986 Reconstruction of African human diet using bone collagen carbon and nitrogen isotope ratios, in Nature 319: Armitage, P Domestication of Animals. In Bioindustrial Ecosystems, edited by D. J. A. Cole and G. C. Brander, pp. 5-30, Elsevier Science Publishers, Amsterdam, Netherlands. Bertram, J. B An Analysis of Archaeological Assemblages, Nonhuman Burials and Worked Bone Items: Hovenweep Laterals Project Appendix F. In Archaeological Excavation on the Hovenweep Laterals, edited by J. N. Moms, pp , Complete Archaeological Service Associates, Cortez, CO. Blinman, E Technology: Ceramic Containers, in Dolores Archaeological Program: Final Synthetic Report, edited by D.A. Breternitz, C.K. Robinson and G.T. Gross, pp United States Department of the Interior, Bureau of Reclamation, Denver, CO. Boessneck, J., and A. von den Driesch 1978 The Significance of Measuring Animal Bones from Archaeological Sites. In Approaches to Faunal Analysis in the Middle East, edited by R. H. Meadow and M. A. Zeder, pp , Peabody Museum Bulletin 2. Peabody Museum of Archaeology and Ethnology, Cambridge, MA. Bokonyi, S Archaeological Problems and Methods of Recognizing Animal Domestication. In The Domestication and Exploitation of Plants and Animals, edited by P. J. Ucko and G. W. Dimbleby, pp London. Boserup, E The Conditions of Agricultural Growth. Aldine Publishing Co., Chicago, IL. Bradley, B. A Excavations at Sand Canyon Pueblo. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. Brand, M. J Zooarchaeology of Sand Canyon Pueblo, Shorlene's Site, Roy's Ruin, Lillian's Site and Troy's Tower, Colorado. Honours Thesis, Department of Archaeology, Simon Fraser University, Burnaby, BC. Braun, D. P., and S. Plog 168

181 1982 Evolution of "Tribal" Social Networks: Theory and Prehistoric North American Evidence. American Antiquity 47(3): Breitburg, E Prehistoric New World Turkey Domestication: Origins, Developments and Consequences. Ph.D. Dissertation, Department of Anthropology, Southern Illinois University, Carbondale IL. Bullard, W The Cerro Colorado Site and Pithouse Architecture in the Southwestern United States Prior to A.D Peabody Museum of Archaeology and Ethnology Papers vol44(2). Peabody Museum, Harvard University, Cambridge, MA. Bullock, M. E Morphological and Functional Analysis of a Pueblo III Anasazi bone Tool Assemblage from Southwestern Colorado. M.A. Thesis, Department of Anthropology, Washington State University. Capone, P. H., and M. J. Schoeninger 1991 Food or Feathers: Amino Acid Racernization Analysis of Turkey Bone. Paper presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans, LA. Cattanach, G. S. J Long House Mesa Verde National Park, Colorado. National Park Service Publications in Archaeology No. 7H. United States Department of the Interior, National Parks Service, Washington, DC. Chao, V. S Stylistic and Functional Variability Within Sand Canyon Pueblo. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO.,.- d Cordell, L. S Prehistory of the Southwest. Academic Press Inc., San Diego, CA. Cordell, L. S., D. E. Doyel, and K. W. Kintigh 1994 The Process of Aggregation in the Southwest. In The Origin and Evolution of Prehistoric Southwest Society, edited by G. Gumerman and M. Gell-Mann. School for American Research Press, Santa Fe, NM. Dean, J. S., R. C. Euler, G. J. Gumerman, F. Plog, R. H. Hevly, and T. N. V. Karlstrom 1985 Human Behaviour, Demography, and Paleoenvironment on the Colorado Plateaus. American Antiquity Decker, K. W., and L. L. Tieszen

182 '' 1989 Isotopic Reconstruction of Mesa Verde Diet from Basketmaker I11 to i/ Pueblo 111. Kiva 55(1): Driver, J. C Medullary Bone as an Indicator of Sex in Bird Remains from Archaeological Sites. In Ageing and Sexing Animal Bones from Archaeological Sites, edited by Bob Wilson, C. Grigson and S. Payne, pp , BAR British Series 109. Oxford, England Crow Canyon Archaeological Center Manual for Description of Vertebrate Remains. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO Identification, Classification and Zooarchaeology. Circaea 9(1): Social Complexity and Hunting Systems in Southwestern Colorado. Paper presented at the Chacmool Conference "Debating Complexity", Calgary, AB. Driver, J. C., M. J. Brand, L. Lester, and N. MUNO 1995 Faunal Studies. In The Site Testing Program, edited by M. D. Varien, Crow Canyon Archaeological Center Occasional Paper, in press. Ducos, P "Domestication" Defined and Methodological Approaches to its Recognition in Faunal Assemblages. In Approaches to Faunal Analysis in the Middle East, edited by R.H. Meadow and M. A. Zeder, pp Peabody Museum of Archaeology and Ethnology, Cambridge, MA Defining Domestication: A Clarification. In The Walking hrder, edited by J. Clutton-Brock, pp , Unwin Hyman, London, England. Earle, T. K A Model of Subsistence Change. In Modeling Prehistoric Subsistence Economies, edited by T. K. Earle and A. L. Christenson, pp Studies in Archaeology. Academic Press, New York, NY. El-Najaar, M The Etiology of Porotic Hyperostosis among the Prehistoric and Historic Anasazi Indians of Southwestern United States. American Journal of Physical Anthropology Emslie, S. D Interpretation of Faunal Remains from Archaeological Sites in Mancos Canyon, Southwestern Colorado. M.A. Thesis, Department of Anthropology, University of Colorado.

183 1987 Prehistory of the Americas. Cambridge University Press, Cambridge, England. Ford, R. I Gardening and Farming before A.D. 1000: Patterns of Prehistoric \- Cultivation North of Mexico. Journal of Ethnobiology 1(1):6-27. Gilbert, B. M., L. D. Martin, and H. G. Savage 1985 Avian Osteology. Modern Printing Co., Lararnie, WY. Gilman, P. A Architecture as Artifact: Pit Structures and Pueblos in the American Southwest. American Antiquity Gleichman, P. J., and C. L. Gleichman 1992 The Lower Canyon Survey. In The Sand Canyon Archaeological Project A Progress Report, edited by W. D. Lipe, pp Crow Canyon Archaeological Center Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. Gnabasik, V. R Faunal Utilization by The Pueblo Indians. M.A. Thesis, Department of Anthropology, Eastern New Mexico University. Gould, R. R The Mustoe Site: The Application of Neutron Activation Analysis in the Interpretation of a Multi-component Archaeological Site. Ph.D. Thesis, Department of Anthropology, University of Texas at Austin. Grayson, D. K On the Quantification of Vertebrate Archaeofaunas. In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp Academic Press, San Diego, CA Quantitative Zooarchaeology. Academic Press, New York, NY. Greenfield, H. J The Origins of Milk and Wool Production in the Old World: A Zooarchaeological Perspective from the Centrid Balkans. Current Anthropology Fauna From the Late Neolithic of the Central Balkans: Issues in Subsistence and Land Use. Journal of Field Archaeology 18: Haas, J., and W. Creamer 1990 The Irritating Role of Warfare in the Pueblo I11 Period. Paper presented at the Conference for Pueblo Cultures in Transition: A.D in the American Southwest, Crow Canyon Archaeological Center, Cortez, CO.

184 Hallasi, J. A Archaeological Excavations at the Escalante Site, Dolores, Colorado, 1975 and Cultural Resource Series No. 7(2) Bureau of Land Management, Colorado State Office, Colorado. Hargrave, L. L Turkey Bones from Wetherill Mesa. American Antiquity 3 1 : Feathers From Sand Dune Cave: A Basketmaker Cave Near Navajo Mountain, Utah. Technical Series. Museum of Northern Arizona, Flagstaff, Arizona. Harvey, E. B., H. E. Kaiser, and L. E. Rosenburg 1968 An Atlas of the Domestic Turkey. U.S. Atomic Energy Commission, Division of Biology and Medicine, Washington, DC. Hawley Ellis, F An Interpretation of Prehistoric Death Customs in Terms of Modem Southwestern Parallels. In Collected Papers in Honor of Lyndon Lane Hargrave, edited by A. H. Schroeder, pp Museum of New Mexico Press, Santa Fe, NM. Hayden, B Subsistence and Ecological Adaptations of Modem HunterIGatherers. In Omnivorous Primates: Gathering and Hunting in Human Evolution, edited by G. Teleki and R. Harding, pp Columbia University Press, New York, NY Nimrods, Piscators, Pluckers, and Planters: The Emergence of Food Production. Journal of Anthropological Archaeology 9: Models of Domestication. In Transitions to Agriculture in Prehistory, edited by A. B. Gebauer and T. D. Price, pp Prehistory Press, Madison, WI. Hayes, A. C., and J. A. Lancaster 1975 The Badger House Community: Mesa Verde National Park. National Park Service Publications in Archaeology No. 7E. United States Department of the Interior, National Parks Service, Washington, DC. Hegmon, M Six Easy Steps to Dating Pueblo III Ceramic Assemblages: Working Draft. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO. Hesse, B Slaughter Patterns and Domestication: The Beginnings of Pastoralism in Western Iran. Man 17: Hesse, B., and P. Wapnish

185 1985 Animal Bone Archaeology, From Objectives to Analysis. Taraxacum, Washington, DC. Higgs, E. S., and M. R. Jman 1972 The Origins of Animal and Plant Husbandry. In Papers in Economic Prehistory, edited by E. S. Higgs, pp Cambridge University Press, Cambridge, England. Hill, J. N., and W. N. Trierweiler 1986 Prehistoric Response to Food Stress on the Pajarito Plateau, New Mexico: Technical Report and Results of the Pajarito Archaeological Research Project, Final Report to the National Science Foundation for Grant #BNS Huber, E. K., and W. D. Lipe 1992 Excavations at the Green Lizard Site. In The Sand Canyon Archaeological Project A Progress Report, edited by W. D. Lipe, pp Crow Canyon Archaeological Center Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. Hunter- Anderson, R Explaining Residential Aggregation in the Northern Rio Grande: A Competition Reduction Model. In Adaptive Change in the Northern Rio Grande Valley, edited by J. V. Biella and R. C. Chapman, pp Archaeological Investigations in Cochiti Resevoir, New Mexico Volume 4. Office of Contract Archeology, University of New Mexico, Albuquerque, NM. Jarman, M. R., and P. F. Wilkinson 1972 Criteria of Animal Domestication. In Papers in Economic Prehistory, edited by E. S. Higgs, pp , Cambridge University Press, Cambridge, England. \ Johnson, G. A Dynamics of Southwestern Prehistory: Far Outside-Looking in. In Dynamics of Southwestern Prehistory, edited by L. S. Cordell and G. J. Gumerman, pp Srnithsonian Institute, Washington, D.C. Kane, A. E Prehistory of the Dolores River Valley, in Dolores Archaeological Program: Final Synthetic Report, edited by D.A. Breternitz, C.K. Robinson and G.T. Gross, pp United States Department of the Interior, Bureau of Reclamation Engineering and Recreation Center, Denver, CO. Kent, S The Influence of Sedentism and Aggregation on Porotic Hyperostosis and Anaemia: A Case Study. Man 21.:

186 1989 Cross-cultural Perceptions of Farmers as Hunters and the Value of Meat. In Farmers as Hunters, edited by S. Kent, pp. 1-17, Cambridge University Press, Cambridge, England Excavations at a Small Mesa Verde Pueblo I1 Anasazi Site in Southwestern Colorado. Kiva 57(1): Kintigh, K Settlement, Subsistence and Society in Late Zuni Prehistory. Anthropological Papers No. 44. University of Arizona Press, Tucson, AZ. Kohler, T. A Introduction. In Bandelier Archaeological Excavation Project: Research Design and Summer I988 Sampling, edited by T. A. Kohler, pp Reports of Investigations 6 1. Washington State University, Pullman, WA Field Houses, Villages, and the Tragedy of the Commons in the Early Northern Anasazi Southwest. American Antiquity 57: Kohler, T. A., and M. H. Matthews 1988 Long-Term Anasazi Land Use and Forest Reduction: A Case from Southwestern Colorado. American Antiquity Kuckelman, K. A., J. Kleidon, M. D. Varien, and R. R. Lightfoot Sand Canyon Project Site Testing Program: Preliminary Report on the Excavations at Saddlehorn (5MT262), Mad Dog Tower (5MT181), Castle Rock Pueblo (5MT1825), Lester's Site (5MT10246), Lookout House (5MT10459) and Cougar Cub Alcove (5MT1690). Ms. on file, Crow Canyon Archaeological Center, Cortez, CO. Kuckelrnan, K. A., and J. N. Monis 1988 Archaeological Investigations on South Canal, vol. 2. Bureau of Land Management, Salt Lake City, UT. Lang, R. W., and A. H. Harris 1984 The Faunal Remains from Arroyo Hondo Pueblo. School of American Research Press, Santa Fe, NM. Lange, C. H Notes on the Use of Turkeys by Pueblo Indians. El Palacio 57: LeBlanc, S. A Settlement Patterns in the El Morro Valley, New Mexico. In Investigations of the Southwestern Anthropological Research Group, edited by R. C. Euler and G. J. Gumerman, pp , Museum of Northern Arizona, Flagstaff, AZ.

187 Lekson, S. H Settlement Pattern and the Chaco Region. In Chaco and Hohokam Prehistoric Systems in the American Southwest, edited by P. L. Crown and J. W. Judge, pp School of American Research Press, Santa Fe, NM Scale and Process in the Southwest. Paper presented at the Third Southwest Symposium, Tucson, AZ. Lightfoot, K. G Prehistoric Political Dynamics: A Case Study from the American Southwest. DeKalb, IL, Northern Illinois University Press. Lightfoot, K. G., and S. Upham 1989 Complex Societies in the Prehistoric American Southwest: A Consideration of the Controversy. In The Sociopolitical Structure of Southwestern Prehistoric Societies, edited by S. Upham, K. G. Kent and R. A. Jewett, pp Westview Press, Boulder, CO. Lightfoot, R. R Sand Canyon Locality Context: Research Design and Preliminary Results. Ms. on file, Crow Canyon Archaeological Center, Cortez CO. Linares, 0. F Garden Hunting in the American Tropics. Human Ecology 4: Lipe, W., D. 1992a Introduction. In The Sand Canyon Archaeological Project A Progress Report, edited by W. D. Lipe, pp Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. 1992b Summary and Concluding Remarks. In The Sand Canyon Archaeological Project A Progress Report, edited by W. D. Lipe, pp Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. Lyman, R. L Faunal Analysis: An Outline of Method and Theory with Some Suggestions. Northwest Anthropological Research Notes 13: Martin, D. L., A. H. Goodman, and G. J. Armelagos 1985 Skeletal Pathologies as Indicators of Diet. In The Analysis of Prehistoric Diets, edited by R. I. Gilbert and J. H. Mielke, pp Academic Press, Inc., Orlando, FL. Matson, R. G The Origins of Southwestern Agriculture. University of Arizona Press, Tucson, AZ.

188 Matson, R. G., and B. Chisolrn,1991 Basketmaker I1 Subsistence: Carbon Isotopes and Other Dietary Indicators from Cedar Mesa, Utah. American Antiquity 56: McKusick, C. R The Casas Grandes Avian Report. In Casas Grandes, A Fallen Trading Center of the Gran Chichimeca, edited by C. C. DiPeso, J. B. Rinaldo and G. J. Fenner, pp , Amerind Foundation, Inc., Dragoon. 1980a Three Groups of Turkeys from Southwestern Archaeological Sites. Contributions to Science 330: b Avian Remains. In Long House Mesa Verde National Park, Colorado, edited by G. J. Cattanach, pp Publications in Archaeology 7H. National Park Service, United States Department of the Interior, Washington, DC The Faunal Remains of La Humanas. In Contributions to Gran Quivira Archeology, edited by A. C. Hayes, pp National Park Service, United States Department of the Interior, Washington, D.C Avifauna from Grasshopper Pueblo. In Multidisciplinary Research at Grasshopper Pueblo, Arizona, edited by W. A. Longacre, S. J. Holbrook and M. W. Graves, pp University of Arizona Press, Tucson, AZ Southwest Indian Turkeys, Prehistory and Comparative Osteology. Southwest Bird Laboratory, Globe, AZ. Meadow, R. H Osteological Evidence for the Process of Animal Domestication. In The Walking Larder, edited by J. Clutton-Brock, pp , Unwin Hyman, London, England. Medlock, R. C Faunal Analysis. In The Cache River Archaeological Project: An Experiment in Contract Archaeology, edited by M. B. Schiffer and J. H. House, pp Arkansas Research Series No. 8. Arkansas Archaeological Survey, Fayetteville, AK. Micozzi, M. A Postmortem Change in Human and Animal Remains: A Systematic Approach. Charles Thomas, Springfield, IL. Minnis, P. E., 1985 Social Adaptations to Food Stress: A Prehistoric Southwestern Example. Prehistoric Archaeology and Ecology Series. University of Chicago, Chicago, IL. &

189 % 1989 Prehistoric Diet in the Northern Southwest: Macroplant Remains from Four Comers Feces. American Antiquity Moms, E. H Archaeological Studies in the La Plata District, Southwestern Colorado and Northwestern New Mexico. Carnegie Institution of Washington, Publication No. 19, Washington, DC. ' 1954 Basketmaker III Sites Near Durango, Colorado. Carnegie Institute of Washington, Publication 604, Washington, DC. Moms, J. N Archaeological Excavations on the Hovenweep Laterals. Four Corners Archaeological Project Report No. 16. Complete Archaeological Service Associates, Cortez, CO. Nash, D. T., and M. D. Petraglia 1987 Natural Formation Processes in the Archaeological Record: Present Problems and Future Requisites. In Natural Formation Processes and the Archaeological Record, edited by D. T. Nash and M. D. Petraglia, pp BAR International Series 352, Oxford, England. Neusius, S. W Faunal Analysis from the 1984 Excavation at 5MT765: A Preliminary Report. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO The Dolores Archaeological Program Faunal Data Base: Resource Availability and Resource Mix. In Dolores Archaeological Program: Final Synthetic Report, edited by D. A. Breternitz, C. K. Robinson and G. T. Gross, pp , United States Department of the Interior, Denver, Colorado. Nickens, P. R Pueblo III Communities in Transition: Environment and Adaptation in Johnson Canyon. The Colorado Archaeological Society in cooperation with the University of Colorado at Boulder, Boulder, CO. Olsen, S. J The Osteology of the Wild Turkey, Appendix. In Peabody Museum Papers v. 56 #2. Peabody Museum Papers. Peabody Museum of Archaeology and Ethnology, Boston, MA Archaeologically What Constitutes an Early Domestic Animal? In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp Academic Press, San Diego, CA. Orcutt, J. P., E. Blinman, and T. A. Kohler.

190 1990 Explanation of Population Aggregation in the Mesa Verde Region Prior to A.D In Perspectives on Southwestern Prehistory, edited by P. E. Minnis and C. L. Redman, pp Westview Press, Boulder, CO. Palkovich, A. M Arroyo Hondo Skeletal and Mortuary Remains. Arroyo Hondo Archaeological series. School of American Research Press, Santa Fe, NM. Parsons, E. C Taos Pueblo. General Series in Anthropology No. 2., Johnson Reprint Corporation, New York, NY. Pinkley, J. M The Pueblos and the Turkey: Who Domesticated Whom? American Antiquity 3 1 : Plog, F., G. J. Gurnerrnan, R. C. Euler, J. S. Dean, R. H. Hevly, and T. N. V. Karlstrom 1988 Anasazi Adaptive Strategies: The Model, Predictions and Results. In The Anasazi in a Changing Environment, edited by G. Gumerman, pp Cambridge University Press, Cambridge, England. Plog, S Spatial Organization and Exchange: Archaeological Survey on Northern Black Mesa. Southern Illinois University Press, Carbondale, IL. Prudden, T. M The Prehistoric Ruins of the San Juan Watershed in Utah, Arizona, Colorado and New Mexico. American Anthropologist 5(2): Rea, A. M Late Pleistocene and Holocene Turkeys in the Southwest. In Papers in Avian Paleontology Honoring Hildegarde Howard, edited by K. E. Campbell Jr., pp Natural History Museum of Los Angeles County, Los Angeles, CA. Reed, A. D The Dominquez Ruin: A McElmo Phase Pueblo in Southwestern Colorado. M.A. Thesis, Department of Anthropology, University of Colorado. Reed, E. K Turkeys in Southwestern Archaeology. El Palacio 58: Rick, A. M Bird Medullary Bone: A Seasonal Dating Technique for Faunal Analysts. Bulletin of the Canadian Archaeological Association 7:

191 Rindos, D The Origins of Agriculture. The Academic Press, London, England. Rohn, A. H Mug House, Mesa Verde National Park, Colorado. National Park Service Archaeological Research Series 7D. United States Department of the Interior, National Parks Service, Washington, D.C Cultural Change and Continuity on Chapin Mesa. The Regents Press of Kansas, Lawrence, KS Northern San Juan Prehistory. In Dynamics of Southwestern Prehistory, edited by L. S. Cordell and G. J. Gumerman, pp School of American Research Advanced Seminar Series. Smithsonian Institution Press, Washington, DC. Schlanger, S. H Population Change in the Dolores Area. Paper presented at the 50th Annual Meeting, Society for American Archaeology, Denver, CO Patterns of Population Movement and Long-term Population Growth in Southwestern Colorado. American Antiquity 53: Schorger, A. W An Ancient Pueblo Turkey. Auk 78: The Wild Turkey. University of Oklahoma Press, Norman, Oklahoma A New Subspecies of Meleagris gallopavo. Auk 87: Schroeder, A. H Birds and Feathers in Documents Relation to Indians of the Southwest. In Collected Papers in Honor of Lyndon Lane Hargrave, edited by A. H. Schroeder, pp Museum of New Mexico Press, Santa Fe, NM. Schwab, A. L., B. A. Bradley, and R. L. Lightfoot 1989 Laboratory Manual Crow Canyon Archaeological Center. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO. Senior, L. M., and L. J. Pierce.-, Turkeys and Domestication in the Southwest: Implications from & '/, Homol'ovi 111. Kiva 54: Shaffer, B.S Interpretation of Gopher Remains from Southwestern Archaeological 6- Assemblages. American Antiquity 57: Sherratt, A. G. 179

192 1983 The Secondary Exploitation of Animals in the Old World. World Archaeology l5(1):9o Speth, J. D., and S. L. Scott 1985 The Role of Large Mammals in Late Prehistoric Horticultural Adaptations: The View from Southeastern New Mexico. In Contributions to Plains Prehistory, edited by D. Burley, pp Occasional Paper #26. Alberta Culture Historical Resources Division, Alberta Horticulture and Large-mammal Hunting: The Role of Resource Depletion and the Constraints of Time and Labor. In Farmers as Hunters, edited by S. Kent, pp Cambridge University Press, Cambridge, England. Speth, J.D. and K.A. Spielmann 1983 Energy Source, Protein Metabolism, and Hunter-Gatherer Subsistence Strategies. In Journal of Anthropological Archaeology 2: Steadman, D. W A Review of the Osteology and Paleontology of Turkeys (Aves: Meleagridinae). In Papers in Avian Paleontology Honoring Hildegarde Howard, edited by K. E. Campbell Jr., pp Natural History Museum of Los Angeles County, Los Angeles, CA. Stiger, M. A Mesa Verde Subsistence Patterns from Basketmaker to Pueblo 111. The Kiva 44: Swannack, J. D Big Juniper House, Mesa Verde. National Park Service Archaeological Research Series 7C. United States Department of the Interior, National Parks Service, Washington, D.C. Upham, S Politics and Power: An Economic and Political History of the Western Pueblo. Academic Press, New York. Van der Merwe, N.J Carbon Isotopes, Photosynthesis, and Archaeology. American Scientist 70: Van West, C Modeling Prehistoric Climatic Variability and Agricultural Production in Southwestern Colorado: A G.I.S. Approach. Ph.D. dissertation Thesis, Department of Anthropology, Washington State University, Pullman, WA. Van West, C. R., and W. D. Lipe

193 1992 Modeling Prehistoric Climate and Agriculture in Southwestern Colorado. In The Sand Canyon Archaeological Project A Progress Report, edited by W. D. Lipe, pp Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. Varien, M. D., K. A. Kuckelrnan, and J. H. Kleidon 1992 The Site Testing Program. In The Sand Canyon Archaeological Program A Progress Report, edited by W. D. Lipe, pp Occasional Paper No. 2. Crow Canyon Archaeological Center, Cortez, CO. Varien, M. D., and R. L. Lightfoot 1989 Ritual and Nonritual Activities in Mesa Verde Region Pit Structures. In The Architecture of Social Integration in Prehistoric Pueblos, edited by W. D. Lipe and M. Hegmon, pp Occasional Paper No. 1. Crow Canyon Archaeological Center, Cortez, CO. Varien, M. D., W. D. Lipe, B. A. Bradley, M. A. Adler, and I. Thompson 1994 Southwest Colorado and Southeast Utah Mesa Verde Region Settlement A.D to In Pueblo Cultures in Transition A.D , edited by M. A. Adler. In press. Vierra, B. J Archaic Hunter-Gatherer Archaeology in Northwestern New Mexico. In Perspectives on Southwestern Prehistory, edited by P. E. Minnis and C. L. Redman, pp Westview Press, United States. von den Driesch, A A Guide to the Measurement of Animal Bones from Archaeological Sites. Peabody Museum Bulletins. Peabody Museum of Archaeology and Ethnology, Boston, MA. Walker, D. N Duckfoot Site (5MT765) Zooarchaeology. Ms. on file at Crow Canyon Archaeological Center, Cortez, CO. 1990a Preliminary Report of Faunal Remains from Sand Canyon Pueblo (5MT765), Colorado. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO. 1990b Zooarchaeology of the Green Lizard Site (5MT3901), Colorado. Ms. on file, Crow Canyon Archaeological Center, Cortez, CO. Wilcox, D. R., and J. Haas 1991 The Reality of Competition and Conflict in the Prehistoric Southwest. Unpublished Ms. on file, Crow Canyon Archaeological Center, Cortez, CO. Wis, W. H Early Agriculture and Sedentism in the Prehistoric Southwest. In Journal of World Prehistory 2:

194 Wills, W. H., and T. C. Windes 1989 Evidence for Population Aggregation and Dispersal during the Basketmaker I11 Period in Chaco Canyon, New Mexico. American " Antiquity 54: c, Wilshusen, R. H Unstuffing the Estufa: Ritual Floor Features in Anasazi Pit Structures and Pueblo Kivas. In The Architecture of Social Integration in Prehistoric Pueblos, edited by W. D. Lipe and M. Hegmon, pp Occasional Paper No. 1. Crow Canyon Archaeological Center, Cortez, CO. Windes, T. C The Use of Turkeys at Pueblo Alto Based on the Eggshell and Faunal Remains. In Investigations at the Pueblo Alto Complex, Chaco Canyon, " ' I edited by F. J. Mathien and T. C. Windes, pp National Parks Service, Santa Fe, NM. Winship, G The Coronado Expedition. 14th Annual Report, Bureau of American Ethnology, Washington, DC.

195 Appendix A: Elemental Distribution of Turkey by Site

196 I lsite Number I I - Appendix A: Frequency of turkey elements in Sand Canyon Locality sites.