The Pennsylvania State University

The Graduate School

College of the Liberal Arts

SMALL ANCESTRAL PUEBLO SITES IN THE MESA

VERDE REGION:

LOCATION, LOCATION, LOCATION

A Dissertation in

Anthropology

by

A’ndrea Elyse Messer

©2009 A’ndrea Elyse Messer

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2009 This dissertation of A’ndrea Elyse Messer was reviewed and approved* by the following:

George Milner Professor of Anthropology Dissertation Advisor Chair of Committee

David Webster Professor of Anthropology

Dean Snow Professor of Anthropology

Christopher Duffy Professor of Civil Engineering

Nina Jablonski Professor and Head, Department of Anthropology

*Signatures are on file in the Graduate School

ii ABSTRACT

A basic theoretical issue in settlement archaeology is the effect that environment on the one hand, or large centers on the other, have on the placement of small habitations. A key methodological issue is whether old survey data, collected at a time when today’s questions had not yet been formed, can still prove useful. This dissertation investigates the viability of old survey data, some of the environmental influences – landform, elevation, temperature, precipitation – and settlement population density, on site location choice in the Mesa Verde Southwest. I also investigated the effects of large sites on small site location and looked at all these factors with respect to reinhabited sites versus pristine sites. My results suggest a method to determine which old surveys can be used, and which cannot. The site populations in the mesa surveys are similar but differ from the site populations of the non-mesa surveys, indicating a possible difference in settlement pattern between mesas and other areas. Site population in general increased through time. Wetherill Mesa is the only location where Late Pueblo III site population dipped dramatically, probably due to movement into aggregated cliff locations. Mesa-top sites remained the majority through time, but population pressure caused people to move sites to the talus slopes. No large site influence pushes sites up or down in elevation, but population pressure does expand the range of elevations inhabited. Distances to large sites from small sites do not change the appearance of large sites. Small sites move away from or toward large sites without pattern. The environment appears to influence small site locations far more than proximity to large sites. While large sites might have influenced the myriad small-site inhabitants by altering their social and political life, they did not seem to alter the locations people chose for their homes. Population pressure seems to be the largest influence on the settlement patterns in the Mesa Verde area.

iii TABLE OF CONTENTS

LIST OF TABLES ...... vi

LIST OF FIGURES ...... vii

ACKNOWLEDGEMENTS ...... xiii

Chapter 1. INTRODUCTION ...... 1 Variables ...... 8 Rehabitation of Sites ...... 10 Old Data ...... 13 Study Area ...... 15 Ancestral Puebloan World – Mesa Verde ...... 18 Mesa Verde Archaeology ...... 22 Organization of Chapters ...... 28

Chapter 2. THE SURVEYS ...... 31 Approach to Survey Data ...... 32 Chapin Mesa Survey ...... 34 Wetherill Mesa Survey ...... 37 Mockingbird Mesa Survey ...... 41 Sand Canyon/Goodman Point Survey ...... 42 Mesa Verde Area Study 2000 Survey ...... 44 Database Structure ...... 45 Chronology and Its Problems ...... 50 Normalization of Cultural Phase Lengths ...... 56

Chapter 3. SITE POPULATIONS ...... 58 Broad Time Scales – BM, PI, PII, PIII...... 62 Narrow Time Scales – BM, PI, EPII, MPII, LPII, EPIII, LPIII ...... 68 The Chapin Mesa Problem ...... 71

Chapter 4. LANDFORMS ...... 73 Broad Time Scales ...... 76 Narrow Time Scales ...... 80 Comparison of Landforms ...... 84 The Chapin Mesa Problem ...... 87

Chapter 5. ELEVATION AND CLIMATE ...... 91 Broad Time Scales ...... 95 Narrow Time Scales ...... 99 Temperature and Elevation ...... 105

iv Temperature Analysis ...... 108 Broad Time Scales ...... 108 Narrow Time Scales...... 114 Precipitation Analysis ...... 119 Broad Time Scales...... 119 Narrow Time Scales...... 124

Chapter 6. DISTANCE TO LARGE SITES ...... 129 Large Sites ...... 130 Wetherill Mesa Survey ...... 137 Sand Canyon/Goodman Point Survey ...... 145 Mockingbird Mesa Survey ...... 154

Chapter 7. REINHABITED SITES ...... 163 Population ...... 163 Landforms ...... 169 Elevations ...... 175 Distance to Large Sites ...... 181

Chapter 8. SUMMARY AND CONCLUSIONS ...... 186 Ancestral Puebloan World – Mesa Verde ...... 187 Site Populations ...... 188 Landforms ...... 189 Elevation ...... 190 Distance to Large Sites ...... 191 Reinhabited Sites ...... 192 Old Surveys ...... 194 Future Work ...... 196 Conclusions ...... 198

BIBLIOGRAPHY ...... 200

Appendix A. MODIFIED DATABASE OF MOCKINGBIRD MESA SURVEY, BUREAU OF LAND MANAGEMENT – HABITATION SITES ONLY ...... 214

Appendix B. MODIFIED DATABASE OF SAND CANYON STUDY AREA SURVEYED SITES – HABITATION SITES ONLY ...... 219

Appendix C. MODIFIED DATABASE OF PENN STATE MESA VERDE AREA SURVEY – HABITATION SITES ONLY ...... 227

v LIST OF TABLES

Table 1 Phases and dates in the Mesa Verde Southwestern U.S...... 7 Table 2 General survey information ...... 32 Table 3 Site counts of surveys used ...... 38 Table 4 Distribution of sites by site type ...... 41 Table 5 Variables Used in Survey Databases ...... 47 Table 6 Pecos classifications ...... 51 Table 7 Dates and phases used in the Wetherill Mesa Survey ...... 53 Table 8 Phases and Dates Currently in Use in the Mesa Verde Southwestern U.S...... 54 Table 9 Comparison of dates across surveys ...... 55 Table 10 Phase lengths and multipliers for normalization of data ...... 56 Table 11 Landform categories by survey...... 73 Table 12 Student T test of Pueblo I elevations vs. Pueblo II elevations ...... 103 Table 13 Student T test of Pueblo II elevations vs. Pueblo III elevations ...... 103 Table 14 Student T Test of Early Pueblo III Elevations vs. Late Pueblo III Elevations . 104 Table 15 Sites with 50 or more rooms in survey areas...... 131 Table 16 Numbers of pristine and reinhabited sites in Pueblo III ...... 163 Table 17 Distribution of Pueblo III sites between pristine and reinhabited sites, percentages...... 164

vi LIST OF FIGURES

Figure 1 Map of the Four Corners area. Enclosed area approximately encloses the range of surveys discussed (MapSource 1999) ...... 14 Figure 2 Major vegetation found in the study area by elevation (Lipe 1999)...... 16 Figure 3 Map of the five surveys used in this research. A is Mockingbird Mesa, B is Sand Canyon/Goodman Point survey, C is the Mesa Verde Area 2000 survey, D is the Wetherill Mesa survey and E is the Chapin Mesa survey...... 17 Figure 4 Escalante Ruin in Dolores, Colorado outside Anasazi Heritage Center (BLM). This is a small, Chacoan style Late Pueblo II site...... 19 Figure 5 Dominguez Ruin in Dolores, Colorado outside Anasazi Heritage Center (BLM). This is a small Pueblo II site...... 19 Figure 6 Graph of small site populations for five surveys in four time periods. Cultural phases are normalized so that periods are equal and data are comparable...... 61 Figure 7 Graph of small site populations for four surveys in four time periods. Chapin Mesa survey is excluded. Cultural phases are normalized so that periods are equal and data are comparable...... 63 Figure 8 Graph of small site populations for five surveys and combined total database for four time periods. Cultural phases are normalized so that periods are equal and data are comparable...... 64 Figure 9 Graph of small site populations for four surveys and combined total database for four time periods. Chapin Mesa Survey is excluded. Cultural phases are normalized so that periods are equal and data are comparable...... 65 Figure 10 Graph of small site populations for five surveys in seven time periods. Cultural phases are normalized so that periods are equal and data are comparable...... 66 Figure 11 Graph of small site populations for four surveys in seven time periods. Chapin Mesa is excluded. Cultural phases are normalized so that periods are equal and data are comparable...... 67 Figure 12 Graph of small site populations for four surveys and combined total database for seven time periods. Chapin Mesa is excluded...... 68 Figure 13 Graph of small site populations for five surveys and combined total database for seven time periods...... 69 Figure 14 Site distribution by landform for Sand Canyon survey. Four cultural phases. ..76 Figure 15 Site distribution by landform for Mockingbird Mesa survey. Four cultural phases ...... 77 Figure 16 Site distribution by landform for MVAS survey. Four cultural phases...... 77 Figure 17 Site distribution by landform for Wetherill Mesa survey. Four cultural phases . 78 Figure 18 Site distribution by landform for Chapin Mesa survey. Four cultural phases. ..78 Figure 19 Site distribution by landform for Sand Canyon survey. Seven cultural phases. . 80 Figure 20 Site distribution by landform for MVAS survey. Seven cultural phases...... 81 Figure 21. Site distribution by landform for Mockingbird Mesa survey. Seven cultural phases...... 81

vii Figure 22. Site distribution by landform for Wetherill Mesa survey. Seven cultural phases...... 82 Figure 23. Site distribution by landform for Chapin Mesa survey. Seven cultural phases. ..82 Figure 24 Comparison of mesa distribution of five surveys in seven cultural phases. Only the percentage of mesa sites is shown for each survey...... 84 Figure 25 Comparison of talus distributions of five surveys in seven cultural phases. Only the percentage of talus sites is shown for each survey...... 85 Figure 26 Comparison of cliff distributions of five surveys in seven cultural phases. Only the percentage of cliff sites is shown for each survey...... 86 Figure 27 Comparison of elevation ranges and means across the Wetherill Mesa, Mockingbird Mesa, Mesa Verde 2000 and Sand Canyon/Goodman Point surveys...... 92 Figure 28 Elevation distribution of sites through time in Mesa Verde 2000 survey ...... 94 Figure 29 Elevation distribution through time of Wetherill Mesa survey...... 95 Figure 30 Elevation distribution through time of Sand Canyon/Goodman Point survey. ..96 Figure 31 Elevation distribution through time of Mockingbird Mesa survey...... 97 Figure 32 Elevation distribution through time of Mesa Verde 2000 survey using narrow time cultural phases...... 99 Figure 33 Elevation distribution through time of Wetherill Mesa using narrow cultural time periods...... 100 Figure 34 Elevation distribution through time of Sand Canyon/Goodman Point using narrow cultural time periods...... 101 Figure 35 Elevation distribution through time of Mockingbird Mesa using narrow cultural time periods...... 102 Figure 36 Comparison of average temperatures in the Southern Colorado Plateau.. (A) annual temperatures, (B) five-year averages, (C) 30-year averages, (D) averages based on broad cultural periods, (E) averages based on narrow cultural periods. (Salzer 2005) ...... 106 Figure 37 Comparison of average precipitation in the Southern Colorado Plateau. (A) annual precipitation, (B) 5-year averages, (C) 30-year averages, (D) averages based on broad cultural periods, (E) averages based on narrow cultural periods (Salzer 2005)...... 107 Figure 38 Elevation, population and temperature comparison for Wetherill Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 110 Figure 39 Elevation, population and temperature comparison for Mockingbird Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 111

viii Figure 40 Elevation, population and temperature comparison for Mesa Verde Area survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 112 Figure 41 Elevation, population and temperature comparison for Sand Canyon/ Goodman Point survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 113 Figure 42 Elevation, population and temperature comparison for Wetherill Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 115 Figure 43 Elevation, population and temperature comparison for Mockingbird Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 116 Figure 44 Elevation, population and temperature comparison for Mesa Verde Area 2000 survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 117 Figure 45 Elevation, population and temperature comparison for Sand Canyon/Goodman Point survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature...... 118 Figure 46 Elevation, population and precipitation comparison for Wetherill Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 120 Figure 47 Elevation, population and precipitation comparison for Mockingbird Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 121 Figure 48 Elevation, population and precipitation comparison for MVAS survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 122 Figure 49 Elevation, population and precipitation comparison for Sand Canyon Goodman Point survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 124 Figure 50 Elevation, population and precipitation comparison for Wetherill Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 125

ix Figure 51 Elevation, population and precipitation comparison for Mockingbird Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 126 Figure 52 Elevation, population and precipitation comparison for Mesa Verde Area 2000 survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 127 Figure 53 Elevation, population and precipitation comparison for Sand Canyon/ Goodman Point survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation...... 128 Figure 54 Map of Wetherill Mesa with large sites marked, PIII sites red square, Badger House in red circle. (MapSource 1999) ...... 132 Figure 55 Map of Sand Canyon/Goodman Point large sites. Sites shown in red squares. (MapSource 1999) ...... 133 Figure 56 Map of Mockingbird Mesa large sites indicated with red squares. (MapSource 1999) ...... 135 Figure 57 Map of all habitation sites recorded in Wetherill Mesa Survey with sites with over 50 surface rooms identified. Sites in upper right corner are not actually on Wetherill Mesa but were included in the survey. Large sites are all on canyon rims; the sparser distribution is on the talus slope...... 136 Figure 58 Average distances through time of small sites to Double House and Mug House on Wetherill Mesa using broad cultural time periods...... 137 Figure 59 Average distances through time of small sites to Long House, Ruin 16, Kodak House and Spring House on Wetherill Mesa using broad cultural time periods...... 138 Figure 60 Average distances through time of small sites to Double House and Mug House on Wetherill Mesa using narrow cultural time periods...... 139 Figure 61 Average distances through time of small sites to Long House, Ruin 16, Kodak house and Spring House on Wetherill Mesa using broad cultural time periods...... 140 Figure 62 Map of all habitation sites recorded in Sand Canyon/Goodman Point Survey with sites with over 50 surface rooms identified...... 145 Figure 63 Average distances through time of small sites to Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra in Sand Canyon/Goodman Point locality using broad cultural time periods...... 146 Figure 64 Average distances through time of small sites to Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra in Sand Canyon/Goodman Point locality using narrow cultural time periods...... 147 Figure 65 Average distances through time of small sites to Sand Canyon Pueblo and Goodman Point Pueblo in Sand Canyon/Goodman Point local using narrow cultural time periods...... 148

x Figure 66 Comparison of average distances through time of small sites to Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra and to Sand Canyon and Goodman Point only in Sand Canyon/Goodman Point local using broad cultural time periods...... 150 Figure 67 Comparison of average distances through time of small sites to Sand Canyon, Goodman Point Pueblos and Casa Negra and to Sand Canyon and Goodman Point Pueblos only in the Sand Canyon/Goodman Point local using narrow cultural time periods...... 153 Figure 68 Map of all habitation sites recorded in Mockingbird Mesa Survey with sites with over 50 surface rooms identified...... 155 Figure 69 Average distances through time of small sites to Seven Tower Ruin, 5MT1541 and 5MT1512 in Mockingbird Mesa survey using broad cultural time periods...... 156 Figure 70 Average distances through time of small sites to Seven Tower Ruin, 5MT1541 and 5MT1512 in Mockingbird Mesa survey area using broad cultural time periods...... 157 Figure 71 Average distances through time of small sites to Seven Tower Ruin and 5MT1512 in Mockingbird Mesa survey area using broad cultural time periods...... 158 Figure 72 Average distances through time of small sites to Seven Tower Ruin and 5MT1512 in Mockingbird Mesa survey area using narrow cultural time periods...... 159 Figure 73 Comparison of average distances through time of small sites to Seven Tower Ruin and 5MT1512 when 5MT1541 is included or excluded using broad cultural time periods...... 160 Figure 74 Comparison of average distances through time of small sites to Seven Tower Ruin and 5MT1512 when 5MT1541 is included or excluded using narrow cultural time periods...... 160 Figure 75 Percent of pristine and reinhabited sites during Pueblo III across four surveys. 165 Figure 76 Percent of pristine and reinhabited sites during Early Pueblo III across four surveys ...... 166 Figure 77 Percent of pristine and reinhabited sites during Late Pueblo III across four surveys...... 167 Figure 78 Comparison of PIII Landforms across four surveys for pristine and reinhabited sites...... 169 Figure 79 Comparison of PIII Landforms across four surveys for pristine and reinhabited sites...... 170 Figure 80 Comparison of LPIII Landforms across four surveys for pristine and reinhabited sites...... 171 Figure 81 Comparison of four surveys with respect to landform and reinhabited and pristine sites during Pueblo III ...... 172 Figure 82 Comparison of pristine and reinhabited sites on Wetherill Mesa for all Pueblo III, Early Pueblo III and Late Pueblo III elevations...... 175

xi Figure 83 Comparison of pristine and reinhabited sties on Mockingbird Mesa for all Pueblo III, Early Pueblo III and Late Pueblo III elevations...... 176 Figure 84 Comparison of pristine and reinhabited sties in Mesa Verde 2000 survey for all Pueblo III, Early Pueblo III and Late Pueblo III elevations...... 177 Figure 85 Comparison of pristine and reinhabited sties in Sand Canyon/Goodman Point survey area for all Pueblo III, Early Pueblo III and Late Pueblo III elevations...178 Figure 86 Comparison of distances to large sites of pristine and reinhabited sites on Wetherill Mesa – all sites, Kodak House, Double House, Long House...... 179 Figure 87 Comparison of distances to large sites of pristine and reinhabited sites on Wetherill Mesa – Mug House, Ruin 16, Spring House ...... 180 Figure 88 Comparison of average distances to large sites of pristine and reinhabited sites on Mockingbird Mesa, all sites...... 182 Figure 89 Comparison of distances to large sites for pristine and reinhabited sites in the Sand Canyon/Goodman Point Survey – Sand Canyon and Goodman Point sites only...... 183 Figure 90 Comparison of distances to large sites for pristine and reinhabited sites in the Sand Canyon/Goodman Point Survey – Sand Canyon, Goodman Point and Casa Negra.the pristine and reinhabited sites show no significant difference. ..184

xii ACKNOWLEDGEMENTS

I would like to thank the Pennsylvania Space Grant Consortium for their support through two Space Grant Fellowships, without which neither the survey nor the archival work at Mesa Verde National Park could have been done. I would also like to thank the Penn State Department of Anthropology for the Hill Foundation grant. The research personnel at Mesa Verde National Park allowed me access to the original field cards for the Wetherill and Chapin Mesa surveys and cheerfully put up with my intrusion. The archaeologists at Crow Canyon Archaeological Center hooked me on anthropological archaeology and have offered their support throughout the years. I am not sure I would have done this work or sought any degree in anthropology if it were not for the fourth pit structure at Duckfoot. Ricky Lightfoot and Mark Varien have been totally supportive since I ruined their Fall in 1987. Scott Ortman taught me to recognize the pottery types in the Mesa Verde Southwest and let my field crew disturb his laboratory. Kristin Kuckleman was always solidly rooted in reality and a friend and Donna Glowacki frequently assured me that I was not repeating other peoples work and encouraging me to finish. All the other people at Crow Canyon, who helped me along the way are greatly appreciated. Cortez, Co., is a small town and in only four months, faces became familiar and people became friends. Catherine Pratt allowed me to share her home, her ear and her cats for four months, Bruce and Cindy Bradley and Larry and Nancy Hammack allowed me to wander all over their land. Bruce and Cindy also put up with my insane ideas as the project evolved. Survey cannot be done by one person. Judy Cooper and Steve Barry were invaluable crew and I could not have completed the survey without them. Richard Alley always had a kind word and lent an ear on climate questions and Andrew Comrie found me the latest southwestern climate information. I am glad you both were so willing to share your expertise. The Penn State Department of Public Information has always lent their support. By allowing me to take a four-month, albeit unpaid, educational sabbatical, they made this project possible. Bill Mahon and Vicki Fong both encouraged me. Barbara Hale covered for me when I was away. I know it was not easy. I also want to thank Cathey Chaffee for copy editing and proofreading, something I could not have done myself. The Penn State Department of Anthropology faculty was willing to put up with a graduate student with a more than full time career in a totally different discipline. While I tried not to let work interfere with school, there were times when it inevitably did. The faculty was always willing to accommodate the vagaries of a job that sometimes deals with unscheduled, emergency situations. The encouragement from all the faculty members who have been on my committee was amazing. Alexander Joffe, Frances Hayashida and Anna Barros left before I could complete my work, but they were always encouraging and supportive. My committee, David Webster, Dean Snow and Chris Duffy also helped and encouraged me. I am certain that George Milner wishes I had completed this work a long time ago. He graciously put up with the time scale on this and guided me throughout this entire project. I know that I would not have finished this without his help.

xiii While I only met Al Hayes a few times in person, I feel that I know him through his work. His impressive ability to record the landscape and set the tone for future southwestern survey continues long after his passing. While I appreciate all the help that has been offered me, I am the only one responsible for this work. All errors, missteps and misconceptions are entirely mine.

xiv Chapter 1. INTRODUCTION

Travel through the Southwestern United States and the picture presented by some large sites – sandstone walls; curved towers; multistoried, windowed buildings nestled in cliff alcoves – is often breathtaking. Because of their size and location, cliff dwellings in the Mesa Verde area are often easily viewed and visited. Throughout the Four Corners area – Southwestern Colorado,

Southeastern Utah, Northwestern New Mexico and Northeastern Arizona – National Parks and

Monuments attest to the visibility and majesty of large, cliff alcove sites. Other large sites, such as those found at Chaco Canyon National Historic Park (Lister and Lister 1981, Sebastian 1992a,

Sebastian 1992b, Vivian 1990), or dotted throughout the Mesa Verde area are not “cliff dwellings,” but monumental enough to be noticed and, in some cases, excavated early in the history of Southwestern archaeology.

Whether large aggregated sites are located in cliff alcoves or at canyon heads on the mesa top, they are difficult to miss. It was not only those first archaeologists who focused on the large sites. Archaeologists to this day study large sites and their influence and importance to the social and political organization of an area. Large sites in the Mesa Verde area are considered focal points of social integration (Varien and Wilshusen 2002b) at least during the Pueblo III period.

These aggregated sites existed only during two periods in the Mesa Verde area, but inhabitants of small habitation sites had to decide where to locate their homes through the entire time.

A basic theoretical issue in settlement archaeology is the effect that environment on the one hand, or large centers on the other, have on the placement of small habitations. A key methodological issue is whether old survey data, collected at a time when today’s questions had not yet been formed, can still prove useful.

1 This dissertation investigates the viability of old survey data and some influences – site population density, landform, elevation, temperature, precipitation – on site location choice in the

Mesa Verde Southwest. I also investigated the effects of large sites on small site location and looked at all these factors with respect to reinhabited sites versus pristine sites. These objectives will shed some light on the environmental influences on site location and whether the large aggregated villages had influence over site placement.

This work attempts to answer the question: What influenced the decisions that ancestral

Puebloan small site dwellers made when they chose a place to live? By extension, this issue should provide insight on small site locations in any area where large aggregated habitation sites appear only sporadically, but small sites dot the landscape through time. An investigation of the landscape and its relationship to the small habitation sites can eliminate the natural environmental influences, leaving the effects of sociopolitical forces. Undoubtedly large sites, when they exist, do exert some social and political influence, but how much do they actually affect small site location? I suggest that the choices made for the location of small sites is, in large part, unaffected by the sudden appearance of large aggregated sites. This is not to say that the changes in social, economic and political structures in the area did not influence the inhabitants of the small sites, but that they organized themselves on the landscape in basically the same way regardless of whether large sites existed at the time or were absent. In effect, there would be two organizing structures in play: one originating in the large sites and perhaps influencing trade, ritual and the sociopolitical structure of the overall community, and one originating in the landscape, heavily influenced by the environment and by the propinquity of small sites.

Although much research on social organization has been done in the Mesa Verde area

2 (Adler 1990, Adler 1996b, Adler 2002, Adler and Varien 1991, Cordell 1994, Cordell, et al.

1994, Hegmon 2002, Kohler 1995, Mahoney, et al. 2000, Neitzel 1999, Peterson and Drennan

2005, Varien 1999b, Varien and Wilshusen 2002b) much of it still focuses on times when large sites existed and views settlement patterns as a byproduct of these sites’ presence. Over time, researchers have suggested the social structure of communities in the Southwest and the political organization ranged from disaggregated agrarian villages to state-level societies (McGuire and

Saitta 1996, Saitta 1997, Sebastian 1992a, Vivian 1990). Determination of whether political structure was highly organized and hierarchical, less organized and egalitarian, or somewhere in between, was often based on exactly the same data (Van West, et al. 1987, Vivian 1990). Adler

(1990) suggests that the foci of communities in the Sand Canyon locality were socially integrated locations such as great kivas or large, aggregated sites. These socially integrating places served as the nexus for changing land tenure approaches and community structure. Adler also suggests that some locations where aggregated sites occurred late in the occupation series were socially integrating sites before aggregated settlement occurred (Adler 1990). These late Pueblo III aggregated pueblos were built around large springs. Undoubtedly, people would have visited the springs and they would have been socially integrating locations before settlement. Many areas contiguous to springs were only built up during the late Pueblo III period (Kuckelman 1997,

Kuckelman 2003, Kuckelman and Coffey 2007), even if fairly large sites were nearby before then.

Small sites dot the landscape throughout the entire Mesa Verde area. Even areas sometimes thought to be vacant, such as the intermediary land between the talus slopes and the flat valleys, have been shown to be populated by small sites (Messer 2001b). Excavations of small farmsteads are often done with an eye to their relationship with the large sites (Van West, et al.

3 1987, Varien 1990, Varien 1991). Small sites are often looked at in the aggregate, subdivided through time, but rarely considered for their individual or combined impact on the community or social organization absent large aggregated sites (Petersen, et al. 1986, Powell 1983).

Saitta and McGuire (1996, Saitta 1997) suggest that communalism and a form of hierarchy can coexist because the hierarchy is subsumed in the community. Saitta illustrates his argument by suggesting that the control of labor need not be coercive to support a variety of hierarchical functions and that no “coercive exploitative elites” (Saitta 1997:14) need be postulated to account for this duality of structure. The population can still have autonomy and supply labor for community functions. McGuire and Saitta (1996:201) suggest that “while not egalitarian, (the pueblos) were not stratified either; in fact, they were simultaneously both.”

While I do not see large aggregate pueblo inhabitants as the evil overlord, nor small site denizens as happy, carefree proletariat, an intertwined but two-tiered system of organization does appear to be a reasonable solution to how Mesa Verde regional society functioned. The overall mosaic of small sites and its relationship to the environment is important and its inherent social structure – absent the large sites – has value. The changes in land use patterns must reflect changes in social structure, even when viewed separately from the overall regional pattern or nearby large sites (Adler 1990, Adler 1994, Adler 1996a, Adler 1996c, Varien 1997). If we have an understanding of the environmental forces inherent influence on the location of small sites, then it will be much easier to identify the influences of large sites sociopolitical, ritual or economic spheres.

The Mesa Verde area is an ideal location to look at the influences of environmental variables on the location of small habitation sites, the influence of large sites on location and the

4 possible use of old surveys for a number of reasons. Full coverage survey in the Southwest is exceptionally useful because many of the later sites are visible on the surface due to the large amounts of stone used in construction. Even in earlier periods, some stone was used in architecture and it is often visible. Mesa Verde area pottery is abundant and remains fairly undamaged by the elements. Using pottery, good site dating linking sites to established Ancestral

Pueblo phases is possible. While large portions of the Mesa Verde region have not been surveyed, large block surveys do exist for the area. The site cards for old surveys in Mesa Verde National

Park are archived at the park and recent GPS coordinates for much of those surveys exist. They are a valuable resource, but because they were done more than 50 years ago must first be evaluated for their usefulness. Of great importance for a study of this type is that the Mesa Verde area is very densely blanketed with archaeological sites.

Small sites – sites under 50 rooms – occur on mesa tops, talus slopes, in alcoves or cliffs and even in canyon bottoms. Small sites also exist in every Ancestral Pueblo time period. While some portion of the population might move to aggregated sites at a specific time, some fraction of the population remains in small sites continuously. How much of the settlement pattern of small sites is influenced by the existence of a large site in the area and how much of the small site distribution pattern is influenced by the landscape? How much the existence of large sites influenced sociopolitical organization can only be seen once we take into account the natural landscape and environment.

While land tenure issues, hierarchy, ritual practices and political considerations undoubtedly influence site location and settlement patterns, it seems more logical to look first at the environment for clues to site choice.

5 To answer some of these questions I look at a variety of full coverage site surveys carried out over the past 50 years in the Central Mesa Verde Area. These include two surveys done in the late 1950s and early 1960s in Mesa Verde National Park on Chapin and Wetherill Mesas

(Hayes 1964, Rohn 1977), the Sand Canyon/Goodman Point Survey (Adler 1988, Adler 1990,

Van West, et al. 1987), the Mockingbird Mesa Survey (Fetterman and Honeycutt 1987) and the

Mesa Verde Area Survey 2000 (Messer 2001b). Ancestral Pueblo sites throughout the region occur on mesa tops, talus slopes, in valley bottoms and cliff caves (Varien 1991). The average duration of occupation of small sites in the Mesa Verde region is short. According to Cordell

(1984), permanent sites had short periods of occupation even when heavy labor efforts had been expended for architectural construction. She suggests that some studies indicate that as long as there was sufficient land for cultivation, villagers would relocate in response to minor climatic changes or resource depletion. With movement of location perhaps every generation, the choice of location for a small site was a routine rather than occasional task.

The exact factors used to choose site locations are unknown. Ancestral Puebloans needed to consider a wide range of variables to allow them to farm and survive. Obviously, the environmental variables we can see today that influence small farms also influenced Ancestral

Puebloans. Those factors we can recognize and which they undoubtedly also took into consideration, can be investigated. I will investigate some of these environmental and social factors in this work. Hopefully, the factors considered represent a significant part of the environmental influences.

What we can never know are the intangible aspects of site location selection. Perhaps a location had a wonderful view. Perhaps the choice was ancestral; grandma lived here once.

6 Table 1 Phases and dates in the Mesa Verde Southwestern U.S. (Lipe 1999)

Cultural Stage Dates Basketmaker III (BMIII) 500 – 750 C.E. Pueblo I (PI) 750 – 900 C.E. Pueblo II (PII) 900– 1150 C.E. Early Pueblo II (EPII) 900 – 1050 C.E. Late Pueblo II (LPII) 1050 – 1150 C.E. Pueblo III (PIII) 1150 – 1300 C.E. Early Pueblo III (EPIII) 1150 – 1225 C.E. Late Pueblo III (LPIII) 1225 – 1300 C.E.

Perhaps there was some ritual, magic or sacred reason to choose a site. These intangible aspects that affect all human decisions cannot be considered. Only the information gathered from the landscape and environment and any information garnered from archaeological evidence can ever be included in any study. Although we will never know exactly how they decided where to live, we can make some headway in what variables influenced their choices.

I will also look at the data in two time frames, broad cultural categories of Basketmaker,

Pueblo I, Pueblo II and Pueblo III and narrower categories that break the Pueblo II and Pueblo III designations into smaller time segments (Table 1). In general, the broader time periods are more accurate because it is easier to place sites into these categories without error. The narrower categories are more precise in terms of their smaller time span, but less accurate because it is more likely error is made when allocating place to sites in narrower categories. Another difficulty

7 is determining if there is a break in habitation between the narrower categories, especially during

Pueblo III when looking at reinhabited sites. The narrower time scales also pose a problem because dividing the sites up into smaller groups sometimes creates datasets that are too small for statistical analysis.

Variables

Site population is the number of sites that exist during any time interval. Changes through time in number of sites provide a proxy for the number of people and allows an assessment of population pressure’s influence on site location choices. Several people have used site populations to estimate number of inhabitants (Adler 1990, Schlanger 1988, Wilshusen 2002) and site numbers are a good proxy for trends in population.

Landform is the type of land where the site is located. Whether the site location is on a mesa top or flat area, talus slope, cliff alcove or in a canyon bottom can say something about the choices made in site location. Are they choosing the best location to live, or are they, for some reason, choosing a less optimal site? How does the choice of landform shift through time and what does that shift mean?

Elevation is the height above sea level that a site is located. Elevation is an important factor for farmers. Agriculture becomes more difficult with cooler temperatures, dryer areas and shorter numbers of frost free days. The elevation of site locations change through time and must be considered with changes in regional temperature and precipitation.

Temperature can be investigated in a variety of ways. Mean annual temperature is the mean temperature in an area during an entire year. Temperature can also be averaged over a

8 number of years to show extended trends. These trends are often more instructive about the effects of temperature on agricultural productivity and an agricultural community’s ability to survive. Annual, 5-year and 30-year averaged temperatures are considered as are averages for both broad and narrow time periods. Temperature datasets for the Colorado Plateau are reconstructed from data made available by the U.S. National Oceanic and Atmospheric

Administration (http://www.ncdc.noaa.gov/paleo/pubs/salzer2005/salzer2005.html).

Precipitation includes rain, snow and sleet. An important requirement for all agriculture, differing amounts of precipitation can spell the difference between a great or a failed crop.

Precipitation is presented for annual, 5-year and 30 year averaged means and averages for both broad and narrow cultural time periods. These precipitation datasets are reconstructed for the

Colorado Plateau from data made available by the U.S. National Oceanic and Atmospheric

Administration (http://www.ncdc.noaa.gov/paleo/pubs/salzer2005/salzer2005.html).

Large Sites are any site during any time period that has 50 or more above-ground rooms

(Adler and Johnson 1996). Sites of this size or larger are considered aggregated sites and did not exist at all times in the Mesa Verde Southwest. Only large sites that fall within a survey area, or near enough to exert influence, are considered.

Distance to large sites is the straight line distance from a small to large site. For the purpose of this study, the location of a large site was used as a focal point whether the site actually existed at the time or not. By comparing the distances between large and small site locations when the large sites did not exist to the distances when they did exist, any organizational changes around the large site’s communities can be identified.

If large aggregated sites do become the focal point for social organization, exchange,

9 ritual and political organization in their area and exert some type of control or influence on the communities, then when the large aggregated sites appear, the distribution of small sites should shift toward the aggregated sites to indicate the importance of these large sites.

Rehabitation occurs when people from a later period return to a location that was previously used. The difference between sites that are reinhabited and sites that are pristine during a time period is a useful variable to investigate population pressure, temperature, precipitation, elevation and landform changes through time. Comparing the choices in location made at reinhabited sites to those made at pristine sites during the same time period might make the reasons behind decisions clearer.

Rehabitation of Sites

Since the first habitation was built, small houses were replaced by slightly larger ones and larger ones by higher ones and villages grew. Houses are demolished so that others can reinhabit the same piece of land. Subsequent houses, villages, cities are built on the ruins of previous houses, villages and cities. In the Middle East, they become tels (Aharoni 1962) – raised hills easily identified on the landscape – consisting of layer upon layer of habitation. In the modern

New World, especially in the cities, old houses become the base for larger buildings.

In cities of all sizes and even in small villages, rebuilding after fire, flood, abandonment or simply to update seems sensible. In rural areas, why reuse the same land, when so much other land is available?

The Mesa Verde region is no exception. Multiple component sites occur in the area consistently (Hayes 1964, Reher 1977, Varien 1990, Varien 1991), but the frequency of

10 reinhabitation is not known. Because of the heavy habitation in the Mesa Verde area through time, it is not at all unusual that some sites overlap. What is unusual is that sites are often completely superimposed on each other, when adequate, apparently equally acceptable and unused land is available nearby.

Schiffer (1987:125) notes that “In the American Southwest it is not uncommon to find pithouses that have been excavated into previous deposits (and into earlier pithouses).” Varien comments:

“The degree to which this small site (5MT3918) was reoccupied raises an interesting question. It is unlikely that settlements repeatedly locate in exactly the same place by accident. The repeated occupation of Shorelene’s Site may have come about because of the ease with which new settlers could salvage raw materials from the previous use of the site, but one would not have to locate new structures directly inside old structures to salvage raw materials.” (Varien 1990:91)

And Schiffer (1987:104) comments that “because abandoned sites are resource areas, their presence might influence the settlement decision of later people.”

Rehabitation in the Mesa Verde area occurs. Comparing sites that were occupied both during the Pueblo III and at an earlier time – reinhabited sites – and sites occupied only during

Pueblo III – pristine sites – might lead to understanding the pattern of environmental and cultural forces that come into play during site location decisions.

The reasons behind the rehabitation of one site but not another remain unknown. Patterns of occupation in the Mesa Verde Southwest show some Basketmaker III, Pueblo I and II sites are never reinhabited, while others are reused at least once and often more than once (Varien 1990).

Through time, the population increases (Adler and Metcalf 1992, Cordell 1984:283) but, in most cases, the occupation of the landscape never approaches anything more than a very small village,

11 except during Pueblo I when large aggregated sites occurred (Breternitz, et al. 1986) and in the

Pueblo III when large villages appear (Bradley 1987, Bradley 1992, Kuckelman 1997, Kuckelman

2003, Kuckelman and Coffey 2007)

Schlanger (1992:92) refers to persistent places in the Ancestral Pueblo landscape as “a place that is used repeatedly during the long-term occupation of an area..” While she does not limit her persistent places to habitation sites, she does suggest that they have unique qualities making them especially suited for specific purposes. In other words, persistent places are reinhabited because they have some quality that makes them attractive. These might include stretches of farmland, marshland, riparian bottoms or stands of timber, as well as workable stone or clay outcrops. Other locations might become persistent places because of streams, springs or other topographic features. Schlanger also believe that such places may have hearths, storage spaces or other attributes either natural or anthropogenic that attract rehabitation. Finally, she suggests that, over time, persistent places become important because they are repeatedly used and the cultural materials accumulated there become attractive resources in their own right and are scavanged.

“Settling directly on top of sites occupied by one’s ancestors may have been a means of legitimizing territorial boundaries and marking systems of land tenure” according to Varien

(1990:91). He adds that determining ritual and religious significance will also be extremely difficult.

12 Old Data

Since the late 1800s archaeologists have been exploring the Southwest, excavating and collecting data. Archaeological sites are a fragile and nonrenewable resource and so any information collected might be the only data ever available from a site. Certainly a site once excavated cannot easily be excavated again and even if it can, much of the data are not recoverable. However, if and when a location is reexcavated it is possible to combine the knowledge from both excavations and determine, at least to some extent, what of the old data are incorrect (Kuckelman 2003).

It would seem that nondestructive full coverage survey would pose less of a problem, because nothing is destroyed and resurveying would not be difficult. In truth, often survey is done as mitigation before water behind a new dam floods the area (Breternitz, et al. 1986, Toll

1971), an airport is built or a building erected. In those cases survey cannot be redone, at least not until and unless the reservoir is drained, airports close or buildings fall. Even in cases where survey could physically be redone, the high cost in time and money to carry out full coverage survey generally means that any survey done for non-mitigation reasons will not be conducted in an area already surveyed.

So old survey data exist and the chances of obtaining new surveys in those areas are minuscule. Researchers in the area can then simply pick and choose which surveys they wish to use, sticking to the most recent believing that those are the surveys that convey the most accurate information and disregarding any surveys deemed old or out of date. This could be a great waste of previous effort and information. Some older surveys might be quite reliable, others might have portions or datasets that are usable and others, unfortunately, might be useless. Archaeologists are

13 used to preserving precious resources, but are losing them if they simply write off old survey data as useless.

More of a potential problem is using old survey data without consideration of their utility.

SHPO databases make access to large amounts of data easily available, but the data are not usually in usable form, nor are they necessarily checked for accuracy (Ortman, et al. 2007).

Archaeologists are beginning to make large scale computer programs to do modeling, simulations and reconstruction such as the Village Ecodynamics Project (Kohler and Leeuw 2007b) and some are employing a variety of methods to determine usable data (Ortman, et al. 2007).

Computers now allow analyses considered impossible in the past, but these applications require a large amount of information to create, for example, simulation models and validate them (Snow, et al. 2006a, Snow, et al. 2006b).

Figure 1 Map of the Four Corners area. Enclosed area approximately encloses the range of surveys discussed (MapSource 1999)

14 No survey data are perfect. Errors occur in every one. Sites are missed, misdated, double numbered, undercounted, over counted and sites that are not really there are incorporated into the data. These, however, are random occurrences. Systematic errors caused by preconceived biases, habitual practices and limitations in the survey are not random and signal a problematic survey. Those old surveys and perhaps newer ones under suspicion can perhaps be identified by comparison to surveys in the same area that are considered reliable. Statistical methods should be able to determine which survey is not like the others.

Study Area

The study area is in the southwestern corner of Colorado in Montezuma County in the

Four Corners area of the United States (Figure 1). This area is designated the Southern Colorado

River Basin area and encompasses parts of the Mesa Verde-Mancos and Monument-McElmo drainages (Lipe, et al. 1999). In the distant northeast are the San Juan Mountains, but the area is better known for and dominated by Mesa Verde and the Ute Mountain to the south. Ute

Mountain is not within the study area, but is visible in the southern portion of the area. Portions of the study area border on the Ute Mountain Ute Reservation, notably portions of Mesa Verde

National Park and the Mesa Verde Area Survey 2000. Mesa Verde National Park is administered by the National Parks Service (NPS), Mockingbird Mesa is on Bureau of Land Management land

(BLM), but it now resides within the NPS designation of Canyon of the Ancients National

Monument. Other spots of NPS and BLM land dot the area including Goodman Point Pueblo administered by the NPS. The remaining land is privately owned. In the center of the area is the modern town of Cortez, Colorado, in the Montezuma Valley. At the north is the town of Dolores,

15 Colorado. East, but not inside the area, is Mancos, Colorado. The topography ranges from about

2460 m on the north rim of Mesa Verde to 1550 m in the Mancos Canyon. In the south, the

Mancos River flows year round and in the north the Dolores River, although now dammed to form the McPhee Reservoir, was also a year round water source. Neither river is in or touches any of the surveys in this study. Currently a seasonal water source, McElmo Creek, is in the study area. Springs and seeps that form where sandstone and shale meet are located throughout the area occurring at canyon heads, in alcoves and at other locations. These springs and seeps “were likely sufficient to satisfy prehistoric needs, at least when populations were low.” (Adams and

Peterson 1999:22).

Historically, the area was dry-land farmed by European settlers producing a variety of crops, but relying heavily on the production of beans (Connolly 1990, Connolly 1992). Currently, because of irrigation from McPhee Reservoir other more water dependent crops are also grown in

Figure 2: Major vegetation found in the study area by elevation (Lipe 1999).

16 those areas where the canal system can reach. Land that cannot be irrigated is still farmed using dry land methods. Much of the private land, especially that in the Sand Canyon/Goodman Point area and on the McElmo Dome is currently being farmed.

The natural environment in the study area is mostly pinyon juniper woodland with small areas of gambol oak scrub land. The center of the area, near Cortez, Colorado, is sagebrush- saltbush (see Figure 2). Soils in the area are eolian-derived upland soils and alluvial soils along canyon terraces and in flood plains. The eolian-derived soils are considered good soils for dry- land farming and in many locations the soils have depths of up to 2 m. The alluvial soils are found mostly in lower floodplains and on canyon terraces and are considered good locations for other

Figure 3 Map of five surveys used in this research. A is Mockingbird Mesa, B is Sand Canyon/Goodman Point survey, C is the Mesa Verde Area 2000 survey, D is the Wetherill Mesa survey and E is the Chapin Mesa survey. (modified from MapSource 1999).

17 crops. Below the soil as a consequence of the low levels of rainfall is a layer of calcium carbonate or caliche which can be thick in places (Jenny 1941). This area of the Southwest has experienced arroyo cutting for the past 100 years. This type of erosion is usually cyclical and it is unknown if rapid arroyo cutting took place during Ancestral Puebloan times (Jenny 1980).

The dry-land agricultural potential for both the pinyon juniper woodland and sagebrush- saltbush areas is considered high. The dry-land agricultural potential for gamble oak scrub land is very low. Most of the study area falls within pinyon juniper woodland and sagebrush-saltbush areas. The potential for agriculture in both Monument/McElmo and Mesa Verde/Mancos drainages is high (Lipe, et al. 1999).

Precipitation in the study area has a biseasonal pattern. Snow falls in the winter months and summer thunderstorms between July and September supply other moisture. This pattern of moisture has been around since at least the 20th century C.E. (Dean 1996).

Figure 3 shows the areas of the five individual surveys in topographical context. The two

Mesa Verde National Park surveys – Wetherill Mesa and Chapin Mesa – are side by side on the mesa. The remainder of the mesa, which is not green, is Ute Mountain Ute Reservation land.

The Sand Canyon Goodman Point Survey area is north of Ute Mountain and the Mockingbird

Mesa survey is to the west of the Sand Canyon/Goodman Point Survey. The smallest survey, the

Mesa Verde Area Survey 2000, can be seen at the edge of Mesa Verde in the transition zone.

Ancestral Puebloan World – Mesa Verde

During the Pueblo time periods, Ancestral Puebloans were agriculturalists raising corn, beans and squash (Blinman 2008, Ford 1981). They lived in the Mesa Verde area on the

18 Colorado Plateau in the Figure 4 Escalante Ruin in Dolores, Colorado outside midst of sagebrush flats Anasazi Heritage Center (BLM). This is a small, Chacoan style Late Pueblo II site. and pinyon juniper forests occupying mesa tops and talus slopes

(Lipe, et al. 1999).

Their small habitations usually consisted of a room block of anywhere from 2 to 20 above ground rooms built of dry laid sandstone blocks

(Figures 4 and 5). Some of these room blocks were two stories high. The room blocks usually had front rooms with entrances and sometimes hearths and back rooms, probably used for storage. The entrance to the room block faced south toward the plaza and the kiva. The kiva was usually a circular room dug into the ground with only a smoke hole and ladder visible from above.

Associated with the room block and kiva Figure 5 Dominguez Ruin in Dolores, Colorado outside Anasazi Heritage Center (BLM). This is a was a midden. Some of these small sites small Pueblo II site. had towers (Cordell 1984). Depending on

their size, a room block kiva complex could

house a single family – husband and wife

with children, or a range of extended

families including grandparents, aunts,

uncles, married sisters.

The Ancestral Puebloans could

19 easily walk to their nearest neighbors and on the sage flats could probably see one or two neighboring homes in the distance. On Wetherill Mesa, for example, the entire community covered only 25 km2 (Hayes 1964), the communities around Sand Canyon and Goodman Point

Pueblos were about the same size (Adler 1990) and Mockingbird Mesa was only about 16 km2

(Fetterman and Honeycutt 1987).

No large permanent water source existed in the area, but seeps and springs occurred at canyon heads and where sandstone and shale layers meet. The Ancestral Puebloans could obtain drinking water easily but, because they did not irrigate their crops, they were not bound to locations near seeps and springs (Lipe, et al. 1999). Arroyos were often dammed with check dams to slow rainwater runoff. Small fields might be planted uphill of the check dam providing that small field with a bit more water. Larger reservoirs were created at larger sites (Wilshusen, et al. 1997).

Clay for pottery was widely available and grayware corrugated utilitarian pots and painted black and white whitewares were made in each family. It is likely that these coiled whiteware pots were fired communally (Ermigiotti 1997).

Pottery firing was not the only community activity that occurred through the year.

People probably gathered at great kivas – solitary ones or those situated within aggregated villages – during the year. This afforded occasions for feasting and social engagement community wide (Varien and Wilshusen 2002b).

Sandstone was abundant for building (Griffitts 1990). Morrison chert and Dakota quartzite were the rocks used for projectile points, scrapers and other lithic tools. Large metates and manos were generally made from local sandstone or conglomerate and used to grind corn

20 (Lightfoot and Etzkorn 1993) . Obsidian was seen only in small amounts and was imported as there is no volcanic glass in the immediate area. Projectile points were used on arrows, and the bow and arrow was the predominant hunting tool. Mule deer, rabbits, jack rabbits and turkeys were hunted. Turkeys were also domesticated at times, but more for feathers than for food.

Pinyon nuts were a staple food from pre-agricultural times and prickly pear cactus along with sego lily, cattail and other root-type plants were gathered and eaten (Dunmire and Tierney 1997).

However, by Pueblo III times, corn, beans and squash were the staple crops (Blinman 2008).

Soils in the area are deep and fertile, wind deposited in the distant past. One of the best indications of good agricultural soil is the existence of sagebrush in the field. This slow growing bush likes the same locations that maize does (Blinman 2008). Because of the marginality of the environment for growing maize, small fields would have been planted in a number of different locations with varying elevations, exposures and slopes to ensure that some of the corn would mature. Planted in one field, a hail storm could wipe out an entire years work.

Maize was probably planted in foot-deep recessed wells dug into the cleared fields. Seeds were planted 3 to 4 in. below the soil surface in the well. About six plants were allowed to remain in each well and the wells were spaced three paces apart on all sides. Weeds were probably not allowed to survive in Ancestral Puebloan fields. (Brown, et al. 1952). They might or might not have planted beans in the same holes with maize and squash as is done on some of the Northern

Rio Grande Pueblos, but is not at Hopi. The varieties of maize grown in any specific location would be adapted to that elevation or moisture regime. Harvested maize was stored for three years (Blinman 2008, Brown, et al. 1952) to mitigate starvation from potential droughts. Three years worth of maize would be on hand for eating and as seed. The maize was stored either in

21 underground bell shaped cysts (Varien 1991) or in small storage rooms tucked into cliff edges.

Another crop probably grown was sunflowers for seed and dye (Ford 1981).

Ancestral Puebloan existence was not idyllic. During the Pueblo III warfare and cannibalism occurred (LeBlanc 1999, Wilcox and Hass 1994, Turner II and Turner 1992, White

1992) although whether the local inhabitants needed protection from outside forces or from their neighbors is unclear. In some cases, near the end of the Late Pueblo III entire large villages were apparently abandoned after major battles (Kuckleman 2000).

Mesa Verde Archaeology

From “windshield” to full coverage, survey plays an important part in archaeology today, especially in settlement pattern research (Billman 1999). The use of survey to establish the basis of settlement patterns is often traced to the Virú Valley Project (Willey 1953, Willey 1999).

Sanders (1999:12) admits that the Basin of Mexico project “was visualized as the first stage of a long-term, ambitious plan to apply Gordon R. Willey’s Virú Valley methodology (i.e., the study of regional prehistoric settlement patterns using a surface survey strategy).”

However, recording of sites and excavation had of course been going on far earlier than

1946 when the Virú project took place. The first mention of archaeological sites in the Central

Mesa Verde Area occurred in the journals of the Dominguez-Vélez de Escalante expedition in

1776, but undoubtedly the indigenous Ute, Navajo and Pueblo peoples were aware of the ruins, cliff dwellings and other artifacts in the area.

Perhaps one could say that survey in the United States West began in the mid to late

1800s when geographical and geological survey teams roamed the area recording geology,

22 biology, botany, fossils and archaeological sites. The first expedition into the area was led by

Captain John N. Macomb. His geologist, J. S. Newberry, visited Escalante Ruin and climbed to the top of Mesa Verde. He recorded a number of small pueblo sites as well (Lipe, et al. 1999).

Although the most famous expedition leader, Ferdinand Vandiveer Hayden and his photographer

William H. Jackson, who explored the Colorado area, might have missed the ruins at Mesa Verde, they did explore, record and photograph more than 40 sites along the Mancos River (Jackson and

Holmes 1981). They, along with William H. Holmes, conducted scientific survey of the area from

1867 to 1878. The Hayden Survey, through the use of maps, sketches and photographs and a variety of publications including annual reports and monographs, documented a wide swath of the area. They did, however, not explored Mesa Verde nor see the cliff dwellings that were to become so famous.

Richard Wetherill, for whom Wetherill Mesa is named, and Charlie Mason, local ranchers from the Mancos area who were searching for lost cattle, spotted Cliff Palace on what is now

Chapin Mesa in December of 1888. This discovery brought the large sites of the Central Mesa

Verde area to light (Lipe, et al. 1999, McNitt 1991, Snead 2001). From the discovery of Cliff

Palace to his murder in Chaco Canyon, New Mexico, in 1910, Wetherill was to play an important part in the archaeology of the Southwest. After Cliff Palace, Wetherill along with his four brothers continued to explore Mesa Verde. They quickly looked at 182 cliff dwellings, collected artifacts and produced exhibits, including one at the Columbian Exposition in Chicago in 1893

(McNitt 1991). Wetherill conducted an expedition into southeastern Utah in the 1890s where he conducted work for the American Museum of Natural History. Although not a trained archaeologist, his scientific bent led him to collect stratigraphic information that became the basis

23 of southwestern chronologies (Lipe, et al. 1999). Eventually, Wetherill met and guided many of the explorers and archaeologists who came to the Four Corners area. This included a rather long list of people who came very shortly after discovery of Cliff Palace on the Mesa Verde Area. He subsequently moved to a ranch in Chaco Canyon and fell out of favor with local archaeologists.

He died while herding cattle in Chaco Canyon in 1910, ambushed by a group of Navajo (McNitt

1991).

One of the first to visit with Wetherill was Frederick Hastings Chapin for whom Chapin

Mesa is named (Chapin 1988). He came in the summers of 1889 and 1890 and shortly thereafter wrote an article, “Cliff-dwellings of the Mancos Cañons,” that described the landscape and antiquities in the area. In 1892 he published The Land of the Cliff Dwellers as a more in depth report of his explorations (Chapin 1988). Gustaf Nordenskiöld, in his work on Mesa Verde, named the mesa containing Cliff Palace after Chapin because he published the first works on the area in the early 1890s (Lister 1988).

Although not an archaeologist, Nordenskiold conducted the first scientific excavations in the area. Trained as a geologist and physical scientist he came to the area from Sweden in 1891.

He stayed to excavate some of the Mesa Verde cliff dwellings. His excavation work was mostly on Wetherill Mesa, exploring the large cliff dwellings, and he exported the majority of his finds to

Europe. His collection eventually ended up in the National Museum of Finland. In 1893 he published The Cliff Dwellers of Mesa Verde in both Swedish and English (Nordenskiöld 1893).

Most of what Nordenskiold suggests about the culture of Mesa Verde was incorrect as we know things today. His major contribution was to publish a complete, detailed report of his work in a timely manner (Lipe, et al. 1999).

24 While not carrying out survey, per se, Chapin, Wetherill and Nordenskiold began the long process of identifying, classifying and dating archaeological sites in the northern Southwest, but they mostly identified the large sites – the cliff dwellings or large mound that stood out on the landscape. The practice of full coverage survey, or any truly systematic survey, did not exist in their time.

In 1903, T. Mitchell Prudden, an eastern pathologist, visited the Southwest frequently and conducted the first formal surveys in the area (Kidder 1973, Lipe, et al. 1999). Prudden was also one of the first to look at the small sites and his major contribution is the Prudden Unit. A

Prudden Unit consists of a pit structure, above ground rooms and a midden. Prudden identified these small sites as stand-alone habitations as well as parts of larger aggregated settlements.

In 1907, Alfred Vincent Kidder, Sylvanus Griswold Morley and John Gould Fletcher, all

Harvard students, were assigned by Edgar Hewitt to survey an area near McElmo Canyon. They recorded many Pueblo III sites, but were not interested in sites from other time periods. Morley later excavated Cannonball Ruins, a Pueblo III canyon head aggregated site (Givens 1992). His analysis of the site would be considered questionable now, but he did publish immediately after returning from the field (Lipe, et al. 1999).

Jesse Walter Fewkes excavated and repaired or reconstructed sites at Mesa Verde

National Park from 1908 to the 1920s. He, as with most of the previous archaeologists, was unconcerned with chronology (Lipe, et al. 1999). While he worked at many of the large pueblos and did publish his work, many of his interpretations were based on misconceptions (Kidder

1973). Fewkes continued with the same approach throughout his career, never adopting more modern approaches (Lipe, et al. 1999). His work, however, had a great impact on the Park and

25 some of it is still visible today.

During the first part of the 20th century, Southwestern United States archaeology was changing. The Pecos Conference in 1927, organized by Kidder, produced the Pecos classification, the culmination of the previous 20 years work. During that time, archaeologists in the Southwest looked at applying a variety of scientific approaches to their work. Included in these were stratigraphic excavation, recording of provenience, ceramic seriation, quantification and architectural typologies.

Crucial to a study of southwestern prehistory was a workable chronology. Kidder, using ceramic seriation from Pecos Pueblo, Nels Nelson using careful stratigraphy and Earl H. Morris excavating at Aztec ruin combined to produce a chronology for the area (Cordell 1984, Kidder

1927). By 1921, Morris had published the first chronology and in 1924, Kidder (1973) published

An Introduction to the Study of Southwestern Archaeology which focuses on his work at Pecos

Pueblo, New Mexico, but includes an extensive chapter on the San Juan Basin. By 1927, when

Kidder organized the Pecos Conference, most archaeologists in the area were ready to produce a synthetic, formal chronology that could be applied throughout the Southwest (Lipe, et al. 1999).

From the Pecos Conference to 1957, archaeologists in the Central Mesa Verde Area and the rest of the Southwest focused on obtaining information that was missing (Lipe, et al. 1999).

The number of players multiplied and they began to flesh out the descriptive components of archaeology. During this time, Lancaster carried out the survey of Chapin Mesa (Rohn 1977).

Yellow Jacket Pueblo (5MT5), one of the largest aggregate pueblos in the region, and the smaller pueblos surrounding it were the first sites in Colorado to receive state numbers. In 1931, the Museum of Western State College in Gunnison, Colorado, operated a field school at Yellow

26 Jacket Pueblo (Site 5MT5) (Kuckelman 2003). This large pueblo was estimated at more than 600 rooms (Adler and Johnson 1996), but after limited excavation by Crow Canyon Archaeological

Center in the late 1990s it is now believed to have had as many as 1200 surface rooms, 195 kivas and 19 towers covering about 100 acres (Kuckelman 2003).

In the mid-1950s, archaeology in the area again changed with the beginning of the Navajo

Reservoir salvage project and the Wetherill Mesa Survey. Other projects, large and small, focused on salvage. Theoretical approaches changed, settlement pattern studies and community social organization came to the forefront and cultural ecology was the primary theoretical framework. Julian Steward’s cultural ecology (Steward 1972) also affected Willey in his Viru survey (Willey 1953, Willey 1999). Influenced by Steward and Leslie White (Lipe, et al. 1999),

Southwestern archaeologists moved on to the “New” or processual archaeology. Much of the effort in the 1960s focused on Mesa Verde National Park and on such large sites as Yellow Jacket

Pueblo. By the early 1970s inventories were undertaken on many federally held lands to satisfy the

1966 National Historic Preservation Act and the 1969 Executive Order 11593. With the passage of the Archaeological and Historic Preservation Act in 1974, the range of federal agencies required to take into account any possible damage to archaeological sites increased.

In 1978 the Dolores Archaeological Project (DAP), one of the first large projects funded under the Archaeological and Historic Preservation Act, began. This effort ran from 1978 into the early 1990s and was a mitigation project to recover data from the area to be flooded by the

McPhee Reservoir (Breternitz, et al. 1986, Kane and Robinson 1988, Lipe, et al. 1988, Toll

1971). Although this project, which included survey and site excavation, was located north of the areas considered in this study, it greatly influenced some of the subsequent research in the area.

27 Colorado has an extensive survey database administered by the State Historic Preservation

Officer (SHPO). The sites from the DAP, the canal surveys and irrigation ditches became part of the database, as had all the surveys done in conjunction with carbon dioxide pipelines, oil and gas exploration, roads and canals in the area since. With all that, only 13 percent of Montezuma

County has been surveyed and much of that in small pieces (Lipe, et al. 1999).

Organization of Chapters

Chapter 2. The SURVEYS includes a brief discussion of survey in the Mesa Verde area.

Each of the five surveys discussed is individually described, including the survey’s basic parameters, history and the dataset’s pedigree. Included is a discussion of the variables recorded in the databases and my attempts to make all five databases compatible. Also included in this chapter is a discussion of dating in the Southwest including a brief history, the common cultural stages and my accommodations in making the surveys comparable. A formula for normalizing the cultural period lengths so that not only can surveys be compared, but the various cultural phases – time periods of unequal length – can be compared is included.

Chapter 3. SITE POPULATION explores the numbers of habitation sites across the five surveys in various time periods from Basketmaker to Pueblo III. The site populations are graphed and compared to each other and to a database composed of the five surveys combined. This chapter not only looks at the progression of population through time, but also is used as a check of the validity of the individual databases. All analysis is done for both broad cultural periods –

Basketmaker, Pueblo I, Pueblo II and Pueblo III and narrow cultural periods – Basketmaker,

Pueblo I, Early Pueblo II, Middle Pueblo II, Late Pueblo II, Early Pueblo III and Late Pueblo III.

28 Chapter 4. LANDFORM looks at the topographical locations of habitation sites and how they changed through time. Included are the arbitrary definitions of landform and the rationale for these imposed by me on the various surveys. Again, both broad cultural time periods and narrow cultural time periods are investigated. All five surveys are included in this chapter.

Chapter 5. ELEVATION AND CLIMATE looks at both the elevation of sites in the survey and at the regional temperature and precipitation reconstructions. Elevations are shown as ranges with mean elevations recorded. Elevation ranges are compared across surveys and within surveys through time. Chapin Mesa is not included in this analysis because the data from that survey are not considered accurate. This chapter also looks at temperature and precipitation reconstructions on an annual, 5-year and 30-year averaged basis. Both temperature and precipitation are averaged over broad and narrow time periods as well. Elevation and population data are plotted against both precipitation and temperature information.

Chapter 6. DISTANCE TO LARGE SITES explores the geographical relationship between the small habitation sites and the aggregated large sites with 50 or more rooms.

Communities of small sites most closely attached to large sites are identified and the distances between small sites and their nearest large site neighbor are calculated. The contraction and expansion of these communities are investigated. Only three surveys are considered in this chapter. Again the Chapin Mesa database is not used because of its inaccuracy. The Mesa Verde

2000 database is not used because there are no large sites within the survey boundaries and the distance to the nearest large site is large enough to make the difference in distances trivial.

Chapter 7. REINHABITED SITES compares Pueblo III sites that are reinhabited to those that are built on pristine ground. This dichotomy of sites is explored for the variables previously

29 considered – population, landform, elevation and distance to large sites. The variables are investigated in both broad cultural time scales and narrow cultural time scales.

Chapter 8. SUMMARY & CONCLUSIONS sums up the analysis of survey usefulness and presents a method for vetting old or suspect surveys. The variables of site population, landform, elevation, temperature, precipitation and distance to large sites are also evaluated.

Using reinhabited versus pristine sites as a tool for discovering trends in settlement patterns is explored and conclusions as to the basis for choice of habitation locations are presented.

30 Chapter 2. THE SURVEYS

With a history of exploration dating to the mid-19th century, the assumption is always that large amounts of work have already been done in the Four Corners area and that Southwestern archaeology must be quite mature. Nothing could be further from the truth. While surveys of

Chapin Mesa (Rohn 1977) and Wetherill Mesa (Hayes 1964) were done in the 1950s and 1960s and surveys of the other mesas in the park (Smith 1987) were completed in the 1970s, little full coverage intensive survey has been done in areas outside the park.

A request for surveys recorded from the Colorado State Historic Preservation Office yielded 42 surveys in the Montezuma Valley near Mesa Verde. Of these surveys, 21 were linear surveys and 21 were block surveys. Linear surveys are very prevalent in the area. They are usually performed under mandate of Section 106 of the National Historic Preservation Act of 1966

(Public Law 89-665). This law requires federal agencies to plan their undertakings so as to take into account any possible effects on archaeological and historic sites. A large portion of the archaeology in the U.S. today is done to comply with Section 106.

In the Mesa Verde region there are the usual road right of ways, irrigation ditches and airport expansions that all require federal funding and consequently require archaeological survey and mitigation. Also in this area, large amounts of carbon dioxide are pumped from the ground and piped to oil fields in the south. These pipelines also require survey. This requirement accounts for all the linear survey and some of the block survey.

Even among the block surveys, large full coverage surveys are rare. Only 6 of the 21 block surveys were more than 20 acres, the largest being 320 acres. The total of all 21 block surveys was 938 acres. While there are a few large surveys outside the Mesa Verde National Park, the

31 area outside the park is mostly unsurveyed. With so little of the area surveyed and yet well over

13,000 sites recorded in the county as a whole, many sites were recorded without formal survey and settlement patterns in the Mesa Verde area are not yet fully understood.

Approach to Survey Data

Working with a variety of surveys conducted at different times and with different people creates problems in uniformity. No one expects the data types to be identical, but the absence of any standards or expectations through time in archaeological survey makes the problem stickier than in other fields. Table 2 illustrates a few of the problems encountered when working with data from the five Class III – full coverage or “total” – surveys used in this study.

As can be seen, the first survey was begun in 1951 while the last was completed in 2000.

Over 50 years many things have changed in archaeology. How good is the information gathered by people in the field in the middle of last century compared to those in the field today? Was Class

Table 2 General survey information

Recorde Area in Survey Year Archaeologist d km2 Sites Area

Chapin Mesa 1951+ James A. Lancaster unknown unknown 987

Wetherill Mesa 1958-60 Alden C. Hayes 6,274 acres 25.4 km2 807

Sand Canyon/ Goodman Point 1986-9 Michael A. Adler 44 km2 44 km2 442

Mockingbird Jerry Fetterman Mesa 1981-4 Linda Honeycutt 3,976 acres 16 km2 834

Mesa Verde Area 2000 A’ndrea Elyse Messer 320 acres 0.5 km2 56

32 III survey the same then as it is now? If the reports written then (Hayes 1964, Hayes 1981) are any indication, survey done properly at any time can be of value not just for the survey’s initial purposes, but also in the future. On the other hand, surveys done incompletely or with little care might serve no future purpose as their reliability is questionable.

In order to work accurately with dissimilar surveys, the database categories must all match even if some of them are empty or some variables have to be recharacterized to fit the database requirements. This is especially important if databases are to be combined at any point in the project. As shown in Table 2, one problem is that some surveys used the metric system, while others used the British system of measurement. Operationally this is easily corrected, but can be time consuming and if not done carefully can become a source of errors. Designations of landforms vary from survey to survey forcing interpretation of meaning that might not always be clear.

Five surveys completed by five different people or groups of people, each with their own quirks and biases, can be daunting. While no one can ever completely remove personal bias from the survey data, and while anyone will automatically inject their own bias into reinterpretation, one way at least to equalize these peculiarities is to become familiar with the raw data whenever possible.

In today’s world of computer databases and available information, it is easy to contact the

State Historic Preservation Office (SHPO) and request the data from X or Y survey. Although not yet available online, the SHPO will quickly return the information. While fast and fairly easy, I chose to obtain the large portion of my data from the actual field notes or forms whenever possible. This was partly because not all data are recorded by the SHPO, and partly because

33 simply accepting columns of numbers does not really tell me anything about the surveys. Some data, GPS coordinates and elevations, were obtained from other sources or from maps. Beginning with raw data increased the time necessary to create the databases, but also allowed me to get to know the surveys intimately. Because I was completely familiar with the data, I could tweak and adjust portions for uniformity with some knowledge of exactly what I was changing, rather than arbitrarily changing information for conformity.

Chapin Mesa Survey

Chapin Mesa is located in the southwestern corner of Colorado in Montezuma County. It is the largest mesa in the Mesa Verde National Park and is bounded on the east by Soda Canyon and Little Soda Canyon and on the west by Navaho Canyon (Griffitts 1990). The Mesa Verde formation is a cuesta that rises from south to north at about 7 degrees (MVNP 2005). Mesa

Verde consists of a series of north-south oriented mesas including the mesas of the national park and those mesas residing in the Ute Mountain Tribal Park on the Ute Mountain Ute Reservation.

Part of Chapin Mesa and part of Wetherill Mesa form the public areas of Mesa Verde National

Park. The remainder of Chapin Mesa is on the Ute Reservation.

In 1906, 22 days after Theodore Roosevelt signed the Antiquities Act, he signed a separate act creating Mesa Verde National Park (Lee 2000). Serious, organized investigation and excavation were part of the park plan. This included a series of excavations and restorations intended for the education of the visitors to the park (Rohn 1977).

Organized survey of the park did not begin until 1951 when James A. Lancaster began a survey of Chapin Mesa that he completed in 1954. A report of this survey supposedly included

34 sketch maps and site characteristics (Uhler 2007), but no citation for this report can be found and in Lancaster’s obituary (Judge 1996) there is no mention of his involvement with the survey.

Arthur H. Rohn, in Cultural Change and Continuity on Chapin Mesa (Rohn 1977), does not make reference to the report. However, it is generally understood that because of the time, the labor costs and the financing, this survey was unable to determine accurately the site locations although each site was numbered and marked with a metal stake (Rohn 1977).

Information from the survey was recorded on field cards and stored at the park in the archives. In 1958 Rohn joined the park and as one of his duties was assigned “general collation of the mass of data that had accumulated since the beginning of survey activities in 1951,” (Rohn

1977:v). The outcome of that was Cultural Change and Continuity on Chapin Mesa (Rohn 1977) that provides information on habitation sites and non-habitation sites, as well as ceramics, cultural sequences and an evaluation of the sites through time. One important contribution of this work is a finer determination of the placement of the habitation sites into time periods.

In recent years, efforts on Chapin Mesa have included GPS survey of the numbered sites.

Since a series of fires in the past 12 years (www.nps.gov/archive/meve/fire/index.htm), more survey has been carried out on the mesa, but those data are not available.

The Database of Chapin Mesa Survey (Messer 2005a) was created from a variety of sources. I received permission from Mesa Verde National Park to review and copy information from the field cards for both Chapin Mesa and Wetherill Mesa. During a two-week period, field cards from 987 sites on Chapin Mesa yielded the basis for the Chapin Mesa database. Those field cards contained the site numbers assigned to each site, the date surveyed, the map number for the map on which the site was recorded, the site type (habitation, burned stones, shrine, check dam,

35 petroglyph, pithouse, etc.), estimated habitation period, landform and whether a sketch existed.

Other information that might appear on the card was placed in a comment column. The majority of “comments” consisted of common site names which helped to correlate the site information from Rohn (1977) and the GPS survey. Other comments included materials such as adobe, slabs, upright ovals or the presence of a kiva or the number of rooms.

In general, the field cards for each site were sparsely recorded. Often the description was mound or house mound, sometimes “small mound,” “medium mound” or “large mound.” While some cards indicated the site was a habitation site, others while describing a house mound did not indicate it was a habitation site. There was little uniformity in recording the archaeological information although information on landform seemed to be uniform and consistent. Little information appeared on any of the cards about architecture except for indicating the presence of slabs or adobe. Very few sites show any room counts or estimation.

While Rohn’s ( 1977) description of the sites is not clear cut, it was possible to tease out from his tables more information on the sites. By his own description, certain burned mound sites with certain types of refuse classify as Pueblo II, Late Pueblo II or Pueblo III. Sites without specific architectural remains are listed in a miscellaneous-sites table and specifically dated if possible. Rohn also estimates the number of rooms for each habitation site and lists the actual sizes of the mounds. For sites located on the talus slope, determination of time period was made from a combination of architectural features and pottery counts which are included (Rohn 1977).

While Rohn suggests that Jesse L. Nusbaum found some early pottery and perishable remains in at least three cliff sites (Rohn 1977:233), no Basketmaker III sites are recorded in cliff locations.

Once the information from the field cards and from Rohn (1977) were input into the

36 standard database structure, the database was missing only locational information. Mesa Verde

National Park supplied me with the GPS coordinate list for the sites on Chapin Mesa and these data were input into the database as UTM Northings and Eastings. All locations are within Zone

12. I personally input and data checked all the data from the field cards, Rohn’s (1977) book and the GPS survey.

The final database for Chapin Mesa (Messer 2005a) contained 987 separate site entries.

This does not mean that there are 987 recorded sites on Chapin Mesa, but that there are 987 site components. Sites might be recorded twice if there are multiple dates or widely varying usage information. The Chapin Mesa database (Messer 2005a) contains 897 entries that are considered habitation sites or a 90.9 habitation site percentage as shown in Table 3. Table 4 shows the distribution of site types for the survey. The Chapin Mesa Database (Messer 2005a) is on record with Mesa Verde National Park.

Wetherill Mesa Survey

Wetherill Mesa is located on the western most side of Mesa Verde National Park and is bounded on the east by Long Canyon and on the west by Rock Canyon (Griffitts 1990). Along with Chapin

Mesa, Wetherill makes up the visitor accessible area of the park. The survey of Wetherill Mesa began in 1958 as part of a program of survey and excavation in the park. Alden C. Hayes began a small portion of the work in 1958 and in 1959 work began on a large scale. The survey was completed in the fall of 1960 having recorded approximately 800 sites. In 1964, The

Archeological Survey of Wetherill Mesa (Hayes 1964) was published.

37 Table 3 Site counts of surveys used

Survey Database Total Site Habitation % Components Site Habitation Components Chapin Mesa Database 987 897 90.9 Wetherill Mesa Database 807 614 76.0 Sand Canyon/ Goodman Point 444 290 65.3 Database Mockingbird Mesa Database 818 267 32.6 (Lithic Scatters Removed) 538 267 49.6

MVAS 2000 Database 56 40 71.4

Total 3112 2108 67.7 (Lithic Scatters Removed) 2832 2108 74.4

As was the case for the Chapin Mesa survey, information from the survey was recorded on field cards and stored at the park in the archives. I was also given access to these cards by Mesa

Verde National Park and they became the basis of the Database of Wetherill Mesa Survey

(Messer 2005b). These field cards were much more detailed than those for the Chapin Mesa

Survey. Most contained sketch maps, nearly all had some explanation of architecture and assigned a date to the site. Many more non-habitation sites were included in the survey. Hayes was clear to explain in his report (Hayes 1964) the extent of his survey including the fact that the talus slopes and canyon bottoms were surveyed completely. It amazed me to realize that the information collected in this survey fairly well satisfied the requirements made by the Colorado SHPO for survey carried out today.

38 Another aspect of this survey that was unusual for its time was the use of radio triangulation to determine site locations. In February 1961, Hayes reported in American Antiquity

(Hayes and Osbourne 1961) on the development and use of this equipment. These radio instruments produced a signal readable, with some adjustment, for a half mile and, with the use of two radio source beacons, the site location could be determined. This direction finder allowed the surveyors to locate sites in dense forest, increased location accuracy and decreased the number of survey days by half, according to Hayes.

Better accuracy not withstanding, the locational data used in the Wetherill Database

(Messer 2005b) were taken from the recent GPS survey done by Mesa Verde Park’s staff.

Additional information for the database was taken from the extensive tables included in Hayes

(1964) which included architectural information, pottery data and dates.

Because one aim of this work is to show the value of old survey data and to find a way to determine which old survey data can be used, it is important to consider the hands-on feel of the data and the reputation and history of the data gatherers. Hayes went on to survey a variety of locations for the National Park Service, including Chaco Cultural Historic Park in New Mexico

(Hayes 1981). His obituary (Woodbury 1999:616) states that “he perfected the exacting techniques of the “total” archaeological survey, making possible at Mesa Verde and Chaco

Canyon much more adequate planning for the subsequent stages of major research programs.”

Woodbury (1999:616) notes that “his promptly published and detailed reports became the cornerstones of the future research of others.” This is obviously true as Linda Cordell’s (1972) dissertation, Settlement Pattern Changes at Wetherill Mesa, Colorado: A Test Case for

Computer Simulation in Archaeology relied on Hayes’ work (Hayes 1964) published only eight

39 years before.

Perhaps Hayes himself summed up his approach to “total” archaeological survey when he wrote, “We have included details which will be of little interest to many readers. We have done this in the belief that much that may seem irrelevant today may, in the light of future discoveries, provide a clue.” (Woodbury 1999:617).

The Wetherill Mesa Survey was not the first archaeology done on that mesa, but Hayes was meticulous in correlating Nordenskiöld’s site numbers (Nordenskiöld 1893), Watson-Franke

Survey numbers and those assigned during the Gila Pueblo survey with the numbering scheme used during his survey and with the, by then, common names of sites.

One problem with the Wetherill Mesa data is that the project used a cultural dating system different from anything used in the area prior to when the Wetherill Mesa survey took place.

Fortunately, Hayes (1964:88) included in his report a chart correlating the Wetherill Survey period and phase system with that used by the Pecos Classification. It was then possible to convert Wetherill dating information to the system of modified Pecos Classification used by

Michael Adler in dating his survey (Adler 1990, Adler and Varien 1991). The final database for

Wetherill Mesa (Messer 2005b) contained 807 separate site entries (Table 3). Again, this does not mean that there are 807 recorded sites on Wetherill Mesa, but that there are 807 site components.

Sites may be recorded twice if there are multiple dates or widely varying usage information. The

Wetherill Mesa database (Messer 2005a) contains 614 entries that are considered habitation sites or a 76.0 habitation site percentage as shown in Table 3. Table 4 shows the distribution of site types for this survey. The Wetherill Mesa Database (Messer 2005a) is on record with Mesa Verde

National Park.

40 Table 4 Distribution of sites by site type Site Type Wetherill Mockingbird Sand Canyon/ Chapin MVAS Mesa Mesa Good Point Mesa 2000 Habitation 614 267 290 897 40 Limited Activity 144 528 103 80 12 Storage 24 17 27 8 1 Ceremonial 7 6 5 1 2 Unknown 13 0 0 0 0

Mockingbird Mesa Survey

The Mockingbird Mesa survey was carried out between 1981 and 1984 for the U.S.

Bureau of Land Management in Montezuma County by Woods Canyon Archaeological

Consultants, Inc. This survey covered 3,976 acres of land on the top of Mockingbird Mesa. While the Mesa Verde National Park surveys and the Sand Canyon locality survey were done on lands in near proximity, the Mockingbird Mesa survey is about 10 miles northwest of Cortez, Colorado. A report on this project was published in 1987 (Fetterman and Honeycutt 1987). The project included four different crews over the course of the three field seasons. While the authors of the report acknowledge that the time span and crew differences did influence some information, they have, as much as possible, accounted for that in their report.

The information in the database (Messer 2006) was obtained from the Fetterman and

Honeycutt report (Fetterman and Honeycutt 1987). The Appendix: Site Data was especially helpful in creating the database. Although I have no first hand knowledge of this survey, the

41 report is clear, complete and concise and I have no difficulty including this information in my study.

The final database for Mockingbird Mesa (Messer 2006) contained 834 separate site entries (Table 3). Again, this does not mean that there are 834 recorded sites on Mockingbird

Mesa, but that there are 834 site components. The Mockingbird Mesa database contained 16 sites dated to the Archaic which are not included in this study. The Mockingbird Mesa database contains 267 entries that are considered habitation sites or a 32.6 habitation site percentage. This survey also contains a very high number of undated sites that are predominantly lithic processing or fire processing sites (Table 4). This may be because of the heavier Archaic influence identified in the area, or because the survey was better equipped to identify these limited activity sites. If those sites are removed from the database, then the percent of habitation sites becomes 49.6 as shown in Table 3. A modified version of the Mockingbird Mesa Database (Messer 2005a) is in

Appendix A.

Sand Canyon/Goodman Point Survey

The Sand Canyon/Goodman Point Survey was carried out in the Montezuma valley in the areas surrounding the Ancestral Pueblo sites of Goodman Point and Sand Canyon. As part of the

Crow Canyon Archaeological Center’s Sand Canyon Project (Lipe 1992) three surveys were carried out within what is called the Sand Canyon Locality. These surveys were all intended to be

“total” archaeological surveys. The first survey was conducted in the summer of 1987 and covered a two-mile radius around Sand Canyon Pueblo (Van West, et al. 1987). The second survey, which expanded the area of the first survey was conducted in the summer of 1988 by

42 Michael Adler ( 1988, 1990, 1992, Adler and Varien 1991). Together, these two surveys are referred to as the Upper Sand Canyon Survey.

A third survey, conducted during the summer of 1990 (Adler and Metcalf 1992) focused in the lowlands south of Sand Canyon Pueblo. While this survey is included in Adler’s original database, it is not used in this study because it was not done under the same protocols as the upper Sand Canyon surveys (Varien 2004).

While these data were used in a variety of research projects between 1990 and 1993, it was agreed that I would combine the information from two existing computerized datasets, site reports and field notes to create a complete computerized database of this project. This database

(Messer 1994) was completed and delivered to Crow Canyon Archaeological Center and Adler in

October 1994. It was also used as the basis for my master’s work (Messer 1995).

Because the two upper Sand Canyon surveys were not done by the same people nor were they done to the same specification, the two datasets needed to be modified before they could be joined together. The original 1987 survey field notes were not available, so data were taken from the report filed with the Bureau of Land Management (Van West, et al. 1987). The BLM report contained synopsis of the actual field notes. The original site survey forms for the 1988 field season were available.

Whenever possible, data were taken from the site reports and input into the database and then checked against the existing databases and reports. This approach not only supplied a check on the data, but also allowed me to begin to understand the way the forms were filled out and the nature of the survey. It became evident that there were patterns in the field notes. At one time I was certain that I could take all the site forms filled out by a single field crew on a specific date

43 and order them by time of day simply by how carefully and seriously the forms were completed.

This is not to say that forms filled out later in the day were any less complete or accurate, but the crew personalities certainly were more evident on some forms than on others.

The final database for Sand Canyon/Goodman Point Survey (Messer 1995) contained 444 separate site entries (Table 3). The Sand Canyon locality database (Messer 1995) contains 290 entries that are considered habitation sites or a 65.3 habitation site percentage as shown in Table

3. Table 4 shows the distribution of site types for the survey. A modified version of the the Sand

Canyon Database (Messer 1994) is in Appendix B.

Mesa Verde Area Study 2000 Survey

The Penn State Mesa Verde Area Survey was conducted by me during the summer of

2000. A crew of three archaeologists worked for eight weeks on a half section (0.5 mi2) of land.

The purpose of this survey was to fill in some of the gaps in geography in the knowledge base of the area. With the Chapin Mesa, Wetherill Mesa and Mockingbird Mesa surveys covering uplifted areas and the Sand Canyon/Goodman Point area covering the lower canyonland area and the

McElmo Dome, the areas on the periphery had not been surveyed to any extent.

This transitional area between mesa and canyonlands, as far as I could determine, was unexplored. Several people suggested that there would be few archaeological sites in this area.

Others people assured me that there would be sites in this area, just as there are many sites on the talus slopes (Varien, et al. 1996, Varien 1990, Varien 1991) where once archaeologists thought none existed.

The only formal survey previously done in this area was completed for Reach II of the

44 Towaoc Canal, an 11.6 mile reach running along the Highline Ditch and feeding the Rocky Ford

Laterals, a series of underground irrigation pipes. The Class III survey for this project located 44 prehistoric sites in this narrow band, with none in the vicinity of the survey area (Errickson 1995).

The canal, which has been completed, did not run in the Mesa Verde Area 2000 study area, but did, at times, afford access to the power line road.

The survey took place on a half section in the area below the talus slope of Mesa Verde.

This particular section is below the Ute Mountain Tribal Park portion of Mesa Verde and, in fact, borders to the north on Ute land. The southern extent of the area slopes into the flat, cultivated lands of the valley. While there is no talus in the area, the northernmost portion is a wave of ridges or fingers coming out from beneath the talus and pointing toward the south. Nearly half of the area covered by the survey included these ridges, while the southern portion sloped down into the sage flats.

The survey was conducted using the site recording forms required by the Colorado

SHPO’s office and a report (Messer 2001b) was filed with the SHPO including the survey database (Messer 2001a). The survey recorded 55 sites of which one was a historic site and 40 were habitation sites. The percentage of habitation sites is 71.4 percent as shown in Table 3. Table

4 shows the distribution of site types for the survey. A modified version of the Mesa Verde Area

Survey Database (Messer 2001a) is in Appendix C.

Database Structure

Comparability across surveys was the main goal when developing the five databases used in this project. While it was understood that not all data would be available for all surveys or even

45 for all sites in each survey, it was decided to make all databases identical and include all columns in all databases. The list of 87 attributes can be found in Table 5. Most of the attributes are self explanatory, but others are peculiar to the Southwest or were chosen for specific reasons. The attributes fall into topical groupings.

The first set of five attributes are identifiers. CCSite# refers to the temporary field numbers assigned to sites before state numbers are allocated. State Site # refers to the numbers assigned to the sites by the SHPO. Sites in Montezuma County receive 5MT numbers, but sites within Mesa Verde National Park, even though within the county, receive 5MV numbers. The

Site Name is the common or map name of a site if one exists. Date Surveyed indicates the date the actual site was surveyed and Survey indicates which survey or portion of a multi-year survey in which the site was recorded. Also listed are U.S.G.S. Quad Map designating the 7.5' quadrangle map on which the site is located and the Map Date.

The next group of attributes is locational: the next five – County, Section, Quarter,

Township and Range – are the designators for the Federal Township and Range System of geographic location. The UTM designations of Zone, Easting and Northing are the requirements for location using the Universal Transverse Mercator coordinate system. Both types of mapping information, UTMs and Township and Range, were included because Colorado site recording requires them. Location and Location/Access indicates where the site is in relationship to current roads. Location/Access is only used if the site is not reachable directly, but must be accessed in a circuitous manner.

The next group of information is bookkeeping type data. Sketch Map indicates if a map of the site was drawn, Photo if a photo of the site exists, Hours Surveyed indicates how much

46 Table 5 Variables Used in Survey Databases

CC Site # Soil Description Midden Area State Site # Soil Formation Depth of Deposit Site Name Landform 1 Other Date Surveyed Landform 2 Boundary Description Survey Landform 3 Size Artifact Scatter U.S.G.S. Quad Map Nearest Water Name Shape Artifact Scatter Map Date Nearest Water Name Distance Artifacts Tallied County Nearest Permanent Water 1st Core Date Section Nearest Permanent Water Dist. 1st Extended Date Quarter Site Vegetation Basis of 1st Township Area Vegetation 2nd Core Date Range Site Type 1 2nd Extended Date Aerial Photo # Site Type 2 Basis of 2nd UTM Zone Site Type 3 Archaic UTM Easting Type Comment BM UTM Northing Structure Type PI Location Size of Rubble Scatter EP2 Location/Access Est. # Rooms MP2 Sketch Map Room Sizes LP2 Photo Towers EP3 Hours Surveyed Tower Diameters LP3 Government Land Pitstructures Historic Agency Pitstructure Type Unknown Land Owner Pitstructure Diameter Land Owner Address Pitstructure Depth Slope Feature Type 1 Exposure or Aspect Feature Size 1 Elevation Feature Type 2 Cultivated Feature Size 2 Bedrock Feature Type 3 Ground Visible Feature Size 3 Soil Depth Midden time it took to record the site. Government Land and Agency indicates first if the site is on federal land and, if so, the agency that owns the land – NPS, BLM, etc. If the land is privately owned then Land Owner and Land Owner Address are filled in. The next set of variables is

47 environmental. Slope indicates the percentage from flat the land slopes, Exposure or Aspect indicates either the direction the site faces or that the site is open on all sides. This is important because many sites are on quite steep hills with limited exposures. The Elevation is the height of the site location above sea level. Cultivated is whether or not the site sits on land that is or recently was cultivated to provide an indication of disturbance in historic times. Bedrock is the percent of the site covered in bedrock and Ground Visible is an indication of what portion of the ground was so covered in vegetation it was not surveyable. Soil Depth, Soil Description and Soil

Formation are all measures of the amount and type of soil at the site.

Environmental Information is recorded either in metric or British units depending on the dates of the surveys. When those variables are used in analysis they are converted to the metric system. The conversions were not made in the database to preserve the original information.

There are three Landform variables to accommodate a variety of information. Because each survey recorded this information in a different manner and some of the assigned values were difficult to define, the main operational landform choices are mesa, talus, canyon bottom and cliff which appear in Landform 1. Landform 2 and Landform 3 contain other landform information recorded in the surveys.

The location of water is important in the Mesa Verde Southwest although many surveys do not record where water is located. The databases include Nearest Water Name, Nearest Water

Distance, Nearest Permanent Water, Nearest Permanent Water Distance indicating the nearest water – seasonal or permanent, the distance to that water, the nearest permanent water and the distance to that water, respectively. The next two variables, Site Vegetation and Area Vegetation indicate the types of growth on the site specifically, such as sage or cheat grass and the type of

48 vegetation common in the area such as sage flats or pinyon juniper forest.

Following the environmental information is the site cultural information. There are three

Site Type designations set up so that Site Type 1 always contains either Habitation, Limited

Activity, Storage, Ceremonial or Unknown. Site Type 2 and Site Type 3 include more specific information such as water control feature or field house. In this way the database can also accommodate sites that fall into up to three categories. An open comment category follows the

Site Types for any unusual information.

Structure Type records the type of construction. This is diagnostically important in the area and can range from jacal to loose stones to dry set walls to dressed stone. The Size of Rubble

Scatter is an indication of the size of the structure. If rooms are visible then the Est. # Rooms and

Room Sizes will be recorded. If towers are located on the site than Towers will contain the number of towers and Tower Diameter will indicate their size. For pitstructures, Pitstructure records the number, Pitstructure Type whether a pithouse, kiva or something else, Pitstructure

Diameter and Pitstructure Depth provide the size of the pitstructure.

The database has room for three pairs of feature information, Feature Type and Feature

Size, because many sites easily have three recordable features, such as a burned hearth, grinding stones and upright slabs.

The next set of data describes the midden. Midden indicates if one was located, Midden

Area, the extent of the midden and Depth of Deposit, an estimate of the accumulation of artifacts and dirt in the midden. Other is another note field to record any cultural information that does not fit in the previous categories.

The overall site area is described by Site Boundary which indicates how the edges of the

49 site were determined. In most cases this is by artifact scatter, but in some areas it is defined at least on one side by cliff edge, or in the case of the MVAS 2000 survey, by the property boundary with the Ute Mountain Ute Reservation. The Size Artifact Scatter and Shape Artifact Scatter provides a description of the area encompassed by the site. Artifacts Tallied indicates if there were artifacts found and cataloged from the site.

The initial core dating – 1st Core Date, 1st Extended Date, Basis of 1st, 2nd Core Date, 2nd

Extended Date, Basis of 2nd, was derived by Michael Adler (1990) and provides a range of habitation time. The two sets of variables allow for multiple, separate habitation events for each site. The Basis category might be tree ring dates, architectural form or pottery style or any combination. The Core dates are narrow time periods and the Extended Dates are broader.

Adler’s Sand Canyon/Goodman Point survey is the only survey that contains this information and so for the purposes of this study, the actual dating periods are used. These time periods are

Archaic, BM – Basketmaker, P1 – Pueblo I, EP2 – Early Pueblo II, MP2 – Middle Pueblo II, LP2

– Late Pueblo II, EP3 – Early Pueblo III, LP3 – Late Pueblo III, Historic and Unknown.

However, not all surveys classified the Pueblo II period into thirds – early, middle and late. Some surveys simply use early and late PII. In those cases other information was used to determine if a middle PII designation was warranted.

Chronology and Its Problems

Determining time in an archaeological context includes both actual dating and relative dating, and both can be dependent on a variety of methods or artifacts. The first method to jump to mind is often radiocarbon dating for actual dates, but little of that was done in the Southwest in

50 Table 6 Pecos classifications (Hayes 1964:88, Cordell 1984:55, Lipe, 1999:65, Plog 1997:9)

Cultural Stage Description Early Basket Maker (postulated) pre-agriculture Basket Maker I (BMI) Now considered Archaic Basket Maker agriculture, atlatl, no pottery Basket Maker II (BMII) Late Basket Maker pottery, slab houses Basket Maker III (BMIII) Proto-Pueblo cranial deformation, true masonry, Pueblo I neck-banded pottery Pueblo II small villages, corrugated pottery Pueblo III large communities Proto historic corrugated pottery disappears, area Pueblo IV contracts Historic 1600 C.E. to present Pueblo V

the past because radiocarbon dating could only give an approximate date with an error range of plus or minus hundreds of years. Southwestern archaeologists had more precise methods of dating sites.

Tree ring dating in the Southwest is very advanced and has been ongoing since before A.

E. Douglass established the Laboratory of Tree Ring Research at the University of Arizona in

1937 (Lipe, et al. 1999). Tree ring dating is essential in the Ancestral Pueblo area, but is of little value in survey work because little undamaged, intact wood is found during survey.

Tree ring dating has, however, been the key in refining and securing in time the dating

51 efforts done throughout the Southwest in pottery analysis and seriation. Distinctive pottery, especially painted ceramics, is essential in dating sites (Breternitz 1966, Breternitz, et al. 1974,

Lucius and Breternitz 1992, Lucius 1992, Nordenskiöld 1893, Oppelt 1998) While broad placement as early or late Pueblo can be made from the utilitarian, everyday greywares, it is the black and white pottery in the area that stylistically narrows the time period (Wilson and Blinman

1991).

The effort to establish time periods in the Southwest began early. By August 1927, Alfred

V. Kidder, who had been excavating at Pecos Pueblo in New Mexico since 1915, invited archaeologists and other interested parties to Pecos to try to solve some of the problems existing in Southwestern archaeology at the time (Cordell 1984:55). Among those problems was one of nomenclature for cultural stages. This first Pecos Conference established a chronological ordering of the Southwestern cultures, called the Pecos Classification (see Table 6). Because tree ring dating was not tied to the calendar until 1929, only the latest portion of the classification, that occurring after European contact, could in any way be tied to the actual dates. However, relying heavily on pottery analysis and seriation, the Pecos Classification established an order and definition of the stages of culture from what is now considered the late Archaic, to the present.

The Pecos Classification continues to influence chronological dating in the Southwest.

Although the original Basket Maker I stage has shifted to an Archaic phase, the remainder of the classification scheme remains. However, many archaeologists alter the stages to fit their needs. As noted above, Hayes did not use the Pecos Classification in his Wetherill Mesa survey, but did include a table in his report to convert his phases, which were somewhat more refined, to the established classification (see Table 7).

52 In general, the cultural phases are now tied to actual dates through the use of tree ring dating and other absolute dating techniques. However, the fine details seem to be constantly shifting. In 1999, the Colorado Council of Professional Archaeologists published, Colorado

Table 7 Dates and phases used in the Wetherill Mesa Survey (Hayes 1964:88, Cordell 1972:32)

Phase Cultural Dates Other Pottery Types Stage C.E. Indications La Plata BMIII 574-700 Chapin Grey La Plata B/W Piedra PI 700-900 mesa top Chapin Grey 7000-7300 ft. Ackmen EPII 900-975 Mancos Grey Cortez B/W Mancos LPII 975-1050 terraces Mancos B/W retaining walls McElmo EPIII 1050- McElmo B/W 1150 Mesa Verde LPIII 1150- rectangular Mesa Verde 1300 kivas, caves B/W

Prehistory: A Context for the Southern Colorado River Basin (Lipe, et al. 1999). The dates for the various phases established in this book can be seen in Table 8.

Working with five surveys that took place over 50 years presents all of the dating strategies used at various times and places in the Southwest. A comparison of the dates used in the surveys can be seen in Table 9. Obviously, the later date assignments are more accurate than the earlier, but even now, beginning and end dates for cultural stages shift depending upon the archaeologist and, to some extent, the exact area within the region (Lipe, et al. 1999).

53 Table 8 Phases and Dates Currently in Use in the Mesa Verde Southwestern U.S. (Lipe 1999).

Cultural Stage Dates Pottery Types, Etc. Basketmaker II (BMII) 1000 B.C.E. – 500 C.E. no pottery, some maize Basketmaker III (BMIII) 500 – 750 C.E. early brown wares, plain gray, Chapin Gray, Chapin B/W Pueblo I (PI) 750 – 900 C.E. neck banded grayware Piedra B/W, Rosa B/W, Bluff B/R Pueblo II (PII) 900– 1150 C.E. Early Pueblo II (EPII) 900 – 1050 C.E. corrugated graywares, Cortez B/W, Deadman’s B/R Late Pueblo II (LPII) 1050 – 1150 C.E. corrugated graywares, Mancos B/W Pueblo III (PIII) 1150 – 1300 C.E. Early Pueblo III (EPIII) 1150 – 1225 C.E. corrugated grayware, mesa verde corrugated appears, McElmo B/W early, mesa verde B/W later, non-local redwares Late Pueblo III (LPIII) 1225 – 1300 C.E. Mesa Verde corrugated, Mesa Verde B/W

Conveniently, because of architectural attributes, changing the beginning or ending dates of a phase does not change data for most of this study, especially if we are looking at the broad designations of Basketmaker, Pueblo I, Pueblo II and Pueblo III. However, when looking at the finer grading of Early, Middle and Late Pueblo II or III, everything is not equal. The Pecos

54 Table 9 Comparison of dates across surveys. Gray areas have no entries for that survey.

Cultural Wetherill Chapin Sand Canyon Colorado SHPO Stage [Hayes, 1964:21] [Rohn, 1977:41] [Adler, 1990:4] [Lipe, 1999:235] BMIII 575-700 590 - 750 575 - 725 500 – 750 PI 700-900 750- 900 725 - 930 750 – 900 PII 900-1050 900 -1100 930 - 1150 900– 1150 EPII 900-975 900 - 1000 930 - 980 900 – 1050 MPII 980 - 1060 LPII 975-1050 1000 - 1100 1060 - 1150 1050 – 1150 PIII 1050 - 1300 1100 - 1300 1150 - 1300 1150 – 1300 EPIII 1050-1150 1100 - 1200 1150- 1225 1150 – 1225 LPIII 1150-1300 1200 - 1300 1225 - 1300 1225 – 1300

Classification used only entire phases. Hayes broke those phases further into smaller units. The general consensus today is that assignment to a cultural phase depends primarily on pottery types and secondarily there are definitively Early and Late Pueblo II and Early and Late Pueblo III divisions. Adler (1988, 1990, Adler and Varien 1991) added a Middle Pueblo II designation as well. This project will include a Middle Pueblo II designation whenever possible. Consideration will also be given to intermediate phases between Pueblo I and Pueblo II and between Pueblo II and Pueblo III. Recently, researchers attempting to create computer models of population histories in the Southwest have used a more finely divided chronology (Kohler, et al. 2007).

55 Normalization of Cultural Phase Lengths

Table 10 Phase lengths and multipliers for normalization of data

Cultural Phase Length Multiplier Basketmaker III 250 1.00 Pueblo I 150 1.66 Pueblo II 250 1.00 Early Pueblo II 125 2.00 Middle Pueblo II 75 3.33 Late Pueblo II 50 5.00 Pueblo III 150 1.66 Early Pueblo III 75 3.33 Late Pueblo III 75 3.33

The Pecos Conference, establishing cultural time periods without connection to actual calendrical dates, has led to a routine problem of comparing apples to oranges – two things that are unlike – in the Southwest. The number of habitation sites during Pueblo I is compared to the number of sites during Pueblo III. Occupation during Early Pueblo II is compared to that during

Late Pueblo II. Acceptable approaches in the early days when actual date ranges were unimportant and the cultural changes were the most important consideration. However, this simple comparison approach continued once the Pecos phases were tied to the calendar.

Unfortunately, cultural phase duration appears to vary. The length of time a group of people exhibits Pueblo I attributes is not equivalent in length to that of Pueblo II or even Early Pueblo II.

While the exact lengths of each cultural period are always under consideration and constant

56 revision, I normalized the data on the date ranges used during the Sand Canyon/Goodman Point

Survey (Adler 1988). The phase lengths and multipliers are in Table 10.

57 Chapter 3. SITE POPULATIONS

When we look at survey data, we see a palimpsest of what came before. Teasing out when each site existed is often difficult. As noted above, the cultural phases designated by the

Pecos Classification (Lipe, et al. 1999) are flexible. On a broad scale, the designations of

Basketmaker (BM), Pueblo I (PI), Pueblo II (PII) and Pueblo III (PIII) are the coarsest breakdown in the time periods under study. A finer breakdown of Basketmaker (BM), Pueblo I

(PI), Early Pueblo II (EPII), Middle Pueblo II (MPII), Late Pueblo II (LPII), Early Pueblo III

(EPIII) and Late Pueblo III (LPIII) is often employed by Southwestern archaeologists. I will first look at the broad time scales. While these phases are longer (see Table 8), the likelihood of sites being correctly dated to these periods is greater. The trade off between coarseness of timescale and accuracy of dating makes this an acceptable approach with survey data. I will then look at the finer phases and, while the potential for error is greater, knowing that more sites are likely to be misclassified, the narrow cultural periods still provide a useful way to look at changes of habitation in the area.

Of course, some sites at the end of one period or the beginning of the next will be wrongly classified because the transition from one cultural time period to the next is gradual.

Hopefully, no bias in one direction or the other exists and error introduced by misclassification becomes acceptable. I will assume that half the incorrectly dated sites are assigned earlier dates and half are assigned later dates.

To look at site population numbers I extracted all the habitation sites from each of the five survey databases. I eliminated any habitation sites dated as historic and any dated as Archaic or early Basketmaker. While all the databases recorded historic sites, not all the surveys recorded

58 Archaic sites. It is unknown whether those surveys not recording Archaic sites were truly devoid of Archaic habitation or whether the surveys simply did not look for or record pre-agricultural settlements. The number of habitation sites for each survey are in Table 3.

Determining what is a habitation site from surveys that someone else performed is not easy. In some surveys, very specific indications of habitation and type of habitation were recorded. In others, the indications of an architectural site were often there, but whether it was a habitation site, a field house, a storage site or a workshop of some kind were not always clear.

When creating the databases, I tried my best to tease out exactly what function the site had, using all available information including masonry information, pottery and sketch maps to classify habitation sites.

One area where there may be some confusion or errors is that of field houses. Some field houses are single room dwellings used only to tend the fields and never lived in year round.

Sometimes habitation sites later become field houses (Varien 1999a, Varien 1999b). It is very difficult to know if this occurs, especially in sites that show rehabitation. When an existing site later becomes a field house, only one or two rooms from that site would be reused on an occasional or seasonal basis. Many of the reused sites are small, two or three rooms or less to begin with.

For locations that show no sign of permanent habitation, but appear to be field houses, one accepted indication of a field house rather than a small habitation is seasonal occupations and scarcity of ceramics. However, because some of the surveys did not include pottery counts (or if they did they were not filed with the survey records or included in the reports) some structures may be misidentified.

59 A field house is defined as a small temporary habitation site separate from the permanent residence and used periodically and seasonally while tending the crops. The reuse of abandoned sites for this purpose undoubtedly occurred and I have no problem with a secondary site designation of field house. The problem is more complicated when sites are designated primarily as field houses. One indication of a primary field house designation is absence of pottery, but as noted above, not all surveys recorded pottery counts. Also, depending on the location of the site, surface pottery may well have been removed by those searching for mementos. Also, depending on the terrain, surface pottery may have washed down into an arroyo or be covered by silt.

Another supposed characteristic of primary field houses is that they are not associated with kivas. Unfortunately, locating a kiva with a small site, while expected from a purest point of view, is not always possible. Kiva depressions are filled in, grown over and otherwise masked from discovery by survey crews. Thus identification of a field house versus a farmstead becomes difficult.

Some archaeologists suggest that field houses occur only with later sites (Fish 2000,

Hayes 1981). Others (Kohler 1992, Wilcox 1978) see the existence of field houses as early as

Pueblo I. There have also long been discussions as to the purpose of field houses and whether field houses are really field houses at all, what their purpose is and how one interprets them – field house or small site (Kohler 1992, McAllister and Plog 1978, Vivian 1990, Gregory 1991,

Moore 1978, Wilcox 1978). Kohler (1992) suggests that field houses are a manifestation of village aggregation and statement of land ownership. Wilcox (1978) considers field houses an indication of shifting agriculture and agricultural pressure in a community. Both Kohler and

60 Wilcox consider field houses a manifestation of an aggregate village settlement pattern.

However, through much of time in the Mesa Verde southwest there were no aggregated sites.

Large villages simply did not exist. One difficulty with the idea of primary filed houses is that with a disaggregated community, the distance to the field was nowhere near that requiring overnight or week long stays. Wilcox (1978), although he believes that primary field houses existed, notes that morphologically there is little difference between a farmstead and a field house. There are cheerleaders on the side of most small sites being permanent habitation sites or

Figure 6 Graph of small site populations for five surveys in four time periods. Cultural phases are normalized so that periods are equal and data are comparable.

61 farmsteads (Gregory 1991, Vivian 1990). In the Chaco Canyon area, the significance of small sites has been discussed continually and both sides, for field houses and against field houses, still exist.

I tend to err toward the side of small sites as farmsteads rather than field houses because of the difficulty of actually identifying primary field houses. Single room structures were probably limited activity sites, so I classifying these as field houses or granaries. Two or three room buildings are very difficult to catagorize as field houses and so unless there was very good evidence that the structure was a field house, the site was classified as a habitation site.

Broad Time Scales – BM, PI, PII, PIII

One of the quickest and simplest ways to compare site populations during any specific time is to graph the surveys against each other by cultural phase. The time periods were normalized as indicated in Chapter 2 and the site population numbers are shown as a percent of the entire number of sites. Figure 6 shows the five surveys -- Chapin Mesa, Wetherill,

Mockingbird Mesa, Sand Canyon Goodman Point survey and Mesa Verde Area Survey 2000

(MVAS) – through the four time periods of Basketmaker, Pueblo I, Pueblo II and Pueblo III graphed together.

The first thing that can be seen is that three of the lines, those for Sand Canyon,

Mockingbird Mesa and MVAS, closely match each other in form and amplitude. The line for

Wetherill Mesa, while mirroring the form to some extent, has a larger amplitude. The line for

Chapin Mesa does not match the form or the amplitude of any of the other surveys. A quick glance at the graph does not suggest anything useful in analyzing the site populations. General

62 population trends in the Mesa Verde Southwest indicate a decrease or at least a halt in increase during the Pueblo I followed by increasing numbers of habitation sites until near 1300 C.E. at the end of the Pueblo III when populations decline and then are absent completely (Adler 1988,

Figure 7 Graph of small site populations for four surveys in four time periods. Chapin Mesa survey is excluded. Cultural phases are normalized so that periods are equal and data are comparable.

63 Adler 1990, Adler 1994, Adler 1996b, Cordell 1984). We will see that when using a finer time scale there are more dips and recoveries in the population line, but in general from Pueblo I until the end, gradual increases in number of sites per time period occur. The Chapin Mesa line increases dramatically during the Pueblo I, flattens at Pueblo II and drops dramatically at Pueblo

III. The possible reasons for this difference will be discussed later.

Throughout my recording and manipulation of the data from the Chapin Mesa survey I felt uncomfortable with the survey database. As noted in Chapter 2, the field notes were minimal

Figure 8 Graph of small site populations for five surveys and combined total database for four time periods. Cultural phases are normalized so that periods are equal and data are comparable.

64 Figure 9 Graph of small site populations for four surveys and combined total database for four time periods. Chapin Mesa Survey is excluded. Cultural phases are normalized so that periods are equal and data are comparable.

and the descriptions of habitation sites vague. Those were intuitive feelings about datasets and how they were collected. However, looking at the five graphs in Figure 6, the Chapin Mesa line does not appear to belong with the other lines.

Figure 7 shows what happens to the graphs if the Chapin Mesa line is removed. Now the graphs can be seen to cluster better and, while not a perfect fit, do show some similarity.

The benefit of having five large survey databases in exactly the same form is that they can be combined to form one composite database, as if the survey areas were contiguous and all part of the same overall survey. I have named this combined database Total. I graphed the Total

65 database, containing a compilation of all five databases, in the same way that I graphed the individual databases. Figure 8 shows the five databases with the Total database line in black.

What can be seen from this graph is that the distance from the Total line, which is an average, is very large for the Chapin Mesa line and moderately large for most of the other lines. The Total line is also rather flat showing very little change in slope through time.

When I created a Total without Chapin database and plotted it against the four survey databases that it includes, the picture becomes very different. Figure 9 shows the four surveys

Figure 10 Graph of small site populations for five surveys in seven time periods. Cultural phases are normalized so that periods are equal and data are comparable.

66 Figure 11 Graph of small site populations for four surveys in seven time periods. Chapin Mesa is excluded. Cultural phases are normalized so that periods are equal and data are comparable.

and the Total of those four surveys plotted through four time periods. As can be seen, the fit is much better. The distances from the Total line to the various survey lines are small and the lines are clustered. The total line also has a slight shift at Pueblo I and continues increasing until the end of PIII. Comparing Figures 8 and 9 It is easy to see that Figure 9 is a better fit.

One problem with using only four broad cultural phases is that fine points of settlement pattern become obscured. It is widely known that the Mesa Verde Southwest becomes completely empty of habitation at about 1300 C.E., but the graphs just shown do not indicate

67 this. There are also subtle site population changes that are not seen because of the broad time spans.

Narrow Time Scales – BM, PI, EPII, MPII, LPII, EPIII, LPIII

While using the seven commonly used cultural phases for the northern Southwest allows a finer look at site populations, it also allows more room for error. Determining if a site is Early Pueblo

III or Late Pueblo III, Middle Pueblo II or Late Pueblo II or even Late Pueblo II or Early Pueblo

III is more difficult than simply placing the sites in the broader time periods. The finer

Figure 12 Graph of small site populations for four surveys and combined total database for seven time periods. Chapin Mesa is excluded.

68 Figure 13 Graph of small site populations for five surveys and combined total database for seven time periods.

delineations of pottery styles tend to blur at the edges, especially for sites that do not have large amounts of surface pottery, making the proper designation fraught with difficulty. However, the decisions were made with as much accuracy as possible, especially for the two oldest surveys,

Wetherill Mesa and Chapin Mesa.

Figure 10 shows the graph of small site populations spread across the seven cultural

69 phases for the five surveys. Immediately it can be seen that there is much more variability in the settlement pattern than when looking only at four time periods. A finer grain of variability can be seen when modelers chose shorter, even intervals in which to assign sites (Kohler, et al. 2007,

Kohler and Leeuw 2007a), but this is much more difficult to do when working directly with existing surveys. Again, even with the finer grain of these time periods, the Chapin Mesa data look out of place. As seen in Figure 10, the only times when Chapin Mesa is like the other surveys is during the Basketmaker and Late Pueblo II. During the Early Pueblo II, Chapin data do mirror that of the Wetherill Mesa and Mockingbird Mesa databases, perhaps indicating that there is some mesa top versus lowland effect taking place. Even though there is a bit more convergence of Chapin Mesa with the other surveys during parts of this time span, it still appears that the Chapin Mesa data do not fit with the data from the other four surveys.

Figure 11 shows the site population graphs for the finer time scale without the data from

Chapin Mesa. As can be seen, the graph once again looks more compact and consistent among surveys. Also, once the Chapin Mesa data are removed, it does seem that there is a mesa top versus canyonlands effect with the Sand Canyon and MVAS surveys closely tracking and the

Wetherill and Mockingbird Mesa surveys tracking as well. The pattern does seem to converge during the Middle and Late Pueblo II and in the Late Pueblo III.

There are certainly environmental reasons why mesa top site distribution patterns would differ from those in the lower canyon lands and broad valleys. Elevation is the immediate consideration. That coupled with particularly dry times would place more sites in the upper

elevations where there is likely to be more precipitation. Mesa Verde in general in known to have far more rainfall and snowfall than the Montezuma Valley.

70 Once again I created a Total survey combining all five specific surveys and plotted the

Total survey against the other five datasets. Figure12 shows the Total survey line in black. Again it becomes obvious that the distance from the Total line is in some portions of this graph large.

The Total line flattens out during the Pueblo I and Early and Middle Pueblo II. As expected, there is good fit for the Total dataset in the Basketmaker and Late Pueblo II, but otherwise the graph seems to have no pattern.

Figure 13 shows the Total line for the surveys once the Chapin Mesa data have been removed. Again, the distances between the Total data line and the other surveys is now small and the lines appear fairly well behaved. Except for the Early Pueblo II, the graph lines follow the same path and indicate good fit between the datasets.

The Chapin Mesa Problem

As shown above, Chapin Mesa data appear anomalous when graphed with the other survey data. If all survey technique and data collection were equal, then one could say that there was some unusual phenomenon taking place on Chapin Mesa and that there was no problem with the data. However, because of the age of the data, the condition of the field cards and my concerns about the data while I was working with the Chapin Mesa database (Messer 2005a) it makes more sense to consider that the Chapin Mesa database is not an unusual phenomenon but that the Chapin Mesa database is basically flawed.

I looked at the standard deviation from the mean of the five surveys in the seven time periods used above. That is the standard deviation from the mean of the five data points for

Basketmaker sites, Pueblo I sites, Early Pueblo II sites etc., for all five surveys.

71 While most of the data in this attempt fall within one standard deviation of the mean, with some surveys having only one time period falling within two standard deviations from the mean, the Chapin Mesa data fall two standard deviations or more from the mean in four of the seven time periods. In my second attempt, I eliminated Chapin Mesa from the datasets and found that only two of the surveys had a time period that fell more than one standard deviation from the mean. Without the Chapin Mesa data, there is much less variability in the datasets.

It appears that the Chapin Mesa dataset, at least when it comes to numbers of sites and the time periods they are attributed to, is much less reliable than the other four surveys. While

Chapin Mesa will be used in some of the following analysis, caution in using this dataset for settlement pattern and land use studies is advised. If the basic site counts and date attribution are incorrect, then the remainder of the data, which partly relies on those variables, will also be incorrect.

72 Chapter 4. LANDFORMS

Table 11 Landform categories by survey. Gray areas have no entries in those surveys.

Sand Canyon MVAS Mockingbird Wetherill Chapin Standar Goodman Point 2000 Mesa Mesa Mesa d Mesa Upland flat area Upland slope Drainage head Upland knoll crest Upland slope side Upland Ridge top Upland slope base Tableland Top Mesa Mesa Talus Canyon bench Upland bench Canyon ledge Hill Canyon rim bench Hill slope Canyon Ridge Slope/bench Canyon Slope Talus Talus Canyon rim bench/terrace Canyon Arroyo Canyon Canyon Bottom Canyon floor Gully Bottom bottom bottom Cliff Alcove Cliff Cliff alcove alcove

Between the tops of the mesas and the bottoms of the canyons a myriad of places exist for ancestral pueblo sites. The large, aggregated Late Pueblo III villages in cliff alcoves at Mesa

Verde seem impossibly located. Listen a bit to the questions asked by visitors to the park and you hear, how did they get up to the mesa? How did they get down to the water? How did they keep their children from falling off the edge of the cliff? A more practiced visitor knows that climbing up and down the cliff side was an everyday occurrence, but why choose to live in such a

73 difficult place when you can live on the mesa top?

More puzzling to me are the sites built into the talus slope. Having worked on excavating more than a few of these, I wonder why anyone would choose those locations. Sometimes the slope is so steep that having one leg shorter than the other seems like a good idea. Even some mesa top sites seem misplaced. Why locate near the edge of the mesa when you can be in the middle? Then there are the few canyon bottom sites. Certainly it does not rain often in the Mesa

Verde area, but when it does, canyon bottoms flow like rivers and arroyos fill rapidly. Yet throughout the northern Southwest, habitation sites are located all over mesa tops, talus slopes, cliffs and canyon bottoms, although canyon bottoms do seem to be the least popular locations.

The landform of the area where a site is located can tell us something about the choices that ancestral pueblo peoples made when setting up a farmstead. An analysis of where the sites are located and how that changes through time can shed light on the settlement patterns and how they change.

Looking at landforms across five site surveys can be a difficult task. While every survey recorded the type of terrain where a site was located (Adler 1988, Fetterman and Honeycutt

1987, Hayes 1964, Messer 2001b, Rohn 1977), there is no uniform vocabulary to describe the actual terrain. I have therefore forced the varying terminology into one simple set of landform types. The variety of terminology used is shown in Table 11. Rohn (1977) clearly identifies those sites in the Chapin Mesa survey that are located in cliff alcoves, but other sites are less clearly sorted by landform and more often by rubble mound/architectural type. Notes on the site survey cards for Chapin Mesa were interpreted to fall into the categories of mesa, talus, cliff and canyon rim. There were no sites recorded as canyon bottom in this survey. The Wetherill Mesa survey

74 (Hayes 1964) uses the location designations of mesa, talus, cliff and canyon bottom. The

Mockingbird Mesa survey used (Fetterman and Honeycutt 1987) tableland, slope, alcove and canyon bottom. The original Sand Canyon Goodman Point database from Michael Adler used more than eight designations including upland flat area, upland slope, canyon floor, canyon bench, etc. The Penn State Mesa Verde Area Survey 2000 used tableland, ridge, hill slope, arroyo/gully, etc. Some of the variation occurs because the geography of the land surveyed differed. On the mesas there were flat areas on the mesa tops, but in the Sand Canyon Goodman

Point area, while there are fairly flat areas, they can not legitimately be called mesa tops. The

MVAS survey was perhaps the most difficult to characterize as it was done in a transitional area where there is no real talus slope and where finger ridges cut across portions of the area.

All of the flat upland areas were designated as mesa. All of the sloping areas where designated as talus and the areas identified as cliff, alcove or cliff bench were designated cliff.

Everyone seems to agree on where canyon bottom is and what is canyon bottom. The rationale behind these designations is that mesa signifies all of the flat or semi-flat areas: the places that appear, at least at first glance by a 21st century eye, as the most ideal locations to live, plant and harvest. Talus becomes the sloping areas. Places not impossible to live, but apparently less desirable because they pose obstacles to agriculture and living in general. These locations are occasionally very steeply pitched, often have more slick rock, shallower soil and more large rocks. However, not all talus slopes have shallow soils. Lookout House, part of Crow Canyons

Small Site Testing Project (Varien 1990) has a number of kivas built into an area of the slope with an incline of 30 percent. Also in that area is a subterranean room that reaches a depth of about 1.5 m beneath the modern ground surface at the time the room was built. Certainly that

75 depth of soil is sufficient for agriculture as well as building kivas and subterranean rooms.

Canyon bottom designates the lowest areas and the apparently least ideal locations. In all of the surveys, canyon bottom locations are the least prevalent.

Cliff locations are often considered fortified locations protected from direct assault by the difficult paths necessary for entrance or exit. Some cliff locations are very close to the mesa rim, while others are midway between canyon bottom and mesa top. Some cliff sites have water sources in the back of the cave-like alcoves and others would have required inhabitants to travel some distance for water. Some talus slope sites might also be considered fortified, many have enclosing walls and steep approach slopes. Lookout House has an extremely steep approach from beneath, as much as a 35 percent grade (Varien 1990).

Broad Time Scales

Figure 14 Site distribution by landform for Sand Canyon survey. Four cultural phases.

76 Figure 15 Site distribution by landform for Mockingbird Mesa survey. Four cultural phases

Figure 16 Site distribution by landform for MVAS survey. Four cultural phases.

77 Figure 17 Site distribution by landform for Wetherill Mesa survey. Four cultural phases

Figure 18 Site distribution by landform for Chapin Mesa survey. Four cultural phases.

78 When looking at the distribution of sites on landforms in the five surveys over four cultural phases, we see the data in Figures 14 through 18 have some similarities and some differences.

Wetherill Mesa, Chapin Mesa and Mockingbird Mesa, Figures 16-18, all show cliff site locations, while Sand Canyon and MVAS (Figures 14-15) do not because there are no true cliff alcoves in these survey areas. For the most part, the majority of sites during the earliest three time periods are designated mesa. During the Pueblo III, all surveys show a decrease in mesa locations, but some much more than others. In the Sand Canyon/Goodman Point survey (Figure 14), the percentage of sites in flat setting remains above 90 percent throughout the four cultural phases with a slight shift to talus slope during Pueblo III. In the MVAS survey (Figure 15), there is somewhat more movement to talus slopes through time, but the majority of sites are still on the flattest land around. The three mesa surveys, Wetherill Mesa, Mockingbird Mesa and Chapin

Mesa, show early mesa occupation, but in the case of Mockingbird and Wetherill Mesas (Figures

16-17), there is more of a diversity of landform use early on. Both these surveys show

Basketmaker use of cliff sites with Wetherill showing nearly half the Basketmaker sites in cliff locations and Mockingbird showing only about 10 percent Basketmaker cliff sites. Mockingbird

Mesa survey does, however, show nearly 30 percent of the Basketmaker sites on the talus, while

Wetherill shows almost none in that location. While both Mockingbird and Wetherill are classified as mesa areas, all mesas are not the same and the absence of cliff sites in both the

Basketmaker and Pueblo III suggest that perhaps cliff locations were not as available on

Mockingbird. Surprisingly, Chapin Mesa (Figure 18) recorded no Basketmaker sites in cliff locations, but then that survey only recorded cliff locations during Pueblo III. It is only on the

Mesa Verde surveys at Wetherill and Chapin that we find fewer than 50 percent of the sites for

79 any period on the mesas and then only for the Pueblo III. While the Wetherill survey shows a small canyon bottom occupation and decent talus slope occupation in the later periods, the

Chapin Mesa survey records no canyon bottom sites at all and few talus sites through all time periods. In general, this broad look at the various time periods show a gradual shift from mesa top to talus and cliff locations, but to less an extent than the aggregated large sites on Wetherill and Chapin mesas would suggest. Only the mesa sites show major movement from the mesa.

Again, the Chapin Mesa data are unusual and will be discussed separately.

Narrow Time Scales

Figure 19 Site distribution by landform for Sand Canyon survey. Seven cultural phases.

80 Figure 20 Site distribution by landform for MVAS survey. Seven cultural phases.

Figure 21. Site distribution by landform for Mockingbird Mesa survey. Seven cultural phases.

81 Figure 22. Site distribution by landform for Wetherill Mesa survey. Seven cultural phases.

Figure 23. Site distribution by landform for Chapin Mesa survey. Seven cultural phases.

82 Looking at the landform distribution of the five surveys over seven cultural phases the placement of sites on the landscape does not change very much. The Sand Canyon survey data

(Figure 19) show a very gradual decline in mesa sites through time of about 20 percent with a gradual increase in talus slope sites of the same 20 percent. Gradually, through time, about 20 percent of the population moved from the flat areas to the sloping areas. The MVAS survey

(Figure 20) Shows a similar pattern with a few more canyon bottom sites listed. The one exception is in the Early Pueblo II when half the sites are on the mesa and half are on sloping areas. Again, the MVAS survey was conducted on unusual terrain where there are no true canyon bottoms, but arroyos cut the entire area and where there are rippled ridges coming out from beneath the actual talus slope. More sites are on sloping areas because by percentage, more of the land is actually sloping.

The Mockingbird Mesa survey (Figure 21) on a finer time scale shows a gradual decline in mesa sites after the Basketmaker and Pueblo I times with a matching rise in talus sites.

Mockingbird also has fairly constant, but small numbers of sites located in cliff areas. This cliff presence is smaller, later than that of the Basketmaker and Pueblo I. Wetherill Mesa (Figure 22) also shows a gradual decline in mesa sites from the Pueblo I on, with a rise in talus slope sites until the Late Pueblo III. On Wetherill Mesa, cliff sites are high in the Basketmaker period and decrease dramatically in the Pueblo I. They increase gradually from then until they jump to the maximum of more than 70 percent in the Late Pueblo III. The Chapin Mesa survey (Figure 22) shows constant decline in mesa sites with initial increase in talus slope sites until the Early Pueblo

III when talus sites peak and then dip during the Late Pueblo III. Cliff sites suddenly jump to

83 nearly 90 percent during the Late Pueblo III and this is the only time that the Chapin survey records cliff sites. Using the finer time scales in this graph allow us to see that the cliff locations occur only during the Late Pueblo III. At the same time, mesa top sites dwindle to only a few percent of all the sites. Again, the Chapin Mesa data are unusual and will be discussed separately.

Comparison of Landforms

Figure 24 Comparison of mesa distribution of five surveys in seven cultural phases. Only the percentage of mesa sites is shown for each survey.

Figure 24 shows the distribution through time of sites on mesa tops. It is obvious that the mesa tops were the most desirable areas. Fifty percent or more of the sites in any one period are on the mesa except at Wetherill Mesa during the Pueblo III and Chapin Mesa during the Late

84 Figure 25 Comparison of talus distributions of five surveys in seven cultural phases. Only the percentage of talus sites is shown for each survey.

Pueblo III. The dip in flat sites during the Early Pueblo II in the MVAS survey is unusual and unaccountable. Why half the sites are on the talus and half on the mesa during this period is unknown. However, the unusual, transitional topography in the area may account for this.

Looking at the talus sites in Figure 25 we see that Chapin and Wetherill Mesa surveys have nearly identical form. Mockingbird Mesa survey shows much more Basketmaker activity on the talus slope than any of the other surveys and the Sand Canyon survey shows no sites at all during Pueblo I. The MVAS survey shows a spike during the Early Pueblo II that does not appear anywhere else. The talus slope site distribution is less uniform than the mesa top distribution, but in general there tend to be more use of the talus slope the later the time period

85 except at the very end during Late Pueblo III when some talus use declines in some areas.

Figure 26 Comparison of cliff distributions of five surveys in seven cultural phases. Only the percentage of cliff sites is shown for each survey.

The cliff sites shown in Figure 26 represent only three of the surveys as the other two were not mesa surveys and had no real cliff or alcove locations to record, although some of the

Sand Canyon talus slope sites with slopes as great as 35 degrees were probably quite difficult to access in at least one direction. Immediately it is apparent that something is wrong with the

Chapin Mesa survey data in this landform. While the other two surveys show some cliff activity throughout the time covered, the Chapin Mesa Survey shows only cliff occupation during the

Late Pueblo III. Seeing any similarity when looking at the two surveys is difficult. The Wetherill data show high activity during the Basketmaker and very high activity during Late Pueblo III with precipitous decline during Pueblo I and a gradual increase to Early Pueblo III. Mockingbird

86 shows more activity early and fairly flat activity during Pueblo II and III. Occupation of cliff areas is obviously not straightforward, judging by the difference between these two surveys.

The Chapin Mesa Problem

The landform data again show that the Chapin Mesa survey is unlike the other surveys.

Chapin and Wetherill Mesas are both on Mesa Verde and are close enough that their occupation patterns should be very similar. There is today easy movement from one mesa to the other as

Chapin Mesa makes up the main portion of the Mesa Verde National Park and Wetherill Mesa is the other area that is open to the public. The Chapin data look too uniform, almost contrived. I do not think that Lancaster purposefully skewed his data collection or that Rohn misinterpreted the data (Rohn 1977). What I am suspicious of is that those who carried out the Chapin Mesa survey had a preconceived idea of where sites were and that this bias heavily influenced where sites were found. If no one looks in the canyon bottom because everyone knows that there are no sites in the canyon bottom, then no sites will be found there. In this case, because the exact extent of the survey is unknown, it is possible that the canyon bottoms were not surveyed, explaining the absence of canyon bottom sites in the Chapin Mesa survey. Because there was a general consensus for a long time that sites were not prevalent on the talus slopes, this may have affected the survey of that zone. However, the one truly problematic area of the landform data for Chapin Mesa is cliff sites. The data for Chapin Mesa may be a self-fulfilling prophecy for the

Late Pueblo III. The Chapin Mesa survey data seems to prove the idea that nearly everyone moved from the mesa tops to the cliffs during the last period of Ancestral Pueblo occupation.

Perhaps the assumption that this was the case created the pattern, rather than that the pattern

87 supported the case. Wetherill had occupation throughout the time periods in the cliffs, and still had habitation on the mesa during the Late Pueblo III, although vastly reduced to about 20 percent.

A P2 test of the progression of landform choices through time will show if the changes were the product of chance, or if the inhabitants were in some way changing their preferences for specific locations of their small farmsteads. When looking at the landform data, some time periods and some categories do not meet the requirements of a P2 test – at least 80 percent of the categories with 5 items. In some cases the canyon bottom category is very small and so is collapsed into the talus slope category leaving only two categories for the non-mesa surveys. The

Wetherill survey has four categories.

Looking at the Mesa Verde Area Survey 2000, the only time periods that are valid for a

P2 test are the Pueblo II to Pueblo III grouping. That transition is not significantly different. In the Sand Canyon/Goodman Point survey, none of the transitions before Pueblo II are valid for a

P2 test. The Pueblo II to Pueblo III transition shows significant difference with a P2 value of

5.084 (df=1, p=0.0241). However, the transition from Late Pueblo II to Early Pueblo III and the transition from Early Pueblo III to Late Pueblo III are not significantly different at "=0.05. This would suggest that any changes that were being made were gradual.

The Mockingbird Mesa survey data allow P2 testing of a variety of transitions from

Basketmaker to Pueblo I, Pueblo I to Pueblo II, Pueblo II to Pueblo III and Early Pueblo III to

Late Pueblo III. All of the transitions show no significant difference except for the transition from Pueblo II to Pueblo III which shows a P2 value of 9.458 (df=1, p=0.002). On Wetherill

Mesa, again the early periods, Basketmaker to Pueblo I, Pueblo I to Pueblo II are not valid for a

88 P2 test. The transition from Pueblo II to Pueblo III does show significant difference with a P2 value of 201.02 (df=3, p=0.0001). On Wetherill Mesa, the transition from Early Pueblo II to

Middle Pueblo II is not significantly different, but the transitions from Middle Pueblo II to Late

Pueblo II, Late Pueblo II to Early Pueblo III and Early Pueblo III to Late Pueblo III are all significantly different with P2 values of 7.987 (df=2, p=.018), 10.659 (df=3, p=0.014) and 36.723

(df=3, p=<0.0001).

As a check on whether the Chapin Mesa data in any way resembles the Wetherill Mesa data, P2 tests and Fisher’s Exact test were run where possible and appropriate comparing the landforms in each time period. The Basketmaker and Late Pueblo III periods were not suitable for statistical testing, however, all the other time periods were suitable. A problem does arise because while both surveys have entries for mesa and talus, the Chapin Mesa survey has no canyon bottom sites and no cliff sites until the Pueblo III while the Wetherill Mesa data has both canyon bottom and cliff sites through most of the time periods. I have therefore compared only the mesa and talus records between the two surveys and then compared the mesa and combined non-mesa tallies.

During the Pueblo I, no comparison could be made of the mesa and talus sites because there are insufficient talus sites in the Wetherill Mesa data, however a mesa versus non-mesa evaluation was not significantly different. The total Pueblo II was significantly different with a P2 value of 30.013 (df=1, p<0.0001) and a similar finding for the Fisher Exact test. Considering mesa versus non-mesa, the results were the same with a higher P2 value of 70.138 (df=1, p<0.0001). The breakdown of the Pueblo II data showed all three categories were significantly different.

89 For the Pueblo III, the P2 value of 5.261 (df=2, p= 0.072) is significantly different. For the Early Pueblo III the P2 value of 6.091 (df=1, p=0.014) is significantly different. Looking at the mesa versus non-mesa in this time period, we have a P2 value of 24.626 (df=1, p<0.0001) and the datasets are significantly different.

90 Chapter 5. ELEVATION AND CLIMATE

In the southwestern U.S. elevation, precipitation and temperature play important parts in agricultural productivity. Some places are normally colder than others, with elevation playing a large part in such variation. Other places are dryer with some elevations receiving more rainfall than locations nearby. Insufficient frost-free days in many places would doom any prospect of reliable crops. Yet all is not always as it seems in the Mesa Verde area. Wetherill Mesa and the rest of the Mesa Verde mesas, while the highest areas, receive more rainfall and are better suited for dry land agriculture than any location in the region. Canyons anywhere in the area become cold sinks sucking down cold air and causing frosts when nowhere else is hoary, so that elevation alone does not dictate temperatures. The result is a complex mosaic of land that is either good or bad for agriculture with some land on the cusp, one year receiving sufficient moisture and warmth and the next year stunting plant growth with an early freeze or parching unripened corn through an absence of precipitation. To grow maize – the primary crop from Basketmaker to

Pueblo III – an area needs sufficient precipitation, temperatures and frost-free days. All three variables are elevation related.

Precipitation in the area is biseasonal with precipitation in the winter falling mostly as snow and lesser rains falling in the summer. Estimations based on actual precipitation readings for May to September show that, in general, areas above 2134 m (7000 ft) receive 9.6 in. of rain, those lying between 1829 m (6000 ft) and 2134 m (7000 ft) average 6.2 in. during that time.

Nearly all the land under consideration in this study falls in these two precipitation belts except for small portions of Wetherill Mesa and Mockingbird Mesa, which fall in the band receiving only 4.5 in. of summer rain (Adams and Peterson 1999) (see Figure 27).

91 Figure 27 Comparison of elevation ranges and means across the Wetherill Mesa, Mockingbird Mesa, Mesa Verde 2000 and Sand Canyon/Goodman Point surveys.

Annual rainfall for the various survey areas ranges from Wetherill Mesa which receives

18.4 in. of rain at its mean elevation, to Mockingbird Mesa that receives 10.1 in. of annual rainfall at its mean elevation. Sand Canyon/Goodman Point survey area receives 16.0 in. of annual rainfall at its mean elevation and the Mesa Verde Survey 2000 receives 14.7 in. at its mean elevation. While there is some disagreement about exactly how much annual rainfall is necessary for maize growth on the Colorado Plateau and whether annual or summer rainfall is the better indicator of agricultural success, everyone would agree that the rainfall in these four

92 areas is sufficient for cultivating maize (Adams and Peterson 1999) (see Figure 27).

While elevation influences rainfall, it also influences the temperatures in the area. And although in general the higher the elevation, the higher the rainfall, the converse applies to temperatures. To grow maize, an area needs a minimum of 2,500 Corn Growing Degree Days

(CGDD), a measure of temperature in an area. CGDD units are the difference between daily average temperature and the base temperature of 60. A unit accrues for each degree Fahrenheit the daily average is above the base temperature. So, if the daily average is 70, that equals 10

CGDD units. Researchers measure CGDD during the growing period from May to September.

The growth range for maize is between 50 and 86 degrees Fahrenheit so numbers below 50 and above 86 are considered 50 and 86, respectively. Average temperatures below the base temperature are ignored because they do not add to the growth of the maize, so negative units are ignored (Adams and Peterson 1999). Most areas in the study have almost sufficient corn growing degree days for maize production, but would probably require the added tilt or south facing slope to ensure good harvests. However, even with its elevation, Mesa Verde (Wetherill

Mesa) has close to the necessary corn growing degree days and with its abundant water does appear to be the best location to grow maize.

The last elevation-related parameter for growing maize is number of frost-free days. The assumption is that maize needs 120 days above 28.5 degrees Fahrenheit for growth and maturation, although maize can actually survive temperatures as low as 26.5 degrees Fahrenheit.

Even lower temperatures at some points in the growing cycle do minimal harm. There are maize varieties that need only 60 to 90 days to mature (Salzar 2000), but using the 120 day

93 requirement, the climate data show that above 2134 m, there are only 104 frost-free days in this area. Between 2134 m and 1829 m, where most of the survey’s elevations fall, there are only

117.6 days, a marginal number to ensure good harvests. From 1829 m to 1524 m, there is an average of 155 frost-free days. While it looks as if maize agriculture would be a very dicey endeavor in this area with sufficient frost-free days occurring only in areas with insufficient precipitation, the exceptions seem to prove otherwise. On Mesa Verde, coupled with the good rainfall and corn degree growing days, there are 153 frost-free days. In fact, Mesa Verde has the least number of cold days, on average than any location in the area (Salzar 2000). The other surveys have about 120 frost-free days and would mature crops about 90 percent of the time.

Figure 28 Elevation distribution of sites through time in Mesa Verde 2000 survey

94 Figure 29 Elevation distribution through time of Wetherill Mesa survey.

Broad Time Scales

Figures 28-31 show changes in elevation through time for the Mesa Verde 2000,

Wetherill Mesa, Sand Canyon/Goodman Point and Mockingbird Mesa surveys, respectively.

Chapin Mesa will not be considered. What is notable is that while usually the range of elevations increases from Basketmaker through Pueblo III, the means, in all cases, remain very much the same. Not surprisingly, the Mesa Verde 2000 survey (Figure 28) has one of the narrowest elevation ranges. This survey has the smallest area and is confined, by definition, to a transitional location between the bottom of the talus slope and the valley. While the range of elevations goes from 1899 to 1981 m, a change in elevation of 82 m, the range of elevation means goes from

1921 to 1932 m, a change of only 11 m. While these elevations are slightly below the optimum

95 for corn, this area lies in the shelter of the mesa. Through time, the elevation ranges increase, however the majority of the increase in this area is toward higher elevations, undoubtedly because the lowest sites even during the Basketmaker period are already on the valley floor and there is no where within the area to go lower.

Wetherill Mesa has the largest variation in elevation and encompasses a large survey area.

The inhabited elevations on Wetherill Mesa rise from 1798 to 2463 m with a range of 665 m.

This is the largest range of elevations among the surveys. The mean elevations through time,

Figure 30 Elevation distribution through time of Sand Canyon/Goodman Point survey.

however, show a much smaller variation as shown in Figure 29, although this is also the largest variation in the means among the surveys. The means through time for Wetherill Mesa go from

2141 to 2196 m or a span of 55 m. Figure 27 shows that all of the elevation ranges of the other

96 surveys fall within the minimum and maximum elevations of Wetherill Mesa habitation sites, however, all ranges also fall predominantly below the mean of the Wetherill Mesa site elevations.

Wetherill Mesa has the most variation in inhabited elevations, the highest elevations and perhaps the best locations for growing maize.

Figure 31 Elevation distribution through time of Mockingbird Mesa survey.

Wetherill Mesa habitation sites during Basketmaker times had a broader range than the

Basketmaker sites found in the other surveys in this study, and those sites are found at a much higher elevation than the other surveys. Wetherill Mesa also had the largest number of

Basketmaker sites in cliff alcoves with more than 50 percent of the sites located in sheltered

97 areas. Through time, the elevations change in both directions, but the elevations seem to be limited in the higher elevations showing a variation of only 14 m while the variation in the lower elevations ranges through 248 m. This suggests that while it was not physically possible to locate habitation sites higher on the mesa, there was a sizable reserve of land lower on the mesa that was used during later periods.

The Sand Canyon/Goodman Point survey has habitation sites from elevations of 1997 to

2164 m (see Figure 30), with a range of 167 m. The elevation means rise from 2069 to 2094 m with the difference in the mean elevations through time is just 25 m. The upper limits of elevations stay consistent in the Sand Canyon/Goodman Point survey area with the changes in elevation moving downward, but only slightly. Undoubtedly, the highest sites were already on top of the McElmo Dome with no where higher available.

Mockingbird Mesa is the most uniform in elevation because the mesa itself is relatively flat. The elevations on Mockingbird Mesa (see Figure 31) go from 1800 to 1950 m with a range of 150 m. The range of elevation means is from 1891 to 1899 m with a change of only 8 m toward lower elevations because there is little higher land on the mesa. It is interesting to note that in all but the Mockingbird Mesa Survey, the total elevation range ever used closely matches the range of elevations occupied during the Pueblo II, and not the latest period, Pueblo III. In the

Mockingbird Mesa Survey area, however, the closest match to the total range is during Pueblo

III. As with the Sand Canyon/Goodman Point data, through time, elevation ranges broaden the changes in elevation range moving downward, but only slightly. This is because the sites are already at the highest possible locations.

98 Narrow Time Scales

Figure 32 Elevation distribution through time of Mesa Verde 2000 survey using narrow time cultural phases.

Looking at the finer grained, narrow cultural phases, Figures 32-35 show the elevations used during seven cultural phases for the Mesa Verde 2000, Wetherill Mesa, Sand

Canyon/Goodman Point and Mockingbird Mesa surveys, respectively. The Mesa Verde 2000 survey (Figure 32) shows the most difference from the broader cultural periods showing a definite broadening of elevation range beginning in the Late Pueblo II. Before this time, the elevation ranges were fairly uniform. As with the broader time periods, the mean elevations remain uniform.

99 Figure 33 Elevation distribution through time of Wetherill Mesa using narrow cultural time periods.

For the narrow cultural time periods on Wetherill Mesa (Figure 33), elevations are fairly uniform beginning in the Middle Pueblo II period. Early Pueblo II is only slightly different with the range ending at a higher elevation, a little above the lowest elevation. The Basketmaker and

Pueblo I ranges exhibit narrower ranges than the later periods. The elevation means on Wetherill

Mesa show little movement.

In the Sand Canyon/Goodman Point locality (Figure 34) there is little difference in elevation range during the later time periods. The period that matches the total range is the Late

100 Figure 34 Elevation distribution through time of Sand Canyon/Goodman Point using narrow cultural time periods.

Pueblo II. The elevation mean shows a slight tendency to be lower during the later time periods, but with a difference overall of only 28 m, it is not much lower.

At Mockingbird Mesa ( Figure 35), the elevation distributions look somewhat different from those in the other three surveys. The range of locations for Middle Pueblo III looks narrow and shrunken, but this is probably an artifact of the way the survey was done and suggests that insufficient sites were assigned to the Middle Pueblo II period. If we ignore this anomaly, then

101 Figure 35 Elevation distribution through time of Mockingbird Mesa using narrow cultural time periods.

we see that the Late Pueblo II is not the broadest range, Instead, both Pueblo III time periods show the broadest range and match the total range occupied through time.

It appears that in all but the Mockingbird Mesa survey the ranges contract slightly after the Late Pueblo II. Whether this is due to a climate change or some other variable is not known from this data.

Looking at the data from a statistical point of view, Student T tests should show if there

102 Table 12 Student T test of Pueblo I elevations vs. Pueblo II elevations Survey Site Count Site Count p Significance PI PII

Wetherill Mesa 132 348 .01 N.S.

Mockingbird Mesa 19 75 .54 N.S.

MVAS 2000 7 21 .42 N.S.

Sand Canyon/Goodman Point 26 157 .07 N.S. is any change in the site elevations through time. The change from Pueblo I to Pueblo II and then

Pueblo II to Pueblo III and tests of the change from Late Pueblo II to Early Pueblo III and from

Early Pueblo III to Late Pueblo III were completed. In Table 12 none of the surveys show significant difference in site elevations from Pueblo I to Pueblo II. However, The MVAS 2000 survey has far too few sites to successfully do statistical analysis and the Sand Canyon/Goodman

Point survey has slightly fewer sites than required during Pueblo I for analysis.

Table 13 Student T test of Pueblo II elevations vs. Pueblo III elevations Survey Site Count Site Count p Significance PII PIII

Wetherill Mesa 348 199 .22 N.S.

Mockingbird Mesa 75 65 .84 N.S.

MVAS 2000 21 29 .45 N.S.

Sand Canyon/Goodman Point 157 122 .02 S

Table 13 shows the results of two-tailed Student T tests with " = .05. Again the data for

MVAS 2000 have both site counts below n = 30, although the Pueblo III count is 29. The other three surveys have large enough sample to complete a test of significance. The elevation changes for Wetherill Mesa and Mockingbird Mesa are not significant, however, the changes in elevations

103 from Pueblo II to Pueblo III in the Sand Canyon/Goodman Point locality are significant.

Table 14 Student T Test of Early Pueblo III Elevations vs. Late Pueblo III Elevations Survey Site Count EPIII Site Count EPIII D Significance

Wetherill Mesa 878 199 .01 S

Mockingbird Mesa 47 49 .46 N.S.

MVAS 2000 16 13 .98 N.S.

Looking at the differences in elevation between Early and Late Pueblo III, Table 14 shows that once again the MVAS 2000 survey has too few sites for statistical analysis. Both the

Mockingbird Mesa and Sand Canyon/Goodman Point Survey data show no significant difference in the elevations of the Early Pueblo III and Late Pueblo III sites. Wetherill Mesa, however, shows a significant difference between Early and Late Pueblo III, which may relate to the large increase in cliff locations during the Late Pueblo III. The increase in cliff sites on Mockingbird

Mesa was less pronounced.

If only the Pueblo III sites that are not located on cliff are considered, then there are 110 locations during Pueblo III that are not on cliffs and 64 and 60 non-cliff sites during Early Pueblo

III and Late Pueblo III, respectively. A two-tailed Student T test of the non-cliff Early Pueblo III sites vs. the Late Pueblo III sites shows no significant difference with a p = .06. This suggests that the significant difference in the change in elevations on Wetherill Mesa between Early and

Late Pueblo III are due to the cliff locations and that other locations where chosen in the same way as they had been.

104 Temperature and Elevation

Considerations when looking at elevations should include the environmental variables that influence local growing situations. The ideal way to look at temperature and precipitation would be on a local level because of the patchwork of microclimates that tile the area.

Unfortunately, temperature and precipitation reconstructions for individual locations in the Mesa

Verde area are not available. General Palmer Drought Severity Index information is not available before 1200 C.E. (Comrie 2008, Cook 2000). The only available data covering the years between 625 and 1300 C.E. are average temperature and precipitation for the entire area (Salzer and Kipfmueller 2005). This data is compiled from tree ring data.

Richard Alley, a climatologist, suggests that PDSI and precipitation data are probably more accurately reconstructed from tree ring data than temperature (Alley 2008). Because PDSI is not available, temperature and precipitation are the only climate variables used here. Annual temperature and precipitation graphs, while interesting, do not show trends well. Temperature graphs for annual, 5-year and 30-year temperatures are shown in Figure 36. Also shown are graphs of the temperature averaged over both the broad and narrow cultural time periods.

Precipitation graphs for annual, 5-year and 30-year temperatures are shown in Figure 37, as are precipitation averaged over both the broad and narrow cultural time periods.. Although the elevation graphs for each of the four surveys discussed in this chapter were compared to the annual temperature and precipitation graphs, those comparisons are not shown because it is impossible to see trends in the annual temperature graphs. The comparisons with 5 and 30-year averaged temperature and precipitation graphs will be shown.

105 Figure 36 Comparison of average temperatures in the Southern Colorado Plateau. (A) annual temperatures, (B) five-year averages, (C) 30-year averages, (D) averages based on broad cultural periods, (E) averages based on narrow cultural periods. (Salzer 2005)

106 Figure 37 Comparison of average precipitation in the Southern Colorado Plateau. (A) annual precipitation, (B) 5-year averages, (C) 30-year averages, (D) averages based on broad cultural periods, (E) averages based on narrow cultural periods (Salzer 2005).

107 It should be noted in these comparisons that the temperature and precipitation data have much better time resolution than any of the surveys being used. The temperature and precipitation data are on yearly, 5 or 30-year intervals, while the Ancestral Pueblo cultural phases are from 50 to 250 years in length. The inability to tie sites to narrower time scales probably obscures any affect of temperature and precipitation. An attempt to equalize time resolutions can be seen in D and E in Figures 36 and 37. Both temperature and precipitation were averaged in the same time intervals as the cultural phases and plotted at the center point of each phase. For precipitation, this serves to totally obliterate any variation for both broad and narrow time scales.

While the temperature averages over broad and narrow time scales show a little more variation, the differences are minuscule. None of the averages over cultural time periods will be compared to elevation and population.

Temperature Analysis

Broad Time Scales

Figure 36 shows the annual, 5-year and 30-year averaged temperatures for the overall area (Salzer and Kipfmueller 2005). Because low temperatures are of primary concern, the periods of unusual high temperature will not be discussed. In the 5-year averaged temperature data there appear to be three periods of cooler temperatures, one just before 700, one in the mid-

800s and one beginning in about 1195 and ending in about 1245, a period of 50 years of below average temperatures. The 30 year averages show these three cooler periods and show the persistence of a fourth at about 1125 years. That cool period is narrower and smaller than the others and is probably an artifact of one very cool year in 1120 with an average temperature of

108 13.7 degrees Celsius. Figure 38 shows a comparison of elevation, population, 5-year and 30-year average temperatures for Wetherill Mesa for the broad cultural areas. While the first two cool periods do not appear to influence either elevation location or population, the last cool period does appear to correspond to a leveling off in elevation range and a slight decrease in site populations. However, the population line begins to decrease long before the low temperatures.

Also, the elevation had already reached the limits of either available elevation or ability to grow crops. There is no correlation between temperature and elevation. What is evident is a correlation between site population increase and the broadening of the elevation range.

On Mockingbird Mesa (Figure 41), there appears to be no correlation between temperature and elevation as the elevation range continues to expand through all occupation periods after Pueblo I. There does however appear to be a correlation between site population increase and elevation range increase.

Figure 42 shows the elevation, population and temperature data for the Mesa Verde Area

2000 survey. While the elevation range does contract a bit during the later period, corresponding to the decrease in temperature, the population continues to increase through time. In the Sand

Canyon/Goodman Point locality (Figure 41), the elevation ranges do contract slightly, but again, the site population continues to increase regardless of lower than average temperatures.

109 Figure 38 Elevation, population and temperature comparison for Wetherill Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

110 Figure 39 Elevation, population and temperature comparison for Mockingbird Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

111 Figure 40 Elevation, population and temperature comparison for Mesa Verde Area survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

112 Figure 41 Elevation, population and temperature comparison for Sand Canyon/Goodman Point survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

113 Narrow Time Scales.

Looking at the narrower data may show closer relationships between the three variables of elevation, population and temperature. For the Wetherill Mesa data (Figure 42), there again seems to be no relationship between temperature and elevation ranges. There also seems to be no relationship between temperature and population. With respect to elevation range and population, the elevation range does expand when the population begins to increase, however, with the population spiking and falling and increasing again, it is difficult to see any relationship in in the narrow data.

Mockingbird Mesa data for the narrow time spans (Figure 43) show no relationship between elevation and temperature, but does show a slight relationship between elevation range and population. There is an increase in range at the same time the population begins to increase.

For both the Mesa Verde 2000 (Figure 44) and Sand Canyon/Goodman Point surveys (Figure

45) the relationships are the same as in the previous cases. There is no relationship between temperature and elevation range, but a slight relationship between elevation range and population. For the Mesa Verde Area 2000 survey the highest site population corresponds to the widest elevation range. For the Sand Canyon/Goodman Point survey, the site population peak also corresponds to the broadest elevation range.

It should be noted that any change in population or shift in elevation range should come after the dip in temperatures. This does not happen in either of the two early cool periods and for the late cool period, half of the surveys show a decrease in population after the cool period and half show an increase in population. It should also be noted that by 1300 this entire area is abandoned, which might be a reflection of the low temperatures in the 1200s.

114 Figure 42 Elevation, population and temperature comparison for Wetherill Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5- year average temperature.

115 Figure 43 Elevation, population and temperature comparison for Mockingbird Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

116 Figure 44 Elevation, population and temperature comparison for Mesa Verde Area 2000 survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

117 Figure 45 Elevation, population and temperature comparison for Sand Canyon/Goodman Point survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30 year average temperatures, (C) shows 5-year average temperature.

118 Precipitation Analysis

Broad Time Scales.

Figure 37 shows the comparison of annual, 5-year and 30-year averaged precipitation for the overall area (Salzer and Kipfmueller 2005). With precipitation in this area it appears more common to have unusual increases in precipitation than drastic decreases. While this may be true if we just look at the differences from the mean, low levels of precipitation may still impact crops because the amounts of normal precipitation in this area are so low. Beginning in 625 there is a period of alternating above average precipitation and below average precipitation, a seesaw that goes on for about 15 years. In the 30 year averaged data, this seesaw pattern disappears and in fact shows a slight increase in precipitation over that time period. The low precipitation from 700 to 725 on the 5-year average precipitation graph shows as a slight decrease in precipitation on the 30-year graph. On the 5-year graph, around 925 there is a downward spike of precipitation and then again in 975. Neither of these two spikes lasts longer than five years. In the late 1000s there is a slightly broader downward spike lasting about 10 years and then beginning in 1217 there is another seesaw that lasts until 1300. The 30-year averages show almost no time when precipitation is severely below average and, in fact, show four time periods with greater than average precipitation at about 725, 925, 1030 and 1110. It does appear that precipitation during the late 11th century was very variable, but although the precipitation rates swing above and below the mean, this time period seems to have fewer highs and lows than previous times, except for the last five years when the precipitation level reaches its lowest point.

Looking at the remaining surveys, Wetherill (Figure 46), Mockingbird Mesa (Figure 47),

Mesa Verde Area Survey (Figure 48) and Sand Canyon/Goodman Point (Figure 49), there does

119 Figure 46 Elevation, population and precipitation comparison for Wetherill Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5- year average precipitation.

120 Figure 47 Elevation, population and precipitation comparison for Mockingbird Mesa survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

121 Figure 48 Elevation, population and precipitation comparison for MVAS survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

122 Figure 49 Elevation, population and precipitation comparison for Sand Canyon Goodman Point survey for broad cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

123 not appear to be any relationship between precipitation and elevation changes. When it comes to population and precipitation, there also does not seem to be a relationship. While in all cases population increases when the precipitation is highest and most variable, population does not decrease when precipitation shows less variability.

Narrow Time Scales.

Looking at the narrow time scales for precipitation, elevation and population shown in

Figures 50-53, relationship between population and precipitation is evident. On Wetherill Mesa, during the Pueblo III, rather than a slight decline, we see a peak at the Late Pueblo II that decreases to Pueblo I and Middle Pueblo II levels in the Early Pueblo III. It then rises in the Late

Pueblo III while the precipitation seesaws. Data for Mockingbird Mesa, Mesa Verde Area 2000 survey and Sand Canyon/Goodman Point surveys, look similar in that they all peak at Late

Pueblo II, but do not show the dramatic decrease in Early Pueblo III. These three surveys indicate little change or loss of site populations. There appears to be no correlation between precipitation and population or elevation when looking at the narrow cultural periods.

It would be nice to be able to statistically compare elevation and population, elevation and temperature, and elevation and precipitation, but with at most seven data points in each set no statistically valid method of comparison is possible. From the graphs, it appears that population change is not associated with either temperature or precipitation change. It also appears that the Ancestral Puebloans change the elevation of their habitation sites to take advantage of free space within the constraints of maize agriculture.

124 Figure 50 Elevation, population and precipitation comparison for Wetherill Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

125

Figure 51 Elevation, population and precipitation comparison for Mockingbird Mesa survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

126 Figure 52 Elevation, population and precipitation comparison for Mesa Verde Area 2000 survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

127 Figure 53 Elevation, population and precipitation comparison for Sand Canyon Goodman Point survey for narrow cultural periods. (A) shows elevation in comparison with population, (B) shows 30-year average precipitation, (C) shows 5-year average precipitation.

128 Chapter 6. DISTANCE TO LARGE SITES

Large aggregated sites occurred during two time periods in the Mesa Verde area –

Pueblo I and Pueblo III. During Pueblo I (Breternitz, et al. 1986, Varien and Wilshusen 2002b,

Wilshusen and Ortman 1999), large aggregated villages of pithouse structures came into existence. Many of these were excavated during the Dolores Archaeological Project. These pueblos dated between 750 and 900 C.E. (Wilshusen 1999) and had between 50 and 400 rooms.

While these early large villages could be used to investigate settlement patterns and the relationship of large sites to small sites, this study will not include them for two reasons. First,

Pueblo I villages like those found during the DAP (Kane and Robinson 1988, Lipe, et al. 1988) are scarce in other parts of the region. Second, the area where most of the large Pueblo I villages have been surveyed or excavated was abandoned when the Pueblo I villages were abandoned and do not show later occupation (Coffey 2006, Schlanger 1988, Schlanger and Wilshusen 1993).

The areas covered by surveys in this study have only one large Pueblo I village.

Sites like Yellow Jacket Pueblo with 1,200 plus rooms and Sand Canyon Pueblo and

Goodman Point Pueblo with 400 plus rooms are obviously large, but the accepted definition of a large site in this geographic area is a site with more than 50 surface rooms (Adler 1996c). This number was arbitrarily established as the threshold for large sites at the “Pueblo Cutures in

Transition: A.D. 1150-1350 in the American Southwest,” conference held at Crow Canyon

Archaeological Center in 1990 (Adler 1996c).

Although many aggregated sites existed only during Pueblo III and some only during

Late Pueblo III, the location of the large aggregated sites will be used to compare distances in all time periods. In this way, any change in distance to the large sites should be apparent.

129 Large Sites

Of the five surveys in this study, only the Penn State Mesa Verde Area Survey does not have a large site within its boundaries (Messer 2001b). Although the surrounding area in the valley has not been surveyed, it is doubtful that a pueblo of 50 or more surface rooms would go unnoticed. The closest known large site is located at Yucca House (5MT5006), 6.7 km from the survey area. While there might be closer large sites on the mesa top on the Ute Mountain Ute

Reservation, those sites cannot be considered as part of the community. During the eight weeks I surveyed this area (Messer 2001b), I spent quite a bit of time staring at the steep side of the mesa and could find no path from the valley to the mesa top. Discussions with Bruce Bradley, one of the land owners who lived there for many years and an archaeologist, indicated that he had never been able to find any path from the bottom to the top of the mesa (Bradley 2000). No one else I discussed this with was able to suggest a path to the top, so it is difficult to see how sites on the mesa top could have been part of the same community as sites in the valley in this particular location.

With only one large site anywhere near, and that at a distance of nearly 7 km, it is doubtful that an analysis of distance to the site would shed any light on the locations of the small sites in the Penn State Mesa Verde Area Survey. Thus this survey is not considered in the remainder of this chapter.

As can be seen in Table 15, the Chapin Mesa Survey area does contain large sites. In fact, this survey has seven sites with over 50 rooms. However, one site, 5MV782, is undated. Because

130 Table 15 Sites with 50 or more rooms in survey areas. Survey Site Name Site # # of rooms Landform Time

Sand Sand Canyon 5MT765 400+ Mesa, Canyon LPIII Canyon head Goodman Point 5MT604 400+ Mesa, Canyon LPIII head Casa Negra 5MT3925 60 Mesa PI- LPIII Mockingbird Seven Tower Ruin 5MT1000 120 Talus, Canyon LPIII head? Unnamed 1 5MT1541 99 Mesa EPIII Unnamed 2 5MT1512 50 Talus, Canyon LPIII head? Wetherill Long House 5MV1200 150 Cliff PI, LPIII Mug House 5MV1229 94 Cliff EPII- LPIII Double House 5MV1385 75 Cliff EPIII- LPIII Kodak House 5MV1212 60 Cliff LPIII Spring House 5MV1406 55 Cliff LPIII Ruin 16 5MV1241 50 Cliff LPIII Dog 5MV1676 50 Mesa PI- House(Badger) EPII Chapin Cliff Palace 5MV625 220 Cliff LPIII Spruce Tree 5MV640 114 Cliff LPIII House Unnamed 1 5MV907 75 Mesa EPIII Square House 5MV650 70 Cliff LPIII Oak Tree House 5MV523 55 Cliff LPIII Unnamed 2 5MV782 50 Mesa NA

Far View Ruin 5MV808 50 Mesa EPIII

131 this site cannot be assigned to a specific time period and because of the evaluations in Chapters

3 and 4 that the survey data for Chapin Mesa are unreliable, the Chapin Mesa survey will not be used in the remainder of this chapter.

The remaining three surveys, Wetherill Mesa, Sand Canyon/Goodman Point and

Mockingbird Mesa, all have large sites within their survey areas. Table 15 shows these sites with their room numbers, dates and landform locations.

On Wetherill Mesa (Figure 56) there are seven sites with more than 50 surface rooms,

Figure 54 Map of Wetherill Mesa with large sites marked, PIII sites red square, Badger House in red circle. Green denotes National Parks Service land. (MapSource 1999)

132 however, Badger House (5MV1676) is a Pueblo I village with Early Pueblo II habitation and will therefore not be considered. Long House (5MV1200) and Mug House (5MV1229) both have early occupations, but also have Late Pueblo III occupation so will be considered. The remainder of the large sites in the Wetherill survey all have Pueblo III habitation and most are only Late

Pueblo III habitations. All of the large sites on Wetherill Mesa are cliff sites. Although

Nordenskiöld (1893) notes a seep sufficient for his crew’s needs beneath Kodak House

(5MV1212), the only “good” spring he notes is on the North end of the mesa about 3 km from

Figure 55 Map of Sand Canyon/Goodman Point large sites. Sites shown in red squares. Green denotes National Parks Service land. (MapSource 1999)

133 Long House (5MV1200). However, Spring House also has a spring in the back of the alcove.

There probably are or were other springs and seeps on the mesa.

Three large sites exist In the Sand Canyon/Goodman Point survey (Figure 57), Sand

Canyon Pueblo (5MT765), Goodman Point Pueblo (5MT604) and Casa Negra (5MT3925).

Both Sand Canyon and Goodman Point have more than 400 surface rooms but Casa Negra has only 60 rooms. Both Sand Canyon and Goodmen Point Pueblos are dated to the Late Pueblo III period while Casa Negra shows habitation from Pueblo I through Late Pueblo III. However, it is likely that Casa Negra was a Pueblo II great house (Varien 1999a) and might have had only slight occupation during the Late Pueblo III. While all three of the sites are on the mesa top,

Sand Canyon and Goodman Point Pueblos are built around canyon heads. Goodman Point still has a decent spring within the pueblo confines and Sand Canyon had a spring within recent memory (Connolly 1990, Connolly 1992). The spring apparently disappeared in recent years when someone decided to dynamite it to increase flow.

Because of the differences between Sand Canyon and Goodman Point Pueblo and Casa

Negra, the Sand Canyon Goodman Point survey area will be investigated in two ways. First as if only the two largest sites were important as community centers and second as if all three were community centers.

On Mockingbird Mesa (Figure 58) there are also three large sites, Seven Tower Ruin

(5MT1000), 5MT1541 and 5MT1512. Two of the sites, Seven Tower Ruin and 5MT1512 date to the Late Pueblo III, while 5MT1541 dates to the Early Pueblo III. The two Late Pueblo III sites are on the talus slope while the third site is listed as mesa top. However, because of the way landforms were recorded in that survey and because the two “talus slope” sites are both located

134 at springs, it is likely that these are actually near the canyon head. The Early Pueblo III site is on the mesa top about 400 m from a spring. Mockingbird Mesa will be investigated with all three sites as central places and then with only the two Late Pueblo III sites as central locations.

Figure 56 Map of Mockingbird Mesa large sites indicated with red squares. (MapSource 1999)

If large sites did have great sociopolitical impact, then they would form the center of communities. For this study, I have established community size and distribution by associating each small site with the closest large site. I figured the distance from each small site to each of

135 the large sites in its survey and then chose to associate the small site with the large site it was closest to. The mean distance between large and small sites in each time period (whether the large site existed in that spot at that time or not) were recorded using both the broad time scales of Basketmaker, Pueblo I, Pueblo II, Pueblo III and the narrow time scales of Basketmaker,

Pueblo I, Early Pueblo II, Middle Pueblo II, Late Pueblo II, Early Pueblo III and Late Pueblo III.

By sorting the sites in this way many of the communities consisting of less than 30 sites at

Figure 57 Map of all habitation sites recorded in Wetherill Mesa Survey with sites with over 50 surface rooms identified. Sites in upper right corner are not actually on Wetherill Mesa but were included in the survey. Large sites are all on canyon rims; the sparser distribution is on the talus slope.

136 specified times and therefore most standard statistical analysis were inappropriate (Clark and

Hosking 1986). I did carry out complete descriptive statistics on all communities and time periods including mean and standard deviation. I have graphed the means of each large site community through time assuming that, on average, if the large sites were indeed magnets, then the average location of the small sites would tend toward the large site locations, only when the large sites existed.

Wetherill Mesa Survey

Figure 58 Average distances through time of small sites to Double House and Mug House on Wetherill Mesa using broad cultural time periods.

137 Figure 57 shows the distribution of all habitation sites on Wetherill Mesa in all time periods. The figure also shows all sites with more than 50 surface rooms. It is clear from Figure

54 that these sites are all cliff alcove sites and that they are all on the edges of canyons. The distribution in Figure 57 shows the drop off in sites past the cliff edge, but does show some sites on the talus slope. The sites in the northeastern corner, while included in the Wetherill Mesa survey, are not actually on Wetherill Mesa.

Figure 59 Average distances through time of small sites to Long House, Ruin 16, Kodak House and Spring House on Wetherill Mesa using broad cultural time periods.

Figures 58 and 59 show the mean distances to a large site from the small sites associated with it. I have grouped Long House, Ruin 16, Kodak House and Spring House

138 together because they are in the middle of the mesa and are contained on both sides. That is, there is a boundary, either the mesa edge or another large site, surrounding these sites. They have total average distances of associated sites at 466 m, 623 m, 622 m and 662 m, respectively.

Double House and Mug House while also having mesa edges defining two sides of their territory, each have one territorial edge that is unrestricted by either another large site or a relatively nearby mesa edge. The average distance of small sites to these large sites is 1491 m and 2680 m, respectively. Consequently, while the four constrained sites have fairly small communities ( Long

Figure 60 Average distances through time of small sites to Double House and Mug House on Wetherill Mesa using narrow cultural time periods.

139 House, with 50 sites; Ruin 16, with 49 sites; Kodak House, with 61 sites, and Spring House, with 52 sites) and are fairly small in area, the two large sites on the edges have much larger communities (Double House, with 176 sites and Mug House, with 218 sites) and cover a much larger area. Because we know that the entirety of Wetherill Mesa was surveyed and there are, in fact, small sites on both ends of the mesa with no large sites in those locations, it seems that the sizes of the communities were dictated more by the happenstance of the location of large sites

Figure 61 Average distances through time of small sites to Long House, Ruin 16, Kodak house and Spring House on Wetherill Mesa using broad cultural time periods.

140 than the large sites as attractors. This could imply that some small sites formed communities with large sites as their focus while others did not, or it could mean that small sites were located on the landscape to meet environmental constraints. That is, the spaces defined as communities really are not communities, but are simply the normal and natural way that small sites were positioned on the landscape to make the most convenient and beneficial use of the land.

If we look at the mean distances for the four central large sites using the broad time scales, we see that distances to the locations of the large sites are mostly unchanged through time. Figure 59 shows that for Long House, the largest large site on the mesa, the difference from the closest mean distance to the furthest mean distance is only 24 m. Because Long House is a Late Pueblo III site, it would be expected that the Pueblo III mean distance would be the shortest, however, the Pueblo III distance is 8 m further away than the Pueblo II mean. The small difference in distances cannot be assumed to influence the lives of the people who lived in the small sites surrounding Long House. Also, the number of total sites in the area is about average for the four sites and the number of sites during the Pueblo III is only 6 sites less than in the

Pueblo II.

Ruin 16, like Long House, lacks any Basketmaker habitation, but in general, during the other time periods, there is little change in mean distance. In this case, the largest mean occurs during the Pueblo III and is 14 m more than the next furthest mean which occurred during the

Basketmaker period, but 53 m further than the mean for the Pueblo II. Certainly the small sites around Ruin 16 were not moving toward the large site. At Kodak House, the Pueblo III time period mean is 4 m further away than the Pueblo II mean and both the Pueblo II and III means are further away than either the Basketmaker or Pueblo III means.

141 At Spring House, the furthest mean distance to small sites occurs during the Pueblo I period. During the Pueblo II and III periods, there is only one meter difference between the mean distances. While Spring House does show a bit more variation than the other three sites, during the later half of Pueblo occupation there is little change in mean distances.

The relative standard deviations or coefficient of variation of these four sites through time show only moderate to small variation with Spring House showing the least variation followed by Long House and Ruin 16. Of these four communities, Kodak House has the most variability, however, the values are still moderate.

Looking at Double House and Mug House, the large sites located on the outside of the four discussed above, shows much more variation in the mean distances to small sites. At Double

House, the mean of the small sites is closest during Pueblo I and then moves further away by 703 m during Pueblo II. During Pueblo III, the mean distance is smaller, by 373 m. Obviously the community around Double House appears to contract and expand at different time, with Pueblo

II being an expansion phase. Mug House also has its closest phase during Pueblo I with a jump of 1407 m during Pueblo II and another 273 m during Pueblo III. At least during the later phases of occupation it appears as if Mug House is a repellant rather than a magnet.

The variability of the mean distances for the Double House and Mug House sites is larger than for the other four sites. Mug House has the largest variability in mean distances. The variability of Double House, while slightly more than the four middle sites, is still less than that for Mug House.

Figures 60 and 61 show the mean distances from the large sites to the small sites in their communities using the narrow, more finely divided cultural phases. While this is beneficial for

142 many analyses, in the case of distances to large sites it becomes problematic. By dividing both time and space into smaller units, the numbers of small sites associated with large sites becomes much smaller yielding, in some cases, site samples too small to indicate anything. Also, while sites can reliably be assigned to broad time periods, assigning sites to the finer time periods is less accurate. Given these cautions, it is still valuable to look at the sites in the narrower cultural phases.

As seen in Figure 61, again, the four large sites located in the middle seem unaffected by the appearance of large sites in their midst. At Long House, the mean distances are again fairly unchanging except for a decrease in the mean distance during the Early Pueblo III and a slight increase in the distance during Late Pueblo III. Ruin 16 sees the opposite with an increase in the

Early Pueblo III and a decrease in the Late Pueblo III. However, this decrease is still 22 m further from Ruin 16 than the mean distance was during the Late Pueblo II. Kodak House sees a large increase during the Early Pueblo II of 243 m, but then a decrease of 291 m during the Late

Pueblo III. This, however, is only 42 m closer to Kodak House than the mean distance during the late Pueblo II. Again the existence of a large site in the area during Late Pueblo III does not seem to alter site distribution a great deal. The mean distances for Spring House are nearly flat throughout the Pueblo II and Pueblo III time periods, with the largest difference in mean distance

15 m.

Figure 60 shows the mean distance distribution for Double House and Mug House using the narrow cultural phases. In this case, Double House does appear to be less variable than Mug

House with its largest mean distance during Late Pueblo II and a drop of 570 m during Early

Pueblo III followed by another drop of another 170 m in Late Pueblo III. At Mug House, the

143 changes are larger with a drop during Late Pueblo II of 1156 m, and increase in Early Pueblo III of 759 m and a drop again of 1285 m in the late Pueblo III. Looking at the graph of Mug House in Figure 60, the changes in mean distance appear without pattern.

If we look at the coefficients of variation for Mug House we see the highest level of variability of any of the six sites. Spring House has the lowest variability with Kodak House,

Long House and Ruin 16 showing only slightly more variability during the Pueblo II and Pueblo

III phases. Double House is slightly more variable than the middle four large sites.

Do the large sites on Wetherill Mesa act as central places for the location of small sites when they appear in the Late Pueblo III? While it is impossible to answer that question definitively with the data available, I think that what is happening is something different from a central place attraction. In the vicinity of the four middle sites, the site distribution pattern fills the landscape. Sites are spread out within the area to take advantage of the space, which still leaves the small sites with easy access to the large ones. One indication of this is that during the

Pueblo III phase, the four middle sites have approximately 5 fewer sites in their communities than they did during Pueblo II. This would happen because the large sites can accommodate more people than a small site and the population is also thought to have declined. However, if the large sites were the only factor in landscape use, then one would expect there to be about three times fewer, or 15 fewer small sites surrounding Long House which is three times as large as the other three sites. However, there is only one less site around Long House than around the other three sites. The small sites occupy the landscape in this area in the same way they did when no large site was present.

With respect to the two large sites on the edges, the variation in mean distances simply

144 reflects the larger, unbounded area to be filled. The difference between Mug House and Double

House can also be explained by geography. Mug House has more open land than the four middle sites, but it is also near the edge of Mesa Verde National Park and therefore the edge of the survey area. Double House, however, is near an open area where no other large sites exist. These sites can either be considered part of the Double House community or they could be considered unaffiliated sites. Either way, the distribution around Double House is more dispersed than any of the other sites and does not indicate that Double House was an attraction.

Sand Canyon/Goodman Point Survey

Figure 62 Map of all habitation sites recorded in Sand Canyon/Goodman Point Survey with sites with over 50 surface rooms identified.

145 Figure 62 shows the distribution of all sites in all time periods in the Sand

Canyon/Goodman Point survey. Also shown are the three sites with 50 or more surface rooms –

Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra. In this case, Sand Canyon Pueblo and Goodman Point Pueblo are approximately the same size, and, as can be seen from Figure 55, are both at canyon heads, are both at or were at springs and are both Late Pueblo III habitations.

Casa Negra shows habitation from Pueblo I through Pueblo III, but is thought to be the location of a Pueblo II Chacoan great house. It is located on flat land away from the canyons.

The distribution of sites in Figure 62 does not look very different in most locations from that of Wetherill Mesa. However, there is one area about one third of the distance from

Figure 63 Average distances through time of small sites to Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra in Sand Canyon/Goodman Point locality using broad cultural time periods.

146 Goodman Point Pueblo to Sand Canyon Pueblo that is unusual in that it contains no habitation sites at all through all time periods. Figures 57 and 68 show the distribution of sites on Wetherill

Mesa and Mockingbird Mesa and do show some areas that appear to be empty of sites, but a closer look finds that these empty areas are at canyons or other geographic or geological features precluding sites. Most of these empty spaces are vaguely circular or irregular.

On the Sand Canyon/Goodman Point map, the empty space cuts a large swath from northwest to southeast. A look at the topography shows no canyons or other geographic or geological entities that would explain an absence of habitation sites. In fact, the location is on the top of the McElmo Dome, a raised geological formation that is mostly unbroken level ground.

Figure 64 Average distances through time of small sites to Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra in Sand Canyon/Goodman Point locality using narrow cultural time periods.

147 Why are there no sites in this particular location? One explanation could be that that area is the dividing line between the two communities, a semi-no man’s land. However, with six large sites on Wetherill Mesa there are no empty areas. In fact, one cannot draw the boundaries of the assumed communities without identifying the sites closest to each site and circling them. While small sites might have in some way formed communities around large aggregated sites, in other areas, the aggregation does not form empty zones between communities. Also, no sites occur in this area during any time period. Another possible explanation would be that the area was not surveyed, but we know that all land within the Sand Canyon/Goodman Point survey area was

Figure 65 Average distances through time of small sites to Sand Canyon Pueblo and Goodman Point Pueblo in Sand Canyon/Goodman Point local using narrow cultural time periods.

148 covered. A clue to what is really happening might be found in the land use patterns in the area during modern times. The McElmo Dome area has been farmed by bean farmers since the early

1900s (Connolly 1990, Connolly 1992). These farmers use dry land techniques which include substantial plowing. In many areas, when driving down the road the locations of Pueblo sites can be seen in the middle of fields where they have been plowed around. In other places farmers took the time and effort to remove the rubble mounds. This entire area was impacted by plowing

(Adler 1990) and sites in the empty area might no longer be visible from the surface. If the cause of the empty area is obliteration of the sites in that area, the absence of sites in this area will certainly decrease the number of sites associated with Goodman Point Pueblo. However, because the missing area cuts evenly across the study area, it should not necessarily influence the pattern of mean distances to the pueblo, but will influence the results because the most distant sites are no longer evident. The Sand Canyon/Goodman Point survey will be analyzed as if the cause of the missing sites were plowing, or the empty area occurred naturally. If these are not the reasons for the empty space, these data would be unusable for this portion of the study.

The distribution of the mean distance from large sites considering all three large sites in the area is shown in Figure 63 for the broad cultural phases. The total mean distances when considering all three large sites are 1391 m for Sand Canyon, 1299 m for Goodman Point and

1205 m for Casa Negra. The mean distances for Sand Canyon Pueblo peak during Pueblo I and taper off during Pueblo II and Pueblo III. The mean distance of 1606 m for Pueblo I decreases by 210 m to 1396 m in Pueblo II and then by 50 m in Pueblo III. The pattern for Goodman Point

Pueblo looks very similar except the largest mean distance is during Basketmaker with the subsequent time periods decreasing afterwards. The mean distance for Pueblo I is 1578 m. The

149 mean distance from small sites decreases by 196 m to 1382 m during Pueblo II and another 330 m during Pueblo III for a mean distance of about 1052 m.

The pattern for Casa Negra differs from that for Sand Canyon and Goodman Point

Pueblos. Pueblo I is actually the time period with the nearest mean distance between the small sites at 1108 m. The mean distance to small sites then increases by 72 m to 1180 m in Pueblo II and by 213 to 1393 m by Pueblo III. If the Casa Negra site was a Chacoan style great house and

Figure 66 Comparison of average distances through time of small sites to Sand Canyon Pueblo, Goodman Point Pueblo and Casa Negra and to Sand Canyon and Goodman Point only in Sand Canyon/Goodman Point local using broad cultural time periods.

150 formed the center of a Pueblo II community and if community centers dictated a physical movement toward the center, it is puzzling why the mean distance to small sites increases during

Pueblo II rather than decreases (Doyel and Lekson 1992, Sebastian 1992a, Sebastian 1992b,

Vivian 1990).

The variation of mean distances in the broad time scale appears to be similar to those seen in the Wetherill Mesa data if we assume Sand Canyon and Goodman Point are the edge sites and Casa Negra is the middle site. The variation of the mean distance to small sites from

Casa Negra is very small. However, the variation of the mean distances from small sites to both

Sand Canyon and Goodman Point Pueblo is large, even more so than on Wetherill Mesa. One reason for this variability pattern using broad time scales is, I believe, that while Casa Negra in the middle is constrained by the presence of the two sites on either side, Sand Canyon and

Goodman Point Pueblos are not constrained by other large sites. Although they are both located at canyon heads, these canyons are much smaller than those on Mesa Verde. These two large sites are constrained only in the direction and by the small arc of their canyons, and attract sites from the rest of the landscape. The mean total distances when considering only two large sites in the area are 1725 m for Sand Canyon and 1341 m for Goodman Point.

Figure 64 shows the distribution of mean distances to large sites using the narrow cultural phases. For Sand Canyon Pueblo, the distances remain fairly even for Early and Middle Pueblo II and then increase by 125 m during Late Pueblo II. The mean distance decreases by 158 m in

Early Pueblo III but then increases by 6 m in Late Pueblo III. In essence, there is no change during Pueblo III. The Goodman Point data show a similar pattern but with a small increase of

65 m between Early and Middle Pueblo II and another increase of 42 m to Early Pueblo III.

151 Again, the Early and Late Pueblo III phases differ by only 6 m. At Casa Negra, the closest period is in Early Pueblo II with a continual rise throughout the remaining phases. The difference from

Early Pueblo II to Middle Pueblo II is 201 m with another 185 m increase to Late Pueblo II. The

84 m increase to Early Pueblo III is followed by a decrease of 26 m for Late Pueblo III. While

Casa Negra had a Pueblo I occupation and an Early Pueblo II occupation, the Casa Negra great house probably did not come into existence until Middle or Late Pueblo II. However, this is the time period during which the mean distance to the site from small sites was increasing.

Using the narrow cultural phases, the variation of the means is slightly greater for Casa

Negra than it is for the broad time scales, but both are small. The variation for Sand Canyon and

Goodman Point remain the same as they are for the broad cultural phases.

Figure 65 shows the mean distance distribution for small sites to larger sites considering only Sand Canyon and Goodman Point Pueblos. Casa Negra was probably not a major force during the Pueblo III time period and both Sand Canyon and Goodman Point Pueblos have been dated to Late Pueblo III exclusively (Bradley 1987, Bradley 1992, Coffey and Kuckelman 2006,

Kuckelman and Coffey 2007). If there was any geographical move toward large aggregated sites, this should appear when looking at only these two sites as central places for the Sand

Canyon/Goodman Point survey area.

When only two large sites are considered, the data for Sand Canyon become very uniform. While there is a slight decrease from Pueblo I to Pueblo II and then Pueblo III, this decrease in total amounts to 51 m with a decrease of 47 m from Pueblo I to Pueblo II and another decrease of 4 m to Pueblo III. The Goodman Point data do show a steady decline from

Pueblo I to Pueblo III, decreasing 183 m from Pueblo I to Pueblo II and then 332 m from Pueblo

152 II to Pueblo III. This is the first indication that there might have been movement toward

Goodman Point during the Pueblo III period. However, if that happened because of sociopolitical reasons, it did not appear to take place at Sand Canyon Pueblo.

Figure 65 shows the mean distances between small sites and Sand Canyon and Goodman

Point Pueblos using narrow cultural phases. The Sand Canyon data do show more changes using the finer time scales including a drop of 393 m from Early Pueblo III to Late Pueblo III perhaps indicating that Sand Canyon was a magnet during that time period. However, now the Goodman

Point data show nearly identical distances for Early and Late Pueblo III and flat distances for all three segments of Pueblo II.

Figure 67 Comparison of average distances through time of small sites to Sand Canyon, Goodman Point Pueblos and Casa Negra and to Sand Canyon and Goodman Point Pueblos only in the Sand Canyon/Goodman Point local using narrow cultural time periods.

153 The variability of the Goodman Point data are small to moderate with these time periods, but the Goodman Point data are still very variable. Figure 66, using the broad cultural phases, compares the Sand Canyon and Goodman Point Pueblos when Casa Negra is considered a large site in the area and when only the two large pueblos are considered. The differences between the

Sand Canyon graphs are marked. With Casa Negra, the Sand Canyon and Goodman Point

Pueblos have similar patterns. Without Casa Negra, Goodman point essentially remains the same while Sand Canyon becomes very uniform in mean distances. It appears that Casa Negra only takes small sites away from Sand Canyon and not from Goodman Point. This could be an artifact of the empty area near Goodman Point, it could be an indication that Casa Negra played an important part in Pueblo III organization, or it could simply be a chance occurrence.

Figure 67 compares Goodman Point and Sand Canyon Pueblos when Casa Negra is considered and when Casa Negra is omitted using the narrow time scale. Again, the absence of

Casa Negra does not seem to affect Goodman Point Pueblo very much at all. While it is obvious that Sand Canyon garners more sites from the absence of Casa Negra, the pattern of mean distances for Sand Canyon does not change very much using narrow time phases.

Mockingbird Mesa Survey

Figure 68 shows the distribution of all sites through all time periods on Mockingbird

Mesa. Also shown are the three large aggregated sites. South of 5MT1512 and North of

5MT1541 are locations free of sites because these areas are on the edge of canyons or are greatly sloping locations. Seven Tower Ruin (5MT1000) does have some small sites surrounding it, but is also fairly near the mesa’s edge. Woods Canyon runs up the east side of the map area. The

154 Figure 68 Map of all habitation sites recorded in Mockingbird Mesa Survey with sites with over 50 surface rooms identified.

average total distances from large sites to small sites are 1475 m for Seven Tower Ruin, 1250 m for 5MT541 and 2114 for 5MT1512.

Figure 69 shows the distributions of mean distances from small sites to the three large sites on Mockingbird Mesa. Seven Tower Ruin shows an increase of 102 m in mean distance from Pueblo I to Pueblo II and a 399 m drop in mean distance from Pueblo II to Pueblo III. At site 5MT1541, there is a gradual rise in mean distance from Pueblo I with an increase of 61 m to

Pueblo II and an increase of 78 m to Pueblo III. These increases are small and show a gradual move away from the site. At site 5MT1512 there is also an increase in mean distance from

Pueblo I through Pueblo III, but the distances are much greater. The change in mean distance

155 Figure 69 Average distances through time of small sites to Seven Tower Ruin, 5MT1541 and 5MT1512 in Mockingbird Mesa survey using broad cultural time periods.

from Pueblo I to Pueblo II is 457 m and the change from Pueblo II to Pueblo III is 132 m.

The variability of the means at 5MT1541 and 5MT1512 are small in the broad time scale.

The variability for Seven Tower Ruin is larger, but much less than seen in the Sand Canyon

Pueblo data. One problem with the Mockingbird Mesa data are the small number of sites associated with the large sites during these time periods. This only becomes more pronounced when dividing the sites into the narrow cultural phases.

Figure 70 shows the distribution of mean distance for the Mockingbird Mesa data using narrow cultural phases. Little changes in the Seven Tower Ruin distribution. For the 5MT1541

156 site, the Early and Late Pueblo III time periods are slightly nearer than the Pueblo II time periods, but the difference between Early Pueblo III and Late Pueblo III is 49 m. Using the narrow time periods, 5MT1512 takes on a ragged look with the furthest time period during the

Figure 70 Average distances through time of small sites to Seven Tower Ruin, 5MT1541 and 5MT1512 in Mockingbird Mesa survey area using broad cultural time periods.

Middle Pueblo II dropping 252 m in the Late Pueblo II and then increasing 286 in the Early

Pueblo III and decreasing 100 m in the Late Pueblo III. Again, 5MT1512 shows very small variability with 5MT1541 showing only a slight bit more variability. Seven Tower Ruin, however, is more variable in the narrow time periods than in the broad time periods.

157 Figure 71 Average distances through time of small sites to Seven Tower Ruin and 5MT1512 in Mockingbird Mesa survey area using broad cultural time periods.

Even though 5MT1541 is a fairly large site with 99 surface rooms, it is an Early Pueblo

III site so possibly exerted no influence in the Late Pueblo III when the other two sites became aggregated. Figure 71 shows the distribution of mean distances when only Seven Tower Ruin and 5MT1512 are considered for broad time periods. At Seven Tower Ruin, the mean distance increases from Pueblo I to Pueblo II by 83 m and then drops by 199 m from Pueblo II to Pueblo

III. At 5MT1512, using broad time intervals the mean distance increases by 406 m from Pueblo I to Pueblo II and then by 172 m from Pueblo II to Pueblo III. The variability when only Seven

158 Figure 72 Average distances through time of small sites to Seven Tower Ruin and 5MT1512 in Mockingbird Mesa survey area using narrow cultural time periods.

Tower Ruin and 5MT1512 are considered is again very small for 5MT1512, but only slightly reduced from when all three large sites are considered.

Figure 72 shows the mean distance from the small sites to large sites on Mockingbird

Mesa using narrow time periods. The average total distance for Seven Tower Ruin when only two sites are considered is 1475 m. For 5MT541 it is 2062 m. For Seven Tower Ruin there is a decided steady decline in distance from the site from Middle Pueblo II through Late Pueblo III.

For 5MT1512 the distances rise from Middle Pueblo II through Early Pueblo III and then decline

159 Figure 73 Comparison of average distances through time of small sites to Seven Tower Ruin and 5MT1512 when 5MT1541 is included or excluded using broad cultural time periods.

Figure 74 Comparison of average distances through time of small sites to Seven Tower Ruin and 5MT1512 when 5MT1541 is included or excluded using narrow cultural time periods.

160 in Late Pueblo III, however, the closest community occurs during Pueblo I. Even though the decline in distance from Early Pueblo III is 223 m, it is smaller than the increase from Late

Pueblo II to Early Pueblo III. The variability of 5MT1512 is also small using narrow time periods and the variability at Seven Tower Ruin is slightly less using narrow times with broad times.

Figure 73 compares the distribution of the mean distances from small sites to Seven

Tower Ruin and 5MT1512 when site 5MT1541 is present and omitted using broad time periods.

For the Seven Tower Ruin, the only real change seen in the pattern is a shift closer to the aggregated site during the Pueblo III. For the 5MT1512 site, the pattern of distribution for the two scenarios is similar with a bit more distance from the aggregated site to the small sites when only two large sites are considered.

Figure 74 compares the distribution of the mean distances from small sites to Seven

Tower Ruin and 5MT1512 when site 5MT1541 is present and omitted using narrow time periods. Again, even with the narrow time periods, there is little difference in pattern for the

Seven Tower Ruin although the closest distance is 163 m further away when only two sites are considered. Looking at 5MT1512, the patterns also remain pretty much the same with little difference in distances.

Thus, the presence or absence of 5MT1541 has little effect on the pattern of site distribution around 5MT1512. It also appears that the only effect that 5MT1541 has on Seven

Tower Ruin is the small piece of territory from which it captures nearest small sites.

Looking at the mean distance from small sites to large, aggregated sites, there does not seem to be any evidence that the appearance of large aggregated sites changed settlement patterns in most cases. Except for Seven Tower Ruin on Mockingbird Mesa and Goodman Point

161 Pueblo in the Sand Canyon/Goodman Point locality, the other large sites showed either a neutral effect in regard to the mean distance from small sites during the Late Pueblo III or a distancing of small sites from large sites during that time period.

162 Chapter 7. REINHABITED SITES

An investigation of reinhabited sites, their distribution with respect to site numbers, landforms, elevation and distance to large sites can shed light on how ancestral Puebloans chose where to place their houses and what did or did not influence them.

Population

Table 16 Numbers of pristine and reinhabited sites in Pueblo III Survey PIII PIII PIII EPIII EPIII EPIII LPIII LPIII LPIII and only and only and only Wetherill Mesa 253 83 170 85 61 24 192 38 154 Sand Canyon 122 58 64 118 57 61 106 102 3 Mesa Verde 2000 29 16 15 24 14 10 20 15 5 Mockingbird Mesa 61 27 34 47 27 21 49 36 13 Chapin Mesa 213 17 195 69 15 54 159 17 143

Because the majority of large sites occur during the Late Pueblo III (Cordell 1984, Varien 2002,

Varien and Wilshusen 2002b), only sites inhabited at that time are considered. Sites are divided between locations that were previously occupied and those established on pristine land. Pristine sites are referred to as “PIII only” or pristine, and previously inhabited sites are “PIII and” or reinhabited in tables and graphs. Table 16 shows the distribution of reinhabited sites for all

Pueblo III sites also broken down between Early and Late Pueblo III. In order to compare the

163 various surveys, Table 17 shows the same data as percentages of all PIII sites in each survey.

Looking at Table 17, it is obvious that once again the Chapin Mesa data do not fit with the other four surveys. Fully 92 percent of all Pueblo III sites are classified as PIII only sites, while the other surveys range from 33 percent to 48 percent reinhabited. Chapin Mesa during the

Early Pueblo III shows only 15 percent reinhabited, while the other surveys show anywhere from

53 percent to 72 percent reinhabited. The highest rehabitation location for Early Pueblo III is

Wetherill Mesa, the closest area to Chapin Mesa. During the Late Pueblo III, Chapin Mesa has a

10 percent reinhabitation level while the other surveys range from 20 percent to 97 percent.

It appears that once again the data from the Chapin Mesa Survey are different. Survey personnel can certainly miss an early occupation of a site when looking at the obvious remains of a late settlement, but pottery seriation should show at least some of the previous occupations. On

Chapin Mesa, earlier occupations were not recorded to the extent they were in other surveys,

Table 17 Distribution of Pueblo III sites between pristine and reinhabited sites, percentages. Survey PIII PIII PIII EPIII EPIII EPIII LPIII LPIII LPIII and only and only and only Wetherill Mesa 100 33 67 100 72 28 100 20 80 Sand Canyon 100 48 52 100 49 51 100 97 3 Mesa Verde 2000 100 55 45 100 53 47 100 64 36 Mockingbird Mesa 100 44 56 100 53 47 100 73 27 Chapin Mesa 100 8 92 100 22 78 100 10 90

164 suggesting a flaw in the survey technique. So once again, Chapin Mesa will not be considered in this analysis.

A table of numbers, even percentages, does not convey a sense of how much of the

Pueblo III world was actually built on previously used land. Figure 75 shows the ratio of reinhabited sites to pristine sites in the Pueblo III time period across the Wetherill Mesa,

Mockingbird Mesa, Mesa Verde Area Survey 2000 and Sand Canyon/Goodman Point surveys.

Figure 75 Percent of pristine and reinhabited sites during Pueblo III across four surveys.

The survey with the lowest percentage of reinhabited sites during the Pueblo III is the

Wetherill Mesa survey. This probably occurs because Wetherill also has the largest number of cliff or alcove sites of all the surveys during the Pueblo III and except during Basketmaker, cliffs and alcoves were not used much as habitation sites. Consequently the move to cliff locations in

165 the Pueblo III on Wetherill Mesa opened up a large area of previously unused ground. Wetherill

Mesa has the largest percentage of pristine Pueblo III sites of all the surveys considered.

Of the remaining three surveys, Mockingbird and Sand Canyon/Goodman Point have similar percentages and Mesa Verde 2000 has a slightly higher number of reinhabited sites. The

Mesa Verde 2000 survey is the only one with more than half the sites reinhabited, but the actual difference in reinhabited sites is just one with a 16 reinhabited to 15 pristine ratio. A P2 test of the

Pueblo III distributions between reinhabited and pristine sites shows that the surveys are significantly different from each other with a P2 of 10.534 (d=3, p=0.014).

Figure 76 Percent of pristine and reinhabited sites during Early Pueblo III across four surveys

166 The distribution of reinhabited and pristine sites during the Early Pueblo III is shown in

Figure 76. Wetherill Mesa has far more reinhabited sites than pristine sites and the Sand

Canyon/Goodman Point area has about the same ratio of pristine to reinhabited sites as during the entire Pueblo III time period. The Mockingbird Mesa survey and the Mesa Verde 2000 survey show about 60 percent reinhabited sites during the Early Pueblo III. The large proportion of reinhabited sites during Early Pueblo III for the Wetherill Survey might indicate that there are sufficient sites available to reuse, or that fewer cliff sites were occupied during the Early Pueblo

III. There are fewer cliff and alcove sites on Wetherill Mesa during Early Pueblo III. A P2 test of the Early Pueblo III distributions between reinhabited and pristine sites shows that the surveys

Figure 77 Percent of pristine and reinhabited sites during Late Pueblo III across four surveys.

167 are significantly different from each other with a P2 of 11.2 (d=3, p=0.010).

Figure 77 shows the population distribution of reinhabited and pristine sites during the

Late Pueblo III. It becomes obvious that something has happened in the Late Pueblo III but the different surveys initially suggest that whatever occurred did so haphazardly. In the Sand

Canyon/ Goodman Point area, there are almost no sites built in pristine places. Nearly all Late

PIII sites seem to be reinhabited. This might be because if a site was inhabited during the Early

Pueblo III it was considered a reinhabited site during the Late Pueblo III. However, it is impossible in most cases, to tell if there was a break in habitation or the site was actually inhabited in the late Early Pueblo III and early Late Pueblo III. While this can be a problem, it applies to all the surveys and the Sand Canyon/Goodman Point area survey still looks odd. For the Mockingbird and Mesa Verde surveys, they once again have similar ratios of reinhabited to pristine sites, but there are far fewer pristine sites. For all but Wetherill Mesa, the number of pristine sites is less than that during Early Pueblo III. Again, this is undoubtedly due to the fact that anything that was a site during Early Pueblo III was considered a reinhabited site in Late

Pueblo III. A P2 test of the Pueblo III distributions between reinhabited and pristine sites shows that the surveys are significantly different from each other with a P2 of 178.86 (d=3, p<0.0).

Total Pueblo III reinhabited sites numbers are probably more accurate than the narrow time periods because of the difficulty in determining if there is a break in habitation between

Early Pueblo III and Late Pueblo III. However, looking at the data in this way does show the anomaly with Wetherill Mesa. Fully 80 percent of the Late Pueblo III sites on Wetherill Mesa are on pristine sites, undoubtedly indicating the move to alcoves and cliffs. The next section on landform supports this inference.

168 Landforms

Figure 78 Comparison of PIII Landforms across four surveys for pristine and reinhabited sites.

The locations of sites with respect to landform change through time as shown in Chapter

4. Figure 78 shows the locations of sites with respect to landform and whether they are reinhabited sites or pristine sites. The first block of bars shows the reinhabited sites and the second set of bars shows the pristine sites with respect to landform. Mockingbird Mesa and the

Mesa Verde 2000 survey show the least difference between reinhabited and pristine sites. For

Mockingbird, a few cliff sites appear as locations for pristine sites and for the Mesa Verde 2000 area, the few sites located in canyon bottoms disappear. In general, however, for these two surveys, the distribution of pristine and reinhabited sites is pretty much the same.

169 Figure 79 Comparison of PIII Landforms across four surveys for pristine and reinhabited sites.

For the Wetherill Mesa data, the differences are profound. Far fewer pristine sites are on the mesa top and many more are in cliffs or alcoves. The number of pristine talus slope sites also declines, but not as drastically as those on the mesa top. It is not so much that sites are not built on the mesa top during Pueblo III as that sites are not built on new territory on the mesa top during that time period. Old, previously inhabited locations are still used, but new sites move to the cliffs.

Looking at the Early Pueblo III (Figure 79) we see that for both the Sand

Canyon/Goodman Point and Mesa Verde 2000 surveys, there are more talus slope sites on virgin land than there are reinhabited talus slope sites. On Mockingbird Mesa, cliff sites are all

170 reinhabited with no new locations chosen for habitation. Bottom sites, while very limited in number, suddenly appear and are all in pristine locations. The relative distributions of mesa top and talus slope locations remain pretty much the same.

On Wetherill Mesa, the talus slope sites remain steady in percentages between reinhabited and pristine sites, but there are fewer pristine mesa top sites and more pristine sites located in canyon bottoms.

Figure 80 Comparison of LPIII Landforms across four surveys for pristine and reinhabited sites.

171 Figure 80 shows the distribution of landform locations for the four surveys during the

Late Pueblo III for reinhabited and pristine sites. The obvious difference is in the Wetherill Mesa data where there is a huge jump in the percentage of pristine cliff and alcove sites and a decrease in the number of mesa top and talus slope sites. At the Sand Canyon/Goodman Point locality, there is an even distribution of mesa top and talus slope sites between reinhabited and pristine sites. In the Mesa Verde 2000 survey, there are no pristine talus slope sites and all new

Figure 81 Comparison of four surveys with respect to landform and reinhabited and pristine sites during Pueblo III

172 habitations are on flat areas. At Mockingbird Mesa again the cliff sites are all previously inhabited and there are a few pristine bottom sites that appear. There are also more pristine sites on the talus slope and fewer new locations on the mesa top.

Looking at all the surveys with respect to landform and pristine and reinhabited sites

(Figure 81), we see the general trend through time as in Chapter 4 is for fewer mesa top sites, but that for mesa top surveys, Wetherill Mesa and Mockingbird Mesa, the reduction in mesa top sites actually comes during Late Pueblo III and then only for the pristine sites. There the similarity between the mesa surveys ends. Mockingbird Mesa has very few cliff sites in general and no Late Pueblo III pristine cliff or alcove sites. Wetherill Mesa has many more Pueblo III cliff or alcove sites and almost twice as many pristine cliff sites as reinhabited cliff sites.

For the Sand Canyon/Goodman Point and Mesa Verde 2000 surveys, the picture is quite different. There is no difference between pristine and reinhabited sites during the Late Pueblo III forSand Canyon, and only mesa top and talus slope sites exist. However, there are slightly more pristine talus slope sites during Early Pueblo III than reinhabited talus slope sites, or there are more reinhabited mesa top sites than reinhabited talus slope sites during the earlier time period.

For the Sand Canyon/Goodman Point area, the percentage of mesa top sites that are pristine and those that are reinhabited during the Early Pueblo III are close, but the percentage of reinhabited talus slope sites is about half that of the pristine sites because there are canyon bottom sites that are reinhabited, but no new canyon bottom locations occur. During the Late Pueblo III, all of the pristine sites are on the mesa top, while nearly a quarter of the reinhabited sites are still on the talus slope.

P2 tests could not be performed on the Mesa Verde 2000 survey because the numbers of

173 sites were too small to be statistically valid. P2 tests were performed for the Wetherill Mesa,

Mockingbird Mesa and Sand Canyon/Goodman Point surveys. In some cases the categories of talus and cliff or bottom had to be conflated to meet the requirements of more than five entries per category and more than 80 percent of the total data above five entries.

For Wetherill Mesa, the P2 test for Pueblo III pristine and reinhabited sites showed significant difference with a P2 = 48.338 (d=3, p=0.0001). For Mockingbird Mesa, the P2 test for

Pueblo III pristine and reinhabited sites showed no significant difference with a P2 = .3610 (d=1, p=0.548). For Sand Canyon/Goodman Point, the P2 test for Pueblo III pristine and reinhabited sites showed a significant difference with a P2 = 5.202 (d=1, p=0.022).

When looking at the relationship between pristine and reinhabited sites during the Early

Pueblo III, on Wetherill Mesa, the locations of pristine and reinhabited sites are not significantly different. A P2 test cannot be applied to the Late Pueblo III because of insufficient numbers in the categories. When comparing pristine or reinhabited sites to all Pueblo III sites, there was a significant difference in both cases. There was no significant difference when looking at all Late

Pueblo III sites and reinhabited or pristine sites. Only during the Early Pueblo III is there a significant difference between all sites and the reinhabited or pristine sites.

On Mockingbird Mesa, there is no significant difference between pristine and reinhabited sites in all cases using the narrow time periods. In fact, there is no significant difference for all

Mockingbird Mesa landform data except for the relationship between LPIII reinhabited and

LPIII pristine where there is a P2=5.054 (d=1, p=0.044). In the Sand Canyon/Goodman Point locality, looking at the entire PIII, there is a significant difference between pristine and reinhabited sites with a P2=5.202 (df=1, p=0.023). When looking at the Early PIII sites, there is a

174 significant difference between pristine and reinhabited sites with a P2=4.595 (df=1, p=0.032).

There are insufficient data to apply a P2 test to the Late Pueblo III data. When comparing pristine or reinhabited sites to all Pueblo III sites, there is no significant difference among all the combinations, including all Pueblo III sites to pristine Pueblo III sites and reinhabited Pueblo III sites.

Elevations

Figure 82 Comparison of pristine and reinhabited sites on Wetherill Mesa for all Pueblo III, Early Pueblo III and Late Pueblo III elevations.

175 Figure 83 Comparison of pristine and reinhabited sties on Mockingbird Mesa for all Pueblo III, Early Pueblo III and Late Pueblo III elevations.

Figure 82 shows a comparison of elevations of reinhabited and pristine sites during the

Pueblo III on Wetherill Mesa. The range in elevation of the new sites is in all cases narrower than the range of reinhabited sites. The means shift slightly from 2176 m to 2142 m for the entire

Pueblo III and from 2184 m to 2215 m during Early Pueblo III and 2179 to 2129 m during Late

Pueblo III. Employing a Student T test with " = .05, the pristine and reinhabited sites during the entire Pueblo III are not significantly different from each other. During the Early Pueblo III, the

176 pristine and reinhabited sites are also not significantly different from each other. However, during the Late Pueblo III, the pristine and reinhabited sites are significantly different from each other.

Again, this is probably an artifact of the move on Wetherill Mesa to cliff and alcove locations.

Removing the cliff sites from the Late Pueblo III data to see if they are the influence shows that the pristine and reinhabited sites show no significant difference.

Figure 84 Comparison of pristine and reinhabited sties in Mesa Verde 2000 survey for all Pueblo III, Early Pueblo III and Late Pueblo III elevations.

177 Figure 85 Comparison of pristine and reinhabited sties in Sand Canyon/Goodman Point survey area for all Pueblo III, Early Pueblo III and Late Pueblo III elevations.

On Mockingbird Mesa the pristine sites all have the same range although the range of the

Late Pueblo III reinhabited sites contracts to almost equal that of the pristine sites (see Figure

83). It appears that anything below 1850 m was abandoned during the Late Pueblo III. A

Student T test with " = .05 comparing the deviation of Pueblo III pristine and reinhabited sites shows that there is no significant difference between the two sets of sites. Although the Student

T test shows no significant difference with a p = 0.095, the test was run on 34 pristine sites and

178 Figure 86 Comparison of distances to large sites of pristine and reinhabited sites on Wetherill Mesa – all sites, Kodak House, Double House, Long House.

only 28 reinhabited sites and so the validity of this finding is in question. The numbers for the

Early and Late Pueblo III are of course even smaller, and while they do show no significant difference are not statistically valid.

Figure 84 shows the elevations of pristine and reinhabited sites in the Mesa Verde Area

Survey 2000 survey location. Unlike Wetherill Mesa and Mockingbird Mesa, the elevation range for reinhabited sites in this area expands rather than contracts during Pueblo III and Early Pueblo

III but contracts slightly and moves downward during Late Pueblo III. Compared to the other three surveys, the Mesa Verde 2000 survey area is flat with a change in elevation of only 63 m.

The lowest elevation at 1899 m is on the valley floor. The highest elevations are at the fence line for the Ute Mountain Ute Reservation and although sites can be seen above the fence, yet still

179 Figure 87 Comparison of distances to large sites of pristine and reinhabited sites on Wetherill Mesa – Mug House, Ruin 16, Spring House

not in the talus, they were not recorded. The number of sites in this area during Pueblo III is only

29. Dividing them by pristine and reinhabited produces numbers far too small for statistical purposes.

Sand Canyon/Goodman Point locality elevations for pristine and reinhabited sites appear in Figure 85. As in the Mesa Verde 2000 survey, the lower elevations are all the same except for the pristine Late Pueblo III sites. Discrepancies with these sites occur because there are only three pristine sites during this time period. The sites in the other time periods all seem to reach the same lower elevation. Surprisingly, this is not the single canyon bottom site at 2036 m, but a non-canyon bottom site that is lower.

A Student T test at " =.05 comparing pristine and reinhabited sites showed no significant

180 difference for Pueblo III sites in the Sand Canyon/Goodman Point locality. The same holds for pristine and reinhabited sites in this locality during the Early Pueblo III. Because of the very small number of pristine Late Pueblo III sites – three – no statistical test is appropriate.

Distance to Large Sites

Again, because the distance from the Mesa Verde Area Survey 2000 area to any large site is so large and the number of sites in the survey small, that particular survey is not included in this analysis. Chapin Mesa is also excluded and only Wetherill Mesa, Mockingbird Mesa and

Sand Canyon/Goodman Point surveys are included.

Figures 86 and 87 show the distribution of mean distances from pristine and reinhabited sites to large sites on Wetherill Mesa. Included in Figure 86 is the distribution of all sites, showing that looking at all Pueblo III sites, reinhabited sites are farther away on average than pristine sites. However, that is not the case when looking at the individual large sites. For Spring

House, Double House, Mug House and Kodak House, the pristine sites are all further away from their respective large sites. For Ruin 16, the pristine and reinhabited sites are practically the same. It is only for Long House that the Pueblo III pristine sites are closer to the large site.

Looking at all Pueblo III sites definitely obscures the picture.

Analysis of the relationship of the distances to large sites between pristine and reinhabited

Pueblo III sites and between all Pueblo III sites and both pristine and reinhabited sites might show how sites were positioned during the Pueblo III and determine what, if any, influence the large sites had on the location of the small sites, both new and reinhabited.

Looking at all the Pueblo III sites, a Student T test of pristine vs. reinhabited sites shows

181 a significant difference at "=.05 and p=.005. While comparing the pristine sites to all the sites shows no significant difference, comparing the reinhabited sites to all the sites is significant at p=.04. While looking at all the sites is interesting, it does not really show any propensity for either pristine or reinhabited sites to gravitate toward large sites during the Pueblo III. For

Double House, Kodak House, Long House, Spring House and Ruin 16, there is no significant difference in any of the relations – pristine vs. reinhabited, pristine vs. all and reinhabited vs. all.

At Mug House, analysis of pristine vs. reinhabited site is significantly different at p=.001. The

Figure 88 Comparison of average distances to large sites of pristine and reinhabited sites on Mockingbird Mesa, all sites.

182 relationship between reinhabited sites and all Pueblo III sites is also significantly different at p=.01. For Mug House, the reinhabited sites are moving further from the location of Mug House.

Figure 89 Comparison of distances to large sites for pristine and reinhabited sites in the Sand Canyon/Goodman Point Survey – Sand Canyon and Goodman Point sites only.

It is difficult to interpret the pristine and reinhabited sites for Early and Late Pueblo III because the numbers are so small. For all large sites except Mug House, the numbers are too small to be statistically valid. For Mug House, the relationship of all Early Pueblo III sites to the pristine sites is statistically valid and significantly different at "=.05 and p=.027 however the number of pristine sites is only 18. Again the reinhabited sites are further away from the location of Mug House. During the Late Pueblo III, no comparison can be made between pristine and

183 reinhabited sites as at least one dataset is too small in every case. A few comparisons to total

Late Pueblo III sites can be made, but they are all not significantly different.

Figure 90 Comparison of distances to large sites for pristine and reinhabited sites in the Sand Canyon/Goodman Point Survey – Sand Canyon, Goodman Point and Casa Negra.

On Mockingbird Mesa, simply dividing the Pueblo III sites between pristine and reinhabited sites produces a pristine site count of 30 and a reinhabited site count of 29. Further dividing these sites to form communities around large sites produces numbers that are much too small for statistical analysis. Fig. 88 shows the breakdown between pristine and reinhabited for all

Pueblo III sites. The sites distances are nearly equal and a Student T test at " = .05 shows that

184 the distance relationship between pristine and reinhabited sites, all sites and pristine sites and all sites and reinhabited sites are not significantly different. An attempt to look at Early and Late

Pueblo III sites on Mockingbird Mesa is not statistically valid as the site numbers are too small.

Figures 89 and 90 show the distances to large sites in the Sand Canyon/Goodman Point locality. The analysis was done twice, considering all three large sites and only the two Late

Pueblo III large sites. In all cases the average distance to the large sites is slightly less for pristine sites. However statistical analysis of the case with three large sites is impossible because of small site numbers. Looking at only Sand Canyon Pueblo and Goodman Point Pueblo, the two Late

Pueblo III large sites, there is no significant difference between the distance to pristine and reinhabited sites although the numbers are slightly too low for solid statistical inference.

185 Chapter 8. SUMMARY AND CONCLUSIONS

How are small sites distributed on the landscape and what are the factors that influence this placement? Is there more of a sociopolitical component or do the environmental concerns override all else? Can these settlement pattern questions be answered with a combination of old survey data collected in a past era of archaeological research and new survey data collected according to today’s standards? Can we discern which old surveys are useful and which are not?

The choices made when deciding the location of small sites are seemingly unaffected by the sudden appearance of large aggregated sites during the Pueblo III. Changes in social, economic and political structures undoubtedly occur when large sites appear, but environmental variables influence the locations of the small sites more. Ancestral Puebloans in the Mesa Verde area organize themselves on the landscape in basically the same way regardless of whether large sites exist at the time or are absent.

Working with old survey data can be difficult, but it is possible as long as the validity of the data can be ascertained. A method for comparing old survey data to new survey data shows promise in weeding out unusable information, while enabling researchers to use reliable old data that would be difficult to reproduce. Major systematic error in survey data can be identified by comparison to other, similar surveys.

Using broad cultural phases rather than narrow phases, the dating of the sites is less prone to error, although the data are then coarser. The narrower cultural phases can show some interesting information, but the numbers of data points are often too small for valid statistical analysis. A balance between accuracy and precision is therefore needed. While I show the data for both broad and narrow cultural phases, much of the conclusions and statistical analyses are

186 based only on the broad cultural information.

This study is the first to compare small site data from five full-coverage surveys and to combine them for analysis. It is also the first to use Mesa Verde data from Wetherill Mesa and

Chapin Mesa. Other studies have taken available small sites in the SHPO database (absent Mesa

Verde) and created computer and mathematical models (Kohler and Leeuw 2007b, Ortman, et al.

2007). This study was also undertaken with an eye to the locations of small sites rather than the importance or centrality of large aggregated sites. Much research has been done with the Sand

Canyon data, but most of that work places emphasis on the large sites as community centers

(Adler 1996b, Adler 1996c, Adler and Johnson 1996, Adler and Varien 1991, Bernardini 2005,

Lipe 1994, Varien 2002, Varien, et al. 2007, Varien and Wilshusen 2002a). The Small Site

Testing Program run by Crow Canyon Archaeological Center in the 1990s did look at and excavate small sites in the Sand Canyon/Goodman Point locality (Varien 1990, Varien 1991,

Varien 1999a), however, this was done with an eye to understanding the communities located around the large sites of Sand Canyon and Goodman Point. Other research in the Mesa Verde region also focuses on the large aggregated sites as community organizers (Cordell 1994, Cordell

1996, Kohler 1995, Kohler, et al. 2007, Lekson 1997).

Ancestral Puebloan World – Mesa Verde

In the more than 500 years from Pueblo I to the abandonment of the Mesa Verde region in the 1270s, some things changed and some things remained the same. The Ancestral Puebloans were still agriculturalists raising maize, beans and squash. The sagebrush flats and pinyon juniper forests still existed but parts of the sage were cut for agricultural fields and did not have time to

187 grown back. Ancestral Puebloans still lived in small habitation sites composed of room blocks, kivas, middens and sometimes towers and their homes still faced southward. But large sites appeared on the landscape with upwards of 400 surface rooms and accompanying kivas. Some of these sites surrounded what were once easily accessible, permanent water sources.

On the mesas and flat areas, neighbors’ homes were closer together and many built their room block-kiva homes in cliff alcoves or on the talus slope. Living on the slopes, they built check dams to control water runoff. They still spread their farm fields across the landscape to mitigate the effects of a varying microclimate regime, but field locations were more and more difficult to find. Pinyon nuts, other plants and small animals were still gathered, but the same amounts had to be distributed to a larger population. Hunters traveled further to find mule deer, elk and turkey. The soils were still deep and fertile, but there were more mouths to feed. The goal of keeping three years of maize on hand could not always be met.

Site Populations

What does this analysis say about the four reliable surveys and the numbers of sites on the landscape at any time? When looking at the site populations in the broad cultural phases, as expected, the numbers of sites increase through time. Because these are all settled agriculturalists, this increase indicates that the population increased through time. The

Mockingbird Mesa and Sand Canyon/Goodman Point site populations are very similar in the number of sites and in the rates of change of sites. The Wetherill Mesa survey data do differ slightly. There are more Pueblo I and II sites and fewer Pueblo III sites than the other surveys.

The Wetherill Survey is the only survey that shows a definite decrease during the Pueblo III. This

188 is probably due to the shift to the cliffs and the aggregation of sites at that time. It is not surprising to see higher populations on Mesa Verde during the Pueblo I and II because the mesa is one of the most favorable locations for agriculture in the area.

The narrow cultural periods show a bit more discrimination between the surveys. The

Mockingbird and Wetherill Mesa survey seem to follow each other in shape. The Sand

Canyon/Goodman Point Survey and the Mesa Verde Area Survey 2000 also follow similar paths.

This might indicate a difference in population growth through time between the mesas, the canyonland and valley areas.

Landforms

Regardless of whether large sites existed or not, small sites remained on the mesas and flat areas at all times. Sites slowly move off the mesas and onto the talus slopes. In the Late

Pueblo III the greatest percentage of sites remain on the flatter areas everywhere except on

Wetherill Mesa where cliff sites become a major Late Pueblo III location. Although the Chapin

Mesa data are unreliable, they too shows a sharp increase in cliff sites at this time and share the

Wetherill Mesa geography. It appears that if cliffs are available they are used in the Late Pueblo

III. The narrow cultural phase data show a gradual fall off in mesa sites for all surveys and a gradual increase in talus and cliff sites. The difference in landform distribution appears to be caused by the landforms that are available.

The shift to more talus slope sites occurs before the large aggregated sites appear, so they cannot be the impetus for the move. This move from mesas and flat areas to talus slopes seems to be pushed by the increase in number of sites. Increased population would make

189 agricultural land scarcer, moving off the best farming land would provide more good land to farm. There also might not have been any usable places left on the mesa top, all habitation sites being occupied by others with prior claim to the mesas. Because the Ancestal Puebloans probably farmed in a variety of small fields, moving to the talus slope would not hinder access to those fields. Cliff locations and talus slopes also might have been chosen as defensive positions.

It is clear that Ancestral Puebloans did not change the landform location of their sites because of the existence of large, aggregated sites. In fact, it is difficult to imagine how moving a small site to a cliff location would make it closer to an aggregated cliff site. In most cases, one would have to climb up to the mesa top to travel to the large cliff site and then climb down to the large site, making access much more difficult than if it were from the mesa top or a talus slope.

Elevation

People lived at a wide range of elevations in the Mesa Verde Area beginning at about

1798 m all the way up to 2463 m. With sites located throughout this 765 m range, what is interesting is that within each survey, the means and medians of elevation remain the same through time. Ranges might expand or contract, but the central elevation remains much the same.

It appears that the optimal choice of location is in the center of the range, but that other circumstances, probably increased population, cause people to locate sites either higher or lower on the landscape depending on which land is available and still will produce maize. As more space was required, the habitation sites moved to fill the elevations available where maize could still be grown. The elevations most occupied are the flattest areas. These areas are also where sagebrush naturally grows and are the best areas for agriculture. Because water is not used for

190 irrigation, it does not seem to be a deciding factor in location. Movement up or down always moves to the less flat, more sloping sites, suggesting that no flat land is available. As habitation sites appear in marginal elevations – higher, wetter but cooler or lower, dryer but warmer – other environmental aspects might have been manipulated to accommodate agriculture i.e., check dams on slopes to slow runoff, southward facing slopes to increase temperatures or terracing to provide flat planting areas.

It appears that the movement in elevations, except during Late Pueblo III, was dictated in a way similar to landform choice. While a comparison of elevations to both temperature and precipitation showed no influence of climate on elevation changes, the influence of site population on elevation was strong. As site populations begin to rise, the range of elevations where the sites are located begins to expand and continues to what might be the limits of either geography or the ability to grow crops. The lack of any obvious influence from the temperature and climate record is probably due to the mosaic of microclimates that tiles the areas. No temperature or precipitation re-creation for the total area can really be applied to a landscape that changes rapidly from location to location.

Distance to Large Sites

If large, aggregated sites are influential in organizing the settlement pattern, then when these large sites exist small sites should, on average, move closer to large sites. They do not. On

Wetherill Mesa only the community around Double House moves closer. In the Sand

Canyon/Goodman Point locality, sites through time move toward Goodman Point but not toward

Sand Canyon Pueblo except during Late Pueblo III. On Mockingbird Mesa sites only move

191 slightly toward Seven Tower Ruin. There is no pattern of increasing movement toward large, aggregated sites in any of these surveys. For most communities there is little difference in distance through time; for the remainder, some move further away and others move closer. The two large sites in the Sand Canyon/Goodman Point Survey and the Mockingbird Mesa Survey are permanent water locations and sites might have been located near these locations long before large sites existed.

Certainly all of the small sites surrounding a large site could not have shifted toward the aggregated site because there would not have been sufficient available land close by, but if nearness was desirable some shifting on the landscape should have occurred. The large sites undoubtedly housed a portion of the population, some of the small site locations should have been abandoned and available for use. Site populations do decline in Late Pueblo III for all but

Wetherill Mesa and Wetherill sites decrease in Early Pueblo III suggesting there should have been room for sites to move closer. Also, the distances in these areas are not great, habitation locations could have been located in clusters or rings around large sites with fields further away.

This, however, was not the case.

Reinhabited Sites

The investigation of reinhabited versus pristine sites shows the effects of landform, elevation and large sites on settlement pattern. Rehabitation of sites was a common occurrence, but the extent of rehabitation varied among surveys. Comparison of reinhabited to pristine sites showed that for Wetherill Mesa and the Sand Canyon/Goodman Point locality, a difference between pristine and reoccupied sites exists. The fact that most of cliff sites on Wetherill are

192 pristine sites explains the difference there. The difference in the Sand Canyon/Goodman Point locality suggests that the reinhabited sites are in good locations and the pristine sites are in less than optimal areas because of the higher percentage of pristine talus slope sites. It appears that small site inhabitants went back to the flatter sites to reinhabit them and the less optimal talus and cliff sites were chosen as new sites, perhaps because that was all that was available.

When it comes to elevation, while pristine sites, for the most part, occur in a narrower range of elevations than the reinhabited sites, none of the elevation differences are significant.

The small contraction of the ranges, eliminating the marginal elevations for growing, might have occurred to compensate for the less optimal landform choices available for the pristine sites, but elevation does not seem to change much whether the sites are reinhabited or pristine.

If the large sites are attracting the small sites, and if the pristine sites represent a choice from among a variety of less than optimal locations, then the pristine sites should be, on average, closer to the large sites than the reinhabited sites. In most cases they are not. On Wetherill Mesa, for Mug House, the pristine sites are significantly different, but further away from the large site.

There is no significance in the difference between distances on Mockingbird Mesa or in the Sand

Canyon/Goodman Point locality.

The idea that small sites move toward large sites when large sites appear is not shown to be true. Whether or not large sites are present, the location on the landscape with respect to distance does not change much. This is true regardless of whether one looks at all the sites or compares the difference between reinhabited sites and pristine sites. For the landform and elevation variables, the large sites again do not seem to have any impact on these variables, but

193 rather, the pristine sites are located on less optimal landforms and in narrower elevation bands.

This is more an indication of population pressure than any attraction to the large sites.

Old Surveys

A comparison of the suspect survey to other surveys taken in the area can provide an answer to whether a survey can be used today, even if the data were collected 50 years ago.

Evaluation of the Chapin Mesa data determined that the database developed from the field cards is not useful in any analytical way. When comparing the numbers of sites in the Chapin Mesa survey to four other surveys from the area it became clear that the numbers of sites recorded on

Chapin Mesa do not follow the pattern of sites recorded in other nearby surveys. When I combed all five surveys’ data into a “total” dataset the Chapin Mesa data fell systematically more than one standard deviation from the mean, while the other surveys only varied from the mean more than one standard deviation once or twice. Removing the Chapin Mesa data from the total shows the remaining surveys in total had only two data points more than one standard deviation from the mean.

This site population data suggests that the number of sites located during the Chapin

Mesa survey and the time periods assigned to those sites are incorrect. Further investigation showed that there are other problems with the Chapin Mesa data.

When looking at the landform locations of sites, the Chapin Mesa survey is again an outlier. It has sites located on cliffs only during the Pueblo III time period, while the Wetherill

Mesa survey and Mockingbird Mesa survey – the only other surveys that have cliff locations – show cliff habitation throughout the habitation period. Wetherill Mesa, located closest to Chapin

194 Mesa, also shows a fairly large Basketmaker cliff/alcove habitation. The Chapin Mesa survey shows little habitation on the mesa tops during the Late Pueblo III while all other surveys show at least 20 percent and as much as 80 percent mesa top sites during that time period. Those surveying Chapin Mesa might have found sites only where they expected them to be. The P2 tests on a comparison of Wetherill Mesa and Chapin Mesa indicate that there is, in general, a significant difference between the two surveys in the distribution of sites. The Wetherill Mesa survey is in line with other surveys in the area, so it is assumed that the Chapin Mesa data are unreliable.

On Chapin Mesa the survey data indicate a very low number of sites that were ever reinhabited in Pueblo III. This is true for the total Pueblo III sites as well as for Early and Late

Pueblo III sites. The survey data show that a larger percentage of sites in the survey area were missed entirely, that the sites that were found were dated incorrectly to both final date of occupancy and any earlier occupancy, and that the identification of sites was biased by expectations for time periods and landform. Consequently, it is conclusive that the Chapin Mesa survey is not a usable survey.

There is, of course, the possibility that the Chapin Mesa dataset does reflect the actual habitation history of the mesa, in which case the area had a habitation pattern totally different from that of Wetherill Mesa nearby and any of the other areas studied in this comparison of Mesa

Verde Area surveys. While this is possible, it is not probable because the inhabitants of Chapin

Mesa have the same artifacts, live in the same types of buildings and use the same architecture.

They also live on the same mesa, albeit a different section of it, and experience the same environmental influences. In all ways, they appear to be culturally the same as those people who

195 lived on Wetherill Mesa, Mockingbird Mesa, in the Sand Canyon/Goodman Point locality and at the base of the mesa in the Mesa Verde Area Survey 2000 area.

It should be noted that the Wetherill Mesa Survey, done not many years later using many of the same techniques and in a climate where many of the earlier biases still prevailed, appears to be a perfectly usable survey. Data collected in the Wetherill Mesa survey closely mirrors data collected in later surveys when adjustments for geography and geology are considered.

So, it is possible to use old surveys today to answer questions about settlement patterns and land use in the Mesa Verde Southwest, as long as the old surveys are vetted to test the accuracy of the data. The method of vetting would be to compare site populations and time periods, site locations on the variety of landforms and the prevalence of reinhabited sites during a time period. As long as the surveys compared are in the same cultural and climatic area, this comparison should identify those old surveys that can be used and those that cannot.

Future Work

Certainly consideration of one set of surveys in one area of the Southwest does not ensure that this method of validating datasets will hold for every location in every place. Another sample of sets of data in the Southwestern United States should be tested in the same way, by comparing old survey data to new, somewhat standardized surveys, to see if the method holds true in the area. Similar comparisons could and probably should be made in other geographic areas to see if they can identify datasets that are useful and those that are not.

One problem encountered in this study is the inability to date sites to finer time periods.

In surveys where pottery collections exist or where pottery counts and descriptions were

196 recorded, it should be possible to create finer breakdowns. These pottery collections and counts should be revisited. These phases might show more of a correlation with temperature and precipitation reconstructions. This would also eliminate some of the problems with the differences in scale between temperature and precipitation and site dates. Kohler, et.al. (Kohler, et al. 2007, Ortman, et al. 2007) for their Village Ecodynamics Project used a mathematical model to place sites within ten year time spans. While an approximation, this approach might help to more finely place sites in time.

Although this work explored the influence of population, landform, elevation, temperature, precipitation and the rehabitation phenomenon, there are many more environmental variables that could be investigated. Foremost should be slope and aspect which might be used to compensate for less optimal locations in small site placement. Soil type, distance to water, check dams and Palmer Drought Indices would also shed light on the location of agricultural fields and how that relates to small site location.

Another area for further study is that of micro-level climate data which would clarify the environmental impact on agriculture and site location. Analysis of other surveys, in other parts of the Ancestral Puebloan area would also provide a comparison and help identify the influences of environment as compared to the influence of social and political forces.

Of utmost value would be a resurvey of Chapin Mesa with careful attention to identifying sites and placing them in time periods. Resampling of the other surveys used in this study would ensure the accuracy of those surveys as well. Other surveys have been conducted in other portions of Mesa Verde National Park (Smith 1987) and those surveys should be explored as well.

197 Conclusions

Old survey data should not be discarded or filed in archives never to see the light of day.

With a methodical examination of the information and a comparison to other surveys in the area, determining the reliability of the data is possible.

This study has looked at the influences on small sites and their placement on the landscape and not at the placement of the large sites. The environmental variables investigated influence the site locations far more than the placement of the large sites. Sometimes, the large sites are actually at permanent water sources which are environmental attractors to begin with, and so the attraction might be due to the water that is always there and not the large sites.

Organization on the landscape and political and social organization undoubtedly coexisted in the Central Mesa Verde Area. The inhabitants of small sites had to be influenced by the existence of large, aggregated villages in their midst. However, the large sites existence, while it might have influenced the myriad small-site inhabitants by altering their social and political life, did not alter the locations they chose for their small habitation as much as the numbers and increase of small sites did. By extension, this work should provide insight on small site locations in any area where large aggregated habitation sites appear only sporadically, but small sites dot the landscape through time.

Landform, elevation, temperature, precipitation and site population have more influence on the location of small sites than proximity to large sites. The most important force in the small site settlement pattern appears to be site population. Choices change as the population increases, with those choices tend toward somewhat less ideal locations. As population increases, the slope or accessibility of the land becomes worse, the elevations move further from the optimal mean of

198 settlement elevation. Small sites do not move closer to large sites when they appear. In effect, there are two organizing structures in play, one originating in the large sites and perhaps influencing trade, ritual and the sociopolitical structure of the overall community, and one originating on the landscape, heavily influenced by the environment and by propinquity of other small sites.

199 BIBLIOGRAPHY

1906 American Antiquities Act of 1906 16 USC 431-433. U.S. Government, www.nps.gov/history/local-law/anti1906.htm.

1974 Archeological and Historic Preservation Act of 1974. U.S. Government, Washington, D.C.

2006 Federal Laws and Archaeology: Summary, pp. 1-7. N.C. Archaeology, cita.chattanooga.org/fedlaws.html.

Adams, K. R. and K. L. Peterson 1999 Environment. In Colorado Prehistory: A Context for the Southern Colorado River Basin, edited by W. D. Lipe, M. D. Varien and R. H. Wilshusen. Colorado Council of Professional Archaeologists, Denver, Co.

Adler, M. A. 1988 Archaeological Survey and Testing in the Sand Canyon Pueblo/Goodman Point Ruin Locality, Montezuma County, Colorado, 1987 Field Season. Report submitted to the Bureau of Land Management, San Juan Resource Area Office, Durango, Co.

1990 Communities of Soil and Stone: An Archaeological Investigation of Population Aggregation Among the Mesa Verde Region Anasazi A.D. 900-1300. Ph.D., University of Michigan, Ann Arbor, Mich.

1992 The Upland Survey. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 11-23. Crow Canyon Archaeological Center, Cortez, Co.

1994 Population Aggregation and the Anasazi Social Landscape: A View from the Four Corners. In The Ancient Southwestern Community, edited by W. H. Wills and R. D. Leonard, pp. 85-100. University of New Mexico Press, Albuquerque, N.M.

1996a "The Great Period": The Pueblo World During the Pueblo III Period, A.D. 1150 to 1350. In The Prehistoric Pueblo World A.D. 1150 -- 1350, edited by M. A. Adler, pp. 279. The University of Arizona Press, Tucson, Az.

1996b Land Tenure, Archaeology, and the Ancestral Pueblo Social Landscape. Journal of Anthropological Archaeology 15:337-371.

1996c The Prehistoric Pueblo World A.D. 1150 --- 1350. University of Arizona Press, Tucson, Az.

200 2002 The Ancestral Pueblo Community as Structure and Strategy. In Seeking the Center Place: Archaeology and Ancient Communities in the Mesa Verde Region, edited by M. D. Varien and R. H. Wilshusen, pp. 344. Univeristy of Utah Press, Salt Lake City.

Adler, M. A. and A. Johnson 1996 Appendix: Mapping the Puebloan Southwest. In The Prehistoric Pueblo World A.D. 1150 -- 1350, edited by M. A. Adler, pp. 255-272. The University of Arizona Press, Tucson, Az.

Adler, M. A. and M. Metcalf 1992 Archaeological Survey of Lower East Rock and Sand Canyons, Montezuma County, Colorado. Report submitted to the Bureau of Land Management, San Juan Resource Area Office, Durango, Co.

Adler, M. A. and M. Varien 1991 The Changing Face of the Community in the Mesa Verde Region A.D. 100 -- 1300. Paper presented at the Anasazi Symposium 1991, Mesa Verde National Park, Co.

Aharoni, Y. 1962 The Land of the Bible: A Historical Geography. Translated by A.F. Rainey. The Westminster Press, Philadelphia, Pa.

Alley, R. B. 2008 Personal Communication.

Bernardini, W. 2005 Reconsidering Spatial and Temporal Aspects of Prehistoric Cultural Identity: A Case STudy from the American Southwest. American Antiquity 70(1):5-30.

Billman, B. R. 1999 Settlement Pattern Research in the Americas. In Settlement Pattern Studies in the Americas: Fifty Years since Viru, edited by B. R. Billman and G. M. Feinman, pp. 1- 5. Smithsonian Institution Press, Washington, D.C.

Blinman, E. 2008 2000 Years of Cultural Adaptation to Climate Change in the Southwestern United States. Ambio (14):489-598.

Bradley, B. A. 2000 Personal Communication

201 1987 Annual Report of Excavations at Sand Canyon Pueblo 5MT6765, Montezuma County, Colorado, 1986 Field Season. Crow Canyon Archaeological Center, Cortez, Co.

1992 Excavations at Sand Canyon Pueblo. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 79-98. Crow Canyon Archaeological Center, Cortez, Co.

Breternitz, D. A. 1966 An Appraisal of Tree-Ring Dated Pottery in the Southwest. Anthropological Papers of the University of Arizona. The University of Arizona Press, Tucson, Az.

Breternitz, D. A., J. Arthur H. Rohn and E. A. Morris 1974 Prehistoric Ceramics of the Mesa Verde Region. Museum of Northern Arizona Ceramic Series 5. Northern Arizona Society of Science and Art, Inc., Flagstaff, Az.

Breternitz, D. A., C. K. Robinson and G. Gross 1986 Dolores Archaeological Program: Final Synthetic Report. U.S. Department of Interior, Bureau of Reclamation, Engineering and Research Center, Denver, Co.

Brown, W. L., E. G. Anderson and J. Roy Tuchawena 1952 Observations on Three Varieties of Hopi Maize. American Journal of Botany, 39:597-609.

Chapin, F. 1988 The Land of the Cliff-Dwellers. The University of Arizona Press, Tucson, Az.

Clark, W. A. V. and P. L. Hosking 198 Statistical Methods for Geographers. John Wiley & Sons, New York, N.Y.

Coffey, G. 2006 Reevaluating Regional Migration in the Northern San Juan During the Late Pueblo I Period: A Reconnaissance Survey of the East Dove Creek Area. Kiva 72(1):55-70.

Coffey, G. D. and K. A. Kuckelman 2006 2006 Report of 2005 Research at Goodman Point Pueblo (Site 5MT604), Montezuma County, Colorado. Crow Canyon Archaeological Center, Cortez, Co.

Comrie, A. 2008 Personal Communication.

Connolly, M. 1990 A Historic Land Use Study of Goodman Point. Crow Canyon Archaeological Center, Cortez, Co.

202 1992 The Goodman Point Historic Land-Use Study. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 33-44. Crow Canyon Archaeological Center, Cortez, Co.

Cook, E. R. 2000 Southwestern USA Drought Index Reconstruction. International Tree-Ring Data Bank. IGBP PAGES/World Data Center for Paleoclimatology, NOAA/NGDC Paleoclimatology Program., #2000-053.

Cordell, L. S. 1984 Prehistory of the Southwest. New World Archaeological Record. Academic Press, Inc., Orlando, Fla.

1972 Settlement Pattern Changes at Wetherill Mesa, Colorado: A Test Case for Computer Simulation in Archaeology. Ph.D, University of California -- Santa Barbara, Santa Barbara, Calif.

1994 Introduction: Community Dynamics of Population Aggregation in the Prehistoric Southwest. In The Ancient Southwestern Community, edited by W. H. Wills and R. D. Leonard. University of New Mexico Press, Albuquerque, N.M.

1996 Big Sites, Big Questions: Pueblos in Transition. In The Prehistoric Pueblo World A.D. 1150 -- 1350, edited by M. A. Adler, pp. 228-240. The University of Arizona Press, Tucson, Az.

Cordell, L. S., D. E. Doyel and K. W. Kintigh 1994 Processes of Aggregation in the Prehistoric Southwest. In Themes in Southwest Prehistory, edited by G. J. Gumerman, pp. 109-133. School of American Research Advanced Seminar Series, D. W. Schwartz, general editor. School of American Research Press, Santa Fe, N.M.

Dean, J. S. 1996 Demography, Environment and Subsistence Stress. In Evolving Complexity and Environmental Risk in the Prehistoric Southwest, edited by J. A. T. and B. B. Tainter, pp. 25-56. Santa Fe Institute Studies in the Sciences of Complexity. vol. Vol 24. Addison-Wesley, Reading, Mass.

Doyel, D. E. and S. H. Lekson 1992 Regional Organization in the American Southwest. In Anasazi Regional Organization and the Chaco System, edited by D. E. Doyel, pp. 15-21. Anthropological Papers No. 5. Maxwell Museum of Anthropology, University of New Mexico, Albuquerque, N.M.

203 Dunmire, W. W. and G. D. Tierney 1997 Wild Plants and Native Peoples of the Four Corners. Museum of New Mexico Press, Santa Fe, N.M.

Ermigiotti, P. 1997 The Kiln Conference at Crow Canyon: A Summary Report 1991 to 1996. Crow Canyon Archaeological Center, Cortez, Co.

Errickson, M. 1995 Four Corners Archaeological Project: Archaeological Excavations on Reach II of the Towaoc Canal. Cultural Resource Program: Bureau of Reclamation, Upper Colorado Region, Report No. 20, Salt Lake City, Utah.

Fetterman, J. and L. Honeycutt 1987 Mockingbird Mesa Survey, Southwestern Colorado. Cultural Resources Series 22. Bureau of Land Management, Denver, Co.

Fish, S. K. 2000 Farming, Foraging, and Gender. In Women & Men in the Prehispanic Southwest, edited by P. L. Crown, pp. 197-220. School of American Research Advanced Seminar Series, D. W. Schwartz, general editor. School of American Research, Santa Fe, N.M.

Ford, Richard I. 1981 Gardening and Farming Before AD 1000. Journal of Ethnobotany :6-12.

Givens, D. R. 1992 Alfred Vincent Kidder and the Development of Americanist Archaeology. University of New Mexico Press, Albuquerque, N.M.

Gregory, D. A. 1991 Form and Variation in Hohokam Settlement Patterns. In Chaco & Hohokam: Prehistoric Regional Systems in the American Southwest, edited by P. L. Crown and W. J. Judge, pp. 159-194. Advanced Seminar Series, D. W. Schwartz, general editor. School of American Research Press, Santa Fe, N.M.

Griffitts, M. O. 1990 Guide to the Geology of Mesa Verde National Park. Mesa Verde Museum Association, Inc., Mesa Verde National Park, Co.

Hayes, A. C. 196 The Archeological Survey of Wetherill Mesa, Mesa Verde National Park -- Colorado. Archaeological Research Series #7-A. National Park Service, U.S. Department of Interior, Washington, D.C.

204 1981 A survey of Chaco Canyon Archeology. In Archaeological Surveys of Chaco Canyon, New Mexico, edited by A. C. Hayes, D. M. Brugge and W. J. Judge, pp. 1- 68. University of New Mexico Press, Albuquerque, N.M.

Hayes, A. C. and D. Osbourne 1961 Fixing Site Locations by Radio-Direction Finder at Mesa Verde. American Antiquity 27:110-112.

Hegmon, M. 2002 Concepts of Community in Archaeological Research. In Seeking the Center Place: Archaeology and Ancient Communities in the Mesa Verde Region, edited by M. D. Varien and R. H. Wilshusen, pp. 263-279. University of Utah Press, Salt Lake City, Utah.

Jackson, W. H. and W. H. Holmes 1981 Hayden Survey 1874-1876: Mesa Verde and the Four Corners. U.S.G.S. Survey Report, Ouray, Co..

Jenny, H. 1941 Factors of Soil Formation: A system of quantitative pedology. McGraw-Hill Book Company, Inc., New York, N.Y.

1980 The Soil Resource: Origin and Behavior. Springer-Verlag, New York, N.Y.

Judge, J. 1996 Obituary: James Allen Lancaster 1894-1992. American Antiquity 61(2):265-268.

Kane, A. E. and C. K. Robinson 1988 Dolores Archaeological Program: Anasazi Communities at Dolores: McPhee Village. U.S. Department of Interior, Bureau of Reclamation, Denver, Co..

Kidder, A. V. 1927 Southwestern Archaeological Conference. Science 66:489-491.

1973 An introduction to the Study of Southwestern Archaeology. 7th ed. Yale University Press, New Haven, Conn.

Kohler, T. A. 1992 Field Houses, Villages, and the Tragedy of the Commons in the Early Northern Anasazi Southwest. American Antiquity 57(4):617-635.

205 1995 The Calculus of Self-Interest in the Development of Cooperation: Sociopolitical Development and Risk Among the Northern Anasazi. In Evolving Complexity & Environment in Prehistoric Southwest, edited by J. Tainter and B. B. Tainter, pp. 169-196. Addison-Wesley Publishing Co., Santa Fe, N.M.

Kohler, T. A., C. D. Johnson, M. Varien, S. Ortman, R. Reynolds, Z. Kobti, J. Cowan, K. Kolm, S. Smith and L. Yap 2007 Settlement Ecodynamics in the Prehispanic Central Mesa Verde Region. In The Model-Based Archaeology of Socionatural Systems, edited by T. A. Kohler and S. E. v. d. Leeuw, pp. 290. School for Advanced Research Resident Scholar Book. School for Advanced Research, Santa Fe.

Kohler, T. A. and S. E. v. d. Leeuw 2007a Introduction: Historical Socionatural Systems and Models. In The Model-Based Archaeology of Socionatural Systems, edited by T. A. Kohler and S. E. v. d. Leeuw, pp. 290. School for Advanced Research Resident Scholar Book. School for Advanced Research, Santa Fe, N.M.

2007b The Model-Based Archaeology of Socionatural Systems. School for Advanced Research, Santa Fe, N.M.

Kuckelman, K. A. 1997 Preliminary Results of Testing and Report of 1997 Excavations at Yellow Jacket Pueblo (Site 5MT5), Montezuma County, Colorado. Crow Canyon Archaeological Center.

2000 The Archaeology of Castle Rock Pueblo: A Thirteenth-Century Village in Southwestern Colorado Available: http://www.crowcanyon.org/castlerock. 5/22/09. Crow Canyon Archaeological Center, Cortez, CO.

2003 The Archaeology of Yellow Jacket Pueblo (Site 5MT5): Excavations at a Large Community Center in Southwestern Colorado [HTML Title]. Available: http://www.crowcanyon.org/yellowjacket. 5 May 2008. Crow Canyon Archaeological Center, Cortez, Co.

Kuckelman, K. A. and G. D. Coffey 2007 2007 Report of 2006 Research at Goodman Point Pueblo (Site 5MT604), Montezuma County, Colorado. Crow Canyon Archaeological Center.

LeBlanc, S. A. 1999 Prehistoric Warfare in the American Southwest. University of Utah Press, Salt Lake City.

206 Lee, R. F. 2000 Creating Mesa Verde National Park and Chartering the Archeological Institute, 1906. In The Story of the Antiquities Act. National Parks Service, U.S. Dept. of Interior, http://www.nps.gov/archeology/PUBS/LEE/index.htm.

Lekson, S. H. 1997 Anasazi Communities in Context. In Anasazi Architecture and American Design, edited by B. H. Morrow and V. B. Price, pp. 241. University of New Mexico Press, Albuquerque, N.M.

Lightfoot, R. R., M. C. Etzkorn 1993 The Duckfoot Site, Descriptive Archaeology, Occasional papers of the Crow Canyon Archaeological Center, Cortez, Colorado.

Lipe, W. D. 1992 Introduction. In The Sand Canyon Archaeological Project: A Progress Report, edited by W. D. Lipe, pp. 1-10. Crow Canyon Archaeological Center, Cortez, Co.

1994 Comments on Population Aggregation and Community Organization. In The Ancient Southwestern Community, edited by W. H. W. R. D. Leonard, pp. 141-143. University of New Mexico Press, Albuquerque.

Lipe, W. D., J. N. Morris and T. A. Kohler 1988 Dolores Archaeological Program: Anasazi Communities at Dolores: Grass Mesa Village. U.S. Department of the Interior, bureau of Reclamation.

Lipe, W. D., M. D. Varien and R. H. Wilshusen 1999 Colorado Prehistory: A Context for the Southern Colorado River Basin. Prehistory of Colorado: A Publication Series. Colorado Council of Professional Archaeologists, Denver, Co.

Lister, R. H. 1988 Forward. In Land of the Cliff Dwellers, pp. vii-xxii. University of Arizona Press, Tucson, AZ.

Lister, R.H. and F. C. Lister 1981 Chaco Canyon: Archaeology and Archaeologists. University of New Mexico Press, Albuquerque, New Mexico.

Lucius, W. A. and D. A. Breternitz 1992 Northern Anasazi Ceramic Styles: A Field Guide For Identification. first ed. Publications in Anthropology 1. Center for Indigenous Studies in the Americas, Phoenix, AZ.

207 Mahoney, N. M., M. A. Adler and J. W. Kendrick 2000 The Changing Scale and Configuration of Mesa Verde Communities. Kiva 66(1):67- 85.

McAllister, S. P. and F. Plog 1978 Small Sites in the Chevelon Drainage. Paper presented at the Small Sites Conference, University of Arizona, Tucson, Az.

McGuire, R. H. and D. J. Saitta 1996 Although They Have Petty Captains, They Obey Them Badly: The Dialectics of Prehispanic Western Pueblo Social Organization. American Antiquity 61(2):197- 216.

McNitt, F. 1991 Richard Wetherill, Anasazi, Pioneer Explorer of Southwestern Ruins, A Biography. 7th paper ed. University of New Mexico Press, Albuquerque.

Messer, A. E. 1994 Database of Sand Canyon Study Area Surveyed Sites, Crow Canyon Archaeological Center.

1995 Pueblo III Rehabitation in the Sand Canyon Locality: Population Pressure and the Environment. Masters paper, Penn State, University Park, Pa..

2001a Database of Penn State Mesa Verde Area Survey 2000, University Park, Pa.

2001b Penn State Mesa Verde Area Survey 2000, SHPO, Denver, Co.

2005a Database of Chapin Mesa Survey, Mesa Verde National Park. University Park, Pa.

2005b Database of Wetherill Mesa Survey, Mesa Verde National Park. University Park, Pa.

2006 Database of Mockingbird Mesa Survey, Bureau of Land Management. University Park, Pa.

Moore, B. M. 1978 Are Pueblo Field Houses a Function of Urbanization? Paper presented at the Small Sites Conference, University of Arizona, Tucson, Az.

MVNP 2005 Geology of Mesa Verde, edited by Geocorps America Program, Geological Society of America, pp. 1-7. Mesa Verde National Park, Mesa Verde, Colorado.

208 Neitzel, J. 1999 Examining Societal Organization in the Southwest: An application of Multiscalar Analysis. In Great Towns and Regional Polities in the Prehistoric American Southwest and Southeast, edited by J. E. Neitzel, pp. 81-94. Amerind Foundation New World Studies Series, A. I. Woosley, general editor. University of New Mexico Press, Albuquerque, N.M.

Nordenskiöld, G. 1893 Cliff Dwellers: The Mesa Verde, Southwestern Colorado, Their Pottery and Implements. 1990 Reprint ed. Translated by D. Lloyd Morgan. P.A. Norstedt & Soner, Stockholm, Sweden.

Oppelt, N. T. 1998 Earth Water Fire: The Prehistoric Pottery of Mesa Verde. Oppelt Publications, Greeley, Co.

Ortman, S. G., M. D. Varien and T. L. Gripp 2007 Empirical Bayesian Methods for Archaeological Survey Data: An Application from the Mesa Verde Region. American Antiquities 72(2).

Petersen, G. W., L. K. Fenstermaker, D. L. Mimms and M. L. Imhoff 1986 Remote Sensing of Desert Soils, Final Report to the National Aeronautics and Space Administration. Office of Remote Sensing of Earth Resources, Institute for Research on Land & Water Resources, The Pennsylvania State University.

Peterson, C. E. and R. D. Drennan 2005 Communities, Settlements, Sites, and Surveys: Regional-Scale Analysis of PrehistorIc Human Interaction. American Antiquities 70(1):5-30.

Powell, S. 1983 Mobility and Adaptation: The Anasazi of Black Mesa, Arizona. Publications in Archaeology. Southern Illinois University Press, Carbondale & Edwardsville, Illinois.

Reher, C. A. 1977 Basic Site Data in Settlement and Subsistence Along the Lower Chaco River: The CGP Survey. The University of New Mexico Press, Albuquerque, N.M.

Rohn, A. H. 1977 Cultural Change & Continuity on Chapin Mesa. Regents Press of Kansas, Lawrence, KS.

209 Saitta, D. J. 1997 Power, Labor, and the Dynamics of Change in Chacoan Political Economy. American Antiquity 62(1):7-26.

Salzar, M. W. 2000 Temperature Variability and the Northern Anasazi: Possible Implications for Regional Abandonment. Kiva 65(4):295-318.

Salzar, M. W. and K. F. Kipfmueller 2005 Reconstructed Temperature and Precipitation on a Millennial Timescale from Tree- Rings in the Southern Colorado Plateau, U.S.A. Climatic Change 70(3):465-487.

Sanders, W. T. 1999 Three Valleys: Twenty-Five Years of Settlement Archaeology in Mesoamerica. In Settlement Pattern Studies in the Americans: Fifty Years since Viru, edited by B. R. Billman and G. M. Feinman, pp. 12-21. Smithsonian Series in Archaeological Inquiry, B. D. Smith and R. M. Adams, general editor. Smithsonian Institution Press, Washington, D.C.

Schiffer, M. B. 1987 Formation Processes of the Archaeological Record. University of New Mexico Press, Albuquerque, N.M.

Schlanger, S. H. 1988 Patterns of Population Movement and Long-Term Population Growth in Southwestern Colorado. American Antiquity 53(4):773-793.

1992 Recognizing Persistent Places in Anasazi Settlement Systems. In Space, Time and Archaeological Landscapes, edited by J. Rossignol and L. Wandsnider, pp. 91-112. Plenum Press, New York, N.Y.

Schlanger, S. H. and R. Wilshusen 1993 Local Abandonment and Regional Conditions in the North American Southwest. In Abandonment of Settlements and Regions: Ethnoarchaeological and Archaeological Approaches, edited by C. M. Cameron and S. A. Tomka, pp. 85-98. New Directions in Archaeology. Cambridge University Press, Cambridge, UK.

Sebastian, L. 1992a The Chaco Anasazi. Cambridge University Press, Cambridge, UK.

1992b Chaco Canyon and the Anasazi Southwest: Changing Views of Sociopolitical Organization. In Anasazi Regional Organization and the Chaco System, edited by D. E. Doyel, pp. 23-32. Anthropological Papers No. 5. Maxwell Museum of Anthropology, University of New Mexico, Albuquerque, N.M.

210 Smith, J. E. 1987 Mesas, Cliffs, and Canyons: The University of Colorado Survey of Mesa Verde National Park 1971-1977. Mesa Verde National Park, Mesa Verde, Co.

Snead, J. E. 2001 Ruins and Rivals: The Making of Southwest Archaeology. The University of Arizona Press, Tucson, Az.

Snow, D. R., M. Gahegan, C. L. Giles, K. G. Hirth, G. R. Milner, P. Mitra and J. Z. Wang 2006a Cybertools in Archaeology. Science 311:958-959.

Snow, D. R., K. G. Hirth and G. R. Milner 2006b Envisioning an Archaeological Cyberinfrastructure. Archaeological Record 6(5):15- 16. Steward, J. H. 1972 The Theory of Cultural Change: the methodology of multilinear evolution, University of Illinois Press, Champagne, Il.

Toll, H. W. 1971 Dolores River Archaeology: Canyon Adaptations as Seen Through Survey. Bureau of Land Management, Denver, Co.

Turner II, C. G. and J. A. Turner 1992 The First Claim For Cannibalism in the Southwest: Walter Hough's 1901 Discovery at Canyon Butte Ruin 3, Northeastern Arizona. American Antiquity 57(4):661-682.

Uhler, J. W. 2007 Mesa Verde National Park. In Survey History of Mesa Verde National Park. Hillclimb Media, www.mesa.verde.national-park.com/info.htm.

Van West, C. R., M. A. Adler and E. A. Huber 1987 Archaeological Survey and Testing in the Vicinity of Sand Canyon Pueblo, Montezuma County, Colorado. Bureau of Land Mangement, San Juan Resource Area Office and Colorado Historical Society, Denver, Durango, Co.

Varien, M., W. D. Lipe, M. A. Adler, I. M. Thompson and B. A. Bradley 1996 Southwestern Colorado and Southeastern Utah Settlement Patterns: A.D. 1100 to 1300. In The Prehistoric Pueblo World A.D. 1150 -- 1350, edited by M. A. Adler, pp. 86-113. The University of Arizona Press, Tucson, Az.

Varien, M. D. 1990 1988 Small Site Testing: Preliminary Descriptive Report on the Excavations at Lillian's Site (5MT3936, Roy's Ruin (5MT3930), Shorlene's Site (5MT3918), and Troy's Tower (5MT3951). Crow Canyon Archaeological Center, Cortez, Co.

211 1991 1989 Sand Canyon Project Testing Program: Preliminary Report on the Excavations at Troy's Tower (5MT3951), Catherine's Site (5MT3967) and Stanton's Site (5MT10508). Crow Canyon Archaeological Center.

1997 New Perspectives on Settlement Patterns: Sedentism and Mobility in a Social Landscape. Ph.D. dissertation, Arizona State University, Phoenix, Az.

1999a The Sand Canyon Archaeological Project: Site Testing, On line publication. Version 1. Crow Canyon Archaeological Center, Cortez, Co.

1999b Sedentism and Mobility in Social Landscape: Mesa Verde and Beyond. University of Arizona Press, Tucson, Az.

2002 Persistent Communities and Mobile Households: Population Movement in the Central Mesa Verde Region, A.D. 950 to 1290. In Seeking the Center Place: Archaeology and Ancient Communities in the Mesa Verde Region, edited by M. D. Varien and R. H. Wilshusen, pp. 163-184. University of Utah Press, Salt Lake City.

Varien, M. D., S. G. Ortman, T. A. Kohler, D. M. Glowacki and C. D. Johnson 2007 Historical Ecology in the Mesa Verde Region: Results from the Village Ecodynamics Project. American Antiquity 72(2).

Varien, M. D. and R. H. Wilshusen (editors) 2002a Seeking the Center Place: Archaeology and Ancient Communities in the Mesa Verde Region. The University of Utah Press, Salt Lake City, Utah.

Vivian, R. G. 1990 The Chacoan Prehistory of the San Juan Basin. New World Archaeological Record. Academic Press, Inc., Harcourt Brace Jovanovich, Publishers, San Diego, Ca.

White, T. D. 1992 Prehistoric Cannibalism at Mancos 5MTUMR-2346, Princeton University Press, Princeton, N.J.

Wilcox, D. R. 1978 The Theoretical Significance of Field Houses. Paper presented at the Small Sites Conference, University of Arizona, Tucson, Az.

Wilcox, D. R. and J. Hass 1994 The Scream of the Butterfly. In Themes in Southwest Prehistory, edited by G. J. Gumerman. School of American Research Press, pp. 211-238. Santa Fe, N.M.

212 Willey, G. R. 1953 Prehistoric Settlement Patterns in the Viru Valley Peru #155. Bureau of American Ethnology Bulletin, Smithsonian Institution, Washington, D.C.

1999 The Viru Valley Project and Settlement Archaeology: Some Reminiscences and Contemporary Comments. In Settlement Pattern Studies in the Americans: Fifty Years since Viru, edited by B. R. Billman and G. M. Feinman, pp. 9-11. Smithsonian Series in Archaeological Inquiry, B. D. Smith and R. M. Adams, general editor. Smithsonian Institution Press, Washington, D.C.

Wilshusen, R. H. 1999 Pueblo I. In Colorado Prehistory: A Context for the Southern Colorado River Basin, edited by W. Lipe, M. Varien and R. Wilshusen, pp. 196-241. Colorado Council of Professional Archaeologists, Denver, Co.

2002 Estimating Population in the Central Mesa Verde Region. In Seeking the Center Place: Archaeology and Ancient Communities in the Mesa Verde Region, edited by M. D. Varien and R. H. Wilshusen, pp. 101-120. University of Utah Press, Salt Lake City.

Wilshusen, R. H., M. J. Churchill and J. M. Potter 1997 Prehistoric Reservoirs and Water Basins in the Mesa Verde Region: Intensification of Water Collection Strategies during the Great Pueblo Period. American Antiquity 62( 4): 664-681.

Wilshusen, R. H. and S. G. Ortman 1999 Rethinking the Pueblo I Period in the San Juan Drainage: Aggregation, Migration, and Cultural Diversity. Kiva 64(3):369-399.

Wilson, D. and E. Blinman 1991 Ceramic Types of the Mesa Verde Region, Handout prepared for the Colorado Council fo Professional Archaeologists Ceramic Workshop. Office of Archaeological Studies, Museum of New Mexico, Santa Fe, N.M.

Woodbury, R. B. 1999 Obituaries: Alden Cary Hayes (1916-1998). American Anthropology 101(3):616- 619.

213 Appendix A. MODIFIED DATABASE OF MOCKINGBIRD MESA SURVEY, BUREAU OF LAND MANAGEMENT – HABITATION SITES ONLY State Site UTM Zone UTM Easting UTM Northing Elev Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM P1 EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT0930 12 686310 4148680 1900 mesa Habitation 12 X X 5MT0931 12 684900 4144000 1850 talus Habitation Field House 0 X X X 5MT0932 12 684960 4144220 1850 talus Habitation 4 X 5MT0947 12 684350 4147930 1900 cliff Habitation 8 X 5MT0948a 12 686620 4149010 1930 mesa Habitation 6 X X X X 5MT0949 12 686200 4148870 1900 canyon Habitation 4 X 5MT0950 12 685550 4148190 1900 talus Habitation 12 X 5MT0952a 12 685450 4147720 1900 mesa Habitation 6 X X 5MT0953a 12 685260 4147640 1900 talus Habitation 12 X X 5MT0970a 12 685000 4145500 1900 mesa Habitation 12 X X X X 5MT0971a 12 685070 4145760 1900 mesa Habitation 4 X X X 5MT0972 12 684900 4145780 1900 mesa Habitation 12 X 5MT0986 12 684740 4145120 1900 mesa Habitation 12 X 5MT0987a 12 684720 4145020 1900 mesa Habitation 6 X 5MT0992a 12 684540 4146360 1880 mesa Habitation 6 X X 5MT0993a 12 684480 4145300 1900 mesa Habitation Water Control Check Dam 12 X X 5MT0994 12 684640 4145690 1900 mesa Habitation 6 X 5MT0995a 12 684700 4145850 1900 mesa Habitation 6 X X 5MT0996 12 684840 4145920 1900 mesa Habitation 18 X 5MT0997 12 684990 4146160 1900 mesa Habitation 12 X 5MT0998a 12 684320 4144480 1875 mesa Habitation 6 X X X 5MT0999a 12 684100 4145380 1875 mesa Habitation 6 X X X 5MT1000 12 683600 4145720 1860 talus Habitation 120 X 5MT1512 12 686060 4146060 1900 talus Habitation Water Control Check Dam 50 X 5MT1531a 12 685400 4146180 1900 mesa Habitation 6 X X X X 5MT1532a 12 685280 4146280 1900 talus Habitation 4 X 5MT1534 12 685360 4146650 1900 talus Habitation 12 X 5MT1536 12 685200 4146560 1900 mesa Habitation 12 X 5MT1541a 12 683990 4146000 1870 mesa Habitation 90 X X 5MT1544a 12 684860 4146950 1900 talus Habitation 36 X X X 5MT1548a 12 684640 4146450 1880 mesa Habitation 6 X X X 5MT1549a 12 686700 4147100 1900 mesa Habitation 18 X X 5MT1550a 12 686940 4147260 1900 talus Habitation 24 X X 5MT1554a 12 686640 4146650 1840 mesa Habitation 24 X X 5MT1559 12 686750 4146160 1840 mesa Habitation 6 X 5MT1560 12 686700 4147190 1900 mesa Habitation 4 X 5MT1561a 12 686320 4147390 1900 talus Habitation 6 X X

214 State Site UTM Zone UTM Easting UTM Northing Elev Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM P1 EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT0930 12 686310 4148680 1900 mesa Habitation 12 X X 5MT1580 12 683500 4145050 1860 cliff Habitation 4 X 5MT1581 12 683590 4145090 1850 mesa Habitation 6 X 5MT1583a 12 683720 4145120 1870 cliff Habitation 6 X X X X 5MT1595 12 684200 4143880 1850 talus Habitation 30 X 5MT1598a 12 686980 4148250 1800 talus Habitation 12 X X X 5MT1599a 12 687160 4148320 1900 talus Habitation 6 X X X 5MT1601a 12 687110 4149110 1950 mesa Habitation 12 X X 5MT1602 12 686600 4148610 1890 mesa Habitation 12 X 5MT1604a 12 686740 4149150 1950 mesa Habitation 8 X 5MT1604d 12 686740 4149150 1950 mesa Habitation 24 X 5MT1606a 12 687250 4150470 1950 mesa Habitation 4 X X X 5MT1607a 12 687180 4150020 1950 mesa Habitation 0 5MT1607a 12 687180 4150020 1950 mesa Habitation 18 X X 5MT1609a 12 687400 4150600 1950 mesa Habitation 6 X X X 5MT1610 12 687340 4152020 1950 talus Habitation 6 X 5MT1617 12 686700 4147690 1840 talus Habitation 6 X 5MT1618 12 686760 4147100 1900 cliff Habitation 6 X 5MT1619a 12 686140 4147220 1900 mesa Habitation 12 X X 5MT1620 12 686350 4147000 1890 talus Habitation 6 X 5MT1621 12 686450 4146860 1830 mesa Habitation 6 X 5MT1622 12 685970 4146570 1900 talus Habitation 24 X 5MT1624a 12 685780 4147260 1900 mesa Habitation 6 X X 5MT1626a 12 687110 4148230 1850 talus Habitation 6 X X 5MT3029 12 686500 4145940 1840 mesa Habitation 8 X 5MT3036a 12 686600 4146760 1830 mesa Habitation 12 X X 5MT3048 12 686690 4146440 1840 cliff Habitation 18 X 5MT3049a 12 686760 4146390 1840 cliff Habitation 12 X 5MT3050a 12 686740 4146370 1840 mesa Habitation 8 X X 5MT3052 12 686640 4147130 1900 mesa Habitation 6 X 5MT3059 12 685910 4146980 1900 talus Habitation 12 X 5MT3060a 12 685879 4146940 1900 talus Habitation Water Control Check Dam 12 X X 5MT3062a 12 685840 4146890 1900 cliff Habitation 4 X 5MT3065 12 686270 4146810 1900 mesa Habitation 4 X 5MT3066 12 685670 4146690 1890 mesa Habitation 4 X 5MT3072 12 686170 4146600 1887 talus Habitation 8 X 5MT3082 12 686260 4146470 1875 talus Habitation 6 X 5MT3083a 12 686200 4146430 1900 mesa Habitation Water Control Check Dam 12 X X

215 State Site UTM Zone UTM Easting UTM Northing Elev Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM P1 EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT0930 12 686310 4148680 1900 mesa Habitation 12 X X 5MT3101 12 684900 4144360 1875 mesa talus Habitation 4 X 5MT3103 12 684840 4144480 1875 mesa Habitation 6 X 5MT3105 12 684900 4144060 1850 talus Habitation 4 X 5MT3108 12 684980 4143750 1850 mesa Habitation 4 X 5MT3111 12 684960 4143590 1850 talus Habitation 4 X 5MT3113 12 684830 4143460 1850 mesa Habitation 4 X 5MT3130a 12 684490 4142900 1850 mesa Habitation 12 X 5MT3134 12 684500 4142960 1850 talus Habitation 4 X 5MT3145 12 684458 4144490 1900 mesa Habitation 6 X 5MT3173a 12 684200 4144920 1875 mesa Habitation 0 X X X 5MT3177a 12 684610 4145280 1900 mesa Habitation 4 X 5MT3180a 12 684050 4145260 1875 mesa Habitation 4 X 5MT3214a 12 684860 4145500 1900 mesa Habitation 4 X 5MT3216b 12 685200 4145640 1900 cliff Habitation 4 X 5MT3217 12 685200 4145740 1900 talus Habitation 12 X 5MT3224 12 684280 4145700 1860 mesa Habitation 12 X 5MT3232 12 684970 4145860 1900 mesa Habitation 6 X 5MT3240 12 685210 4146060 1900 talus Habitation 6 X 5mT3255a 12 685550 4147000 1900 talus Habitation 4 X 5MT3262 12 685650 4146420 1870 mesa Habitation 6 X 5MT3282 12 684380 4145800 1900 mesa Habitation 4 X 5MT3283 12 684450 4145940 1900 talus Habitation 8 X 5MT3287a 12 684820 4146280 1900 mesa Habitation 6 X X 5MT3288 12 684860 4146380 1900 talus Habitation 6 X X 5MT3291 12 683700 4145830 1860 mesa Habitation 6 X 5MT3303a 12 685020 4146879 1900 mesa Habitation 0 X X X 5MT3304 12 685040 4147040 1900 mesa Habitation Ceremonial Stone Circle 6 X 5MT3313 12 684920 4148200 1900 mesa Habitation 4 X 5MT3314 12 684820 4148130 1900 mesa Habitation 12 X X 5MT3318 12 684680 4148000 1900 mesa Habitation 6 X X 5MT3319 12 684680 4147890 1900 mesa Habitation 4 X 5MT3320 12 684500 4147780 1900 mesa Habitation 4 X 5MT3322a 12 683860 4147879 1900 mesa Habitation 6 X X X 5MT3323 12 683860 4147870 1900 talus Habitation 4 X 5MT3324 12 684170 4148150 1900 talus Habitation 6 X 5MT3980 12 687330 4150960 1950 mesa talus Habitation 4 X 5MT4386 12 684970 4146560 1900 talus Habitation 6 X

216 State Site UTM Zone UTM Easting UTM Northing Elev Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM P1 EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT0930 12 686310 4148680 1900 mesa Habitation 12 X X 5MT4993a 12 685410 4148420 1900 mesa Habitation 12 X X X X X 5MT5001 12 686860 4149700 1950 mesa Habitation 12 X 5MT5011a 12 684600 4147800 1900 mesa Habitation 4 X X 5MT5018 12 684180 4147760 1900 mesa Habitation 4 X 5MT5823a 12 687250 4149510 1950 mesa Habitation 6 X X 5MT5826 12 687100 4149220 1950 mesa Habitation 12 X 5MT6019 12 683920 4147750 1900 mesa Habitation 6 X 5MT6739a 12 684600 4144040 1850 mesa Habitation 4 X X 5MT6740 12 684600 4144320 1875 mesa Habitation 6 X 5MT6741a 12 684500 4144500 1875 mesa Habitation 12 X X X X 5MT6742 12 684540 4144900 1900 mesa Habitation 6 X 5MT6744a 12 684720 4145660 1900 mesa Habitation 6 X X 5MT6746a 12 685320 4146900 1900 cliff Habitation 8 X 5MT6754 12 687000 4149149 1950 mesa Habitation 6 X 5MT6827 12 686200 4148470 1900 mesa Habitation 6 X 5MT6860 12 687230 4150350 1950 talus Habitation 4 X 5MT6861 12 687270 4150640 1950 talus Habitation 6 X 5MT6958 12 686940 4149560 1950 talus Habitation 12 X 5MT6967a 12 687500 4151010 1950 mesa Habitation 36 X 5MT6967a 12 687500 4151010 1950 mesa Habitation 36 X 5MT6967a 12 687500 4151010 1950 mesa Habitation 36 X 5MT7283 12 685880 4147520 1900 mesa Habitation 6 X 5MT7284a 12 686430 4147850 1890 mesa Habitation 0 X X 5MT7292a 12 686910 4148090 1880 talus Habitation 6 X X X X 5MT7295 12 686000 4147630 1900 mesa Habitation 6 X 5MT7296a 12 686040 4147650 1900 mesa Habitation 12 X 5MT7309a 12 686210 4148050 1900 mesa Habitation 6 X X 5MT7313 12 686380 4148170 1890 canyon Habitation 6 X 5MT7317 12 685700 4147700 1900 mesa Habitation 6 X 5MT7323a 12 686590 4148540 1880 mesa Habitation 4 X X 5MT7336a 12 686920 4147760 1915 talus Habitation 24 X X X 5MT7345a 12 685840 4148870 1900 mesa Habitation 6 X X 5MT7347 12 685820 4148380 1900 talus Habitation 4 X 5MT7351 12 686170 4148690 1900 talus Habitation 4 X 5MT7354c 12 685460 4148090 1900 talus Habitation 6 X X X 5MT7356a 12 685670 4147120 1900 mesa Habitation 6 X X 5MT7358a 12 686040 4147280 1900 mesa Habitation 4 X 5MT7360a 12 686260 4147320 1900 talus Habitation 6 X X X X X

217 State Site UTM Zone UTM Easting UTM Northing Elev Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM P1 EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT0930 12 686310 4148680 1900 mesa Habitation 12 X X 5MT7363 12 686650 4147210 1900 talus Habitation 4 X 5MT7370a 12 685510 4147190 1900 mesa Habitation 6 X X X 5MT7371 12 683400 4145800 1860 talus Habitation 6 X 5MT7372a 12 686900 4147240 1900 talus Habitation 24 X X 5MT7385 12 683680 4145600 1860 talus Habitation 4 X 5MT7386 12 683710 4145500 1860 mesa Habitation 4 X 5MT7395 12 683900 4144900 1865 canyon Habitation Water Control Check Dam 12 X 5MT7400a 12 683340 4144300 1860 talus Habitation 24 X X X 5MT7405a 12 687270 4147400 1900 talus Habitation 36 X X X 5MT7406 12 685240 4147270 1900 mesa Habitation 4 X 5MT8463 12 687150 4149250 1950 talus Habitation 12 X 5MT8465 12 687340 4150380 1950 mesa Habitation 0 X 5MT8510 12 685160 4148420 1900 mesa Habitation 8 X 5MT8511 12 685120 4148300 1900 mesa Habitation 8 X 5MT8529 12 687260 4151520 1950 mesa Habitation 4 X 5MT8534a 12 687600 4151840 1950 talus Habitation 12 X X 5MT8535 12 687220 4149400 1950 mesa Habitation 6 X 5MT8546 12 687000 4149740 1950 mesa Habitation 6 X 5MT8554 12 685060 4147950 1900 talus Habitation 12 X 5MT8557 12 685030 4147790 1900 talus Habitation 12 X 5MT8560 12 685300 4147940 1900 mesa Habitation 4 X 5MT8567 12 687200 4149730 1950 mesa Habitation 6 X 5MT8581 12 684560 4147200 1850 talus Habitation 6 X 5MT8665a 12 684800 4147550 1900 mesa Habitation 4 X X 5MT8774a 12 686080 4147120 1900 talus Habitation 36 X X X

218 Appendix B. MODIFIED DATABASE OF SAND CANYON STUDY AREA SURVEYED SITES – HABITATION SITES ONLY CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC86-001 12 695520 4142080 2073 Mesa Habitation 0 X CC86-003 12 695290 4142060 2073 Mesa Habitation 0 X CC86-004 12 695380 4142090 2073 Mesa Habitation 0 X CC86-005 12 695840 4141200 2073 Mesa Habitation 0 X CC86-009 12 695360 4141660 2073 Mesa Canyon Habitation 10 X X CC86-010 12 695640 4142100 2073 Mesa Habitation 0 X CC86-011 12 695699 4142030 2073 Mesa Habitation 0 X X CC86-012 12 695980 4141200 2109 Mesa Habitation 0 X CC86-014 5MT3918 12 695700 4141960 2115 Mesa Habitation 10 X X CC86-016 5MT3951 12 695700 4140760 2121 Mesa Habitation 0 X X CC86-022 12 696440 4141540 2103 Mesa Habitation 0 X CC86-025 5MT760? 12 695950 4140490 2103 Talus Habitation base camp 0 X CC86-026 12 696500 4141840 2091 Mesa Habitation 0 X CC86-027 12 696440 4141980 2085 Mesa Habitation 5 X X CC86-028 12 696530 4141720 2091 Mesa Habitation 0 X CC86-029 12 696530 4141770 2091 Mesa Habitation 0 X CC86-030 12 696560 4141500 2094 Mesa Habitation 0 X CC86-031 12 696610 4141540 2094 Mesa Habitation 0 X CC86-032 12 696680 4141510 2091 Mesa Habitation Limited 0 X CC86-033 5MT5151 12 695420 4141140 2097 Mesa Habitation 0 X CC86-034 5MT3935 12 695450 4140170 2103 Mesa Habitation 14 X X CC86-036 12 695400 4140450 2115 Mesa Habitation 0 X CC86-037 12 695420 4140320 2120 Mesa Habitation 3 X X CC86-038 12 695380 4140120 2131 Talus Habitation 0 X CC86-039 12 695150 4140150 2124 Mesa Habitation 0 X CC86-041 12 695180 4140450 2109 Mesa Habitation 0 X CC86-042 12 695160 4140560 2109 Mesa Habitation 0 X CC86-043 5MT3943 12 695070 4140550 2115 Talus Habitation 4 X X CC86-044 5MT1670 12 695030 4140750 2103 Mesa Habitation 14 X X X X CC86-045 12 695080 4140360 2115 Mesa Habitation 0 X CC86-046 5MT3823 12 695580 4140720 2121 Mesa Habitation 0 X X CC86-047 12 696860 4141890 2085 Mesa Habitation 0 X X CC86-050 5MT5152 12 695540 4141320 2100 Mesa Habitation 12 X CC86-051 5mt3945 12 694990 4140900 2097 Mesa Habitation 10 X CC86-052 12 697180 4141640 2091 Mesa Habitation 0 X CC86-053 5MT3930 12 697160 4142180 2076 Mesa Habitation 7 X X

219 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC86-055 12 697020 4141890 2085 Mesa Habitation 0 X CC86-056 5MT3932 12 696770 4142120 2079 Mesa Habitation 0 X X X X CC86-057 12 696720 4142120 2079 Mesa Habitation 0 X X CC86-058 12 696730 4141670 2087 Mesa Habitation 0 X CC86-060 5MT3926 12 697200 4141700 2091 Mesa Habitation 0 X X CC86-061 5MT6265 12 697200 4141420 2099 Mesa Habitation 20 X X X X CC86-063 12 695060 4141140 2085 Talus Habitation 0 X CC86-071 12 696280 4141320 2109 Mesa Habitation 0 X CC86-072 5MT3920 12 696160 4141140 2109 Mesa Habitation 6 X X CC86-073 12 697840 4142200 2076 Mesa Habitation 0 X X CC86-075 12 698080 4142140 2079 Mesa Habitation 0 X X X CC86-076 12 698200 4142000 2082 Mesa Habitation 0 X CC86-077 12 698200 4141860 2086 Mesa Habitation Limited 0 X CC86-078 12 698200 4141660 2091 Mesa Habitation 0 X CC86-079 12 698240 4141630 2091 Mesa Habitation 0 X CC86-080 12 698200 4141600 2091 Mesa Habitation 0 X CC86-081 12 698300 4141650 2085 Mesa Habitation 0 X CC86-082 12 698270 4141730 2088 Mesa Habitation 0 X CC86-083 12 698300 4141800 2088 Mesa Habitation 0 X CC86-084 5MT10246 12 698370 4141560 2088 Mesa Habitation 0 X X X X CC86-085 12 698220 4141160 2103 Mesa Habitation 0 X CC86-086 5MT10459 12 698160 4141320 2097 Mesa Habitation 0 X X X X CC86-087 5MT3942 12 695490 4141790 2091 Mesa Habitation 12 X X X CC86-088 12 697940 4141250 2109 Mesa Habitation 0 X X X CC86-090 12 697780 4141220 2096 Mesa Habitation 0 X CC86-092 5MT3923 12 697660 4141360 2101 Mesa Habitation 6 X X CC86-093 5MT3923 12 697680 4141300 2103 Mesa Habitation 11 X X X CC86-094 5MT3923 12 697630 4141240 2103 Mesa Habitation 0 X X X X X CC86-095 12 698800 4141060 2097 Mesa Habitation Limited 0 X CC86-096 12 698680 4140820 2103 Mesa Habitation 0 X CC86-097 12 698580 4140720 2106 Mesa Habitation 0 X CC86-098 5MT3925 12 697820 4140880 2121 Mesa Habitation 60 X X X X X X CC86-100 5MT6724 12 695820 4142150 2073 Mesa Habitation 0 X CC86-101 5MT6723 12 695660 4142140 2071 Mesa Habitation 0 X CC86-103 12 695700 4142280 2067 Mesa Habitation 0 X CC86-104 5MT3936 12 695730 4142580 2073 Mesa Habitation 9 X X CC86-107 5MT766? 12 697040 4140360 2093 Mesa Habitation 3 X CC86-111 5MT6288 12 697460 4141400 2100 Mesa Habitation 0 X X X X

220 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC86-113 12 697420 4141220 2097 Mesa Habitation 4 X X CC86-115 12 697660 4140880 2115 Mesa Habitation 0 X X CC86-119 12 695930 4142560 2048 Talus Habitation 10 X X CC86-121 12 696720 4142180 2079 Mesa Habitation 0 X CC86-123 5MT11338 12 696780 4142240 2076 Mesa Habitation 10 X X CC86-124 5MT765 12 696700 4140300 2072 Mesa Canyon Habitation 400 X CC86-126 12 697430 4139800 2152 Mesa Habitation 4 X CC86-127 5MT3799 12 697400 413800 2152 Mesa Habitation 13 X X X CC86-131 12 697750 4140070 2137 Mesa Habitation 0 X X CC86-132 12 697390 4140050 2137 Mesa Habitation 14 X X CC86-133 5MT3798 12 697380 4140110 2137 Mesa Habitation 18 X X X CC86-135 5MT3927 12 697290 4140050 2137 Mesa Habitation 0 X CC86-136 12 697450 4140250 2128 Mesa Habitation 2 X X CC86-137 12 697480 4140230 2128 Mesa Habitation 0 X CC86-140 12 697260 4140350 2118 Mesa Habitation 6 X X CC86-141 5MT3914 12 697380 4140440 2115 Mesa Habitation 0 X X X CC86-143 12 697780 4140440 2121 Mesa Habitation 0 X CC86-145 12 697420 4140540 2125 Mesa Habitation 0 X CC86-147 12 697460 4140600 2109 Mesa Habitation 0 X CC86-149 5MT3916 12 696320 4140640 2024 Talus Habitation 10 X X CC86-150 5MT3901 12 696240 4140620 2012 Talus Habitation 15 X X CC86-152 5MT3929 12 697020 4140340 2121 Mesa Habitation 8 X X X CC86-156 5MT3949 12 697020 4140500 2118 Mesa Habitation 12 X X X CC86-157 5MT604 12 701400 4142700 2033 Mesa Canyon Habitation 400 X X X CC87-158 5MT3924 12 697590 4142010 2079 Mesa Habitation 0 X X X X CC87-160 12 697440 4141900 2079 Mesa Habitation 0 X CC87-161 12 697300 4141480 2097 Mesa Habitation 0 X CC87-162 12 695850 4142000 2079 Mesa Habitation 0 X X CC87-163 12 695960 4141680 2091 Mesa Habitation 0 X CC87-164 12 696000 4141600 2092 Mesa Habitation 0 X CC87-166 5MT6283 12 696320 4141380 2097 Mesa Habitation 0 X X CC87-168 12 696240 4141810 2085 Mesa Habitation 0 X CC87-169 12 697569 4139650 2152 Mesa Habitation 0 X CC87-170 5MT6268 12 696000 4141400 2099 Mesa Habitation 0 X CC87-171 12 696200 4141450 2099 Mesa Habitation 0 X CC87-172 12 696210 4141550 2091 Mesa Habitation 0 X X CC87-173 12 697350 4139110 2158 Mesa Habitation 2 X X CC87-176 5MT3791 12 697460 4139650 2146 Mesa Habitation 6 X

221 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC87-178 12 697280 4139000 2164 Mesa Near Rim Habitation 0 X CC87-179 12 697340 4138980 2164 Mesa Habitation 4 X X CC87-180 12 697970 4140310 2128 Mesa Habitation 0 X CC87-181 12 698150 4139840 2134 Mesa Habitation 0 X CC87-183 12 698350 4140320 2128 Mesa Habitation 0 X CC87-184 12 698400 4140000 2128 Mesa Habitation 10 X CC87-186 12 697150 4142680 2036 Canyon Habitation field house 0 X CC87-188 5MT3957 12 698000 4142230 2079 Mesa Habitation 15 X X X CC87-189 12 697120 4139920 2118 Talus Canyon Overhang/s Habitation 6 X X X CC87-197 5MT3929 12 697080 4140420 2087 Mesa Habitation 5 X CC87-198 12 697630 4143000 2089 Mesa Habitation 0 X X X CC87-199 12 697650 4138970 2164 Mesa Habitation 0 X X CC87-200 12 697600 4142890 2067 Mesa Habitation 0 X CC87-201 12 697800 4139780 2143 Mesa Habitation 0 X X CC87-202 12 697650 4143100 2067 Mesa Habitation 0 X X CC87-203 12 697650 4139400 2158 Mesa Habitation 0 X X CC87-204 12 697750 4142330 2073 Mesa Habitation field house 0 X X CC87-205 12 697900 4139630 2151 Mesa Habitation 0 X X CC87-206 12 697750 4142280 2074 Mesa Habitation 0 X CC87-207 12 697850 4139050 2158 Mesa Habitation 0 X CC87-208 12 697900 4143010 2060 Mesa Habitation 0 X CC87-209 12 697980 4139130 2151 Mesa Habitation 0 X CC87-210 12 698050 4142570 2076 Mesa Habitation 0 X CC87-211 12 698010 4139500 2146 Mesa Habitation 0 X CC87-212 12 698080 4140560 2128 Mesa Habitation 0 X CC87-214 12 698100 4140470 2128 Mesa Habitation 0 X X CC87-215 12 698760 4140500 2115 Mesa Habitation 0 X CC87-217 12 697480 4143150 2067 Mesa Habitation 8 X X CC87-218 5MT3958 12 697320 4142780 2048 Talus Habitation 16 X X CC87-222 5MT6293 12 698760 4142200 2079 Mesa Habitation 0 X CC87-223 5MT3959 12 697470 4142580 2048 Talus Habitation 8 X X CC87-224 12 698680 4142270 2079 Mesa Habitation 14 X X CC87-225 5MT6291 12 698650 4141700 2088 Mesa Habitation 8 X X CC87-226 12 700640 4143260 2067 Mesa Habitation 0 X X CC87-228 12 700790 4143100 2054 Mesa Habitation 0 X X CC87-229 5MT3960 12 695650 4140000 2048 Talus Habitation 6 X X CC87-230 12 700720 4143160 2054 Mesa Habitation 0 X X CC87-232 5MT3808 12 698880 4140120 2118 Mesa Habitation 0 X X

222 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC87-233 12 699000 4140560 2103 Mesa Habitation 0 X CC87-234 12 699140 4140040 2109 Mesa Habitation 0 X X CC87-235 12 699330 4140060 2103 Mesa Habitation 0 X X X CC87-243 12 698860 4141310 2094 Mesa Habitation 0 X X CC87-244 12 698900 4141000 2094 Mesa Habitation 0 X X X X CC87-244 12 698900 4141000 2094 Mesa Habitation 0 X X X X CC87-245 12 699120 4140690 2103 Mesa Habitation 0 X X CC87-246 12 699040 4141180 2088 Mesa Habitation 0 X CC87-247 12 699200 4140760 2121 Mesa Habitation 0 X CC87-248 12 699370 4140250 2103 Mesa Habitation 0 X CC87-249 12 699310 4140820 2115 Mesa Habitation 0 X X CC87-250 12 699360 4140540 2112 Mesa Habitation 0 X CC87-251 12 699280 4141380 2086 Mesa Habitation 0 X X CC87-252 12 698940 4140120 2115 Mesa Habitation 0 X X CC87-253 12 699540 4140590 2097 Mesa Habitation 12 X CC87-254 5MT3961 12 699450 4140580 2100 Mesa Habitation 0 X CC87-255 12 699580 4140840 2088 Mesa Habitation 0 X X CC87-256 5MT3962 12 699870 4142190 2067 Mesa Habitation 0 X X X CC87-257 12 700300 4142860 2048 Mesa Habitation 0 X X CC87-259 12 700170 4142860 2060 Mesa Habitation 0 X CC87-260 12 700120 4142240 2067 Mesa Habitation 0 X X CC87-261 12 699900 4142830 2063 Mesa Habitation 0 X X CC87-262 12 700320 4142170 2060 Mesa Habitation 10 X CC87-263 12 699870 4142400 2070 Mesa Habitation 0 X X X X X X X CC87-264 5MT3831 12 700550 4142000 2054 Mesa Habitation 0 X CC87-265 12 700300 4142220 2060 Mesa Habitation 0 X X CC87-266 5MT3834 12 700550 4142050 2054 Mesa Habitation 5 X X ? CC87-268 12 700840 4142020 2042 Mesa Habitation 0 X CC87-269 5MT3927 12 701300 4142240 2045 Mesa Habitation 16 X X CC87-270 12 700880 4142200 2048 Mesa Habitation 10 X CC87-276 5MT3825 12 701170 4142260 2048 Mesa Habitation 15 X X CC87-278 5MT3927 12 701010 4142200 2045 Mesa Habitation 8 X X CC87-279 5MT10508 12 697040 4140060 2109 Talus overhang/shelter Habitation 5 X X CC87-280 12 701010 4142200 2045 Mesa Habitation 6 X X CC87-281 5MT3967 12 696940 4140000 2060 Talus Habitation 14 X X CC87-284 5MT10246 12 696760 4141170 2067 Talus Canyon overhang Habitation 9 X X CC87-286 12 696640 4141100 2067 Talus Canyon terrace Habitation 6 X X CC87-288 5MT3842 12 696430 4141020 2097 Talus terrace Habitation 7 X X

223 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC87-303 12 702060 4142460 2030 Mesa Habitation 0 X X CC87-305 12 700900 4144330 2051 Mesa Habitation 0 X X CC87-306 12 702400 4142860 2039 Mesa Habitation 0 X X CC87-307 12 700500 4144000 2048 Mesa Habitation 0 X CC87-308 12 702540 4142440 2036 Mesa Habitation 0 X CC87-309 5MT3814 12 700900 4144000 2048 Mesa Habitation 0 X CC87-310 5MT3963 12 700860 4140760 2057 Mesa Habitation 14 X X CC87-310 5MT3963 12 700860 4140760 2057 Mesa Habitation 14 X X CC87-311 12 700970 4141400 2048 Mesa Habitation 0 X X X CC87-312 12 700830 4140740 2060 Mesa Habitation 0 X X CC87-313 12 700840 4143420 2051 Mesa Habitation 0 X X CC87-314 12 701090 4140910 2054 Mesa Habitation 8 X CC87-315 12 700500 4143250 2057 Mesa Habitation 0 X X CC87-316 12 701180 4140910 2051 Mesa Habitation 11 X X CC87-317 12 701100 4143320 2048 Mesa Habitation 0 X X CC87-319 12 701100 4143180 2048 Mesa Habitation 0 X X CC87-320 12 701190 4141160 2045 Mesa Habitation 3 X X CC87-321 12 701120 4143100 2048 Mesa Habitation 0 X X X CC87-322 12 699700 4143390 2064 Mesa Habitation 0 X X CC87-323 5MT3807 12 701000 4143050 2055 Mesa Habitation 0 X X CC87-324 12 699650 4143080 2073 Mesa Upland Habitation 0 X X X X CC87-324 12 699650 4143080 2073 Mesa upland Habitation 0 X X X X CC87-324c 12 699650 4143080 2073 Mesa upland Habitation 0 X X X X CC87-326 12 702020 4143260 2036 Mesa Habitation 0 X X CC87-329 12 700280 4142910 2061 Mesa Habitation 0 X X CC87-330 12 702070 4143500 2043 Mesa Habitation 0 X X CC87-330 12 702070 4143500 2043 Mesa Habitation 0 X X CC87-330c 12 702070 4143500 2043 Mesa Habitation 0 X X CC87-331 12 701150 4143100 2064 Mesa Habitation 0 X X CC87-332 12 702250 4143430 2040 Mesa Habitation 0 X X CC87-333 12 702620 4143620 2028 Mesa Habitation 0 X X X X CC87-339 12 700640 4140960 2067 Mesa Habitation 2 X CC87-339 12 700640 4140960 2067 Mesa Habitation 2 X CC87-339c 12 700640 4140960 2067 Mesa Habitation 2 X CC87-341 12 700720 4140910 2064 Mesa Habitation 5 X CC87-342 5MT3807 12 700800 4142900 2054 Mesa Habitation 0 X X X X CC87-343 5MT3803 12 700570 4140440 2070 Mesa Habitation 16 X X CC87-344 12 700500 4141300 2061 Mesa Habitation 8 X X

224 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC87-346 12 700800 4141200 2054 Mesa Habitation 0 X X X CC87-347 12 700740 4140810 2064 Mesa Habitation 0 X CC87-349 12 700570 4140680 2073 Mesa Habitation 0 X CC87-353 5MT10247 12 701630 4141840 2012 Talus Canyon overhang Habitation 7 X X X X CC87-355 12 701700 4141500 1997 Talus terrace Habitation 3 X CC87-356 5MT1747 12 700800 4141870 2054 Mesa Habitation 6 X X CC87-357 12 702390 4141890 2031 Mesa Habitation 0 X CC87-358 5MT10248 12 700980 4141680 2018 Mesa Habitation 17 X X X CC87-359 12 702410 4141980 2034 Mesa Habitation 0 X X CC87-360 5MT10249 12 700930 4141680 2048 Mesa Habitation 8 X X CC87-361 12 702660 4141920 2028 Mesa Habitation 0 X CC87-362 5MT10250 12 700900 4141700 2048 Mesa Habitation 12 X X CC87-364 5MT1752 12 700890 4141670 2048 Mesa Habitation 8 X X CC87-365 12 702140 4141050 2031 Mesa Habitation 0 X X CC87-367 12 702170 4140820 2040 Mesa Habitation 0 X X CC87-368 12 701040 4141750 2045 Mesa Habitation 4 X CC87-369 12 702220 4140870 2036 Mesa Habitation 0 X CC87-371 12 702310 4141080 2031 Mesa Habitation 0 X X X CC87-375 12 702750 4140770 2036 Mesa Habitation 0 X CC87-377 12 702640 4140730 2036 Mesa Habitation 0 X CC87-379 12 702740 4140850 2040 Mesa Habitation 0 X X X X CC87-380 12 702210 4141650 2028 Mesa Habitation 0 X CC87-381 5MT10251 12 702100 4141480 2012 Talus Canyon overhand Habitation 8 X X CC87-382 12 702240 4141620 2025 Mesa Habitation 0 X X CC87-384 12 701250 4142910 2048 Mesa Habitation 0 X X X CC87-386 12 701260 4142970 2051 Mesa Habitation 0 X X X CC87-387 12 701730 4142960 2045 Mesa Habitation 0 X X X CC87-388 12 701620 4143400 2030 Mesa Habitation 14 X X CC87-389 12 701610 4143480 2030 Mesa Habitation 10 X CC87-390 12 701160 4143500 2048 Mesa Habitation 10 X X CC87-391 12 701310 4143570 2045 Mesa Habitation 0 X X CC87-392 12 701200 4143990 2043 Mesa Habitation 10 X X X CC87-393 12 701560 4144210 2043 Mesa Habitation 0 X X X X CC87-394 12 701810 4143870 2038 Mesa Habitation 15 X X CC87-395 12 701250 4144380 2049 Mesa Habitation 8 X X CC87-396 5MT3829 12 700510 4142760 2054 Mesa Habitation 14 X X CC87-397 12 701260 4144330 2045 Mesa Upland slope Habitation 8 X X CC87-398 5MT3829 12 700440 4142810 2054 Mesa Habitation 4 X X X X

225 CC Site # State Site # UTM Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown CC87-399 12 701520 4144260 2043 Mesa Habitation 0 X X X CC87-400 5MT3829 12 700500 4142810 2054 Mesa Habitation 6 X X CC87-401 12 700920 4142870 2054 Mesa Habitation 28 X X CC87-402 5MT3829 12 700430 4142860 2054 Mesa Habitation 6 X CC87-403 5MT604? 12 701400 4142870 2045 Mesa Habitation 5 X X CC87-404 12 700840 4142850 2054 Mesa Habitation 27 X X CC87-404c 12 700840 4142850 2045 Mesa Habitation 27 X X CC87-405 12 701200 4140860 2045 Mesa Habitation 40 X X CC87-406 12 700960 4142820 2052 Mesa Habitation 20 X X CC87-407 12 700930 4142330 2048 Mesa Habitation 6 X CC87-408 12 701450 4142830 2043 Mesa Habitation 8 X CC87-409 12 698870 4141730 2092 Mesa Habitation 0 X CC87-410 12 701470 4142790 2040 Mesa Habitation 18 X X X CC87-415 12 700720 4142440 2025 Mesa Habitation 25 X X CC87-416 12 698880 4142200 2079 Mesa Habitation 0 X X X CC87-417 12 700750 4142380 2052 Mesa Habitation 16 X CC87-418 12 699100 4141730 2085 Mesa Habitation 0 X CC87-419 12 700050 4141460 2067 Mesa Habitation 0 X CC87-420 12 699190 4141870 2094 Mesa Habitation 0 X CC87-421 12 700410 4141740 2052 Mesa Mesa Habitation 10 X CC87-422 12 699330 4142010 2094 Mesa Habitation 0 X X CC87-423 12 700300 4141600 2054 Mesa Habitation 0 X X CC87-424 12 699170 4142430 2085 Mesa Habitation 0 X CC87-425 12 700150 4141200 2073 Mesa Habitation 8 X CC87-426 12 700690 4142480 2054 Mesa Habitation 30 X X CC87-427 12 700380 4141640 2052 Mesa Habitation 0 X X

226 Appendix C: MODIFIED DATABASE OF PENN STATE MESA VERDE AREA SURVEY – HABITATION SITES ONLY State Site UTM Z Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT13931 12 710985 4126340 1933 Talus slope Habitation Rubble Mound X X 5MT13932 12 710978 4126372 1933 Mesa slope Habitation Rubble Mound X X 5MT13933 12 710932 4126410 1915 Mesa Habitation 2 X 5MT13934 12 710853 4126370 1917 Mesa Habitation Limited Activity Field House ? X 5MT13935 12 710818 4126505 1921 Mesa Habitation X 5MT13937 12 710660 4126406 1924 Mesa Habitation Limited Activity Field House X X 5MT13939 12 710246 4126425 1908 Mesa Habitation X X 5MT13940 12 710145 4126425 1905 Mesa Habitation 3 X X 5MT13943 12 711069 4126179 1936 Mesa Habitation 4 X 5MT13944 12 711175 4126183 1943 Talus Habitation 6 X X 5MT13945 12 711134 4126202 1938 Talus Habitation X 5MT13946 12 710546 4126200 1924 Mesa Habitation 0 X 5MT13947 12 710501 4126202 1921 Mesa Habitation X X X 5MT13948 12 710335 4126213 1928 Mesa Habitation X 5MT13949 12 710755 4126153 1925 Mesa Habitation X 5MT13950 12 710781 4126131 1925 Mesa Habitation X X X 5MT13951 12 710824 4126059 1941 Mesa Habitation X 5MT13952 12 710824 4126059 1941 Mesa Habitation X 5MT13953 12 710943 4126159 1926 Mesa Habitation Limited Activity Water 4 X X 5MT13954 12 710897 4126121 1942 Mesa Habitation X X 5MT13955 12 710967 4126117 1937 Mesa Habitation Water Control X 5MT13960 12 711074 4126027 1952 Mesa Habitation X X 5MT13961 12 711141 4125983 1949 Talus Habitation X 5MT13962 12 711206 4125830 1955 Canyon Habitation X X 5MT13964 12 711020 4125679 1981 Talus Habitation X 5MT13966 12 710865 4125731 1962 Talus Habitation X X 5MT13967 12 710850 4125755 1943 Talus Habitation X X X X 5MT13968 12 710595 4125743 1925 Talus Habitation X 5MT13969 12 710492 4125694 1948 Mesa Habitation X 5MT13972 12 709900 4126071 1904 Mesa Limited Habitation X 5MT13975 12 710667 4126042 1927 Mesa Habitation X X X X 5MT13976 12 710630 4126067 1931 Mesa Habitation X 5MT13978 12 710009 4126224 1907 Mesa Habitation X X 5MT13979 12 709908 4126251 1899 Mesa Habitation X 5MT13980 12 709932 4126315 1905 Mesa Habitation X X 5MT13981 12 709665 4125930 1910 Mesa Habitation X 5MT13982 12 709643 4125870 1906 Mesa Habitation 8 X

227 State Site UTM Z Zone UTM Easting UTM Northing Elevation Landform 1 Landform 2 Landform 3 Site Type 1 Site Type 2 Site Type 3 Est. # Rooms Archaic BM PI EP2 MP2 LP2 EP3 LP3 Historic Unknown 5MT13931 12 710985 4126340 1933 Talus slope Habitation Rubble Mound X X 5MT13983 12 709779 4125870 1912 Mesa Habitation 16 X X X X 5MT13984 12 709971 4125826 1931 Talus Habitation X 5MT13985 12 710370 4126193 1923 Mesa Habitation X

228 Vita A’ndrea Elyse Messer, 318 W. Prospect Ave.,State College, Pa. 1680, 814-867-1774, 814-883-3307 Education Dec. 2009 The Pennsylvania State University, Ph.D. Anthropology (Archaeology) May 1995 The Pennsylvania State University, M.A. Anthropology (Archaeology) May 1976 Boston University, M.S. Journalism: Science Communication Dec. 1973 Purdue University, B.A. in Science & Culture Work June 2009 - present Senior Science & Research Information Officer Pennsylvania State University, Research Communications, University Relations July 2006 - June 2009 Senior Science & Research Information Officer Pennsylvania State University, Science, Research & Engineering Information Unit, Public Information Sept. 1988-July 2006 Science & Research Information Officer Pennsylvania State University, Science, Research & Engineering Information Unit, Public Information June 1983-Sept 1988 American Society of Mechanical Engineers, Technical News Writer/Editor, Public Information June 1983-Sept 1988 American Society of Mechanical Engineers, Technology writer, Public Information Dec 1979-Sept 1980 Freund Publishing, Tel Aviv, Israel, Editor Sept 1976-July 1979 Bell Telephone Laboratories, Whippany, N.J.,Editor, Technical Publications, Science & Technical Writer June - Sept 1975 Attleboro Sun Chronicle, Attleboro, Mass., Reporter Field Work Summer 2000 In charge of Penn State Mesa Verde Area Survey 2000, 8 week block survey, supervised team of two.. 1987 to Present Crow Canyon Archaeological Center, Cortez., Co, more than 45 weeks in two week segments excavation & lab work. July 1994 Kostenki, Russia, 3 weeks, excavation, St. Petersburg Institute of Material Culture July 1980 Halutza, Negev, Israel, 4 weeks, excavation, Dr. Negeb, Hebrew University. August 1973 Tel Sheva, Field School in Archaeological Technique, Y. Aharoni, Tel Aviv U. Grants & Awards 1998 Hill Foundation Grant, Penn State 1999-2000 NASA Space Grant Fellowship, Pennsylvania Space Grant Consortium, $20,000 2001-2002 NASA Space Grant Fellowship, Pennsylvania Space Grant Consortium, $20,000 2008 Fellow, American Association for the Advancement of Science Publications 2001 Penn State Mesa Verde Area Survey 2000, Montezuma, Colo., Filed with SHPO, Denver. 1995 Pueblo III Rehabitation in the Sand Canyon Locality: Population Pressure and the Environment, unpublished master’s paper, on file Dept. of Anthropology, Penn State. Presentations 2009 “Reinhabited Small Sites in the Mesa Verde Southwest: Location, Location, Location, Society of American Archaeology, April 24. 2009 2009 “It must be true, I saw it on video,” Society of American Archaeology, April 23. 2009 2007 “There and Back Again: Rehabitation of Small Sites in the Mesa Verde Southwest,” Society of American Archaeology, April 25-29, 2007, Austin, Texas. 2007 “Don’t Run & Hide: Meet the Press Head On, Media Relations Workshop,” Society of American Archaeology, April 25-29, 2007, Austin, Texas 2006 “Small Sites, Old Surveys and What they Tell Us,” Society of American Archaeology, Apr. 26-30 2006, San Juan, Puerto Rico 2005 “Sunken Treasure, Pirate Ships and the Media,” Society of American Archaeology, March 30- Apr. 1, 2005, Salt Lake City Utah. 2005 Chair & Discussant, “Stealing the Past,”Society of American Archaeology, March 30- Apr. 1, 2005, Salt Lake City Utah. 2003 “Small site Settlement Patterns in the Mesa Verde Area of Montezuma County: The more things change, the more they stay the same, Society of American Archaeology, April 9-13, 2003, Milwaukee, WI. 2001 “In the Shadow of the Mesa: Block Survey in the lower Montezuma Valley, Co., Society of American Archaeology, April 21, 2001, New Orleans, La. 2001 Chair & Discussant, “When Archaeology Hits the News,” Society of American Archaeology, April 19, 2001, New Orleans 2000 Invited Paper, “Voyeurs on Both Sides: Anthropologists and Journalists,” American Anthropological Association Conf., Nov. 17, 2000, San Francisco. 2000 Panel on Dealing with the Press, “Voyeurs on Both Sides: Anthropologists and Journalists,” American Anthropological Association Conf., Nov. 17, 2000, San Francisco. 1998 Panel on Dealing with the Press, “Voyeurs on Both Sides: Anthropologists and Journalists,” American Anthropological Association Conf., Dec. 4, 1998, 2000, Philadelphia. 1996 “Pueblo III Rehabitation in the Sand Canyon Locality: Environment and Material Culture,” Society of American Archaeology, April 13, 1996, New Orleans.