INTERMOUNTAIN CULTURAL RESOURCES CENTER ANTHROPOLOGY PROGRAM OCUMENTS JAN 1 1995 An Analysis of Variability and Condition of Cavate Structures in Bandelier ^National — J Monument H. WOLCOTT TOLL 1939 Intermountain Cultural Resources Center 1986 Q G-008?-p Professional Paper No. 53 Intermountain Cultural Resources Center Professional Paper No. 53 An Analysis of Variability and Condition of Cavate Structures in Bandelier National Monument By H. Wolcott Toll With contributions by Peter J. McKenna and June Crowder Contribution #3 of the Bandelier Archeological Survey Anthropology Program U.S. Department of the Interior National Park Service 1995 Mission As the Nation's principal conservation agency, the Department of the Interior has responsibility for most of our nationally-owned public lands and natural and cultural resources. This includes fostering wise use of our land and water resources, protecting our fish and wildlife, preserving the environmental and cultural values of our national parks and historical places, and providing for the enjoyment of life through outdoor recreation. The Department assesses our energy and mineral resources and works to assure that their development is in the best interests of all our people. The Department also promotes the goals of the Take Pride in America campaign by encouraging stewardship and citizen responsibility for the public lands and promoting citizen participation in their care. The Department also has a major responsibility for American Indian reservation communities and for people who live in Island Territories under U.S. Administration. CONTRIBUTIONS OF THE BANDELIER ARCHEOLOGICAL SURVEY 1 M ATfflEN, FRANCES JOAN, CHARLIE R. STEEN, AND CRAIG D. ALLEN 1993 The Pajarito Plateau: A Bibliography. Southwest Cultural Resources Center Professional Paper No. 49. Santa Fe. 2 KOHLER, TIMOTHY A. (editor) 1989 Bandelier Archaeological Excavation Project: Research Design and Summer 1988 Sampling. Reports of Investigations No. 61. Department of Anthropology, Washington State University, Pullman. 3 TOLL, H. WOLCOTT 1995 An Analysis of Variability and Condition of Cavate Structures in Bandelier National Monument. Intermountain Cultural Resources Center Professional Paper No. 53. Santa Fe. 4 MATHIEN, FRANCES JOAN 1991 Glimpses into the History of the 1908 Fieldwork at Yapashi, Bandelier National Monument. In Puebloan Past and Present: Papers in Honor of Stewart Peckham, edited by Meliha S. Duran and David T. Kirkpatrick, pp. 121-132. Papers of the Archeological Society of New Mexico: 17. Albuquerque. 5 KOHLER, TIMOTHY A. (editor) 1990 Bandelier Archaeological Excavation Project: Summer 1989 Excavations at Burnt Mesa Pueblo. Reports of Investigations No. 62. Department of Anthropology, Washington State University, Pullman. 6 KOHLER, TIMOTHY A., and MATTHEW J. ROOT (editors) 1992 Bandelier Archaeological Excavation Project: Summer 1990 Excavations at Burnt Mesa Pueblo and Casa del Rito. Reports of Investigations No. 64. Department of Anthropology, Washington State University, Pullman. 7 WHITE, JOSEPH COURTNEY 1992 In the Land of the Delight Makers: An Archaeological Survey in the American West. University of Utah Press, Salt Lake City. 8 KOHLER, TIMOTHY A., and ANGELA R. LINSE (editors) 1993 Papers on the Early Classic Period Prehistory of the Pajarito Plateau, New Mexico. Reports of Investigations 65. Department of Anthropology. Washington State University, Pullman. 9. POWERS, ROBERT P., and JANET D. ORCUTT (editors) In The Bandelier Archeological Survey. Intermountain Cultural Resources Center preparation Professional Paper No. 57. Santa Fe. VI Foreword In 1916 Bandelier National Monument was established by proclamation of President Woodrow Wilson to protect and preserve for public enjoyment and education the large Pueblo settlements and spectacular cave dwellings of the southern Pajarito Plateau. At the time, the monument and its archaeological resources enjoyed considerable national prominence both in the public eye and within the discipline of archaeology, largely as a result of the pioneering explorations of Adolph Bandelier and the later excavations and preservation efforts of Edgar L. Hewett. Since then the monument has ceded much of its prominence in southwestern prehistory, as the focus of archaeological research has shifted to other regions. Although sporadic investigations have occurred over the last 75 years, the extent to which Bandelier has been forgotten is exemplified by the modest number of documented sites. In 1985 fewer than 500 were known in the 51 square miles of the monument. Knowledge of even these was poor at best. The present volume by H. Wolcott Toll represents the third of several National Park Service and Washington State University contributions that report the findings of the Bandelier Archeological Survey. Through these publications we hope to reestablish publicly and professionally the monument's important place in late Pueblo prehistory. The ten-year Bandelier Survey was begun in 1985 with the goal of recovering both research and cultural resource management data, so that the Park Service may better understand and interpret the monument's archaeology, and also better preserve it. These objectives provided the impetus for the present study of cavate architecture, undertaken in 1986 under the able direction of Wolky Toll. Cavates-cave rooms excavated into tuff cliff deposits-occur along the entire eastern slope of the Jemez Caldera from Tsiping in the north to the vicinity of Cochiti Pueblo in the south. Within Bandelier, cavates are concentrated primarily, but not exclusively, in Frijoles Canyon and at Tsankawi, where they are primarily late (post- a.d.1400 to 1600) and contemporaneous with the nearby, large Classic Period pueblos of Tyuonyi and Tsankawi. Outside of these settings, in the prehistoric as well as modern backcountry, cavates are isolated and smaller, ranging from single rooms to modest pueblos, and temporally span virtually the entire Pueblo occupation of the plateau (a.d. 1 190 to 1600). Because these smaller sites were not known in 1986, the study necessarily concentrated on the larger, Frijoles and Tsankawi groups. Although cavates have drawn more popular attention than any other Pajaritan settlement type, they have received comparatively little scholarly study, despite the wealth of architectural features preserved within their walls. Because past work had frequently focused on a few rooms in a particular cavate group, we felt that a new, more expansive approach was necessary-combining recovery of relatively detailed information with a larger, multi-site sample of cavate rooms. Such a study would identify the range of room variability and function, and also document the condition of the rooms as a basis for future preservation. The sample of more than 350 cavate rooms from 4 cavate groups in Frijoles Canyon and 1 cavate Vll group at Tsankawi provides a strong foundation for archaeological inference and preservation planning. Because the focus of the investigation is the architecture of cavate rooms, the study differs in important ways from the 1987-1991 inventory survey. The inventory goal of surveying 40 percent of the park required investigation at the broader scales of component and site, an approach that minimized our ability to examine smaller units, such as individual rooms and their features. In this respect the two studies are at once different and complementary. Four of the five cavate groups recorded in 1986 and reported here were later inventoried by the survey. At these groups the availability of both small- and large-scale architectural data provides a level of architectural documentation rivaled only at inventory sites later excavated by Tim Kohler of Washington State University. Plane table plan and profile views of three of the Frijoles cavate study groups (Groups F, I, and M) prepared during the inventory have been included in the present volume. Original sketches produced during 1986 for the remaining two sites (Group A and Tsankawi) have been retained, since the first of these sites was not included within the inventory sample areas, while the second was recorded, but not instrument-mapped . Rapid, systematic, and comprehensive recording of several hundred rooms containing not only floor and wall features but also roof features is a professional challenge most of us have never faced, but one that Toll has met with ingenuity. One particular problem was how to document hundreds of often dark interior surfaces in a manner that would establish a condition baseline that could be used by future investigators to measure deterioration. Videotaping each room, with a running audio commentary, was the solution. Many field projects should evaluate this technique as a backup to notes, maps, and still photography. Metal-based videotape, now widely available, provides long, nearly archival tape life, and tape digitization ensures a virtually permanent record. Robert P. Powers, Director Bandelier Archeological Survey June 1994 vm Acknowledgments The number of people participating in the cavate recording project was pleasantly small, so that each person's contribution forms a substantial part of the result. The fieldwork was carried out full time by Barbara Mills, Bruce Panowski, and myself. Working with Barbara and Bruce was a genuine pleasure. Barbara was a steady source of thoughtful input, hard work, and good spirits; Bruce accepted his consignment to the field and commuting with his usual good cheer. Especially warm and heartfelt thanks are due to the four volunteers who all traveled great distances in various ways to contribute their skills and time. Bill and June Crowder, Betsy Fuller, and Liz Bayer all helped us accomplish a great deal more than would have otherwise been possible. The park staff was uniformly pleasant and helpful, making Bandelier an even nicer place to work. Ed Greene's rapelling into a few hard-to-reach cavates kept us all in suspense in more than one way. Peter McKenna and Bob Powers conducted surface inventories of the cavates to attempt to date the cavate groups studied. Peter provided the section on ceramics and dating and continued to provide advice on chronology and ceramics during preparation of this report. The Crowders quickly put all their photos and records in good order and June prepared the summary of her detailed rock art study. Bruce Panowski continued to spend a good deal of time on the project after we left the field. He coordinated the data-entry phase, helped to locate and fix problems, and generated the preliminary outputs. He and Tony Tagliaferro also helped bring the data back to the NPS system after its sojourn at UNM. Mary Padilla, Suzette Lopez, Sophia Ulibarri, and Betsy Fuller all struggled nobly with entering the forms into the computer (especially those left- handed ones). Several members of the staff of the University of New Mexico Computing Center were very helpful. Without Mike Prine's help, there is some question as to whether the data could have been transferred to the UNM system at all. Dusty Teaf and Sandy Robinson cheerfully and energetically assisted with tape and disk management problems and SAS questions many times. Robert Preucel made extensive and helpful comments on the draft and shared data and information from his involvement with the Pajarito Archaeological Research Project of the University of California at Los Angeles. Dr. Lys Ann Shore provided a careful, in-depth technical edit of the whole volume. Jerry Livingston drafted the final figures, and, assisted by Ernesto Martinez, helped with interim maps. Anne Goldberg did extensive typing and formatting of tables for the final version, and Sarah Chavez put the whole thing into camera- ready format. Kathleen Havill did the indexing. Bob Powers was involved with the project throughout. As director of the Bandelier Survey, he initiated and administered it. He oversaw the contracts for write-up and completion, read drafts and made suggestions. In spite of ever-increasing duties at the Park Service, Bob continued to put in a great deal of time and thought during the editing and production phases of the report, exhaustively IX checking the text, and carefully redrafting figures. I especially appreciate his having arranged for follow-up contracts to cover my time during final edits. The Office of Archaeological Studies, Museum of New Mexico allowed me to take time off to work on this report on numerous occasions. I was gone or grouchy at several points along its way (they were presumably glad when the two were simultaneous) Mollie Toll and our sons Nick and Spencer deserve thanks for their patience, understanding, and support during all phases of my involvement with the project and its write-up. Spencer was not yet born when the fieldwork was done for this report; he can now ride a bicycle. CONTENTS FOREWORD BY R. P. POWERS vii ACKNOWLEDGMENTS ix LIST OF FIGURES xiii LIST OF TABLES xv 1. INTRODUCTION TO CAVATES AND THEIR STUDY 1 What Is a Cavate? 1 Why Study Cavates? 2 Past Work in Cavate Features 2 The Present Project 14 2. CONTEXT, DESCRIPTIONS, AND CHRONOLOGY 15 The Setting 15 The Sites 17 Cavate Chronology 61 Incorporation of Ceramic Dates and Further Dating Potential 61 Analysis of Surface Ceramics from the Study Areas by Peter J. McKenna 64 3. RECORDING PROCEDURES, GROUP ATTRIBUTES, AND CAVATE CONDITION 77 Recording Procedures 77 Data Manipulation 90 Group Attributes 93 Cavate Condition 99 4. CAVATE AND NONCAVATE FEATURES: DEFINITIONS, DISTRIBUTIONS, AND DIMENSIONS 107 Structural Features 108 Floor Features 138 Wall Features 150 Summary of Detailed Rock Art Study by June Crowder 197 5. PRELIMINARY FUNCTIONAL ANALYSIS OF CAVATE CHAMBERS 201 Feature Co-occurrence 201 Plastering and Smoking 206 Cluster Analysis 206 XI 6. INTERPRETATION AND CONCLUSION 213 Cavate Use and Relationship to Large Surface Pueblos 216 Conclusion 217 APPENDICES 219 1. Forms, Coding, and Materials Collected 221 2. Data Sets, Volume Calculation, Output Listing, and Photographic Data 237 3. Base Information, Threatened Cavates, and Room Stability 245 4. Detailed Listing of Rock Art and Historical Correlation with Chapman by June Crowder 267 5. Chamber Cluster Membership 275 REFERENCES 285 INDEX 293 XII LIST OF FIGURES 1.1. Map showing areas of the Pajarito Plateau in which cavates have been studied, and the extent of the Bandelier Tuff 3 2.1. Map of Frijoles Canyon showing the location of the groups recorded in 1986 . . 18 2.2. Map of the Tsankawi section, including the vicinity of the group of cavates recorded in 1986 19 2.3. Cross-canyon view of Group I 21 2.4. Cross-canyon view of upper Group M, the portion recorded in 1986 21 2.5. Room plan view for the Group A sample 23 2.6. Elevation sketch for the Group A sample 25 2.7. Comparison photos for A-13, 1939-1986 30 2.8. Comparison photos for A-60, 1939-1987 31 2.9. Masonry front wall of A-10, 1986 32 2.10. Room plan view for the Group F sample 33 2.11. Elevation sketch for the Group F sample 35 2.12. Comparison photos for F-31, 1939-1986 38 2.13. Comparison photos for upper Group F, Room 12, 1939-1986 39 2.14. Room plan view for Group I 41 2.15. Elevation sketch for Group I 43 2.16. Comparison photos for 1-22, 1939-1986 46 2.17. Room plan view for the Group M sample 47 2.18. Elevation sketch for the Group M sample 49 2.19. Room plan view for the Tsankawi sample . 55 2.20. Elevation sketch for the Tsankawi sample 57 2.21. Comparison photos for TS-53, 1939-1986 60 3.1. Schematic drawing of a cavate showing examples of several of the79 features recorded 79 3.2. Example of digging stick marks in A-10 82 3.3. Evidence of chamber expansion in M-10 83 4.1. Histogram showing volume distribution for all chambers at Frijoles and Tsankawi 117 4.2. Histogram showing volumes for all chambers with assigned functions 118 4.3a. Histogram comparing volumes for habitations and "kivas" at Frijoles and Tsankawi 119 xiii 4.3b. Histogram comparing volumes for storage rooms at Frijoles and Tsankawi .... 120 4.4. Floor depressions/pot rests and large wall niche in A-47 142 4.5. Floor pit complex in TS-55 144 4.6. Metate rest and floor ridge in M-40 145 4.7. Grinding complex in M-60 146 4.8. Row of loom anchors in M-59 148 4.9. Wooden loom anchor loop in TS-59 148 4.10. Adobe collar and wall niche in F-37 151 4.11. Deflector in upper Group F, Room 15 152 4.12. Large floor-level niches 153 4.13. Large floor-level niche volumes, showing Frijoles and Tsankawi cases 156 4.14. Plot of height above floor and volume of wall niches 160 4.15. Example of a slot in M-59 162 4.16. Three-dimensional plot of viga hole height, depth, and diameter 166 4.17. Three-dimensional plot of "latilla" hole height, depth, and diameter 168 4.18a. Three-dimensional plot of beam seat height, depth, and diameter 169 4.18b. Three-dimensional plot of viga hole and beam seat height, depth, and diameter 172 4.19a. Diameters of cavate indeterminate holes, showing Frijoles and Tsankawi cases 175 4.19b. Depths of cavate indeterminate holes, showing Frijoles and Tsankawi cases . . . 176 4.20a. Three-dimensional plot of height, depth, and diameter of indeterminate holes . . 177 4.20b. Three-dimensional plot of indeterminate holes with diameters and depths of 10 cm or less 178 4.21. Plot showing overlap of viga, latilla, and indeterminate distributions 181 4.22. Narrow wall incisions in TS-24 189 4.23. Cliff niches on extramural cliff at Tsankawi 191 4.24. Rock art, viga holes, and plaster dado in TS-59 192 4.25. Large bird figures beside door to TS-40 193 4.26. Bird figure in F-23 193 4.27. Awanyus in M-13 194 4.28. Plastered-over masks in M-60 194 xiv LIST OF TABLES 1.1. Group M ceramic types reported by Turney (1948) with counts from cavate surface material analysis 10 2.1. Chronometric dates and ceramic associations from Bandelier 63 2.2. Date ranges used by various analysts for Rio Grande ceramic types 2.3. Ceramic samples for Bandelier cavates 68 2.4. Ceramic form and ware data and significance tests 73 3.1. Distribution of noncavate feature types 85 3.2. Time spent and forms completed by cavate group, 1986 89 3.3. Room type and mode of recording by site 94 3.4. Evidence for construction 95 3.5. Estimate of excavated versus natural space (cavates only) 96 3.6. Masonry presence and type 97 3.7. Room level, cavates and noncavates combined 98 3.8. Fill type, cavates only 98 3.9. Fill depth, cavates only 98 3.10. Tuff characteristics, cavates only 99 3.11. Nonhuman use of cavates 100 3.12. Overall stability of cavates and noncavates 100 3.13. Human damage by group and chamber location 102 3.14. Natural damage by group and chamber location 104 4.1. Overall occurrence of cavates, noncavates, and features by group 108 4.2. Occurrence of all individually recorded features in cavates and noncavates, by cavate group (number and percentage) 109 4.3. Chamber occurrence and dimensions 113 4.4. Exterior door occurrence and dimensions 121 4.5. Exterior opening occurrence and dimensions 123 4.6. Occurrence of doors and openings by group and cavate type 124 4.7. Interior door occurrence and dimensions 125 4.8. Natural wall occurrence and dimensions 127 4.9. Plaster coats by group, wall, and function 129 4.10. Mean plaster height and number of coats by group and wall 131 4.11. Plaster color by group 132 4.12. Masonry wall occurrence and dimensions 133 4.13. Floor occurrence and dimensions 135 xv 4.14. Floor plaster coats by group and function 136 4.15. Ceiling occurrence and smoking by group and function 137 4.16. Firepit occurrence and dimensions 139 4.17. Floor pit occurrence and dimensions 141 4.18. Floor depression occurrence and dimensions 143 4.19. Metate rest occurrence by location and group 147 4.20. Co-occurrence of possible mealing complex features 147 4.21. Loom anchor occurrence and dimensions 150 4.22. Large floor-level niche occurrence and dimensions 154 4.23. Wall niche occurrence and dimensions 158 4.24. Co-occurrence of floor-level and wall niches by group 161 4.25. Slot occurrence and dimensions 163 4.26. Viga hole occurrence and dimensions 164 4.27. "Latilla" hole occurrence and dimensions 167 4.28. Beam seat occurrence and dimensions 170 4.29. Indeterminate hole occurrence and dimensions 173 4.30. Results of cluster analysis on round wall holes 180 4.31. Discriminant analysis classification of feature types 180 4.32. Loom support occurrence and dimensions 183 4.33. Smokehole occurrence and dimensions 184 4.34. Vent occurrence and dimensions 186 4.35. Groove occurrence by shape and group 186 4.36. Wall depression occurrence and dimensions 187 4.37. Rock art occurrence and chamber location 195 4.38. Cavates containing rock art by group and motif 198 4.39. Rock art nomenclature 199 5.1. Summary of chamber attribute occurrence 202 5.2. Co-occurrence of feature categories in chambers 203 5.3. Spearman rank-order correlations of feature category co-occurrence 204 5.4. Plaster coat- feature number co-occurrence 207 5.5. Smoking of plaster coats 208 5.6. Means, membership, and correspondence to assigned function for chamber cluster analysis 209 6.1. Frequencies of features by assigned function 214 xvi Introduction to Cavates and Their Study The true character of the so-called "Cavate lodge" has not been fully understood. E. L. Hewett, 1909 The Pajarito Plateau of New Mexico was formed largely by a series of gigantic volcanic ash flows. The ash consolidated into a soft rock called tuff, which was gradually dissected into many deep canyons by runoff from the Jemez Mountains to the Rio Grande. In the sheer cliffs formed by the tuff are literally thousands of "cavates," chambers hollowed out of the tuff by the prehistoric Pueblo people who flourished there in the twelfth through sixteenth centuries. Bandelier National Monument contains several canyons with abundant cavates. This study is the result of a pilot project to investigate this one type of archaeological feature in detail, as part of a larger project to inventory the park's archaeological resources. What Is a Cavate? It seems likely that archaeology provided the word cavate to the language, at least as a noun. Webster's definition of the term is "cut in soft rock: EXCAVATED < -cliff dwelling > ." Recent dictionaries, however, show the word only as a verb meaning "to hollow out" and list its use as "rare"; given this apparent trajectory, it seems archaeologists had better use the term or lose it. It is not known when the term first came into archaeological use, though Mindeleff (1896:217) provides a useful discussion of it as early as 1896: Cavate lodges comprise a type of structures closely related to cliff houses and cave dwellings. The term is a comparatively new one, and the structures themselves are not widely known. They differ from the cliff houses and cave dwellings principally in the fact that the rooms are hollowed out of cliffs and hills by human agency, being cut out of soft rock, while the former habitations are simple, ordinary structures built for various reasons within a cove or on a bench in the cliffs or within a cave. The term seems to have had considerable currency around the turn of the twentieth century (e.g., Powell 1886, 1891; Mindeleff 1896; White 1904; Bierbower 1905; Beam 1906). In this study, cavate features are defined as cavities in the canyon wall that are primarily the result of excavation of the rock. Both Mindeleff and Hewett recognized that there is variability in these features, how they are incorporated into structures, and how they relate to other sites lacking cavates. Hewett (1909a:438), however, contended that "this term is one that should be rejected from the nomenclature of Southwestern archeology." Siding with Mindeleff (1896) and Fewkes CAVATE STRUCTURES (1913:193 Note 1), our position is that cavates are sufficiently different from other archaeological remains to be a useful separate category (see also Hall 1992:23-24). This is especially so in an area such as the Pajarito Plateau where these features abound. Although it is not possible at present to draw the absolute limits of cavate distribution in northern New Mexico, there is little doubt that the line would follow the boundaries of the Bandelier Tuff, which is more or less coextensive with the Pajarito Plateau (Figure 1.1; Bailey et al. 1969; Ross et al. 1961; Mathien et al. 1993). On the north, cavates are present at the Tsiping Ruin at the north end of the Jemez Mountains near Cerro Pedernal (Dougherty 1980:17; Stuart and Gauthier 1981:104). It is unlikely that any cavates exist east of the Rio Grande. They are present as far west as the Jemez River and appear to extend south of 3andelier National Monument (Fliedner 1975; R. Preucel, personal communication, 1988), at least to Peralta Canyon northwest of Cochiti Pueblo. Why Study Cavates? The study of cavates serves two main purposes: adding to archaeological knowledge of these features and assessing their current condition as a means of caring for them. These concerns depend on the basic archaeological goal of recording these features in a systematic way in order to assess their variability. Knowledge of the variability of cavates will help archaeologists and interpreters understand the prehistoric functions of these features and may enable them to date cavates more precisely and determine whether the features changed through time. Cavate structures are unusual in the prehistoric Pueblo record, and they are better suited to recording without excavation than most architectural features. Cavates are much like dry caves, so that many fragile features- including plastering and even organic materials- are remarkably well preserved in cavates. Cavates also preserve information that is seldom available in excavated sites, since ceilings and full walls are present. This permits better estimation of room volume than is usually possible. Cavates have a great deal of archaeological information available for collection with little disruption of deposits and relatively little effort. In spite of the attention cavates have received, there has been surprisingly little formal recording of them. They are so numerous on the Pajarito Plateau that a thorough yet efficient means of recording them is essential to a well-rounded survey of Bandelier and the Pajarito Plateau. On the management level, recording the cavates provides a basis for monitoring them for deterioration and for devising and assessing countermeasures. Cavates attract considerable attention from both visitors and vandals, subjecting them to both casual and intentional attrition. They are also probably more subject to several types of natural degradation than are other archaeological sites: more features are either exposed or very shallowly buried, the tuff is very friable, and there is possibility of collapses of canyon wall segments. Past Work in Cavate Features Cavates in Other Areas Ideally, the cavates found in Frijoles Canyon should be placed in a broad anthropological perspective, but that endeavor is beyond the scope of this study. Rooms excavated out of the rock exist elsewhere in the Southwest and around the world. Studies of cavates in other places may contain information on labor, ventilation, storage, insulation, and maintenance relevant to those of the Pajarito Plateau. Kempe (1988) provides a more global view of cave dwellings, including substantial material on artificial caves, emphasizing Old World sites. Although he includes a chapter on INTRODUCTION San Juan Santa Clara San lldefonso it Tsankawi Pueblo FIG. 2-2 Tshirege Mortendad Canyon FIG. 2-1 -+■ Peaks ^^ Limits of Bandolier Tuff A Cavate Groups in This Study A Cavate Groups Mentioned Modern Pueblos Towa S Keres KS Tewa Santo Domingo o 5 M l l I l 25 30 kilometers _l Figure 1.1. Map of the Pajarito Plateau and environs, showing the locations of study areas (areas of figures 2.1 and 2.2 are indicated by dotted lines) and other recorded cavates, the distribution of the Bandelier Tuff, the Valle Grande Caldera, major streams, and modern pueblos. The elevation of the Rio Grande at the mouth of the Rito de los Frijoles is 1646 m. (Compiled from Bailey et al. [1969]; Dane and Bachman [1965]; Dougherty [1980]; Hyland [1986]; Ross et al. [1961]; USDA Forest Service Santa Fe National Forest Map). CAVATE STRUCTURES the Four Comers area of the American Southwest, he does not mention the cavates of the Pajarito Plateau, in spite of their similarity to excavated structures he discusses in other parts of the world. Some striking parallels to the Pajarito Plateau can be found in Cappadocia, an ancient province in what is now central Turkey. The parallels are primarily geological rather than archaeological: Cappadocia has large expanses of tent rocks formed in a thick layer of volcanic tuff (Blair 1970; Heiken 1979; Fewkes 1910; Kempe 1988; Riboud 1958; Severy 1983:737). Cut into this tuff are thousands of rooms, ranging from small monastic cells to elaborate churches. From a distance many of the smaller rooms look very much like Bandelier cavates, but on the whole they are probably larger and more elaborate and ornate (e.g., Heiken 1979:568-569). Some of these sites are truly grand in scale; for example, "Vardzia numbers at least five hundred rooms and apartments, including chapels, banqueting halls, wine cellars, stables, all connected with a labyrinth of stairs and passages. It is made up of a number of storeys [sic], being cut out of a high vertical cliff face" (Lang 1966:125, also Plate 55). Cosmos Mindeleff (1896) lists four areas of the Southwest where cavates are found: the Rio Grande, the lower San Juan River drainage including the lower Mancos River, the Flagstaff area, and the Verde Valley. Both the San Juan and Verde examples are excavated out of soft sedimentary deposits, while those in the Flagstaff area are dug into "cinders." Mindeleff (1896:217-235) notes that there are thousands of cavates in the Verde Valley, usually in small groups, but sometimes in groups of several hundred rooms. He devoted considerable care to recording some of the cavates in the Verde, including several individual structure plans, measurements, a map of a very large cluster, and photographs. Fewkes (1913:188-193) recorded other cavate rooms further up the Verde drainage, and discussed their relationship to nearby masonry structures. Mindeleff reports rooms as large as 3.7 x 6.7 m and series of rooms extending up to 14 m back into the bluff, far deeper than any artificial caves we encountered on the Pajarito Plateau. Recording of cavates in the Verde Region, desultory since the time of Mindeleff and Fewkes (see Hall 1992:50-66), took a major step toward rigor with Susan Hall's (1992) recording and analysis. Hall's thesis focusses on the Mindeleff Cavate Site, providing plans and elevations for a large number of cavate suites and data on feature types and room sizes. In all, she collected data from 343 rooms in 89 cavate suites (Hall 1992:69). Hall also provides theoretical direction for the architectural study of cavates. Principle differences between the Verde cavates (at least at the Mindeleff Cavate Site) and those on the Pajarito Plateau seem to be that there is less use of exterior masonry on the Verde, and that the Verde examples are larger and have much more complex cavate plans, with many chambers linked together from a single exterior opening. Single chambers with exterior openings are the norm in the cavates we studied, with three chambers the maximum and the exception; in Hall's (1992:104) sample, most suites contained 2 to 5 rooms, and she recorded two cases with 10. Some of these rooms are small enough that we might have called them large niches, but the differences are marked. Quoting the peripatetic Major Powell (1891), Mindeleff (1896:223-224) gives a brief description of two groups of structures east of San Francisco peak near Flagstaff. One of these includes about 150 rooms dug into a cinder dome, the top of which was walled and levelled to form a plaza. The cavate rooms were 3 to 4 m in diameter and 2 to 3 m high, arranged as a larger chamber central to smaller ones. The second group was built around the crater of a larger cone. Here a combination of free- standing masonry, utilized natural caves, and excavated caves formed a village which also had a formal plaza. Powell postulated that the INTRODUCTION builders of these villages were ancestors of the Havasupai. Fewkes (1904:35-39; Whittaker n.d.) describes three different techniques of cavate construction in the Flagstaff area, providing some photos, a plan, and more detail than Mindeleff. At Turkey Tank Caves there are alternating layers of hard lava and breccia and sections of breccia have been removed, with the resulting cavities partitioned with walls. At the New Caves site, a number of cavate rooms were dug into the steep walls of an extinct crater, with masonry rooms in front. At Old Caves, there is an extensive single-story masonry pueblo many rooms of which have a subterranean component. Some of these "cellars" contain several chambers (see Fewkes 1904:37). Colton (1932:25; 1946:37-39) also recorded and mapped Old Caves, noting that "the curious underground chambers in nearly every room hollowed out of the cinders make it unique in pueblo architecture" (Colton 1932:25). He dates these structures to Late Pueblo III, a.d. 1250-1300. Cavates are uncommon in the San Juan area. Prudden (1903:252-253) observes that "these examples [in the Mancos River drainage] are so small or so weathered away that one who should be tempted to make the long journey to the San Juan or lower Mancos for the sake of a study of cavate lodges would risk disappointment, especially in view of the more extensive, varied, and typical groups in the Verde valley ... or those in the valley of the Rio Grande upon the eastern slope of the Valles." Prudden includes a photograph of a "weathered remnant of one of the cavate dwellings, showing the soft shale in which the shallow shelters were dug" (Prudden 1903: Plate XXX). Pajarito Plateau For a variety of reasons, people have been in and out of cavate structures on the Pajarito Plateau ever since their abandonment. Hendron (1943) found evidence of reuse in the seventeenth century (ca. Pueblo Revolt?) and later. Both the Tewa and the Keres have clearly used the area for centuries, and there can be little doubt that the caves were visited periodically after the large sites were abandoned and before there was a significant non-Indian presence in the area. Extensively recorded visits probably began with Adolph Bandelier, who first visited Frijoles Canyon in 1880, making collections and observations. On his second visit in December, he slept in cavates in Alamo Canyon and near Tyuonyi. He found them quite comfortable in spite of some cold weather (see Lange and Riley 1966, especially 225-228). He appears to have selected large chambers in which to stay: he gave two height measurements of more than 2 m. He referred to the cavate where he stayed in Frijoles as the Room of the Cacique. At that time the rooms were well preserved and the ruins in general "very rich in fine fragments of pottery and manos"; the pottery was "prevailingly glazed." He visited many cavates and measured several of them. Many of his observations still hold, though artifacts and masonry are now considerably less abundant. The rooms are remarkably well preserved in most cases, and much stonework used. The goats have filled them with their dung. . . The floor is perfect in most cases, also the yellow plastering. The ceiling is generally smokey [sic] and sooty. . . There are, lower down, several of these large circular rooms like our present quarters. Were they estufas? The Indians say not; they are all houses, and the estufas were those below in the valleys. . . Every room has its fireplace, except such as were evidently used as storehouses. The ruins are in groups, and the deep recesses and reentering angles of the cliffs are avoided. CAVATE STRUCTURES ... In general the rooms of the eastern half [presumably of the cavate group in which they were staying] are larger than those of the western section, there are even a number of very large ones. They are all plastered yellow, and smoky above. . . Many of the rooms contain carved walls, but while the carvings may have been made by Indians, they are certainly posterior to abandonment of the caves, as they are carved in the plastering. (Lange and Riley 1966:226- 228) Bandelier returned in 1885, checking his descriptions for Die Koshare (later translated and expanded into The Delight Makers). At that time there were three ranchos in the valley. One of his companions, named Pacifico, "found a black olla, entire, in one of the caves; also two stone axes" (Lange et al. 1975:76). They proceeded further north, noting large artificial caves in "Canada Ancha" and many caves at "Tzirege." In 1886 Bandelier visited Puye, where he further showed his interest in caves: On the whole they are an exact repetition of the Rito, only their situation is different. They are plastered with yellow clay, and there is a smoke escape cut out above the doorways. Floors are everywhere black and about two inches thick. Many holes for beams. On the average, they are only one story, but I saw two and three stories also. There are also beam holes indicating porches in front of the rock. The caves are singularly distributed, and they are high, over all timber and plainly visible at a great distance. It is a good position for defense and watch. (Lange et al. 1975:160) Bandelier detailed his vision of life and social organization in the cavates in The Delight Makers (1971), published in 1890. At about the same time that Bandelier was pursuing his investigations, members of the Bureau of American Ethnology visited cavates on two occasions and briefly stated their findings in annual reports of that organization. James Stevenson spent a month in Frijoles Canyon in 1882. In many of the caves which were examined a flooring of fine red clay, very neatly and smoothly spread in several thin layers, is still seen, as also a plastering of red or yellow clay upon the walls. In some of them the lower part of the wall is of one color and the upper part and ceiling of another, the two colors being separated by a broad line of dark brown or black which runs around the cave about two feet from the floor. In the walls were found small niches. Beneath some of these caves, which were situated higher in the face of the cliff, were evidences of the former existence of annexed exterior chambers below. The cliff walls beneath these apertures had evidently been hollowed out to form the rear wall of the annexed chamber, and were nicely plastered with red and yellow clay. Rows of small round holes were seen which, it was thought, had been used as rests for the rafters, while large quantities of roughly squared stones used in building lay scattered about the base of the cliff. In some cases there appeared to be two and even three tiers of houses constructed in this manner. (Powell 1886:xxxvi- xxxvii) J. W. Powell, the director of the bureau, visited the cavates (which he called by that term) in 1886. Powell (1891:xxi-xxiv) devoted several pages to discussing cavates. He considered them to be dwellings reached either by ladders or artificial terraces. He noted the presence of firepits, niches, and abundant potsherds. He INTRODUCTION made some chronological interpretations that differ from those now held: On more careful survey it was found that many chambers had been used as stables for asses, goats, and sheep. Sometimes they had been filled a few inches, or even two or three feet, with the excrement of these animals. . . Altogether it is very evident that the cliff houses have been used in comparatively modern times; at any rate since the people owned asses, goats, and sheep. The rock is of such a friable nature that it will not stand atmospheric degradation very long, and there is abundant evidence of this character testifying to the recent occupancy of these cavate dwellings. . . Every mesa had at least one ancient pueblo upon it, evidently far more ancient than the cavate dwellings found in the face of the cliffs. It is, then, very plain that the cavate dwellings are not of great age; that they have been occupied since the advent of the white man, and that on the summit of the cliffs there are ruins of more ancient pueblos. . . It was at once noticed that the potsherds of these cliff dwellings are, both in shape and material, like those now made by the Santa Clara Indians. . . While encamped in the valley below, the party met a Santa Clara Indian and engaged him in conversation. From him the history of the cliff dwellings was soon obtained. His statement was that originally his people lived in six pueblos, built of cut stone, upon the summit of the mesas; that there came a time when they were at war with the Apaches and Navajos, when they abandoned their stone pueblos above and for greater protection excavated the chambers in the cliffs below; that when this war ended part of them returned to the pueblos above, which were rebuilt; that there afterward came another war, with the Comanche Indians, and they once more resorted to cliff dwellings. At the close of this war they built a pueblo in the valley of the Rio Grande, but at the time of the invasion of the Spaniards, their people refused to be baptized, and a Spanish army was sent against them, when they abandoned the valley below and once more inhabited the cliff dwellings above. Here they lived many years, until at last a wise and good priest brought them peace, and persuaded them to build the pueblo which they now occupy-the village of Santa Clara. . . It is therefore evident that the cavate dwellings of the Santa Clara region belong to a people still extant; that they are not of great antiquity, and do not give evidence of a prehistoric and now extinct race. (Powell 1891:xxiii-xxiv) In the early twentieth century a visit to the Pajarito Plateau became a vacation adventure. Susan Bierbower (1905) reported her visit to Puye. Her Santa Clara guide was apparently an astute, early cultural-resource conservationist, and Bierbower herself something less than a cultural relativist: The next morning . . . armed with a good staff and my kodak [sic], I again ascended to the dwellings. We had provided ourselves with a pick and shovel for excavating, and you can judge of our disgust when we ascertained that Juan had left them in Santa Clara. It is my firm belief that this was done with malice aforethought, as we learned that the Indians are very superstitious and unwilling to disturb these places. A small trowel was all we had. (Bierbower 1905:232) Beam (1906) produced a more widely circulated report on the Pajarito, including 8 CAVATE STRUCTURES several photographs of cavates at Puye (despite his title, "The Prehistoric Ruin of Tsankawi"). Both Bierbower and Beam concluded that the use of the cavates and pueblos in the region had ceased due to cataclysms, respectively earthquakes and "fierce and implacable enemies" or perhaps earthquake or eruption (Beam 1906:813). The work done by Hewett in 1908 and 1909 in Frijoles Canyon included "clearing" cavate rooms, for some of which he gives mea- surements and locations (Hewett 1909a,b). Hewett also defined 13 groups of cavate/talus sites (A-M) along the north side of the canyon. These designations have been used by most later workers in Frijoles Canyon, and they are shown on the remarkable and still useful map and ele- vation prepared by Kenneth Chapman and published both by Hewett (1909a, 1938) and by Hendron (1940). Hewett discussed cavates in several places. In the first of two 1909 papers in American Anthropologist he presented some ex- ceptional panoramic photos of large segments of the north wall of Frijoles, showing three of the groups in which our crew later worked (A,F, and I). He also reproduced Chapman's recon- struction of Long House and the interior of a cavate in use. An interior photo illustrated many features of cavate rooms, such as digging stick marks, differential plastering leaving a band, a large floor-level niche, and a possible upper loom support. In the second article (Hewett 1909b) Hewett gave plans and mea- surements for some cavates in the eastern part of Group E (Sun House), but not for the others that he dug in the western part of Group E (Snake Village). Aside from a burial in the Snake Village portion, he made no mention of the con- tents or features of these rooms. He gave more attention to two "kivas," larger, heavily smoked rooms with loom anchors in the floor. The Snake Village received its name from an Awanyu painted on the plaster of the associated "kiva." Hewett also partially excavated an even larger cave kiva in the area but did not give its exact location. Hewett's later publication (1938) contains much of the material in the 1909 art- icles; in addition, it presents a sort of develop- mental sequence based on his estimation of quality of workmanship, in which a chrono- logical element is implied but not specified. The following discussion of burial placement suggests what Hewett may have found but did not report in detail: Crypt or cave burial was here secon- dary. Mortuary crypts were posterior chambers to pueblo-like cliff dwellings. They were receptacles for great quan- tities of disjointed bones, the rooms being filled with these unrelated remains to a depth of several feet. No utensils accompanied them. I consider these crypts to have been depositories for bones removed from, or washed out of, the cemeteries above. In individual cave burial as practiced in this region the dead are found in embryonic position and usually wrapped in feather robes or matting of yucca fiber. (Hewett 1938:134-135) Chapman (1916, 1938) made a survey of the "cave art" of the "region of the Rito de los Frijoles;" he gives a breakdown of 106 prehistoric "works" by subject. He seems to have kept more extensive records as well, and he mentions an illustrated presentation, published at least in part as an appendix to Hewett's (1938) book on the Pajarito (see also Chapter 4 and appendix 4 of the present study). Chapman was especially intrigued by naturalistic figures and scenes scratched into plaster. While Bandelier attributed this style of rock art to postoccupational visits, Chapman clearly thought that at least some examples were done by residents of the cavates. Hewett and Morley also worked at Puye, where a large number of cavate rooms are INTRODUCTION located below the mesa-top room blocks (Hewett 1908; Morley 1910). Stewart Peckham, formerly of the Laboratory of Anthropology, has devoted considerable effort to assembling Morley's notes from the Puye work, but in all his searches he has found no notes relating to the cavates (S. Peckham, personal communication, 1987). M. R. Harrington of the Southwest Museum worked in cavates in 1926; Peckham has found a map of several cavate rooms on three levels, with masonry rooms in front of the first level. Once again, however, extensive research and inquiry have revealed no further notes. Harrington's map shows floor ridges (which he calls sleeping ridges), "cook tables," firepits, partitions, vents, and three burials, one in a small, sealed back chamber. Cavates were brought to the attention of the profession in other contexts in the early twentieth century. J. W. Fewkes made several references to Pajarito cavates in his presidential address to the Anthropological Society of Washington and included some photographs of cavates (Fewkes 1910). W. B. Douglass (1917) presented several sketch plans and sections of cavates in a paper for the nineteenth Congress of Americanists. His profiles show several cavate features, some quite common and some, such as benches and altars, rare. Hendron (1940, 1943) did some of the most carefully recorded work in cavates. He excavated five masonry rooms and the four cavates associated with them in the center of Group M, above the NPS Residence Area. His intention was to thoroughly examine some structures in order to be better able to stabilize them. The rooms contained roofing material, cow bone, and Tewa blackware, indicating historic (ca. 1680 Pueblo Revolt) use of these particular structures. The uppermost walls were underlain by other wall alignments, some of which may have been used as foundations for the later walls. Hendron was very impressed by the fragility of the natural canyon walls, which led him to emphasize the danger of living in cavate rooms and to suggest that the walls probably receded very rapidly. He believed that the rooms visible at that time may have been preceded by earlier rooms later obliterated by erosion. Although the Bandelier Tuff is clearly a very soft material, it seems probable that Hendron overestimated the rate of cutting (see Carlson and Kohler 1989:59). Forty-five years after he worked there, the structures showed relatively little change. Hendron estimated the period of use of the sites as 400-500 years. He described firepits, plaster "dados," smoke vents, a basalt threshold, and depressions in the floor, which he said suggest sleeping spaces. Given the smoke blackening and the method of ventilation, Hendron was inclined to doubt that the cave rooms were used for habitation. He made some extremely nice drawings and sections (now in the park archives), but said little about recovery of cultural material. The rooms he stabilized are clearly visible today. Maxon (1969) reported a single tree-ring date of a.d. 1493 from Group M. In his master's thesis, J. F. Turney (1948) wrote up the artifacts from Hendron's excavation as well as some further material from the excavation of a drainage trench in Group M. He noted that "it has been necessary to remove this material from its unfinished status and bring it to a conclusion as an aid to further research" (1948:i). The majority of the thesis is devoted to a discussion of pottery classification and description of types; the types identified by Turney are listed in Table 1.1. In connection with the present study, surface sherd counts from the same area were conducted and are presented in Table 2.3. The fauna! material includes deer, bison, turkey, and bear bone and a few worked pieces. Manos, metates, and axes were recovered, as well as bifaces and at least 10 projectile points. Obsidian is the most abundant chipped stone material (though Turney stated that "obsidian is brittle and not too well-suited for chipping" [1948:64]). Though the perishable materials 10 CAVATE STRUCTURES Table 1.1. Group M Ceramics Reported by Turney (1948) with Counts from Cavate Surface Material Analysis (See Table 2.3). Ceramic Type Turney Group M Sample Frijoles Sample Santa Fe Black-on-white 57 5 16 Wiyo Black-on-white 24 1 3 Abiquiu Black-on-gray 30 - - Bandelier Black-on-gray a 26 62 b Sankawi Black-on-cream 56 Tewa Polychrome 61 - - Agua Fria Glaze-on-red 2 4 9 Cieneguilla Glaze-on-yellow 5 1 1 Glaze B yellow 1 1 2 Glaze C 5 1 5 Glaze D 11 7 34 Glaze E 15 3 7 Glaze F 5 - - Zia Glaze 4 - - Early glaze 32 - - Middle glaze 203 - - Late glaze 55 - - Zia Polychrome 13 - - Kapo polished blackware 73 1 1 Culinary c 469 1144 Total 652 519 1284 •Not given; presumably abundant: shown as "major occupation." b Biscuit B. c "By far the largest type"; "culinary ware is of little value" (1948:47). INTRODUCTION 11 seem to have come from the masonry rooms rather than the cavates, they are abundant and remarkably well preserved. The inventory includes a digging stick, two weaving tools, two bow fragments, several arrow shafts, "a meager three" fire hearths, some carved sticks, a cradle board, and other wood attributed to craft wastage. Basketry, cordage, yucca fiber, feathers, a feather blanket, and a piece of woolen textile were also present. Turney listed 24 pieces of leather, including moccasin fragments, sewn buckskin, and a thong. A small bowl of tobacco was found, analyzed, and partially smoked(!) by Hendron (Hendron 1946). Also recovered were corn plant parts (bundles of leaves, more than 600 cobs) and cucurbit peduncles and rinds. Following Mera (1932, 1934) and presumably Hendron, Turney concludes that a reoccupation, possibly during the Pueblo Revolt, was probable. The types of perishable materials and the degree of preservation indicate that they may relate to the later occupation suggested by some of the pottery types. In 1939 and 1940, R. H. Lister conducted extensive stabilization work in Bandelier. Much of his time was spent at Long House (Lister 1939), but he also worked at Otowi and Tsankawi, west of Camp Hamilton, in Pueblo Canyon, and along the Rito de los Frijoles. In the area outside Frijoles he located 567 caves and worked on more than half of them. Most of his work consisted of building dams at the entrances of the cavates or diversions above them to prevent water from running through them. In Frijoles Canyon he did similar preventive maintenance and rechinked some masonry walls. Lister's dams are difficult to see today, but they are probably doing their job since water damage seems to be less a problem than human impact. The stabilization records for Long House deal only with masonry walls. The records for the work in "caves" (Lister 1940a,b) contain many before- and-after photographs of the cavates in which Lister worked (see Chapter 2), as well as brief descriptions of the tasks performed. The work done outside Frijoles (Lister 1940b) was concerned almost entirely with landform and drainage, though Lister did note and photograph some rock art and disperse some rock corrals in front of cavates. The site group west of Camp Hamilton has an especially high density of cavates; Lister's map shows 161 in less than half a kilometer. Lister assigned numbers to cavates in each of the areas where he worked. In each area the numbers start with one, and C is used as a prefix in all areas (e.g., C35). He prepared a map for each area showing the configuration of the cliff. Lister's (1940a) descriptions of cavates in Frijoles Canyon are somewhat more detailed than those for the other areas, and the work he performed included some masonry pointing. The most common modification he made involved rearranging fill in front of cavates, but Lister makes no mention of artifacts in or outside of them. He summarized his cavate work in several notes in the Southwestern National Monuments newsletter. In 1960 C. Johnson, a graduate student at the University of New Mexico, removed a secondary burial of a child from "Cave Room C54 Tsankawi Ruin." This room is located in the upper, gray tuff cliff, just below the main pueblo, and is part of the group our crew recorded at Tsankawi (LA 50976). Matting, cordage, a few sherds and lithics, and a Bandelier Black-on-gray bowl were present in the back corner of the room, under about 8 cm of fill. Since small bones were missing and the long bones were stacked below the cranium, this was quite clearly a reinterment. A floor pit, apparently unrelated to the burial, was beneath it (Johnson 1960). In 1962 James Maxon, the park archaeologist at Bandelier, cleared the floor of a cavate in Mortandad Canyon, then in the Otowi Section of the monument (Maxon 1962). This is a fully enclosed, 3 x 3.6 m cavate with two 12 CAVATE STRUCTURES doors, one blocked up; a firepit near the blocked door; four large niches; and abundant rock art. Because this room is the largest in its group, Maxon called it a kiva. Some corn remains and a few sherds were recovered from the excavation; the decorated sherds from the room and the area outside are Santa Fe Black-on-white and Wiyo Black-on-white. The petroglyphs in this cavate largely defined Steen's Mortandad style of rock art (Steen 1979). Maxon analyzed material from two Los Alamos Archaeological Society excavations for his master's thesis (1969). One of the sites was the Tshirege Cave Site, consisting of 13 masonry rooms, 6 cave rooms, and a court area. Tshirege is located near the modern community of White Rock and is thus not far from Tsankawi. Like Tsankawi, Tshirege is a large masonry pueblo with cavates below it, including the "Tschirege Cave Site" (Maxon's spelling). Maxon stated that: available photographs indicate that these cave rooms were also typical of other cave rooms in the area. The rooms rarely exceed 10 feet square in size and ceilings are rarely more than 6 feet in height. Occasionally cave chambers which were used as ceremonial rooms or kivas were somewhat larger. . . Usually [they are] behind surface dwellings . . . rarely used alone. Cave rooms typically had adobe plastered floors, often showing several layers of plaster. The walls were also plastered to a height of about 30 inches. The plaster was often colored red or white. Above the plaster the walls are heavily soot coated from fires, either in the caves themselves or from adjoining rooms to the outside. In addition to the doorways, sometimes a small ventilator hole was located near the ceiling of the cave. The cave rooms were often interconnected... Apparently cave rooms were not excavated more than one row deep into the cliffs. Aside from their good insulation from both heat and cold, the caves had little to offer for day to day living. The lack of light, ventilation, and their generally small size made them less desirable than the rooms built in front. Nevertheless, evidence of use of fire, storage niches, and repeated refurbishing of the floors suggest that the caves were utilized as much as the outside rooms (Maxon 1969:49-50). The ceramics suggest that this part of Tshirege was used from the late fourteenth through the early sixteenth century. Material consisted of fairly abundant pottery, some milling stones, mauls, and chipped stone dominated by obsidian. The only perishable mentioned from the collections is some leather. Charlie Steen (1977, 1982) has assembled information on work done over many years on the lands of Los Alamos National Laboratory. Although he made "no particular effort ... to locate groups of cavate rooms" (1977:3), he made several observations concerning cavates. He believes they relate to the period on the Pajarito Plateau when the larger sites were being occupied. He argues that cavates served primarily for storage and for ceremony. The argument for ceremony rather than habitation is based on several contentions: that cavates are intentionally smoke blackened rather than blackened by heating fires (replastered walls are sooty but not black), that hearths near doors are kiva features, and that cavates often have artwork in them (1977:15- 17). He believes the very small, blackened examples were used for individual meditation. The larger blackened and plastered cavate rooms with rock art "served as religious or ceremonial rooms for basic families, and each was similar to a family chapel" (Steen 1979:42). Steen also defined a rock art style, the Mortandad style INTRODUCTION 13 mentioned above, which he says is limited to a small subarea of the Pajarito Plateau and found strictly in cavates. He attributes it to the late fourteenth century. The style is characterized by incisions into blackened tuff and by the presence of Awanyus, birds, Kokopelli, and an anthropomorphic figure that Steen likens to the Toltec sun god. Steen's viewpoint concerning the Pajarito Plateau cavates found a wide audience in a 1982 issue of National Geographic, which has long been an outlet for discussions of cavates (Beam 1906; Canby 1982:578-579, 592). The Pajarito Field House Project~a part of the Pajarito Archaeological Research Project (PARP) of the University of California, Los Angeles-excavated the shallow deposits in an isolated cavate (LA 52333) in the area south of Puye (Preucel 1985, 1986a; Hill and Trierweiler 1986). This single room contained three hearths. Both squash and corn remains were recovered. The few sherds associated with it and those on the talus in front were Santa Fe Black-on-white and Tesuque corrugated. Robert Preucel suggested that because of its isolation this cavate could have functioned as a field house. As part of the PARP, Justin Hyland collected data on features and room dimensions for a total of 44 cavates in 2 adjacent groups in Garcia Canyon, between Tsankawi and Puye. He presented the results of this recording and its analysis in his honors thesis (Hyland 1986). Although some of the variables measured were different, and the means of measurement also differed, Hyland's study is by far the most nearly comparable to the present one, and it monitored many of the same attributes. In addition to the cavates recorded by Hyland, PARP recorded the locations and some measurements for 431 other cavates in a total of 35 locations (R. Preucel, personal communication, 1988). The PARP survey also documented several large, Late Coalition pueblos with extensive associated "cavate villages" in the area south of Puye (Preucel 1986b: 8; 1987). As part of an excavation program related to the Bandelier survey, a Washington State University field school under the direction of Timothy Kohler excavated a single cavate chamber in Frijoles Group M near the rooms excavated by Hendron and outside the area recorded by our crew. The fill of this room was primarily disintegrated tuff, with sloughed wall plaster near the floor. Materials recovered were similar to those we observed during recording: a few cultivar remains and very sparse artifacts. Seven features were recorded, of which the three floor features-two bins and a cist-were rare or absent in our sample (in which there are relatively few floor features). A deep, heavily modified niche is unlike any recorded in this study (Carlson and Kohler 1989). Cavates have long attracted attention, and it seems that in the past hundred years they have been found to contain little material. That situation, however, may well relate as much to their visibility as to the condition in which they were left (in some senses they have never been completely "abandoned"). Although most of the people who have studied cavates have noted replasterings, smoke blackening, and domestic features, they generally agree that for reasons of space and ventilation, cavates would be an undesirable place to live. Three functional categories have been defined: small, unsmoked rooms used for storage; rooms with smoke blackening and other features that may have been habitations; and the largest cavates, smoke- blackened and containing rock art and sometimes loom anchors, which have been called "kivas." Although the last category in particular is loaded with assumptions and inferential leaps, these categories may at least be tested with detailed data from cavates. The PARP study and the sample reported here are a beginning, but expanded research is needed if we are to make 14 CAVATE STRUCTURES meaningful regional statements about function and variability. The Present Project In the summer of 1986 a crew of three National Park Service (NPS) archaeologists, assisted by four volunteers on varying schedules, spent six weeks recording measurements and features of cavate rooms in Frijoles Canyon and in the Tsankawi section of Bandelier National Monument. The locations of cavate features in Frijoles Canyon and at Tsankawi have been known for a long time (e.g., Hewett 1909a; Lister 1940b). This project recorded data on condition, measurements, associations, and visible features for a sizable number of cavate rooms. These data were collected as consistently as possible, recorded in a coded format, and entered into a computer data base so that they could be readily manipulated and easily recalled. The project produced location sketches and maps, along with photographs of feature openings and selected features, and videotapes of all rooms recorded. Immediately following the fieldwork, we entered the data into the computer and prepared a preliminary report, but we lacked funding and personnel for either checking or analyzing the data. A later contract permitted us to check for errors in coding and data entry and to prepare baseline data on cavate feature occurrence and measurement. This study is the result of the latter analysis combined with the preliminary report. In addition to the foregoing oudine of previous archaeological work done in cavates, it describes the sites in which recording was done, discusses the recording procedures followed, and presents the results of descriptive and interpretive analyses using the data gathered. Beyond the goals of better cavate management and description, several higher level questions motivated this study. Several interesting questions arise from the project. Can we group cavates based on size and features present, and can we interpret the possible functions of these groups? Do certain types of cavate tend to occur in certain locations? How do Tsankawi cavates differ from those in Frijoles, and can we attribute those differences to the cultural boundary between the Tewa and the Keres, which those groups traditionally consider to have existed between those areas (e.g., Hewett 1938; Steen 1977)? The data assembled here allow us to begin discussing these questions; answering them will require following them much further. In capsule, then, this project was designed to establish: (1) an estimate of the ranges of variability and covariation of attributes in cavates; (2) a procedure for recording these features; (3) the current condition of a sample of structures as a baseline for future maintenance and for the types of disintegration that occur in cavates; and (4) the coherence and logic of various interpretations of cavates. Context, Descriptions, and Chronology Edgar Hewett (1938:27, 34) takes credit for naming the east flank of the Jemez Mountains the Pajarito ("little bird") Plateau, in spite of his peculiar statement that "the country is almost devoid of birds" (Hewett 1938:30-31). As he defined it, the plateau extends from the Chama River on the north to Canada de Cochiti on the south and is bounded by the Rio Grande on the east. The name comes from Pajarito Canyon, which in turn takes its name from the translation of the Tewa word Tshirege (Harrington 1916:282-283; Lange et al. 1975:58, 77; Hendron 1946:89; also spelled "Tsirege," "Tschirege," and "Tzirege"), a large, Classic period pueblo ruin on the outskirts of modern White Rock. The Pajarito Plateau is a place of great drama and beauty. Its geological history culminates in a huge explosion; its great elevational range and resulting moisture give rise to biological diversity and splendid panoramas; its human history is long and highly varied. This study examines a small spatial and temporal portion of this broader context. In addition to summarizing the setting, this chapter describes the areas in which we worked and how they were selected. It concludes with a discussion of the means and problems of the chronological placement of cavates. The Setting Geology and Environment Even more than in most prehistoric settings, the geology of the Pajarito Plateau was of critical importance to how its inhabitants adapted to life there. The massive, relatively easily excavated tuff deposits exposed by the canyons of the plateau made possible the construction of cavate dwellings. The source of this tuff was a major geological event during the Pleistocene: the explosive eruption of the Valles Caldera followed by an ash flow that spewed forth a couple of hundred cubic kilometers of volcanic ash (Ross et al. 1961:141; Heiken 1979; Mathien et al. 1993). The singularity of the event accounts in large part for the limited occurrence of cavate dwellings. The Jemez Mountains were formed by millions of years of volcanism, but it was the "climactic and terminal stage" that formed the Tewa Group of tuffs and lavas in which the cavates were constructed (Figure 1.1; Bailey et al. 1969:12-15). The Tewa Group contains both the Valles Rhyolite and the Bandelier Tuff; the Bandelier Tuff is the formation into which the cavates were carved. The caldera left by the eruption, the ash flow blanket and its later dissection, are the main features of the landscape occupied by the Pueblo peoples in the twelfth through sixteenth centuries. The Jemez Mountains rise to 3500 m (11,500 ft.). They form an effective moisture trap, and several streams flow out of them, including the Rito de los Frijoles. The presence of water, the sharp elevational differences, and the softness of the tuff have predictably led to dissection of the Pajarito Plateau by numerous deep, often sheer-walled canyons radiating from 15 16 CAVATE STRUCTURES the caldera. The Bandelier Tuff has been divided into two members-the Otowi being overlain by the Tshirege--each of which consists of ash flow units resting on thinner pumice beds (Bailey et al. 1969:12-14). Cavates occur primarily in the Tshirege (upper) member, which consists of "a series of cliff-forming welded ash flows" (Bailey et al. 1969:13; Kelley et al. 1961:56-59). The lowest cavates recorded at Tsankawi are cut into a distinctive reddish tuff that is much softer than the overlying gray tuff and contains larger chunks of pumice. In this location Bailey and others (1969) describe the Tsankawi Pumice Bed at the base of the upper member of the Bandelier Tuff. The layer into which the cavates are cut, however, is thicker than their description of the pumice bed (they measure it at around 1 m). In any case, these cavates are also in the Tshirege member of the Bandelier Tuff, very near its base. Grant Heiken (1979:569) notes that deposits at the base of pyroclastic flows are easier to excavate. In response to an inquiry about patina formation and, more specifically, the variability in the tuff at Tsankawi, he wrote the following: Most of the variations in the tuff at Tsankawi ... are related to postdepositional processes. Obvious facies within the tuff, which are soft and easy to excavate, include nonwelded bases and pumice falls. Harder tuffs include those cemented by secondary minerals in the vapor phase zone (that zone near the top of the composite section of pyroclastic flows affected by hot gases rising through the cooling tuff deposit) and the thin, resistant layers cemented by zeolite cement. The latter were located at the top of the ground water table that was present before the canyons were excavated by erosion. Multiple resistant beds can represent a record of declining ground water as the canyons were growing deeper and wider with time. The flat benches at Tsankawi are tops of the more resistant zeolite- cemented tuff. Nearly all cavates are located within distal regions of the pyroclastic flows where the tuff is nonwelded. Closer to the source, for example above Ponderosa Campground, the tuffs are welded; these would have been impossible to excavate, being hard and dense. (G. Heiken, personal communi- cation, 1986) The well-watered uplands contain lush vegetation and associated montane fauna (Mathien et al. 1993:6-8). Because the elevation drops rapidly to the Rio Grande at around 1600 m, there is considerable biotic diversity in a fairly small area. Alpine tundra, spruce-fir- aspen, ponderosa-oak, pinon-juniper, and riverine plant associations are all found within 25 km of the study areas, depending on exposure and elevation (see Powers 1988 for more detailed discussion of environmental zones). Even in the lower elevations of the plateau, the growing season varies considerably, from 120 to 180 days (Hubbell and Traylor 1982:29). Cultural Context For an area seemingly well suited to supporting populations subsisting by hunting and gathering, remarkably few sites of the Archaic and early parts of the Pueblo eras are known on the Pajarito Plateau. David Stuart and Rory Gauthier (1981:48-49) found that fewer than 12 percent of all components in the state survey files date to before a.d. 1 175. More sites from the earlier period have been and will be found as more comprehensive archaeological work takes place, such as the preceramic pithouse near Otowi (Lent 1988). Present samples, however, indicate relatively low human population and use in the area in early prehistory. This situation changed dramatically, however, at around 1175. At that time there was a sudden profusion of small pueblos characterized by rectilinear room blocks, pit structures, and ceramics dominated CONTEXT 17 by Santa Fe Black-on-white: the Coalition Period (Stuart and Gauthier 1981:45-51; Cordell 1979:53-64; Preucel 1987; Powers 1988; Mathien et al. 1993:9-34). Judging from ceramics, it is probable that the first cavate structures were dug sometime during the Coalition Period (see below). Following the Coalition Period, between about 1325 and 1540 the population of the area became increasingly aggregated, as seen in the archaeological record for much of the northern Rio Grande, including the Pajarito Plateau (Kohler and Linse 1993:3-5). At this time~the Rio Grande Classic-there were fewer but much larger settlements. Large, free-standing pueblos were built, such as Tyuonyi and Tsankawi, and on the Pajarito Plateau the large pueblos have large groups of cavates nearby. Other, well known sites having both cavate clusters and large, free-standing structures include Otowi, Navawi, Tshirege, and Puye. These associations and the predominance of ceramics, such as Bandelier Black-on-gray and Tsankawi Black-on- cream, on cavate sites suggest that at least clustered cavates are a part of the Classic Period aggregation. In the historic era, following the Classic Period, permanent habitation on the Pajarito Plateau shrank dramatically, and almost all Pueblo settlements were located in the major river valleys. The plateau remained an important subsistence and sacred resource area for the Pueblo peoples, as well as a refuge, but its time as a location for large human populations was over until the coming of the nuclear age to Los Alamos. The Sites Cavate Groups Both Frijoles Canyon and the area around the main ruin at Tsankawi were very densely settled, so that the concept of discrete sites in these areas is somewhat suspect. Hewett and Chapman divided the cavates in Frijoles into Groups A through M based on breaks between clusters, which are often caused by drainages or stretches of cliff unsuitable for cavate construction. The separation between the groups they defined is only a few meters in several cases, and the groups vary considerably in size. Whether or not this long string of cavates was 14 or more settlements, as implied by this topographic grouping, can only be inferred by careful study. Our recording is a early step in making this inference. In 1986 we spent time recording in five groups of cavates (Figures 2.1, 2.2). The Hewett-Chapman groups remain useful to the NPS as a framework for management, and we used them as the first stratum for our recording sample. Because of short time and small crew size, we recorded all of one of the Hewett groups and parts of three others. In addition, we worked at a fifth group in the Tsankawi portion of Bandelier National Monument, 11 straight-line km northeast of the cavates in Frijoles Canyon (Figure 1.1). The sample of cavates selected was designed to assess several dimensions of variability. Within Frijoles Canyon we were interested in whether a number of locational variables influenced cavate morphology: upstream or downstream location within the canyon; vertical and horizontal proximity to the Rito; size of cavate group; location within a group. At the next level, we were interested in differences and similarities between the Frijoles cavates and a group of cavates outside Frijoles Canyon (the Tsankawi sample). In Frijoles, dimensions of locational variability are to some degree correlated; that is, cavates in the upper end of the canyon are closer to the Rito because of the canyon topography and the intersection of the stream with the tuff strata. Tsankawi Mesa has at least three major tuff types, and we studied some rooms from each. The rationale for recording a stratified sample of groups rather than a random sample is grounded on both practicality and information yield. First, to accurately sample the whole Frijoles or Tsankawi population would require 18 CAVATE STRUCTURES Cliffs Studied Cavates GROUP M 500 meters RAINBOW HOUSE 2000 feet I Figure 2.1. Map ofFrijoles Canyon showing the location of the cavate groups discussed in this study and other major sites. Only selected contours from the USGS Frijoles Quadrangle are shown, with the cliffs of the north side and the cliffs and steep slopes of the south side shown by shading. CONTEXT 19 Figure 2.2 Map of the Tsankawi section ofBandelier National Monument, showing the main pueblo of Tsankawi. LA 50976 and three groups ofcavates mapped by Lister are scattered along the south edge of the mesa. By request ofBandelier National Monument, the locations of these fragile cavate groups are not shown. Selected contours and features are taken from the USGS White Rock Quadrangle. 20 CAVATE STRUCTURES complete inventories of each, and those inventories did not exist in 1986. Second, locating randomly selected features would be time-consuming and it would be difficult to determine which other features should be included. Third and more important, there are good archaeological reasons for recording groups. It is extremely unlikely that any single cavate in either of these locations was a site unto itself in terms of prehistoric use. By recording groups, it is possible to gain some idea of whether size or functional groups are associated in a regular way. It might be possible to infer, for example, whether or not smaller groups of cavates can be considered individual-use units (see Figures 2.3, 2.4.). Through an unfortunate oversight, the field crew was given an incorrect set of Laboratory of Anthropology (LA) site numbers for use in the field. These numbers were used on all forms, photo records, and notes, and on photo boards. Although the correct, official numbers are used throughout this text, the presence of the field numbers in so many places requires a concordance between official numbers and field numbers: Group Official No. Field No. M LA 50972 LA 50020 A LA 50973 LA 50021 I LA 50974 LA 50022 F LA 50975 LA 50023 Tsankawi LA 50976 LA 50024 The field numbers have been retained in the computer data base for purposes of matching field records and photographs with final records. A plan map and elevation or profile sketch is included for each study group in the following descriptions. The variability in these figures is the result of three different recording techniques. During the 1986 fieldwork each Frijoles group was mapped using tapes and compass. Cavates in the Tsankawi group were placed on Lister's map. Rough field elevations were drawn showing locations, and then drawings were made on Mylar overlays on photographs. Groups F, I, and M in Frijoles were included in the later sample survey of the monument. During recording for the survey, plane table maps and careful elevations were made. When available, the more detailed maps and drawings from the park survey have been included here. When these were unavailable, drafted versions of the 1986 recording have been used. Frijoles Group A, LA 50973 (Held no. LA 50021) Group A is located in a lovely, park-like wide spot in the canyon. It is a large group, running a couple of hundred meters along the base of the cliff and consisting of around 130 cliff-associated rooms and several substantial rubble mounds representing masonry room blocks (Figures 2.5, 2.6). Group A is separated from Group B by a projection in the cliff, but the two are quite close together. In spite of its designation, which would seem to indicate that it is the first group of cavates in the canyon, there is a substantial cluster of cavates 1.1 km upstream from Group A (see the description of Cuevitas Arribas, below). Group A is located at the upper end of the fairly continuous string of locations with cavates that stretches for about 2 km through the central occupation area of Frijoles Canyon. Much of this distribution may be explained by canyon width: Group A is located in the last wide spot in the canyon as one proceeds upstream, the upper end of the wide part of the canyon where the Rito is not deeply entrenched. Like Long House (Group D), Group A is located very near the canyon floor rather than at the top of a talus slope, the location of most of the other Frijoles cavates. The U.S. Geological Survey (USGS) quadrangle indicates that the creek is at an elevation of 6150 ft (1875 m) where it passes Group A. Altimeter readings at the cavates indicate a difference of 150 ft (46 m) from the Visitor Center benchmark, or about 6210-6220 ft (1893-1896 m). The cavates of Group A are all at about the same elevation because the base of the cliff is at the same level for most of the CONTEXT 21 Figure 2.3. View of Group I from across the canyon. Masonry rubble is visible in front of the concentration of cavate rooms. The relatively recent rockfall on the righthand side of the group partially covers rooms 30-34. The group extends from the lower chamber on the left and there are two rooms out of the frame to the right. -* 1 •^fe^ Figure 2.4. Upper Group M viewed from across Frijoles Canyon. The area recorded extends from out of the picture on the left to the small tent rocks at the far right. Note both the extensive rubble and the heavy use of cavate chambers that had masonry closings. (63)65(66) ^Ue, -' 4 167^8/J^ CONTEXT 23 73 ^ Ceramic Sample Area Extends 25 Meters — LA 50973 GROUP A PLAN VIEW I 44 J i i Upper Story Cavate tefe;zP~ 1st Story Cavate fl(n ^ Extent of Rubble >$£&, Bed Rock ^ Boulder 10 i i i i i i_ 15 _1_ 20 10 _1_ m«t»t» 20 30 40 50 ip A sample. Room numbers between dotted lines are I at the bottom of the stacked room numbers. Back rooms Ing a compass and tapes in 1986. CONTEXT 23 (63)65(66) ^ 73 » \ / 62 / I | ' ,'70 '71 \ 7 ?X \ 59,60, \ / .1 -' ! 67 ' Ceramic Sample Area Extends 25 Meters — i LIMIT OF CERAMIC SAMPLE AREA LA 50973 GROUP A PLAN VIEW 43. Upper Story Cavate l?feiT~ 1st Story Cavate "rmir^ £"'•"• of Rubble ■'?t>.: Bed Rock (7/X Boulder 10 i — i — i — i i i_ 10 20 30 40 Figure 2.5. Room plan view for the Group A sample. Room numbers between dotted lines are arranged with lowest story room at the bottom of the stacked room numbers. Back rooms are in parentheses. Drawn using a compass and tapes in 1986. Group at Back of CONTEXT 25 61 63 | 66 9 64 sample. Drawn from field sketches and photos in 1986: distance from Cavate 1 to Cavate 73 is approximately CONTEXT 25 Group at Back of Cleft (same level as 1-5) LA 50973 GROUP A PROFILE sags Wall Alignment ••• Viga Holes 10 i meters 20 30 _l L_ 40 I 20 _l 51 54 ), 56 £H Figure 2. 6. Elevation sketch for the Group A sample. Drawn from field sketches and photos in 1986: scale approximate (straight-line distance from Cavate 1 to Cavate 73 is approximately 93 m). CONTEXT 27 group. The USGS contours are quite confusing here. While converging contours are to be expected given the size of the cliffs, the fashion in which they converge here is enigmatic. Group A, as noted, is located in a relatively wide spot in the canyon. This fact is suggested by the contours, but the 6200 and 6300 foot lines converge just below A and show a much broader bench than appears to exist 200 ft (60 m) above the Rito. While the cavate symbols seem to be properly placed horizontally (there are three for all of Group A), the contours seem wrong, and the altimeter supports our disagreement with the contours. Therefore, we chose 6210 ft (1893 m) as an elevation for Group A and estimated the variations from that baseline. The tuff at the base of the cliff is more variable in a shorter vertical section in the area of Group A than in any other area in which we worked. There are several layers that differ in color, texture, and durability, including a crumbly gray unit containing quantities of black rock, a coarse-grained but fairly compact white layer, a brownish layer, and the gray-white finer texture seen commonly elsewhere in Frijoles cavates. All of these units intersect the structural parts of Group A. The portion of Group A that we recorded extends from the uppermost rooms about halfway through the group. The upper end of the group is quite well defined. There is a steep, deep cut in the canyon wall at the upper end of the wide area in the canyon, which contains several pockets well up the cliff. These could have contained rooms but do not appear to have had any, and no further rooms seem to exist between upper Group A and Ceremonial Cave (however, a thorough examination has not been made). The lower end of our study in Group A was defined more by the exigency of field time than by natural breaks within the group. Since the sample of Group A was small relative to the group and we had an extra day and a half at the end of the field season, the naturally defined subgroup extending to A-36 was expanded to A-73. Our count of 130 rooms showing in some way on the cliff includes 40 rooms recorded in detail in July 1986 and 90 rooms below that stopping point. (The latter figure is based on three counts giving 90, 92, and 81; the count of 81 was made from a point well away from the cliff, from which vantage point many low rooms are not easily visible.) Usually the process of detailed recording revealed more rooms than were recorded with careful but necessarily more superficial counts, and this has a minor effect on the relative positions recorded for some cavates. The overall area of the rooms recorded is about 95 m along the cliff by 20 m, allowing for recesses in the cliff and rubble areas. It is not entirely clear whether the base of the cliff in Group A had a continuous structure along it or whether there were separations between structures. Rows of viga holes directly above one another show that major remodeling did take place in the masonry rooms of Group A, which makes determination of the presence of continuous built structure even more difficult. Whether or not the building was at some time continuous, most of the cliff did have some structure placed against it. The amount of building rubble now visible varies considerably along the base of the cliff, with greater quantities present in recessed areas in the cliff. The total number of rooms is probably more than twice the number visible on the cliff. Considered as a site, then, Group A was probably at least as big~and as long~as Long House and was probably comparable to Group M. Particularly at the lower end, but elsewhere as well, much of the site was multiple stories; in some places it was at least two, and quite often four. Its location near the canyon floor probably made it easier to erect taller structures here than on higher, steeper talus slopes at other cavate groups. Several other archaeological features are present in the vicinity of Group A. The most notable is a masonry-lined kiva below the 28 CAVATE STRUCTURES approximate center of the group. This structure is shown on the Hewett-Chapman map and remains distinctly visible, which indicates that it has probably been at least partially excavated, though Park Archaeologist Bill Sweetland says there is no record of such an excavation (personal communication, 1986). Visible structural remains are otherwise scarce on the relatively flat canyon bottom below the group. This absence of structures is probably related to the conservation of watered bottomland for crops, though the presence of a low terrace would have required a substantial ditch to irrigate much of this part of the canyon; pot irrigation would certainly be easier here than in most of the Southwest. On the first ledge above the base of the cliff are several petroglyphs. We did not venture up, but the ledge looks large enough to have had some structures on it. Two constructed routes seem to extend at least as far up as the ledge at the upper end of the room group (vicinity of rooms A-9 and A-10). Routes to the canyon rim above this ledge as well as slightly up-canyon are suspected, but none of these has been actually climbed. The only other known recorded archaeological work at Group A is the repairs made by Lister in 1939 (Lister 1940a). Considering the size of Group A, there seems to be remarkably little trash present on the surface. Some of the other groups have steep slopes below them, which may accelerate removal of trash by washing, but this is less a problem at Group A. Based on casual observation and McKenna's more systematic study, Group M has more trash on the slope below it than does Group A. Lower visibility due to the denser vegetation of Group A is probably one reason for this, but a more important one is its heavier tourist traffic. Excavated tests of trash quantities (and qualities) at any of the cavate groups would be of considerable interest. This area has been closed to visitation for some time (Sweetland, personal communication, 1986). Nonetheless, Group A seems to have more graffiti than any of the other groups in which we worked, including Tsankawi, which receives heavier, unsupervised visitation. Group A clearly has visitors and has had them for a long time. Our presence there may have reduced illicit visits, but in the five or so days we spent recording there, we saw only two young boys going up to the rooms (they never saw us). It seems unlikely that the current level of visitation will lead to degradation of deposits, though it will continue to wear away at the structures. Further public education concerning the fragility of these resources and more frequent passes by rangers going to and from Ceremonial Cave might help reduce this source of erosion. Perhaps the discreet placement of a rain shelter somewhere in the area would cut emergency visits by walkers caught far from the Visitor Center in summer storms. Natural deterioration at this group takes several forms. The most widespread results from the very friable nature of the lowest tuff unit in several parts of the Group (around A-l- A-9, A-23-A-29, and A-50-A-58, for example). This tuff erodes much faster than does the overlying stratum, and it forms the back wall to many rooms in the first visible story, as well as a possible, speculative story below that. The fact that it undercuts also threatens features in upper tuff levels, though this does not seem to be an imminent threat. Recent slumping of large tuff blocks has occurred in several places, most notably at A-72, parts of which seem to have fallen quite recently, and in front of A-15. In several places erosion and cliff deterioration seem liable to cause loss of features; this is true of the high rooms A-20 and A-39, where doors are very exposed and are becoming quite thin. A more severe loss could occur in rooms A-22 and perhaps A-18, which are connected, complete rooms. The front wall of A-22 is supported by a very fragile-looking pillar, and loss of much of the front wall is conceivable. This condition might be relieved by repair or replacement of the masonry wall that appears to have been present there prehistorically. This CONTEXT 29 course would involve at least partial excavation on both sides of the wall. Finally, there are three rooms with masonry closing portions of their fronts: A-10, A- 13, and A-60, all of which were pointed by Lister in 1939. The small doorjamb in A-13 looks just as it did in 1939, and A-60 seems to be in reasonably good shape (Figures 2.7, 2.8). A-10 appears to be something of a miracle: a large piece of masonry appears to be supported mostly by fit and pressure; most of the wall has no foundation and forms a sort of arch (Figure 2.9). Even more remarkable is the location of the room at the back of a large pour-off, which looks as though it should have washed the whole thing away long ago. An area in the middle of the wall now lacks mortar; at least pointing, and perhaps detailed recording, seems to be indicated here. The rooms along the side of the A-10 drainage cut (A-9, A-ll, A-12) are severely eroded. In summary, the distinguishing characteristics of Group A are its proximity to the stream, its length, the presence of much rubble, an associated round masonry kiva in the flat area in front of the group, and highly variable tuff strata. Frijoles Group F, LA 50975 (Held no. LA 50023) Of the groups recorded in 1986, Group F is central; it lies between Tyuonyi~of which it has a splendid overview~and the "Big Kiva." It also seems to have the most varied topography of the groups we studied, since it includes two large tent rocks and covers a considerable range in elevation (Figures 2.10, 2.11). The portion in which we worked may be considered the lower part in two senses-it is at the downstream end of the group and it is closer to the canyon floor. The area studied extends from a small, naturally sculpted arch at the upper end (this may well have been a room but was not recorded as such because it lacks features) along a stretch of relatively low cliff base behind two large tent rocks to just up the slope in the large embayment containing Group G. The tent rocks have remnants of rooms around their bases, and these rooms were included in the group both by Chapman and by our crew. The rooms from which data were collected can be contained in a trapezoid with a base of about 48 m along the cliff base, a height of 17 m from the cliff to the front of the tent rocks, and a top of about 20 m across the front of the tent rocks. The upper part of Group F is located on higher cliff bases and steeper slopes; we counted 46 rooms in this part of the group. 1 We recorded 48 rooms along the cliff base and 12 rooms around the tent rocks, so that the total number of cliff-associated rooms for Group F is 106. In counting rooms up- and down-canyon from specific rooms, the tent rock rooms were not included. Remnants of intact masonry exist, especially in the unrecorded part of the group; considering the centrality of this group both to Tyuonyi and to tourism, the rooms in this group seem to be in surprisingly good shape. Below where it levels out (at about F- 15), the part of Group F that we studied has a great deal of masonry rubble associated with it. This segment of the group appears to be better suited to building than the upper part because of less slope. Approximately in the middle of our study group (rooms F-19-F-30), there may have been as many as six stories against the cliff (Figure 2.11), and considerable rubble remains in this area. Four cliff levels are clearly visible here, with a fifth suggested by a depression in the cliff above the uppermost. In addition, the viga holes for the lowest visible story are close to the present ground surface, which is the top of a considerable mound, leaving the possibility of yet another, invisible story at the base. This area shows evidence of a considerable expenditure of human energy, in the form of room remodelings, hand-and-toe hold trails, and large petroglyphs. There appear to have been retaining walls made of large blocks between the tent rocks. Terraces may also have been constructed at the extreme downstream end of the group, where the talus becomes quite steep 30 CAVATE STRUCTURES Figure 2.7. Comparison photographs for A-l 3, showing little change during 47 years, a, b. Before- and after-stabilization photographs taken by Lister in 1939. c. Photograph of A-l 3 taken in August 1986. Note the similarity of plaster and masonry condition; somewhat greater wear is probably present at the lower left of the opening, and there may be additional graffiti on the plaster of A-l 4 below the lower right of the opening. CONTEXT 31 Figure 2. 8. Comparison photographs for A- 60. a, b. Before- and after- stabilimtion photographs taken by Lister in 1939. c. Photograph of A-60 taken in April 1987. Very little change is apparent after 48 years. 32 CAVATE STRUCTURES Figure 2.9. Photograph ofA-10 taken in 1986. Lister did stabilization work on A-10 but did not include a photo in his report. It is remarkable that the masonry of this room, located at the head of a drainage, has survived. A 50975 OUP F PLAN VIEW CONTEXT 33 Boulder &.A- Bedrock • Rock Terrace 5 i i i i 10 15 20 2 1 l 10 1 meters 20 30 40 i i i 50 l 1 sample. Shows Group F divided into upper and lower the rooms in lower Group F. Upper Group F has four bers shown for Lower Group F are those used in this '. in 1990 by J. Snead. CONTEXT 33 LA 50975 GROUP F PLAN VIEW Figure 2. 10. Room plan view for the Group F sample. Shows Group F divided into upper and lower parts; this project recorded only the rooms in lower Group F. Upper Group F has four sets of room numbers; the numbers shown for Lower Group F are those used in this study. Redrawn with an alidade in 1990 by J. Snead. CONTEXT 35 ROFILE 5 |inge (view point break) o Room not recorded by this study L — Rooms 19 and 20 obscured by angle in cliff face GROUP F sample. Shows Group F divided into upper and lower le rooms in lower Group F. Changes in cliff angle are s. Tent rocks with their cavate rooms are not shown. ndH. Newman. LA 50975 GROUP F PROFILE CONTEXT 35 Viga Holes Angle Change {view point break) 5 10 15 Room not recorded by this study - Rooms 19 and 20 obscured by angle in cliff face LOWER GROUP F Figure 2.11. Elevation sketch for the Group F sample. Shows Group F divided into upper and lower parts; this project recorded only the rooms in lower Group F. Changes in cliff angle are indicated by vertical dashed lines. Tent rocks with their cavate rooms are not shown. Redrawn in 1990, by J. Snead and H. Newman. CONTEXT 37 as the cliff turns a corner into the recessed area containing the last few rooms of Group F (F-39- F-46) and then Group G. As at Groups A, I, and M, the distribution of rooms along the cliff base would have been very nearly continuous, though numbers of levels would have varied considerably. Especially above our study area, this distribution was accomplished in some places by heroic (if incomprehensible) room placements in drainages, in pour-offs, and on steeply inclined cliff bases. The elevations recorded for Group F rooms are based on an altimeter reading of 6160 ft (1878 m) at the center of the sample. The quadrangle. in this area suggests 6180-6200 ft (1884-1890 m); the 6200 and 6300 foot contour lines are merged in this area. The 6160 ft (1878 m) reading has been used as a baseline with some elevation added for the ends of the study group and some subtracted for the tent rock rooms. The elevations, once again, are approximate. The elevation of the Rito opposite Group F is about 6080 ft (1853 m). To my knowledge there are no records of excavation in Group F. Lister (1940a) did some stabilization in several Group F cavates, which we rephotographed (Figures 2.12, 2.13). Several of the walls he worked on have since collapsed. The proximity of these rooms to Tyuonyi and their ease of access makes it quite likely that they were among those used by field school students and earlier explorers, such as Bandelier. Indeed, in the vicinity of F-20, a row of nails has been driven into the cliff, presumably for clothes hooks. The old tourist trail passes below the Group F rooms, but not far below them. Frijoles Group I, LA 50974 (Held no. LA 50022) Located at the top of a relatively steep and high talus, Group I is a compact group of rooms fronted by a reasonably level area around 10 m wide (Figure 2.3, 2.14, 2.15). The main group of contiguous rooms is 29 m long (rooms I-7-I-35). At the upper end of the group are six rooms located above a precipitous drainage (1-1- 1-6). Most of these rooms are lower than the main group and face more directly east. The lower end of the group is defined by a rockfall consisting mostly of very large boulders. Six rooms were partially obscured by part of this rockfall (I-30-I-35), and it seems likely that the event occurred after the construction, and probably after the abandonment, of the rooms. In addition, three rooms stand almost exactly midway between Groups I and J. It is not entirely clear whether Hewett and Chapman included the sole complete one (1-36) on their map, and since they are slightly closer to Group I, we included them in our version of Group I. (The intact room, 1-36, is the lowest of the three and is 18.2 m from 1-35 and 16.8 m from the highest Group J chamber). By our count, then, 38 rooms can be seen on the cliff in Group I. All the rooms were recorded except 1-15, a small, third-level chamber we were unable to reach. Other rooms may exist at the same level as 1-15, but our inspection of some of these possibilities showed them to have no definable features and to be too ambiguous to be recorded as rooms. The flat area in front of the center of Group I has a considerable amount of rubble on it. There may have been two rows of masonry rooms in front of the cliff rooms. In the area of 1-24, 1-25, and 1-27, the cliff rooms are three levels high, with the bottom level substantially filled. These filled rooms may have well- preserved floor features. By altimeter the elevation of the center of Group I is 6240-6250 ft (1902-1905 m). The map location is once again confusing. Converging contours are shown below the symbol that corresponds to Group I. This indicates a sharper, higher drop than is the case and suggests an elevation of around 6320 ft (1926 m). We used a baseline of 6250 ft (1905 m) for the records; with the exception of I- 1-1-4 38 CAVATE STRUCTURES Figure 2.12. Comparison photographs for F-31. a, b. Before- and after-stabilization photo- graphs taken by Lister in 1939. c. Photograph of F-31 taken in July 1986. There has clearly been a dramatic change since Lister stabilized this feature. In addition to the collapse of the entire masonry fissure closure, the viga hole at each figure's shoulder is broken away in the 1986 photograph, suggesting that vandalism may be involved. Note the lintel stone in the precollapse photographs and the groove left after its removal. CONTEXT 39 Figure 2.13. Upper Group F, Room 12, a room not recorded by this project, as it appeared in 1939 and 1986. a, b. Before- and after-stabilization photographs taken by Lister in 1939. c. Photograph of the same cavate taken in July 1986. Note that the masonry plug to the right of the door is now gone and that the mortar has returned to its prestabilization state. CONTEXT 41 50974 IP I PLAN VIEW <- 1st Story Cavate L 2nd Story Cavate Wall Alignment Extent of Rubble Bed Rock Dripline Boulder 5 10 15 J_J I i_ 40 SO rt all rooms are shown on the plan: some appear only , 21, and 29 are back walls and not shown). Redrawn lead and A. Prieto. CONTEXT 41 Approximate Cliff Face Petroglyph LA 50974 GROUP I PLAN VIEW -J ^ 1st Story Cavafe --' '. 2nd Story Cavafe oo cr ° Wall Alignment /n $s Extent of Rubble .£}£<■ Bed Rock Dripline £3 Boulder -LIMIT OF CERAMIC SAMPLE AREA Figure 2.14. Room plan view for Group I. Not all rooms are shown on the plan: some appear only in the elevation (Rooms 11, 16, 18, 21, and 29 are back walls and not shown). Redrawn using an alidade in 1990 by G. Head and A. Prieto. CONTEXT 43 , Trails , Behind Rockfall LA 50974 GROUP I PROFILE C I Boulders oooo Wall Alignment Viga Holes • ••• 5 _L_L 10 15 20 meters 10 20 30 40 50 60 i i I l I I I I feet drawn in 1990 by G. Head and A. Prieto. CONTEXT 43 Petroglyph Hand- and -Toe Holds .15 Cavates 30-35 Behind Rockfall LA 50974 GROUP I PROFILE Trails Q Boulders OOOO Wall Alignment • ••• Viga Holes I I I I 5 10 I I I 15 I 21 I meters I 1 10 20 30 i i i 40 i 50 i 60 I feet Figure 2.15. Elevation sketch for Group I. Redrawn in 1990 by G. Head and A. Prieto. CONTEXT 45 and a few third-level rooms, the rooms in Group I are at about the same level. The elevation of the Rito de los Frijoles below Group I is about 6100 ft (1859 m). Trash is present on the slopes below the rooms, but as at Group M the quantity is fairly small. The length and steepness of the slope must again be considered, in addition to the possibilities of disposal elsewhere or in the rooms. Few artifacts are to be found on the surface of the masonry rooms or in the cavate rooms. Known previous work in Group I is limited to some stabilization done by Lister in 1939 (1940a). Group I has several extant blocks of masonry, and he replaced mortar in some of these (including 1-10 and 1-22; Figure 2.16). Lister's photos of 1-22 show that a ventlike hole was completely blocked in 1939 but is now only about half blocked (Figure 2.16). The Civilian Conservation Corps (CCC) trail passes considerably below the group. The climb to Group I seems to be arduous enough to discourage visitors, so that recent impacts are much less than at Group A. The graffiti here are about as numerous as at Group M or perhaps slightly more so. Deterioration in central Group I seems not to be severe, but the ends have suffered more damage. The upper end in particular, which is located above steep slopes and presumably lost the protection of masonry fronts early, is severely weathered. Although some damage must have resulted from the rockfall at the lower end of the contiguous rooms, burial by rockfall may have had the net effect of preserving the lower parts of these rooms. There are two sets of hand-and-toe holds within Group I suggesting routes to the canyon rim. The first of these begins at around the middle of the group (above 1-27) and probably relied on rooftops as a starting place (Figure 2.15). The second route is in the drainage in which the three isolated rooms (1-36-1-38) are located. There is a set of holds on each side of the rooms, leading to a bench and apparently on to the rim. Neither route was empirically tested. Other cliff features include sizable petroglyph panels at both ends of the contiguous cavates, a high petroglyph above 1-4, and one above 1-35. Frijoles Group M, LA 50972 (Held no. LA 50020) Group M lies at the down-canyon extreme of concentrated sites of all types in Frijoles Canyon. The group is a long, fairly continuous set of cliff and masonry rooms above a sizable talus. The talus is neither so steep nor so high as that at Group I (Figure 2.4, 2.17, 2.18). In terms of overall length and number of rooms, Group M is probably the largest group in which we worked in 1986, though Group A is in many ways comparable. Kohler estimates Group M to have had more than 200 rooms between a.d. 1325 and 1375 (Kohler and Linse 1993:5), but Group M surface ceramics analyzed by the Bandelier Survey indicate that substantial occupation before 1450 is unlikely (Robert Powers, personal communication, 1994). Rainbow House (LA 217), a 50-60 room pueblo dating to the fifteenth century (see Table 2.2 at the end of this chapter; Caywood 1966), is directly below Group M and clearly visible from it. Some relationship is very likely to have existed between the two. Even closer to Group M is Saltbush Pueblo (LA 4997), an 11 room structure with a single pit structure. In terms of both architecture and ceramics, Saltbush Pueblo fits well into the Coalition Period, and the few absolute dates obtained fall in the thirteenth century (Snow 1974). David Snow suggests that occupation and construction of Group M took place during the latter stages of occupation of Saltbush Pueblo. His report compares the relatively low diversity of the faunal assemblage from Saltbush Pueblo (11 species) to the somewhat richer (14 species), very late Group M assemblage from Hendron's excavation, and to the very rich (23 species) Rainbow House assemblage. These three assemblages have little or no temporal overlap. 46 CAVATE STRUCTURES Figure 2. 16. Comparison photographs for 1-22. a. After-stabilization photo taken by Lister in 1939. b, c. Two views of the door to 1-22 and the opening to 1-20, taken in 1986. Note again the lintel groove in the door; apparently the vent to the left of the door was completely closed in 1939, but most of the masonry is now missing. Much of the 1939 mortar is now gone, but the masonry above the door remains. The room may have had more fill in 1939 than at present; other change appears to be minimal. CONTEXT 47 O 2 0> • n in n 3 3 O cc X o c o c » 5 55 >» 5 35 >> g 0) 5 55 £ £ ■o ■o £ X x c UJ UJ " CM CO » $ / j * 1 J 1 f ( ( 1 1 \ '> 1 I. .5 5 ,60 CONTEXT 49 20 Linear meters of cliff face ommited here; elevation change is to s/l PROFILE ew point break) 15 20 25 40 i 50 i 60 _l sample. Upper Group M with vertical dashed lines Vooms 10, 61, and 65 are back chambers not visible \roup. Possible rooms (not recorded) are indicated 1991 by K. Barthuli and S. Hall. CONTEXT 49 20 Linear meters of cliff face ommited here; elevation change is to scale. Angle change, no cliff face omitted LA 50972 UPPER GROUP M PROFILE i'""\ Possible Room ^ Niche «•• Viga Holes Angle Change (view point break) 10 I 15 20 I 25 _l 10 20 30 40 50 60 I feet Figure 2. 18. Elevation sketch for the Group M sample. Upper Group M with vertical dashed lines showing changes in cliff angle. Rooms 10. 61, and 65 are back chambers not visible when looking at the front of the group. Possible rooms (not recorded) are indicated by dotted outlines. Redrawn in 1991 by K. Barthuli and S. Hall. CONTEXT 51 The cavate group is well defined by large drainage cuts at each end. The portion of the group in which we worked (upper Group M) extends from the up-canyon end, where there are two isolated rooms amid a rockfall, to about midway in the group, where a natural break is formed by some small tent rocks, a steep talus, and a slight lowering in the elevation of the cliff-base rooms. Upper Group M has an overall length of 102 m (78 m excluding the two uppermost rooms, M-l and M-2; Figure 2.17). Below the tent rocks multiple counts give an average of 73 rooms showing on the cliff. A total of 66 rooms out of 67 were recorded in the upper area, giving a total of 140 cliff-associated rooms in the group. Some high pockets in upper Group M have openings that look suspiciously like doors; those we were able to check were not rooms in our estimation, but some others may be. Recording of Group M in 1990 by the Bandelier Survey confirmed the count of 67 rooms in the upper area, but the survey crew saw an additional 24 rooms (for a total of 94) in the lower area. Although it seems likely that their count includes features or pockets we did not consider rooms, the count discrepancy also reflects the intensity of cliff modification in lower Group M, and the potential this maze of features affords for deriving different counts. As at Group I, the width of the flat areas at the top of the talus varies. Toward the upper end (M-3-M-14) the talus is steep and hard and extends directly to the base of the cliff; in this area no evidence of masonry rooms remains, though they were unquestionably present. In front of M-18-M-20 the top of the talus is considerably flatter, and there are many suggested walls and strewn blocks. Here it is first possible to see three levels of rooms (M-17 and M-64). In front of M-25-M-31 the amount of rubble further increases, and in this area three levels of rooms are again visible on the cliff. M-33, once an elaborately decorated room, stands on the up-canyon edge of the largest concentration of rubble for the entire group, including the part below our study area. Here a slight embayment in the cliff contains Rooms M- 35-M-49, and the rubble extends 9-10 m out from the cliff base with a relief of around 2 m. Wall alignments suggest as many as three rows of rooms out from the cliff, and there are as many as five levels of cavates (Figures 2.17, 2.18). Given the height of the mound and the level of rooms such as M-54, a large structure is indicated. Several cavate rooms in this area are larger and very heavily used and have many features. This area is "downtown" for Group M. Past the mounded area, in front of Rooms M-50-M-60, there is once again a steep, disintegrated tuff talus slope, though more masonry blocks are present on this slope than at the upper end. Even in this location, cliff evidence shows rooms on two to three levels, most of which seem likely to have had masonry fronts. Some large pieces of canyon wall have fallen here, which probably removed natural fronts of some of these rooms. The rooms above these steep slopes seem to have suffered the most from natural disintegration, presumably because masonry elements in these locations were far more prone to collapse. Lower Group M contains several notable features. As discussed earlier, the best documented excavations of cavate rooms in the canyon are Hendron's (1943) excavation and stabilization of four cavates and five associated masonry rooms (see also Turney 1948 and Kohler's [Carlson and Kohler 1989] excavation of a single cavate chamber nearby). There are several exposed hearths in Lower M, two of which were sampled in a pilot archaeomagnetic study. At the bottom end of the group is a large chamber located above a major drainage; the chamber was probably subdivided during at least part of its use. This "kiva" contains rock art, including a relatively well-preserved, green Awanyu. Almost all the rock art at this group is confined to room walls. Room M-33 contains remnants of red, yellow, black, and white wall paintings, as well as incised figures, on at least two walls and on several coats of plaster. Its 52 CAVATE STRUCTURES condition is now rather fragmentary and warrants careful removal and expert study of what remains. A route leads out of the canyon not far up-canyon from the group, but none seems to be located inside the room cluster. There is a set of six large holes above the four-level section of upper Group M (above M-41), but these seem to lead only to a ledge where possible room M-64 is located (Ranger Ed Greene reached this area by rope). The cavates and features we recorded in Group M are on the same level to a remarkable degree. We did not use an altimeter for this group, but the contours here are reasonably clear, and the 6180-6200 foot (1884-1890 m) level seems to identify the location of the majority of the rooms (in several places there are multiple levels of rooms). Using 6200 ft (1890 m) as a baseline for the main level of rooms (M-14, M-18, M-21, M-23, M-33, M-39, M-43, M-53, M-56, and-though they are well above the base of the cliff-M-59 and M-60), we estimated other elevations. Most other rooms are within 5 m of this elevation. The general level of upper Group M is around 6 m higher than the lower half. Group M is not indicated by a symbol on the USGS Frijoles Quadrangle, but Group L is suggested and shows as 6300 feet (1920 m) because of merged contours. This is probably too high; the elevation of the Rito below Group M is about 6040 ft (1840 m). At Rainbow House the Rito enters a narrow and fairly deep inner canyon, making it much less accessible to irrigation than in the main part of the canyon. Tsankawi Section, LA 50976 (Held no. LA 50024) Cavate rooms are abundant in many of the dry canyons north of Frijoles Canyon. With a few exceptions, such as Puye and Garcia Canyon, these features have received less attention from archaeologists than those in Frijoles. Lister (1940b) did some stabilization and mapping in several areas that were later divested from the monument, as well as at Tsankawi. Lister's work resulted in several maps, some of which cover long stretches of cliff and great numbers of cavates, as at Otowi and Tsankawi. This area takes its name from Tsankawi Pueblo (LA 211), a major "plaza pueblo" located on the mesa top with a commanding view in all directions. The mesa (or potrero) is formed by Sandia Canyon on the south and Los Alamos Canyon on the north. Tsankawi is a Tewa word meaning 4 the place [or gap] of the round cactus'. The orthography of Native American languages has been modified many times. Thus, a more recent, presumably more nearly correct, rendering of Tsankawi is Tsankawi'i, and earlier ones include Sankawi and Sankewi. The NPS and USGS use Tsankawi, and I have used that spelling for ihe ruin and as shorthand for the cavate group we studied, although there are cavates around Tsankawi Pueblo not included in this study. To make matters even more complicated, there is a pottery type named after LA 211; since it was named early, it has traditionally been spelled Sankawi Black-on-cream. The pottery name, too, has been rendered in many ways, usually without the initial tee (e.g., Sankawi'i [Kohler and Linse 1993:36]); the traditional spelling is used here. Cavates are abundant on the south and east slopes of the mesa and can be separated into reasonably discrete groups on the basis of distribution, as was done long ago for the Frijoles cavates. Such groups include the rooms located at the end of the western lobe of Tsankawi Mesa (not recorded by Lister), the lower rooms in the red tuff (Lister's C-123-C- 165 and C-170-C-174), the rooms in the south- facing rincon below the pueblo to the southwest (Lister's C-69-C-120), and the east-facing rincon below the pueblo to the southeast (Lister's C-15- C-45 and C-50-C-67). Such subdivision may impose some breaks that do not correspond to prehistoric groupings, and other methods could CONTEXT 53 be used to distinguish other groups. Still, this division has the advantage of creating more manageable analytical units for archaeologists. The east-facing rincon group was chosen for recording because it contains many rooms (even more than we anticipated) in both the red and gray tuff units. Also, the rooms in the red unit looked to be in good condition relative to most of the other rooms in that stratum. The "site," as we defined it, is delimited by two points in the mesa side. The point on the north and east extreme is small, and there are scattered cavates quite close to the LA 50976 group in the next rincon to the east. The point on the south is a more distinct break formed by a relatively high, sheer cliff and an absence of cavates for a considerable distance. Lister mapped and mentioned 49 rooms in this area. With a few exceptions, he included only relatively complete chambers and did not assign separate numbers to back chambers. Our count for the whole area is 119 rooms, reached by counting exterior rooms indicated by cliff features and by giving back chambers separate numbers. We retained Lister's numbers but replaced his C prefix with a TS (C indicated only "cave" and was used at all the other groups Lister worked in as well). New TS numbers, beginning with 501, were assigned to the rooms he had not numbered. TS-501 was chosen as a starting point because Lister's numbers for the whole Tsankawi area go up to 180, and it was thought desirable to flag the added numbers and keep them separate (TS-501-TS-570 were used). Lister's map clearly shows the relative locations of structures, and we adopted it as a location map to which we added our new numbers. The map included here (Figure 2.19) is based on the modified Lister map and a detailed sketch map drawn by the Bandelier Survey crew. The only other recorded work in this group is the removal of a secondary burial by Johnson (1960; see chapter 1 of this study). The rooms in the LA 50976 rincon are located on about four different levels, three in the thick gray-and-white tuff layer, and another in the underlying red tuff layer (Figure 2.20). Within the rincon several subgroups can be defined; these could be considered as the equivalent of room blocks in a masonry site or as separate sites. At the north end of the rincon, on the highest bench, is a group of about 25 rooms (TS-15-TS-26 and TS-501-TS-513) located in a porous, frothy tuff layer, overlain by the homogeneous, obviously indurated layer that forms the mesa caprock. They are closely spaced, and most front on a flat shelf 4-5 m wide that may have had structures on it: there are several sets of viga holes for rooms near rooms TS-21-TS-25, even though masonry rubble is scarce. Above most of this stretch of rooms on the rimrock is an elaborate, varied, and well-preserved rock art panel; the good condition of this panel results from the hardness of the top stratum. Among the petroglyphs there are three large rectilinear depressions ground into the cliff face. These features seem to occur only here and with a second cavate group recorded by Lister in a rincon to the west of LA 50976. After some discussion, we decided to call these features cliff niches, though their bases slope somewhat (Figure 4.23). There is also a cleavage in the rimrock with paired beam seats reminiscent of scaffold supports, like those at Scaffold House in Navajo National Monument. The densest cluster of rooms is on the next level down, on the east-facing slope and about in the center of LA 50976. Rooms TS-54- TS-61 and TS-530-TS-533 are packed onto several levels, and rooms TS-62-TS-65 are close by. Down yet another level is a smaller group of rooms, the last cluster in the white and gray tuff: TS-51-TS-53 and TS-540-TS-546. The top of the red tuff layer forms a distinct ledge for most of the length of the rincon. This coarse, very soft tuff, possibly the Tsankawi Pumice Bed (Bailey et al. 1969:14- 15), is exposed for thicknesses of 3 m or less for most of the middle of the rincon, with thicker exposures at each end. The rooms in the red tuff form a subgroup because of their elevation; Petroglyphs Petroglyphs Petroglyph Red Tuff *^?Pf-i \ f ** 43 ^ \«A 565 564 CONTEXT 55 LA 50976 PLAN VIEW D Petroglyph Bedrock Trail ^^ Bedrock Cliff or Outcrop — . — Ceramic Sample Areas Oripline ^ Talus Boulder 5 i , . , i i 10 15 20 meters i 10 20 30 40 50 i i i i i feet Scale Approximate 25 506 507 \ y, Petroglyphs -501 Hand -and -To* Holds / | Limit of Northern Ceramic Sample Unit , ..': Hand -and -To* Holds © u") m -N m ON ■«t »n On «-H U"> d >— i en — i e*> d T* ^H d 1-H ^* d CN VO 2 o CN — 00 o en - 00 5 VO en CN VO - *— < CN PH •* VO o\ i-H ts CN ^ d d VO d d en d d CN — < r» 5 »n >0 en en o\ Os 00 on CN 5 »— c vo vo en <* CN VO i—i m § ^^ CN 00 en 8 en Ti- v© 1 en ON 5? 00 On w-> — CM VO CN en >-H (S en ■TN en en o VO en VO CM q NO en ■^■ en CM oo en CM 00 NO ON i/"> 00 in vo 00 en en 0\' vo q »h »n Tt on r~; en 'H d vo vd ■-! en -h ON —I ON 00 r~ en o CO en r-~ en 5 -H -H »-< U-) VO l/^ en o rN i/"> >tn t* 1 1 S3 03 00 00 2 1? 3 S> "S ^ "8 c a -c "1 E 1 a o a M V &o 5 CO U Oh "S -8 1 T3 ' 1 0QOaooOHOHl> CONTEXT 69 a o * 00 2 o VO 8? oo -H o ~H CS ro 6 6 ON 6 d 6 1* 1-H ^H CS •— 1 ^ -/-> Uh £ K 3 £ CO H o 2 3 VO —l CO © •i & r-8 ■t o m oo cs CO 00 cs co oo CS On CO On CO ON ^ cs 1 3 1 5 o Z % « I § 1 a ft VO CS ^ 6 CO cs tJ- vo vo 6 6 © m ifl m 6 VO 6 Ifl H H N -H ^-, dodo do ^h co co t- o 0\ - h tn " o cs CO NO ON ON d CS 00 cs On CS On CS cs On NO CO d 4 d d cs co VO CO CO VO d cs CO -H CS On On CO ~h d d CO ~* d CO d CO t-» 1 -s 5 « « £ « .2 o ^ a, Oh 3 U o >> •3 J ►, 8. o t-i ^ *^r 3 S *< 5. & JJ OQ o o o Agua Ciene Ciene San C Glaze o o o ^ 00 60 00 4J H 1- «- S j*. >» a 1 o o o i o Q 1 .a m J & d § u n CO 70 CAVATE STRUCTURES 5 6 o a i * 3 o 1 * o H ou * la a * H CO U o © d m ^ , **! *i °. 00 vo On »-H ^H Tf m w> r- o\ ^ « ^ oi d « d oo o On NO (S 00 ON ON m • fN J3 u 3 "3 § own CSV© vnvriON(S ts>nNor^ V/"Nl/-> Nifl 8u">w>© VTNO n ^ c» ■<1- o *-< d d rJ 8 - 1 NO s On ■* Tf ON c4 8 3 fONO en CS T* oo 5 «n 1 do rJ vi ©u-> PHNO *-* *-H NO - 1 - ON NO m 00 ON CO 1 ©>/-> CO NO 3 CONTEXT 71 Black-on-cream suggest either that cavates were not occupied as long or that the use of Sankawi Black-on-cream differs at the cavates. In the Frijoles groups, Group A (1225- 1550) is estimated to have been first occupied at 1225 because of the amount of Santa Fe Black- on-white relative to Wiyo and later black-on-whites. Group F (1100-1175 and 1315-1550) seems to have seen two separate occupations, one during the late Developmental Period and another during the Classic Period; the absence of Santa Fe Black-on-white suggests that Group F was not in use during the Coalition Period. This pattern is consistent in both cliff cavate and tent rock ceramic subgroups at Group F. With only two Santa Fe Black-on-white vessels represented, Group I (a.d. 1275-1525) shows little evidence of a late Coalition Period occupation; at the terminus of its ceramic sequence, Glaze D polychromes are more similar in rim form to earlier Glaze C forms than to the Glaze D examples encountered in other cavate groups. These facts suggest that Group I may have had the shortest period of occupation of all the sampled cavate groups, with a late beginning date and an earlier ending date. Group I also has the fewest ceramics tallied, though the sample area is comparable to those covered at other cavate groups. This low frequency of sherds may also be a reflection of a shorter occupation span; whatever the cause, a sample size problem is evident which is critical in assessing the validity of the temporal placement of the group. Group M (1250-1550 and post-1695 [to 1725]) shows the only clear evidence of early historic ceramics. A single Kapo Black and historic polychrome jar base were the only representatives of this late assemblage in the 1986 sample. Turney's 1948 work (see Table 1.1) presents other evidence for Pueblo Revolt-early eighteenth-century occupation at Group M. The mid-1200s- mid- 1500s occupation at Group M is indicated by the continuum of decorated types from Santa Fe through Biscuit B and Glaze E ceramics. Glaze D seems to have lasted longer than the traditional 20-year span, ending in 1515, while the Glaze E sample is largely finished in muted fawn and tan tones, a style of coloration likely coeval with late Glaze D and predating what may be the later use of light cream slips in Glaze E. This is the basis for the terminal date assignment of 1550 for the initial occupation of Group M. No Glaze F was identified at any of the cavate groups sampled. Ceramic samples from the Tsankawi Pueblo proper suggested a date of 1250-1675 based on conventional time spans for the production of Sankawi Black-on-cream in conjunction with minor but consistent amounts of Santa Fe and Wiyo Black-on-white. Beginning dates nearer 1300 and ending dates nearer 1650, however, may be more appropriate. Cavate samples suggest a similar time span for those features. No confidently identifiable glaze types were found in the cavate samples, but the impression from partial rims and surface attributes of glaze sherds is that glaze import became more common during and after Glaze D; a similar situation was noted in the mesa-top pueblo samples as well. Two other cavate groups in Frijoles Canyon, sampled during the 1985 test survey, also had ceramics indicative of late occupation. Cavates B-70 and B-71 (portions of Groups C and B respectively as defined in the 1985 pilot survey [McKenna and Powers 1986]) had even fewer sherds than the present sample of cavates, but they had early historic types of Kapo Black, Tewa Polychrome, and Glaze E. These cavate groups were dated 1450-1700 and 1500-1700 respectively. Here again, sample size problems undoubtedly affect the assessment of occupation span. Cavate pueblos sampled away from the main clusters in Frijoles Canyon and around Tsankawi show earlier, more discrete occupation spans. LA 50909, south of Frijoles Canyon in the Corral Hill area of Bandelier National Monument, was primarily a Coalition Period site dating to 1200-1425 (McKenna and Powers 1986: Table 9). Coalition Period occupation was also the main period for cavates surveyed north of Bandelier by the Pajarito Archaeological 72 CAVATE STRUCTURES Research Project (Hill and Trierweiler 1986). Especially when viewing sherds on the surface, the long-term aggregation associated with communal pueblos and cavates is a major factor in making occupational episodes at cavate groups indistinct. The trend to aggregation is stronger at lower elevations (such as Tsankawi) than at higher ones (such as at Yapashi). Ceramic Patterns The chi-square statistic was used to test for differences in the ceramic samples from the cavates. Vessel forms, reduced to open ("bowls") or closed ("jars"), ware, and provenience, were the dimensions of the tests. Data and test results are presented in Table 2.4. The 0.05 level was accepted as significant. In the tests of intrasite sample areas for Group F and at Tsankawi, we found no differences in the general vessel populations. Clearly, surface ceramic evidence cannot be used to support special use of cavates at the intrasite level as presently measured. However, considering only decorated ware in the Tsankawi red tuff cavates (south sample area), there is a significant difference in vessel form between the cavates in the western and eastern halves of that subgroup (Test 1): a greater than expected number of whiteware jars is found in ceramics associated with the western red tuff cavates. In the comparisons between areas at Tsankawi (Test 2) we found a significant difference between the northern area and the central and southern areas. The number of jars is higher than expected in the northern area, while bowls are more common in the central and southern areas; northern and southern area distributions are the strongest contributors to this pattern. Distributions of service ware only (Test 3) reinforce this observation, with bowls being more common in the southern and central areas and jars more common in the northern cavates. A test of ware distributions (Test 4) shows culinary (exclusively jars) more than expected in the north area, which again contributes to the value of the chi-square in determining a significant difference in form distribution between these cavate subdivisions at Tsankawi. In comparing cavate ceramic distributions with those of the pueblo, we found general agreement in form ratios (bowls :jars- Tsankawi= 1:2.9; cavates = 1:2.9), but the number of decorated bowls (Test 5) is higher than expected at the cavates, as compared to the main pueblo midden sample, where jars are more common. We found no significant difference in form distributions among the Frijoles cavates. Form distribution, however, does seem to differ between cavate sites and some communal pueblos. A test (Test 6) of distributions of culinary versus service ware shows that culinary ware is significantly more abundant at cavate sites, while service wares are relatively more common at communal pueblos. Since culinary ware consists almost entirely of closed forms, it follows that closed forms will be relatively more common on cavate sites in later tests. Comparison of all forms among the Frijoles cavates, Yapashi, and Tsankawi Pueblo plus cavates (Test 7) shows the Yapashi and Tsankawi samples to be higher than expected in bowls while the Frijoles cavates are higher in jars. A test of the Tsankawi Pueblo and cavate ceramic forms showed no significant difference, permitting us to treat them as a unit in Test 7. A test considering only service wares (Test 8), however, shows that the Tsankawi sample is higher than expected in bowls and the Frijoles cavates are higher in jars, while forms at Yapashi occur as expected. Some differences may be related to trends in ware forms: glazed ware production is more often associated with closed forms and Tsankawi's matte-paint bichrome tradition with bowls. Some differences may also be related to differing emphasis in site functions involving ceramic use. However, the consistently low value for the coefficient of contingency (Table 2.4) indicates a weak association between site types and the ceramic dimensions of form and ware in the present data. CONTEXT 73 Table 2.4. Ceramic Form and Ware Data and Significance Tests. All Ceramic Forms 3 Test No. Decorated Only Site Bowls Jars Bowls Jars Test No. Frijoles cavates A 47 206 47 26 F 119 424 119 80 I 38 136 38 26 M 110 553 110 70 B-70 (C) 14 56 14 15 B-71 (B) 21 54 21 14 Tsankawi cavates North 69 280 2 69 29 3 Central 40 101 2 40 17 3 South 120 285 2 120 21 3 Frijoles cavates A-M 314 1319 7 314 202 8 Tsankawi cavates 229 666 229 67 5 Tsankawi Pueblo 255 645 255 150 5 Tsankawi all 484 1311 7 484 217 5, 8 Yapashi Pueblo 234 572 7 234 124 8 Intra-site Group F Lower 53 210 52 43 Remainder 66 214 66 37 Tsankawi (south sample) East half 83 196 83 9 1 West half 37 89 37 12 1 All Ceramic Wares Site Culinary Decorated Test No. Tsankawi North 251 98 4 Central 84 57 4 South 264 141 4 Tsankawi cavates 599 296 6 Frijoles cavates 1104 540 6 Yapashi Pueblo 448 377 6 Tsankawi Pueblo 495 450 6 74 CAVATE STRUCTURES Table 2.4. (continued) Test No. N x 2 degrees of freedom probability Contingency Coefficient Cells <5 1 141 5.456 1 0.020 2 895 10.250 2 0.006 0.106 3 296 9.216 2 0.010 0.174 4 895 7.929 2 0.019 0.019 5 701 16.596 1 0.000 6 4309 84.469 3 0.000 0.139 7 4234 39.476 2 0.000 0.096 8 1575 8.834 2 0.012 0.075 "Unknown forms excluded. The distinction in ware varieties between Tsankawi and Frijoles needs no test to illuminate it. All samples show culinary ware as the majority of the ceramic assemblage. The culinary wares are distinct between the two areas. The micaceous paste Tesuque series dominates the Tsankawi assemblage, and the sand-tempered utility wares of the Rio Grande series are most abundant in the Frijoles cavate sites. The pattern in decorated or nonculinary ceramics is even more apparent: glaze-paint ware is clearly associated with the Frijoles cavates (23 percent of the assemblage), while nonglaze service wares are associated with Tsankawi (31 percent). Although the ordering of the ware representations is clear, there is relatively more matte-paint material in the Frijoles sample (9 percent) than there is glaze-paint material in the Tsankawi sample (2 percent); see Table 2.3. Summary As might be expected, the surface sherds inventoried at the Bandelier cavates provide date groups too broad to be useful for fine temporal comparison or structuring in analyses of architectural variability. All architecturally sampled cavate groups overlap for some undeterminable period(s) during their occupation; it is not possible to isolate "earlier" from "later" groups of cavates either within or between sites based on the ceramic sample. Temporal differences are present, but the nature of the ceramic samples does not permit their correlation to groups of cavates within sites or the meaningful temporal placement of cavate sites relative to one another. Temporal holes exist which may relate to some cavate variation, but confidence in assessing those "holes" depends on the willingness to accept chronological evidence from very sparse, widely scattered, and heavily picked-over ceramic assemblages. The present sample suggests the following major temporal distinctions: (1) the early, apparently noncontinuous, occupation of cavate Group F; (2) the post-Pueblo Revolt historic reoccupation of Group M; (3) the apparently relatively short, possibly less intense, occupation at Group I. These distinctions generally suggest that occupation of cavates was discontinuous, at least in the Frijoles groups. The spotty evidence of historic reoccupation is clear at Groups M, B, and C. Prehistoric periodicity in occupation is CONTEXT 75 also suggested by stylistic trends in some cavate groups, such as the similarity of Glaze D to C rims in Group I, and by the apparent discontinuity in some type sequences (though this may be reading more into the sample/type tabulations than is warranted). At Tsankawi no temporal distinctions are evident among the cavates or between the cavates and the main pueblo. Some functional differences are suggested between cavates and between cavates and nearby communal pueblos. A difference in functional emphasis at cavate sites is implied by the greater-than-expected number of culinary vessels. The association of closed forms with the Frijoles cavates is particularly strong, since both culinary and service- ware jars occur more often than expected. Such a difference might indicate use of cavate sites for living and cooking, or it might result from an emphasis on storage in cavates. Better samples and more detailed analysis of excavated samples are needed to refine this interpretation. While culinary ware forms the majority at all sites examined here, we suspect that this may be at least in part a function of long collection of decorated sherds. Some differences in the proportion of service ware in sections of the red tuff cavates and in vessel forms between the cavate subgroups at Tsankawi also suggest some functional variation within cavate groups. Economic and possibly ethnic differences are certainly reflected in the relative proportions and types of wares in the Tsankawi and Frijoles samples. The relative proportions of micaceous to sand-tempered culinary, and matte-bichromes to glaze-paint ware, are reversed in the two areas. This clearly indicates differing patterns in ceramic circulation and, by extension, in spheres of production and exchange. Nevertheless, functional patterns of ceramic use, when compared with a limited sample of community pueblos, tend to crosscut these differences in both the Tsankawi and Frijoles cavate samples. This indicates a measurable difference in site types, which may be more clearly expressed in architecture than in ceramics. 1. This count includes all traces of former rooms. Figure 2.10 shows at least 32 excavated rooms. 2. Santa Fe Black-on-white is also a minority type at Tyuonyi and, as at the cavates, its presence appears to reflect early occupation. Although Tyuonyi was probably constructed between the late 1300's and early 1500's, tests conducted by Onstott (1948) below the earliest surface in Tyuonyi's plaza revealed Santa Fe Black-on-white and contemporary utility wares suggestive of a late Coalition occupation. A higher plaza surface with Santa Fe Black-on-white and early glazes is either associated with the remains of an early structure or the early building stages of the present pueblo (Onstott 1948; Van Zandt 1994). Recording Procedures, Group Attributes, and Cavate Condition Recording cavates presents an interesting archaeological problem. On the positive side, many features not normally visible without excavation are exposed without the time and effort required for digging. On the negative side, cavates are difficult to survey because they are numerous, because they are exposed to various types of degradation, and because recording all the data available would consume much field time. To take advantage of these data-rich, well-preserved features, a survey on the Pajarito Plateau must compromise between recording a few cavates in extreme detail and recording many of them in insufficient detail. A major goal of the 1986 fieldwork was to develop a procedure for efficiently recording enough cavate data to understand variability in their construction and features. Although it was not so used, the recording method was intended for dealing with cavates encountered by the survey. The sample of cavates and associated features recorded by this pilot study provides baseline information on cavate variability, as well as on the condition of this particular set of features. Recording Procedures Cavate and Noncavate Forms The prehistoric Pueblo people had many techniques for using the tuff cliffs of the Pajarito Plateau. In some cases they hollowed entire chambers out of the cliff, leaving a small door. In others they seem to have excavated a large chamber and then built the front wall of the room entirely of masonry. The proportion of room excavated into the cliff ranges continuously from a fully excavated chamber with a tuff front wall, to a masonry-fronted completely excavated chamber, to masonry-tuff combination rooms, to mostly masonry rooms with back walls somewhat indented, to rooms that used the otherwise unmodified cliff with some viga holes bored in it. Edgar Hewett recognized and described this range at Puye: We note here three classes of dwellings. 1. Excavated, cavelike rooms, serving as domiciles without any construction in front. 2. Excavated rooms with open rooms or porches built on in front . . . 3. Houses of stone, one to three stories high, with corresponding terraces, built upon the talus against the cliff. In these groups the excavated chambers now seen in the cliff wall were simply back rooms of the terraced buildings. . . . An examination of the talus discloses remains of several villages of considerable extent that were built against the cliff. (Hewett 1908:19) 77 78 CAVATE STRUCTURES This variability adds another dimension to the problem of recording sites with cavates. The situation is further complicated by processes of erosion and falling blocks of cliff, and by the fact that masonry is more susceptible to disintegration than are cavate rooms. To attempt to address the problem, we used two types of forms in the field, one for cavates and the other for noncavates. The noncavate form was designed for rapid recording of rooms showing little cliff excavation. Such rooms clearly lack many of the features that more completely excavated rooms may have (Figure 3.1). The difficulty comes when deciding which form to use for borderline features. Copies of the forms as they were used in the field, the coding conventions, and the dates of added categories are given in appendix 1 . Our use of these forms was influenced to some extent by the evolution in our perception of the project's goal. Initially the focus was on cavate structures as unusual architectural features in need of better recording, with a nod to other associated features. Early in the field season we adopted a somewhat different and, we believe, better approach: the recording of whole sites. In two respects, we did not attain the goal of full- site recording. First, we noted but did not record in detail architecture that had no manifestation on the cliff, such as rubble mounds and walls. Second, we did not record artifact samples. The latter shortcoming was partially remedied by a later field inventory of surface ceramics (see chapter 2). The use of two forms allowed us to separate features into analytical classes and also shortened the time spent recording partial features. Reduced recording of partial features is acceptable because they contain less aggregate information and waste many form entries. As our working definition of a cavate in this study, we specified that a feature must be enclosed on at least three sides and have some remaining sheltering qualities. Some judgment is required to gauge the completeness of a cavate room, and the decision is complicated in some cases by erosion of the exterior edge. We recorded cavates we considered reasonably complete and made an ordinal estimate of the completeness of all cavates. The noncavate category includes badly collapsed former cavate rooms, excavations in the tuff for rooms that were largely built rather than excavated, and other prehistoric, artificial cliff features. Completely natural pockets or overhangs showing prehistoric use should be recorded with their natural origin noted, but during our recording we found no examples of cavates we regarded as completely natural. In the field the distinction between a cavate and a noncavate usually centered on whether the feature was judged to be sufficiently intact and to contain sufficient information to merit detailed recording. Early in the recording (especially at Group M) there was greater variability in application of differentiation criteria. On the whole, as the season progressed we tended to make greater use of the noncavate form in the case of partial rooms. Group M has a relatively low number of noncavates, which is partly an artifact of this adjustment period and partly the result of the occupants of Groups F and A having made greater use of the cliff as a back wall with no excavation than did the builders of the other three groups. A standardized list of noncavate feature types did not exist during the field recording. The categories we used, however, fell into a fairly restricted range and were placed into the following groups during the data correction phase: back wall, partial cavate, filled cavate, chamber, hand-and-toe hold route, cliff niche (see also Table 3.1). The cavate form includes an estimation of room completeness, which thus becomes an important screen for chamber analyses, since it allows sorting of observations into reliability categories. Another carryover from the initial inclination to treat the two categories differently is that features within noncavate rooms were RECORDING PROCEDURES 79 Figure 3.1. Schematic drawing of three cavates and two noncavates, showing feature types. At upper left is a row ofviga holes defining a noncavate backwall; at lower left is a mostly filled cavate that would have been recorded on a noncavate form in 1986. At mid-left and upper right are the most common form of cavate, which would have been closed by masonry, and at lower right is a fully enclosed cavate with an exterior door. Note the truncated pyramid shape, and how wall plaster extends only partway up the wall. Ceilings are smoke-blackened and show digging stick marks. The features shown, with code numbers, are as follows: 2 Exterior door 4 Firepit 6 Floor ridge 7 Floor pit 8 Large floor-level niche 9 Wall niche 10 Slot 11 Viga holes 13 Loom anchors 14 Upper loom supports 15 Smokehole 16 Vent 19 Interior door 26 Exterior opening 27 Ceiling 28 Indeterminate holes 37 Wall depression 40 Metate rest 80 CAVATE STRUCTURES sometimes measured by ranges (which was our initial intent) and sometimes individually (which we concluded was preferable). Features recorded only within a range obviously cannot be used the same way as individually recorded ones (though ranges are decidedly faster to measure and record). In the spirit of full-site recording and monitoring of the full range of variability, the two-form approach had its uses, but the forms should have been fully congruent in terms of codes and feature recording. It is probably preferable to have a single form with a rigorous set of room types, along the lines enumerated by Hewett (1908; see above). Although we had both computer and paper forms for recording data, only the paper form is presented here and in appendix 1 . The computer form was slightly different, but all the same information was recorded in both formats, and all data are, of course, now equivalent in the various data bases. Certain classes of information do not apply to certain sets of features (firepits do not lead to other rooms, and a wall does not need a precise location). These classes of features are separated on the paper form, with only the variables relevant to each included. We stressed the avoidance of redundancy in recording, so that notes relative to specific features, such as condition of walls, were entered with the features rather than with general room notes. portion was usually a straightforward assessment, though partitioning walls in curved chambers sometimes required subjective judgments. Azimuth location of features and measurement of height above floor give a relative location of all features relative to all others. The compass was usually located by establishing the intersection of the diagonals across the room. Taking an azimuth to a feature requires reducing it to a point; we shot to the center of features. Long Awanyu petroglyphs and whole walls are not amenable to this treatment, so we did not take azimuths to them. Ideally, every feature would also have a distance measured to the "Brunton station" (the compass location), but this procedure would greatly increase the recording time. As a compromise and a means of better locating walls, we adopted the practice of taking angle and distance to distinct chamber features. Chamber corners were given preference, but other features occasionally had to be used. Given these basic measurements it would be possible either to relocate the measurement point with considerable accuracy or to create a rough plan of a room from measurements. The measurements would be especially useful for reconstructing feature layout to record station-to-feature distances to all floor features, though we did not do so in 1986. Rooms that did not have enough space or floor for setting up the compass lack azimuth locations for features. Discussion of Codes Shapes Although all the "variable states" are presented in appendix 1 with the forms, several variables require further explication, and some need reconsideration after field use. Chamber Location Locations of features within cavates were recorded with a two-part system: by chamber portion, such as right or left wall or floor, and by means of azimuths from a point as near as possible to the center of the room. The chamber Given the variety of chamber and feature shapes present in the cavates, it is clear that measurements are far more meaningful if associated with a shape. This requires visualizing features as regular geometric figures, which is often not easy. The effort is worthwhile, however, because it makes the measurements more accurate and appropriate while also serving to describe the features. The use of fractions further refines the information collected. RECORDING PROCEDURES 81 Geometric Fraction The two fraction codes-geometric and feature—probably caused more discussion and confusion than any other part of the recording process. Geometric fraction is intended to be a descriptive code and to complement the assignment of a shape to a feature. Because features often do not conform well to purely geometric shapes, geometric fraction is an estimate of how much of the shape assigned and measured is actually present. For example, if a chamber most nearly approximates a rectangular solid, but a portion of the whole solid is missing because of curvature of the cliff at the opening of the room, estimating the amount of the solid that is present tells the analyst that the volume of the room is perhaps only 80 percent as large as the measurements suggest. It is often possible to refine the estimate of how much of a shape is present by measuring the "missing" portion. Portions of circles were used to record semicircular shapes. By multiplying the figure derived for the shape measured by the geometric fraction, actual volumes and areas can be more closely approximated. Feature Fraction In contrast to the descriptive nature of geometric fractions, feature fraction is interpretive. The recorder assesses the feature as it now exists and estimates how much is missing. There is a subjective element here, of course, but-particularly in the case of fairly complete features-it does not require wild guessing. These estimates can be used as confidence limits in analysis of feature measurements. That is, measurements for whole or nearly whole features can be used with confidence; the more of the feature that is missing, the more suspect the measurements become. In cases where it is very difficult to estimate how much of the feature is present, usually because so much of it is missing, assignment of a very low feature fraction flags the measurements as probably unreliable for inclusion in size assessments. Feature fraction values of greater than 0.7 were usually required for inclusion in feature dimension analysis. As an example of how the two fractions interact, consider a viga hole that is basically cylindrical but has a portion of the cylinder missing because of its location in a slanted wall. It might be recorded as having a geometric fraction (GF) of 0.8. Inspection of the edges of the viga hole show that the feature is still intact as it was used, so the feature fraction (FF) is 1.0. The GF and the FF will be the same if it is apparent that the feature was originally a regular shape, none of which has been removed through deterioration. One cause of confusion is the tendency of some features (openings, especially doors) to become larger rather than smaller when they deteriorate. Whenever possible we measured what we considered to be the original feature. This was possible when, for example, two sides of a door remained but the other two were broken away; in such a case, the FF would be entered as 1 .0. Where we measured an opening that was clearly larger than the prehistoric fact, we could assign FF greater than 1.0. Again, as values approach 1.0, they are more reliable. To reiterate, geometric fractions describe what is present and help correct area and volume calculations; feature fractions are interpretive judgments of how much of the original feature is present and place a rating on the reliability of measurements. FF values close to 1.0 could be used to estimate full measurements of original features. Evidence for Construction This variable involves the straight- forward observation of various construction features. It does, however, result in some mixture of phenomena. That is, digging stick marks (Figure 3.2) are clearly evidence of part of the construction of a cavate. The other code values deal more with features that were made, 82 CAVATE STRUCTURES Figure 3.2. Example of digging stick marks in the ceiling of A-10. These are especially visible because they cut through the smoking of the ceiling. Presumably they are evidence of cavate construction and/or expansion. such as shaping of corners or doorways. Although these are somewhat inconsistent, I do not see this as a major problem, especially in view of the scarcity of evidence of construction other than digging stick marks. Some chambers contain fairly clear evidence of having been enlarged after having been used for a while (Figure 3.3). Although we had no code for this (it was mentioned in the verbal notes), one might be useful in future recording. similar, though it can vary considerably within a site area, notably at Tsankawi and Group A. Slightly redefining this variable would have permitted a meaningful entry for each cavate, rather than lumping many cavates into an ill- defined "normal" group. Somewhat by default, "normal" in our observations means basically the fine-grained, white to very light gray tuff seen particularly in locations such as Frijoles Groups Fandl. Unusual Tuff Characteristics A more fruitful approach to this variable would have been to call it something like tuff type. The tuff in given areas tends to be Rooms Up/Rooms Down This observation places a room within its group by specifying how many rooms are up- canyon and how many are down-canyon to the RECORDING PROCEDURES 83 7 ■ Figure 3.3. Evidence for chamber expansion in M-10. The plaster and smoking of the back wall stop at the juncture with the right wall. A single coat of plaster is present on the right wall, while the back wall has several coats. ends of the group. It turned out to be a difficult blank to fill in the field, so the values were all entered after the field season when complete recording and maps were available. Counting rooms is far from exact in these sites because cliff features are subject to several interpretations. We made a careful count as we began to record each group in order to assign room numbers, but we invariably found that we had to add rooms as we studied features carefully during recording. Thus, counts from parts of groups not recorded (as at Groups A, F, and M) are likely to be somewhat low. In determining the entries for this variable, we treated columnar sections of cliff, so that multistory rooms above one another and rooms that opened off the back of other chambers all had the same number of rooms up- and down- canyon. This means that the counts are not a steady progression as one moves through the group unless rooms are single-level without back chambers. In addition to the potential for seeing whether location in group seems to relate to size and function, this observation is helpful in locating rooms. Level An observation that could better and more easily have been made in the field was added after we returned from the field. Each room was placed according to its level (or story). We also noted how many levels were present at that location. Thus each room now has an observation of the type "second level of three." In some instances intervening or underlying levels may have been present but not visible, so we assumed some other means of access to the upper levels. Such an assumption is more plausible for fully excavated cavates than for rooms evident only as back walls. Recording this attribute has the added practical advantage of helping to locate rooms, as does the rooms up/rooms down variable. This variable is also relevant to the possibility of assessing the distribution of functions and room sizes. Noncavate Data Sets As noted, the separate form for recording noncavates was a good idea that would have produced more useful results if it had been more completely compatible with the fuller cavate forms. Anyone using noncavate records individually should be aware that the first feature notes (rather than the base notes) often contain most of the observations on the noncavate; this is because the first feature is often most of what remains to be recorded. Several of these 84 CAVATE STRUCTURES compatibility problems could be fixed by examination of the records. Here again, we made some later modifications to the data set. Feature Type On the field form this was a verbal entry, which gave rise to a variety of classes. The variability was reduced and the responses were made to follow a more consistent set of criteria. I would recommend that future cavate recording use a coded list, such as this, rather than verbal entries. Possible responses now include: 1. Back wall. The category for which the noncavate form is most appropriate is the common instance where a masonry room was clearly placed against the cliff. Evidence for this feature includes rows of viga holes, wall features (niches, other holes), and plaster. In most cases these features involve little or no excavation into the cliff. 2. Partial cavate. Features included in this category have at least portions of one or more walls in addition to the back wall. As a result of this definition and the nature of the archaeological remains, this category contains a wide variety of features. The use of the noncavate form was intended in part to cover remains that, for whatever reason, had little visible information. Thus features that might literally be "partial cavates" were in many cases recorded as cavates. Partial cavates probably result in about equal measure from the loss of fronting masonry structures from only partially excavated rooms and from the loss of natural exterior walls. Although we cannot know with certainty, I believe the former is probably the more common cause. There is no doubt that cavate rooms are present, but they are largely unmeasurable. 4. Chamber. In a few cases the noncavate form was used for more or less intact chambers. These should probably have been recorded on cavate forms. They were placed on these forms because there was some question as to their prehistoric use and artificial origin. Because of the absence of features, however, loss of data and comparability was minimal. 5. Cliff niches. These features are peculiar to Tsankawi (see chapter 4). They appear to have been outside rooms so that assignment of room numbers to them is inappropriate. They are recorded on forms both individually and in groups; they have all had their room-number values set to zero. 6. Trails and hand-and-toe hold routes. There are two instances of extramural routes on noncavate forms from Group A. Other examples are present at Group I and Tsankawi, but they were not recorded on forms. Although recording these is difficult, perhaps it should have been done. These features also have room number values of zero. Table 3.1 shows the distribution of these noncavate features at the groups studied. Connected with Cavate This is filled in only when a noncavate and a cavate are physically linked by a feature. The most common example occurs where a door to a cavate goes through a back wall or partial cavate. In a few cases, a vent passes through a wall between a noncavate and a cavate, and these have been coded as connected. 3. Filled cavate. Certain areas, particularly at Tsankawi and Group M, have a great deal of architectural and natural fill against the cliff base. In these areas only the top of an opening to a chamber is sometimes visible. Condition/Damage The first version of the noncavate form had only a single code for condition. During the recording of Group M (M-53 and after) this was Table 3.1. Distribution ofNoncavate Feature Types. RECORDING PROCEDURES 85 Feature Type Group A Group F Group I Group M Tsankawi Total Back wall Partial cavate Filled cavate Chamber Hand & toe routes Cliff niches 37 10 1 2 38 7 14 2 9 7 3 27 125 21 45 2 7 2 3 2 4 4 Total 50 45 16 19 56 186 split into human and natural damage; we reconstructed the damage codes for the early forms from the information on the forms. Cavate Type This variable was added after the fieldwork to link the cavate and noncavate records. All cavate records received the code "cavate," while noncavate records were coded as one of the noncavate feature types listed above. In retrospect, it would have been useful to classify both "fully enclosed cavate" and "excavated portion of partially masonry room" as types of cavate, since many of the current noncavate "partial cavate" features are excavated portions of masonry rooms. The two cavate subtypes would have somewhat different information recorded, particularly for doors: enclosed cavates have true "exterior doors," while the back portions of partly masonry rooms would be "exterior openings." Collections Materials collected were limited to perishable items that seemed likely to carry information and to be possibly at risk of disintegration (see appendix 1). Two possible coprolites and a fused clump of corn were collected from Group M. A possible squash rind was collected from Group A. Most important, several bones from a very small infant were collected at Tsankawi. These remains were exposed in the disturbed fill of a room. Given the heavy visitation at Tsankawi, we decided to collect the exposed elements pending decisions on how best to deal with the vandalism of the deposits and how to most properly treat the burial. Rock Art Recording rock art is a time-consuming and specialized process. In recognition of this fact, we initially planned only to note the general size, location, and subject matter of rock art as features during recording. We were extremely fortunate in having June and Bill Crowder to do supplemental recording of the rock art. They took photographs, recorded locations, and compiled a summary of the rock art that was present in the sections where we worked, providing a much more detailed record of rock art than would have otherwise been available (see chapter 4). 86 CAVATE STRUCTURES Mapping We made a plan map for each of the Frijoles groups. The procedure was to establish Brunton compass stations, run a long tape along a recorded azimuth, and take right-angle measurements (right angles verified by Brunton) to edges of rooms along the line. Later Brunton stations were established on the line as made necessary by cliff curvature and topography. The first group we mapped was Group M, and we attempted to place all rooms at all levels, which proved time-consuming and resulted in a confusing map. The remaining maps measured only the rooms at cliff base, with higher rooms recorded in their relative locations, with level (or story) keyed. The maps locate only the edges of openings of rooms and do not reproduce floor plans. As discussed in chapter 2, this volume includes plane table maps made by the Bandelier Survey crew when they are available. The map for Group A (Figure 2.5) was made by the technique described here; the maps for Groups F, I, and M (Figures 2.10, 2.14, and 2.17) were made using a plane table and alidade by crews from the park survey. The plane table maps do render chamber shape, but not floor plan. We did not draw a plan for the Tsankawi group (LA 50976) because of the availability of Lister's map (1940b). The map included here is based on Lister's enlarged and updated map and a park survey detailed sketch map to show rooms added by this project (Figure 2.19). Frontal elevation or profile maps were drawn to compensate for the lack of measured plan locations (Figures 2.6, 2.11, 2.15, 2.18, 2.20). Initially, we hoped that the NPS Branch of Remote Sensing would produce metrically correct elevations for each group; this plan was not followed, however. We therefore had to produce frontal elevations using field sketches and combinations of distant and close-up photographs. Because they attempt to render very irregular surfaces flat, and because of the perspective problems inherent in transferring from uncorrected photos, scales are approximate. Although not metrically correct, these elevations do show relative locations and shapes of rooms. The survey field crews also made more detailed elevation maps of Groups F, I, and M. The Group F survey map and elevation have been modified to reflect cavate project room numbers; the survey crews used cavate project numbers for the Group I and Group M maps. Individual maps were drawn for only a few rooms, all at Tsankawi. One of these rooms had a complex floor plan due to chamber expansion, one had many floor features revealed by extensive brushing and is heavily visited, and one contained the infant remains and some disturbance, as noted above. Although it would be desirable to have plans and profiles for each feature, angle and size measurements and complete video coverage help compensate for that lack. Photography and Video Recording Photographic recording fell into four categories. First, Bill Crowder made photographs (mostly 4 x 5) of the fronts of all the Frijoles cavates that we recorded. Detailed photography of cavate fronts is complicated by the steep slopes in front of many of the rooms, as well as by the high vegetation in some areas. Second, as noted above, the Crowders took photographs of each rock art panel in the areas recorded. Rock art photography can also be difficult, but through experimentation and darkroom techniques, most figures are visible in the photographs. Third, the archaeologists photographed specific examples of features and room groups. Such record photographs were kept to a minimum because of the complete videotape coverage. Lister (1940a,b) had taken some pictures of groups in which we worked. In Groups A, F, and I and at Tsankawi we took pictures from the same angle to show change after 46 years (Figures 2.7, 2.8, 2.12, 2.13, 2.16, 2.21). Fourth, we made videotape RECORDING PROCEDURES 87 recordings of each room recorded. Videotape has proven extremely useful in mapping of underwater archaeological sites, and it seemed likely to be useful in cavates as well. After the completion of standard recording in each group, two people (one handling camera and color, and one serving as coach and carrying props) proceeded back through each room filming each feature inside and out and attempting to point out features in each. The videotapes form a complete, economical, readily archived, and easily referenced record of the current condition of the cavates included in this preliminary project. Videotape has other advantages, as well. It allows detailed mapping of plaster, features, and rock art if such drawings become necessary, without the great expenditure of field time required to draw such maps on site. Since the completion of the cavate fieldwork, videotape technology has advanced to allow direct digitizing for computer mapping and recording of features like cavates, although this process is expensive. It is also possible with the proper equipment to produce slides from videotape (C. Schaafsma, personal communication, 1988). Videotape redefines the position of traditional still photography. If the video recording is of good quality, the videotape will hold more information than still photographs. But photographs still have their uses. In taking photographs for this project we emphasized their use for illustrating reports with examples of features and giving an idea of setting, and the Crowders made extensive use of them for recording rock art. As Bill Crowder's work demonstrates, photographs can be used to bring up detail not visible in videos and can be of far higher quality. Video technology is evolving rapidly and has advanced since 1986, but still photography retains its important place in recording, especially given the levels of expertise and types of equipment available to most archaeological projects. Photographs are more accessible than videotapes, since they do not require special equipment for viewing; however, they are also an archival headache. Since we were fairly satisfied with the results of videotaping, we took fewer still photographs than we would have done otherwise. Assessment of "High Tech" Techniques During the cavate fieldwork we used two tools that could, in 1986, be considered high- tech: direct entry of data into a laptop computer, and use of a video camera for recording features. (As noted above, the possibility of using remote sensing for producing elevations was eliminated before we even went into the field.) Both laptops and video cameras have great potential for fieldwork, and this is an evaluation of their usefulness as applied in Bandelier in 1986. Direct Data Entry Computer technology and capability are advancing even faster than video recording, and many of our experiences in 1986 could be avoided in 1994. Only the portions of what we learned that remain relevant are presented here. Data entry in the field has some well-known advantages: it circumvents a lengthy phase of data entry after leaving the field (time was projected to be short at the end of the cavate project field season and key-punching personnel scarce); it reduces keystroke errors because there are fewer generations of input; it creates legible and easily searched records; and it makes summary information and preliminary analysis available in the field. If the field workers are competent with the hardware and software, data recording can be faster than conventional methods. Our experience with field data entry was less than a resounding success. File transfers to the mainframe computer were far more difficult than we had been led to expect, and field conditions were harsher than the computer could easily withstand. File transfer problems result from the use of several unrelated software 88 CAVATE STRUCTURES packages (WordPerfect, CEO Write, Oracle) and staff inexperience in file translation. Use of compatible software and participation in field form creation by someone familiar with the program to be used in analysis will prevent these problems. The computer we had was a fancy, expensive piece of equipment--for 1986. It was an 8088-based machine with a single floppy drive, no hard drive, and 512 K of RAM. On the whole it worked well, although we had to be extremely cautious with it. The dusty and gritty nature of cavates would never have been so evident to us if we had not had the computer with us: the keyboard collected grit at an alarming rate, and in spite of careful cleaning after every day in the field, it was always noticeable. The single disk drive necessitated continual insertion and removal of disks, unavoidably introducing more dirt into the machine. Disks went bad at an alarming rate. Even today's smaller, far more capable laptop computers will require means of dust and grit protection if they are to survive and be useful in archaeological fieldwork. The actual process of data entry went fairly well. There was often considerable difficulty in reading the screen, but an adequate angle could usually be found. Speed of entry was about comparable to pencil and paper form; if anything, I found a tendency to take more complete verbal notes. We found one drawback, at least with the entry format we used: it was more difficult to check an earlier part of the form for previous entries and measurements than it was with a paper form. Videotape Recording The advantages of video recording are listed above. Overall, we were pleased with the quality of the picture. The cameras did a remarkable job of recording in the low light of many of the cavates. They also seemed to be less susceptible to the ubiquitous dust and grit than the computer, though that was again a concern. Of course, not every last detail of every wall is visible and well recorded on our tapes. It would probably be preferable to make the videotaping process part of the recording procedure for each room, so that all details could be remarked. This would considerably increase the recording time, however, and would also increase the risk of damage to the equipment, as it would have to be in the field every day instead of on selected videotaping days. As recording becomes more complex, there is a greater possibility that something will go wrong, and that is what happened here. The microphone of the first camera stopped working partway through the season, leaving large stretches of tape without sound. We remedied this by dubbing in the sound later, but the detail of the commentary undoubtedly suffered. This sort of event can considerably increase the time involved. On the whole, however, the video recording offers the advantages of completeness of coverage available in no other way. Archiving videotapes presents several problems. As of 1994 the NPS still did not have a systematic method for storing and retrieving tapes, although they are used by several units. More troubling, the technology is still new enough that no one knows how long images survive, although there is no doubt that they do degrade and that their storage life is probably quite short (less than 10 years?). To be useful as archival material, tapes of mid-1980s vintage must either be rerecorded onto longer lasting metal-based tape or, ideally, digitized. Digitization does ensure a virtually permanent, archival-quality record. Work Schedule and Field Time Spent The approximate distribution of time expenditure for this project is shown in Table 3.2. Adjustments have been made for the varying hourly schedules among the people who worked on the project, but all days have been rounded down to standard work days (that is, some "overtime" does not show). It is quite RECORDING PROCEDURES 89 Table 3.2. Time Spent and Forms Completed by Cavate Group, 1986. Frijoles Group M LA 50972 July 8-part of July 17 field work man-hours: 236 cavate records: 46 noncavate records: 19 total rooms: 140 % recorded: 46.4 personnel: B. J. Mills and B. P. Panowski, 7Vfe days H. W. Toll, 6Vi days; B. Fuller, 3 (recording) Bill and June Crowder, 3 days each (photos & rock art) Group A LA 50973 Group I LA 50974 Group F LA 50975 part of July 17-part of July 23, August 14-15 fieldwork man hours: 170 cavate records: 25 noncavate records: 50 (48 rooms) total rooms: 130 % recorded: 56.2 personnel: BJM, 4 days; BPP, 3; HWT, 5 BF, 3 days (recording) JC, 3 days; BC, 2 (photos and rock art) E. R. Bayer, 2 days (recording) part of July 23-part of July 25 fieldwork man hours: 106 cavate records: 21 noncavate records: 16 total rooms: 38 % recorded: 97.4 personnel: ERB, BJM, HWT, 2Vi days; BPP, 2 BF, 1 day (recording) Crowders, 1 day each (photos and rock art) part of July 25-July 30 fieldwork man hours: 126 cavate records: 15 noncavate records: 45 total rooms: 106 % recorded: 56.6 personnel: ERB, BJM, BPP, HWT, 3 days BF, 1 day (recording) Crowders, 1 day each (photos and rock art) 90 CAVATE STRUCTURES Table 3.2. (continued) Tsankawi LA 50976 late July 30-most of August 13 field work man hours: 268 cavate records: 63 noncavate records: 56 (54 rooms) total rooms: 117 % recorded: 100.0 personnel: BJM, BPP, HWT, 8 days ERB, 1V4 days BF, 3 days (recording) Crowders, Vh days each (photos and rock art) clear that the groups that have more cavates than noncavates take considerably more time to record. The break-in period at Group M is also evident, though some of that extra time resulted from a more detailed mapping, and some from revisitation of rooms to record variables added after the rooms had been recorded. Data Manipulation After we returned from the field, Bruce Panowski coordinated a group of five people (including himself) who entered first the paper forms and then the personal computer records into the version of the Oracle database form that he had devised for the mainframe. None of those who entered data worked full time, and the process spanned a month (from August 18 to September 18, 1986). During this time and later considerable further effort was invested by Toll and especially Panowski in doing initial tabulations and locating various errors and omissions. After entry and checking, the files were transferred among several systems for further processing and for analysis. These included the University of New Mexico and NPS mainframe computers and personal computers. This section discusses various ways in which the data were used. Appendix 2 contains information on actual data set transformations and storage. Data Set Modifications and Formats We made some modifications to the data set during data correction and analysis. We outline these changes here for future users of the data and as part of the description of the data. Laboratory of Anthropology Numbers As discussed in the description of cavate groups, LA numbers that had been assigned to another project were given to the field crew. Replacement "official" numbers are given in chapter 2. In the computer data sets, the numbers used in the field and on all records are retained under the name FIELDLA, while the numbers on record with the Laboratory of Anthropology Survey Room ARMS File, are under the variable NEWLANO. Locations Universal Transverse Mercator (UTM) grid locations were assigned to all the study groups. It is not possible to locate every cavate RECORDING PROCEDURES 91 on 1:24,000 USGS maps, or to realistically specify UTM coordinates at an accuracy greater than 10 m. UTMs, therefore, were assigned to upper and lower halves of Groups A, F, and M; Group I was assigned a single UTM value; and the Tsankawi (LA 50976) group was divided into northern, middle, and southern groups for UTM assignment. Assignment of Section, Township, and Range location for the Tsankawi group was straightforward, since that entire location falls within a single quarter-quarter section. Township 18N, Range 6E, however, has never been platted. Moreover, the platted sections on the northern part of the Frijoles Quadrangle are somewhat irregular. Projections were made from the existing platted areas on the Frijoles Quad to the cavate locations, and quarter-quarter-section locations have been provided in the computer files for the groups studied. The section locations and probably the quarter sections are likely to be correct, but the locations must be recognized as projections and estimates only. Calculation of Volumes and Areas The measurements we took were designed to allow us to calculate volumes of chambers and areas of plane surfaces. Assumptions and compromises are necessary in calculating and using such values both because of the nongeometric shapes of cavates and because they are often only partially present. The procedures and formulas we used to obtain volume and area measurements are given below. An annotated version of the SAS program (SAS Institute 1985) is presented in appendix 2. Plane Shapes Formulas for all the plane shapes except the two varieties of ovals are well known and straightforward: Rectangle: Area equals width times length. Same formula used for bowed rectangle. Oval (and "natural oval"): As a compromise approximation, the area of oval features was calculated as the mean of the axes divided by two (giving the equivalent of a circle's radius), which was then squared and multiplied by pi. Trapezoid: A trapezoid may be thought of as a rectangle with either one or two triangles appended. The heights of the triangles are equal to the rectangle. The formula for finding the area is half the height times the sum of the bases (see, e.g., Kern and Bland 1934:vii). Circle: The area is the square of the radius times pi. Since we measured diameters rather than radii, the measurements were halved to find the radius. Triangle: The area is the product of the base and one-half the height. Areas were not calculated for linear or irregular shapes. The areas found using the above formulas were all multiplied by the geometric fraction to correct for partial and irregular forms. This correction is especially important in "round" shapes, which were often semicircular. We took all measurements in meters (generally measuring to the nearest centimeter) and rounded the areas only to four decimal places, since this allows expression of a square centimeter. Zeros were used for missing values when the data were originally entered. All areas of zero have been changed to missing values, as have all zero values for measurements, except for height above floor where zero is a valid measurement. Solid Shapes Solids are more complex to measure and thus involve more assumptions. Nonetheless, the resulting volume estimates are unquestionably closer to the actual volume than they would be if based on purely rectilinear 92 CAVATE STRUCTURES measurements. The formulae used are taken from Kern and Bland (1934:14-34). Rectangular solid: The volume is the product of the length and width of a base and the height. Cylinder: The volume is the area of the base times the area of the height. This shape was used most often for holes in the walls, so that the "base" is the aperture of the hole. Because the area of the base requires a radius, the aperture diameter had to be halved. Hemisphere: Measuring the volume of a hemisphere (or a spherical segment of one base) requires knowing the diameter of the whole sphere. We assumed that the diameter taken in the field is the diameter of the whole sphere, which is unlikely to have been the case. For the height cf the segment the depth measurement was used, which is the best approximation of the segment height. The formula for the volume is one-third pi times height squared times the quantity three times the sphere's radius minus the segment's height. In contrast to our practice with other area and volume measurements, we did not use the GF on hemispheres, since the two measurements should approximate the fraction of a sphere present. This method of estimating hemispherical volume may give values somewhat greater than the actual values because of inflation of the radius value. Using this approach, we found that some volumes result in negative values, which shows the approximate nature of this calculation and indicates that hemispherical volumes should be treated with extra caution. We set volume values to missing for features whose measurements resulted in negative values (about 12 in all). Truncated cone: This shape was used mostly for wall holes. The upper diameter was often difficult to measure; the cones are assumed to be right cones (that is, the altitude from the apex to the center of the base is a right angle). The volume formula is one-third pi times the height times the sum of the square of the basal radii and the product of the radii. Truncated pyramid: Several assumptions are active here: that the figure is the frustum of a right pyramid, and that the bases are rectilinear (and thus proportional). We measured two sides of the lower base of the figure, the height of the figure, and the length of the upper base. The volume formula requires that the area of both bases be known, which in turn requires that both sides of the upper base be known. To arrive at the width of the upper base, the lower base width was divided by the lower length, and the result was multiplied by the upper base length, generating the upper base width. The volume is then found by multiplying the sum of the areas of the two bases and the mean proportional of the two bases (the square root of the product of the two areas) by the height divided by three. Cone: Like the truncated cone, cones are assumed to be right. The volume is calculated as one-third pi times the basal radius squared times the height (basal radius found by halving the diameter taken in the field). Sphere: The volume formula is pi times the cube of the diameter, all divided by six. Pyramid: Pyramids were measured as four-sided right figures. The volume of such figures is calculated as one-third the area of the base times the altitude. In some cases we were best able to describe a feature as a plane figure with depth (usually triangular or oval). In these cases the area of the plane figure was multiplied by the depth to give a feature volume. For hemispheres and the "plane figures with depth," the GF was applied only to the area value. For all other shapes the volume values were also multiplied by the GF to correct for nonconformance to the geometric shape used. RECORDING PROCEDURES 93 After the calculation of volumes, rectangles with depth were converted to rectangular solids for purposes of mean volume calculations. "Plane figures with depth" presumably do not conform easily to the selection of geometric solids because of such things as irregular depth and rounded corners, and this treatment of volume may somewhat inflate the volume derived. The change to solid figures was done only for calculation, not as a permanent change to the data sets. Volume values were rounded to seven decimal places (of a cubic meter), which allows expression to the nearest 10 cc. Again, field measurements are generally to the nearest centimeter, but given the difficulties and imprecisions in measuring volumes, the nearest 10 cc is the closest realistic estimate, and in many cases even that level is probably false precision. Photographic Entries Four sets of photographs exist from the 1986 cavate recording: 4x5 frontal photographs of individual cavates (or groups of a few closely spaced rooms) taken by Bill Crowder, 35 mm black-and-white negatives of rock art taken by Bill and June Crowder, color slides of rock art taken by the Crowders, and 35 mm black-and- white shots of features and cavates taken by the archaeologists both during the field season and in April 1987. Some of the photographs in each of these groups are of activities or are overviews of large portions of cavate groups. The more specific photographs have been entered on the appropriate data lines, using the numbering conventions listed in appendix 2. Computer listings of photographs sorted by feature type and photograph and, for base information, by group and cavate number are on file with the National Park Service. Although not every feature or cavate that shows in every photograph has been entered, the coverage should be good enough to permit users to locate photographs of examples and specific instances. Users can also find specific instances by referring either to general features, such as the wall on which a feature is located, or to nearby cavates as indexed by the frontal drawings. Location and Processing of Computer Data Sets The field forms for the cavate data were entered by various NPS personnel using the Oracle data base at the regional office in Santa Fe. After several attempts, the data were transported via tape to the University of New Mexico mainframe IBM system. Once there, Statistical Analysis System (SAS) formats (SAS Institute 1985) were generated, and the data sets were checked line by line and corrected as necessary. Much of the data analysis for this study was done in SAS on the university system, though many analyses were conducted on a personal computer using the Number Cruncher Statistical System (NCSS) and SAS-PC. During the project, NPS acquired SAS for its Santa Fe system, so we returned the data sets in SAS format for use in Santa Fe and did much of our analysis on that system. Appendix 2 lists the data sets and the various outputs saved and now in the care of NPS. Group Attributes As discussed above, we recorded two sets of information for each cavate and noncavate feature: first, general or "base" information, followed by recording of specific features. Examination of the base information shows the distribution of construction, condition, and other attributes in the various groups. The distribution of room types among groups as reflected by form types is quite different. Groups A and F both show strong predilection for rooms built against the cliff, while the other three groups show more frequent use of rooms at least partially excavated into the cliff (Tables 3.1, 3.3). 94 CAVATE STRUCTURES Table 3.3. Room Type and Mode of Recording by Site. Recording Format Group A Group F Group I Group M Tsankawi Total Cavate 25 15 21 47 66 174 Noncavate 48 45 16 19 52 180 Total 73 60 37 66 118 354 Construction and Use Based on our observations, we can only speculate about how the cavates were actually made, and about whether the builders converted the tuff removed into building material. In Turkey people excavate rooms with picks, splitting off blocks for use in construction elsewhere; one man expected to spend a year working part-time by himself to complete a dwelling of five rooms and a cellar (Blair 1970:142-145). The most common evidence of construction in our sample-observable in more than half of the cavates-is the presence of grooves, which are especially visible in the ceilings (Table 3.4, Figure 3.2). These are modally 2-3 cm wide, around 1 cm deep, and perhaps 30-40 cm long; they look very like the digging stick marks sometimes seen in the hard earth walls of pits revealed by excavation. They are at numerous different angles, though they tend to have some patterning. They look more like a means of final shaping of the chamber than a means of removal of large chunks. At least partly because of different tuff types, digging stick marks are less common at Tsankawi, but at both Tsankawi and Frijoles it was common practice to smooth the walls of a cavate as high as the plaster was to go (often around 1 m) and leave the upper wall and the ceiling rough. Remodeling and housekeeping seem not to have included removing the smoked layer of tuff from the ceiling; replastering the walls seems to have sufficed. In contrast, recent Turkish occupants of cavate rooms cut into tuff similar in appearance to that in Frijoles, remove smoke-blackened rock from the ceilings with hammers, and replaster on an annual basis (Riboud 1958:138-139). Some form of grinding or polishing similar to that used to smooth lower walls probably also finished the shaped openings and doors, but we noted no evidence other than smooth, regular surfaces. As far as we could determine, 95 percent of the cavates we examined resulted more from artificial excavation of the cliff than from use of natural pockets (Table 3.5). The use of masonry was the least at Tsankawi (Table 3.6). Among the Frijoles groups, there was less use of masonry in upper Group M than in the other groups, though M-60 contains one of the most substantial remaining walls recorded. Masonry remains in about 20 percent of the cavate features of Groups A, F, and I; if the upper part of Group F had also been recorded, Group F would have been likely to stand out as containing more masonry than the other groups, because of the number of masonry features and dividing walls present there. The greatest use of multilevel rooms occurs at Group M, where a large central cluster of cavate rooms is located above a substantial house mound. Group F has a similar high central cluster, but it forms a smaller percentage of the rooms recorded. There are also many multilevel rooms at Group A, though upper Group A contained no areas with four levels. The part of A we did not record contains further Table 3.4. Evidence for Construction. RECORDING PROCEDURES 95 Type of Evidence Group A Group F Group I Group M Tsankawi Total Cavates Evidence lacking 2: Digging stick marks 3: Shaped opening 4:2 + 3 5: Shaped corners 6:2 + 5 7: 2 + 3 + 5 8: Floor leveling 4 3 2 4 7 4 1 1 3 6 3 1 1 1 3 1 6 5 5 15 24 47 4 13 29 2 5 7 6 8 18 13 3 24 6 6 26 1 7 22 1 Total 25 15 21 47 66 174 Total digging stick marks 16 13 15 17 30 91 Evidence lacking 43 36 14 18 51 162 Digging stick marks 2 9 1 4 16 Shaped opening 1 1 Shaped corners 4 1 6 Total 49 45 16 19 55 185 Combined total 74 60 37 66 121 359 96 CAVATE STRUCTURES Table 3.5. Estimate of Excavated versus Natural Space (Cavates Only). Portion Excavated Group A Group F Group I Group M Tsankawi Total Completely excavated 12 9 13 20 41 95 > Half excavated 10 6 7 26 22 71 > Half natural 2 1 2 5 Completely natural 1 1 1 3 Total 25 15 21 47 66 174 multilevel areas, and it is likely that of the five groups, Group A has the highest frequency of stacked rooms. Although the percentage of upper-level rooms is somewhat less at Group I, it still far outstrips that at Tsankawi (Table 3.7). The stepped nature of the tuff outcrops at Tsankawi made it more difficult for us to assign levels, and in most areas it precluded construction of more than one or two levels. Tsankawi has a markedly lower frequency of rooms built at least partly above other rooms. Fill The paucity of fill in a great majority of the cavates is evident in Tables 3.8 and 3.9: one-fifth were found to have no fill at all. Most of those containing fill in the Frijoles groups contain a mixture of disintegrated tuff and a very fine, floury dust, presumably mostly aeolian in origin. There is frequently some mixture of organic materials, such as grass and leaves, which often look as though they may also have blown in or may have been brought in by rodents. Animal dung was sometimes present in Frijoles, but never in quantity. In the lowermost tier of rooms at Tsankawi, several rooms had a thick, hard layer of dung. This tier of structures also had many of the most deeply filled rooms because of its location at the base of a low cliff. The few deeply filled rooms in Frijoles are next to the rockfall in Group I or have openings below the level of the exterior ground surface so that fill runs into the structure. A few chambers of which only the top few centimeters were visible were recorded on noncavate forms at Tsankawi and Group M, so that the deeply filled chamber count in Table 3.9 is somewhat depressed. The consistency with which cavate rooms with raised entries have very little fill leads me to suspect that they may never have had much. Alternatively, earlier visitors may have been very thorough in their "investigations" of these features. Small fill amounts make it difficult to detect whether there has been disturbance of the fill, because every entry into a room with a few centimeters of fill is a substantial disturbance. Our forms did not have a specific entry for noting the presence of pothunting activity, and generally we saw little evidence of it. Two of the deeply filled, red tuff rooms at Tsankawi do have major "potholes" in them (one exposing the bones of an infant). One or two Group F rooms have dirt and roots clinging to the walls, suggesting that fill has been removed, and another has had its floor cut through with metal tools. The few rooms we examined that were especially difficult to get into did have more cultural material in them. Four smaller, higher rooms had quantities of chunks of tuff in them RECORDING PROCEDURES 97 Table 3.6. Masonry Presence and Type. Masonry Type Group A Group F Group I Group M Tsankawi Total Cavates Masonry absent 20 12 17 43 63 155 Large blocks, simple 4 4 4 1 13 Small rock, much mortar 1 2 3 Coursing absent 1 2 3 Total 25 15 21 47 66 174 Percent with masonry 20.0 20.0 19.0 8.5 4.5 10.9 Non-cavates Masonry absent 47 44 15 19 56 181 Large blocks, simple 1 1 2 Small rock, much mortar 1 1 Coursing absent 1 1 Total 49 45 16 19 56 185 Combined total Percent of group with masonry 74 60 9.5 6.7 37 66 13.5 6.1 122 2.5 359 6.4 along with relatively hi gh artifact frequencies Natural Features (rooms A-4, A-57, F-56, and M-15). The likeliest explanations for the unusual rock fill are either the collapse of portions of walls or "basketball" practice after abandonment, probably during the last hundred years. The greater variability in tuff found at Group A and Tsankawi is shown in Table 3.10. These assessments of tuff are somewhat impressionistic. Still, the tuff at Group M 98 CAVATE STRUCTURES Table 3.7. Room Level, Cavates and Noncavates Combined. Level Group A Group F Group I Group M Tsankawi Total First Second Third Fourth Total 38 34 27 29 107 235 29 20 9 22 11 91 6 4 1 12 23 2 2 4 73 60 37 65 118 353 Table 3.8. Fill Type, Cavates Only. Fill Type Group A Group F Group I Group M Tsankawi Table 3.9. Fill Depth, Cavates Only. Total Clear floors 7 2 2 16 8 35 Disintegrated tuff 7 3 4 18 5 37 Dung and tuff 1 1 2 Aeolian/alluvial 7 3 10 7 20 47 High organic content 1 1 2 Tuff, dung, organic 3 3 3 11 20 Aeolian/alluvial, 4 2 21 27 tuff, organic Total 25 14 20 45 66 170 Fill Depth Group A Group F Group I Group M Tsankawi Total < 10 cm 22 6 11 38 28 105 10-25 cm 1 4 6 4 17 32 25.1-50 cm 2 12 14 > 50 cm 1 1 6 8 Total 23 11 20 42 63 159 RECORDING PROCEDURES 99 Table 3.10. Tuff Characteristics, Cavates Only. Tuff Type Group A Group F Group I Group M Tsankawi Total Ordinary white 12 15 19 23 18 87 Softer than usual 1 1 Stratified, variable 9 1 1 11 Large fibrous chunks 4 1 23 1 29 Highly porous 24 24 Soft red 22 22 Total 25 15 21 47 66 174 contains abundant soft, fibrous chunks of tephra, which probably leads to an increased rate of deterioration. The most common nonhuman users of the cavates are insects (Table 3.11). Wasp nests are probably the most common evidence of insects, and larval casings are also found. A bee swarm was noted at Group A, but it was happily in a natural hole well above where we were working. A single room in Group M was almost completely plastered with some form of larval casing that we did not see in any other place. We saw no rodent nests, but tracks and scat were commonplace. Pack rats seemed to be frequenting only two features at Tsankawi and one at Group A. The absence of bats was surprising, but the shallowness of many cavates, and the rounded, perchless nature of most of the ceilings may account for the bats seeking the higher cracks rather than the cavates. Birds seem to be primarily casual, though not infrequent, visitors to the cavates, except for the vultures that use the low, dark, dank, mostly filled rooms at one end of Group I. We watched with bemused indignation as a rock wren brazenly removed our room tags from Group M very shortly after we had put them in place. Cattle and sheep have made the greatest impact on cavates, but fortunately their use is fairly limited. Livestock (and perhaps wild burros or even deer) has entered many cavates, especially at Tsankawi. Where livestock has been abundant, a thick layer of dung is present, which may well protect room fill; heavy use by large animals has apparently worn grooves 20-30 cm wide ("incised dados") into cavate walls just above the top of the fill. We recorded four instances of these grooves at Tsankawi and one at Group M. "Combination" use, as noted on our forms, generally meant use by several of the common categories of nonhuman users, especially insects, rodents, and ungulates. We noticed no visitation or use by carnivores, though we found some tracks in Group I that we could not identify. Perhaps incorrectly, we did not include pothunters as nonhuman users. Cavate Condition On the whole, extremely fragile cavates are likely to have fallen apart long ago, becoming, for our purposes, noncavates or nonexistent. In comparing the estimated stability of cavates and noncavates, we see that relatively more cavates seemed to be stable and more that seemed to be facing major problems, while a higher percentage of the noncavates fall in the middle ranges (Table 3.12). Noncavates tend to be more exposed and either have fewer features 100 CAVATE STRUCTURES Table 3.11. Nonhuman Use ofCavates. Animal Use Group A Group F Group I Group M Tsankawi Total None visible 10 3 4 21 30 68 Ungulates 2 2 10 14 Pack rats 1 2 3 Insects 7 4 6 15 9 41 Combinations 5 3 3 6 8 25 Raptors and vultures 4 4 Other birds 1 1 5 1 8 Other rodents 1 2 2 6 11 Total 25 15 21 47 66 174 Table 3.12. Overall Stability of Ca votes and Noncavates. Degree of Threat Group A Group F Group I Group M Tsankawi Total Cavates Apparently stable 4 1 5 8 13 31 Lesser threat 8 9 6 9 35 67 Greater threat 7 4 6 17 13 47 Major problem 5 1 3 7 4 20 Total 24 15 20 41 65 165 Noncavates Apparently stable 3 1 1 7 12 Lesser threat 13 22 7 1 41 84 Greater threat 25 22 8 10 7 72 Major problem 7 1 7 1 16 Total 48 45 16 19 56 184 Combined total 72 60 36 60 121 349 RECORDING PROCEDURES 101 at risk (as with back walls) or have already experienced most of what loss was possible. The majority of both types falls into the middle ranges. Appendix 3 lists all cavates and noncavates by their apparent stability and includes the notes for each room considered to have an imminent possibility of collapse. In spite of its greater visitation, Tsankawi seemed to have the highest percentage of stable and "lesser threat" cavates, while Group M seemed to have the greatest deterioration problem. Of course, Group M has a large number of partial rooms, which accounts in part for its lower stability (see Table 3.12). Tsankawi and Group F have the highest relative frequencies of complete or mostly complete cavate chambers, with the absolute number being by far the greatest at Tsankawi. The amount of graffiti in noncavates is really quite small (Table 3.13), but complete cavates attract far more attention than do back walls. Of the areas in which we worked, Group A is the most seriously affected by graffiti, which is reflected in both the cavate and the noncavate data. Signs of serious wear are most evident at Tsankawi. The exposed nature of noncavates is reiterated in the assessment of natural impacts presented in Table 3.14, where severe weathering combined with severe cliff deterioration is the most common category. The presence of natural and human damage was assessed on two levels. First, we estimated the overall stability of each cavate and noncavate; for each cavate structural feature (wall, floor, ceiling), we recorded human and natural damage, and we used the same codes for an overall assessment of each noncavate. Tables 3.13 and 3.14 summarize this information by group and by chamber location. Of the groups we studied in 1986, Tsankawi has the most complex situation regarding the condition of the rooms. The chambers at Tsankawi are visited daily, sometimes by rather large numbers of people, while access to Groups A, F, I, and M is limited to varying degrees. In the two weeks when we worked at LA 50976 on a regular basis, one substantial portion of shaped tuff was broken off in the area of TS-57 and TS-60, campers were present in TS-57, and numerous rocks were thrown and dislodged in the midst of the room concentration (some too close for our comfort). In rooms where we swept away shallow fill, a very clear wear pattern was present: the plaster floors are missing completely around the door and into the middle of the room, while aboriginal floor, sometimes in remarkably good condition, is present around the edges. Even more remarkable is the presence of intact loom anchors in some of the rooms in spite of their being covered by only a few centimeters of fine fill. The loom anchors seem to be on the margin of the worn area, probably because the looms were placed at the edge of the prehistoric traffic flow, as well. On days of high traffic and bad behavior, complete closure of the site seemed the best course. On the other hand, many of the visitors to this area genuinely appreciated the opportunity to see these structures up close. Powers favors importing fill for shallowly filled rooms, but while this would help protect remaining floor features, it would be difficult to do and would have its own impacts. Finding fill that is not part of an archaeological deposit will require going to the canyon bottom, and transporting it will be laborious and potentially damaging to the structures. Closure of especially fragile areas, such as the TS-54-TS-66 area, has some appeal, but closure would only keep out law-abiding people. The best alternative may be to carefully, systematically, and completely clear ("excavate," if you will) shallowly filled rooms, collect samples, fully record the floors, and then refill them to the extent possible. This should not be a hurried, under recorded, and under reported salvage job. Rather, it should receive the full excavation status warranted by its location in a national monument. Such a project would not be cheap, 102 CAVATE STRUCTURES Table 3.13. Human Damage by Group and Chamber Location. A. Human Damage to Cavate Walls, Floors, Ceilings Type of Damage Group A Group F Group I Group M Tsankawi Total No damage 74 48 74 144 197 537 Minor graffiti 13 15 12 17 19 76 Major graffiti 19 2 6 4 31 Graffiti & other 1 1 1 3 Obvious wear 6 4 2 2 17 31 Tourist blasting 2 2 Minor vandalism 4 1 3 3 11 Major vandalism 3 3 Total 117 70 94 167 246 694 B. Human Damage to Noncavates Type of Damage Group A Group F Group I Group M Tsankawi Total No damage 3 42 12 16 11 2 83 Minor graffiti 3 1 1 5 Major graffiti 1 1 Obvious wear 1 1 Tourist blasting 3 3 Minor vandalism 1 1 Major vandalism 1 1 Total 48 13 16 12 6 95 RECORDING PROCEDURES 103 Table 3.13. (continued) C. Human Damage by Chamber Location, Cavates Only h Right Back Left Exterior Type of Damage Wall Wall Wall Wall Floor Ceiling Total No damage 112 96 109 60 50 109 537 Minor graffiti 23 18 22 8 5 76 Major graffiti 9 13 5 3 1 31 Graffiti & other 1 1 1 3 Obvious wear 2 2 5 11 10 1 31 Tourist blasting 2 2 Minor vandalism 1 7 2 1 11 Major vandalism 1 1 1 3 Total 148 140 145 83 60 117 694 a In 91 cases there was no entry; probably no damage. b Cavate structural features only; 14 features with damage not recorded are not shown. but it would likely be no more expensive than some of the other alternatives. It is also a sensible plan in archaeological terms, given the current visitation regime. Another impact that is more severe at Tsankawi than at the other sites is the effect of livestock. The lower, red tuff rooms were heavily used by sheep and/or cattle; combined with the friable nature of this geological unit, the livestock use has ensured that few wall features remain in the majority of the red tuff rooms. On the whole these low rooms have mon fill than most cavates, which may have preserved features closer to the floors. The rate at which cavates deteriorate can only be discussed generally, given the lack of any precise baselines. Lister's photographs from the late 1930s allow us to make some comparisons (Figures 2.7, 2.8, 2.12, 2.13, 2.16, 2.21). They show clearly that the elements most at risk are masonry: all the major changes visible involved masonry collapse, in spite of the attempts to slow that process through stabilization. On the whole, it seems that there is very slow (on a human scale) erosion of natural features, punctuated by occasional, much larger but relatively infrequent events, such as the rockfall at Group I, the dislodged blocks at Group A and Tsankawi (A-72 and TS-22), and 104 CAVATE STRUCTURES Table 3.14. Natural Damage by Group and Chamber Location. Type of Damage A. Natural Damage to Cavate Walls, Floors, Ceilings Group A Group F Group I Group M Tsankawi Total No damage 19 10 17 28 37 HI Moderate erosion 21 16 7 31 35 110 Severe erosion 4 2 3 10 9 28 Moderate cliff 15 4 2 16 18 55 deterioration Severe cliff 7 5 3 18 9 42 deterioration Moderate erosion & 8 16 17 12 66 119 deterioration Severe erosion & 43 17 45 52 74 231 deterioration Total 117 70 94 167 248 696 B. Natural Damage to Noncavates, Including Nonrooms Type of Damage Group A Group F Group I Group M Tsankawi Total Absent 1 1 2 Moderate weathering 1 1 4 6 Severe weathering 2 5 2 6 8 23 Severe deterioration 2 1 3 Moderate weathering & deterioration 5 1 1 13 20 Severe weathering & deterioration 37 38 13 10 31 129 Total 48 45 16 18 56 183 RECORDING PROCEDURES 105 Table 3.14. (continued) C . Natural Damage by Chamber Location, Cavates Only Right Back Left Exterior Type of Damage Wall Wall Wall Wall Flo( No damage 25 34 22 8 3 Moderate erosion 25 29 27 10 4 Severe erosion 5 6 4 4 5 Moderate cliff 15 12 12 4 2 deterioration Severe cliff 7 6 7 6 7 deterioration Moderate erosion 26 27 22 17 7 & deterioration Severe erosion & 45 27 51 34 32 deterioration Ceiling Total 19 111 15 110 4 28 10 55 9 42 19 118 42 231 Total 148 141 145 83 60 118 695 fallen boulders at Tsankawi with evidence of former rooms. Robert Preucel, formerly of the PARP team, reports that between 1980 and 1985 three cavates collapsed in the Garcia Canyon area, presumably due to natural processes such as freezing and thawing (R. Preucel, personal communication, 1988). Human visitation compounds natural erosion and, worse, causes disturbance of deposits and features that are subject to little natural damage. The human threat is more focused and destructive to archaeological information than other types of damage. At least in theory, however, it is also more easily averted. Cavate and Noncavate Features: Definitions, Distributions, and Dimensions This chapter defines each feature type recorded in 1986, as well as presenting occurrence and metric data for each. This information forms the basis for the analyses, which use selected features and attributes to look for patterning within and among the various cavate groups. All measurements are in meters, though in the discussions we sometimes refer to centimeters or other smaller units. Volume measurements are given to five decimal places, which might appear to be a pretension to extraordinary precision. The field measurements were taken to the nearest centimeter, however, and 1 cc is 0.000001 m 3 , so that technically the data as presented drop one decimal point of precision. In interpreting all the measurements, and especially the volumes, one must remember that measuring cavate features usually requires fitting complex, curvilinear shapes into simplified geometric shapes. Thus, most of the measurements are only best approximations. In some cases, measurements only from features thought to be reasonably (more than 70-75 percent) complete, or feature heights only in chambers where fill is minimal (15-20 cm is the usual cutoff) are used. In other cases, combining similar shapes gives larger samples and better overall characterization of feature size. Wherever these screens and combinations have been applied, the tables are annotated accordingly. In examining distributions of features across cavate groups and locations within chambers, the varying numbers of observable cases from which the sample is drawn are obviously crucial to understanding what is normal and what is unusual. As discussed in chapter 2, different groups have different compositions, and this can be seen in Table 4.1. Although the gross number of features recorded can generally be ranked in descending order as Tsankawi, Group A, Group M, Group F, and Group I, the pattern certainly does not hold for all categories. The standard of comparison cannot always be number of cavates, partly because of the variability in cavates. Further, while some features can occur only in fairly complete cavates, other features are found in both cavates and noncavates, and still others, such as viga holes, are more likely to occur in noncavates than cavates. As discussed in chapter 3, the noncavate form allowed us to record groups of features and size ranges for those groups. The metric data presented here are only for individually recorded features. For each feature type containing examples recorded as groups, the distribution table includes an entry for "noncavate, grouped" showing the numbers of features so recorded at each cavate group. A total of 362 features were group- recorded, and 83 percent of these fall into three feature types: 234 viga holes (65 percent), 37 107 108 CAVATE STRUCTURES Table 4.1. Overall Occurrence ofCavates, Noncavates, and Features by Group. Group A Group F Group I Group M Tsankawi Total Total rooms recorded 73 60 37 65 118 353 Row percentage 20.7 17.0 10.5 18.4 33.4 Cavates recorded 25 15 21 46 65 172 Row percentage 14.5 8.7 12.2 26.7 37.8 Noncavates recorded 48 45 16 19 53 181 Row percentage 26.5 24.9 8.8 10.5 29.3 Total features 667 402 318 616 1394 3397 Row percentage 19.6 11.8 9.4 18.1 41.0 Total walls recorded 137 104 89 123 265 718 Row percentage 19.1 14.5 12.4 17.1 36.9 indeterminate holes (10 percent), and 29 hand- and-toe holds (8 percent). The 53 feature types recorded are discussed individually in the following pages (Table 4.2). They are grouped into the following categories: structural features, such as chambers, doors, and walls; floor features; wall features; and rock art. Except for feature types of which only a few examples were found, the occurrence across groups, the shape distributions, and appropriate metric data are tabulated for each type. Particularly where classifications were difficult or subjective-as in the case of holes in walls, for example- comparisons across types have also been made. This chapter includes some discussion of co- occurrence of functionally related feature types, and chapter 5 further explores feature associations. Structural Features Chamber (Code 1) This feature type was used to give the best approximation of the entire space enclosed by a cavate. The components enclosing the space (walls, floor, ceiling) were also recorded, but this feature type is best suited to describing the room as a whole. The shape-by-group frequencies show whole chambers that were recorded either as single shapes (most cases) or as combined shapes (Table 4.3). The combined shapes are shown as the base shape plus the top shape; one chamber at Tsankawi resulted from removing the wall between two truncated-pyramid chambers. The single cylindrical "chamber" is a partial cavate at Tsankawi, of which only the base was recordable. A clear trend toward rounded shapes is seen at Tsankawi, though the most common shape overall, the truncated pyramid, is also abundant there. Of the 182 cases shown, 166 were recorded as cavate features and 17 as noncavates (1 1 partial cavates, 4 filled cavates, and 2 chambers). The mean volume of chambers is also larger at Tsankawi, but variability in chamber size is also much greater there. Thirteen chambers at Tsankawi have volumes greater than 8 m 3 , which is larger than the largest for any FEATURES 109 Table 4.2. Occurrence of All Individually Recorded Features in Cavates and Noncavates, by Cavate Group (Number and Percentage). Feature Group A (n) (%) Group F (n) (%) Group I (n) (%) Group M (n) (%) Tsankawi (n) (%) Total (n) (%) Structural features Chamber 26 13.5 18 9.3 22 11.4 48 24.9 79 40.9 193 5.7 Exterior door 10 15.4 11 16.9 6 9.2 6 9.2 32 49.2 65 1.9 Exterior opening 16 22.2 5 6.9 2 2.8 10 13.9 39 54.2 72 2.1 Interior door 4 10.3 8 20.5 8 20.5 19 48.7 39 1.1 Passage 1 33.3 1 33.3 1 33.3 3 0.1 Natural wall 137 19.0 104 14.4 89 12.4 124 17.2 266 36.9 720 21.2 Masonry wall 1 16.7 3 50.0 2 33.3 6 0.2 Ceiling 20 16.4 12 9.8 14 11.5 28 23.0 48 39.3 122 3.6 Masonry and tuff wall 3 100.0 3 0.1 Chamber corner 33 22.0 18 12.0 22 14.7 14 9.3 63 42.0 150 4.4 Compass location point 14 100.0 14 0.4 Floor features Floor 18 27.7 3 4.6 5 7.7 31 47.7 8 12.3 65 1.9 Firepit 12 42.9 1 3.6 3 10.7 7 25.0 5 17.9 28 0.8 Floor burn 1 16.7 4 66.7 1 16.7 6 0.2 Subfloor pit 4 19.0 2 9.5 6 28.6 9 42.9 21 0.6 Floor depression 18 27.7 3 4.6 5 7.7 31 47.7 8 12.3 65 1.9 110 CAV ATE STRUCTURES Table 4.2. (continued) Feature Group A (n) (%) Group F (n) (%) Group I (n) (%) Group M (n) (%) Tsankawi (n) (%) Total (n) (%) Floor pit complex 1 25.0 3 75.0 4 0.1 Posthole 9 100.0 9 0.3 Floor ridge 3 27.3 7 63.6 1 9.1 11 0.3 Metate rest 1 1 4 6 16.7 16.7 66.7 0.2 Loom anchor 1 12 16 29 3.4 41.4 55.2 0.9 Step 1 33.3 2 66.7 3 0.1 Axe groove 5 100.0 5 0.1 Adobe collar 1 100.0 1 0.0 Deflector 1 33.3 2 66.7 3 0.3 Wall features Large floor-level niche 23 25 8 24 48 128 18.0 19.5 6.3 18.8 37.5 3.8 Wall niche 29 28 9 26 64 156 18.6 17.9 5.8 16.7 41.0 4.6 Slot 3 1 1 5 10 30.0 10.0 10.0 50.0 0.3 Viga hole 97 34.4 37 13.1 19 6.7 55 19.5 74 26.2 282 8.3 Possible latilla hole 21 5 6 16 48 43.8 10.4 12.5 33.3 1.4 Beam seat 39 7 9 45 76 176 22.2 4.0 5.1 25.6 43.2 5.2 Indeterminate hole 91 78 51 74 277 571 15.9 13.7 8.9 13.0 48.5 16.8 Possible loom support 1 3.6 3 10.7 3 10.7 21 75.0 28 0.8 Table 4.2. (continued) FEATURES 111 Feature Group A (n) (%) Group F (n) (%) Group I (n) (%) Group M (n) (%) Tsankawi (n) (%) Total (n) {%) Smokehole 12 20.7 11 19.0 8 13.8 2 3.4 25 43.1 58 1.7 Vent 5 15.6 8 25.0 4 12.5 3 9.4 12 37.5 32 0.9 Groove 2 25.0 1 12.5 1 12.5 4 50.0 8 0.2 Wall depression 24 35.3 12 17.6 6 8.8 7 10.3 19 27.9 68 2.0 Wall ledge 4 36.4 1 9.1 2 18.2 1 9.1 3 27.3 11 0.3 Vertical ceiling hole 1 1.8 55 98.2 56 1.6 Narrow wall incisions 1 4.0 24 96.0 25 0.7 Hand-and-toe hold 4 100.0 4 0.1 Incised dado 1 20.0 4 80.0 5 0.1 Cliff niche 5 100.0 5 0.1 Rock art Geometric petroglyph 2 12.5 4 25.0 4 25.0 6 37.5 16 0.5 Geometric pictograph 2 28.6 4 57.1 1 14.3 7 0.2 Zoomorphic petroglyph 1 3.1 4 12.5 6 18.8 9 28.1 12 37.5 32 0.9 Zoomorphic pictograph 2 40.0 3 60.0 5 0.1 Indeterminate petroglyph 3 9.7 4 12.9 4 12.9 8 25.8 12 38.7 31 0.9 Indeterminate pictograph 8 36.4 1 4.5 1 4.5 7 31.8 5 22.7 22 0.6 Handprint 3 100.0 3 0.1 112 CAVATE STRUCTURES Table 4.2. (continued) Feature Group A (n) (%) Group F (n) (%) Group I (n) (%) Group M (n) (%) Tsankawi (n) (%) Total (n) (%) Anthropomorphic petroglyph 1 7.1 2 14.3 4 28.6 7 50.0 14 0.4 Total 665 401 318 616 1395 3395 Percent 19.6 11.8 9.4 18.1 41.1 100.0 Note: Row percentages are shown for each feature, and percentages of the total count are shown in the "Total" column and row. Frijoles group (see Figures 4. 1-4. 3b). At least two very large chambers at Tsankawi (TS-36 and TS-66) resulted from the prehistoric removal of walls dividing what once were two separate chambers. Six of the rooms greater than 8 m 3 , and three of the five largest (larger than 13 m 3 ), are located in the top tuff stratum, which suggests that that location was favorable to large chamber construction (TS-25, TS-26, TS-15, TS-27, TS-24, TS-20 in ascending order of size; see Figures 2.19, 2.20). The vesicular nature of the tuff there may have provided natural beginnings for rooms. In addition, the tuff in the top stratum seems to be indurated, perhaps making the chambers more stable. Three of the large chambers are located in the soft, red, lowermost tuff layer (TS^4, TS-29, and TS-36), and all are less than 10 m 3 . TS-36, an expanded chamber, contains much fill and its volume may have been underestimated. The remaining four large chambers are in and adjacent to the concentration of structures in the middle levels of the group (TS-50, TS-66, TS-64, and TS-59). Both TS-50 and TS-66 have two doors, and TS- 59, TS-64, and TS-66 contain diverse features and rock art. The largest chamber (TS-59) is remarkable for the number of features, elaborate rock art, and unusual height (nearly 2 m). Among the rooms larger than 8 m 3 , there is a clear mode (and median) of 9-10 m 3 : 5 of the 13 fall into that size range, and 8 fall between 9 m and 15 m 3 (see Figure 4.1). Using a t-test to compare the 29 relatively complete Tsankawi chambers designated as habitation rooms with the 44 from Frijoles, the two groups are significantly different (means of 6.28 and 4.26 m 3 , t= -2.59, p =0.014). A similar comparison of rooms designated as storage rooms, however, finds no significant difference between groups of 19 and 23 rooms (means of Tsankawi [1.42 m 3 ] and Frijoles [1.34 m 3 ], t= -0.24, p=0.811). When the chambers greater than 8 m 3 are removed from the calculation, the Tsankawi mean becomes 2.7095 m 3 , which is smaller than the mean for all other groups but Group I. Perhaps large chambers were used for most functions, so that fewer activities were carried out in the remaining rooms, and they could be smaller. Alternatively, with access to larger rooms for habitation, the residents could devote more of the smaller rooms to storage. Assigning functions to chambers is a matter of speculation, but we based the FEATURES 113 Table 4.3. Chamber Occurrence and Dimensions. A. Chamber Occurrence by Shape and Group Shape Group A Group F Group I Group M Tsankawi Total Oval 2 2 4 with hemispherical top 1 1 Rectangular 3 1 4 8 8 24 with hemispherical top 1 1 Cylindrical Hemispherical 4 2 19 33 58 with rectangular top 1 1 with pyramidal top 1 1 Truncated cone 2 2 4 Truncated pyramid 17 11 11 17 21 77 with hemispherical top 2 2 with conical top 1 1 with 2nd truncated 1 1 pyramid Spherical 10 3 4 Irregular 2 2 Total 23 17 21 47 74 182 114 CAVATE STRUCTURES Table 4.3. (continued) B. Mean Chamber Volumes in Cubic Meters Grouping Coefficient Standard of Variation Mean Deviation Minimum Maximum (CV) By shape Rectangular 12 3.86317 2.34067 0.7358 9.1728 60.6 Hemispherical 37 3.72179 3.99955 0.3445 15.4956 107.5 Truncated Pyramid 69 4.24722 3.22900 0.3512 18.6326 76.0 By assigned function Habitation 73 5.04244 2.69408 0.5906 15.1495 53.4 Storage 42 1.37660 1.04885 0.3445 5.5545 76.2 Kiva 7 8.98803 6.74858 2.7211 18.6326 75.1 By group Group A 18 3.60704 2.45064 0.5576 7.5840 67.9 Group F 13 3.32241 1.87310 0.3512 6.2586 56.4 Group I 14 2.85180 1.35703 0.7358 5.5313 47.6 Group M 29 3.14225 2.19974 0.3445 7.3764 70.0 Tsankawi 57 4.78904 4.41563 0.4523 18.6326 91.0 By group and function Group A Habitation 12 4.92763 1.83442 2.2717 7.5840 37.3 Storage 5 0.69419 0.14392 0.5576 0.9307 20.7 Group F Habitation 10 4.11853 1.25625 2.2362 6.2586 30.5 Storage 3 0.66867 0.3512 1.2032 FEATURES 115 Table 4.3. (continued) Standard Coefficient of Variation Grouping n Mean Deviation Minimum Maximum (CV) Group I Habitation 7 3.18887 1.34459 1.5239 5.5313 42.2 Storage 5 1.96018 0.99595 0.7358 3.0620 50.8 Kiva 1 2.86365 Group M Habitation 15 4.31373 2.03163 0.5906 7.3764 47.1 Storage 10 1.55611 1.57855 0.3445 5.5545 101.4 Kiva 2 2.88122 2.7212 3.0413 Tsankawi Group Habitation 29 6.27539 3.40041 1.3315 15.1495 54.2 Storage 21 1.41990 0.82667 0.5422 3.2121 58.2 Kiva 4 13.05380 6.22100 4.8256 18.6326 47.7 Grouping n C. Mean Heights in Meters Mean Coefficient Standard of Variation Deviation Minimum Maximum (CV) By assigned function Habitation 70 1.58 0.441 0.95 3.93 28.0 Storage 38 1.11 0.242 0.75 2.05 21.7 Kiva 7 1.62 0.232 1.25 1.94 14.3 116 CAV ATE STRUCTURES Table 4.3. (continued) Grouping n Mean Standard Deviation Minimum Maximum Coefficient of Variation (CV) By group Group A 16 1.42 0.271 0.85 1.81 19.1 Group F 12 1.39 0.241 0.97 1.74 17.3 Group I 14 1.48 0.411 0.82 2.05 27.8 Group M 28 1.32 0.351 0.69 2.12 26.6 Tsankawi 51 1.47 0.539 0.75 3.93 36.6 Note: Chambers with indeterminate function assigned are not shown. Table includes only chambers judged to be more than 70% complete. Heights have been corrected by adding fill depths. Volumes have been given for all shapes, heights for simple regular shapes. assignments on criteria including numbers and types of features, plastering, and size (Figure 4.2). Although the "kiva" features are present in larger chambers at Tsankawi, those at the Frijoles groups are within the range of habitation room sizes recorded here (Figures 4.2, 4.3a). Rooms called storage rooms tend to be smaller, but there are some large examples as well (Figure 4.3b). Clearly, the distribution for "habitation" rooms shows a great deal of overlap with the other two categories. Most of the rooms to which no function was assigned are smaller (less than 3 m 2 ), though there are three larger ones as well (Figure 4.2). Hyland's (1986:101) results from Garcia Canyon show chamber volumes and heights considerably greater than those we found. He gives a mean volume of 9.62 m 3 with a range of 0.99-24.42 m 3 , and of 25 measured cases his data contain 9 chambers greater than 12 m 3 . This disparity is partly a result of his having calculated volumes from the product of the maximum dimensions of the cavates involved, without taking into account how cavate rooms constrict toward the ceiling or how rarely they approximate rectilinear solids. If we calculated the volume for the largest chamber at Tsankawi following Hyland's procedure, the result would be more than 25 m 3 , larger than any of the volumes Hyland found. Still, the apparent tendency to build larger chambers at both Tsankawi and Hyland's more northern cavates may have some cultural significance. Exterior Door (Code 2) This feature type was used for all intentionally shaped openings to what is currently outside. Many such doors probably opened into masonry rooms when the sites were in use, but we assigned them to this type anyway (Table 4.4). Measurable doors are clearly mostly rectangular with some oval and trapezoidal variants (Figure 4.25); of the "circular" doors, three were considered full circles and one a half circle. Not surprisingly, 60 of the 65 recorded doors are found in the exterior wall of their rooms, with the other 5 occurring in exterior corners (2), the right wall (1), the top of the chamber (1), and inside another feature (1). FEATURES 117 1 1 8 C AV ATE STRUCTURES FEATURES 119 R to CO c o 52 n E CO O (0 > c CO c o • MB s ■ ■■» -Q (0 .2 s Sjff CO CO 5 S o o c c ILILhh (O © 00 CO * cqN(qu>«tcocMT^q r^-^dddddddddd too\i 3Aoqe m6;8H S d o ■ *: Si FEATURES 161 Table 4.24. Co-occurrence of Floor-level and Wall Niches by Group. Rooms with: Group A Group F Group I Group M Tsankawi Total Floor-level niche only 10 11 6 8 19 54 Wall niche only 13 16 6 10 13 58 One of each type 2 2 2 6 2 14 Multiple of each 2 1 1 1 5 Floor-level > wall niches 2 3 2 3 10 Wall > floor level niches 2 1 2 9 14 Total rooms with niches 31 34 14 29 47 155 Total floor-level niches 23 25 8 24 48 128 Total wall niches 29 28 9 26 64 156 Maximum n floor-level 2 3 1 2 5 niches Maximum n wall niches 2 2 2 2 7 holes we recorded individually are associated with noncavate back walls, and almost as many viga holes were recorded in sets as individually. The combined counts demonstrate the importance of cliff-backed masonry rooms at these sites, particularly Groups A and F. The relatively high frequency of viga holes at Group A fits with the large number of "back wall" features there (Table 4.26). Why use vigas in a cavate? Viga holes are present in "cavates" largely because many cavates had masonry closing structures. Also, the counts by location show only individually recorded viga holes, not those recorded in groups on noncavate forms. Still, features that are to all appearance viga holes do occur in closed cavates, suggesting either some ceiling or rack structure. Some of these may have been upper loom supports lacking other recognition criteria (Figure 4.24). Many such features have smoke-blackened interiors, suggesting that they may have served only to hold some form of beam for part of the period of use of the cavate. Figure 4. 16 gives an indication of the variation in viga hole height, diameter, and depth. There is a concentration of holes 10-15 cm in diameter, 15-25 cm in depth, and 1.2-1.6 m high. We recorded height above floor only for cavates; if we had included noncavate back walls, the mean would be larger. The height above floor is fairly consistent; indeed, some height above the floor is necessary for a hole to be assigned to this feature type. Although a few cases fall in the 1.8-2.0 m range that most moderns would consider a comfortable ceiling height, 56 percent of the included cases fall between 1.2 and 1.6 m above the floor (Figure 4.16). Given the values below 1 m and the 162 CAVATE STRUCTURES Figure 4.15. Good example of a slot in cavate M-59 (scale =30 cm). This room also contains loom anchors and a large floor level niche, but no recognizable milling features. placement inside cavates, it is likely that if some of these viga-shaped holes did in fact contain vigas, the beams had functions other than ceiling support. Possible Latilla Hole (Code 12) This feature type is more ambiguous than the viga hole. Size and placement again contribute to its use. Occasionally series of small holes are found at about ceiling height, which may have been for latillas; these are the ideal case, but the category also includes less clear examples (Table 4.27). The feature category was originally inspired by a structure by the Tsankawi trail, where there is a series of 10 closely spaced, flattened, hemispherical (or "ovals with depth") holes that clearly supported a roof above an area in front of a cavate (probably Lister's C-119). After all the present measurements had been taken, I returned to this set of "ideal" latilla holes and measured them; they are 0.07-0.13 m wide, 0.064.11 m high, and 0.06-0.09 m deep. They are thus somewhat larger than the average for other features recorded in this category, though the depths are similar. Figure 4.17 shows that the features placed in this category are generally smaller and shallower and the measurements more dispersed than viga holes (compare Figure 4.16). Given the mean viga hole diameter and that for possible latillas, it appears that the wood being used in cavates tended to be fairly small. Comparing the height FEATURES 163 Table 4.25. Slot Occurrence and Dimensions. A. Occurrence by Type and Group Shape Group A Group F Group I Group M Tsankawi Total Oval 1 1 Linear 2 1 3 Rectangular 1 5 6 Total 3 1 1 5 10 Shape B. Chamber Location by Shape Right Wall Back Wall Left wall Total Rectangular Oval Linear 4 1 1 6 1 1 3 3 Total 10 C. Mean Dimensions of Oblong Shapes in Meters Standard Dimensions n Mean Deviation Minimum Maximum CV Width 10 0.06 0.027 0.02 0.11 42.4 Height 10 0.26 0.080 0.14 0.40 30.7 Depth 3 0.04 0.02 0.07 Height 5 0.14 0.14 -0.07 0.25 98.4 above floor Note: C, oblong includes rectangular, oval, trapezoidal, and triangular shapes. Only features judged to be more than 70 percent present are included. All available cases and shapes are included in height-above-floor measurements. 164 CAVATE STRUCTURES Table 4.26. Viga Hole Occurrence and Dimensions. A. Viga Hole Occurrence by Shape and Group Shape Group A Group F Group I Group M Tsankawi Total Rectangular Oval Cylindrical Hemispherical Truncated cone Conical 3 1 59 34 2 1 11 8 42 17 8 7 4 1 5 2 2 2 1 19 34 9 9 7 3 139 100 19 13 Total 97 37 19 54 74 281 Noncavate, grouped 67 87 48 26 234 B. Chamber Location by Shape Right Back Left Exterior Wall Back Shape Wall Wall Wall & Corners Corners Other Total Rectangular 1 4 2 7 Oval 1 1 1 3 Cylindrical 15 91 15 3 7 8 139 Hemispherical 7 77 5 4 5 2 100 Truncated 2 9 2 2 4 19 cone Conical 13 Total 26 186 27 10 21 11 281 / FEATURES 165 Table 4.26. (continued) Dimension C. Mean Dimensions by Grouped Shape in Meters Mean Standard Deviation Minimum Maximum CV Diameter 255 0.14 0.056 0.04 0.38 40.5 Depth 251 0.16 0.097 0.03 0.71 62.2 Volume 245 0.0026 0.0038 0.00003 0.0431 148.5 Height 121 1.39 0.266 0.85 1.94 19.2 above floor Note: Shapes included are cylindrical, hemispherical, conical, and truncated cone. Only features judged to be more than 70 percent complete are included. Heights are given only for chambers with 15 cm or less of fill. measurements with those for viga holes shows that this feature class is on average lower than viga holes, contrary to what would be expected if "latillas" were placed on top of vigas. In only a few cases did these two types of holes occur together in the "proper" sequence; of 150 cavates or noncavates where either feature type occurs, only 10 have both types. As with viga holes, we can only speculate about the actual function of apparently structural holes inside cavates. Holes for small beams do seem to have been used, and it seems more likely that they supported shelves than the prehistoric equivalent to false ceilings. Beam Seat (Code 30) This code differs from the viga hole primarily in that placement and pairing do not suggest roof support. Size still suggests that a substantial piece of wood would have been inserted (Figure 4.18a). The distinction is somewhat subjective, and some crossover between the two categories would be likely if we were to reclassify all cases. Other than a low frequency at Group F, there is little remarkable in the distributions of beam seats (Table 4.28). Compared to viga holes, features recorded as beam seats are much more often oval or rectangular, though the majority of openings are round; the mean diameter for beam seats is about two-thirds that for viga holes, and they are somewhat shallower on average (Figure 4.18b). Beam seats are more evenly distributed on chamber walls than are viga holes, though the majority are again found on back walls. Although a few possible beam seats were located near the chamber floor, around 70 percent are 0.8-1.4 m above the floor, with most of those in the 0.8-1.0 m range. Figure 4.18a shows a concentration of beam seats around 9 cm in diameter, 20 cm in depth, and 60-100 cm above the floor. Indeterminate Hole (Code 28) Cavates and noncavates have a great many small holes in their walls, and since almost none of the holes now contains identifiable remains, it is impossible to know what their function was, resulting in heavy use 166 CAVATE STRUCTURES ^J *> "3 1 Si FEATURES 169 170 CAVATE STRUCTURES Table 4.28. Beam Seat Occurrence and Dimensions. A. Beam Seat Occurrence by Shape and Group Shape Group A Group F Group I Group M Tsankawi Total Oval 12 3 15 Rectangular 7 2 4 19 32 Cylindrical 6 3 16 27 52 Hemispherical 12 1 2 11 13 39 Truncated cone 1 3 2 3 7 16 Truncated pyramid 3 1 4 Conical 1 3 4 4 12 Irregular 3 2 5 Total 39 44 76 175 Noncavate, grouped B. Chamber Location by Shape Right Back Left Exterior Wall Back Shape Wall Wall Wall & Corners Corners Ceiling Total Rectangular 12 6 7 5 1 1 32 Oval 3 6 3 1 2 15 Cylindrical 10 21 13 5 2 51 Hemispherical 5 19 3 7 2 36 Truncated 2 9 4 1 16 cone Truncated 1 1 1 3 pyramid Conical 3 6 1 2 12 Irregular 3 2 5 Total 39 68 33 18 5 7 170 FEATURES 171 Table 4.28. (continued) C. Mean Dimensions by Shape in Meters Standard Shape n Mean Deviation Minimum Maximum CV Round Diameter 114 0.09 0.035 0.06 0.90 72.4 Depth 114 0.14 0.140 0.01 1.15 101.6 Volume 113 0.0010 0.0023 0.00001 0.0208 116.2 Rectangular Width 29 0.13 0.051 0.04 0.24 38.4 Height 29 0.09 0.045 0.06 0.17 30.8 Volume 29 0.0010 0.0009 0.0001 0.0040 96.7 Height 137 1.01 0.034 0.01 1.78 33.2 above floor 8 Note: B, two features of unknown shape and three located within other features are not shown. "Heights above floor from chambers with less than 15 cm fill only. of this feature type during recording. This category certainly includes features that had a wide variety of functions, as well as some holes that are either natural or postoccupational. While they often occur in groups at similar heights above the floor, they are distinguished from possible latilla holes in arrangement and in- ability to support a small beam; there is again overlap, as measurements and multivariate analyses clearly show. Most are fairly small, but a few large, truly indeterminate holes are also included in this category (Table 4.29). The numbers of walls and features recorded in the five study areas can be used as a means for estimating how many walls would be expected at a group given an even or random distribution of wall features on walls. At Tsankawi there are considerably more holes than "wall expectation" and fewer at Groups A, I, and M. Some of this difference may result from differences in tuff at Tsankawi as compared to Frijoles: the top stratum at Tsankawi contains a great number of vesicles. We tried to be conservative in what we recorded as features (indeed we had to be, given the number of holes in some walls), but the tuff may have inflated the hole count. The poorer preservation of plaster at Tsankawi may also add to the higher count there, since fewer natural and/or disused holes remain covered by plaster. The measurement data show that this feature category does indeed cover a variety of holes in the wall, including some very large ones (Table 4.29, Figures 4.19a, b-4.20a, b), and that they are located at all heights. The means, however, show that most of these holes are fairly small and tend to occur a little less than 1 m above the floor. Modally (Figures 172 CAVATE STRUCTURES FEATURES 173 Table 4.29. Indeterminate Hole Occurrence and Dimensions. A. Indeterminate Hole Occurrence by Type and Group Shape Group A Group F Group I Group M Tsankawi Total Rectangular Oval Circular portion Cylindrical Hemispherical Truncated cone Truncated pyramid Conical Irregular 2 1 3 1 4 5 1 2 1 5 2 43 22 29 49 26 20 13 10 1 1 3 1 2 1 14 24 2 5 1 26 9 5 147 39 8 41 2 33 21 13 290 108 13 4 86 3 Total 91 78 51 74 277 571 Noncavate, grouped 15 10 10 37 B. Chamber Location by Shape Shape Right Wall Back Wall Left Wall Exterior Wall & Corners Back Corners Ceiling Total Rectangular 8 7 6 11 32 Oval 5 10 2 3 1 21 Circular portions 2 4 2 5 13 Cylindrical 56 119 62 35 4 11 287 Hemispherical 15 46 33 9 2 105 Truncated cone 5 4 3 1 13 Truncated pyramid 3 1 4 Conical 26 29 14 9 3 5 86 Irregular 1 1 1 3 Total 117 222 124 72 10 19 564 174 CAVATE STRUCTURES Table 4.29. (continued) C. Mean Dimensions by Grouped Shape in Meters Shape n Mean Standard Deviation Minimum Maximum CV Round Diameter 489 0.05 0.037 0.01 0.40 78.5 Depth 476 0.08 0.134 0.01 2.50 170.1 Volume 472 0.0008 0.0144 0.0000 0.3142 1812.5 Rectangular Width 33 0.09 0.056 0.02 0.32 58.9 Height 33 0.06 0.024 0.02 0.12 42.4 Volume 32 0.0004 0.0005 0.00002 0.0026 130.0 Oval Width 21 0.10 0.523 0.03 0.24 51.1 Height 21 0.12 0.151 0.01 0.60 125.2 Volume 11 0.0007 0.0014 0.00004 0.0050 188.5 All height above floor 411 0.86 0.399 -0.05 2.32 46.2 Note: B, seven cases in floor, unknown feature and unknown part are not included. C, includes only features that are 70 percent or more present. Round includes cylindrical, conical, and hemispherical shapes. Height above floor is only from chambers with less than 15 cm of fill. 4. 19a, b), indeterminate holes are only 3 cm in diameter and 5 cm deep, and while the various sizes are fairly evenly split between Tsankawi and Frijoles, there is some tendency for smaller holes to be found at Frijoles and larger ones at Tsankawi (Figure 4.20a). Of 370 cases that met the criteria for inclusion in Figure 4.20a, 323 (87%) had diameters and depths of 10 cm or less. Even with removal of cases with extreme diameter and depth values, the mean diameter remained at 4 cm, and the mean height above floor remained at 84 cm. The mean height above floor is very close to what might be considered a "typical" plaster height in the cavates, and many smaller indeterminate holes do occur at the plaster line. In that location it is easy to visualize them as containing pegs on which things (canteens, clothes, and so forth) could be hung, and once in a while they do contain sticks. Another indication that some may have been for pegs is that the plaster around the openings is often broken away, presumably from removal of the peg. Multivariate Analyses of Holes in Walls As an alternative means of examining the many round holes in cavate walls, two multi variate techniques were applied to a set of features conforming to the following specifications: Feature types: viga holes, latilla (?) holes, beam seats, and indeterminate holes FEATURES 175 (0 0) o X o 5 •c E o "O c I ' I ' I ■ I ■ I ' I ' I ■ I ' I ' I ' I ■ I 8 00QOOOQQOOOO r- ^ t- tunoo > *a 6 ^ § 13 $ .5 S 176 CAVATE STRUCTURES FEATURES 177 178 CAVATE STRUCTURES CO 0) c • MB E ■o c "55 73 E *v z\ ao o 1 I f I I * •v* ««, «*i 8 .60 FEATURES 179 Shapes: round orifices, cylindrical, conical, truncated cone, and spherical (a maximum diameter of 40 cm was used to exclude a few very large cases) Other criteria: only features judged to be at least 70 percent complete and located in chambers having less than 25 cm of fill (to control for completeness and accurate height above floor) This pruning procedure and exclusion of features with missing values resulted in a group of 650 holes. Cluster Analysis A cluster analysis (SAS FASTCLUS procedure) was run requesting six clusters, based on diameter, depth, and height above the floor. Repeated passes were made through the data readjusting the cluster "seeds" and locating cases in the closest cluster. The six clusters identified by this analysis vary from 5 to 242 members (Table 4.30). Six clusters were requested, to allow the program to identify two clusters more than the four categories presumably present. Although the clusters do not correspond precisely to the feature types, the fact that cluster membership is greatly reduced by cluster 6 suggests that searching for more than six clusters would be inappropriate. Predictably, indeterminate holes are spread the most evenly over the most clusters, though two of the clusters comprise mostly features from this category. Viga holes also dominate a cluster, though they are split between two clusters. The groups of holes created by this analysis (it should be remembered that the feature types were not provided to the cluster analysis) are recognizable by their means and form what may be useful-even functional- subdivisions of circular holes. Thus cluster 1, the group with the most members, consists of holes 2-8 cm in diameter located near the top of the plaster in many rooms. Cluster 2 contains considerably larger and deeper holes much higher on the wall, and cluster 3 is also well up the wall, though the holes are smaller, shallower, and lower. Cluster 4 is composed of much smaller holes lower on the wall. Clusters 5 and 6 are uncommon shapes and locations; both are deep, especially cluster 6, but cluster 6 is quite high (similar in height to cluster 3) and cluster 5 is near the floor. As might be predicted, the holes identified as viga holes are fairly consistent, though they come in two sizes. Discriminant Analysis This large group of round hole features was also analyzed using discriminant analysis (the SAS DISCRIM procedure; 645 cases analyzed, 5 omitted due to missing values). For the discriminant analysis the program was provided with types assigned and then calculated a profile for each feature type. Once again the variables used to describe the features were diameter, depth, and height above floor. The individual cases were then compared to the profiles and placed in the one to which they most closely conform. For this analysis the prior probability that a feature would fall into a given type was set at equal, which is not the case for the actual distribution, since indeterminate holes form 62 percent of the total. Because it is of interest whether or not "indeterminate" holes form an identifiable category, this is a reasonable prior condition. Based on a test of covariance matrix homogeneity performed by the program, within covariance matrices were used in the discriminant function. The discriminant analysis gives an idea of the metric overlap among the feature types (Table 4.31, Figures 4.18b, 4.21). Thus, the majority of features called viga holes and beam seats are described by similar measurements, though there are probably at least two subgroups: one higher, larger, and deeper (viga hole means) and the other lower (about 1 m) and smaller. Each of these also overlaps with the 180 CAVATE STRUCTURES Table 4.30. Results of Ouster Analysis on Round Wall Holes. A. Members and Means for Hole Clusters Cluster n Near Diameter Depth Height Above Floor 1 242 3 0.050 0.081 0.853 2 74 3 0.098 0.180 1.639 3 186 2 0.067 0.099 1.268 4 126 5 0.049 0.062 0.421 5 17 4 0.097 0.196 0.100 6 5 3 0.164 0.838 1.182 B. Ouster Membership by Assigned Feature Type Cluster Viga Latilla Indeterminate Beam Seat Total 1 16 18 173 35 242 2 47 1 18 8 74 3 49 15 80 42 186 4 1 117 8 126 5 1 13 3 17 6 3 1 1 5 Total 116 35 402 97 650 Table 4.31. Discriminant Analysis Classification of Feature Types. Computer-Assigned Type Original Type Viga Latilla Indeterminate Beam Seat Total Viga hole 80 16 1 15 112 LatUla (?) hole 1 30 4 35 Indeterminate hole 12 206 158 25 401 Beam seat 27 22 12 36 97 Total 120 274 175 76 645 FEATURES 181 0) a 0) o 0) ffV ?V 90 S 5 «u « ?> *• £ .2 -^ 1 J "1 .2P - §o e 5 |l| ° i * IS* ©:-**■* 182 CAVATE STRUCTURES other groups, which are smaller holes. Beam seat, as a category for features that are more difficult to interpret, shows considerably more overlap with the other two categories than does viga hole. The possible latilla hole group is the most consistent, probably at least in part because it is the smallest group. Indeterminate holes appear to fall into at least two main groups, more of which are like "latilla" holes than the smaller holes described by the measurements for the entire indeterminate hole group. Preservation and feature variability will always mean that a large number of wall features will have to be placed in a category such as "indeterminate holes," but as used here, this category is too inclusive. In future recording of features in cavates more types of small wall holes should be recorded, subdividing the "indeterminate hole" category used in this study. Additional types suggested by the analyses and by observation include: —peg holes. This type should be used for holes still containing portions of pegs or holes that show evidence for the removal of a peg. The group is likely to be fairly small, but isolating features with a known function will help place other, less easily interpreted features. -small holes at or near the top of the wall plaster. As noted, these may have been for pegs, but that use is less clear. It can be compared to the peg hole category. Possible Upper Loom Support (Code 14) The clearest examples of upper loom supports are deep, vigalike holes near the top of a chamber. The holes occur in pairs at a considerable angle to one another. Hewett's reconstruction of these features shows forked logs projecting from the seats; I do not know if this is based on intact examples or on speculation. Features associated with these angled supports include grooves in the ceiling (presumably to accommodate a cross-bar) and rows of loom anchors in the floor. As can be seen in Kent's (1983b) photographs of traditional looms, ordinary vigas can suffice as upper loom supports for looms close to walls, and it is quite possible that some de facto loom supports were called something else. The low ceilings and inclined walls of many cavates, however, may have required more specialized upper supports. The scarcity of identified upper loom supports at Groups F and A combined with the lack of loom anchors at those groups suggests that this pattern may be more than a sampling artifact (Table 4.32). Loom supports tend to be cylindrical holes in or near the ceiling of the chamber in which they are found. As we recorded these features, we came to recognize that in addition to their location, their angle of entrance into the ceiling was an important attribute. Because of the lateness of the realization and the difficulty of measuring it, however, we did not record this angle. The observed cases suggest that they are on average larger than viga holes and can be very deep. Smokehole (Code 15) Features called smoke holes are very commonly found near the chamber entrance. It seems likely that some smoke holes were separated from the door by a lintel, since some probable smoke holes now show as enlargements of the tops of doors, sometimes associated with grooves for rock or wood lintels. Smoke holes are usually quite large in diameter and angle up as they pass to the outside. Size, angle, and placement well up the wall are the most important criteria in assignment of this feature type. Smoke holes are slightly more abundant at Groups F and I than might be expected based on overall feature counts, and they are markedly infrequent at Group M (Table 4.33). This emphasizes the fact that Group M rooms made especially heavy use of masonry closing walls. The chamber locations show the strong (and in FEATURES 183 Table 4.32. Loom Support Occurrence and Dimensions. A. Loom Support Occurrence by Shape and Group Shape Group A Group F Group I Group M Tsankawi Total Trapezoidal 1 1 Rectangular 1 1 Cylindrical 1 3 2 14 20 Hemispherical 2 2 Truncated cone 1 1 Conical 3 3 Total 1 3 3 21 28 Shape B. Chamber Location by Shape Right Wall Back Wall Left Wall Ceiling Total Rectangular Trapezoidal Cylindrical 3 Hemispherical Truncated cone Conical 1 4 1 2 1 11 2 3 1 1 20 2 1 3 Total 17 28 C. Mean Dimensions of Circular Shapes in Meters Dimension n Mean Standard Deviation Minimum Maximum CV Diameter Depth 26 26 0.13 0.38 0.070 0.317 0.06 0.08 0.35 1.75 48.2 83.4 184 CAVATE STRUCTURES Table 4.33. Smokehole Occurrence and Dimensions. A. Smokehole Occurrence by Type and Group Shape Group A Group F Group I Group M Tsankawi Total Oval Trapezoidal Triangular Rectangular Cylindrical Truncated cone 11 1 3 2 1 1 7 5 2 1 1 21 7 1 2 1 46 1 Total 12 11 25 58 Shape Exterior Wall Back Wall Left Wall Rectangular solid Oval 7 Triangular 1 Trapezoidal 1 Cylindrical 35 1 4 Truncated 1 B. Chamber Location by Shape Ceiling Left Exterior Corner Floor Total 1 4 1 1 1 1 7 2 1 46 1 cone Total 45 58 C. Mean Dimensions for Circular Shapes in Meters Standard Dimension n Mean Deviation Minimum Maximum CV Diameter 39 0.23 0.073 0.06 0.35 32.2 Depth 39 0.49 0.226 0.15 1.01 46.2 FEATURES 185 a majority of cases necessary) preference for placement of smoke holes in the exterior wall. Clearly, they tend to be cylindrical holes bored through the cliff, but as with upper loom supports, the fact that most angle upward needs to be noted in addition to the recorded data. Vents (Code 16) Vents differ from smoke holes in two respects: they are horizontal rather than angled upward and they are generally found lower on the wall (see Figure 2.21b to left of door). Any hole passing through a wall that is not either a door or a smokehole was called a vent. Openings into other rooms as well as to the outside are included in this feature type. Like smoke holes, vents are predominantly cylindrical holes in the exterior wall. They, too, are infrequent at Group M and relatively abundant at Group F (Table 4.34). Groove (Code 31) Judging from differing size, location, and orientation, this feature category covers several probable functions. We observed some examples in ceilings, where they may have accommodated loom cross-bars; examples high on walls might have supported roofing ("wall ledge" would have been a more appropriate code for these); some may have helped support vertical partitions; and some may have been decorative. This code, then, describes a feature of questionable function (Table 4.35). Wall Depression (Code 37) Wall depressions are usually shallow concavities in tuff walls. Some of these features are plastered over, while others are clearly abraded through the plaster into the tuff. These features rarely have enough of a shelf at the base to hold anything (indeed, features with any shelf at all would probably have been called niches rather than wall depressions). Wall depressions are found at varying heights above the floor, but it seems likely mat those close to the floor in the vicinity of mealing features resulted from grinders' feet pushing against the wall (Figure 4.7, Table 4.20). Mills made this observation in the field, independent of Chapman's reconstruction (Hewett 1909a:451) showing the same probable function. Wall depressions are somewhat more abundant at Groups A and F (including the noncavate observations) than would be expected given numbers of features recorded (Table 4.36). These features also tend to be placed in back walls and corners and right walls as opposed to other chamber locations. The associated depression and metate rest in M-60 are located in the back corner of the room (Figure 4.7). The great variability in size, and thus probably in source and function, is visible in the mean volumes and areas and their coefficients of variation (Table 4.36C). Wall Ledge (Code 41) Occasionally a substantial ledge is found near the top of a chamber. Most likely, these features were an alternative roof support to viga holes, though they are much less common (see Figure 2.16b above door). The examples we encountered did not seem to have suitable width or flatness for storing items. Such ledges may have paralleled the main roof support of the room, serving to support secondary roof and/or floor members. Vertical Ceiling Hole (Code 47) Of 59 examples of this feature (also known as Panowski holes), we recorded 58 at Tsankawi and one at Group A. The code was added during the Tsankawi recording, so it was not available at Groups F, I, and M, and the first half of A, but few if any vertical ceiling holes are likely to be present in those groups. The nature of the tuff at Tsankawi may in part 186 CAVATE STRUCTURES Table 4.34. Vent Occurrence and Dimensions. A. Vent Occurrence by Shape and Group Shape Group A Group F Group I Group M Tsankawi Total Oval 2 2 Trapezoidal 1 1 Rectangular 1 1 Cylindrical 4 4 3 3 12 26 Truncated cone 1 1 2 Total 5 8 4 3 12 32 Shape B. Chamber Location by Shape itWal tack Wall Left Wall Exterior Wall Total All Shapes 2 2 6 20 1 1 32 Dimension n C. Mean Dimensions for Round Shapes in Meters Standard Mean Deviation Minimum Maximum CV Diameter Depth 23 23 0.18 0.35 0.068 0.153 0.04 0.33 0.09 0.60 38.6 43.2 Table 4.35. Groove Occurrence by Shape and Group. Shape Group A Group F Group I Group M Tsankawi Total Linear 2 1 3 6 Rectangular 1 1 Cylindrical 1 1 Total 2 1 1 4 8 Note: Grooves were located on exterior, right, and back walls (two or three on each wall type), and one case was on a ceiling (a loom support). FEATURES 187 Table 4.36. Wall Depression Occurrence and Dimensions. A. Wall Depression Occurrence by Shape and Group Shape Group A Group F Group I Group M Tsankawi Total Rectangular Rectangular solid Oval Cylindrical Hemispherical Conical Trapezoidal Circular portion Linear, irregular 3 7 5 1 5 1 2 1 2 2 3 1 1 1 2 3 1 2 1 1 1 2 1 2 2 5 8 1 1 9 14 11 3 19 4 3 2 3 Total 25 11 19 68 Noncavate, grouped 16 3 19 B. Chamber Location by Shape Shape Right Wall Back Wall Left Wall Exterior Wall & Corners Back Corners Other Total Rectangular 1 18 2 2 23 Oval 3 5 2 1 11 Trapezoidal 2 1 3 Cylindrical 3 3 Hemispherical 8 7 1 1 2 19 Conical 1 2 1 4 Circular portion 1 1 2 Linear, irregular 1 2 3 Total 16 35 8 4 5 68 188 CAVATE STRUCTURES Table 4.36. (continued) C. Mean Dimensions by Shape in Meters Standard Shape n Mean Deviation Minimum Maximum CV Rectangular 14 0.03167 0.0711 0.0006 0.2676 224.6 solid volume Hemispherical 18 0.01905 0.0313 0.0001 0.1031 164.1 volume Oval area 10 0.1531 0.1379 0.0154 0.4779 90.1 Rectangular 7 0.7004 1.2910 0.0391 3.6100 184.3 area Height above 23 0.34 0.218 0.04 0.86 63.4 floor" "Heights above floor include only chambers with less than 15 cm fill. explain their presence there, and it is unlikely that we overlooked their presence in the Frijoles sample recorded before the Tsankawi recording. The features consist of cylindrical vertical holes in the ceiling, sometimes extending to remarkable depths: the deepest recorded is 97 cm. The average depth is 18 cm, and the majority are 5-10 cm deep. The holes are usually fairly small in diameter (mean of 4 cm, with 95 percent of cases less than 10 cm) and are often extremely regular. They sometimes look mechanically created, but many have smoke-blackened interiors and thus appear to be ancient. They seem generally to occur in groups; the occurrences range from 1 to 12, averaging 4 per cavate. The function of this feature type is unknown; they may be related to weaving, partitions, or construction. The vesicular nature of the upper stratum of Tsankawi tuff also opens the possibility that some are natural, though some are unquestionably artificial. Narrow Wall Incisions (Code 48) These features are also almost unique to Tsankawi, with the exception of a single case at Group A. The code was also added during the Tsankawi recording, so it was not available at Groups F, I, and M. As is true of vertical ceiling holes, codes were added as new features were observed, and we did not note these features before working at Tsankawi. Differences both in tuff type and in prehistoric activity, rather than recording differences, account for most of the distributions in this sample. If this feature were as common in the Frijoles cavates as it is at Tsankawi, we would have added it sooner. The incisions are usually vertical or close to it, and are 1-2 cm wide, 1-3 cm deep, and up to 30 cm long (Figure 4.22). They appear to be incised by grinding. They generally occur in groups and were usually recorded as groups rather than individually; 26 cases are recorded. They seem most often to be FEATURES 189 in smaller, unfinished rooms, though they are not confined to such rooms. Many more may be covered by plaster. It is possible that they were meant to help hold plaster, but this seems doubtful; they are more likely some artifact of construction, perhaps of cutting out blocks of tuff to be knocked free. Eleven of the cases recorded are on back walls, and eight are on right walls. Hand-or-Toe hold (Code 33) Generally these are not features one would expect to find inside rooms. The cases we observed either were cliff features (e.g., trails at Group A) or may result from remodeling. At Group F in the area of F-21, F-22, F-24, and F-25, for example, rooms appear to have extended quite far up one part of the cliff; probably after the rooms were gone (but perhaps before or even intermediate to building phases) a hand-and-toe hold route was apparently put across the same part of the cliff. The only individually recorded cases with rooms in our sample are four examples from Group M; two extramural sets were recorded as noncavates at Group A. A group of 13 holds was recorded at Group A, and 16 more were recorded as a group at Group F. Incised Dado (Code 29) We invented this feature type to include an observed phenomenon. In five cavates we Figure 4.22. Narrow wall incisions in TS-24. This feature type was observed almost exclusively at Tsankawi, and may. be the result ofcavate excavation or may have been to help plaster adhere to the wall. The vesicular nature of the top layer of the tuff at the Tsankawi group is visible here. 190 CAVATE STRUCTURES recorded a distinct abraded band around substantial portions of the base of the chamber as defined by the top of the fill; in most cases the chambers containing this "feature" also had substantial amounts of fill. These bands are consistently 30-40 cm wide by about 6 cm deep, so that they seem intentional. We now believe, however, that incised dados are postoccupational damage, probably caused by large animals (probably sheep and goats, maybe burros or cattle) bedding down in chambers and rubbing against the walls. This interpretation is supported by the frequent association of dung with this "feature," and by its location at the top of the fill. Cliff Niche (Code 49) Cliff niches are distinctive features found only at Tsankawi. They consist of rectangular depressions in the tuff, mostly apparently outside of rooms, though quite possibly adjacent to rooftops. They look a great deal like fold-down metates, but hinges are lacking and gravity appears to be quite normal in the area. A group of five is located near TS-25 (Figure 4.23), where there is also a great deal of rock art. Indeed, they might even be considered rock art, as it is hard to impute a function to them. They are arranged so that three are next to one another and two more are widely spaced but close to the same level. Though they are several centimeters deep, the bases slope enough that it would be impossible to put anything in them. In the right light they look like doors, and they may have been designed to make the settlement look larger than it was from a distance (much as a cat tries to look big when frightened or a moth feigns to have large, scary eyes). The mean dimensions of the seven cliff niches we recorded are 0.62 m high by 0.37 m wide by 0.18 m deep. Mean door dimensions are fairly similar (0.70 m high by 0.56 m wide), and the cliff niche ranges fall well within the ranges for doors (Table 4.7). Another pair is located in the Tsankawi cavate group west of LA 50976, in the area of Lister's C-88-C-91. These niches are also located at the head of the rincon, on the caprock stratum, above cavates and probable masonry rooms. Of all the cliff niches observed, only one of these two is deep enough at the base (36 cm) to hold something. Rock Art More often than not the rock art visible appears to be incomplete due to combinations of plaster deterioration, weathering, and vandalism. Usually rectangular boundaries of figures were estimated and measured for recording, though it was occasionally possible to measure actual features. June Crowder recorded the rock art separately (see her summary at the end of this chapter; see also appendix 4 at the end of this study), so the rock art entries on our recording forms were usually quite general. The rock art was categorized by subject matter and means of manufacture, with any form of painting being termed a pictograph and any form of incision or pecking a petroglyph. Petroglyphs thus include the fine-line scratching in plaster that seems to have been quite common; Chapman (1916; in Hewett 1938) found this style of rock art especially interesting (see also Schaafsma 1980:285). Pictographs include any figures painted on the walls. We observed several colors of paint, including yellow, red, black, white, and green, and there are some figures that appear to have been done in a thin wash of plaster different in color from the wall plaster. Pictographs of all forms are less well preserved than incised or pecked rock art, harder to discern, and less frequent. The various rock art codes were applied as follows. Where two codes are present, both a petroglyph and a pictograph code were used. Geometric figures with definable geometric layout (codes 20, 22) include terraced figures, FEATURES 191 Figure 4.23. Four of the Jive cliff niches in the cap rock behind the Tsankawi cavates recorded by this project. Note that the bases of the niches are not flat enough to stand anything on. rectangles, circles, and crosses. Zoomorphic figures most commonly include parrots and other birds and Awanyus or horned (or plumed) serpents (codes 21, 23; Figures 4.24-4.28; Schaafsma 1980:255-288). Frequently it is clear that rock art exists, but it is either no longer a recognizable form or it may include several categories; these cases were called indeterminate (codes 24, 25). Early in the recording we observed handprints, which are a common prehistoric Pueblo form, and added a code for them (code 32); we did not, however, observe any more of them. The case we observed would be considered a pictograph since the prints are negative figures apparently done by spraying paint around an outspread hand. Anthropomorphs, on the other hand, proved to be fairly common (codes 34, 38); included in this category were fairly naturalistic figures (dancers and hunters) and supernatural figures (masks, katsinas, ogres). (See Figures 4.24- 4.28.) The rock art tallied here does not include the examples observed at Groups A and I and at Tsankawi that are not associated with rooms (June Crowder did record these panels; see appendix 4). Since the majority of the rock art present consists of some small (and usually undeterminable) fraction of the original, we have not included measurements for the various panels, though measurements are present in the data base. 192 CAVATE STRUCTURES Figure 4.24. Large katsina-like figures in TS-59. There is a terraced figure between them and an Awanyu line continues from them to the left. This chamber also contains a mask or katsina incised in the plaster next to a wall niche. This is the largest chamber recorded and that with the most features; it contains an array of loom features, floor features, and niches; viga holes and a low plaster "dado " are visible. The viga holes visible in this photo are of the variety that do not seem necessary for roofing; the presence of loom anchors and other upper loom supports in this room suggest that these "vigas * may have been involved with weaving. Not surprisingly, Group M and Tsankawi, the two groups with the most cavates have the preponderance of rock art (Table 4.37). Although they attract vandalism, cavates provide good conditions for preservation of rock art. Vandalism is most severe at Group A, increasing the chance that examples were overlooked there. Group M is notable for having the greatest variety of figures and media, including some elaborate polychrome panels, both within and outside the recorded sample. Group A is unusual in containing nearly equal numbers of painted and pecked or incised examples, while painted elements are infrequent at Group F. Though it is not apparent in the tables, the rock art at Tsankawi is notable for its larger scale. June Crowder summarizes her recording of rock art at the end of this chapter. Two preservation variables complicate the differences noted by Crowder between Tsankawi and Frijoles. First, the rimrock at Tsankawi is harder than any of the exposed cliffs in Frijoles-presumably that is why it is FEATURES 193 Figure 4.25. Large bird figures remain on either side of the door of TS-40, and a probable cloud motif is over the door. These figures are in the open on the very soft red tuff layer at Tsankawi, and must have been quite deep to have survived. 1 GF C23 L._ ._ .. ... ... ... 1. Figure 4.26. Probable bird figure incised through some of the smoked tuff wall; F-23. 194 CAVATE STRUCTURES Figure 4.27. Parallel zigzag lines, possibly Awanyus, in M-13; deeply chipped and abraded lines like this were seen elsewhere in Frijoles and also at Tsankawi. Patches of white paint are visible on the wall (this chamber also contained a handprint outlined in white). Vandalism to plaster is visible (scale =10 cm). Figure 4.28. Two masks or faces side by side in M-60. Notice how the figures appear to have been plastered over. This room also contains two metate rests (see Figure 4. 7). FEATURES 195 Table 4.37. Rock Art Occurrence and Chamber Location. A. Rock Art Occurrence by Type and Group Rock Art Code Group A Group F Group I Group M Tsankawi Geometric petroglyph Zoomorphic petroglyph Indeterminate petroglyph Anthropomorphic petroglyph 3 9 2 6 6 11 7 13 Total Geometric 4 a 4 4 6 18 petroglyph Zoomorphic petroglyph 3 a 4 6 9 12 34 Indeterminate 3 4 4 8 12 31 petroglyph Anthropomorphic petroglyph 2 a 1 2 4 7 16 Geometric 2 4 1 7 pictograph Zoomorphic pictograph 2 3 5 Indeterminate 8 1 1 7 5 22 pictograph Anthropomorphic pictograph 3 1 4 Hand print pictograph 3 3 Total 22 14 17 41 46 140 B. Chamber Location by Type Back Left Exterior Type Right Wall Wall Wall Wall Ceiling Other Total 16 32 14 31 196 CAVATE STRUCTURES Table 4.37. (continued) Type Back Left Exterior Right Wall Wall Wall Wall Ceiling Other Total Geometric pictograph Zoomorphic pictograph Indeterminate pictograph Anthropomorphic pictograph Hand print pictograph 7 4 13 1 1 5 22 4 3 Total 28 62 23 134 "Each includes 2 group-recorded examples. caprock. The remarkable extramural rock art at Tsankawi is confined to this caprock layer; the extramural rock art in the groups in Frijoles is now uniformly eroded and difficult to see, making comparisons difficult. Second, the plaster in Frijoles is more often in better condition than that at Tsankawi. Much of the Frijoles rock art we encountered was incised in the plaster, so that Frijoles and Tsankawi are again not quite comparable. Moreover, the rock art inside Frijoles chambers is not comparable to Frijoles cliff figures. These further cavate caveats do not imply disagreement with the differences noted by Crowder: there are several petroglyphs inside cavates at Tsankawi that are like nothing we saw in Frijoles, and it is likely that the extramural art was also different when it was all new. Steen (1979) gives the name Mortandad Style to bold figures incised through smoke blackening into the light tuff, creating a strong contrast between the white figure and the black background. He says this style occurs in a small area from Bayo to Ancho Canyons, the immediate vicinity of Tsankawi. Figures include Awanyus, Kokopellis, dancers, birds, and the Toltec sun god. Several cavates at LA 50976 contain rock art that fits this definition. Steen considers this style to have been only briefly used in the late fourteenth century, though the ceramics found by James Maxon (1962) in a "cave kiva" containing Steen's prime examples of the style suggest to him a date of 1325. An important criterion of Steen's definition of the style seems to be that the incised lines are free of soot. As Maxon notes, "Apparently either there were few fires in the kiva after the figures were carved, or they were periodically cleaned, as the incised areas have little or no soot remaining in them" (Maxon 1962:2). It is unclear whether an expert in the Mortandad style would classify the sooted figures in TS-59 (Figure 4.24) as belonging to that style. It may be that a later date for the style is correct, and that it represents use of the cavates after their primary occupation. Rock art executed in this fashion is either much rarer or absent in Frijoles FEATURES 197 cavates, perhaps lending some support to the notion of a territorial boundary between Frijoles and Tsankawi, at least in more recent times. Summary of Detailed Rock Art Study June Crowder The data recorded during this investigation are summarized in Table 4.38 and the tables in appendix 4 at the end of this study. The types of rock art and individual cavate contents are shown for both petroglyphs and pictographs for cavate Groups A, F, I, and M in Frijoles Canyon and for the selected group at Tsankawi (LA 50976). In addition, the cliff- face petroglyphs associated with the recorded groups are noted. Appendix 4 concludes with a brief correlation of the rock art recorded during this survey with that done by Chapman in 1916. Table 4.38 presents a breakdown of cavates containing rock art types as defined in Table 4.39. Though I made no formal compari- sons between this rock art and the examples described by Polly Schaafsma, our examples fall within the area and style she defines as the Tewa Division of the Rio Grande Style (Schaafsma 1975, 1980). Many of the cavate walls and ceilings have multiple petroglyph designs. Because it is difficult to enumerate motifs and desirable to simplify the presentation of the recorded information, Table 4.38 and the group data in appendix 4 present the frequency of cavates containing each design type occurring within each cavate group, rather than the absolute frequency of each motif. The few pictographs found were for the most part unidentifiable. They are included in the summary table of each group as the number of cavates containing pictographs (appendix 4), with a brief description of each in the individual cavate table. The objective of this investigation was to record the type, number, and condition of petroglyphs and pictographs in selected areas of Frijoles Canyon and in a sample at Tsankawi. The photographs we took can serve as a baseline against which to measure the effects of natural erosional forces, pollution, and vandalism over time. For example, in this survey, I matched 23 drawings from Frijoles Canyon done by Kenneth Chapman in 1916 (later published in Hewett's Pajarito Plateau and Its Ancient People [1938]) with the originals. Of these 23, 3 show severe deterioration and 3 show slight deterioration resulting from loss of wall plaster. There are three major differences be- tween the cavate wall drawings in Frijoles Canyon and those in Tsankawi. First, several two-horned serpents were found at Tsankawi but none in Frijoles Canyon. Second, Frijoles Canyon cavates contained many stylized parrots and other birds, while none was found at Tsankawi. Third, Tsankawi contained a larger number of anthropomorphic figures. In Frijoles Canyon, there was a notice- able difference between the art on the cavate walls and that on the cliff faces. The cavates contain a richer variety of art and more cere- monial figures than do the cliff faces. Art on the cliff faces at Tsankawi shows more variety than that at Frijoles Canyon and corresponds more closely to the cavate wall drawings. The total number of cavates that con- tained rock art was 46, of which 17 were in Group A, 7 in Group F, 4 in Group I, 6 in Group M, and 12 in LA 50976 (Tsankawi). In these cavates, 98 percent of the rock art was in- cised petroglyphs, 1.7 percent was abraded petroglyphs, and 0.3 percent pictographs. The cliff-face rock art from both Frijoles Canyon and Tsankawi consists only of petroglyphs. 198 CAVATE STRUCTURES Table 4.38. Cavates Containing Rock Art by Group and Motif. Motif Group A Group F Group I Group M Tsankawi Total Abstract 8 4 3 4 7 26 Geometric 4 3 3 3 2 15 Geometric abstract 1 1 2 4 Cross 1 1 2 Zigzags 1 1 2 Pictograph 4 1 1 3 1 10 Anthropomorphs 1 6 7 Ceremonial figure 2 2 2 6 Human figure 1 1 2 Mask 2 1 3 1 7 Hunter 1 1 Flute player 2 2 Stick figure 1 1 2 Realistic animal 1 1 2 Realistic bird 1 2 3 Realistic snake 1 1 2 4 Serpent 1 1 Two-horned serpent 2 2 Serpent motif 2 2 Snake motif 2 2 Stylized bird 1 1 2 Stylized parrot 1 1 Stylized insect 1 1 Terrace 1 1 2 Total aboriginal motifs 21 18 16 19 34 108 Cavates with art 17 12 46 Modern graffiti 11 18 FEATURES 199 Table 4.39. Rock Art Nomenclature. Abstract Geometric Geometric abstract Terrace Zigzags Anthropomorphic Ceremonial figure Human figure Masks Quadruped Realistic animal Realistic bird Realistic snake Serpent Serpent motif Snake motif Stylized bird Stylized parrot Pictograph Modern graffiti Includes curvilinear forms (different circular patterns), straight and curved lines, miscellaneous forms (dots, etc.) Connected straight lines forming a design Design that has abstract elements and geometries connected Geometric with stepped sides Lines with sharp angles in alternate directions "Boxlike" drawing of the human figure Elaborately decorated (especially headwear) human figure Realistic rendering of human figure (modern?) Both simple and stylized masks All animal types not falling into the realistic category (cliff-face figures only in this sample) Drawings of identifiable animals Figures easily identified as birds Figures easily identified as snakes Extremely long and wide snake figure Extremely long and wide snake figure without a head Any double-line figure of a snake without a head Elaborate drawings of birds Elaborate drawings of birds with definite parrotlike beaks Painted figures Names initials, dates, etc. Preliminary Functional Analysis of Cavate Chambers Ultimately, one of the main things we want to know about cavates is their prehistoric function. Archaeological determination of function is ideally based on several categories of evidence, such as associated artifacts, debris from activities, well-grounded analogy to ethnographic groups, and architectural morphology. In the present study, we are limited almost entirely to the last category. Functional inferences must therefore be tentative; they are based on the study of feature co-occurrence, chamber size, and estimates of intensity of use. Feature Co-occurrence To examine feature co-occurrence, the cavate feature data set was used to generate a second data set through a series of merges from various data sets; each case of the new data set is a cavate (noncavates are excluded). The variables are chamber volume, coats of plaster, and the occurrence of ten feature categories (see Table 5.1). Variable values for the feature categories are the sum of all the occurrences of features in a particular category. Number of plaster coats was taken from the back wall of each chamber, or added manually from a consciously selected wall for cavates without back walls recorded. Volumes were included by merging information from the feature lines for chambers. The result was a data set with 175 cases profiling the cavates recorded, though fewer cases had data for the plaster and volume variables (Table 5.1). The categories contain the following feature types: Walls: all masonry and natural walls Holes: indeterminate holes and possible latilla holes Beam features: viga holes, beam seats, and wall ledges Niches: wall niches and large floor-level niches Floor features: floor, firepit, floor burn, floor ridge, and floor pit Rock art: each recorded group of rock art (but not every figure) Doors: interior and exterior doors Vents: smokeholes and wall vents Other features: grooves, wall depressions, deflectors, narrow wall incisions Loom features: loom anchors and upper loom supports 201 202 CAVATE STRUCTURES Table 5.1. Summary of Chamber Attribute Occurrence. Attribute Mean Standard Deviation Minimum Maximum Median Number of 175 13.3 11.51 1 64 10 features" Plaster coats 165 1.8 1.97 9 1 Volume 156 3.62 3.10 0.3 18.6 2.9 Walls 175 3.0 1.20 6 3 Holes 175 3.3 4.30 24 2 Beam features 175 1.8 2.76 15 Niches 175 1.1 1.49 8 1 Floor features 175 0.9 1.31 6 Rock art 175 0.7 1.95 13 panels Doors 175 0.6 0.78 4 Vents 175 0.5 0.98 6 Other features 175 0.5 1.07 9 Loom features 175 0.3 2.03 24 "Excluding walls. Some rare feature types and ceilings are not included. The minimum occurrence column in Table 5.1 shows that no feature type occurs in every cavate. Ten chambers have no walls recorded; these "chambers" are mostly very irregular or partial (a few at Group M were "remotely recorded" by means of a ranger calling down observations). These ten have been excluded from the rank-order correlations of occurrence (Table 5.2), as have floor features, since recording them depends on depositional accident. The same could be said for loom anchors, but they are included because loom supports are visible. A matrix of Spearman rank-order correlations (r^ for feature counts was generated for the chamber data set (Table 5.2). Given that many chambers will be tied at zero or one for some features, the rank-order correlations of feature occurrence cannot be considered reliable predictors of the likelihood of finding feature & if feature b_ is present. Siegel (1956:210) states that the effect of ties is to raise the value of r s , though there is relatively little impact, at least on small data sets. Although the nonparametric correlation is more appropriate to these data, experimental runs with Pearson Correlations give similar orderings of association. The correlations provide some ordering of associations among features, volume, and plaster coats, with the proviso that infrequent feature types tend to have lower maximum r s values. Since feature counts are retained, the FUNCTIONAL ANALYSIS 203 Table 5.2. Co-occurrence of Feature Categories in Chambers. Feature 1 Feature 2 Neither Fl Only F2 0nly Both Holes Beam support 50 43 16 66 Holes Niches 49 32 17 77 Holes Vents 60 64 6 46 Holes Other features 60 73 6 36 Holes Rock art 59 74 7 38 Holes Loom features 65 100 1 10 Niches Rock art 74 59 7 35 Niches Floor features 58 41 23 53 Niches Vents 70 54 11 40 Niches Loom features 80 85 1 9 Beam supports Niches 58 23 35 59 Beam supports Vents 75 49 18 33 Beam supports Loom features 90 75 3 7 Vents Floor features 77 22 47 29 Rock art Floor features 84 15 49 27 Rock art Loom features 128 37 5 5 Floor features Loom features 96 69 3 7 Other features Loom features 127 38 6 4 correlations also give added dimension to the straight co-occurrence data (Table 5.3). Quite understandably, we find a fairly good association between chamber volume and the number of features present, but there is less likelihood that a larger chamber will have been plastered many times. Generally, and again predictably, the matrix shows that large numbers of features tend to occur together (see especially the feature number correlations in Table 5.2). Among the many correlations, several deserve note. The presence of niches seems to correlate well with holes, beams, volume, rock art, and plaster coats. There are somewhat surprising correlations between floor features and rock art, and floor features and plaster coats. These correlations may result in part from the fact that chambers showing floor features tend to be well-protected, well- preserved ones. The correlation between art and vents may be in part similarly explained, and the correlation between vents and doors is related to the presence of intact exterior walls. The 204 CAVATE STRUCTURES Table 5.3. Spearman Rank-Order Correlations of Feature Category Co-occurrence. A. Plaster Coats, Number of Features, Walls, and Holes with Other Feature Types Feature Volume Plaster Coats Feature n Walls Holes Volume n 1.000 148 0.378 143 0.631 148 0.247 148 0.485 148 Plaster coats n 1.000 160 0.460 160 0.064** 160 0.386 160 Feature number n 1.000 165 0.389 165 0.798 165 Walls n 1.000 0.275 165 Holes n 1.000 165 Beam supports nt 0.411 148 0.386 160 0.596 165 0.171* 165 0.373 165 Niches 0.543 0.467 0.685 0.216 0.5220 Vents 0.445 0.320 0.556 0.451 0.403 Rock art panels 0.266 0.384 0.527 0.111** 0.300 Floor features 0.225 0.472 0.371 -0.069** 0.118** Loom features 0.222 0.134** 0.274 0.071** 0.184 Doors 0.234 0.093** 0.426 0.469 0.309 Other features 0.226 0.060** 0.390 0.180* 0.246 FUNCTIONAL ANALYSIS 205 Table 5.3. (continued) B. Beams, Niches, Vents, Rock Art, and Floor Features with Loom and Door Features Feature Beams Niches Vents Rock Art Floor Beam supports 1.000 0.443 0.225 0.245 0.208 nt 165 165 165 165 165 Niches 1.000 0.335 0.379 0.289 Vents 1.000 0.309 0.200 Rock art panels 1.000 0.343 Loom features 0.184 0.273 0.134** 0.171* 0.235 Doors 0.168* 0.247 0.480 0.168 -0.029** Other features 0.117** 0.194 0.174* 0.221 0.189 C. Loom Features and Doors with Other Features Feature Loom Doors Other Door Other 0.046** 0.100** 1.000 0.147** Note: The feature numbers used do not include walls as features; values for volumes are rounded to the nearest 0.1 m 3 for the calculations. ♦Associated probability of no correlation greater than 0.01 (.01MB) THEN ME=((MB/MA)*MC) IF (SHAPE=14 AND MA0 THEN VOL=AREA*MD; Calculates the volume for plane figures with depth measurements recorded. This procedure was carried out for all cavate features and all individually measured noncavate features. 240 CAVATE STRUCTURES Retained Cavate Computer Outputs Note: These are on file at the National Park Service Southwest Regional Office, Santa Fe. I. Base Data A. Early Oracle output, uncorrected (1986) 1 . Cavbase variables 2. Noncavbase variables 3. Cavate levels 4. Noncavate levels 5. Cavate exposures, fill depths 6. Listing of cavate base information 7. Tabulations of cavate variables 8. Tabulations of noncavate variables B. Listings following successful conversion to UNM/IBM format 1 . Cavate base data 2/ 19/87 2. Noncavate base data 2/19/87 C. SAS formatting 1. Cavate formatting 3/9/87 2. Creation of notes variable, trial list 3/13/87 3. Catalog base data set 3/15 4. Format noncavate base data 3/12/87 D. Listings-Correction copies 1. Listing of base data with corrections marked; notes A,F,I 3/14 2. Base data notes for Group M and Tsankawi 3/15/87 3. Noncavate base data and notes, corrections marked 3/16/87 4. Addition of UTMs 5. Addition of new LA number, renames field LA 3/4/88 6. Line corrections 3/14/88 7. Listing of cavate base data for Groups F and I 3/16/88 8. Photo listing for cavate base data 5/3/87 E. Tabulations 1 . Cross-tabulations of cavate base variables 4/1 1/87 2. Cavate HPGS by group and level 5/3/87 3. Noncavate base data cross-tabulations 4/11/87 II. Feature Data A. Oracle outputs, before data checking 1 . Formats for CAVFEAT, NONCAVFEAT 9/86 2. Frequencies, type shape, count, measurements, with and without nulls 3. (4 outputs) 4. Listing of cavate features, uncorrected 9/86 B. Listings following conversion to UNM/IBM format 1. Cavate features 2/20/87 2. Noncavate features 2/20/87 3. Generation to disk C. SAS formatting 1. Cavate formats 3/11/87 2. Noncavate formats 3/11/87 APPENDIX 2 241 D. SAS listings, correction copies 1 . Cavate feature data 3/17/87 2. Cavate feature notes 3/19/87 3. Noncavate feature data 3/28/87 4. Noncavate feature notes 3/28/87 5. Listing of features with shapes 13 and 14 by BJM to check/correct measurement order 6. Feature photo generation and listing by photo 5/1/87 7. Feature photos sorted by feature type 5/6/87 E. Feature tabulations 1. Cavate and noncavate features by shape by group, by shape by part, by shape by cavate type, all features 4/19/87 2. Cavate features, type by group 5/9/87 3. Comparison of CFEAT+NCFEA with CAVEAT +NONCV data sets, feature type by group 3/3/88 4. Counts of all noncavate features 2/24-6/88 5. Cavate features group by feature type PC-SAS 3/3/89 6. Noncavate features group by feature type 3/4/89 7. Occurrence of viga and latilla holes by record no. 2/26/88 + other notes 8. Niches-type co-occurrence, height above floor 2/24/88 9. Milling feature co-occurrence 3/3 and 3/13/89 F. Feature metrics 1. Generation of volumes and areas, fill groups-test 4/18/87 2. Volume and area means by type and shape 4/19/87 3. Feature metric replacements, refined shape breakdown, counts by shape 5/1/87 4. Door and niche dimensions 5/8/87 5. Niche, beam seat, indeterminate hole dimensions 6/19/87 6. Indeterminate hole dimensions and breakdowns 2/88 7. Mean dimensions for conical and cylindrical features, wall and floor areas 5/7/87 8. Firepit and floor pit dimensions; plaster height to HAFL ratio by feature type 6/19/87 9. Fire pit and floor pit, floor depression dimension breakdowns 2/88 G. Feature condition 1 . Feature type by natural and human damage 2/26/88 2. Wall damage by group and part; hearth listing 5/5/87 3. Smoking on walls and ceilings by plaster, function, group 3/10/88 4. Moderately and unweathered walls: plaster coats and heights, and plaster color by part, group, function 3/25/88 5. Plaster color by group: all walls, back walls 6/9/89 6. Floor plaster coats by group and function; wall plaster coats by features, smoked coats 1988 H. Multivariate analysis of features 1. NCSS Cluster analysis of indeterminate and latilla holes 6/18/87 2. Cluster analysis attempt for several hole types 7/19/87 3. Cluster & discriminant analysis: several hole types 3/4/88 242 CAVATE STRUCTURES III. Chamber data A. Listings and attributes 1 . Creation of chamber data set, some frequencies 6/10/87 2. Listing of chamber data set, frequencies of attributes by group and function 7/19/87 3. Generation of location index, tabulations 2/26/88 4. Repairs to data set 3/14/88 5. Listing of corrected and updated chamber data set 3/23/88 B. Tabulations and analyses 1 . Chamber metrics by group, function, shape (from feature, rather than chamber data set) 6/19/87 2. Shape by niche 2/26/88 3. Means, breakdowns, group comparisons, t-tests for chamber volumes 2-3/88 4. Means for chamber volumes by group combining habitation and "kiva" classes 3/5/89, 3/9/89 5. Tables of feature number by plaster coats, chamber data 3/14/88 6. Co-occurrence of feature groups 3/21/88 7. Feature occurrence correlations, all and selected chambers 3/23/88 8. Location index cross-tabulations 4/4/89 C. Multivariate analyses 1. Discriminant analysis by function and group, first attempt 7/19/87 2. Discriminant analysis by function and group, different variables 3/25/88 3. Cluster analysis, Ward's method 3/3/88 4. Feature cooccurrence and cluster analysis 3/14/88 5. Fastclus with 6 and 10 clusters, different variables 3/16/88 6. Chamber clusters with listings 3/18/88 7. Fastclus analyses, listings different clusters and iterations 3/21/88 IV. Data management A. Transfer from NPS format to IBM format 1. First attempts 2/5/87 2. ASCII tape with incorrect block sizes 2/18/87 3. UNIX tape index showing incorrect block sizes from NPS 2/18 4. Regeneration of UNIX data (M. Prine) 2/19/87 5. Generation of data from tape to UNM disk 2/19/87 B. Tape index for UNM formats 1 . Tape on file at NPS in IBM format (CAVUNX) 2/20/88 C. Transfer SAS formats and data to NPS system 1. Six data sets to tape 12/16/87 2. Seventh set to tape 1/14/87 3. XCOPY for export data set 2/1 1/88 V. Listings of corrected files A. Contents and listing of CAVBASE.SSD-cavate base data B. Contents and listing of NCBASE.SSD-noncavate base data C. Contents and listing of CFEAT.SSD-cavate feature data D. Contents and listing of NCFEA.SSD-noncavate feature data E. Contents and listing of CAVEAT. SSD~cavate feature analysis set APPENDIX 2 243 1 . Frijoles data 2. Tsankawi F. Contents and listing of NONCV.SSD-noncavate feature analysis (single features only) G. Contents and listing of CHAMBER. SSD--chamber analysis set VI. Various SAS GRAPH programs for generating figures 244 CAVATE STRUCTURES Photographic and Videotape Data Photo Rolls, Formats, and Numbering Conventions 4x5 These were numbered only by group and exposure number, and were so entered. Photo A-2, for example, corresponds to the number written on the negative and the photo record. Rock art (35 mm) The Crowders assigned roll numbers for this series; they have had the prefix CRA (Cavate Rock Art) attached. CRA10:12,13 indicates black-and-white rock art roll 10, exposures 12 and 13. Color slides (35 mm) There is only one roll of slides; it has been labeled CC (Cavate Color); CC1:5 is Cavate Color Roll 1, exposure 5. NPS rolls (35 mm) Four rolls were used during the recording session, two were shot before the season started, and three were shot in April 1987. All these rolls were given the prefix CAV; the four used during the season were CAV1-CAV4, while the two taken before were CAV8 (taken by Toll in April) and CAV9 (taken by A. Ireland, also in April 1986). CAV9 is all distant shots of whole groups from across the canyon, and was not entered on any data lines. CAV8 includes a number of shots from groups in which we did not record. Also included with this group are three rolls (CAV5-CAV7) taken in April, 1987 by B. Crowder and W. Toll to prepare frontal drawings, to rephotograph a few more cavates shown in Lister's photos, to fill in some gaps, and to reproduce Lister's 1939 photos of cavates in which we worked in 1986. Only a few of the April 1987 (CAV5-CAV7) shots were entered on data lines. Video Tapes Videotapes were taken for all cavate areas recorded. These provide color shots of all walls and features as of summer 1986, along with verbal commentary, except for a few tapes on which there were some audio problems (these problems were partially corrected by voice-over after recording). The tapes are in VHS format and are kept at the Southwest Region offices of the National Park Service, Santa Fe. Appendix 3: Base Information, Threatened Cavates, and Room Stability Computer Output of Most Base Information for All 1986 Cavate Forms Variable abbreviations are as follows: RC# Record number GP Cavate group cv# Cavate number ELEV Elevation EXP Exposure in degrees true north PGS Height above present ground surface FT Fill type FILDPT Fill depth (cm) COM Completeness CON Evidence for construction NES Natural vs. excavated space M Masonry UP Rooms up-canyon DOWN Rooms down-canyon STAB Stability TUFF Tuff type NHUSE Nonhuman use FUNC Function assignment On this listing zeros show as blanks Note: Variable values and definitions can be found in Appendix 1. 245 246 CAVATE STRUCTURES U.3ZU I — — CM — Ift — — — W — — — CM — — (0 — — CM — CMCM — — — — — — CMCM — — — — — — — — CMinCMCMCMCM — CM — IrtCM — I ZI3MU i * vios * « « in in in in « « cm co co «- oo — in»<*invwN^-in * t -<■<■* i t-3u.ii. i n co co *» co co co co » co co 9 * <■ i to »- « » i -OMOnnci 3ti «m(m(mm — — — ooooanatBocthhhsitiBiBitttnnnnNr — — cocococococmcmcmcm — ol ___________ - i a i — ^oio^ioooo^ — cocouj^^r>>«jr»ocouoo>- tooior^cMcooof^cMcowo — cm v I — — — CM CM CM CM — CM — CM CM CO ^ — CM — — CM CM CO — — — — — — CM — — CM — — CM — CM — — MM — — — CM — CM — — NN I uoz i '<«- in — * r^ ^ * in mo oitmig co in in in co - i — — — co co — coco co — id co id co — toco— — to — co od in co cm minco— in co * co — coco co w i i cd o in r- ooooino — ininininincMocMincoininoa>cooininoininoo)oo ooooo ooin Oio a>t>-oo) co — oino>*r^o«)cocot>-r^inincocMa> — oocm — co«M^r>-ooo> a> ^ o ^ cm ooin a. I— CM — — — in — CO — mCM — CM — CM— CM — — — I — — II— CM (N CM— CM— — I a. i in — fooMnaiaininooNiniooN^iDiDNinnoaio- aootf>oinina>ininoooino x i r- •- o — ^ ^ co *• w UJ I — — — — CMCMCMCM — CM — — CMCMCMCMCM — — CMfMCM — CM — CMCM — CMCMCM — CM — CM — CM CM CM — — — CMCMCMCMCMCMCMCM I > i oooooooooooooooooooooooooooooooooooooooooooooooooo uj i — — cmcmcm — — — co — — — cm — — co — — — — — — — — — oooouJr»N(C(0(0l — — — — — CMCMCMCMCOCOCOCO*inWtO<0<0 a. a. a. a. ofu.2u.su. u. 3 3 a or jaatraijLCcau-crTQractctaccccceo; x 1 ami x oou o 00 uo I o I o in 01— xuioujuioujujoujoujujujujoujujiijuj u 1 i-i-o.i-ino.t-i-ijhiahaiaiaiiiiit-ii-iifflhiizi-iiaiffllllii-iiil oc 1 xxoxoxooxxaoxooxooaonnaaaxaxooujxaaoxaooaujooaoxoooo I (D1O333D3D_ UJ I -»->->->-»->->->->->-J->->->->-5<<<<<<<<<->->->->->->-»->->->->->-»->->->->->->->->->-J->-> I- I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I < i o>o>(jiO)0>o>cococotococococococo Q I — — CM— — CMCMCMCMCMCMCMCMCMCMCM — — — — — — — — — CMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMNCMN f- «o ^ o — co

»'r*ifliniOiji(OiflMniBiDa)*ooo--t--NNN a. i — — — cococococococococo — — — — — — — — — — — — — — cm — — — — — — — — — — i APPENDIX 3 247 U.3ZOI co — — — cm win — — cm — cm — — cmco — inwin — in — — in — --mm — co — co — cm — — — cm — — in cm — cm — — i •-MM I3MUJ i io-tii) icwmuiin r»T* * v p» » v » ^ r» r» in » in in «• «■ ^ v i 10 \- 3 u. u. i co ************ <» * « — t * »- < co i — * «r — co co — *• cm co * -^•-Nn^n^nnfxnno'-NNn-NB'-nn^nnoonnNn^ i cm — — zi en m — — «»'-oiB^»oo«)ooioi"-NiOMO((ino)oiMfliflN«oUN o I CO CM CO ZU1V) I •-•■ — — CO CM — CM N CM •"N-'-NNNN- CMCM — CM — — — — CM — CM — — — CMCM — CMCMCMrMCMCMrMCM — I CM — — uoz i itiAioiDin — ioism in — co — — * co * *■ co co co ^ io * ^ — * ^ 'f it'* i UOII <■'•■•■« l>l-NNNt — CM — — ^^CM^-^^CM**^* — — ^COCMCO — CMC0CM^C0***^CM*CM I n — — ji- i ninvrto eninooowioinocMcocM noN-ooN-oN'- o o o cm — oooo — — — co — — ■-• 0. I O CM O O O WO — CMOOOOCMOOO OOOOOOOOOOOOOO— OOOOOOOOO OO U.QI CROO I O — CO I ... u-Ki in « co — — — co in in — co — — — — — — — — — co — — w — — — • -co i co in co i/i i o o o o o oenwcMOooooooooooaooiniDooinoooiocMoo* moooooooooo d i aio-'h NoimooooaiaiaiNOiNiflovnMiin^NOaqiAvn^ cooo- i^oicoo — co — co a i ooo I— — — CM O O CM I — — CM CM CM CM 10 CM CM «» CO — CM CO CO CM — CM (N ^ — CM CM ^ in ^T CM I o. i t in in in in o in — — o an in — ^ » — — in ao in co o cm ^ m-eooois- in — oinoin cm x i in ^ ^ ^ <■ cm r» 10 co co CM^-cotcom cm — in in o o co r» coco — ^coooJcocMmio^iooQ v UJICMCMCMCMCMininCMCMCMCMCMCM (MCMCMCMCMCM CMCMCMCMCMCMCM CM CM CM CM CM «M CM — N N N N (M N N CM I CM U> CM CM — CM > i ooooo ooooooooooinoinooointnoooinooooaininoinininmoinominoino uii in in en in to kii/ii(inn-----oo'-nnoooO'-'-0'-0"00o-^-0'-o^oo-oo _J I NNNNNOOONNNNNNNNNNIMNNNNNIMNNNNNNNNNNNNNNNNNNNNNNNNN uj i (o I — CM CM CM CM C0COC0 — — — — — — — — CMCMCMCMCMCMC0C0C0C0C0COC0COVVfV« z a. a z t- t- hi- a a a a i a -> or a u.u.u.ora.o.-)a,u-u- a -> 32 )) a a. u_ u. a aciiDffliiJiuintDnuiisiocDcDfiOD co co XX xx co oo co id O I >» 'V "* >» V, ^ V ^ •>• ^ >» ^ ^ ^ v. V ^ v. >^ ••». v. v. ■v. -^ u 1 »-aazza.aaazi-)-(-i-zzi-ah-t-azzZA.zzo.i-Ka.ZZi-i-CLZZt-ai-t-ZZa.aa.CLa.ai- a 1 xuiDcoocoococaxxxxomxoxxocflocacDiDmcoxxocDoxxcocQcaxcoxxiDaicDiDCDaicficox I ->->-y-i-y-y->->-)~>->-»->-y-)->-i->->->-i->-y~i->~>-i->~)->->->-i~>~>->->-i-i~>~i-i-)->~>~i~i~>~> t- I I I I I i I I I I i I I I I I i i i i i i i i i i i i i i i i i i i i i i i i i i i i i I i I I I I < i nnn7««Y4tt«aoaooaoocomioiciDi»(iio>oioao)aia)0oo — cMcoMfooc»onvin^cMin 1 CMCMCMCOMCOCOMCOO* — — — (010(0— — — — CMCMCMCMCMCMCOCOCMCOCOCO^MrCOCO^^^-^inWin OK I — ______,-.-,()- I z a -> a co CO ■N* >. a z K K H- H- a. -> * 2 i * CO CO X I I X 248 CAVATE STRUCTURES U.3ZUI— CM — — — CM CM .- .-.-m — — CM — — — — — — — CM — — CM — — CM — — CM — — - — r- — NNO-- — -CO — CM I l3Mui i r>- v « * — in < 10 in co v *• — — * in in * — — — — in — -NN(M<- — CM NN(Mt-N»NNWIM^nr)NNNN>-nN(Me»r»*co*cocM 5 i r-.r-r-r-r-f-acM q i — i a I O)0IN7IA«N« — 3 i mmtotototo* — i i Si— co - ZIUI/l I N<-NN>-- — — CM — CMCM — — CMCMCMCM' I CM I CM CM CM i (N-(N (N — C4 uoz i iflNiniflinn m — «o — co — — — in — «» — * in — co in — cm co cm — co co co in to — to w cm i OOSMV — — CM — — — CM CM — IM — (OMCMCON — — WV — — — M — — <•> — — CM — — — MCM — — — — CM CM CM — CO CM CM I j(- i on cm — r» mo in to cm cmocococmuicoo oinvininr»ooinooooina>iDoininrM--o- ooio -• a. i o o ooo o— o— o ocmcmooooo ocm — f» r» — ^moNininNO- ocm — ooooococo — u.f-i co cococo t OO in to — ooaoinoooininococominoMCMininoinofoooico- cm m co m k cocooh-* — — oo>cM — N^mincMO — o — o>oocfto>r^o)oo>o> — cmooi — in^cMointa r-^^^.».-CMCMCMCM- — — — — — — — — — — — — — — — w- w- — ■-»-.- > i ooooooooo oooooooooooooooooooooooooooooooooooooooo uj i oooooococmo oo — — oooocnooooooooooooootscocoo- t^t^t^aoaaxDaasaxB _i i cMCMCMCMcMCMCMCMioocoioiOieiOiototoiniotoininininininininininin^^^^inininininininininininininin UJ I t0l0(Ol0(Ot0l0l£l0O» oo o> o — v in w r»a>o>o — covinton»coeno — CMcowintor-cocno- cm co ^ in o — Nra^mtomoio- co > i mininintoioioto — — — — cMCMCMCMCMCMCMCMCMcocococococococococo**«^**ininininininininin i/)ioi/ii/)vii/ivivii/)i/ivii/)i/iv)toi/ivii/iv)i/)vii'ii/ti/)i/ii/ivii/ii/ii/ii/ii/ivii/ii/ii/ii/iini/ii/) KI-t-l-»-»-KI-»-(-l-l-»-»-l-l I I I a i uj i a i I o- 3 ■ DIBIl com & a a. »- s t- Q- »- n- >>te a s: u. a a * -> u. 3 as 33 z UJ (0 03 ■ uj ■ x id so I oa r IX UJ'vlL.'V^v.'*.^ v. x. ^. v. >.%» V.V. :luiaiiiiia.jiigit.iiaiiiLiiaii >->oco.->oK"^-^ao:->a:a->-»a-»-»a.->->Q.Q. IO3OflDiO3UJO3flDOBUJAuJfl3O9flDCDO3(DO3O3fl303O3 Z O. Z -> a -7 03 03 03 os a Z Z 2 03 03 CD U. X U. BID >, >. ^. Z a. z Z -> a. -> "j 03 03 03 03 Z a Z t- 03 03 03 I I- a. a. * a. a X 03 03 3 3 2 III ■v. v. x. Q. Z Z Z Z 0. -> -> -> -» CD CD CD CD CD CO fp TE CO CD 00 CD CD CD CD CD CD QD CD CD CD OO Q3 CO CD CD CD ^JD ^D QD 00 00 CD CD CD ^0 35 OB 00 DP * T) CD CD CD 00 00 ^ fl CO OO O B ^^ 00 00 CD CO i i i i i i i i i • i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i JJJJJJJJJJJJJJJt3l90Jl3l}Ol30aeOt9Dl3t]l]l)OOl3l]lll}t}Ul3l}i:OOOt}0 33333333333333333333303333333333333333333303333333 ->->->">-»->->->->->->->->->-><<«-><<«««<«<<«««•«<<<<«•««<<««<«•«<<<< I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I in^r>-inininio«o — — — — — — — — — — — ^^k — — — — — — — — — cMCMCMcocMCMCMt^r^«r«-tototointoininin — OO — — — — — COCOCMCOCOCOPIOOOCOOOO — — — — — — — — — — — — o— — — oooooooooooo i ro co «■» t in to - coinior^cococMOD- cMco«oifli>inooo>o-cMco«rtnco»tina>ofxin *• co cm i inininininmiotooooo — — — cmcmcm — cococoa>a>o>oicncnci>cnoioo- — — — cMior»r»r» H < CD I — CM CO — NONNOBNT-N- — ^ON- — —CM Z I 2 I O I I a i 3 i S I CM — CM CO CM CO CM - V — — CM — — — — — CN uoz i lonieiS'- — cm i to m co UOII "->-<-n" o o> coo> r- o I OOOOOOOOOOOOOOOOOOOOQOOO uj i aaoNo-oooooo^ifiaioMOoooooo _i i inininin<0 i «oinir>inu?mir>mir>tnu)inmmif)inu)inui (/)(/)(/)(/)(/>(/)t/> t-t-t-t-t-t-t-t-y-t-t-t-t-t-t-t-t-t-t-t-t-t-t-t- a iu Q a o u u a a. u. a. to cd 3 2 U. I ICO a. u. a m co ao.sszi-zsa.so.t-zzai-1-t-ixsszzz CDGDCDCDCOXCOCOCOCOCQXCOCOCOXXXCOCDCOCQCOfD I I I I I I I I I I I I I I I I I I I I I I I I I I OOOO-1-J-IOUOOOOOOOOOOOOOOO I 3333333333333333=)33=)3333 Ul I <<<<->->-)<<<<<<<<<<<<<<<<< t- I I I I I I I I I I I I I I I I I I I I I I I I I < i ***co — — — — — — ****«o-N«OO»**(fl»(0MB0)-OOCM I CMCMCM l a © CM 0\ CO c o r i 8 B* .s — "■* 8 5 .S if cd J3 T3 g ff 2 3 Q- o §■8 * > > .g 4> 2 ^ * eg £ O Q •8-1 2? 2 « 18s • O £ 1,11 It-S o 68 © — CM CM a 2 "8 S S J i-s §• I s3 § OS o O o CM tu ~ •5 3 2 -a w •- P 8 "3 CM APPENDIX 3 251 o S j2 » O g * .9g| - £-5 K * * <* £ cs s s s s s 252 CAVATE STRUCTURES o Z E o M E o o 03 > U a. 2 a at % u i> s i S op 1 g 8 < "C ■" =3 J= © P a i 8 5 u J2 "3 5 o. >>m >. 3 8 -a -° 8 w O 4> v «-• Z, 00 I- u_ o 5 c « «- o bOJO 5 2 o c -^ 2 *3 M "> c o u- c U O O '" 3 = S 5 M 65 5 ~ O « <„ u y jo 5 8 £ 2 « S J = H o 35 -S 15 C CO H «n 11 3JJ 00 APPENDIX 3 253 Breakdown o/Cavate Room Stability by Cavate Group Record Cavate Amount Excavated Human Natural Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage GROUP A Apparently stable 1 106 4 1 1 3 2 100 34 1 3 1 1 3 334 50 1 3 1 9 9 4 340 57 1 4 2 1 Lesser threat 5 67 1 2 3 1 9 9 6 68 2 2 1 9 9 7 80 13 2 2 9 9 8 85 18 2 4 2 3 5 9 330 47 2 3 1 2 3 10 345 63 2 1 5 11 354 66 2 1 9 9 12 346 67 2 4 1 2 5 Greater threat 13 82 15 3 3 2 2 6 14 89 23 3 3 2 6 15 104 30 3 3 9 9 16 99 32 3 3 4 1 6 17 342 60 3 2 9 9 18 343 62 3 3 6 19 347 73 3 1 9 9 Major problem 20 75 10 4 2 9 9 21 86 20 4 4 2 9 9 22 90 22 4 3 1 9 9 23 87 39 4 4 2 9 9 24 352 71 4 1 9 9 254 CAVATE STRUCTURES Record Cavate Amount Excavated Human Natural Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage GROUP F Apparently stable 25 147 4 Lesser threat 26 144 2 27 155 12 28 161 15 29 159 16 30 167 23 31 166 27 32 174 31 33 183 38 34 186 45 Greater threat 35 152 9 36 180 35 37 197 56 38 200 59 Major problem 39 181 37 2 2 2 2 2 2 2 2 2 3 3 3 3 9 9 9 9 9 1 9 9 5 5 9 9 1 5 5 9 9 6 9 6 GROUP I Apparently stable 40 122 19 1 9 9 41 131 26 1 42 135 31 1 6 43 136 32 1 9 9 44 500 34 1 9 9 Lesser threat 45 108 2 2 1 5 46 109 3 2 1 1 APPENDIX 3 255 Record Cavate Observation Number Number Amount Excavated Stability Tuff Type vs. Amount Natural Human Natural Damage Damage 6 2 6 6 6 9 9 9 9 9 9 9 9 6 6 9 9 9 9 9 9 47 113 7 48 114 8 49 117 12 50 124 17 Greater threat 51 111 5 52 116 10 53 120 13 54 123 14 55 132 27 56 137 33 Major problem 57 125 20 58 127 22 59 140 37 2 2 2 2 3 3 3 3 3 3 4 4 4 4 3 2 1 1 2 1 2 1 2 3 2 1 1 2 irently stable 1 60 8 8 61 10 10 62 23 29 63 30 34 64 36 37 65 58 57 66 56 61 67 63 65 zr threat 68 1 1 69 4 4 70 13 13 71 24 30 GROUP M 4 4 4 4 4 4 4 1 1 1 2 1 1 1 1 2 1 1 1 9 9 9 9 9 9 9 9 9 1 9 9 9 3 9 9 1 9 9 256 CAVATE STRUCTURES Record Cavate Amount Excavated Human Natural Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage 72 25 31 2 2 9 9 73 28 35 2 4 1 9 9 74 51 51 2 4 2 9 9 75 57 58 2 2 5 76 54 59 2 2 5 77 55 60 2 1 9 9 78 61 64 2 Greater threat 79 2 2 3 4 2 9 9 80 14 14 3 4 2 9 9 81 64 16 3 4 2 6 82 15 18 3 1 9 9 83 17 20 3 4 2 9 9 84 29 23 3 4 2 6 85 27 24 3 1 1 9 9 86 32 33 3 1 4 87 33 36 3 1 9 9 88 34 38 3 4 2 9 9 89 40 40 3 2 9 9 90 39 44 3 2 9 9 91 42 46 3 2 9 9 92 45 48 3 2 9 9 93 46 49 3 2 6 94 47 50 3 4 2 6 95 52 52 3 4 2 6 Major problem 96 3 3 4 4 2 4 97 9 9 4 4 2 9 9 98 65 15 4 4 2 6 99 62 17 4 2 9 9 100 41 41 4 1 9 9 101 50 55 4 1 9 9 APPENDIX 3 257 Record Cavate Amount Excavated Human Natural Observation Number Number Stability Tuff Type vs. Amount Natural Damage Damage 102 53 56 4 TSANKAWI 2 6 Apparently stable 103 237 28 5 1 1 104 293 35 6 1 3 105 274 52 1 9 9 106 256 58 5 1 1 107 259 59 1 9 9 108 248 64 5 1 3 109 322 67 1 1 110 241 521 1 9 9 111 265 530 5 1 112 264 531 1 9 9 113 316 564 6 1 114 503 567 6 1 9 9 115 305 568 6 1 9 9 Lesser threat 116 205 15 2 5 1 9 9 117 206 16 2 5 1 5 118 207 17 2 5 2 9 9 119 208 18 2 5 1 5 120 213 19 2 5 2 5 6 121 218 21 2 5 1 9 9 122 222 24 2 5 2 9 9 123 223 25 2 5 2 5 124 216 26 2 1 5 6 125 235 27 2 2 9 9 126 289 31 2 6 2 5 127 290 32 2 6 1 5 128 291 33 2 6 1 9 9 258 CAVATE STRUCTURES Record Cavate Observation Number Number Stabi 129 292 34 2 130 295 37 2 131 296 38 2 132 309 40 2 133 313 41 2 134 314 42 2 135 315 43 2 136 275 51 2 137 276 53 2 138 262 55 2 139 261 56 2 140 253 61 2 141 252 63 2 142 242 65 2 143 204 501 2 144 217 508 2 145 224 509 2 146 236 517 2 147 244 523 2 148 266 533 2 149 297 553 2 150 321 570 2 Greater threat 151 285 29 3 152 288 30 3 153 294 36 3 154 319 44 3 155 320 45 3 156 267 50 3 157 263 54 3 158 254 60 3 159 247 66 3 Amount Excavated Human Natural Stability Tuff Type vs. Amount Natural Damage Damage 6 6 6 6 6 6 6 5 5 5 5 3 5 5 6 6 6 6 6 6 1 2 2 3 4 2 1 2 1 1 1 2 2 1 2 1 1 9 9 9 9 9 9 1 9 9 9 9 9 9 9 9 9 9 9 9 9 9 7 6 6 9 9 9 9 9 9 5 1 9 9 3 5 1 3 6 9 9 9 9 9 9 9 9 9 9 9 9 APPENDIX 3 259 Record Cavate Observation Number Number Amount Excavated Human Natural bility Tuff Type vs. Amount Natural Damage Damage 3 5 2 9 9 3 5 3 6 3 2 6 3 6 1 6 4 5 1 5 6 4 6 1 9 9 4 4 2 9 9 4 9 9 160 219 507 161 232 514 162 233 515 163 287 548 jor problem 164 221 23 165 304 39 166 239 519 167 270 536 Note: Variable values and definitions can be found in appendix 1 . 260 CAVATE STRUCTURES Noncavate Stability Sorted by Rating, Group, and Number Record Number Cavate Number Stability Natural Damage Human Damage Apparently stable 69 3 71 6 103 36 Lesser threat 70 5 74 9 95 27 101 33 102 35 325 42 328 45 329 46 336 53 338 55 339 56 350 59 354 74 Greater threat 83 16 91 21 92 24 93 25 94 26 96 28 98 31 77 37 105 40 326 43 327 44 GROUP A 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 1 4 2 4 6 6 5 5 6 5 6 5 2 6 4 6 6 6 6 6 6 6 6 6 6 6 1 6 1 APPENDIX 3 261 Record Number Cavate Number Stability Natural Damage Human Damage 331 48 3 6 332 49 3 6 333 51 3 6 335 52 3 6 337 54 3 6 1 341 58 3 6 344 61 3 6 349 64 3 6 348 65 3 6 355 68 3 6 356 69 3 6 353 70 3 6 351 72 3 6 Major problem 72 7 4 6 73 8 4 6 79 12 4 6 7 81 14 4 6 2 88 19 4 6 78 38 4 6 324 41 4 6 84 17 9 GROUP F 9 9 Lesser threat 143 1 2 6 148 5 2 6 153 10 2 6 158 14 2 6 163 18 2 5 165 19 2 6 164 20 2 6 262 CAVATE STRUCTURES Record Number Cavate Number Stability Natural Damage Human Damage 169 21 2 2 172 22 2 6 179 26 2 6 1 168 29 2 6 171 30 2 6 173 32 2 6 177 33 2 6 178 34 2 6 184 39 2 6 191 41 2 6 187 44 2 2 189 46 2 6 145 47 2 6 193 50 2 6 194 52 2 6 Greater threat 146 3 3 6 149 6 3 6 150 7 3 6 151 8 3 6 154 11 3 6 156 13 3 6 162 17 3 1 176 24 3 2 175 25 3 6 170 28 3 2 182 36 3 6 185 40 3 6 190 42 3 2 188 43 3 6 157 48 3 6 203 49 3 6 192 51 3 6 APPENDIX 3 263 Record Number Cavate Number 195 53 196 54 198 55 199 57 202 61 Major problem 201 58 Apparently stable 115 9 Lesser threat 118 16 119 18 126 21 133 28 139 35 141 36 142 38 Greater threat 107 1 110 4 112 6 121 11 128 23 129 24 130 25 134 29 Lesser threat 26 32 Stability Natural Damage Human Damage 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 6 6 6 6 6 6 6 6 5 6 6 2 6 6 6 6 6 6 6 6 GROUP I GROUP M 264 CAVATE STRUCTURES Record Number Cavate Number Stability Natural Damage Human Damage Greater threat 5 5 3 2 7 7 3 9 9 19 22 3 2 21 26 3 6 31 27 3 2 35 39 3 6 43 45 3 6 44 47 3 6 66 66 3 6 Major problem 6 6 4 2 11 11 4 2 12 12 4 6 22 28 4 4 48 53 4 6 1 59 62 4 6 60 63 4 6 TSANKAWI Stable 249 525 2 250 526 6 251 527 6 257 528 6 278 541 6 302 556 5 Lesser threat 220 22 2 6 210 503 2 6 211 504 2 6 215 505 2 6 5 214 506 2 6 5 APPENDIX 3 265 Record Number Cavate Number Stability Natural Damage Human Damage 225 510 2 5 226 511 2 1 227 512 2 5 231 513 2 6 234 516 2 6 238 518 2 6 243 522 2 2 246 524 2 5 265 532 2 6 5 269 535 2 6 272 538 2 1 273 539 2 2 277 540 2 6 279 542 2 6 280 543 2 6 281 544 2 6 283 545 2 6 282 546 2 6 284 547 2 5 288 549 2 6 298 551 2 5 299 552 2 2 303 557 2 1 306 558 2 5 307 559 2 5 308 560 2 5 310 561 2 5 311 562 2 5 312 563 2 6 317 565 2 5 318 566 2 6 323 569 2 6 266 CAVATE STRUCTURES Record Number Cavate Number Stability Natural Damage Human Damage Greater threat 255 62 3 6 209 502 3 6 260 529 3 6 268 534 3 6 271 537 3 6 286 550 3 6 300 554 3 6 Major problem 258 57 4 2 4 Note: Variable values and definitions can be found in appendix 1. Appendix 4: Detailed Listing of Rock Art June Crowder Summary of Rock Art Tables by Cavate Group Group A Number of cavates in Group A containing the indicated rock art: Abstracts 8 Cross 1 Geometries 4 Geometric abstracts Modern graffiti 11 Realistic bird Realistic snake Zig-zags Pictographs 4 Total 32 s of the individual cavates: Petroglyphs: Cavate 1 modern graffiti 2 abstracts 3 geometric (enclosed hourglass) 5 modern graffiti 10 zig-zags 12 modern graffiti 13 abstracts, geometries, modern graffiti 14 modern graffiti 15 abstracts, realistic bird 16 abstracts, crosses (2), geometries 18 modern graffiti 22 abstracts, modern graffiti 32 abstracts, modern graffiti, realistic snake 50 abstracts, geometries, modern graffiti 60 abstracts, modern graffiti ? (no number assigned; 2-3 cavates east of #75) abstracts, complex geometric abstract, modern graffiti Pictographs, or possible remains of pictographs: Cavate 15 small group of red splotches 50 well executed red circle outlined in black, lower one-quarter and interior missing ? (1-2 cavates east of #73) 4 small white stripes, lower one-quarter of wall has remnants of a white border and a small vertical stripe with horizontal slashes 267 268 CAVATE STRUCTURES ? (2-3 cavates east of #73) small red stripes and red splotches A73 large white figure, white stars, etc. Group F Number of cavates on Group F containing the indicated rock art: Abstracts 4 Ceremonial figure Geometries 2 3 Geometric abstracts 1 Human figure Masks 1 2 Modern graffiti Realistic snake 3 1 Stylized bird Stylized insect Terrace 1 1 1 Pictographs 1 Total 21 Contents of the individual cavates: Petroglyphs: Cavate 2 24 25 26 27 31 38 Pictograph: Cavate 38 abstracts, geometries, complex stylized bird, terraces, ceremonial figure simple mask, stylized mask, stylized insect (?) geometric abstract modern graffiti complex curvilinear abstract, realistic snakes (4), geometries, abstracts abstracts, modern graffiti ceremonial figure, modern graffiti, abstracts, geometric designs, human figure (armless), masks (4) possible remnants of pictographs composed of red paint splotches with superimposed incised lines APPENDIX 4 269 Group I Number of cavates in Group I containing the indicated rock art: Abstracts 3 Anthropomorphic figure 1 Ceremonial figure 2 Geometries 3 Geometric abstracts 2 Mask Modern graffiti Realistic animal Stick figure Stylized parrots Pictograph Total 17 Contents of the individual cavates: Petroglyphs: Cavate 8 12 19 26 Pictographs: Cavate 19 abstracts, geometries, ceremonial figures (3), stick figure geometric abstract stylized parrots (7), ceremonial figures (3), abstracts, geometries, modern graffiti geometries, abstracts, geometric abstract, squirrel profile, anthropomorph, masks (2) white circle (23 cm diameter), possible white parrot-type beak, two white solid rectangles, tan boxes (2) 270 CAVATE STRUCTURES Group M Number of cavates in Group M containing the indicated rock art: Abstracts 4 Geometries 3 Human figure Hunter 1 1 Masks 3 Realistic animal 1 Stylized birds Stick figure Zig-zags Pictograph 1 1 1 3 Total Contents of the individual cavates: 19 Petroglyphs: Cavate 13 18 33 40 41 60 Pictographs: Cavate 13 33 41 large zig-zags, abstracts abstracts geometries geometries, abstracts, stylized birds (4), mask profile walking-stick figure deer profile, hunter, small masks (9), geometric, outline of human torso (no features) abstracts, geometries, stylized mask, simple masks (3) white paint outline of right hand several design remnants composed of a partial red headdress, red stripes, yellow stripes, red and yellow paint, yellow paint with red dots, small black and yellow interlocking fret, some superimposition of incised lines red vertical realistic snake with superimposed incised lines APPENDIX 4 271 Tsankawi (LA 50976) Number of cavates in LA 50976 containing the indicated rock art: Abstracts 7 Anthropomorphic figures Bird, realistic 6 2 Ceremonial figures Cross 2 1 Flute player Geometries 2 2 Masks 1 Modern graffiti Realistic snake 3 2 Serpent Serpent motif Serpent, two-horned Snake motif 1 2 2 2 Terrace 1 Pictograph 1 Total Contents of the individual cavates: 37 Petroglyphs: Cavate 16 20 26 33 40 41 53 54 59 61 64 66 Pictographs: Cavate 59 flute player, anthropomorph, abstracts anthropomorph, ceremonial figure, two-horned serpent anthropomorphic figure circle and dots on ceiling, abstracts outside entrance: 2 realistic birds, wavy line abstracts two-horned serpents (2), two-horned anthropomorph, head of two- horned serpent unfinished flute player, modern graffiti, abstracts cross, serpent motifs, abstracts, geometric, concentric circles, realistic bird figures (2), ceremonial figures (4), realistic snake, masks (4), terrace, snake motif ceiling snake motif abstracts, anthropomorph, serpents (one with a rattle), modern graffiti, geometries geometries, abstracts, serpent motifs, realistic snake, modern graffiti, anthropomorphic figure five small white stripes 272 CAVATE STRUCTURES Cliff-face Petroglyphs Type and location of the various petroglyphs found on the cliff faces at Frijoles canyon and Tsankawi: Group A anthropomorphic figure, geometric design, bird, concentric circle, terrace Group I birds, snake, quadrupeds LA 50976 corn symbol, quadrupeds, anthropomorphic figures, birds, abstracts, arrow, snake motif, Kokopelli, sun symbol APPENDIX 4 273 Historical Correlation with Chapman Twenty-three of Chapman's drawings (Hewett 1938) from Frijoles Canyon were matched with the originals during this survey. The table below lists the location of these specific petroglyphs. Also included is a brief description of noticeable changes from the drawings. Chapman (Hewett 19381 Location Plate III h 1-19 (photo 8) Plate IV f A- 13 (photo 2)— slight loss of bottom design due to deterioration of wall plaster k F-2 (photo 4) k F-38 (photo 4)~severe deterioration of middle design area due to loss of wall plaster Plate V a F-2 (photo 4) c A- 13 (photo 5) Plate VII a 1-19 (photo 9)~head shows severe deterioration due to loss of wall plaster b 1-19 (photo 3)-severe deterioration of body design due to loss of wall plaster, head marred with an incised X e 1-19 (photo 9) Plate VIII a 1-19 (photo 6) g 1-19 (photo 7) h 1-19 (photo 5) Plate IX a M-33 (photo 3) c M-33 (photo 7) d M-33 (photo 4) f M-33 (photo 3) h M-33 (photo 3)~slight deterioration of edge of body design due to loss of wall plaster Plate X a F-2 (photo 3)~slight deterioration of design due to loss of wall plaster Plate XI c 1-19 (photo 2)~incised X across face Plate XII c F-2 (photo 5) Plate Xin d 1-19 (photo 8) f M-60 (photo 2) j F-38 (photo 3A)-incised lines through smallest mask Appendix 5: Chamber Cluster Membership Chamber Cluster Analysis Listings Cluster Analysis Results Sorted by Cluster and Distance from Seed Cavate Cluster Volume Feature n Plaster From Seed TS 53 1 4.83 39 2 4.3 TS 55 1 6.14 43 1 4.3 TS 50 2 9.77 29 3 3.7 TS 26 2 9.47 20 4.0 TS 27 2 13.15 43 2 4.1 TS 64 2 11.75 50 3 4.4 TS 15 2 10.46 27 5.1 TS 66 3 17.00 58 4 3.3 TS 20 3 15.50 41 3 6.4 TS 59 3 18.63 64 2 6.9 M 59 4 7.11 36 6 1.9 M 35 4 2.72 28 7 5.1 A 50 4 6.42 32 7 5.2 A 18 4 6.17 36 3 9.4 TS 16 5 4.79 24 3 2.9 F 31 5 6.26 19 5 3.1 F 35 5 3.51 21 4 3.4 A 22 5 5.18 26 3 3.7 M 38 5 5.65 18 2 3.8 TS 30 5 5.93 29 2 3.8 A 13 5 5.33 27 5 4.4 F 45 5 4.05 18 6 4.6 F 23 5 3.56 25 3 4.9 A 32 5 4.07 24 7 5.0 275 276 CAVATE STRUCTURES Cavate Cluster Volume Feature n Plaster From Seed 5 7.58 21 7 5.6 5 7.47 25 8 6.4 5 4.78 29 9 7.6 6 .21 0.8 6 2.22 5 1 0.8 6 1.59 3 1 0.9 6 2.83 3 1.1 6 1.29 5 1 1.2 6 1.44 6 1 1.2 6 1.82 4 1.3 6 .20 1.4 6 .40 1.4 6 .40 1.4 6 1.02 2 1 1.4 6 0.93 5 1 1.5 6 1.32 6 2 1.5 6 3.43 5 1 1.5 6 1.17 4 1.6 6 3.12 6 1.6 6 1.20 2 1.7 6 2.32 4 1.7 6 3.06 4 1.7 6 0.45 2 1 1.8 6 0.48 2 1 1.8 6 0.56 3 1 1.8 6 0.60 3 1 1.8 6 0.86 2 1.8 6 1.49 5 1 1.8 6 0.59 8 1 1.9 6 0.72 4 1.9 A 67 A 47 F 38 I 5 I 20 TS 501 TS 515 I 7 M 51 TS 19 I 3 TS 509 TS 517 M 37 A 30 TS 521 M 2 TS 519 TS 536 F 56 A 39 I 14 F 12 M 10 A 4 TS 508 TS 570 M 61 TS 531 A 63 APPENDIX 5 277 Cavate Cluster Volume Feature n Plaster From Seed TS 533 6 0.73 6 1.9 TS 37 6 0.83 5 1.9 I 17 6 2.76 6 1 1.9 I 32 6 2.85 7 1 1.9 F 15 6 0.35 5 1 2.0 TS 507 6 0.45 6 . 2.0 A 57 6 0.64 4 2.0 M 8 6 3.26 4 1 2.1 M 29 6 0.34 5 2 2.2 TS 530 6 1.22 5 2.2 TS 40 6 0.85 6 2.3 TS 548 6 0.99 6 2.3 M 34 6 1.35 19 5 2.3 A 66 6 0.62 7 2.4 M 41 6 0.59 9 1 2.5 TS 23 6 2.84 9 2.9 TS 523 6 5.11 6 2.9 A 10 6 2.65 17 1 3.0 M 57 6 0.51 8 1 3.1 I 33 6 5.23 3 1 3.2 A 2 6 . 14 4 3.4 TS 41 6 4.09 15 1 3.4 I 2 6 . 1 1 3.5 M 1 6 4.76 5 3 3.6 A 73 6 2.27 19 3 4.5 F 4 6 5.11 11 4 4.7 TS 18 6 1.33 12 4.8 I 12 6 3.07 8 6 5.3 TS 44 6 8.18 4 3 6.4 TS 25 6 9.25 14 7.2 M 31 6 7.38 7 6 7.3 278 CAVATE STRUCTURES Cavate Cluster Volume Feature n Plaster From Seed 7 15.15 28 1 2.7 7 22 1 3.0 8 3.04 18 1 3.5 8 4.94 43 5 3.5 9 3.21 12 1.9 9 3.14 10 2.6 9 5.49 16 2 2.7 9 0.74 10 1 3.0 9 0.86 8 1 3.0 9 1.23 7 3.1 9 4.00 16 1 3.1 9 2.69 24 3.2 9 13 2 3.3 9 4.28 16 4 3.3 9 2.48 17 2 3.6 9 4.10 14 4 3.7 9 1.32 21 3.8 9 1.48 15 4 3.8 9 3,17 25 3 3.9 9 4.49 16 3.9 9 3.02 20 3 4.0 9 2.24 16 1 4.6 9 4.35 22 2 4.6 9 2.04 10 1 5.2 9 2.86 14 6 5.2 9 3.25 22 5.5 9 0.94 23 . 5.7 9 6.69 25 2 6.1 TS 24 TS 21 M 13 I 19 TS 34 TS 58 F 16 I 31 I 34 TS 63 TS 33 TS 56 I 30 A 34 A 71 A 23 TS 52 M 36 F 27 TS 17 F 2 F 37 I 26 F 59 I 37 TS 61 TS 51 TS 54 APPENDIX 5 279 Cavate Cluster Volume Feature n Plaster From Seed I 13 10 2.86 12 1 2.7 A 60 10 5.25 20 2 3.3 I 22 10 1.52 14 1 3.5 M 55 10 2.69 20 3 3.7 TS 28 10 2.95 29 3.7 TS 65 10 5.08 21 1 3.7 M 30 10 1.84 10 1 3.8 M 44 10 1.54 17 3 4.0 M 9 10 6.39 23 1 4.0 M 58 10 2.87 25 1 4.3 M 60 10 5.25 27 2 4.6 M 40 10 3.81 15 6 4.8 M 20 10 6.31 24 3 4.8 TS 45 10 7.02 14 4.9 TS 29 10 9.17 19 1 5.2 280 CAVATE STRUCTURES Ouster Analysis Results Sorted by Group and Cavate Cavate Cluster Volume Feature n" Plaster From Seed 6 14 4 3.4 6 0.56 3 1 1.8 6 2.65 17 1 3.0 5 5.33 27 5 4.4 4 6.17 36 3 9.4 5 5.18 26 3 3.7 9 4.10 14 4 3.7 6 0.93 5 1 1.5 5 4.07 24 7 5.0 9 4.28 16 4 3.3 6 2.32 4 1.7 5 7.47 25 8 6.4 4 6.42 32 7 5.2 6 0.64 4 2.0 10 5.25 20 2 3.3 6 0.72 4 1.9 6 0.62 7 2.4 5 7.58 21 7 5.6 9 2.48 17 2 3.6 6 2.27 19 3 4.5 9 3.02 20 3 4.0 6 5.11 11 4 4.7 6 0.45 2 1 1.8 6 0.35 5 1 2.0 9 5.49 16 2 2.7 5 3.56 25 3 4.9 9 3.17 25 3 3.9 5 6.26 19 5 3.1 5 3.51 21 4 3.4 A 2 A 4 A 10 A 13 A 18 A 22 A 23 A 30 A 32 A 34 A 39 A 47 A 50 A 57 A 60 A 63 A 66 A 67 A 71 A 73 F 2 F 4 F 12 F 15 F 16 F 23 F 27 F 31 F 35 APPENDIX 5 281 Cavate Cluster Volume Feature n a Plaster From Seed F 37 9 2.24 16 1 4.6 F 38 5 4.78 29 9 7.6 F 45 5 4.05 18 6 4.6 F 56 6 1.20 2 1.7 F 59 9 2.04 10 1 5.2 2 6 1 1 3.5 3 6 . 2 1.4 5 6 . 2 1 0.8 7 6 1.29 5 1 1.2 12 6 3.07 8 6 5.3 13 10 2.86 12 1 2.7 14 6 3.06 4 1.7 17 6 2.76 6 1 1.9 19 8 4.94 43 5 3.5 20 6 2.22 5 1 0.8 22 10 1.52 14 1 3.5 26 9 4.35 22 2 4.6 30 9 . 13 2 3.3 31 9 0.74 10 1 3.0 32 6 2.85 7 1 1.9 33 6 5.23 3 1 3.2 34 9 0.86 8 1 3.0 37 9 2.86 14 6 5.2 M 1 6 4.76 5 3 3.6 M 2 6 3.43 5 1 1.5 M 8 6 3.26 4 1 2.1 M 9 10 6.39 23 1 4.0 M 10 6 0.48 2 1 1.8 M 13 8 3.04 18 1 3.5 282 CAVATE STRUCTURES Cavate Cluster Volume Feature n a Plaster From Seed M 20 10 6.31 24 3 4.8 M 29 6 0.34 5 2 2.2 M 30 10 1.84 10 1 3.8 M 31 6 7.38 7 6 7.3 M 34 6 1.35 19 5 2.3 M 35 4 2.72 28 7 5.1 M 36 9 1.48 15 4 3.8 M 37 6 1.02 2 1 1.4 M 38 5 5.65 18 2 3.8 M 40 10 3.81 15 6 4.8 M 41 6 0.59 9 1 2.5 M 44 10 1.54 17 3 4.0 M 51 6 1.44 6 1 1.2 M 55 10 2.69 20 3 3.7 M 57 6 0.51 8 1 3.1 M 58 10 2.87 25 1 4.3 M 59 4 7.11 36 6 1.9 M 60 10 5.25 27 2 4.6 M 61 6 1.49 5 1 1.8 TS 15 2 10.46 27 5.1 TS 16 5 4.79 24 3 2.9 TS 17 9 4.49 16 3.9 TS 18 6 1.33 12 4.8 TS 19 6 1.82 4 1.3 TS 20 3 15.50 41 3 6.4 TS 21 7 . 22 1 3.0 TS 23 6 2.84 9 2.9 TS 24 7 15.15 28 1 2.7 TS 25 6 9.25 14 7.2 TS 26 2 9.47 20 4.0 APPENDIX 5 283 Cavate Cluster Volume Feature n' Plaster From Seed 2 13.15 43 2 4.1 10 2.95 29 3.7 10 9.17 19 1 5.2 5 5.93 29 2 3.8 9 4.00 16 1 3.1 9 3.21 12 1.9 6 0.83 5 1.9 6 0.85 6 2.3 6 4.09 15 1 3.4 6 8.18 4 3 6.4 10 7.02 14 4.9 2 9.77 29 3 3.7 9 0.94 23 5.7 9 1.32 21 3.8 1 4.83 39 2 4.3 9 6.69 25 2 6.1 1 6.14 43 1 4.3 9 2.69 24 3.2 9 3.14 10 2.6 3 18.63 64 2 6.9 9 3.25 22 5.5 9 1.23 7 3.1 2 11.75 50 3 4.4 10 5.08 21 1 3.7 3 17.00 58 4 3.3 6 1.59 3 1 0.9 6 0.45 6 2.0 6 0.60 3 1 1.8 6 4 1.4 6 2.83 3 1.1 6 4 1.4 TS 27 TS 28 TS 29 TS 30 TS 33 TS 34 TS 37 TS 40 TS 41 TS 44 TS 45 TS 50 TS 51 TS 52 TS 53 TS 54 TS 55 TS 56 TS 58 TS 59 TS 61 TS 63 TS 64 TS 65 TS 66 TS 501 TS 507 TS 508 TS 509 TS 515 TS 517 284 CAVATE STRUCTURES Cavate Cluster Volume Feature n a Plaster From Seed TS 519 6 1.17 4 1.6 TS 521 6 1.32 6 2 1.5 TS 523 6 5.11 6 2.9 TS 530 6 1.22 5 2.2 TS 531 6 0.59 8 1 1.9 TS 533 6 0.73 6 1.9 TS 536 6 3.12 6 1.6 TS 548 6 0.99 6 2.3 TS 570 6 0.86 2 1.8 "Feature n is all features except walls. Not all of the features in the count were included in the cluster analysis. References Bailey, R. A., R. L. Smith, and C. S. Ross 1969 Stratigraphic Nomenclature of Volcanic Rocks in the Jemez Mountains, New Mexico. U.S. Geological Survey Bulletin 1274-P. U.S. Government Printing Office, Washington. Bandelier, Adolph F. 1971 The Delight Makers. Harcourt Brace Jovanovich, New York. Originally published 1890. Beam, George L. 1906 The Prehistoric Ruin of Tsankawi. National Geographic 20:807-822. Bierbower, Susan 1905 Among the Cliff and Cavate Dwellings of New Mexico. Records of the Past 4:226- 233. Blair, Jonathan S. 1970 Keeping House in a Cappadocian Cave. National Geographic 138:126-146. Breternitz, David A. 1966 An Appraisal of Tree-Ring Dated Pottery in the Southwest. Anthropological Papers of the University of Arizona 10. University of Arizona, Tucson. Canby, Thomas Y. 1982 The Anasazi: Riddles in the Ruins. National Geographic 162:562-592. Carlson, Ingrid K., and T. A. Kohler 1989 LA 50972 (Cavate M77). In Bandelier Archaeological Excavation Project: Research Design and Summer 1988 Sampling, edited by T. A. Kohler, pp. 50-59. Reports of Investigations 61. Washington State University, Pullman. Caywood, Louis R. 1966 The Excavation of Rainbow House, Bandelier National Monument, New Mexico. National Park Service, Southwestern Archeological Center, Globe, Ariz. Chapman, Kenneth M. 1916 Graphic Art of the Cave Dwellers. Palacio 3 (2):37-41. El 1938 The Cave Pictographs of the Rito de los Frijoles. In Pajarito Plateau and Its Ancient People, by E. L. Hewett. University of New Mexico Press, Albuquerque. Colton, Harold S. 1932 A Survey of Prehistoric Sites in the Region of Flagstaff, Arizona. Smithsonian Institution Bureau of American Ethnology Bulletin 104. Washington, D.C. 1946 The Sinagua: A Summary of the Archaeology of the Region of Flagstaff, Arizona. Museum of Northern Arizona Bulletin No. 22. Flagstaff, Arizona. Cordell, Linda S. 1979 A Cultural Resources Overview of the Middle Rio Grande Valley, New Mexico. USDI Bureau of Land Management and USDA Forest Service, Santa Fe and Albuquerque. 285 286 CAVATE STRUCTURES Dane, Carle H., and George O. Bachman 1965 Geologic Map of New Mexico. U.S. Geologic Survey, Washington, D.C. Dougherty, Julia D. 1980 An Archeological Evaluation of Tsiping Ruin. Cultural Resources Report 1, Santa Fe National Forest. USDA, Forest Service, Santa Fe. Douglass, William Boone 1917 Notes on the Shrines of the Tewa and Other Pueblo Indians of New Mexico. Proceedings of the XIX Congress of Americanists, pp. 344-378. Washington, D.C. Fewkes, J. Walter 1904 Two Summers' Work in Pueblo Ruins. In 22nd Annual Report of the Bureau of American Ethnology, pp. 3-195. Government Printing Office, Washington, D.C. 1910 The Cave Dwellings of the Old and New Worlds. American Anthropologist 12:390- 416. 1913 Antiquities of the Upper Verde and Walnut Creek Valleys, Arizona. In 28th Annual Report of the Bureau of American Ethnology, pp. 181-220. Government Printing Office, Washington, D.C. Fliedner, Dietrich 1975 Pre-Spanish Pueblos in New Mexico. Annals of the Association of American Geographers 65:363-377. Hall, Susan D. 1992 An Architectural Analysis of Cavate Dwellings in the Verde Valley, Arizona. M.A. thesis, Northern Arizona University, Flagstaff. Harrington, John P. 1916 Ethno geography of the Tewa Indians. Bureau of American Ethnology, 29th Annual Report. Smithsonian Institution, Washington, D.C. Heiken, Grant 1979 Pyroclastic Flow Deposits. Scientist 67 (5):564-571. American Hendron, J. W. 1940 Prehistory of El Rito de los Frijoles: Bandelier National Monument. Technical Series 1. Southwestern Monuments Association, Coolidge, Ariz. 1943 Group M of the Cliff Dwellings, Rooms 1, 2, 3, 4, and 5, Caves 1, 2, 3, and 4, Frijoles Canyon, Bandelier National Monument, New Mexico. Ms. on file, Bandelier National Monument. 1946 We Found America's Oldest Tobacco. New Mexico Magazine 24 (11): 11-13, 33-37. Hewett, Edgar L. 1908 Excavations at Puye in 1907. Papers of the School of American Archaeology 4. Also published as Archaeology of Rio Grande Valley, Out West 31, no. 2 (August 1909). 1909a The Excavations at Tyuonyi, New Mexico, in 1908. American Anthropologist 11:434-455. 1909b Excavations at El Rito de los Frijoles in 1909. American Anthropologist 1 1:651 -673 . 1938 Pajarito Plateau and Its Ancient People. University of New Mexico Press, Albuquerque. Hill, James N., and W. Nicholas Trier weiler 1986 Prehistoric Response to Food Stress on the Pajarito Plateau, New Mexico: Technical Report and Results of the Pajarito Archaeological Research Project, 1977-1985. National Science Foundation Grant REFERENCES 287 BNS-78-08118. Department of Anthro- pology, University of California at Los Angeles. Hub bell, Lyndi, and Diane Tray lor 1982 Bandelier: Excavations in the Flood Pool ofCochiti Lake, New Mexico. National Park Service, Southwest Cultural Resources Center, Santa Fe. Hyland, Justin R. 1986 A Preliminary Investigation of a Prehistoric Cavate Site on the Pajarito Plateau, North Central New Mexico. Honors thesis, Department of Anthropology, University of California, Berkeley. Johnson, Chester 1960 Report on the Excavation of a Child's Burial in Cave Room C54, Tsankawi Ruin, Bandelier National Monument. Ms. on file at Bandelier National Monument. Kelley, V. C, E. H. Baltz, and R. A. Bailey 1961 Road Log: Jemez Mountains and Vicinity. In New Mexico Geological Society Twelfth Field Conference, pp. 47-60. New Mexico Bureau of Mines and Mineral Resources, Socorro. Kempe, David 1988 Living Underground: A History of Cave and Cliff Dwelling. The Herbert Press, London. Kent, Kate P. 1983a Prehistoric Textiles of the Southwest. University of New Mexico Press, Albuquerque. 1983b Pueblo Indian Textiles: A Living Tradition. School of American Research Press, Santa Fe. Kern, Willis F., and James R. Bland 1934 Solid Mensuration. John Wiley Sons, New York. and Kohler, Timothy A., and Angela R. Linse 1993 Papers on the Early Classic Period Prehistory of the Pajarito Plateau, New Mexico. Reports of Investigations 65. Department of Anthropology, Washington State University, Pullman. Lang, David M. 1966 The Georgians. Praeger, New York. Lang, Richard W. 1982 Transformation in White Ware Pottery of the Northern Rio Grande. In Southwestern Ceramics: A Comparative Overview, edited by A. H. Schroeder, pp. 153-200. Arizona Archaeologist 15. Arizona Archaeological Society, Phoenix. Lange, Charles H., and Carroll L. Riley 1966 The Southwestern Journals of Adolph F. Bandelier, 1880-1882. University of New Mexico Press, Albuquerque. Lange, Charles H., Carroll L. Riley, and Elizabeth M. Lange 1975 The Southwestern Journals of Adolph F. Bandelier, 1885-1888. University of New Mexico Press, Albuquerque. Lekson, Stephen H. 1988 The Idea of the Kiva in Anasazi Archaeology. The Kiva 53:213-234. Lent, Stephen C. 1988 Excavations of a Late Archaic Pit Structure (LA 51912), near Otowi, San Ildefonso Pueblo, New Mexico. Lab- oratory of Anthropology Notes 52. Museum 288 CAVATE STRUCTURES of New Mexico, Office of Archaeological Studies, Santa Fe. Lister, Robert H. 1939 Stabilization of Long House, Bandelier National Monument. Ms. on file, Bandelier National Monument. 1940a Stabilization of Frijoles Cave Ruins, 1940, Bandelier National Monument. Ms. on file, Bandelier National Monument. 1940b Stabilization of Cave Ruins, Bandelier National Monument, 1940. Ms. on file, Bandelier National Monument. McKenna, Peter J., and Robert P. Powers 1986 Preliminary Report on the 1985 Test Survey of Bandelier National Monument, New Mexico. Ms. on file, Intermountain Cultural Resource Center, Anthropology Program, National Park Service, Santa Fe. Magers, Pamela C. 1986 Weaving at Antelope House. In Archeological Investigations at Antelope House, compiled by D. P. Morris, pp. 224- 276. National Park Service Publications in Archeology 19. National Park Service, Washington, D.C. Mathien, F. Joan, Charlie R. Steen, and Craig D. Allen 1993 The Pajarito Plateau: A Bibliography. Southwest Regional Office, Division of Anthropology, Southwest Cultural Resources Center Professional Paper 49. Southwest Cultural Resources Center, Santa Fe. Maxon, James C. 1962 A Cave Kiva in Bandelier National Monument. Ms. on file, Intermountain Cultural Resource Center, Anthropology Program, National Park Service, Santa Fe. 1969 A Study of Two Prehistoric Pueblo Sites on the Pajarito Plateau of New Mexico. M. A. thesis, University of Wisconsin. Mera, Harry P. 1932 Style Trends of Pueblo Pottery in the Rio Grande and Little Colorado Culture Areas in the Sixteenth to Nineteenth Centuries. Laboratory of Anthropology Technical Series Bulletin 4. Laboratory of Anthropology, Santa Fe. 1934 A Survey of the Biscuit Ware Area in Northern New Mexico. Laboratory of Anthropology Technical Series Bulletin 6. Laboratory of Anthropology, Santa Fe. Mills, Barbara M. 1986 Temporal Variability in the Ceramic Assemblages of the Eastern Slope of the Black Range, New Mexico. In Mogollon Variability, edited by C. Benson and S. Upham, pp. 169-180. University Museum Occasional Papers 15. New Mexico State University, Las Cruces. Mindeleff, Cosmos 1896 Aboriginal Remains in Verde Valley, Arizona. In Thirteenth Annual Report of the Bureau of American Ethnology, by J. W. Powell, pp. 179-261. U.S. Government Printing Office, Washington, D.C. Morley, Sylvanus G. 1910 The South House, Puye\ Papers of the School of American Archaeology, O.S. 7. Archaeological Institute of America. Onstott, Thomas B. 1948 Excavations at Tyuonyi, Bandelier National Monument, 1947-48. Southwestern Monuments Library, Western Archaeological and Conservation Center, Tucson. Orcutt, Janet D. 1994 Chronology. In The Bandelier Archaeological Survey. Edited by Robert P. Powers and Janet D. Orcutt. Ms. on file, REFERENCES 289 Intermountain Cultural Resource Center, Anthropology Program, National Park Service, Santa Fe. Peckham, Stewart 1979 When Is a Rio Grande Kiva? In Collected Papers in Honor of Bertha Pauline Dutton, edited by A. H. Schroeder, pp. 55- 86. Papers of the Archaeological Society of New Mexico 4. Archaeological Society of New Mexico, Albuquerque. Powell, John Wesley 1886 Annual Report of the Director. In Fourth Annual Report of the Bureau of Ethnology, 1882-1883. U.S. Government Printing Office, Washington, D.C. 1891 Annual Report of the Director. In Seventh Annual Report of the Bureau of Ethnology, 1885-1886. U.S. Government Printing Office, Washington, D.C. Prudden, T. Mitchell 1903 Ruins of the San Juan Watershed in Utah, Colorado, and New Mexico. American Anthropologist n.s. 5:224-288. Riboud, Marc 1958 Cappadocia: Turkey's Country of Cones. National Geographic 113:122-146. Robinson, William J., John W. Hannah, and Bruce G. Harrill 1972 Tree-Ring Dates from New Mexico I, O, U: Central Rio Grande Area. Laboratory of Tree-Ring Research, University of Arizona, Tucson. Ross, C. S., R. L. Smith, and R. A. Bailey 1961 Outline of the Geology of the Jemez Mountains, New Mexico. In New Mexico Geological Society Twelfth Field Conference, pp. 139-143. New Mexico Bureau of Mines and Mineral Resources, Socorro. Powers, Robert P. 1988 Final Archeological Research Design for a Sample Inventory of Bandelier National Monument. Ms. on file, Intermountain Cultural Resource Center, Anthropology Program, National Park Service, Santa Fe. Preucel, Robert W. 1985 Preliminary Report of the Pajarito Field House Project: 1985 Season. Report prepared for the Santa Fe National Forest, Southwestern Division, Albuquerque. 1986a The Pajarito Field House Project. Archaeology at UCLA 2 (19). 1986b Preliminary Report of the Pajarito Field House Project: 1986 Season. Report prepared for the Santa Fe National Forest, Southwestern Division, Albuquerque. 1987 Settlement Succession on the Pajarito Plateau, New Mexico. The Kiva 53 (l):3-33. SAS Institute 1985 SAS User's Guide: Basics. Version 5 edition. SAS Institute, Cary, N.C. Schaafsma, Polly 1975 Rock Art in the Cochiti Reservoir District. Museum of New Mexico Papers in Anthropology 16. Museum of New Mexico, Santa Fe. 1980 Indian Rock Art of the Southwest. University of New Mexico Press, Albuquerque. Severy, Merle 1983 Gifts of Golden Byzantium. Geographic 164:722-737. National Siegel, Sidney 1956 Nonparametric Behavioral Sciences. York. Statistics for the McGraw-Hill, New 290 CAVATE STRUCTURES Smiley, Terah L. 1951 A Summary of Tree-Ring Dates from Some Southwestern Archaeological Sites. Laboratory of Tree-Ring Research Bulletin 5. University of Arizona, Tucson. Smiley, Terah L., Stanley Stubbs, and Bryant Bannister 1953 A Foundation for the Dating of Some Late Archaeological Sites in the Rio Grande Area, New Mexico: Based on Tree-Ring Methods and Pottery Analyses. Laboratory of Tree-Ring Research Bulletin 6. University of Arizona, Tucson. Smith, Watson 1952 Excavations in Big Hawk Valley, Wupatki National Monument, Arizona. Museum of Northern Arizona Bulletin 24. Museum of Northern Arizona, Flagstaff. 1972 Prehistoric Kivas of Antelope Mesa, Northeastern Arizona. Papers of the Peabody Museum of Archaeology and Ethnology 39, 1. Harvard University, Cambridge. Snow, David H. 1974 The Excavation of Saltbush Pueblo, Bandelier National Monument, New Mexico. Laboratory of Anthropology Note 97. Museum of New Mexico, Santa Fe. 1982 The Rio Grande Glaze, Matte-Paint, and Plainware Tradition. In Southwestern Ceramics: A Comparative Overview, edited by A. H. Schroeder, pp. 235-278. Arizona Archaeologist 15. Arizona Archaeological Society, Phoenix. Steen, Charlie R. 1977 Pajarito Plateau Archaeological Survey and Excavations. Los Alamos Scientific Laboratory LASL-77-4. 1979 The Mortendad Style of Rock Art, Pajarito Plateau, New Mexico. In Collected Papers in Honor of Bertha Pauline Dutton, edited by A. H. Schroeder, pp. 41-53. Papers of the Archaeological Society of New Mexico 4. Archaeological Society of New Mexico, Albuquerque. 1982 Pajarito Plateau Archaeological Surveys and Excavations, II. Los Alamos Scientific Laboratory LA-8860-NERP. Available from National Technical Information Service, U.S. Department of Commerce, Springfield, Va. Stuart, David E., and Rory P. Gauthier 1981 Prehistoric New Mexico: Background for Survey. New Mexico Historic Preservation Bureau, Santa Fe. Thompson, Raymond H., editor 1990 When Is a Kiva? and Other Questions about Southwestern Archaeology by Watson Smith. University of Arizona Press, Tucson. Turney, John F. 1948 An Analysis of the Material Taken from a Section of Group M of the Cliffs, Frijoles Canyon, Bandelier National Monument, New Mexico, 1943. M.A. thesis, Adams State College, Alamosa, Colo. Van Zandt, Tineke R. 1994 Architecture and Site Structure. In The Bandelier Archeological Survey. Edited by Robert P. Powers and Janet D. Orcutt. Ms. on file, Intermountain Cultural Resource Center, Anthropology Program, National Park Service. Warren, A. Helene 1979 Ceramic Studies in Cochiti Reservoir, 1976-1977. In Archaeological Investigations in Cochiti Reservoir, New Mexico, vol. 3, edited by J. Biella, pp. 27-42. Office of Contract Archaeology, University of New Mexico, Albuquerque. REFERENCES 291 White, G. E. Whittaker, John C. 1904 The Cavate Dwellings of Cappadocia. n.d. J. W. Fewkes's Photographs of New Records of the Past 3:66-73. Caves in 1900. Ms. submitted to Kiva. Index adobe collar, 149 adobe ramp, 142. See also mealing complexes Alamo Canyon, 5 American Anthropologist, 8 archaeomagnetic dating, 62 artifacts and collections, 9, 11, 85, 97, 236 Awanyu, 8, 51; illustrated, 192, 194 axe groove (axe-sharpening groove), 140, 149, 235 back wall features, 84, 161, and exterior openings, 122; in noncavates, 151, 157 Bailey, R. A., R. L. Smith, and C. S. Ross, 2, 15, 16, 53 Bandelier, Adolph F., 5-6, 216 Bandelier Black-on-gray, 11, 17 Bandelier National Monument, 1, 2, 11, 17. See also Frijoles Canyon; Tsankawi Bandelier Tuff, 3, 9, 15, 16. See also tuff types Barthuli, K., and S. Hall, 49 Barthuli, K., J. Vint, and W. Bustard, 47 Beam, George L., 1, 7-8, 13 beam features, as functional, 201; holes for beams/vigas (beam seat), 107, 157,165, 182, 233; niches and beams, 203. See also ceilings beam seat, 165, 182, 233; viga holes, 107, 157. See also beam features Bierbower, Susan, 1, 7 Big Kiva, 29, 63 bins, plank slots for, 157 birds, illustrated designs for, 193 Biscuit B ceramic, 64 Blair, Jonathan S., 4, 94, 138 boundary, cultural, 217; cultural context, 16,59 Bretemitz, David A., 67 Bureau of American Ethnology, 6 burials, 8, 11, 53 burns, floor, 138; firepits, 138, 201, and deflectors, 150; hearths, 51, 62 C-14 Sample, 64 Camp Hamilton, 1 1 Canada de Cochiti, 15 Canby, Thomas Y., 13 Cappadocia, Turkey, 4, 94, 134, 138 Carlson, Ingrid K., and T. A. Kohler, 9, 13, 51, 126, 157 cavate type, 85; definitions, 1-2, 17, 78, and functional categories, 13; filled cavate, 84, (number of) 108; partial cavate, 84, (number of) 108; variation in form, 77. See also chambers Cay wood, Louis R., 45 ceilings, 134, 233; vertical holes in, 61, 185, 235; viga holes in, 157, 161. See also beam features; plastering; roofing ceramic patterns, 72-75. See also ceramic sample ceramic sample, 10, 68-70, 73, 216; surface sherd counts, 9, 64; Turney's types, 9. See also ceramic patterns; ceramics ceramics, dating, 61, 63, for Group M, 45; form distributions, 72, 74, 75, and site function, 75; rock art dating with 293 294 CAVATE STRUCTURES ceramics, 196. See also ceramic patterns Ceremonial Cave, 27, 61 Cerro Pedernal, 2 Chama River, 15, as Rio Chama, 216 chamber corner, 132, 233, 234 chamber location, 80, 90, 231, 233, 234; levels (stories), 83, 98; rooms up/rooms down, 82. See also chambers chambers, defining, 84, 108, and back chambers, 150; combined chambers, 138, and expanded chambers, 59, 83; comparisons, 59, 117-20; functions, 13, 112, 116, 213, and functional clustering, 208-11; recording location, 82-83, and shape and size, 80-81, 91-93, 112, 208, 226, 229, 232, 239; volume as functional, 201. See also cavate type; chamber location; and rooms separately Chapman, Kenneth M., 8, 17, 185, 190, 197; survey of rock art by, 8, 273 chronology, 5, 61, and dating with cei amies, 45, 61, 63, 169; Classic Period, 17, 62, 67, 71; Coalition Period, 17, 45, 62, 67, 71, 71-72; Developmental Period, 71; Early Classic, 216; Late Coalition, 216; Pueblo III, 5. See also occupation Civilian Conservation Corps (CCC), 45 Classic Period (Rio Grande Classic), 17, 62, 67,71 cliff niche, 84, 190, 235 cluster analysis, 206-11, 275 Coalition Period, 17, 45, 62, 67, 71, 71-72 Cochiti Pueblo, 2 collections, 85, and artifacts, 9, 11, 97, 236 Colton, Harold S., 5 combined chamber, 138; expanded chambers, 59, 83 compass location point, 138, 235 computers, recording with (direct data entry), 87, and data manipulation, 90, 93 condition/damage, 84, 99-105 construction and use, 94-96; recording evidence for, 81, 226, 233; reoccupation, 11, 74, and historic use, 7, 9, 71, 74. See also occupation Cordell, Linda S., 17 Corral Hill, 71 cotton, 216; weaving, 215. See also loom features Crowder, Bill, 86, 93 Crowder, June, 190, 191; rock art study by, 197, 267-72 Crowder, June, and Bill Crowder, 85, 93 Cuevitas Arribas, 20, 61 culinary ware, 72, 74, 75 cultural context, 16, and comparisons, 59; cultural boundaries, 217 dado, incised, 189, 233; illustrated, 153, 192 damage, 84, 99-105, 231 data manipulation, 90 deflector, 150, 201, 234; illustrated, 152 Delight Makers, The, Bandelier, 6 Developmental Period, 71 Die Koshare, Bandelier, 6 digging sticks, 94 direct data entry, 87 doors, as functional features, 201, 229; exterior, 116, and exterior openings not doors, 122; interior, 122; niches as false doors, 190; passages, 138 Dougherty, Julia D., 2 Douglass, William Boone, 9 Early Classic, 216 economics, cotton in the, 216; the growing zone, 16 Ellis, F. H., 65 environment, 15 ethnicity, canyon and mesa, 59; cultural boundary, 217; cultural context, 16 exterior door, 116. See also doors exterior opening, 122, 233. See also doors fauna, 9, 45, 99, 227 feature co-occurance, 201 feature fraction, 81, 231 feature types, 84, 228, 231, 233, 235, and canyon/mesa comparisons, 61. See also feature types separately INDEX 295 Fewkes, J. Walter, 1-2, 4, 5, 9 field houses, cavates as, 13, 215 field time, 88-90 FIELDLA, 90 fill, 96, 98, 226; incised dado and, 190; filled cavate, 84, 108 filled cavate, 84, and number of, 108 firepits, 138, 201, and deflectors, 150; floor burns, 138, 201; hearths, 51, 62 Flagstaff, Arizona, 4, 5 Fliedner, Dietrich, 2, 215 floor burn, 138, 201; firepits, 138, 201, and deflectors, 150; hearths, 51, 62 floor depression, 140, 234 floor features, 138-50, as functional, 201; floors, 134. See also floor features separately floor pit, 201 floor pit complex, 140, 235 floor ridge, 142, 201, and slots, 157 floors, 134; floor features, 138-50, 201 Four Corners Area, cavates of the, 4 Frijoles Canyon, 14, 17-18, and comparisons, 59, 112, 116, 157, 197, 217; Bandelier at, 5; Hewett at, 8; Lister at, 11; Stevenson at, 6. See also Group A; Group F; Group I; Group M; and see sites separately Frijolito site, 63 140, 142, 152, 161, 185, 188, 189, and comparisons, 59, 61; other collections for, 85 Group B, 71, 74; unrecorded rooms at, 132 Group C, 71, 74 Group D (Long House), 8, 11, 20, 216 Group E (Snake Village and Sun House), 8 Group F, ceramics and dating for, 62, 65, 71, 74; identification for, 20, and setting, 29-39, 82, 216; rock art at, 192, 197, 268; room plans, 33-35, 86, 91, 94, and features, 94, 122, 132, 134, 138, 149, 150, 161, 185, 189; unrecorded rooms for, 132 Group I, ceramics and dating for, 62, 65, 71, 74, 75; identification for, 20, and setting, 21, 37, 45^6, 82, 84, 216; rock art at, 197, 269, 272; room plans, 41, 43, 86, 91, 94, and features, 45, 122, 132, 134,149, 150 Group J, mentioned, 37 Group M, ceramics and dating for, 62, 63, 65, 67, 71, 74; excavation at, 9, 13, 51, and collections, 85; identification for, 20, and setting, 21, 45, 51, 52, 216; rock art at, 51, 192, 197, 270; room plans, 47, 49, 78, 84, 86, 94, and features, 51, 122, 126, 132, 138, 142, 149, 157, 189, and comparisons, 61 Garcia Canyon, 13, 62, 105, 116 Gauthier, R., 62 geology, 15, and topography compared, 59 geometric fraction, 81, 230 Glaze C ceramic, 71 Glaze D polychrome ceramic, 71 Glaze E ceramic, 71 Greene, Ed, 52 groove, 233; axe-sharpening, 140; wall groove, 185, 201 group attributes, 93-94 Group A, ceramics and dating for, 62, 64- 65, 71; identification for, 20, and setting, 20-28, 82, 84, 216, and deterioration, 28-29; rock art at, 192, 197, 267, 272; room plans, 22-25, 86, 91, 94, and features, 29, 122, 126, 132, 134, 138, habitation rooms, 13, 206, 213, 227, and comparisons, 112 Hall, Susan, 2, 4 hand-and/or-toe holds, 84, 108, 189, 233; routes to canyon rim, 45, 52; stairway, 59; steps, 149; trails, 84 Harrington, John P., 15 Harrington, M. R., 9 Havasupai, the, 5 Head, G., and A. Prieto, 41, 43 hearth, 51, 62; firepits, 138, 150, 201; floor burns, 138, 201 Heiken, Grant, 4, 15, 16, 134 Hendron, J. W., 5, 8, 9, 11, 13, 15, 45, 51, 62; excavation by, 63 Herr, S., and R. Powers, 57 Hewett, Edgar L., 1, 8, 14, 15, 17, 59, 77, 296 CAVATE STRUCTURES 80, 142, 145, 149, 182, 185, 190, 214, 216 Hill, James N., and W. Nicholas Trierweiler, 13, 72 historic use of cavates, 7, 9, 71, 74 hole, indeterminate, 108, 165, 174, 179, 182, 201; holes and niches, 203. See also separately by hole type holes for pegs, 182 holes in walls, multivariate analyses of, 174 Hopi, the, 149, 214-16 housekeeping, 94 Hubbell, Lyndi, and Diane Traylor, 16 Hyland, Justin R., 13, 62, 64, 116 incised dado, 189, 233; illustrated, 153, 192 incisions, narrow wall, 188, 201, 235 indeterminate holes, 108, 165, 179, 182, 201, 233; at the line of plastering, 174 interior door, 122. See also doors irrigation, 28, 61 isolated room (1-36), 37 Jemez Mountains, 2, 15 Jemez River, 2, 215 Johnson, Chester, 11, 53 Kapo Black ceramic, 71 Kelley, V. C, E. H. Baltz, and R. A. Bailey, 16 Kempe, David, 2, 4 Kent, Kate P., 145, 149, 182, 215, 216 Keres, the, 5, 14, 59, 217 Kern, Willis F., and James R. Bland, 91, 92 Kidder, A. V., 145 kiva, 8, 12, 27-28, 51, 116, 206, 213-15, 226; room categorized as, 13, 210, 211, 227 Kohler, Timothy A., 13, 45 Kohler, Timothy A., and Angela R. Linse, 17, 45, 52 LA 211 (Tsankawi Pueblo), 16, 17, 19, 20, 52, 71, 72, 74, 216 LA 217 (Rainbow House), 45, 52, 63, 216 LA 4997 (Saltbush Pueblo), 45, 63 LA 50020. See Group M LA 50021. See Group A LA 50022. See Group I LA 50023. See Group F LA 50024. See Tsankawi LA 50909, ceramic dating for, 71 LA 50972. See Group M LA 50973. See Group A LA 50974. See Group I LA 50975. See Group F LA 50976 (Tsankawi cavate group), 19, 53, 216. See also Tsankawi LA 52333, excavation at, 13 LA 60550 (Tyuonyi Annex), 216 Laboratory of Anthropology, 9, 20; numbering by the, 90 Lang, Richard W., 4, 65 Lange, Charles H., and Carroll L. Riley, 5, 6 Lange, Charles H., Carroll L. Riley, and Elizabeth M. Lange, 6, 15 large floor-level niche, 150 Late Coalition Period, 216 latilla hole, 162, 182, 201 Lekson, Stephen H., 214 Lent, Stephen C, 16 levels (stories), 83, 98; multilevel rooms, 94; Tsankawi, 112 linguistics, 217 lintels, 182 Lister, Robert H., 11, 14, 28, 29, 37, 45, 52, 53, 59, 86, 162 Long House (Group D), 8, 11, 20, 216 loom anchors, 144, 232, and upper loom supports, 182, illustrated, 192; kivas with loom features, 214-15, and comparisons, 215-16. See also loom features loom features, 201, 203, 206; loom anchors, 144, 182, 214-16 loom support, upper, 182, illustrated, 192. See also loom anchors Los Alamos Archaeological Society, 12 Los Alamos Canyon, 52 Los Alamos National Laboratory, 12 McKenna, Peter J., 28, 61; ceramic analysis by, 64, 216 INDEX 297 McKenna, Peter J., and Robert P. Powers, 64,65 Magers, Pamela C, 214 Mancos River, 4, 5 mapping, 86, 90-91 masonry, back wall, 84; comparisons of, 59; partial cavate, 84; presence and type of, 97, 132, 226, 234, for functional analysis, 201; exterior masonry rooms, 140; the use of, 94 masonry and tuff wall, 132 masonry wall, 132; masonry and natural wall, 234 Mathien, F. Joan, Charlie R. Steen, and Craig D. Allen, 2, 15, 16, 17 Maxon, James C, 9, 11-12, 196 mealing complexes, adobe ramp in, 142; floor ridges and, 142; metate rest in, 142; wall depression and, 185 Mera, Harry P., 11 metate rest, 142, 234; other mealing features, 142, 185 Mills, Barbara M., 67, 185 Mindeleff, Cosmos, 1, 4 Mindeleff Cavate Site, 4 Morley, Sylvanus G., 8-9 Mortandad Canyon, 11 Mortandad style of rock art, 196 narrow wall incisions, 188, 235 National Geographic, 13 National Park Service, 14; database for the, 93, 237, 240 natural features, 97 natural wall, 122, 126 Navajo National Monument, 53 Navawi site, 17 New Cave, Arizona, 5 NEWLANO, 90 niche, cliff, 53, 84, 190. See also niches niche, deep modified, 13. See also niches niche, large floor-level, 150, illustrated, 153. See also niches niches, as functional features, 201, and feature correlations with, 203; back wall niche, 151; cliff niche, 53, 84, 190; comparisons of types, 157; deep modified, 13; pairing of, 157, illustrated, 153; wall niche, 152, 157 niche, wall, 152, 157; back wall niche, 151. See also niches noncavates, back wall niches and, 151; back wall slots and, 157; connected with cavates, 84; data sets for, 83-84; exterior openings and, 122; features for, 78, and distribution, 85; nonhuman users of, 99, 100, 227; viga holes in, 161 occupation, a sequence for, 62, 216; ceramic dating and, 67, 74; duration of, 62, 206; Group F patterns of, 71; remodeling, 94, 157; reoccupation, 11, 74, and historic use, 7, 9, 71, 74 Old Caves, Arizona, 5 Old World, cavates of the, 2, 4; in Turkey, 4, 94, 134, 138 Onstott, Thomas B., 75 note 2 openings, as doors, 201; exterior, 116, 122; interior, 122; passages between chambers, 138. See also Doors Orcutt, Janet D., 67 Otowi, 11, 16, 17 overhang, 234; viga holes in cavate, 157 Pajarito Archaeological Research Project (PARP), 13, 72 Pajarito Plateau, described, 1, 3, 5-8, 15; past work on the, 2-13; settlement pattern for the, 216 Panowski, Bruce, 90 Panowski Holes, 61, 185, 235 partial cavate, 84. See also noncavates partitions, vertical room, 185 passage, 138, 234. See also doors Peckham, Stuart, 9, 145, 149, 213, 214 Pecos Pueblo, 145 pegs, holes for, 174, 182 Peralta Canyon, 2 petroglyphs, 12, 233, 234, 273; cliff face, 272; defined, 190; Group A, 28; Group F, 29; Group I, 45; Tsankawi, 53 photography, recording with, 86, 93, 244 pictograph, defined, 190 298 CAVATE STRUCTURES pit complex, floor, 140 pithouse, preceramic, 16 pits, subfloor, 140 plane shapes, 91, 232, 234 planks, 142, 157 plaster and niches, 203 plastering, 94, 126-32, 206, 230; as functional feature, 201; colors of, 126, 132, 233, 235; fine-line scratching in, 190; floor features and, 203; indeterminate holes at the line of, 174, 182; niches and, 150, 157; replastering, 94; smoking and, 126. See also floor features population, aggregated, 17, 62; elevation and, 72 possible latilla hole, 162, 182, 201 possible upper loom support, 182 postholes, 140, 235 pot rests (floor depressions), 140 Powell, John Wesley, Major, 1, 4, 6-7 Powers, R., and T. Chadderdon, 55 Powers, Robert P., 16, 17, 45, 61 Preucel, Robert W., 2, 13, 17, 105, 215, 216 Prudden, T. Mitchell, 5 Pueblo III, Late, 5 Pueblo Canyon, 1 1 Pueblo Revolt, 1 1 pueblos, cavate relationship to surface, 216 Puye, 13, 17; dating, 62, 63; described, 77; early visitors to, 6-9 Rainbow House (LA 217), 45, 52, 63, 216 remodeling, 94; combined chambers, 138; expanded chambers, 83; one from two chambers, 59; viga holes and, 157. See also occupation retaining wall, Group F, 29 Riboud, Marc, 4, 94 Rio Chama, 216, as Chama River, 15 Rio Grande, 2, 4, 15, 16, 213 Rio Grande (Ceramic) Series, 65, 66, 74 Rio Grande Classic, 17, as Classic Period, 62, 67, 71 Rito de los Frijoles, 11, 15 Robinson, William, 62, 63 Robinson, William J., John W. Hannah, and Bruce G. Harrill, 62, 67 rock art, as a functional feature, 201, with other features, 203; Chapman report on, 8, 273; comparisons, 61, 192, 197, and discussed, 190; Crowder Study on, 197, 267-72; Mortandad Style of, 12, 12-13, 196; motifs for, 198, 233, and nomenclature, 199; recording rock art, 85; rooms with notable, (Group M) 51, (Tsankawi) 53, 61, and others, 193, 194,210,211,273 roofing, beam features for, 201; viga holes, 157; wall ledges, 185. See also ceilings Room A- 10, ceiling construction marks, 82; photograph, 32 Room A-13, photographs, 30; rock art, 273 Room A-47, features illustrated, 142, 153; pot rests, 140 Room A-60, photographs, 31 Room F-2, rock art, 273 Room F-23, features illustrated, 193 Room F-31, photograhs, 38 Room F-37, features illustrated, 151 Room F-38, rock art, 273 Room F-47, step, 149 Room 1-15, mentioned, 37 Room 1-19, features, 211; kiva-like, 211; rock art, 273 Room 1-22, photographs, 46 Room 1-36, setting, 37 Room M-10, expansion, 83 Room M-13, kiva-like, 211; rock art, 194 Room M-33, rock art, 51, 273 Room M-40, features illustrated, 145 Room M-44, features illustrated, 153 Room M-59, features, 145, illustrated, 148, 162 Room M-60, masonry, 94; mealing activity, 142, 185, illustrated, 146; rock art, 273, llustrated, 194 Room of the Cacique, 5 Room TS-20, kiva-like, 210 Room TS-24, features illustrated, 189 Room TS-25, features illustrated, 190 Room TS-36, volume, 112 Room TS-40, features illustrated, 193 INDEX 299 Room TS-53, photographs, 60 Room TS-55, feature illustrated, 144 Room TS-59, features, 112, 140, illustrated, 148; kiva-like, 210; rock art, 196, illustrated, 192 Room TS-64, features, 206; kiva-like, 210 Room TS-65, features, 149 Room TS-66, features, 112; kiva-like, 210 Room 12, Group F, photographs, 39 Room 15, Group F, features illustrated, 152 rooms. See chambers; habitation rooms; kivas; noncavates; storage rooms; and see groups separately rooms up/rooms down, 82, 226 Ross, C. S., R. L. Smith, and R. A. Bailey, 2, 15 routes to canyon rim, 45, 52; hand and/or toe holds, 84, 108, 189; stairway, 59; steps, 149; trails, 84 SAS Institute, 93 Saltbush Pueblo (LA 4997), 45, 63 San Juan River, 4, 5 San Lazaro Glaze-on-polychrome, 65 Sandia Canyon, 52, 59 Sankawi Black-on-cream, 52, 67, 71 Santa Clara, the, 7 Santa Fe, New Mexico, 93 Santa Fe Black-on-white, 12, 13, 17, 62, 71 Scaffold House, Navajo National Monument, 53 Schaafsma, C, 87 Schaafsma, Polly, 190, 191, 197 seasonality, 215 settlement pattern, 216. See also occupation Severy, Merle, 4 shapes, 80, 232; plane shapes, 91, 234; solid shapes, 91, 233, 234 Siegel, Sidney, 202 slots, 157 Smiley, Terah L., 67 Smiley, Terah L., Stanley Stubbs, and Bryant Bannister, 62, 63 Smith, Watson, 145, 213, 214 smokehole, 182, 201 smoking, 206 Snake Village (Group E), 8 Snead, J., 33 Snead, J., and H. Newman, 35 Snow, David H., 45, 63 solid shapes, 91, 232, 233, 234 Southwest, American, 4 Southwest Museum, 9 stability, 227, 245, 250, 253; table for, 100. See also condition/damage stairway, 59; hand and/or toe holds, 84, 108, 189; route to canyon rim, 45, 52; steps, 149; trails, 84 Steen, Charlie R., 12, 12-13, 14, 59, 134, 196, 215 steps, 149, 233. See also stairway Stevenson, James, 6 storage rooms, defining, 13, 208, 211, 213, 227, and comparing, 112, 116, 144; milling areas and, 144; plaster and smoking in, 134, 206 stories (levels), multilevel rooms, 94; recording levels, 83, 98; Tsankawi, 1 12 structural features, 108, 116, 122, 132 Stuart, David E., and Rory P. Gauthier, 2, 16,62 subfloor pit, 140 suites of cavates, 4 Sun House (Group E), 8 surface pueblos, cavate use and relationship to large, 216 Sweetland, Bill, 28 tent rocks, 29, 51 terraces, 140 Tesuque (ceramic) series, 74; Tesuque corrugated, 13 Tewa, the, 5, 9, 14, 15, 52, 59, 217; rock art style and, 197 Tewa polychrome, 71 Thompson, Raymond H., 213 trails, 84; hand and/or toe holds, 84, 108, 189; routes to canyon rim, 45, 52; stairway, 59; steps, 149 tree-ring dating, 9, 62-63; for Group M, 62 Tsankawi, 11, 12, 14; ceramics and dating for, 62, 65, 67, 72; collections for, 85; comparisons, 60, 72, 112, and cultural separation of, 217; features for, 82, 84, 300 CAVATE STRUCTURES 116, 122, 132, 134, 138,140, 142, 145, 149, 150, 157, 162, 185, 188, 190, 206; kivas at, 215; rock art at, 192, 196, 197, 271, 272; setting of, 52-59, 61, 216, and room plans, 55, 57, 86, 91, 94, 108, 112, 116,208 Tsankawi Black-on-cream, 17 Tsankawi Mesa, 17, 52 Tsankawi Pueblo (LA 211), 16, 17, 19, 20, 52, 71, 72, 74, 216 Tsankawi Pumice Bed, 16, 53 Tshirege Cave Site, 12, 17, 63 Tsiping Ruin, 2 tuff type, 82, 97, 99, 227, 233, 235; Bandelier Tuff, 3, 9, 15, 16; patina and smoking on tuff, 134 Turkey, comparisons to Cappadocia, 4, 94, 134, 138 Turkey Tank Cave, Arizona, 5 Turney, John F., 9-11, 51, 71 Tyuonyi Annex (LA 60550), 216 Tyuonyi Pueblo, 5, 17, 29, 63, 216 United States Geological Survey (USGS), 20 Universal Transverse Mercator (UTM) grid, 90-91 University of California, 13 University of New Mexico, 11 Unshagi site, 215 Valle Grande Caldera, 3, as Valles Caldera, 15 Van Zandt, Tineke R., 75 note 2 vents, 185, 203; as functional features, 201 Verde Valley, Arizona, 4 vertical ceiling hole, 61, 185, 235 video recording, 86, 88, 244 viga holes, 107, 157, and beam seats, 165; as beam features, 201. See also beam features visitation, 101; at Group A, 28. See also damage volumes and areas, 91, 239; volume and niches, 203 wall depression, 185, 234 wall features, 150 wall ledges, 185, 234, as beam features, 201 wall niche, 152, 157 wall, masonry, 132; masonry and tuff, 132; natural wall, 122, 126 walls, as functional features, 201; back walls, 84, 122, 151, 157, 161; depression in, 185; features for, 150; ledges in, 185, 201; masonry of, 132, and natural, 122, 126; niches in, 152, 157; recording, 230 Washington State University, 13 weaving, 215. See also loom features White, G. E., 1, 134 White Rock, New Mexico, 12, 15 Whittaker, John C, 5 Wiyo Black-on-white, 12, 71 wood, features made of, 144 Yapashi site, 72 oU.S. GOVERNMENT PRINTING OFFICE: 1995-676-953/25132 Clemson University 3 1604 01 357 938 FEDERAL PUBLICATION DATE DUE DEMCO, INC. 38-2931 ■ ^B