Chapter 9

Summary and Conclusions

by Kristin A. Kuckelman

The Crow Canyon Archaeological Center's investigations at Sand Canyon Pueblo (Site 5MT765) have greatly enriched our understanding of the ancestral Pueblo people who constructed and occupied villages in the central Mesa Verde region in the latter half of the A.D. 1200s. This research has been a central focus of Crow Canyon's long-term study of continuity and change in ancient Pueblo society during the final centuries of regional occupation. Many of the previous excavations at sites that dated from this time in the region were conducted at cliff dwellings in Mesa Verde National Park (e.g., Cattanach 1980*1; Fewkes 1909*1, 1911*1; Nordenskiöld 1979*1; Rohn 1971*1); significant research had also been conducted at the thirteenth-century sites in Hovenweep National Monument (Hammett and Olsen 1984*1; Winter 1975*1, 1976*1, 1977*1, 1978*1).

Crow Canyon's excavations at Sand Canyon Pueblo were one part of the Sand Canyon Archaeological Project (Lipe 1992*1), which included large-area survey (Adler 1990*1, 1992*3), excavations at 15 smaller sites in the Sand Canyon locality (Huber 1993*1; Huber and Lipe 1992*1; Kuckelman 2000*1; Varien 1999*1, 1999*7), and many special studies (see the Sand Canyon Pueblo Bibliography). Numerous synthetic and topical works that focus in whole or in part on Crow Canyon's work in the Upper Sand Canyon community and at Sand Canyon Pueblo itself have also been published (Bradley 1993*1, 1996*1; Kuckelman et al. 2000*2; Lipe 1992*2, 2002*1; Lipe and Ortman 2000*1; Ortman and Bradley 2002*1; Varien 1999*1, 1999*7).

The research design that guided work at Sand Canyon Pueblo is discussed in Chapter 2. Research at this site was designed to (1) sample the architectural variability across the site that was observed as differential room-to-kiva ratios across blocks, (2) sample public architecture, (3) obtain a probabilistic sample of artifacts from nonarchitectural areas of the site, and (4) test numerous kivas in areas of the site that were not intensively excavated. These data would shed light on the chronology and use of the site, on possible functional, economic, or social differentiation within the village, and on whether the site served as a political, religious, or economic center for the surrounding Sand Canyon community (Bradley 1992*2:79). Studies of community and regional organization would focus on the dimensions of scale, differentiation, integration, and intensification (Lipe 1992*1, 1992*3; Lipe and Bradley 1986*1).

In the chapters of this volume, the data resulting from these excavations have been used to determine the uses of architectural blocks, suites of structures, and public architecture in the village, as well as the times and sequences of construction. Many inferences that address the research goals, particularly those specific to individual structures, are contained in The Sand Canyon Pueblo Database. Discussions at the level of the architectural block and kiva suite can be found in Chapter 4. In this final chapter, I address research questions and goals, as well as other topics of interest, both at an interblock level and for the village as a whole, drawing from all excavation and analysis data available for the site to date. Findings regarding when, how, and why structures and suites were abandoned and the village depopulated are drawn from abandonment-context data as well as from comparisons of abandonment-context assemblages with midden-context assemblages. These findings are discussed below under "Depopulation of the Village."

Village Chronology

Beginning in the mid-1200s, hundreds of Pueblo people constructed what we now refer to as Sand Canyon Pueblo—a large village situated on and just below the canyon rim of a small side canyon near the head of Sand Canyon. Many families probably relocated from scattered, upland farmsteads within the Sand Canyon community. However, some who came to reside at Sand Canyon might have come from outside the area (Chapter 3). The villagers built many residences that grew through accretion, and they also cooperatively planned and constructed public architecture (buildings that were not ordinary domestic structures), such as a great kiva, a D-shaped bi-wall building, and a massive, masonry wall that enclosed much of the village. In all locations where a room or kiva abutted this wall (Database Map 4001), it would have been a minimum of one story tall—and thus had significant defensive potential.

Tree-ring dates indicate that construction of the village began in the late A.D. 1240s or early 1250s and ended in the late 1270s. Although many timbers from this site were cut in the early A.D. 1200s, the preponderance of evidence suggests that construction of the village did not begin until the late 1240s or early 1250s. The timbers dating from the initial decades of the thirteenth century were probably salvaged from nearby abandoned farmsteads (Varien 1999*1, 1999*7). Some residence groups probably brought roofing timbers from their own farmsteads to build new homes in the village.

The spatial distribution of tree-ring dates for Sand Canyon Pueblo does not suggest a particular pattern of growth within the village, although the great kiva (Great Kiva 800; see Database Map 4281) and two circular towers, Tower 212 (Database Map 4055) and Tower 1203 (Database Map 4159), were constructed before the buildings adjacent to them, and they might have been among the earliest structures built in the village. Kiva Suite 501—which appears to have been constructed in the early A.D. 1250s (Chapter 4, paragraph 77)—might have been constructed near the time the village was founded. Any or all of these structures could predate the village-enclosing wall, which was built in large sections, probably during the A.D. 1250s or 1260s, or both.

That the village was occupied during only the middle and late A.D. 1200s is indicated by the dominance of Mesa Verde Black-on-white sherds in the pottery assemblage. The tree-ring evidence of this fairly short occupation span is corroborated by the presence of only minor architectural remodeling and the fact that few structures contain secondary refuse. The latest tree-ring date available for Sand Canyon Pueblo is A.D. 1277vv. Because this date is among the latest for the region as a whole (Robinson and Cameron 1991*1; Varien et al. 2007*1) it is likely that the village was depopulated within a few years of that date.

Uses of Structures and Kiva Suites

At the outset of excavations at Sand Canyon Pueblo, indications on the modern ground surface suggested that the proportion of rooms to kivas was not uniform among architectural blocks. This gave rise to a theory that blocks with different ratios of rooms to kivas were used for different purposes. Therefore, after initial mapping of the site in 1983 (Database Map 4316), 14 large, more-or-less discrete clusters of structures—called architectural blocks—were categorized on the basis of their room-to-kiva ratios as assessed on the modern ground surface (two additional blocks were designated later, for a total of 16). The typical ratio of rooms to kivas at residential sites in this region is four to 10 rooms per ordinary-size kiva (Lightfoot 1994*1; Lipe 1989*1:Table 1), although a ratio of five to seven rooms per kiva might be a more accurate estimate for Pueblo III habitations that were not constructed in cliff overhangs (Lipe and Ortman 2000*1:113).

The site-wide ratio for Sand Canyon Pueblo was estimated at five rooms per kiva (Bradley 1992*2:Table 7.1). Individual blocks as defined were designated as "kiva dominated" (fewer than four rooms per kiva), "standard" (five to 16 rooms per kiva), or "room dominated"; the single block assigned to this last category (Block 300) was estimated, from indications on the modern ground surface, to have more than 20 rooms per kiva (Bradley 1992*2:80–81). At least one cluster of structures in an architectural block assigned to each of these categories was then excavated intensively.

To some extent, how clusters of structures and kiva suites were used can be inferred from architectural characteristics; associated features, artifacts, and ecofacts constitute vital evidence as well. At Sand Canyon Pueblo, assemblages of artifacts left on floors, roofs, and outdoor surfaces are probably atypical of the assemblages that were present in those locations during most of the occupation of the village. That is, the assemblages are likely to have been modified by the violent events and other anomalous circumstances—as indicated primarily by the human remains in abandonment contexts—that prompted village depopulation (see "Depopulation of the Village," below, and Kuckelman [2007*1]). Therefore, it is the contents of middens that most accurately reflect how suites of structures were typically used during the occupation of the village.

The contents of midden assemblages from intensively excavated "kiva-dominated" (Blocks 100, 200, and 500) and "standard" (Blocks 1000 and 1200) architectural blocks indicate that these structures and suites were residences. Some researchers have proposed that the low ratio of rooms to kivas in kiva-dominated blocks reflects small households or households that were provisioned from central food stores (Lipe 2002*1; Lipe and Ortman 2000*1; Lipe and Varien 1999*1); however, Lipe and Ortman (2000*1:115) point out that many small Pueblo III sites with low room-to-kiva ratios do not exhibit any other evidence of special activity or status. This finding is corroborated by a study of the space-syntax variable of integration of the architectural suites excavated at Sand Canyon, indicating that these structures were residential rather than special-use structures (Goodwill-Cohen 2001*1:134–135). In addition, evidence of year-round residential use of Sand Canyon Pueblo has been contrasted with evidence of the ritual use of Pueblo Alto, a Chaco great house (Ortman and Bradley 2002*1:44–48). In sum, it is clear that Sand Canyon was primarily a residential village and not a cluster of specialized buildings that functioned only as a ceremonial center for the residents of farmsteads in the surrounding community. Nevertheless, the presence of higher relative proportions of particular types of artifacts in some middens, such as the abundance of chipped-stone tools in Midden 103 (Database Map 4021), is suggestive of some production specialization within kiva suites.

The presence of distinctive, possibly ritual, features such as sipapus in kivas, as well as wall niches in kivas and other buildings, suggests that rituals as well as domestic activities were conducted in these residential kiva suites. The location of these features within residential buildings also suggests that the groups participating in these rituals would have been primarily the residents themselves. It is possible, however, that the particular rituals conducted varied somewhat between kiva suites; Till and Ortman (Chapter 8, paragraph 262) infer that the floor assemblage of Kiva 102, for example, suggests unusual ritual use late in the history of this structure.

The actual room-to-kiva ratio of Block 300, originally designated as the only "room-dominated" block, remains unresolved. Early discussions postulated special, nonresidential purposes such as communal storage for this block (Bradley 1992*2:90–91; Lipe 1992*2:125, 2002*1; Ortman and Bradley 2002*1:65). Lipe (2002*1) explores some implications of the presence of possible centralized storage facilities in this village. However, several issues can be raised with regard to this theory. First, the northern boundary of this block was defined more arbitrarily than most other block boundaries at the site—Block 300 was originally defined as including two additional kivas to the north-northeast that later were reassigned to Block 100 (Database Map 4003). Second, a kiva (Kiva 306) was exposed during excavation that had not been recognized on the modern ground surface. This discovery significantly reduced the estimated number of rooms per kiva for the block (Database Map 4003) and also raised the possibility that additional kivas are present in this architectural block that were not recognized on the modern ground surface.

There is also evidence that structures were added to Block 300 over time; this is more typical of residences than of planned, communal structures. In addition, although the excavated structures in this block show some indications of remodeling, the last use of these structures appears to have been residential. Lastly, midden materials from Block 300 are characteristic of domestic refuse, although, because only a small sample of refuse was obtained in the vicinity of this block, the use, or uses, of the structures in the block cannot be demonstrated conclusively. The paucity of refuse recovered might indicate either that little refuse was deposited or merely that we failed to locate the bulk of the refuse associated with these structures. If the former is true, it might constitute evidence that more of the rooms in Block 300 (as defined) were used for storage, at least originally. In sum, although an accurate room-to-kiva ratio and history of construction and use of the structures in Block 300 cannot be determined from the available data—and therefore use of this block might indeed have been different from that of most other blocks in the village—the evidence indicates that the structures excavated in this block were used for residential purposes (Chapter 4).

Room-to-kiva ratios have been used widely in the Southwest as an indicator of how structures and sites were used. As stated above, the relatively low ratio of rooms to kivas at Sand Canyon Pueblo, as determined from evidence visible on the modern ground surface, was key to the early hypothesis that, rather than being a residential village, this was a ceremonial center for a dispersed community (ancient kivas have been interpreted as ritual structures and men's meeting houses because historic and modern kivas are used for these purposes). However, the contents of midden assemblages collected during Crow Canyon's research at the site indicates that Sand Canyon Pueblo was residential. The relatively large number of kivas (for the number of rooms present) and the fact that nearly all hearths found at the site are located in kivas support the inference that ancient kivas, unlike the kivas in historic and modern pueblos, were residential structures. The specific locations of hearths within ancient pueblos are crucial indicators of domestic activities, because hearths are the loci of meal preparation and cooking and are also the places that residents gathered for warmth in winter and for light at night.

Public Architecture

The structural characteristics of two buildings at the site—Great Kiva 800 (with its peripheral rooms) and the D-shaped building (Block 1500) (Database Map 4001)—suggest that they were built as public architecture. As such, the uses of these structures reflect important aspects of ritual and sociopolitical organization at Sand Canyon Pueblo. However, none of these structures was being used for its original purpose when the village was depopulated; therefore, the best artifactual indicators of their original uses are the materials found in their associated middens. The rooms that encircle the two kivas in the D-shaped block and those that surround the great kiva contained little evidence of their original use; this could indicate that they were used for storage, albeit probably of a special, possibly ritual, nature.

The Great Kiva

Great kivas are abundant in the Mesa Verde region and are inferred to have served as community integrative structures from late Basketmaker times (Lightfoot 1988*1) through regional depopulation in the late A.D. 1200s. Similar to other great kivas in the region, Great Kiva 800 at Sand Canyon Pueblo contains features not found in ordinary-size kivas, such as a low masonry bench, floor-based masonry pillars, and a large masonry-lined floor vault (for more about these and other features found in the great kiva, see The Sand Canyon Pueblo Database). Construction of this great kiva might have predated construction of most of the other buildings in the village (Chapter 4).

The encircling peripheral rooms are a distinctive feature of the great kiva at Sand Canyon Pueblo. Although other great kivas with such rooms are known in the northern Southwest (Morris 1921*1:122), the peripheral rooms at Sand Canyon are more formal than most that have been documented (Martin 1936*1:51–53; McClellan 1969*1; Vivian and Reiter 1965*1:94–95)—that is, they form a neatly constructed row of contiguous rooms that abut the outside wall of the great kiva, and they are relatively uniform in size and shape. Little evidence of the original uses of these rooms was found (Chapter 4).

Several lines of evidence suggest that the great kiva at Sand Canyon Pueblo was not roofed, at least not during its final span of use (Chapter 4, paragraph 111), which could reflect a more inclusive, public use of this structure than of roofed kivas (Churchill et al. 1998*1; Lipe 2002*1:223; Ortman and Bradley 2002*1:62–63). Few artifacts were found on the floor; most were associated with a modest deposit of secondary refuse, the presence of which suggests that use of this structure ended before the village was depopulated. The cultural materials in the midden associated with the great kiva (Midden 803) suggest that a variety of activities occurred in that structure. This refuse contains an abundance of turkey bones and a moderate number of rabbit bones. Other taxa represented include great horned owl (one of only two bones of this taxon found at the site), canids, and falconiforms (vulture, hawk, or eagle). Some bones—mostly rabbit scapulae and elements from turkey legs—had been burned, probably the result of roasting for consumption. The flesh of any of these animals might have been cooked in the great kiva or brought to the kiva already prepared, and the bones might be feasting refuse. The relatively high frequency of sherds from white ware bowls is indicative of an emphasis on serving food in this structure (Chapter 8, paragraph 232) and is characteristic of locations where large groups of people gathered (Lipe 1970*1:129–130). Ortman and Bradley (2002*1:67–68) discuss additional evidence of possible feasting at Sand Canyon Pueblo.

The presence of particular types of artifacts in the refuse of the great kiva also suggests other activities. For example, the following artifacts were found in substantial quantities: modified sherds, stone cores and debitage, peckingstones, and ground-stone tools. Also found were bone artifacts (awls, beads, other modified bones, and a needle), two gypsum/calcite/barite items that might have been pendants, and three unmodified pieces of the same mineral. These artifacts indicate that the following activities might have occurred in the great kiva: pottery making, chipped-stone and ground-stone tool production or maintenance, weaving or basket making, and the fashioning of ornaments. Alternatively, the presence of any of these items or of bones from the animals discussed in the preceding paragraph could indicate their use in rituals or ceremonies. Interestingly, a high proportion of peckingstones has been found at sites in the northern Southwest that are thought to have had primarily ceremonial use (Ortman and Bradley 2002*1).

The D-Shaped Building

The multiwalled D-shaped building (Block 1500) at Sand Canyon Pueblo is an intriguing structure. Many D-shaped and multiwalled structures have been found in the northern Southwest (Lipe 2002*1:Table 10.2, 223–224; Lipe and Ortman 2000*1:110–111), in late Pueblo III villages in particular. To date, the excavated structure most similar to Block 1500 at Sand Canyon Pueblo is Sun Temple on Mesa Verde (Fewkes 1916*1), although the latter is substantially larger in plan. The shape and many other architectural characteristics of the D-shaped block at Sand Canyon Pueblo appear to be unique within the site.

Several lines of evidence suggest that the D-shaped block was constructed for special use (Chapter 4, paragraph 155) and designed to restrict access (Goodwill-Cohen 2001*1:134); in addition, the occupants might have held significant social power (Lipe 2002*1:226). Some researchers see evidence of calendrically based rituals in this block (Ortman and Bradley 2002*1:61). It is conceivable that the "D" shape of this structure and other structures of the same plan in the region are analogous to the plan of Pueblo Bonito (see also Bradley 1996*1), which could have been an ancestral site (see Chapter 7, paragraphs 82–85). Or, the "D" shape might signify an early "Bow" warrior society or priesthood. Warrior societies have been documented in many historic Pueblos (e.g., Bunzel 1992*1:525–528; Parsons 1939*1:127, 926; Smith 1952*1:204, 238; Stevenson 1904*1; White 1932*1, 1962*1), and evidence that such groups existed as early as the late A.D. 1200s in this region was found at Cliff Palace on Mesa Verde (Cole 2004*1:70). If the D-shaped block at Sand Canyon was indeed associated with this type of society, the labor investment in the construction of this block and the prominent location within the village could signify that this society or priesthood was of primary importance to the villagers and might have been linked to the warfare that occurred during this time.

At some unknown point in the history of this block, its use apparently changed: the doorways between the bi-wall rooms were sealed, and features and associated refuse either became domestic for the first time or became more heavily domestic. The original use of this block is therefore difficult to discern, but clues to this use might be found in the early refuse from this block. It is likely that the refuse associated with early use of the block was deposited around the outside wall of the building, and refuse found inside the rooms was deposited later. Refuse in abandonment contexts was deposited last.

What activities are indicated by the contents of only that refuse located outside the block? Those deposits contain substantial quantities of "indeterminate" ground-stone artifacts, stone cores, peckingstones, and modified sherds; lesser quantities of bone awls, projectile points, pottery lids, stone disks, and objects of adornment were found. This assemblage suggests that many of the same activities or rituals that occurred in the great kiva also occurred in the D-shaped block—chipped-stone and ground-stone tool production and maintenance, pottery making, and weaving or basket making. Only one metate and a few manos were recovered from this early refuse, in contrast to the more than 50 manos found on the floors and in the fills of the structures within the block; in addition, late in the use of this block, metate bins were added to the two kivas. The relatively lower proportions of corrugated gray sherds in this early refuse suggest that cooking increased during later use (Chapter 8, paragraph 237). The paucity of grinding tools and corrugated gray sherds in outside refuse is not typical of residential use and could be interpreted as evidence that the block was neither designed nor initially used for residential purposes but only for ritual or other nonresidential purposes. Or, the block might have been inhabited, not by a typical family group, but by one or more persons of special status or position who were supplied with prepared food by others in the village.

When viewed within this temporal framework, the faunal data from Sand Canyon Pueblo may offer new insights into the history of the pueblo. Previous researchers (Muir 1999*2; Muir and Driver 2002*2, 2004*1; see also Chapter 5, this volume) have reached numerous conclusions regarding the association of artiodactyl remains with D-shaped towers (towers abutted to the outside face of the village-enclosing wall) in Blocks 100, 200, and 1000. Muir (Chapter 5, paragraph 49), who interpreted the faunal data for this site primarily along spatial parameters, proposes that the structures associated with these towers were used as offices for the organization of communal hunts and for the processing, storage, and distribution of game.

Artiodactyl remains, although quite sparse in the faunal assemblage from the site as a whole, were, in fact, relatively more abundant in Blocks 100 and 1000. However, these remains were found almost exclusively in abandonment contexts—in collapsed roof debris and on kiva floors—not in middens. When separated along temporal parameters, then, the artiodactyl data indicate that very little of this meat was procured during the occupation of the village; most was obtained just before depopulation and therefore reflects activities and behaviors associated only with the end of occupation, a time of atypical conditions and events (see "Depopulation of the Village," below, and Kuckelman [2007*1]). The temporal patterns thus do not support inferences regarding sociopolitical organization (Chapter 5), feasting (Chapter 5), and ceremonial and ritual activities (Muir and Driver 2004*1) that have been drawn for the faunal assemblage on the basis of spatial distribution, unless these activities occurred only at the end of village occupation. In addition, physical evidence of genetic relatedness between the individuals whose remains were found in structures associated with D-shaped towers and the wide range of ages of those individuals (Chapter 7) strongly suggest that these structures were family residences rather than offices for the organization of communal hunts.

The refuse associated with the D-shaped block includes the largest variety and highest proportion of wild bird remains from the site; this, along with the distinctive planned layout of the block, has been interpreted previously as indicating special and ritual use (Chapter 5). Although this interpretation appears to be valid for early refuse outside the block, many wild bird remains were found inside the block in secondary refuse and abandonment contexts. The birds in these later contexts were thus probably procured during the more residential use of the block just before village depopulation. The presence of these remains might have resulted from ritual activity or might reflect the domestic consumption of nonpreferred animals and plants near the end of village occupation that is evidenced in abandonment assemblages in other areas of the site (see "Depopulation of the Village," below, and Kuckelman [2007*1]). For example, the sandhill crane, great horned owl, and turkey vulture, as well as animals such as fox and bobcat, are all represented in late deposits within the block and are not represented in the earlier refuse outside the block.

Village Layout and Organization

Considerable discussion has focused on the possible significance of various aspects of the layout of Sand Canyon Pueblo, including room-to-kiva ratios, the composition and layout of architectural blocks, and the locations of public architecture such as Great Kiva 800 and Block 1500 (the D-shaped building). Because the preponderance of evidence indicates that the pueblo was largely residential, it is probable that much of the village layout and organization was based on kinship ties and other associations between residential and social groups. The whole-site overview afforded at this writing may offer further information relevant to several of these topics, including the pattern of construction of the village-enclosing wall and the role of the central drainage in the plan and organization of the village.

In the areas excavated, the structures that were adjacent to the village-enclosing wall abutted that wall, indicating that the wall was built before those structures. However, evidence indicates that the enclosing wall was not built in one construction episode (contra Bradley 1992*2:95, 1993*1:39; Morgan 1994*1; Ortman and Bradley 2002*1:48, 49) but instead was constructed in large sections (Chapter 4, paragraphs 169–171), perhaps indicating somewhat less planning than has been inferred previously for this village. Thus, although some of the village layout was probably planned as large sections of enclosing wall were built, it is likely that new kiva suites were added by extended families as they grew during the 30-year occupation of the village. It is also possible that some aspects of spatial organization were associated with status or specialization.

It has been suggested that the bisection of the village by the drainage was significant organizationally and that, by design, most of the special-function or community architecture was located west of the drainage (Bradley 1992*2:97; Lipe 1992*2:126; Ortman and Bradley 2002*1) and the majority of the domestic architecture was located east of the drainage (Bradley 1992*2:97). Because the contents of midden assemblages indicate that, with the exception of the D-shaped block and the great kiva, the structures that were excavated were residential, it now appears that most structures on both sides of the drainage were, in fact, used for domestic purposes. Also, no evidence of socioeconomic differentiation was noted between midden assemblages from east of the drainage vs. those from west of the drainage (Chapter 5 and Chapter 8).

Although it is true that the two public buildings were located west of the drainage (Database Map 4001), the significance of this, if any, is debatable. Tree-ring dates indicate that the great kiva was built before the D-shaped block and possibly before most of buildings in the village were constructed. It is plausible that the location of the D-shaped block and the adjacent plaza were planned early in the history of the village. The decision to locate this block west of the drainage might have resulted from the location of the great kiva (and possibly other, unidentified public structures). However, it is just as likely that the location was chosen for its dramatic cliff-edge vantage point and its central location within the settlement.

Another theory that has been proposed for Sand Canyon Pueblo, as well as for other late canyon-rim villages in the Mesa Verde region, is that the drainage reflects a bilateral social division of the settlement, one perhaps akin to the dual division found in many historic pueblos (see discussion in Lipe and Ortman [2000*1:108–109]). Whether this was the case at Sand Canyon Pueblo cannot be substantiated with the data available. If one of the primary purposes of constructing this village was to cluster the community spatially around the canyon-head water source, then the fact that these villages were divided by a drainage might have been simply a result of the topography and of little social significance. However, if the division by the drainage was indeed socially significant, it raises a number of interesting observations and questions.

The substantially greater number of residences west of the drainage would indicate that this social group was substantially more populous than the group who resided east of the drainage. More important, perhaps, is the history of the development of dual divisions in Pueblo society. For example, did this dual-division system already exist when Sand Canyon Pueblo was laid out? When a family decided to relocate from a farmstead to the canyon head, did the members already belong to a wider group, and did that membership determine on which side of the drainage their home would be constructed? Is there evidence that dual division developed or existed among the scattered farmsteads that composed Pueblo II and Pueblo III communities? Of course, it is alternatively possible that dual division developed much earlier, in the Pueblo I settlements of the nearby Dolores River valley or in the villages of Chaco Canyon, and the concept and membership somehow persisted even during the intervening generations of dispersed residences.

A third possibility is that the seeds of the dual division seen in historic pueblos were sown in these late Pueblo III, canyon-head villages of the Mesa Verde region, which just happened to be bisected by a drainage. In other words, when people aggregated around the canyon-head springs, the physical reality of some residential groups being on one side of the drainage and some on the other could have been the basis of the development of a dual system. The clustering of populations in these canyon-head villages might have afforded the first opportunity and purpose in this region for internal organization of this type to develop. Fully exploring this subject is beyond the scope of this site report, but further study may prove fruitful.

The Villagers

The human remains found at Sand Canyon Pueblo are those of Pueblo people whose diet was heavily reliant on maize and who were relatively healthy, as judged by the standard of the times (Chapter 7). Although the represented subset of the residents of Sand Canyon might have suffered slightly more from systemic infectious diseases than did other ancient populations in the Southwest, the average to above-average stature of those represented indicate that the villagers generally enjoyed adequate nutrition and that disease did not adversely affect their growth. Occupational stress among these individuals is indicated by arthritis as well as by unusual tooth wear suggestive of the processing of leather with the teeth. A suite of distinctive occupational stress markers found on the remains of one middle-aged male suggests that this person was a craftsman (Kuckelman 2006*1, 2007*1; see also paragraphs 58–61 in Chapter 7, this volume) ; the relatively high proportion of stone tools, but not debitage, in the midden nearby suggests that these tools were used in his craft.

Evidence of possible intrasite, intercommunity, and interregional genetic relationships is presented in Chapter 7. Possible genetic relationships are indicated between two individuals found in Block 100, between six individuals found in Block 1000, and between the residents of the excavated portions of these two blocks (Database Map 4001). Evidence suggestive of intercommunity biological relatedness between residents of Sand Canyon Pueblo and residents of Woods Canyon Pueblo (10 km northwest) was found as well; these individuals might have been contemporaries. Possible consanguinity was also identified between an individual of elevated status whose remains had been interred in a room at Pueblo Bonito in Chaco Canyon and an individual at Sand Canyon Pueblo who might have been a descendent (Chapter 7, paragraph 82). A great deal more empirical evidence is needed to substantiate suggestions of extracommunity relatedness such as these; this evidence suggests avenues for further research regarding ancient consanguinity and migration patterns in the Southwest.

Subsistence

The villagers at Sand Canyon Pueblo used a wide variety of wild plant materials for food, fuel, and construction material, and they altered the natural environment around the village significantly during the occupation of the pueblo. Food remains found at the site indicate that the villagers relied heavily on resources they produced or raised themselves—primarily maize and turkeys. Although turkey feathers might have been used for ritual purposes, much as they had been for centuries, it is clear that turkeys had attained top status in the subsistence base of Pueblo people of the mid-thirteenth century in this region. The residents of Sand Canyon Pueblo also widely exploited the rapidly reproducing cottontail rabbit and edible plants that flourished in the disturbed soils of crop fields. The abundance and wide distribution of the remains of cottontail rabbits across the site suggest that, during most of the occupation of the village, cottontails were the wild animal consumed with the greatest frequency. Because human skeletal data indicate that the residents were healthy, this subsistence strategy was apparently successful during most of the time the village was occupied.

Depopulation of the Village

The depopulation of the Mesa Verde region in the late A.D. 1200s has been discussed extensively in the archaeological literature. Drought, environmental degradation, population pressures, conflict, and social and religious upheaval are among the causes commonly cited for regional depopulation (Dean and Van West 2002*1; Kuckelman 2002*1; Kuckelman et al. 2002*2; Lipe 1995*1; Lipe and Varien 1999*1; Petersen and Matthews 1987*1).

Although it is likely that a complex array of environmental and sociocultural factors led to the depopulation of Sand Canyon Pueblo and the region, important evidence of events and circumstances associated with depopulation was found in abandonment contexts at this site. This evidence indicates that the village was depopulated relatively quickly about A.D. 1280, when regional migrations were under way. Numerous lines of data contribute to the inferences drawn in this chapter regarding the stimuli, conditions, behaviors, and events surrounding the depopulation. I present this discussion of the depopulation of the village by briefly reviewing the conditions and events in the order in which they occurred, as indicated by evidence from the site; more-detailed data and inferences are contained in the chapters of this volume and in a journal article that is in press as of this writing (Kuckelman 2007*1).

By the A.D. 1250s, numerous villages—many of which enclosed a water source and incorporated architectural features that might have served defensive purposes—had been constructed in the Mesa Verde region. The natural resources of the region had probably been seriously depleted by centuries of occupation (Dean and Van West 2002*1:97; Driver 2002*4:158–160), which might have been a causal factor in these aggregations. Numerous researchers have proposed a causal relationship between competition for resources and aggregation during this period in the Mesa Verde region (e.g., Adams 1996*1:54; Dean and Van West 2002*1:97–98; Lipe 1995*1; Varien 1999*1:212–213; Varien et al. 1996*1:106–107), and others have proposed the same theory for aggregation in other areas of the Southwest during this same time (Crown et al. 1996*1:201; Reid et al. 1996*1:77).

An increased need for defense is also widely cited as a cause of aggregation in the thirteenth century, both for the Mesa Verde region (Haas and Creamer 1996*1; Kuckelman 2002*1; Wilcox and Haas 1994*1) and for others areas of the Southwest during this time (Lipe 2002*1:212; Reid et al. 1996*1:77; Rice and LeBlanc 2001*1). The presence of antemortem (as indicated by evidence of healing) skull fractures on the remains of numerous individuals from Sand Canyon and other settlements (Kuckelman et al. 2002*2:Table 3) suggests that raiding occurred at least sporadically during the middle and late A.D. 1200s. Competition for resources could have been one reason for this violence.

During most of the occupation of Sand Canyon Pueblo, the continuing existence of the village and the good health of the villagers indicates that their subsistence strategy was successful. However, because the diet of the villagers was heavily reliant on maize (Katzenberg 1995*1; Meyer 1995*1) and because the primary source of animal protein was domesticated turkeys (Kuckelman 2007*1; Muir 1999*2; see also Chapter 5, this volume), which were fed maize (Katzenberg 1995*1), the subsistence base was heavily—and thus precariously—dependent on climatic conditions favorable for growing maize (Kuckelman 2007*1). Carbon isotope levels suggest that the diet of the residents of the Sand Canyon community was even more heavily reliant on maize and animals that consumed maize (turkeys) than were the diets of contemporaneous communities on Mesa Verde and in Mancos Canyon, as well as on Cedar Mesa in southwestern Utah (compare Decker and Tieszen 1989*1:Table 1; Katzenberg 1995*1:Table 1; Matson and Chisholm 1991*1:Table 3). Data from Sand Canyon Pueblo indicate that the subsistence strategy of the villagers shifted substantially just before depopulation (Kuckelman 2007*1; Muir 1999*2; see also Chapter 5 and Chapter 6, this volume). For example, the proportion of turkey bones to other identified bones discarded into the middens was highest in the lowermost (earliest) layers of midden, which contained refuse that was probably discarded in the A.D. 1260s; this proportion declined slowly over the following decades and dropped abruptly near the time of village depopulation (Kuckelman 2007*1). In addition, the contents of thermal features suggest that, near the time of village depopulation, much less maize was prepared and a greater diversity of wild plant foods was procured (Chapter 6).

The timing of these subsistence changes correlates well with a known episode of environmental deterioration. Dean and Van West (2002*1:94–95) state that there were several periods of unfavorable conditions in the Mesa Verde region between A.D. 900 and 1500, but that the conditions during the late thirteenth century were probably the most severe, characterized by low groundwater levels, low crop-yield potential, disrupted seasonal precipitation patterns, a reduced dry-farming belt, and, beginning in A.D. 1276 and lasting until 1299, the Great Drought (Douglass 1929*1). These conditions could have drastically reduced or destroyed crops. Wild plant foods and weedy annuals associated with the crops would have also been significantly diminished.

In abandonment contexts at Sand Canyon Pueblo, there is evidence of a fairly abrupt climatic downturn and resulting acute subsistence stress for the relatively large population of the village. The residents of this village and the other residents of the heavily populated central Mesa Verde region (Varien 1999*1) could have been forced to exploit nonpreferred plant resources and to consume precious stored food, which probably consisted primarily of maize. Near the time of village depopulation, the consumption of cottontails increased, and turkey flocks, which required a great deal of maize and water, dwindled (Kuckelman 2007*1; see also Chapter 5, this volume). The available data suggest that one-fourth to three-fourths of the villagers departed before the settlement was completely depopulated (Kuckelman 2007*1); many residents might have planned to return if the climatic conditions improved. Refuse that accumulated on the floor of Great Kiva 800 and changes in the D-shaped block suggest an increase in residential use. Use of these important structures for rituals probably decreased or might have ceased entirely as a result of the difficult times and the drop in population.

As the drought continued, the remaining residents were probably compelled to consume most of their turkeys and their stores of seed maize. The contents of the latest thermal features and other abandonment contexts at the site indicate that, just before the village was depopulated, little maize (Chapter 6) or turkey (Kuckelman 2007*1) was being prepared for consumption. It appears that the residents adopted a largely hunting-and-gathering subsistence strategy and ranged ever farther afield for game and wild plant foods, which also would have been reduced by the drought. Some of the latest contexts at the site contain wild plant foods that were probably not preferred and some of which might have been obtained outside the central Mesa Verde region (Chapter 6, paragraph 68 and paragraph 72). A resident population might have been left in the village to tend the few remaining turkeys and maintain ownership of the community water supply (the spring) and ritual goods, as well as personal possessions, while others hunted and foraged outside the community and brought provisions to their families as they were able to obtain them.

As the exploitation of wild animals increased, the villagers continued to procure cottontail rabbits but also hunted animals they had not procured (or had seldom procured) previously, most notably pronghorn antelope, bighorn sheep, mule deer, and bobcat and other carnivores. A study that compared the expected richness of taxa against the actual richness from different contexts at the site revealed a significantly lower-than-expected richness in midden assemblages and a higher-than-expected richness in "all other assemblages," which, if recent contexts such as the modern ground surface are excluded, roughly equates with abandonment contexts (Chapter 5, Figure 10). These results provide independent corroboration of an overdependence on a few animal taxa (turkeys and cottontails) during most of the occupation of the village and the exploitation of a greater-than-expected variety of wild taxa near the end of village occupation.

Several of these wild taxa are known to have ritual significance among Pueblo people (Chapter 5, paragraphs 60–65), and it is possible that, even near the time of village depopulation, selected animals were used for ritual purposes and were not consumed. However, some of the remains have cut marks and localized burning indicative of butchering and cooking, which suggests consumption. Also, it seems unlikely that, near the time of village depopulation, villagers would have hunted numerous animals for ritual purposes that they had seldom or never before procured for that use.

Very few mule deer had been obtained during most of the occupation of the village; oddly, the greatest deer procurement appears to have occurred just before village depopulation. Game such as mule deer, pronghorn antelope, and bighorn sheep in particular was probably procured during coordinated, long-distance hunting trips or, less likely, obtained through trade. The acquisition of this food, regardless of the means, was probably either a strategy to outlast climatic difficulties or a means of provisioning long-distance travel from the region at a time when the resources at hand were meager.

It is possible that non-Pueblo people who, as a result of the drought and environmental degradation, were migrating through the region were one source of nonlocal plants and anomalous animal remains in the abandonment deposits at Sand Canyon Pueblo. However, it seems unlikely that travelers, be they peaceful or hostile, would transport a substantial amount and variety of large-animal meat that was still attached to the bone. It is more probable that the villagers themselves brought these provisions back from hunting forays during a period of subsistence stress.

Sometime after A.D. 1277, Sand Canyon Pueblo was attacked. Many of those still residing in the village perished, including groups of biologically related individuals in Block 100 and in Kiva Suite 1004. Two victims struck down while on the roofs of rooms along the village-enclosing wall might have been in lookout positions. Three others might also have been on the roof of a structure along the enclosing wall or perhaps were in the second story of a tower (not excavated) adjacent to the outside face of the enclosing wall. Some bodies were dumped down roof hatchways. The expediency and carelessness of this action suggests that it was done by the attackers, apparently during the time they remained in the village.

Several villagers died in kivas, perhaps caught while sleeping or hiding. It is not possible to state how many residents were killed, although, in the portion of the site excavated, the clustered or articulated remains of 23 individuals were found in abandonment contexts, and scattered remains in abandonment contexts suggest that at least 11 additional people died near the end of village occupation. None of these 44 people appears to have been formally interred, and the remains of several of these individuals exhibited direct evidence of violent death.

Among those individuals in abandonment contexts, adolescents appear to be overrepresented, and adult males seem underrepresented. The most robust adult males, those between 20 and 40 years of age, might have survived the assault, or perhaps the attackers chose to strike when the residents who could provide the greatest resistance were absent from the village. There is also evidence that some individuals who died had a diminished capacity to defend themselves, either as a result of immaturity, advanced age, frail heath, ill health, or possibly mental disability (Chapter 7, paragraphs 73–74). The data suggest that at least some attackers remained in the village for a short time after the assault ended, probably seeking hidden food stores; anthropophagy (the consumption of human flesh) might also have occurred and, if so, could be additional evidence of severe subsistence stress (Kuckelman 2007*1). Most proposed incidents of anthropophagy in the Mesa Verde region date from the middle A.D. 1100s and the late A.D. 1200s (Billman et al. 2000*1; Kuckelman 2006*2:134; Kuckelman et al. 2000*1), both of which were times of severe drought in the region.

The roofs of many kivas at Sand Canyon Pueblo were burned; these actions appear to have occurred after the violence. The roof burning might have been part of a "closing" ritual (Billman et al. 2000*1; Kuckelman 2003*1; Lipe 1995*1:157; Lipe and Varien 1999*1:238) effected by allies or residents who returned to the village after the attackers departed. These individuals might have hastily "closed" the kivas before vacating the village and migrating from the region.

Although it is not possible to determine the span of time over which the depopulation of Sand Canyon Pueblo occurred—that is, from the departure of the first group to migrate until the final attack. The evidence does indicate that it was a relatively short process, probably not more than a few years. Secondary refuse was found on the floors or in the fills of comparatively few structures. The fact that the subsistence strategy of the villagers changed dramatically suggests substantial subsistence stress, especially when paired with the independent evidence of severe drought beginning in A.D. 1276 (Douglass 1929*1). And although long-term, or chronic, nutritional stress can result in the slow development of arthritis and osteoporosis, acute nutritional stress does not leave unique markers on human skeletal remains. Therefore, it is possible that some individuals whose remains were found in abandonment contexts at Sand Canyon Pueblo starved rather than having died at the hands of attackers. None of the many thousands of tree-ring dates for the Mesa Verde region date later than A.D. 1281 (Robinson and Cameron 1991*1; Varien et al. 2007*1); thus, A.D. 1280 could be considered a "best guess" for the time when the depopulation of Sand Canyon Pueblo was complete.

The depopulation of this large village thus appears to have been stimulated at least in part by the failure of the subsistence base—heavy dependence on maize and turkeys—to support its sizable population during an abrupt climatic downturn and after probable long-term environmental degradation. It is likely that competition for dwindling resources was at least one catalyst for the violence at Sand Canyon and at other locations in the region during the late A.D. 1200s. The definitive event in the depopulation of the village was a lethal attack that resulted in the deaths of many residents and sealed the fate of this once-prosperous and successful village.

The identity of those who stimulated defensive postures and attacked sites in the Mesa Verde region has long been debated. The preponderance of evidence is of the negative variety; that is, little evidence of a non-Pueblo presence has been found in violence-associated contexts—or in any other contexts, for that matter. As a result, the theory of non-Pueblo invaders has enjoyed few strong supporters (see Cordell 1984*1:305–306; Kuckelman et al. 2002*2:505; Lipe and Varien 1999*1:341; Wilcox and Haas 1994*1:237). Although the cultural affiliation of those who attacked Sand Canyon Pueblo cannot be determined for certain, data from this site do contribute fodder to the debate.

There is no evidence to suggest that any of the human remains found at this site are of non-Pueblo people. All observable crania exhibit the cradleboard flattening that was ubiquitous among the Pueblo inhabitants of the Mesa Verde region by late Pueblo III times (Billman et al. 2000*1:164); the presence of this flattening has been used as evidence of cultural relatedness between individuals in a burial population (Karhu 2000*1:50). The results of stable carbon isotope analysis for numerous individuals from Sand Canyon Pueblo (Human Remains Occurrences 1–8, 10–17, and 19–22) indicate that these individuals all shared the same diet, which was heavily maize reliant (M. Anne Katzenberg, personal communication 2006). Thus, all the human remains found appear to be of Pueblo people—either the residents of Sand Canyon Pueblo or Pueblo attackers.

Artifacts that could be interpreted as evidence of the presence of non-Pueblo people in the village consist of five projectile points that are classified in Crow Canyon's analysis system as "nonlocal" (Chapter 8, paragraph 167). Any of these points might have been produced by a non-Pueblo craftsman; however, none was found in direct association with human remains. Two of the points were deposited sometime late in the occupation of Sand Canyon Pueblo, and three were deposited very near the time of village depopulation. The two points that were deposited late in the occupation consist of a Nawthis point fashioned from a nonlocal material; this point was found in a late refuse deposit on the floor of Room 1219 (a kiva corner room). The other was a Desert Side-Notched point, made from locally available material, which was found in a late refuse layer in Midden 103.

The other three points were found in structures that contained abandonment-context human remains. One Desert Side-Notched point made of nonlocal material was found near the remains of an adult female (HRO 14) in the burned roof fall of Kiva 501 (PL 88, Database Map 4089). A Nawthis Side-Notched point fashioned from a material found locally (PD 593, PL 1, Database Map 4135) and a Parowan Basal-Notched point made of an unknown type of silicified sandstone (PL 46, Database Map 4132) were found near or on a bench surface and in collapsed roofing debris, respectively, in Kiva 1004. The remains of a child and an adolescent who died as a result of violence were found on the floor of that kiva (Database Map 4135). The remains of several other individuals were also found in abandonment contexts elsewhere in Kiva Suite 1004 (Chapter 7).

The presence and contexts of these projectile points indicate either direct or indirect contact with non-Pueblo groups, probably from the west, that occurred late in the occupation of the village. The absence of similar items from earlier excavated deposits at the site also suggests that such contact did not occur before that, although such points might have been found in earlier contexts had more of the site been excavated. It is worth noting here that a Bull Creek projectile point made of red jasper, and possibly from southern Utah, was found in a late context at Castle Rock Pueblo (Kuckelman 2000*1:par. 8; Kuckelman et al. 2002*2:505; Ortman 2000*2:par. 99), a village near Sand Canyon Pueblo that was also fatally attacked in the late A.D.1200s.

The presence of these projectile points in late contexts at Sand Canyon and Castle Rock pueblos could be interpreted as evidence that both sites were attacked by non-Pueblo people, although the evidence is far from conclusive. These data do, however, add to the ongoing debate on the identity of late-thirteenth-century attackers and leave open the possibility that they were non-Pueblo people from areas to the west rather than Pueblo people from the Mesa Verde region.

Conclusions

Our research at Sand Canyon Pueblo has substantially deepened the understanding of community organization, continuity, and change in Pueblo society during the A.D. 1200s and of the conditions that played a role in stimulating the movement of people out of the Mesa Verde region in the late thirteenth century. The evidence suggests that, for the most part, the village that was Sand Canyon Pueblo represented a spatial "collapsing" of a dispersed community into an aggregated settlement that centered tightly on the community's primary water source and its communal ritual structure, Great Kiva 800. Some who became residents of the village might have been new members of the community who settled in the Sand Canyon locality in the mid–A.D. 1200s (Chapter 3).

Many other Pueblo III villages in the central Mesa Verde region were well established by the time the villages in the Sand Canyon locality, such as Sand Canyon, Castle Rock, and Goodman Point pueblos, were constructed about A.D. 1250. Why the Sand Canyon locality communities aggregated so late is not known, although one possibility is that they were compelled, finally, to aggregate for defensive reasons after other villages in the region had been built and violence had begun to escalate. The construction at Sand Canyon Pueblo of arguably the most complete and intimidating enclosing wall in the region suggests that the builders felt the need for such a structure; that is, that they felt more threatened than did the residents of other villages.

If the sample of structures excavated at Sand Canyon Pueblo is representative of the site, then most structures and suites in the village were constructed and used as residences. This conclusion is corroborated by the results of a study that examined the proportions of particular artifacts in the site-wide assemblage; this study indicated that Sand Canyon was a year-round, residential village rather than a ceremonial center (Ortman and Bradley 2002*1:48). The available data also support the theory, growing in popularity among archaeologists in the northern Southwest that, in ancient pueblos, ordinary-size kivas were the primary domestic structures and that only small-scale rituals were conducted in them (see discussion in Kuckelman [2000*5:par. 19–25]). The use of these kivas thus appears to have been more akin to that of Basketmaker pithouses than that of modern Pueblo kivas.

Because such a small proportion of the kivas present at the site were tested, however, the possibility remains that some kivas in the village were not residential but were used for special purposes by societies that crosscut kin groups and that the use of these structures helped integrate residents of the village above the household level and below the community level. It is also possible that "oversize" kivas similar to the suprahousehold, integrative protokivas that were constructed in the Pueblo I period (Kane and Robinson 1988*1; Wilshusen 1989*2) are present at Sand Canyon Pueblo; in fact, oversize depressions were observed at the site. However, the testing of one of these large depressions revealed an ordinary-size kiva (Kiva 1004)—the large depression on the modern ground surface was created by the kiva and a sizable courtyard that surrounded it. Thus, with the data available, it cannot be demonstrated that any oversize kivas other than the great kiva were constructed at Sand Canyon Pueblo.

In sum, at least two buildings—the D-shaped block and the great kiva with its peripheral rooms—were designed and used as public architecture; feasting and other ritual activities appear to have been associated with these structures, although it is possible that the D-shaped block was also a special-status residence. These buildings probably served as community-level integrative structures. Communal storage is also likely to have occurred in the rooms in these buildings, but the types of goods stored are unknown—possibly community ritual paraphernalia or special food stores such as seed corn. Because little evidence of trade was found at the site, it is unlikely that trade goods were stored there.

Among excavated structures and clusters of structures, there is little to indicate that vertical differentiation existed among the residents of the village, either in the form of architectural differences, type or abundance of burial goods, or abundance of trade items or ornaments. However, a slightly greater incidence of nonlocal materials and objects in Kiva 102 and associated Midden 103 could indicate enhanced social or economic power or status of individuals residing in this kiva and the associated structures. Other evidence of possible status or social power and the lack of evidence of "individual political aggrandizement" among the residents of this village is discussed by Lipe (2002*1:230–232).

Although some innovations resulted from this new, spatially compact social setting—for example, communal plaza areas, the D-shaped block, and other cooperative efforts—I think the occupation of this village was too brief (about 30 years) to accommodate the development of elaborate social adaptations and innovations. However, there is evidence of the ability to mobilize large groups for planning and construction and to organize rituals and other gatherings at the community level. Some adaptations to aggregated society might have been borrowed from other communities, such as Yellow Jacket (Kuckelman 2003*4), which were well aggregated decades before the construction of Sand Canyon Pueblo.

If the evidence of acute subsistence stress in the late A.D. 1270s at Sand Canyon Pueblo is representative of conditions and events across the region, then Pueblo residents could well have been largely "pushed" from the region. Drought, crop failure, decimation of turkey flocks, depletion of natural resources, and lethal attacks on villages would have constituted very persuasive "push" factors that would have been sufficient to depopulate the region. I think "pull" factors would have influenced residents primarily in their choice of migration destination. The demise of Sand Canyon Pueblo was thus probably one component in a pattern of subsistence strife and social unrest that led to widespread Pueblo migrations from the Mesa Verde region in the late A.D. 1200s.

References cited | To borrow, cite, or request permission