Go to Table of Contents.
List of Tables
List of Illustrations
Research Objectives and Methods
Architecture and Site Layout
Population Estimates
Faunal Remains
Archaeobotanical Remains
Human Skeletal Remains
Water Control and Subsistence
Abandonment and Emigration
Appendix A

Faunal Remains

by Jonathan C. Driver


The analysis of the Woods Canyon Pueblo fauna is part of a larger project on faunas from archaeological sites in the northern San Juan region. Previous and current work in the region includes a study of domestic turkey (Munro 1994*1); analysis of animal bone from Sand Canyon Pueblo (Muir 1999*2), Castle Rock Pueblo (Driver 2000*1), and a dozen small Pueblo III sites in the Sand Canyon locality (Driver 1996*1, 1997*1; Driver et al. 1999*1; Muir 1999*2); and a general overview of spatial and temporal variation (Driver 2002*1). The Woods Canyon sample is of interest because the site is located in an area with one of the densest concentrations of Pueblo III sites in the region (Varien 1999*1) and therefore allows for comparisons between this site and numerous surrounding sites.

Many of the publications cited above discuss the methodological problems inherent in any zooarchaeological analysis conducted in this particular region. Therefore, this report does not discuss the characteristics of individual taxa, recording methods, or taphonomy. The emphasis instead is on intrasite and intersite variation.

Methods and Results

Specimens were identified and described by the author and research assistants using a standardized recording system (see Driver et al. 1999*1). We analyzed all faunal specimens, except those identified as human, and they included specimens that had been modified as artifacts (for example, awls and tubes). Each specimen was described on a separate line in the database. In addition to recording provenience information, analysts described each specimen in terms of taxon, skeletal element and portion, length, and modification (including breakage, burning, and cut marks). Few specimens were identified to the level of species, but about half could be considered "identifiable" to some extent. The complete database is available on Crow Canyon's Web site. The number of identified specimens is presented, by taxon, in Table 1. I did not calculate values for minimum numbers of individuals, although this could be done using the information in the database.

The animals represented in Table 1 either can be found today within a few kilometers of the site or could have been found until the introduction of modern agricultural practices. The taxa identified at Woods Canyon are normally found in any assemblage of comparable size in the region. As is common in Pueblo III assemblages, the Woods Canyon sample is dominated by turkey and lagomorphs (cottontail and jackrabbit). When compared with assemblages from other sites in the region, the Woods Canyon sample has relatively few rodents, especially those in the squirrel family, Sciuridae, which includes both prairie dogs and various species of ground squirrel.

Intrasite Analysis

The intrasite analysis was driven by two main questions:

  • If different areas of the site were used at different times, is there evidence of change in faunal use through time, perhaps similar to that described for the nearby Sand Canyon locality (Driver et al. 1999*1; Muir 1999*2)?
  • Are there differences among faunal assemblages associated with different types of structures, such as those between domestic and public architecture described by Muir (1999*2) at Sand Canyon Pueblo?

The site was divided into four spatial zones, or sections, by the archaeologists—the canyon bottom, the upper west side, the east talus slope, and the canyon rim (see "Architecture and Site Layout"). Of these, the canyon bottom is thought to have been occupied earliest and longest. The rim complex, on the canyon rim, was occupied latest, and it might include more public architecture (in the form of towers and a D-shaped structure) than is present in other parts of the site. Table 2 compares three major categories of animals from these four main subdivisions of the site. "All lagomorph" includes specimens that were identified as either cottontail rabbit (Sylvilagus) or jackrabbit (Lepus), or simply as "lagomorph" (usually broken bones or bones of immature animals that could not be assigned confidently to either genus). For reasons described elsewhere (Driver et al. 1999*1), all specimens identified as "large bird" are assumed to be from turkey (Meleagris gallopavo), and these are combined into a single category (turkey/large bird) for analysis. The third category, "artiodactyls," includes any specimens from even-toed ungulates such as deer, sheep, or pronghorn antelope. In the assemblage from Woods Canyon Pueblo, only four such specimens were identified, and none could be assigned to a genus.

Table 2 reports the combined totals from structures and nonstructures (areas outside buildings that probably include midden material) for each of the four main sections of the site. The bulk of the assemblage was excavated from the canyon bottom and the upper west side. There is little difference in the collections from these areas, both of which are dominated numerically by the turkey/large bird category. Of the two much smaller assemblages, that from the rim complex is very similar to those from the upper west side and canyon bottom, and the very small sample from the east talus slope has somewhat less turkey/large bird. However, given the variation in sample size, there is little reason to think that differences among the various sections of the site are significant.

Table 2 also shows the relative frequency of taxa from structures and nonstructures. Looking at the two largest assemblages (canyon bottom and upper west side), one can see that the dominance of the turkey/large bird category is more marked in middens, whereas lagomorphs increase in relative importance within structures. This pattern has been observed consistently at sites in the region, and possible explanations have been explored elsewhere (Driver 2000*1; Driver et al. 1999*1). One hypothesis is that lagomorphs (especially cottontails) may have been attracted to abandoned and collapsed masonry structures and that postoccupational natural deaths account for some lagomorph bones in structures. An alternative explanation is that human behavior favored the deposition of lagomorph bones in rooms, perhaps because rabbits were stored or processed there.

Intersite Analysis

In previous studies, I and others have investigated patterns of faunal representation within much of the northern San Juan region (Driver 2002*1), as well as on a much smaller scale, within the Sand Canyon locality (Driver 1996*1, 1997*1; Muir 1999*2). In this analysis, I consider an area of intermediate size where Woods Canyon Pueblo is located. This area is defined by the triangle formed by the modern communities of Montezuma Creek, Utah; Dove Creek, Colorado; and Cortez, Colorado; it is in the center of the area that has been referred to as the "central Mesa Verde region" (Varien 2000*1:Figure 1). It is a plateau dissected by canyons, and during Pueblo III times it supported a relatively dense concentration of large and small settlements (Varien 1999*1; Varien et al. 2000*1).

In the Southwest, it is common for zooarchaeologists to study intersite variability by examining relative frequencies of the lagomorphs (sometimes combined and sometimes as separate taxa), turkey, and artiodactyls. Table 3 presents these data for assemblages in the study area that have at least 200 specimens of "lagomorphs" and "turkey/large bird" combined. Only sites with "pure" Pueblo III assemblages have been included, resulting in the exclusion of sites where some mixing may have occurred. One must contend with a variety of problems when comparing assemblages reported by different analysts; the most serious is probably that different analysts have different criteria for identifying specimens (Driver 1992*1). Intersite comparison therefore must be approached with some caution.

In Table 3, sites are arranged in order of the relative frequency of "lagomorphs" and "turkey/large bird." The sample is not distributed evenly across the study area, and there is a preponderance of assemblages from the Sand Canyon locality. It is evident that all the assemblages that my students or I analyzed form a group in the middle of Table 3 (from Woods Canyon Pueblo to Kenzie Dawn). Green Lizard, however, is also in the middle of the table, and although it is located in Sand Canyon, it was studied by Walker (1990*2), who did not use the same recording system. Thus, the clumping of sites that were studied using the same methods is not necessarily a product of the recording system.

Perhaps the most striking feature of the assemblages is the great range of relative values for "lagomorphs" and "turkey/large bird." For example, turkey dominates the Pueblo III faunal assemblage from Nancy Patterson Village, located in Montezuma Canyon to the west, whereas lagomorphs outnumber turkey at Wallace Ruin to the southeast. Interestingly, there appears to be some sort of relationship between the ratio of lagomorphs to turkey/large bird and the percentage of artiodactyls in the assemblage. All the assemblages in which turkey/large bird makes up at least two-thirds of the combined lagomorph, artiodactyl, and turkey/large bird total have only 1 percent or less of artiodactyl specimens. In assemblages with greater relative frequencies of lagomorphs, artiodactyls become somewhat more common. I should note, however, that all Pueblo III sites in the study area have low artiodactyl values when compared with earlier sites and when compared with Pueblo III sites on Mesa Verde (Driver 2002*1). Thus it appears that sites that emphasized the domestic production of turkey over the hunting of lagomorphs were also the least likely to be involved in hunting large game.

The site where artiodactyl hunting has been investigated most thoroughly is Sand Canyon Pueblo (Muir 1999*2). Muir's analysis of the faunal assemblage demonstrated that the higher frequency of artiodactyls (mainly deer) was not characteristic of all deposits at the site. In fact, assemblages associated with domestic structures and middens were similar in composition to assemblages from nearby sites where there was little evidence of deer hunting. At Sand Canyon Pueblo, artiodactyls were most commonly associated with architecture (notably towers) that did not resemble domestic structures. Muir argued that the structures were associated with ritual organization and that hunting of artiodactyls was undertaken by particular groups of people within the pueblo. The conclusion drawn by Muir is more specific than my earlier argument that the increase in deer hunting at Sand Canyon Pueblo was the result of a larger population using its political power to exclude small settlements from its hunting territory (Driver 1996*1). Although this might have been the case, Muir's analysis demonstrates the likely mechanism whereby this was achieved and also raises the possibility that influential groups within Sand Canyon Pueblo might have had better access to artiodactyl meat.

Hurth's (1986*1) analysis of a small domestic settlement in the Yellow Jacket complex (Site 5MT3) yielded results consistent with what has been documented for many other Pueblo III sites. However, preliminary data from the Crow Canyon Archaeological Center's recent excavations at Yellow Jacket Pueblo (Site 5MT5) indicate that more artiodactyl bones were found in association with towers and D-shaped structures and that the overall Pueblo III assemblage from this site is similar to that from Sand Canyon Pueblo (see Table 3). This contrasts with Woods Canyon and Castle Rock pueblos (Driver 2000*1), which seem to have been community centers (Varien 1999*1:Figure 7.5). Neither of those sites, however, has artiodactyl values approaching those of Sand Canyon and Yellow Jacket pueblos, and the few artiodactyl remains at Woods Canyon and Castle Rock pueblos are not associated with particular types of structures. There appears to be a size threshold in the central Mesa Verde region, with only the very largest sites displaying evidence for more than occasional hunting of artiodactyls in Pueblo III times.


Analysis of the faunal assemblage from Woods Canyon Pueblo reveals that the inhabitants of the village relied on domestic turkey and hunted or trapped lagomorphs for their supply of animal protein and fat. When compared with other sites in the same area, Woods Canyon is notable for a relatively high ratio of turkey to lagomorphs and a low overall frequency of artiodactyls. Intrasite analysis indicates no significant differences among the four sections of the site, suggesting that neither temporal change nor differences in the use of space had much effect on faunal assemblages. Although Woods Canyon Pueblo was a relatively large site and might have functioned as a community center, its faunal assemblage is not distinctively different from the faunas of smaller sites in the region. Faunal data for the two largest pueblos in the central Mesa Verde region suggest that only the very largest sites have faunal assemblages that display significant intrasite variation. These large sites are also notable for the increased quantities of artiodactyls in their faunal assemblages.

References cited | To borrow, cite, or request permission

Copyright © 2002 by Crow Canyon Archaeological Center. All rights reserved.