The best-known examples of probable cannibalism in the prehistoric Southwest all cluster in a very short period of time and in a relatively small geographic area: around AD 1150 in the area surrounding the modern town of Cortez, Colorado. Perhaps the most solidly documented of these assemblages is the one at Cowboy Wash on the southern piedmont of Sleeping Ute Mountain, which has been reported in several publications. There are sites with reputed evidence of cannibalism that fall outside of these narrow temporal and spatial boundaries, but most of them that have been adequately dated seem to fall earlier rather than later. In addition to sites like Cowboy Wash where cannibal incidents occurred during the long drought of the mid-twelfth century AD, there are some sites that seem to date to the tenth century, and perhaps the ninth and eleventh as well.
What has not been as well documented is any evidence of cannibalism later, especially during the major drought in the late thirteenth century that coincided with the total abandonment of the Mesa Verde region. Indeed, the authors of the Cowboy Wash articles note that there is no evidence for cannibalism in the Mesa Verde region after AD 1200. This is actually rather odd, since environmental conditions during the “Great Drought” of AD 1276 to 1299 were surely at least as bad as those during the twelfth-century drought, and the area had a significantly larger population as well. If, as many have argued, assemblages like Cowboy Wash were the result of warfare between communities in the region spurred by dwindling resources, it seems logical that there might be similar assemblages from the later drought period. There is certainly plenty of evidence of warfare and concern for defense. It could be that whatever it was that inspired some communities in the twelfth century to cannibalize others when attacking them failed to inspire later communities to do the same, even when faced with similar conditions.
There is evidence from at least one site, however, indicating that there may actually have been some cannibalism accompanying the warfare of the late thirteenth century. An article by Kristin Kuckelman, Ricky Lightfoot, and Debra Martin published in 2002 reported on evidence for violence at two sites in Canyons of the Ancients National Monument partially excavated by Crow Canyon Archaeological Center. One of these, Sand Canyon Pueblo, we have encountered before; it had an enclosing wall that seemed to indicate concern for defense, and as this article relates there were also eight individuals found there who seemed to have died violently, perhaps defending the community from attack. There was no evidence of cannibalism at Sand Canyon, however, and since very little of this large site was excavated it is hard to tell what the handful of violent deaths signify or whether the attack that apparently killed them led to the abandonment of the community. It does seem to have occurred around the time of abandonment, which was also the time that the whole region was abandoned, so it would make sense if the end of Sand Canyon Pueblo came violently.
The other site, however, was rather different. This was Castle Rock Pueblo, a smaller site a few miles from Sand Canyon. Here at least 41 individuals who died violently were found, at a site that probably only held about 15 households total with a total population of maybe 75. Since only a small portion of the site was excavated and yet remains representing at least half of the probable number of site residents were found, it is very likely that the entire site was destroyed, and all its residents killed or captured, in a raid at the end of site occupation, sometime around AD 1280 or 1285. In addition, many of the bones found showed clear evidence of processing similar to that found at Cowboy Wash and other twelfth-century cannibalism assemblages. This implies that at least some of the victims of the raid were eaten, although the lower number of processed bones relative to the total number of bones suggests that this was only done with some portion of the dead, which may mean that starvation would not have been the main motivation behind the cannibalism in this case. Several tools and pottery vessels at the site also tested positive for human myoglobin, using similar methods to those used to identify the contents of the coprolite at Cowboy Wash, again strongly suggesting that they were used to process and cook human muscle. All this suggests that cannibalism was in fact practiced after AD 1200 in the Mesa Verde region, at least at this one site. Since no other site from this period has shown evidence of cannibalism, however, even though evidence of warfare is common, it seems to have been an unusual occurrence, which does still leave the difference between the numerous incidents of cannibalism during the twelfth century and this isolated one during the thirteenth as an open question.
Debra Martin was one of the authors of the critique of the initial Cowboy Wash article, and she seems to have been responsible for the parts of the critique that addressed the analytical methods and data reporting in the article. Since she was also a coauthor on the Castle Rock article, it serves as a useful glimpse of what she meant in the critique. The tone and structure of this article are very different from the Cowboy Wash one. The tone is restrained, with the word “cannibalism” used sparingly and often replaced by “anthropophagy” (which means the exact same thing but sounds more technical and scientific), and a footnote discusses Crow Canyon’s consultation with its Native American advisory group and notes the objections some members of the group had to the publication of the article. The reporting of the data and analysis is very careful and thorough, and the text pointedly mentions that this is for the benefit of other scholars, to allow them to evaluate the information for themselves. The article never explicitly contrasts this with the approach of the Cowboy Wash team, but the implicit reproach is obvious.
This article also includes a useful discussion of who the attackers might have been. It is generally thought that other Puebloan groups rather than other cultural groups, such as nomadic hunter-gatherers, were responsible for attacks like this, and the authors of this study generally agree, although they do note the presence of a projectile point of a non-local type from east-central Utah as one possible piece of evidence for a cultural difference between attackers and victims. They then discuss which communities in the Mesa Verde region would still have been populous in the 1280s to defeat Castle Rock, which was a relatively small pueblo but was in a pretty defensible location and would have taken a fairly large force to overrun. They conclude that Sand Canyon, Yellow Jacket, Mesa Verde proper, and possibly Hovenweep were still occupied at the time, and that any of them could conceivably have been the attackers. They also note that groups from outside the region, especially in the Kayenta area to the southwest and the Totah area to the south, could have been involved, either on their own or in alliance with local groups. Figuring out which of these groups was actually to blame is tricky, however. With the twelfth-century cannibalism sites there is a division between “victim sites” like Cowboy Wash, where people were apparently cooked and eaten at the location that had been attacked, and “perpetrator sites” where bodies were brought from sites that had been raided and eaten at the attackers’ home base. Castle Rock is clearly a victim site, but since it is the only known cannibalism site dating to this period no known perpetrator sites that might reflect the attackers have been identified.
One of the major contributions of this paper, in addition to documenting thirteenth-century cannibalism in the Mesa Verde region, is in specifically tying cannibalism to warfare. In most documented cannibalism assemblages, including at Cowboy Wash, the bones have been so heavily processed that any evidence of the cause of death is obliterated, so it can be impossible to tell if the victims died violently or died some other way and were then eaten, perhaps by starving members of their own communities. Since the Castle Rock assemblage contains many instances of violent death without apparent cannibalism, in addition to some examples of cannibalism, all of this having happened at the same time, it makes a strong case that cannibalism was in fact associated with warfare during this period, which in turn implies that it probably was during the twelfth century as well. This is an important finding in the attempt to figure out what these gruesomely enigmatic sites mean for the prehistory of the Southwest.
Kuckelman, K., Lightfoot, R., & Martin, D. (2002). The Bioarchaeology and Taphonomy of Violence at Castle Rock and Sand Canyon Pueblos, Southwestern Colorado American Antiquity, 67 (3) DOI: 10.2307/1593823