Archive for the ‘Natural Environment’ Category

Metate Incorporated into Wall Capping, Pueblo Bonito

I’ve written a bit about the recent research, spearheaded by Larry Benson of the USGS, into the sources of the corn found at Chaco.  These studies continue to refine the techniques used to identify source areas, but so far they have shown that corn was almost certainly being imported to Chaco both during and after the florescence of the Chaco system between AD 1030 and 1130.  As they begin to test more potential field areas, of course, the number of matches for the strontium isotope ratios in the corn at Chaco has increased.  While early studies indicated that much of it likely came from the Chuska Valley, it now looks much more likely that it instead came from the area along the Chaco River between there and the canyon.  This is an area with numerous outlying great houses, and it was probably the main route for the many commodities from the Chuska area that were brought to the canyon, and it’s also generally a better area for agriculture than the canyon itself, so this all makes sense.  There has also been some evidence that at least some corn was also coming from the Totah area to the north, again a more productive agricultural area with many Chacoan outliers.

Based on the proveniences of the corncobs from Pueblo Bonito that were tested early on, one tentative suggestion emerging from this research was that the main sources of imported corn changed over time.  The cobs that came from the lower Chaco River were from Rooms 3 and 92 in the northern part of Old Bonito, one of the earliest parts of the building to be built, while the one cob of possible Totah origin came instead from Room 170, in the southeast corner, one of the newest parts of the site.  Since there does seem on other grounds to have been a shift in the emphasis of the Chaco system from south to west to north over time, it would make sense that the early rooms contained early cobs from the west while a later room contained a later cob from the north.

Talus Unit with Snow

A paper published in 2008 by a group of big names in Chacoan studies sought to look at this directly by radiocarbon dating the cobs.  This is an interesting paper which goes beyond that narrow topic to also look at the characteristics of the corn found at the various great houses and other sites.  One of the co-authors is Mollie Toll, a specialist in archaeobotany who has done a lot of research on Chacoan corn.  As part of that research, she had long noted that the corn at Pueblo Bonito generally had bigger ears with more rows of kernels than most other corn known from the prehistoric Southwest.    It was bigger than earlier and later corn, for one thing, but it was also bigger than most other corn from the same period.  Corn from the Chacoan occupation of Salmon Ruin was also unusually large, as was corn from the Talus Unit behind Chetro Ketl, but corn from Pueblo Alto and Pueblo del Arroyo, other contemporary great houses at Chaco, was smaller and more in line with that from earlier and later sites.

Toll came up with three possible explanations for the difference.  Pueblo Bonito corn could be a different variety or “landrace” from the others, which is plausible but not directly testable with current technology.  It could also have been grown outside of the canyon where conditions were better for agriculture, while the corn from other great houses was grown in the canyon where conditions were poorer.  Finally, and problematically, the corn at Bonito might not have been Chacoan at all!  Since modern corn is generally bigger than ancient corn, Toll (when she was first looking at this in the 1980s) couldn’t exclude the possibility that the corn found at Pueblo Bonito had actually been put there by Navajos in the nineteenth century.  Much of it was from George Pepper‘s excavations in the 1890s, so it couldn’t be newer than that, but there was no way for Toll to tell how much older it was.

Room 3a/92/97, Pueblo Bonito

We still can’t tell different ancient landraces apart (although the recent sequencing of the maize genome may make this more feasible in the future), but the strontium isotope testing is giving us a sense of where the corn was grown, and accelerator mass spectrometry now makes directly dating the corn relatively easy.  Seven cobs from Pueblo Bonito that had been used in the strontium studies were dated for this paper.  One was the cob from Room 170 that possibly came from the Totah, one was from Room 92, and the rest were from Room 3.

The results were illuminating, but also challenging.  All the cobs clearly dated to ancient times, so the possibility that the size of Pueblo Bonito’s corncobs represents recent deposition is effectively quashed.  Three of the Room 3 cobs had closely clustered dates with intercepts around AD 1000, which offered some partial support for the idea that the corn in the early rooms was relatively early, but the other two were widely spaced, one at 870 and at 1170.  This is problematic for the idea that the date of corn in a room can be predicted from the date of that room’s construction, but it makes sense that the deposits in a room may date to well after its construction.  Since Room 3 dates very early, probably to the 900s, it’s likely that the deposits there resulted from much later trash dumping once it was no longer used for its original purpose.  Room 3 has a firepit, so it was probably originally a residential room, and it is likely one of those “big square rooms” that I have argued began to take the place of kivas in Chacoan room suites of the tenth century.  Room 92 is part of the maze of confusing rooms next to Room 3.  It had a well-preserved floor with corn and bean bushes on it (it’s not clear from Pepper’s description if this was the second or third floor), which suggests that it was used as a storeroom at the end of the period of occupation in this part of the building.  The cob from this room had the latest date of any in the study, with an intercept of AD 1220, which is consistent with the idea that this room was in use as a storeroom at the end of occupation.

Room 170, Pueblo Bonito

The biggest surprise, however, was the cob from Room 170, which dated to AD 1010.  This is particularly odd, since Room 170 was probably built around 1080 or even later.  Looking at the probability curve for this date, there is some chance that the actual date was around 1100, but the curve as a whole has a much more prominent peak around the intercept at 1010 than any of the other reported dates, which suggests that the probability is quite high that the intercept does in fact represent the true date or close to it.  The authors give various possibilities for why the cob might have been placed in this room long after it was grown, including the idea that it was put there as some sort of ritual offering of continuity with the occupation of earlier parts of the building.  I prefer another explanation they also suggest, which is that it was part of an earlier trash deposit that was redeposited in Room 170 for some reason.  There is very little information on what the deposits in this room were actually like, but many of the rooms in this part of the building were full of trash when excavated, and I think it’s most likely that this one was too.  The trash could have been put there for any number of reasons; if it was redeposited from somewhere else, it may have served as structural fill to support an upper story.  In any case, this puts a damper on the idea that the overall sources of corn changed over time.  Indeed, the sources seem to have been pretty constant through time for cobs left in different areas of the site, which suggests that the real story is much more complicated.

One nice thing about this paper is that the authors do a very good job of properly reporting their radiocarbon dates, particularly in giving point estimates as intercepts, which are meaningful, rather than midpoints, which are not.  Many papers make this mistake, including some of Benson’s reporting these and other dates on corn.  This paper also shows the probability curves for the dates, which give even more information.  This seems to be pretty common these days among Mesoamerican archaeologists, but it’s still quite rare in the Southwest, where radiocarbon dating has only recently become a major focus.  The availability of tree-ring dates, which are much more precise, has generally led Southwestern archaeologists to neglect radiocarbon, but it’s becoming increasingly obvious from studies like this one that the ability to date things other than trees is very useful in interpreting sites.

Obviously this paper just reports a handful of dates, and the authors take pains to point out the tentative nature of any conclusions they draw, but it’s an important contribution to the issue of where the Chaco system, whatever its nature, was getting its means of support.  As is often the case with new avenues of research, at this point papers like this pose more questions than they answer, but there are plenty of corncobs out there to date and analyze in other ways, just as there are plenty of potsherds to test for theobromine.  Once we get a bigger database of dates and strontium (and other) ratios, we’ll start to get a clearer picture of the behavior behind these remains.

Metate Fragment at Pueblo Alto

Read Full Post »

Shannon Bluffs, South of Farmington, New Mexico

One reason for the relative lack of information available on the prehistory of the Totah is that the presence of all those big rivers leads many sites to be buried under alluvium and/or destroyed by flooding and changes in the courses of the rivers.  As a result, many sites are not visible at all on the surface, and this is particularly the case for small sites, especially since the local architecture for much of the Pueblo period relied heavily on adobe and cobble masonry, which is much less durable than the sandstone masonry typical of Chaco and Mesa Verde.  Thus, aside from really big sites like Salmon and Aztec, many Totah sites are only discovered with very deep excavation or erosion.

Linda Wheelbarger’s chapter in the Salmon synthesis volume, reporting on the findings of the Totah Archaeological Project on the Bolack Ranch just south of Farmington, emphasizes this in pointing out how many of the sites on the ranch were not visible in any way from the surface and were only discovered inadvertently, such as when breaches in irrigation ditches lead to swift erosion, revealing sites well below the ground surface.  The most obvious sites are on the terraces above the river, and these are also some of the largest sites (including some probable Chacoan great houses), but it’s not clear if they are actually the largest or if there are larger ones buried somewhere in the floodplain.   Most of the known floodplain sites are small houses, but they are quite numerous, and Wheelbarger is able to define five “communities” along the southern bank of the San Juan between the confluence of the Animas River to the west and the Gallegos Wash to the east.  These tend to be at the confluences of various side drainages (including the Animas and the Gallegos) with the San Juan, which is a pattern noted elsewhere in the San Juan Basin as well, including to some degree at Chaco itself.

Plaza at Salmon Ruin

This is something to keep in mind when evaluating the conventional wisdom that the area around Salmon Ruin was largely uninhabited when construction of the building began around 1090.  The basis for this very common assertion is an extensive site survey done in the area around Salmon by the San Juan Valley Archaeological Project in the 1970s in conjunction with excavations at Salmon.  This survey revealed only four small sites within 1 kilometer of the great house that might have been contemporary with it, and only 12 such sites within 6 km.  In his chapter on the function of Salmon in the synthesis volume, Paul Reed explains the survey and its limitations:

The survey did not entail 100 percent coverage because of the complexity of land ownership and lack of permission to survey some parcels.  Nevertheless, much of the territory in the 1 km area around Salmon was surveyed.  As a caveat, it is likely that flood deposits from the San Juan River, along with alternating cycles of erosion, may have concealed or removed other sites located on the floodplain below Salmon.  We have no way of knowing how many such sites may have been present.  With the data that are available, however, it is clear that Salmon was not the center of a large community of surrounding small pueblos; rather, Salmon largely comprised the entire community.

Reed is clearly aware that it is likely that any sites that may have existed on the floodplain are no longer visible, but he nevertheless concludes that “it is clear that Salmon was not the center of a large community of surrounding small pueblos.”  Well, no, it isn’t clear, even “with the data that are available,” unless you make the totally unwarranted assumption that the available data do in fact reflect the reality despite their obvious shortcomings.  It’s worthwhile to note that the handful of sites that were identified were mostly on the terraces, rather than the floodplain, which means that they don’t have much relevance to the issue of how many sites there were in the region overall.  It’s certainly possible that Salmon was founded in a vacant area, as Reed concludes, but it’s important to note (as he does) that this would make Salmon quite unusual among Chacoan great houses, which usually were built among contemporaneous small sites both in Chaco Canyon itself and at outlying communities.

West Wing of Aztec West and Terrace to the North

A somewhat comparable situation exists at Aztec, although there is evidence of a fairly substantial residential district on the terrace above the West and East great houses.  Very little is known about the extent of settlement on the floodplain around the main “downtown” district, and some have argued that Aztec, too, was founded in an area without substantial prior settlement (with the terrace-top houses presumed to postdate the initial construction of the great houses), whereas others have argued that there probably was some sort of existing settlement there that is no longer visible because of the river-side location.  In either case it is clear that Aztec was a larger and presumably more important community within the region than Salmon.

There isn’t any way to settle this issue without extensive testing and excavation, which is unlikely to happen any time soon, but I just want to flag it to emphasize that a lot of the ideas that get entrenched in the archaeological literature are not necessarily well founded, and it’s important to understand the evidence behind them and how strong it is.

Terrace North of Salmon Ruin with Salmon Museum at Top

Read Full Post »

Totah Theater, Farmington, New Mexico

In comments to my post on Salmon Ruins, John Barton asks for more discussion of this area, which is surprisingly poorly understood given its obvious importance to Southwestern prehistory as a whole and the Chaco system in particular.  Wolky Toll has a chapter in the Salmon synthetic volume discussing the Totah region (named from the Navajo name for the Farmington area), and particularly the La Plata subregion, which is becoming somewhat better understood due to a major salvage archaeology project along New Mexico Highway 170, which parallels the La Plata River from the Colorado border south to its confluence with the San Juan just west of Farmington.  Toll has played a major role in this project, and his chapter has interesting things to say about the Totah in general and the La Plata valley in particular.  I don’t really buy all of his interpretations of Chaco; he’s one of the major proponents of a view of Chaco as a regional ceremonial center drawing pilgrims from throughout the San Juan Basin, including the Totah, but with a minimal population permanently resident in the canyon.  He’s particularly associated with the view that even the small-house residents at Chaco only lived there for part of the year, having other residences in other communities, especially along the Chuska Slope to the west.  I’m more inclined to see Chaco as some sort of hierarchical system with at least a relatively large permanent population, mostly in the small houses, though I’m not sure which version of this idea (and there are many out there) I find the most convincing.

Still, Toll knows a lot about the Totah.  He even introduced the term to archaeological use in an important chapter in a previous edited volume that he coauthored with Peter McKenna.  One of the important points he makes in the newer chapter is that while this region has historically been treated as part of either the Mesa Verde region to the north or the Chaco region to the south, it really has an independent identity and cultural trajectory that has been obscured by seeing it entirely in terms of migration or influence from north or south.  This is not to say that the Totah was isolated from developments to the north and south; far from it.  It’s really more accurate to see the whole San Juan basin as a single cultural region, with remarkable uniformity in many cultural expressions and changes over time.  The specific manifestations of those cultural processes were not necessarily identical, of course, but there’s more similarity than archaeologists are often inclined to say.

Mesa Verde Museum

Part of the problem here is just the way archaeology developed in the Southwest.  As Toll notes, the activities of the Wetherill family had a huge influence on which areas came to be considered most important to the interpretation of regional prehistory.  They were not the only influential figures, of course, but they definitely did a lot to put Mesa Verde and Chaco specifically on the radar of the archaeological profession as well as the general public.  In any case, the way things developed was that Mesa Verde and Chaco became well-studied, with major excavation projects in the late nineteenth and early twentieth centuries producing huge numbers of artifacts and a general understanding of the chronological sequence of pottery types and other artifacts.  Once tree-ring dating provided an absolute chronology for the whole region, the general outline became clear: Chaco flourished in the eleventh century then declined in the twelfth, while Mesa Verde hit its peak later, in the thirteenth century, shortly before the whole region was abandoned around 1300.

This was a bit of a shift from the more evolutionary approach to culture history encapsulated in the original Pecos Classification, developed at the first Pecos Conference in 1927 and described by Alfred Vincent Kidder in a short article in Science at that time.  This system saw both Chaco and Mesa Verde, with their big, impressive masonry “pueblos,” as belonging to the Pueblo III or “Great Pueblo” period.  The tree-ring dates, however, showed that Chaco’s peak actually occurred earlier, coincident with the widespread small sites that marked the Pueblo II period.

Aztec West Great House, Aztec Ruins National Monument

Turning back to the Totah, the main excavation project there in the early twentieth century was conducted by Earl Morris at Aztec Ruins.  This was the largest site complex in the area, and it clearly indicated some level of social and cultural importance.  What Morris found there, however, instead of a unique and clearly indigenous material culture, was a mix of what seemed to be Chaco and Mesa Verde material culture.  The early deposits showed clear similarities to Chaco, as did the architecture of the site, which Morris interpreted as evidence for a close cultural connection to Chaco.  After this period, however, Morris saw evidence for an extended hiatus with little evidence of any sort of occupation or use.  After that there was another, quite different suite of material culture that looked much more like Mesa Verde.  Morris interpreted this sequence as an initial Chaco-affiliated occupation followed by abandonment and reoccupation by immigrants from the Mesa Verde region to the north.  In an important chapter in the Salmon synthesis volume, Gary Brown, Peter McKenna, and Tom Windes argue persuasively that Morris was actually wrong about this, and that while the construction and early occupation of Aztec does indeed show substantial connections to Chaco, there was probably not any abandonment or hiatus, just a period of somewhat reduced construction activity at a time of widespread drought and environmental hardship in the mid-twelfth century.  This lull was followed by extensive occupation and construction in the thirteenth century, especially at the east ruin (which Morris didn’t excavate).  The occupants at this time did have pottery similar to that used at Mesa Verde, but that doesn’t mean they were immigrants from there, and it’s much more likely that they were primarily local people who had been living at Aztec all along.  Everyone in the region at this point was making the type of pottery now known as “Mesa Verde Black-on-white,” and there’s no particular reason to think that any groups in the Totah had links to Mesa Verde, which itself seems to have been remarkably isolated during this period, with few trade goods found at the many excavated sites in the region despite its large population.  A similar story seems to obtain for Salmon, with an early Chaco-affiliated occupation followed by a period of continued occupation but little major activity, then an increase in population and activity before the final depopulation of the entire region.

So why did Morris get this wrong?  One reason, which Toll emphasizes, is that the mere fact that Chaco and Mesa Verde have been much more extensively studied than the Totah means that ceramic types (and other types of material culture, but pottery is the most important for cultural classification) have become associated with one or another of these areas, so that when they are found elsewhere in the region they are taken to indicate influence or migration from Chaco or Mesa Verde rather than a regionwide stylistic trend uniting all of these areas.  The latter is more likely, however, especially for the Totah, which was a major population and cultural center throughout the Pueblo II and III periods.  In her chapter in the Salmon synthesis volume, Lori Stephens Reed describes the discovery that the ceramic types found at Salmon and Aztec that have traditionally been classified as “Cibola” (Chaco) or “Northern San Juan” (Mesa Verde) types based on temper and design were mostly made within the Totah, judging from the type of clay used for the paste and slip of the vessels.  Rather than define new types, she just adds the qualifier “Animas Variety” to the existing type designations to indicate this local origin.  This makes sense from an Ockham’s Razor perspective, but as Toll notes in his chapter it’s really the type names themselves that have led to the downplaying of the local factor in the prehistory of the Totah.

Mesa Verde Escarpment from 2009 Pecos Conference at McPhee Campground

The best example of this is the very widespread thirteenth-century pottery type known as “Mesa Verde Black-on-white,” which is found all over the place but has tended to be interpreted as indicating some sort of influence or migration from Mesa Verde.  This is highly improbable, however, since Mesa Verde was gaining rather than losing people for most of this period (until the very end), and the people there don’t seem to have been very actively engaged in regional trade.  This strongly suggests that Mesa Verde Black-on-white is probably of local origin wherever it is found, despite the name.  Toll even muses more than once about how interpretations of Southwestern prehistory might be different if it were called “Aztec Black-on-white” instead.  It’s quite clear that Aztec was a very important site during this period, perhaps not as important as Chaco had been earlier but certainly more important than any single site in the Mesa Verde area.  And yet, because Mesa Verde has been more intensively studied, until quite recently it has been accorded an enormously important role in regional dynamics during this period that closer examination is revealing to be mostly undeserved.  Chaco has received a similarly privileged position for its period of florescence for similar reasons, but it seems to have actually been roughly as influential as this assumption implied.  (Something of an archaeological Gettier case.)

But why didn’t the Totah get the early attention that would have gained it the pride of place in Southwestern archaeology occupied by Chaco and Mesa Verde?  Ironically, a big part of the answer seems to be tied precisely to the geographic factors that made it such an important area in the first place.  One of the main reasons Mesa Verde and Chaco attracted early attention from archaeologists and pothunters was that their isolated locations left them unbelievably well-preserved.  The sites were very obvious on the landscape, many had stood relatively well due to either their massive construction (at Chaco) or their sheltered locations (at Mesa Verde), and they were sufficiently hard to get to that subsequent inhabitants and explorers hadn’t done them much harm.

Animas River, Farmington, New Mexico

The Totah, however, is an enormously attractive and productive agricultural area.  This is presumably what attracted people to Salmon, Aztec, and other communities in prehistory, and it definitely attracted huge numbers of Anglo settlers in the late nineteenth century who proceeded to plow over, loot, and otherwise damage the numerous archaeological sites they found before archaeologists had even heard of them.  The really big sites, like Salmon and Aztec themselves, managed to remain in relatively good condition until they could be professionally excavated, but innumerable smaller sites have likely been completely destroyed.

The local environment has also led to decreased visibility for these sites directly, by covering them with alluvial silt that makes them difficult or impossible to see from the surface.  As a result, we have little sense of how many sites are out there today, let alone how many were there initially before the farmers and the pothunters got to them.  Again, this is in contrast to the harsh environments of Chaco especially, and Mesa Verde to a lesser extent, where there are no permanent rivers to bury sites so deeply.  Furthermore, modern development in the Totah has been extensive, and there’s very little information about what lies underneath the rapidly growing modern towns of Farmington, Aztec, and Bloomfield.  For all of these reasons, the Totah remains surprisingly understudied, despite its obvious importance for understanding Southwestern prehistory.  Luckily this is starting to change a bit, at least on the conceptual level, with publications like Toll’s and Reed’s that point out the distinctiveness of this area and its independent identity.  The Totah has stood in the shadow of Chaco and Mesa Verde for a very long time, but it now seems to be finally coming into the light.

Chaco Street in Aztec, New Mexico

Read Full Post »

Chuska Mountains from New Mexico Highway 371

In the year AD 1098 a spruce tree was chopped down in the Chuska Mountains, which run roughly along what is now the border between Arizona and New Mexico.  We don’t know who cut it down, exactly, since the people living in the area at the time had no system of writing and have therefore not left us any explanation of their actions.  We do know why it was cut down, however, and what became of it, and that gives us some clue to who might have wielded the axe.  It was likely either the local inhabitants of the Chuskas or the inhabitants of Chaco Canyon, 45 miles to the east, who had close connections to the Chuskas and may have traveled there regularly to do things like cut down trees.

After the bark and branches of the tree were carefully stripped from the trunk, and its ends laboriously ground flat, it was probably left to dry out for a couple of years.  At the end of that period, a group of men carried it carefully, by hand, the 45 miles to Chaco, where it was used in the construction of Kiva G at Chetro Ketl, the prominent elevated round room that can still be seen prominently at the site today.  This room was built over the remains of earlier round rooms, and the architectural history of this part of Chetro Ketl is very complicated and hard to unravel.  This particular beam, now known as CK-168, was used as a “tie beam” in the northeast corner of the square room into which Kiva G was built.  I take this to mean that it was used to span the interstitial space roughly tangent to the outside of the circle of Kiva G as part of the elaborate support structure of beams that was necessary to support the “blocked-in” kivas like this one that are so typical of Chacoan great house architecture.  The stage of construction involving Kiva G, and therefore beam CK-168, was one of the latest at Chetro Ketl, and Kiva G is noteworthy for displaying the “McElmo” style of masonry, which is a relatively late style at Chaco that echoes the style of masonry found to the north at sites such as Aztec Ruins.

Elevated Kiva G at Chetro Ketl from the North

We know when and where beam CK-168 was cut down because of two types of analysis that have been done on this and many other pieces of wood from Chaco and other prehistoric archaeological sites in the American Southwest.  The first is tree-ring dating, which allows the determination of the date a given tree was cut down with amazing precision, potentially to the calendar year.  Not every beam that is analyzed produces what is called a “cutting date,” indicating the exact year in which it was cut, but CK-168 did, and this is how we know it was cut down in 1098.  The other technique, which was developed much more recently, is strontium isotope ratio analysis, which at its current level of development allows the determination of the general area in which a tree (or a person) lived.  Many fewer pieces of wood have undergone this type of analysis than have been tree-ring dated, but CK-168 is one that has, and it is from this testing that we know it came from the Chuskas.  We don’t know exactly where in the Chuskas it came from, but we can be reasonably sure it was somewhere in the range.  These techniques are more or less the state of the art in wringing information from the archaeological remains of the prehistoric Southwest, and the amount of information available from them in this arid region with good preservation of materials like wood is quite remarkable by the standards of prehistoric archaeology in general.

At the same time the tree that became beam CK-168 was being cut down and left to dry out, on the other side of the world a group of heavily armed men was rampaging through a country that was not yet under their control but soon would be.  These were the Crusaders, a group of pious Christian knights and others, mostly from the various lands that are now part of France, who had heeded Pope Urban II’s call in 1095 to aid the Byzantine emperor in fending of the Turks who were threatening his lands and who at some point added the goal of conquering the holy city of Jerusalem from the Turks as well.  This expedition is now known as the First Crusade, and by 1098 it had reached the Holy Land and begun to conquer parts of it.  In June of that year they captured the important coastal city of Antioch (now in Turkey) after an eight-month siege and began to move on toward Jerusalem.  One group, under the command of Count Raymond of Saint-Gilles, arrived at the nearby city of Ma’arra (now in Syria) on November 28 and began to lay siege to it.  This siege, in striking contrast to the long, miserable one at Antioch, lasted only two weeks, and on December 12 the Crusaders entered the city after having destroyed its defenses.  They then remained at Ma’arra for another month as Raymond and the other commanders who had assisted him in the siege argued over how to divide up the spoils of their conquests.  On January 13, 1099 the ordinary soldiers, who were impatient to get to Jerusalem, finally prevailed on Raymond to move on and the army left Ma’arra.  They reached Jerusalem in early June and, assisted by a group of Genoese sailors who landed at Jaffa on June 17 and disassembled their ships to make siege engines, seized the city on July 15.  They then massacred the majority of the inhabitants.

Villages in Syria from the Golan Heights

In the grand scheme of the Crusade, and compared to the brutality that marked the Crusaders’ actions at Antioch and Jerusalem, the siege of Ma’arra may seem like a minor chapter.  There was something distinctive about Ma’arra, however, and it remained a painful memory for both sides quite out of proportion to the short duration of the siege.  The Arab historians recorded that the massacre of the people of Ma’arra after the siege was even worse in total number killed than that at Jerusalem, and some reported that the Crusaders had promised the people of the city safety before treacherously killing them.  According to the chroniclers who accompanied the Crusaders, however, what happened at Ma’arra was even worse than that.  With clear discomfort but remarkable consistency, almost all of the chroniclers recorded that some of the Crusaders had engaged in an activity much less respectable in European eyes than killing Muslims: eating them.

The cannibalism at Ma’arra has long been a difficult issue for historians of the Crusades to understand.  Beginning with the contemporary chroniclers, some of whom were at Ma’arra and apparently witnessed the events personally, they have offered various explanations for the behavior.  In a recent article Jay Rubenstein took a close look at the accounts the chroniclers offered of Ma’arra, and found some interesting patterns.  For one thing, although almost all of the contemporary chroniclers mention the cannibalism, the specifics of their accounts vary considerably.  Some placed it during the siege, while others put it afterward, during the month when the Crusaders were hanging around the city as their leaders argued.  Some attributed it to famine and starvation among the soldiers, who were said to have eaten flesh from dead bodies furtively.  One eyewitness account says that when the leaders of the Crusade found out about this they piled the bodies in a mound and set fire to them to put an end to it.  Others, however, said that the Crusaders ate the flesh of the dead Muslim enemies avidly and publicly, and one even said that some of the Crusaders were so disgusted by the sight that they deserted, blaming the leaders of the Crusade for not providing sufficient supplies and letting things deteriorate to this point.  Some of the chroniclers blamed the cannibalism on an apparently fictitious group of poor Crusaders called “Tafurs.”

Church of the Holy Sepulchre, Jerusalem

Rubenstein makes some interesting connections between all this and various other medieval references to cannibalism, including various episodes of starvation cannibalism during famines in Europe and references to “theatrical” mock-cannibalism involving the ostentatious preparation of dead enemies as if to eat them in view of living enemies, including one instance allegedly during the siege of Antioch and one earlier, in Spain.  He also mentions Biblical references to cannibalism as a punishment for disobeying God, with the context often being siege warfare quite similar to that conducted by the Crusaders, as well as some suggestive references to the possibility that some Crusaders might have resorted to eating each other (or at least come close) during the long, hard months at Antioch.  This is all rather speculative, however, and most of it has one crucial difference with the reports of what happened at Ma’arra, which Rubenstein does acknowledge: in these cases people were generally eating members of their own group, while at Ma’arra all the chronicles are quite clear in saying that the Crusaders only ate the flesh of the Muslim enemy.  This alone suggests that more than simple survival cannibalism is involved here, since there would surely have been plenty of dead Crusaders around to eat and the contempt the Crusaders had for “infidels” makes it odd that they would choose to defile themselves by eating them.  Rubenstein notes this and comes up with an admittedly speculative reconstruction of what happened at Ma’arra in which he proposes that there had been scattered instances of survival cannibalism at Antioch and perhaps earlier but that at Ma’arra, during the siege, some of the starving Crusaders decided to make a virtue of a necessity by ostentatiously eating dead Muslims in full view of the defenders of the city as a form of psychological warfare.  This would presumably have struck fear in the hearts of the Muslims and perhaps led to an easier conquest of the city than had been the case at Antioch.  Upon learning about this, however, the chroniclers, highly uncomfortable with this celebratory cannibalism, finessed their accounts to put the cannibalism after the siege, attribute it to starvation, and make it furtive and discouraged rather than open and aggressive.  Rubenstein puts the celebratory cannibalism in the context of “holy war” and the idea that the Crusaders thought of themselves as tools of God sent to punish the Muslims, which justified anything they did in the course of achieving that aim, regardless of how perverse and taboo it might be under normal circumstances.

This is an interesting interpretation, and it is similar in some ways to the model I have proposed for the outbreak of cannibalism in the Southwest around AD 1150 (about fifty years after the Ma’arra incident) in which I argued that the attackers, whoever they were, may have been motivated by both hunger and the desire to terrify their adversaries.  I ultimately don’t buy it, though.  A careful look at how Rubenstein supports his argument shows that the sources he relies on are not the most likely to be reliable.  The chronicles that put the cannibalism at Ma’arra during the siege rather than after it, as well as those that mention the possibility of earlier cannibalism and various other aspects of Rubenstein’s theory, were mostly written well after the events in question by people who did not witness them.  Most of them did apparently speak to eyewitnesses in preparing their accounts, and in some cases these eyewitnesses may have included some of the cannibals themselves, a fact that Rubenstein leans on heavily in arguing for their validity as sources, but this is a thin reed on which to hang such an elaborate argument.  Looking just at the chronicles written by writers who actually accompanied the Crusade, some of whom were at Ma’arra personally, they generally seem to agree that the cannibalism took place after the siege, when the Crusaders were hanging around Ma’arra and probably still hungry from the months of deprivation and disease at Antioch.  These chroniclers do disagree on some other aspects of the cannibalism, but in general their disagreements are more minor than those between them and some of the later accounts by writers who were not there and relied on secondhand testimony.  The fact that only Muslims were eaten does still require some explanation beyond simple starvation, but a desire to humiliate the enemy, perhaps combined with an aversion to eating people they had known personally in life, could explain it at least as well as an attempt at psychological warfare, which as Rubenstein notes really requires that the cannibalism took place during rather than after the siege.  It may be a better explanation, in fact, in light of the fact that the Arab historians do not mention the cannibalism although they do note the brutality of the siege and massacre.  This suggests either that the Crusaders were unsuccessful in trying to terrorize the Muslims through ostentatious cannibalism or that the cannibalism took place after there were no living Muslims left in Ma’arra to see it.  It could also suggest that the cannibalism was successful in intimidating the defenders, who were then massacred, leaving no one left to tell other Muslims about it, but the fact that the eyewitness chroniclers were in general agreement that the cannibalism took place after the siege makes me think that it’s more likely that there just weren’t any Muslims around.

Sunset at St. Peter's Church, Jaffa

Note the difference here between Rubenstein’s interpretation of the cannibalism at Ma’arra and mine.  This is a context where we have abundant written documentation of the events in question, and yet we have a variety of possible interpretations of what really happened and no way to tell for sure which is closer to the truth.  We must just rely on which fits the evidence best and is the most plausible, a somewhat subjective matter at all times.  This is in contrast to the more “objective” evidence we have of what happened in some contexts in the prehistoric Southwest, such as the time and place that beam CK-168 was cut down.  Note, however, that the information we have about CK-168 is extremely limited compared to what we know about Ma’arra, despite the many conflicting accounts in the chronicles.  These just aren’t comparable levels of knowledge, and this points out the difference between history and archaeology.  Both can tell us stories about the past, but the stories are very different.  Without the written documents about Ma’arra we would likely never know that cannibalism occurred there; unlike in the alleged instance of cannibalism in the prehistoric Southwest, which are based largely (though not entirely) on extensive “processing” of human bones.  The type of cannibalism alleged at Ma’arra consisted largely of the cutting of strips of meat from the fleshy parts of the body, which would likely not have left any marks on the bones and thus no physical evidence visible in the archaeological record.  But because we do have written records we do know about it.  If we had written records of what happened at Cowboy Wash we would presumably understand it better than we do, but the example of Ma’arra shows that we would not necessarily understand exactly what had happened even then.

Archaeology has its advantages, however.  Judging from what people typically write down when they have access to writing, it’s very unlikely that anyone in the Chuskas or at Chaco would have bothered to write down exactly when and where beam CK-168 was cut down even if they could.  Hundreds of thousands of beams were imported to Chaco during its florescence, and it’s unlikely that the specific origin of each of them would have been important enough to anyone to record.  Nevertheless, the analytical techniques available to modern archaeology does allow us to know in some cases when and where individual beams were harvested.  The story of CK-168 and the story of Ma’arra may be very different types of story, known to us in very different ways, but they are both known to us, and knowing both of them enriches our understanding of the past despite the still unanswered and perhaps unanswerable questions the stories raise.  And remember that these two stories happened at the same time; they were literally simultaneous.  Many things were going on in the world in the year 1098, and the stories of most of them are lost forever.  The stories of these two events, however, are not, and we can still tell them today.
Rubenstein, J. (2008). Cannibals and Crusaders French Historical Studies, 31 (4), 525-552 DOI: 10.1215/00161071-2008-005

Read Full Post »

Sleeping Ute Mountain and Surrounding Landscape from Four Corners

If you stand at the Four Corners monument and look in the direction of Colorado you will see Sleeping Ute Mountain dominating the view.  From this direction you are looking at the southwest side of the mountain, and in front of it you see the southern piedmont.  On the right side of the piedmont, though not visible from this distance, is Cowboy Wash.  It’s one of several ephemeral streams running from the mountain itself across the piedmont to the San Juan River.

One thing that might strike you about the view from this perspective is that it looks like an awfully dry, desolate, uninhabitable wasteland.  And you would be correct to think that.  The southern piedmont of Sleeping Ute Mountain is an extremely arid and inhospitable environment even by the standards of the Southwest, which is saying something.  It’s only a few miles from Mesa Verde to the east and the Great Sage Plain to the north, both areas that get relatively abundant rainfall and supported large and prosperous prehistoric communities, but it is worlds away from them environmentally.  While those areas get sufficient rainfall to support dry farming, and the Great Sage Plain is commercially farmed even today, the southern piedmont does not, and any type of agriculture there would have to rely on some sort of irrigation.  Today the Ute Mountain Ute tribe has a large irrigation project in the area, using water brought in from McPhee Reservoir, 45 miles to the north, via the Towaoc Canal.  The construction of the reservoir and the canal was part of the Dolores Project, which involved substantial archaeological excavation of the inundated area that significantly improved archaeological understanding of the prehistory of the region.  This work took place from 1978 to 1985 and was known as the Dolores Archaeological Project, the largest salvage archaeology project in US history.

McPhee Reservoir, Dolores, Colorado

The creation of the irrigated fields on the piedmont resulted in further salvage excavations in the 1990s.  Among the sites excavated was 5MT10010, which contained considerable evidence of a gruesome incident of probable cannibalism around AD 1150.  It is not the only site in the area to show evidence of cannibalism during this period; in fact, three other sites in the same community, excavated slightly earlier in connection with the construction of the canal, also showed evidence of having been destroyed in an incident involving extensive processing of human remains in a way suggesting cannibalism, and there are several other sites in the area showing similar assemblages, most from the same period but at least one from a later period.  It is at 5MT10010 that the most solid evidence for actual cannibalism, as opposed to processing of bones in a way that may or may not indicate actual consumption of human flesh, in the form of a coprolite that tested positive for the presence of human muscle tissue.

There are many questions that arise from these findings, but one of the most puzzling is also one of the simplest: what were people doing living at Cowboy Wash in the first place, and how did they manage it?  After all, they weren’t building giant dams and canals of the sort involved in the Dolores Project.  In many parts of the Southwest, especially upland areas like Mesa Verde, dry farming using only rainfall was standard during this period, and water control techniques were generally used only for domestic water if a nearby spring or other reliable source was not available.  There are a few springs on the southern piedmont that probably would have supplied sufficient domestic water for the small number of people living there, but the rainfall would definitely not have been sufficient to farm with.  The only source of water at all sufficient for agriculture would have been the occasional floods, from spring snowmelt and summer thunderstorms, that would flow through Cowboy Wash itself and the other drainages on the piedmont.  None of these flows permanently today, and there is no evidence that they ever did.  As at Chaco Canyon, then, which is similarly dry, farming would have to have been based on some sort of technique for capturing the floodwater.

Flowing Chaco Wash and Cliffs below Peñasco Blanco

There are a variety of ways this might be done, including diverting the rainwater from cliffs, as was done at Chaco, planting along the sides of the drainage where the floods would regularly overflow the banks, and what is known as “ak-chin” farming, as practiced by the O’odham of southern Arizona, which involves planting right in the path of the runoff at places where the velocity of the water is relatively low, as at the mouths of tributaries to main arroyos.  There are no sheer cliffs on the southern piedmont like the ones at Chaco, so probably a mix of overbank and ak-chin farming would have been practiced at Cowboy Wash.

A paper by Gary Huckleberry and Brian Billman addresses the nature of farming at Cowboy Wash, and also addresses a related issue, which is whether periodic entrenchment of arroyos due to drought played a role in the patterns of abandonment and migration that characterize Southwestern prehistory.  It is pretty clear by now that the paleoclimatological record shows periods of drought corresponding to periods of abandonment of certain parts of the Southwest, and one proposed mechanism for how this would have worked is that drought would have led to increased erosion and/or hydrological changes in the water table that led to the entrenchment of arroyos, which would have been disastrous for populations dependent on certain types of floodwater farming (especially overbank), as the broad floodplains of the local drainages would have been replaced by deep channels that took the water away quickly instead of letting it overflow to water the crops.  Ak-chin farmers would not necessarily have been affected to the same degree, but if the side drainages they used became entrenched as well they would not have been able to use their techniques either.  Thus, drought would lead to arroyo-cutting, which would lead people to leave formerly productive areas for others that were less affected.  This theory has been proposed as an explanation for certain events at Chaco, with the idea being that some of the social changes late in the Chacoan occupation were due to degradation of the Chaco Wash and the need to change agricultural strategies.  The phenomenon of arroyo-cutting in general is richly illustrated in historic times at Chaco.  The early reports of the Chaco Wash from the nineteenth century indicate that it was a shallow, meandering drainage, much like the current condition of the Escavada Wash to the north and the “Chaco River” that is formed by the confluence of the two at the western end of the canyon and flows north to the San Juan.  By the early twentieth century, and accelerating since then, however, the Chaco Wash through the canyon has cut down significantly and there is a very deep arroyo channel apparent today.

Entrenched Arroyo at Chaco

The drought-downcutting-abandonment theory makes sense as far as it goes, but as Huckleberry and Billman point out there are some problems.  For one thing, the extent to which arroyo-cutting is actually linked to drought, rather than other factors including the specific geology of the area, is hotly debated and there is no consensus.  The idea that while drought may be one factor causing arroyo-cutting there are other factors involved as well is supported by the fact that in different drainages in the Southwest that have been studied in depth the periods of arroyo-cutting do not necessarily correspond to region-wide droughts or other climatic changes.  In some areas they do, but in other areas they don’t.  At Cowboy Wash specifically, the available evidence indicates that the wash began to entrench sometime before AD 950, and that it began to refill with sediment sometime between AD 1265 and 1400.  If abandonment does in fact correspond to arroyo-cutting, then presumably the Cowboy Wash area should have been abandoned between 950 and 1265, and possibly occupied before and after this.  If downcutting results from drought, there should also be evidence of drought during the 950 to 1265 period.

The basic upshot of the Huckleberry and Billman paper is that neither of these expectations is met.  The evidence for drought conditions at Cowboy Wash generally matches that for the rest of the region, with the major droughts in the mid-twelfth century and late thirteenth century AD and several smaller droughts at irregular intervals before then.  This doesn’t show any particular relationship to the stratigraphic evidence for arroyo-cutting, which seems to have been going on to some degree throughout the period from AD 950 to at least AD 1265.  Furthermore, the evidence for settlement doesn’t line up either.  The marginal nature of the Cowboy Wash area implies that it would probably not have been occupied for most of prehistory, and this was indeed the case.  There were a few ultimately unsuccessful attempts to colonize the southern piedmont, however, and they don’t show any particular relationship to the periods of arroyo-cutting (although they do perhaps relate to periods of drought).  The first agricultural occupation of the area came during the Basketmaker III period, when a few pithouses were apparently used seasonally as summer fieldhouses, presumably associated with nearby fields, from about AD 600 to 725.  After these were abandoned, at a time which may correspond to a drought, the area does not seem to have been occupied again for more than three hundred years.  Then, around AD 1050, a few permanent, year-round sites were built.  These seem to have been occupied for only a few years, however, as there was no significant buildup of trash associated with them.  After they were abandoned, three larger villages, including one at Cowboy Wash, were established around AD 1075.  These had extensive trash deposits and seem to have been occupied for one or two generations.  These communities were apparently abandoned, however, when the next occupation began in the 1120s by a population with apparent links to the Chuska Mountain area to the south.  This occupation at Cowboy Wash is the community that was apparently destroyed around AD 1150 (again coincident with a major drought) when its inhabitants were mutilated and cannibalized.  After this event, the area was once again abandoned until about AD 1225, when two new communities were founded, including one again at Cowboy Wash.  Within a few decades the population at Cowboy Wash appears to have aggregated at Cowboy Wash Pueblo, following a typical pattern for the region.  Also typical of the region, the whole southern piedmont seems to have been abandoned by AD 1280, at the time of the “Great Drought” that coincides with major changes throughout the Southwest.

Entrenched Chaco Wash from Cliff Top near Pueblo Bonito

So basically, all of the attempts at year-round occupation of the southern piedmont seem to have occurred during the period that Cowboy Wash was being downcut.  While these were all ultimately unsuccessful, some lasted for a few decades, so clearly they were able to grow some food at some times.  This strongly implies that at least in this case, arroyo-cutting was not particularly linked to abandoned, although drought probably was.  Huckleberry address the issue of how farming could have been done during periods of downcutting by looking at Cowboy Wash and its tributaries today.  They find that while some portions of the main wash, especially, are indeed heavily downcut, other portions are not, and they label this type of drainage a “discontinuous ephemeral stream,” which is to say, a normally dry wash with some portions that are severely downcut and others that are not.  On the uncut portions, which include much of the length of the tributaries, overbank or ak-chin farming could easily be done today, and this was presumably the case in antiquity as well.  The hydrology of the area is such that the areas of downcutting would not have been stable, and would have tended to migrate upstream, but the complexity of the system is also such that this would not have made the entire system unusable; while some parts were being newly cut, others would be filling in, and prehistoric farmers would merely have to move their fields around a bit rather than abandoning the area entirely.

All that being said, however, the question of why people were trying to settle this quite harsh and difficult area in the first place.  It is interesting to note that the attempts at settlement generally came during periods of relatively favorable environmental conditions, which would have made this area a bit less forbidding than usual, as well as during times of increased regional population, when all the good land may well have been taken and some people were forced to seek out the more marginal areas.  The violence that appears to have accompanied the drought of the twelfth century, especially, suggests that when the good times came to an end social relations got very bad very fast.  Huckleberry and Billman suggest that the reason people did end up abandoning Cowboy Wash, the times when they were not attacked, was merely drought itself, which they were unable to cope with as well as other populations, even those who also used floodwater farming techniques, because the size of the watershed was relatively small and the amount of rainfall feeding the washes was also small, so the total amount of water they had to work with was much smaller even in good times than at place like Chaco with large watersheds.  In that context, even a small decrease in annual precipitation could be devastating, leading to failed harvests and the need to move away.

Non-Entrenched Escavada Wash from New Mexico Highway 57

Indeed, there is evidence that the time of the massacre at Cowboy Wash was very difficult for the people there.  Archaeobotanical studies of pollen and other plant remains showed that there was apparently little or no maize in or around 5MT10010 at the time of abandonment, which is quite surprising for a Pueblo site.  The plant remains that were there were mostly from wild plants such as chenopod, amaranth, and tansy mustard, all of which would have been available in the spring and likely would have been intensively collected if there were no stored corn available due to a failed harvest the previous fall.  In addition to pinpointing the season in which the incident occurred, this implies that times were very tough for the inhabitants of 5MT10010, and perhaps for their attackers too.  The coprolite showed no sign of having plant material in it, which suggests that whoever left it had not just eaten some corn at home before setting out to attack 5MT10010.

Another paper associated with the project, by Patricia Lambert, suggests another problem the Cowboy Wash inhabitants apparently had: disease.  In this paper Lambert reports on analyses of ribs of individuals at 5MT10010 and other sites in the Cowboy Wash area dating to various periods of occupation that had lesions on them suggestive of those seen in modern collections of individuals known to have died of tuberculosis and (to a lesser extent) other respiratory diseases.  These lesions were found in 11 of 32 individuals from Cowboy Wash that had enough of their ribs left to examine.  One of the individuals with lesions was from 5MT10010.  This was an adult woman who was not one of the victims of the attack at site abandonment but who had instead died earlier and been formally buried.  Lambert also examined comparative collections of remains from Pueblo Bonito at Chaco and Elden Pueblo near Flagstaff Arizona.  Only 3 of the 45 individuals from Pueblo Bonito and 2 of the 20 from Eldon Pueblo had similar lesions, suggesting that this disease was much more prevalent at Cowboy Wash than at these other sites, even though it was not absent at them.  Lambert notes that tuberculosis is an opportunistic disease that tends to strike people whose systems are compromised by other problems such as hunger and stress.  The evidence for physical violence in the Cowboy Wash sample, even setting aside the cannibalism assemblages, was much greater than in the other two samples as well.  Combined with the harsh environment, this suggests strongly that Cowboy Wash was a difficult place to live for several reasons.  Farming was possible but risky, and when conditions turned bad both hunger and violence from other hungry people were constant threats.

Given this context, the occurrence of extreme events such as cannibalism incidents at Cowboy Wash starts to make some sense.  Cowboy Wash is a place of extremes.
Huckleberry, G., & Billman, B. (1998). Floodwater Farming, Discontinuous Ephemeral Streams, and Puebloan Abandonment in Southwestern Colorado American Antiquity, 63 (4) DOI: 10.2307/2694110

Lambert, P. (2002). Rib lesions in a prehistoric Puebloan sample from southwestern Colorado American Journal of Physical Anthropology, 117 (4), 281-292 DOI: 10.1002/ajpa.10036

Read Full Post »

Pueblo Mural at Cortez Cultural Center, Cortez, Colorado

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.

Entrance Sign at Canyons of the Ancients National Monument

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.

Sleeping Ute Mountain from Yellow Jacket Pueblo

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

Read Full Post »

Sleeping Ute Mountain from Cortez, Colorado

In comments to the previous post, Graham King raises an important question: assuming that the assemblages of broken, burned, and otherwise unusually treated bones at sites like 5MT10010 at Cowboy Wash represent incidents of cannibalism, what does this mean culturally and historically?  After all, cannibalism has occurred in various contexts in many societies, including our own, and it can arise from a variety of causes.  One of the most obvious is starvation, seen most often in situations like the plane crash in the Andes made famous in the movie Alive and incidents during the settlement of the American West such as the Donner Party and the Alferd Packer case, all of which involved small groups being trapped in mountains under harsh conditions and resorting to cannibalism to survive.

So were the events at Cowboy Wash around AD 1150 the result of survival cannibalism, perhaps of members of the community who had died due to disease or starvation by other members of the community?  The excavators of 5MT10010 argued that they were not, but the evidence they gave for this conclusion is not particularly convincing.  They basically had two arguments against survival cannibalism:

  • If the individuals at Cowboy Wash had died natural deaths, the demographics of the dead should reflect the differing susceptibility to harsh conditions of different age groups, which generally means a higher death rate among infants and a lower one among older children and adolescents.  Since the 5MT10010 assemblage contained no infants but a higher than expected number of older children, it is unlikely to have resulted from natural death.
  • Unlike in the historically documented cases, the people at Cowboy Wash were not physically trapped.  There was nothing preventing them from moving away from the area if there were insufficient resources.  The southern piedmont of Sleeping Ute Mountain is pretty flat and open, rather than mountainous, and the abandonment of the sites there apparently took place in the spring, when snow would not have been an issue.

Neither of these arguments is very convincing.  The assemblage of cannibalized individuals at 5MT10010 looks demographically unusual, sure, but then again it apparently consists of only seven people.  That’s a minimum number, of course, but since the whole site was excavated and it wasn’t a large site to begin with it seems likely that the recovered bones represent most or all of the people involved.  If the assemblage represents the inhabitants of the site, which the excavators claim and I see no reason to doubt, they likely represent a family or other social unit, and it’s quite possible that they just didn’t happen to have any infants in the group at the time of the event.  In any case, with a population this small there is no reason to expect any sort of “normal” demographic profile.

The second argument is better, but still somewhat unconvincing.  It’s definitely true that the inhabitants of Cowboy Wash wouldn’t have been snowed in at the time of the abandonment event, which apparently took place in the spring, but it’s not necessarily true that they would have had somewhere else to go.  This was a time of drought and resource shortages throughout the northern Southwest, and there is considerable evidence for violence (even leaving aside the alleged cannibalism incidents) throughout the region, including at other sites in the Ute Mountain area, at the time.  Indeed, Cowboy Wash is a very marginal area for agriculture to start with, so the fact that they were there in the first place, and that they had not already left when the harvest failed, implies strongly that they had nowhere else to go.  Under those circumstances, it is certainly conceivable that some community residents might have begun to succumb to hunger or disease and others in the community may have resorted to eating them.  That’s not to say that it’s likely or the best explanation, of course.  One piece of evidence arguing against this interpretation, which the excavators mention but don’t go into much detail on, is that all the processing seems to have taken place at the same time, rather than drawn out over a period of weeks or months as might be expected if people were gradually dying and being eaten.

Sleeping Ute Mountain from Lowry Pueblo

Instead of survival cannibalism, the Cowboy Wash excavators prefer a scenario of intercommunity violence to explain the 5MT10010 assemblage.  This implies that the people who ate the residents of 5MT10010 also killed them.  This sort of explanation does fit well with the other evidence for violence short of cannibalism in the region during this period, but it is tricky to show any correlation between violent death and cannibalism because the processing of the bodies would likely eliminate any skeletal evidence of violent death.  That does seem to be the case in this assemblage, as in most other alleged cases of cannibalism.  There’s plenty of evidence of skeletal trauma, of course, but most of it seems to be directly related to the processing of the bodies shortly after death, and much of the rest seems to represent injuries incurred during life that had healed before death.  In this case, then, there is no direct evidence linking alleged cannibalism to violent death.

One piece of evidence supporting the “raid by other group” theory is that three other sites in the Cowboy Wash community, contemporary with 5MT10010, have also been excavated and all three show similar signs of cannibalism.  Since there are only ten sites in the whole community, the fact that four of them seem to have been rapidly abandoned at the same time that many people in them were eaten seems to imply that the abandonment was the result of a single event.  As the authors of the 5MT10010 paper point out there and in more detail in another article, however, there are some differences among the four assemblages and even between the assemblages in the two pitstructures at 5 MT10010 containing large numbers of bones.  Although all four sites show evidence of processing for consumption, the specific body parts that are most prevalent and the way they were processed suggest that the processing was done separately at each site and not in accordance with a common technique.  For example, at 5MT10010 the bodies found in one pitstructure appeared to have been cut into small pieces and cooked in a pot on the surface, with the bones thrown down the ventilator shaft after consumption.  At the other pitstructure, however, the bodies seem to have been cut into fewer pieces and roasted directly on a fire in the hearth inside the structure.  All this suggests that if in fact the Cowboy Wash community was destroyed by a raid, the raiders who attacked each part of it seem to have acted fairly autonomously in deciding how to deal with the inhabitants, even though they all apparently had the common goal of eating them.

This is actually rather odd, and does seem to undermine the idea that warfare was behind the cannibalism at Cowboy Wash.  It is possible, however, that raiding and survival cannibalism aren’t mutually exclusive, and that the raiders of Cowboy Wash may have attacked it specifically in order to eat the people there because they themselves were starving.  The excavators do mention that the meat from a cannibalistic attack may have been a valued “spoil of war” during harsh environmental times like this, but they generally downplay this aspect and focus more on the idea that cannibalism would have served as a terroristic tactic to strike fear in the hearts of enemies and perhaps scare them away from the area.  (The two are not necessarily mutually exclusive, of course.)  The fact that the people in the different parts of the Cowboy Wash community were apparently prepared differently does seem more compatible with the idea that the attackers were focused on eating them because they were hungry than with the idea that they were concerned with striking fear into any survivors or other nearby villages.  The terroristic theory seems to imply a ritualistic approach to the cannibalism, which is not particularly apparent in this or any other Southwestern cannibalism assemblage.  Any type of ritual cannibalism would likely have been more standardized in execution, as in Mesoamerica where ritual cannibalism was widespread.  In the Southwest, by contrast, the execution seems rather haphazard.

So, tentatively, I think the best explanation at least for Cowboy Wash specifically, and perhaps for other cases as well, is that certain communities in the Mesa Verde area during the extended drought of the mid-twelfth century AD hit upon the idea of compensating for their poor harvests by attacking other communities, not to take their stored food (since they likely didn’t have any either) but to eat their residents.  Which communities these might have been and where they got the idea remain open questions.  This idea led to a rash of cannibalism incidents around AD 1150 which subsided soon thereafter, perhaps as climatic conditions improved (or, alternatively, as other communities got better at defending themselves).  When the next major drought came, in the late thirteenth century, the idea of cannibalistic raiding does not seem to have been taken up again.  The authors of the Cowboy Wash articles take pains to note that there is no evidence of cannibalism in the Mesa Verde area after AD 1200.  Although there is plenty of evidence of violence during the late 1200s preceding the total depopulation of the region by 1300, none of this involved cannibalism.

Or did it?  There is actually some evidence from at least one site suggesting that the idea of cannibalism did not totally disappear from the Mesa Verde area after 1200.  But that’s a matter for another post.
Lambert, P., Billman, B., & Leonard, B. (2000). Explaining variability in mutilated human bone assemblages from the American Southwest: a case study from the southern piedmont of Sleeping Ute Mountain, Colorado International Journal of Osteoarchaeology, 10 (1), 49-64 DOI: 10.1002/(SICI)1099-1212(200001/02)10:13.0.CO;2-B

Read Full Post »

Sunset Crater Volcano

The effect of the eruption of Sunset Crater Volcano on the prehistoric population of northern Arizona has long been a topic of interest to archaeologists.  As I’ve mentioned recently, in the 1930s and 1940s Harold Colton of the Museum of Northern Arizona in Flagstaff came up with a theory to explain the settlement dynamics of the Wupatki area northeast of Sunset Crater.  In Colton’s view, the eruption resulted in a level of volcanic ash falling on Wupatki that acted as a mulch to retain water and make that very arid area suitable for farming for the first time, resulting in a “land rush” in which people from all over the region converged on Wupatki to farm the newly available land.  Over time, however, the ash began to blow away and the land became less productive, so people aggregated into the large pueblos for which Wupatki is best known, then left entirely when the area could no longer support its population.  Dendrochronological evidence from timbers at Wupatki Pueblo later provided a basis for dating the eruption to around AD 1064, which would put the “land rush” shortly after that.  Other evidence has shown that the abandonment of the area probably occurred some time in the thirteenth century, a time when many parts of the Southwest were being abandoned as well.

As I’ve mentioned, recent archaeological survey at Wupatki has cast doubt on some aspects of this model.  The main influx of population seems to have come after AD 1130, a few decades after the eruption, and the scale of that influx was probably quite a bit lower than Colton estimated, since many of the sites he counted to compute his population estimates were probably season field houses or other temporary structures rather than permanent habitations.  This implies that there wasn’t really a “land rush” the way Colton described it, but rather a substantial increase in population at some point after the eruption, perhaps in response to drought or other problematic conditions in other parts of the Southwest.

A few parts of Colton’s model do seem to hold up, however.  Experiments have shown that the levels of ash found at Wupatki do indeed work well as a mulch.  Without this mulch, dry farming in the area with any reasonable measure of reliability is basically impossible, since there just isn’t enough rain, and irrigation or floodwater farming isn’t possible on any substantial scale either due to the geological conditions and the lack of permanent surface water sources.  Furthermore, the Wupatki survey showed that this lack of agricultural suitability made the area essentially uninhabited before the eruption.  Of nearly a thousand datable sites recorded by the survey, only two dated to before the eruption.  The biggest influx of population came after about 1130, but there was already a fairly significant movement of people into the area in the immediate post-eruptive period.  Perhaps these people first experimented with agriculture using the ash as a mulch, and were so successful that when conditions deteriorated elsewhere others joined them.  The ash was liable to blow away in the strong winds, however, and over time the advantages it offered as a mulch would have diminished as a result of this and other factors, so it’s quite possible that it was declining agricultural productivity, perhaps exacerbated by warfare to defend land claims, that led the area to be abandoned in the thirteenth century.

Volcanic Rock in Masonry at the Citadel, Wupatki National Monument

That’s all well and good, but where did the people who moved to Wupatki after the eruption come from?  Colton saw them as coming from all over, but at least in the immediate post-eruptive period a more specific answer is tempting: perhaps they came from the area right around the volcano, which would have been rendered uninhabitable (and certainly unfarmable) by lava flows and massive ash fall.  A relatively recent paper takes a close look at the circumstances of the Sunset Crater eruption and its likely effects on local people, and basically comes to this conclusion.

From a detailed analysis of the details of the eruption, the authors of this paper found that the area of the heaviest ash fall and the largest lava flows was probably densely populated and heavily farmed before the eruption.  They cast some doubt on the tree-ring evidence pointing to an AD 1064 date for the eruption itself, but they argue on other grounds that the eruption likely took place between AD 1050 and 1100 and that it was relatively quick, lasting from a few weeks to a few years at the most.  Because the high-elevation area where the eruption took place gets more precipitation than lower-elevation Wupatki, it would have been the most favorable area for farming at the time, and a large number of homes and farms were likely buried by the lava and ash.  The amounts of ash falling right around the volcano would have been much too thick to serve as a mulch.  The ash itself is sterile, so it could only function effectively as a mulch if plants could reach their roots down through it to the soil underneath.  The few inches of ash cover at Wupatki would have allowed this, but the uplands immediately around the volcano got over a foot of ash, which would have effectively killed any agricultural potential.

Lava at Sunset Crater

Thus, the effects of Sunset Crater on local agriculturalists were two-fold: they were forced to leave a rather large and previously quite productive agricultural area around the volcano, but they were able to go to a previously unproductive area nearby that was made newly fertile by the ash.  Cinder-cone eruptions like the one that created Sunset Crater rarely cause much direct loss of life, and that would have been particularly the case in this context, since the pre-eruption populations lived in dispersed farmsteads and were probably not organized sociopolitically at any level above the household or extended family.  This would have allowed rapid reactions to the eruption, which would primarily have taken the form of migration away from the immediate area.  Since the population was so dispersed, people fleeing the ash-fall zone would likely have had relatives or friends in less affected areas to whom they could go for shelter and assistance in the immediate aftermath of the eruption.  The population movements spurred by the eruption, however, could well have resulted in groups infringing on territory claimed by others and resulting violence and loss of life.  Within this context, the relatively empty Wupatki area may have seemed particularly attractive even before its enhanced potential for farming was discovered.

Another reaction of people in the local area to the eruption, which was documented in an earlier paper by some of the same authors, was the apparent practice of placing corncobs in the path of lava and carrying the resulting “corn rocks,” with visible imprints of the cobs (which were vaporized by the heat) to rather distant settlements.  Given the amount of effort this would have required, it probably had some ritual significance, perhaps to appease the spirits of the volcano or something similar.

Mt. Trumbull from Pipe Spring National Monument

In addition to the Sunset Crater eruption, the authors of this paper also discussed a smaller and less studied eruption that likely took place about the same time at Little Springs, next to Mount Trumbull on the Arizona Strip just north of the Grand Canyon.  Here there was relatively little ash fall, so the loss of productive land and enhanced productivity of other land seen in the Sunset Crater case did not occur.  Instead, the main effect was a lava flow, with the land immediately surrounding it continuing to be largely ash-free and fertile.  The people who had lived and farmed in the immediate area covered by the lava flow would have had to leave, but people clearly continued to live and farm right around the lava, and they also built sites on top of the flow itself.  These sites have few artifacts and likely served defensive purposes, a theory that is supported by the presence of an elaborate system of trails on the lava flow that would have made it an effective refuge in times of war.  The use of defensive refuges or strongholds separate from ordinary living quarters is well-attested in the prehistoric and historic record of the Southwest.  Similar to the corn rocks at Sunset Crater, in this area there were some rocks with potsherds embedded in them, a sign of similar ritual behaviors with respect to the volcano.

These two eruptions and the different reactions to them by local populations show the effects that sudden, catastrophic events can have on human societies.  The eruption of the much larger White River Volcano a bit earlier and its effect on local Athapaskan populations in Alaska and the Yukon is another example.  Unlike many other catastrophes, volcanic eruptions are generally pretty visible in the archaeological record, which makes them a useful source of information on how societies adapt to sudden shocks.
ORT, M., ELSON, M., ANDERSON, K., DUFFIELD, W., & SAMPLES, T. (2008). Variable effects of cinder-cone eruptions on prehistoric agrarian human populations in the American southwest Journal of Volcanology and Geothermal Research, 176 (3), 363-376 DOI: 10.1016/j.jvolgeores.2008.01.031

Read Full Post »

Ledge at Wukoki Pueblo, Wupatki National Monument

Wupatki is a very dry place even by the standards of the Southwest, with annual precipitation averaging about 8 inches.  Human habitation in such an arid landscape is therefore highly dependent on capturing as much available moisture as possible.  It appears that the prehistoric inhabitants took advantage of the volcanic ash laid down over the area by the eruption of Sunset Crater in AD 1064 for farming purposes since it acted as a mulch, retaining water from the summer rains that would otherwise have evaporated in the heat and strong winds.  For other purposes such as drinking, cooking, and construction, however, water trapped in the soil isn’t very useful, so other sources needed to be found.  As at Chaco Canyon, which is similarly dry, some of this water would have come from a few springs in the area, especially in the dry season, but it would also have been useful to capture as much of the runoff from the summer rains as possible.  Due to the geology of the Wupatki area, this water could only be used for floodwater farming in a very few places, but there were other ways to take advantage of it.

One such way was apparently shown by a discovery made by two National Park Service archaeologists in the 1940s.  While out evaluating sites for stabilization needs, Albert Schroeder and Philip Van Cleave found some potsherds on the ground in sufficient number to make them think that they might be reconstructible into something approaching the original vessel.  They picked up the sherds and dug a bit into the ground beneath them to see if there were any more.  Sure enough, just under the surface of the ground there was a whole ring of sherds in place, indicating the presence of a broken but substantially complete jar that had apparently been deliberately buried.  They excavated it and took some pictures, and Schroeder wrote up a short article on the discovery for American Antiquity which was published in 1944.

Small Site on Ledge, Box Canyon, Wupatki National Monument

The jar was of the ceramic type Moenkopi Corrugated, which Schroeder dates to AD 1075 to 1275.  This is unfortunately a quite wide date range, encompassing almost the entire period of substantial prehistoric occupation of Wupatki, so it is not possible to say at what point during the occupation the jar was buried.  From its position, however, Schroeder was able to determine that it was likely placed to capture runoff from the summer rains.  It was buried in the sand underneath one of the sandstone ledges that are so common at Wupatki, so one possibility is that it was placed to capture runoff from the ledge.  Indeed, it seemed that the part of the ledge above the jar naturally collected runoff from a wide area of the sandstone outcrop.  At the time Schroeder and Van Cleave found the jar, however, the water pouring off the ledge fell somewhat short of where the jar was.  Schroeder suggested that there may have been some erosion in the period between the time the jar was buried and the time it was found, such that at this time of placement the ledge extended further out and the runoff may have poured directly onto the jar.  If this was not the case, however, the jar was probably buried with the sand level with or a bit higher than the rim, so that runoff from the sandy ground around the jar rather than the ledge above would flow into the jar.

Either way, it seemed apparent to Schroeder that the purpose of the jar was likely to collect water, which makes sense in such an arid environment.  He admitted to being somewhat unsure of the details of his proposal, and he did not venture any theories as to what the water would have been used for or why a jar was used in this way to collect it.  Obviously the amount of water in a single jar would not have been much for agricultural purposes, so I suspect the water was used for household use.  To be so used, depending on how close the household in question was (which Schroeder unfortunately did not mention), the jar could either have been dug up after filling or left in place.  In the latter case, the water could have been taken out with a ladle and transferred to a canteen or some other sort of vessel for transportation.

I don’t have any sort of major point to make about this paper, but it’s interesting as an example of the kinds of adaptations people make to harsh environments.  Wupatki would have been a hard place to live in prehistoric times, but people gave it their best shot.
Schroeder, A. (1944). A Prehistoric Method of Collecting Water American Antiquity, 9 (3) DOI: 10.2307/275790

Read Full Post »

T-Shaped Doorway at Lomaki, Wupatki National Monument

The paper by Glenn Davis Stone and Christian Downum that I mentioned in the last post, which evaluated the archaeological record of the Wupatki area of northern Arizona in the light of Ester Boserup‘s theory of agricultural intensification, was based largely on the data from an extensive archaeological survey of Wupatki National Monument done by the National Park Service in the 1980s.  This data is presented in a more complete form in an earlier paper that Downum cowrote with Alan Sullivan.  This paper looks at the previous models proposed for the settlement and abandonment of Wupatki in the context of the new data from the survey.

Cinder Cones from the Citadel, Wupatki National Monument

The most influential model for the prehistory of Wupatki has been that presented by Harold Colton of the Museum of Northern Arizona based on work done in the 1930s and 1940s.  Colton saw the extreme aridity of Wupatki as having discouraged settlement there until the eruption of Sunset Crater Volcano in AD 1064 spread a layer of volcanic ash over the area.  This ash acted as a natural mulch to retain water from the infrequent rains which would otherwise evaporate from the thin soil.  Colton looked at the large number of sites that seemed to have been built in the aftermath of the eruption and saw a “land rush” in which people from all over the local area come to Wupatki to take advantage of the improved conditions for farming from the ash fall.  Over time, however, the ash cinders began to blow away in the strong winds and the productivity of the land declined, so the people began to aggregate into the large pueblos for which the Wupatki area is best known.  Once in these aggregated villages, the poor sanitary conditions of living in such close quarters, combined with the continuing decline of agricultural conditions, forced the abandonment of the whole area some time in the thirteenth century.

Wall Abutment, Wupatki Pueblo

This is a plausible story on the face of it, but Colton’s account has been challenged more recently by other archaeologists who point out that a great many of the structures built soon after the ash fall that Colton included in calculating the population increase were small, ephemeral structures that probably served as field houses or other special-use locations rather than year-round dwellings.  This implies that Colton was double-counting both these impermanent structures and the actual permanent houses of the people who used them, thus coming up with inflated population figures on which he based his “land rush.”  The systematic nature of the survey in the 1980s provided the opportunity to determine just how many sites there really were and how many actually served as permanent dwellings.

The Citadel and Sunset Crater from Lomaki, Wupatki National Monument

As Downum and Sullivan tell it, the results basically vindicate Colton’s critics.  The vast majority of the structures found were small and relatively impermanent, with few artifacts.  In addition, a careful tabulation of sherd types at most of the sites showed that the immediate post-eruption period, far from being the land rush of Colton’s theory, was actually a time of relatively limited occupation.  There were more sites from this period than from the pre-eruption period, when the area was nearly uninhabited, but still not very many.  It was not until a few decades later, starting around AD 1130, that building began to really pick up, as indicated by both sherd types and tree-ring dates.  The high point of construction didn’t come until the 1160s, a century after the initial eruption.  (It is actually not clear how long the eruptions continued after the beginning around 1064, and there may well still have been occasional activity by the volcano this late or even later.)  Construction seems to have effectively ceased by 1220, and the area was probably abandoned not long after that.

Beam Sampled for Tree-Ring Dating, Wupatki Pueblo

The upshot of all this for Colton’s theory is that, while it does seem to be true that the ash improved the suitability of Wupatki for agriculture, people didn’t immediate act to take advantage of this.  Downum and Sullivan propose that this may have been because it took some time for the effects of the ash fall on the soil to manifest, but I think a more plausible explanation for this can be found by looking outside the immediate area to the larger region.  The decades after 1130 were a time of extensive drought throughout the northern Southwest.  This is when Chaco collapsed (or at least declined), and there were likely extensive migrations all around the region.  In this context, people may have come to Wupatki less from the “pull” factor of the beneficial effects of the volcanic ash and more from the “push” factors of drought and/or political instability elsewhere.  Of course, there were at least some people farming at Wupatki before this, so the fertility of the area may have become well known at the same time as things were deteriorating elsewhere, making both push and pull factors part of the regional dynamics.

Great Kiva at Wupatki Pueblo

In line with the arguments in the later paper by Downum and Stone, Downum and Sullivan here argue that agriculture for most of the period of occupation of Wupatki was extensive rather than intensive.  They do claim, however, that intensification came right at the end of the occupation period, after 1220, on the basis of more intensive usage of the sites from that period based on sherd counts.  This is kind of dubious, and it appears that Downum changed  his mind about it in the eight years between this paper and the later one.  Intensification at this can, however, be incorporated into the argument made in the later paper that intensification was impossible in this area due to ecological conditions.  Once people began to leave the area, perhaps spurred by increased warfare and/or continuing climatic instability, those who remained would not necessarily have been able to secure access to the large amounts of land they had had claimed earlier as part of the consolidated political groups associated with the large pueblos in the Stone and Downum model.  These few remaining farmers may then have attempted to intensify production on the smaller amounts of land available to them.  Given the aridity of the area, however, this would not have worked reliably enough to allow them to stay, so within a few decades or less they would leave as well, leaving the entire Wupatki area abandoned by 1275.  Note that this is when the famous “Great Drought” associated with the abandonment of Mesa Verde and other areas began, so the aggregation and abandonment processes associated with Wupatki may well have been different from the similar processes elsewhere in the Southwest.

Upper Walls Built on Rock Outcrop, Wupatki Pueblo

Since I’ve been taking note of the scholarly context of the papers I’ve been discussing lately, I should point out that this one is very much an archaeology paper, and a classic processual one at that, with lots of statistics and an explicit model of interactions between people and the environment.  This certainly makes it more “scientific” than, say, the later paper by Downum and Stone, which is more anthropological and not very scientific at all, but as with many such archaeological papers the scientific trappings are somewhat superficial.  This is definitely not as rigorous an attempt at quantitative social science as the economics paper on plowing and gender roles I discussed a little while ago, for instance.  I would therefore argue that this is only science in a somewhat questionable expansive sense, and not necessarily anthropology at all, despite the frequent claims of processualists to be doing “archaeology as anthropology.”  Again, however, that doesn’t mean it’s not worthwhile scholarship.  Regardless of how it’s classified, this is interesting research that can serve as a useful source of data for a variety of other studies such as the one Downum later did with Stone.
Sullivan, A., & Downum, C. (1991). Aridity, activity, and volcanic ash agriculture: A study of short-term prehistoric cultural-ecological dynamics World Archaeology, 22 (3), 271-287 DOI: 10.1080/00438243.1991.9980146

Read Full Post »

« Newer Posts - Older Posts »