Archaeology Southwest - Exploring and protecting the places of our past

The Center’s Current Research: Kayenta and Salado Migrations

Jeffery Clark, Center Preservation Archaeologist
Deborah Huntley, Center Preservation Archaeologist
J. Brett Hill, Hendrix College and Center Research Associate
Robert Jones, Center Preservation Fellow and University of Arizona Graduate Student

One of the few places in the southern Southwest where large populations are evident in the mid- to late 14th century is the greater Upper Gila region, including the Upper Gila River, its tributaries, and the Mimbres River Valley. We propose that a late prehistoric migration corridor followed the Gila River upstream into southwestern New Mexico. Our current research is directed at understanding the nature and timing of waves of migration, as well as their impact on community organization and identify formation.

Field Research at the 3-Up Site

One component of our research involves test excavations at a few poorly understood Salado sites in the greater Upper Gila region. One of these is the 3-Up site along Mule Creek at the Arizona-New Mexico border. This large, multi-component site was occupied as early as the Late Archaic period (2100 B.C.–A.D. 50; also known as the Early Agricultural period in the Tucson Basin) and continuing through the Cliff phase (A.D. 1300–1450). The site had a large Classic Mimbres (A.D. 1000–1050) component, and may have been occupied through the 12^th and 13^th centuries. Significantly, the site sits practically on top of what appears to be the largest and most widely circulated Upper Gila obsidian source in the period preceding European contact.

Initial work at the site was conducted by researchers from Arizona State University (ASU) as part of the Mogollon Prehistoric Landscapes Project, and we deeply appreciate their willingness to share information and help us make initial contacts in the area. The ASU researchers mapped the site and excavated four test units, including two along a bulldozer cut in the main mound.

In the summer of 2008, the Center and Hendrix College jointly conducted a field school at the site. Our goal was to maximize spatial coverage and expand upon ASU’s work, with particular attention to documenting the site’s latest occupation episodes. In keeping with the Center’s focus on preservation archaeology and minimizing site impacts, we set out to sample trash concentrations outside of—but still associated with—structures. We excavated eight test units in the four loci where we anticipated late deposits, each near evidence of adobe architecture. Four of these units contained variably dense and deep trash deposits, and one (Unit 103) had a very productive, deep midden beneath the remains an adobe structure—we exposed just the corner of a room with a section of eroded floor.

One area, Locus C, proved to be the most interesting, if not exactly what we expected. We had some trouble finding good midden deposits here, as two of the units ended up being in relatively shallow sheet trash. In Unit 106 we encountered roof fall just above a well-plastered floor at about 30 cm below surface. There were several crushed reconstructable vessels and a perforated plate in contact with the floor. One of the crushed vessels is Dinwiddie Polychrome, a late, smudged-interior type of Roosevelt Red Ware recently defined by Patrick Lyons. We also recovered several beam portions from the roof fall and from floor contact. We are awaiting the results of analysis of these and other dendrochronological samples collected by ASU.

Artifact Analyses

Analysis of obsidian from the site and surrounding region is ongoing, but we have some preliminary results and some hypotheses that remain to be verified. First—and consistent with its location near the large Mule Creek source area—abundant obsidian is found on the surface of the 3-Up site and in the adjacent creek bed. Obsidian comprises more than half of the flaked stone assemblage. Together with the obsidian distribution patterns we have already identified, this provides further support for our idea that the Mule Creek area was a node in the obsidian distribution network controlled by Kayenta colonies beginning in the late 1200s.

The ceramics are also quite interesting, and include many Roosevelt Red Ware types (Gila, Cliff, and other late variants) consistent with a Cliff phase occupation (A.D. 1300–1450). There are also a smaller number of Classic Mimbres types, White Mountain Red Ware, El Paso Polychrome, and some Playas Red Incised. Notably, Maverick Mountain Polychrome, one of the hallmarks of Kayenta migrant groups, occurs in relatively high frequencies—in fact, it seems to be more common than expected based on our understanding of other collections. ASU researchers found none in the areas they excavated.

Current Working Model

Based on our previous research and very preliminary data from our investigations at the 3-Up site, we have developed a model incorporating at least two major waves of migration. The first was a Kayenta migration in the mid- to late 1200s; the second was a more substantial late 14^th-century Salado migration from eastern Arizona into the Upper Gila.

We think that the greater Upper Gila region was essentially an empty frontier in the mid- to late 13^th century. The first wave of Kayenta immigrants to the region likely formed small colonies where they continued to practice their traditions in relative isolation. At this point, we speculate that the Mule Creek colony was more substantial than we suspected it would be, based on previous work in the area. It is difficult to spatially resolve the migrant footprint, and we really need a better idea of the size of the pre-migration population following the Classic Mimbres (A.D. 1000–1050) occupation to be able to adequately evaluate this.

The second wave was focused most intently on the Cliff Valley within the Upper Gila. We clearly have evidence for this in the Mule Creek area, too, as both we and the ASU researchers found evidence for Cliff phase occupations (A.D. 1300–1450) just about everywhere we looked at the 3-Up site. We still don’t understand how substantial that late migration was, however, given the superposition of episodic occupations and site disturbance. We are very interested in the smaller, spatially discrete occupation of Locus C, which we think was ONLY occupied during the Cliff phase. Recall that we found late Roosevelt Red Ware types, a perforated plate, shallow trash deposits, and an adobe room very near the surface.

Comparisons of regional demographic trends are somewhat hindered by the fact that very late occupations may remain poorly documented because newly recognized late Roosevelt Red Ware types have not previously been differentiated from Gila Polychrome. There is also a real need for tree-ring dates to more securely establish dates for these newly recognized types.

Next Steps

First, we would like to determine if there was really an occupational hiatus in the Upper Gila/Mimbres area in the 1200s. How large were resident (Black Mountain or Tularosa Phase) populations in the area when the first wave of Kayenta migrants arrived? Local population size would obviously have implications for the challenges and opportunities available to migrant communities.

Second, we’d like to better understand the nature of 14^th-century migration and community organization in the greater Upper Gila region. We infer—based on the evidence for an archaeologically visible Kayenta migration sometime in the 1200s—that a relatively large population was already in place, at least at Mule Creek, when later groups arrived. How did the later Salado migrants interact with the Kayenta colonies established only a few generations earlier?

We see the Upper Gila area, and the Mule Creek area in particular, as an ideal test case for examining concepts of diaspora—shared culture separated by space—and coalescence—shared place separated by culture. Over the next several years we will be assessing the relative strengths of Kayenta diasporic and Salado coalescent connections among and within Upper Gila region settlements using several lines of evidence:

• improving regional and site-level chronologies
• evaluating ceramic type frequencies indicative of variable social networks
• sourcing late Roosevelt Red Ware types using INAA and petrography
• stylistic analyses of these late types
• analysis of utility ware technological attributes to assess technological styles used by closely interacting groups of potters
• further obsidian sourcing studies

Some of the questions we hope to answer include:

• During the demographic upheavals of the late 1300s, were migrations an attempt to reconstitute a Kayenta diasporic community in the Upper Gila region that had been variably altered by contact (or lack thereof) with local groups?
• Conversely, did this migration generate myriad coalescent communities comprised of settlements each with a distinct identity?
• Do we see a proliferation of local variants in Roosevelt Red Ware in the late 1300s? Do these variants reassert Mogollon or Pueblo affiliations in technology as larger traditions break down?

Finally, there has been much speculation about Salado contributions to protohistoric Pueblo groups. In a sense, the 1300s Salado migration represents one step in a return migration to the Pueblo world that may eventually be traced to the protohistoric Hopi, Zuni, and Rio Grande pueblos.

Perhaps this will be the next phase of the Center’s research…