Atomtrl2.gif (2767 bytes)

A report from the Archaeological

EDXRF Laboratory

University of California, Berkeley


M. Steven Shackley, Ph.D.

Phoebe Hearst Museum of Anthropology

University of California, Berkeley


Report Prepared for

The Roosevelt Platform Mound Study

 The Office of Cultural Resource Management

Department of Anthropology

Arizona State University


16 September 1996


The page is best viewed with Netscape 2+






Location of the Tonto Basin and Arizona topography











Site groups and specific features in the Tonto Basin








Obsidian marekanites (upper left), flakes, and projectile points from project sites
(approximately actual size at 1024X768 screen resolution)


Recent chemical analyses of archaeological obsidian recovered from pre-classic and classic period contexts in Tonto Basin indicated a simple spatial dichotomy of obsidian source procurement. This strategy included obsidian derived from either the Superior (Picketpost Mountain) source to the south in presumably Salado held territory, or Coconino Plateau sources likely procured from the Sinagua. This much larger sample (n=178) of analyzed obsidian artifacts reported here indicates a much more complex procurement of obsidian in pre-classic, and particularly classic periods. Settlements in the southern portion of the basin (the Salt arm) are dominated by obsidian procured from the Mogollon or San Carlos Valley in southeastern Arizona or western New Mexico, while the settlements in the northern portion of the valley (Tonto arm) are dominated by obsidian procured from both the Mogollon/San Carlos Valley regions and the Hohokam core area. Both areas exhibited obsidian derived from Superior (Picketpost Mountain) and the Coconino Plateau. These data suggest the probability of considerable intra-basin exchange as well as differing interaction with external groups signaling possible differences in kin and social organization. This web site documents the analytical methods used for this study followed by a discussion of the results.

To go to a discussion of the analytical methods

Further information on the laboratory instrumentation can be found on the World Wide Web at:




Archaeological knowledge of Early Classic and Classic period obsidian procurement in the Southwest is coming of age (Bayman 1995; Mitchell and Shackley 1995; Peterson et al. 1996; Shackley 1995; Simon et al. 1994). The complexity of obsidian exchange both within an intra-site kin group and a regional system is greater than could have been predicted even five years ago (Shackley 1995). The rather large assemblage analyzed here is certainly as complex as the Classic period Pueblo Grande collection from the Phoenix Valley, although the pattern of procurement indicated is quite different (Mitchell and Shackley 1995; Peterson et al. 1996). The discussion here will first center on the source provenance in the assemblage as a whole and follow with a more focused discussion of the intra-site comparisons in the basin proper.

Source Assignment (data plots are hyperlinked below)

Source discrimination was made by comparison with source standard data at Berkeley reported most recently in Shackley (1995). Various bivariate and three-dimensional plots are presented to graphically distinguish the sources. Since the number of sources in the assemblage was relatively high, four plots are needed to discriminate the sources. The 3-D plot is helpful in discriminating Government Mountain from Cow Canyon, but the rest of the source data are clustered in the Zr wall, although Superior clusters at the bottom. Two simple bivariate plots of Rb, Sr, Y, and Zr effectively separate the sources. Cow Canyon which is difficult to distinguish between Sauceda Mountain samples in Y/Sr, is easily separated in the Rb/Zr plot. Zirconium is consistently higher in Sauceda Mountains obsidian (Shackley 1995). The Mule Creek specimens are difficult to separate with these elements (barium, not measured here, is a more effective element in this regard), but this is not necessarily an issue in this study, both due to the distance to source, and the presence of all these sub-sources in the Gila River alluvium (Shackley 1992, 1995). The Mule Creek/North Sawmill Creek chemical group is always distinctive based on very high rubidium. Refer to the source standard table in Shackley (1995:538-543) for elemental compositions.

Source Provenance

The small sample first analyzed from the Livingston Group (n=17) was certainly not representative (Simon et al. 1994). This analysis indicated only Coconino Plateau and Superior sources in the assemblage, suggesting contact with only Sinagua and Salado groups outside the basin area. The raw data (231K) or crosstabulation (29K) and figures show the distribution of obsidian source provenance basin wide. Quite apparent, is that while Superior the nearest source still dominates as a single source (35.1%), obsidian from the Mogollon/San Carlos Valley (Mule Creek sources and Cow Canyon=42.5%), and the Hohokam Sonoran Desert Sources (Sauceda Mountains and Vulture=12.7) actually contribute more obsidian to the collection. Coconino Plateau obsidian (Government Mountain and RS Hill/Sitgreaves), while present in sites in both the Tonto and Salt arms of the basin comprise only 9.2% of the basin assemblage.

To go to the data table (careful it's 231K)!

or better yet a crosstabulation of site by source only 29K

The Livingston, Schoolhouse Mesa and Mound, and Rock Island sites contained most of the Superior material, but substantial proportions of the Mule Creek and Cow Canyon obsidian as well as three specimens at Schoolhouse derived from the Sauceda Mountains source also occur. The Cline Mound and Cline Terrace sites in the Tonto arm exhibited the greatest diversity of archaeological obsidian sources in this analysis. Obsidian from Superior (20.5%), Mule Creek/Cow Canyon (43.2%), the Coconino Plateau sources (13.6%), and the Sonoran Desert sources (21.6%) are relatively evenly distributed compared with the Salt arm sites that exhibit a greater proportion of Superior (70.5%), and virtually no obsidian most likely derived from the Gila/Salt Basin Hohokam.

Obsidian Exchange in Tonto Basin

It is important to stress that the following inferences are, at this point hypothetical, since the pre-Classic and Classic material is pooled and other data sets (i.e. ceramics or architecture) are not taken into consideration. Given the proportion of obsidian source procurement evident between sites in the two arms discussed above, it appears that the inhabitants of the Cline Terrace area had access to the greatest number of sources, either because exchange partners or kin ties were more diverse, the control of intra-basin exchange was in their hands, or more likely a combination of both. Additionally, exchange with the Hohokam to the south was much more common here compared with the Salt arm where obsidian from the Mogollon/San Carlos Valley suggests greater ties to that region. It is certainly possible that at least some of the Mule Creek/Cow Canyon obsidian in Cline Terrace sites was derived from intra-basin exchange with partners on sites on the Salt arm. This is difficult to discern with only the obsidian data.

It is tempting to posit an exchange model where the Cline Terrace groups in the Classic Period were more tightly aligned with the Hohokam of the Phoenix Basin, while the groups in the Salt arm were more closely aligned with the Mogollon/San Carlos Valley groups. The evidence for contact with groups "outside" their exchange partners posited here could be due to intra-basin exchange between the Salt and Tonto arms.

Recent analysis of obsidian from Desert Archaeology's work in Early Classic contexts along Tonto Creek just north of Cline Terrace indicated a very similar obsidian procurement pattern as the Cline Terrace material (Shackley 1996). These sites exhibited a mix of obsidian from Superior, Vulture, Sauceda, and the Coconino Plateau sources, but none from the Mogollon/San Carlos Valley, although one obsidian projectile point was produced from obsidian procured from the Valle Grande source in northern New Mexico (Shackley 1996:8). While the sample was relatively small (n=25), the presence of obsidian derived from Hohokam sources in the Sonoran Desert and the absence of Mogollon/San Carlos Valley sources could be rallied as evidence favoring the assumption that the Cline Terrace groups procured this obsidian through intra-basin exchange. This, however, is obviously weak evidence; other data sets will have to be played against these inferences. Parenthetically, all these sources of obsidian are essentially equally good media for tool production, so raw material inequality is not an issue.

Importantly here, there is evidence suggesting differential procurement of obsidian for groups inhabiting the Salt versus Tonto arms. Whether this was part of a larger kin and social network is difficult to determine with the obsidian studies alone. I would suggest that it is, and individuals or kin leaders employed different regionally based kin agreements in the Salt arm than did their neighbors in the Tonto arm. This is interesting, given that many of these sites were likely inhabited contemporaneously, and certainly were in contact if not consanguineally related. This is not unexpected given very similar patterns evident at Classic platform mound complexes in the Phoenix and Tucson Basins (Bayman 1995; Bayman and Shackley 1996; Mitchell and Shackley 1995; Peterson et al. 1996; Shackley 1995). In some cases it appears that individuals or kin groups in the platform mound blocks had very different access to obsidian, and likely other resources, than individuals or kin groups outside of the platform mound. Indeed, those outside the "confines" of the platform mound appear to have access to a greater diversity of sources. Something very similar seems to be occurring in the Tonto Basin, at least between the groups living on the northern and southern portions of the basin. Both Bayman (1995) and McGuire and Saitta (1996), for different reasons, have suggested that both an elite community and an "egalitarian" community could have existed in the Classic and post-Classic period. While this does not directly explain the procurement discussed here, it may provide a framework for further investigation, and can explain differences between the distribution of obsidian sources in a platform area that are different from areas outside the platform. In any event, the analysis of archaeological obsidian here offers, as usual, more questions than answers.

To cited references

This page maintained by Steve Shackley (shackley@berkeley.edu).

Copyright 1997 M. Steven Shackley and Arizona State University . All rights reserved.
Revised:20 March 2015

All materials appearing on this Web (http://www.swxrflab.net//tontobs) may not be reproduced or stored in a retrieval system without prior written permission of the publishers, except for educational purposes, and in no case for profit.