The Lithic Thin-Sectioning Project

Mashantucket Pequot Museum and Research Center

Brian D. Jones, Patricia Coombs, and Karen Morando

Lithic Thin-Sectioning Methods

Images of Lithic Thin-Sections from the Archaeological Samples

Comparative Bedrock Thin-Section Samples

A Visual Comparison of Archaeological and Bedrock Samples

Introduction to the Project

This essay is based on the paper "The Creation of a Petrographic Thin-Section Library of Materials Recovered from Archaeological Sites on the Mashantucket Pequot Reservation" presented at the Geological Society of America Northeastern Section Meeting, Westin Hotel, Providence, RI, March, 1999.

By Brian D. Jones, Field Archaeology Supervisor, Mashantucket Pequot Museum and Research Center

The Archaeology Department of the Pequot Museum and research Center has begun to assemble a petrographic thin-section reference library of stone materials recovered from sites excavated over the past fifteen years on the Mashantucket Pequot Reservation, as well as bedrock materials common in Northeastern archaeological sites. This introduction begins by summarizing the lithic raw material data as it currently exists based on the visual inspection of artifacts. While certain patterns are evident in this data, there are a number of reasons why one must be suspicious of any conclusions drawn from them. This is followed by a discussion of the three-part goals of the lithic thin-section library effort. These are to identify materials to rock-type, to identify potential sources of rock types found at Mashantucket, and finally, to assess diachronic changes in lithic material use at Mashantucket and explain this within the framework of dynamic prehistoric social and settlement systems. This final goal is the crux of the effort to produce the comparative thin-section library. Because of this I will spend a short time discussing ways in which knowledge of rock sources and paths of movement can aid the reconstruction of past settlement and exchange systems. I will also discuss a number of difficulties underlying such reconstructions. I hope to make clear that rock identification and sourcing provide important information, but for that information to be imbued with meaning, it must be combined with standard macroscopic methods of lithic analysis.

I: Summary of Mashantucket Lithic Assemblages: The Record as it Stands
Close to 35,000 stone artifacts have been recovered from over 80 prehistoric sites on the Mashantucket Pequot Reservation. These sites date from at least 10,000 years ago to the present and provide one of the most complete records of Native use of a small landscape in the Northeast. Over twenty lithic materials have been field-identified among this assemblage of artifacts. The majority (41%) is made of a brown to tan quartzite, presumed to be of the relatively local Plainfield Formation. Twenty-five percent are manufactured of quartz, also presumed to be of local origin. Chert makes up over 16% of the assemblage and must have been transported over 150 km from sources to the west or north. Argillite represents just over 6% of the artifacts, while rhyolite makes up about 5%. Unidentified materials comprise over 3% of the assemblage (1200 artifacts), while seventeen or so other materials account for the remaining 3%. It should be emphasized that these material identifications are based largely on very subjective field criteria, and many are likely false. It is believed, however, that the overall trend observed in the most common materials is approximately correct.
Additional patterns are visible upon examination of diagnostic artifacts. Close to 300 projectile points have been classed to probable stone material type and prehistoric stylistic tradition. This data set provides a look at changing patterns of stone raw material use over time, at least as reflected in the manufacture of diagnostic bifaces. Based on the current identifications, a number of patterns are apparent which may shed light on prehistoric mobility, exchange and subsistence or political territories. There is good evidence of material use preference during different periods. For example, chert is the dominant Paleoindian material, quartzite and rhyolite were most commonly used during the Middle Archaic, quartz is strongly associated with Squibnocket and Narrow-Stemmed components, and argillite was the dominant material of the Terminal Archaic. More subtle patterns (such as a shift from argillite to quartzite use from the Snook Kill to Mansion Inn phases of the Terminal Archaic period) also appear to exist.
Unfortunately, until stone material identification at Mashantucket can be more objectively pursued, any conclusions drawn concerning prehistoric raw material use and movement are suspect. To rectify this, I have begun to sample selected materials from sites on the reservation and prepare petrographic thin-sections which should help in material identification, and ideally in the identification of possible source areas. Samples have focused primarily on unidentified specimens. It is hoped that their identification will lead to the most efficient use of the thin-sectioning effort in terms of field application. I anticipate that at least 200 samples will need to be prepared before we have a good understanding of some of these more difficult to identify materials. As the project continues, I hope to do additional work with the quartzites, rhyolites and cherts as well in order to identify variations in these materials and pin down a number of possible sources. A preliminary investigation of the cherts and rhyolites has been initiated.

II: The Petrographic Thin-Section Library Effort
The above observations suggest that important information concerning prehistoric mobility and exchange patterns can be gleaned from the existing lithic assemblages from Mashantucket. These have the potential to provide a wealth of information regarding both local and regional prehistoric social and settlement systems and how these have changed over time. As it stands, however, the data is at best suggestive of probable trends, but is far from indicative of them. It is evident that a more objective method of rock type identification is needed before the data can be used to its potential.
Previous studies have established the poor capacity of archaeologists to identify most rock types based on macroscopic hand sample inspection. Barbara Calogero (1992) has demonstrated an 88% misidentification rate in some tests. While these often involved "trickier" highly metamorphosed rock types such as hornfels, a number of potentially more obvious rock types such as cherts, rhyolites, and basalts were commonly misidentified as well. If the data presented above suffers from over 80% misidentification then the interpretations of the lithic assemblage are moot or downright false. If, however, at least quartz and quartzite have been correctly identified, then at least 66% of the identifications are correct, and some of the observations may still apply.
Clearly, there exists a need for a more objective method of identifying rock types than the field judgements used to date. Calogero indicated the need to "…employ more specific and accurate methods of analysis" during rock identification of archaeological assemblages (Calogero 1992:89). By creating a lithic thin-section library of materials from most sites on the reservation, we hope to meet her challenge.
The goals of the project are three-fold. First, is to create sample thin-section specimens from many sites and identify them to probable rock type. Second, is to assess possible source areas for each rock type in the collection, wherever possible. And third, is to analyze lithic assemblages with regard to the movement of materials from quarries to Mashantucket, with special consideration of changes in assemblage patterning over time. This is a long-term project, and I am currently focused on the selection of thin-section candidates from existing assemblages, the manual creation of the thin-sections, and their initial assessment. The first phase of the selection process has focused on unidentified rock types which make up 1,200 artifacts within the total assemblage of Mashantucket sites. Clearly, not all of these will require thin-sectioning. Preliminary work indicates that these are primarily rhyolites, cherts, hornfels and siltstones. I anticipate that this project will continue over many years to come. Once the data is compiled, and even after many quarry locations have been resolved, only then will the real task of assessing the data begin. While it is easy to become engrossed in lithic identification issues, it is only after the geologist has been put to bed that the anthropologist inside must rise and get to work. I wish then, to briefly review a few issues regarding the interpretation of lithic assemblages, particularly in light of issues concerning exchange and mobility.

III: Rock Sourcing and Reconstructing the Past: Prospects and Problems
While a number of excellent reviews of the difficulties of sourcing rock types in New England have been written since the nineteen seventies, less has been spoken of the difficulties of interpreting the sourcing data once it has been compiled. However, a number of exceptions to this over-generalization appear in Ellis and Lothrop's 1989 "Eastern Paleoindian Lithic Resource Use." Identifying rock types and sources is fraught with enough difficulties, that once one is reasonably sure of a rock type and source, it is understandable to hope that the information should then speak for itself in terms of its archaeological implications. Unfortunately this is far from true. Most reviews and sourcing efforts emphasize the elucidation of exchange mechanisms as the fruit of their labors. There is less open debate in the literature concerning whether this is really the prime mechanism of stone material movement in the past. This is unfortunate, because it is likely that the trade of lithics may have played a relatively minor role in past exchange systems. Binford, for example, suggested many years ago that most lithic raw materials were acquired directly during the course of annual residential moves (Binford 1979). There is also evidence that some stone quarries were in fact maintained as neutral territory to insure open access to an important distant resource (Wright 1967).
It is evident, however, that complex exchange systems existed in the past, and that some stone materials were acquired through barter or gift-giving. The presence of a distantly sourced material on an archaeological site has profoundly different implications for the reconstruction of past patterns of mobility, territoriality, social interaction spheres if it was acquired through exchange or quarried directly by the site's occupants. It is critical, therefore, to be able to distinguish between materials acquired directly and those acquired indirectly through exchange before one can benefit from information gathered through rock type sourcing. Rock sourcing can tell us that a material moved from here to there, but says nothing concerning the mechanism of that movement (Webb 1974: 360). I believe that it is precisely the mechanism of movement that most concerns anthropologically minded archaeologists.
How does one distinguish between exchanged and directly acquired elements of a lithic assemblage? The task is certainly not simple, but I will suggest a few approaches. I encourage debate over these and anticipate that some may prove less helpful than others.

  1. I expect that materials acquired through exchange should never represent the majority of a lithic assemblage at habitation sites. To become dependent on exchanged stone strikes me as economically risky behavior not typical of hunter-gatherer societies. At what point a material becomes uncommon enough to be a possible trade item is unclear, however. Is it 20%, 5%? - this matter requires further investigation.
  2. Because of the lower abundance of exchanged lithics, debitage associated with them should be poorly represented, and when present should be typified by small retouch flakes. (This may not apply well to trade cores, however, whose intended product was large, usable flakes.)
  3. Exchanged goods are expected to be acquired as special forms quite visible to archaeologists, for example, as finished points, prepared blanks, or prepared, easily transportable cores.
  4. Exchanged goods may receive special treatment, that is, they may appear more commonly in caches and burial contexts.
  5. When present on habitation sites, tools manufactured from exchanged materials are expected to be heavily curated, especially if prestige is attached to the possession of such objects.
  6. In situations where the society is stratified, the distribution of exchanged goods is expected to be uneven (though there is risk of circular argument-making here).
To evaluate any of the above requires the careful examination of full lithic assemblages in terms of the quantities of various materials represented (1), debitage analysis by material representation, size and type (2), typological analysis (3), tool rejuvenation and curation assessment (5), and spatial distribution analysis (4, 6).

Looking at the data from the Mashantucket diagnostics, for example, it is reasonable to ask at what point one might consider chert to have become a trade commodity rather than a material acquired during the course of annual residential movements between resource areas. During the Paleoindian period the great majority of tools recovered are manufactured from chert. Based on the above criteria, this material was most likely acquired directly by the sites' occupants. The quantity of debitage at the Hidden Creek site supports the hypothesis that this material was acquired in some abundance and transported rather rapidly from its source. Numerous tools and large flakes were discarded in a useable state without evidence of heavy curation or an attempt to preserve what might be seen by the archaeologist as an uncommon material. There is a sense that the site's occupants had a very secure feeling that there was "a lot more where that came from." During the latter Early Archaic, chert is still used in the manufacture of nearly 50% of the projectile points from Mashantucket. This suggests that it was still likely being acquired directly during the course of seasonal movements. By the early fifth millenium BP Laurentian phase of the Late Archaic, chert was used in only 36% of the diagnostics recovered. This suggests that while it may still have been acquired directly, the source area was perhaps becoming more marginal. By the late fifth millennium, during the Narrow-Stem phase of the Late Archaic, chert becomes so uncommon that its presence is more simply explained as an expression of trade with neighboring peoples. Assuming that this material was being acquired primarily within or west of the Hudson River drainage, the regional resource territory appears to have shrunk to within 100 km from Mashantucket at this time.

IV: Summary and Conclusions
The thrust of this paper is simple. Petrographic thin-section data can provide archaeologists with information critical to better understanding prehistoric settlement and exchange dynamics. The information acquired by such data is, however, of limited anthropological value without a multi-faceted approach that includes a variety of lithic analysis methods. Stepping back from the microscope and examining the full assemblage of associated materials provides the opportunity to convincingly argue whether stone from a site was acquired through direct means by the site's inhabitants or was acquired indirectly through exchange. Knowing this, it is then possible to provide a context for the site's artifacts in terms of the social dynamics, residential mobility, subsistence region, and possibly territory of its past occupants.

Binford, Lewis
1979 Organization and Formation Processes: looking at curated technologies. Journal of Anthropological Research 35(3):255-273.

Calogero, Barbara
1992 Lithic Misidentification. Man in the Northeast 43: 87-90.

Ellis, Christopher J. and Jonathan C. Lothrop (eds.)
1989 Eastern Paleoindian Lithic Resource Use. Boulder: Westview Press.

Webb
1974    Exchange Networks: Prehistory.  Annual Review of Anthropology 3:357-383.

Wright, Gary A.
1967. Some Aspects of Early and Mid-Seventeenth Century Exchange Networks in the Western Great Lakes. Michigan Archaeologist 13 (4): 181-197.