A definition of raw material analysis is required. First it is important to understand that a raw material analysis may be conducted at different levels. The first and most fundamental step is description: identifying the raw materials represented in an assemblage, determining their relative abundance, and determining the relative abundance of such characteristics as heat treatment or the presence of cortex. The second step is establishing raw material context: examining the geographic origin of the raw materials in relation to where the assemblage was found. The next step is establishing archaeological context: comparing the assemblage with other assemblages, whether from nearby or distant sites. The final step is interpretation based on the above information: discussing how raw materials were procured, circulated, modified and utilized.
The analysis must include identification of the raw materials represented in an assemblage and their relative abundance; it should also include whenever possible an evaluation of the relative abundance of characteristics such heat treatment and presence of cortex. Publication of this type of descriptive information in a report provides useful raw data for other investigators, even if no additional interpretation is attempted. However, once this set of descriptive information is in hand, examining raw material context is usually fairly straightforward. This paper's initial discussion of raw material resource regions is intended to support such an examination. At the most basic level, establishing raw material context allows the raw materials represented in an assemblage to be characterized as local, nonlocal or exotic (or sometimes indeterminate). The direction and relative distance of sources for nonlocal and exotic raw materials can also be discussed on a basic level. Additional detail can be added to such an evaluation, as appropriate or possible, by considering natural raw material distribution in more detail. This can be accomplished by referring to additional literature sources (such as some of the papers cited above) or by undertaking additional field research on the natural distribution of one or more raw materials.
The most interesting step is usually establishing archaeological context by comparing your assemblage with other assemblages, whether from nearby or distant sites. This step is not always possible or practical, because it requires collecting a body of comparative data either by literature research or undertaking a descriptive raw material analysis of other assemblages. Both of these options can be time consuming. In addition, it can be difficult to find useful comparative information in published reports because identification and description has not been based on any consistent set of raw material categories. As better information on Minnesota raw materials has become available in the last few years, however, consistency and comparability have improved.
The final step, discussing how raw materials were procured, circulated, modified and utilized, is beyond the scope of many site reports. In addition, many assemblages do not provide an adequate sample to support this kind of discussion. Such interpretations can, however, be built over a longer course of research with reference to many collections. Among other results, such a research program should allow definition of technological "industries" -- patterns of lithic raw materials procurement, circulation, modification and utilization with specific geographic and chronological limits.
It should be noted that raw material analysis, as discussed here, is limited to chipped stone technology. It does not extend to groundstone, carved stone or other stone technologies, and does not normally extend to the related but separate issues of technology and typology. It is also distinct from the more specific task of raw material sourcing. Any of these other analyses, however, may serve to complement a raw material analysis.
Methods. The approach I advocate is based on identification of lithic raw materials by macroscopic visual characteristics, supplemented by occasional examination under low power magnification (hand lens or binocular dissection microscope).(4) In part this is a practical consideration. Macroscopic examination is the method available to and utilized by most archaeologists in the course of a regular site analysis. I also believe that this approach can provide substantially accurate results in most cases. This identification technique depends on consideration of multiple characteristics, including apparent color, color of transmitted light, fracture surface characteristics, color patterning, inclusions, relative opacity or translucency, cortex characteristics, texture, and alterations produced by heat treatment. The use of several characteristics provides an adequate "multidimensional" conceptual space within which most raw materials may be separated. The approach is not unlike that used by geologists to distinguish various formations and members. And, critically, this approach serves to establish a broad research framework, within which more explicit (and expensive) identification techniques can be utilized to best result. Such methods, including trace element analysis and thin section analysis, can provide a valuable complement. The important consideration, as is always the case in archaeology, is to recognize and respect the interpretive limits of the resulting information.
I also advocate making distinctions and identifications within the context of regional raw material availability. Consider where the assemblage was found in relation to the lithic raw material regions discussed above. Because most sites contain a high percentage of locally available raw materials, this will give you a good indication of several raw materials you should encounter at the site. For example, if a site is located in west central Minnesota (i.e., well within the Western Resource Region), you would expect to find substantial amounts of Swan River Chert and Red River Chert, together with smaller amounts of Tongue River Silica, etc. In contrast, if a site is located in central Minnesota (near the transition between the Western and Eastern Resource Regions), in addition to the materials just mentioned you might expect to find Jasper Taconite, siltstone and other eastern materials.
Conducting a raw material analysis in the context of regional raw material availability also helps with another difficulty commonly encountered in a raw material analysis: distinguishing between similar materials. For example, Red River Chert (and, for that matter, essentially every other lithic raw material) has a range of physical characteristics. Occasionally, RRC may resemble some forms of Prairie du Chien Chert, Galena Chert, Hudson Bay Lowland Chert or possibly other materials. For this example, we will assume that we are examining the lithic artifacts from a site in northwestern Minnesota. The site contains Swan River Chert, Rhyolite, and Tongue River Silica. There is also a set of artifacts definitely identified as Red River Chert and another set of artifacts which may be Red River Chert. Note first that there are not positive identifications of distinctive southern materials. This suggests that the questionable material is more probably Red River Chert than Prairie du Chien Chert or Galena Chert. If there had been clearly identified pieces of some southern material (such as oolitic Prairie du Chien Chert), than the questionable material would require more careful examination to see if it might be nonoolitic PDC. Stated in another way, if a particular artifact might be plausibly identified as either a local or nonlocal materials, it is preferable to identify it as a locally available material. This avoids implying trade networks, long distance travel or other phenomena which cannot be adequately substantiated.
None of this is meant to imply that sites will normally not contain raw materials from outside the lithic raw material region in which they are located. In fact, this is very common. But the abundance of a particular raw material at a site is related to its availability, and therefore "locally" available raw materials are usually more common as artifacts. In this context, exceptions to the pattern are all the more remarkable and deserve careful consideration and explanation. They should also be subject to more rigorous proof.
When conducting a raw material analysis, it is very important to remember that the goal of the analysis is to produce a raw material characterization of an entire assemblage. Even though you are working with individual pieces, your goal is not to produce a list which accurately identifies each piece in terms of raw material. This is not practical; in fact, it is not possible -- except in rare cases. Lithic raw materials are too variable in character and there is too much overlap between the characteristics of various materials. This problem is accentuated as individual pieces become smaller. There is a point where a piece is too small to give an adequate indication of the characteristics of the constituent raw material. At this scale, many raw materials cannot be distinguished from one another with any degree of accuracy.
Thus absolute accuracy is not possible. If, however, you think of your task as characterizing an entire assemblage, you have set an achievable goal. You will be able to produce, first of all, a substantially accurate list of raw materials represented in the assemblage. Second, you will be able to produce a substantially accurate evaluation of the relative proportion of each raw material. This is due in part to the fact that examining many pieces gives you a better idea of the range of characteristics represented; the range of characteristics represented among the pieces compensates in part for the lack of range apparent in individual pieces. It is further due to the fact that a few individual errors of identification will not significantly affect your results. Misidentifying one or two dozen pieces in a collection of several hundred pieces, for example, results in an acceptably small margin of error. As long as most pieces are identified accurately, the misidentification of a few pieces does not significantly change the total characterization.
An assemblage may consist of all the lithics from a site, or of some subset such as all the lithics from a single stratum or feature. When sorting and identifying raw materials, it is best to work with an entire assemblage at once. If you can, dump the entire assemblage -- tools, flakes, shatter and other pieces -- in a single terrifying heap. Then start sorting This produces betters result than examining pieces individually, in serial, as happens during regular cataloging. Sort by placing materials with similar characteristics together. Having pieces side by side helps in comparing them more effectively. Make the easy identifications first; set aside the difficult pieces. When you come back to them later, you will be more familiar with the characteristics of raw materials present in the assemblage and many of the difficult pieces will no longer seem so mysterious.
As you sort, remember that your decisions are not final. You can easily change your mind. And as you work, think in terms of tentative raw material identifications for each batch you have separated. You can also easily change your mind on these identifications. In making tentative raw material identifications, consider where the site is located. What are the locally available raw materials? In most cases, these will constitute a substantial portion of the material you are examining. This is especially useful in distinguishing similar materials. For example, Galena and Red River may both be opaque, slightly chalky, light colored cherts. But if you are conducting a raw material analysis of a site in southeastern Minnesota, a chert of this appearance is more likely to be Galena. At a site from northwestern or west central Minnesota, such a material is more likely to be Red River Chert. Consider the raw material resource context of the site and allow this to guide your identifications.
Also consider which nonlocal raw materials are likely to be found, on the basis of proximity to the sources or on the basis of what has been found and identified at other sites in the region. For example, Knife River Flint is more common at sites in western Minnesota, Hixton Quartzite at sites in eastern Minnesota, and Burlington Chert at sites in southern Minnesota. The presence of exotic raw materials is influenced by, although certainly not completely limited by, proximity to the source of the exotic raw material. After all these procedures, it may be that you will still encounter other nonlocal materials. But identification of these materials should be made only on the basis of good evidence, after other more probable identifications have been ruled out.
As you work, compare the texture of fracture surfaces. Many materials have a smooth break, while others have a rough break. These textures can be very distinctive. For example, Knife River Flint usually has a smooth, somewhat satiny fracture surface that shows individual flake scars fracture features quite clearly. Gunflint Silica has a fracture texture not unlike a broken ice cube. Swan River Chert has a fracture surface reminiscent of orange peel, and individual flake scars are distinguished only with difficulty on this raw material.
It is better to view materials in the light from an incandescent lamp than under fluorescent light. The incandescent bulb provides a fuller spectrum of light, thus better approximating natural light and allowing the color of the material to be seen more accurately. Bright, diffuse daylight (not direct sunlight) is better yet.
Try holding each piece up to a bright light so you can see the "transmitted" light coming through the raw material. The color of the light as it comes through the material can be diagnostic. For example, the translucent variety of Cedar Valley Chert usually shows a golden or yellow color. Hudson Bay Lowland Chert shows a distinctive reddish color. Viewing the light coming through an artifact will also help you determine how opaque or translucent the material is, and reveal the presence of inclusions. Diagnostic characteristics such as these are discussed below in the raw material reference list.
Access to a comparative collection of lithic raw material samples is imperative. Even a small collection can be extremely helpful. A large collection with multiple samples of each material is better because it allows you to get a clearer idea of the range of variation for a given raw material. Many of the raw material characteristics which you will observe are rather subtle and difficult to remember with sufficient precision. Refer to the samples periodically.
Remember that overheating, especially burning, can greatly alter the appearance of a raw material, sometimes to the point where it cannot be specifically identified. At many sites a classification such as "burned chert" may be useful.
When your raw material identification is nearly complete, you should still have a few unidentified pieces. Nature is not categorically tidy: there is simply too much variation in each material and too much overlap of characteristics between materials to confidently identify each piece. Don't expect to be able to do this, except in extraordinary cases. Depending on how many unidentified pieces there are and what percentage they constitute of the total assemblage, there are various ways to handle them. If there are just a few unidentified pieces, leave them in a single category. If you have a larger number of unidentified pieces, you may wish to sort them into smaller groups on the basis of observed characteristics. It is preferable to use as few additional categories as possible. The proliferation of nonspecific categories provides little useful information and tends to make an analysis more confusing.
Unidentified pieces may be included in an analysis under general, descriptive terms rather than under specific material names. Use of general terminology allows an unidentified material to still be classified and described in useful fashion. It also provides an alternative to guessing at raw material identifications in order to fit all samples into specific categories. Match the specificity of your terminology to your degree of certainty. Consider using broad terms like "chert," fossiliferous chert," "other silicates," etc.
When you have separated the various raw materials, you should sort each individual raw material into subcategories according to whether cortex is present or absent and, when possible, whether or not the material has been heat treated. This will produce up to four subcategories for each raw material: no cortex, not heat treated; no cortex, heat treated; cortex, not heat treated; cortex, heat treated. These subcategorizations can be useful in examining reduction strategies. Examples are discussed below.
Cortex is usually not difficult to identify, even though the appearance of cortex varies from one raw material to another. Cortex refers to the weathered or otherwise distinctive exterior surface of a cobble or other block of raw material. It may be a distinct, separate rind which varies in color, texture or other characteristics from the chert on the interior of the cobble. It may also be a chemically and mechanically weathered surface on the chert itself. It may even consist of adhering parent material such as limestone. Because of this variability, a reference collection that includes whole cobbles can be helpful. Even without this kind of reference material cortex is usually not difficult to identify. There are exceptions, of course. Siltstone is an example. This raw material tends to develop a heavy patination which may be difficult to distinguish from cortex.
Heat treatment can be more difficult to identify. The effects of heat treatment vary from one raw material to another, and also vary from blatant to subtle to undetectable. Materials like Tongue River Silica and opaque Cedar Valley Chert undergo obvious color changes. Swan River Chert exhibits more subtle but still detectable changes in texture. Siltstone shows no noticeable changes whatsoever. As a rule, some combination of changes in color, texture or luster results from heat treatment. Many raw materials will redden in color, either subtly or dramatically, although other color changes occur. Mottling or other color patterning may be enhanced or diminished, and fossils or other inclusion may become more or less obvious. Many pieces become waxy or even glassy when they are heated. The degree of change depends partly on the individual raw material and partly on the maximum temperature reached. Changes in luster may be seen by comparing heated and unheated specimens. Sometimes differences can also be seen on a single piece between fracture surfaces created before and after heating. In some cases heating might also produce changes in translucency or other characteristics. Given this kind of variability, I recommend two ways of learning to recognize heat treatment. First, examine the range of variability for each raw material in an assemblage and look for consistent variations in color, texture, luster or other characteristics. Sometimes the differences between heated and unheated pieces will be obvious. Second, refer to experimentally heat treated samples. You might even attempt heat treating samples yourself, using a kiln, a fire or even the kitchen stove. Temperatures of 400o to 500o F are enough to produce changes in many raw materials.
You will not be able to identify heat treatment on some raw materials. With other raw materials, you will not be able to consistently identify heat treatment. In such cases it is adequate to state in your descriptive analysis that identification of heat treatment was not attempted for a particular raw material. Also remember that overheating or outright burning can produce radical changes in appearance, often to the point where the raw material cannot be identified. The most obvious signs of overheating or burning are crazing (a network of fine cracks across the surface) or potlid fractures (small round spalls that have popped off the surface).
The results of a raw material analysis are usually expressed in some quantitative way, commonly as a series of counts and weights for each raw material (and subcategory) identified. For example, supposes that when you are done sorting you have a set of heat treated Swan River Chert flakes that lack cortex. This would be recorded as: 12 SRC flake, heat treated, no cortex, 24.40 grams. In contrast, you might also have: 1 SRC flake, not heat treated, cortex present, 18.20 grams. Count and weight provide complementary means of evaluating the amount of a raw material present. There is a rough correspondence between the number of individual pieces (especially flakes) and the amount of reduction undertaken (i.e., each flake represents an individual blows delivered to a core). Weight, in contrast, gives an approximation of the total amount of a raw material present: a single decortication flake may outweigh a dozen biface trimming flakes. Correlation of count and weight can provide useful information on reduction stage, strategy, etc., especially when further correlated with characteristics of heat treatment and presence or absence of cortex.
Establishing Raw Material Context. With this information obtained by using the procedures described above, you can evaluate the complement of raw materials present in the assemblage and draw conclusions about how raw materials and raw material sources were being utilized. First consider the geographic location of the site. Does it fall within the Western, Eastern or Southern Resource regions? Is it in one region but close to another? Remember that there are no clearly defined boundaries between regions, especially when glacial drift serves as a raw material source. If a site is, for example, within the Western Resource Region but close to the Eastern Resource Region, raw materials characteristic of both regions can be considered locally available.
When you have placed your site in relation to the raw material resource regions, you should be able to evaluate whether individual raw materials are local, nonlocal or exotic. Remember that exotic refers only to materials which were widely circulated and probably deliberately traded. For example, within the Western Resource Region, KRF qualifies as an exotic raw material; Cedar Valley Chert or Gunflint Silica would be considered exotic.
At many sites there will be raw materials which cannot be characterized as local, nonlocal or exotic. These may include generic categories of raw materials which have not been specifically identified (e.g., "chert," "burned chert," "silicate," etc.) or distinctive but unidentified raw materials. An example of the latter occurs the Gulseth site in Rock County, southwestern Minnesota (Skaar et al. 1994). At this site, the predominant raw material is a distinctive, translucent grey silicate. Although it can be distinguished from other raw materials at the site, its source and distribution are not known. Thus it is not possible to characterize this raw materials as local, nonlocal or exotic; instead, it is characterized as "indeterminate."
Next note the directions that nonlocal and exotic raw materials are originating from, and the relative distance for each source area. For example, at some sites you may discover that nonlocal and exotic raw materials indicate fairly strong ties in a particular direction or with a certain region. At another site the raw materials may indicate connections in several directions but at a much more limited distances. Other sites evidence almost total reliance on immediately available local raw materials.
Site function may also come into consideration. For example, long term or repeated habitations tend to have a greater mix of raw materials and present a broader general picture or raw material procurement, utilization and circulation patterns. In contract, a quarry or procurement site will probably contain one raw material almost exclusively, and offer little information on general patterns of raw material circulation, etc. A single episode, brief period habitation may also offer limited information on general trends, although it may provide other types of valuable information.
In evaluating the complement of raw materials present in an assemblage, also consider the location of your site within the region. For the Western Resource Region, an important distinction is whether the site is located within or outside of the Agassiz lakebottom. Natural supplies of raw material are essentially absent at most locations within the lakebottom; all lithics present were carried in, and exotic raw materials (especially KRF) may be more abundant. For the Eastern Resource Region, sites in the northern part of the region are more likely to be in proximity to bedrock raw material sources. For the Southern Resource Region, relative east and west position is important. Raw materials at sites in the western to central part of this region are more likely to be derived from glacial drift. Sites in the eastern end of the region, however, are more likely to lie in proximity to focal, primary-context raw material sources. Such sites may contain a preponderance of a single local material such as Galena Chert of Cedar Valley Chert.
Examining Archaeological Context. Once these characteristics have been examined, your assemblage may be compared with other assemblages, whether from nearby or distant sites. This is often most interesting part of the analysis. This step is not always possible or practical, because it requires collecting a body of comparative data either by literature research or undertaking a descriptive raw material analysis of other assemblages. Both of these options can be time consuming. It can also be difficult to find useful comparative information in published reports because identifications and descriptions have not been based on any consistent set of raw material categories. As better information on Minnesota raw materials has become available in the last few years, however, consistency and comparability have improved. The more similar the analyses and raw material characterizations are between two assemblages, the more informative the comparison should be.
On a basic level, such a comparison involves the presence or absence and relative amounts of various raw materials found in the assemblages being compared. Even this simple step is a valuable exercise, because it contributes to building up a large scale picture of overall utilization and circulation patterns of various raw materials.
Rather than discussing at length the rationale and methods of this type of comparison, two examples will be presented in summary. Both are taken from previous analyses which I have undertaken. These methods and results come from an ongoing research program, one which will be refined based on initial results. The final results will hopefully include an expanded body of comparative information, an elaborated explanation of methods, and discussion of the history of lithic raw material procurement, circulation, modification and utilization patterns in Minnesota prehistory.
The first example comes from a survey conducted by the Archaeology Department, Minnesota Historical Society in Rock and Pipestone counties of southwestern Minnesota (Skaar et al. 1994; the following discussion draws heavily on this source). Twelve sites were identified during the course of the survey. Due in part to poor surface visibility, the lithic sample from most of the sites was fewer than 10 lithic artifacts; at many sites, only one or two lithic artifacts were recovered. The largest assemblage, totalling 71 pieces, came from the Gulseth Site. Because of the small size of the assemblages, and because the sites were contained in a limited area of relatively uniform environment, the assemblages from the various sites were grouped for the purpose of analysis. Lithic assemblages from sixteen other sites in the region were also examined in order to provide a broader context for evaluating and understanding the limited sample from the survey sites.
Certain interesting characteristics were noted during the analysis. The assemblage from the 12 sites identified by the survey contained an unusual complement of lithic raw materials, including a large sample of a previously undescribed raw material (at the Gulseth site), fusulinid cherts, Prairie du Chien Chert, Galena Chert and Knife River Flint. These materials would not be expected to occur locally.
Whether this represented a significant trend or a sampling problem could not be determined from the small survey sample. When information from other sites in the region was included in the analysis, however, it seemed clear that this variety was not coincidental. The larger sample contained additional examples of Prairie du Chien Chert, Knife River Flint, Galena Chert and fusulinid cherts. It was especially significant that at least eight of the other sites also contained pieces of the same unidentified silicate found at the Gulseth site. In addition, the larger sample contained examples of Grand Meadow Chert, Moline Chert, Burlington Chert, Bijou Hills Silicified Sediment, Hudson Bay Lowland Chert, jaspers which are probably of Black Hills origin and possibly even Arkansas Novaculite. This included raw materials originating from nearly every direction and from distances of up to several hundred miles. The sample contained several distinctive but unidentifiable raw materials. It is fairly clear that they are not of local origin, but specific identifications were not possible. It is probable that these raw materials also represent exotic materials; that they are unfamiliar suggests that they may originate from relatively distant sources.
The diversity of this sample is unusual and requires some explanation. Two possible factors were considered. First, this small region is located between and in reasonable proximity to three riverine drainage systems. Each of these could represent important transit routes. The region is south of the Minnesota River, near tributaries of the Missouri, and north of major streams leading up from the Middle Mississippi through Iowa. Any of these could have provided gateways to the region -- but not reasons for entering the region. Such motivation could have been provided by a second factor.
This part of southwestern Minnesota and adjacent parts of southeast South Dakota contain a unique and valued resource. Pipestone, also known as Catlinite, was and continues to be a material of considerable ritual significance to many Native American peoples. Archaeological evidence for the importance of this material comes from a number of sources, one of the most concise and persuasive of which is Sigstad's (1972) sourcing study involving 193 pipestone artifacts from across the continent. He determined that 60 percent of the artifacts examined were traceable to pipestone sources in southwestern Minnesota or adjacent parts of South Dakota; about 37 percent were made from pipestone mined at the quarries in the Pipestone National Monument. He was able to establish associations between the pipestone quarries of southwest Minnesota and pipestone artifacts from areas as distant as Anasazi and Hohokam sites in the desert southwest; Adena and Hopewell sites in the Ohio area; Fort Ancient sites in Kentucky; Great Bend sites in Kansas; Post-Contact sites in Michigan, Idaho and Quebec; and sites of uncertain chronology as far away as Kentucky and Pennsylvania (Sigstad 1972:42-47). This is in addition to a large number of sites within the Upper Midwest, with various cultural and chronological affiliations. In addition, by carefully consideration of cultural and chronological associations, he showed that the quarries had served as important sources of pipestone as long ago as 3000 BP (Sigstad 1972:51).
The distribution of Catlinite provides evidence that the southwest Minnesota region was at the focus of a unique, transcontinental resource distribution system. That the archaeological sites of this small region should contain a diversity of widely derived lithic raw materials is perhaps not so surprising. The importance of such exotic lithic raw materials should also not be underestimated in this relatively stone-poor region. Approximately two-thirds of the artifacts examined for this region are locally available raw materials. Of these material, Swan River Chert is by far the most common. Other local materials represented include Tongue River Silica, Red River Chert, rhyolite, quartz, quartzite derived from glacial drift, and possibly chalcedony derived from glacial drift. These materials are generally of marginal quality. The one-third of the artifacts consisting of nonlocal and exotic raw materials constituted most of the high quality stone.
This interpretation is based on a limited study and must be considered tentative. The possible associations are intriguing, however, and provide research questions for examining other sites in this part of the state.
A second example comes from excavation of the Roosevelt Lake Narrows site conducted by Woodward-Clyde Consultants in Cass County, north central Minnesota (Bakken 1995a). The lithic assemblage from the site (including only chipped stone artifacts) totalled over 4300 pieces, including tools and debitage. This sample was large enough to support an extensive intrasite analysis. The assemblage was also compared with assemblages from 14 other sites, both within and beyond the region, in order to place it in a broader context. Comparative information was obtained both by first hand examination of collections and from published site reports. The basic comparative analysis was limited to four raw materials: quartz, Knife River Flint, Tongue River Silica, and obsidian. Quartz, KRF and obsidian were included because they are distinctive, widely recognized, and should have been accurately identified in most cases. Tongue River Silica (TRS) is included because it is also fairly distinctive, and because the substantial variation in percentage may be significant. However, TRS is not often specifically identified in reports which are more than a few years old, so information on this material was lacking for some sites.
The analysis showed that the percentages of quartz and KRF at Roosevelt Lake Narrows are comparable to other sites in the region. At most of the sites, quartz is the single most abundant material, often constituting over half of the total lithic assemblage. Most of the sites within the study region also contain a minimal amount of KRF. The comparative information further shows that small amounts of obsidian are not unusual at Woodland sites in the study area. These characteristics seem to be part of a regional raw material utilization pattern.
Assemblages from sites outside the region provide an instructive contrast. (These sites were located to the west; several were in the Red River Valley.) The range of percentages for KRF barely overlaps between regional and extraregional sites. The western (extraregional sites), which are closer to the KRF primary source area, contain significantly higher amounts of KRF In this respect, it is also interesting that very small amounts of Hixton Quartzite are found at sites in the region. Hixton is much less common than KRF, even though the Hixton source area is closer. In fact, in Woodland sites even obsidian is more abundant than Hixton, in spite of the fact that the source area is over 800 miles away. The scarcity of Hixton requires some explanation. One possible reason is that some social barrier to trade existed between this part of north central Minnesota and the region where Silver Mound is located. Another potential explanation is that Hixton Quartzite, for whatever reason, was not as desirable as KRF (or obsidian).
The analysis also shows that the percentage of Tongue River Silica varies widely, especially within the region, with Roosevelt Lake having the highest percentage of any of the sites examined for this study. In contrast, the sample of sites outside the area of study shows much greater consistency in the amounts of TRS present. At this point, the reasons for this variation are not clear. It might be due to the abundance of TRS in the immediate vicinity of the site or to a paucity of other, better quality raw materials.
The amount of quartz in the sites outside the region is less by an order of magnitude. The abundance of quartz in study area sites requires some explanation. Part of such an explanation may be the flaking characteristics of quartz. Although it could be classified as a marginal quality raw material, that would ignore some significant characteristics. Quartz does not fracture evenly in any direction; even when it is "homogenous," it is not isotropic. Fracture is harder to control, which means that quartz is less suitable for making patterned tools. This characteristic is reflected in the relative scarcity of patterned tools made of quartz. It might, however, be well suited to making expedient flake tools -- flakes which were used without further modification and discarded when the edge was dulled (cf. Sather 1989).
Hopefully these examples serve to illustrate some of the potential of this kind of comparative analysis, based in large part on the methods of raw material analysis described above. As additional sites are examined and the body of comparative information continues to grow, interpretations should become more secure and elaboration of a contextual overview should become possible.
(4) Another approach to raw material identification which I have recently seen applied with promising results is macroscopic examination of samples under ultraviolet light. This technique seems to have especially good potential for distinguishing materials which look similar under visible light, although a systematic evaluation is needed.
| Continue to Next Section | Return to Index |