Abstract. This paper provides a synthesis and overview of current knowledge concerning lithic raw material resources in Minnesota. A large part of the state is covered with glacial drift which served as a diffuse source of raw materials. Primary geological sources such as outcrops and lag deposits also served as raw material sources in some parts of the state. Three raw material resource regions are defined and described, each containing a different set of raw materials and types of raw material sources. Strategies for conducting a raw material analysis within this context are discussed, including both identifying individual materials and interpreting the significance of a lithic assemblage. A cross referenced list of raw materials found at regional sites is included.
INTRODUCTION
Stone artifacts are probably the most abundant and certainly the most enduring traces of the prehistoric human presence in Minnesota. In order to glean as much information as we can from these traces, we analyze any number of subtle variations in artifact form and composition. Of the many approaches to analyzing lithic artifacts, one of the most common is the identification of raw material. In Minnesota, however, this approach has been hampered by a scarcity of published information on lithic raw materials and by an advanced state of confusion over nomenclature and terminology. In the following pages, an attempt is made to assemble what is known about lithic raw materials in the state, to organize this information and place it in a broader context, and to provide a common vocabulary for discussing lithic raw materials. Hopefully this will contribute to more productive archaeological analyses.
In many areas, describing the local lithic raw material base consists of identifying geological formations which contain knappable silicates, describing the silicates, describing the extent of the formations, and so forth. In these cases raw material sources tend to be restricted and relatively well defined. Such situations lend themselves reasonably well to sourcing studies, such as those which compare trace element signatures of archaeological specimens with lithic samples from known sources in order to identify probable place of origin. In Minnesota, however, the situation is qualitatively different.
Most of the state is covered by glacial drift which served as a lithic raw material source. Silicate rocks from various sources are mixed in the drift, and this relatively homogenous mixture is spread over vast areas -- sometimes from one end of the state to the other. Traditional sourcing studies are not generally useful in this situation. Although such methods might allow archaeological specimens to be associated with a geological context, there is no assurance that the raw material was quarried from the original context rather than from a drift related, redeposited context. A different sort of analytical approach is therefore outlined in this paper. It is based on the gross distribution of various kinds of drift in the state, on the regional distribution of outcropping bedrock, and on the kinds of knappable lithic raw materials contained in both these types of sources. The types and amounts of raw materials represented at any given site may then be studied in comparison to regional patterns of raw material availability.
The information presented comes from a variety of sources. In part it is gleaned from archaeological and geological literature. Some information has been gleaned from conversations with archaeologists (both professional and avocational) working in the region. A substantial part comes from provenienced comparative samples collected from across the state; I am especially grateful to colleagues at the Minnesota Historical Society for their help in collecting these samples in the course of conducting archaeological surveys. The information on prehistoric use of raw materials comes from examining collections at the Minnesota Historical Society, the University of Minnesota, Moorhead State University, other institutions and in private collections.
Terminology. Discussion of a few important concepts and terms will facilitate this discussion. First is the notion of "focal" versus "diffuse" raw material sources. A focal raw material source is one which has clearly defined limits, and is usually relatively restricted in extent. An example would be the Nora Member of the Shell Rock Formation. This member occurs in discontinuous locations in north central Iowa and possibly in south central Minnesota; surface or near surface exposures are even more restricted. It contains a fossiliferous chert which, within Minnesota, is thought to occur only in parts of two south central Minnesota counties. In contrast, a diffuse raw material source is one which has vaguely defined limits, and which is relatively extensive. An example would be the Des Moines drift, which cover large areas from Manitoba to the Iowa border, and from the borders of North and South Dakota into east central Minnesota and even western Wisconsin. This drift contains a number of knappable raw materials, distributed more or less at random throughout this large region.
Second is the idea of "primary" versus "secondary" geological context. A material in primary geological context is in or in close association with to the geological context where it was formed. Examples could include lenses of chert in a limestone bed, lag deposits of chert which have been left behind as their limestone matrix disintegrated but which have not been otherwise dislocated or transported, or chert in talus deposits at the base of a limestone bluff. In the latter case, although the raw materials have begun to move away from their primary context, the association is still clear and the distance is negligible. A material in secondary geological context is no longer in or in proximity to the geological context where it was formed. In some cases, the primary geological context may no longer even exist. Raw materials in a secondary geological context include, for example, chert in glacial till, chert in reworked glacial sediments such as fossil beaches and outwash fans, or chert in fluvial gravels (whether ancient or modern).
This discussion is also based to some degree on considerations of human scale. Phrases such as "extensive distribution" and "restricted distribution" are meaningful only in reference to human experience, including such factors as how far a person can travel in a day or how widely a person may wander is a season and lifetime. This also relates to factors like the lifespan of a stone tool: will it be used for a day, a season, or longer? On one scale, an outcrop of chert bearing limestone a mile long might seem enormous. In reference to the distance covered during a cycle of seasonal rounds, however, the mile is small. Chert harvested from that outcrop might be abandoned at a another location a hundred miles away. This definition of scale is still unfortunately imprecise. Until some meaningful, more objective scale can be articulated, however, the notion of "human scale" allows some calibration of the following discussion.
Regarding the proximity of site to lithic raw material source, three categories are used in this report -- local, nonlocal, and exotic. The distinction is based on relative distance from source area and on trade patterns. Local materials are those which are available in the immediate vicinity of the site or in the surrounding area. For the purposes of this discussion, "surrounding area" includes territory that might be covered during normal seasonal rounds. Although this definition is hardly satisfactory, it must suffice because it is not presently feasible to provide a more specific definition. In contrast, nonlocal and exotic material are not found in the surrounding area. They may come from neighboring regions or distant sources. The distinction between nonlocal and exotic depends on circulation patterns. Nonlocal materials were traded incidentally and occur only sporadically outside of the source area. In contrast, exotic raw materials were deliberately quarried for trade or export, and were widely circulated. They are also of high quality. The most common examples at sites in Minnesota are Knife River Flint, Hixton Quartzite and Burlington Chert.
In the following discussion, some raw material names are abbreviated. This is done not to save space but to acquaint the reader with the abbreviations. These abbreviations are in fairly common use because they have proven convenient in cataloging and discussion. Many of the raw materials discussed are known by more than one name or have a name with multiple spellings. In such cases I have used the name or spelling which seems preferable, with some explanation of my choice. Alternate names and spellings are cross referenced in the "Raw Material Reference List" section.
A few technical terms also require some discussion. Most of the terminology archaeologists use to discuss lithic raw materials is borrowed from the large and finely articulated vocabulary of geology. Generally this works well. But the two disciplines have different goals, and occasionally geological vocabulary does not meet the needs of the archaeologists. In such cases, what was originally a geological term may acquire a specifically archaeological usage.
This is reminiscent of the way archaeologists have borrowed the concepts and vocabulary of stratigraphy from the field of geology. These terms and ideas have been adapted and have acquired specific archaeological connotations. Perhaps this can guide our use of geological terminology in lithic studies also. While it is desirable to maintain a degree of agreement between the two fields, it is not always necessary to strictly follow geological conventions. Rather, we should feel free to make intelligent and well considered departures from such conventions when this allows us to better meet the needs of the discipline of archaeology. Such adaptation may, in fact, be necessary. Geologists use multiple systems for describing and classifying rocks and minerals, including the kinds of materials discussed in this paper. I have tried to select a coherent, applicable set of terms to use here. The following discussion will adhere, when practical, to standard geological definitions. Exceptions ar noted.
Types of Silicates. Most rocks used for flaking consist principally of a material called silica, SiO2. It is silica which gives these silicate (1) rocks their desirable characteristics. They are hard. The break in a predictable fashion and produce a sharp edge. They are brittle enough to flake, yet strong enough to have a durable edge.
The differences between various silicates result principally from differences in how the silica is structured, and partly from the amounts and kinds of other constituents. Both depend largely on how the rock was formed. On the basis of these considerations, silicates may be divided into a few basic groups.
The first and most important group is the cryptocrystalline silicates, which consist principally of microscopic silica crystals and only minor traces of other elements or minerals. These materials usually occur in sedimentary formations, and are thought to result from the chemical precipitation of silica from seawater, the gradual replacement of other minerals by silica, or other processes (see Calvert 1983; Hesse 1989). If the crystals are arranged in a fibrous structure, the material is called chalcedony. If the crystals have a random orientations, somewhat like a felt of silica crystals, the material is called chert. These two materials are easily distinguished by petrographic analysis, but not by macroscopic visual inspection. Agate is a variety of chalcedony usually having concentric bands of different colors, although other patterns (such as moss agate) do occur. Jasper is a variety of chert which is relatively opaque and usually yellow, olive, brown or red in color as a result of the presence of iron. Flint may be used as a synonym for chert, or may refer to varieties of chert which are opaque to translucent and dark grey, dark brown or black in color. The dark color may come from the inclusion of organic matter.
In practice, archaeologists often use flint, chert, jasper and chalcedony as mutually exclusive, descriptive terms rather than according to their more technical definitions. In this type of usage, chalcedony is a high quality, translucent to transparent silicate, often with a waxy texture. Flint is high quality, translucent to opaque rock which is dark in color. Jasper is good quality, usually opaque and sometimes waxy textured rock which is yellow, brown, red or olive in color. And chert is more opaque, often lighter colored or somewhat coarser cryptocrystalline material which does not look like chalcedony, flint or jasper. This descriptive usage is acceptable, especially when a petrographic identification is not available.
A second group of silicates consists of indurated, or solidified, sediments. The particles may be fused by temperature and pressure or may be cemented by silica. In some cases silica may have replaced other, nonsilicate minerals. These materials may be classified on the basis of the size of the particles which they contain. When the sediment consisted of sand sized quartz grains, the resulting material is called quartzite. When the grains have been fused by metamorphosis, the material is known as metaquartzite. When the grains have been cemented by silica, the material is known as orthoquartzite. The suitability of the material for flaking depends on how strongly the grains are joined. If they are so strongly bonded that the fracture passes through rather than around individual grains, flaking quality is better.
Siltstone is a sedimentary rock containing much finer particles. The terms argillite and siltstone are both established in archaeological usage, and are generally used synonymously. Geologists use these terms differently, distinguishing a number of fine grained sedimentary rocks based on subtle distinctions of particle size, trace mineral content or other characteristics. Such specific distinctions and terminology are not generally useful to archaeologists.
The third group of silicates is igneous extrusive rocks, which occur when molten rock rich in silica has cooled at or near the surface. Because the rock was able to solidify quickly, there was little time for crystals to form. When cooling was very rapid, with no crystal formation taking place, the result is obsidian. This material is comparable in structure to glass (which also consists of noncrystalline silica). When the cooling takes place more slowly, resulting in the formation of very small crystals, rhyolite is produced. If the same molten rock had solidified farther below the surface, slowly enough to produce visible crystals, the result would be granite (which is not flakable). Each of these rocks contains minerals besides silica, although the silica dominates in the case of flakable rocks. (Molten rock which is low in silica produces other materials, such as diorite and basalt.) Geologists distinguish many varieties of high silica, igneous extrusive rock on the basis of which minerals are present and in what proportions. Such subtle distinctions are generally not useful to archaeologists.
In contrast to these other materials, quartz has a large scale crystal structure. In some cases an individual crystal will be large enough to cut and polish into a sphere, the infamous crystal ball. In most cases quartz consists of a mass of conjoined crystals with different orientations. Quartz is usually less suitable for flaking than other silicates because it tends to breaks along crystal plains. The fracture is harder to control, and it is more difficult to produce a desired shape. Quartz still provides a useful sharp edge when broken, however, and was widely utilized.
Descriptive Terms. A number of descriptive petrographic terms are encountered in discussions of lithic raw materials. A vug is a small cavity in a rock. Vugs are often lined with druse, a crust of small crystals. Druse may also occur on exterior surfaces. Ooliths are small, generally round accretionary bodies which are cemented together in a rock. They are sometimes described as resembling fish eggs. The term is also used in a loose sense for any small, generally round bodies which occur in a contrasting matrix in a rock, although this usage is not recommended. An oolite is a rock containing ooliths. Chert and other rocks which contain ooliths are also called oolitic. Phenocrysts are small, often glassy crystals which are scattered in a finer rock matrix. Cortex is the outer, mechanically and chemically weathered surface of a rock. Other kinds of rock may adhere to this surface. Chert cobbles, for example, often retain a crust of limestone on their cortical surface. Patination refers to a distinct kind of surface alteration, usually the chemical weathering of a fracture surface; mechanical weathering is less important. Patination is typically very thin, often different in color than the interior of the rock, and sometimes marked by increased porosity.
Raw Material Names. A specific raw material is often named for the geological stratum in which it originates. For example, that chert which comes from the Galena Formation is called Galena Chert. In some cases the geological associations are not known, and an arbitrary name is given. One material which has never been found in a bedrock association is especially abundant around the Knife River of western North Dakota, and has come to be known as Knife River Flint.
In a few cases, a nonstratigraphic name has become well established in archaeological usage even though the geological context of the material is known. For example, a distinctive chert which was quarried near the town of Grand Meadow in south central Minnesota has become known as Grand Meadow Chert (GMC). GMC provides a good example of why a nonstratigraphic name may sometimes be preferable. Trow (1981:102) associated this raw material with the Rapid Member of the Cedar Valley Formation, which suggests that either Cedar Valley Chert or Rapid Chert might be a suitable name for this material. The first name, however, is used to designate a different and quite distinct chert found in the Cedar Valley Formation. Problems also exist with the second name.
Witzke et al. (1988) proposed an extensive revision of Devonian stratigraphy for Iowa and southeastern Minnesota. The Rapid Member was restricted to southeastern Iowa and adjacent areas, the Cedar Valley Formation became the Cedar Valley Group, and several new formations were named (with associated new members). Presently it is not even clear whether Grand Meadow Chert comes from the Cedar Valley Group or Wapsipinicon Group, much less which formation or member it is associated with. If this material had been called Rapid Chert, it could be argued that the name should be changed to reflect the stratigraphic revisions. While I believe it is generally a good practice to provide stratigraphic names to newly described materials (e.g., "Shell Rock Chert" in Olmanson et al. 1994), I emphatically do not believe it is a good idea to abandon an established name in order to keep current with changing interpretations of geological stratigraphy. Such periodic changes can only serve to introduce more confusion to an already confused situation; they also make older archaeological reports difficult to read and understand. Changes in nomenclature should be made only for very good reason and as infrequently as possible. Established, widespread usage should be given priority over stratigraphic association as a matter of convention.
There has not been a standard regarding capitalization of lithic raw material names. An often used form is to capitalize the unique part of the name, but leave the generic element uncapitalized (e.g., Swan River chert, Knife River flint). In this paper I recommend and use another form. The terms for specifically named and described raw materials are proper names, and each element should be capitalized: Knife River Flint, Swan River Chert, Knife Lake Siltstone. This is consistent with some other capitalization conventions in English, for example "a river" but "the Mississippi River." The names of categories of raw materials, however, are not proper names and do not need to be capitalized: flint, chert, silicified wood, siltstone.
(1) In archaeological discussions, the word "silicate" is commonly used to signify what geologists might call a "siliceous rock." In geological usage, the term "silicate" actually covers a broader category of compounds based on the silicon-oxygen tetrahedron.
| Continue to Next Section | Return to Index |