Duck Lake Site and Implications for Late Archaic Copper Procurement and Production in the Southern Lake Superior Basin, The

Duck Lake Site and Implications for Late Archaic Copper Procurement and Production in the Southern Lake Superior Basin, The

Hill, Mark A

ABSTRACT

Few sites relating to Archaic procurement of copper have been archaeologically investigated in the southern part of the Lake Superior basin. The Duck Lake site, located in the Ontonagon River watershed of Michigan’s Upper Peninsula, is within the source area for native copper. Investigations at this site reveal a field camp consisting of one or more occupations by Late Archaic groups. Lithic analysis demonstrates a strong reliance on non-local raw materials, and suggests a degree of residential mobility within a region known to the occupants. Lithic material was obtained through embedded procurement and exchange with more distant groups. Copper likely entered the early Late Archaic exchange system in the Midwest through frequent interactions and trade between groups, and appears consistent with a down-the-line model of exchange.

While copper has long been of interest in the archaeology of the upper Midwest and Great Lakes, an understanding of the role of copper procurement and exchange within the Late Archaic societies of the region has been slow to develop. Recent research in the southern Lake Superior basin has begun to open approaches to studying Late Archaic societies and their mobility, resource acquisition, and exchange, and to examine copper procurement and production within these larger issues. In this paper, one such site is examined in depth, and lithic analysis is used to address issues of resource procurement and group mobility.

Previous research concerning copper focused on broad morphological studies (Hedican and McGlade 1993; Wittry 1951), mortuary occurrences (Hruska 1967; Pleger 1998), sourcing and trace element analysis (Pletka 1991; Rapp et al. 1980, 2000), and detailed metallurgical studies (Vernon 1990). Exchange systems involving copper have been used to address the emergence of social complexity and the formation of social networks (Bender 1985; Brown 1985). Copper studies have also played a role in the development of systems and processual theory in archaeology (Binford 1962). However, an understanding of the procurement, transport, and processing of copper is, as yet, poorly developed. How was copper obtained and by whom? How did copper procurement fit within the mobility strategies of Archaic societies? How did copper enter the exchange system in the upper Midwest during this period?

These questions have a long history in the archaeological literature of the Midwest. One of the most explicit early studies of copper acquisition and exchange was conducted by Fogel (1963), who examined the distribution of copper artifacts in the upper Midwest and found it consistent with a directional trading model (Renfrew 1972) in which people at a distance from the source area preferentially obtained copper while those closer to the source had lesser quantities. Two obvious omissions leave this result in question. First, archaeological research has preferentially focused on areas farther from the source while those areas closer to the source have (until recent years) received less archaeological attention (Martin 1999:198-199). Second, Fogel’s study ignored any potential differences in population density that may have existed from north to south in the Archaic Great Lakes, and the consequences of such variability in the deposition of copper in the archaeological record. More recently, Clark (1995) demonstrated that copper acquisition on Isle Royale was embedded in a broader set of subsistence related activities. The direct procurement model endorsed by Fogel thus begins to give way to perspectives that favor embedded procurement and exchange systems emphasizing down-the-line models.

A clearer view of Archaic copper acquisition and exchange may be obtained through a closer examination of sites in and around the native copper source area. Within the midcontinent, that source area is largely limited to the Lake Superior region where copper is concentrated along the Keweenaw Highlands in the southern basin, and Isle Royale, the Minnesota and Ontario north shores, and Michipicoten Island in the northern basin (Martin 1999:25-30; Rapp et al. 2000).

Archaeological research in this area has been uneven. While survey and testing of Isle Royale have revealed several Woodland and some Archaic period copper procurement locations, the Archaic components align more closely with north shore cultural developments than those of Michigan and Wisconsin to the south (Clark 1996:132). In the southern basin, historic mining and development along the Keweenaw Highlands damaged or destroyed many sites in the nineteenth century thus limiting the number of intact sites available for research. Furthermore, the western Upper Peninsula of Michigan received little archaeological attention prior to the inception of cultural resource management activities during the past few decades. The limited number of excavated components that relate to Archaic period procurement and processing of copper in the southern basin, and further south in Wisconsin, Michigan and surrounding states, is partially a result of these historical factors.

However, recent research at several sites in the western Upper Peninsula and adjacent northern Wisconsin put researchers on the threshold of these issues. Several sites in the southern basin have now revealed Late Archaic components in which copper was either obtained or processed (Hill 2005; Martin 1993; Moffat and Speth 1999). One of these sites, Duck Lake, contains a discrete Late Archaic component directly related to copper acquisition and processing.

In this paper, the Duck Lake site will be described and lithic analysis will be used to explore procurement and mobility strategies of the Archaic period residents. Materials such as lithics and copper may be obtained and moved throughout the region in three primary, yet non-exclusive, means. First, materials may be obtained through direct procurement by the deliberate movement of people to the source area. Subsequent population movements then distribute materials throughout the region as use and discard lead to the deposition of these materials in the archaeological record. Second, acquisition of resources such as lithics and copper may be embedded within other subsistence and social activities of these populations such as hunting and seasonal aggregation and dynamics of micro and macroband formation. Finally, materials may be obtained through trade, gift exchange, or related social interactions with other populations who have direct access to those resources.

By examining the Duck Lake lithics, a better understanding of the types of mobility and resource procurement activities can be developed. Locating and exploiting copper resources by the Archaic occupants is not expected to vary from those methods used for lithics. Copper is found in the same types of geological and environmental settings as lithics as both primary deposits in bedrock outcrops and secondary sources in glacial features, riverbeds, and lakeshores. Activities directed toward finding and exploiting lithic resources will also effectively function to locate copper resources. Therefore, it is proposed that acquisition, use, and exchange of lithics may serve as a model for those same activities directed toward copper.

The Duck Lake Site

The Duck Lake site is located in Ontonagon County, Michigan along the East Branch of the Ontonagon River (Figure 1). The site lies approximately 22 km south of the Keweenaw Highlands, or Trap Range, and approximately 50 km upstream from Lake Superior.

The area is one of generally low relief and sandy soils. The East Branch of the Ontonagon River flows through a deeply entrenched valley to the north and east of the site, while Duck Lake and a series of small marshy ponds are located immediately south of the site. On-site vegetation is largely pine, with lesser quantities of maple, birch, and aspen and an understory dominated by bracken fern, wintergreen, and blueberry.

Late Pleistocene and early to mid-Holocene glacial and postglacial events are responsible for much of the modern landscape. Glacial advances deposited a terminal moraine immediately north and northeast of the site, while remnants of other terminal moraines are located south and east of the site. This glacial advance blocked drainage in the area, forming proglacial Lake Ontonagon between 10,500 and 11,000 B.P. (Farrand and Drexler 1985). As the glacial lobe subsequently receded, Lake Ontonagon drained northward and melt water from the receding ice-sheet carried fine sandy sediments that were deposited over the clay lakebed soils around the eastern edge of the Lake Ontonagon plain and the margins of the rockier moraines. An early stage of the East Branch of the Ontonagon River flowed over one of these outwash plains as it drained northward to Lake Superior.

As the glaciers receded, lake levels of the newly formed Lake Superior dropped to levels up to 180 feet below modern by about 9,500 radiocarbon years before present (RCYBP) (Farrand and Drexler 1985; Fitting 1975:11-18). This significantly lowered the base level for the rivers of the region, and the East Branch of the Ontonagon River eroded through the fine sandy sediments of the outwash plains and the denser day lakebed soils to form an entrenched meandering river (Haywood 1998:8). Lake levels subsequently rose to approximately 20 feet above modern levels during the Nipissing transgression around 4500 B. P., thus reducing the rate of downcutting and resulting in the deep steep-sided gorge through which the river currently flows.

The Duck Lake site is located on one of the sandy outwash plains, and the entrenched East Branch of the Ontonagon forms a deep gorge along the eastern and northern margins of the site. A large terminal moraine, with its rockier soils and higher relief, is located on the opposite side of the river from the site (Figure 2).

The site was originally found by collectors who noted the presence of cultural material on the surface of two dirt roads that cross the cultural deposit’s northwestern and eastern portions. Mapping of the archaeological deposit later revealed three areas, localities A, B, and C, where cultural material appeared to be concentrated (Figure 1).

Locality A is found on a small southward projecting peninsula on the shore of one of the small ponds. Locality B occupies a low ridge overlooking the river to the northeast, while locality C is located between A and B at the intersection of two of the dirt roads and along the north and west shore of another of the small ponds on the southwestern part of the site.

Formal investigations began in 1994, when archaeologists with the Ottawa National Forest conducted a systematic shovel-testing project to document site boundaries and determine the nature of the cultural deposit. In 1996 and 1997, Ottawa National Forest archaeologists and volunteers with the Forest Service’s Passport in Time program returned to the site to conduct limited excavations at locality B (Figure 3). Twenty-eight m^sup 2^ were opened, mostly located in a five m^sup 2^ block excavation centered on an area of lithic and copper concentration that had been identified through shovel testing. All excavation was conducted by hand troweling in three-cm levels, fill was screened through one-quarter-inch mesh, and all artifacts found during excavation were piece plotted before removal.

Controlled metal detector surveys were conducted to identify areas of copper concentration. An attempt was made to distinguish between recent metal artifacts, such as bullets and beverage can tabs, and the prehistoric copper artifacts. Copper artifacts were flagged and mapped as diey were encountered and most were not excavated. However, copper artifacts along the dirt road on the south side of locality B were collected after mapping as collectors and metal detector enthusiasts have traditionally exploited this area.

Excavation revealed a relatively straightforward alfisol, based upon a parent material of fine, well-sorted sand. A thin organic horizon, consisting of a mat of decaying leaf litter and plant material, extended from the surface to a depth of between one and two cm. A four-to-five cm thick dark gray sandy loam A horizon was located beneath the organic O horizon. A thicker eluvial strata of light grayish-brown to pinkish-gray sandy loam, varying from 8 to 18 cm in thickness, was found beneath the A horizon. Finally, the transition from the light grayish-brown E horizon to the underlying strong brown finely sorted sands of the B horizon was pronounced and readily observed.

The cultural deposit began at the base of the A horizon and was most commonly associated with the light colored sandy loam soils of the E horizon. Cultural material was found in a single cultural stratum that began around 12 cm below surface and extended to a maximum depth of 32 cm. Cultural material frequency rapidly diminished at the bottom of the E horizon and materials were largely absent from the fine sands of the underlying B horizon.

Three features were identified, induding two unprepared surface hearths and a concentration of copper artifacts and scrap (Table 1). All three were found near the center of the block excavations in locality B.

Feature A was a poorly defined area of charcoal staining and reddened soil found at a depth of 18 cm below surface. No pit was identified, and the feature is interpreted as a surface hearth. Materials associated with this hearth include worked copper nuggets, copper preforms, a comer-notched projectile point fragment, and Prairie du Chien and Galena chert debitage. One wood charcoal sample was recovered from this feature for radiocarbon dating.

A 3.5-liter soil sample was collected from Feature A and submitted to the Museum Archaeology Program, State Historical Society of Wisconsin, for flotation and analysis. Wood, bark, charcoal, and other floral remains were recovered from the sample after sorting under 10x binocular magnification. Identified taxa include wood charcoal from maple, spruce, red pine, white pine, and red oak (Table 2). One seed of bittersweet and one unidentifiable seed were also found.

A second hearth, Feature B, was found approximately 140 cm south of Feature A starting at a depth of 18 cm below surface and extending to 24 cm. This was a well defined oval surface hearth with charcoal, charcoal staining, reddened soils and fire-cracked rock. Approximately half of the feature was excavated; the other half extended into the west walls of the excavation unit. A worked copper nugget rested on the pitted upper surface of an anvil stone three cm east of the feature. Other materials in association included lithic debitage of Prairie du Chien and Galena cherts and worked copper scrap.

A seven-liter soil sample was taken from the hearth and submitted for flotation and analysis. This produced 778 pieces of botanical material (Table 2). Most consisted of wood charcoal, induding spedes such as maple spruce and red oak.

A small concentration of worked copper, fragmentary bone, and lithic material was found between the two hearths at a depth of 20 cm below surface and labeled Feature C. The purpose and formation process of this feature are not dear; it may represent a small activity area, or it could be the result of the cleaning of one of the nearby hearths.

A soil sample was recovered from the area around this feature and again submitted for flotation and analysis. Botanical materials found in association with this feature include wood charcoal from white pine, red oak and unidentified coniferous species, along with one maple seed and an unidentified seed (Table 2).

Excavation also produced 155 small highly fragmentary, mostly burned, pieces of bone (Table 3). These all appear to represent mammal species, but most are unidentifiable. One fragment represents a long bone of a raccoon-to-deer-size mammal, while a second is a right phalanx of a fisher- or marten-sized mammal. Identification of genus and species is not possible for any of this material.

Two charcoal samples were collected for radiocarbon dating. The first was wood charcoal collected from a depth of 24 cm below surface near the base of Feature A. Accelerator mass spectrometry (AMS) analysis of this sample produced a conventional C14 age of 3420 +/- 50 RCYBP (Beta 099777). Calibration of this date using the Pretoria Calibration Procedure (Vogel et al. 1993) yielded an intercept date of 1705 B.C. and a 1σ intercept range of 1755 to 1660 B.C.

The second sample was taken from a depth of 17 cm below surface immediately north of Feature B. It was found in association with an area thought to represent the production of copper artifacts. Materials associated with the charcoal sample include copper scrap, an anvil stone, lithic debitage, and fire-cracked rock. This wood charcoal sample was submitted for extended counting which produced a conventional age of 3400 +/- 110 RCYBP (Beta 124454). Calibration yielded an intercept date of 1685 B.C., and two ranges within the 1σ probability; 1870 to 1830 B.C., and 1780 to 1530 B.C.

Excavation within locality B suggests that this part of the extensive site consists of overlapping small activity areas or components. The compressed stratigraphy of the site did not allow for separation of components. Nonetheless there is an impression of spatially distinct dusters of material that may represent separate contemporary groups or different occupations of the site. One or perhaps two of these activity areas are represented in the five m^sup 2^ block excavation, where the two hearths served as the focus of activities, and lithic tools, debitage, and copper artifacts were densely concentrated within two m of these features (Figure 4).

Lithic Analysis

Stone tool and debitage assemblages are sensitive to group mobility and the availability of raw material (Andrefsky 1994; Bamforth 1986; Binford 1979). Binford (1979) proposed that high mobility correlates with a “curated” lithic assemblage featuring bifaces and other formal tools, while higher sedentism correlates with assemblages featuring expedient tools such as utilized and retouched flakes. The correlation between bifaces and mobility is related to the ability of bifaces to function both as tools and as a source of material for the production of other tools (Kelly 1988), while expedient tools are seen as more efficient in sedentary contexts since less energy is expended in their production.

Since its introduction, the concept of a “curated” lithic assemblage has undergone considerable discussion and debate (e.g. Andrefsky 1994; Bamforth 1986; Short 1996). The availability of raw material has been determined to affect the efficiency of expedient tools (Bamforth 1986), while Short (1996) has proposed that “curation” is not a nominal value, but rather a continuous one best applied to individual artifacts in terms of the degree of utility extracted from the tool.

Several aspects of the Duck Lake lithic assemblage are thus examined to explore issues of mobility and exchange. These include a) identification of lithic raw material types to examine source areas and potential differences in assemblages composed of different materials; b) frequencies of expedient versus formal tools; c) a measure of “degree of utility extracted” for formal tools; and d) examination of debitage attributes such as flake type, debitage size, frequencies of biface thinning flakes, and platform types to understand reduction strategies and their variance based on raw material types.

Raw materials

The vast majority of Duck Lake’s lithics were of non-local materials such as Prairie du Chien chert, Galena chert, Knife River Flint and Onondaga chert. This assemblage stands in marked contrast to other Late Archaic sites in the Lake Superior and northwestern Lake Michigan drainages, where locally available quartz typically dominates assemblages (Hill 1994; Salzer 1974).

High quality lithic materials are limited in the southwestern Lake Superior basin. In the western Upper Peninsula and adjacent northern Wisconsin, quartz is the most commonly encountered lithic material in archaeological contexts dating from the Archaic until the advent of European trade materials in the historic period. Quartz occurs as primary deposits in veins within the mafic rocks of the Keweenaw Highlands, Gogebic Range, and other bedrock exposures of the region. Glacial activity and subsequent erosion and stream flow have deposited quartz as rounded cobbles frequently found in streambeds, lakeshores, and glacial till.

Hudson Bay Lowland chert is also available along streambeds, lakeshores and in glacial till. This material is of highly variable quality, and was deposited through glacial action from a source north of Lake Superior. Hudson Bay Lowland chert typically occurs as very small cobbles and pebbles, and is of limited utility for tool production.

Basalt was only rarely used for tool production, but is available locally. This mafic rock is found in bedrock deposits along the Keweenaw Highlands, Gogebic Range, Porcupine Mountains, and other upland landforms in the region. Basalt also commonly occurs as cobbles along streambeds, lakeshores, and in glacial till.

Most of the tools and debitage at Duck Lake were produced from raw materials found outside of the Lake Superior basin. Several different lithic materials occur in nearby northeastern Minnesota, including Knife Lake siltstone, Gunflint silica, and jasper taconite, yet none of these north shore materials occur in the Duck Lake lithic assemblage. Instead, the assemblage seems to have its primary sources to the south as represented by significant quantities of Prairie du Chien and Galena cherts, along with lesser quantities of Hixton Silicified Sandstone. Prairie du Chien and Galena cherts primarily outcrop in bedded deposits and stream valleys approximately 200 miles southwest of the site in the Mississippi River valley of southwestern Wisconsin and adjacent regions in Minnesota, Iowa and Illinois (Klawiter 2001). Hixton Silicified Sandstone is an orthoquartzite found primarily in west central Wisconsin, also approximately 200 miles to the southwest. Other lithic materials at the site originate well outside the Lake Superior Basin. These include Knife River Flint from 700 miles to the west in western North Dakota (clayton et al. 1970) and Onondaga chert from 500 miles to the east in the Lake Ontario region of New York and eastern Ontario (Hammer 1976; Wray 1948).

To identify these materials, lithics were first sorted into raw material classes by visual characteristics. Comparative collections from southeast Wisconsin were used to identify Prairie du Chien and Galena cherts, while collections from near Silver Mound in Wisconsin were used to visually identify Hixton Silicified Sandstone. Likewise a comparative collection from Dunn County, North Dakota, was used to identify Knife River Flint, while collections of raw material from the western Lake Ontario basin were used to tentatively identify Onondaga chert. Quartz was readily identifiable, while Hudson Bay Lowland samples from stream and beach gravels of the western Lake Superior region were used to aid in the identification of this locally available chert. Several quartzites, such as Penokee and Chequamegon, are found in the western Upper Peninsula and northern Wisconsin, but these are currently poorly defined and sourced, so non-Hixton quartzites were grouped into a single category, which may actually represent more than a single source.

Additional analysis was performed to definitively identify Knife River Flint and Onondaga chert. Instrumental neutron activation analysis (INAA) was performed at the University of Toronto’s SLOWPOKE Reactor Facility on 28 debitage samples selected from the Duck Lake assemblage to represent the variety of materials found at the site (Bury 1997). The only reference material used in the INAA was for Knife River Flint, Hudson Bay Lowland chert, and Onondaga chert. Chemical composition verified the initial visual identification of Onondaga and Knife River Flint. The remaining samples visually identified as other materials for INAA, were not available, including Prairie du Chien and Galena cherts.

Overall, most tools were manufactured from non-local materials, principally Prairie du Chien chert, Galena chert, and Knife River Flint (Table 4). Local lithics, including quartz and Hudson Bay Lowland chert, are represented but are in the minority. Local materials were largely used in the production of cores and expedient tools, although two cores are of Prairie du Chien chert.

Non-local materials also dominate the debitage assemblage (Table 5), most of which consists of Prairie du Chien and Galena cherts. Knife River Flint, Hixton Silicified Sandstone, and Onondaga chert are also represented. Local materials make up less than 12 percent of the debitage assemblage, and are represented by quartz, Hudson Bay Lowland chert, and basalt. The remainder consists of unidentified cherts and quartzites that may represent both locally available lithics and non-local materials.

Tools

Twenty-one formal tools (Table 6), including projectile points, drills, scrapers and bifaces, along with 76 expedient tools such as utilized and retouched flakes were recovered during formal investigations at Duck Lake. Two projectile point base fragments were recovered (Table 6), representing one small corner-notched point (Figure 5a) and one stemmed point. Both were manufactured from Prairie du Chien chert.

Scrapers are most commonly small thumbnail shaped endscrapers made on flakes (Figure 5g-l). Five are manufactured from Prairie du Chien chert, while the sixth is made from an unidentified gray chert.

Bifaces tend to be fragmented and small. When identifiable, bifaces are either oval or irregular in shape. A few complete and several fragmentary specimens are present in the collections (Figure 5e-f). One is manufactured from locally available quartz, while the rest are of identified non-local materials or unidentified chert.

Two fragmented Prairie du Chien chert drills from the Duck Lake site are thin, narrow bifaces and exhibit both lenticular and diamond-shaped cross sections (Figure 5b-c). A third drill was manufactured on a flake of Knife River Flint and exhibits side notches that appear to have facilitated hafting (Figure 5d).

Most formal tools were heavily used, broken, and worn, as expected from an assemblage of curated tools at the end of their use life. The degree of “utility extracted” (Shott 1996) was estimated using the “scraper index” developed by Kuhn (1991:82). This index measures the maximum centerline thickness of a scraper and compares it to the maximum vertical thickness of the working edge, and thus provides a relative measure of the remaining utility of the tool in which a value of zero represents unmodified flakes and one indicates tools that were extensively reduced and exhausted. The scrapers from Duck Lake were all manufactured of non-local material, and their scraper index averages .96 (range = 0.8 to 1.0, s.d. = .07, n = 6) thereby indicating that these tools were almost at the end of their utility. In other words, the scrapers, and by extension and visual confirmation many of the non-local tools, were highly worn and depleted.

Informal, or expedient, tools, consisting of retouched flakes and utilized flakes, comprise the majority of tools recovered from the site. Retouched flakes were defined as non-patterned expedient tools generally manufactured on a flake. One or more edges have been deliberately retouched to form a working edge. Flake scars from this retouch are small and do not exceed one-quarter of the width of the tool. Analysis of edge angles for working edges shows a median value of 67° with a central range of 55° to 75° (Ferone 1998:11). This was interpreted as indicating heavier woodworking activities associated with these artifacts.

Utilized flakes differ morphologically from retouched flakes in that they exhibit use wear rather than deliberate retouch along one or more edges. No deliberate shaping, flaking, or retouch is evident, but use wear is indicated by dulled, shattered, or micro-flaked edges. Ferone (1998) also demonstrated a difference in edge angle and possibly function between these tools and retouched flakes. Edge angles for utilized flakes are shallower than for retouched flakes, with a median value of 46.5° and a central range of 30° to 65°. The shallow median angle and wide central range is interpreted as indicative of the multi-function nature of utilized flakes, with light tasks such as skinning, hide scraping, and light woodworking being likely activities associated with these tools. A few utilized flakes exhibit steeper edge angles (72° to 90°, n=10), and these were likely used for heavier processing tasks.

Lithic debitage

Debitage was divided into two broad categories of 1) cores, and 2) flakes and shatter. Cores and core fragments are represented by seven pieces. At least three of these artifacts represent bipolar reduction based upon their polyhedral shape, small size, and crushing evident on both the upper and lower surfaces. While bipolar reduction is a common element of quartz technology in the region, all three bipolar cores from Duck Lake are of chert.

Analysis of flakes and shatter followed a modified “triple cortex” typology (Andrefsky 1998:112-114; Morrow 1984). In this approach, materials were sorted into categories of primary, secondary, and tertiary flakes based upon the presence and extent of cortex on the dorsal flake surface. Additional categories included flake shatter and angular shatter, while the technologically defined category of “bifacial reduction” flake was also used.

Primary flakes were defined by the presence of cortex or patina covering the entire dorsal surface of the flake, and are often thought to represent initial reduction stages of tool manufacture. secondary flakes represent an early, but intermediate, reduction stage. These flakes exhibit weathering or cortex over a portion of the dorsal flake surface. However, other flake scars representing earlier stages offtake removal are also present on the dorsal surface. Tertiary flakes do not exhibit cortex or weathering on any surface. These may be the result of later reduction stages and tool maintenance.

Flake shatter is defined as fragments of obvious flakes based upon presence of a single ventral surface, thinness, ripple marks, and terminations that are either feathered, stepped or hinged. However, they are incomplete in that they lack bulbs of percussion, striking platforms, or other surfaces of applied load. Controlled experiments have demonstrated that broken and incomplete flakes more often result from core reduction than tool production (Amick and Mauldin 1997:20; Andrefsky 1998:123-124).

Angular shatter consists of blocky fragments of chert or quartz that lack striking platforms, bulbs of percussion, or other definitive flake elements. This category is a common byproduct of the early stages of reduction, especially with regard to quartz materials.

Biface reduction flakes represent thinning, shaping, or maintenance activities associated with bifaces. They are typically a specialized type of tertiary flake, with a characteristic small, sharp lip that occurs at the junction of the striking platform and the bulb of percussion. This lip and the striking platform represent the remnant edge of a biface. Recent research into lithic fracture mechanics demonstrates that these flakes result from a non-hertzian fracture, and are now referred to as “bending flakes” (Andrefsky 1998:23-29). They often result from application of load on the edge of an objective piece such as a biface. Loads that produce bending flakes are most typically consistent with soft hammer and pressure techniques. Flakes categorized as biface reduction flakes are bending flakes that exhibit a remnant of the original biface edge.

Finally, all striking platforms were coded as cortical, flat, complex, abraded, crushed, or unidentifiable. These platform types have been correlated with type of reduction, with flat and cortical representing core reduction, and complex and abraded more commonly associated with reduction of bifaces (Andrefsky 1998:88-96).

Each of these categories represents a different aspect of tool production and use. To examine the potential differential use of local versus non-local lithics, debitage was examined by material type, typological category, and platform type. Significant differences were noted (Tables 7 and 8).

Materials from the early stages of lithic reduction, especially angular and flake shatter, dominate the quartz debitage assemblage (Table 8). Quartz assemblages are often skewed towards early reduction stage materials due to the poor flaking qualities of the material (Meinholz and Kuehn 1996). Nonetheless, given the nature of the quartz debitage and the local availability of this lithic material, it is likely that the early stages of quartz reduction occurred on site. Striking platforms found on quartz flakes are most frequently flat, further indicating reduction of cores as the primary activity associated with this material.

In addition to flake and angular shatter, the Hudson Bay Lowland chert debitage consists of cores, secondary flakes and tertiary flakes, with primary flakes relatively prominent. As with the quartz assemblage, this also appears to represent early stage reduction activities. Platform analysis also supports this conclusion with 40 percent of the flakes exhibiting either flat or cortical platforms. However, 60 percent of the Hudson Bay Lowland chert flakes exhibit complex platforms suggesting that later stages of reduction and perhaps biface production were taking place with this material.

Regional non-local materials, including Prairie du Chien chert, Galena chert, and Hixton silicified sandstone, show a composition of debitage attributes that is significantly different from that of the local materials, particularly in the higher frequencies of late stage reduction debitage (χ^sup 2^=84.46, df=4, p

Knife River Flint and Onondaga chert debitage exclusively represent the later stages of artifact use life. This small assemblage is fairly divided between secondary, tertiary and biface reduction flakes with a complete lack of shatter and primary flakes. With the exception of one flat platform, all other striking platforms are complex. These materials were most likely brought to the site in the form of bifaces or other tools.

Debitage size also reveals differences between non-local and local materials. For local materials (quartz, quartzite, basalt, Hudson Bay Lowland chert) the mean weight of non-shatter debitage is .9 grams, while the same non-shatter debitage of exotic materials (Prairie du Chien, Galena, Hixton, Knife River Flint, Onondaga) is nearly half the size at .5 grams. This difference lends further support to the conclusion that non-local materials are more often subjected to the latter stages of tool production and maintenance, whereas the local materials are relatively more often used in early stages of lithic reduction.

In sum, this lithic assemblage is consistent with a moderate amount of residential mobility within a defined territory. Materials such as Knife River Flint and Onondaga chert were likely obtained through exchange as tools, and their limited frequency and late reduction stage at Duck Lake is expected within a down-the-line model of exchange. Regional materials such as Prairie du Chien chert, Galena chert, and Hixton silicified sandstone were obtained either through trade or embedded procurement, and carried as both formal tools and raw material. The presence of cores and early stage reduction debitage shows that while formal tools were produced in anticipation of group mobility, moves were likely conducted over short distances and within a known region, thus facilitating the limited amount of logistical planning and preparation observed at this site. Also consistent with this view is the use of locally available materials such as quartz and Hudson Bay Lowland chert for the production of expedient tools. The regional and long distance lithic materials comprise a formal toolkit of “curated” tools and small cores, as would be expected with a degree of mobility. However, the utilization of non-local raw materials for cores and of local materials as expedient tools suggests a degree of familiarity with the locally available resources. Taken together, the picture that emerges from the lithic assemblage is one of residential mobility within a known region-perhaps associated with seasonal aggregation and disaggregation of macro and microbands. Lithic procurement is characterized by two processes: embedded procurement of locally available materials and trade for exotic materials.

Copper Analysis

Eighty-seven copper items were recovered during investigations at the Duck Lake site (Table 9). Approximately half of these were recovered through formal excavation, the remainder found through controlled metal detector survey. Most of the artifacts consist of worked, flattened, or otherwise modified but unshaped nuggets. Another 22.2 percent consist of unmodified nuggets of float copper imported to the site. The remaining 22.2 percent are tools, ornaments, preforms, or incompletely shaped artifacts.

The copper assemblage represents a wide range of copper artifact production. Many intermediate forms are present which represent preforms at various stages of completion. In addition, while the site is located only about 22 km southeast of a primary source of copper in the Keweenaw Uplands, copper does not naturally occur within the well sorted sediments of the sandy outwash plain on which the site is located. Therefore copper found on the site was likely imported. Thus, the copper assemblage provides insights into the procurement and production strategies employed in Late Archaic copper technologies.

Artifacts were divided into seven classes based on the presence and amount of modification of the native copper, as well as on morphology of finished or nearly finished tools. These categories include unworked copper nuggets, worked copper nuggets and waste, worked and shaped copper pieces, projectile point preforms, edged tools, bipointed objects, and beads.

Twenty one unworked copper nuggets were recovered during excavation and controlled metal detector surveys. Unworked copper nuggets tend to be small, averaging only 2.2 grams, and appear to represent glacially redeposited copper.

Worked copper artifacts vary from hammered copper nuggets to small waste pieces. This is the largest category of copper artifacts, and the pieces tend to be small, ranging in weight from less than .1 grams to 35.7 grams. A few different classes of materials may be represented, including hammered nuggets and waste pieces, or scrap. Hammered nuggets represent worked pieces of float copper, while small waste pieces originate from hammering and removal of small projections from other worked pieces. Waste pieces are quite small, with a mean weight of .3 grams. Hammered nuggets are larger, with a mean weight of 5.1 grams, and may have been intended for later production of finished materials.

Copper pieces that have undergone greater modification and shaping yet do not dearly represent an identifiable tool type are designated as “worked and shaped.” Two of these were found (Figure 7f, o). One is a sub-rectangular to lunate shaped item, with a rectangular cross section. This is a relatively large artifact, with a weight of 52.4 grams. A second worked and shaped piece is the largest copper artifact recovered from the site with a weight of 185.4 grams. It is approximately rectangular with a rectangular cross-section.

Five artifacts are tentatively identified as representing intermediate stages in the production of projectile points. Two to three morphological classes are represented, including stemmed artifacts (Figure 7j, k), small triangular preforms (Figure 7c, d), and a quadrilateral or diamond-shaped artifact (Figure 7b). These artifacts are often rectangular to lenticular in cross section, with blunt lateral margins. The stemmed form resembles the “ace of spades” points found at the Riverside site in Menominee County, Michigan (Hruska 1967), yet have not been fully drawn out to their final width. The triangular preforms resemble specimens recovered on the Keweenaw Peninsula where they were referred to as “wedges” (Martin 1993). At the Riverside site as well as at 20KE20, triangular points with the beginnings of sockets were found (Hruska 1967; Martin 1993) that might be intermediate between these triangular preforms and conical points.

Two edged tools, both apparently complete, were found during investigations at Duck Lake. The first of these is a tanged “butter knife” form with a thin, slightly tapering rectangular shape and rounded tip (Figure 7e). The tang is approximately 18.6 mm in length and tapers to a blunt point. The second tool is a leaf-shaped blade with a lenticular cross section (Figure 7a). One edge is sharpened, while the opposite edge appears to be blunted.

Five small bipointed objects or fragments were recovered. These artifacts are small, round to rectangular in cross-section, and terminate on each end in rounded to pointed tips (Figure 71-n). The function of these artifacts is not readily apparent although they tend to co-occur with small beads and possible bead manufacturing debris. They appear too slender and gracile to have functioned as pressure flaking tools, but their use as awls cannot be ruled out. These artifacts are similar to those described as “pins” from 20KE20 (Martin 1993:148).

Two small round beads and one flattened tubular bead were recovered. The two small beads were found in the concentration of copper artifacts in Feature C (Figure 7g, h). These two beads are very similar, and both seem to have been manufactured by bending a small rectangular band of copper around a mandrel or possibly cordage, and then butting the ends together. The tubular bead has been flattened, and was found approximately 1.5 m north of Feature A.

One unusual artifact consists of a long, narrow and very thin band of copper. When first found it was in one piece but was broken into five pieces during excavation due to its highly fragile condition. Combined, the five pieces measure a total length of 41.2 mm, and it has a rectangular cross-section that is 2.4 mm wide and 1 mm thick. This artifact may represent a blank for the manufacture of small round beads.

Copper was found in several concentrations on the site. Controlled metal detector surveys were used to map the locations of individual artifacts by flagging all locations that exhibited readings consistent with copper. Metal detector surveys covered all of locality B and portions of locality C near the intersection of the two dirt roads. A metal detector survey was also conducted along the dirt road below the ridgetop at locality B, where individual artifacts were excavated by trowel and mapped in place to prevent loss due to unauthorized collecting.

Copper was found to be concentrated in a roughly 10m wide by 12Om long area along the low ridgetop in locality B (Figure 3). Copper materials were also found at the base of the ridge along the road and on the relatively level area between the ridge and the river valley to the northeast. Materials were not found on the moderate slopes of the ridge. Metal detector surveys were conducted to the south and west of the road with no items identified, and around the road intersection in locality C where few items were noted.

Within the main distribution on the ridgetop, there were three smaller locations where copper materials appeared to be more densely clustered. No excavation was conducted around the first concentration, while a single unit in the second produced one piece of worked copper and one bipointed object. The block excavation exposed much of the third concentration.

Within the main block excavation, copper was concentrated around the three features as were most other cultural materials (Figure 4). Most of the copper was found within approximately 60 cm of Feature A. This area produced several pieces of worked and unworked copper, one triangular preform, two bipointed objects, and the possible bead blank.

Fewer pieces were found around Feature B. However, the association between copper, the hearth, and an anvil stone was very clearly represented here. Worked and unworked copper pieces were present, and one piece of hammered copper was found directly over an anvil stone (Figures 4 and 8) next to the hearth.

A concentration was dearly observed during excavations and was designated Feature C. This feature is located almost halfway between the two hearths, and produced the two stemmed preforms, one bipointed object, and the two small beads.

The area around features A, B and C appears to be a small copper-working activity area where initial working of nuggets was done near the hearths as evidenced by the anvil stone and worked and unworked nuggets. This component is quite similar to that created by LaRonge (2001:380) during experimental recreations of copper production methods. A corner-notched projectile point and two radiocarbon dates of 3420 +/- 50 BP (Beta 099777) and 3400 +/- 110 BP (Beta 124454) are associated with this well dated Late Archaic activity area.

Copper Production

At this point, a tentative copper production sequence can be proposed for the Duck Lake materials. While lithic reduction strategies have a longstanding analytical history, few have been developed for copper production technologies. The presence of an entire copper production sequence at the Duck Lake site makes it possible to propose an initial impression of copper tool and ornament production methods.

Distinct stages in copper tool production appear in the Duck Lake collection. Categories such as worked copper, preforms, and finished tools appear to represent a continuum of production with significant and notable objectives for each stage of production. Copper artifacts were produced from small nuggets of glacially transported and deposited copper typical of till plains and glacial moraines. While no geochemical analysis has yet been conducted on the copper, the source for Duck Lake copper is thought to be a ground moraine located across the East Branch of the Ontonagon River immediately northeast of the site. Based on the discarded unworked nuggets, these pieces of float copper tended to be relatively small. At Duck Lake, the average mass of unworked copper nuggets was found to be 1.7 +/- 1.7 grams. However, selection factors may have led to discard of nuggets that were either too small or large, so this figure may be a biased representation of the nugget size selected for production of tools and ornaments.

Copper nuggets were initially hammered and annealed to remove projections and to begin rough shaping and flattening. Methods used to shape these pieces include hammering and annealing, grinding, cutting and twisting. The result of this step is a Stage I blank and small copper waste pieces or scrap. The average mass of Stage I blanks is 10.2 +/- 28.7 grams, and given the wide variance it is likely that these represent several different types of potential finished tools and ornaments.

The second stage of production is rough shaping and forming, leading to the production of preforms and blanks. At this point, Stage II artifacts are morphologically distinct, and dearly represent various tool/ornament types. Production of Stage II artifacts is done using hammerstones, anvil stones, and hammering and annealing methods. It is unlikely that much scrap is produced between Stage I and Stage II, although evidence for this conclusion is not readily apparent in the data.

Stage III is final shaping and sharpening of blade margins. The product is either a finished tool as seen in the edged tools represented in Figure 7 (a, e), or the preforms for socketed tools and beads. No significant waste material is expected from the production of Stage III tools, except perhaps copper dust from grinding and sharpening.

A fourth manufacture stage may be present with regard to socketed tools where the socket may be formed in Stage IV. Also, bead production may take place at this point after forming a long thin blank. In Stage IV, bead production would consist of folding the blank around a mandrel or piece of cordage, then hammering and grinding the overlapping edges. This stage is only tentatively identified based upon apparent trajectories of manufacture for conical points, and is not dearly represented in the Duck Lake data.

While artifacts deposited in the Duck Lake site appear to represent somewhat distinct stages of production, it is more likely that they represent significant points on a continuum. Some sharpening or blade forming may take place in Stage II, some final removal of waste pieces may occur in Stage III. Yet overall, there seems to be a process involved with relatively sequential stages of production.

Discussion and Summary

Duck Lake provides an opportunity to examine Late Archaic mobility, resource procurement, and copper production in the upper Great Lakes. Excavation revealed a small activity area around two hearths with domestic debris consisting of lithic tools, lithic debitage, and faunal remains. Copper is dosely associated with the hearths, and a wide variety of copper forms are present. Unworked copper nuggets were transported to the site by human action, and the nearest source for copper is the glacial deposits assodated with a ground moraine located immediately north of the site. Worked copper indudes initially hammered forms, preforms for triangular and stemmed points, beads, and bipointed objects.

The chronology of the Late Archaic component is well established. Uncalibrated dates place the radiocarbon age of this component at approximately 3400 B.P. Calibration of the two dates places the occupation at around 1685 to 1705 B.C. Many chronologies of the upper Great Lakes 5region would place that in the later portion of the Middle Archaic period, which is often dated up to around 3000 B.P. (e.g., Fitting 1975:68; Martin 1999:162; Stoltman 1997:134). More recent examination of Middle Archaic dates in the upper Midwest suggests the termination of that period by 2000 B.C. (Kuehn 2002). Further, small comer-notched points are often a hallmark of the early Late Archaic in the upper Midwest (Salzer 1974; Stoltman 1997:134) and the presence of a small corner-notched projectile point similar to Monona Stemmed and Preston Notched in direct association with the dating samples, copper production area and features strongly suggests that this Duck Lake component is associated with the early Late Archaic period.

Seasonality of site occupation is tentatively suggested by seeds of bittersweet (Celastrus scandens) and maple (Acer sp.) recovered from feature fill. Bittersweet flowers in late May and June, and sets seeds in July to August. Sugar maple, the most likely maple species represented by the Acer sp. seed from Feature C, has seeds that mature in autumn. While the numbers of these seeds are small, their presence in Late Archaic cultural features suggests occupation during the late summer or autumn.

The Archaic component at Duck Lake is largely consistent with the Burnt Rollways Phase first defined by Salzer (1974) in Oneida and Vilas counties of northern Wisconsin. Burnt Rollways was defined in the Highland Lakes region of northern Wisconsin, and additional investigations have extended it into the same physiographic region of adjacent Michigan (Hill 1994). Lithic tools and debitage from Burnt Rollways Phase sites indicate a significant dependence on locally available quartz, yet tools are often made of chert and projectile points are extensively manufactured from chert (Salzer 1974:46-47). Hixton and Knife River Flint are present in small quantities, and suggest trade or other connections outside the local area. Burnt Rollways had an established copper industry, and Salzer tentatively assigned this phase to the period from 2000 to 1000 B.C. Additional Burnt Rollways components have been identified at the Rainbow Dam East and Rainbow Dam West sites in northern Wisconsin, where locally available quartz was found to dominate the lithic assemblages along with lesser quantities of non-local materials such as Prairie du Chien, Galena, and Silurian cherts and Knife River Flint (Moffat and Speth 1999). Recent research (Bruhy et al. 1999; Hill 1994; Moffat and Speth 1999) has provided several radiocarbon dates for the Burnt Rollways phase, securely placing it from 1700 B.C. to perhaps as recent as 600 B.C.

The physiographic region in which the Burnt Rollways phase was defined did not extend northward to Duck Lake, and the nearest identified Burnt Rollways phase component is approximately 30 miles to the south of Duck Lake. No Late Archaic phases have yet been defined in the western Upper Peninsula, and only a few Late Archaic sites are known despite extensive surveys. One of these, the Ottawa North site at the headwaters of the Menominee River, is a short-term field camp with an uncorrected date of 1370 +/- 220 B.C. (Hill 1994:27-28). No diagnostic artifacts or copper materials were recovered from this site, and the lithic assemblage is dominated by locally available quartz. A second western Upper Peninsula site, Alligator Eye, is a quartz quarry location with two uncorrected radiocarbon dates of 1690+/- 150 B.C. and 1540+/- UO B.C. (Hill 1994:39). At the tip of the Keweenaw Peninsula, 20KE20 is a multicomponent site associated with copper procurement and production. Two of the dates from the site (1350 +/- 60 B.C. and 1310 +/- 70 B.C.) suggest that at least one of the components is roughly contemporary with Duck Lake (Martin 1993:175). As at Alligator Eye and Ottawa North, the lithic assemblage at this site is dominated by quartz, with the majority of the chert artifacts apparently derived from locally available Hudson Bay Lowland chert (Pauketat 1993:176-180).

Occupants of the Duck Lake site appear to have participated in an extensive regional exchange system involving lithic materials. Non-local raw materials, including Prairie du Chien chert, Galena chert, Knife River Flint, Onondaga chert, and Hixton Silicified Sandstone, dominate both the tool and debitage assemblages. Local lithics, including quartz and Hudson Bay Lowland chert, are most common among the expedient tools and retouched flakes, yet even here they are a minority.

Lithic analysis indicates that these non-local lithics were brought to the site as both formal tools and cores as part of a “curated” lithic toolkit. What is less dear is how those materials were acquired. Prairie du Chien and Galena cherts comprise most of the assemblage, and these materials have their nearest source approximately 200 miles to the southwest along the Mississippi Valley of west central Wisconsin and adjacent Minnesota, Iowa, and Illinois (Bury 1997; Klawiter 2001; Morrow 1994). Prairie du Chien and Galena cherts are often found as minor components of other sites in the region during the Late Archaic (Moffat and Speth 2001; Pleger 2000), suggesting that their acquisition is somewhat commonplace at this time. Given the distances involved it is unlikely that direct acquisition of Prairie du Chien and Galena cherts was common. Instead, interactions between the northern highlands and the Mississippi Valley may have taken place in the form of regular contact and exchange between local populations.

It is also unlikely that direct acquisition or embedded procurement accounts for the presence of materials such as Knife River Flint and Onondaga chert. Onondaga chert has its source in the Lake Ontario region of New York state and Ontario, Canada, approximately 500 miles from Duck Lake, while Knife River Flint has its source in the Golden Valley Formation of western North Dakota some 700 miles to the west. Acquisition of these materials is best explained through reliance on exchange networks.

Knowledge of lithic exchange systems provides important clues to the nature of Late Archaic copper exchange in the region. Copper acquisition at Duck Lake initially appears to fit either a model of direct or of embedded procurement. However, an examination of the Duck Lake lithics suggests regular interaction between populations in the north and those of the Mississippi Valley to the soudiwest. While this may have involved actual movement of populations, exchange is deemed more likely. Exchange of lithics appears to have included at least some unworked raw material, again suggesting some frequency of interaction and relatively small social distance between these regions. The same situation may hold true for the movement of copper from its source region in the Lake Superior basin to the soutii. More distant interactions may have been less frequent and may have featured more complete goods such as tools of Knife River Flint, Onondaga chert, and, reciprocally, copper. If this is the case, copper exchange does indeed fit a down-the-line model. If so, copper moved out of the Lake Superior basin through the regular interaction between local and regional populations.

Acknowledgments. Many individuals have contributed substantially to this study. Personnel of the Ottawa National Forest graciously provided much of the time and funding needed to conduct this project. Forest Service archaeologists Troy Ferone and Jill Ferone assisted with project planning, excavation, and analysis and their efforts are much appreciated. Matt Thomas, Veronica Bury and Pamela Haywood also assisted with excavation and specialized analysis of lithics and geomorphic setting. Susan R. Martin provided valuable insights into the archaeology of copper procurement and exchange. Kelly Davidson assisted with the laborious task of debitage analysis. Elizabeth Horton graciously lent her skills to the faunal analysis. William Andrefsky’s comments on an earlier draft of this paper led to important new perspectives. Finally, the project could not have been conducted without the hours of time, effort and interest dedicated to it by volunteers under the Forest Service’s Passport in Time program. While many have contributed to this work, any errors of fact or interpretation belong solely to the author.

References Cited

Amick, D. S. and R. P. Mauldin

1997 Effects of Raw Material on Flake Breakage Patterns. LiMc Technology 22(l):18-32.

Andrefsky, W. Jr.

1994 Raw Material Availability and the Organization of Technology. American Antiquity 59(l):21-34.

1998 Lithics: Macroscopic Approaches to Analysis. Cambridge Manuals in Archaeology, Cambridge University Press.

Bamforth, D. B.

1986 Technological Efficiency and Tool Curation. American Antiquity 51(1):38-50.

Bender, B.

1985 Emergent Tribal Formations in the American Midcontinent. American Antiquity 50(l):52-62.

Binford, L R.

1962 Archaeology as Anthropology. American Antiquity 28(2):217-225.

1979 Organization and Formation Processes: Looking at Curated Technologies. Journal of Anthropological Research 35(3):255-273.

Brown, J. A.

1985 Long-term trends to Sedentism and the Emergence of Complexity in the American Midwest. In Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity edited by T. D. Price and J. A. Brown. Academic Press, New York.

Bruhy, M. E., K. Egan-Bruhy, and K. L. Potaracke

1999 Summary Report of Excavations at the Butternut Lake Inlet Site (47-FR137): Exploring Seasonality and Diet at a Multicomponent Site in Forest County, Wisconsin. USDA Forest Service, Chequamegon-Nicolet National Forests, Northern State Regional Archaeology Center, Report of Investigations No. 19. Rhinelander.

Bury, V.

1997 Chemical Characterization and Identification of Michigan Upper Peninsula Lithics by INAA. Manuscript on file and copies available, Ottawa National Forest. Ironwood, Michigan.

Clark, C. P.

1995 Archaeological Survey and Testing, Isle Royale National Park, 1987-1990 Seasons. Occasional Studies in Anthropology 32. Midwest Archaeological Center, National Park Service, Lincoln, Nebraska.

1996 Old Fort and Old Copper: The Search for the Archaic Stage. In Investigating the Archaeological Record of the Great Lakes State: Essays in Honor of Elizabeth Baldwin Garland, edited by M. B. Holman, I. G. Brashler, and K. E. Parker, pp. 101-138. New Issues Press, Kalamazoo, Michigan.

Clayton, L, W.B. Bickley Jr., and W. J. Stone

1970 Knife River Flint. Plains Anthropologist 15:282-290.

Farrand, W. R. and C. W. Drexler

1985 Late Wisconsinan and Holocene History of the Lake Superior Basin In Quaternary Evolution of the Great Lakes, edited by P. F. Karrow and P. E. Calkin, pp. 17-31. Geological Association of Canada, Special Paper 30. St. John’s, Newfoundland, Canada.

Ferone, T. J.

1998 Quantitative Analysis of an Informal Tool Assemblage: The Duck Lake Site (20ON21). Manuscript on file and copies available, Ottawa National Forest. Ironwood, Michigan.

Fitting, I. E.

1975 The Archaeology of Michigan: A Guide to the Prehistory of the Great Lakes Region. Cranbrook Institute of Science, Bloomfield Hills, Michigan.

Fogel, I. L.

1963 The Dispersal of Copper Artifacts in the Late Archaic Period of Prehistoric North America. The Wisconsin Archeologist 44(3):129-180.

Hammer, J.

1976 Identification and Distribution of Some Lithic Raw Materials from New York State. Man in the Northeast 11:39-62.

Haywood, P. A.

1998 The Archaeological Geology of the Duck Lake Site, Ottawa National Forest. Unpublished Honors Thesis. Department of Geology, Albion College.

Hedican, E. J. and J. McGlade

1993 A Taxonomic Analysis of Old Copper Projectile Points. Man in the Northeast 45:21-38

Hill, M. A

1994 Ottawa North and Alligator Eye: Two Late Archaic Sites on the Ottawa National Forest. USDA Forest Service, Ottawa National Forest Cultural Resources Management Series 6. Ironwood, Michigan.

2005 Late Archaic Copper Procurement and Production in the Southern Lake Superior Basin: An Example from the Duck Lake Site. Paper presented at the 70th Annual Meeting of the Society for American Archaeology, Salt Lake City.

Hruska, R.

1967 The Riverside Site: A Late Archaic Manifestation in Michigan. The Wisconsin Archeologist 48(3): 145-260.

Kelly, R. L.

1988 The Three Sides of a Biface. American Antiquity 53:717-734.

Klawiter, B. N.

2001 Geochemical-Geoarchaeological Provenance of Prairie du Chien Chert. The Minnesota Archaeologist 60:119-130.

Kuehn, S. R.

2002 Denning the Temporal Boundaries of the Middle Archaic: Old and New Evidence from Southern Wisconsin. The Wisconsin Archeologist 83(1):19-44.

Kuhn, S. L.

1991 “Unpacking” Reduction: Lithic Raw Material Economy in the Mousterian of West-Central Italy. Journal of Anthropological Archaeology 10:76-106.

LaRonge, M.

2001 An Experimental Analysis of Great Lakes Archaic Copper Smithing. North American Archaeologist 22(4):371-385.

Martin, S. R. (editor)

1993 20KE20: Excavations at a Prehistoric Copper Workshop. Michigan Archaeologist 39(3-4):127-193.

Martin, S. R.

1999 Wonderful Power: The Story of Ancient Copper Working in the Lake Superior Basin. Wayne State University Press, Detroit.

Meinholz, N. M. and S. R. Kuehn

1996 The Deadman Slough Site: Late Paleoindian/Early Archaic and Woodland Occupations along the Flambeau River, Price County Wisconsin. Archaeological Research Series, No. 4. Museum Archaeology Program, State Historical Society of Wisconsin.

Moffat, C. R. and J. M. Speth

1999 Rainbow Dam: Two Stratified Late Archaic and Woodland Habitations in the Wisconsin River Headwaters. The Wisconsin Archeologist 80(2):111-159.

Morrow, C. A.

1984 A Biface Production Model for Gravel-Based Chipped Stone Industries. Lithic Technology 13:20-28.

Morrow, T.

1994 A Key to the Identification of Chipped Stone Raw Materials Found on Archaeological Sites in Iowa. Journal of the Iowa Archaeological Society 41:108-129.

Pauketat, T.

1993 Lithic Analysis. In 20KE20: Excavations at a Prehistoric Copper Workshop, edited by S. R. Martin. Michigan Archaeologist 39(3-4): 176-180.

Pleger, T. C.

1998 Social Complexity, Trade, and Subsistence During the Archaic/Woodland Transition in the Western Great Lakes (4000-400 B.C.): A Diachronic Study of Copper Using Cultures at the Oconto and Riverside Cemeteries. Ph.D. Dissertation, Department of Anthropology, University of Wisconsin-Madison. University Microfilms, Ann Arbor.

2000 Old Copper and Red Ocher Social Complexity. Midcontinental Journal of Archaeology 25(2):169-190.

2001 New Dates for the Oconto Old Copper Culture Cemetery. The Wisconsin Archeologist 82(l-2):87-100.

Pletka, K.

1991 Chemical Trace Element Analysis of Native Copper. Michigan Archaeologist 37(4):283-297.

Rapp, G. Jr., E. Henrickson, M. Miller, and S. Aschenbrenner

1980 Trace-Element Fingerprinting as a Guide to the Geographic Sources of Native Copper. Journal of Metals 32(l):35-45.

Rapp, G. Jr., J. Allert, V. Vitali, Z. Jing. and E. Henrickson

2000 Determining Geological Sources of Artifact Copper. University Press of America.

Renfrew, C.

1972 The Emergence of Civilization: The Cyclades and the Aegean in the Third Millennium B.C. Methuen & Co., London.

Salzer, R.

1974 The Wisconsin North Lakes Project: A Preliminary Report. In Aspects of Upper Great Lakes Anthropology, Papers in Honor of Lloyd A. Wilford. Edited by Eiden Johnson. Minnesota Prehistoric Archaeology Series, No. 11.

Shott, M. J.

1996 An Exegesis of the Curation Concept. Journal of Anthropological Research 52(3):259-280.

Stoltman, J. B.

1997 The Archaic Tradition. The Wisconsin Archeologist Special Issue 78(1-2): 112-139.

Vernon, W. V.

1990 New Archaeometallurgical Perspectives on the Old Copper Industry of North America. In Archaeological Geology of North America, edited by N. P. Lasca and J. Donohue, pp. 479-498. Geological Society of America, Centennial Special Volume 4. Boulder.

Vogel, J. C., A. Fuls, E. Visser, and B. Becker

1993 Pretoria Calibration Curve for Short Lived Samples. Radiocarbon 35(1):73-86.

Wittry, W. L.

1951 A Preliminary Study of the Old Copper Complex. The Wisconsin Archeologist 32(1):1-18.

Wray, C. F.

1948 Varieties and Sources of Flint Found in New York State. Pennsylvania Archaeologist 18(1-2):25-47.

Mark A. Hill

Department of Anthropology, Washington State University, Pullman, WA 99164 (mark_hill@wsu.edu)

Copyright Office of the State Archaeologist Fall 2006

Provided by ProQuest Information and Learning Company. All rights Reserved