https://orcid.org/0000-0002-6084-8420
Figures
Abstract
North America’s ancient copper use, predicted to originate as early as 9000 cal BP, represents the earliest use of native copper for utilitarian tool production in the world. Although recent work has focused on establishing the first use of copper in the western Great Lakes region, little attention has been paid to determining the age ranges of subsequent copper using groups or to the identification of broader trends in copper use during the Archaic Period (10,000–3000 RCYBP). Here we address this issue by applying Bayesian modeling to a comprehensive suite of 76 radiocarbon dates directly associated with copper use. Our results identified two distinct peaks in copper usage, ca. 5500 cal BP and ca. 3300 cal BP. Age ranges for the three Archaic Period traditions and practices associated with copper use of the western Great Lakes are revised using modern calibration curves. Bayesian revisions of age ranges from sites where copper tools and/or production debris have been found provide insight into the historical relationships between, and cultural interactions among, these early copper using groups. This study provides an updated, refined chronology based on the most recent calibration curve (IntCal20) for the varied cultural contexts of copper use across the western Great Lakes.
Citation: Bebber MR, Buchanan B, Holland-Lulewicz J (2022) Refining the chronology of North America’s copper using traditions: A macroscalar approach via Bayesian modeling. PLoS ONE 17(4): e0266908. https://doi.org/10.1371/journal.pone.0266908
Editor: John P. Hart, New York State Museum, UNITED STATES
Received: September 13, 2021; Accepted: March 29, 2022; Published: April 26, 2022
Copyright: © 2022 Bebber et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
North America’s ancient copper use, predicted to originate as early as 9000 cal BP [1], represents the earliest use of native copper for utilitarian tool production in the world. However, although recent work has focused on establishing the “first” use of copper in the western Great Lakes region, little attention has been paid to addressing the temporalities of subsequent copper using traditions and practices. Here we address this issue by applying Bayesian modeling to a comprehensive suite of 76 radiocarbon dates, directly associated with copper use, to identify temporal patterns over the Archaic Period (10,000–3000 RCYBP). Age ranges for the three Archaic Period traditions and practices associated with copper use of the Great Lakes—traditionally defined by the culture-historical monikers Old Copper, Burnt Rollways, and Red Ocher—are revised using modern calibration curves. We recognize the heuristic nature of these essentialized groupings, though we also recognize that each of these are associated with particular traditions and practices that give varying social, cultural, and historical context to the long-term use of copper in the region. As such, through the refinement of the age ranges associated with these traditions, we build towards a more chronologically sound and temporally complex landscape of copper using societies, traditions, and histories that defined the western Great Lakes region between ca. 10,000–3000 RCYBP. In particular, we attempt to resolve issues related to the interactions of diverse copper-using societies, to the varied historical trajectories of copper-using traditions, and to the potential movements, evolutions, and transformations of particular practices across a complex sociocultural matrix of groups living around the western Great Lakes. To further this goal, we include 12 dates from Archaic Period copper sites in the Ottawa River Valley, Canada, which are often excluded from broader discussions of early copper use in the Americas. In this regard, this research program is situated within a broad, macroscalar perspective to better temporally situate the peoples involved in the production, use, and exchange of copper materials at geographic scales that are not often considered.
The Archaic Period
The onset of the Archaic Period in North America varies by region, beginning earlier in the southern areas of the Eastern Woodlands (ca. 10,000 BP [2]) compared to regions surrounding the Upper Great Lakes, where certain Paleoindian technologies continue until at least 8000 BP and possibly as late as 7545 BP [3, 4]. Indeed, archaeological evidence suggests overlap in material culture, namely stone projectiles, associated with both Late Paleoindian and Early Archaic groups [5–12]. Whether these data signal a co-mingling between discrete peoples, or a historical relationship between Late Paleoindian and Early Archaic groups is a topic not fully understood. This problem generates uncertainty regarding the onset of the traditions associated with the Archaic Period in the Upper Great Lakes.
Given the long duration of the Archaic Period (10,000–3000 BP), it is commonly discussed in terms of three subphases, termed simply Early (10,000–8000 BP), Middle (8000–5000 BP), and Late (5000–3000 BP) Archaic. However, as with the onset and termination of the Archaic Period in general, the precise timing of these subphases varies by region, with diagnostic characteristics occurring earlier in the more southerly regions and expanding northward in a time-transgressive fashion. Broadly, the onset of the Early Archaic is signaled by the presence of stemmed and notched bifacial flint tools. Surprisingly little is known about human lifeways during this era due to a paucity of archaeological sites, but it is generally accepted that groups were small, dispersed, highly mobile, and occasionally used rockshelters [13–16]. Aquatic and wetland resources were heavily exploited in the early Holocene as the climate of the Upper Great Lakes was wetter and cooler than in the Middle Archaic when the climate became warmer and began to stabilize [17, 18]. During the Middle Archaic archeological materials increase, and the first mortuary focused sites appear in North America. Based on assemblage data, researchers have inferred that Archaic groups were broad-spectrum foragers who would often seasonally aggregate to share resources and exchange items. In the western Great Lakes, fishing and aquatic resources were important, along with a wide variety of woodland game, fowl, nuts, and wild plants [13, 19].
Of central interest here are the temporalities of the various traditions, practices, and histories of groups who inhabited this region and were the direct users of native copper as tools or ornaments. Despite sustained interest in establishing the timing for copper usage via radiocarbon dating throughout the second half of the twentieth century, this topic has received little attention in the past decade. Early research on North American Archaic Period copper use focused on broad topics of cultural origins and delineation, primarily using excavation data from cemeteries [20–23] and analysis of artifacts collected from the surface or other non-mortuary archaeological contexts [24–27]. Later studies focused on the role of copper in Archaic Period societies [28], the nature of tool use [29, 30], tool production techniques [31–33], and copper sourcing [34–38]. Recent research has focused on the nature of exchange-based relationships [39, 40], population dynamics and changing social structure [41], as well as experimental research on copper tool functional efficiency and the interrelated factors affecting technological change [42–45].
Indeed, although North American copper usage has been a topic of interest for centuries, the chronological relationships and temporalities for the various copper using groups are still relatively uncertain, mainly due to small sample sizes and the vagaries of dating [46]. Here we use Bayesian modeling to establish updated chronologies for three distinct cultural phenomenon differentially inscribed across the western Great Lakes landscape, 1) Middle Archaic mortuary sites located far from copper sources, referred to herein as “Old Copper” traditions, 2) Late Archaic mortuary sites located nearer to copper sources referred to herein as “Red Ocher” traditions, and 3) Late Archaic production sites located near copper sources, referred to herein as “Burnt Rollways” traditions. In contrast to extant work that treats these traditions as separate “cultures” or “peoples,” we instead conceptualize these as merely three different kinds of archaeological sites that represent different and varied cultural expressions and practices. These differences, described in detail below, are primarily defined by distinct mortuary practices, unique uses of copper, and the presence or absence of associated habitations and tool-working areas. The goal is to model age ranges for each of these site types, which have only to this point used the “eye ball” approach, masking important temporalities and chronological relationships between distinct social, cultural, and economic practices [47]. Once revised age ranges have been established, we can move beyond simple associations of the archaeological record with distinct “cultures” to better determine the historical relationships, as well as clarify when/if there were opportunities for interrelationships to have existed between the peoples who frequented these functionally distinct sites. Here we provide a Bayesian revision for the age ranges for the three main copper using cultural phenomena of the Archaic Period in the Great Lakes region in order to better clarify their temporal position and to identify broader patterns in native copper tool occurrence.
Copper-using groups of the Great Lakes
“Old Copper”.
North America’s copper-using peoples of the Archaic Period were initially classified together as one broad technological phenomenon that was referred to by a variety of designations such as the “Old Copper Culture” [21, 48, 49] the “Old Copper Complex” [23, 50], or the “Old Copper Industry” [51, 52]. Regardless of naming convention, the “Old Copper” phenomenon was formally defined by the presence of large utilitarian copper tools which were forged from native copper and found in the Great Lakes region [15].
In recent decades, researchers have delineated other culturally and historically distinct uses of copper and now generally consider Old Copper to be the first cultural manifestation of groups linked by copper tool technology, both utilitarian and decorative. The age range for Old Copper is generally thought to extend from approximately 6780–3100 cal BP [15] (dates [14] calibrated using IntCal20). This suggests that the Old Copper traditions and the earliest copper use in the mortuary contexts began in the Middle Archaic (ca. 8000–5000 BP) and may have extended into the Late Archaic (ca. 3000–1000 BP). The Old Copper temporal range was initially defined based on evidence from Reigh [21, 53], Osceola [20, 21] Oconto [21, 54], and Price III [55]. These four sites represent the type sites for early investigations of copper use in the western Great Lakes and provided the basis for the definition of Old Copper [23] as a specific set of early practices.
“Red Ocher”.
Red Ocher sites are defined as a set of mortuary practices that are both more complex and found nearer to copper sources than earlier Old Copper mortuary practices. These sites are located throughout the Midwest, spanning Iowa and the Upper Mississippi valley in the west, across southern Michigan, and eastward into southern Ontario and Ohio. Red Ocher traditions were defined largely on a suite of burial practices that are in many ways similar to those of Old Copper sites but exhibit increased mortuary complexity. Specifically, Red Ocher cemeteries contain remains that have been covered with hematite powder, representing a distinctly new practice associated with societies using copper. These burials often include caches of burnt ceremonial bifacial blades and in some instances contain marine shell beads [41, 56]. In the upper western Great Lakes, Red Ocher mortuary contexts contain copper materials as grave offerings. Of interest is that these Red Ocher burial sites evince a shift in copper goods away from the large utilitarian tools that were regularly associated with Old Copper burials, towards smaller copper items that are interpreted as objects of personal ornamentation. In addition to the shift in copper forms, there was an overall increase in prestige goods, exemplified by the presence of worked marine shell, exotic cherts, and obsidian [41].
The Early Woodland tradition—delineated by the presence of burials in human-made mounds, the use of pottery, and the adoption of horticulture—appeared around 3000 BP in regions south of the Great Lakes. However, such patterns do not appear in the Upper Great Lakes until sometime after 2000 BP. Given the time-transgressive nature of material culture traditions, there is ambiguity in precisely how the Red Ocher practices fit into the Late Archaic-Woodland transition of the Upper Great Lakes and how these traditions articulated with those further south. Indeed, in some instances, late Red Ocher sites contain pottery and burials in human-made mounds [57, 58] both formal characteristics of Woodland traditions [15]. However, the majority of Red Ocher sites do not have ceramic materials or human-made mounds. As with the preceding Old Copper tradition, little is known regarding Red Ocher social organization or subsistence strategies as no habitation sites have been excavated. However, mortuary data has provided insight into emerging social complexity during this time as mortuary centers include a greater frequency of exotic goods and personal ornaments, many of which occur with infant and young adult female burials [41].
“Burnt Rollways”.
The Burnt Rollways expression was first defined by Salzer [11] and has been further clarified with subsequent investigations [39, 59–61]. In contrast to other copper using traditions, which were defined primarily through their associations with particular mortuary programs, Burnt Rollways sites frequently contain evidence of copper tool production along with the associated debris [11, 39, 59].
Although not nearly as well-known as the Old Copper or Red Ocher traditions, Burnt Rollways sites are vital for understanding the trajectory of cultural change and interaction among copper-using groups, as Burnt Rollways sites are located nearer the copper sources in the southern Lake Superior Basin (Fig 1). Likewise, in contrast to sites from Old Copper and Red Ocher, the Burnt Rollways sites are associated with evidence for habitation that provides insight into the lifeways of copper producers during the Late Archaic period–data that are lacking for the earlier Middle Archaic Old Copper components, where most data come from mortuary contexts several hundred kilometers away from the copper deposits. Burnt Rollways sites were situated farther north than both Old Copper and Red Ocher sites with many of the sites located near the headwaters of the Wisconsin, Wolf, Menominee, and Ontonagon Rivers [40].
Radiocarbon dates from the archaeological sites shown here were used to generate age ranges for three Archaic Period copper-using traditions—Old Copper, Burnt Rollways, and Red Ocher. Republished from MapChart.net under a CC BY license, with permission from Minas Giannekas, original copyright 2021. Modified by MRB using Adobe Photoshop version 13 x64.
Therefore, given the proximity of Burnt Rollways sites to copper sources, and their temporal position which may have overlapped with both the Old Copper and Red Ocher mortuary complexes, it is possible that the Burnt Rollways sites functioned as the distributional hub for copper during the Late Archaic to Woodland transition. However, to fully understand the relationships between copper procurement in the mining region and possible distribution to the southerly ceremonial centers, contemporaneity between these different kinds of sites must be established.
Bayesian chronological modelling
Bayesian statistics are uniquely suited for the analysis of radiocarbon data because of their focus on probabilities. As extensive overviews of Bayesian analysis of radiocarbon dates have been published elsewhere [62–67] only a brief introduction is provided here and below. A more thorough review of the methodology, concepts, and terminology are also included with the S1 File. The results of “scientific dating are always interpreted contextually” and Bayesian statistics “provide an explicit, quantitative method which can combine raw dates with other prior information included in a model to produce formal statistical date estimates which combine both sets of evidence” [63]. In the following case, radiocarbon determinations represent likelihoods. The association of radiocarbon dates with each of the three distinct kinds of copper-bearing archaeological sites represent our prior beliefs. The critical assumption made here by grouping radiocarbon dates by particular practices (identified through their associations with particular tool technologies, settlement patterns, and mortuary traditions) is that these resulting groups can be used to then model the starts, ends, and spans of these associated traditions and for each of the three kinds of copper-bearing sites. Bayesian techniques allow us to formally produce such temporal estimations in place of “eye-balling” large radiocarbon datasets. Importantly, we employ formal techniques to assess the order and contemporaneity of temporal events (e.g., the starts and ends of the three copper-using traditions) rather than through less rigorous, visual assessments of radiocarbon datasets.
Data and methods
The dataset used here (Table 1) includes available radiocarbon dates associated with three distinct kinds of copper-bearing sites or traditions associated with the use of copper: Old Copper (n = 21), Burnt Rollways (n = 5), and Red Ocher (n = 16). A total of 42 dates from 11 sites associated with these three traditions are used to build a series of chronological models. Radiocarbon dates were used only from sites that contained copper artifacts or evidence of copper working. Additionally, 34 radiocarbon dates from 21 sites associated with copper, but not specifically associated with any of three sets of traditions explored here, were also used to investigate the total dataset of copper-associated radiocarbon dates from the region. Of the 76 measurements, 32 were made on charcoal, 17 on unidentified wood or wood bark, 13 on human bone, 4 on unidentified animal bone, 4 on cordage or plant fiber, 3 on unidentified charred wood, 2 on leather, and 1 on subcarbonized organic material. Of the 76 available dates, 9 dates were excluded for having error ranges over 150 years. This is an arbitrarily defined cut-off, but most of the dates excluded exceeded error values between 250 and 600. Though these dates were excluded from inclusion in the primary models, all models were also run with these dates and are included as alternative models in the S1 File.
All dates come from contexts directly associated with copper.
It is important to acknowledge that the dataset of available radiocarbon dates is relatively small given the geographic and temporal ranges expected for the cultural practices investigated here. While a small sample size does represent a constraint on the confidence of model outputs, the Bayesian framework employed here actually provides a more formal, empirical assessment of chronological placement given the currently available data. In contrast to approaches that merely "eye-ball" temporal ranges and chronological trends, Bayesian approaches provide a formal framework for assessing probabilities, even from small datasets. In fact, the model structures provided here, and in the S1 File, might serve as a framework for incorporating both new radiocarbon dates and new prior information as either becomes available. Additionally, the simulations we describe below are meant to offer full transparency on the potential effects of our small sample sizes on model outputs and on the potential for our models to change as new information is added. In sum, regarding sample size, the models presented here represent the best and most empirical chronologies and temporal ranges that can be estimated given currently available data and provide a formal framework for future chronological investigations.
Three separate models, one for radiocarbon dates associated with each of the three sets of traditions, were constructed independently of one another. Dates associated with the same kind of copper-associated sites were grouped first by site and then in an overall phase representing the practice of particular traditions (e.g., Old Copper). Each of the three models (Old Copper, Burnt Rollways, and Red Ocher) incorporated trapezium boundaries to estimate start and end boundaries. Trapezium boundaries are commonly used to model cultural phenomena that have an unknown length of period across which they began, persisted, and ended [68–71]. Trapezium boundaries allow for varied temporalities of adoption, use, and abandonment of particular traditions or practices. In general, they represent a more flexible set of modeling parameters compared to other kinds of boundaries (e.g., sigma boundaries; see Alternative Model D in the S1 File for outputs produced using sigma boundaries for comparison). A charcoal outlier model was applied to dates obtained from charcoal to account for potential discrepancies between unidentified wood charcoal dates within the model [72]. Additionally, a general outlier model was applied to all non-charcoal dates to identify any potential outliers and to down weight those determinations that exhibit heightened potentials to be outliers.
In addition to the primary models described above, a series of alternative models were included as a form of sensitivity analysis. By altering model parameters, we were able to assess the effects of different modeling decisions on chronological estimations and to assess the propensity for the model to change given different parameters. The results of the Bayesian modeling efforts (modeled starts, ends, and spans) represent the posteriors. All models were built using OxCal v 4.4 [65, 73] and the IntCal20 calibration curve [74]. Full descriptions and OxCal code for all primary, alternative, and simulated models are included as S1 File.
An additional step was taken using simulations to determine the representativeness of our own radiocarbon dataset and to model chronological scenarios given an ideal dataset [64, 75–79]. As such, in addition to the results gleaned from our radiocarbon dataset, we simulated how the results might change with the addition of a more robust set of radiocarbon dates. To do this, we ran a number of simulations to determine the number of dates needed to produce models that 1) exhibit high levels of precision in start and end boundary determinations and that 2) do not change with the addition of more radiocarbon dates. The full procedures and methods of determinations are presented in the S1 File.
Results
The three primary models were used to produce start boundary ranges, end boundary ranges, and spans for Old Copper, Burnt Rollways, and Red Ocher (Figs 2–7). The results at both the 68% and 95% confidence intervals are presented in Table 2. Modeled results in text are cited at the 68% confidence interval and presented in italics. Old Copper traditions are estimated to have begun between cal BP 7520–6180, ended between cal BP 4840–4190, and spanned a length of 1,520–3,200 years. Burnt Rollways is estimated to have begun between cal BP 4780–3880, ended between cal BP 2610–1630, and spanned a length of 1,500–3,120 years. Red Ocher is estimated to have begun between cal BP 3240–2980, ended between cal BP 2260–1870, and spanned a length of 790–1,290 years.
Modeled start and end boundaries from both the primary models (top) and simulated models (bottom) for Old Copper, Burnt Rollways, and Red Ocher.
Modeled spans (produced using the interval command) from both the primary models (left) and simulated models (right) for Old Copper, Burnt Rollways, and Red Ocher.
The order command in OxCal was used to statistically determine the appropriate sequence of start and end boundaries for each of the three traditions (Table 3). While it is quite clear that Old Copper practices began far earlier than both Burnt Rollways and Red Ocher practices, the timing for the end of Old Copper and the beginning of the two other traditions remains ambiguous. The probability that Old Copper practices ended before the start of Red Ocher practices is high. There is, however, just a 45% chance that Old Copper practices ended before the start of Burnt Rollways practices. Similarly, the temporal relationship between the copper working and habitation Burnt Rollways sites and the mortuary Red Ocher sites is ambiguous. While it is clear Burnt Rollways sites begin to be used before Red Ocher practices develop, it is difficult to determine whether or not Red Ocher mortuary traditions outlast the habitation and tool-working Burnt Rollways sites. It may be the case that habitation sites, copper-processing sites, and the Red Ocher mortuary practices all cease at roughly the same time.
To increase the precision of potential chronological placements based on available information and model parameters, a simulation experiment was undertaken to determine how many dates would be needed to resolve some of these sequencing issues and to achieve a more robust chronological framework that would be unyielding to new radiocarbon data. The full description of these procedures is reported in the S1 File along with all the OxCal code for each of the simulated models. In short, the modeled chronological ranges produced using the primary models were used to produce simulated chronological ranges. To do so, simulated dates that fell within the ranges produced through the primary models were added until the modeled age estimates 1) ceased to substantively change with new data and 2) provided a precision for start and end boundaries of less than 200 years. It should be noted that these simulated results are representative of actual observations only if new observations (dates) indeed follow the trends identified in the currently available datasets. New radiocarbon dates that fall substantively earlier or later than currently available dates would more substantively change estimated chronological boundaries. The simulated results here thus represent a scenario consistent with extant datasets and currently available observations.
Given the length of these temporal periods and their association with particular parts of the radiocarbon calibration curve, it was determined that ca. 90 additional radiocarbon dates for each of the three traditions (a total of 270 additional radiocarbon dates) would be necessary for achieving a desired precision and representativeness. The point of this exercise is not necessarily to call for such a program of data collection, but rather to create an environment of transparency and to explore the limitations of the currently available dataset [78]. The simulation results should be used as a hypothetical baseline against which to assess the representativeness and robustness of the radiocarbon dates that do exist. Importantly, the only experiments run here were those that added more dates. Further simulations that build in stratigraphic relationships between radiocarbon dates, radiocarbon dates from more sites, or new contextual associations between dates could certainly produce representative models that necessitate far fewer than 270 dates. Nevertheless, the simulated models provide refined ranges for each of the three traditions and lend insight into potential scenarios of their temporal relationships to one another. The results of these simulations are presented next to the results of the primary models in Figs 5 and 6 and in Table 2. Using these simulations based on currently available information, Old Copper is simulated to have begun between cal BP 7200–7070, ended between cal BP 4530–4410, and spanned a length of 2,570–2,740 years. Burnt Rollways is simulated to have begun between cal BP 4510–4400, ended between cal BP 2090–1990, and spanned a length of 2,380–2,530 years. Red Ocher is simulated to have begun between cal BP 3170–3070, ended between cal BP 2140–2070, and spanned a length of 930–1,050 years.
Once again, the order command in OxCal was used to statistically determine a simulated sequence of start and end boundaries for each of the three traditions (Table 3). It remains clear that the Old Copper practices began much earlier than the other two sets of practices. There is still ambiguity as to whether Burnt Rollways sites began to be occupied before Old Copper practices ceased, with there being a 60% chance that Old Copper practices ended before the start of Burnt Rollways occupations. This likely indicates a potential scenario in which the end of Old Copper mortuary practices and tool-making traditions are contemporaneous with the beginning of Burnt Rollways habitation sites. In any case, Old Copper practices ended, and Burnt Rollways sites began to be occupied, before the Red Ocher traditions began to develop. However, Red Ocher mortuary practices and copper-working traditions certainly began before Burnt Rollways sites were all abandoned, indicating that the two overlap significantly in time. In fact, there is a 90% chance that the end of Red Ocher practices preceded the end of Burnt Rollways practices, suggesting that the entirety of the Red Ocher tradition was contemporaneous with Burnt Rollways habitation and copper-working sites.
It is important to reiterate, however, that the simulations are constrained by what can be gleaned from the extant dataset of radiocarbon dates. As such, new data could certainly alter these results. New radiocarbon dates that fall within the maximum and minimum bounds produced using the current dataset will likely not have a substantive effect on results and may potentially match the results produced through simulations. New dates that fall beyond the minimum and maximum ranges represented by the extant dataset may be more likely to substantively alter the results presented here. This caveat may be most relevant to the timing of Burnt Rollways, where the current dataset consists of just five radiocarbon dates associated with copper materials. The likelihood of obtaining new radiocarbon determinations that expand the range of Burnt Rollways thus remains high. That said, the current dataset of Burnt Rollways dates (five dates from four sites) does represent a potentially robust set of dates that produce generally unchanging results when modeled multiple times, exhibiting little variance in model results over many runs (e.g., modeled boundaries vary by less than 200 years over ten runs; see S1 File for a test of variance for Burnt Rollways modeled boundaries). Unfortunately, Burnt Rollways sites, representing relatively inconspicuous practices like ephemeral habitations and copper-working/processing, are likely least represented both spatially and temporally of the three kinds of sites explored here. While these five dates from four sites consistently produce similar modeled results, new archaeological work may likely produce a wider temporal range for such sites than presented here.
All 76 available radiocarbon dates were also used to produce a kernel density estimate plot. The purpose of the KDE plot is simply to assess the overall distribution of the dataset. The plot itself represents the distribution of dates along a temporal gradient (the means of those dates are represented as white circles below the x-axis in Fig 7). Kernel density estimates take into consideration the frequency of observations (radiocarbon dates), not simply the raw sum of their probability distributions, as in the case of summed probability distribution plots [80]. In this way, a plot is produced that represents the probability of a radiocarbon date at a specific point in time. Such frequentist approaches like kernel density estimates have routinely been used for spatial analyses to estimate the probability that something will occur at a specific place given the frequency of observations of that thing occurring there and elsewhere.
The 76 dates associated with copper from across the region span a great deal of time from ca. 8500 cal BP to ca. 250 cal BP. In the KDE plot presented here (Fig 7), there are three conspicuous features, two broad peaks and a subtle valley between them. The first peak is centered at ca. 5500 cal BP, the second peak at ca. 3300 cal BP, and the valley spanning roughly 500 years between 4400 and 3900 cal BP. It is clear that these peaks are associated with dense clusters of radiocarbon dates whereas the valley is associated with a paucity of dates between these two dense clusters. Whether or not this gap is real, or a product of sampling and archaeological coverage, remains unclear, but its investigation may present fruitful directions for future research.
Discussion
The results of the models reveal several thought-provoking possibilities, and in some ways, introduce more questions than they can fully answer. For example, although copper use likely began, at least in small scale, by Late Paleoindian times [1], Bayesian models show that a unifying cultural tradition does not become evident until sometime after 8000 years ago. This is not to say that shared traditions involving copper did not begin earlier, but simply that, based on current evidence, the first archaeologically identifiable cultural manifestation linked by an abundance of utilitarian copper tools is that of Old Copper traditions. Indeed, the results of the Bayesian analyses suggest that Old Copper practices began and ended substantially earlier than previously thought and also had a much shorter span (Table 2). These modeled results are informative, however, many questions remain. In particular, given that there are so few early dates from archaeological sites located near the copper sources [81], it is still uncertain precisely who conducted the earliest copper mining in this region. Nor is it fully understood how or why the predominance of copper tools have been found in areas far removed from the copper mines, while there is a paucity of copper tools found near the mining sources. This gap in our understanding is exacerbated by the ephemerality of archaeological materials that could provide insight into the lifeways of Early-Middle Archaic groups in the upper Great Lakes.
Based on the available archaeological data, we suggest that during the Middle Archaic, groups with modes of mortuary practices involving the inclusion of large, utilitarian copper tools (“Old Copper”) made seasonal trips to copper sources for procurement, and then may have returned south to the larger mortuary sites as part of their seasonal rounds. If these groups were small and mobile, it would explain the relative lack of habitation sites during the Early-Middle Archaic period in northern Wisconsin and the Upper Peninsula of Michigan. As successfully argued by others, evidence suggests that during the Middle Archaic, long-distance, task-specific movements were the norm, and groups engaged in such task-specific activities are nearly archaeologically invisible [16].
Alternatively, we acknowledge it is possible that there were larger scale habitation sites in the mining district, possibly even antecedent to the Burnt Rollways sites, that simply have not yet been found. Of interest is that the dates modeled here push the initial start date of copper-bearing habitation sites and tool-working sites (“Burnt Rollways”) back by nearly a millennium (Table 2). It is certainly possible given the age range for Burnt Rollways sites, that these practices developed from Old Copper traditions. However, the modeled age ranges show that, even though there was possibly temporal overlap, Burnt Rollways sites were not contemporary with the apogee of Old Copper practices. Thus, due to temporal disparities, it is unlikely that these northern Burnt Rollways sites, would have been involved in the distribution of copper to the Old Copper mortuary centers to the south. Though, once again, this temporal disparity may be the product of a paucity of dated habitation and tool-working sites more generally, and future work may reveal a more intensive temporal overlap between the Old Copper mortuary practices and these habitation sites located near copper resources.
Considering the geographic location it is possible that the Burnt Rollways sites represent an entirely different group of copper users in the northern region—one that is currently not well understood—who were focused on copper production and possibly distribution to other Late Archaic communities and groups in the surrounding region. Certainly, it is quite possible that Burnt Rollways is a later manifestation of an Early-Middle Archaic group who had inhabited the upper Great Lakes area but is currently unknown from the material record. Given that there were at least two groups with distinct practices coexisting in the area during the Late Paleoindian/Early Archaic transition, it is possible that these groups persisted in the area as separate entities who eventually formed and maintained their own unique traditions. Of interest to untangling these possible historical relationships, is evidence showing that as early as Late Paleoindian times, cremation was being practiced as a mortuary tradition [5] in the upper Great Lakes, possibly by people who later practiced communal mortuary ritual evinced in the Middle-Late Archaic Periods. Perhaps Late Paleoindian groups, who were the first to begin using native copper, developed in situ cultural traditions focused on copper mining and production, which eventually became archaeologically visible as Burnt Rollways sites. In fact, although habitation sites are largely missing, the earliest dated copper isolates [1, 82] are found in the northern regions close to the mining districts, and these copper projectiles date to the Early-Middle Archaic Period. Likewise, in the Keweenaw mining district, lead pollution from native copper processing first appears around 8000 cal BP but then decreases dramatically after 5000 cal BP [83], just prior to the initial occupations of Burnt Rollways sites.
According to the Bayesian models, Red Ocher traditions in the copper region only spanned a millennium. Given that Red Ocher mortuary practices and copper-working traditions began after the initial occupations of Burnt Rollways sites and was largely contemporaneous with these sites, multiple scenarios are possible regarding their potential relationships.
It is reasonable to conclude that there may have been relationships based on exchange of materials between peoples producing Burnt Rollways sites and peoples engaging in the complex Red Ocher mortuary practices. Or it could simply be that these various site types represent one group of people who frequented temporary production sites and small seasonal habitation sites, but were ultimately linked by a larger mortuary complex such as that at the Riverside site.
Regarding peaks in radiocarbon dates as gleaned from the KDE plot (Fig 7), the first peak at 5500 cal BP likely represents a period when copper was being extensively mined, processed, and used for a wide variety of large, utilitarian items by groups who are unfortunately not yet fully understood. Of interest is that by the second peak at 3300 cal BP, evidence for intensive copper mining has largely ceased [83]. A possible interpretation is that there were no longer seasonal trips to the mining district for large-scale copper extraction and production efforts. Unfortunately, the precise reasons for this change are currently unknown. It is possible that groups became more regionally restricted around 4500 years ago, at the same time that Old Copper mortuary practice and tool-making styles were disappearing from cultural repertoires. Limited evidence, in the form of a stone projectile found embedded in a human vertebra at the Price III site [55], suggests intergroup violence and competition were factors affecting human lifeways by the end of the Middle Archaic. Indeed, increasing regionalization and territorialism could also explain the disparities in material culture between Burnt Rollways and Red Ocher sites [40].
Likewise, if those participating in Old Copper practices who had previously practiced high levels of seasonal mobility—heading north in the warmer months to exploit aquatic resources and extract copper, and then south to mortuary centers and rockshelter hunting camps during cooler months—could no longer do this, subsequent groups would be forced to rely on one of two options for acquiring copper 1) sourcing “float” copper from local glacial till, and/or 2) recycling. Likewise, change in mobility patterns could also explain shifts in copper tool forms. If groups partaking in the complex mortuary practices and social traditions of the Red Ocher variety could no longer source large amounts of copper from the north, they would be forced to rely on recycling of copper material from large Early-Middle Archaic copper tools to produce new copper items. It follows that, recycling copper materials would ultimately lead to smaller artifacts being made from copper (i.e. awls and beads), consistent with Red Ochre traditions’ preference for smaller, ornamental copper pieces, which happen to be the two most common types of copper artifacts to persist into the Late Pre-contact Period [44, 45]. Additionally, as access to copper became more restricted and the raw material became less readily available, the social value of copper may have increased due to the scarcity principle [84]. This in turn may have intensified the prestige value placed on copper goods, which could explain the subsequent proliferation in copper ornaments after 3000 BP and shaped the social institutions and practices that defined the Red Ocher traditions [41].
Conclusion
The trajectory of native copper use in the western Great Lakes of North America is a fascinating topic with global significance, however, it is also a topic that presents unique challenges. Nevertheless, these problems can be solved via the application of new techniques in dating, modeling, paleoenvironmental research, and experimental archaeology. Here, in an effort to better understand the timing of the various social, cultural, and economic practices associated with copper-use that occurred throughout the Archaic Period, and to elucidate the potential interrelationships between the copper-using groups involved in these expressions, we provide a revised and updated chronology via Bayesian modeling. The updated age ranges inform our understanding of the historical relationships between, and cultural interactions among, North America’s earliest copper-using groups. Unlike previous work focusing on fine-scale sequences of copper use at particular localities, the research presented here has been situated within a macroscale perspective to better temporally situate the peoples involved in the production, use, and exchange of copper materials across a socially, culturally, and historically complex landscape.
Supporting information
S1 File. Supplementary material file including alternative models, simulations, terminology, and code.
https://doi.org/10.1371/journal.pone.0266908.s001
(PDF)
References
- 1. Bebber MR, Key AJM. Optimal Linear Estimation (OLE) Modeling supports early Holocene (9000–8000 RCYBP) Copper Tool Production in North America. Am Antiq. In Press.
- 2.
Sassaman KE. The Eastern Archaic, Historicized. Lanham: Altamira Press; 2010. https://doi.org/10.1212/WNL.0b013e3181c7197b pmid:19940271
- 3.
Buchner AP, Pettipas LF. The early occupations of the glacial Lake Agassiz basin in Manitoba 11500 to 7700BP. In: Lasca NP, Donahue J, editors. DNAG, Centennial Special Volume 4: Archaeological Geology of North America. Boulder: The Geological Society of America, Inc; 1990. p. 51–59.
- 4.
Martin SR. Wonderful power: the story of ancient copper working in the Lake Superior Basin. Detroit: Wayne State University Press; 1999.
- 5. Buckmaster MM, Paquette J. The Gorto site: preliminary report on a late Paleo-Indian site in Marquette County, Michigan. Wis Archeol. 1988;69(3):101–24.
- 6. Clark CP. Plano traditional lithics from the Upper Peninsula of Michigan. MidCont J Archaeol. 1989 Jun(35):88–112.
- 7.
Harrison C, Redepenning E, Hill CI, Rapp G, Aschenbrenner S, Huber JK, et al. The Paleo-Indian of southern St. Louis County, Minnesota: the reservoir lakes complex. University of Minnesota Interdisciplinary Archaeological Studies No. 4, Dubuque: Kendall/Hunt Publishing Company; 1995. 235 p.
- 8. Hunzicker DA. Little Rice Lake Site (47V1272): An Old Copper Complex manufacturing site in Vilas County, WI. Wis Archeol. 2002 83(1):45–62.
- 9.
Mason RJ. Great Lakes Archaeology. New York: Academic Press; 1981
- 10.
Meinholz NM, Kuehn SR. The Deadman Slough Site (47 Pr-46). Wisconsin Historical Society; 1996.
- 11.
Salzer RJ. The Wisconsin north lakes project: a preliminary report. In: Johnson E, editor. Aspects of Upper Great Lakes Anthropology. Minnesota Prehistoric Archaeology Series 11. St. Paul: Minnesota Historical Society, 1974. p. 40–54.
- 12. Steinbring J. A Copper Blade of Possible Paleo-Indian Type. Manitoba Archaeological Newsletter. 1968;5(1–2):3–12.
- 13.
Emerson TE, McElrath DL, Fortier AC, editors. Archaic societies: diversity and complexity across the midcontinent. Albany: SUNY Press; 2012.
- 14.
Prufer OH, Pedde SE, Meindl RS, editors. Archaic transitions in Ohio and Kentucky prehistory. Kent: Kent State University Press; 2001.
- 15.
Pleger TC, Stoltman JB. The Archaic Tradition in Wisconsin. In: Emerson TE, McElrath DL, Fortier AC, editors. Archaic societies: Diversity and complexity across the midcontinent. Albany: SUNY Press. 2009. p. 697–724.
- 16.
Lovis WA. Hunter-gatherer adaptations and alternate perspectives on the Michigan Archaic: Research problems in context. In: Emerson TE, McElrath DL, Fortier AC, editors. Archaic societies: Diversity and complexity across the Midcontinent. Albany: SUNY Press. 2009. p. 725–54.
- 17. Kuehn SR. New evidence for Late Paleoindian-Early Archaic subsistence behavior in the western Great Lakes. Am. Antiq. 1998 Jul;63(3):457–76.
- 18.
Lovis WA. The Middle Archaic: learning to live in the woodlands. In: Linebaugh JR, editor. Retrieving Michigan’s buried past: the archaeology of the Great Lakes state. Bloomfield Hills: Cranbrook Institute of Science; 1999. p. 83–94.
- 19.
Vickery KD. Archaic manifestations in southwestern Ohio and vicinity. In: Redmond BG, editor. Transitions: archaic and early woodland research in the Ohio country. 2008. p.1–28.
- 20. Ritzenthaler RE. The Osceola Site: An ’Old Copper’ Site near Potosi, Wisconsin. Wis Archeol. 1946;27(3):53–70.
- 21.
Ritzenthaler RE, editor. The Old Copper Culture of Wisconsin. Madison: Wisconsin Archaeological Society; 1957.
- 22. Wittry WL. A Preliminary Study of the Old Copper Complex. Wis Archeol. 1951;32(1):1–18.
- 23. Wittry WL, Ritzenthaler RE. The old copper complex: an Archaic manifestation in Wisconsin. Am Antiq. 1956;21(3):244–54.
- 24.
Griffin JB, editor. Lake Superior copper and the Indians: miscellaneous studies of Great Lakes prehistory. U of M Museum Anthro Archaeology; 1961.
- 25. Johnson E. Copper artifacts and glacial Lake Agassiz beaches. Min Archaeol. 1964;26(1):4–22.
- 26. McHugh WP. New Archaeology" and the Old Copper Culture. Wis Archeol. 1973 Jun;48:145–260.
- 27. Phillips GB. The primitive copper industry of America. Am Anthropol. 1925 Apr 1;27(2):284–9.
- 28. Binford LR. Archaeology as anthropology. Am Antiq. 1962 Oct;28(2):217–25.
- 29. Fregni G. A Study of Copper Spearheads in the Old Copper Complex. Min Archaeol. 2009;67:121–30.
- 30. Penman J. The Old Copper culture: an analysis of old copper artifacts. Wis Archeol. 1977;58(4):3–23.
- 31. LaRonge M. An experimental analysis of Great Lakes Archaic copper smithing. North Am. Archeol. 2001 Oct;22(4):371–85.
- 32. Peterson DH. Red metal poundings and the “Neubauer Process”: copper culture metallurgical technology. Cent States Archaeol J. 2003 Apr 1;50(2):102–6.
- 33. Peterson DH. Native copper characteristics demonstrated in the Neubauer Process. Cent States Archaeol J. 2003 Jul 1;50(3):168–23.
- 34. Ehrhardt KL. Copper working technologies, contexts of use, and social complexity in the eastern woodlands of native North America. J. World Prehist. 2009;22(3):213–235.
- 35. Hancock RG, Pavlish LA, Farquhar RM, Salloum R, Fox WA, Wilson GC. Distinguishing European trade copper and north‐eastern North American native copper. Archaeometry. 1991 Feb;33(1):69–86.
- 36. Levine MA. Determining the provenance of native copper artifacts from northeastern North America: Evidence from instrumental neutron activation analysis. J Archaeol Sci. 2007 Apr 1;34(4):572–87.
- 37. Mauk JL, Hancock RG. Trace element geochemistry of native copper from the White Pine Mine, Michigan (USA): Implications for sourcing artefacts. Archaeometry. 1998 Feb;40(1):97–107.
- 38. Rapp G, Henrickson E, Miller M, Aschenbrenner S. Trace-element fingerprinting as a guide to the geographic sources of native copper. JOM. 1980 Jan;32(1):35–45.
- 39. Hill MA. The Duck Lake site and implications for late Archaic copper procurement and production in the southern Lake Superior Basin. MidCont J Archaeol. 2006 Jan 1;31(2):213–47.
- 40. Hill MA. Tracing social interaction: Perspectives on Archaic copper exchange from the Upper Great Lakes. Am Antiq. 2012 Apr;77(2):279–92.
- 41. Pleger TC. Old copper and red ocher social complexity. MidCont J Archaeol. 2000 Oct 1:169–90.
- 42. Bebber MR, Eren MI. Toward a functional understanding of the North American Old Copper Culture “technomic devolution”. J Archaeol Sci. 2018 Oct 1;98:34–44.
- 43. Bebber MR, Key AJ, Fisch M, Meindl RS, Eren MI. The exceptional abandonment of metal tools by North American hunter-gatherers, 3000 BP. Sci Rep. 2019 Apr 8;9(1):1–4 pmid:30626917
- 44. Bebber MR, Norris JD, Flood K, Fisch M, Meindl RS, Eren MI. Controlled experiments support the role of function in the evolution of the North American copper tool repertoire. J Archaeol Sci Rep. 2019 Aug 1;26:101917.
- 45. Bebber MR. The Role of Functional Efficiency in the Decline of North America’s Copper Culture (8000–3000 BP): an Experimental, Ecological, and Evolutionary Approach. J Archaeol Method Theory. 2021 Jan 4:1–37.
- 46. Buchanan B, David Kilby J, Hamilton MJ, LaBelle JM, Meyer KA, Holland-Lulewicz J, et al. Bayesian Revision of the Folsom Age Range Using IntCal20. PaleoAmerica. 2021 Apr 3;7(2):133–44.
- 47. Holland-Lulewicz J. From Categories to Connections in the Archaeology of Eastern North America. J Archaeol Res. 2021 Jan 4:1–43.
- 48. Steinbring J. Old copper culture artifacts in Manitoba. Am Antiq. 1966 Apr;31(4):567–74.
- 49. Quimby GI, Spaulding AC. The old copper culture and the Keweenaw waterway. Fieldiana. Anthropol. 1957 Nov 8;36(8):189–201.
- 50.
Martin S, Pleger TC. The complex formerly known as a culture: The taxonomic puzzle of Old Copper. In: Williamson RF, Watts CF, editors. Taming the taxonomy: Toward a new understanding of Great Lakes archaeology. Provincetown: East End Books.1999. p.61–70.
- 51. Miles SW. A revaluation of the old copper industry. Am Antiq. 1951 Jan;16(3):240–7.
- 52. McKern WC. The first settlers of Wisconsin. Wis Mag Hist. 1942 Dec;26(2):153–69.
- 53. Baerreis DA, Daifuku H, Lundsted JE. The burial complex of the Reigh site, Winnebago County, Wisconsin. Wis Archeol. 1954;35(1):1–36.
- 54.
Ritzenthaler RE, Wittry WL. The Oconto site: An old copper manifestation. Madison: Wisconsin Archaeological Society; 1952. https://doi.org/10.1016/0002-9378(52)90073-2 pmid:14902972
- 55. Freeman JE. Price site III, RI-4, a burial ground in Richland County, Wisconsin. Wis Archeol 1966;47:33–75.
- 56. Ritzenthaler RE, Quimby GI. The red ocher culture of the upper Great Lakes and adjacent areas. Fieldiana. Anthropol. 1962 Mar 27;36(11):243–75.
- 57. Beaubien PL. Cultural variation within two Woodland Mound groups of northeastern Iowa. Am Antiq. 1953 Jul;19(1):56–66.
- 58.
Esarey D. Red ochre mound building and Marion phase associations: a Fulton County, Illinois perspective. In: Farnsworth K, Emerson TE, editors. Early Woodland Archaeology. Kampsville: Center for American Archaeology Press; 1986. p. 213–43.
- 59. Moffat CR, Speth JM. Rainbow dam: Two stratified late archaic and woodland habitations in the Wisconsin river headwaters. Wis Archeol. 1999;80(2):111–59.
- 60.
Hill MA. Ottawa North and Alligator Eye: Two Late Archaic sites on the Ottawa National Forest: Ottawa National Forest Cultural Resources Management Series 6. Ironwood: U.S. Department of Agriculture Forest Service; 1994.
- 61. Hill MA. New dates for old copper: Contemporaneity in the Archaic western Great Lakes. Wis Archeol. 2011;92:85–92.
- 62. Bayliss A. Quality in Bayesian chronological models in archaeology. World Archaeol. 2015 Aug 8;47(4):677–700.
- 63. Bayliss A. Bayesian buildings: an introduction for the numerically challenged. Vernac Archit. 2007 Feb 1;38(1):75–86.
- 64. Bayliss A, Bronk Ramsey C, Van der Plicht J, Whittle A. Bradshaw and Bayes: towards a timetable for the Neolithic. Camb. Archaeol. J. 2007 Feb;17(S1):1–28.
- 65. Bronk Ramsey C. Bayesian analysis of radiocarbon dates. Radiocarbon. 2009;51(1):337–60.
- 66.
Buck CE, Cavanagh WG, Litton CD, Scott M. Bayesian approach to interpreting archaeological data. West Sussex: Wiley; 1996. 382 p.
- 67.
Whittle AW, Healy FM, Bayliss A. Gathering time: dating the early Neolithic enclosures of southern Britain and Ireland. London: Oxbow Books; 2011.
- 68. Barrier CR. Town aggregation and abandonment during the era of urban transformations in the Cahokia region: Bayesian modeling of the Washausen mound-town. J Archaeol Sci: Rep. 2017 Feb 1;11:523–35.
- 69. Lee S, Bronk Ramsey C. Development and application of the trapezoidal model for archaeological chronologies. Radiocarbon. 2012;54(1):107–22
- 70. Manning SW, Birch J, Conger MA, Dee MW, Griggs C, Hadden CS. Contact-era chronology building in Iroquoia: age estimates for Arendarhonon sites and implications for identifying Champlain’s Cahiagué. Am Antiq. 2019 Oct;84(4):684–707.
- 71. Quinn CP, Ciugudean H, Bălan G, Hodgins G. Rethinking time, culture and socioeconomic organisation in Bronze Age Transylvania. Antiquity. 2020 Feb;94(373):44–61.
- 72. Bronk Ramsey C. Dealing with outliers and offsets in radiocarbon dating. Radiocarbon. 2009;51(3):1023–45.
- 73.
Bronk Ramsey C. OxCal Program Version 4.4. Oxford: Oxford Radiocarbon Accelerator Unit, University of Oxford; 2021.
- 74. Reimer PJ, Austin WE, Bard E, Bayliss A, Blackwell PG, Bronk Ramsey C, et al. The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon. 2020 Aug;62(4):725–57.
- 75. Bayliss A. Rolling out revolution: using radiocarbon dating in archaeology. Radiocarbon. 2009;51(1):123–47.
- 76. Buck CE, Christen JA. A novel approach to selecting samples for radiocarbon dating. J Archaeol Sci. 1998 Apr 1;25(4):303–10.
- 77. Christen JA, Buck CE. Sample selection in radiocarbon dating. Journal of the Royal Statistical Society: Series C (Applied Statistics). 1998 Apr;47(4):543–57.
- 78. Holland-Lulewicz J, Ritchison BT. How Many Dates Do I Need?: Using Simulations to Determine Robust Age Estimations of Archaeological Contexts. Adv Archaeol Pract. 2021 Jul 15:1–6.
- 79. Krus AM, Cobb CR. The Mississippian fin de siècle in the middle Cumberland region of Tennessee. Am Antiq. 2018 Apr;83(2):302–19.
- 80. Bronk Ramsey C. Methods for summarizing radiocarbon datasets. Radiocarbon. 2017 Dec;59(6):1809–33.
- 81. Martin SR. 20KE20: Excavations at a prehistoric copper workshop. Mich Archaeol. 1993;39(3):4.
- 82. Pompeani DP, Steinman BA, Abbott MB, Pompeani KM, Reardon W, DePasqual S, et al. On the timing of the Old Copper Complex in North America: a comparison of radiocarbon dates from different archaeological contexts. Radiocarbon. 2021 Apr;63(2):513–31.
- 83. Pompeani DP, Abbott MB, Steinman BA, Bain DJ. Lake sediments record prehistoric lead pollution related to early copper production in North America. Environmental Science & Technology. 2013 Jun 4;47(11):5545–52. pmid:23621800
- 84. Lynn M. Scarcity effects on value: A quantitative review of the commodity theory literature. Psychol Mark. 1991 Mar;8(1):43–57.
- 85. Reardon B. Oldest carbon-14 dated copper projectile points from Wisconsin. Wis Archeol. 2014;95(1):86–7.
- 86. Clermont N, Chapdelaine C, Cinq-Mars J. L’île aux Allumettes: L’Archaïque supérieur dans l’Outaouais. Collection Paléo-Québec. 2003.
- 87.
Pilon JL, Young J. Ottawa Valley burial patterns spanning six millennia. In: Keenlyside DL, Pilon J, editors. Painting the Past with a Broad Brush: Papers in Honour of James Valliere Wright. Gatineau: Mercury Series Archaeology Paper 170. Canadian Museum of Civilization; 2009. p. 181–211.
- 88.
Bradford ME. The Old Copper Complex Component at the Sandy Lake Dam Site. St. Cloud: St Cloud State University. 2017. 142 p.
- 89. Morris GK, Steinbring J. Radiocarbon dates for Old Copper in Rusk County, Wisconsin. Cent States Archaeol J. 2020;67(1):20–2.
- 90. Clermont N, Chapdelaine C. Île Morrison: lieu sacré et atelier de l’Archaïque dans l’Outaouais. Collection Paléo-Québec. 1998;(28).
- 91. Libby WF. Chicago radiocarbon dates V. Science. 1954 Nov 5;120(3123):733–42. pmid:17842266
- 92. Ritzenthaler RE. Another Radiocarbon Date for the Oconto Site. Wis Archeol. 1970;51:77
- 93.
Kennedy CC. Preliminary report on the Morrison’s Island-6 site. Gatineau: National Museum of Canada, Bulletin. 1967;206: p.100–125.
- 94. Crane HR, Griffin JB. University of Michigan radiocarbon dates IX. Radiocarbon. 1964;6:1–24.
- 95. Crane HR, Griffin JB. University of Michigan radiocarbon dates X. Radiocarbon. 1965;7:123–52.
- 96. Hruska R. The riverside site: a late archaic manifestation in Michigan. Wis Archeol. 1967;48(3):145–260.
- 97. Crane HR. University of Michigan radiocarbon dates I. Science. 1956; 123(3223):664–672. pmid:17800889
- 98. Crane HR, Griffin JB. University of Michigan radiocarbon dates IV. Radiocarbon. 1959;1:173–98.
- 99. Ritzenthaler RE. Some Carbon 14 Dates for the Wisconsin Old Copper Culture. Wis Archeol. 1958 (3):173–174.
- 100.
Papworth, ML. Cultural Tradition in the Lake Forest Region During the Late High Water Stages of the Post Glacial Lakes. 1967. Ph.D. dissertation, Department of Anthropology, University of Michigan, Ann Arbor. University Microfilms, Ann Arbor.
- 101. Marois R. Souvenirs d’antan: les sépultures archaïques de Coteau-du-Lac, Québec. 1987. Rech Amerindien Que XVII(1–2):7–35.
- 102.
Kennedy CC. The Upper Ottawa Valley: a glimpse of prehistory. Pembroke: Renfrew County Council; 1970.
- 103. Beukens RP, Pavlish LA, Hancock RG, Farquhar RM, Wilson GC, Julig PJ, Ross W. Radiocarbon dating of copper-preserved organics. Radiocarbon. 1992;34(3):890–7.
- 104. Rubin M, Suess HE. US Geological Survey radiocarbon dates III. Science. 1956 Mar 16;123(3194):442–8. pmid:17775408
- 105. Steventon RL, Kutzbach JE. University of Wisconsin radiocarbon dates XXIII. Radiocarbon. 1986;28(3):1206–23.