In the current world context of migration and globalization, archaeological studies of population movements, intercultural interactions, and colonialism can provide insight into similar situations in the past, contributing to a clearer understanding of their causes while potentially reshaping views of present-day outcomes of those events (Hayes and Cipolla Reference Hayes, Cipolla, Cipolla and Hayes2015; Schaepe et al. Reference Schaepe, Angelbeck, Snook and Welch2017; Whittington Reference Whittington2017; Williamson and MacDonald Reference Williamson, MacDonald, Biehl, Comer, Prescott and Soderland2015). To investigate archaeological sites that inform on these critical topics of past migrations, exchange, and interaction, firm chronologies are required. Methodological advances in the dating of archaeological materials over the past 70 years have greatly improved regional and site chronologies worldwide, resulting in more refined interpretations of past behavior (e.g., Kennett et al. Reference Kennett, Culleton, Dexter, Mensing and Thomas2014; Manning et al. Reference Manning, Birch, Conger, Dee, Griggs, Hadden, Hogg, Ramsey, Sanft, Steier and Wild2018; Sanchez et al. Reference Sanchez, Rick, Culleton, Kennett, Buckley, Erlandson and Losey2018; Waters and Stafford Reference Waters and Stafford2007; Yates et al. Reference Yates, Smith and Bertuch2015). Meanwhile, analytical techniques for identifying artifact composition to trace patterns of exchange and migration have become increasingly affordable and less invasive (e.g., Carter et al. Reference Carter, Dussubieux, Polkinghorne and Pottier2019; Hill Reference Hill2012; Reslewic and Burton Reference Reslewic, Burton and Jakes2002; Tykot Reference Tykot2016). Although the value of excavating new sites need not be stated, curated archaeological assemblages can also provide new information about past lifeways when reevaluated with updated dating techniques (Lovis Reference Lovis1990; Sanchez et al. Reference Sanchez, Rick, Culleton, Kennett, Buckley, Erlandson and Losey2018). This is particularly important when original site interpretations have never been critically reassessed and have the potential to render regional chronologies inaccurate. Our recent collection-based research pairs accelerator mass spectrometry (AMS) dating of food residues on pottery with compositional analyses of glass and metal trade items to investigate the long-term occupation history of a key Upper Great Lakes site.
The Cloudman site (20CH6) is a multicomponent habitation occupied by northeastern North American Indigenous groups for more than 1,500 years. The site yielded pottery from the Middle Woodland (200 BC–AD 500/600), early Late Woodland (AD 500/600–1000), and late Late Woodland (AD 1200–1600) periods, as well as European-made trade goods from the seventeenth century (ca. AD 1600–1700). The Cloudman site is, therefore, important for understanding diachronic developments of ancient social dynamics and regional lifeways in the Upper Great Lakes and has been a subject of exploration by a number of scholars (Branstner Reference Branstner1995; Dunham Reference Dunham2014; Kooiman Reference Kooiman2018; Walder Reference Walder2015).
The original site chronology established in the 1990s was based on relative stylistic cross-dating of pottery vessels (Branstner Reference Branstner1995). The interpretation of either a “Protohistoric” or “Contact” period occupation of the Cloudman site in the early seventeenth century was based on the stratigraphic association of Iroquoian or Iroquoian-style pottery with European-made trade goods believed to date to circa AD 1615–1630 (Branstner Reference Branstner1995:12–13; Cleland Reference Cleland and Halsey1999). Early, indirect instances of “contact” may be evidenced archaeologically through down-the-line exchange of European-made goods. The presence of such goods at a Native American site, without historical or material evidence of Europeans themselves, can be generally described as a “protohistoric” context. This term should be used with caution (Brown and Sasso Reference Brown, Sasso, Brose, Cowan and Mainfort2001:211–212), not in the least because the term may emphasize the perspective of ahistoricity of Indigenous peoples (sensu Wolf Reference Wolf1982:24–25). We recognize that an uncritical use of “Contact period” to describe the time of early colonial encounters itself still perpetuates the power imbalances of colonialism (Silliman Reference Silliman2005), but in the absence of any direct or sustained European colonial presence until the eighteenth century in this region, “contact” most accurately describes the nature of Native-European interactions between AD 1630 and 1670, when power dynamics were still favorable to local Anishinaabe groups (White Reference White1991; Witgen Reference Witgen2012). We also use contact to refer to direct rather than indirect interactions.
Our recent, independent studies of the Cloudman site artifact assemblage produced evidence that prompted critical reevaluation of the timing of the latter occupation of the site. AMS dating of carbonized food residues associated with Iroquoian pottery vessels revealed ages centuries older than the proposed early seventeenth-century occupation. These are the first direct dates on Iroquoian pottery in the Straits of Mackinac and St. Mary's River region of the Upper Great Lakes. The compositional analysis of European-made glass beads from the Cloudman site indicated that these artifacts postdated the early seventeenth century. We therefore argue that the seventeenth-century or “Contact period” component is later than originally proposed, closer to circa AD 1670 than 1630. The evidence for Late Precontact (AD 1300–1600) activities supports a long history of reoccupation of Drummond Island, Michigan, as an important resource-procurement and trading center.
Methodological and Theoretical Framework
While European colonialism has been the context of many investigations of regional and interregional interaction, exchange, and migration in North America, examining these topics in precontact contexts is likewise informative, especially when returning to reassess previous archaeological interpretations using current methods (e.g., Comstock and Cook Reference Comstock and Cook2018; Pluckhahn et al. Reference Pluckhahn, Menz, West and Wallis2018; Price et al. Reference Price, Burton and Stoltman2007; Slater et al. Reference Slater, Hedman and Emerson2014; Trabert et al. Reference Trabert, Eiselt, Hill, Ferguson and Beck2017; Tubbs Reference Tubbs2013). For example, a recent AMS dating of curated Early and Middle Woodland ceramics highlights previously unrecognized complexity in the development of Kansas City Hopewell, once thought to be an example of simple migration from the Havana-Hopewell heartland (Keehner and Adair Reference Keehner and Adair2019). Similarly, Manning and coauthors’ (Reference Manning, Birch, Conger, Dee, Griggs, Hadden, Hogg, Ramsey, Sanft, Steier and Wild2018) exploration of three sites in southern Ontario with a suite of 86 new AMS dates has shifted local chronologies later, from a “Late Prehistoric” sequence circa AD 1450–1550 into the “contact era” of the sixteenth century (ca. AD 1530–1615). This revision challenges existing assumptions about the timing of the arrival of European-made trade items in the region and the nature of Indigenous economic and sociopolitical networks of eastern North America. Examining precontact Indigenous political economies also informs our understanding of interregional differences in later European colonial strategies. Lightfoot and colleagues (Reference Lightfoot, Panich, Schneider, Gonzalez, Russell, Modzelewski, Molino and Blair2013) have shown that the size, structure, and land-management practices of Indigenous polities influenced the colonial practices of European colonizers, with implications reaching into present issues of federal tribal recognition.
The advent of AMS radiocarbon dating has improved the applicability and accuracy of absolute dating, and in recent decades it has become more cost-effective and accessible. Despite advances in the calibration of AMS dates by wiggle matching in combination with dendrochronology (Bronk Ramsey et al. Reference Bronk Ramsey, van der Plicht and Weninger2001), chronometric methods of dating sites after circa AD 1500 are limited by standard errors and issues with calibration. For the fifteenth century onward in North America, European-made trade items offer other opportunities to obtain dates through investigations of stylistic and technological change over time. These methodological advances facilitate new investigations of the variability of human responses to the intercultural interactions of the colonial era in the Americas (e.g., Allard Reference Allard2018; Forde Reference Forde2017).
We reiterate, however, that this focus has the potential to obscure the complexity of intergroup social interactions prior to the arrival of Europeans in the New World (Lightfoot Reference Lightfoot1995; Wolf Reference Wolf1982). Such is the case at the Cloudman site, originally interpreted as one of the earliest “Contact” period sites in Michigan (Cleland Reference Cleland and Halsey1999). References to the Cloudman site as the first Contact period site in Michigan have infiltrated the cultural resource management background literature for Michigan and the broader region (e.g., O'Leary and Bergman Reference O'Leary and Bergman2005:Appendix B, “Cultural Overview”).
Although the value of excavating new sites need not be stated, our work with the Cloudman collections demonstrates that curated archaeological assemblages, or “legacy collections,” can also provide new information about past lifeways when reevaluated with updated dating techniques. This is particularly important when original site interpretations have never been critically reassessed and have the potential to render regional chronologies inaccurate. In-house contract reports of Michigan State University investigations (C. Branstner Reference Branstner1992, Reference Branstner1995) and a small portion of a regional synthesis (Cleland Reference Cleland and Halsey1999) were the only published information available for the critically important but underreported Cloudman site. By “excavating the collections” (S. Terry Childs, quoted in Bawaya Reference Bawaya2007:1026) and applying analytic techniques that were too expensive, not invented, or otherwise unavailable during earlier investigations, we can reevaluate previous work using new and independent lines of evidence, such as glass bead compositions and AMS dates of food residues on pottery. Reassessment of curated artifacts from key sites used as chronological indicators can verify and improve regional understandings of the timing and nature of intercultural exchanges, both prior to and after Native-European contact.
Such work highlights the critical need for and meaningful outcomes of maintaining high curation standards in the face of shortages of space and funding (Kersel Reference Kersel2015; Voss Reference Voss2012). In cultural resource management today, “no-collection” strategies of in-field analysis are making a resurgence (American Cultural Resources Association 2019; Archaeological Collections Consortium 2019; Heilen and Altschul Reference Heilen and Altschul2013) as a cost-saving measure, despite long-standing critique of the accuracy and ethicality of no-collection methods (Beck and Jones Reference Beck and Jones1994; Butler Reference Butler1979). Our recent collections-based research reevaluates artifacts from the Cloudman site using a suite of modern analytic techniques, including AMS dating of adhered pottery residues and elemental analysis of glass trade beads and copper objects. This work highlights a complex Late Precontact occupational history of the site that was previously obscured by relative dating and stratigraphic issues as well as the importance placed on identifying “the earliest of Michigan contact sites” (Cleland Reference Cleland and Halsey1999:280).
Site Background
The Cloudman site is located on Drummond Island in Lake Huron, in the eastern Upper Peninsula of Michigan (Figure 1). The site was surveyed and excavated between 1991 and 1995 by the Michigan State University St. Mary's River Archaeological Survey, under the supervision of Charles Cleland. The original analysis of the Cloudman pottery assemblage, summarized in a technical report by Christine Branstner (Reference Branstner1995), grouped diagnostic sherds by vessel, which were then classified according to existing regional typologies. This included Middle Woodland Laurel wares and Late Woodland Juntunen sequence wares local to the Straits of Mackinac and found throughout the Upper Great Lakes region (Branstner Reference Branstner1995; Janzen Reference Janzen1968; McPherron Reference McPherron1967).

Figure 1. Cloudman site location (map by E. Beachey and H. Walder).
Pottery from the most recent component at the Cloudman site was categorized as a mix of nonlocal Iroquoian/Huron-style vessels as well as locally made “Huron imitation pots” from Ontario, which are stylistically distinct from Woodland vessels (Branstner Reference Branstner1995:12). These vessels were located primarily in the third, lowest terrace of the site, found in association with European-made trade goods, which included glass beads, gunflints, a copper knife, and an iron harpoon head (C. Branstner Reference Branstner1992, Reference Branstner1995). Cleland and Branstner concluded that these artifacts in aggregate represented a “Protohistoric” or early “Contact” period occupation dating to circa AD 1615–1630, rooted in visual identification of the trade beads according to a standard bead typology (Kidd and Kidd Reference Kidd and Kidd1970) and identification of similar associations of Iroquoian pottery and European trade items at Providence Bay (Bk-Hn-3) and Hunter's Point, two nearby sites in Ontario (Branstner Reference Branstner1995:13). They concluded that the Contact period occupants were most likely Odawa, who were mapped as living on Manitoulin and Drummond Islands circa 1630 (Trigger Reference Trigger1976:294) and known to be middleman traders between the Ojibwe to the west and Ontario Iroquois groups to the east, thus explaining the presence of nonlocal Iroquois pottery (Branstner Reference Branstner1995:13). Sites in this area containing European trade items and Iroquoian pottery are traditionally interpreted as Odawa sites (Fox Reference Fox1987, Reference Fox, Ellis and Ferris1990; Fox and Garrad Reference Fox and Garrad2004).
As part of a research initiative to understand changing Woodland resource exploitation, cuisine, and pottery technology in the northern Great Lakes, Kooiman (Reference Kooiman2018) reevaluated the Cloudman pottery vessel typologies and subjected samples to radiocarbon dating to refine the site chronology. Iroquoian pottery, identified by incised decoration techniques not used by Woodland potters, make up 18% (n = 37) of the total assemblage of identified vessels. These vessels display many characteristics of common Iroquoian pottery types, such as Huron Incised, but generally do not correspond with traditional descriptions, as is common among Iroquoian-style pottery found within traditionally “Algonquian” territories (Fox and Garrad Reference Fox and Garrad2004). Without association with well-dated established typologies, the relative ages of these vessels could not be determined.
Adhered carbonized food residue samples from five pottery vessels were submitted for AMS radiocarbon dating. Four Woodland vessels fall within the time ranges expected for their style (Kooiman Reference Kooiman2018; Kooiman et al. Reference Kooiman, Stephenson and Dunham2019). Residue from the single Iroquoian vessel submitted for dating, Vessel 162, produced a median date of cal AD 1433 (Table 1), a full two centuries earlier than the date of the proposed Contact period occupation, indicating that Iroquoian wares were in use at the Cloudman site long before the era of European contact (Kooiman Reference Kooiman2018; Kooiman et al. Reference Kooiman, Stephenson and Dunham2019).
Table 1. Accelerator Mass Spectrometry Radiocarbon Ages and Calibrated Ages of Pottery Residues from the Cloudman Site, with Relationships Determined by the Test for Contemporaneity and Averaging.

a Stuiver et al. Reference Stuiver, Reimer and Reimer2018.
b t-value = 0.08; df = 1; α = 0.05.
c t-value = 0.18; df = 1; α = 0.05.
Walder's ongoing efforts to refine regional chronologies and avenues of exchange and population movement in the early historic Upper Great Lakes region likewise sheds light on the Cloudman occupations. Her initial research (Reference Walder2015, Reference Walder2018, Reference Walder2019) has investigated these questions by identifying temporally diagnostic European glass trade bead recipes using the minimally invasive elemental analysis technique of laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) and through a stylistic analysis of copper-base metal trade items from a total of 38 archaeological sites, including Cloudman. Following up on this research, she conducted additional analysis of artifacts from Cloudman in 2017, determining that the glass bead recipes present were most similar to those from sites dated to circa AD1670 but that the use of native (not smelted) copper objects may have persisted at that time, despite the arrival of limited quantities of European-made metal materials (Walder and Perry Reference Walder and Perry2018). Together, these discoveries cast doubt on the existence of a single “Contact” period occupation of the Cloudman site.
Chronometric Dating
To further define and refine the occupational history of the Cloudman site, adhered carbonized food residues from an additional eight vessels from the later occupations of the site were collected and submitted for AMS dating to the W. M. Keck Carbon Cycle Mass Spectrometry Laboratory at the University of California, Irvine. The results corroborate expected dates from vessels associated with Middle Woodland, early Late Woodland, and middle Late Woodland occupations of the site (Kooiman Reference Kooiman2018; Kooiman et al. Reference Kooiman, Stephenson and Dunham2019) and reveal an intriguing and complex occupation history of the site post–AD 1200 (Table 1). Calibrations of AMS dates and testing for contemporaneity were accomplished with CALIB 7.1 (Stuiver et al. Reference Stuiver, Reimer and Reimer2018).
Late Late Woodland Occupation circa cal AD 1200–1300
Juntunen ware, the local late Late Woodland pottery variety, was primarily manufactured between AD 1200 and 1400 (Brashler et al. Reference Brashler, Garland, Holman, Lovis, Martin, Emerson, McElrath and Fortier2000; Brose and Hambacher Reference Brose, Hambacher and Halsey1999; Lovis Reference Lovis1973, Reference Lovis2014; McHale Milner Reference McHale Milner1998; McPherron Reference McPherron1967). A Juntunen Drag-and-Jab vessel (Vessel 102; Figure 2a) and a Juntunen Linear Punctate vessel (Vessel 205; Figure 2b) were statistically contemporaneous (see Table 1). This establishes a solid late Late Woodland occupation by Juntunen groups circa cal AD 1250.

Figure 2. Late Late Woodland vessels from the Cloudman site: (a) Juntunen Drag-and-Jab (Vessel 102, exterior surface); (b) Juntunen Linear Punctate (Vessel 205, exterior surface).
Late Precontact circa cal AD 1300–1500
Subsequent occupation of the Cloudman site seems to have been characterized by a series of distinct, short-term habitations or encampments, as demonstrated by our presentation here of the first direct dates on Iroquoian pottery in the Straits of Mackinac/St. Mary's River region. Vessel 146 (Figure 3) yielded date ranges from cal AD 1282–1322 (0.448) to cal AD 1348–1392 (0.552), with a median probability of cal AD 1354 and no statistical contemporaneity with the Juntunen vessels. Sometime in the early fourteenth century, either Iroquoian people or their trade goods made an appearance on Drummond Island. This occurred after occupation of the site by local Late Woodland Juntunen groups, with no dated contemporaneous Woodland vessels present during this habitation episode.

Figure 3. Iroquoian vessel (V146) from the Cloudman site. Left: exterior surface; right: rim profile with interior surface facing left.
Another distinct occupation is represented first by Vessel 70, which produced a median age of cal AD 1393 (Figure 4). The date range for Vessel 70 is statistically contemporaneous with that of Vessel 150, a Traverse Plain variety Scalloped vessel (Figure 5; see Table 1), a late Late Woodland ware local to the northern Lower Peninsula of Michigan (Hambacher Reference Hambacher1992). Similar scalloped wares have been described as “Algoma ware” in Ontario (Conway Reference Conway1977, Reference Conway1988), but this typology is poorly defined. Vessel 150 produced a median date of cal AD 1406, which suggests either occupation of the Cloudman site by distinct social communities during the same interval of time or the use of traded/imported Iroquoian pottery by local Woodland groups.

Figure 4. Iroquoian vessel (V70) from the Cloudman site. Left: exterior surface; right: rim profile with interior surface facing left.

Figure 5. Traverse Plain variety Scalloped vessel (V150) from the Cloudman site. Left: exterior surface; right: rim profile with interior surface facing left.
Vessel 162 (Figure 6a), as discussed above, represents yet another statistically distinct occupation by users of Iroquoian pottery circa cal AD 1433. Finally, Vessel 179 (Figure 6b) produced an AMS date of cal AD 1439–1518 (2σ). This represents the final occupation of the Cloudman site by pottery-producing/using groups for which we have solid evidence. Altogether, the early dates produced by the Iroquoian pottery vessels were unexpected based on the original site interpretation of a Contact period occupation. Early Ontario Iroquois (AD 900–1300) and Middle Ontario Iroquois (AD 1300–1400) pottery, however, is not unknown west of Georgian Bay, Ontario, and the frequency of Huron-Petun vessels (post–AD 1400) found throughout the Upper Great Lakes is even higher (Fox and Garrad Reference Fox and Garrad2004; McHale Milner Reference McHale Milner1998:212); therefore, the early occurrence of Iroquoian pottery at Cloudman is not singular.

Figure 6. Iroquoian vessels from the Cloudman site: (a) Vessel 162; (b) Vessel 179—left: exterior surfaces; right: rim profiles with interior surfaces facing left.
Some scholars have cautioned that the freshwater reservoir effect can result in a freshwater reservoir offset (FRO), which might affect radiocarbon dating of food residues obtained from pottery vessels that have cooked fish or other aquatic resources, due to the metabolic fixing into their tissue of ancient carbon from dissolved bedrock carbonates (Fischer and Heinemeier Reference Fischer and Heinemeier2003; Keaveney and Reimer Reference Keaveney and Reimer2012; Stott et al. Reference Stott, Berstan, Evershed, Hedges, Bronk Ramsey and Humm2001, Reference Stott, Berstan and Evershed2003). Kooiman (Reference Kooiman2018) subjected absorbed and adhered residues from Woodland and Ontario Iroquois pottery from the Cloudman site to stable isotope analysis. The stable isotope signatures display enriched δ15N values, with a mean of 11.69‰. Analysis of experimental cooking residues has demonstrated that δ15N values higher than 9.00‰ are associated with the processing of aquatic resources, including fish (Craig et al. Reference Craig, Forster, Andersen, Koch, Crombé, Milner, Stern, Bailey and Heron2007, Reference Craig, Saul, Lucquin, Nishida, Taché, Clarke, Thompson, Altoft, Uchiyama, Ajimoto, Gibbs, Isaksson, Heron and Jordan2013; Morton and Schwarcz Reference Morton and Schwarcz2004). However, significant offsets (>100 14C years) in experimentally produced residues generally occurred only when freshwater fish made up greater than 50% of the food mix (Hart et al. Reference Hart, Lovis, Urquhart and Reber2013). Thus, the presence of fish in pottery residues may not always result in a significant FRO.
Total alkalinity values of greater than 90 mgCaCO3/L are also needed for fish to cause significant FROs (Hart et al. 2013, Reference Hart, Taché and Lovis2018). The alkalinity of the St. Mary's River, which separates Drummond Island from the mainland Upper Peninsula (Figure 7), is only 40 mgCaCO3/L (Bratzel et al. Reference Bratzel, Thompson and Bowden1977:272), while the average alkalinities of the North Channel of Lake Huron and Central Lake Huron are approximately 56 mgCaCO3/L and 76 mgCaCO3/L, respectively (Bratzel et al. Reference Bratzel, Thompson and Bowden1977:408–409, 420–421). Furthermore, Hart and Lovis (Reference Hart and Lovis2007) compared AMS dates obtained from pottery residues with those derived from charcoal from 29 coastal and inland sites across the Great Lakes, Midwest, and Northeast and found no statistical difference between the two datasets. The AMS dates for residues from Woodland period vessels at the Cloudman site correspond with expected time ranges for their respective typologies, indicating that despite the possible presence of fish in the dated residues, local FROs are low (Lovis and Hart Reference Lovis and Hart2015). Although FRO can vary in a given body of water over time, the AMS dates for the Iroquoian pottery at the Cloudman site, although earlier than their initial interpretation, are also within the range of dates associated with similar vessels found throughout the Upper Great Lakes (McHale Milner Reference McHale Milner1998).

Figure 7. Map of Drummond Island and surrounding waterways (map by E. Beachey and H. Walder).
Ultimately, the AMS dates indicate that the occupational deposits present in the third terrace of the Cloudman site represent multiple short-term uses of the site over the course of a 200-year span, rather than a single “Contact/Protohistoric” occupation. The role of the site may have transformed from that of a significant coastal aggregation locale during the Late Woodland period to a short-term campsite along an important travel route after AD 1250.
Elemental Analysis of European-Made Trade Items
Investigation of glass and copper artifacts from Cloudman provides insight into both the timing and the activities of the seventeenth-century occupants of the site. Walder (Reference Walder2015) initially analyzed 16 blue glass trade beads using LA-ICP-MS (Table 2) and recorded formal attributes of 11 copper-base metal artifacts from feature contexts. She has also determined that it is possible to use portable X-ray fluorescence (pXRF) to qualitatively differentiate native from smelted copper with no surface cleaning of corrosion products prior to analysis (Dussubieux and Walder Reference Dussubieux and Walder2015).
Table 2. Glass Beads from the Cloudman Site Analyzed with Laser Ablation–Inductively Coupled Plasma–Mass Spectrometry.

a Walder Reference Walder2018.
b In this instance, IVa11 refers to outer clear “coat” over inner opaque white core; this layering is not readily visible without use of a microscope.
Glass Trade Beads
A total of 21 glass beads were recovered from the Cloudman site, and 18 were analyzed with LA-ICP-MS. Fifteen of the beads are type IIa55/56 in the Kidd and Kidd (Reference Kidd and Kidd1970) typology, the widely used classification system for these artifacts. Type IIa55/56 is a small, drawn, cobalt or navy blue bead type that was mass-produced for trade; another bead is a larger cobalt blue variety, type Ia57. It should be noted that these typological identifications differ from those made in the initial report on the site (Branstner Reference Branstner1995:101–103), but most differences in type are the result of different perceptions of bead color, demonstrating the unreliability of visual classifications. Chemical composition provides a more objective way of differentiating and classifying beads. The LA-ICP-MS compositional analyses of the glass beads were conducted at the Elemental Analysis Facility of the Field Museum, Chicago, Illinois.
The glass recipe of the blue beads from Cloudman is similar to that of beads recovered from other sites dated to before circa AD 1700 (see Walder Reference Walder2015, Reference Walder2018). In general, most trade beads of this period found in the Midwest are soda-lime glass, with small differences in proportions of elements from key glass ingredients such as fluxes, stabilizers, or colorants. Specifically, levels of magnesium (Mg) and phosphorus (P) were recorded (as weight percent of oxides) for the Cloudman site cobalt-colored blue glass beads in a range of MgO = 3.4% ± 0.4 and P2O5 = 0.3% ± 0.1. Similar values were identified in typologically similar beads from other occupations dated to the later part of the seventeenth century, including Rock Island Period 3a, circa AD 1670–1700 (Mason Reference Mason1986); Marquette Mission, circa AD 1670–1701 (S. Branstner Reference Branstner, Walthall and Emerson1992); and Zimmerman, circa AD 1650–1690 (Rohrbaugh et al. Reference Rohrbaugh, Stelle, Emerson, Walz and Penman1999). Additional analysis of white beads, which are also temporally diagnostic (Hancock et al. Reference Hancock, Aufreiter, Kenyon, Vandiver, Druzick, Merkel and Stewart1997; Sempowski et al. Reference Sempowski, Nohe, Moreau, Kenyon, Karklins, Aufreiter and Hancock2000), was needed to narrow the possible date range of glass bead production.
In 2017, Walder used LA-ICP-MS to analyze the two white beads (Figure 8, top) recovered from the Cloudman site (see Table 2) and found that they were opacified with antimony (Sb), which is a temporally diagnostic ingredient that occurs in glass beads produced after AD 1650 and more commonly after circa 1670 (Hancock et al. Reference Hancock, Aufreiter, Kenyon, Vandiver, Druzick, Merkel and Stewart1997; Sempowski et al. Reference Sempowski, Nohe, Moreau, Kenyon, Karklins, Aufreiter and Hancock2000). If the trade beads at Cloudman were deposited circa AD 1615–1630, as previously suggested, then these beads would have been opacified with tin (Sn), not antimony. One of the two white beads analyzed (CL_17) and a blue glass bead (CL_15) both were recovered from a water screening sample from the same level and quadrant of a 2 × 2 m excavation unit (Level 3, 10–15 cmbs, S20E104, SE quadrant; see Table 2). Therefore, we consider the blue and white glass beads to be part of the same glass bead assemblage, deposited contemporaneously rather than in separate events. Together, the glass bead analyses indicate dates of production circa AD 1670 to 1700 for the Cloudman bead assemblage. The full compositional data for the two Cloudman white beads are included in Table 3.

Figure 8. White glass trade beads (top; photo by H. Walder) and copper knife (bottom; photo by W. A. Lovis) from the Cloudman site (photos used with permission).
Table 3. Composition of the White Glass Beads from the Cloudman Site Analyzed with Laser Ablation–Inductively Coupled Plasma–Mass Spectrometry.

Copper-Base Metal
Differentiating native from smelted copper artifacts in protohistoric and colonial contexts is a useful method of tracing the spread of European-made materials. Native copper in North America is very pure; the presence of additional trace elements, including nickel, tin, arsenic, lead, antimony, and/or silver, indicates that the metal was smelted. The absence of these elements, with the exception of arsenic, which sometimes occurs naturally, indicates pure native copper (see Dussubieux and Walder Reference Dussubieux and Walder2015:Table 2 for a summary of published values for these trace elements in native and smelted European copper). This has been examined at many eastern North American sites, including several sites in northeastern Ontario (Hancock et al. Reference Hancock, Pavlish, Farquhar, Salloum, Fox and Wilson1991; Hancock et al. Reference Hancock, Fox, Conway and Pavlish1993, Reference Hancock, Pavlish, Fox and Latta1995). The method of analysis in these earlier studies was neutron activation analysis, which is destructive when used for analysis of metal artifacts; LA-ICP-MS is now more commonly used to make this distinction (e.g., Dussubieux et al. Reference Dussubieux, Deraisme, Frot, Stevenson, Creech and Bienvenu2008; Hawkins et al. Reference Hawkins, Petrus, Anselmi and Crawford2016). Interestingly, Hancock and colleagues (Reference Hancock, Fox, Conway and Pavlish1993) found that seven of the 10 copper-base metal artifacts analyzed from the Providence Bay site, an early seventeenth-century Odawa village on Manitoulin Island dated to circa AD 1600–1620, were native, not smelted, copper. The rest were presumably cut from European-made trade kettles, constituting a small “early contact period” assemblage that also includes glass trade beads, an iron knife, and a French iron trade ax (Conway Reference Conway1988:229–234, cited in Smith and Prevec Reference Smith and Prevec2000:77). It is notable that the dating of this Providence Bay assemblage to the “latter part of the sixteenth and early part of the seventeenth centuries” (Smith and Prevec Reference Smith and Prevec2000:77) was highly influential in assigning the original early seventeenth century date to the Cloudman site's “protohistoric/contact period component” (Branstner Reference Branstner1995:13).
The analysis of copper-base metal objects during Walder's initial research highlights additional discrepancies with the original interpretation of early seventeenth-century technology and activities at the Cloudman site. The results of the stylistic comparison and attribute analysis of the large sample (n = 3,705) of copper-base metal artifacts from across the Upper Great Lakes region were recently published, identifying regional, social, and temporal patterns in copper-working technology (Walder Reference Walder2019). The copper-base metal artifacts from the Cloudman site (n = 11) included in that study were of limited usefulness for stylistic and attribute comparison with much larger assemblages but offered potential for compositional analysis (Walder and Perry Reference Walder and Perry2018). Based on their visual appearance and association with precontact artifacts, Branstner described many of the metal artifacts from Cloudman as native, rather than smelted, copper (e.g., a rolled copper bead from a feature containing Middle Woodland ceramics). Our current findings support this interpretation. Branstner (Reference Branstner1995:97) also identified two pieces of “scrap” metal as well as a copper bar (Catalog No. 7282.1554.04.01) from possible “Contact” period contexts. The scraps (Nos. 7282.770.03.01 and 7282.989.03.06) were originally cataloged as “kettle scrap,” while the rest of the copper assemblage was assumed to be made of native copper because of working methods and thicknesses of the metal. Visually differentiating native copper and smelted copper based on working methods is unreliable at best, so nondestructive compositional analysis using pXRF was undertaken. A semiquantitative and qualitative investigation of all probable native copper artifacts from the Cloudman site using pXRF (Walder and Perry Reference Walder and Perry2018) has indicated that, indeed, none of the copper-base metal artifacts analyzed, including the “copper bar,” are made of smelted copper.
Some of the native copper artifacts come from contexts clearly attributed to the Woodland period occupations, but others come from proveniences where other European-made trade items, such as glass beads, were recovered. These include an artifact (see Figure 8, bottom) once tantalizingly described as “an exact replica of a French clasp knife” made of native copper (Cleland Reference Cleland and Halsey1999:280). This artifact is from the “Garden Collection”; that is, uncontrolled collections from the Cloudman family garden. Two blade gunflints “manufactured from locally available Fossil Hill chert” (Branstner Reference Branstner1995:101) came from that same context and were originally used as evidence to support the presence of the “Contact period” component.
In 2018, the two copper-base metal pieces interpreted as “kettle scrap” (Nos. 7282.770.03.01 and 7282.989.03.06) were also subjected to pXRF using the Michigan State University Department of Anthropology Handheld XRF Bruker Tracer III-SD instrument, which employs energy-dispersive X-ray fluorescence. This instrument uses a 4W Rh (rhodium) anode tube as a primary X-ray source. It includes a silicon drift detector with a full width height maximum of 142 eV at the Mn-K-alpha line. Both artifacts were analyzed with a 120-second assay using a foil filter (12 mil Al, 1 mil Ti; Bruker yellow filter), which allows for the detection of heavier elements and is ideal for analyzing metals. The primary X-ray source was set at 40 kV, 9.6 μA, and no vacuum was used. These were the same instrument and settings used for the earlier examination of the Cloudman native copper assemblage (Walder and Perry Reference Walder and Perry2018).
The result of this analysis method is a semiquantitative dataset that was examined using Bruker's S1PXRF (version 3.8.30) and ARTAX (version 7.4.0) software. Data are visualized as different peaks on an energy spectrum generated as the data output. Spectral peaks (representing counts per second) of both scraps were visually compared with a known native copper sample (Archaic period knife) and smelted copper sample (copper kettle lug) from the Michigan State University collections. This analysis indicates that one of the Cloudman “kettle scraps” (No. 7282.989.03.06) is in fact smelted copper, due to the presence of spectral peaks at the energies for lead (L-alpha, L-beta), arsenic (K-alpha, K-beta), and silver (K-alpha, K-beta). The other piece of “kettle brass” (No. 7282.770.03.01) has no such peaks for these trace elements introduced during the smelting process and is characterized as native copper.
Two additional irregular “copper scraps” (Nos. 7282.880.02 and 7282.989.029.03) came from units that also yielded lead shot and other pieces of indeterminate historic metal, as well as decorated and undecorated ceramic sherds and lithic artifacts. These two irregular scraps were likewise characterized using pXRF as native, not smelted, copper (Walder and Perry Reference Walder and Perry2018). These results indicate that native copper was recovered alongside European-made trade items in the seventeenth-century occupation of the site. Although Walder has previously observed that this finding shows potential for continuity of native copper working into the seventeenth-century occupation (Walder and Perry Reference Walder and Perry2018), this interpretation is not as strongly supported given the new AMS dates reported above and a reconsideration of site stratigraphy.
Site Context and Stratigraphy
The morphology of the Cloudman site deposits likely contributed to the misidentification of its most recent component as an early Contact period occupation. Archaeological materials were encountered in three distinct river/lake terraces on the site, each generally corresponding to the major site components (Branstner Reference Branstner1995). The third and lowest terrace was described as containing “a mixture of Protohistoric/Contact and Late Woodland period materials,” with most material culture yielding from the stratum below the humus, an 8–10 cm layer of black sandy loam (Branstner Reference Branstner1995:23–24). The lower terrace was excavated in 5 cm levels, and no additional stratigraphic controls were undertaken during excavation. The compaction of 700 years’ worth of occupational history into a single 10 cm natural stratum would severely complicate the distinction of different occupations from one another during excavation.
Examination of artifact contexts indicates that postdepositional mixing may indeed have affected the Cloudman site stratigraphy. Vessel 103 (No. 7282.018.05), a Mackinac Banded pot, dating to cal AD 941–997, and Vessel 102 (No. 7282.018.05), a Juntunen Drag-and-Jab vessel dating to cal AD 1216–1275, were both found in the same 5 cm level (20–25 cmbs) of Feature 18. The sandy soil and small size of certain artifacts would have likewise contributed to vertical displacement and mixing. One of the blue glass trade beads (sample CL_07; No. 7282.880.03.03), likely manufactured circa AD 1670–1700, was found in the same unit (S16E98) as—but within a 5 cm level below—Vessel 179 (No. 7282.880.03.03), an Iroquoian vessel dating to cal AD 1439–1518. The glass bead was found in the level 10–15 cmbs, while the vessel above it came from 5–10 cmbs.
The smelted copper “kettle scrap” (No. 7282.989.03.06) was recovered in Unit S23E108, Level 3, 13–15 cmbs. From this same provenience came 27 Native-made ceramic body sherds, 72 pieces of lithic debitage, seven pieces of fire-cracked rock, and one white seed bead (CL_18, described above). Given that this assemblage comes from a 2 cm partial level of a 1 × 1 m unit, it is unsurprising that the excavators considered this a “protohistoric” assemblage.
Although the co-occurrence of Iroquoian-style vessels and European-made trade goods is seen at other sites in the Upper Great Lakes (Conway Reference Conway1977; Mason Reference Mason1986; Mazrim Reference Mazrim2011; McHale Milner Reference McHale Milner1998:214; McPherron Reference McPherron1967), stratigraphic and dating issues have often impeded proper evaluation of these contexts (McHale Milner Reference McHale Milner1998:211). Throughout much of the Upper Great Lakes, the soil is rather sandy and slow to develop, often precluding clear differentiation of occupational episodes at archaeological sites. Drummond Island itself is composed of dolomitic limestone bedrock and covered with generally thin, poorly developed soils. Although alluvial sedimentation has resulted in greater soil development at the Cloudman site than elsewhere on the island, the soils are still relatively shallow (Branstner Reference Branstner1995). Podzolic soils associated with temperate woodlands are generally thin, shallow, compressed, and easily disturbed (Gordon Reference Gordon2013). Without lenses of “sterile” soil to clearly separate and differentiate occupations in stratified sites, as often observed in areas with greater soil development, the impact of soil compression is a factor that must be considered when assessing sites in these environments, which may represent a contiguous record of numerous human occupations rather than isolated occupational events (Gordon Reference Gordon2013). Therefore, site stratigraphy in this region requires careful excavation and/or application of methods such as AMS dating and compositional analysis of artifacts to properly assess site use chronologies.
Discussion
The application of a suite of advanced dating and material analysis techniques for investigating archaeological materials to the Cloudman artifact assemblage has enhanced understanding of the late occupational history of the site. AMS dating of three late Late Woodland and four Iroquoian pottery vessels and LA-ICP-MS analysis of glass trade beads confirm disparate Late Precontact and late seventeenth-century occupations of the Cloudman site by local Indigenous peoples. Because of postdepositional mixing and compression, native copper artifacts must be cautiously interpreted as most likely related to the precontact components, rather than as evidence of continued use of this technology into the “Contact” period at this site (cf. Cleland Reference Cleland and Halsey1999:280). The “replica clasp knife” is not an imitation of French technology but, instead, appears to be a native copper knife with a bent tang. Although the presence of Indigenous groups on Drummond Island at the time of the arrival of Europeans during the first half of the seventeenth century is likely, the archaeological evidence at present provides no clear, direct confirmation of an occupation at the Cloudman site during that time. Therefore, the original argument for a possible early Contact period occupation of the site dating to AD 1615–1630 is unsupported. The presence of antimony-opacified glass trade beads indicates that a reoccupation of the site by Native people with access to European-made beads and other items occurred after circa AD 1670. It is noteworthy that this coincides with the establishment of the nearby mission at Sault Ste. Marie by Father Marquette in AD 1668 (Thwaites Reference Thwaites2000:54:X [1890]), beginning the “Colonial” period in what is now upper Michigan.
The unexpected early presence of Iroquoian pottery at the Cloudman site suggests an intriguingly complex sequence of changing social relationships and/or population movements in the Upper Great Lakes prior to the arrival of Europeans. Precontact Indigenous movement at the regional scale has been a topic of recent discussion among scholars studying midwestern societies farther to the south, such as the Mississippian, Oneota, and Fort Ancient groups (e.g., Bardolph Reference Bardolph2014; Comstock Reference Comstock2017; Comstock and Cook Reference Comstock and Cook2018; Cook and Price Reference Cook and Douglas Price2015; Price et al. Reference Price, Burton and Stoltman2007; Ritterbush Reference Ritterbush2002; Ritterbush and Logan Reference Ritterbush and Logan2000; Slater et al. Reference Slater, Hedman and Emerson2014; Tubbs Reference Tubbs2013), but this phenomenon has received less attention in the Upper Great Lakes. Social interaction and trade between various Woodland groups (assumed Algonquian speakers) and Iroquoians during the precontact era are indicated in the Great Lakes, from the presence of native copper from the western Upper Peninsula of Michigan at early Iroquoian sites in Ontario and New York (Trigger Reference Trigger1985:104) to the distribution of Iroquoian pottery in parts of the eastern Upper Great Lakes with increasing frequency after AD 1000 (Brose Reference Brose, Trigger and Sturtevant1978; Dawson Reference Dawson1979; Fox and Garrad Reference Fox and Garrad2004; McHale Milner Reference McHale Milner1998:210–215; McPherron Reference McPherron1967; J. Wright Reference Wright1967).
Drummond Island is geographically located in a logical thoroughfare for trade and travel between southeastern Ontario and the Upper Peninsula of Michigan, forming the western terminus of a convenient route through Lake Huron that includes the Bruce Peninsula and Manitoulin Island (see Figure 1). Major sites such as Providence Bay on Manitoulin Island and Hunter's Point on the Bruce Peninsula have been interpreted as contact-era Odawa habitations, conveniently located at the heart of this route (Conway Reference Conway1988; Molnar Reference Molnar1997; Smith and Prevec Reference Smith and Prevec2000). Upon the arrival of the French at Manitoulin Island in AD 1615, the Odawa occupying this region were widely known as middlemen traders between the Wendat-Petun of Ontario and groups in the western Great Lakes (Feest and Feest Reference Feest, Feest, Trigger and Sturtevant1978; Smith and Prevec Reference Smith and Prevec2000; Trigger Reference Trigger1976). Occupying an ecological transition zone between the northern forests and the agriculturally productive south, the Odawa facilitated the trade of meat and furs from the west and maize from Huronia (Smith Reference Smith1996; G. Wright Reference Wright1967), in addition to the transportation of chert and pottery throughout the region (Fox Reference Fox1987; Fox and Garrad Reference Fox and Garrad2004). This system of trade was established prior to the arrival of the French and undoubtedly had roots in networks of even greater antiquity. Evidence of Iroquoian goods and/or people prior to the arrival of Europeans, therefore, should not be unexpected at the Cloudman site.
Data obtained from this research serve as a starting point for a deeper discussion of Late Precontact population movements and social interactions and the methods used to observe these phenomena at the scale of both sites and regions. Researchers studying Late Precontact/Contact period sites should use caution during evaluation of temporal components and not let the thrill of finding the potential “earliest” or “oldest” of any kind of site or artifact obscure the data. Likewise, although the subject of early interactions between Europeans and Indigenous North Americans is a productive line of anthropological inquiry, emphasis on this time period should not come at the expense of understanding the full complexities and dynamism of Late Precontact trade and social relationships, especially when investigating multicomponent sites. As our case study demonstrates, modern methods and techniques applied to curated collections can help clarify old archaeological problems and highlight the nuances of ancient Indigenous lifeways.
Conclusion
The application of a suite of modern analytic methods (including AMS radiocarbon dating, LA-ICP-MS, and pXRF) to the Cloudman site artifact assemblage established a new Late Precontact and seventeenth-century occupational history of the site. It revealed a series of occupational episodes of the site post–AD 1300, with evidence of new population movements or new trade relationships and a reoccupation of the site by Indigenous groups circa AD 1670–1700, several decades after the arrival of Europeans and European trade goods into the region. Site formation processes obscured the distinction of these episodes, and assessment of the Cloudman artifact assemblage using relative dating of ceramic vessels and visual categorization of glass trade beads and copper objects provided inaccurate interpretations of the site chronology. The results of our study demonstrate how reexamination of extant archaeological assemblages using updated techniques is often necessary for resolving old archaeological problems and discovering new topics for future exploration.
Acknowledgments
Accelerator mass spectrometry dating was conducted by the W. M. Keck Carbon Cycle Accelerator Mass Spectrometry Laboratory, University of California, Irvine. Funding for accelerator mass spectrometry dating was provided by the Michigan State University (MSU) College of Social Science and MSU Alumni and Friends of Archaeology Fund. Kooiman's dissertation research was funded by the National Science Foundation, MSU College of Social Science, MSU Alumni and Friends of Archaeology Fund, and MSU Graduate School University Distinguished Fellowship. We are grateful to the archaeologists who excavated and reported their findings at the Cloudman site, especially project director Christine Stephenson (formerly Branstner), under the supervision of Dr. Charles Cleland. Original analysis of the Cloudman pottery assemblage, including vessel groupings and typological identifications, was conducted by Stephenson.
Funding for the glass bead analysis in 2017 (Walder) was supported with a Nell Mondy Fellowship and Vessa Notchev Fellowship from Graduate Women in Science. The analysis of all glass beads took place at the Elemental Analysis Facility of the Field Museum, Chicago, under the supervision of Laure Dussubieux. The copper-base metal investigation using portable X-ray fluorescence was undertaken as part of Walder's teaching of an undergraduate Lab Methods course at MSU; both Dussubieux and William A. Lovis provided necessary technical and logistical support for that course. Two additional “kettle scrap” samples described in this article were analyzed by Frank Raslich (MSU), and we are grateful for his assistance. Walder's doctoral research was funded in part by a National Science Foundation dissertation improvement grant.
We both thank William A. Lovis for assistance with this project and for his comments on the manuscript and John Hart for his contributed expertise on freshwater reservoir offsets. Additionally, we acknowledge the MSU Museum for making the Cloudman site assemblage available for analysis and the MSU Department of Anthropology for providing both equipment and facilities. Nancy Paniche May provided the Spanish abstract, Erin Beachey drafted Figures 1 and 7, and Jack Biggs assisted with pottery photography. We would like to thank Gary and Barbara Cloudman for accommodating various investigations of the Cloudman site and for their ongoing support of research initiatives concerning the Cloudman artifact assemblage. Finally, we thank Michael Hambacher and three anonymous reviewers for their comments and suggestions, which resulted in a stronger manuscript.
Data Availability Statement
The data used for this article are housed in MATRIX—The Center for Digital Humanities and Social Sciences at Michigan State University. They are also available from the corresponding author upon request. Access to the Cloudman site archaeological material must be requested through the curator of anthropology at the Michigan State University Museum.