INTRODUCTION
It is becoming increasingly clear that large-scale movements of villages and regional populations were part of historical events and possibly adaptive processes that permeated sections of the Eastern Woodlands (Birch Reference Birch2012; Pluckhahn et al. Reference Pluckhahn, Wallis and Thompson2020). Closer to our study area along the Georgia Coast is the postulated abandonment of the Savannah River Valley (Anderson Reference Anderson1994) and possible subsequent movement of people to the coast (Ritchison Reference Ritchison2018b). Indeed, there are even hints that population movements also were part of the regular coastal pattern or were perhaps reactions to early colonial encounters on the Georgia Coast (Figure 1). For example, there appears to be no Altamaha period sites on Ossabaw Island (Pearson Reference Pearson2014), suggesting the possibility that it was abandoned prior to or just at the time when Spanish settlers entered the region.
Figure 1 Location of Sapelo Island along the Georgia Coast.
The issue with evaluating such evidence for population movement for the Georgia Coast is, however, three-fold. First, the ceramic chronologies as they are currently constructed do not provide the temporal resolution necessary to evaluate settlements and abandonments for this era. Second, with regards to radiocarbon (14C) dates, this area of the calibration curve tends to fall on a plateau which has a tendency to produce multiple intercepts (see Thompson and Krus Reference Thompson and Krus2018). Finally, mollusks have historically been the material of choice for archaeologists attempting to date such events. Mollusks, however, due to their need for marine reservoir corrections, have been found to produce less precise dates (Hadden and Cherkinsky Reference Hadden and Cherkinsky2017; Thompson and Krus Reference Thompson and Krus2018; Krus and Thompson Reference Krus and Thompson2019) and be susceptible to the possible uptake of older carbon (Cherkinsky et al. Reference Cherkinsky, Pluckhahn and Thompson2014).
In order to circumvent some of the pitfalls identified above we apply a Bayesian analysis of 14C dates from the late pre-Mission (AD 1300–1580) and Mission period (AD 1580–1700) occupations of the Sapelo Shell Ring Complex (9MC23). Our purpose here is to provide a finer grain temporal resolution regarding both the Mission period occupation and exactly when Native Americans lived in the same locale prior to the establishment of the Spanish Mission system. Understanding these events has implications for the 15th and 17th centuries and is the first step in a broader understanding of the nature of colonial entanglements in the region.
NATIVE AMERICANS AND SPANISH MISSIONARIES AND EXPLORERS ON SAPELO ISLAND AND THE GEORGIA COAST
Late period Native American archaeological sites occupied in the few centuries just before European contact on the northern Georgia coast are identified primarily by the presence of Irene and Savannah ceramic types. The literature on these ceramic series place the Savannah period between AD 1150–1275 (800–675 BP) and the Irene period AD 1275–1580 (650–370 BP) (DePratter Reference DePratter1991; Thomas Reference Thomas and Thomas2008b). More recently Ritchison (Reference Ritchison2018a) produced a Bayesian analysis of the entire coastal sequence using the 14C database developed by Turck and Thompson (Reference Turck and Thompson2014) for coastal Georgia. In his modeling of dates associated with the Savannah and Irene periods he found that his models suggested an earlier start date for the Savannah period and a much later end date for their production (Ritchison Reference Ritchison2018a: 7). His model for Irene ceramics generally agrees with the previous start dates for this series; however, the model indicates that people ceased producing these ceramics earlier than the standard chronologies suggest (Ritchison Reference Ritchison2018a). As Ritchison (Reference Ritchison2018a) points out, his research shows not so much that his new dates for these ceramic chronologies need to be accepted, but rather there are questions regarding the chronologies of these ceramics that need to be addressed by better sample contexts and types (e.g., shorter-lived botanical species). Understanding such chronologies has the potential to inform us about the processes of missionization and the social landscape at the time of European contact, as discussed in more detail below.
In the 16th century when Spanish explorers and missionaries first arrived along the shores of the Atlantic, the Guale of the northern Georgia Coast were organized into ranked political units with inherited leadership and status (Thomas Reference Thomas and Thomas2008c). As Thomas (Reference Thomas and Thomas2008b) argues, inherited leadership may have emerged by around AD 800, which is early for the region in general. By the 16th century, however, this system, at least as practiced among the Guale, may have been less individualizing and more corporate and collective in nature. This is possibly indicated by the changing focus and form of public architecture at the Irene site during this time (Anderson Reference Anderson1994; Saunders Reference Saunders2000; Thompson Reference Thompson2009; Ritchison Reference Ritchison2018b). Despite this shift to a more collective form of leadership, the Guale continued to have formal inherited status roles well into the era of European colonization (Worth Reference Worth2007; Thompson and Worth Reference Thompson and Worth2011).
David Hurst Thomas (Reference Thomas and Thomas2008a) has identified one of the central issues concerning the Guale at the time of European colonization. Dubbed the “Guale Problem,” he lays out some of the contradictions concerning the nature of mobility among Native peoples of the region. Specifically, there are inconsistencies in the historic documents relating to the Mission period and the nature of mobility that was practiced among Guale groups (see also Jones Reference Jones, Thomas, Jones, Durham and Larsen1978). This also relates to Guale subsistence and their reliance on maize agriculture. In short, early descriptions by 16th century Jesuit missionaries in the region describe the Guale as being highly mobile, moving from place to place in search of resources. Subsequent descriptions by 17th century Franciscan missionaries describe the opposite. These accounts suggest the presence of large aggregate villages that engaged in maize agriculture.
There is now a growing corpus of archaeological studies that address the Guale problem. These studies include zooarchaeological, bioarcheological, and isotopic research (Keene Reference Keene2004; Andrus and Crowe Reference Andrus, Crowe and Thomas2008; Thompson and Andrus Reference Thompson and Andrus2013), all of which point to year-round occupation of Savannah and Irene period sites on a variety of site sizes and potentially different types. In fact, studies of the earlier Late Archaic sites suggest that Native inhabitants of the Georgia coast were living in year-round villages over 4000 years ago (Colaninno Reference Colaninno2010; Thompson and Worth Reference Thompson and Worth2011; Thompson and Moore Reference Thompson, Moore, Morehart and Lucia2015; Sanger et al. Reference Sanger, Quitmyer, Colaninno, Cannarozzi and Ruhl2019). One possible explanation for the disconnect among the historic sources is that the Guale were experiencing severe drought conditions in the late 1500s, as documented and discussed by Blanton and Thomas (Reference Blanton, Thomas and Thomas2008), thus engendering a higher than normal degree of mobility.
One other possible hypothesis that would explain the observed degree of mobility of the Guale in the late 1500s is that they were in the middle of larger-scale settlement migrations at this time. As noted earlier, Ossabaw Island, located just two islands north of Sapelo Island, appears to be abandoned prior to or during the Mission period, as there are no Altamaha sites currently recorded on the island (Pearson Reference Pearson2014). In addition, Spanish accounts for Sapelo Island indicate that it too was largely devoid of Native American settlement by 1595 (Worth Reference Worth2007; Jefferies and Moore Reference Jefferies, Moore, Thompson and Thomas2013). We do know that there was substantial population movement and aggregation during the Mission period where mission towns moved wholesale around the coast, shifting their location to different islands and annexing mainland populations at times (Worth Reference Worth2007; Thomas Reference Thomas and Thomas2008a; Jefferies and Moore Reference Jefferies, Moore, Thompson and Thomas2013). These movements were due, in part, to largely external factors, such as attack from pirates, as well as other Native groups intent on capturing slaves (Worth Reference Worth2007; Jefferies and Moore Reference Jefferies, Moore, Thompson and Thomas2013). However, it is possible that these larger-scale movements were part of a much older institution of population migration among islands as a way of dealing with scalar and resource stressors. At this point this is merely a hypothesis and the first step to see if it is possible to identify settlement and island-wide abandonments that might indicate this historical pattern.
The Sapelo Shell Ring Complex
As a way to explore some of the larger issues regarding the nature of abandonment of islands, settlement continuity of Mission period sites, and when these were occupied by Spanish missionaries, we have begun a large-scale 14C dating project of the Sapelo Shell Ring Complex on Sapelo Island. While mainly known for its Late Archaic shell rings, the complex also has overlapping Savannah, Irene, and Spanish Mission period settlements. For the past several years, Jefferies and Moore (Reference Jefferies, Moore, Thompson and Thomas2013, Reference Jefferies, Moore and Burns2018) have excavated a host of Mission period and earlier contexts at the site. The Mission period for the Georgia coast begins in AD 1568 and continues until around AD 1684 (Worth Reference Worth2007).
Large-scale surveys of the site by Jefferies and Moore have documented a number of large shell midden piles that contain Savannah, Irene, and Altamaha series ceramics, suggesting that these midden piles date to the few centuries leading up to the Mission period, as well as some being contemporaneous with the mission on Sapelo. Exactly when, however, prior to the mission these earlier piles date is uncertain (e.g., 14th–16th centuries).
Jefferies and Moore suggest that the Sapelo Shell Ring Complex’s Mission period occupation is that of the 17th century mission of San Joseph de Sapala. To further refine the exact temporal components of the excavated deposits, artifacts, and features of this site, we conducted a Bayesian analysis of 10 14C dates associated with Spanish period features. While the mission itself was established in the early 1600s, the results of this research indicate that deposits sampled represent the very tail end of the Mission period along the Georgia coast, AD 1660–1684 (Thompson et al. Reference Thompson, Jefferies and Moore2019). At this point, we have a good understanding of the temporal position of these deposits; however, there is still the possibility of an earlier Spanish mission on the island, as well as a later Yamasee occupation. Part of the issue with identifying these earlier and later occupations, as well as the dates of the earlier shell midden piles at the site, is that both the European and Native ceramics found in these deposits have production periods of 100 years or more (Thompson et al. Reference Thompson, Jefferies and Moore2019); therefore, we once again turn to 14C and Bayesian analysis to refine our temporal understanding. In what follows, we present the results of nine new dates modeled with the previous 10 to both further refine the Mission period occupation and begin to understand the nature of the Native American occupation just prior to the arrival of the Spanish.
MATERIALS AND METHODS
As we outlined in our earlier study (Thompson et al. Reference Thompson, Jefferies and Moore2019), our model building for the 10 earlier and 9 new dates started with a selection of samples that we thought would return the most reliable dating of the events we were interested in evaluating (Table 1). To select the samples, we relied on three main criteria which consisted of (1) if the sample was from a short-lived botanical remain, (2) if the sample was a European botanical or animal introduction, and (3) the frequency of ceramic types and other artifacts associated with the samples.
Table 1 Radiocarbon samples and dates.
Two of the main sample types that we knew must date the early historic era or later were species that were brought over and quickly adopted when Spanish missionaries arrived on the Georgia Coast. For Sapelo these include peach trees (Prunus persica) and pigs (Sus scrofa), and fragments of both are represented in our samples. Next, we chose maize samples (Zea mays) from feature contexts that we knew had European made artifacts in them, which included objects like wrought nails. We grouped together all of these samples in the model, as will be described below.
We had two other groupings of dates that we incorporated into our sampling procedures. These included maize samples and grape seeds (Vitus spp.) that were found in direct association with Altamaha ceramics, which are the Mission period Native American ceramics of the Georgia Coast. These ceramics have a production end date in the region of around AD 1715 (Thomas Reference Thomas, Deagan and Thomas2009). The final grouping of dates in our sample are hickory nuts, maize, and one unidentified carbonized wood sample associated with Savannah and Irene series ceramics. These dates are thought to represent the pre-mission period.
Based on our knowledge of the types of samples, their contexts, and historic dates for the end production of Altamaha ceramics, and Mission period and post-mission period events, we constructed a series of models in OxCal 4.3; however, space here only permits presentation of our final model. Variations of the model, however, did not produce dramatically different results. In fact, our inclusion of new dates did not significantly alter the dates from our previous modeling of the Mission period dates from the site (see Thompson et al. Reference Thompson, Jefferies and Moore2019). These dates were grouped in a phase as samples with known associated Native American Mission period ceramics.
The structure of the model can be seen in Figure 2 and Table 2. We used a three-phase model for these dates with the Group 1 Savannah-Irene being in a sequence with the other two, Group 2 Altamaha Samples and Group 3 Spanish Samples, which are independent phases within a larger phase. In a few cases in both Group 1 and Group 3, we are able to use the sequence command to order dates by their stratigraphic order, which further constrained them to tighter ranges. All dates were calibrated and modeled using the IntCal13 curve in OxCal 4.3 (Reimer et al. Reference Reimer, Bard, Bayliss, Beck, Blackwell, Ramsey, Buck, Cheng, Edwards and Friedrich2013) and were rounded to the nearest 5-year interval. For the lone sample of unidentified carbonized wood, we used the Charcoal Outlier Model which includes an offset for “wood charcoal samples” (Bronk Ramsey Reference Bronk Ramsey2009; Hamilton and Krus Reference Hamilton and Krus2018). As with our previous work, we did not include a TPQ for the model, as establishing exactly when the site was initially settled is still a matter of ongoing research. We did, however, include a TAQ, which is critical to the modeling of the dates. In our previous study, we used a conservative date of 1750 for the end production of Altamaha ceramics, as this was the upper end of the range of the 14C dates associated with these dates. Again, archaeologists usually assign a production end date to Altamaha ceramics of 1715 as we state above. For Sapelo Island, we know from historic records that by 1733 it was used only for hunting by Native American groups, providing another possible TAQ to be used (Sullivan Reference Sullivan2001). We modeled all of the dates using each of these as TAQs. None of these produced significantly different results with most modeled dates being the same, except for a larger range for two dates with the TAQ of 1750. Therefore, we adopt in the current study the TAQ of 1733 with an error range of 15 years, as it is likely the most reflective of the actual use of the island.
Figure 2 Probability distributions for the pre-mission village and mission period context samples from the Sapelo Island Shell Ring Complex. The light gray and dark gray together represent calibrated distributions and the dark gray alone represents the posterior density estimates based on the model that incorporates at a TAQ of 1733 ± 15 based on historic documents and archaeological information.
Table 2 Modeled pre-mission and mission dates from Sapelo Shell Ring Complex.
RESULTS
The results of our modeling of the dates indicate good agreement. Both the Amodel (98) and Aoverall (93.4) for the model indicate statistical significance, exceeding the 60-threshold established for Bayesian analysis (Bronk Ramsey Reference Bronk Ramsey2009; Hamilton and Krus Reference Hamilton and Krus2018). Due to the long tails in the distribution of these dates we focus on the 68% probability range, but 95% ranges are provided as well (Table 2). All dates indicate good convergence (i.e., >95).
The modeling provides a start date for the pre-mission village at cal. AD 1265–1370 (68% probability) and cal. AD 1200–1385 (95% probability). In looking at the dates for the Group 1: Savannah- Irene, the modeled ranges all end around AD 1500. The only exception is one date whose range extends into the 16th century (UGAMS 39000), which has a range of cal. AD 1465–1515 (68% probability) and cal. AD 1450–1585 (95% probability). The boundary end estimation for these dates is cal. AD 1475–1555 (68% probability) and cal. AD 1465–1615 (95% probability).
Dates associated with Mission period deposits were modeled as a single overarching phase with two independent subphases, Group 2: Altamaha Samples and Group 3: Spanish Samples. For these dates, the model-estimated start for the Mission period occupation falls between cal. AD 1605–1655 (68% probability) and cal. AD 1550–1660 (95% probability). The end date for the Mission period falls between cal. AD 1660–1685 (68% probability) and cal. AD 1650–1705 (95% probability).
In looking at each of the modeled dates, all of them seem to fall around the later part of the 17th century (Figure 2). Visually there appears to be a gap between the Irene pre-mission village and the currently dated deposits for the Mission period on Sapelo. Using the interval command, the possible gap between the end of the pre-mission village and the start of mission period activities is around 50–150 (68% probability) or 0–160 years (95% probability).
DISCUSSION AND CONCLUSION
The result of our modeling study with new dates for the Mission and pre-mission periods for the Sapelo Shell Ring Complex provides additional insight into the nature of the human geography of the pre-mission landscape and the temporality of missionization on Sapelo Island. Our new dates and models for this site indicate that some of the observations made about the island specifically in the 16th century may be an accurate assessment of the nature of Guale occupation, and this work may possibly point to larger population movements and processes that were in operation prior to the establishment of Spanish missions on the coast.
At this point we can say that our modeling has confirmed our previous work at the Sapelo Shell Ring Complex. The newly incorporated dates from the site do not change substantively our observation that most of the Mission period deposits sampled thus far date to the later part of the Mission period on the island, ca. AD 1640–1684, which corresponds to the presence of majolica pottery as well. It is possible that one of the dates relates to an early mission component; however, it is difficult to evaluate this without additional samples from the same context. Again, our samples from the later end of the mission occupation bracket a critical time at the close of Spanish efforts on the Georgia Coast (Thompson et al. Reference Thompson, Jefferies and Moore2019). It was at the end of this timeframe that Mission San Joseph de Sapala was abandoned.
One of the interesting aspects of our new dates and modeling indicates that the Guale village, even with deposits that have Savannah period ceramics, appears to start only as early as the late 1200s or early 1300s, cal. AD 1265–1370 (68% probability). The latest modeled date for this grouping of dates at the 68% confidence interval is cal. AD 1475–1555 (68% probability). Most other modeled dates do not date later than the mid- to late 1400s. Comparing this grouping of dates to the Mission period modeled dates suggests the possibility of a gap in the occupation of the general locale sometime during the 1500s that may have lasted upwards of 50 years.
What is interesting regarding the possible gap in occupation for the Sapelo Shell Ring Complex is that an island-wide hiatus is suggested in the Spanish documents for this time frame. As we previously noted, Spanish missionaries suggest that Sapelo was abandoned sometime before 1595. A recent large 14C dating project of Kenan Field by Ritchison (Reference Ritchison2019) returned only one sample that spanned this time period in question. Further, no 14C dates in the coastal database assembled by Turck and Thompson (Reference Turck and Thompson2014) span this period for Sapelo Island. It may be that nearby St. Catherines Island was the location where Sapelo, and potentially Ossabaw, groups moved and aggregated during this time, as this island currently has 14C dates that span this timeframe (Thomas Reference Thomas and Thomas2008c; Turck and Thompson Reference Turck and Thompson2014). Alternatively, populations may have aggregated at other locations like Creighton Island, which has yielded evidence for both large Irene and Mission period occupations and is suspected to be the lost town of Mission Santa Clara de Tupiqui (Porter et al. Reference Porter, Thompson, Thompson and DePratter2016).
In looking at the larger Guale landscape just prior to contact, it appears that a large number of sites containing Irene series ceramics were established during and after the 1300s (Thomas Reference Thomas and Thomas2008c; Pearson Reference Pearson2014). During the Mission period we see a reduction in numbers of sites, in part due to the fact that Native peoples were aggregated into mission towns (Jefferies and Moore Reference Jefferies, Moore, Thompson and Thomas2013), with this being especially the case for the latter part of the Mission period due to external slave raids and other threats.
While the aggregation of Native peoples into mission towns certainly occurred, it is possible that some of the reorganization with a greater focus on aggregation and coalescence of villages occurred just prior to the Mission period. The question is why this would be so. One possible explanation is that groups aggregated due to a constriction of resources caused by possible droughts during the 1500s. This may also account for the observed or interpreted higher mobility of the Guale by the Jesuits during this time as well. Dendrochronological studies indicate several high amplitude droughts during this time (Blanton and Thomas Reference Blanton, Thomas and Thomas2008). This process of aggregation may be a way to buffer against resource depression by allowing for more collective acquisition of aquatic resources, as has been argued for southwestern Florida (Thompson et al. Reference Thompson, Marquardt, Walker, Thompson and Newsom2018).
Currently, the large-scale population shifts hinted at in the 14C record, historic documents, and distribution of sites across the landscape provide an interesting laboratory to develop research methodologies to test for abandonments and aggregation during a critical time period. Our work at Sapelo and our modeling of these pre-mission and Mission period dates point towards a process by which some of the ideas presented here can be tested. Currently, these hypothesized movements and abandonments need to be tested empirically. Large-scale dating of several sites on these islands and careful sample selection will be needed to see if these processes of abandonment and reoccupation are really at work during this time period. We now, however, have the tools to evaluate these important human scale events at the time Europeans entered the American Southeast.
ACKNOWLEDGMENTS
We extend our gratitude to Bryan Tucker and the State Historic Preservation Division of the Georgia Department of Natural Resources for their support of research on Sapelo Island. Funding for the 14C dates was made possible by a University of Indianapolis InQuery Collaborative Grant and by grants from the Georgia Department of Natural Resources. Institutional support for this project was provided by the University of Georgia’s Center for Applied Isotope Studies, the University of Georgia’s Laboratory of Archaeology, and the Department of Anthropology. We would like to extend thanks to the Historic and Cultural Preservation Department of the Muscogee (Creek) Nation, especially Raelynn Butler, for commenting on this manuscript and for allowing us to conduct research on their ancestral lands. Finally, we thank the two anonymous reviewers whose comments improved this work.
