Megadrought in the Carolinas: The Archaeology of Mississippian Collapse, Abandonment, and Coalescence

Megadrought in the Carolinas

ARCHAEOLOGY OF THE AMERICAN SOUTH: NEW DIRECTIONS AND PERSPECTIVES

Series Editor

Christopher B. Rodning

Editorial Advisory Board

Robin A. Beck

John H. Blitz

I. Randolph Daniel Jr.

Kandace R. Hollenbach

Patrick C. Livingood

Tanya M. Peres

Thomas J. Pluckhahn

Mark A. Rees

Amanda L. Regnier

Sissel Schroeder

Lynne P. Sullivan

Ian Thompson

Richard A. Weinstein

Gregory D. Wilson

Megadrought in the Carolinas

THE ARCHAEOLOGY OF MISSISSIPPIAN COLLAPSE, ABANDONMENT, AND COALESCENCE

JOHN S. CABLE

THE UNIVERSITY OF ALABAMA PRESS

TUSCALOOSA

The University of Alabama Press

Tuscaloosa, Alabama 35487-0380

uapress.ua.edu

Copyright © 2020 by the University of Alabama Press

All rights reserved.

A Dan Josselyn Memorial Publication

Inquiries about reproducing material from this work should be addressed to the University of Alabama Press.

Typeface: Caslon

Cover image: Jacques Le Moyne de Morgues (1533?–1588), Floridae Americae provinciae recens . . . (detail); Library of Congress

Cover design: Michele Myatt Quinn

Cataloging-in-Publication data is available from the Library of Congress.

ISBN: 978-0-8173-2046-1

E-ISBN: 978-0-8173-9276-5

Contents

List of Illustrations

Preface

Acknowledgments

Abbreviations

Introduction

1. The Central South Carolina Coast: At the Margins of South Appalachian Mississippian Interaction

2. A Model of Ceramic Change for the Eastern Wing of South Appalachian Mississippian

3. The Fifteenth-Century Depopulation of the Central South Carolina Coast

4. The Cultural and Natural Geography of Megadrought

5. Regions of the Greater Desert of Ocute

6. Migration to the Ring of Drought Resilience

7. Drought-Related Indigenous Disease Epidemics

8. The Broader Implications of Late Prehistoric Societal Collapse and Transformation in the Southern Latitudes of the United States during an Age of Global Warming

References

Index

Illustrations

FIGURES

I.1. Distribution of Mississippian culture areas

1.1. Locations of Central Coast Mississippian site clusters

1.2. Mississippian sherd densities, Fogarty Creek site

1.3. Comparison of farmstead site structure between the Royall and Lindsey sites

2.1. Central Coast culture chronology

2.2. Typical sherd cores of McClellanville and Jeremy series paste variants

2.3. Typical sherd cores of Pee Dee series paste variants

2.4. Central Coast Mississippian paste seriation

2.5. Central Coast Mississippian subassemblage seriation

2.6. Wateree Valley decorative mode sequence

2.7. Central Coast decorative modes I

2.8. Central Coast decorative modes II

2.9. Decorative modes from McDowell I and II phase contexts at the Mulberry site

2.10. Probability densities of accelerator mass spectrometry dates obtained from deer bone at the Mulberry site

3.1. Proxy population curve for the Central Coast Mississippian settlement

4.1. Late prehistoric abandonment regions in eastern North America

4.2. Geographic patterns of twentieth-century drought frequency for the continental United States

5.1. Greater Desert of Ocute in relation to major Mississippian sites discussed in the text

5.2. Lamar Incised sherds from the Buie mound and the Tidewater site

6.1. Fifteenth- and sixteenth-century polities making up the Ring of Drought Resilience

6.2. Bar chart of Mississippian site frequencies by phase for the Oconee Valley, Georgia

7.1. Plan map of the Mulberry site

7.2. Spring rainfall curve for central South Carolina in relation to the hypothesized pestilence at Talimeco

8.1. Generalized map of Canal System 2 in the Hohokam Core Area Phoenix, Arizona

8.2. Palmer Drought Severity Index time series curve for the Hohokam Core Area between AD 1300 and 1550

8.3. Stratigraphic sections of the geomorphic history of the middle Gila River

8.4. Heuristic demonstration of the bimodal structure of radiocarbon calibrations from AD 1200 to 1600

8.5. Palmer Drought Severity Index time series curve for the Black Warrior River valley between AD 1150 and 1350

TABLES

2.1. Radiocarbon Dates from South Carolina Central Coast Contexts

2.2. Comparison of Bayesian End Boundary Ages with Those of the Proposed Chronology

2.3. Ceramic Type Percentages for Contexts Used to Construct the Surface Treatment Seriation

2.4. Brainerd-Robinson Coefficients and Sampling Error Probabilities for the Segregated Contextual Slices

2.5. Similarity Occurrence Seriation of the Segregated Contextual Slices

2.6. Occurrence Seriations for Ceramic Decorative Modes from Central Coast Slices

2.7. Decorative Mode Percentages by Phase, Central Coast

2.8. Comparison of the Town Creek and Central Coast Decorative Mode Occurrence Seriations

2.9. Wateree Valley Decorative Mode Seriations from Excavated Contexts

2.10. Occurrence Seriation Model for the Eastern Wing of South Appalachian Mississippian

3.1. Decorative Mode Distributions for the Central Coast Regional Sample

3.2. Brainerd-Robinson Test of Similarity between the Central Coast Regional and Seriated Samples

5.1. Comparison of Decorative Mode Percentages between Wachesaw Landing and the Late Fourteenth- and Fifteenth-Century Phases from the Central Coast and the Wateree Valley

5.2. Contact Period Ceramic Type Percentages from North Carolina

5.3. Frequencies of Late Prehistoric Ceramic Types from Yauhannah Bluff

5.4. Ceramic Type Percentages and Decorative Mode Frequencies from Excavated Contexts at the Tidewater Site

5.5. Radiocarbon Dates from the Tidewater Site

7.1. Calculation of Brainerd-Robinson Similarity Coefficient Comparison of Ceramic Assemblages from the Mound Sector Shovel Test Sample and Test Unit 1 at the Mulberry Site

8.1. Twenty-Five-Year Increments for Mean PDSI Values in the Hohokam Core Area for the Period AD 1300 to 1499

8.2. Twenty-Five-Year Increments for Mean PDSI Values in the Hohokam Core Area for AD 1100 to 1299

8.3. Cleansed List of Radiocarbon Assays Associated with Architectural Features in Canal System 2, with the Addition of Polvorón Samples from the Queen Creek Drainage

8.4. Estimated Bayesian End Dates for Phases within the Hohokam Classic Period Chronology, Results for Runs 1 and 2

8.5. Estimated Bayesian End Dates for Phases within the Hohokam Classic Period Chronology, Results for Runs 3 and 4

Preface

This book is the result of a personal research project that I undertook with the goal of understanding what happened to cause the collapse of numerous agriculturally dependent, complex societies in the southern latitudes of the continental United States during a period of megadrought and other extreme climate events in the first half of the last millennium. The germ of the project was planted in my mind when I was 11 years old, soon after my family moved from Kansas to Scottsdale, Arizona, in 1960. One Sunday afternoon my dad loaded us up in our black Rambler sedan and drove out to visit Pueblo Grande Museum, where the sole above-ground monument of Hohokam civilization in the Phoenix metropolitan area—an adobe-walled platform mound—had been preserved and protected from urban development. At the time, the museum stood as a dusty and dilapidated remnant of the public works programs of the 1930s and 1940s. This was of little concern to me, however, as I was fascinated by both the mound, the top of which could be accessed by deck stairs, and the exposed ball court next to it that was encased in a crumbling veneer of cement. My dad told me that such ruins were at one time spread all across the valley and that the Hohokam had once thrived in the desert by virtue of networks of canals that watered vast agricultural tracts. Naturally, I asked whatever happened to these people, to which he replied that no one really knew, that they had just disappeared. Unsatisfied with his reply, I tucked this mystery away in my memory, vowing one day to solve it.

Shifting forward in time, I became an archaeologist, but my childhood enterprise remained dormant because opportunity dictates its own goals and subject matter, especially in cultural resource management. I conducted excavations at a stratified Archaic site on the Haw River in North Carolina and then returned to Phoenix to investigate and collaborate on a range of Hohokam sites, not least of which Pueblo Grande itself. It was not until I returned to the Southeast and became engaged in long-term regional research in the Francis Marion National Forest on the South Carolina coast that the mystery sprang forth again in an entirely new light. During my years of surveying the swamplands and coastal fringes of the forest, I became gradually aware that the prehistoric inhabitants of this region had also experienced societal collapse and regional abandonment. Moreover, this happened at virtually the same time as the Hohokam event, in the early to mid-fifteenth century. David Anderson had already demonstrated a contemporaneous abandonment scenario for the nearby middle and lower Savannah River valley, which led me to believe that more was afoot here than simple coincidence. I then set out to see just how big this zone of abandonment on the Southeast Coast might have been by reviewing the occupation sequences of adjacent regions. This was a difficult task, but after a long period of collections reanalysis and inter-regional correlations I was able to ascertain that a contiguous zone of roughly 34.5 million acres along the coast from the Cape Fear River in North Carolina to the Ogeechee River in Georgia had been abandoned or depopulated during this same time frame.

I reasoned that kindred events of such grand scale and synchronicity must surely have had a common root cause, and it was my hunch that it was linked to the global patterns of climate change that were manifested during what paleoclimatologists identify as the Medieval Climatic Anomaly, or MCA (AD 1000 to 1500). This was a period of warmer temperatures and extended droughts that would have exerted extreme pressures on the productivity of prehistoric farming systems and other natural resources and could well have led to societal collapse, regional abandonments, and migrations. The idea was not new, as Anderson had suggested previously that droughts were a principal factor in the collapse of the Savannah River chiefdoms; other scholars had already proposed that the prehistoric metropolis of Cahokia at St. Louis may have been the victim of extended drought. However, Cahokia had collapsed in the thirteenth century, much earlier than the events I was specifically interested in and interjecting noise into an otherwise neatly ordered synchronicity. A similar contradiction to my thesis was presented by the large-scale abandonments and settlement reorganizations that occurred on the Colorado Plateau and Mimbres regions of the southwestern United States at about the same time as Cahokia was collapsing. Undeterred, I consulted the paleoclimatic literature to see whether there might be geographic and chronological patterns of extreme drought that could explain these differences. Sure enough, there were indeed different geographic patterns of megadrought on a continental scale, but it was Greg McCabe’s study of twentieth-century drought distributions that brought the whole abandonment scenario into fine focus. From his maps, I could see that there are two geographic patterns of drought associated with global warming, of which the first is caused by hot sea surface temperatures in the Atlantic Ocean and cool sea surface temperatures in the Pacific. This pattern is most intense/frequent over Cahokia and the Colorado Plateau. The other pattern is caused by hot temperatures in both oceans and its effects are most intensely felt along the South Atlantic coast and the Ohio River valley, another region that experienced depopulation during the fifteenth century. If one more leap of faith is to be taken, the late prehistoric collapse and abandonment scenarios, then, can be explained by chronologically offset patterns of megadrought conditions, with one scenario corresponding to the twelfth and thirteenth centuries and a second occurring in the fifteenth century.

Eureka, synchronicity is re-established and explained by contrasting patterns of global warming! I left out one small detail—the fifteenth-century climate pattern presupposes a period of wet conditions for the Hohokam. It should be appreciated that the consequences of global warming are not always drought. They are sometimes manifested by other extreme climate events, and in the Hohokam case there is strong evidence that their civilization was brought down by alternating periods of drought and megafloods in the late fourteenth and early fifteenth centuries due to the teleconnected relationships of hot Pacific sea surface temperatures and the El Niño-Southern Oscillation pattern. I spend much of this book’s last chapter defending this hypothesis as originally formulated by Donald Graybill and colleagues.

It is not the charge of anthropologists and archaeologists to reconstruct climate and I hope those who know much more about the subject will not be too critical of my foray here. In our field, we are most interested in the implications of climate change patterns for explaining the context of social change and transformation. I endeavor in this book to rearrange what we know about the late prehistoric landscapes of the continent so that we can begin to formulate new questions about the demographic shifts of migration and immigration in times of great climatic upheaval. I demonstrate that such processes were in play within and between the regions examined here, but the hard work of piecing the puzzle together and understanding how the coalescence of different cultural groups acted as a catalyst for culture change remains to be done.

Acknowledgments

A project built over decades could not have been accomplished without the contributions of many people too numerous to thank for their individual efforts. Here I will acknowledge those who have had the most direct impact on the final product. First, I thank the reviewers of the various renditions of this manuscript, including Charlie Cobb, Chris Rodning, and Michael Trinkley. Their suggestions and advice were particularly instrumental in improving the presentation of the arguments and inferences made; I am especially grateful to Charlie Cobb for his important feedback throughout the process of writing and rewriting. Wendi Schnaufer of the University of Alabama Press is to be thanked for guiding me through the publication process and offering advice on how to present the subject matter. My copy editor, Irina du Quenoy, is to be thanked for expertly improving the flow of the text and ensuring a clearer presentation of the ideas and arguments offered herein. I benefitted from discussions with my longtime friend and colleague, Chester DePratter, who commented on certain aspects of the book and delivered his thoughts to me in the form of one of his signature three-ring binder notebooks, a treasure I will curate and always appreciate in my personal library. I would like to thank Steve Davis for steadfastly and earnestly responding to my many queries about late prehistory in North Carolina and granting me early access to the digital archive housed at the Research Laboratories of Archaeology. Others who made important contributions include, in no particular order, Joe Herbert, Chuck Cantley, Brett Riggs, Mark Williams, Gary Huckleberry, Gail Wagner, Chris Judge, Greg McCabe, Jeffery Clark, and Brett Hill. I would also like to acknowledge the support of the many individuals who sweated out the fieldwork with me in the swamps and remote areas of Francis Marion National Forest. In particular, the efforts of George Price, Ray Talley, Ken Pinson, Jason Smith, and Sean Taylor will always be appreciated and fondly remembered. Thanks are extended to Chris Judge and Gail Wagner as well for providing me the opportunity to work on three wondrous and magnificent sites in the Wateree Valley. A special note of gratitude is extended to my friend Bob Morgan of the United States Forest Service for his sponsorship and encouragement of my effort to understand regional abandonments in South Carolina. I am grateful to my son Patrick for contributing most of his free summer months while in college to help manage my collections and databases. Finally, I would like to thank my wife, Nita, whose loving support ensured that my project came to fruition.

Abbreviations

Introduction

We learn by rearranging what we know.

—attributed to Ludwig Wittgenstein

This book represents the culmination of a personal research quest that began nearly three decades ago with a series of cultural resource investigations into the archaeological record of the Francis Marion National Forest located just north of Charleston, South Carolina. Although the forest has never been the site of large-scale excavations, survey coverage of systematic shovel testing has been extensively and consistently implemented, yielding an unusually clear picture of prehistoric settlement and occupation in a sizeable geographic area scarcely duplicated anywhere in the United States. Throughout the years, as I became more familiar with this record, I began to suspect that the late prehistoric Mississippian occupation was abruptly terminated sometime prior to European contact. I knew that several neighboring coastal regions were also suspected to have been abandoned during this same time frame, most notably the middle and lower Savannah River valley so admirably described and documented by David Anderson (1990a, 1994, 1996). He attributed the Savannah River abandonment to a combination of the natural instability of Mississippian chiefdoms (i.e., chiefly cycling) and mid-fifteenth-century drought as reconstructed from tree-ring sequences. Adding the Mississippian occupation of the Francis Marion National Forest (henceforward to be referred to as the Central Coast region) to the mix, however, suggested to me that we were looking at an event that was macroregional in scale and whose primary cause was most likely an external factor such as prolonged drought and not the normal instability of chiefly rivalry and factionalism. Indeed, such organizational oscillations had occurred frequently over several centuries prior without leading to regional evacuations. Moreover, there was no good evidence that the Central Coast Mississippian population had ever achieved chiefdom-level organization, meaning that it would probably not have been subject to the destabilizing effects of chiefly cycling, and yet, it was included in this macroregional abandonment event.

In order to prove my thesis, however, I needed first to demonstrate rigorously that the Central Coast had indeed been abandoned at the same time as the middle and lower Savannah River valley. Ceramics provided the vehicle for this endeavor because there is not a more sensitive analytical method for assigning relative ages to occupations. When this information is combined with associated absolute dates such as radiocarbon ages and further integrated into a system of correlated ceramic sequences from adjacent regions, it is transformed into an even more powerful instrument for precisely dating archaeological deposits. Of course, this simple solution was not that easy to execute, because it required satisfying several levels of closure. First, I needed to understand which attributes of the ceramic assemblage would be most conducive to chronological analysis. This was accomplished by consulting the models of decorative change developed by Chester DePratter and Chris Judge (1986) and Tony Boudreaux (2005) for the closely related cultures of the middle Wateree and upper Pee Dee valleys. Next, I needed to ensure that the relevant attributes were consistently recorded in the survey collections that I had earlier analyzed as well as in those reported by other researchers who were conducting investigations in the Francis Marion National Forest. Finally, I needed to obtain radiocarbon dates from selected contexts that would provide a basis for assigning absolute date ranges to the ceramic phases I had constructed. Once these matters were ironed out over a period of six or seven years, I was able to demonstrate rigorously that the Central Coast was abandoned, or at least severely depopulated, between AD 1425 and 1450, coinciding with the Savannah River abandonments.

The first three chapters of this book focus on the structure of the Central Coast Mississippian occupation. Chapter 1 describes the subsistence base and settlement pattern of the Central Coast Mississippian system, which is representative of a marine adaptation dependent on estuarine resources as opposed to the intensive corn agriculture economies of the classic Mississippian systems of the interior Southeast and Midcontinent. Chapter 2 presents the steps I took to construct a cultural phase chronology for the Central Coast and to develop an interregionally correlated model of ceramic decorative change capable of undergirding a reconstruction of the occupation history of the Central Coast. Chapter 3 demonstrates statistically that the Mississippian system that had flourished in the region for some two centuries was abruptly terminated in the early fifteenth century and that its sizeable population either perished or out-migrated. Since it is my thesis that this event was forced by prolonged drought, I consider how dry climatic conditions might have adversely affected the productivity of the estuaries on which the Central Coast Mississippian populations depended.

Reasoning that similar abandonment events might have occurred at this time across other coastal regions of the South Atlantic Slope, I set out to examine the archaeological record of these adjacent areas to see what they might be able to tell us about the geographic scale of the hypothesized disaster. This endeavor was supported by the model of ceramic decorative change developed earlier and presented in chapter 2. Although the evidence is spotty and not conducive to statistical demonstration, this review of the archaeological record led me to determine that there is a high likelihood that fifteenth-century abandonment events occurred over a vast contiguous area of the South Atlantic coastline, from the Cape Fear River in North Carolina to the Ogeechee Valley in Georgia, covering roughly 140,000 km² or 34.5 million acres. The Soto entrada encountered the western boundary of this unoccupied and tangled zone as the Spaniards pushed forward between the native chiefdoms of Ocute and Cofitachequi in 1540; the entrada’s chroniclers were so impressed by its deserted condition that they called it the Desert of Ocute in reference to its deserted condition. Building on this convention, I named the larger zone the Greater Desert of Ocute (GDO).

To more fully understand how megadrought might have caused the abandonment of the GDO, I consulted available studies of paleoclimatic reconstructions for North America. The timing of the prehistoric abandonment events of the South Atlantic Slope coincides with the Medieval Climatic Anomaly (MCA), which is best described as a period of recurring multidecadal droughts (megadroughts) associated with increased global temperatures occurring between AD 1000 and 1500. Within this period, the continental distributions of megadrought shifted in accordance with the teleconnected patterns of sea surface temperatures in the Atlantic and Pacific oceans (B. Cook et al. 2014). Although these paleoclimatic distributions have not been precisely mapped, related modern studies have identified two distinct drought concentrations in the Southeast and Southwest regions of the United States (McCabe et al. 2004, 2008) that correlate with major prehistoric abandonment events. The abandonments of Chaco Canyon in New Mexico, the Four Corners region of the Colorado Plateau, Cahokia, and the middle Mississippi valley during the eleventh through thirteenth centuries correspond to a pattern of concentrated drought frequency associated with high Atlantic and low Pacific sea surface temperatures. For their part, the abandonments of the lower Tennessee and Ohio River valleys and the Greater Desert of Ocute in the fifteenth century conform to similar patterns brought about by high sea surface temperatures in both oceans. Clearly, there is more to these correlations than mere coincidence and the fact that local tree-ring sequences identify the fifteenth century on the Atlantic Slope as a period of increased drought intensity strongly supports the inference that megadrought caused the depopulation of the GDO.

Chapter 4 presents a paleoclimatic overview of MCA megadrought across the continental United States and discusses the geographic correspondence of this information with the major prehistoric abandonment events examined in this book. Chapter 5 details the state of knowledge concerning the settlement histories of the GDO. Both archaeological and ethnohistorical evidence is reviewed, including the sixteenth-century landings of the Ayllón voyages along the Southeast Coast, in which the colonists’ concerted search for human occupation resulted in only sparse evidence of the latter.

Another objective of this project was to strengthen the overall argument by examining the archaeological record of the regions immediately surrounding the GDO, a composite area that I call the Ring of Drought Resilience (RDR) because it was less affected by drought due to its fortunate geographic position and more favorable geophysical structure relative to evapotranspiration rates. I reasoned that if populations were out-migrating from the GDO, then they were moving to zones already occupied, for which the RDR is a prime candidate because we know that this area teemed with native populations at the time of the Soto entrada. I further proposed that the arrival of refugees from the GDO should be evidenced by crowding, conflict, and unusually high population growth rates in the host territory. I examine this proposition in chapter 6. Although I limited my search to three regions within the RDR, in each case there was ample evidence of expanding population and increased hostilities marked by the construction of palisade fortifications. Together these factors satisfy the conditions conducive to the formation of coalescent societies as described by Stephen Kowalewski: heavy and continuing population loss, warfare, people on the move as refugees and migrants, and abandonment of large areas (2006:118). Most important to the work I have outlined here is Kowalewski’s hypothesis that the process of coalescence can result in a weakening of hierarchical authority structures in societies such as chiefdoms and a transformation of the prevailing social structure into one that is more communal or collective in nature. I return to a consideration of this hypothesis in the last chapter.

Chapter 7 explores the possible impacts of indigenous drought-related disease epidemics that might have afflicted Mississippian systems prior to and during initial European contact. I focus this discussion on the ethnohistoric and archaeological details related to the pestilence that purportedly beset the great town of Talimeco just prior to the arrival of the Soto entrada in the chiefdom of Cofitachequi on the Wateree River. The origin of the disease, if indeed the Soto accounts describe an actual event, has always been an issue of great controversy. I show that the criteria for an abrupt abandonment of a great town in the 1530s appear to be met if the Mulberry site is assumed to be the archaeological manifestation of Talimeco. This argument runs counter to the prevalent view that Mulberry in fact represents the town of Cofitachequi where Soto met the casica of the chiefdom, who in turn directed him to the abandoned town of Talimeco (cf. DePratter 1989). My reasons for contesting this interpretation are discussed at length, based on a review of both the ethnohistoric accounts and archaeology of the Mulberry site. Speculating as to what kind of disease outbreak might have caused the pestilence, I draw on the work of Acuña-Soto and colleagues (2002), who have argued that the cocoliztli epidemics that raged during the sixteenth century in central Mexico were caused by an indigenous form of hemorrhagic fever. Contagions of this type are transmitted to concentrated human populations by rodent reservoirs that expand during wet spells that succeed prolonged droughts. A wet spell linked to a prior prolonged drought occurred in central South Carolina in the 1520s and 1530s, a climatic episode that could well have exposed Talimeco to an outbreak fueled by a growing rodent reservoir centered in the agricultural fields and town environs. Certainly, these coincidences by no means demonstrate on their own that an indigenous epidemic caused the abandonment of Talimeco, but if this was a chronic condition of dense human population aggregations under drought cycles, then such diseases might have been endemic to Mississippian towns with previously long, continuous occupation histories like Talimeco.

The eighth and final chapter provides an opportunity to, in Glenn Davis Stone’s (2001:164) turn of phrase, square the chickens. That is, to step back, hit the pause button, and ask what parts of the narrative as presented do not fit within the nice neat box I have constructed to contain it and what additional factors might require consideration if we are to arrive at a more satisfactory explanation. As I see it, four topics emerge that need further elaboration. First, I would like to know more about why some regions in North America seem not to conform to the expectations for abandonment based on the PDO/AMO model of chrono-geographic drought intensity that I developed in chapter 4. I further consider this issue in relation to the Hohokam Core Area of southern Arizona, where a fifteenth-century collapse and abandonment scenario occured under wet conditions. Second, my inference that Talimeco was abandoned because of an indigenous disease outbreak would seem to imply that this might have been an endemic problem for Mississippian towns with highly concentrated and large populations, especially in an age of megadroughts. This led me to search for correspondences between dry-wet periods and other seemingly abrupt town abandonments. I show that the abandonment of Moundville in Alabama at the turn of the fourteenth century is a good candidate for this type of event scenario and suggest some ways we might go about confirming the occurrence of viral diseases, which rarely manifest their symptoms on human bone because of the short interval between contraction and death.

The third topic concerns Kowalewski’s (2006) hypothesis, which is only tangentially addressed in earlier chapters but demands further development. He expects that coalescence should result in breakdowns in centralized authority structures as clans reassert their corporate power. In the Southeast, his model has been used primarily to explain the processes that led to the collapse and transformation of Native American cultures in the context of the depredations of European contact in the seventeenth and eighteenth centuries (Beck 2013; Ethridge 2009; Ethridge and Hudson 2002). What I am suggesting, along with some others (for example, Regnier [2006, 2009, 2014] and Kowalewski himself), however, is that disruptions of similar magnitude were already underway in the Southeast prior to European contact, in the fifteenth century, and that these changes most likely involved similar circumstances of coalescence. In this case, however, I argue that the external driver was megadrought. This interpretation extends the chronological depth of the fracture zone (Ethridge 2009) well into the prehistoric era and provides a more informed basis for understanding the structure and condition of Native American societies at the time of first European contact. As a starting point, I examine what evidence there might be to support the hypothesis that the hierarchical authority structure of late prehistoric and protohistoric chiefdoms was weakening.

Fourth, I consider how we might go about structuring the study of coalescent contact and assimilation on the South Atlantic Slope. I begin by examining some archaeological instances of coalescence and discussing the stages of assimilation that such culture contact situations produce. In doing so, I rely on the excellent archaeological models of migrant-host relations generated from years of investigation by other scholars in regions of the southwestern United States, along with Amanda L. Regnier’s book (2014) on the remarkable coalescence of three distinctively different cultures in the Alabama River valley in the fifteenth and sixteenth centuries. Based on these findings, I provide an initial strategy for studying these processes in places where material culture differences are much more subdued, such as sites related to the Mississippian culture of the South Atlantic Slope.

Before closing this introduction, it will be beneficial to describe the larger cultural context in which this story plays out. The Mississippian settlement of the Central Coast, the other groups of the GDO, and many of the groups within the RDR are part of a culture area referred to as South Appalachian Mississippian (SAM). In his original formulation, W. H. Holmes (1903:130) defined what would later be called SAM as a pottery tradition of distinctive stamped designs found in a province covering the states of Georgia and South Carolina and contiguous portions of Alabama, Florida, North Carolina, and Tennessee. Holmes was referring more specifically to the archaeological groups that inhabited the South Appalachian Province after about AD 1000, whose members typically stamped the exteriors of many of their pots with unique geometric motifs and built temple mounds in their main towns. Historically, this province was home to ethnographic tribes belonging to three linguistic stocks: Siouan, Iroquoian, and, most prominently in Holmes’s opinion, Muskogean, suggesting that the ancestral macroculture was to some degree multiethnic in composition. We know today that the geographic boundaries of the South Appalachian Province supported a long sequence of earlier pottery-producing cultures, some of which exhibited ceramic traits suggesting continuity with the culture Holmes described. To distinguish the later cultural pattern from these earlier associates and ancestors, Leland Ferguson (1971:11), following James B. Griffin (1967), promoted the concept of South Appalachian Mississippian. By this construction, he meant a broadly defined cultural system that shared the material attributes of complicated stamped pottery and temple mounds (and raised elite residences) but also incorporated a wide range of other traits reflecting the Mississippian adaptive complex evident throughout much of the eastern United States in late prehistory (cf. Muller 1997). These included corn agriculture, large villages, high population densities, palisade fortifications, and Southern Cult ceremonialism, or what is referred to today as the Southeastern Ceremonial Complex (see King 2007) or Mississippian Ideological Interaction Sphere (Reilly and Garber 2007:3–4).

A modified version of Griffin’s (1967) map of the larger Mississippian macrocosm shows the boundaries of the various types of Mississippian culture as they stood fifty years ago (Figure I.1). Based on the current state of knowledge, the northern boundaries of the SAM culture area should be adjusted slightly northward to include the upper Catawba valley. Mississippian culture spread eastward over a relatively time span from its origins in the American Bottom on the middle Mississippi River to its maximum geographic extent on the Atlantic coast. Anderson (1999:226) mapped this spread at the turn of the millenium. Modern radiocarbon calibrations suggest that the specific times of arrival are about 100 to 120 years later than those depicted on his map, but the rate of progress remains unchallenged. This process of spread, or Mississipianization, began at about AD 1050 in and around Cahokia, reached the Black Warrior River valley (Moundville) at around AD 1120 and emerged at Etowah some 20 to 40 years later. Adam King (2003) would position the beginnings of Mississippian in the Etowah Valley as much as 70 to 150 years earlier, but I agree with Pauketat’s (2007:115–116) argument that the radiocarbon dating King presents as evidence is too late to support his chronology. If the calendrical ages of the dates are accepted at face value, a beginning date for Mississippian at Etowah is more likely to fall around AD 1150, which is consistent with what is understood to be the timing of Mississippianization’s spread eastward. If we take an estimated start date of AD 1150 for fully Mississippianized Etowah, then the calibrated radiocarbon dates from the Central Coast to be discussed herein suggest that SAM’s march to the sea was complete by AD 1275 to 1300, some 125 to 150 years after it was initiated at Etowah.

Most SAM societies appear to have been chiefdoms of varying complexity and geographic reach (see Anderson 1990; Hally 1993; Widmer 1994). From a cultural evolutionary perspective (Sahlins 1963; Service 1962), chiefdoms are situated within the broad adaptive niche of sociopolitical evolution bridging simpler egalitarian tribes and complex archaic states such as those that emerged in Peru, Mesoamerica,