NORTHWESTERN UNIVERSITY

Monumentality During the Mid-Holocene in the Upper and Middle St. Johns River Basins,

A DISSERTATION

SUBMITTED TO THE GRADUATE SCHOOL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

for the degree

DOCTOR OF PHILOSOPHY

Field of Anthropology

By

Virgil Roy Beasley III

EVANSTON, ILLINOIS

December 2008 2

© Copyright by Virgil Roy Beasley III 2008 All Rights Reserved 3

ABSTRACT

Monumentality During the Mid-Holocene in the Upper and Middle St. Johns River Basins, Florida

Virgil Roy Beasley III

This dissertation reviews the contexts and conditions for the appearance of monuments in the upper and middle St. Johns basins of peninsular during the Mt. Taylor period. Beginning with an attempt to determine if mortuary monuments built of shell could be distinguished from shell middens, the of four preceramic shell sites is analyzed using Harris matrices and summarized. It is argued that particular portions of these sites represent intentional constructions. A hermeneutical approach to the study of sacred architecture is employed to bridge understanding between particular archaeological signatures and behaviors. Finally, the development of monumentality is presented as a process of continuous feedback between ecological conditions, personal decisions, and community devotion. 4 ACKNOWLEDGMENTS

I cannot express enough gratitude to all the people who have contributed to this journey. First, and most importantly, I thank my wife Marina and my son Marko. Without my family, this would not have happened. Marina has been unwavering in her support. Marko has reminded my many times that if I was finished, we could be outside playing baseball. You have made me a better man my son. To my mother and father, thank you for giving me a home. Mama, if I don’t show my love for you enough, just know how deep it really is. Daddy, you have been a real father. James Allison Brown, thank you. You are the finest mentor a student could have. Jim provided me the intellectual field where seeds could grow. Tim Earle and Bill Leonard served as the other members of my committee, and have always strived to make me a better scholar. Tim taught me so much about economy and complexity, and Bill made me explain myself without unneeded verbiage. Thank you both. To every teacher I’ve ever had, from Ms. Maxine in kindergarten to the faculties of Central Florida Community College, the University of Central Florida, the University of Alabama, and Northwestern, I thank you for your sharing and for your efforts. I learned from you all. I especially appreciate you Ian Brown. Many, many friends have contributed to the ideas in this dissertation. Rob Beck, it has been an honor to be your colleague. Doug, Sorenson, Schwartz, Rana, Andy; it was good to be your classmates. Keith Little, thanks for the coffee and smokes. You have made this a much better dissertation. Tom Lewis, it was fun running down the road with you; glad we didn’t wreck. To all my UA friends, thanks for listening to my ramblings. John Lieb and Myron Estes are true salt of the earth fellers, and have helped me so many times and in so many ways, I could never express the depth of my thanks. Michael Stallings, you said on a river bank in 1991 that you would stand by me to the end, and you have. You are the essence of a friend. I dedicate this to you. 5 Ta b l e o f Co n t e n t s

Chapter 1: Introduction 12 Organization of the Study 16 Future Research 17 Chapter 2: Theoretical Background 20 Conjectural History Perspectives 20 Culture‑Area Perspectives 23 Culture‑History Perspectives 24 Processual Perspectives 26 Contemporary Perspectives 27 Shell Heaps as Monuments 27 Hunter-Gatherer Social Complexity 29 Chapter 3: Theoretical Approach 31 Monumentality and Sacred Architecture 31 A Hermeneutical Approach to Sacred Architecture 32 Temporalities 35 The Making of 36 Discussion 41 Chapter 4: A History of Mt. Taylor Archaeology 44 Initial Formulation 44 Formalization 47 Maturation 55 Chapter 5: Mt. Taylor as a Temporal Construct 60 Mt. Taylor Period Specifics 60 Burial Modes 61 Grave Inclusions 62 Burial Strata Matrix 63 Non-local Stone 63 Shell Artifacts 64 6 Background to the Thornhill Lake Period 65 Thornhill Lake Period Specifics 68 Burial Modes 69 Grave Inclusions 70 Burial Strata Matrix 71 Non-local Stone 72 Shell Artifacts 73 Radiocarbon Chronology 74 Methodology 75 Results 76 Discussion and Conclusions 79 Chapter 6: Environment and Subsistence 81 General Characteristics 81 Geology 81 Soils 82 Sea-levels 83 Pollen 86 Temperature 88 Subsistence 89 Discussion 94 Chapter 7: Midden and Monument 96 Persistent and Transitory Surfaces 96 Chapter 8: Site Descriptions 100 Introduction 100 Methodology 101 Harris Creek (8Vo24) 102 Stratigraphy and Mortuary Deposits 106 Palmer-Taylor Mound (8Se18) 112 Setting 112 Previous Research 114 7 Moore’s 1892 Excavations 116 Excavation I 116 Excavation II 117 Excavation III 117 Excavators’ Club Excavations 118 Rollins College 1975 Excavations 120 Rollins College 1976 Excavations 123 Author Visit 130 Construction of the Palmer-Taylor Mound 130 Orange Mound (8Or1) 134 Setting 134 Previous Research 136 Moore’s 1892 Excavations 137 Excavation I 138 Excavation II 141 Excavation III 142 Excavation IV 143 Moore’s 1893 Excavations 144 Other Visits 145 Construction of the Orange Mound 146 (8VO4367) 151 Setting 151 Previous Research 152 Moore’s 1892 Excavations 155 Construction of Persimmon Mound 157 Persimmon Mound II 159 Moore’s 1892 Excavations 159 Construction of Persimmon Mound II 159 Shared Stratigraphic Attributes 160 8 Areally Extensive and Continuous Strata 160 Human Bone Entombed by Extensive Shell Strata 162 Grave Pits into Extant Strata 163 Group/Community Burial 166 Associations of Fire and Mortuary Contexts 167 Associations of Food and Mortuary Contexts 168 White Sand 169 Artifact Sparsity 170 Discussion 171 Chapter 9: A Basic Model of Midden Accumulation 173 Thompson=s Evolution of Place Model 175 Midden Matrix 179 Discussion 181 Chapter 10: Ritual-Architectural Reception Histories of the Upper St. Johns Mt. Taylor Burial Monuments 182 Review of the Stratigraphy of the Orange Mound 182 Review of the Stratigraphy of the Persimmon Mound 183 Review of the Stratigraphy of the Persimmon II Mound 184 A Ritual-Architectural History of the Palmer-Taylor Mound 184 A Ritual-Architectural Sequence 188 Harvesting 188 Heirlooming 188 Siting 190 Initiating 191 Inciting 191 Reproducing 192 Discussion 193 Chapter 11: Conclusions 194 Palmer-Taylor as Sacred Architecture 195 Discussion 196 9 Works Cited 201 Appendix A: The Creation of Nunih Waya 219 10 Li s t o f Fi g u r e s

Figure 1: Sites Discussed in Text 13 Figure 2: Ritual-Architectural Reception History of Nunih Waya 42 Figure 3: Bone Pins 63 Figure 5: Busycon Shell Tools 64 Figure 4: Archaic Stemmed Point, Buzzard’s Roost 64 Figure 6: Sketch Map of the Thornhill Lake Site 65 Figure 7: Shell Beads 70 Figure 8: Bannerstones, Thornhill Lake 71 Figure 9: Bluffton Burial Mound Section 72 Figure 10: Stone Beads 73 Figure 11: Strombus Celts 73 Figure 12: Microliths 74 Figure 13: Seriation of Radiocarbon Dates 78 Figure 14: Holocene Sea Levels 84 Figure 15: Windover Aboreal Pollen 87 Figure 16: Windover Non-aboreal Pollen 87 Figure 17: Mean July Temperature Anomaly of the Past 10,000 cal. yr. BP. 89 Figure 18: Sites on Tick Island 103 Figure 19: Harris Creek Sections 106 Figure 20: Area of Palmer-Taylor Mound 112 Figure 21: View East from Palmer-Taylor Mound 113 Figure 22: View from East towards Palmer-Taylor Mound and Suspected Spring 114 Figure 23: Excavations at Palmer-Taylor Mound 115 Figure 24: Moore Excavation I Section 116 Figure 25: Moore Excavation II Section 118 Figure 26: Excavators’ Club Section 119 Figure 27: 1975 Unit 6 West Section 121 Figure 28: 1975 Unit 2 West Section 122 11 Figure 29: 1976 Unit 4 North Section 124 Figure 30: 1976 Unit 9 North Section 125 Figure 32: 1976 Unit 3 West Section 127 Figure 31: 1976 Unit 2 North Section 127 Figure 33: 1976 Unit 5 North Section 128 Figure 34: 1976 Unit 6 North Section 128 Figure 35: Unit 7 West Section 129 Figure 36: Unit 8 North Section 129 Figure 37: Harris Matrix of Palmer-Taylor Mound 131 Figure 38: Area of Orange Mound, Persimmon Mound, and Persimmon II 134 Figure 39: Orange Mound 135 Figure 40: Location of Moore’s Excavations at Orange Mound 138 Figure 41: Moore’s Excavation I 139 Figure 42: Moore’s Excavation II 142 Figure 43: Moore’s Excavation III 143 Figure 44: Moore’s Excavation IV 144 Figure 45: Harris Matrix of Orange Mound 147 Figure 46: Sketch Map, Persimmon Mound and Persimmon II 152 Figure 47: Moore’s Excavation I, Persimmon Mound 156 Figure 48: Harris Matrix, Persimmon Mound 158 Figure 49: Moore’s Excavations, Persimmon Mound II 160 Figure 50: Burial Pit, Palmer-Taylor Mound 165 Figure 51: Time Slice Model of Resistance Profile, Ring III, Sapelo Shell Ring Complex 176 Figure 52: McGee’s Model of Mt. Taylor Site Configuration 177 Figure 53: Shell Crescents 178 Figure 54: Sassaman et al. Model of Mt. Taylor Site Structure 180 Figure 55: Mortuary Strata of Palmer-Taylor Mound 186 Figure 56: Palmer-Taylor Ritual-Architectural Reception History 189 12 Li s t o f Ta b l e s

Table 1: Chronology of the Mid-Holocene in the Upper St. Johns 61 Table 2: Selected Radiocarbon Dates from Mt. Taylor, Thornhill Lake, and Contexts 77 Table 3: Comparisons between Mt. Taylor and Thornhill Lake Periods 79 Table 4: Shared Structural Attributes of Mt. Taylor Mounds 161 13 Ch a p t e r 1: In t r o d u ct i o n

AI would like to add as a proviso here that in matters of religion, as of art, there are no >simpler' peoples, only some peoples with simpler technologies than our own. Man's >imaginative' and >emotional' life is always and everywhere rich and complex. Just how rich and complex the symbolism of tribal ritual can be, it will be part of my task to show@ (Turner 1969: 3). The Mt. Taylor period (7300‑5600 cal 14C yr BP) of the St. Johns river in peninsular Florida has played a central role in the development of shell site archaeological theory and interpretations (Figure 1). These enigmatic large heaps of shell, distinguished by the presence of human remains in some, have proved theoretical Rorschah tests for arguments of primitive and endowed. Interpretations range from basest savagery to first , from conditions of miserable abjectness to hyper‑abundance. Not unexpectedly, archaeological narratives serve the theoretical agendas of author and era. The region is characterized by extensive marshes, swamps, and slow-moving waterways that began to become entrenched during the mid-Holocene, leading to the development of extensive freshwater fisheries. These fisheries provided an economic base that was both highly productive and highly predictable and encouraged populations to become tethered to smaller and specific territories. Restricted mobility left more obvious physical traces of perpetuity on the landscape in the form of extraction and processing facilities and large scale middens, a process that further strengthened groups’ ties to bounded taskscapes (Ingold 1993). However, people interring their dead in highly visible monuments indicates a striking commitment to the history of a particular place and the expectation of continued access to the honored dead and to local resources. These monuments are the foundational sacred architecture of emergent corporate groups. As a materialization of claims, these monuments demonstrate a willingness to invest not only time, economic goods, and even ancestors into their construction, but also submission to a new type of community and a devotion to defense of those monuments and what they represent. Building a burial monument binds people to places in new and persistent ways. Mt. Taylor burial monuments are materializations of a new social contract with ancestral remains as collateral. 14

MOUNT TAYLOR MOUND

BLUFFTON MIDDEN HARRIS CREEK MOUND

HONTOON DEAD CREEK MOUND LIVE OAK MOUND OLD ENTERPRISE/GROVES’ ORANGE MIDDEN LAKE MONROE OUTLET MIDDEN BUZZARD'S ROOST PERSIMMON II

THORNHILL LAKE MOUND PERSIMMON MOUND

PALMER-TAYLOR MOUND WINDOVER POND RATTLESNAKE HUMMOCK ORANGE MOUND

INDIAN FIELDS

V B NorthR 0 150 KM

Figure 1: Sites Discussed in Text

For nineteenth century evolutionists, the shell heaps1 were clear examples of the stage of savagery and confirmation of developmental forms implicated from colonial observations. Human remains found in the heaps were read as evidence of cannibalism. Enduring themes of behavioral, mental, and material poverty were cast at this time. Early twentieth century

proponents of culture‑area found strong correspondence between the sparse material record and

1 In reference to the upper St. Johns shell-bearing sites, shell heap is used to indicate the composite site, i.e. the total midden, monument, and any other shell constructions present. It is a general category. Midden is used spe- cifically in reference to deposits demonstrably shown to be food refuse. I use midden in the original Scandinavian sense of kjoekkenmoeddings, or kitchen middens, not as a general nominal for shell-bearing sites. Monument, in regards to the upper St. Johns sites, refers to mounded shell with demonstrable mortuary strata. Mound is occa- sionally used as a synonym to monument. 15 diffusionary prediction, since Mt. Taylor sites are found far from nuclear centers of complexity and advancement. Adoption of core‑periphery principles by post‑Depression archaeologists continued to encourage an end‑of‑the‑line geographical position for the St. Johns. Coevally, preceramic shell sites of the southeast were congruent to a return of stage constructions and a greater integration of environment as an explanatory device. In the mid‑twentieth century, the St. Johns and most of the Florida peninsula was central to issues of trans‑oceanic contacts. During the later decades of the twentieth century, processual approaches emphasized local developments and environmental adaptations, and shell heaps were conceived of as ideal datasets, while ethnographic data were newly re‑embraced as a source of analogy. More recently, the attribution of intentionality during the mid‑Holocene has drawn the Mt. Taylor shell heaps into a lively debate regarding the origins and conditions of social complexity and capability. Read as intentional burial mounds, the Mt. Taylor heaps are one of the earliest dated examples of mortuary architecture in North or Central America. However, while all these models have utilized the presence of burials in shell matrices as evidence in their conclusions, particularly as indicative of complexity or the lack thereof, none have addressed the fundamental question: Can we distinguish between a shell midden and a burial monument built with shell? If so, and we can demonstrate that at least some of the Mt. Taylor shell heaps represent intentional constructions, what are the implications for understanding the emergence of monumentality and the potential of fisher-hunter-gatherer economies? Addressing these questions, the archaeological record of five aceramic shell heaps with human remains is intensively reviewed and analyzed. In particular reference to the Mt. Taylor period shell heaps, some of the archaeological criteria used to secern monument‑antimonument are that monuments: have consistent, deliberate stratigraphic structure, present a patterned ritual component, and display siting regularity. Mortuary layers are not surmounted by contemporary middens; when middens are found atop burial remains, they are significantly later in time. There is a grammar of deposition, a pattern repeated at multiple sites and Mt. Taylor burial monuments are part of a total “cultural complex,” found in concert with material consequences of all activities, including middens. 16 Throughout this dissertation, the primary assumption is that human remains interred with care are always considered sacred, worthy of veneration. Contexts that include human remains treated respectfully are expectedly sacred. By this reasoning– that the shell heaps can be shown to be intentionally constructed and that contexts with human remains shown care and attention are sacred– shell deposits that meet these criteria are conceived of as sacred architecture. With this qualification of Mt. Taylor mortuary monuments, a theoretical approach tothe study of sacred architecture forwarded by historian of religion Lindsay Jones (2000a; 2000b) is employed as a protocol of investigation. Beginning with a thesis of superabundance of meaning, the method reframes sacred architecture in a background independent way and provides a language of mapping the archaeological data to the transformative event of monument creation. Jones describes a hermeneutical and experiential approach to sacred architecture, foregrounding human experience as the tempo of investigation, with the writing of ritual‑architectural reception histories as one goal. These are reconstructions of people’s encounters with sacred architecture, informed by the materializations of these encounters. In this dissertation, the method of Jones is employed as a middle-range theory; it provides linkage between specific archaeological contexts and specific behaviors. If it can be demonstrated that Mt. Taylor period burial monuments exist, there is warrant to reconsider the interpretations of mortuary practice/non‑practice in previous models and give an alternative approach. Additionally, if a model of intentional construction is supported, we have opportunity to bring new insights into the capabilities of fisher‑hunter‑gatherers. Tempo and materiality are of central importance throughout this dissertation. Human experience as the tempo of investigation runs contradictory to general archaeological practice. Change over time, as conceptualized by many archaeologists, is not the focus of this study. Instead, tempo is treated as an attribute of material data. To apply Jones= prescription for studying sacred architecture to archaeological data requires recognition that contexts excavated sometimes represent millennium, sometimes moments. Chronology, environment, subsistence: none are called upon for explanation in a hermeneutical model, but all are necessary to bring the tempo of investigation to that of human 17 experience. There is therefore a lengthy recitation of such data as an adducer to temporality and a descriptor of the specific case.

Or g a n i z a t i o n o f t h e St u d y Chapter 2 is a history of theoretical approaches applied to Mt. Taylor archaeological remains. Discussion begins with Jeffries Wyman and extends into recent research. Emphasis is on changing interpretations of human bone found buried in shell deposits. Chapter 3 outlines my usage of Lindsay Jones’ hermeneutical approach to sacred architecture. After summarizing Jones’ program, I introduce two temporal concepts that relate to the experiences

of people apprehending architecture. A Choctaw story of mound creation is outlined, and then used to illustrate my technique of creating a ritual‑architectural reception history of a sacred building. Chapter 4 reviews the history of Mt. Taylor archaeology, culminating with the contemporary model of Mt. Taylor society. There is presented a history of research concerned with preceramic shell heaps of the St. Johns, emphasizing the upper St. Johns and beginning with how the nineteenth century excavations of Wyman and Moore were transformed and interpreted by culture‑historical, salvage, environmental, processual, and more recently, ideational archaeologies. Chapter 5 details an updated chronology of Mt. Taylor period archaeology. A reassessment of Mt. Taylor as an archaeological period, utilizing published excavations and radiocarbon dates, reintroduces a restriction of the period temporally and materially and describes a new archaeological period encompassing the later portion of the preceramic shell heaps. Once so defined, further discussion in this dissertation emphasizes the Mt. Taylor period. Chapter 6 discusses two main arenas: the general environmental context and the basic subsistence strategy of the Mt. Taylor period. Physical characteristics of the region are summarized.

Contemporary reconstructions of the mid‑Holocene conditions and local data are reviewed, concentrating on the 7300‑5600 cal 14C yr BP time frame. The state of understanding of Mt. Taylor subsistence strategies is given summary, accenting the role of fishing and how the data do not support dietary models with a primary focus on shellfish or terrestrial fauna, but instead, 18 the evidence indicates a wide spectrum strategy emphasizing aquatic resources. These data are important for contextualizing the archaeological remains. Chapter 7 is a brief introduction to the following two chapters. Contrasting monuments and middens is done by focusing on a few attributes. In addition, a distinction between surfaces based on the time involved in their deposition is made and elaborated upon. In Chapter 8, the excavation data for four Mt. Taylor burial monuments are presented. I use four main sources for these data: the published works of C. B. Moore, the unpublished field notes of C. B. Moore, the published summary of one project, and the unpublished field notes from two Rollins College field schools. There are four monuments from the upper St. Johns described and discussed; these are compared to the middle St. Johns Harris Creek site, the most completely reported Mt. Taylor burial mound. Harris matrices are used to present the stratigraphic data from the upper St. Johns sites. The majority of excavation data presented have not been previously published. Chapter 9 gives a basic model of midden accumulation at Mt. Taylor sites. This model is based in part on recent work with shell rings. It is argued that the bulk of any individual Mt. Taylor site represents accretionary midden, but at some of these sites, burial monuments are found as a component to the total site. Chapter 10 is application of the Jones’s theoretical program to the archaeology of the Mt. Taylor period shell heaps of the upper St. Johns river basin and a summary of interpretations regarding the development of monumentality during the Mt. Taylor period. Specific discussion emphasizes the Palmer‑Taylor Mound, the site for which we have the greatest amount of excavation data. The primary intent is the creation of a ritual‑architectural reception history of the site and relating of this history to emergent monumentality as a processual phenomenon. Finally a review of this study and how this study serves as a datum for further research is given.

Fu t u r e Re s e a r c h The construction of monumental architecture is one of the watershed events in the development 19 of complex society. Monumentality is indicative of a dramatic shift from the impermanent, egalitarian social constructs associated with a foraging mode of subsistence to sedentary, stratified formations indicative of permanence on the landscape. There is a radical reformation of the social landscape and the development of new social organizations that differ dramatically with those constructed and experienced by highly mobile hunter‑gatherer populations (Bradley 1993). The transformation of people’s relationship with their dead, from a personal, household level retention and interaction to the inclusion of those remains in a promiscuous community facility indicates the development of a radically new type of community, submission to a new social order, and commitment to defense of monument, village, and resources. While monuments have traditionally been associated with agriculturalists, the recognition that some foragers constructed monuments, particularly groups whose economy focused on aquatic resources, provides opportunity to expand the discourse of the conditions and social implications of large‑scale mortuary architecture. However, it has proven beyond the depth of this dissertation to fully engage these ideas. Demonstrating Mt. Taylor monumentality is the first step towards a more robust model of fisher‑hunter‑gatherer complexity and the conditions for monumentality. This dissertation has set out to demonstrate unequivocally that Mt. Taylor shell heaps with burials are monuments by conceptualizing those deposits as sacred architecture and detailing a history of ritual interaction. Concomitantly, a review of the temporal placement, environmental context, and subsistence strategies during the Mt. Taylor period shows that data are immensely informative yet limited in these arenas, and further research is required to understand the role these factors have in emergent monumentality and fisher‑hunter‑gatherer complexity. Material conditions are key for understanding the infinite variety of human experience. Immediate concerns are fleshing out details of locally specific ecological conditions, subsistence strategies, chronology, sedentism, and the siting of villages and monuments. A regionally encompassing program of acquiring core samples from sites and the extensive peats found in the upper St. Johns will provide specific data, including: radiocarbon dates, palynological specimens, floral specimens, and stratigraphic data. Isotopic data from shells may prove a way to demonstrate 20 if shell in mortuary strata is gathered from different locations. Further development of a digital paleoenvironmental model will allow for intensive interrogation of the intersection of culture and environment during the Mt. Taylor period, statistical analysis of the concurrence of resources, middens, and burial monuments, and how these findings relate to other areas locally and globally. A productive emphasis on monuments as components of built landscapes has emerged (Ashmore and Knapp 1999). The interface between people and land is often recoverable archaeologically, even in situations of mobile hunter‑gatherers. Taking advantage of the capabilities of geographical technologies and a traditional archaeological concern with spatial distributions, investigations of monuments as foci of economic and/or ideational (Knapp and Ashmore 1999) landscape studies will allow for comparative approaches and more applicability to broader processual and postprocessual theoretical discussions (Buikstra and Charles 1999; Joyce 2001). This work has demonstrated that often, the primary consideration for monument location is not economic, but instead monuments are situated in reference to and nearby natural features that have supernatural importance. The Mt. Taylor monuments are found in a variety of locations, but appear to be focused on springs, landscape features that may have been perceived as portals to the otherworld (Crumley 1999). In Florida, the use of springs and aquatic features for burial has roots in Late Paleoindian time, extending through the early and middle Holocene (Clausen et al 1979; Beriault et al 1981; Doran 2004). A main facet of future research is siting of burial monuments, including the collection of geophysical data to test a thesis of monument proximity to now-relic springs. Findings of this dissertation have highlighted the importance of critical engagement of the history of any particular idea. The archaeology of shell heaps began in a particularly racist theoretical environment and those first pictures of shell site life perpetuated myths of Victorian superiority. The overt racism is gone, but the narrative that mental and spiritual ideas had not yet evolved remains implicit in contemporary archaeological reconstructions of shell site life. Failure to consider the biases of our models perpetuates error and is contradictory to anthropological principle. Future research will continue a critique of theoretical schools of shell site anthropology and advancement of alternative approaches. 21 Ch a p t e r 2: Th e o r e t i c a l Ba c k g r o u n d

This section reviews the theoretical paradigms that have shaped our understanding of Mt. Taylor period archaeology. People of shell heaps bear many interpretive crosses, not the least of which is the undercurrent that they are living in their own filth and spilth. Partly resultant of essentialist assumptions when encountering shell in archaeological contexts, explanations that begin with denigration and assumptions of poverty are not anthropologically productive. Casting the material conditions of shell heap people as unclean obviates a possibility of sacredness and defending interpretations of simplicity by attacking the straw of personal olfactory preferences is distracting and ethnocentric. Applications of anthropological theory to the finding of human remains in the Mt. Taylor shell heaps is emphasized below. Much of the effort of earlier researchers seeks to explain the seeming incongruity of skeletal material and assumptive discard. Again, the central question of this dissertation is can we distinguish between a midden and a monument. Previous theoretical approaches have tended to presume ubiquity of the former. More recently, archaeologists are considering the possibility of the later (Sassaman 2003, 2005), but without making an explicit distinction, spirituality can be dismissed and only gross materialistic factors considered (Milner 2004; Crothers and Bernbeck 2003).

Co n j e c t u r a l Hi s t o r y Pe r sp e c t i v e s Jeffries Wyman was the first professional archaeologist to systematically excavate St. Johns sites, working in the 1860‑1875 conjectural history environment of Darwin, Huxley, Lubbock, Maine, Morgan, and Tylor (Trigger 1989; 1998). Conjectural history used observations of living and historic groups that were commonly collected unsystematically and with tremendous ethnocentric biases. Non-European societies were ranked by their shared similarities with the Victorian ideal and then collated as representations of prehistoric developmental stages. The standard model of shell midden life had been necessitated by the excavation of Danish kjoekkenmoeddings (Morlot 1861). Archaeologists needed analogs to explain the record, and 22 exploited two primary ethnographic examples: the people of Tierra del Fuego and the Andaman Islands, societies considered the lowest type of mankind in the world. The Fuegians, as called, are described as “remarkably stupid” and “disgusting” (Lubbock 1913). Darwin asserted that “These were the most abject and miserable creatures I anywhere beheld” and “Viewing such men, one can hardly make oneself believe they are fellow‑creatures and inhabitants of the same world. . .” (1839). The basics of this conjectural history model of people living on shell sites, particularly the essence of abjectness, remains today. Both the ethnographic people and the archaeological sites were representative of the Middle Status of Savagery, the next to lowest possible stage of human evolutionary development (Morgan 1877), the point just above absolute zero (Wyman 1876). Central to this paradigm is that evolution is not just material, but also mental. The most basic institution of mankind is subsistence: the earlier a form of subsistence in the scheme, the more primitive the mental and technological development of the humans, archaeologically or ethnologically. Attributes of the hypothetical Lower status include a fruit and nut centered, sedentary existence and almost no technology or material culture, i.e., pure gatherers2. A Lamarckian chain of development whereby humankind observes fire, thinks of possessing fire, and then possesses fire by evolutionary leap, leads to the ability and opportunity to exploit fish, the first animal intensively collected. Fire is one of humanity’s first good ideas, fishing allowed humanity to move from the garden and spread across the globe (Morgan 1877: 21). For late nineteenth century theorists, the Mt. Taylor human remains being found in a shell heap are an expectation of conjectural theory. Human bone must be part of the refuse from cannibalistic feasting, intermixed with other secular discard, since living populations who create shell middens are cannibals. The thesis of cannibalism has never been directly challenged by later researchers. It is being contested in this dissertation. Religion of any sort, even a simple notion of the afterlife, was said by the writers to not exist in these societies. Cannibalism is alleged to have occurred in both ethnographic examples cited,

2 There is a usually a strong Genesis tone to this developmental sequence (e.g., Wyman 1876). 23 but, as directly pointed out by several, human remains or indications of cannibalism are not found in the Danish sites (Morlot 1861: 301). Therefore, Wyman’s encountering human remains in St. Johns shell sites fulfilled the expectations of the conjecture and the mere presence of burials is considered evidence of the practice of cannibalism, (1868a, 1868b, 1868c, 1874, 1875, 1876). However, a review of the examples Wyman provides mainly describes burials with some bones missing, typically because it is an eroding or disturbed context. His evidence given for cannibalism is terribly weak (1875: 60-78). Fortunately, conjectural history provided a narrative:

AOne shudders with horror at the prolonged tortures which preceded death and the feast among these savage people. Every device cruelty could suggest was practised. . . finally his body was divided limb from limb, roasted or thrown into the seething pot, and hands and feet, arms and legs, head and trunk, were all stewed into a horrid mess, and eaten amidst yells, songs, and dances@ (Wyman 1875: 71). C.B. Moore’s excavations a few decades later, which provide a large fraction of the archaeological data we have for Mt. Taylor mortuary contexts, were interpreted by him as confirmation of the earlier theories. In the first archaeological paper published by Moore (1892a), he addresses cannibalism:

AThat the makers of many of the shell heaps of the River were cannibals is everywhere admitted since the researches of Professor Wyman; and the writer, in January 1873. . ., where they recovered burnt and broken human bone, split . . .presumably more readily to extract the marrow@ (Moore 1892a:137, emphasis added). This would have been Moore’s senior year at Harvard (Wardell 1956). Nineteen years later, Moore began his own professional career. The thesis of cannibalism remained a current throughout the seasons of excavation and reporting by Moore at sites discussed in this dissertation. Speaking of the Palmer‑Taylor mound:

“In this shell heap the hypothesis of burials, even of disconnected bones, would seem untenable, as absolutely nothing found in association with them pointed to interments. A portion of the remains of the probably cannibalistic feast can be seen at the Wagner Free Institute, Philadelphia@ (Moore 1893c: 612). Moore’s excavations into the preceramic sites that define the Mt. Taylor did produce burnt and broken human bone, in contexts with other fauna. Cannibalism is a reasonable thesis. But, the 24 only explanation considered was that the remains represented cannibalism. The heaps originated with Aa savage people hardly capable of respect for the dead@ (FN:5:181)3. Fortunately, Moore was a competent observer and very punctual disseminator, allowing for thorough dissection of his assertions. Before the beginning of the twentieth century, preceramic cannibalism had achieved the status of demonstrated. Following Moore, almost no professional excavation of Mt. Taylor sites would occur for nearly 70 years.

Cu l t u r e ‑Ar e a Pe r sp e c t i v e s Reaction to nineteenth century evolutionary schemes took form most fully as culture‑area theorizing of ethnological and archaeological data (Trigger 1998). Again, the question of Mt. Taylor monumentality had direct implications for some of the central interpretations forwarded by scholars and the sites were used in development of culture-area ideas (Wissler 1917, 1926). Most fundamental to the culture‑area approach is the thesis that developments and advancements (traits) originate where found in greatest density. The Law of Diffusion (Wissler 1926: 182‑187) instead of unilinear evolution explained differences in social complexity. The converse is that development is latest on the frontiers; the “Principle of Peripheral Retardation,” an idea that appears rooted in Klemm’s Kulturvolker, culturally creative people, and Naturvolker, culturally passive people. “Rather, we must conceive of a line or an axis along which societies and cultures, ..., can be ranged from the one extreme or pole of greatest folk‑like or tribal backwardness to the opposite pole of greatest sophistication@ (Kroeber 1948: 281). All relations of power and dominance are top‑down, as are all facets of ideation and materialization; nothing originates at the periphery (Wissler 1917, 1926; Kroeber 1919, 1939, 1948). This model is applied to explain, in its entirety, native development and experience in the Americas. "New World culture is thus a kind of pyramid whose base is as broad as the two Americas and whose apex rests over Middle America@ (Wissler

3 The field notebooks of C.B. Moore are referred to throughout by FN#, corresponding to the micro- film roll number originally assigned by Davis (1987) and how they are still indexed at Cornell Univer- sity. 25 1917: 361). As the culture‑core theoretical school matured, ecology and subsistence were given greater explanatory weight (Kroeber 1939). "We have found the answer or given an explanation as to why it is that human traits have geography. The cause is ecological@ (Wissler 1926: 221‑222). Whereas the conjectural historians gave substantial depth of time for the archaeological record, culture‑area proponents greatly compressed the occupation of the Americas. Monumentality, recognized as mound building, was dated in reference to Mesoamerican epigraphic data. Mound building is a trait with prerequisites, two of the most important being pottery and agriculture. Traits accrete into civilizations, not greatly different than the basic model of conjectural history being railed against. Whatever the advancement, it should be found latest and simplest on the fringes. Even with a brief chronology, Mt. Taylor shell heaps are stratigraphically early, materially impoverished, and are found far on the periphery, therefore, the possibility of their intentional or meaningful creation is dismissed by rule. Cannibalism as a practice, one of the dirtiest brushes of the conjectural historians, is taken as fact (Wissler 1917: 250). Radiocarbon dating, not discovered when these ideas were germinating, indicates a relatively early date for the heaps, approximately

7300‑5600 cal 14C yr BP, which would make the Mt. Taylor heaps one of the earlier examples of mortuary monumentality in North or Central America, the exact opposite expected by culture‑core principals. If intention can be shown for the Mt. Taylor shell sites in question, we have legitimate warrant to question basic assumptions of the culture‑area school.

Cu l t u r e ‑Hi s t o r y Pe r sp e c t i v e s Post‑Depression era students of American prehistory embraced culture‑area theorizing, wholeheartedly at first (Ford and Willey 1941: 326; Goggin 1948b: 13‑16), and later, with an integration of evolutionary stage constructions (Willey and Phillips 1958). Throughout, the nuclear model remained as the central explanatory mechanism (Willey 1955). Classification and segregation of time and space were the primary operational goals and diffusion was viewed as the primary mechanism of change. For culture‑historians, complexity fluoresces over time as advancements accrete into complexes: people become more civilized. Two individuals associated 26 with this school, John Goggin and , most directly impacted Mt. Taylor archaeology, contributing to and participating in broader culture‑history debates, while writing the chronologies that continue to define Florida archaeology. Their concern with Florida was primarily in testing the idea of an Antillean interaction sphere (Rouse 1951), and despite their long‑lasting influence, they were not primarily concerned with the St. Johns. Neither Goggin nor Rouse did much excavation on the St. Johns or Indian River, and the definition of the Mt. Taylor period primarily uses Moore’s data from six decades previous. In reference to mortuary ceremonialism, Rouse (1951) only noted the presence of burials in what he termed camps or villages and restated the cannibalism thesis. Goggin’s reading of Archaic Tradition mortuary practice followed received wisdom:

“Ceremonial life does not appear to have been complex, nor does there yet appear a cult dedicated to the dead. Primary burials were made in middens, but the scattered human bones in these sites suggest that little interest was felt for the remains; perhaps, as Wyman has suggested, there was even cannibalism. No offerings accompany the interments (Goggin 1949: 21).” This is the entire discussion of Mt. Taylor burial or ceremonial practice by Goggin. He and Rouse do point out an element of Mt. Taylor mortuary behavior not emphasized previously, the presence of primary burials. But, burial does not seem to imply care of the dead, only the presence of grave goods would show humanistic concern: “This indifference towards the dead is emphasized by the lack of burial offerings” (Goggin 1952: 66). In the 1950’s, the first professional excavation of a Mt. Taylor site since C. B. Moore took place at the Bluffton midden prior to the destruction of the site by shell mining (Bullen 1955). Important information was gained from this project, but no burial remains were encountered. Soon after, one of the two recognized burial mounds at Bluffton was excavated (Sears 1960). Interpretations of the Bluffton Burial Mound had little impact on the standard model of preceramic mortuary ceremonialism, mainly due to its being a paradox in the model. This mound consisted of a single individual lain atop a heavily burned and calcined shell surface and surmounted by several discrete strata. The excavator had little doubt that the mound was intentionally constructed. However, despite there being no artifacts of any sort dating to later than the preceramic, the restrictions of 27 contemporary models of monumentality forced the excavator to give a ceramic period date to the mound (Sears 1960). What should be one of the clearest examples of preceramic monumentality is instead one of the clearest examples of the faults of cultural‑historical stage constructions.

Pr o c e ss u a l Pe r sp e c t i v e s Not long after the work at Bluffton came the most important excavation of a Mt. Taylor site to date. Again in the face of shell‑mining, salvage excavations at the Harris Creek site in 1961 by Ripley Bullen (Jahn and Bullen 1978) encountered over 180 individuals in a complex shell deposit that would cause consternation for decades (Russo 1994b; Aten 1999). The archaeology of Harris Creek is summarized later in this dissertation, but the difficulties that Bullen had in interpreting the site illustrate warrant for the present project. Initial indications were that the interments were preceramic in date, as no artifacts associated with later periods were encountered during reconnaissance. Once excavation commenced, difficulties of interpretation arose. The intentionality of construction of the mound was obvious to Bullen, but since this contradicted the inherited models, he posited a ceramic period date for the mound, despite not a single sherd being found. Once analyzed, radiocarbon dates indicated a preceramic construction, and the structure was no longer said to be a mound (Jahn and Bullen 1978; Russo 1994b; Aten 1999). Eventually, Bullen came to accept the possibility of Middle Archaic mortuary ceremonialism, but it was the assumptions of the conditions of monumentality that created his problems of interpretation (Aten 1999). For the next several decades, Mt. Taylor mortuary practices received little further attention. Several cases of Middle Archaic subaqueous cemeteries were excavated, providing greater substance to basic models of mortuary practice (Carr 1981; Wharton et al. 1981; Purdy 1991;

Beriault et al. 1981; Sigler‑Eisernberg 1984; Doran 2002a) but these were not Mt. Taylor sites, and the standard model of simple people with simple practices continued to dominate. The puzzle of Harris Creek tended to set Mt. Taylor somewhat apart (Milanich and Fairbanks 1980: 151‑152), but the issue of monumentality received little attention until the 1990’s. Wyman’s thesis of cannibalism 28 did receive a peripheral dismissal (Purdy 1991), but overall, Mt. Taylor archaeology followed the same trajectory as most hunter‑gatherer research, with environment and ecology becoming dominate sources of explanation. Models of Mt. Taylor society were again explained by reference to living groups, who were the essence of simplicity (Milanich and Fairbanks 1984). Subsistence and adaptation became the focus (e.g. Cumbaa 1976; Sigler‑Eisenberg 1985b; Russo 1985), and, with exceptions (e.g., Purdy 1991), mortuary practices commanded little attention. In Florida, the portrayal of shell site people as primitive and simple, concerned almost exclusively with food acquisition, mitigated against a possibility of complex achievements, including monumentality.

Co n t e m po r a r y Pe r sp e c t i v e s A recent summary article by Sassaman (2004) and a volume edited by Gibson and Carr (2004) focusing on the topics of complex hunter-gatherers and Archaic monumentality provide snapshots of contemporary thought on these subjects. The Mt. Taylor period deposits are part of these conversations, though they have previously played only a peripheral or implicating role (Sassaman 2004: 257), in part due to an infrequency of modern excavation. Current work is changing this situation (Sassaman 2002; Randall and Sassaman 2005). The theoretical approach employed in this dissertation and detailed in the next chapter is unconcerned with complexity. Of any particular position within the current debates of Archaic complexity and monumentality in the southeastern , the strongest principle is adherence to the caution of Saunders (2004) of not extending models of agriculture organization to hunter-gatherer constructions and the concept of beneficent obligation of Gibson (2004). Brief reviews of complex hunter-gatherers and shell heap monumentality are essential to understanding contemporary positions.

Shell Heaps as Monuments

In the late 1980’s‑early 1990’s, a few researchers began to question the standard paradigm of shell‑bearing site research, in particular the notion that abundant shell in archaeological sites solely represented subsistence debris (Widmer 1988). That shell heaps were not exclusively middens was not a new idea (e.g. Jones 1873; Dall in Thomas 1894), but those themes had little traction, and 29 the shell heaps are called middens ab initio. In the southeastern United States, the most prominent practitioners of this new perspective were Cheryl Claassen working on preceramic Archaic sites in the Tennessee‑Green rivers area (1988, 1991a, 1991b, 1996), and Michael Russo, who worked on preceramic Archaic sites in southwest Florida (1991, 1994a, 1994b, 1996). Both argued that in at least some circumstances, shell heaps represented intentional constructions. This radical departure injected monumentality into the discussion of shell sites. Mt. Taylor period sites were part of this discussion early on (Russo 1994b), but no archaeologist actively interrogated the relevant archaeological record until Aten (1999). While the idea of monumentality among shell site people has now become a current of possibility, demonstrating intentionality in shell heaps has proved more difficult than stating the hypothesis. Russo has most fully stated the conundrum of monumentality as an expectation of complexity. In what he terms the “operative evolutionary paradigm” monuments are not compatible with explanatory models, meaning that the evidence does not support a model of complex society, therefore, mounds are impossible by rule. The onus is demonstrating convincing evidence of intentionality. Structure is traditionally shown by layers of earth (e.g. Saunders et al. 1997), but shell matrices, regardless of content, including burials, require a more convincing methodology and theoretical approach (Russo 1994a, 1996). While preceramic earthen mounds with no burials are commonly accepted as built structures, shell heaps with burials are still disputed as being intentional architecture. The major difficulty is in making explicit distinctions between intentional construction and incidental accumulations. Focusing specifically on the structure of the Mt. Taylor shell heaps, I attempt to show how those deposits with human remains are regularly configured in contrast to arenas of secular activity. A historically and spatially particular approach is presented that includes a thick description of material conditions. If Mt. Taylor heaps are shown as monuments, the material conditions documented provide a context for monumentality, though not the essential conditions. It is expected that monumentality will arise in many situations, including foraging and 30 agriculture economies.

Hunter-Gatherer Social Complexity Complex hunter‑gatherer has emerged as the default shorthand term for intensive foraging economies of many stripes (Phillips 1983; Brown and Vierra 1983; Marquardt 1985; Arnold 1996; Prentiss et al. 2007), though the applicability to Mt. Taylor seems taken as a given, rather than something to be demonstrated (Sassaman 2004). Interpreting Mt. Taylor in a comparative framework that includes complex hunter‑gatherers has not been fully engaged, but the theoretical insights from other regions potentially provide a reservoir of research direction (Sassaman 2004). Consequent to modern datasets and theoretical progress, the archaeology of St. Johns shell middens should be approached as material remains of fishing people in a high productivity environment, in contrast to an outdated model of shellfish eaters exploiting a marginal resource and having a marginal existence. Comparing Mt. Taylor material culture to examples of complex hunter‑gatherers reveals affinities and disconformities. Robert Kelly’s characterization of nonegalitarian hunter‑gatherers, with its strong emphasis on material correlates, provides a datum:

“Ethnographically, nonegalitarian hunter‑gatherer societies are characterized by high population densities, sedentism or substantially restricted residential mobility, occupational specialization, perimeter defense and resource ownership, focal exploitation of a particular resource (commonly fish), large resident group sizes, inherited status, ritual feasting complexes, standardized valuables, prestige goods or currencies, and food storage” (1995: 302‑303). During the Mt. Taylor period there is a strong focus on fish and aquatic resources, and a likely, if not demonstrated, restricted mobility. These features could provide the economic conditions for a nonegalitarian structure. Ritual feasting is found in association with mortuary deposits (Aten 1999; Chapter 8). Population density for the Mt. Taylor is unknown; the sites are large but time spans are long (see Stein et al. 2003) . However, indicators of inherited statuses, standardized valuables or prestige goods are not readily found in Mt. Taylor contexts. Monumentality is not part of the definition (Sassaman 2004). 31 If greater social complexity precedes or is ancillary to monumentality is a separate issue. Does the presence of a monument indicate complexity? As Saunders (2004: 147) points out, this is circular reasoning. This is the converse of the paradigm that predicted complexity as a prerequisite of mound building and led to interpretive dead ends for Goggin, Bullen, and Sears. People if the Mt. Taylor period do not simply react to increased aquatic richness with an increase in the rate of accumulation of shell at exploitative camps, but respond to increased richness in part by greater tethering to local environments, more investment in essential facilities, and eventually, decisions to create new corporate organizations materialized as mortuary architecture. 32 Ch a p t e r 3: Th e o r e t i c a l Ap p r o a c h All generalizations are dangerous, even this one. Alexandre Dumas the Younger

This chapter details the theoretical perspective utilized in the analysis of Mt. Taylor shell heaps in testing the primary thesis that these heaps, in part, represent intentional constructions. I make use of a program of analysis proposed by Lindsay Jones (2000a; 2000b) that emphasizes the existence of the hermeneutical experience all persons undergo when apprehending sacred architecture. I also present two temporal concepts that relate to the time frames of these hermeneutical experiences. Finally, I utilize a Choctaw story of the creation of a burial mound to illustrate the creation of a ritual‑architectural reception history as applied to Native American monument construction.

Mo numentality a n d Sa c r e d Architecture Sacred architecture and monuments are related but not equivalent concepts. Not all monuments are sacred architecture, and not all sacred architecture would be considered a monument. Both are defined in reference to the social actors who create and interact with the subject. In this dissertation, the same criteria are being applied as qualification of the Mt. Taylor shell heaps as monument and sacred architecture: the presence of burials and evidence of intentional construction; they are both. The primary contrast between monument and sacred architecture is the temporal relation between social actors and the subject construction. Monument, by implication, refers to a past, a memorial or commemoration . A monument is encountered and interacted with, but is considered extant, a static structure and a message to be broadcast (OED). This is the sense of monumentality most used in archaeology. Sacred architecture is comparatively active, entered into and reflected upon, itself a participant in a sacred performance (Jones 2000a). Sacred architecture is not a messenger, but a place where messages are transmitted and received. The theoretical approach used here is accenting the shell heaps as sacred architecture. In employing a hermeneutical and experiential model, the focus is shifted “from the built form per se to the circumstances that arise in relation to that form, the notion of an architectural event (Jones 2000a: 50)”. Central to this theoretical perspective is the action of building the monument and 33 interacting with the monument, not the monument itself. Sacred architecture can be central to daily life, particularly when the architecture is located in close proximity to homes and activity areas. Mt. Taylor burial monuments are not set apart from the community, but are located within the village and encountered everyday.

A Hermeneutical App r o a c h t o Sa c r e d Architecture Lindsay Jones, a historian of religion who focuses on post‑Classic Yucatec Maya architecture, presents a conceptualization of sacred architecture that encourages application to the archaeology of the Mt. Taylor burial monuments in the upper St. Johns (2000a; 2000b). Jones terms his procedure hermeneutical, as it starts with the premise that architecture is always interpretive and reflective. All sacred architecture, at all times and all places, shares certain properties, the most important of which is that all sacred architecture is superabundant in meaning. Interpretive reflection always occurs, regardless if the sacred architecture is Notre Dame cathedral oran Amazonian hut (Sullivan 2000: xiv). It is this shared property of superabundant meaning that allows for comparison of all sacred architecture, regardless of time or place. In this experiential perspective, there is no hierarchy of sacred architecture; all examples have the same potential of meaning. We cannot know the meaning of sacred architecture, but we do know it is meaningful. Rather than revealing an essential interpretation, a certitude, “. . . understanding must be conceived as a movement of history, a process or sequence of occasions in which neither the interpreter nor the work of art can be thought of as autonomous parts” (Jones 2000a: 41, emphasis in original). This includes the indigenous and the academic. Jones is proposing to problematize architectural interpretation in terms of events or occasions (2000a: 41). To instigate this program, it is required that we move away from the study of sacred architecture as buildings, purely physical phenomena, and instead focus on the human experience of architecture (2000a: 29). Jones asserts that this procedure can be empirical, and suggests a comparative projet that emphasizes the creation, as a researcher, of “ritual‑architectural reception histories” documenting over time the series of uses and interactions, the “monumental occasions” 34 that occur with sacred architecture, focusing not on the built form, but on the built form’s participation in these eventful occasions (2000b: xiii) in pursuit of: “. . . the ascertainment of the actual cosmological and ritual priorities of specific historical contexts” (2000a: 194). The aim is not to classify material culture, but to explore the behaviors and events occurring as a component of hermeneutical reflection and participation, to construct a history of actual experience from material remains. Persons experience, interact with, and interpret sacred architecture, during what Jones terms ritual‑architectural events. These are the monumental occasions referred to above, the intersections between people, buildings, and activity, when indigenous appreciations are formed. Greatly simplifying the details, ritual‑architectural events can be conceived as having a twofold pattern of enticement and engagement. Architecture can form part of the invitation to participate in its creation and history; labor and material resources may be invested heavily in sacred places, yet given without demand. While alluring, sacred architecture is also obviously a potential component of coercion. Once engaged in the hermeneutic experience, people are transformed in their apprehension, changed by experience. Each reception is unique and productive, while still contextualized in past and present. Ritual‑architectural events can have multi‑scalar material consequent. A single devout pilgrim burning incense at a shrine may leave little material while five centuries of pilgrims burning incense at the shrine may accumulate substantial deposits. Some of these happenings include large‑scale investment and material consequent in architecture: they are more materially impressive. Other happenings may leave less of a material signature, but are no less meaningful. Inferring “sacredness” of archaeologically recorded architecture is a preliminary effort towards extrapolation of a class of happenings, but there may be substantive reasons for such a premise. The presence of human remains, particularly as burials and in other contexts of care and celebration, should be reasonably indicative of hallowed architecture. If a sacred nature can be attributed to an architectural feature and we accept that all sacred architecture is superabundant in meaning, occasioned by eventful hermeneutical reflection, then it is reasonable to expect that we 35 can recognize material evidence of ritual‑architectural events (see Johanson 2004). Determining serial order to ritual‑architectural events is part of the strength of archaeological investigation – there is potential for creating ritual‑architectural reception histories. Jones views classification through the lens of Foucault. Classification is not a product of any inherent nature or property of the thing being classified, but instead, is a byproduct of the classifiers own foreknowledge, presumptions, and biases (2000a: 151). He is not proposing however a casting off of classification – comparison is inescapable and potentially academically productive. Instead, he urges a “highly self‑conscious” form of comparison (2000a: 152), i.e., we must be aware that we as scholars are engaging in a hermeneutical exercise, but accept that we still find utility in examination of wider patterns and tendencies. For Jones, it is indigenous apprehensions of sacred architecture during ritual‑architectural events we should compare. Pursuing the event, Jones has developed a historical‑particular methodology as ritual‑architectural historiography that problematizes and rigorously contextualizes “specific occasions of meaningful action wherein specific people and specific buildings are brought into hermeneutical dialogue” (2000a: 191, emphasis in original, 2000b). This is Gertzian thick description, with the explicit intent of cross‑culturally comparing those specific occasions, but not necessarily forwarding generalizations. Jones also proposes that we deal with time differently than in traditional analyses. He views time as one of several means of organizing the diversity of hallowed architecture, but from my perspective of archaeology, it is important that we explore this dimension of understanding. Time is part of our currency. The nature of interaction of person and sacred architecture, i.e., the hermeneutic of the ritual‑architectural event, is by definition situated in the present of the individual, basically synchronic, though contextual and always comparative (2000a: 171‑173). Ritual‑architectural reception histories are by implication diachronic; monuments are constantly revalorized and some have a tenacity of relevance. Experiences are not directly retrievable: that is implied by superabundance. Retrievable are the temporal, serial, and material aspects of experiences, the context of experience. I am attempting below to mediate understanding of ritual‑architectural events at scales of experience. 36 Temporalities In application of Jones’ framework to the archaeology of upper St. Johns burial monuments, i.e., melding experience and vector, it is heuristically useful to cast time into two contrastive forms: transitory and persistent temporalities. Pacing the level of investigation at the tempo of human experience demands we find a means to recognize something representative of this timing. Considering that time, along with space and material culture, is one of the three primary sources of archaeological inference, it is requisite that we treat time as a form of data instead of restricting time to a means of organizing other data. Concurrently, if the aim of analysis is to reconstruct the history of actual and specific interactions of people with sacred architecture, it is necessary that we consider time as it relates to experience (cf. Sassaman 2005). In this dissertation, temporalities are defined as referential, relative time frames of limited duration. Transitory temporalities are defined as maximally not extending beyond the lifetime of one individual. This is analogous to Bailey’s sensational time (1983). Persistent temporalities are defined as minimally extending beyond the lifetime of any one individual. This is analogous to Bailey’s representational time (1983). Temporalities of experience are therefore referential, relative time frames of personal and social experience. Transitory experiences are “the conscious events that make up an individual life” (Merriam‑Webster 1999: experience 3a). Transitory experiences can co‑occur as a group during an event (eyewitness), but can never be shared. Only a relation of the experience can be given or taken (if given directly by an eyewitness, the receiver is an indirect witness). Material culture is formed by transitory experiences (Gamble 1999). Transitory temporalities of experience are the relative time frames of transitory experiences. A transitory temporality of experience must be within the lifetime of an individual. “I am a survivor of World War I” is an example of a transitory experience of co‑occurrence. Focusing on transitory experience, a maximal temporal span of communal events would terminate with the passing of the last participant in a specific event. The event may be brief, but experience lasts a lifetime. Persistent experiences must be shared. Persistent experiences are information received and 37 broadcast transgenerationally by a cohort of identity, the “events that make up the conscious past of a community or nation or mankind generally” (Merriam‑Webster 1999: experience 3b). Material culture is in part informed by and translated through persistent experiences. Persistent temporalities of experience are the relative time frames of persistent experiences. The time referred to is generally “unspecific” in that it is not necessarily related to particular calendrical dates or even to actual historic events. “Our nation fought gallantly during World War I” is an example of a persistent experience – it is a matter of identity. Once an experience passes out of living memory, it can only be persistent.

The Making of Nanih Waiya

To illustrate these two types of temporalities of experience and their relation to ritual‑architectural events, I will describe and discuss a historical text that was collected from a Mississippi Choctaw informant by Gideon Lincecum in the late 19th century (Swanton 1931: 11‑27). This text, which exists as a fragment, relates a journey by the antecedents of the Mississippi Choctaw from “the west” to their historic home range near Philadelphia, Mississippi and the establishment of a permanent community. Along the journey, the people carried bones of their ancestors. After settlement near Nanih Waiya in modern Winston County (spelled Nunih Waya in the relation and herein), a burial mound was established to hold the bones of their ancestors and themselves. The text also describes the practice of the , the establishment of another mound, and battles with the Chickasaw. The first 52 paragraphs (76 total) of this text are given in Appendix A. Synopses of relevant passages are given in this section, with references to the paragraph numbers. This ethnohistoric document is being used for illustration, not analogy. This relation begins somewhat abruptly, with the journey from the west already underway. Travel was halted to give time for the people who are carrying the bones of their ancestors to catch up. The load of bones is great; some families are said to be able to carry nothing else. Individual skeletons outnumber the living. People are devoted to their ancestors though and their own death is preferable to abandonment of the bones (1). The minko, or “chief”, the terms are used interchangeably in the text, looks upon the burden of bones as an evil and begins to ponder a 38 means of getting rid of the burden (2). Following the indications of the sacred pole, which is placed into the ground each night and each morning points the way forward, the people continued on to the southeast, until reaching a point of good lands, water, and foods. The pole had at sunrise after arrival punched itself deeper into the ground and remained erect. The chief orders a stop for winter and the arrangement of camp (3‑5). Several days are required for all the people to make it to the camp, due to the duty of packing the bones. Complaints are rumbling among the people due to despair of continuing on the journey with such loads (6). Before winter, preparations are made for corn plots and seed corn gathered. Life is good at this place. A hill with a hole in the side above one of the camps is given the name Nunih Waya, leaning hill. The encampment is given this name from this time forward (8‑9). Land around is explored and found to be fruitful, the land of vison. The first winter passes and the seed corn returns the first yield (8‑10). “All were filled with gladness” (10). The first Green Corn dance is held at Nunih Waya, after 42 years in the wilderness. The corn is not sufficient to eat, but a pole is erected and a single ear of corn suspended from this pole. People are concerned that the sacred pole will indicate that they should move on, and are fearful, due to the large burden of ancestral bones they would have to transport if a move is mandated. At the same time, they cannot imagine leaving the bones behind (11‑12). The chief is pondering what to do about the burden of the bones. He is aware that the bones are considered sacred relics, and any action must be taken cautiously. The chief is apparently unconcerned with devotion, but only with the bones as a hindrance. “Yet, the oppressive, progress‑checking nature of the burdens was such that they must be disposed of in some way” (13). A council is called by the chief with the leaders of the iksas, appealing for a means to deal with the transportation and adjudication of the bones (14). A short period of time is given for consideration of the issue (15). The council is reluctant to deal with the issue of the bones. People have packed the bones for a long time and over great distances, but for some, there is little connection to the bones 39 themselves. A prohibition against speaking the names of the dead had resulted in a situation where the descendants had no memory of the ancestors. Leaders of the iksas (clans) were fearful of disrespecting the ancestors or bandying about such topics in council. Spirits of the dead are thought to be present, hovering about to insure that they were cared for and not disrespected (16). One Isht ahullo had taught secretly that the bones were sacred relics, not to be labeled as burden or incumbrance. Holding an exclusive council, this man pleads against casting away of the bones, as he suspects the chief wishes to do. The ancestors have charged the people to take care of their bones and carry them whenever they move. If the people were to cast away the relics, then vengeance would come from the spirits, and hunters would fail, disease and hunger, confusion and death would come through the community, until wild dogs feasted on the carcasses of the dead (18). The chief is angry about this secret council, described as being made by bad men with “dark and mischievous influence on the minds of the people” (19) who are sowing discontent. The minko calls forth a meeting of all the people with exception of the “leaders of the clans and conjurers of all grades” (21). Once the people are assembled, he gives a speech against those who encourage carrying the bones. His appeal is to secular concerns and is explicitly not spiritual (23‑25). “I speak not to please or benefit the dead; there is nought I can do or say, that can by any possibility reach their conditions. I speak to the living for the advancement and well being of this great, vigorous, live multitude” (25). The pole appears to have settled permanently, and the journey through the wilderness over (26). But once given this message, according to the chief, the leaders of the iksas and the conjurers spoke of the dangers of disobeying the demands of the ancestors, while also laying claim to exclusivity of communication with the spirits. The chief tells of his dismissing the lot, and has deliberately excluded them from this democratic council. His appeal is to the nation (27). The journey is over in his judgement. The land is good, prophecies have been fulfilled (28). A call is made to become settled (29). In order to take advantage of their improved and providential situation, the chief suggests that 40 a suitable location be chosen, and the bags of ancestral bones piled together and covered with cypress bark. After the pile is created, as appeasement and respect for their dead kin, they should be covered with a mound of soil. The argument is made by the chief that the bones are all from the same kindred, or iksa, and will be pleased to be buried together (29). In the future, after he dies, the chief wants his bones interred in the monument. After his proposals are presented to the people, the chief asks for comment and intercourse (30‑31). An elderly gentlemen who would have been seventeen at the beginning of the migration tells the story of carrying the bones of his father since his boyhood. He loved and respected the bones as sacred, ready to give his life to protect them. But, as his own life is coming to its end, he is reflecting on what will become of his father’s bones in future generations. He has thought of a resting place for his father, but like all the people, was embarrassed by his desires. The proposal of the chief seems a reasonable solution (32). The elder counsels the people to follow the suggestion of the chief rather than leave their ancestors to be tumbled in “greasy packs” and bones lost accidently (33).

The people are convinced (34‑35). A group of men is chosen to find a place for the monument and direct its building. Finding a suitable spot, on sandy land, an oblong square is laid out and soil dug for the first stage of mound building. The soil comes from a spot to the north some ways off. The soil is piled to as high as a man’s head, and then tamped and leveled. A floor is made of cypress bark, and then the sacks with the bones in them are piled up. People are glad to bring their ancestors, and an enormous heap of bones is created. Cypress bark is used to cover the bones, and then covered with soil until there was a “mound half as high as the tallest forest tree” (36). After the population covered the piled human bones with soil, the minko kindles a fire and calls the people together again. He praises their work and the respect demonstrated towards the ancestors. A halt was called to the work for exploitation of the autumnal mast harvests and preparation of fields. After the next Green Corn dance, work should resume on the mound, year after year, until the top was as high as the tallest tree in the forest, with a level platform on top sixty‑by‑thirty steps large, tamped, and planted with nut‑bearing trees (37). 41 Autumn is abundant, and a fall feast is held. After preparation of planting grounds, an elderly gentleman suggests that people return to working on the mound, and not wait for the Green Corn ceremony. After this, people would in honor carry soil to the mound when they had free time (39). The spring brought good conditions, and after planting, people again contributed to the volume of the mound (40). After seven years, everything looked good. Explorations revealed good land and no enemy. Harvests were abundant, and work continued each year on the mound. After the eighth Green Corn ceremony, the committee of men who had guided the building of the mound decreed that the dimensions sought had been reached, and the time had come to plant the trees (41). The minko calls for the man who had carried the emblem of the golden sun during the journey to bring it to the top of the mound and plant it center (42). The people beheld the sun emblem and the monumental work they had accomplished, and were glad. Songs and feasting celebrated their accomplishment, the magnitude of the monument, and the honor given to their ancestors. The time has come to settle, and it is to be at Nunih Waya (43‑44). After feasting for five days, each iksa bought peeled pine poles and ornamented them. For a lunar month, they performed a solemn cry three times a day. After the moon, they gathered for a feast and dance for two days. Mourning, and the monument, were considered completed (45‑46). From then on, after death and cleaning of bones, people “were deposited into a great cavity which had been constructed for that purpose” (47) and the mound became the vault for the nation’s dead. In special circumstances, there were proscribed appropriate actions. If a hunter died too far away to be brought back home, the body and personal effects of the dead were left in place and a conical mound placed over them. Other situations are also addressed (47). The text diverges into a discussion of conical mounds and the recorder’s (Lincecum) excavations of mounds along the Tombigbee (spelled archaically) and his findings there. He claims that the archaeological situation matches the tale teller’s description (48‑49). Following the completion of the mound, the people felt they were permanently settled at Nunih Waya. A speech by the minko summed up the accomplishments of the nation: 42 “We are a brave and exceedingly prosperous people. We are an industrious people. We till the ground in large fields, thereby producing sustenance for this great nation. We are a faithful and dutiful people. We packed the bones of our ancestors on our backs, in the wilderness, forty‑three winters, and at the end of our long journey piled up in their memory a monument that overshadows the land like a great mountain. . . We have traveled over a pathless wilderness, beset with rocks, high mountains, sun‑scorched plains, with dried up rivers of bitter waters; timbered land, full of lakes and ferocious wild beasts. Bravely we have battled and triumphed over all. We have not failed, but are safely located in the rich and fruitful land of tall trees and running brooks seen in a vision of the night, and described by our good chief who is missing. . . Assuredly we are a wonderful people. A people of great power. A united, friendly people. We are irresistibly strong” (52). The remainder of the relation discusses the erection of another mound for the sacred pole and conflicts between the Choctaw and Chickasaw. The story ends as abruptly as it began.

Discussion We can chart this story as it relates to transitory and persistent temporalities, ritual‑architectural events, practice, and depositional correlates by focusing on the transformation of ancestors to burial mound (Figure 2). The story itself is a persistent experience and is referenced to a persistent temporality of experience, i.e., the reference is to the society’s received past. Within the story are references to persistent experiences and temporalities (received ancestors, received burial monument), as well as transitory experiences and temporalities (journey, debate, monument creation, triumph of permanence). The central theme of the story is transformative B a wandering life bearing the bones of their ancestors to a sedentary life with an actively engaged burial monument and claims of territoriality. This transformative event also serves to connect the time before the monument, the time of the making of the monument, and the time since the monument. Following the theme of Jones’ framework, we can focus on particular aspects of the story to form a ritual‑architectural reception history (Figure 2). In this chart, a persistent temporality of experience encompasses an unspecified span from ancestral time to “present”. The four described ritual‑architectural events are termed: Heirlooming, Initiating, Inciting, and Reproducing. The creation of Nunih Waya is encompassed mostly by a transitory temporality, but transitory experiences are contexualized by and contextualize persistent experiences, hence a dashed boundary line 43

Time Persistent Temporality of Experience Burial Ancestors Monument Transitory Temporality of Experience

Nunih Waya Ritual-Architectural Reception History

Ritual-Architectural Event- Heirlooming Initiating Inciting Reproducing Carrying of Surface Preparation Building Continued Bone Sacks, Stacking of Bones, of 1st Mound Burial, Mound Practice- Transporting Covering with Cedar Stratum Constructions During Migration Loose Human Promiscuous Mound Stratum Burials Into Depositional Correlates- Bone, Random Secondary Over Secondary Mound Stratum, Bones in Secular Burials Burials Additional Strata Contexts Figure 2: Ritual-Architectural Reception History of Nunih Waya divides the first and last described ritual‑architectural events, sort of a liminal temporality. Heirlooming, referring to the direct care of ancestral remains, while not being “architecture” itself, is still here nominated as a component of this example of sacred architecture since these remains and their adjudication are the “priority” of the monument in Jones’ sense (2000b). The ancestral remains are the material inheritance of a persistent temporality of experience, being acted on and experienced by contemporary individuals. Heirlooming is part of the allurement of the Nunih Waya monument. Practice is transporting ancestral remains in bags wherever they move. Possible depositional correlates include random human bone in secular contexts. Initiating, facilitating the beginning of, refers to the architectural event of creating a heap of the ancestral bones. Properly, this includes the preparation of the site and collection of cedar. The idea of a communal burial mound was rhetoric until this moment, but again, the material priority is the received ancestral bones. Promiscuous burials, meaning composed of many individuals, would mark this event. Inciting, moving to action, is the communal building of the monument, the piling of the soil over the bones, not only as deemed by the chief, but as a duty and sign of respect. People continue the piling of soil for a total of eight years, after which the mound is considered “finished” as finalizing 44 the transformation of ancestors to architecture. The bone mound is covered with a large stratum of earth. Depositional correlates would be a layer of soil encapsulating promiscuous burials; the monument is built over the ancestors. The horizontal inclusion of community members may be greater than those who deposited ancestral bones in the heap. Reproducing, to produce again, is the continued use of the monument for the inclusion of human remains from the community. The idea is contained in the chief’s wish to have his bones interred and in the architecture becoming a sepulchral vault for the nation. Reproduction is not replication, but continued, unique, ritual‑architectural occasions. Human bones are being placed into the monument. Burials originating from the surface are not intrusive, but a depositional correlate of this event. Exploring this story of Choctaw genesis, it is clear that ritual‑architectural events of the sort Jones refers to are describable. It is also apparent that these events can have material consequent. Additionally, it is possible that time frames of these hermeneutical reflections are describable. Given the potential of this framework to the archaeological record, we can now assess the upper St. Johns shell heaps as records of experience. 45 Ch a p t e r 4: A Hi s t o r y o f Mt. Ta y l o r Ar c h a e o l o g y

Initial Fo r m u l a t i o n The first published notices of the large shell deposits along the St. Johns river come with the work of John Bartram. Appointed by King George III as Royal Botanist in the North American colonies, John Bartram explored much of the southeastern seaboard, including a sojourn up the St. Johns in the company of his more famous son William just prior to the American Revolution (J. Bartram 1942). The expedition of John Bartram is relevant for this dissertation as he navigated to the Upper St. Johns Basin, a feat that was rarely accomplished by whites for the next century. He described the landscape, including soils, waterways, general characteristics of topography, and his immediate surroundings, in addition to the botanical specimens he was sent to observe and collect.

He recognized the shell‑bearing sites as originating with native people, and refers to Indian sites consistently when encountered. He describes shell bluffs, fields, and mounds, all as forms of Indian sites. Contemporary (Seminole) people are not usually mentioned, though it is certain that many local natives were encountered and assisted in the expedition. Of special note, John Bartram sometimes recorded the thickness of archaeological, geological, and soil strata. William Bartram primary interest was in the natural world he encountered, not archaeology. He did no excavation, and beyond florid descriptions of above ground features and the mention of artifacts on the surface, he gives no attention to material remains. Nevertheless, his are useful accounts for general site locations and configurations. But for the purposes of this dissertation, the utility of William Bartram’s account is limited, as he did not go further south than Astor, in the Middle St. Johns, but instead turned west towards the Creek Nations, never returning to the upper St. Johns. Daniel Brinton of the University of Pennsylvania traveled much of the lower and middle St. Johns during the 1850’s, but only brushed the upper St. Johns. His published work includes some exploratory excavations of shell sites. Brinton however did not initially recognize the riverine shell heaps as being culturally constructed, a bias that seriously effects his interpretations. He 46 later did come to accept a cultural creation of the St. Johns shell mounds, but only with minimal comment (Brinton 1859; 1872). As first director of the Peabody Museum at Harvard University, Jeffries Wyman selected the St. Johns basin for his initial collecting foray. Making several journeys to the region, Wyman explored a large portion of the Lower, Middle, and Upper St. Johns, conducting research in 1860, 1867, and intermittently between 1871 and 1874 (1868a, 1868b, 1868c, 1874, 1875: 18, 1876 ). Wyman was a mostly thorough and conscientious scientist, observer, and recorder. Often Wyman is given great credit for early application of hypothesis testing in archaeology, as regards a natural or cultural origin of the shell heaps by undertaking a program of survey and limited excavation (Willey and Sabloff 1993). Wyman=s work resulted in several important findings. He noted that there were preceramic strata at several of the sites encountered. Wyman also noted the transition from fiber‑tempered to sponge‑spicule tempered pottery, and gave accounting of the animal remains found in the shell heaps. Most notably, Wyman described all the shell heaps as middens, without exception. He also developed, based on analogy with historically observed groups, the thesis that St. Johns people were cannibalistic. Moore is often given most credit for developing the lasting ideas about the behavior of St. Johns people, but most every one of his interpretations can be traced back to Wyman. The largest source of data for this dissertation is the work of Clarence Bloomfield Moore. Moore made several forays into the St. Johns basin, excavating dozens of sites, and faithfully reporting his results. Having become intimate with his notebooks, it is admitted that at times, Moore's methods make one twinge, and claims that all of these sites would have been destroyed are too universal. Regardless, it is unarguable that because of him that we have any archaeological information for a large number of sites. Moore excavated more sites in the St. Johns than anyone before or since, and is the only archaeologist to excavate extensively in the upper St. Johns (1892b; 1892c; 1892d; 1893a; 1893b; 1893c; 1893d) . It is his work that provides the sole source of data for most of the sites in the 47 region, and is the southernmost extension for major archaeological excavations of Mt. Taylor shell heaps in the St. Johns basin still today. The large amount of work done by Moore has a large and at times perhaps undue influence on all later interpretations. He has been taken as the final word on many subjects without question. Detailing of Moore's excavation at the four upper St. Johns sites discussed in this dissertation is found in Chapter 8. His field notes are used extensively in this dissertation, and no review of the area is complete if stopping with the published record. Briefly, he made much more explicit the existence and detail of the preceramic strata within the shell heaps, and provided the particulars of relative chronology, soil matrices, artifacts and mortuary remains missing from Wyman's publications. Moore is also clear in distinguishing shell middens from other types of deposits he recognized, mainly sand burial mounds. Importantly, he never equates the presence of human remains in strata that are dominantly shell with intentional burial, meaning human remains in preceramic deposits are not considered intentional. He argues directly against it. However, there are several cases where it is clear that mounds with no pottery represent intentional mortuary constructions. That preceramic burials are intentional is the central question of this dissertation. Moore managed to explain away obvious contexts, such as Persimmon Mound, as anomalies and not being certain proof of any contemporaneity of shell heaps and burial mounds (1894c: 209‑211). William Henry Holmes (1903:120‑121) comments on the absence of pottery in the early layer of St. Johns shell heaps, accepting that this is time before the invention/usage of pottery in the area. In the early 20th‑century, Nels Nelson described a preceramic horizon at the Oak Hill shell mound on the west bank of the Indian River lagoon (1918). In this article, Nelson explicitly seeks to use the stratigraphy of the site to demonstrate chronology, and provides a general preceramic, plain‑ware, check‑stamped ware sequence. Nelson reports a separate nearby shell mound that contained burials and an adjacent spring (1918: 81). This was one of the few professional archaeological endeavors for nearly forty years, with only minimal attention paid to peninsular Florida until depression‑era relief excavations in the 1930's. Matthew Stirling, who directed several of the relief projects, published the first attempt of synthesis for Florida, a brief essay that divided the state into four 48 archaeological regions (1936). John Goggin marks the work of Stirling as the beginning of the modern period of Florida archaeology (1949). Several thrusts of Stirling's article deserve discussion, as they provide some of the main planks for the structure of Florida archaeology. The notion that Florida served as a "huge cul de sac, which has received and blended a variety of impulses from the north" (351) encouraged the perception that all local prehistoric cultural accomplishments were adventitious, with little or no original contribution to developments elsewhere. The presumption that only a brief period of time was represented by the archaeological record was then current throughout Americanist archaeology (O'Brien and Lyman 2001), and Stirling felt that mound building across eastern North America represented a brief "golden age" with most archaeological remains being assignable to historic tribes (1936; 351). An implication of this brevity is that if preceramic populations were present, the population must have been small and still relatively recent (352). The creation of archaeological regions or areas, largely derived from Boas‑Wissler‑Kroeber principles of organizing cultures and geographies (Kroeber 1939) cast the die for Florida archaeological approaches.

Fo r m a l i z a t i o n Goggin took Stirling’s scheme further, creating more formal definitions of the archaeological areas and periods in Florida, including the first organized statement for preceramic time frames in the St. Johns. Within five frantic publishing years, the basic temporal framework for the Native American occupation in all of Florida was sketched out, with almost all the data coming from limited survey and fieldwork, C. B. Moore’s explorations and the nine Federal Emergency Relief Administration excavations of the 1930’s (Goggin 1947; 1948a; 1948b; 1949; 1952; Goggin and Sommer 1949; Rouse 1951; Willey 1948; 1949a; 1949b). Still today, Goggin’s 1952 designations for the middle and lower St. Johns basins remain the core of the accepted temporal scheme (Milanich 1994). In the initial formulation (Goggin 1947), Florida was divided into eight archaeological areas or regions. The Upper St. Johns, the area of concern in this dissertation, includes portions of 49 two of these regions: the Northern St. Johns Region, stretching from the northern state boundary and including the St. Johns river basin and adjacent coast to just south of Cape Canaveral. The southern extremity of this region would include all the shell heap sites described originally in this

dissertation. The remainder of the Upper St. Johns, basically all that below Lake Cane1 would fall within Goggin’s Melbourne Region. The Melbourne Region included the St. Johns drainage and all lands to the coast from Cape Canaveral to the Glades Area in his scheme, i.e., to the true headwaters of the St. Johns. For the Northern St. Johns Region, a non‑ceramic horizon was designated. In the Melbourne Region, a non‑ceramic horizon was considered possible, but not

given definitive status. According to Goggin (1947: 122), no evidence of a preceramic period had been shown for the Melbourne region. Though the work of Wyman and Moore had encountered extensive prepottery deposits, little was known about the culture (123). As an illustration of the compressed time perspective current in some Americanist archaeology at this time, the non‑ceramic period was thought to date between 200 A.D. and 500 A.D., only six thousand years too late. In a revision of his temporal framework the next year, Goggin introduces the name Mt. Taylor to encompass the preceramic period in the Northern St. Johns Region, continuing to include any deposits stratigraphically precedent to ceramics. The northern region continued to include the area to a Cape Canaveral parallel. The Mt. Taylor period concept was not applied to the Melbourne Region, but there no longer remains a question mark to the existence of what is now referred to as a preceramic period (Goggin 1948a: Fig. 12). This paper also marks the usage of Malabar for the post‑Orange ceramic traditions in the upper St. Johns in contrast to the northern region, a direct influence of Rouse (1951). Goggin’s dissertation, Culture and Geography in Florida’s Prehistory, completed at Yale under Rouse, is a remarkable effort at imposing order on Florida’s archaeological record, encompassing the entire state and including data from nearly 1400 sites. Goggin clearly states he has an explicit debt to the culture area concept of Wissler and Kroeber, while recognizing some of the problems

1 The same lake is also known as Lake Cone and Lake Clement. 50 with the method (1948: 13‑16). A formal split is made of the Northern St. Johns and Melbourne regions at the coastal site of Rockledge (1948b). This is further south than the boundary from the above articles, and now bisected Lake Poinsett. The Rockledge division was considered a stopgap measure, as the Melbourne region was under study by Irving Rouse while Goggin was at Yale, and was soon forthcoming in published form (Rouse 1951). Including the area above Rockledge continued to mean that Palmer‑Taylor, Persimmon, and Orange mounds were part of the Northern St. Johns region and are used in his formulation of Mt. Taylor. It was in his dissertation that Goggin first fully developed the organizing principles of his temporal‑cultural constructs. As Weisman demonstrates (2002), Goggin’s system constituted a paradigm, constructed of three main elements, space, time, and tradition (called pattern in his 1948 dissertation), coming together in three dimensions, analogous to a cube. At the same time, the system is hierarchical, with time and space intersecting in areas as periods, with periods being subsumed under traditions. Space and time are generally self‑explanatory; some word need to be said regarding tradition as utilized by Goggin. The concept is most concerned with long‑term material and organizational stability at a general level, with small‑scale change being a constant, but not transformative trait. When formulated as pattern, Goggin defined the term as “a distinct cultural complex which may in the course of time pass through some changes but not enough to alter the basic configuration” (1948b: 214). Soon afterward Goggin (1949) embraced the term tradition, perhaps by Rouse’s influence (Weisman 2002). Goggin distinguished his cultural tradition from the related concept of ceramic tradition, and defined it as:

“A cultural tradition is a distinctive way of life, reflected in various aspects of the culture; perhaps extending through some period of time and exhibiting normal internal cultural changes, but nevertheless throughout this period showing a basic consistent unity. In the whole history of a tradition certain persistent themes dominate the life of the people. These give distinctiveness to the configuration” (Goggin 1949: 17). It was Goggin’s attempt at synthesis for the whole state, i.e., the need for an organizational mechanism, which spurred him to create his constructs; it was the perception of stability, attributable mainly to there being little perceived change in the artifactual record, which helped to determine 51 the features of the constructs. The Archaic tradition, which includes the Mt. Taylor period, is marked by semi‑sedentary populations subsisting through hunting, fishing, and gathering. Pottery is present only in the later part of the tradition, and is only considered an addition to the existing Archaic pattern. The large St. Johns shell heaps are a major site type, but “camp” sites with less refuse are also found. Artifacts are rare, mainly bone worked into utilitarian forms, chipped stone points, blades, scrapers, and shell tools (1949: 21‑22). Mt. Taylor is referred to as a period, as the construct was used by Willey and Woodbury (1942: 236). The hope was that periods would prove to represent distinct cultures, i.e., that there would be a direct correspondence between material and ethnic identities. In the formal definition, there is little detail about Mt. Taylor beyond that already provided for the Archaic tradition. The area of Mt. Taylor is charted as the middle and lower St. Johns basins, with the sites of focus in this dissertation originally included (Goggin 1948b). The vast majority of data come from the Wyman and Moore excavations, and thereby was heavily biased to burial mound contexts, which are without question interpreted as middens. These “middens” are described as being composed of shellfish, varying amounts of sand, ash, humus, and faunal remains, including terrestrial and aquatic species. “There is some stratification in the deposits but it apparently has little cultural significance” (Goggin 1948b: 157). Some attention is paid to the diverse character of the shell deposits in the shell heaps and to the presence of . The Mt. Taylor Mound in the middle St. Johns basin served as the type site for the Mt. Taylor period. This was a large shell heap, nearby to the Bluffton complex. Moore had recovered a relatively abundant artifact assemblage from the preceramic deposits, and this collection provided data towards explicating the period. Artifacts associated with the Mt. Taylor period are few, and are interpreted as being indicative of a simple cultural complex. The same types that are described for the Archaic tradition are characteristic of Mt. Taylor, with no additions. Tool types include bone, shell, and stone artifacts, but all are sparse. The only artifacts that Goggin recognizes as having artistic value are the occasional engraved/incised bone pins, but these are still rare (Goggin 52 1948a: 159‑160; 1952: 42). There is minimal discussion of burials and mortuary patterns for the Mt. Taylor period. Again describing burials as being made in middens, Goggin lists only three sites with burials: Orange Mound, Persimmon Mound, and Osceola Mound (1948b: 160‑161). Two of these would be dropped from inclusion later when the division between the Northern St. Johns and Indian River (Malabar) regions was shifted to Lake Harney (Rouse 1951; Goggin 1952). Eight burials are recognized at Persimmon, one at Orange, and a sawed femur is reported for Osceola. Rollins College in Winter Park, Florida, hosted an important conference in 1949 focusing on the archaeology of Florida and relationships to other geographical areas: the southeastern United States, Middle America, and South America (Griffin 1949). Goggin’s paper at this conference was a synopsis of his dissertation work and outline for Florida’s prehistory. In this paper, Goggin laid out his concept of tradition and sketched the details of the traditions in Florida (1949). Combined with the other papers of this conference, the rough framework of Florida archaeology was felt to be generally worked out, with a call for future work to focus on particular problems and issues. As stated, with the publication of A Survey of Indian River Archeology in 1951, the southern boundary of the Northern St. Johns region was shifted northward to the outlet of Lake Harney and the name for the region was changed from Melbourne to Indian River. From a geographic standpoint, this is a more sensible division, as it segregates the upper St. Johns basin from the middle and upper St. Johns. Archaeologically, this division was meant to accent a perceived material‑culture difference: the greater abundance of Glades ware pottery in the Indian River region and adjacent upper St. Johns basin. This new division also separated Palmer‑Taylor, Buzzard’s Roost, Orange, and Persimmon mounds from other Mt. Taylor sites, being placed into the Preceramic period by Rouse (1951: 238‑239). Relatively extensive excavations at South Indian Field (Br 23) provided a solid stratigraphic sequence of cultural materials spanning the Orange through Malabar II periods (Ferguson 1951; Rouse 1951). The data as reported do not include a preceramic component. South Indian Field is far up the headwaters of the upper St. Johns on the edge of glades, near a small lake with the 53 appropriate name of Hell‑N‑Blazes, often found on modern maps in the genteel Helen Blazes version. If traveling by river to get to this point, which few actually do, you are quite certain that Hell‑N‑Blazes is the place destined. A political science Ph.D. candidate at Columbia University, Edwards conducted excavations soon after the Yale project at the very nearby Br 27, the Helen‑Blazes site, for his dissertation (Edwards 1954). This work recovered mainly preceramic artifacts, mostly chipped stone bifaces. Edwards makes an argument for at least some of the projectile points being types found in Late Pleistocene contexts. At least one of the points appears to be a Middle Archaic type (Edwards 1954: Fig. 27 A). This is the last professional excavation in the extreme upper St.

Johns for over 20 years (Sigler‑Eisenberg 1985). Though now segregated from other Mt. Taylor period sites by geography and category, it was the excavations at Palmer‑Taylor by the Excavators’ Club that allowed for a somewhat more robust characterization of the preceramic. Rouse includes Palmer‑Taylor, Buzzard’s Roost, Orange, and Persimmon mounds in the preceramic period, as well as the so‑called “Early Man” sites Br 44, Br 45, Br 47, and Ir 9. The “Early Man” sites were instances of likely fortuitous associations of human remains and Pleistocene fauna, and are not considered to be clearly contemporary with Mt. Taylor. An additional site, Br 36, is a coastal shell heap with no pottery reported from an avocational excavation, but there is not sufficient data to be certain of the assignment. Rouse noted that the known sites were all found in marshy/swampy areas. Making an astute observation from the data at hand, he proposed that this was a time of lower water levels. Based on the Excavators’ Club data from Palmer‑Taylor, he noted that the preceramic deposits lay on top of the natural soil surface, and were capped by muck. Atop this muck, artifacts from later pottery‑using cultures are found. Rouse also advanced that a lack of preceramic shell heaps on the Indian River or immediate coast was due to a different ecological regime that was not conducive to shellfish propagation. He noted that if Br 36 does date to the preceramic period, it would be an exception. It is not clear why he did not include Oak Hill (Nelson 1918) in his reasoning. He makes the lack of shellfish/different ecology inference based on the “Early Man” sites listed above, but the archaeological situation of these sites is muddled. 54 Reviewing habitation sites, Rouse asserts that: “All the sites apparently consist of the remains of small camps or villages” (1951: 238). This assessment includes all the shell heaps and the Early Man sites. The shell mound sites discussed in this dissertation are included, apparently in the category of villages. Mortuary remains are listed for Palmer‑Taylor, Persimmon, Orange mounds and the Early Man sites. No other sites are listed as having human remains during the preceramic period. Rouse points out that individual bones have been found at Palmer‑Taylor, and while at least some are from disturbed burials, there are burnt and broken human remains, interpreting this as support for Wyman’s thesis of cannibalism.

The picture of Mt. Taylor put forward during these formative years is of a very simple low‑level foraging culture, with little material possessions and no consideration of the afterlife. In this, Goggin and Rouse did not differ from their predecessors Wyman and Moore. Wyman forwarded the thesis of cannibalism to explain the presence of broken and burnt human bone (1875) and Moore endorsed the view, also doubting that the primitive groups who accumulated the heaps had a sense of interring the dead, particularly in reference to the preceramic deposits (Moore 1893c: 612). Goggin and Rouse both felt that cannibalism was a possibility, and that there was little reverence for the dead. “This indifference towards the dead is emphasized by the lack of burial offerings (Goggin 1952: 66).” With the publication of Space and Time Perspective in Northern St. Johns Archaeology, Florida (Goggin 1952), the details were reified, with little change from the dissertation other than the single area focus, the terminological switch to tradition instead of pattern, and the exclusion of the upper St. Johns sites from consideration in the Northern St. Johns region. The next reported professional work at a Mt. Taylor site is Bullen’s short article on the stratigraphy of the Bluffton shell midden (1955). Virtually all the large shell heaps in the middle and lower St. Johns have been mined for shell to be used in modern construction projects, and Bluffton, which was likely the largest freshwater Mt. Taylor heap areally, was no exception. During the mining of Bluffton, John Griffin and William E. Edwards arranged for salvage operations, conducted by Edwards and Bullen. Bullen fails to provide either a section drawing or photograph of the heap in 55 his report, using a table to illustrate the stratigraphy (Bullen 1955: Table 1). According to this table, the highest point of the shell heap that was recorded was 18 feet (5.5 m) above the ground surface, of which nearly 14 feet (4.3 m) were preceramic deposits. Based on Bullen’s description, this portion of the Bluffton site qualifies as midden (Chapter X), with two adjacent burial mounds. Mining continued at the Bluffton site, and in 1959, William Sears excavated one of the two burial mounds. The mound was to the east side of the large shell heap, away from the river, bordered on the east by two circular ponds. Sears’ excavations revealed clearly that this was a burial mound. A large fire was placed onto an extant shell layer, heavily calcining the shell; atop this burned area a single extended burial was placed. A mound of sand was placed atop this burial, capped by a thin mantle of shell, followed by another lens of sand. This sand was then capped by a layer of black muck, capped by another layer of sand. The mound was then capped by a thick layer of secondarily deposited preceramic midden achieving the final approximate dimensions of sixteen feet (4.9 m) above the ground surface and with a rough ovoid shape measuring nearly 100 feet (30.5 m) across the greatest dimension (Sears 1960). At least four ceramic era burials intruded into the mound, but absolutely no pottery was recovered from the mound itself below 18 inches (48 cm) of the surface. No artifacts dating to the ceramic‑producing periods were found within the mound. Ten shell tool artifacts were found, along with a single thought to be associated with the “old midden” rather than the burial (Sears 1960: 59). Sears immediately recognized this shell/sand structure as an intentionally constructed mound. He points out that the mound was likely built in a single intensive event, as there was no feathering of the strata at the ground surface, indicative of weathering, nor any indication of tramping of the strata. The final shell layer was clearly basket‑loaded. The combination of materials is what is thought to have provided the stability to the structure (Sears 1960: 58). All of Sears’ interpretations of the stratigraphy are reasonable and supported by the evidence. Then Sears assigns the mound to an unspecific “before” the ceramic St. Johns Ib or Weeden Island periods (1960: 59). Confident of his logic, Sears gave no possible explanation to this obvious violation of the no Archaic mounds 56 model, other than to claim “ambiguous results” and the statement: “This excavation produced some new problems, and contributed to the solution of none” (1960: 56). This mound is now accepted to date to the Mt. Taylor period (Wheeler et al. 2000). The next major excavation of a Mt. Taylor site was the salvage excavations at Tick Island (Harris Creek Mound) in the face of shell mining. Nearly contemporary with the mining of Bluffton was the removal of shell from Harris Creek. An avocational archaeologist conducted surface collections from the site during removal, delineating several areas (Bushnell 1960). Bullen received an emergency grant to conduct salvage work at the site. A posthumous report (Jahn and Bullen 1978) gave some details of the excavations, including general interpretations of the stratigraphy. This report encouraged speculation over the exact archaeological situation (Russo 1994a, 1994b) eventually convincing Lawrence Aten, who had worked with Bullen at the site, to publish a full report of the work (1999).

Ma t u r a t i o n For the next twenty years, archaeology at Mt. Taylor period sites in the St. Johns river basin was limited. Ripley Bullen conducted excavations at the Sunday Bluff site near the Ocklawaha, a tributary of the St. Johns (Bullen 1969). This site, a large shell deposit on a relict oxbow, included artifacts spanning the Mt. Taylor through St. Johns I periods, but did not include human interments. Mound excavations were limited to the field schools at Palmer‑Taylor mound by Rollins College, work has not been commonly reported prior to this dissertation. Interpretations of the Mt. Taylor remained fairly static during this time, though there are some cautious ventures regarding sedentism, village life, and possible mounds (Milanich and Faribanks 1980: 150). Sites with Orange components, but lacking clear Mt. Taylor deposits, excavated during this time include: Summer Haven (Bullen and Bullen 1961), Castle Windy, and the (Bullen and Sleight 1959, 1960). During this interim, there were several important salvage excavations of Middle and Early Archaic mortuary ponds and cemeteries. Dredging of wet areas throughout the southern half of the 57 Florida peninsula encountered human remains in several locations, including probably several that have never been recorded (Purdy 1991). Of note are the Republic Groves site in Hardee County (Wharton et al. 1981; Purdy 1991), the Bay West site in Collier County (Beriault et al. 1981; Purdy 1991), and two sites immediately east of the area of the upper St. Johns of concern in this thesis: the Gauthier site near Cocoa Beach (Carr 1981; Sigler‑Eisernberg 1984) and the initial discovery of the Windover site near Titusville (Doran 2002a). At least some of these sites are contemporary with Mt. Taylor, but I would not refer to these sites as Mt. Taylor. In 1984, Brenda Sigler‑Eisenberg lead a stratified‑systematic survey of a portion of the upper

St. Johns basin. This is a very well executed project, accomplished in a difficult topography with a talented staff. Using insights gained from work at the Gauthier site regarding subsistence strategy (overwhelmingly aquatic resource oriented), she postulated a linear trend to the catchment areas. A methodology was devised and executed that would attempt to test a hypothesis of linear settlement structure along the marsh/mesic hammock fringe. Stratified and judgmental transects were utilized, using local macro‑ecological categories for stratification, including the pine flatwoods, mesic hammock, marsh, periodically inundated prairie, and swamps. At 30‑meter intervals along the transects, 25‑50‑100cm test units were excavated and all materials screened through 1/4 inch screen (Sigler‑Eisenberg 1985b: 53‑56). This survey remained one of the only truly systematic surveys anywhere in the St. Johns valley until very recently. At least 11 sites with preceramic deposits were encountered during this survey (Russo 1985) Chronologically, the next important Mt. Taylor research is the excavations at Groves’ Orange Midden. Originally discovered in 1987, archaeological investigations led by Barbara Purdy into saturated deposits on the edge of Lake Monroe near the Old Enterprise site in the middle St. Johns basin provided the first opportunity revisit the concept and content of Mt. Taylor (Purdy 1994a). Revealing a fuller array of material culture than usually found in Middle Archaic contexts due to the preservation of organic artifacts, the results from this site transformed notions of Late Archaic subsistence (Russo et al. 1992; Wheeler and McGee 1994b) and provided opportunity to robustly describe the artifact assemblage (Wheeler and McGee 1994a, 1994c; Purdy 1994b). 58 Firstly, the Groves’ Orange Midden data from test pit excavations in 1989 indicated that subsistence models for the Late Archaic (Cumbaa 1976) that emphasized terrestrial game as the primary protein source were flawed. Assumption of poverty of potential of aquatic resources was not founded, and a model of supplemental use of coastal and riverine in a subsistence strategy that was semi‑nomadic and heavily dependent on terrestrial resources did not fit the data being recovered. Fish and shellfish instead were major components of the diet, and likely the primary resource base; botanical data indicated the gathering of mast and wild fruits and the use of Cucurbita pepo and bottle gourds (Russo et al. 1992). Further excavations at the site, including units in the completely saturated deposits offshore, revealed inundated two primary aceramic midden strata separated by a layer of peat, with localized sand, shell, and organic deposits. A generally stratigraphically consistent series of radiocarbon dates spanned the range from 6210 +/‑ 60 to 3160 +/‑ 65 uncalibrated radiocarbon years before present (McGee and Wheeler 1994). An impressive artifact assemblage was recovered, including numerous wooden (Wheeler and McGee 1994c), bone, dental, antler, and shell examples, as well as fired‑clay biconcial, spherical, and amorphous objects (Wheeler and McGee 1994a). Chipped stone tools were similar to contemporary contexts excavated elsewhere (Purdy 1994b), while a few fragments of non‑local stone artifacts indicate at least some limited exchange with other areas of the southeast United States (Wheeler and McGee 1994a). A more comprehensive listing of utilized botanical species indicates a wide spectrum of wild fruits, vines, seedy plants, trees, and herbs (Newsom 1994). Groves’ Orange Midden has substantially increased our knowledge of Mt. Taylor material culture and subsistence. In the early 1990’s, Michael Russo conducted survey near the mouth of the St. Johns river that began to overturn more of the assumptions of the standard paradigm of Mt. Taylor. Extant models asserted that there was an absence of coastal usage prior to the Orange period, due mainly to climatic conditions (e.g., Rouse 1951; Goggin 1952; Griffin and Miller 1978). Russo located five preceramic sites, along with another dozen with possible preceramic deposits (Russo 1992). This work, along with the reassessment of several other coastal sites (Ste. Claire 1990), clearly 59 demonstrated that the coast was extensively used during the preceramic.It was during this era that the idea that Archaic monuments were a reality coalesced into some whole that allowed it to enter the realm of discussion in Eastern Woodlands archaeology (e.g., Charles and Buikstra 1983; Hofman 1986; Claassen 1991a, 1991b; Russo 1991; Russo 1994b; Piatek 1994; Saunders et al. 1997) eventually spuring Lawrence Aten to revisit the 1961 excavations of Harris Creek (1999). Aten provides the first robust description of a Mt. Taylor burial mound, detailing the stratigraphy, burials, artifacts, and construction history. With the publication of Aten’s report, the structure of a Mt. Taylor burial monument was demonstrated, if not always accepted. Soon after the excavations at Groves’ Orange Midden, the acquisition of the extant remnants of two of the largest Mt. Taylor sites by the state of Florida, Bluffton and Mt. Taylor, provided opportunity for Ryan Wheeler and associates to present a thorough reassessment of Mt. Taylor as

an archaeological construct (Wheeler et al. 2000)2. The Wheeler et al. paper is a comprehensive overview of the state of knowledge of Mt. Taylor, and at present is the datum for the archaeology of the period. The authors review the chronology, settlement patterns, site arrangement and form, artifacts, subsistence, exchange networks, decorative arts, mortuary patterns, including the potential for intentional mounds, contemporary mortuary ponds, and possible origins of Mt. Taylor (2000: 142‑155). As would be expected, for some arenas, we know very little, but for others, particularly artifacts, a great deal is now known. Beginning in 2000, Kenneth Sassaman of the began a program of survey and excavation in the middle St. Johns executed through field schools (Sassaman 2003; Randall

2 Unfortunately, the authors refer to Mt. Taylor as a culture in the paper title, and as a period in most instances of the text. This conflation of terminology is not unique: Milanich also uses the two terms interchangeably and at- tributes the naming of Mt. Taylor culture to Goggin (Milanich 1994: 88). Weisman (2002: 157) traces the thrust of the revision to Milanich (1971, 1973) and his graduate students, but the effort has been primarily limited to chang- ing the names, not defining the concepts. Other authors have used Mt. Taylor as culture (Randall 2005: 18), and in a more egregious misapplication, Mt. Taylor is sometimes referred to as a phase by Johnson (2002: Table 1), and at other times as a culture (e.g., 2002: 17). Dickel and Doran (2002: 56) also refer to Mt. Taylor as a phase, citing Goggin 1952. I know of no published formal definitions of Mt. Taylor as a culture or a phase. While this may seem a trifling issue, archaeological constructs are not interchangeable. In the original definition Goggin (1948) refers to Mt. Taylor as a period: that construct is used in this dissertation. 60 and Sassaman 2005). Sassaman’s research interests soon led to the preceramic, and starting with limited work at Live Oak Mound in 2001, this timeframe has become a central focus of the UF program. Extremely important work on Hontoon Island is ongoing, and is already changing our ideas about the Mt. Taylor period. Several students of Sassaman have assisted with the field schools, conducted research, or are actively conducting research concerned with the Mt. Taylor period, including Asa Randall, Megan Blessing, Jon Endonino, and Peter Hallman. This is the modern era of research, with a conscious effort to shrug off the standard paradigm of Mt. Taylor culture and society and a willingness to engage the possibility of intentional construction of burial mounds, sedentism, emergent complexity, and built landscapes. 61 Ch a p t e r 5: Mt. Ta y l o r a s a Te m p o r a l Co n s t r u ct “Time doesn’t exist... there is only a perpetual present” Claude Chabrol, La Fleur du mal (2003)

This section redefines the preceramic Mt. Taylor period of the upper and middle St.Johns River basins in Florida and presents a new temporal construct termed the Thornhill Lake period. Significant differences in mortuary modes, artifact types, the presence of non-local stone, and radiocarbon dates warrant creation of a new period with the aim of providing greater temporal resolution to the archaeology of the Middle and Late Archaic (Mid-Holocene era) in the region. During this period of time we find the earliest evidence for monument construction, long-distance exchange, incipient agriculture, social complexity, and warfare (Anderson, Russo, and Sassaman 2007). Through the efforts of this paper, it is hoped we can gain a better understanding of the timing and conditions for these important phenomena. Each period is summarized with a general background followed by brief descriptions of the burial modes, grave inclusions, nature of burial strata, non-local materials, and shell artifacts for each. Following these summaries, an analysis of published radiocarbon dates is detailed and related to the period descriptions. Material details for the Mt. Taylor period follows Goggin=s original formulation (1948) while the two burial mounds at Thornhill Lake excavated by Moore (1894a; 1894b) serve as the type site for the Thornhill Lake period. Mound and non-mound contexts are utilized, with information coming from published excavation reports and the four upper St. Johns sites studied (Figure 1). I propose to restate the chronological span of the Mt. Taylor period and develop a temporal definition of the Thornhill Lake period, based in part on the calibration of the knownsetof radiocarbon dates. Specific criteria are found below, but time spans for each period are estimated as: Mt. Taylor: 7300-5600 cal 14C yr BP, Thornhill Lake: 5600-4500 cal 14C yr BP, Orange: 4500- 2500 cal 14C yr BP (Table 1).

Mt. Ta y l o r Pe r i o d Sp e c i f i c s John Goggin provided the original definition of the Mt. Taylor period (1948, 1949, 1952). The 62 period was marked by the presence of large shell heaps composed primarily of local shellfish, an absence of pottery, and occasional stone and bone tools. Populations were likely semi-sedentary, practicing a mixed subsistence strategy including hunting, fishing, and gathering. According to Goggin, domestic and ritual life were very simple, and mortuary practice was at the crudest level of development, with cannibalism likely practiced. Preceramic sites with bannerstones, stone beads, extended burials, and mounds were not part of the Mt. Taylor period, but were originally placed in an Unclassified Complex, (Goggin 1952: 51-53). It is now clear that these sites date to the Late Archaic, contemporary with other sites with bannerstones in the Eastern Woodlands (Jefferies 1996; Piatek 1994) and recently, the tendency has been to include all upper and middle St. Johns preceramic shell heap sites in the Mt. Taylor period, including those originally included in the Unclassified Complex (Wheeler et al. 2000; Mitchem 1999).

Burial Modes Aten=s report on the excavations at Harris Creek (1999) provides excellent documentation and discussion of Mt. Taylor mortuary modes, and this paper throughout is greatly indebted to his work. Interments are found in six basic configurations: flexed burials, broken and commingled bones, bundled burials, isolated crania, with or without longbones, isolated bones, and most rarely, extended. There may be burning of the bone with any of these configurations, but it does not appear that cremation is ever the intent. Flexed burials are the most common mode found: 98% of the burials recognized at Harris Creek are flexed (Aten 1999). Flexed burials are also present at Palmer-Taylor, Orange, and Persimmon Mounds. Isolated human bones seem liable to show up anywhere. Solitary crania, usually with some of the long bones, are found at Harris Creek, and may

Archaeological Period Calibrated 14C Date Range Mt. Taylor 7300-5600 BP Thornhill Lake 5600-4500 BP Orange 4500-2500 BP

Table 1: Chronology of the Mid-Holocene in the Upper St. Johns 63 be at other sites. Bundled burials are quite rare, two are known from Harris Creek, and one from Orange Mound. Broken and commingled human bones are found at Harris Creek, Palmer-Taylor and Persimmon II; evidence at Orange Mound is tenuous, and the group of eight burials directly on sand at Persimmon Mound is a community burial, if not commingled. A single extended burial was found at Harris Creek, and comes from the upper levels of the mounds. Extended burials were found in the uppermost burial stratum at Orange mound, but based on factors discussed below, this stratum likely falls within the Thornhill Lake period.

Grave Inclusions Clear Agrave goods@ are very rare in Mt. Taylor mortuary mounds. At Harris Creek, only 22 burials (12%) have possible artifact association, primarily simple bone pins (Figure 3). Some burials have multiple associations, and some associations include multiple artifacts. Five of these associations are certain, fourteen are considered possible, and there are nine instances of artifacts found near to burials (Aten 1999: Table 5). For four upper St. Johns mound sites studied, there are probably less than a dozen artifacts in clear Mt. Taylor contexts, and only a fraction of those were associated with burials. Of the artifacts at Harris Creek associated with burials, most were biface chipped stone tools, but bone beads, antler hair pieces, and bone pins were also found. Aten postulates that burials may have been wrapped in textiles or skins when placed in grave pits, held closed with the bone pins. Faunal remains are commonly found with burials. At Harris Creek, animal bone was found with several burials, as well as on a mound summit where it is interpreted as evidence of repeated feasting (Aten 1999: 180). Moore commonly reports animal bones with human remains in upper St. Johns sites. This includes a repeated association of faunal remains, fire places, and commingled human bone at Palmer-Taylor and Persimmon II, the lower burial stratum at Orange, and perhaps the mortuary stratum at Persimmon Mound. The association of fire places and burnt and broken human and animal bones at Palmer-Taylor and Persimmon II affirmed Moore=s opinion of no original intent of burial and that these were Acannibalistic feasts@ (1893b: 612). Though there is a perception that faunal remains (other than shell) are commonly found throughout Mt. 64 Taylor mortuary mound strata, that is not the case. Faunal remains are almost always localized and in association with burials.

Burial Strata Matrix Moore regularly refers to the matrix of Mt. Taylor sites as composed of the same material as the shell heaps, i.e., middens, and shell is the dominate matrix constituent in all burial-bearing strata of the Mt.

Figure 3: Bone Pins Taylor mortuary mounds discussed here, but rarely is it pure shell. At Palmer-Taylor, Orange, Harris Creek, and Persimmon II varying amounts of sand or soil was mixed with the shell. Large and small deposits of white sand are found at Persimmon Mound, Orange Mound, and Harris Creek. Other localized deposits may also be present. Hearths/ fire places/charcoal concentrations are commonly found, but are mainly either in direct association with grave pits or are restricted to a stratigraphic interface. In strata matrix, Pomacea sp. and Vivaparus sp., both freshwater snails, generally dominate the shell species, with occasional Unios. Excepting the presence of burials and white sand, there is little to distinguish the matrix of Mt. Taylor burial strata from midden deposits. However, there are substantial distinctions between these strata and those found at Thornhill Lake and discussed below. Full descriptions of stratigraphy are found in Chapter 8.

Non-local Stone In Mt. Taylor contexts, non-local stone, meaning stone from outside Florida, is very rare. This includes chipped stone and ground stone artifacts (Figure 4). Of the chipped stone artifacts at Groves= Orange Midden, only Florida stone, mostly from Tampa Bay area sources, was recovered (Purdy 1994). Aten (1999) does not report any extra-Florida chipped stone artifacts at Harris Creek, nor does Bullen report non-local chipped stone at Bluffton (1955). Groundstone artifacts of non-local material are also exceptional in Mt. Taylor contexts. 65 One steatite bannerstone fragment is reported from the sand layer underneath the lower shell midden stratum at Groves= Orange Midden (Wheeler and McGee 1994: 372- 373), but the published provenience information does not make it clear if the artifact is actually from under the shell or beyond the edge of the midden. Bullen reports a single steatite bannerstone fragment from an upper level of the preceramic strata at Bluffton (1955: Table 1), but this is from Thornhill Lake era deposits. No non-local groundstone is reported for Harris Creek, Palmer-Taylor, Orange, Persimmon Mound, or Persimmon II. Stone from outside Florida seems to be absent in Mt. Taylor contexts Figure 4: Archaic Stemmed Point, throughout the upper St. Johns. By contrast, non-local Buzzard’s Roost groundstone artifacts are relatively common in Thornhill Lake period contexts, described below.

Shell Artifacts Shell artifacts found in Mt. Taylor contexts are generally utilitarian, and most all are from Busycon shell. Edged and pounding tools dominate (Figure 5; Aten 1999: 159-160). It is the absence of two particular shell artifact types that help distinguish Mt. Taylor from Thornhill Lake period deposits: Strombus gigas celts and shell beads. S. gigas celts are found in preceramic contexts, but tend to be Figure 5: Busycon Shell Tools concentrated in the upper levels. A fuller 66 discussion of contexts producing S. gigas celts is below. There is no reference to S. gigas artifacts of any type from the Mt. Taylor deposits at Harris Creek (Aten 1999; Wheeler et al.2000 ), or from the upper St. Johns mortuary monuments reviewed. While less certain, it also appears that shell beads are absent from Mt. Taylor period strata. No shell beads were reported with burials at Harris Creek or any of the upper St. Johns mounds reviewed.

Ba c k g r o u n d t o t h e Th o r n h i l l La k e Pe r i o d Description of the Thornhill Lake period will be based primarily on the type site of the same name. The Thornhill Lake site complex (Figure 6) consists of three recognized components, a large sand burial mound (8Vo58), a small sand burial mound (8Vo59), and a shell heap (8Vo60). This site is located on Thornhill Lake, a small lagoon or slough on the east side of the St. Johns.

Depression/ Swampy

1892 Exc. 4 Shell Field 1892 Exc. 3 1892 Trench

1892 Exc. 2 1892 Exc. 5 Near End 1892 Exc. 1 of Crescent 1894 Trench Near Apex 7 Burials of Crescent Small Mound

1892 Trench 7 Burials 50 Yards Crescentic Ridge

Large Mound 1894 “Demolished” 42-47 Burials Depression

Thornhill Lake

Figure 6: Sketch Map of the Thornhill Lake Site (not to scale) 67 Mitchem (1999: 30) has assigned the two burial mounds to the Mt. Taylor period, as have Wheeler et al. (2000). Recently, John Endinino of the University of Florida has conducted excavations at Thornhill Lake. However, this data has not been published at the time of this writing, and derivation of Thornhill Lake period is independent of his research. As preceramic sites in the St. Johns basin go, the Thornhill Lake site complex qualifies as Aspectacular.@ At least 57 individuals were encountered between the two sand mounds, mostly extended burials. A relative abundance of Aprestige goods@ was recovered during excavation, including profuse shell beads, eight bannerstones, two pendants from bannerstone wings, and a large number of stone beads. In an area where non-local artifacts are generally exceptional, A. . . the mounds at Thornhill Lake stand alone@ (Moore 1894b: 170). Exceptional is often taken as enigmatic, and that has certainly been the case for the Thornhill Lake site. While the artifacts found were quite notable, it was what was absent in the mounds that caused the greatest disconcertment. ANot one particle of pottery plain or marked@ (Moore

FN1: 95)1was found below the surface, but the mounds were constructed primarily of sand, a trait thought to be associated with later pottery-producing periods. The possibility that the mounds truly predated ceramic technology was unacceptable to Moore, despite the obvious explanation:

AThis almost total absence of sherds and earthenware, fragmentary or otherwise, is entirely novel in our investigation of Indian mounds devoted to the purpose of sepulture, and it is evident that, with the makers of the mounds at Thornhill Lake, the custom of interring earthenware with the dead did not obtain B an unlooked for departure@ (Moore 1894b: 173). Moore would suggest, that contrary to the implications of his own work, he could not determine if the makers of the shell-heaps built sand burial mounds. Having reviewed Orange, Persimmon, and Thornhill Lake as sand mounds without pottery, he says this does not mean that they were built by the same folks as who built the aceramic shell-heaps. Instead, using the finding of two sherds from Thornhill Lake as his only evidence, two sherds he says are of accidental introduction, Moore

1 Field notebooks of C.B. Moore are referred to throughout by FN#, corresponding to the microfilm roll number originally assigned by Davis (1987) and how they are still indexed at Cornell University. 68 argues that the absence of earthenware is no proof that the builders did not posses the technology (1894b: 210). Sixty years later, the Thornhill Lake site remained problematic. John Goggin placed the site into an Unclassifed Complex, along with other sites producing bannerstones: Mt. Oswald (Tomoka

Mound) Shields Mound, Coontie Island, the surface of Bluffton2 (of Strombus shell), Dillon=s Grove, Salt Run Midden 2, and Stokes Landing midden. Bannerstones were also known from elsewhere in the state, including the Indian River area , the Manatee Region, the Glades area, central Florida, and the central Gulf Coast area. Goggin knew that bannerstones and stone beads dated from preceramic times throughout the Eastern Woodlands including the Tennessee Valley, Green River, and . However, the Thornhill Lake mounds were without a doubt sand burial mounds, and sand burial mounds were only supposed to be post-Archaic, forcing Goggin into the position of placing the site in an unclassifed complex with a likely temporal placement in the St. Johns I period or later. An explanation of heirlooming or imitation of earlier practice was suggested. The issue seems to have nagged Goggin however, and he did not think it settled: AThis represents one of the most interesting problems of the region@ (Goggin 1952: 51-53). Beginning in 1992, Bruce Piatek commenced limited excavations at the Tomoka Mound

Complex, from where Douglass had recovered eight bannerstones in the 19th century. Working in Mound 6 Piatek encountered strata composed of sand and shell. An uncalibrated radiocarbon date of 4060 +/- 70 on coquina shell from the base of the mound indicated an Archaic construction, an interpretation congruent with the bannerstones and absence of pottery at the site (Piatek 1994). The Tomoka Mound date is the only radiocarbon date in the St. Johns region in a context that produced bannerstones. However, a recent project at the Lake Monroe Outlet Midden (8V053) produced artifacts that align the site closely with Thornhill Lake and other preceramic sites with lapidary items. This site, approximately 15 km down river from the Thornhill Lake site, was subjected to salvage operations prior to the expansion of a bridge over the St. Johns. Excavations

2 Mistakenly called Mt. Taylor by Goggin (1952: 52). 69 included three large test units and eleven trenches. Within Unit C, an assemblage was recovered that included 266 disc shell beads and a tubular stone bead. In Unit B, well over 100 microliths suitable for shell bead production were found. A few tubular shell beads were also found (Johnson 2001: Tables 5.7, 7.3, Figure 5.21). These artifacts recall the findings at the Thornhill Lake site of profuse shell beads and stone beads. Radiocarbon dates from the site cluster in the 4600-4800 uncalibrated radiocarbon years range (Johnson 2001: 9.1). It seems reasonable to cross-date the Thornhill Lake site with the Lake Monroe Outlet Midden. A tubular stone bead and radiocarbon dates from the upper midden at Groves= Orange Midden (Wheeler and McGee 1994b; McGee and Wheeler 1994) correlates with the data from Lake Monroe and is assigned to the Thornhill Lake period.

Th o r n h i l l La k e Pe r i o d Sp e c i f i c s This section is based on Moore= published descriptions of his excavations at Thornhill Lake (1894a: 88-89; Moore 1894b: 167-173) and his unpublished field notes from the site (FN 1: 88-96; FN2: 62-67; FN5: 71-76; FN6: 73-76). Moore excavated several units into the shell deposits that are not directly informative to the issue at hand, but deserve further attention. The focus here is on his findings from the two burial mounds (Figure 6). Moore excavated twice at the Thornhill Lake site, in 1892 and again in 1894. He described the site initially as a shell field with a crescent shaped shell ridge with the points toward the water, but later scratched through the crescent shaped shell ridge mentions, perhaps thinking that the description implied too much intentionality. At the center and apex of the ridge was the smaller of the two sand mounds. He describes it as 8 feet, 10 2 inches (2.7 m) above the ground surface, but this includes the underlying shell deposit. The mound measured 295 feet (90 m) in circumference (28.6 m diameter). A trench was excavated on the west side of this smaller mound in 1892, measuring 22 feet (6.7 m) by 7 feet (2.1 m) and 3.5 feet (1.1 m) deep to the underlying shell ridge. In this trench was found an alligator tooth with cut marks, a quartz pebble, and a single flint flake. No bones of any sort were found, nor any pottery. The lack of bone in this test spurred Moore to 70 strike out all references to this small mound as a burial mound in his notes. Moore=s description of this work is in contrast to the opinion of Mitchem that he did not dig into the smaller mound in 1892 (1999: 27). Upon Moore=s return to the site in 1894, he conducted additional trenches along the fringes of the smaller mound, again finding nothing. A trench into the central portion of the smaller mound however encountered 7 extended burials. On the chest of three of these burials (1894b: 170), Moore describes the finding of shell beads and Arude ceremonial implements@ (FN5: 73) of soft stone, much softer than the bannerstones from the large mound described shortly. The larger burial mound is described as 11 feet (3.4 m) in height, 425 feet (130 m) in circumference (41.23 m diameter). Moore locates the mound between the horns of the crescentic ridge, approximately 50 yards (45.8 m) from the small mound and upon the shell field. There was a marked depression south of the mound (towards the water) that may be the source of the mound material. Moore conducted several excavations into the large mound in 1892, beginning with a large unit in the apex of the mound and multiple trenches. A total of seven burials were encountered during this 1892 operation, including two that were flexed, two that were positively extended, and three that were likely extended. At least some of the extended burials had the hands folded across the abdomen. He also found an isolated burnt human humerus fragment. With one of these burials were found shell beads and a chipped stone biface. Returning to the site in 1894, Moore completely excavated the larger mound, encountering at least 42 additional burials, including a few with extensive grave furniture discussed further below. It appears that all the burials encountered in 1894 were extended.

Burial Modes In contrast to the rarity of extended burials in Mt. Taylor mortuary mounds, the dominate interment configuration at Thornhill Lake is extended burial. Of the approximately 57 burials found at the site, only two are reported as flexed. These were found relatively shallow, and may be intrusive from much later (cf. Sears 1960). A preponderance of extended burials is in stark contrast 71 to the situation at Mt. Taylor mortuary monuments, where only one extended burial is reported out of the hundreds known. I place the upper burial stratum at Orange Mound in the Thornhill Lake period, based in part on the presence of only extended burials in this stratum. No lapidary artifacts were recovered by Moore from the burials at Orange Mound, but his excavations were limited, especially compared to his near demolition of the mounds at Thornhill Lake. Fortunately, Orange Mound still exists, so this can be tested in the future. With trepidation, I would place the Bluffton Burial Mound 2 in this period, based in part on the presence of an extended burial, but also due to other specific similarities with Thornhill Lake, namely: burning of the underlying shell, mound strata of sand and shell (including redeposited midden in both cases), and the burial being found in a sand matrix (Sears 1960).

Grave Inclusions At Thornhill Lake, one of the burials in the larger mound was found with an enormous amount of material, including large circular shell beads (Figure 7), tubular shell beads, Agreat quantities@ of small shell beads, and 19 red stone beads. At the wrists of this individual were found miniature bannerstones. Upon the chest was a winged bannerstone. Near the neck, Awith beads in great profusion@ (1894c: 168) was found a decaying stone gorget and a pendant shaped from the wing of a bannerstone (Figure 8). Also found with this burial was a perforated shark tooth and fragments of charcoal. With another burial was found a few shell beads and another pendant shaped from the wing of a bannerstone. Several other burials had stone beads, others were found with just shell beads, a few described as tubular. There are marked difference between the artifacts associated with burials in Mt. Taylor period contexts and Thornhill Lake period contexts. Not all burials have grave accouterments, but many do. One burial had a tremendous amount of material, Figure 7: Shell Beads others have abundant material, some have only a 72 little, and some have no additional artifacts buried with them. While the presence of funerary objects is one of the main features that serve to distinguish Thornhill Lake from Mt. Taylor period burials, it is likely that most Thornhill Lake period burials do not have associated elaborate artifacts. If the upper burial stratum at Orange Mound and the Bluffton burial mound excavated by Sears do fall in the Thornhill Lake period, no grave inclusions were found at either location. There are however a fair number of bannerstones and stone beads that have been found in the St. Johns region, and other Figure 8: Bannerstones, Thornhill Lake examples of burials with Asumptuous@ mortuary furniture may still be found.

Burial Strata Matrix The smaller mound at Thornhill Lake is described as being composed of brown sand and shell heap material. It seems that these are alternating strata of unknown number, but there are some frustratingly difficult ambiguities with Moore=s stratigraphic descriptions from this site, in contrast to most of his work. It is clear though that neither mound was of pure sand. In the larger mound, strata of white sand, brown sand and shell, and shell are described. At the central base of the mound was a calcined shell/charcoal surface, apparently across the whole mound base, and charcoal was found throughout the mound fill (FN 6: 73-76). Brown sand with shell, white sand, and shell strata were present in the mound; the burial with the multiple bannerstones is in shell with white sand above and below. A greater abundance of brown sand within the burial-bearing strata appears to contrast Thornhill Lake period mortuary monuments with Mt. Taylor period mortuary monuments (Figure 9). White sand continues to be an element of the monument building program (and remains a component of the 73

Figure 9: Bluffton Burial Mound Section

St. Johns mound building tradition through the historic period), but brown sand is found in greater abundance. Thornhill Lake period mounds also are more likely to be of a composite character, with alternating strata of sand with a little shell, redeposited shell midden, and occasionally local clays. During both periods, a large fire is typically built atop an extant shell midden surface immediately prior to mound construction. In the Mt. Taylor period, commingled human bone may be piled on this fire while burning. At Thornhill Lake period sites, human remains are not placed until the fire is extinguished, but there was limited burning on a burial at the Thornhill Lake site.

Non-local Stone There is a substantially greater presence of stone from outside Florida in Thornhill Lake period contexts than Mt. Taylor period contexts. With the exception of the steatite bannerstone fragment from beneath the lower midden stratum at Groves= Orange Midden, there does not appear to be any lapidary items of non-local lithic sources in any of the Mt. Taylor levels of sites reviewed, including steatite vessel fragments. The material types of the bannerstones found at the Thornhill Lake site is not given in most instances. One gorget is described as limestone, and two of the stone artifacts found in the smaller burial mound are described as being of phosphate rock, the other referred to as serpentine (Figure 8). The other bannerstones and pendants are not described as to material, but are made from a hard stone not native to the state. The stone beads are described as catlinite by Moore, and jasper by Goggin (Figure 10). In either case, they are certainly non-local (Moore 1894b: 169-170; Goggin 74 1952: 52). A stone bead found in the upper midden stratum at Groves= Orange Midden is made of steatite (Wheeler and McGee 1994), and a stone bead found at the Lake Monroe Outlet Midden is from a non-local stone, perhaps claystone or shale (Johnson 2001). Stone beads and bannerstones from Coontie Island are also of non-local Figure 10: Stone Beads material (Wheeler et al. 2000: 151- 152).

Shell Artifacts Wheeler et al. (2000: 151-152) have noted the absence of S. gigas and groundstone artifacts at Harris Creek, and think that there may be a temporal component to this pattern, as well as being an indicator of participation in larger areal exchange networks that included southeast Florida (Figure 11). I heartily agree, and assert that these artifact types only appear during the Thornhill Lake period. Strombus artifacts are consistently restricted to the upper levels of preceramic deposits in the St. Johns. Of relevance, a bannerstone crafted of a S. gigas shell lip was recovered at the Bluffton site (Wheeler et al. 2000: 152; Moore 1898). Figure 11: Strombus Celts 75 At Bluffton, S. gigas celts are found only in Layer III (Bullen 1955: Table 1). At Groves= Orange Midden (Wheeler and McGee 1994: Table 4), the artifacts are only found in the upper midden stratum and above (Thornhill Lake-Orange periods). Two celts were recovered at the Lake Monroe Outlet Midden (Johnson 2001: Table 7.3) in deposits that date to the Thornhill Lake period. At the Hill Cottage Midden, on the west coast of Florida, Bullen recovered a S. gigas celt in strata postdating Mt. Taylor, but based on calibrated radiocarbon dates, contemporary with Thornhill Lake. Shell beads are also be of temporal importance. At the Thornhill Lake site, shell beads were the most common artifact found with burials. In Mt. Taylor burials, shell beads are not typically found. One possible shell bead from Palmer-Taylor (Stewart 1976) is the only example I am aware of, and it was not directly associated with a grave. No shell beads are reported at Harris Creek (Aten 1999). Shell Figure 12: Microliths bead production appears to have been undertaken at the Lake Orange Outlet Midden, with hundreds of shell beads found in one unit, and well over one hundred microliths found in another (Figure 12; Johnson 2001). Tubular and disc forms are found, but shell species for either is not yet determined. Shell beads in preceramic contexts are consistently found exclusively in Thornhill Lake period deposits.

Ra d i o c a r bo n Ch r o n o l o g y Current definitions date the Mt. Taylor period from between 6000 BP to 4000 BP (Wheeler et al. 2000: 143). As originally stated, Mt. Taylor included everything prior to appearance of pottery in the northern St. Johns region (Goggin 1952: 40-41), but continued research has refined the chronology, with Paleoindian, Early Archaic, Middle Archaic, and Late Archaic periods described (Milanich and Fairbanks 1980; Milanich 1994). As previously defined, the Mt. Taylor period ends 76 with the appearance of fiber-tempered ceramics, the type artifact of the Orange period. Sassaman has placed the earliest use of fiber-tempered pottery in the southeast at approximately 4500 BP (1993: 22).

Methodology In order to test the hypothesis that the material differences in the archaeological deposits described above have chronological implications, I assembled radiocarbon assays from contexts representative of the proposed temporal constructs. These dates were then calibrated and seriated based on the oldest value at the 2-sigma range. Once seriated, the ordering of the sites was reviewed for stratigraphic integrity and compared to the archaeological contexts from which they originated. A robust correlation between radiocarbon sequence, stratigraphic sequence, and material culture as described is interpreted as supporting the hypothesis. I have attempted to use only dates from secure contexts in the St. Johns basin, with one site from a coastal lagoon in the St. Johns region: the radiocarbon date on coquina shell from the Tomoka Mound site complex in northeast Florida of ca. 4600 BP (Piatek 1994: 115) that has become the datum for the end of the Mt. Taylor period (Table 2). A set of four dates from Harris Creek and five samples from the lower levels of Groves Orange Midden have served to reinforce the 6000 BP date for the inception of Mt. Taylor (Wheeler et al. 2000; Jahn and Bullen 1978; Aten 1999; McGee and Wheeler 1994). The other assays used in this analysis include: the eight dates from the upper levels at Groves= Orange Midden, one date on charcoal from a fire place in Ocala Forest Midden 2 (Bullen and Bryant 1965: Table 4), a set of six dates from a recent salvage archaeology project on Lake Monroe (Johnson 2001: Table 9-1), one date from Live Oak Mound (Sassaman 2003: Table 4-1), one date from the Hontoon Dead Creek Mound (Randall and Sassaman 2005: Appendix A), and one previously unpublished date from Palmer-Taylor (Stewart 1983; Calvert et al. 1978: 281). Also included are two dates from the fiber-tempered pottery bearing levels at Groves= Orange Midden (Russo et al. 1992) and one date from a fiber-tempered context at Bluffton (Bullen 1972). Dates from freshwater shell are not included in this analysis. A total of 28 radiocarbon dates are from preceramic contexts in the upper St. Johns (n=1), the middle St. Johns (n=26), and northeast 77 Florida (n=1). Eight of the dates are from mortuary mound contexts, the remainder are from midden contexts. Five of these dates are from marine shell, the remainder are from charcoal, nutshell, or wood. A total of three dates from Orange period contexts, all within the middle St. Johns, are included for comparison and bracketing; one of these dates is from shell and the other two are from charcoal; all come from likely midden contexts (Table 2). Dates from the St. Marys region are not included, but would be contemporary with the Thornhill Lake period (Russo 1992). There has been no attempt at prejudice when selecting radiocarbon dates for this analysis. I have tried to use every date from clear preceramic contexts available in the published literature for the middle and upper St. Johns. All dates were entered into a spreadsheet and formatted for analysis and calibrated with OxCal,

version 3.10 (Ramsey 2005)3. Each date was calibrated as specified, and the resultant 1 and 2-sigma ranges copied to a spreadsheet. The samples were then seriated by the earliest date of the calibrated 2-sigma range of each sample. Results were then arranged as a chart and presented as probability spectrums in descending order (Figure 13). Dates were calibrated and then sorted; no dates were dropped and no specific arrangement was attempted.

Results Comparison between groupings produced by the radiocarbon seriation process and the material traits of the Mt. Taylor and Thornhill Lake periods shows a very strong correspondence. Mt. Taylor sites, characterized by flexed burials in strata dominated by shell, very few non-local materials, few artifacts found with burials, and an absence of Strombus gigas artifacts and shell beads date consistently to the 7300-5600 BP range. Sites with bannerstones, non-local stone beads, and shell beads are all found to date from 5600-4500 BP. Combined with extended burials, abundant grave accouterments with some burials, and burial strata dominated by sand, and we have strong contrasts between the two periods (Table 3). Orange period specifics are not directly researched in 3 For dates on wood/charcoal, the IntCal04 curve and a δ13C per mil value of -25 +/- 2 is used (Reimer et al. 2004). Marine shell is calibrated with the Marine04 curve (Hughen et al. 2004), using a ΔR correction factor of -5 +/- 20, and adjusted for isotopic fractionation by the addi- tion of 390 years prior to calibration (Walker et al. 1994: Table 2). Sample names correspond to those in the Sample Code column of Table 2. Radiocarbon samples analyzed from decades ago have relatively large associated errors, but for most samples, the error is on the order of 100 years or less. 78

Lab Sample 14C BP 14C Age SD d13C d13C SD Delta R Delta R SD Marine Carbon Description Material code code (CRA) years per mil years years years percentage UNKN GOM_OR_1 3159 65 -25 2 0 0 0 ZII_L2 CHARCOAL UNKN GOM_OR_2 3851 70 -25 2 0 0 0 ZII_L3 CHARCOAL RL-32 BLT_RL32 4050 110 0 0 -5 20 100 BUSYCON SHELL BETA_65865 GOM_7 4080 60 -25 2 0 0 0 ZII_L7 CHARCOAL UNKN_1 GOM_1 4115 75 -25 2 0 0 0 ZII_L5 CHARCOAL BETA_65864 GOM_8 4190 60 -25 2 0 0 0 ZII_L13 CHARCOAL UNKN_2 GOM_2 4400 125 -25 2 0 0 0 ZII_L8 CHARCOAL BETA_54622 TMKA_MD_6 4850 70 0 0 -5 20 100 MIDDEN_BASE_MD_6 COQUINA BETA_146752 LMOM_UA_L19 4640 50 -25 2 0 0 0 CHARCOAL BETA_146748 LMOM_UA_L9 4650 110 -25 2 0 0 0 CHARCOAL BETA_146751 LMOM_UA_L14 4710 40 -25 2 0 0 0 CHARCOAL I_563 OF_MID_2_33INBD 4725 180 -25 2 0 0 0 CHARCOAL BETA_146753 LMOM_UC_L9 4760 40 -25 2 0 0 0 CHARCOAL BETA_146756 LMOM_UC_L15 4760 40 -25 2 0 0 0 CHARCOAL BETA_65859 GOM_9 4810 70 -25 2 0 0 0 ZIV_L10 CHARCOAL BETA_65860 GOM_10 4930 80 -25 2 0 0 0 ZIV_L14 CHARCOAL BETA_65862 GOM_11 5040 60 -25 2 0 0 0 ZIV_L16 CHARCOAL BETA_146749 LMOM_UA_L13 5080 80 -25 2 0 0 0 CHARCOAL BETA_59801 GOM_3 5100 70 -25 2 0 0 0 ZIV_L11 CHARCOAL BETA_59802 GOM_4 5160 80 -25 2 0 0 0 ZIV_L14 CHARCOAL M-1270 HRS_CRK_7 5420 200 0 0 -5 20 100 ABOVE MORTUARY ZONES BUSYCON SHELL M-1265 HRS_CRK_2 5320 200 -25 2 0 0 0 ABOVE MORTUARY ZONES CHARCOAL M-1264 HRS_CRK_1 5450 300 -25 2 0 0 0 BOTTOM OF MORTUARY A CHARCOAL M-1268 HRS_CRK_5 5450 180 -25 2 0 0 0 BOTTOM OF BURIAL 12 CHARCOAL BETA_59803 GOM_5 5580 80 -25 2 0 0 0 ZIV_L19 CHARCOAL BETA_59804 GOM_6 5930 80 -25 2 0 0 0 ZIV_L21 CHARCOAL BETA_202281 HNTN_D_CRK_MD 6070 70 -25 2 0 0 0 AUGER_1 HICKORY_NUT UM-1155 PLMR_TYLR 6450 105 0 0 -5 20 100 ABOVE MORTUARY ZONES BUSYCON SHELL BETA_65863 GOM_12 6200 70 -25 2 0 0 0 ZV_L20 CHARCOAL BETA_65861 GOM_13 6210 60 -25 2 0 0 0 ZV_L22 CHARCOAL BETA_??? LIVE_OK_MD_2000 6260 50 -25 2 0 0 0 STRATUM_XIII CHARCOAL

Key: GOM= Groves’ Orange Midden BLT= Bluffton Midden TMKA_MD= Tomoka Mound OF= Ocala Forest Midden LMOM= Lake Monroe Outlet Midden HRS_CRK= Harris Creek HNTN_D_CRK= Hontoon Dead Creek Mound PLMR_TYLR= Palmer Taylor Mound LIVE_OK_MD= Live Oak Mound Table 2: Selected Radiocarbon Dates from Mt. Taylor, Thornhill Lake, and Orange Period Contexts

this work, but a date of 4500 BP for the ending of the Thornhill Lake period and the beginning of the manufacture of fiber-tempered pottery is supported by these data. A vertical line was visually plotted near the upper and lower 2-sigma point on the probability spectrum of the oldest and youngest date within each postulated group (Mt. Taylor and Thornhill Lake) and another line plotted at 2500 cal 14C yr BP based on Sassaman=s estimate (1993) for the approximate end of fiber-tempering as a pottery tradition in the southeast. Given a little leeway of placement of the lines to intersect even 100 year intervals, the time spans derived are: Mt. Taylor

Period: 7300-5600 cal 14C yr BP, Thornhill Lake: 5600-4500 cal 14C yr BP, and Orange: 4500-2500 cal 14C yr BP. Contexts assigned to the Mt. Taylor period based on this process include: Live Oak Mound, Palmer-Taylor Mound, Hontoon Dead Creek Mound, Harris Creek Mound, and the lower shell midden stratum at Groves= Orange Midden. One date from the Lake Monroe Outlet Midden falls 79 Mt. Taylor Thornhill Orange Period Lake Period Period GOM_OR_1 3159±65BP BLT_RL32 4050±110BP GOM_OR_2 3851±70BP

Orange Period GOM_7 4080±60BP GOM_1 4115±75BP GOM_8 4190±60BP GOM_2 4400±125BP TMKA_MD_6 4850±70BP LMOM_UA_L19 4640±50BP LMOM_UA_L9 4650±110BP LMOM_UA_L14 4710±40BP OF_MID_2_33INBD 4725±180BP Thornhill Lake Period LMOM_UC_L9 4760±40BP LMOM_UC_L15 4760±40BP GOM_9 4810±70BP GOM_10 4930±80BP GOM_11 5040±60BP LMOM_UA_L13 5080±80BP GOM_3 5100±70BP GOM_4 5160±80BP HRS_CRK_7 5420±200BP HRS_CRK_2 5320±200BP HRS_CRK_1 5450±300BP HRS_CRK_5 5450±180BP GOM_5 5580±80BP

Mt. Taylor Period Mt. Taylor GOM_6 5930±80BP HNTN_D_CRK_MD 6070±70BP PLMR_TYLR 6450±105BP GOM_12 6200±70BP GOM_13 6210±60BP LIVE_OK_MD_2000 6260±50BP

8000CalBP 7000CalBP 6000CalBP 5000CalBP 4000CalBP 3000CalBP 2000CalBP

Figure 13: Seriation of Radiocarbon Dates within this group, but this date is likely older than its recovered context. There are two later dates from levels of the same excavation unit that are inferior to the level from which this date originates (Table 2). Contexts assigned to the Thornhill Lake period include: Lake Monroe Outlet Midden, Ocala Forest Midden 2, Tomoka Mound 6 base, and the bulk of the upper shell midden stratum 80

Archaeological Flexed Extended Commingled Non-local Strombus gigas Shell Shell Dominate Sand Dominate Premound White Sand Period Burial Burial Human Bone Lapidary Items artifacts Beads Burial Strata Burial Strata Fire Deposits

Mount Taylor PR P A A A P A P P Thornhill Lake RP A(?) P P P A P P P A=Absent P=Present R=Rare

Table 3: Comparisons between Mt. Taylor and Thornhill Lake Periods at Groves= Orange Midden. Orange period contexts include: the uppermost undisturbed level at Groves= Orange Midden and the fiber-tempered pottery strata of Bluffton.

Di s c u ss i o n a n d Co n c l u s i o n s This analysis has found that there is a strong correlation between particular material features of the late preceramic in the St. Johns region and calibrated radiocarbon dates. Distinctions between preceramic sites that have all previously been grouped into the Mt. Taylor period has allowed for a greater temporal discrimination in the later preceramic, spurring the creation of a new archaeological period, the Thornhill Lake period, and restriction of the Mt. Taylor period in time and content. There is direct implication from this analysis for Middle-Late Archaic preceramic contexts throughout the Eastern Woodlands. Evidence from the St. Johns region in Florida indicates that the construction of burial mounds likely predates the participation of St. Johns people in intra- regional exchange networks, including non-local stone artifacts and marine shell from southeast Florida, by over 1500 years. For example, comparison with the Hill Cottage Midden at the Palmer site near Sarasota Bay on the west coast of Florida (Bullen 1972) finds parallel trends, and further investigation would likely show that this trajectory is relatively common. Populations are also expanding into coastal ecological niches during the Thornhill Lake period (Russo 1992). Calibration and analysis of radiocarbon dates and comparison with artifact assemblages and mortuary practices in other areas of the southeastern United States is expected to show similar trends. At the same time, there are important continuities between Mt. Taylor and Thornhill Lake period practices. Many Mt. Taylor midden sites continued to be occupied during the Thornhill Lake 81 period, with little indication of dramatic shifts in subsistence strategies or toolkits. Mt. Taylor and Thornhill Lake mortuary mounds are both typically built on extant aceramic middens, and the mounds share specific structural attributes. For at least one site considered here, Orange Mound, there are Mt. Taylor and Thornhill Lake period mortuary strata. Persistence of a basic pattern is the essence of Goggin=s tradition concept, and such continuity is found between these two periods. The substantial peat layer at Groves= Orange Midden separating the Mt. Taylor and Thornhill Lake deposits could be an indicator of a climatic event that dates to the transition. McGee and Wheeler (1994) point out that this peat layer is indicative of a high water period in Lake Monroe. Climatic data for the southeastern region are discussed in detail by Anderson, Russo, and Sassaman (2007), and deserve further attention in Florida. In this chapter I have attempted to provide a greater resolution to Mt. Taylor as an archaeological construct. Using material traits and radiocarbon dates, Mt. Taylor is restricted to the time span of 7300-5600 BP, and argued to be characterized in part by flexed burial and a general absence of grave furniture and non-local artifacts. Based on the evidence given, the tradition of mortuary monuments in the St. Johns region predates the appearance of non-local materials, differential burial treatments, and adornments, artifacts and traits that would be indicative of rank and hierarchy, by well over a millennium. Findings presented here should encourage us to reassess our assumptions regarding mound building, hierarchy, and power relationships (Saunders 2004). Perhaps, respect of and obligation to ancestors is Aenough@ to encourage mound building (Gibson 2004). 82 Ch a p t e r 6: En v i r o n m e n t a n d Su b s i s t e n c e Give a man a fish. . . Anonymous

This section reviews the general ecological conditions during the mid-Holocene in the upper St. Johns basin and provides a rough outline of the subsistence economy during the Mt. Taylor period (7300-5600 cal 14C yr BP). During this period, the low relief of the Florida peninsula and post-Pleistocene sea-level rise combine in a progressively drowned landscape. Highly productive local environments, composed of a mosaic of aquatic and terrestrial resources, were established and matured. Exploitative strategies focused on fishing, gathering, and hunting. Bountiful natural resources resistant to over-exploitation encouraged a reduction in mobility. Persistent usage of certain locales over generations resulted in the accumulations of large shell middens. Mortuary monuments were constructed, some of the first such structures in the southeastern United States.

Ge n e r a l Ch a r a c t e r i s t i c s Three factors most shape the environment of peninsular Florida and the St. Johns basin: water, youth, and very low relief. The St. Johns is a slow-moving, northward flowing river, one of the few major rivers in North America to do so. The river heads near Blue Cypress Lake west of Vero Beach and flows generally northward for approximately 275 miles (443 km) to the mouth near Jacksonville (Figure 1). An extremely low-gradient characterizes the entire valley, with only ca. 15 feet (4.6 m) of drop for the entire river length, or ca. .005 feet (15 mm) per mile. The lower (northern) portions of the river are daily affected by tides, and tidal influences can be felt as far as the upper St. Johns (Cooke 1945; Brown et al. 1962; White 1970).

Geology In the upper St. Johns basin, most surface is later-Pleistocene to Holocene in age, with earlier geological units generally only exposed in sinkholes or springs, neither of which is commonly exposed surficialy in the region, though they are likely buried (Kindenger et al. 2000). Underlying Eocene limestones form most of the Floridian aquifer. Miocene-Pliocene deposits overlay these limestones, forming part of the Floridian aquifer and creating an aquiclude that separates the 83 nonartesian aquifer (Bermes et al. 1963). Surface geology is primarily siliciclastic sediments, derived from Appalachian sources, deposited and altered by repeated sea-level high and low stands (Scott 2001: 1). Complexes of late Pleistocene and Holocene sands, muds, and organics form the basin (Scott 2001), with likely buried Hawthorne Formation impermeables­ affecting expressions of surface water (Kindenger et al. 2000). Stone suitable for either flaking or grinding is mostly absent throughout the upper St. Johns. Sources for flaked stone artifacts are almost always cherts from north-central Florida or cherts and silicified corals from Tampa area (Purdy 1981). Major terraces formed during higher sea-levels border the Eastern Valley physiographic feature that forms the upper St. Johns. Interpretations (Miller 1998; Cooke 1945: 301) of the origin of the broad, low valley that the St. Johns flows through in this area are primarily that the valley represents a relic lagoon formed during Pamlico (interglaical recession in Wisconsin glacial), but White (1970) gives an alternate origin. The general model is that during a late Pleistocene sea- level high stand, a large coastal lagoon, similar to the present Indian River lagoon, formed between what is today the Osceola Plain and the Atlantic Coastal Ridge. Lowered sea-levels stranded this lagoon, while increasing precipitation and again rising sea-levels eventually created sufficient volume to flow. White (1970) points out that this section of the St. Johns is formed of connected lakes, and may have a more complex formational history. An additional terrace, the Silver Bluff, is found coastward of the upper St. Johns and near the present coast. The age of this terrace could be very young, almost certainly post-Pleistocene in origin, and possibly as recent as 6000-7000 years ago. This scarp would represent an additional sea-level high stand (Cooke 1945; Brown et al. 1962). Recent sea-level research, discussed shortly, supports the thesis of several low violence Holocene sea-level high stands. Formation of post-Pleistocene terraces and barrier islands on the nearby Atlantic coast may better explain the seeming abrupt emergence of mid-Holocene settlement (Brech 2004).

Soils Present soils in the upper St. Johns are dominantly mucks and sands, with little area considered suitable for agriculture without drainage. Several small attempts at cultivation are evident in the 84 region, mainly orange orchards, but large-scale agriculture has not been commonly practiced. Extensive peat deposits are present from the area near Lake Wilmington and southward, with localized peats found in the remainder of the basin, including near the archaeological sites under consideration and in ponds and lakes. The peats are mostly autochthonous, formed due to eutrophication in lowmoor environments, and some deposits are closer to muck than peat in character (Davis 1946). Peat/muck deposits at Groves= Orange Midden and Palmer-Taylor indicate a high-level stand postdating the Mt. Taylor period (Aten 1999; Rouse 1951). These peats are highly informative regarding past ecological and climatic conditions at a local level.

Sea-levels The importance of sea-level changes in the St. Johns basin has been understood by archaeologists since the work of C. B. Moore, and formally discussed by Goggin (1948b) and Rouse (1951). There are a multitude of sea-level reconstructions at global and regional scales with Holocene data from which curves and trends are derived. Specifics of each reconstruction can be contradictory with others, but general trends are usually similar. Balsillie and Donoghue (2004) have analyzed and summarized sea-level curves pertinent to the Gulf of Mexico and Florida. The researchers assembled the relevant sea-level information, and then attempted to eliminate spurious data points, resulting in 341 radiocarbon dated sea-level indicators dating from ca. 18,000 years BP to present. A seven-point floating average was established and a curve produced from this statistic. The sea- level data are presented using uncalibrated and calibrated dates. The calibrated data are used for comparison in this dissertation. Particular attention was paid by Balsillie and Donoghue to the post- 6000 years BP period, i.e., the mid- to late-Holocene. Review of the sea-level curve based on the onshore samples demonstrates a gradually aggrading water level, reaching modern levels at approximately 5800 cal 14C yr BP (Figure 14). This gradual rising of the sea-level is generally congruent with current models of mid- Holocene conditions (Miller 1992, 1998), though earlier than thought, with near-modern conditions appearing nearly contemporary with the initiation of shell heap construction and the Mt. Taylor period. Additional research supports a mid-Holocene sea-level rise (Faught 2002). However, the 85 Balsillie and Donoghue curve shows the sea-level at the beginning of the Mt. Taylor period much lower than traditional reconstructions, as much as 7 meters below modern mean sea-level. Without more regionally specific data, it is difficult to be sure of the local conditions in the upper St. Johns at this time. Water levels have at least gotten to the point where local environments can support shellfish populations, and likely did reach a critical point in the St. Johns valley at approximately 7300 cal 14C yr BP to support more intensive occupation (Miller 1992). But at 7 meters below modern levels, water would have been much more rare and likely restricted to channels. Much more local data is needed to understand changing conditions in the upper St. Johns.

 Mt. Taylor Thornhill Lake #9EARS"0 















                !BSOLUTE9EARS"0

Tanner, Stapor, et al. - St. Vincent Isl. Morton et al. (2000) Blum et al. (2002) Stapor and Stone (2004) 7-Pt Floating Average McFarlan (1961) Fairbridge (1961, 1974) Behrens (1966) Schnable and Goodell (1968) Figure 14: Holocene Sea Levels (modified from Balsillie and Donoghue 2004)

According to the Balsillie and Donoghue curve, there is a high stand at approximately 5600 rcypb (2004: Figure 7), with sea-levels ca. 1.5 m higher than present. This high stand is congruent with the proposed ending date for the Mt. Taylor period, and likely has significance for the 86 archaeological time frame under consideration here. Additional evidence for this high-stand is present at Groves= Orange Midden, where a layer of peat ca. 25-35 cm thick separates a lower shell midden stratum dated to the Mt. Taylor period from an upper shell midden stratum (McGee and Wheeler 1994), dated to the Thornhill Lake period by this author. McGee and Wheeler however associate this peat with a lowered sea-level (1994: 340). An additional period of higher sea- levels at approximately 4200 cal 14C yr BP may represent the mucks found at Palmer-Taylor that separates the Preceramic-Orange deposits from later St. Johns deposits (Rouse 1951). At present, sea-level reconstructions can at best provide directions for research in the upper St. Johns. The lack of local specifics and equivocal correlations creates a situation where possible correlations between cultural shifts and stabilities can be forwarded, but any assertions of certainty should be treated with suspicion. Despite a general scheme of climatic stability during the mid-Holocene (Miller 1998), climatic variability may be quite high (Viau et al. 2006). An organized program of coring would go far in helping to elucidate climatic details of the mid-Holocene. Rouse notes that the mucks found in off-mound units at the Palmer-Taylor mound dated to after the Preceramic and succeeding Orange periods (1951). Rouse dates the Preceramic and Orange periods to a geological event termed the Melbourne-Van Valkenburg interval. In Rouse=s scheme, this is a period of sea level regression. Rising sea levels between the Orange and St Johns II/Malabar II periods, encompassing the time between ca. 500 B.C. and A.D. 500, created these mucks. St Johns II/Malabar II material above the muck is taken to indicate this climatological event. The scheme used by Rouse does not enjoy common currency today, but the trend of rising sea levels following the Late Archaic is borne out by contemporary evidence (Little 2003; Balsillie and Donoghue 2004). An archaeological survey of the Tosohatchee State Preserve south of Palmer- Taylor on the west bank of the St. Johns River located no Archaic sites or St Johns II sites, but several small St. Johns I mounds were located. Nearly all of the mounds tested during this survey and other mounds mapped by local informants, totaling ca. 44 sites, were located on the 4.5 meter contour, and those with diagnostic materials present all had St. Johns I components. No sites were found below this line, but the survey was not systematic (Stewart 1985). This is in contrast to other 87 time periods, where most sites are found in the St. Johns basin and upland usage is transitory at most.

Pollen The nearest site from which a pollen core has been obtained and analyzed is the mortuary pond at the Windover site (Holloway 2002). This site is in the upper St. Johns drainage, 13.6 km southeast of the Orange Mound; the mortuary component of the site predates the Mt. Taylor. Two radiocarbon samples from the uppermost peat zone (Zone 1: 0-89 cm) were analyzed, returning stratigraphically valid date ranges (Doran 2002b: Table 3.1). The older of the two dates (Beta- 13910) is approximately 40 cm below the younger date (Beta-10763), and returns a range of 7170- 6690 cal 14C yr BP at the 2-sigma range, contemporary with the early Mt. Taylor period (Chapter 5). The upper date is calibrated (2-sigma) to 5730-5310 cal 14C yr BP (Table 1), corresponding with the Thornhill Lake period. Review of the arboreal pollen diagram from Windover shows a continuing presence of Pinus and Quercus from the previous depositional episode, with both species showing a marked decline at approximately the 5800-5600 mark, followed by a rapid rebound. This disappearance corresponds to the sea-level high stand described above. A rapid but short-lived rise in sea-level would create permanent standing water in this near-coastal region, preventing the growth of these trees in much of the upper St. Johns east of Osceola Ridge except for scattered high spots. Salix is the one of the few arboreal species with any marked change from earlier deposits, showing a rapid decline at the Mt. Taylor period interface (Figure 15). There are starker contrasts between non-arboreal pollens found in Zone 2, dating to ca. 9500- 7300 cal 14C yr BP and Zone 1, contemporary with Mt. Taylor (Figure 16). There is a tremendous decline in Chenopodiacae/Amaranthaceae pollen, with a corresponding increase in Poaceae and Asteraceae pollens, but this happens slightly later than the initiation of peat formation in Zone 1. Holloway interprets the pollen diagram as perhaps showing salt flats during the deposition of Zone 2, with a return of more arboreal conditions in Zone 1. This is interpreted as a return to cooler and moister conditions (2002: 223-224), and is similar to the trends from the Mud Lake pollen core in 88

Figure 15: Windover Aboreal Pollen

Figure 16: Windover Non-aboreal Pollen 89 central Florida (Watts 1969). The non-arboreal pollen can also be interpreted as a shifting away from salty to freshwater conditions, associated with maturation of coastal features and a higher non-artesian water table due to rising sea-level.

Temperature The general opinion is that climatic conditions became cooler and drier at approximately 7000 rcyb (Holloway 2002; Miller 1998). A recent study by Viau et al. (2006) provides greater detail to temperature fluctuations in North America during the Holocene. This study utilized 752 fossil pollen records that included over 2500 radiocarbon dates. Using 89 taxa and 4590 modern pollen records, the authors created estimates of climatic conditions using the Modern Analogue Technique (Viau et al. 2006: 1-2) for the past 14,000 years, with particular emphasis on the past 10,000 years. The series was interpolated to 100 year intervals using a simple linear interpolation technique and average temperatures found for each interval. Comparison with other datasets (GISP2 ice core, several δ18 records, and others) showed good coherence, encouraging confidence in the reconstruction. This is a continental reconstruction. The authors point out several temperature spikes and trends, including the previously observed high variability during the last 3000 years and, a greater climatic oscillation during the period from 8000-5000 cal 14C yr BP (Hypsithermal) than previously suspected (Viau et al. 2006). There are several high temperature peaks during this period, between which there are smoother troughs of lower temperatures (Figure 17). Three peaks are of relevance to this dissertation. A high temperature spike at ca. 7200 cal 14C yr BP corresponds closely with the beginning of the Mt. Taylor period (7300 cal 14C yr BP). This peak is followed by a rapid drop in temperature, congruent with the pollen data discussed above. At ca. 5800 cal 14C yr BP, there is another high temperature spike, followed again by rapid cooling. The ending of the Mt. Taylor period and the beginning of the Thornhill Lake period (Chapter X) is dated at 5600 cal 14C yr BP. At ca. 4500 cal 14C yr BP, there is an additional temperature spike. This is the date given for the termination of the Thornhill Lake period and the initiation of fiber-tempered ceramics. Though peripheral to this dissertation, there is another temperature peak and rapid cooling at ca. 3100 cal 90 14C yr BP, corresponding closely with the beginning of major mound

Poverty Point? construction at Poverty Orange Point (Gibson 1997). It Thornhill Lake should be emphasized that these peaks fall near the Mt. Taylor inception/decline of the archaeological periods, but do not necessarily characterize the bulk Figure 17: Mean July Temperature Anomaly of the Past 10,000 cal. yr. BP. of time for the periods, which for all is greater than a millennium. A relative stability within the periods is most characteristic of the data, but climatic instability seems to be the rule through most of the Holocene.

Su bs i s t e n c e Russo et al. (1992) have provided a thorough critique and realignment of Archaic subsistence research in the St. Johns, illustrating how the speculations of nineteenth century scholars continue to inform general evolutionary schemes for the region. Brinton, Wyman, and Moore, the most important of the early workers, cast the preceramic in terms of the faunal content of the shell heaps. For Brinton, the St. Johns people were ichtyophagous, i.e., fisheaters, near the bottom of Morgan=s evolutionary scale (Brinton 1872, cited in Russo et al. 1992: 96). Wyman provides what must be one of the earliest tabulations of faunal remains from an archaeological site in Florida (1868), and discusses the contribution of terrestrial and aquatic fauna to the diet. He felt that shellfish were the major element of the diet for the St. Johns inhabitants, an opinion that would be generally accepted for the next century (Goggin 1952; Rouse 1951). 91 Stephen Cumbaa forwarded the first real challenge to this view in 1976. Using somewhat less than ideal methods, Cumbaa deduced that terrestrial mammals provided a greater contribution to the diet than thought. He also arrived at a contribution of plants to the Archaic diet from thin-air, subtracting more from the contribution of aquatic resources. Cumbaa=s conclusions fit in well with the generalized settlement model of seasonal migration between the central Florida highlands, the river basins, and the coast. As clearly demonstrated by Russo et al. (1992), Cumbaa put the hunter and the gatherer into the St. Johns Archaic, at the expense of the fisher. Applying more appropriate quantitative analyses and integrating botanical remains into their assessment, Russo and colleagues (1992) argued that terrestrial mammals formed a much smaller fraction of the diet than previously thought. Aquatic animals, shellfish, fish, and reptiles, were most represented in the faunal assemblage by meat weight based on multiple methodologies. The contribution of plants to the diet is not quantified, but this is one of the first reports of the variety of flora exploited during the Archaic (see also Newsom 1985). Further work at Groves= Orange Midden provided more substantial data sets specifically concerned with the Mt. Taylor and Thornhill Lake periods. Wheeler and McGee (1994b), using the Russo model of St. Johns subsistence, analyzed multiple column samples from the submerged midden using the same basic methodology. The Groves= Orange Midden data indicate that aquatic animals provided the vast majority of the meat weight, greater than 85% for all samples. Aquatic animals, including fish, shellfish, and turtles, provided the greatest amount to the diet, while terrestrial mammals are rare and provide only a small portion of the meat (Wheeler and McGee 1994b: Tables 6-10). Lee Newsom conducted the archaeobotancial analyses for Groves= Orange Midden (Russo et al. 1992; Newsom 1994). A broad variety of plant remains were recovered, most of which were economically useful. Among the important species encountered are Cucurbitaceae pepo and bottle gourds seeds, both found in relative abundance throughout the excavation levels. The seeds do not differ greatly from wild specimens, so are not thought to be domesticated varieties, but may still indicate encouragement of their growth. Starchy seed plants, including amaranth and 92 chenopodium, may also be indicators of incipient horticulture. (Newsom 1994). Though the data are more limited, research in the upper St. Johns basin shows a somewhat different strategy than in the middle basin. Russo (1985) analyzed faunal samples from a number of sites encountered during archaeological survey of the upper St. Johns, including eleven stratified midden sites with preceramic deposits (75-106). Great attention and care is given to the bias of screen size and its effect on representativeness of faunal samples. Excavations at the Gauthier site lead by Brenda Sigler-Eisenberg and earlier work by Elizabeth Wing and others showed that finer mesh gauge could easily treble or quadruple the fish remains recovered during analysis. Not all sites discovered during survey could have all the soils screened through fine mesh (1.6 mm), but all soils were screened through large (6.35 mm) mesh and a small sample retained from each 10 cm level of midden material throughout the survey for laboratory processing. Experiments with finer mesh screens were used to provide some assessment of the recovery bias. Fine mesh screens trebled the individual fish identified, while not substantially increasing the amount of terrestrial fauna recovered, mainly commensal species. Russo estimates that only 34 percent of the faunal remains are recovered with 6.35 mm mesh screens. This level of recovery consistently and drastically underestimates the contribution of fish to the dietary spectrum (see also Marquardt 1996; Reitz and Wing 1999). Overall, the upper St. Johns midden sites show little emphasis on terrestrial mammals or shellfish. At the eleven sites that are believed to have preceramic components, none had shellfish remains in the preceramic layers. The one faunal class thought to be most representative of the Middle Archaic is missing from the non-mound archaeological sites. Shellfish remains begin to appear in the middens with the Orange period (bivalves), and continued throughout later ceramic periods, but are not of high frequency in most of the sites found during survey, less than 40 percent of ceramic sites have shellfish, and only 10 percent of the Orange/Preceramic sites have shellfish remains, while fish remains are consistently found. There are differences between the fish sizes exploited during the different periods, smaller fish from shallow water environments dominate the earliest assemblages, while later periods have larger fish from deeper waters (Russo 1985). Local 93 environments in the upper St. Johns may not have been conducive to shellfish growth. At the mortuary monuments of focus in this dissertation however, it appears that there are middens with shell underlying the monuments and a substantial midden was present at Persimmon Mound. All these monuments are at the northern or lower end of upper St. Johns, and the ancillary middens may indicate different local conditions during the Mt. Taylor period, but the scale of preceramic middens never seems to have approached those found in the middle St. Johns basin. Fish and turtle are likely the dominate protein and calorie source during the mid-Holocene, with lesser contributions from other aquatic and limited terrestrial sources. Excavations at Groves= Orange Midden recovered a sizeable sample of wood, bone, and shell artifacts related to fishing and processing. Based on this assemblage, fish harvesting wasan intensive activity. Nets were one strategy employed, reflected in the recovery of fids for net- weaving and shell net gauges (Wheeler and McGee 1994a). There is not a large sample of net gauges, but those recovered indicate small-mesh screens, ca. 6 cm maximum orifice (Wheeler and McGee 1994a: Fig. 17). Finer mesh fishing screens are indicative of a generalized, high- recovery strategy, with most fish and ancillary aquatic fauna recoverable. (Brown and Vierra 1983). Milanich and Fairbanks (1980: 151) suggest that bone points and awls may have been used in shellfish processing, but this idea has not been tested. Transportation along waterways was facilitated by canoe, for which we have abundant evidence. Low water levels in 2000 exposed organic artifacts throughout Florida, including at Newnans Lake near Gainesville where over 100 canoes and fragments were exposed. Of 55 studied, 41 dated to the Late Archaic (ca. 5500-2300 14C yr BP), contemporary with the Thornhill Lake and Orange periods (Wheeler et al. 2003). Further specific research is needed to determine the contribution of any particular resource to the Mt. Taylor period people=s diet but a picture is emerging of a primary reliance on aquatic fauna as a protein and caloric source with an adjunct reliance on gathered plant foods, terrestrial mammals, and fowl. The composite assemblage of plants and animals supports a thesis of limited transhumance (Russo et al. 1992: 105-106), in contrast to earlier models of high mobility and 94 drifting between sustained upland hunting and seasonal visits to riverine, slackwater, or coastal environments. A model of sedentism is only weakly tested at present, but the majority of data support a Acollector@ settlement system (Binford 1980). Marine fishing has attracted substantial archaeological attention, including considering marine resources as an economic substrate of complex social organizations (e.g., Mosley 1975; Martinez 1979; Yesner 1980; Acheson 1981; Jones 1991; Arnold 1992; Erlandson et al. 1998; Rick et al. 2005; Braje and Erlandson 2007), while freshwater fishing has seen little research (Plew 1996; but see Marquardt 1985; Beckerman 1994). Shell middens are a common feature of all fisher-people, but these middens are not the defining characteristic of the people, only material evidence of a particular adaptation. Most non-marine fish archaeological research in North America has been along the Northwest Coast interior rivers where populations focused seasonally on anadromous species (e.g. O=Leary 1992, 1996; Hayden 1992), with small attention given to the economic importance of vertebrate fish in interior aquatic adaptions in eastern North America (but see Limp and Reidhead 1979; Garson 1980; Brown and Vierra 1983; Price and Brown 1985; Claassen 1991; Erlandson 2001). Paradoxically, sites of freshwater fishing people are in great abundance and have been excavated since the beginning of scientific archaeology in the Americas. Shell-bearing sites, primarily referencing the classic freshwater shell midden, have been and are still a favorite subject of excavation, but in the Eastern Woodlands, are most always conceptualized as the remains of shellfishing people who occasionally fish (Waselkov 1982; Claassen 1991; Crothers 2004), rarely as the remains of fishing people who harvest shellfish as a component of a wide-spectrum aquatically focused strategy (but see Russo 1985; Russo et al. 1992; Marquardt 1985; Milner 1998; Blitz and Mann 2000; Marquardt and Watson 2005). The are several elements that help explain this hesitancy to recognize shell-bearing sites as possible components of a freshwater piscatorial economy; a few are discussed in this dissertation. Many of these issues apply to marine and nearshore shell-bearing sites, but most of these biases have been overcome, due largely to the recognition that marine fisher-hunter-gatherer complexity existed and that the Northwest Coast 95 was not a singular anomaly (Widmer 1988; Arnold 1992, 1993; Marquardt 1992).

Di s c u ss i o n This section is intended to provide an environmental context for the appearance, persistence, and eventual transformation of Mt. Taylor material culture and society. Explicitly, I do not adhere to a perspective that treats environment as explanatory or deterministic. However, I do take a perspective that similar conditions give rise to similar adaptions (Cushing 2005), so material culture will tend to change with alterations in the local environment. Variability in climatic conditions may also encourage realignment of intra- and inter-societal relations. People actively respond to

dramatic climate change and responses may include restructuring relations of material, power, and habitus. Environment, static or dynamic, is always correlative with society. Causality must be demonstrated, not assumed. A general characterization of the environmental conditions at the inception of and during the Mt. Taylor period can be forwarded as follows: A rapid temperature rise, followed by rapid cooling, marks the beginning of the period. There appears to be a drastic shift in the local flora scheme, with a short spike in Cheno-ams followed by a rapid decline in the same genera and a coeval increase in Poaceae and Asteraceae. Sea-levels aggraded slowly throughout the Mt. Taylor period, reaching modern mean level by ca. 5800-6000 cal 14C yr BP. During the Mt. Taylor period, temperatures continued to oscillate, culminating in another peak at ca. 5800 cal 14C yr BP, slightly before the proposed ending date for the Mt. Taylor period. Plant life shows some variation, but remains generally consistent until ca. 5600 cal 14C yr BP, when there is a short by punctated reduction in oak and pine pollens. Sea-levels remained fairly constant until ca. 5500 cal 14C yr BP, when then begin to rise again, peaking at ca. 5300 cal 14C yr BP and then regrading for the next 500 years. Subsistence strategies emphasized aquatic fauna, dominantly freshwater fish. Shellfish and terrestrial fauna supplemented fish as protein sources. The contribution of plants is less researched, but a wide variety of species were exploited for dietary and practical needs. Overall, indications 96 are that a rich local environment emerged during the mid-Holocene and was successfully exploited. The material consequence of these conditions and subsistence practices is the large shell middens that help to define the Mt. Taylor period. 97 Ch a p t e r 7: Mi d d e n a n d Mo n u m e n t One man’s garbage. . . Anonymous

The chapter is primarily a preface to the following two which are intended to accent some material distinctions between Mt. Taylor midden deposits and Mt. Taylor mortuary monument deposits. Two basic tasks are being undertaken in this dissertation: 1) Demonstrating the material and structural distinctions between Mt. Taylor middens and monuments, and 2) Constructing ritual- architectural reception histories of Mt. Taylor mortuary architecture. Both rely on stratigraphy in their explication. As a first undertaking, there is a review of the stratigraphic record from both types of deposits. Focus is on Mt. Taylor burial monuments, and they are first reviewed. Utilizing Harris matrices, stratigraphic reconstructions are presented that demonstrate building sequences and techniques and modes of human interment. The next chapter is devoted to a brief characterization of Mt. Taylor middens. There is an emphasis on contrasts between midden deposits and the previously described monuments. In attempting the task of hermeneutical engagement, there is a demand to make human experience the tempo of investigation. Towards that goal, two stratigraphic unit concepts are introduced as heuristic devices in this chapter: persistent surface and transitory surface. These concepts are in reference to the tempo of deposition and are often recognized archaeologically as interfaces (see Gibson 2006). Shell deposits are frequently cast solely as gradual accruement, but some shell deposits owe their origin to intensive short-term activity (Stein et al. 2003). A distinction between the two is needed to engage a hermeneutical method to monumentality and concurrently aid conceptualization of the stratigraphic distinctions made above.

Pe r s i s t e n t a n d Tr a n s i t o r y Su r f a c e s Shell sites are often essentialized as homogenous features, obscuring diversity within deposits, even when the diversity is known to be intentional:

“These deposits of shells are sometimes spoken of as mounds, but they are rarely to be considered as works of art in the sense that their conformation is the result of design. The accumulating refuse generally increased the habitability of the sites, and 98 distribution of the shells was no doubt in cases intelligently supervised with this end in view. It further appears that actual building sometimes took place, that shape was modified and height was increased for domiciliary and defensive purposes, and when the sites became places of sepulture the shells were utilized in building mounds. It is not, however, as works of art that these deposits are to be considered in this place ‑their use as constructions being a secondary consideration ‑but as accumulations of refuse inclosing in their mass reliable records of the food supply, the arts and industries, and, in a measure, the habits and customs of the people” (Holmes 1907). Making a distinction between two general kinds of activities, refuse accumulation and intentional building, takes us beyond the limits imposed on conceptualizing shell sites as fossils of secularity. A persistent surface is any activity surface exposed for an extended period of time; a surface that is not normally intentionally buried following activity. There may or may not be purpose in leaving the surface exposed. Most persistent surfaces are simply incidentally so. There is no intent to leave most of a village landscape as a surface, but there is not intent to cover it. Some persistent surfaces are intentionally kept so. An earthen house floor would be a persistent surface during its use life, despite daily accumulating debris and artifacts and regular cleaning. A ceremonial plaza may be maintained and cleaned continually to prevent surface accretions. Persistent surfaces are recognizable archaeologically by the presence of weathered surfaces, often an accumulation of debris associated with activity on the surface, and usually, horizontally extensive interfaces. In lithostratigraphic terminology, an unconformity would be a persistent surface (Gibson 2006; Aten 1999). Transitory surfaces are those activity surfaces that are meant to be covered/buried, surfaces not meant to be left exposed for extended periods of time. In contrast to persistent surfaces, with transitory surfaces there is brevity of time involved in their exposure. Most transitory surfaces are intentionally so, and a persistent surface can host a transitory surface. A house burned after the death of the owner and then covered with a mantle of earth would represent a transitory surface surmounted by a persistent surface, as would the purificatory burning of an existing mound surface. Other examples of transitory surfaces would be the intermediate surfaces during brief 99 halts of mound construction, the area of a funeral pyre that is then buried, or a mound summit on which mortuary remains are displayed prior to encapsulation with soil. Transitory surfaces do not become persistent surfaces except by accident or interruption. Importantly, there can be an inverse relationship to the thickness and depth of archaeological deposits and the total amount of time involved (Stein et al. 2003). A substantive mound stratum may be mainly burial of transitory surfaces and in total represent very little elapsed time, perhaps on the order of weeks or months. Conversely, a persistent surface could represent years, decades, or even centuries of time, but be represented by very little material accumulation. Confusion between transitory and persistent surfaces is the greatest impediment to understanding the mechanics of construction in shell mounds, because in contrast to middens, the thickest deposits accreted very quickly, sometimes within days or weeks, while some stratigraphic interfaces, with very little vertical representation, represent centuries of accumulation. Mt. Taylor monuments are conceptualized as a combination of transitory and persistent surfaces. A generalized sequence of monument construction begins with an extant midden (a persistent surface). A large fire is built atop this midden (a transitory surface). Secondary burials are piled atop the fire/charcoal (a transitory surface). After the bones are piled, a mound mantle is deposited rapidly, and then that surface is left persistently exposed. Into this surface, burial pits were excavated and people placed, often with ancillary feasting and activity (transitory surfaces). The initial making of the monument is a brief episode, but once made, the monument is a constant and central feature of everyday life. Mt. Taylor middens are conceptualized as primarily persistent surfaces. Accumulation is gradual, incremental, and mostly undirected. Large crescentic middens are formed by a few households that regularly relocate short distances. Generations of repeated usage contribute to a mounting pile of detritus. The shell underfoot is evidence of community success and deep ties to place. Massive middens are incidental but significant persistent features. With these devices, data are now presented that illustrate distinctions between monuments and 100 middens. The data are extensive, but that is a necessary consequence of providing a radically different model for shell heaps than usually provided and in exploring behavior at the level of the individual and the society. 101 Ch a p t e r 8: Si t e De s c r i p t i o n s “There is some stratification in the deposits but it apparently has little cultural significance.” (Goggin 1952: 41)

In t r o d u c t i o n This chapter reviews the archaeological stratigraphy at five Mt. Taylor shell heaps where human remains have been excavated. Archaeological data are presented detailing the depositional history of the shell heaps, focusing on the mounded areas where human skeletal remains are found. By showing the regularity and repetition of specific attributes of stratigraphy at multiple sites, it is argued that the deposits were intentionally built- that these Mt. Taylor shell mounds are mortuary monuments and sacred architecture, constructed atop and coeval with extant middens and but one component of village life. The focus is on stratigraphic data in this synthesis, utilizing graphic and textual descriptions of the stratigraphic units encountered to derive sections. Stratigraphy is recorded by all investigators of the sites reviewed, in terms that allow comparisons. The descriptions themselves vary widely; none of the projects (excepting Harris Creek) used any type of standardized system of recording the site matrix. I have however strived to use the descriptions of the original recorders. When altered, it is a change for brevity, not content. Harris Creek will serve as the archetype site since it is the most completely reported Mt. Taylor mortuary mound (Jahn and Bullen 1978; Aten 1999). The other four sites, the Palmer-Taylor Mound, the Orange Mound, the Persimmon Mound, and Persimmon Mound II, are located in the lower half of the upper St. Johns river basin. Selection of these sites was based on an extensive review of all the archaeological reports obtainable for the upper St. Johns. These four sites were the only examples of burial loci in the upper St. Johns that could be certainly attributed to the Mt.

Taylor period.1 Each of these sites was excavated by C. B. Moore in the late 19th century, and

1 Many more mounds in the region are assigned to the Mt. Taylor period by Wheeler et al. (2000), but I was unable to verify any but these four. There is a possibility of Mt. Taylor period interments at Indian Fields, Duda Ranch, and Persimmon Mount (distinct from Persimmon Mound), but the available information is too limited to be certain. 102 all but one are previously reported and recognized as dating to the Mt. Taylor period (Mitchem 1999). In the upper St. Johns, the Palmer-Taylor Mound is the only site discussed to have experienced excavation in the last hundred years. One small project conducted immediately prior to World War II is previously reported. There are additionally two archaeological field schools that have not been generally reported prior to this dissertation. The Orange Mound is famous in its own right, but no professional work has occurred since Moore. Orange Mound is located on the west side of the St. Johns, across from the outlet of Salt Creek. Persimmon Mound may have been visited by Wyman. This site is located on the east bank of the St. Johns, just below the outlet of Salt Creek and directly north of Orange. Persimmon II, a small shell heap just west of Persimmon Mound, has not been previously reported anywhere but in Moore=s field notes. Eight stratigraphic attributes are compared between the five localities. The attributes compared are: 1) Areally extensive and continuous strata, 2) Human bone entombed by those strata, 3) Grave pits that are dug into those strata, 4) Group/community burial, 5) Associations of fire and mortuary contexts, 6) Associations of food and mortuary contexts, 7) Associations of white sand and mortuary contexts, and, 8)An extreme sparsity of utilitarian artifacts. Not all of these are universal to every site, but most are found in the five sites reviewed.

Me t h o d o l o g y Referencing sections of the upper St. Johns shell heaps, a modified Harris matrix system is used to analyze stratigraphic data and provide a means of reconstructing depositional structure. A full discussion of the background and theory of Harris matrices can be found in Harris (1979, 1997). Harris matrices are two-dimensional representations of stratigraphic units, used for sorting and correlating the units by assigning a code to and then charting those recognized units. Three basic relationships between units are sought: units above, units below, and units contemporary. Content and composition combined with superposition allow for the collation of stratigraphic units. The most basic rule of the Harris matrix is that the units must follow the laws and principles 103 of stratigraphy; any violation of these rules will be apparent in the matrix. In this dissertation, I subscribe to the rules of the Harris Matrix, but do not utilize all the available options of presentation or analysis. Correlated units (i.e., units that were originally parts of a single unit) are not linked by horizontal lines, but instead are included in a box. A unique coding system is used, primarily to allow the reader to trace the relevant excavation pit idealized in the matrix. Finally, I do not employ phasing or periodization as is possible in the Harris matrix system. The concept of interface (Harris 1979: 43-48) is integral to this analysis, but only those with direct bearing on the

stratigraphic sequence and collation are given notice in the matrix.2

Each designated stratigraphic unit was assigned a unique code, consisting of year of excavation, unit number, section direction, and stratigraphic unit number. Each section was entered into both the Arched and Stratify programs. Obviously contemporary units, including the humic pottery bearing upper layer and the sterile subsoil, were linked. Using the criteria of superposition, content, and character, likely corresponding intermediate units are progressively linked and ordered. As the individual stratigraphic units are integrated into the matrix, continuous layers become apparent and are considered strata. Stratigraphic units that do not extend across the site, i.e., are localized within a single section, are Aset off@ and their relationship to strata determined. This process allows for the determination of the stratigraphy across the site and the horizontal extent of each stratigraphic unit. Each of these strata is given a roman numeral designation and the stratigraphic units arranged to demonstrate relationship.

Ha r r i s Cr e e k (8Vo24) Harris Creek is a large shell heap located on Tick Island in the middle St. Johns basin (Figure 18), one of a number of substantial sites located on the island (Miller and Griffin 1978). Harris Creek was originally investigated by C.B. Moore as the adjunct to his major investigations at

2 The availability of software to draw Harris matrices eases the task. This research used two of these packages: Arched 1.41 and Stratify 1.4. Arched is used to draw matrices while Stratify is a relational database for stratigraphic data that is primarily aimed at organizing and analyz- ing that data using principles of the Harris Matrix. Stratify automatically draws the matrix based on optional conditions. In both programs, no automatic means of assigning relationships between stratigraphic units is provided; all decisions of relationship is left to the archaeologist. The matrices provide a way of visualizing what relationships are determined and which are not. The software only assists in collating many individually documented stratigraphic units. 104

Hardscrabble

Tick Island Complex

Harris Creek

0 1km

V B NorthR Figure 18: Sites on Tick Island the Tick Island Mound (8Vo25) and adjacent shell heaps (8Vo221 and 8Vo222). During 1961 salvage operations at the site while it was being commercially mined for shell, a portion of a Mt. Taylor mortuary mound was excavated and recorded. A full accounting of the excavations was not completed before the death of the primary archaeologist: a short report (Bullen 1962) and a posthumous manuscript, concerned mainly with artifacts from the site and later excavations (Jahn 105 and Bullen 1978), comprised the bulk of information regarding this mound until a 1999 report by Lawrence Aten. Inspired by alternate interpretations of monumental shell constructions during the Archaic period in Florida (Russo 1991; 1994; 1996), Aten, who had assisted Ripley Bullen during the 1961 excavations at Harris Creek, decided to provide a fuller report, integrating the earlier work of C.B. Moore (including unpublished field notes) and Francis Bushnell with the field notes of Bullen and an analysis of the materials recovered in 1961. The product of Aten=s efforts provides the fullest picture yet of Mt. Taylor mortuary practices, and goes a tremendous way towards understanding the complex stratigraphic situation found at Mt. Taylor mounds. This paper represents a watershed moment in the archaeology of harvester-gatherer monumentality, and provides a robust model of mortuary traditions and construction during the Mt. Taylor period. Therefore, it is the most appropriate preface to the Upper St. Johns site descriptions. Aten uses geoarchaeological principles and lithostratigraphic terminology to describe and present the site stratigraphy as recorded by Bullen in 1961. This terminology is meant to provide standard descriptors of stratigraphic units, similar to the principle of the Harris matrix as employed in this dissertation. Without expressing an explicit theoretical stance towards monumentality, Aten forwards the stratigraphy as his primary evidence to argue for the presence of a Mt. Taylor burial mound composed of at least seven construction layers and in the form of a generally flat platform subsumed by later shell deposition. He does not take a final stance on the origination of the material used in construction, but seeks to demonstrate that the materials used were placed to create and expand the mound form. Definitions of Aten=s lithostratigraphic terms and their relation to the Harris matrix system used in this dissertation follow:

Interfaces/Depositional Discontinuities: One of the most important and productive stratigraphic concepts used is the importance of interfaces/depositional discontinuities. The contact between units of deposition, Aten describes three types of interfaces, described below. Bullen recorded interfaces on the Harris Creek stratigraphic sections, representing interruptions of deposition (143).

Unconformity: Significant breaks in deposition due either to erosion or to extended non- 106 deposition and stability of the ground surface. Processes that produce unconformities include incipient soil development and cultural usage of a surface (143). I would add that examples of unconformities in shell deposits would include layers of naturally deposited soils, organic lenses, charcoal/ash rich lenses, and thin layers of crushed shell. In the Harris matrix system, unconformities are described as horizontal layer interfaces (Harris 1979: 124). Unfortunately, specific use of this concept can only be used occasionally in the upper St. Johns site descriptions presented here. C.B. Moore and other archaeologists rarely recorded any description of most interfaces, but all excavators of the upper St. Johns sites would note any remarkable characteristics of these interfaces. There is a presumption that such interfaces existed, as the presence of depositional discontinuities is one of the primary means of distinguishing archaeological strata. Conversely, the lack of these discontinuities in off-horizontal planes of matrix demonstrates the unity of strata and the effort and intention of mound strata construction (Gibson 2006).

Angular Unconformity: A product of the Aremoval of substantial underlying material cutting across the principal planes of deposition followed by placement of new material bedded or oriented in another direction@ (Aten 1999: 143). This type of interface is referred to as an interface of destruction in the Harris matrix system and is described using the Law of original continuity (Harris 1979: 124).

Disconformity: A result of limited sediment removal or disturbance. Examples listed by Aten include: tree roots, tree throws, animal burrows, and postholes (1999: 143). Cultural and natural transforms can apparently produce disconformities. Within the Harris matrix system, a disconformity can be an interface of destruction, and/or a vertical feature interface. In this dissertation, disconformities are not typically discussed, though recognized in some of the original documentation. There are exceptions where relevant, and cultural disconformities are considered types of features, e.g., grave pits and postholes. Within Aten=s system, the fill within a disconformity can be a type of:

Deposit: A stratigraphic unit of limited extent, and used for local features and lithologic variations (Aten 1999: 138). A deposit may or may not be associated with a disconform­ ity. For example, the matrix of a refuse filled pit, the ash, charcoal, and faunal material of a shallow pit , and a pile of sand intentionally heaped on a surface and subsequently buried are types of deposits. There is not direct correlate to deposit in the Harris matrix system, but all would be regarded as a unit of stratification. An obviously useful concept, I utilize Aten=s notion of deposit for the relevant site descriptions.

Layer: Used for Asediments between interfaces when the units could be mapped across all or much of the excavated area, or that have implications for the whole area; they are based largely on continuity, geometry, superposition, and to some extent, lithologic unity (Aten 1999: 138)@. In the Harris matrix system, the man-made layer and natural layer, both units of stratification, would be generally equivalent to Aten=s layer. In this dissertation, the layer of Aten is equivalent to my use of stratum. I use layer to refer to generally horizontal units of stratification as revealed within individual excavation 107 units.

Stratigraphy and Mortuary Deposits The Harris Creek site, like many, if not most, shell mound sites in the middle St. Johns basin has been subjected to intensive commercial mining of the shell. During the last major mining episode at Harris Creek, Ripley Bullen acquired emergency funding for three weeks to document the immense number of Archaic burials that were being exposed and destroyed. The operator who owned Amining rights@ to the shell burial mound allowed Bullen to record what he could. What he could included ca. 73 meters of section drawings and the location of 184 burials. The operator estimated digging a thousand burials and sending them to the shell shaker. In the words of Aten, this can safely translated to Aa lot@ (1999: 17). A total of eight stratigraphic layers and two Aplaceholder@ layers are described by Aten (Figure

Figure 19: Harris Creek Sections 108 19). The placeholder layers represent all intact deposits below those recorded, approximately 1.5 meters (Layer 1) and all deposits above those recorded, i.e., those removed prior to Bullen=s arrival (Layer 10). The remaining eight layers described are all associated with the Mt. Taylor period, and represent intentional construction layers. The mound seems to have generally taken the form of a kidney bean, and at its latest manifestation and greatest dimension, measured an approximate 40 meters in length, 26 meters in width, and greater than 2 meters in height (Aten 1999: Table 1). Layer 2 is a large contiguous stratum of shell that was found across the excavation and is considered to be a basal layer. No mortuary features are found in this layer, and it is composed of two sequential deposits: and earlier snail and brown Adirt@ matrix and a later clean small snail matrix. This layer is estimated to be approximately 15-by-30 m measuring 90 cm in height. Layer 3 rests on Layer 2 unconformably. Above Layer 2 is the first mortuary layer, Mortuary A, a component of Layer 3. Two major structural features, including a Awhite sand zone@ and a ABlack Zone,@ both described shortly, and three or four levels of postholes were recorded as part of this layer. The white sand is associated directly with the mortuary deposits, the result of two principle behaviors: the piling of white sand atop Layer 2 that included the covering of a large grave pit that was eventually surrounded by the shell component of Layer 3, and, the digging of pits into the shell of Layer 3 that were then filled and/or covered with white sand. All the white sand deposits were localized, most being in the range of less than a meter in diameter and a few centimeters thick, with at least one large sand deposit measuring 2.5 to 3 m across and as much as 60 cm thick. Layer 3 draped over all of Layer 2, but particular components of the layer, including the white sand and black matrix mentioned above, are localized. Approximate dimensions of Layer 3 are 34-by-20 m, raising the mound height to 1.2 m at the thickest. A total of 78 burials are associated with Mortuary B. Remains of fires, including charcoal/ash and calcined shell, were found in association with the mortuary deposits. In some of the small grave pits, a fire was built at the base of pit and human skeletal remains, perhaps in wrappings, are placed on this fire, and white sand thrown on top of the fire/bones, extinguishing the fire. Occasionally, evidence of burning is found on the bone. Such an 109 association was found with a large grave pit, the only such pit found during the project. Located at just over a meter east of the northeast margin of the ABlack Zone@ and near the base of Level 3, the pit was ovoid in shape, measuring about 1.1 m in the long axis, and dug into Layer 2 to a depth of approximately 90 cm. Near the base of the pit the unspecified bones of a child and the skulls of two adults, along with a few post-cranial adult bones, were deposited in a fill containing shell and sand. A fire was placed immediately above these bones, calcining the shell and eventually cementing the shell, bone, and sand into a mass. Above this fire, a fill of white sand, brown mottled sand, a few mussel shells and very few snail shells was placed in the pit. Above this fill, the remains of eight individuals was placed in a fill of white sand with some charcoal. The crania of these eight individuals nearly touched each other, and a substantial amount of charcoal found between two of the burials. An unidentified black material was found with these skeletons. A large pile of white sand was placed atop the pit, likely directly over. Some of the posts found in Layer 3 may have formed a screen, fence, or structure surrounding this pit. Other burials were placed around the large pit within the fill of Layer 3, but none were placed directly above the pit. Layer 3 included the only extended burial encountered (Aten 1999: Table 2). Included in Layer 3 was a deposit referred to as the ABlack Zone@ (Aten 1999: 167). Described by Bullen (1978: 21) as a Alens of black charcoal-impregnated sand, six inches thick (15 cm), which covered an area about five by ten feet (1.5 m-by-3 m). Bullen interpreted this black zone as possibly the remains of a fire that was underneath a structure or platform that held wrapped bodies for drying or smoking, which led him to the conclusion that a charnel house may have been built here, and that bodies were buried over time in the accompanying Layer 3 matrix (Aten 1999: 167). Despite the description of this black matrix as charcoal-impregnated and the observation that it was underlain in at least part by calcined shell, Aten is of the opinion that there was not much charcoal based on the inability of a radiocarbon lab to find enough charcoal for dating in soil samples sent in and originating from this deposit in 1964. Instead, Aten suggests that this black deposit originated on a mound summit and was composed of various organics, some animal bone, and including localized abundances of charcoal. Three individual burials are found directly 110 under the black lens, which Aten finds as a significant lack of grave density. Based on the plot of Mortuary A burial locations (1999: Fig. 17), there may be a comparable density under and near the black lens, with 13 crania being found under and within 1.5 m of the deposit limits. Burials are found wherever excavations were made in Mortuary A. Horizontal excavations in 1961 are mainly restricted to 3 10-by-10 foot squares. The large grave pit is in the northeast square, while the black lens is found in the west majority of the two units west and southwest of the large grave pit. A total of 34 postholes were recorded in the excavations of Layer 3, 32 of which are plotted in a sinuous line along the east fringe and immediately beyond the black lens. In doing, the line of posts divide the area between the black lens and the large grave pit. There is no contact surface between the black lens and any of the postholes, but all lie stratigraphically­ beneath the deposit. Based on the top elevation of the postholes, Aten groups these features into four levels. In each level, he reconstructs any evident structural patterns. No pattern was found in Level 2. Two tentative right-angled alignments are found in Level 1, the lowest, and an arc is found in both Levels 3 and 4. None of the 32 postholes described are found in the square of the large grave pit. The lack of horizontal exposure at this layer prevents determining if further post holes existed. Overlying and resting unconformably on Layer 3 is Layer 4, described as dark brown sandy matrix with small snail shell and fairly uniform throughout. Mortuary B was found as a component of Layer 4, and included nearly half of the burials found at the site (n=67). The discussion of Mortuary B is limited compared to Mortuary A, mainly due to the lack of major distinguished deposits including large grave pits, large white sand deposits, black organic/charcoal-impregnated lenses, or posts (Aten 1999: 173). Burials, mainly displaying flexed treatments, were placed into the Layer 4 fill. It appears that white sand was not associated with burials in Mortuary B, but artifacts are somewhat more common (Aten 1999: 179). The maximum dimensions of Layer 4 are estimated as 40-by-21 m and raised the mound to a maximum height of 1.5 meters. Layer 5 is only briefly described. Resting unconformably atop Layer 4, the matrix is described as dark Adirt@ with small snail shells. A total of eight burials are associated with Layer 5, and Aten 111 does not consider there to be a cemetery or designated mortuary area in the layer. This layer seems to have been primarily an expansion of the mound to the east, resulting in a larger mound footprint and increased the size of the summit platform (Aten 1999: 177). The maximum dimensions of the mound following the deposition of Layer 5 is estimated at 40+-by-24 m, with the mound remaining at a height of 1.5 meters. Resting unconformably on Layer 5, Layer 6 is described as primarily clean shell, with a single localized deposit of Ashell and dirt@ (Aten 1999: 145). This layer is mainly an expansion of the mound to the east and increasing the height of the summit platform. Interpreted by Aten as midden refuse and an area of intentional dumping Asince the dark-colored sediment and decayed organic debris of domestic use areas was largely absent@ (1999: 144), I would assert that the lack of organics and soil indicate that this was a rapid and intentional accumulation, but not the dumping of midden refuse. Instead, I would interpret this layer of clean shell as an example of rapid shell construction. No burials are associated with Layer 6. The dimensions of Layer 6 are estimate at greater than 40-by-26 meters, and raised the mound to a height of 2 meters. Layer 7 appears to be another expansion of the mound vertically and to the east. This layer rested unconformably on Layer 6 in most areas, but there are some local disconformities. This layer included several hearths with burned shell and ash, fishbone, and a lens of burned mussels a meter across (Aten 1999: 147), some of the clearest indications of consumptive behavior in the mound. The context of these food preparation areas on the summit of a mortuary monument supports an interpretation of feasting (Beasley 2007). There is only one mortuary deposit associated with this layer, but is a bit unusual and worth comment. This was a large grave pit that originated in Layer 7 extending down into Layer 4 (Aten 1999: Figure 5), and is one of two large grave pits, the other being the large multi-person pit found in Layer 2. A single flexed skeleton is reported for this pit, but no artifacts were reported. Resting upon an angular unconformity on the northern side of the deposits, Layer 8 is composed of a mixture of deposits. The basic matrix is a dark brown soil with small snail shell, lenses of mussels and larger snail shell, occasional lenses of white sand, faunal remains including fish, deer, 112 bird, small animals, and turtle, as well as marine shells. Aten interprets this as another construction layer, perhaps being redeposited material (1999: 147). An alternate possibility is that this deposit represents the discard of special-use refuse on a mound flank (Jackson and Scott 2003). Four burials were associated with Layer 8, including Burial 51, which had what may have been deer antler headgear that could not be recovered (Aten 1999: Tables 2, 5). A layer of muck that may have originated from a marsh bottom and transported to the mound overlay Layers 4 and 7, referred to as Layer 9. Ceramics were found in this layer, including Orange and St. Johns types. This layer appears to cap the Mt. Taylor mound and is associated with a later site occupation. The Bluffton Burial Mound, another Mt. Taylor mound in the middle St. Johns basin and containing a single extended individual, has a similar layer of gumbo muck as a mound stratum (Sears 1960: Figure 2). Aten=s description of the history of cultural deposition at Harris Creek is the fullest exposition of the stratigraphy of a multi-interment Mt. Taylor mound to date. Despite the immense tragedy of the destruction of the Harris Creek site, the lengthy sections revealed and the large area investigated allows for major strata to be followed across and through a single mound and to better understand the presence of a large number of human interments and their relationship to mound strata and structure. Many of the features of Harris Creek are found at other Mt. Taylor sites, including the four mortuary monuments discussed in this dissertation: Palmer-Taylor Mound, Persimmon Mound, Persimmon Mound II, and Orange Mound. Aten discusses the similarities between Harris Creek and Orange Mound, listing 6 specifically:

1) Same local environment, shell midden matrix, and subsistence strategies

2) Mound orientation

3) Pre-Orange period burials

4) Bundled burial

5) White sand layers, associated with mortuary deposits

6) White sand surmounted by brown sand with additional skeletal remains (1999: 176) 113 These are important features of comparison between Mt. Taylor sites. Below, I make comparison between Harris Creek and four upper St. Johns mounds, using some of the similarities pointed out by Aten, and including several other specific structural features found at Harris Creek.

Pa l m e r -Ta y l o r Mo u n d (8Se18)

Setting The Palmer-Taylor Mound is located on the west bank of the St. Johns river just north of Puzzle Lake and directly above the confluence of the Econlockhatchee river and the St. Johns (Figure

20). 3 Several other sites are located in close proximity to Palmer-Taylor, including the small rise

Palmer-Taylor Mound

Great Kilbee Mound

Buzzard’s Roost

0 175 350 700

Meters V B NorthR Figure 20: Area of Palmer-Taylor Mound

3 The approximate UTM coordinates of the mound are UTM Zone 17S, Easting 496591, Northing 3174557, found on the Geneva 7.5’ USGS quadrangle. 114 immediately adjacent to the mound referred to as the Shapfeld Mound and the Mt. Taylor period shell heap Buzzard=s Roost. Other sites found within a 1km radius of Palmer-Taylor include the Cabin Mound, Tozzer Mound, Twin Tree Mound, Great Kilbee Mound, Catfish Mound, and Hog Island Mound. Data are insufficient to determine if any of these sites, excluding Shapfeld and Buzzard=s Roost, date to the Mt. Taylor period. Palmer-Taylor is situated near a small grouping of low hills surrounded by wet prairie. The site regularly floods today, but is accessible by land during low water. The surrounding geology is late Pleistocene or early Holocene in origin, consisting mainly of relic beach and dune sands, with silts and clays from being lagoonal or estuarine interfaces (Scott 2001). Rock of any sort is rare on the surface, but underlying karstic limestones could be exposed in springs or caverns (White 1970). Utilitarian stone is always non-local. The nearest access to water is at the present Econlockhatchee River channel, approximately 200 meters distant. Surrounding the site to the east is freshwater marsh, suitable for airboat traversing, but difficult to navigate by other means (Figure 21). A large pond is found just to the east of the site

Figure 21: View East from Palmer- Taylor Mound 115

Figure 22: View from East towards Palmer- Taylor Mound and Suspected Spring that may have been a spring during the Middle Archaic, when sea levels were lower (Balsillie and Donoghue 2004; Figure 22). This pond exhibits characteristics that encourage this interpretation: the pond retains water during dry periods and the circular shape of the pond is reminiscent of sinkholes. Further research is needed to test this thesis. The site consists of a semi-conical shell heap, approximately 3 meters in height, a thin shell apron to the north, and an additional small shell heap to the northeast, known as the Shapfeld Mound. Along the east-west and north-south axes, the primary mound measures approximately 88 meters in diameter in each direction (Figure 23). Including the shell apron and Shapfeld Mound, the site covers approximately .95 hectares. Despite some damage due to mining of shell for road fill and previous excavations, the site remains in excellent shape, no doubt in part to its inaccessibility.

Previous Research John and William Bartram passed the site in 1766 and likely came within eyesight of the mound twice, but it is not mentioned (Bartram 1942). Jeffries Wyman does not mention the site in his works on the St. Johns shell heaps (1868; 1875). The first archaeological notice of the Palmer- 116

1 meter

1

2 meter 2 3 4 4N 5 3N 6 2N 7 1N 1W 3 meter 1 4W 2W 2 3W

9

1 2 3 4 5 6 7 8

Moore 1892 Units Excavators’ Club 1941 Units 10 0 10 20

Rollins 1975 Units meters Rollins 1976 Units Figure 23: Excavations at Palmer-Taylor Mound

Taylor site is by Francis LeBaron, who located a shell heap in the vicinity of the Econlockhatchee confluence on his 1877-78 map of archaeological sites in Florida, but it is difficult to be certain that Palmer-Taylor is the corresponding site. In February of 1892, C.B. Moore spent two days excavating at the Palmer-Taylor site, referred to as the AShell Heap Near Econlockhatchee Creek@ in his publications and notes (1892b; 1893c; FN1; FN3). The publication describes the mound as approximately 6 feet high, sloping in all directions towards surrounding wet prairie. The site Alooms like an island in the prairie which in wet weather becomes a marsh@ (Moore 1893c: 609). 117 Moore’s 1892 Excavations

Excavation I

The following description is derived from Moore=s publications and his field notebooks. Excavation I is described by Moore in detail. Located at the apex of the mound, the excavation was 6 feet (1.83 m) by 4 feet (1.22 m) by 7 feet (2.13 m) deep (Figure 24). The first foot (30 cm) was composed of sand and shell, which contained pottery and faunal remains (a in Moore=s

Horizontal Arbitrary

3AND POWDEREDSHELL M92U1_1 POTTERY FAUNA SHELLDISK

"ROKENANDFRAGMENTED 3AND POWDEREDSHELL HUMANBONEShINGREAT SLIGHTADMIXTUREWHOLE NUMBERvTHROUGHOUT SHELL HUMANBONES STRATA INCLUDINGON M92U1_2 FAUNA hHEARTHvATINTERFACE WHERETHEYWERE hTREATEDINRESPECTTO BREAKAGEASWEREBONES { M92U1_3 #HARCOAL OFLOWERANIMALSv

M92U1_4 3HELLWITHSLIGHT ADMIXTUREOFSAND

M92U1_5 3ANDANDWHOLESHELL

3!.$

  CM

Moore Palmer-Taylor Excavation 1 Feb.. 4, 1892. Section based on written description, FN1: 46-47, FN3: 101, 1893c Figure 24: Moore Excavation I Section 118 designation system). No pottery was found by Moore below this point. The second layer (b) was described as 2 feet (60 cm) thick, with powdered, broken, and whole shells. Throughout this layer were found many human bone fragments, some with burning, in association with a dog jaw that Moore detailed throughly, and animal feces. The finding of a dog mandible may be the first occasion of discovering canine remains in Middle Archaic contexts in Florida. Due to the scattering of the bones, Moore could not make an estimate to the number of individuals, but did recover four crania fragments. Likely, many more individuals were represented in the deposit. Between the second and third layers, Moore found a lens of charcoal he describes as a hearth. Human bones, broken and burnt, were found in association with this charcoal. The third layer (c), was 1.5 feet (46 cm) thick and described by Moore as pure shell and sand, with little else. The fourth layer (d) recognized by Moore consisted of 2.5 feet (76 cm) of sand and whole shell, with no bones or pottery. At the base of this layer he encountered Aswamp muck@.

Excavation II

Moore=s Excavation II is described as being on the east side of the mound, 10 yards (9.1 m) down the slope from Excavation I and 1.5 feet (46 cm) lower vertically. The excavation measured 6 feet (1.83 m) by 3 feet (.91 m), by 4 feet (1.22 m) deep, and only penetrated the first two layers. Again, pottery was found only in the upper layer (Figure 25). Human bone was found in small numbers, along with animal bones. Two crania fragments were recognized. The unit was taken no further, and no specific mention is made of the charcoal lens.

Excavation III

A third excavation is recorded in Moore=s field notebook as aA small hole about 2 2 feet deep.@ No human remains and Anothing of particular interest@ were found (FN1). There are no clues to the location of this excavation unit and no indication that Moore returned to Palmer-Taylor during any of his later sojourns on the St. Johns. Moore felt that the osteological assemblage was material evidence of a cannibalistic feast and that the Ahypothesis of burials, even of disconnected bones, would seem untenable, as absolutely nothing found in association with them pointed to interments@ (1893c: 612). Clearly, Moore did 119 Horizontal Arbitrary

3AND POWDERED SHELL ANDPOTTERY M92U2_1

3AND POWDERED SHELL SLIGHT ADMIXTUREOF M92U2_1 WHOLESHELL FRAGMENTED HUMANBONE

$EPOSITS#ONTINUED

  CM

Moore Palmer-Taylor Excavation 2 Feb.. 4, 1892. Section based on written description, FN1: 46-47, FN3: 101, 1893c Figure 25: Moore Excavation II Section not feel that the human remains at Palmer-Taylor represented an intentional mortuary deposit.

Excavators’ Club Excavations During the Christmas break of 1940-41, the Excavators= Club, a group of undergraduate and graduate students from Harvard and Radcliffe Universities, drove from Cambridge to the marshes of the upper St. Johns to excavate for one week at Palmer-Taylor (Davis 1999). This endeavor used what were then innovative recording techniques to excavate in whole or part eight 2x2 meter units in an L-shaped pattern at the apex of the mound and three test pits near the margins of the mound. The group also conducted limited testing at nearby sites (Rouse 1951). The goals of the field crew were laudable, and they accomplished much in such a short time. World War II interrupted the reporting of the Excavators= Club work. A manuscript based on the Palmer-Taylor data was written by the Anthropology Club of the Harvard Peabody Museum 120 in 1947 (Dyson and Tooker 1947; Rouse 1951; Davis 1999). I have as yet failed to locate a copy of this manuscript, so the description of the Excavators= Club excavation must rely on Rouse=s synopsis of the materials. The next published reference to Palmer-Taylor after Moore is in Rouse=s Survey of Indian River Archeology (1951: 116-126). Thanks to the work by the Excavators= Club, Palmer-Taylor is one of the most detailed accounts of upper St. Johns sites in this publication. He reviews the published work by Moore, and treats the 1940 material thoroughly. Rouse acquired the Palmer-Taylor

,OAMWITHFINELY CRUSHEDSHELL ABORIGINAL ANDHISTORICARTRIFACTS 41_1_1 DISTURBEDHUMANREMAINS

#OARSELOOSESHELL WITHADMIXTUREOF 41_1_2 LOAM DISTURBED BURIAL !SHYGREYSOILWITH SHELL SOMECONCRETIONS hPIT LIKEvCONCENTRATIONS 41_1_3 OFSHELL PRIMARYFLEXED BURIALWITHASSOCIATED #HARCOAL 41_1_4 ANTLERPROJECTILEPOINT NOTCHEDBONE TURTLESHELLS 3HELLWITHVARYING ADMIXTUREOFCHARCOAL ANDASH SOMEPARTS LARGELYASHORCHARCOAL 41_1_5 SHELLSCOMPACTBUT MAINLYWHOLE

'REYSANDWITH 41_1_6 SOMESHELL ASH PIT SOMEFAUNA

"LACKSANDYSOIL 41_1_7 WITHSCATTERED 7ATERTABLE SHELL BONEARTIFACTS

41_1_8 &INEGREYSAND STERILE

  CM

Excavators Club 1941 Composite Section Section based on written description, Rouse (1951: 119-124) Figure 26: Excavators’ Club Section 121 materials from Harvard and transferred them to . He distilled the Excavators= Club eight 2x2 meter units, designating 6 cultural levels and one sterile level to the site (Figure 26), with the four lowest cultural levels dating to the Preceramic (1951: 124). The first two levels correspond to Moore=s first level, Rouse=s Level 3 correlates with Moore=s b, and contained a primary flexed burial as well as human bone fragments. Several pits are reported in this layer, but their nature is uncertain. Again, a lens of charcoal was found to separate the burial strata from the underlying shell layer. The remaining levels correspond to Moore=s layers, with the addition of the sterile sand at the base. In contrast to Moore=s statements, artifacts were found in the preceramic strata, though not numerous, and not elaborate. Plates and drawings of the few artifacts found, mostly modified bone, can be found in Rouse (1951). Faunal remains were found throughout the mound, though in few numbers below the burial layer. Animal remains found in the burial layer include cottontail rabbit, domestic dog, Florida raccoon, two turtle species, deer, opossum, sandhill crane, fish, and beaver.A test pit was excavated into the Shapfeld Mound. Shell, turtle, and terrestrial faunal remains were recovered, but no artifacts were found. Rouse cautiously assigns the Shapefeld Mound to the preceramic period.

Rollins College 1975 Excavations Davis (1999) reports that Dudley DeGroot led two classes of Rollins College undergraduates to dig at Palmer-Taylor in 1964 and 1965, but I have not been able to locate any further reference to these excavations. There are a few excavations at the site that are unaccounted for, and these intrusions may represent the DeGroot work. In January of 1975, Burton Williams of Rollins College conducted a field school with students from Rollins College. At least one student paper was generated from this work (Irving and Norris 1975). Seven 2 x 2 m units were excavated in whole or part down the northern slope of the mound, along with two smaller off-mound tests. Each unit was separated by a 50 cm balk, resulting in a 17 x 2 m trench. Soils were screened only for the upper levels. 122 The stratigraphy encountered in the southernmost units, those nearest the apex, is similar to that recorded by Moore and the Excavators= Club (Figure 27). The northernmost units are also similar, with the exception of the charcoal lens. The uppermost layer consists of humic soil mixed with some sand and shell, and includes historic and aboriginal artifacts. Various undocumented diggings had disturbed this layer, and penetrated into underlying layers. Underlying this layer is a shell matrix with some ash. No human remains are reported for Unit 6, but some displaced osteological material was recovered in Unit 2 (Figure 28). A thick lens of charcoal was found beneath the shell layer in Units 6 and 7 but did not extend further north. Another layer of shell, with almost no admixtures, was found under the charcoal. This shell was so densely packed that excavators reported finding shell still in the green state. Underneath were layers of sand with some shell, underlain again by sterile sand (Williams 1975). Faunal remains were found throughout the mound, but particularly concentrated in the first layer. The faunal assemblage is consistent with that reported by the Excavators= Club. Terrestrial mammals, avian bones, fish, and turtle were found throughout (Irving and Norris 1975). Conch was found in the upper layers but not reported below the second layer. Artifacts below the disturbed

Soil and shell, aboriginal and historic artifacts, 75_3W_1 disturbed

Shell with some ash

75_3W_2

75_3W_3 Heavy shell with concretions

75_3W_4 Sand and shell

75_3W_5 Sand with little shell

0 50 100cm

Rollins College 1975 Unit 3 West Section Figure 27: 1975 Unit 6 West Section 123

75_2W_1 Sand and shell

Sand and shell 75_2W_2

Heavy shell with concretions 75_2W_3

75_2W_4

75_2W_5 Sand and shell

75_2W_6

Sand with little 75_2W_7 shell

0 50 100cm

Rollins College 1975 Unit 2 West Section Figure 28: 1975 Unit 2 West Section zone were rare. Of the 1182 artifacts found, only 26 were below the first layer, and only 5 of those were found below the second layer. Williams recorded some landscape features near the mound to the northwest that deserve mention, though his details are skimpy. The first day on site, he describes several ring-shaped features that had a slightly mounded perimeter, slightly mounded in the middle, and filled with water and water plants. Williams does not give any measurements of these rings. Later during the project, an undescribed trench was dug through the perimeter of one of these features on both sides of the pool. Artifacts and faunal material was sparse, but apparently present, and the excavations were halted without any real documentation. The general description of these features is reminiscent of Awells@ recorded by Rouse at South Indian Field (1951: 94-104) and the Helen Blazes site nearby (Edwards 1954: 72). These remarkable features are usually larger than 5 meters across and dug down to underlying clay, often to depths approaching 2 meters. Water fills these features during the wet season. Orange pottery is regularly found at the bottom, with occasional later materials and occasional human remains recovered (Rouse 1951). Edwards (1954: 72) makes 124 a incredible report of finding bone, shell, basketry, rope, gourds, and a dwelling support pole among the three wells he excavated. He also reports the depth of one well as near 3 m, and reported Orange period refuse nearly filling the well. There are no photos or drawings of these materials, but as the ceramic period was not Edwards= concern, and these results have no bearing on his thesis or conclusion, it would seem strange for him to make such claims falsely.

Rollins College 1976 Excavations In 1976, Marilyn Stewart, with the assistance of Paul Zeph, led another group of undergraduates from Rollins College to excavate at the Palmer-Taylor Mound (Stewart 1976). Eight units were excavated in whole or part, and an additional unit was designated for the cleaning of a profile where backhoe removal had exposed a substantial section of shell. A flexed burial was found at the top of this exposed section. An uncalibrated radiocarbon date of 6060 +/- 60 rcybp was obtained from conch shell from the preceramic strata (UM‑1155). One conference paper (Stewart and Zeph 1976) and an incomplete manuscript (Stewart 1977) were produced from this work. A sketch map by Stewart that shows extant disturbances has allowed a reconstruction of a site map showing previous excavations (Figure 23). The eight test units followed a single baseline across the southern slope of the primary shell heap. Three of the units measured 1x1 meters, five measured 2x1 meters. Two of the units, 1 and 8, were off the mound; the remainder were on the mound proper. Unit 4 serves as an example of the stratigraphy encountered in the mound (Figure 29). A humic layer with pottery and historic artifacts was encountered first, underlain by a shell layer with some disturbances. Concreted shell under the first shell layer likely belongs to the burial stratum encountered by Moore and the Excavators= Club. Under the concreted shell is a lens of charcoal. A heavy shell layer under this charcoal rests atop another shell layer with proportionally less shell. An unusual clam layer is under this shell, followed by a transition to sterile sand. Units at the periphery of the mound showed a slightly different stratigraphic situation, discussed below. The excavation of Unit 9, which was primarily the cleaning of the profile left after previous backhoe shell mining, provides an exposure along the west of the apex of the mound, where other 125 Many roots, finely crushed shell, dark colored Humic layer 76_4N_1 Shell layer Shell layer 76_4N_3 76_4N_2 76_4N_15 76_4N_4 Charcoal layer 76_4N_4 Part cement layer, 76_4N_5 more shell, light color Light gray with 76_4N_6 76_4N_7 clam and snails Shell layer 76_4N_10 Moon shell lens Compact shell, 76_4N_8 76_4N_9 moon snail Clam and 76_4N_11 snail layer

76_4N_12 Clam layer Transitional layer 76_4N_13 Sterile dirt 76_4N_14 and sand

0 50 100cm

Rollins College 1976 Unit 4 North Section Figure 29: 1976 Unit 4 North Section

previous apical units are located to the east and north of the mound=s highest point. This is the steepest side of the mound, and would have risen more abruptly than the rest of the mound. In this unit, a primary flexed burial was documented at what would have been the upper portion of the second level; the humic layer was removed previously (Figure 30). It is not clear if the burial is properly assigned to the Mt. Taylor period or to a ceramic period, but no artifacts were recovered, suggesting the former. This burial and the flexed burial found by the Excavators= Club (Rouse 1951: 122-123) illustrate some of the diversity of interments found in the mound. Several thin lenses of different colored materials, not found elsewhere, underlies the burial pit, which rests in the burial bearing matrix found elsewhere through the mound. The north section of Unit 2 demonstrates more of the structure of the shell heap. This unit was located on the western fringe of the mound and appears to have been dug at the shell heap/ off-mound interface (Figure 31). Underlying the humic layer, on the eastern end of the section, are the two shell layers found elsewhere. These shell layers transition into loams and sands. Two unique depositional layers were found under the shell. Immediately under the lower shell layer and a transitional loam layer, excavators found a thick layer described as grey sand, soil, and big 126 "URIAL0IT

5NDESCRIBED 76_9N_1 76_9N_2 3ANDANDSHELL CONCENTRATE 7HITETOGRAYAND 76_9N_3 "LACKLAYER SNAILCONCENTRATION 76_9N_4 76_9N_5 76_9N_6 #EMENTED3AND ANDSHELL $ARKGREYWITH SNAILSANDMUSSELS 76_9N_7 "LACKLENSCHARCOAL 76_9N_8

"LACKLENS ,OOSESHELLS 76_9N_9 76_9N_10 MOSTLYMUSSELS 3MALLSNAILS 76_9N_12 $ARKGREY LOOSE ANDMUSSELS 76_9N_11 SHELL LOOSE 76_9N_13

76_9N_14 $ARKGREY FEWERSHELLS 76_9N_15 3TERILESAND

  CM

2OLLINS#OLLEGE5NIT.ORTH3ECTION Figure 30: 1976 Unit 9 North Section rocks. Under this layer is one described as white sand with rocks. Unfortunately, no greater detail beyond Arocks@ is given, but several pertinent observations are recorded in the field notes. Primary among these is that Unit 2 was thought to have Athe edge of the mound there and it seems to have started in a depression. Half the square is black, the other is shell@ (Stewart 1976: 21, emphasis in original). These are the only deposits that are said to include rocks in any of the Palmer-Taylor excavations. Concreted shell, a post-depositional phenomenon common to shell sites that is found by all excavators since Moore, is on occasion affectionately called Ashell rock@ by the Rollins undergraduates in the field notes. But, no shorthand of rock for shell rock is found in any other documentation, and when concretion is found, it is restricted to the first two shell layers and is not in layers primarily composed of sand. In Stewart=s field notes, there is a mention of small iron nodules at about the midpoint of excavation of Unit 2. Unit 2 is described as being full of iron and Astuff that resembles decayed feces@ (Stewart 1976: 26). The amount of Airon@ was so great that the excavators first thought that this was a historic feature, but eventually realized that no historic materials were being recovered, the Airon@ was more like rock, and that this deposit underlaid what 127 appeared to be undisturbed preceramic layers. Throughout the upper St. Johns, stone of any sort is infrequently found. Raw material for chipped-stone artifacts likely originates in either Marion County to the north-northwest or the Tampa Bay area to the west. Limestone can be found, but the rock encountered in the mound does not fit the description for limestone. The rock encountered could be limonite, a loose term for an iron rich amorphous aggregate that is found in subtropical and tropical climates, sometimes forming sizeable masses in littoral zones due to the flocculation of iron hydroxides. Ochre is a form of limonite (Prinz et al. 1978), and Florida limonite Arusts,@ having a bright red color. Smaller masses are found throughout Florida in sandy soils, often referred to in the field as iron nodules. The limonite could be a natural constituent of the submound soils, but is not found in other units. The deposit could be concentrated in this area alone. Large fragments usually form at littoral zones, and may have been deposited in the geological past. Another possibility, given the argument for intentional construction of the shell heaps, is that limonite was mined elsewhere and purposefully placed at the edge of the mound prior to construction. During the late Archaic, bannerstones of limonite are found in Florida and Georgia (Ford 1966). Recognized features were not commonly encountered during excavation, but the remains of a possible structure were found in Unit 9 consisting of a post stain and associated floor observed in profile. No structures are recorded for Mt. Taylor sites in the upper St. Johns. The depth of this possible structure indicates it is contemporary with the earliest occupation/use of the mound. An additional possible feature was located in Unit 2, consisting of ash, a rectangular deposit of Asolidified but rotten shell,@ black soils, and Aespecially lumps of iron-rich soil, sometimes scattered and sometimes concentrated@ (Stewart 1976: 21, 26). The feature was immediately above the limonite bearing layers described above. The function of this feature is postulated by Stewart to be a special purpose hearth. I am not sure of the function, but it may be related to the initial mound dedication. The remaining sections from the 1976 field school are included as Figures 32-36. 128

Humic, topsoil, roots

76_2N_1 Shell Undescribed, likely black muck 76_2N_2

76_2N_4 76_2N_3 Hard-packed shell

Transitional 76_2N_5 grey loam

Grey sand, Orange(?) with 76_2N_71 76_2N_6 soil, “big rocks” grey, black

Orange(?) and 76_2N_9 white sand White sand, 76_2N_8 rocks

0 50 100cm

Rollins College 1976 Unit 2 North Section Figure 31: 1976 Unit 2 North Section

Gray silt with 76_3W_1 snail shells

Shell with some silt 76_3W_2

Dark gray silt and shell 76_3W_3

Black stratum small snails 76_3W_4

76_3W_5 Brown mussel shell

76_3W_6 Gray and black silt and shell

Sterile sand, no shell 76_3W_7

White sand

76_3W_8

0 50 100cm

Rollins College 1976 Unit 3 West Section Figure 32: 1976 Unit 3 West Section 129

Humic, topsoil, 76_5N_1 roots

Shell concrete 76_5N_2 layer

76_5N_3 76_5N_5 Concrete masses 76_5N_4 Concrete masses

Moon snail shell concentrate Snail and mussle with some 76_5N_6 concrete masses

0 50 100cm

Rollins College 1976 Unit 5 North Section Figure 33: 1976 Unit 5 North Section

Humic, topsoil, roots 76_6N_1

Little shell Sand and mussel mostly dirt 76_6N_2 Concreted mussel 76_6N_3 with sand and dirt 76_6N_4

Dirt with snails 76_6N_5 and mussel

Dirt with little shell 76_6N_6

White sand 76_6N_7

0 50 100cm

Rollins College 1976 Unit 6 North Section Figure 34: 1976 Unit 6 North Section 130

Humic, topsoil, 76_7W_1 roots

Light dirt with snail and clam 76_7W_2

Medium dirt with snail and clam 76_7W_3

Darker dirt with 76_7W_4 snail and clam

76_7W_5 Sterile dark dirt

76_7W_6 White sand

0 50 100cm

Rollins College 1976 Unit 7 West Section Figure 35: Unit 7 West Section

Humic layer

76_8N_1

Dark brown with some black and gray sand (muck?) 76_8N_2 Transitional to sand 76_8N_3

76_8N_4 Sterile dirt and sand

0 50 100cm

Rollins College 1976 Unit 8 North Section Figure 36: Unit 8 North Section 131 Author Visit In August of 2005, the author, in the company of Myron Estes, John Lieb, and Michael Stallings, visited several sites in the upper St. Johns, including Palmer-Taylor. A general reconnaissance of the site was made, including a determination of the mound elevation. Measurements in the field indicate that the mound apex is ca. 3 meters, as stated by others. The previous excavations reported above are evident, as is the borrow pit on the western edge. Little appears to have changed at the site since the 1976 field school, but extensive ground cover made it difficult to thoroughly assess the site during our visit.

Co n s t r u c t i o n o f t h e Pa l m e r -Ta y l o r Mo u n d The sum of the archaeological projects described above results in Palmer-Taylor being the most extensively investigated Mt. Taylor period mound site in the upper St. Johns. Over 93 square meters have been exposed at the site, nearly 1 percent of the total site area, including the apron and Shapefeld Mound, or 1.6 percent of the mound proper, a substantial sample far beyond that for other upper St. Johns sites. The lack of publication of the more recent excavations has resulted in the neglect of Palmer-Taylor in considerations of St. Johns archaeology (e.g., Milanich 1994). In this section, an attempt is made to integrate the excavated data, hopefully producing an adequate synthesis of what is known about the mound. The only previous attempt to synthesize the archaeology of Palmer-Taylor is Rouse=s (1951) description of the site. He used the published work of C.B. Moore and reanalyzed the manuscript reporting the Excavators= Club excavations. In this analysis, I utilize the published works of C.B. Moore (1893c), the unpublished field notes of C.B. Moore (FN 1; FN3), the synthesis by Rouse (1951) of the Excavators= Club work, the field notes of Burton Williams= 1975 field school and the student paper from that work (Irving and Norris 1975), the field notes of Marilyn Stewart from the 1976 field school, and the unfinished manuscript from that work (Stewart 1977). This discussion references the Harris matrix constructed for Palmer-Taylor, Figure 37. The uppermost stratum, Stratum I, consists of the humus/topsoil and all soils where ceramics were 132

76_2N_1 1 M92U1_1 1 Stratum I M92U2_1 1 41_1_2=41_1_1 1 75_6E_1 1 76_4N_1 1 76_3W_1 1 75_3W_1 1 76_5N_1 1 76_6N_1 1 76_7W_1 1 76_1N_1=76_1N_2 1 76_8N_1 1 76_4N_15

76_7W_2 76_1N_3=76_1N_4 2 76_8N_2 2 Ceramic Finds

Burial Pit 76_9N_1 Stratum II 76_9N_2 Mortuary Mound Fringe Muck 76_9N_3 Stratum 76_9N_4 Primary Burials 76_9N_5

Secondary 76_2N_2 4 M92U1_2 4 76_9N_6 4 M92U2_2 4 41_1_3 4 75_6E_2 4 76_4N_2=76_4N_3=76_4N_4=76_4N 4 76_3W_2 4 75_3W_2 4 76_5N_3=76_5N_5=76_5N_5 4 76_6N_2 4 Burials

76_5N_4

M92U1_3 5 76_9N_7 5 Charcoal 41_1_4 5 75_6E_3 5 76_4N_6 5

Stratum III 76_4N_7 Shell Building Episode (?) 6 6 76_2N_3 6 M92U1_4 6 76_9N_8 6 41_1_5 6 75_6E_4 6 76_4N_8 6 76_3W_3 6 75_3W_3 6 76_5N_6 5 76_6N_3 76_7W_3

Mt. Taylor Deposits 76_4N_9 76_4N_10

76_2N_4

76_2N_5

76_9N_9 7 76_4N_11 7 76_3W_4 7 Stratum IV 76_6N_4 7 76_2N_ 6

Rock-Bearing 76_2N_7 Layers 76_9N_10 76_9N_12 8 76_4N_12 8 76_3W_5 8 Structure ? Stratum V

76_9N_11 76_2N_8

76_2N_9

M92U1_5 9 76_9N_13 9 41_1_6 9 Stratum VI 76_3W_6 9 76_6N_5 9 76_7W_4 9

10 76_9N_14 10 41_1_7 10 75_6E_5 10 Stratum VII 76_3W_7 10 75_3W_4 10 76_6N_6 10 76_7W_5

Sterile 76_1N_5 3 76_8N_3 3 Sand 76_2N_10 11 76_9N_15 11 41_1_8 11 75_6E_6 11 76_4N_13=76_4N_14 11 76_3W_8 11 75_3W_5 11 76_6N_7 11 76_7W_6 11 76_1N_6 11 76_8N_4 11 Figure 37: Harris Matrix of Palmer-Taylor Mound recovered. Historic disturbances, including arboriculture, previous archaeological excavations, shell borrowing, and surreptitious diggings, are included in this stratum; further analysis would provide segregation of this layer into intact aboriginal areas and historically disturbed areas, but would not be pertinent to this research. Human remains, displaced by previous diggings, were found by twentieth-century excavators, but appear clearly to be out of context (Rouse 1951). Historic and aboriginal artifacts were found in abundance in Stratum I, and pottery is restricted to this layer. During the 1975 Rollins excavations, 99.97 percent of the artifacts recovered were in this layer (Irving and Norris 1975). The horizontal extent of this stratum can be considered to encompass the entire preeminence. All deposits below Stratum I are preceramic and assigned to the Mt. Taylor period. Undisturbed osteological material was found only in Stratum II. Three modes of interment are recorded within this stratum, all of which are secondary treatments: 1) flexed interments, 2) commingled human 133 bones, and 3) burnt and broken commingled human bones. Burials were found in a matrix of shell, primarily snail with some mussels, ash, mixed and concentrated charcoal, and occasional animal bone. Terrestrial and aquatic faunal remains (excluding shell) are found with some of these burials. Shell is found in Apit-like concentrations@ in some areas (Rouse 1951), reminiscent of basket loading of shell observed in other Mt. Taylor sites (e.g., Wyman 1875; Sears 1960). The sequence of interment represented in Stratum II is burnt and broken commingled human bones stratigraphically first directly on a thick bed of charcoal, followed by unburnt commingled human bones. This bone was entombed by the shell matrix and later, pits dug into this shell and primary flexed interments placed in the pits. Layers of crushed shell indicative of tramping and buried A soil horizons are absent throughout Stratum II, as are major disconformities. The horizontal extent of Stratum II is estimated at over 3300 square meters. Burials are not found throughout the entire stratum, the majority are found at the apex of the mound and to the western fringe. A rough estimate of the horizontal distribution of human remains is approximately 330 square meters. It is impossible to estimate the number of individuals represented in this deposit, but dozens, if not hundreds, are likely to have been buried. Underlying Stratum II is a contiguous lens of charcoal. This lens is found primarily at the apex of the mound, contained within five excavation units. It was encountered by all excavators who worked at the site, varying between a few centimeters to ca. 15 centimeters in thickness. This lens is thought to represent a single major burning episode, engulfing a horizontal surface of approximately 569 square meters and probably closely congruent to the distribution of human osteological material. The burnt and broken commingled human bone was found atop this charcoal lens. Stratum III has a slightly greater horizontal extent than Stratum II. This stratum has the characteristics of an intentional mounding event, with almost pure shell found throughout and little other faunal remains or artifacts recovered, and could be a major mound building episode prior to the burning of the surface at the highest point of this stratum (Charcoal Lens) and then the deposition of human remains (Stratum II). An incidental accumulation of shell, i.e., a deposit 134 created by domestic activity and discard, would be expected to have many thin layers of crushed shell from walking and surface activity and substantially more in the way of artifact content (Meehan 1982). This stratum, by contrast, contained shell with little admixture, almost no artifacts, possible basket loading of shell, and densely packed shell that was occasionally found in a still green condition. The probable limonite deposits found in Unit 2 of the Rollins 1976 excavations underlay this stratum. The four strata under Stratum III possibly represent earlier occupational/activity episodes. Generally, a greater proportional amount of soil is found with depth, though Stratum V represents a localized lens of mussel shells. Artifacts were still rare but were recovered in greater proportion in these lower strata during the Excavators= Club excavations (Rouse 1951). A possible structure was encountered in Unit 9 of the 1976 excavations, in close relationship to Stratum V. Stratum VII appears to represent the initial usage of the mound area. A layer of sterile white/grey sand is found across the site underlying the cultural deposits. Total stratigraphic evidence from the Palmer-Taylor Mound indicates that the majority of the heap is an intentionally built preceramic mortuary mound. Following a period of perhaps sporadic usage of the mound area, a major building episode took place followed soon by a large conflagration on the summit and then piling of commingled human remains. Those near the bottom of the pile of bone were charred by the fire, but those higher up were not damaged. This pile of secondary burials was then entombed by a horizontally expansive shell stratum. Later, flexed burials were deposited episodically within this shell matrix. Secondary burials are concentrated on the apex of the mound, but the areal extent of the mortuary stratum is much greater, and burial pits could be found across the mound. A limited later occupation of the mound during the St. Johns II periods included no mortuary activity or apparent linkage to the burial mound and marks the final usage prior to the late nineteenth century. 135 Or a n g e Mo u n d (8Or1)

Setting The Orange Mound is located atop a small rise adjacent to a lagoon west of the west channel of the St. Johns River near a incredibly braided confluence of several major and minor drainages. The site is opposite the outlet of Salt Creek, which itself drains a complex of lakes that extends eastward to just west of the city of Titusville. North of the site the river becomes more organized, though it remains sinuous, slow, and far reaching. South of Orange Mound, the river is divided

into several small channels in a broad basin (Figure 38).4 There are no other archaeological sites presently recorded in a 1 km radius, but several sites, including the Persimmon Mound, a.k.a.,

Persimmon Mound

Persimmon II

Rattlesnake Mound

Indian Fields

Orange Mound

0 750

V B Meters NorthR Figure 38: Area of Orange Mound, Persimmon Mound, and Persimmon II 4 The approximate UTM coordinates of the mound are UTM Zone 17S, Easting 502910, Northing 3163452, found on the Titusville SW 7.5’ USGS quadrangle. 136 the Baxter Mound, Arrowhead Ranch, Possum Bluff, and Rock Island are found in a 2 km radius. Persimmon Mound is positively preceramic. The other sites have pottery, at least on the surface, but if Mt. Taylor deposits are present, it is not determinable with the present state of data. At present, Orange Mound is an island surrounded by shallow water (Figure 39). The small rise on which the site rests (Bear Island) is isolated from the uplands to the west by an extensive marsh nearly 2 km wide. The surrounding geology is late Pleistocene or early Holocene in origin, consisting mainly of relic beach and dune sands, with silts and clays from formal lagoonal or estuarine interfaces (Scott 2001). No outcrops of stone are found nearby. A large pond to the west of the site could be a relic spring. The circular shape is reminiscent of springs and sinkholes, and most indicative, there is an outlet but no inlet. This would tend to indicate that the pond is a spring, but this is supposition.

Figure 39: Orange Mound

The site is described by Moore (1893c) as a crescent, with a maximum height of ca. 14 feet (4.27 m). Following the ridge of the crescent, Moore measured the site as 560 feet (171 m) along a north-south axis and 260 feet (79 m) wide at it=s greatest width. LeBaron incorrectly states the mound height as 40 feet (12 m), but this is clearly a misprint. The opening of the crescent is to the 137 west-southwest. No accurate map has ever been produced for the site.

Previous Research John and William Bartram arrived at Lake Ruth directly east of Orange Mound in January of 1766, and this is as far as the expedition went. John claimed to have accomplished his mission of finding the headwaters of the St. Johns and turned around. This determination is as much an indication of the change in the character of the river as the difficulties with going further into the glades. The location of the site away from the main channel likely mitigated the chance of their encountering the shell heap (Bartram 1942). Jeffries Wyman does not mention the site in his works on the St. Johns shell heaps (1868; 1875). The first archaeological notice of the Orange Mound is by Francis LeBaron, who visited the site in 1877, making a small collection from an earlier excavation by alligator hunters, reporting pottery and bones. LeBaron recommended the site for further research (LeBaron 1884: 776). The site is included on LeBaron=s map of Florida sites (1891). C.B. Moore first conducted archaeological excavations at Orange Mound on February 7-9, 1892, digging four units (FN1: 54-61; FN3: 119-137), and providing the data for his primary publication regarding the site (1893c). He returned on February 25, 1893, but the notes from this digging consist of a single detached paragraph (FN4: 105; FN7: 53). It is his excavations that form the entire corpus of knowledge regarding the Orange Mound. The site is somewhat exceptional for Moore in that he drew perhaps the only true section of an excavation unit found anywhere in his St. Johns work (1893c: 622), which he included in the large field notebook (FN7). Rouse=s description of Orange Mound is based on LeBaron=s visit (1884), the published description of the site by Moore (1893c), a review of the artifacts, and a site visit (1951: 127-129). There is little to add to what was previously known about the site. Rouse does dispute Moore=s assertion that the site represented a sand burial mound, as he notes the presence of shell and animal bones, found in other sites, including Palmer-Taylor, calling the site a midden rather than a mound. A description of the stratigraphy as reported by Moore is given, and Rouse reports on the few artifacts he could locate in collections. 138 Since Moore=s excavations, no further archaeology has taken place at Orange Mound, but the site has become well known to researchers, primarily for the fiber-tempered pottery component. W.H. Holmes, using Moore=s material and observations, recognized the priority of fiber-tempered wares in the St. Johns, referring to the pottery as Midden Ware, and citing Orange Mound as one of the primary locations (1903: 120-121). Later, the site served as the type-site for the Orange period and Orange series (Goggin 1952: 43-47; Griffin 1945; Sears and Griffin 1950; Bullen 1955: 15).

Moore’s 1892 Excavations C.B. Moore excavated at least twice at Orange Mound, in February of 1892 and February 1893, but the information from the 1893 work is sparse. In this section, I describe the four excavation units from 1892, and include the fragment of documentation from 1893. The descriptions follow closely with the original wording found in the field notebooks and published works. A map showing the estimated location of Moore=s 1892 units is provided as Figure X. Without an extant accurate topographic map of the site, the location of the excavation units is based on a postulated topographic model generated by interpolating a local extract of the USGS 7.5' DEM to .1 foot contour lines (Figure 40). Therefore, unit locations are only approximate. We have a limited archaeological dataset from Orange Mound, with four excavation units described shortly being the only information at present. The sections from these units do provide means of illustrating structure and intentionality in the shell heap and provide evidence of architecture. Excavation 1 is the sole unit detailed in published references, and has been until now the only excavation available for study. That section provides a picture of the stratigraphy at the highest point of the mound, but gives no notion of the structure of the site beyond this point. The other units are spread across the mound: Excavation 2 is located 26.8 meters north of Excavation 1, on the northern slope. Excavation 3 is located 30.48 meters south of Excavation 1 on the southern slope, and Excavation 4 is located 2.44 meters south of Excavation 1 (Figure X: Site Map). Incorporating the other three excavation units gives a much more robust picture of the site stratigraphy and the scale of construction and building efforts. Using a Harris Matrix to analyze the structure of the mound stratigraphy helps to demonstrate the sequence of construction 139

Excavation 2

Excavation 1

Excavation 4

Excavation 3

10 0 70 Meters

Orange Mound Figure 40: LocationOrange of Moore’sCounty, FL Excava‑ V B Site Map NorthR tions atC.B. Orange Moore 1892 Mound Excavations and mortuary activities.

Excavation I

This is the primary excavation at Orange Mound, excavated February 7-9, 1892. In the published report on the site, this is the only unit detailed, though others are mentioned. The excavation was located at the highest point of the mound/crescentic ridge. The excavation started as 12.5 feet (3.81 m) by 8.5 feet (2.59 m), and converged to 5 feet (1.52 m) by 3 feet (.91 m) at the bottom (Figure 41). 140

Horizontal Arbitrary

Humic/Topsoil OM92U1_1

Hearth/fireplace (charcoal)

OM92U1_2 Shell, almost pure apple snail

Orange OM92U1_3

Brown sand and shell, mixture Human bones, of species extended and OM92U1_4 bundled burials{ White sand OM92U1_5 Thornhill Lake?

Brown sand and shell, mixture OM92U1_6 of species Approximate depth of human bones

Hearth/fireplace (charcoal) OM92U1_7

Crushed shell OM92U1_8 Mt. Taylor Mt. Taylor “Dirty” sand and shell, some broken

OM92U1_9 Somewhat purer shell Shell and loam OM92U1_10 OM92U1_11 Sterile white sand OM92U1_12

Water Encountered

0 50 100cm

Moore, Orange Mound, Excavation 1 Feb.. 7-8, 1892. Section based on written description, FN1: 55-58, FN3: 121-133, 1893c Figure 41: Moore’s Excavation I

The first layer (Rouse=s Layer 1) is described as loam, with abundant pottery that was the same type as found on the surface, likely St. Johns plain and check-stamped. The layer was approximately 1 foot (30 cm) in thickness, and a shell Achisel@ was found at the base. The second 141 layer (Layer 2) is a 3 feet (91 cm) thick layer of almost pure apple snail with abundant pottery, the pottery Abeing thick and coarse@ (FN 3: 121). Much of this pottery was incised, and based on the illustrations in the published report (1893c), is clearly Orange types. A Afireplace@ was found at 1.5 feet (46cm) below surface. Below this second layer, a layer of brown sand with a slight admixture of various shell species, considered by Moore to be reminiscent of sand burial mounds found on the St. Johns that generally date to the St. Johns I and II periods, is found that contained human remains (Layer 3). This layer varied in thickness between 1 foot (30 cm) and 1.75 feet (53 cm). At 4.5 feet (1.37 m) below surface, human and animal bones were encountered. The bones were not broken, but the excavation had disturbed any indication of burial mode. Moore dug into the south section of this unit with trowels, and found an extended primary burial, of which the majority was recovered. The skeleton is identified as female. Underneath this burial was found a bone implement of a type Anot uncommon in the shell heaps,@ meaning the preceramic mortuary mounds and middens (FN3: 127). At the same level, and another foot into the unit wall, a secondary bundle burial was found. A disassociated femur was found next to the cranium of this burial. There is no mention of burial accompaniments. This third layer is generally thought to be an Orange period burial stratum, based on Moore=s statement that no pottery (Anot a particle@) was found below 5 feet (1.5 m), and pottery was abundant above this depth. Such abundance is not clear in the field notes. Moore comments that pottery begins abundantly and gradually decreases in occurrence. There is no mention of pottery with any of the burials encountered in this layer, a notation that Moore most always makes, even if there are only a few sherds. A sparsity of ceramics in the third layer was encountered in Excavation 2, described below. This layer may date to the Orange period, but I am uncertain of such placement. An early Orange placement, marking the transition from the Mt. Taylor to the first production of pottery, is possible. The next layer (Layer 4) is pure white sand. At its thickest, on the apex of the mound in the 142 northwest corner of the unit, this sand is 1.5 feet (46 cm), but Moore finds it to be locally restricted, diminishing in thickness in the east and west sections, but absent in the southern section. Rouse (1951) considers this layer to be a hiatus in occupation between the preceramic and ceramic site usages, implicitly a result of natural accumulation. However, this seems an unlikely scenario, given that the sand is found at the highest point of the mound, but not along the flank. Alluvial or eolian origin would not likely have deposited the thickest portion at the peak, and little or no sand on the flank. A non-local origin for the sand is possible, as found in other sites. Under the white sand layer was found another brown sand layer (Layer 5) with a slight admixture of sand and bearing human remains. The layer was 2 feet (61 cm) thick, and the human bones were found initially at 1.5 feet (46 cm) and continuing to the base of the layer. Faunal remains were present throughout, including in association with the burial. A hearth/fireplace was found under and some distance apart from the osteological materials, which had a conch, bone awl, and turtle bones in association. A 2.75 feet (83 cm) thick layer of crushed shell (Layer 6), was found under the second burial layer. Two bone implements, Aresembling those of the Swiss lakes,@ (FN3: 131) were found in this layer. Under this crushed shell was found a 2 foot (61cm) thick layer of Adirty sand@ and shell (Layer 7). Two thin lenses of shell and soil mixed were under this dirty sand layer (Layers 8 and 9). At the base was sterile sand (Layer 10).

Excavation II

This excavation was located 88 feet (27 m) north of Excavation 1. The horizontal dimensions of the unit were 7 feet (2.13 m) by 5 feet (1.5 m). The unit was terminated at 7 feet (2.23 m) in depth, but had not reached sterile soil at this point (Figure 42). The first layer is described as 1 foot (30 cm) of humus, underlain by 1 foot (30 cm) of sand with a slight intermingling of shell. Two hearths were found in this layer Below this layer was 1.75 feet (53 cm) of shell and sand, underlain by 4 inches (10 cm) of white sand. Under the white sand layer was a layer described as 1 foot, 5 inches (43 cm) of sand and shell. A hearth was found within this layer. A foot (30 cm) of crushed shell was found to underlay the 143 sand and shell, followed by 1.25 feet (38 cm) of loam and shell. A large point was recovered near the base of this layer. The unit was terminated at this point, but the deposits likely continued. Pottery was found to a depth of 5 feet (1.5 m), but only two sherds were found below 4 feet (1.23 m). It seems possible that these sherds were disturbed from the above stratum, particularly given the excavation techniques. The layer of white sand seems to correspond to the one found in Excavation I. There is no mention of human remains being found in this unit.

Horizontal Arbitrary

OM92U2_1 Humic/Topsoil

Sand, slight admixture OM92U2_2 of shell Orange

Hearth/fireplace (charcoal) OM92U2_3

OM92U2_4 Hearth/fireplace Shell and sand (charcoal) OM92U2_5 Thornhill Lake?

White sand OM92U2_6

Brown sand and shell OM92U2_7 Hearth/fireplace (charcoal)

OM92U2_8

Broken/crushed shell OM92U2_9 Mt. Taylor Mt. Taylor

Loam with shell OM92U2_10

Deposits Continue

0 50 100cm

Moore, Orange Mound, Excavation 2 Feb.. 7-8, 1892. Section based on written description, FN1: 59, FN3: 133, 1893c Figure 42: Moore’s Excavation II Excavation III

This excavation is described as being located 100 feet (30.5 m) from Excavation I on the southern slope. The horizontal dimensions of the unit are given as 6.5 feet (1.98 m) by 5 feet (1.52 m) to 7 feet (2.13 m) deep, where the remainder of the unit is a Asmall hole@ to 10 feet (3.04 m), 144 where water was encountered, but sterile soil had not been reached (Figure 43). The first layer of Excavation III is described as 6 inches (15 cm) of humus. The second layer is described as 1.5 feet (46 cm) of loam and shell. Under this layer was a 3 foot, 8 inch (1.11 m) of apple snail, with a small amount of other shells. The remainder of the unit was in loam and powdered shell. Pottery was not found below 6.5 feet (1.98 m).

Horizontal Arbitrary

OM92U3_1 (UMIC4OPSOIL

OM92U3_2 ,OAMANDSHELL

3HELL MAINLYAPPLESNAIL OM92U3_3

,OAMANDCRUSHED POWDEREDSHELL OM92U3_4 Preceramic (?)

7ATER%NCOUNTERED $EPOSITS#ONTINUE

  CM

Moore, Orange Mound, Excavation 3 Feb.. 8, 1892. Section based on written description, FN1: 60, FN3: 135 Figure 43: Moore’s Excavation III

Excavation IV

This excavation unit was located 8 feet (2.4 m) south of main excavation. The unit measured 6.5 feet (1.98 m) by 7 feet (2.13 m) horizontally. The first 6 inches (15 cm) is described as humus. The next 5 feet (1.52 m) consisted of almost pure apple snail. The third layer was a thin lens of white sand, but Moore=s discussion of this layer is confusing. Below this sand was excavated an additional 3 feet (91.44 cm) of loam and powdered/crushed shell. The unit was terminated at this 145 point while still in this layer (Figure 44). Within this layer was found disassociated human bones. Animal bones and fireplaces were found in this unit, but no specifics are given. Implements of bone were found in the unit, but again, no specifics were given. No pottery was recovered after 5.5 feet (1.68 m).

Horizontal Arbitrary

(UMIC4OPSOIL OM92U4_1

3HELL MAINLYAPPLESNAIL OM92U4_2

OM92U4_3 3AND

h$ISASSOCIATEDv HUMANBONES OM92U4_4 ,OAMANDPOWDERED CRUSHEDSHELL Mt. Taylor

$EPOSITS#ONTINUE   CM

Moore, Orange Mound, Excavation 4 Feb.. 9, 1892. Section based on written description, FN1: 60-61, FN3: 135-137 Figure 44: Moore’s Excavation IV

Moore’s 1893 Excavations Moore returned to the Orange Mound during February 1893, stating A...two days of the winter of 1893 were devoted to serious excavations@ (1893c: 615). Unfortunately, little information exists from this work. In Field Notebook 4, the entire February 25, 1893 entry consists of the heading Orange Mound, Excavations 4 & 5, followed by the short paragraph: 146 AIn Excavation 4 no pottery was met with after 3 ft. from the surface. Human bones were encountered in fragmentary condition scattered in various portions of the excavation. Also burial in anatomical order covered with small layer of white sand 2 2 feet down (105).@ In Field Notebook 7, an equally brief notation for the 1893 work is provided. Following the headings, Feb. 25th, Orange Mound, Excav. 4, the notes read:

AX on summit ridge N of main ex & running (struck out by Moore) extend down W slope. No pottery after 2 2 ft. Scattered human bones 2 ft. down all over exc. a(sic) bout same level burial 2 ft. covered with white sand (53).@ No further details are given in the field documentation. There seems to be no means to locate

this one partially described unit. It is unclear if Moore is continuing the numbering system of 1892, and accidently started with 4, having forgotten the number of units previously excavated, or if the numbering system starts again afresh and there are three completely unknown excavation units. If there were two days of serious excavations, Moore, given his methodology, could have easily excavated several units. Moore apparently suffered some depredations during these few days. His log entry for February 24, 1893 discusses the journey from Lake Cone to Orange Mound. The steamship Alligator that Moore was using at the time became wedged in vegetation while turning a corner within sight of the mound. A long day of effort to dislodge the vessel was finally successful after 7 2 hours, by

which time the day was wasted. The next day, the 25th, was spent excavating, but it rained all day. Perhaps the stress of the previous day combined with a long day of rain was too much, and notes took on a secondary importance (FN 7: 204-202)5.

Other Visits Rouse, during the course of compiling his survey of the region, apparently visited Orange Mound. He describes the site as an Aisland in the midst of marshy country,@ greater than 3 miles from the nearest dry land and only reachable by watercraft. Rouse immediately dismisses the idea that the heap represents a mound, stating the site is actually a shell midden, without any evidence

5 Moore employed the practice of beginning his daily logs at the end of the pre-numbered ledgers he used for his large notebooks, turning the notebook upside down and beginning with the last page. The page numbers are thus reversed. 147 to support this assertion (1951: 127-128). In August of 2005, the author, in the company of Myron Estes, John Lieb, and Michael Stallings, visited several sites in the upper St. Johns, including Orange Mound. A general reconnaissance of the site was made, including a determination of the mound elevation. Measurements in the field indicate that the mound apex is ca. 4.5 meters. Some previous excavations are evident, maybe Moore=s units. The surface is littered with pottery sherds, in remarkable numbers, primarily fiber-tempered types, with a lesser amount of St. Johns types. A few flakes were observed on the surface. No collection was made.

Co n s t r u c t i o n o f t h e Or a n g e Mo u n d Orange Mound is considered by some to be in part a Mt. Taylor mortuary mound, primarily due to the sand in the burial layers (Wheeler, Newman, and McGee 2000: 153), but others deny that there is intentional construction (Rouse 1951). Moore=s interpretation of the sand layer at Orange Mound is instructive.

AThat burials took place at Orange Mound in a regular stratified mound of sand is beyond the shadow of a doubt, and that the burial mound was not made upon an abandoned shell-heap, perhaps long after the eaters of shell-fish had passed away, is irrefutably shown by the three feet of ampullariae piled above. It must therefore admitted that the aborigines of the shell heaps, in one instance at least, interred their dead in a sand mound. . ., the writer considers it so unlikely that the stratified burial mound . . . at Orange Mound is an isolated case. . . (Moore 1892c: 622-623). Stratums VIII, IX, and X appear to be the first usages of the site, and are presumed Mt. Taylor period middens. Only Excavation I extended into these strata, and this hole had Aconverged@ to a smaller excavation prior to reaching this level, 5 feet (1.52 m) by 3 feet (.91 m), and given the depth of the unit at this point, would have been difficult to excavate and record throughly. Little can be said about these lowest strata. Stratum VIII is described as loam and shell, with faunal remains, but little other detail is provided. A tentative stratigraphic linkage for this stratum between Excavations 1 and 2 is shown in the Harris Matrix (Figure 45), but Moore gave slightly different descriptions of the respective layers. The lowest cultural layer found in Excavation I (OM92U1_12) is another candidate for 148

Stratum I OM92U2_1 1 OM92U1_1 1 OM92U4_1 1 OM92U3_1 1 Topsoil

Stratum II OM92U2_2 2 Loam and OM92U3_2 2 Shell

Ceramic Finds

OM92U1_3 3 OM92U4_2 3 OM92U3_3 3 Localized Stratum III OM92U1_2 Charcoal Apple Snail

OM92U2_4 4 OM92U1_4 4

OM92U2_3 Localized Stratum IV Charcoal Human OM92U2_5 Remains Thornhill Lake (?)

Stratum V OM92U2_6 5 OM92U1_5 5 OM92U4_3 5 White Sand

OM92U2_7 6 OM92U1_6 6 OM92U4_4 6 OM92U3_4 6 Stratum VI

OM92U2_8 7 OM92U1_7 7 Human Charcoal Remains

Stratum VII OM92U2_9 8 OM92U1_8 8 Crushed Shell

Mt. Taylor Deposits Stratum VIII OM92U2_10 9 OM92U1_9 9 Loam and Shell

Stratum IX OM92U1_10 Shell

Stratum X OM92U1_11 Shell and Loam Sterile Sand Figure 45: Harris Matrix of Orange Mound 149 being contemporary with OM92U2_10, with two intervening layers found only in Excavation I. Unfortunately, Excavation II does not extend into the sterile underlying sand, so the relationship remains ambiguous. Stratum VII is a layer of crushed shell, characteristic of midden accumulation. Moore reports finding scattered hearths and animal bones. This stratum is found in Excavations 1 and 2, varying between 30 and 75 cm in thickness. Water was encountered in Excavation 3, and it is doubtful if this layer extended far south. Excavation 4 quite probably would have penetrated this stratum if digging had continued. Stratum VII is a Mt. Taylor midden, primarily found in the northern area of the site crescent. Stratum VI is the earlier of the two burial strata. This layer is composed of Aloam@ and powdered/ crushed shell, and is 65 cm thick at a minimum, and substantially thicker in some areas. This stratum is found within all the units excavated by Moore, with human remains being found in Excavations 1 and 4. At least two burial modes are represented in this stratum. In Excavation 1, a primary burial was found, in a decayed state. If this burial was extended or flexed is unclear. Secondary burial of human remains was found in Excavation 4. Burials appear to be concentrated on the apex of the mound, while the accompanying matrix is across most, if not all, the mound.

This layer is conservatively estimated to have a horizontal extent of greater than 2800 m2. In Excavations 3 and 4, Moore ceased digging while still within this stratum or immediately after penetrating this layer. Charcoal lenses were found in Excavations 1 and 2. The charcoal in Excavation 1 was underneath the human remains, but ambiguity in Moore=s documentation clouds the exact association. The charcoal in Excavation 2 is at the same approximate level within Stratum VI, but there no evidence that this is a charcoal lens contiguous across the mound. More reasonably these are small fires associated with the interment of primary human remains as found at Harris Creek. No ceramics were found in Stratum VI. A Mt. Taylor period placement is estimated for this stratum and associated human remains. The atypical matrix of this stratum, when compared with other Mt. Taylor burial contexts, could be a regional or temporal variation. Possibly, the 150 mortuary ritual represented in Stratum VI is relatively late in the Mt. Taylor sequence, but without radiocarbon dating, such assertions are merely speculative. Stratum V is a localized layer of white sand that underlies the burials found in Stratum IV, and overlies the earlier burial layer, Stratum VI. The white sand is thickest on the peak of the mound, and thins downslope. This stratum must be considered purposeful in the absence of any reasonable natural process to explain its formation. No artifacts were found in this stratum, and it is presumed to date to the Mt. Taylor period, but could be associated with the above stratum. This layer caps the lower burial stratum, or alternatively, is a preparation of the mound surface. Stratum IV is the uppermost of two strata with human remains. This stratum consists primarily of sand intermixed with shell, and is found in Excavation 1 and 2. This would indicate that the stratum is concentrated on the north and center of the mound. Within Excavation 1, Moore encountered extended burials and a bundle burial. Two charcoal lens were found in this stratum; one near the upper interface of the stratum and a second immediately beneath the burials, a pattern found in other Late Archaic burial mounds. No burials were found in Excavation 2, but the soil matrix and stratigraphic position indicate that this is the same stratum. As at Palmer-Taylor, the burials seem to be concentrated at the highest point of the mound, while the stratum covered a

much greater area. At a conservative estimate, over 1000 m2 would be covered with this deposit. The presence of burials and the use of sand mixed with shell for the stratum matrix indicates intentional deposition. The similarity of this stratum to other, later sand burial mounds is explicitly noted by Moore (1983c; 1894b: 96; 1894c: 210-213), who had excavated dozens of these mounds by this time in his archaeological career, and felt that Orange Mound was one of the earliest examples of this practice in the St. Johns basin. Stratum IV may date to the Orange period, but ceramics were not as abundant as in Stratum III. As stated previously, no ceramics are noted as being in association with the burials. Another interpretation is that the upper mortuary stratum at Orange Mound dates to the Thornhill Lake period (Chapter 5). Stratum III is a thick layer of primarily Pomacea shell with a slight admixture of other species. 151 The layer varied in thickness between 95 and 145 cm, with the thickest deposits in Excavation 4. This stratum is not present in Excavation 2. At least one lens of charcoal, referred to by Moore as a fireplace/hearth, is present in this stratum in Excavation 1, but no others are recorded. Fiber- tempered pottery was apparently abundant in this stratum. At a very conservative minimum, this stratum has a horizontal extent of at least 1000 m2. Stratum III could be characterized as a constructed stratum, though not unambiguously so. The lack of crushed shell layers within this stratum, the capping of a layer with mortuary remains, the presence of primarily a single species, and the horizontal and vertical scale of this depositional event could indicate a intensive episodic deposition, rather than a long term accretion of subsistence refuse. However, the apparent inclusion of substantial amounts of ceramics within this stratum and at least one tentative hearth can be interpreted as a evidence of midden accumulation. Alternatively, the stratum could represent the use of midden material for building, a common practice at other Late Archaic St. Johns sites (Harris Creek, Bluffton). Stratum II is a loam and shell layer of varying thickness, approximately 30-45 cm. This stratum is found in two units, 2 and 3. Both of these units are along the slope of the mound. This stratum either represents a site usage, likely during the St. Johns II period, that was confined primarily to only the lower portions of the site, or the stratum represents eroded/displaced Stratum I materials, i.e., eroded topsoil. Orange Mound, like most major sites in the St. Johns, was used for agriculture, specifically citrus. Farming practices may have encouraged erosion. No human remains were discovered in this stratum, but faunal material was likely found throughout. Across the mound is found a layer of topsoil that varies from 15 to 30 cm in thickness. This layer is designated as Stratum I in the Harris Matrix (Figure X). Artifacts recovered by Moore include plain and check-stamped pottery, likely St. Johns types. During the 2005 site visit, Orange Plain and Orange Incised was also observed on the surface. This stratum represents the latest aboriginal usage of the Orange Mound. Stratigraphic analysis of the Orange Mound demonstrates successive usages of the site during the Mt. Taylor and the following Thornhill Lake and Orange periods. A substantial midden is 152 deposited, primarily in the northern portion of the site. Atop this midden, secondary and primary burials are found within a stratum of sand with some shell admixture. This stratum extends across the site where explored, and may be the major building episode that gives the mound its initial crescentic shape, and is the largest construction effort represented in these excavations. A layer of white sand was deposited on the highest point of the crescent, immediately above the mortuary remains. Another sandy burial stratum, dating either to the Thornhill Lake or early Orange period, was deposited on the apex, mainly above the white sand layer. An Orange period deposition of shell capped a portion of the upper burial stratum and adjacent slopes. A final usage of the site by St. Johns II people may have been purely utilitarian.

Pe r s i m m o n Mo u n d (8VO4367)

Setting The Persimmon Mound, also known as the Baxter Mound, is located on the east side of the St. Johns river adjacent to an unnamed lagoon approximately 1.5 km down river from the outlet of Salt Creek/Lake Ruth (Figure 38). The mound is located near the community of Baxter Point at the southern terminus of a large highland remnant that is surrounded by several smaller drainages and

marsh6. Two sites are recorded within 1 km of Persimmon Mound, Arrowhead Ranch (8BR2), a multicomponent site with Orange through St. Johns occupation (Rouse 1951: 144), and Rattlesnake Hammock (8BR3/8BR4), with Orange through St. Johns occupation documented (Rouse 1951: 144). Persimmon Mound is located 2 km directly north of Orange Mound on the opposite bank of the river. The site was described by Moore as a large shell deposit covering approximately 5 acres (2.02 ha). A mound or eminence on the east end of this deposit was the mound excavated and reported (1894b: 95). In his field notes, Moore describes a substantial shell ridge north of the mound and an additional nearby shell heap with human remains that has not been recorded in the Florida Master Site File. This second site is described as an oval shell heap 400 yards west of Persimmon Mound

6 The approximate UTM coordinates of the site center are: UTM Zone 17S, Easting 502970, Northing 3165400, found on the Titusville SW 7.5’ USGS quadrangle. 153 (FN 2: 56-57) and will be referred to here as Persimmon II (Figure 46).

%XCAVATION

3HELL2IDGE

0ERSIMMON "URIAL-OUND

%XCAVATION 0ERSIMMON-OUND))

%XCAVATION

,!'//.

Persimmon Mound Volusia County, Florida 25 0 75 Meters Sketch Map based on written description

V B (Moore FN1: 83-85; FN2: 54-57)

R North and Randolph USJ Survey (1920: 211) Figure 46: Sketch Map, Persimmon Mound and Persimmon II Previous Research Francis Harper places John and William Bartram on Persimmon Mound on the night of January 12, 1765, where they camped prior to heading back downstream (Harper 1942: 73). All likelihood is that the Bartrams camped either here or Rattlesnake Hammock. Jeffries Wyman definitely explored the mouth of Salt Creek, but assessing if he actually visited this particular site requires an aside. Rouse (1951: 143) may be correct that the site called by Wyman (1875) Possum Bluff is not Persimmon Mound, but it appears instead that Wyman was originally referring to a site called Rattlesnake Hammock (1868), a separate site (8Br3) in Rouse=s survey (1951: 144). Sometime soon after, the name Rattlesnake Hammock was changed by Wyman to Possum Bluff for unknown 154 reasons (1875), but it does not appear that Wyman ever referred to Persimmon Mound as Possum Bluff. Moore (FN 2: 52) thinks that Wyman=s Possum Bluff is Indian Fields (8Br5), not to be confused with the North, Middle, and South Indian Fields further south and that the real Possum Bluff is located to the south: 8Br14. Wyman excavated an additional site on the Salt Creek drainage, identified as 8Br9, Indian Mound Station by LeBaron (1884), Thomas (1891) and Rouse (1951: 145), a site near the railroad at Titusville, and there is special mention of the distance of the Salt Creek drainage to access to the Atlantic (Wyman 1875: 16), so Wyman probably approached from Titusville going down the Salt Creek drainage to the confluence with the St. Johns. He refers to a site ARattlesnake Hammock, on Salt Creek, right bank, and the union of the creek and the St. Johns (1868: 26)@. In his major publication in reference to the St. Johns sites, Wyman states:

AThe most southerly of any of the deposits of shell we have see are two, both comparatively insignificant, both on the right bank of Salt or Moccasin Creek near its union with the St. John=s and near the waters edge. The higher is known as Possum Bluff (italics in original), the other is a shell field, sloping towards the water, and has been under cultivation” (1875: 16). If the right bank is referring to the downstream right, then Rattlesnake Hammock is an unverified site on the southwest bank of Lake Ruth, as estimated in the Florida Master Site File (FMSF), and assigned the number 8Br3. The presumption in the FMSF is that this is also the location of Wyman=s Amisnamed@ Possum Bluff (8Br4), and the extra site number is redundant. Two site numbers are given to two unverified sites at this point, based on Wyman=s two accounts. Just to add to the confusion, there is another site in the upper St. Johns, 8Br17, listed in the FMSF as Persimmon Mount. This site is referred to as Persimmon Mound by Russo (1985) and on the 1920 Randolph Survey (Sheet 115). There may be a Mt. Taylor component at this site (Russo 1985: 87- 89), but no mortuary remains were reported in the preceramic depths. A mound is mapped at this location on the 1920 Randolph Survey (Sheet 211) and would correspond to Wyman=s Rattlesnake Hammock/Possum Bluff, but is now a boat landing. If this is one of the two shell deposits on the right bank, then Persimmon Mound could be the shell field sloping towards the water. Moore describes Persimmon Mound as a large shell field that had been 155 under cultivation and that the mound was only a small eminence on the eastern edge of this field, probably plowed down somewhat. No prominence is shown on the 1920 Randolph Survey sheet for the area, but the mounds were not always recorded, and shell fields never are in this survey. A high point is plotted near where the mound should be on the Randolph Survey map. However, if by right bank Wyman is referring to upstream right, then Moore would be right in concluding that Indian Fields was Wyman=s Possum Bluff. The site has a similar configuration as Persimmon Mound, with a large shell field adjacent to a mound. There is not a second site recorded, though Wyman could have been referring to the overall Indian Fields site, with a mound and shell field. The likely conclusion seems to be that Rattlesnake Hammock and Persimmon Mound were the two sites reference by Wyman, who found abundant pottery and animal bones in both sites (Wyman 1875: 16). The calling of the site Possum Bluff appears to be either a temporary change in the site name, misinformation, or a simple mistake. A Busycon gouge is in the R. A. Mills collection at the U.S. National Museum which was donated in 1890 (Rouse 1951). This gouge is listed as coming from Persimmon Mound. A socketed antler point was found by Rouse in the collections of the Wagner Free Institute, and is listed as coming from Moore=s Excavation 2 at Persimmon Mound (Rouse 1951: 143). Rouse finds that this conflicts with Moore=s statement that Anot a single fragment of pottery was found, nor was an implement of any sort brought to light (1894b: 96)@. Mitchem suspects that Moore made a later unreported visit (1999: 28). However, as described below, Excavation 2 was in a shell ridge north of the mound, and there is nothing in the field notes to contradict the statement that no material was found in the burial mound or to indicate that a second trip to the site was made. Moore describes Persimmon Mound as a Ashell field cultivated & cleared by the later Indians on a lagoon about 20 miles S. Lake Harney (FN1: 83)@. In Field Notebook 1, the site is described as being 300 yards (274 m) from the river bank, but in Field Notebook 2, the placement is 400 yards (366 m) from river. In the published report, Moore describes the site as 2 miles (3.2 km) north of Indian Fields (8Br5) and 400 yards (366 m) from the channel (1894b: 95). The site as placed in the FMSF is close to this location, being about 360 meters from the historic channel, 156 but on the shore of a lagoon to the west and south of the site. The FMSF indicates that the site has not been visited, and Rouse does not mention visiting the site, though he and Goggin did visit the adjacent Arrowhead Ranch site (1951: 144). Immediately west of the site as plotted is Baxter Point, presently a small community that extends into the lagoon. The alternate name Baxter Mound for Persimmon Mound derives from this geographic feature. There is no specific indication of Persimmon Mound on the Randolph Survey map (Sheet 211). A small plateau is mapped above the 12 foot (3.7 m) contour line adjacent to the lagoon, dropping steeply to the water at the bank. The area suspected to be the site is shown as under cultivation, likely for citrus. One high point is indicated on the survey sheet, near the center of the estimated site polygon. This point is recorded as 17.3 feet (5.3 m) in height, and may be the Aprominent elevation@ encountered by Moore (FN1: 83). A sketch map based on Moore=s description presumes this point to be the mound location (Figure 46).

Moore’s 1892 Excavations Excavation I is described as being at the apex of a prominent elevation located on the east end of a shell field approximately 5 acres (2.02 ha) in extent (Figure 46). Agricultural activities had likely diminished the height of the mound considerably. The excavation measured 9 feet (2.7 m) by 8 feet (2.4 m) and 7 feet (2.1 m) in depth until through the shell. Surface pottery was abundantly found, but no pottery was recovered in the excavation (FN1; FN2). ABroken human bones as though battered by plough from start (FN1: 83).@ The mound matrix is described as being composed of shell, crushed shell, and sand. A layer of white sand was encountered on the north side of the excavation at 1 foot (30 cm) below the surface. At this point, the white sand was 3 inches (7.6 cm) in thickness. The sand layer sloped to the south, while increasing in thickness. In the southern section, the white sand layer was 3 feet (91 cm) below surface and 1 foot (30 cm) in thickness. The sand layer pinched out in the southwest corner of the unit. Some concretion was encountered at the base of the mound (Figure 47). There appears to be two or more mortuary layers, though Moore does not explicitly designate more than one. Human remains are encountered from 6 inches (15 cm) below surface to 3 feet (91 157

Surface Contour Arbitrary Pottery restricted to plowzone Whole, crushed, and powdered PM1892_EX1_7 shell and sand PM1892_EX1_8

White Sand PM1892_EX1_6 Fragmented human bone, localized lenses PM1892_EX1_5 of white sand Burials on Interface

} PM1892_EX1_4 PM1892_EX1_2 PM1892_EX1_3 Interface

Whole, crushed, and powdered shell and sand

PM1892_EX1_1

Sand

Persimmon Mound 0 50 100cm -Schematic Burial Location C.B. Moore Excavation 1, East Section February 22-23, 1892 Section based on written description, FN1: 83-85; FN2: 54-57 Figure 47: Moore’s Excavation I, Persimmon Mound

cm) below surface. Some confusion arises from Moore=s emphasis on ca. 8 skeletons that were laid directly on the white sand layer described above (1894b: 95-96). These burials were lain on the white sand and covered with shell heap material. The mode of burial could only be determined in one instance, a flexed interment. Other burials were encountered above this sand, marked by small pockets or localized deposits of white sand that were across the mound where excavated (FN2: 54). ASmall irregular layers & pockets of white sand begin at 1 foot down & extend to 3' down these (sic) mark the burials (FN1: 83).@ An additional burial was found 1 foot (30 cm) below the white sand layer on the north side of the excavation. Adjacent to the burial was a pocket of bluish sand, likely altered by contact with heat. Burials were found only near the level of the sand. Excavation to the base of the deposit revealed no further stratigraphic divisions other than the presence of concretion immediately above the sand. It is not clear if the underlying sand was sterile. Moore=s Excavation II was located on a shell ridge ca. 50 yards (45.72 m) north of the burial mound. The entire 5 acres (2.02 ha) was covered with shell, but Moore remarks that the distribution 158 of the shell is unequal across the field. However, he only specifically notes the mound and ridge. On the summit of this ridge was dug an excavation measuring 8 feet (2.4 m) by 6 feet (1.8 m) by 7.5 feet (2.3 m) in depth, but failing to reach the bottom of the shell. Pottery was only found within the plow zone, and none below this point. At 38 cm below surface was the socketed antler point found in the collections of the Wagner Free Institute by Rouse (1951). Only fireplaces and animal fauna were found in the remainder of the deposit. No human remains were encountered. The shell ridge north of Persimmon Mound did not reveal human remains in the limited excavations conducted. Composed of midden material, it is the prominence of this feature that encouraged Moore to excavate. Shell ridges and causeways are a common feature of shell sites throughout Florida, including many Mt. Taylor sites (Wheeler et al. 2000). The role of these shell ridges is not yet determined, but there are similarities with shell rings and crescents found at village sites (Thompson 2006).

Co n s t r u c t i o n o f Pe r s i m m o n Mo u n d Data are limited from Persimmon Mound. We have the single excavation into the heap and the excavation into the separate shell ridge. A Harris matrix has only been constructed for Excavation I (Figure 48). Three strata are designated in the matrix, and to assist in elucidation, the burial below the white sand, the interface of the white sand and the overlying shell, the burials on this interface, and two burials with localized white sand have been designated as stratigraphic units. All eight burials on the interface are for the purposes of the matrix considered a single unit, while the choice of two localized burial/white sand associations is arbitrary, as Moore does not give any estimate of the number of this type of interment encountered. Stratum III has the characteristics of a midden. Found directly atop the sand, Moore only described the matrix as being the same as the shell heaps. One burial was found in this stratum, near the very top, and directly below a layer of white sand. With the present state of data, nothing certain can be said. Lain atop Stratum III is a substantial deposit of white sand, Stratum II. Moore did not provide 159

Localized PM1892_EX1_6 Localized Burial-White Burial-White Sand Deposit Sand Deposit PM1892_EX1_7 PM1892_EX1_8

Burials Stratum I Atop White PM1892_EX1_5 Mortuary Sand Stratum

Interface PM1892_EX1_4

Stratum II PM1892_EX1_3 White Sand

Secondary Stratum III Burial Near Midden or PM1892_EX1_2 Fired Sand Mortuary Deposit Stratum? PM1892_EX1_1

Figure 48: Harris Matrix, Persimmon Mound much description of this deposit, but it was similar enough to sand burial mounds elsewhere in the St. Johns for him to consider it an intentional, if mystifying, burial mound (1889c: 173). Similarities between the white sand layer at Persimmon Mound, Harris Creek, and Orange Mound support a thesis that the sand is part of a regularized mortuary program. Resting on the interface of Stratum II and the overlying shell of Stratum I were a minimum of eight skeletons, at least one of which was a flexed primary burial. The packing and concreting of the shell prevented determination of the mode of interment for the other burials. No grave goods are reported being found with the burials. Entombing these burials was a layer of primarily shell, Stratum I. Within these shell were an unspecified number of burials and localized deposits of white sand. An association of white sand and individual burials is also noted for Harris Creek. Plowing in historic times had broken up the bones near the surface. Ceramic finds were restricted to this same plowzone. 160 Evidence for construction is limited for Persimmon Mound, due mainly to the small amount of excavation data available. However, the layer of white sand, Stratum II, is intentionally deposited. Burials were placed atop this sand and then must have been buried by the shell, as there is no indication that the bones were left exposed. The continued use of the stratum (Stratum I) for burial is evident, and is parallel to practice at Harris Creek and the Palmer-Mound. The evidence from Moore=s excavations indicate that the entire shell field, mound, and ridge all date to the preceramic, probably the Mt. Taylor period. Pottery was restricted to the surface and plowzone in all contexts excavated. There is no direct evidence for the depth of the deposits in the shell field, but the depth of 2.3 m within the shell ridge without reaching the base of the shell would seem to indicate that a substantial amount of material was in place and is likely indicative of a substantial midden.

Pe r s i m m o n Mo u n d II

Moore’s 1892 Excavations An excavation was made in an oval shell heap ca. 400 yards (366 m) west of Persimmon Mound. This mound is not previously reported in any of Moore=s publications or in the Florida State Site File. This distance would place the heap (Persimmon Mound II) beyond the 2 ha shell field surrounding Persimmon Mound and likely to the west of the lagoon. The heap measured 200 feet (61 m) by 25 feet (7.6 m) and approximately 6 feet (1.8 m) in height. An excavation in the summit of this shell heap (Figure 49) measured 8 feet (2.4 m) by 5 feet (1.5 m) to a depth of 4.5 feet (1.4 m) through an unspecified shell matrix and then continued as a smaller hole through black sand mixed with shell another 1.5 feet (46 cm). Between 30 and 46 cm below surface human bone intermingled with animal bone was recovered. At least some of the human bone was found lying upon a Afireplace and showing the marks of fire (FN2: 57)@. No pottery or other artifacts were found anywhere in the excavation.

Co n s t r u c t i o n o f Pe r s i m m o n Mo u n d II The description of the excavations at Persimmon II is limited compared to Persimmon Mound, 161

Horizontal Arbitrary Pottery restricted to plowzone

Broken human bones in shell conglomerate, associated with { hearths and animal bones Whole, crushed, and powdered PMII1892_EX1_1 shell and sand

Black sand PMII1892_EX1_2 mixed with shell

Deposits Continue?

0 50 100cm

Persimmon Mound II C.B. Moore Excavation 400 yds west of Persimmon Burial Mound February 23, 1892 Section based on written description, FN2: 54-56 Figure 49: Moore’s Excavations, Persimmon Mound II but some important details are provided. The presence of burned and broken human bone is similar to the situation found at the Palmer-Taylor Mound. Evidence of fire underlying the human bone is paralleled at Palmer-Taylor, Orange, and Harris Creek. From the limited description it cannot be determined if grave pits were present. Of the sites reviewed, Persimmon Mound II best fits a scenario of accretive midden with human remains.

Sh a r e d St r a t i g r a p h i c At t r i b u t e s

Areally Extensive and Continuous Strata The presence of large-scale continuous strata is usually considered a prerequisite of intentional mounding (Russo 1991; 1994a; 1996, Gibson 2006), but volume is not essential for monumentality (Bradley 1993). The Mt. Taylor heaps considered here have substantial continuous strata that typically covers hundreds of square meters (Table 4). Some strata at these sites are accretive middens, in particular the shell bases of the mounds, but those directly related to a regularized 162 c i t n e c s e r C / n a e B e a n / C r e s c e n t i c - y e n d i K i d n e y - e p a h S h a p e * * s u o u n i t n o C y l l a n i g i r O r i g i n a l l y C o n t i n u o u s d n a e v i s n e t x E y l l a e r A r e a l l y E x t e n s i v e a n d a t a r t S t r a t a * * * * * s t i s o p e D y r a u t r o M o r t u a r y D e p o s i t s h t i w d n a S e t i h W h i t e S a n d w i t h * * * o t n i s t i P e v a r G r a v e P i t s i n t o a t a r t S y r a u t r o M o r t u a r y S t r a t a * * * ? ? y b d e i r u B / d e d n u o r r u S u r r o u n d e d / B u r i e d b y s n i a m e R n a m u H u m a n R e m a i n s x i r t a M a t a r t S t r a t a M a t r i x * * * * ? p u o r G / y t i n u m m o C o m m u n i t y / G r o u p l a i r u B u r i a l * * * * * t c a f i t r A r t i f a c t y t i s r a p S p a r s i t y * * * * * g n i t s a e F / d o o F o o d / F e a s t i n g s n i a m e R e m a i n s * * * * * l l e h S n a e l C l e a n S h e l l a t a r t S t r a t a * * * * Shared Structural Attributes of Mt. Taylor Mounds Attributes of Mt. Taylor Structural Shared l a o c r a h C h a r c o a l s e s n e L e n s e s * * * * ? Table 4: Table n e k o r B / t n r u B u r n t / B r o k e n e n o B n a m u H u m a n B o n e * * * h t i w d e t a i c o s s A e r i F i r e A s s o c i a t e d w i t h s t i s o p e D y r a u t r o M o r t u a r y D e p o s i t s * * * * * d n u o M r o l y a T - r e m l a P a l m e r - T a y l o r M o u n d d n u o M k e e r C s i r r a H a r r i s C r e e k M o u n d d n u o M n o m m i s r e P e r s i m m o n M o u n d e m a N e t i S i t e N a m e d n u o M e g n a r O r a n g e M o u n d I I n o m m i s r e P e r s i m m o n I I 163 mortuary program are horizontally expansive without recorded discontinuities. Continuity within archaeological strata relies on a simple principle of lithographic studies: Asoil changes along a stage boundary indicate prolonged exposure; if none is detected or if they are thin and chemical, then weathering has been slight or non‑existent@ (Gibson 2006: 318). The lack of facies or interfaces at angles from horizontal within strata, excluding post-depositional disturbances (e.g., posts, burial pits, modern disturbances), is a common feature between Mt. Taylor mound sites and evidence of single episode large-scale mounding events. One difficulty in recognizing construction at shell sites is that the pace and timing of building is inversely related to the volume of deposits. In built layers, the most abundant matrix can be and often are deposited rapidly, with little indication of interruption or weathered facies. Rapid construction and burial is necessary over the commingled broken and burnt human bone deposits at Palmer-Taylor and Persimmon II and over the large grave pit below Layer 3 at Harris Creek; exposed human bone would be scavenged and scattered if left to the elements. The white sand layers at Orange and Persimmon Mound are most likely very rapid accumulations, and were likely rapidly buried themselves by extensive strata of brown sand and shell.

Human Bone Entombed by Extensive Shell Strata Rapid burial and stratum accumulation is evident at most, if not all, the upper St. Johns mound sample. There is clearly the entombment of human bone by stratum matrix at the Palmer-Taylor Mound and Persimmon Mound. At Palmer-Taylor, a large fire was built on a surface and human bones placed into this fire, resulting in a bed of broken and burnt commingled osteological material. Human remains continued to be piled onto this bed, as we find secondary remains directly above those on the fire, but these bones were unburned. This burnt surface, covered in promiscuous human bone, was then quickly buried by a matrix of sand and shell. Moore reports at least eight burials lain directly on a localized deposit of white sand at the Persimmon Mound, which are then covered with an unspecified shell matrix (FN 1, FN 2, 1894b). Additional burials were found above and below this white sand. At Harris Creek, at least some of the burials were placed on the upper surface of Layer 2, 164 covered with white sand, and then, after an unknown period of time, all the piles of white sand, including that surmounting the large grave pit, were covered in a matrix of brown sand and shell. The large white sand pile covering the large grave pit and the additional burials covered with piles of white sand represent quick deposition. Subsequent Coverage of the white sand/initial burials by a single-stage construction (Layer 3) effectively encapsulated all these burials into a single, horizontally expansive stratum, Mortuary A (Aten 1999: Figure 6, 144). At Orange Mound in the lower mortuary layer, Stratum VI, it appears that the initial burials were placed on the crushed shell surface of Stratum VII and then covered with a brown sand and shell matrix. All the human bone encountered by Moore was found within 15 cm of the base of the stratum. If any human bone was buried by stratum matrix in the upper mortuary layer is unclear, as is the timing of the placement of the white sand deposit that separates the two mortuary layers. The placement of at least some burials on the white sand and then the encapsulation of these burials with a large single stage stratum construction is expected.

Grave Pits into Extant Strata Recorded interfaces indicate that the interim between construction episodes likely represent several years, if not generations. Interfaces in Mt. Taylor mounds are likely exposed for years, though there are little indications of habitation, refuse accumulation, or extensive tramping of these interfaces. The digging of grave pits into preexisting strata is the dominant interment strategy at the Mt. Taylor mound sites discussed here. These pits are not typically dug into extant middens, but into strata that entombed previously deposited mortuary remains; these were already burial strata, the inclusion of graves into these strata was a continuing use for funerary activities. The two exceptions, the earliest large grave pit at Harris Creek and the lowest burial at Persimmon Mound, are associated with major mortuary construction, and are the earliest interments of this construction. At Harris Creek, a rapidly deposited shell and soil stratum (Layer 3) was placed over and encapsulated an undetermined number of burials covered with discrete mounds of white sand. 165 After the construction of this stratum, a persistent surface existed. Into this persistent surface, numerous grave pits were dug into the underlying matrix. In some cases, small fires were built or transferred into the grave pits, and flexed burials put into the pit while the fire still burned. Some bones have limited scorching/burning, but this is likely incidental, and not an attempt to cremate the skeleton. At least one mass of disarticulated human bones appears to have been buried by Layer 3. In Mortuary B at Harris Creek, it is not clear if all the burials encountered were dug into the underlying stratum (Layer 4). Aten states the Afor the most part@ the grave pits were excavated into Layer 4 (1999: 173-174). Layer 4 does cap Mortuary A, and may have encapsulated burials placed on the surface of Layer 3, but this is somewhat speculative. Aten=s analysis indicates that there is important variability in the kind of flexing of the burials at Harris Creek. Approximately 60 percent of the flexed burials were placed into the pits vertically, i.e., with the long axis of the flexed bones oriented to the vertical, generally with the crania up. The vertical burials were most always tightly flexed, indicating either decay or dehydration of flesh, but not active removal of flesh. The burials were probably wrapped in textiles or skins; the occasional find of proximal bone pins may be a signature of this wrapping. Horizontal flexed burials comprised 38 percent of the mortuary population where mode of interment was identified. These burials tend to be more loosely flexed, and are suggested to have been interred soon after death. Two bundled burials and one extended burial (found in Mortuary A) were the remaining modes of interment specified (Aten 1999: 174). At Palmer-Taylor, there is one extensively documented instance of a grave pit dug into the persistent surface of an extant mortuary stratum that encapsulated commingled human remains (Stratum II). Encountered by the 1976 Rollins field school and referred to as Burial 1, this pit was found partially exposed in the vertical cut of a recent mechanical excavation for shell material (Figure 50). The body was placed horizontally and flexed into the base of the pit. No burial accompaniments were recovered. Extensive investigation failed to find the cranium of this individual, and the opinion of the archaeologist was that the skull had not been removed by a later excavation, either recently or in the archaeological past (Stewart 1976). This was the only pit 166

Figure 50: Burial Pit, Palmer-Taylor Mound burial reported at Palmer-Taylor, but the Excavators= Club may have encountered a flexed burial at the approximately same stratigraphic position (Rouse 1951), and both Rollins field schools recovered solitary human teeth and bones in Stratum II. The presence of multiple extended burials in the upper mortuary stratum (Thornhill Lake period) at the Orange Mound demonstrates clearly that a persistent surface existed at the upper interface of Stratum IV, into which graves were dug and the recently dead placed. In the lower Mt. Taylor period mortuary stratum (Stratum VI), there is one possible example of a grave pit dug into the matrix. Moore (FN1; FN3) reports a single decayed extended burial in Excavation 1, but scattered human bones in Excavation 4. All the osteological material was found within 15 cm of the base of the stratum. No burials are reported for the upper portion of Stratum VI, but there is a reasonable possibility of later burial through a persistent surface of Stratum VI. If that is the case, then the white sand layer (Stratum V) would most reasonably be closer in time to the upper mortuary stratum. 167 While excavating at Persimmon Mound, Moore encountered burial remains at a shallow depth, and continued finding human bones until finding the eight discrete burials atop an underlying white sand layer (1894b: 95-96). At least some of these upper remains could represent excavation into the persistent surface of the shell stratum covering the eight burials on the white sand. There are not clear indications of burial pits being dug into any stratum at Persimmon II. Moore only reports human bone intermingled with animal bones between 30 and 46 cm below surface, with some of the human bone at the base lying on hearth and with signs of burning (FN2: 57).

Group/Community Burial There are indications of multiple burials in a single deposit at all the sites described in this dissertation. At Harris Creek, at least three individuals were placed together in the base of the large grave pit associated with Mortuary A. At the top of this pit, an additional eight persons were buried at a single time, and then covered with white sand. Additional multiple graves are reported, though not all are thought to represent actual multiple interments; some may be later graves dug into preexisting grave pits. There is at least one clear example, Burial 58, of a grave pit with multiple individuals separated by thin lenses of white sand. There was additionally a pile of disarticulated human bones, or bone mass, measuring 90-by-90 cm with an overall height of 40 cm (Aten 1999: 174). At Palmer-Taylor and Persimmon II, the presence of commingled human bones is a clear indication of community burial. Intermixed osteological material, with no signs of later disturbance or rearrangement, parallels the bone mass at Harris Creek. Moore interprets these deposits as evidence of Acannibalistic feasts@ (1893c: 612) and felt that the feature in no way represented intentional burials. Obviously, I disagree. There is no clear evidence of group or community burial in the upper mortuary stratum at Orange Mound (Stratum IV). There is a near association of burials, but Moore does not note any masses of bone or combined interments of other burial modes, though a bundle burial is reported (1893c). In the lower Mt. Taylor period mortuary zone, Stratum VI, Ascattered@ human remains are reported in Excavation 4. While not certain, these disarticulated bones could be an example of 168 communal burial. At Persimmon Mound, the eight burials on the white sand layer are interpreted as an example of a community burial. Moore was able to determine the mode of interment for only one of these eight burials, flexed. The close association of the burials, all within a 3.8-x-2.5 m excavation unit, and the placement of the burial directly onto the white sand and then encapsulating those burials supports this interpretation.

Associations of Fire and Mortuary Contexts The association of fire and mortuary ritual is a widespread feature of Native American culture (Yarrow 1878). At all the sites under review, there is direct association between human remains and fire. Fire associated with burial is found in several contexts at Harris Creek. In the large grave pit in Layer 3, a fire was built directly above the skeletal remains of two adults and a child. At the top of this pit, charcoal was found in the matrix containing eight crania. Fires had been placed in several of the small grave pits immediately prior to interment, on occasion resulting in limited burning of the bones. The large black lens atop Layer 3 may have been primarily charcoal, either as a result of a large conflagration or several smaller fires. At the Palmer-Taylor Mound, burned human bone was found in Stratum II and in direct association with an underlying charcoal lens. Based on the observed occurrence of this charcoal lens in excavation units from all the projects conducted at the site, this charcoal is estimated at nearly 570 meters square and at times as thick as 15 cm. Below and in contact with the charcoal were human remains. This layer was found by all excavators and there are no breaks recorded in the lens across the mound. At Orange Mound, charcoal lenses are reported by Moore in Excavations 1 and 2. It is unlikely that there is one large layer of charcoal as at Palmer-Taylor, instead, it appears the charcoal is localized. Moore does not report any lenses of charcoal at Persimmon Mound, but he may have simply failed to note their presence. At Persimmon II, a lens of charcoal underlay and was in contact with the human remains. The extent of the charcoal cannot be interpolated beyond the unit limits, but the basic configuration is reminiscent of Palmer-Taylor, and the charcoal may have 169 extended much further. In Stratum VI at Orange Mound, human remains were found in close association with Afireplaces,@ i.e., localized charcoal lenses. At Persimmon Mound, there is no report of burned human bone, but the presence of heat-altered sand with at least one burial indicates a fire/human remains association. At Persimmon II, human bone was found lying on a hearth with indications of burning. Direct association of fire and mortuary deposits can also be found at the Bluffton Burial Mound in the middle St. Johns basin (Sears 1960).

Associations of Food and Mortuary Contexts Despite the almost universal presumption that shell in most any context is a byproduct of consumption (but see Claassen 1991; 1996), the occurrence of other food remains at the sites described, primarily aquatic and terrestrial fauna, is more limited in distribution. One of the most important distinctions between Amiddens and mounds@ is the association of non-shell faunal remains. In the mortuary strata of Harris Creek and the upper St. Johns Mt. Taylor sites, these other fauna are found in direct association with burials; as a general principle, a diversity of fauna throughout a stratum tends to be a trait of midden. Plant remains would surely be found in burial contexts if looked for, but such studies have yet to occur. Paleobotanical remains from the domestic midden adjacent to the Old Enterprise site included a great diversity of collected species (Newsom 1994; Russo et al 1992). At least some of the consumable plants would be expected in the burial assemblages. Food remains were found in three types of deposits at Harris Creek. Limited amounts of animal bones were found in grave pits, large and small. The ABlack Zone@ is thought by Aten to be a summit platform deposit, and likely resulted in part from organic remains, either from on platform food preparation, or from redeposited domestic refuse (1999: 167). Finally, several hearths consisting of deposits of burned shell and ash, often accompanied by fish bones, were found in Layer 7. All of these deposits are components of the mortuary mound, and all may reflect feasting during funerary rites (Brown 1995: Hall 1997: 32, 35-39). At Palmer-Taylor the upper burials, which are likely pits dug into a preexisting mortuary strata, 170 occasionally include animal remains in the grave pit, but this is not the case for all. In a burial pit excavated by the 1976 Rollins field school, masses of fish bones were found in the pit matrix (Stewart 1976). Faunal material is reported in direct association with the lowest burnt and broken commingled human bone, and may be with found with the intermediate commingled human bone. Some faunal material was found throughout, but the vast majority of animal bone is found in the more recent topsoil layer, with a substantial reduction in the strata below this (Irving and Morris 1975; Stewart 1983). In the lower strata of the mound, faunal material is almost totally absent. Excavations at Orange Mound revealed a similar pattern. In the uppermost burial layer, human bones were found by Moore in direct association with animal bones. The human bone is whole; no further details regarding the animal bone is given. In the lower burial layer human remains and animal bones were again found in Aimmediate association@ (Moore FN 3: 129). A nearby hearth had conch and turtle bones in direct association. Little other faunal material is reported. The layer of apple snail (Stratum III) above the uppermost burial stratum contained almost pure shell, and Moore does not report finding faunal material in any underlying strata. No faunal remains are reported for the Persimmon Mound, the only mortuary context in this sample without this data class. There are hearths and faunal remains found in Excavation 2 at Persimmon in the ridge north of the burial mound, but no human remains were found in the excavation. At Persimmon II, the second shell mound west of Persimmon Mound, human and faunal remains were found in association, at least at the interface of charcoal and commingled burnt and broken human bone but the documentation is not clear if faunal materials are found with the burials above this interface.

White Sand The presence of white sand in immediate context with mortuary remains is a structural characteristic of Harris Creek, Orange Mound, and Persimmon Mound. White sand is associated with mortuary contexts throughout the entire tradition of mound building in the St. Johns from Middle Archaic through the historic period, and localized or extensive deposits of white sand can be found in most burial mounds, including those of Mt. Taylor age. These sands are likely non- 171 local, transported and deposited to the mound to be used in burial ritual (Russo 1991: 407; Jahn and Bullen 1978; Aten 1999). Transportation and deposition of sand and the direct association with mortuary deposits clearly demonstrates intentionality. The absence of white sand at Palmer- Taylor and Persimmon II is noteworthy. At both sites, the initial mortuary deposit is commingled burnt and broken human bones directly atop charcoal. There may be a temporal component to this difference, the one radiocarbon date available comes from Palmer-Taylor (UM-1155), and using the same parameters as Aten (1999: Table 4) for the Harris Creek material returns a calibrated one- sigma date range of 4672 BC - 4431 BC. The earliest end of this range is nearly a century before the earliest Harris Creek date, which may indicate precedence for Palmer-Taylor. However, this is a single date on marine shell and the earliest deposits at Harris Creek were not excavated, so this relationship is tenuous.

Artifact Sparsity Artifacts are generally rare at Mt. Taylor mortuary mounds. This is certainly due in large part to the absence of ceramic technology; very few artifacts resistive to degradation are produced compared to ceramic using populations. There is however a substantive material culture documented for the Mt. Taylor period, including bone, shell, shark teeth, stone, and wood in a large array of artifact types (Wheeler and McGee 1994a). These artifacts are mainly related to several specific endeavors: textile/hide working, fishing, wood carving, personal adornment, food preparation, and occasional hunting. A wide spectrum of artifacts types were found at Grove=s Orange Midden in waterlogged deposits (Wheeler and McGee 1994a, 1994c, Purdy 1994b), and the extremely high rate of preservation reveals some of the diversity present in the material culture. With the exception of wooden artifacts, most of the artifacts recovered at Grove=s Orange Midden do survive in shell deposits. Bone and shell, two of the most common artifact materials during the Mt. Taylor period, survive very well, as the presence of human bone in a shell matrix clearly demonstrates. In addition, shark teeth and of course, stone artifacts survive very well. Occasional finds of artifacts from these materials in the mounds attest to this survivability. However, there is still a rarity of artifacts in mortuary contexts. At Harris Creek, of the 184 burials recovered, only 172 22 burials had any artifact association, and only 5 of these burials had artifacts that were definitely associated, the rest being possibly associated or near a burial (Aten 1999: Table 5). Considering the volume of mound excavated and explored at Harris Creek, and especially when compared with the volume of excavation at Grove=s Orange Midden, there is a significant paucity of artifacts. This pattern seems to hold at the other sites in the upper St. Johns sample, with the caution that most of the information comes from Moore, who did not use screens. At Palmer-Taylor, a few bone artifacts were found during the Excavators= Club excavations in the preceramic strata, but only a few. A report of artifacts from the 1975 Rollins College work reports a total of only 26 artifacts below the humic layer in seven 2-by-2 m units, and only 5 were below Stratum II. The 21 artifacts in Stratum II are all sherds, and probably result from historic disturbances and excavations. Of the five artifacts below Stratum II, four are bone tools, and one is a almost certainly a displaced St. Johns check-stamped sherd found near the base of the mound (Irving and Norris 1975). At Orange Mound, pottery was found in the upper strata. Below this layer, the only artifacts reported are a bone implement found under an extended burial in Stratum IV, a piece of conch shell that could be an artifact at the base Stratum VI, the lower burial stratum, two bone implements in Stratum VII from Excavation 1 and Asome@ implements of bone in Stratum VII in Excavation 4. No artifacts are reported for Persimmon Mound or Persimmon II, though a shell gouge in the R. A. Mills collection at the U. S. National Museum is recorded as coming from Persimmon Mound (Rouse 1951). A socketed antler point from the shell ridge north of the burial mound is the only other known artifact.

Di s c u ss i o n In this chapter four shell heaps in the upper St. Johns are compared to the Harris Creek and reviewed as possible monuments. Using the Harris Matrix system as an analytical tool, stratigraphic data acquired through excavation are ordered and compared between the sites. Eight stratigraphic attributes are compared between the five localities: 1) Areally extensive and continuous strata, 2) Human bone entombed by those strata, 3) Grave pits that are dug into those strata, 4) Group/ 173 community burial, 5) Associations of fire and mortuary contexts, 6) Associations of food and mortuary contexts, 7) Associations of white sand and mortuary contexts, and, 8) An extreme sparsity of utilitarian artifacts. It is shown that these attributes and their sequence is repeated at multiple sites. Such patterning and regularity is an expectation of intentional construction, not incidental accumulation. 174 Ch a p t e r 9: A Ba s i c Mo d e l o f Mi d d e n Acc u m u l a t i o n

Emphasis in this dissertation is on the upper St. Johns burial monuments, but these are not isolated archaeological features, just a particular sort. Mt. Taylor period people were fisher‑folk, and substantial middens accumulated at all of the sites reviewed in this dissertation prior to the establishment of burial monuments, and likely continued long after the initial construction of monuments. Whereas the monuments show the combined effort of several people and large‑scale material consequent, middens are, as classically framed, the material consequent of many thousands of events; they are the detritus of habitus. Middens are not meaningless though, they are part of the daily landscape and evidence of persistence, at times reaching monumental proportions. A communities success is evidenced underfoot. Large‑scale shell middens are not piles of garbage, they are materialization of a community’s long‑term successes. In the framework of temporalities of experience, the upper St. Johns middens are composed of the material consequence of an untold number of transitory experiences, they constitute a persistent experience. Much is made of the grandiosity of St. Johns preceramic shell sites, with reason. The massive heaps at the Bluffton and Mt. Taylor sites in the middle St. Johns are two of the grandest, but neither contained reported burials (there were associated burial mounds at Bluffton). Preceramic deposits at Bluffton measured 4.3 meters thick at the highest point (Bullen 1955) and covered nearly 20 ha. Only 3 km away, the Mt. Taylor site was composed of 8.3 m of shell deposits, all clearly preceramic, and clearly midden (Moore 1893a: 12‑13, 1893b: 113‑115). We should not make the mistake of equating all these deposits with the builders of the Mt. Taylor burial monuments though. Based on the published radiocarbon dates, the total time of the preceramic, shell‑heap sites (Mt. Taylor and Thornhill Lake periods) was ca. 2800 years. Most of the Bluffton midden dates to the Thornhill Lake period, and I suspect the one burial mound excavated there (Sears 1960) does as well. Even if all of the Mt. Taylor site dated to the Mt. Taylor period, this would still represent 1700 calibrated radiocarbon years.

In near‑aquatic environments, shell is an easily accessible natural resource that is abundant, 175 simple to aggregate, plastic (able to be crushed, ground, or broken down into smaller fragments with simple technology), structurally strong and stable when aggregated, resistant to wear and weathering, has high reflectivity, is flexible in its usage, and is found in a variety of settings. Shell is a composite resource, containing meat that can be consumed and shell that can be utilized in artifact manufacture and for building. Different types of shell have different engineering properties. Thick, robust shells, such as oyster, are difficult to break down, but have bulk and strength. Thinner shells, such as mussels or clams, are not as strong or robust, but the ability to easily fragment these materials makes them attractive for other reasons, such as surface preparation. Shell can be used in an infinite variety of ways (Ceci 1984; Blukis‑Onat 1985), and shell‑bearing sites should be recognized as having such potential. In this dissertation, I am referring to two general categories of shell deposits, middens and monuments. The various ridges, causeways, channels and other shell constructions in the St. Johns are part of the diversity of use, but are not dealt with here (Beasley 2001). Distinguishing between midden and mound cannot be done with a set of simplistic rules. Particularly, classing shell‑bearing sites as midden, without problematizing the issue, is an apriority. Reference to ethnohistoric examples of shell accumulations can be informative, but are quite limited, and researchers are constantly inserting burial into the trait lists (May 2005a: 80). There is, as far as I can discover, in the literature only a single claimed ethnohistoric reference to burial in a shell bearing site (May 2005b: 175), but this is not a direct witness of the practice. This is an archaeological example. Cipriani (1966) describes finding five graves in an old midden in the Andaman Islands. He did not observe burial into middens. Ethnographically contemporary

“pygmy” populations created middens, but kept ancestral bones in homes and on their person.1 We cannot as archaeologists infer from or appeal to the ethnohistoric record regarding a behavior that has never been observed ethnographically.

1 Cipriani=s is a continuation of racist attitudes towards people who created shell middens. Beyond the rhetoric, Andaman Islanders were conceived of as a living example of a Avirtually unchanged or even Palaeolithic or even pre-lithic culture (1966: 153)@. The Onges people are depicted as on one of the lowest rungs of the developmental ladder. A recursive interpretive cycle of primitive attributed to present and primitive attributed to past is still for- warded as evolutionary explanation. 176 Focus in this chapter is on two facets of Mt. Taylor middens. The first is extent and configuration, and the St. Johns middens are compared to the expectations of Thompson=s Evolution of Place model (2006). The second is character and content of the deposits. Two Mt. Taylor sites are discussed here as examples of midden sites, Groves= Orange Midden and Hontoon Island North. Groves= Orange Midden was excavated in the late 1980's and early 1990's under the direction of Barbara Purdy (Purdy 1994a). Hontoon Island North was recently surveyed and tested under the direction of Kenneth Sassaman (Sassaman et al. 2005).

Th o m pso n =s Ev o l u t i o n o f Pl a c e Mo d e l Victor Thompson has recently reviewed models of the accumulation of shell rings found along the lower Atlantic seaboard and circum‑Gulf regions. These rings are several millennia later than the Mt. Taylor period, usually have pottery, and have been the subject of intensive investigation and interpretation (Russo 2004; Russo and Heide 2001, 2002, 2003; Russo, Heide, and Rolland 2002; Saunders 2002, 2004). Thompson reviews the basic models for explaining these rings of shell, classifying the models as gradual accumulation, feasting/mounding, and habitation with monuments. He then gives his own model of shell ring formation that he terms the “evolution of place model” (2006: 272‑273). In this scenario, midden material accumulates around structures that are arranged in a circular or arcuate community arrangement. Continued usage and accumulation at one location eventually results in a ring‑shaped or crescentic midden. Utilizing resistance survey of the Sapelo Shell Ring III and ground‑truthing, there is excellent correspondence between high resistance signatures and shell density (Figure 51). Comparing these data with other archaeological and ethnographic examples of shell midden accumulation in other locations (see also May 2005b), Thompson provides a strong case for gradual accumulation for at least some shell rings (but see Russo 2004). I am by no means taking an essentialist position that all rings, or shell sites of any sort, are midden accumulation. There is substantial debate if these rings are domestic, ritual, or both. In Thompson’s model, the rings begin as domestic debris and “evolve” into monumental architecture. This is not the place 177

Figure 51: Time Slice Model of Resistance Profile, Ring III, Sapelo Shell Ring Complex (Thompson 2006) to engage in this exchange regarding the usage and social roles of shell rings of the southeastern United States, but the general model of accumulation outlined by Thompson has potential for understanding the formation processes of upper St. Johns middens. Restating ethnographic observations, Thompson illustrates how kitchen middens accumulate around individual households (Thompson Fig. 7.1). Structures are rebuilt, material continues to accumulate, and middens around houses eventually become connected into a single feature. A similar pattern of community spatial organization and shell‑heap accumulation can be seen at Jamon shell‑bearing sites in Japan. Generations of households, maintaining a similar intra‑community spatial relation, can over time produce a substantial topographic feature. If a site was used this way generally continuously throughout the Mt. Taylor period, there would be 1700 years of community accumulations. In the upper St. Johns areally extensive middens are found in hundreds of locations along the river channel and tributaries. What number of these date to the Mt. Taylor period is a question for future research, but there are numerous middens with preceramic components and without mortuary 178 monuments throughout the St. Johns basin. Survey in a portion of the upper St. Johns discovered that a large percentage of non‑mound middens have preceramic deposits (Sigler‑Eisenberg 1985; Russo 1985). Middens adjacent to mortuary monuments may reflect villages, or at least restricted mobility, but these are not exclusive harvesting locations. Ray McGee developed a basic model of St. Johns shell heap configurations based in part on work at Hontoon Island (Purdy 1991) experience with Groves= Orange Midden (Wheeler et al. 2000) and the descriptions of Wyman and Moore. Nineteenth century researchers often mentioned shell fields and ridges at the shell heaps, ridges that were often arranged as crescents (e.g., Wyman 1875; Moore FN1). In McGee=s reconstruction (Figure 52), the ridges and fields form site configurations similar to those found with the shell rings and crescentic middens documented by Thompson (2006). There are few accurate maps of Mt. Taylor sites overall, and most sites have later occupations that obfuscate earlier configurations. Fortunately, recent efforts by the anthropology department of the University of Florida have resulted in accurate maps of two middle St. Johns shell heap sites with substantial Mt. Taylor Figure 52: McGee’s Model of Mt. Tay‑ components, Live Oak Mound lor Site Configuration and Hontoon Dead Creek Mound (Figure 53). Substantial differences exist between St. Johns shell heaps and southeastern shell rings, not the least of which most of the shell heaps with Mt. Taylor components continued to be utilized by later populations for another five millennia. Shell rings are utilized for long periods, but not on the scale of Mt. Taylor, and therefore, the vertical scale of Mt. Taylor shell heaps is 179

Rollins Guana Oxeyo Island

Bonita Bay Joseph Reed Horr’s Island

0 50M

Jamon

Hontoon Dead Creek Live Oak Palmer-Taylor Figure 53: Shell Crescents 180 greater. Shell rings also do not have associated burials in shell matrix. However, in general form and horizontal scale, there are specific similarities. Comparing "shell rings" found in Florida with three Mt. Taylor shell sites, Live Oak Mound, Hontoon Dead Creek Mound, and Palmer‑Taylor Mound, we find that in horizontal extent, the sites do not differ greatly (Figure 53). All of the sites have a generally similar orientation, i.e., a rough crescentic shape with a ridge of shell and a relatively clear area within the arms of the crescent. Away from the highest portion of the heaps, burials are not usually found. Instead, there is dense shell and, similar to the shell rings, numerous hearths and faunal material throughout the matrix (Sassaman et al. 2005). The overall configuration of these Mt. Taylor sites is similar to that expected from long‑term occupation and usage of one area by a small community: a village midden with a complimentary burial monument.

Mi d d e n Ma t r i x A stratum may be composed of a consistent matrix, but there are often strikingly different constituents between strata. Some of these strata contain animal remains other than shellfish, but such presence does not indicate domestic activity midden by default (Aten 1999). Relatively pure shell, sand mixed with varying amounts of shell, pure sand, and occasional mucks or redeposited midden characterize monument strata, strata indicative of primarily deposited accretionary midden are usually much more heterogenous, contain many small features, and artifacts are more commonly found in the matrix. Middens of periodic shellfish gathering and processing are characterized by palimpsests of refuse accumulation, thousands of small deposits, expectedly including hearths with each consumptive event (Meehan 1982; Stein 1992). Mass harvesting and processing of shellfish can leave large deposits with little admixture to the shell (Meehan 1982), but will include stage boundaries, and will almost always be repeated at regular intervals. The scale of the strata at the sites under review would require incredible episodes of mass harvesting to create such large unbroken deposits (contra Nodine 1987; Crothers 2004; Crothers and Bernbeck 2004). During 2003 and 2004, a program of mapping, shovel testing, stratigraphic excavations, and 181 limited horizontal excavations was undertaken at Hontoon Island North, a partially disturbed midden located on Hontoon Island, a mid-river island in the St. Johns where several Mt. Taylor sites are found. Hontoon Island North was visited and described by Jeffries Wyman (1875) and Barbara Purdy conducted several seasons of research at the site in the 1980’s. Excavations into the preceramic levels of the site encountered features, organically enriched middens, paleofeces, and relatively abundant artifacts (Sassaman et al. 2005). Also found were localized lenses of burned shell. These lenses have calcined shell, artifacts, and faunal material. Sassaman et al. (2005: 54-56) interpret these lenses as house floors. They also note that these lenses differ from those found at the Hontoon Dead Creek Mound, a nearby shell heap with a probable Mt. Taylor mortuary monument. Burned lenses found there did not include non-shell fauna or artifacts. Based on the work on Hontoon Island, Sassaman et al. have forwarded a hypothetical model of site growth (2005: Figure 3-25). Their model follows the same basic trajectory as the Thompson model described above (Figure 54). At Groves Orange Midden, several units were excavated into water saturated deposits (Russo et al. 1992; Purdy 1994; McGee and Wheeler 1994). A large assemblage of artifacts was recovered, including bone, wood, teeth, shell, and stone (Wheeler and McGee 1994a). Many of these artifact types were found due to preservation in wet conditions, but many Figure 54: Sassaman et al. Model of would also preserve in shell Mt. Taylor Site Structure 182 matrices. Comparing the number of artifacts recovered with those reported from the upper St. Johns burial monuments (Chapter 8), there is a striking difference. Less than a dozen artifacts of any sort are reported from the monuments, from numerous excavations. Comparatively, several hundred artifacts were found at Groves= Orange Midden (Wheeler and McGee 1994a).

Di s c u ss i o n In this chapter an argument is made that the major portion of many Mt. Taylor sites is midden, but not the entire sites. The crescentic arrangement characteristic of Mt. Taylor sites is similar in configuration and content to the sites of fishing people elsewhere, and an expected consequence of practical discard and habitation patterns. Excavations in middens encounter shell with a greater proportion of organics, numerous artifacts, and burned lenses with artifacts and non-shell fauna. In contrast, strata of Mt. Taylor shell monuments contain little organics, very few artifacts, and burning is associated with mortuary practices. While there is no simple checklist to allow one to unequivocally determine if a particular shell deposit is midden, monument, or something else, there are generalities that can help. 183 Ch a p t e r 10: Ri t u a l -Ar c h i t e ct u r a l Re c e p t i o n Hi s t o r i e s o f t h e Up p e r St. Jo h n s Mt. Ta y l o r Bu r i a l Mo n u m e n t s The ultimate objective of archeology is the creation of an image of life within the limits of the residue that is available from the past. Phillips and Willey 1953: 616

This chapter is an application of the program detailed by Jones (2000a) for the study of sacred architecture in the writing of a ritual-architectural reception history of a Mt. Taylor burial monument found in the upper St. Johns basin. A review of the stratigraphy of Orange Mound, Persimmon Mound, Persimmon Mound II, and Palmer-Taylor Mound is given below. Since we have the greatest detail for the Palmer-Taylor mound, that is the site given the greatest attention, and the one site where a detailed ritual-architectural reception history is constructed. Archaeological readings of sacred architecture are always filtered through us, and the sketching of a series of events is, at kindest, a guess. A “narrow narrative,” intended in explaining a few specific archaeological contexts is attempted here in the role of middle-range theory. The basic question of this dissertation is how can we distinguish between a shell midden and a burial monument. Towards achievement of that aim, the data are summarized to give articulation of a local history, with no claim to this being the same history as any other place or time. A potential sequence of events, congruent with the extant data, is what is intended. Nor is a claim made that similar histories are not found elsewhere: each history is singular and comparable. Each of the sites reviewed has a unique history of construction and deposition, but in a given region during a given time period, there are specific similarities and patterning. Like contemporary sacred architecture where each temple, church, mosque, or synagogue will be unique, but will also have certain elements and features that unite it with others of its kind.

Re v i e w o f t h e St r a t i g r a p h y o f t h e Or a n g e Mo u n d Excavations at Orange Mound are limited to the work of C.B. Moore. He describes the site as a crescent oriented north-south. Dimensions for the site are given as ca. 171 meters along long axis and ca. 80 meters at widest point (1893c: 615-616). At the highest point, the mound was ca. 184 4.3 m in height. Human remains were found in Stratum VI and Stratum IV. The lower mortuary stratum is dated to the Mt. Taylor period, and the upper mortuary stratum is dated to the Thornhill Lake period. The primary concern here is with the Mt. Taylor deposits. Atop natural sand, a thin deposit of midden material was initially accumulated (Stratum X), surmounted by another stratum of midden, with more shell than that below (Stratum IX). Above this is a relatively thick stratum of Adirty@ sand and shell midden (Stratum VIII). Stratum VII is nearly 80 cm of crushed shell. Crushed shell is generally indicative of trampling and intensive usage of shell surfaces, and a thick stratum of crushed shell is indicative of multiple episodes of use (Blukis-Onat 1985). Immediately above the crushed shell is the Mt. Taylor mortuary stratum. Above a Ahearth@ was found mixed human bones and associated animal bone in Stratum VI (Moore FN1: 56). Human remains were found throughout at least the bottom 30 cm of the stratum. Based on the Moore=s description, it seems he only found secondary burials. A thick layer of white sand (Stratum V) was lain atop the Mt. Taylor mortuary stratum. Present data do not allow for determining if this sand is more appropriately associated with the Mt. Taylor mortuary stratum below, or with the Thornhill Lake period stratum above. Regardless, the sand was culturally transported and deposited, segregating the two mortuary strata. Stratum IV is the second mortuary deposit, and is marked by extended burials and a much sandier matrix. This stratum postdates the Mt. Taylor period.

Re v i e w o f t h e St r a t i g r a p h y o f t h e Pe r s i m m o n Mo u n d C.B. Moore has conducted the only recorded excavations at the Persimmon Mound site. He describes the site as a large shell field with a knoll at one end somewhat higher than the surrounding shell field but composed of the same material. Additionally, a substantial shell ridge was found on the north side of the field. Cultivation of the whole site by historic Seminole had diminished the mound and disturbed some of the human bone. Two excavation pits were undertaken at Persimmon Mound, one in the eminence and one in the 185 shell ridge. Excavation 1, in the burial monument, reached the end of shell at 2.1 m, presumably this is the underlying sterile sand. Only in this pit was human remains found. Excavation 2, in the shell ridge, encountered over 2 meters of whole, crushed, and powdered shell with some sand before being abandoned before sterile sand was found. No pottery was found anywhere in any excavation. Excavation 1 consisted of between 90 and 180 cm of whole, crushed, and powdered shell, achieving greater thickness towards the apex of the mound. Near the top of this stratum at the mound apex was found portions of a secondary burial in near-association with a Ahearth@ and animal bone. Atop this stratum rests a layer of white sand, 30 cm at its thickest point, downslope. Approximately eight flexed burials were placed directly atop this white sand. These burials were then encapsulated with additional whole, crushed, and powdered shell with some sand admixture. Within this uppermost stratum, human bone broken by plowing was found near the surface. Moore=s description of the stratum is that it is midden, same as surrounding field. It is possible that this is redeposited midden material, a practice found elsewhere.

Re v i e w o f t h e St r a t i g r a p h y o f t h e Pe r s i m m o n II Mo u n d Two strata were recognized by Moore in the single excavation unit undertaken, and no pottery was recovered. It does not appear that sterile sand was reached in this excavation. At the base of the unit, ca. 40 cm of black sand mixed with shell was encountered. This was surmounted by a 140 cm thick stratum described as being composed of whole, crushed, and powdered shell with some sand admixture. Between 30 and 45 cm below surface, burnt and broken human bone was found that had formed a concretion with the shell as well as Ahearths@ and animal bone. This deposit was also interpreted as indicative of cannibalism.

A Ri t u a l -Architectural Hi s t o r y o f t h e Pa l m e r -Ta y l o r Mo u n d Analysis of the stratigraphy of four upper St. Johns mounds and Harris Creek in the middle St. Johns demonstrates a recognizable general sequence of extant midden surmounted by burial monument. The implication from these sites is that the creation of burial monument as ritual 186 architecture is preceded by substantial usage of these locations for harvesting aquatic resources. The most extensively excavated site in the upper St. Johns is the Palmer-Taylor mound near the outlet of the Econlockhatchee river. A detailed accounting of the various excavations at this site is found in Chapter 8. The site plan is somewhat amorphous, but is configured similar to other Mt. Taylor sites, with a large heap at one end and a comma or crescent shape to the site plan. At Palmer-Taylor, the stratum with human remains is underlain by at least multiple recognized midden strata. At present, we have no estimate of calendar time required to accumulate these middens, but can safely assume that multiple generations are represented by the deposits. Initial shell and refuse was deposited over local sand (Stratum VII) over which additional shell and assorted fauna accumulated (Stratum VI). Stratum V is a thin lens of mussel shell, localized to near the area where the burial monument would eventually be centered, and associated with a possible structure (Figure 55). There is an unclear stratigraphic relationship between Stratum V and what appears to be an intentional deposition of limonite at the interface of extant shell deposits and surrounding sand, but it appears that this Arocky@ deposit found in the Rollins 1976 excavation 2 dates to near the time of the possible structure. It may be with the deposition of this limonite that the monument is founded, the siting of the monument but there are what appear to be later middens. Stratum IV is a final large-scale midden deposit, extending over an area that would be encompassed by the later monument=s south slope. Capping this midden, and extending to the full extent of the monument, is Stratum III. This is a deposit of shell and sand, averaging approximately 50 cm in thickness. It is uncertain if Stratum III is intentional shell mounding, a primary midden, or secondarily deposited midden material. Through most excavation units, the stratum is generally homogenous shell with little admixture (Chapter 8), and I tentatively propose that this deposit was rapidly accumulated, but the rate of accumulation is only guessed at. If rapidly deposited, then the siting of the mound can reasonably be associated with the limonite. Atop Stratum III is a large, localized lens of charcoal and blackened matrix (Figure XXX; 187

Stratum I

Primary Burials Stratum II

Secondary Stratum III Deposited Burials Charcoal Deposited

Charcoal

Stratum III

Stratum IV

V B NorthR

0 20 m Secondary Burials Deposited Stratum II Deposited, Primary Burials Into Stratum II

Figure 55: Mortuary Strata of Palmer- Taylor Mound charcoal). This charcoal-rich deposit was found in excavation units on the north, west, and south slopes, and is estimated to encompasses a semi-oval area 25x28 meters. In units near the monument apex, the charcoal was found as thick as 15 cm. In a limited area of this charcoal, broken and fragmented human bones were found along with occasional faunal remains. This is the first deposition of human bone. It is in part this deposit of human bone atop charcoal at the Palmer- Taylor mound that affirmed for Moore and later scholars the assertion by Wyman that cannibalism was practiced by Mt. Taylor period people. In this model, the charcoal and commingled bone is part of the initiating of the building of a burial monument. Moore observed that at the base of the bone layer, in contact with the Ahearth@, human bone was more often marked by fire, but bone found higher was not charred. Animal bone associated 188 with the burial is a reasonable expectation of mortuary practice, which often includes feasting. We do not have an estimate for the number of individuals within this layer of bone, but Moore describes it as being 2 feet thick (Moore 1893c: 612). Comparing with Harris Creek, more than twenty secondary burials were found at the base of the first mortuary layer in a limited excavation (Aten 1999). People were retaining ancestral bones for some time prior to interring them in a monument. The character of this human bone deposit has particular implications. The presence of multiple secondarily buried individuals in this deposit and at other Mt. Taylor sites (e.g., Harris Creek) implies that either people are heirlooming ancestral remains or that there was a mass mortality event. Presence of the trait at multiple sites gives support to the former. At Harris Creek, a large fire was created atop Stratum III and a communal pile of human bones placed atop while still burning. The placement of bone on fires sometimes results in limited charring of the bone, but Aten does not think that cremation is the intent (1999). It is doubtful that a heap of human bone would have remained exposed for long. Stratum II is a large-scale intentional deposition of shell with some sand admixture. This shell encapsulates the promiscuous human bone. I contend that the accumulation of this stratum is the product of short-term, intensive construction by a community of people, by inciting others to participate in the making of sacred architecture. A monument is created. A lack of observed discontinuities and the homogenous nature to the stratum matrix indicate that surface soils did not have time to accumulate and substantial weathering did not occur. It is this type of deposit that some have construed as due to mass harvesting events (e.g., Claassen 1981; Crothers 2004). In the upper St. Johns, these strata may well be due to mass harvesting, yet the intention is not subsistence, but instead the harvesting of monument material. The requirement of harvesting massive amounts of shellfish in a short time is part of the devotion to the dead. Stratum II continued to be used for burial after the communal mortuary and subsequent encapsulation of the bone. Primary flexed burials were placed into shallow pits dug into Stratum II. The 1976 Rollins College field school encountered and documented one of these burial pits 189 (Square 9, Burial 1 in field notes). This pit was found in an exposed section, immediately beneath disturbed topsoil (Figure XXX), excavated into Stratum II. Another burial that probably originates in this same stratum was found by the Excavators= Club (Rouse 1951). The one stratum at the Palmer-Taylor mortuary monument was used for many generations for the interment of their dead. In the model given here, this continued use is termed reproducing, as continued mortuary practice and ritual-architectural receptions of the monument. The final described depositional layer at Palmer-Taylor is 30 cm of disturbed soils. Ceramics found in this stratum date to the St. Johns II period (Rouse 1951). No burials are clearly associated with this stratum, and it dates well beyond the time frame of concern in this dissertation.

A Ri t u a l -Architectural Se q u e n c e Following the methodology of creating a ritual-architectural history for a Choctaw story of Chapter 3, a sequence of events is described here as: harvesting, heirlooming, siting, initiating, inciting, and reproducing (Figure 56).

Harvesting The first stage of the sequence, harvesting, is not considered a ritual-architectural event, but persistent usage of a location precedes establishment of a monument. The middens that underlie and generally flank the monuments may be a product of sedentary village occupation, but that has yet to be demonstrated. Harvesting was an ongoing activity; no estimate of the usage of any site prior to monument construction exists at present, but could represent several generations. It is the repetitive harvesting of aquatic resources that first marks the landscape as utilized and which creates a tenacious linkage between people and places, a history of persistence. There are many locations that were used repeatedly for harvesting, but only a few of these locations became sites of mortuary monuments. Harvesting continued after the monument was constructed, but additional material does not seem to have been deposited on the monuments during the Mt. Taylor period.

Heirlooming There are multiple individuals represented in the communal pile of human bones found atop 190

Ritual-Architectural Event- Some Material Correlates-

Burial Practice- Monument

Primary Flexed Continued Burials into Reproducing Use of Monument Extant Mortuary for Mortuary Activity Stratum

Stratum of Building Homogenous Shell Inciting of Mortuary Over Secondary Stratum Burials

Surface Preparation Commingled Building of Fire Secondary Burials Initiating Piling of Bones Charcoal Lens Feasting Animal Bone

Selection Sacred

Transitory Temporality of Experience Temporality Transitory Siting of Place Deposits Persistent Temporality of Experience Temporality Persistent of Monument Monument

Loose Human Ancestors Retention and Bone, Random Heirlooming Care of Bones in Secular Ancestral Bones Contexts History of Persistence Kitchen Harvesting Daily Life Middens Daily Detritus

Figure 56: Palmer-Taylor Ritual-Architectural Reception History 191 charcoal at Palmer-Taylor and Persimmon Mound II, and multiple individuals represented in single assemblages at Persimmon Mound, Orange Mound, and Harris Creek. One direct implication of the presence of multiple secondary burials in a single assemblage is that people were retaining the bones of the dead as regular practice, and that this was a trans-generational practice. Heirlooming ancestors itself is not a ritual-architectural event proper, but it is this practice that provides the material grist for the monumental occasions to come. In the Choctaw story above, the bones of the ancestors had become burdensome, and formed part of the allurement of the monument, but I am not attributing such to Mt. Taylor. Heirlooming is a persistence experience and for our analysis, a persistent temporality. Retention of ancestors would be marked in part by the occasional find of human bone in domestic contexts, a paucity of burial in the time period immediately preceding interment, broken human bone, and eventual selection of larger skeletal elements. Crania seem to have a particular significance, based on their inclusion and configurations at Harris Creek. While heirlooming, ancestors would be physically intimate with the living, being kept in the home and at least on occasion, actually carried. To depart from this relationship must have required renegotiation of the rules of appropriateness.

Siting Siting (Charles and Buikstra 1983) is the decision of where to physically erect a monument. The temporality of siting of monuments is difficult to determine. The specific decision of where to erect a monument may be rapid, a Amoment in time@, but the development of a particular location as sacred could have a long history that precedes the emplacement of a monument. Some factors that seem important to the siting of Mt. Taylor monuments include proximity to water features (springs, ponds, lakes) and relationship to extant monuments. Monuments are found in highly productive localities, but in much of the St. Johns, many places were highly productive. Mt. Taylor sites are typically found in aggregations, including groups of burial monuments. As an example, we can look at the area of the confluence of Salt Creek and the St. Johns. It is here that three of the four identified Mt. Taylor mortuary monuments in the upper St. Johns are found, 192 along with two other shell mounds with definite ceramic era deposits and possible preceramic components (Figure XXX). These three mounds are within 2 km of each other. Along with Palmer-Taylor at the confluence of the Econlockhatchee and St. Johns 13 km north, these are the only Acertain@ preceramic mortuary monuments in the entire upper St. Johns basin, an area of over 780,000 ha.

Initiating Initiating is the beginning of the creation of sacred place as evidenced materially. This is the first formal ritual-architectural occasion of Mt. Taylor monuments. At Palmer-Taylor, the initiation of mound construction began with the deposition of a cache of limonite at the western edge of the mound and the preparation of the surface of the shell heap. A large fire was placed on the surface, generally achieving sufficient temperatures to calcine the underlying shell. A substantial pile of commingled human bone was placed atop the fire. Feasting co-occurred with this event, producing the animal remains that Wyman, Moore, and Goggin interpreted as side dishes to the cannibalistic main course. As hermeneutical ritual-architectural experience, this initiating of monument construction is part of the allurement, the attraction to participating in the making of a place sacred by the inclusion of something imbibed with the sacred, the bones of their own ancestors. At a place where the persistence of a community was evidenced by the midden underfoot, a structural transformation took place that affected the physical relationship between the living and the dead, from ancestors present in home and on person to ancestors in a communal facility. Without this occasion, there would be no burial monument.

Inciting Incitement refers to the major construction event of the mortuary monument, i.e., the piling of shell over the commingled human bone and charcoal. Architecturally, it is this occasion that has the greatest material consequent by volume. At Palmer-Taylor, a stratum covering an area of approximately 3300 m2 was deposited, encapsulating the communal burial and creating a raised 193 and prominent monument, topographically and socially. There is no necessary demand of coercion or formal leadership to guide the building of a monument. It is the power of allurement of sacred architecture to draw people into participating in a ritual-architectural event, to contribute time, labor, material in its creation. Community and integration are part of the incitement to engage, but the motives of participants are not all altruistic. The self is transformed by ritual-architectural occasion, and that is part of the appeal, though the result is not always what was expected (Jones 2000a). We may posit that building of a mortuary stratum could integrate a greater social network then the piling of bones on a fire. Nearby communities could have contributed to the building, though it is not their direct ancestors being interred. A large amount of labor is involved in this building, but there may be a sizeable force applying labor.

Reproducing Reproducing refers to the continued use of the monument for burial. In terms of scale, these are smaller, more intimate ritual-architectural occasions. Individual, small grave pits were dug into the mortuary stratum. Primary flexed burials, and more rarely, bundled burials, were placed into these pits, occasionally associated with food remains and charcoal. Fish bone was found with Burial 1 at the Palmer-Taylor mound. At Harris Creek, fires were occasionally either built in grave pits or still burning coals were placed into the pit, and bodies placed on these small fires, associated animal bone is thought to be from feasting (Aten 1999). Again, the types of burials inform us to time and practice. Primary burials in individual pits indicate that people are interred soon after death. The numbers of primary burials in a single stratum can be surprisingly large: there were hundreds of burials in the mound at Harris Creek. Since populations are probably small, and most segments of society are represented in the burial population, the number of interments implies usage of the monument for many generations. The scale of these occasions may be smaller than heaping of ancestral bones or the deposition of a mass shell stratum, but the events are no less meaningful. Elements of earlier practice continued, including fire, feasting, and at some sites, white sand, practices that continued throughout later 194 archaeological periods in the St. Johns. People are still engaging with the monument, and there is still connection to ancestors and community. The monument remains active, a feature of the living, transitory community, and is a materialization of the persistent community of past ancestors, though externally, little volume or visible modification took place.

Di s c u ss i o n This chapter has presented a model for creating a ritual-architectural reception history of Mt. Taylor burial monuments. Such a history is considered a “narrow narrative” or restricted archaeological interpretation, but still suitable for comparison as a middle-range theory. Utilizing stratigraphic data, a basic sequence of hermeneutical engagement with the Palmer-Taylor Mound is sketched. A generalized sequence of Mt. Taylor burial monument construction is charted. Beginning with an extant midden, a surface is prepared. The preparation may include the depositing of a stratum of mounded shell. Upon a surface, a large fire is prepared and fired. Secondary burials, ancestral bone that has been previously heirloomed is piled atop this fire. Bone near the base of the pile is charred, that higher is not, but much of it is broken. Immediately after the piling of bone, a large- scale depositional event occurs that entombs the burials in an expansive stratum of shell and sand. The surface of this stratum persists, perhaps with some structures, over many generations, and it continued to be used for mortuary activity. Upon death, a small pit is excavated into the shell-sand matrix and, sometimes in the accompaniment of feasting and fire, flexed primary burials are made. Middens do not accumulate over this stratum during the Mt. Taylor period. These monuments are adjacent to middens that continued to be occupied: they are sacred architecture subject to active hermeneutical engagement on a daily basis. 195 Ch a p t e r 11: Co n c l u s i o n s

This dissertation reviews the contexts and conditions for the appearance of intentional burial monuments during the Mt. Taylor period (7300-5600 cal 14C yr BP) of the St. Johns river in peninsular Florida. The region is characterized by extensive marshes, swamps, and slow-moving waterways that began to become entrenched during the mid-Holocene, leading to the development of extensive freshwater fisheries. These fisheries provided an economic base that was both highly productive and highly predictable and encouraged populations to become tethered to smaller and specific territories. Restricted mobility left more obvious physical traces of perpetuity onthe landscape in the form of extraction and processing facilities and large scale middens, a process that further strengthened groups’ ties to bounded taskscapes. However, people interring their dead in highly visible monuments indicates a striking commitment to the history of a particular place and the expectation of continued access to the honored dead and to local resources. These monuments are the foundational sacred architecture of emergent corporate groups. As a materialization of claims, these monuments demonstrate a willingness to invest not only time, economic goods, and even ancestors into their construction, but also submission to a new type of community and a devotion to defense of those monuments and what they represent. Building a burial monument binds people to places in new and persistent ways.

Mt. Taylor is characterized in part by large heaps of freshwater shellfish, some of which have human remains associated with them. Traditionally, these archaeological features have not been understood as intentional constructions, but instead as heaps of refuse which contain accidental or nearly incidental burials; theoretical approaches applied to Mt. Taylor archaeology assumed monumentality as a consequent of other developments. Cultural evolutionary models considered mound building a local derived phenomenon, but one that appeared only after the development or adoption of specific predicates, including pottery making, basic political hierarchies, and agriculture. All of these features are lacking at Mt. Taylor sites. Diffusionary models expected those same traits to appear at nuclear and urban centers first and only as a later development away 196 from those centers. Chronologically, Mt. Taylor sites are contemporary with the earliest recognized monuments in the Americas. In neither case did theory allow for intentional construction to be first at the peripheries or by preceramic fisher-hunter-gatherers. More recent perspectives have embraced the possibility that fisher-hunter-gatherers can create monuments, but shell sites have struggled to get beyond the possible. The repetition of site plan and organization has led to a ready acceptance of earthen mounds in North and South America as being built by fisher-hunter-gatherers. But for the Mt. Taylor period sites, the reception has been more lukewarm. Despite a faithful devotion by some to the idea that these heaps do represent burial mounds, site plans have been explained away by others as a simple function of community discard patterns and the presence of burials and distinctive stratigraphic evidence are argued to result from particular harvesting strategies. Researchers have called for, and failed to deliver, models that emphasize structure, pattern, and repetition. Demonstrating intentionality has required a reconceptualization of these shell heaps, and a disengagement with the theoretical positions usually taken. I have turned to an approach that emphasizes the building of the monument as a historical sequence of ritual events and interactions.

Using a protocol of investigation advocated by Lindsay Jones, I created a history of the construction of Mt. Taylor shell heaps in the upper St. Johns basin. Beginning with a thesis of superabundance of meaning, the Jones methodology reframes sacred architecture in a background independent way and provides a language of mapping the archaeological data to the transformative event of monument creation; it serves as a middle-range theory to bridge specific stratigraphic contexts with specific behaviors and gives a direction to write narratives of how these archaeological data can be explained as a product of particular actions. This approach insists that the appropriate tempo of investigation is the temporality of human experience.

Pa l m e r -Ta y l o r a s Sa c r e d Architecture Reviewing the stratigraphy of the Palmer-Taylor site as a sequence of events allows us to track the genesis of a permanent monument. Conceiving of the monument as sacred architecture, 197 we find the radical decision was made to change the negotiation between the dead and the living from the direct and personal relationship of ancestral remains cared for and revered in the household to the promiscuous commingling of honored dead in a central structure. Middens accumulated as people returned to the same locations year after year, eventually becoming substantial features of the landscape and demonstrating longitudinal successes. Certain locations became favored, probably due to many factors, and some places were sacredly distinct and monuments sited there. To produce the archaeological situation found at the site, there has be consent between cross-kin groups to include their sacred trusts in a single pile and then communal participation to entomb those remains. The monument is a social contract that binds houses together with sacred collateral. At its building it is the formal recognition of the group and the most sacred artifact of that group.

After the entombment of the commingled human bone, the monument continues as sacred architecture. People continue to be buried in this same shell layer for many generations afterwards, but burials are now represented as small pits into Stratum II and placing of primary burials into those pits. The foundation of the monument occurred within a lifetime but as sacred architecture, these heaps of shell held a central role for as long as a millennium, and while not occurring at Palmer-Taylor, many Mt. Taylor sites continued as villages and community long past the time period, and many of the features of Mt. Taylor burial monuments are found until the historical period nearly 7,000 years later.

Di s c u ss i o n The recent recognition of the construction of sacred architecture by groups of harvesters and hunters across the world is forcing us to reconsider not only presumptions about the potentiality of non-agriculturally based economies, but also the role of monuments in the development of social and political complexity. in the Lower Mississippi Valley, Gobekli Tepe in Turkey, or Mt. Taylor in Florida: realization of the true nature of these sites did not require redefining what is a monument, just discarding assumptions of extant theoretical approaches and an appreciation 198 of how cosmogonic and sacred concerns can provide allurment to participate in their building. Additionally, these monuments stand as contracts of negotiations between kin groups in new matrices of obligation and commitment; the foundations of new kinds of associations. Of the three cases of hunter-gatherer monumentality just mentioned, only Mt. Taylor has been shown to have burials. Including ancestral remains in a monument is a sacrifice of the few to the many and a way of insuring dedication to the community. It seems likely that these other sites could produce human remains, perhaps as cremation or isolated skeletal elements. As we recognize monumentality in more places, times, and conditions, we must continue to rethink what they tell us about the development of society. This study has set out to determine if we can distinguish between Mt. Taylor middens and Mt. Taylor monuments through application of stratigraphic analysis and hermeneutical approach to sacred architecture. Demonstration of monumentality in this temporal, ecological, and technological context gives warrant to readdress basic assumptions about the capabilities of fisher- hunter-gatherers. A history of theoretical approaches applied to Mt. Taylor archaeological remains says much about current understandings. Beginning most explicitly with Jeffries Wyman, the people of the Mt. Taylor period were cast in terms of savagery, poverty, and filthiness. Relying on severely biased ethnographic observations and a model of conjectural history with an overt Victorian climax, human remains found in the shell heaps were early on presumed to be the remains of cannibalism. As theoretical perspectives developed, the image of the people of the Mt. Taylor period as simple, base, and brutal was perpetuated as culture-core ideas were introduced to explain the development and distribution of cultures and societies while the geographic position of the St. Johns was used to explain a perceived lack of cultural development and elaborate traits. Processualists inherited notions of simple and savage. Highly valuable research integrated environment, system, and change into models of Mt. Taylor society. The overt racism that marked the earliest research was expelled, but many of the conceptualizations remained. When confronted with contexts that violated inherited theoretical constructions (Archaic burial mounds), scholars 199 sought convoluted routes to explain away discrepancy. Acknowledging that monuments were built before agriculture and particular technological developments marks part of current opinion, but not all. Mt. Taylor shell heaps have become part of a movement that reimagines shell deposits with burials as intentional constructions. However, without a clear means of distinguishing situations of incidental accretionary middens and shell structures, these ideas have been subject to casual dismissal and disparagement. But, by applying Lindsay Jones’ hermeneutical approach to sacred architecture to Mt. Taylor monuments, we can determine the sequence of construction as activity, the film of time can be slowed down to the tempo of human life. To achieve this retardation, we must know the material and temporal specifics of Mt. Taylor archaeology. Mt. Taylor research began with large-scale excavations, and then is marked by decades of interpretation with no further excavation. In the late 1940’s, Irving Rouse and John Goggin created the basic cultural-historical framework for the St. Johns River area. In the mid- twentieth century a few large salvage projects recorded a large amount of data from mortuary contexts. Later excavations emphasized smaller exposures and were almost all aimed at retrieving subsistence and environmental data. Recent research is more multi-pronged and holistic. Restricting the chronological span of the Mt. Taylor period, using published excavations and radiocarbon dates from preceramic shell heaps, we find that the artifact assemblage to be extremely local. During the Mt. Taylor period there is little evidence for long distance exchange, social hierarchy, or elaborate grave treatment, attributes that are found with the succeeding Thornhill Lake period. The general environmental context and the basic subsistence strategy of the Mt. Taylor period furthers the characterization of society and position. During the Mt. Taylor period, local environments became gradually more inundated and characterized by high productivity. Aquatic resources became the focus of exploitation, and their availability and predictability could have provided for expanding populations. Comparing excavation data shows that Mt. Taylor burial monuments are distinguished from 200 companion middens by particular stratigraphic attributes. By understanding that deposition is not a uniformitarian phenomenon and that surfaces are sometimes intentionally short-lived, we better comprehend how pacing and effort is materially marked as midden and monument. Applying Jones’s theoretical program to the archaeology of the Mt. Taylor period shell heaps of the upper St. Johns river basin permits the writing of a ritual‑architectural reception history of the Palmer‑Taylor Mound. This kind of approach provides a means of linking behavior and stratigraphy and study of sacred architecture as experience. While radically different than many traditional archaeological perspectives, it is only an initial step towards explication of material correlates of monumentality. The accomplishments of this dissertation are the first steps in attempting a fuller understanding of Archaic shell sites and society. In the main, the intent is to provide a measure of certainty to the question of the existence of Mt. Taylor monuments. Having done so, we are now in a better position to consider the contexts where monuments arise. There is little doubt that material conditions are immensely important, and that without adequate resources or population, monuments of the scale found in the St. Johns will not arise. What are those conditions and the timing of their arrival, and, how can we better relate cultural developments at the pace of experience to more general data from subsistence and environmental studies? An extensive program of coring site and off-site locations in the upper St. Johns will be the first step in this direction. Future research will also emphasize the study of emergent complexity and inequality. Mt. Taylor period archaeology is marked by some correlates of complex hunter-gatherers, but others are absent. In particular, indications of hierarchy are not found in Mt. Taylor mortuary contexts. However, in the succeeding Thornhill Lake period, we do find such distinctions. Relating these transformations to changes in cultural and natural environments is a primary concern. Mt. Taylor monuments are part of a program of mortuary ritual and practice, and this dissertation does little more than demonstrate its existence. These burials are not marked by grave goods or elaborate funerary treatments, but are definitely marked by ritual and care. Archaic burial 201 monuments in the St. Johns are almost always located adjacent to circular water features, including springs, ponds, and sinks. This association is not accidental, and continues throughout the native occupation of the St. Johns. Geographic Information System technology will be applied to modeling surfaces of resource availability and the location of monuments will be compared to resource distributions and landscape features. Hunter-gatherer research is marked is some ways by drastic underestimation of the mental, spiritual, and material development of cultures under study. Shell sites people have been vicitimized by a history of bigotry. Theoretical treatments of monumentality among agricultural societies have tended to emphasize engaged actors, while foragers are often portrayed as being subservient to ecological conditions. Speaking specifically of hunter‑gatherers, Steward (1969: 188, emphasis added) states “The social environment, as contrasted with the natural environment, is also a factor shaping the nature of any society, but its role is minimal in most of these cases”. Contrast this with a quote from the same year, and the alternate possibility should be clear: A. . .in matters of religion, as of art, there are no >simpler' peoples, only some peoples with simpler technologies than our own. Man's >imaginative' and >emotional' life is always and everywhere rich and complex". (Turner 1969: 3). Environment is context and possibility, but humans make choices everyday. 202 Wo r k s Ci t e d Acheson, James M. 1981 Anthropology of Fishing. Annual Review of Anthropology 10: 275-316.

Anderson, David, Michael Russo and Kenneth Sassaman 2007 Mid-Holocene Cultural Dynamics in Southeastern North America. In Climate Change and Cultural Dynamics: A Global Perspective on Mid-Holocene Transitions, edited by David G. Anderson and Daniel H. Sandweiss. Elsevier Inc.

Arnold, Jeanne E. 1992 Complex Hunter-Gatherer-Fishers of Prehistoric California: Chiefs, Specialists, and Maritime Adaptations of the Channel Islands. American Antiquity 57: 60-84

1993 Labor and the Rise of Complex Hunter-Gatherers. Journal of Anthropological Archaeology 12: 75-119.

1996 Organizational Transformations: Power and Labor among Complex Hunter-Gatherers and Other Intermediate Societies. In Emergent Complexity: The Evolution of Intermediate Societies, edited by Jeanne E. Arnold, International Monographs in Prehistory, Ann Arbor, MI, pp. 59–73.

Ashmore, Wendy and A. Bernard Knapp (editors) 1999 Archaeologies of Landscape: Contemporary Perspectives. Blackwell Publishers, Malden MA.

Aten, Lawrence E. 1999 Middle Archaic Ceremonialism at Tick Island, Florida: Ripley P. Bullen’s 1961 Excavation at the Harris Creek Site. The Florida Anthropologist 52(3): 131-200.

Bailey, G.N. 1983 Concepts of Time in Quaternary Prehistory. Annual Review of Anthropology 12: 165- 192.

Balsillie, James H. and Joseph F. Donoghue 2004 High resolution sea-level history for the Gulf of Mexico since the Last Glacial Maximum. Florida Geological Survey, Tallahassee.

Bartram, John 1942 Diary of a Journey Through the Carolinas, Georgia, and Florida, From July 1, 1765 to April 10, 1766, annotated by Francis Harper. Transactions of the American Philosophical Society 33(1). 203 Beasley, Virgil Roy III 2001 Middle St. Johns Mound Construction during the Mississippian Period. Paper presented at the 58th Annual Meeting of the Southeastern Archaeological Conference.

Beckerman, Stephen 1994 Hunting and Fishing in Amazonia Hold the Answers, What are the Questions? In Amazonian Indians for Prehistory to Present

Beriault, John, Robert Carr, Jerry Stipp, Richard Johnson, and Jack Meeder 1981 The Archaeological Salvage of the Bay West Site, Collier County, Florida. The Florida Anthropologist 34: 39-58.

Bermes, B.J., G.W. Leve, and G.R. Tarver 1963 Geology and Ground-Water Resources of Flagler, Putnam and St. Johns Counties, Florida. Report of Investigations 32, State of Florida, State Board of Conservation, Division of Geology, Florida Geological Survey, Tallahassee

Binford, Lewis R. 1980 Willow Smoke and Dogs’ Tails: Hunter-Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity 45(1): 4-20.

Blitz, John H., and C. Baxter Mann 2000 Fisherfolk, Farmers, and Frenchmen: Archaeological Explorations on the Mississippi Coast. Mississippi Department of Archives and History, Jackson

Bradley, Richard 1993 Altering the Earth: The Origin of Monuments in Britain and Continental Europe. Society of Antiquaries of Scotland Monograph Series Number 8, Edinburgh.

Braje, Todd J. and Jon M. Erlandson 2007 Measuring Subsistence Specialization: Comparing Historic and Prehistoric Abalone Middens on San Miguel Island, California. Journal of Anthropological Archaeology 26: 474-485.

Brech, Alan 2004 Neither Ocean Nor Continent: Correlating the Archaeology and Geomorphology of the Barrier Islands of East Central Florida. Unpublished Masters Thesis, Department of Anthropology, University of Florida, Gainesville.

Brinton, Daniel G. 1872 Artificial Shell Deposits of the United States. InThe 1866 Annual Report of the Board of Regents of the Board of Regents of the Smithsonian Institution, Government Printing Office, Washington, D.C. 204 Brown, D.W., W.E. Kenner, J.W. Crooks, and J.B. Foster 1962 Water Resources of Brevard County, Florida. Report of Investigations 28, Florida Geological Survey, Tallahassee.

Brown, James A. and Robert K. Vierra 1983 What Happened in the Middle Archaic? Introduction to an Ecological Approach to Koster Site Archaeology. In Archaic Hunters and Gatherers in the American Midwest, edited by James L. Phillips and James A. Brown. Academic Press, New York.

Buikstra, Jane E. and Douglas K. Charles 1999 Centering the Ancestors: Cemeteries, Mounds, and Sacred Landscapes of the Ancient North American Midcontinent. In Archaeologies of Landscape: Contemporary Perspectives, edited by Wendy Ashmore and A. Bernard Knapp. Blackwell Publishers, Malden MA.

Bullen, Ripley P. 1955 Stratigraphic Tests at Bluffton, Volusia County, Florida. The Florida Anthropologist 8(1): 1-16.

1969 Excavations at Sunday Bluff, Florida. Gainesville: Contributions of the Florida State Museum, Social Sciences 15.

Bullen, Ripley P. and William J. Bryant 1965 Three Archaic Sites in the Ocala National Forest. The William L. Bryant Foundation American Studies Report No. 6, Orlando.

1972 The Orange Period of Peninsular Florida. Florida Anthropological Society Publications Number 6, The Florida Anthropologist 25(2-2): 9-33.

Bullen, Ripley P. and Adelaide K. Bullen 1961 The Summer Haven Site, St. Johns County, Florida. The Florida Anthropologist 14: 1-15.

Bullen, Ripley P. and Frederick W. Sleight 1959 Archaeological Investigations of the Castle Windy Midden, Florida. The William L. Bryant Foundation Report No. 1, Orlando.

1960 Archaeological Investigations of Green Mound, Florida. The William L. Bryan Foundation Report No. 2, Orlando.

Calvert, M., Kim Rudolph, and J.J. Stipp 1978 University of Miami Radiocarbon Dates XII. Radiocarbon 20(2): 274-282. 205 Carr, Robert S. 1981 Excavations of an Archaic Cemetery in Cocoa Beach, Florida. The Florida Anthropologist 34: 81-89.

Charles, Douglas K. and Jane E. Buikstra 1983 Archaic Mortuary Sites in the Central Mississippi Drainage: Distribution, Structure, and Behavioral Implications. In Archaic Hunters and Gatherers in the American Midwest, edited by James L. Phillips and James A. Brown. Academic Press, New York.

Cipriani, Lidio 1996 The Andaman Islanders. Edited and translated by D. Tayler Cox, assisted by Linda Cole. Frederick A. Praeger, New York.

Claassen, Cheryl P. 1988 New Hypotheses for the Demise of the Shell Mound Archaic. Paper presented at the annual meeting of the Southeastern Archaeological Conference, New Orleans.

1991a Gender, Shellfishing, and the Shell Mound Archaic. InEngendering Archaeology: Women and Prehistory, edited by J.M. Gero and M.W. Conkey, Basil Blackwell, Oxford.

1991b New Hypotheses for the Demise of the Shell Mound Archaic. In The Archaic Peirod in the Mid-South, edited by Charles McNutt. Archaeological Report No. 24, Mississippi Department of Archives and History, Jackson.

1996 A Consideration of the Social Organization of the Shell Mound Archaic. In Archaeology of the Mid-Holocene Southeast, edited by Kenneth E. Sassaman and David G. Anderson. University Press of Florida, Gainesville.

Clausen, Carl J., A.D. Cohen, Cesare Emiliani, J.A. Holman, and J.J. Stipp 1979 Little Salt Spring, Florida: A Unique Underwater Site. Science 203: 609-614.

Cooke, C. Wyeth 1945 Geology of Florida. Geological Bulletin 29, Florida Geological Survey, Tallahassee.

Crumley, Carole L. 1999 Sacred Landscapes: Constructed and Conceptualized. In Archaeologies of Landscape: Contemporary Perspectives, edited by Wendy Ashmore and A. Bernard Knapp. Blackwell Publishers, Malden MA.

Cumbaa, Stephen L. 1976 A Reconsideration of Freshwater Shellfish Exploitation in the Florida Archaic. The Florida Anthropologist 29: 49-59. 206 Cushing, Frank H. 2005 The Lost Florida Manuscript of Frank Hamilton Cushing, edited by Phyllis E. Kolianos and Brent R. Weisman. University of Florida Press, Gainesville.

Dall, William H. 1891 St. Johns and Volusia Counties. In Cyrus Thomas, Report of Mound Explorations of the Bureau of Ethnology, Twelfth Annual Report, Washington D.C.

Darwin, Charles G. 1839 (1959) The Voyage of the Beagle. Everyman’s Library, London.

Davis, J.H. 1946 The Peat Deposits of Florida, Their Occurrence, Development, and Uses. Bulletin 30, Florida Geological Survey, Tallahassee.

Davis, Mary B. 1987 Field Notes of Clarence B. Moore’s Southeastern Archaeological Expeditions, 1891- 1918: A Guide to the Microfilm Edition. Huntington Free Library, Bronx, New York.

Dickel, D.N. and G.H. Doran 2002 An Environmental and Chronological Overview of the Region. In Windover: Multidisciplinary Investigations of an Early Archaic Florida Cemetery, edited by Glen H. Doran. University Press of Florida, Gainesville.

Doran, Glen H. (editor) 2002 Windover: Multidisciplinary Investigations of an Early Archaic Florida Cemetery. University Press of Florida, Gainesville.

Doran, Glen H. 2002a Introduction to Wet Sites and Windover (8BR246) Investigations. In Windover: Multidisciplinary Investigations of an Early Archaic Florida Cemetery, edited by Glen H. Doran. University Press of Florida, Gainesville.

2002b The Windover Radiocarbon Chronology. In Windover: Multidisciplinary Investigations of an Early Archaic Florida Cemetery, edited by Glen H. Doran. University Press of Florida, Gainesville.

Erlandson, Jon M. 2001 The Archaeology of Aquatic Adaptations: Paradigms for a New Millennium. Journal of Archaeological Research.

Erlandson, Jon M., Mark A. Tveskov, and R. Scott Byram 1998 The Development of Maritime Adaptations on the Southern Northwest Coast of North America. Arctic Anthropology 35(1): 6-22. 207

Edwards, W.E. 1954 The Helen Blazes Site of Central Eastern Florida: A Study in Utilizing the Disciplines of Archaeology, Geology, and Pedology. Ph.D. Dissertation, Department of Political Science, Columbia University, University Microfilms, Ann Arbor.

Faught, Michael K. 2002 Submerged Paleoindian and Archaic Sites of the Big Bend, Florida. Journal of Field Archaeology 29: 273-290.

Ferguson, Vera M. 1951 Chronology at South Indian Field, Florida. Yale University Publications in Anthropology No. 45, New Haven.

Ford, James A. and Gordon R. Willey 1941 An Interpretation of the Prehistory of the Eastern United States. American Anthropologist 43(3): 325-363.

Gamble, Clive 1999 The Paleolithic Societies of Europe. Cambridge University Press, Cambridge, U.K.

Garson, Adam G. 1980 Comment Upon the Economic Potential of Fish Utilization in Riverine Environments and Potential Archaeological Biases. American Antiquity 45(3): 562-567.

Gibson, Jon L. 1996 Poverty Point and Greater Southeastern Prehistory: The Culture That Did Not Fit. In Archaeology of the Mid-Holocene Southeast, edited by Kenneth E. Sassaman and David G. Anderson. University Press of Florida, Gainesville.

2004 The Power of Beneficent Obligation in First Mound-Building. InSigns of Power: The Rise of Cultural Complexity in the Southeast, edited by Jon L. Gibson and Philip J. Carr, The University of Alabama Press, Tuscaloosa.

Goggin, John M. 1948a A Revised Temporal Chart of Florida Archaeology. The Florida Anthropologist 1: 57-60.

1948b Culture and Geography in Florida Prehistory. Ph. D. dissertation, Yale University. University Microfilms, Ann Arbor.

1949 Cultural Traditions in Florida Prehistory. In The Florida Indian and His Neighbors, edited by John W. Griffin. Rollins College, Winter Park, Fl. 208 1952 Space and Time Perspective in Northern St. Johns Archeology, Florida. Yale University Press, London.

Goggin, John and Frank H. Sommer 1949 Excavations on Upper Matecumbe Key. Yale University Publications in Anthropology No. 41, New Haven.

Griffin, John W. and Hale G. Smith 1978 Cultural Resource Reconnaissance of Merritt Island National Wildlife Refuge. MS on file, Florida Department of State, Division of Historical Resources,Tallahassee

Hofman, Jack L. 1986 Hunter-Gatherer Mortuary Variability: Toward and Explanatory Model. Ph. D. dissertation, The University of Tennessee, Knoxville, University Microfilms, Ann Arbor.

Holloway, R.G. 2002 Pollen Analysis of Holocene Sediments. In Windover: Multidisciplinary Investigations of an Early Archaic Florida Cemetery, edited by Glen H. Doran. University Press of Florida, Gainesville.

Holmes, William H. 1903 Aboriginal Pottery of the Eastern United States. Twentieth Annual Report of the Bureau of American Ethnology to the Secretary of the Smithsonian Institution, Government Printing Office, Washington, D.C.

1907 Aboriginal Shell-Heaps of the Middle Atlantic Tidewater Region. American Anthropologist 9(1): 113-128.

Ingold, Tim 1993 The Temporality of the Landscape. World Archaeology 25(2): 152-174.

Irving, Guy, and James Norris 1975 Palmer-Taylor Site. Ms. on file at Rollins College.

Jahn, Otto L. and Ripley P. Bullen 1978 The Tick Island Site, St. Johns River, Florida. The Florida Anthropologist 31(4): i-25.

Jefferies, Richard W. 1996 The Emergence of Long-Distance Exchange Networks in the Southeastern United States. In Archaeology of the Mid-Holocene Southeast, edited by Kenneth E. Sassaman and David G. Anderson, University Press of Florida, Gainesville.

Johnson, Robert E. 2001 Phase III Mitigative Excavation at the Lake Monroe Outlet Midden (8VO53), Volusia 209 County, Florida. Report prepared for U.S. Department of Transportation, Federal Highway Administration. Report prepared by Archaeological Consultants Inc. and Janus Research.

Jones, Charles C. Jr. 1873 Antiquities of the Southern Indians, Particularly of the Georgia Tribes. D. Appleton and Company, New York.

Jones, Lindsay. 2000a The Hermeneutics of Sacred Architecture: Experience, Interpretation, Comparison, Volume 1: Monumental Occasions, Reflections on the Eventfulness of Religious Architecture. Harvard University Center for the Study of World Religions, Harvard University Press, Cambridge, MA.

2000b The Hermeneutics of Sacred Architecture: Experience, Interpretation, Comparison, Volume 2:Hermeneutical Calisthenics, A Morphology of Ritual-Architectural Priorities. Harvard University Center for the Study of World Religions, Harvard University Press, Cambridge, MA.

Jones, T. L. 1991 Marine Resource Value and the Priority of Coastal Settlement: A California Perspective. American Antiquity 56: 419-443.

Joyce, Rosemary A. 2001. Burying the Dead at Tlatilco: Social Memory and Social Identities. Archeological Papers of the American Anthropological Association 10:12-26.

Kelly, Robert L. 1995 The Foraging Spectrum: Diversity in Hunter-Gatherer Lifeways. Smithsonian Institution Press, Washington, D.C.

Kindinger, Jack L., Jeffrey P. Davis, and James G. Flocks 2000 Subsurface Characterization of Selected Water Bodies in the St. Johns River Water Management District, Northeast Florida. USGS Open File Report 00-180, Department of the Interior, U.S. Geological Survey, Washington, D.C.

Knapp, A. Bernard and Wendy Ashmore 1999 Archaeological Landscapes: Constructed, Conceptualized, Ideational. In Archaeologies of Landscape: Contemporary Perspectives, edited by Wendy Ashmore and A. Bernard Knapp. Blackwell Publishers, Malden MA.

Kroeber, A.L. 1919 California. In Encyclopaedia of Religion and Ethics, edited by James Hastings, Charles Scribner’s Sons, New York. 210

1931 The Culture-Area and Age-Area Concepts of Clark Wissler. In Methods in Social Science, edited by Stuart A. Rice, University of Chicago Press, Chicago.

1939 Cultural and Natural Areas of Native North America. The University of California Press, Berkeley.

1948 Anthropology: Race, Language, Culture, Psychology, Prehistory. Harcourt, Brace and Company, New York.

Limp, Frederick and Van A. Reidhead 1979 An Economic Evaluation of the Potential of Fish Utilization in Riverine Environments. American Antiquity 44: 70-78.

Lubbock, John (Lord Avebury) 1913 Pre-historic Times, as Illustrated by Ancient Remains, and the Manners and Customs of Modern Savages. Seventh edition, Henry Holt and Company, New York.

Marquardt, William H. 1985 Complexity and Scale in the Study of Fisher-Gatherer-Hunters: An Example from the Eastern United States. In Archaic Hunters and Gatherers in the American Midwest, edited by James L. Phillips and James A. Brown. Academic Press, New York.

1992 Shell Artifacts from the Caloosahatchee Area. In Culture and Environment in the Domain of the , Institute of Archaeology and Paleoenvironmental Studies, Monograph Number 1, University of Florida, Gainesville.

1996 Four Discoveries: Environmental Archaeology in Southwest Florida. In Case Studies in Environmental Archaeology, edited by Elizabeth J. Reitz, Lee A. Newsom, and Sylvia J. Scudder. Plenum Press, New York.

Marquardt, William H. and Patty Jo Watson 2005 The Green River Shell Mound Archaic: Conclusions. In Archaeology of the Middle Green River Region, Kentucky, edited by William H. Marquardt and Patty Jo Watson, Institute of Archaeology and Paleoenvironmental Studies, University of Florida, Gainesville.

Martinez, Agustin L. 1979 9,700 Years of Maritime Subsistence on the Pacific: An Analysis by Meanso fo Bioindicators in the North of Chile. American Antiquity 44(2): 309-324.

May, J. Alan 2005a Ethnographic and Ethnohistoric Suggestions Relevant to the Middle Green River Shell Middens. In Archaeology of the Middle Green River Region, Kentucky, 211 edited by William H. Marquardt and Patty Jo Watson, Institute of Archaeology and Paleoenvironmental Studies, University of Florida, Gainesville.

2005b An Explanatory Quantification of Midden Constituents to Detect Archaeological Strata at the Carlston Annis Site, 15Bt5. In Archaeology of the Middle Green River Region, Kentucky, edited by William H. Marquardt and Patty Jo Watson, Institute of Archaeology and Paleoenvironmental Studies, University of Florida, Gainesville.

McGee, Ray M. and Ryan J. Wheeler 1994 Stratigraphic Excavations at Groves’ Orange Midden, Lake Monroe, Volusia County, Florida: Methodology and Results. The Florida Anthropologist 47(4): 333-348.

Meehan, Betty 1982 From Shell Bed to Shell Midden. Australian Institute of Aboriginal Studies, Camberra.

Merriam-Webster 1999 Merriam-Webster’s Collegiate Dictionary, 10th edition. Merriam-Webster, Springfield, MA.

Milanich, Jerald T. 1971 The Alachua Tradition of North-Central Florida. Contributions of the Florida State Museum of Anthropology and History 17, Gainesville.

1973 The Southeastern : A Preliminary Definition. Florida Bureau of Historic Sites and Properties Bulletin 3: 51-63.

1994 Archaeology of Precolumbian Florida. University Press of Florida, Gainesville.

Milanich, Jerald T. and Charles H. Fairbanks 1980 Florida Archaeology. Academic Press, Orlando.

Miller, James J. 1992 Effects of Environmental Changes on the Later Archaic People of Northeast Florida. The Florida Anthropologist 45: 100-106.

1998 An Environmental History of Northeast Florida. University Press of Florida, Gainesville.

Milner, George R. 1998 The Chiefdom: The Archaeology of a Mississippian Society. Smithsonian Institution Press, Washington and London.

2004 Old Mounds, Ancient Hunter-Gatherers, and Modern Archaeologists. In Signs of Power: The Rise of Cultural Complexity in the Southeast, edited by Jon L. Gibson and Philip J. Carr, The University of Alabama Press, Tuscaloosa. 212

Mitchem, Jeffrey M. 1999 Introduction Clarence B. Moore’s Research in East Florida, 1873-1896. In The East Florida Expeditions of Clarence Bloomfield Moore, University of Alabama Press, Tuscaloosa.

Moore, Clarence Bloomfield 1891 Field Notebook 1. Manuscript on File, Cornell University.

1892a A Burial Mound of Florida. The American Naturalist 26(302): 129-143.

1892b Supplementary Investigation at Tick Island. The American Naturalist 26(307): 568-579.

1892c Mounds in Florida. The American Antiquarian and Oriental Journal 14: 292-295.

1892d Certain Shell Heaps of the St. John’s River, Florida, Hitherto Unexplored (First Paper). The American Naturalist 26(311): 912-922.

1892e Field Notebook 2. Manuscript on File, Cornell University.

1893a Certain Shell Heaps of the St. John’s River, Florida, Hitherto Unexplored (Second Paper). The American Naturalist 27(314): 8-13.

1893b Certain Shell Heaps of the St. John’s River, Florida, Hitherto Unexplored (Second Paper continued). The American Naturalist 27(314): 113-117.

1893c Certain Shell Heaps of the St. John’s River, Florida, Hitherto Unexplored (Third Paper). The American Naturalist 27(319): 605-624.

1893d Certain Shell Heaps of the St. John’s River, Florida, Hitherto Unexplored (Fourth Paper). The American Naturalist 27(320): 708-723.

1894a Certain Shell Heaps of the St. John’s River, Florida, Hitherto Unexplored (Fifth Paper). The American Naturalist 28(325): 15-26.

1893-1894a Field Notebook 5. Manuscript on File, Cornell University.

1893-1894b Field Notebook 6. Manuscript on File, Cornell University.

1894b Certain Sand Mounds of the St. John’s River, Florida, Part I. Journal of the Academy of Natural Sciences of Philadelphia 10:4-128.

1894c Certain Sand Mounds of the St. John’s River, Florida, Part II. Journal of the Academy of Natural Sciences of Philadelphia 10:129-246. 213

1898 Recent Acquisitions. A “Banner-stone”of Shell. Journal of the Academy of Natural Sciences of Philadelphia 11:187-188.

Morgan, Lewis H. 1877 Ancient Society or Researches in the Lines of Human Progress from Savagery, through Barbarism to Civilization. Henry Holt and Company, New York.

Morlot, A. 1861 General Views on Archaeology. Annual Report of the Smithsonian Institution for 1860: 284-343.

Moseley, Michael E. 1975 The Maritime Foundations of Andean Civilization. Cummings Publishing Company, Menlo Park.

Nelson, Nels 1918 Chronology of Florida. Anthropological Papers of the American Museum of Natural History 22(2): 75-103.

Newsom, Lee 1985 Analysis of Wood Artifacts. In Archaeological Site Types, Distribution, and Preservation within the Upper St. Johns River Basin, Florida, edited by Brenda Sigler-Eisenberg. Florida State Museum Miscellaneous Project and Report Series Number 27, Department of Anthropology, Florida State Museum, Gainesville, FL.

1994 Archaeobotanical Data from Groves’ Orange Midden (8VO2601), Volusia County, Florida. The Florida Anthropologist 47(4): 404-417.

O’Leary, Beth 1992 Salmon and Storage: Southern Tutchone Use of an “Abundant” Resource. Occasional Papers on Archaeology No. 3. Heritage Branch, Government of the Yukon, Whitehorse.

1996 The Structure of a Salmon Resource: The Southern Tutchone Fishery in the Southwest Yukon, Canada. In Prehistoric Hunter-Gatherer Fishing Strategies, edited by Mark G. Plew. Department of Anthropology, Boise State University, Boise.

Phillips, James L. 1983 Introduction. In Archaic Hunters and Gatherers in the American Midwest, edited by James L. Phillips and James A. Brown. Academic Press, New York.

Piatek, Bruce J. 1994 The Tomoka Mound Complex in Northeast Florida. Southeastern Archaeology 13(2): 109-118. 214

Plew, Mark G. (editor) 1996 Prehistoric Hunter-Gatherer Fishing Strategies. Department of Anthropology, Boise State University, Boise.

Prentiss, Anna Marie, Natasha Lyons, Lucille D. Harris, Melisse R.P. Burns, and Terrence M. Godin 2007 The Emergence of Status Inequality in Intermediate Scale Societies: a Demographic and Socio-economic History of the Keatley Creek Site, British Columbia. Journal of Anthropological Archaeology 26: 299-327.

Purdy, Barbara A. 1991 The Art and Archaeology of Florida’s Wetlands. CRC Press, Boca Raton, FL.

1994a Excavations in Water-Saturated Deposits at Lake Monroe, Volusia County, Florida: An Overview. The Florida Anthropologist 47(4):326-332.

1994b The Chipped Stone Tool Industry at Groves’ Orange Midden (8VO2601), Volusia County, Florida. The Florida Anthropologist 47(4): 390-392.

Ramsey, Christopher B. 2005 OxCal Program v3.10, The Manual. University of Oxford, Radiocarbon Accelerator Unit, Oxford.

Randall, Asa R. and Kenneth E. Sassaman 2005 St. Johns Archaeological Field School 2003-2004: Hontoon Island State Park. Technical Report 6, Laboratory of Southeastern Archaeology, Department of Anthropology, University of Florida, Gainesville.

Reimer PJ, MGL Baillie, E Bard, A Bayliss, JW Beck, C Bertrand, PG Blackwell, CE Buck, G Burr, KB Cutler, PE Damon, RL Edwards, RG Fairbanks, M Friedrich, TP Guilderson, KA Hughen, B Kromer, FG McCormac, S Manning, C Bronk Ramsey, RW Reimer, S Remmele, JR Southon, M Stuiver, S Talamo, FW Taylor, J van der Plicht, and CE Weyhenmeyer 2004 Radiocarbon 46:1029-1058.

Reitz, Elizabeth J. and Elizabeth S. Wing 1999 Zooarchaeology. Cambridge Manuals in Archaeology, Cambridge University Press, Cambridge.

Rick, Torben C., Jon M. Erlandson, and Rene L. Vellanoweth 2001 Paleocoastal Marine Fishing on the Pacific Coast of the Americas: Perspectives from Daisy Cave, California. American Antiquity 66(4): 595-613. 215 Rouse, Irving 1951 A Survey of Indian River Archeology. Yale University Publications in Anthropology No. 44, Yale University Press, New Haven.

Russo, Michael 1985 Recovery and Analysis of Faunal Remains. In Archaeological Site Types, Distribution, and Preservation within the Upper St. Johns River Basin, Florida, edited by Brenda Sigler-Eisenberg. Florida State Museum Miscellaneous Project and Report Series Number 27, Department of Anthropology, Florida State Museum, Gainesville, FL.

1991 Archaic Sedentism on the Florida Coast: A Case Study from Horr’s Island. Ph.D. Dissertation, University of Florida, Gainesville, University Microfilms,Ann Arbor.

1992 Chronologies and Cultures of the St. Marys Region of Northeast Florida and Southeast Georgia. The Florida Anthropologist 45(2): 107-126.

1994a A Brief Introduction to the Study of Archaic Mounds in the Southeast. Southeastern Archaeology 13(2): 89-93.

1994b Why We Don’t Believe in Archaic Ceremonial Mounds and Why We Should: The Case from Florida. Southeastern Archaeology 13(2): 93-109.

Russo, Michael, Barbara A. Purdy, Lee A. Newsom, and Ray M. McGee 1992 A Reinterpretation of Late Archaic Adaptations in Central-East Florida: Groves’ Orange Midden (8VO2601). Southeastern Archaeology 11: 95-108.

Sassaman, Kenneth E. 1993 Early Pottery in the Southeast: Tradition and Innovation in Cooking Technology. University of Alabama Press, Tuscaloosa.

2003 St. Johns Archaeological Field School 2000-2001: Blue Spring and Hontoon Island State Parks. Technical Report 4, Laboratory of Southeastern Archeology, Department of Anthropology, University of Florida, Gainesville, FL.

2004 Complex Hunter-Gatherers in Evolution and History: A North American Perspective. Journal of Archaeological Research 12(3): 227-280.

2005 Poverty Point and Structure, Event, Process. Journal of Archaeological Method and Theory 12(4): 335-364

Saunders, Joe 2004 Are We Fixing to Make the Same Mistake Again? In Signs of Power: The Rise of Cultural Complexity in the Southeast, edited by Jon L. Gibson and Philip J. Carr. The University of Alabama Press, Tuscaloosa. 216

Saunders, Joe W., Rolfe D. Mandel, Roger T. Saucier, E. Thruman Allen, C.T. Hallmark, Jay K. Johnson, Edwin H. Jackson, Charles M. Allen, Gary L. Stringer, Douglas S. Frink, James K. Feathers, Stephen Williams, Kristen J. Gremillion, Malcolm F. Vidrine, and Reca Jones. 1997 A Mound Complex in Louisiana at 5400-5000 Years Before the Present. Science 277: 1796-1799.

Scott, Thomas 2001 Text to Accompany the Geologic Map of Florida. Florida Geological Survey Open File Report 80, Florida Geological Survey, Tallahassee.

Sears, William H. 1960 The Bluffton Burial Mound. The Florida Anthropologist 8(2-3):55-60.

Sigler-Eisenberg, Brenda J. 1984 The Gauthier Site: A Microcosm of Biocultural Adaptation in the Upper St. Johns River Basin. Paper presented at the 41st Southeastern Archaeological Conference, Pensacola, FL.

1985a Introduction. In Archaeological Site Types, Distribution, and Preservation within the Upper St. Johns River Basin, Florida, edited by Brenda Sigler-Eisenberg. Florida State Museum Miscellaneous Project and Report Series Number 27, Department of Anthropology, Florida State Museum, Gainesville, FL.

1985b Environmental Structure and Dynamics. In Archaeological Site Types, Distribution, and Preservation within the Upper St. Johns River Basin, Florida, edited by Brenda Sigler- Eisenberg. Florida State Museum Miscellaneous Project and Report Series Number 27, Department of Anthropology, Florida State Museum, Gainesville, FL.

Stewart, Marylin 1976 Palmer-Taylor Field Notes, Winter Term 1976. Manuscript on file, Rollins College, Winter Park, Florida.

1983 The Palmer-Taylor Site. Manuscript on file, Rollins College, Winter Park, Florida.

1985 A Partial Archaeological Survey of the William Beardall Tosohatchee Preserve. The Florida Anthropologist 38(4): 253-260.

Stein, Julie K. 1992 Deciphering a Shell Midden. Academic Press, San Diego.

Stein, Julie K., Jennie N. Deo, and Laura S. Phillips 2003 Big Sites—Short Times: Accumulation Rates in Archaeological Sites. Journal of Archaeological Science 30: 297-316. 217

St. Claire, Dana 1990 The Archaic in East Florida: Archaeological Evidence for Early Coastal Adaptations. The Florida Anthropologist 43(3): 189-197.

Stiriling, Matthew W. 1936 Florida Cultural Affiliations in Relation to Adjacent Areas. In Essays in Anthropology Presented to A.L. Kroeber in Celebration of his Sixtieth Birthday. University of California Press, Berkley.

Sullivan, Lawrence E. 2000 Foreword: Monumental Works and Eventful Occasions. In The Hermeneutics of Sacred Architecture: Experience, Interpretation, Comparison, Volume 1: Monumental Occasions, Reflections on the Eventfulness of Religious Architecture, Lindsay Jones. Harvard University Center for the Study of World Religions, Harvard University Press, Cambridge, MA.

Swanton, John R. 1931 Source Material for the Social and Ceremonial Life of the Choctaw Indians. Smithsonian Institution, Bureau of American Ethnology Bulletin 103, Washington, D.C.

Trigger, Bruce G. 1989 A History of Archaeological Thought. Cambridge University Press, Cambridge.

1998 Sociocultural Evolution: Calculation and Contingency. Blackwell Publishers, Oxford.

Turner, Victor 1969 The Ritual Process: Structure and Anti-structure. Cornell University Press, Ithaca.

Viau A. E., K. Gajewski, M. C. Sawada,1 and P. Fines 2006 Millennial-scale Temperature Variations in North America during the Holocene. Journal of Geophysical Research 111.

Walker, Karen J., Frank W. Stapor, and William H. Marquardt 1994 Episodic Sea Levels and Human Occupation at Southwest Florida’s Wightman Site. The Florida Anthropologist 47(2): 161-179.

Wardle, H. Newell 1956 Clarence Bloomfield Moore (1852-1936). Bulletin of the Philadelphia Anthropological Society 9(2): 9-11.

Watts, W.A. 1969 A Pollen Diagram from Mud Lake, Marion County, North-Central Florida. Geological Society of America Bulletin 80: 631-642. 218

Weisman, Brent R. 2002 Pioneer in Space and Time: John Mann Goggin and the Development of Florida Archaeology. University Press of Florida, Gainesville.

Wharton, Barry, George Ballo, and Mitchell Hope 1981 The Republic Groves Site, Hardee County, Florida. The Florida Anthropologist 34: 59- 80.

Wheeler, Ryan J. and Ray M. McGee 1994a Technology of Mount Taylor Period Occupation, Groves’ Orange Midden (8VO2601), Volusia County, Florida. The Florida Anthropologist 47(4): 350-379.

1994b Report of Preliminary Zooarchaeological Analysis: Groves’ Orange Midden. The Florida Anthropologist 47(4): 393-417.

1994c Wooden Artifacts from Groves’ Orange Midden. The Florida Anthropologist 47(4):380- 392.

Wheeler, Ryan J., James J. Miller, Ray M. McGee, Donna Ruhl, Brenda Swan, and Melissa Memory 2003 Archaic Period Canoes from Newnans Lake, Florida. American Antiquity 68(3): 533- 551.

Wheeler, Ryan J., Christine L. Newman, and Ray M. McGee 2000 A New Look at the Mount Taylor and Bluffton Sites, Volusia County, with an Outline of the Mount Taylor Culture. The Florida Anthropologist 53(2-3): 132-157.

White, William A. 1970 The Geomorphology of the Florida Peninsula. Geological Bulletin 51, State of Florida Department of Natural Resources, Bureau of Geology, Tallahassee.

Widmer, Randolph J. 1988 The Evolution of the Calusa: A Nonagricultural Chiefdom on the Southwest Florida Coast. The University of Alabama Press, Tuscaloosa.

Willey, Gordon R. 1948 Culture Sequence in the Manatee Region of West Florida. American Antiquity 13: 209- 218.

1949a Excavations in Southeast Florida. Yale University Publications in Anthropology No. 42, New Haven 219 1949b Archaeology of the Florida Gulf Coast. Smithsonian Miscellaneous Collections Vol. 113, Washington, D.C.

1955 The Prehistoric Civilizations of Nuclear America. American Anthropologist 57(3): 571- 593.

Willey, Gordon R. and Philip Phillips 1958 Method and Theory in American Archaeology. The University of Chicago Press, Chicago.

Willey, Gordon R. and Jeremy Sabloff 1993 A History of American Archaeology, 3rd edition, Thames and Hudson Ltd., London

Williams, Burton 1975 Field Notes, Palmer-Taylor Mound Excavation. Ms. on file, Rollins College.

Wissler, Clark 1917 The American Indian. McMurtrie, New York.

1926 The Relation of Nature to Man in Aboriginal America. Oxford University Press, New York.

Wyman, Jeffries 1868a An Account of Some Kjoekkenmoeddings, or Shell-Heaps, in Maine and Massachusetts. The American Naturalist 1(11): 561-584.

1868b On the Fresh-Water Shell-Heaps of the St. Johns River, East Florida. The American Naturalist 2(8): 393-403.

1868c On the Fresh-Water Shell-Heaps of the St. Johns River, East Florida (Concluded). The American Naturalist 2(9): 449-463.

1874 Human Remains in the Shell Heaps of the St. John’s River, East Florida: Cannibalism. The American Naturalist 8(7): 403-414.

1875 Fresh-water Shell Mounds of the St. John’s River, Florida. Peabody Academy of Science, Salem, MA.

1876 Primitive Man. The American Naturalist 10(5): 278-282.

Yesner, D.R. 1980 Maritime Hunter-Gatherers: Ecology and Prehistory. Current Anthropology 21: 727-750. 220

Ap p e n d i x A: Th e Cr e a t i o n o f Nu n i h Wa y a (Swanton 1931) 221 1) The chief halted the advance body of Choctaws on a little river to wait until scouts could be sent forward to explore the region of country round about; and to give time for the aged and feeble and those who were overloaded to come up. Many of the families were loaded with so many of the bones of their deceased relatives that they could carry northing else, and they got along very slowly. At this stage of their long journey there were a greater number of skeletons being packed along by the people than there were of the living. The smallest families were heaviest loaded; and such were their adoration and affection for these dry bones that before they could consent to leave them on the way, they would, having more bones than they could pack at one load, carry forward a part of them half a day’s journey, and returning for the remainder, bring them up the next day. By this double traveling over the route, they were soon left a great distance in the rear. They would have preferred to die and rot with these bones in the wilderness, sooner than leave them behind.

2) The minko looked upon the notions of the people in regard to the extraordinary and overwhelming burthen of bones as a great evil; and he cast in his mind for some plausible excuse to rid the people of a burthen that was as useless as it was oppressive to them.

3) And now the scouts had returned and the reports they made of quite an extensive excursion were very favorable and encouraging. They stated that everywhere, and in all directions, they found game of all sorts, fish and fowl and fruits in abundance; tall trees and running brooks; altogether they looked upon it as the most desirable and plentiful region they had found during their pilgrimage. They also stated, that the most convenient place they had found, for a winter encampment, lay in a southeasterly direction at the junction of three large creeks, which coming together at the same point, formed an immense lowland, and a considerable river. In the fork of the first and the middle creeks lay an extensive range of dry, good lands, covered with tall trees of various kinds, grapes, nuts, and acorns; and rivulets (bok ushi) of running water. For the multitude, it was distant eight or ten days’ travel, and the route would be less and less difficult to that place.

4) At the rising of the sun on the ensuing morning, the leader’s pole was observed to be inclining to the southeast, and the people were moving off quite early. The nights were becoming cooler, and they desired to have time to prepare shelter before the winter rains should commence. The chief, with the Isht Ahullo, who carried the sacred pole, went in front, and being good walkers, they traveled rapidly until they came to the place which had been designated. Great numbers of the stronger and more athletic people came up the same day.

5) Early on the next morning the chief went to observe the leader’s pole, which at the moment of sunrise, danced and punched itself deeper into the ground; and after some time settled in a perpendicular position, without having nodded or bowed in any direction. Seeing which the chief said, “It is well. We have arrived at our winter encampment.” He gave instructions to the tool carriers to lay off the encampment for the iksas and mark on posts their appropriate symbols. He ordered them to all sufficient 222 space for the iksas, having particular regard to the watering places.

6) It was several days before the people had all reached the encampment. Those who were packing the double load of bones came in several days later, and they complained of being greatly fatigued. They mourned and said, “The bad spirit has killed our kindred; to pack their bones any further will kill us, and we shall have no name amongst the ikasas of this great nation. Oh! when will this long journey come to an end?”

7) There were plenty of pine and cypress trees and palmetto; and in a short time the people had constructed sufficient tents to shelter themselves from the rain. Their hunters with but little labor supplied the camps with plenty of bear meat; and the women and children collected quantities of acorns and oksak kapko, and kapun (large hickory nuts, and scaly barks). It was an extremely plentiful land, and the whole people were rejoicing at the prospects for a pleasant and bountiful winter. Their camps being completed, the chief gave instructions, to have sufficient ground prepared to plant what seed corn might be found in the camps. Search was made by Isi maleli (Running Deer) for the corn. He found a few ears only; they had been preserved by the very old people, who had no teeth. The corn they found was two years old, and they were very much afraid that it was dead. The minko suggested to Isi maleli, that as the tool carriers had iron [!] implements with which to break the ground, it would be best to detail a sufficient number of them to prepare ground to grow it. So the minko called out twenty of the tool carriers for the purpose, and appointed the wise Isi maleli, to direct them, and to select the soil for growing the corn properly, and to preserve it when it matured.

8) One end of the encampment lay along the elevated ground – bordering the low lands on the west side of the middle creek. Just above the uppermost camps, and overhanging the creek, was a steep little hill with a hole in one side. As it leaned towards the creek, the people called it the leaning hill (nunih waya). From this little hill the encampment took its name, “Nunih Waya,” by which name it is known to this day.

9) The whole people were healthy at Nunih Waya. Full of life and cheerfulness, they danced and played a great deal. Their scouts had made wide excursions around the encampment, and finding no signs of the enemy in any direction, they consoled themselves with the idea, that they had traveled beyond his reach. The scouts and hunters, on returning into camps, from their exploring expeditions, were often heard to say, “The plentiful, fruitful land of tall trees and running waters, spoken of by our great and wise chiefs, who saw it in a vision of the night, is found. We have found the land of plenty, and our great journey is at an end.”

10) They passed their first winter at Nunih Waya quite pleasantly. Spring opened finely. Their few ears of corn came up well and grew off wonderfully. The creeks were full of fish and the mornings rang with the turkeys and singing birds. The woods everywhere were full of buffalo, bear, deer and elks; everything that could be wished for was there, and easily procured. All were filled with gladness. 223 11) And when the time for the green corn dance was near, the hunters brought into the camps wonderful quantities of fat meat, and they celebrated this dance five days. They did not eat of their corn, but that it might be properly called the green corn dance, they erected a pine pole in the center of the dance ground, and upon this they suspended a single ear of green corn.

12) When they had finished this, their forty-third green corn dance in the wilderness, the people began to be concerned as to the probabilities of their having to journey further. Many of them declared that if the sacred pole should indicate a removal, it would be impossible for them to go farther, on account of the great number of bones that had accumulated on their long journey. They could not carry the bones, neither could they think of leaving them behind.

13) The chief had for some time been considering the great inconvenience the marvelous amount of bones had become to the nation. He knew very well, the feelings of the people on the subject, and how difficult it would be to get them to consent to abandon the useless encumbrance. He could see very plainly, that should they have to go further, a portion of the people, under their present impressions in regard to the dry bones, would be most certainly left behind. On hearing the murmuring suggestions of so many of the people, every day, about the bones of their deceased relatives, and the sacred duty incumbent on the living to preserve and take care of them, he was convinced that the subject must be approached with caution. Yet, the oppressive, progress-checking nature of the burdens was such that they must be disposed of in some way.

14) He called a council of the leaders of the iksas, and in a very prudent and cautious manner, consulted them in behalf of the suffering people, enquiring of them at the same time, if it was possible to invent any means that would aid them in the transportation of their enormous packs of useless dry bones. It was a subject they had not before thought of, and they required a day or two to make up their minds.

15) Time was granted to them, and in the meantime the minko convened with many of the people [to consult] about it.

16) The council met again, and there was some discussion, but nothing conclusive. They were loath to speak of the bones of their deceased friends and relatives. They had packed their bones a great way, and for years; but there had been no conversation, no consultation, on the subject. There were among the young people, many who were carrying heavy packs of bones, who had never heard, and who really did not know, whose bones they were carrying. They had grown up with the bones on their backs, and had packed them faithfully, but never having heard the name of their original owners, they could tell nothing, no did they know anything about them. That the spirits hovered about their bones to see that they were respectfully cared for, and that they would be offended and punished with bad luck, sickness, or even death for indignities, or neglect of their bones, every one knew. It was a great indignity to the spirits to repeat the names they 224 were known by during their mundane existence. The greater part of the living who were then in the camp, had been born and reared in the wilderness, and were still packing the bones of those who had lived long before and of whom they knew nothing. Yet they worried along with heavy loads of these dry bones on their journey, in good faith, and in a full belief of its necessity as a sacred duty. The leaders of the iksas, who were not in council, were carrying heavy loads of bones themselves which they could not consent to part with; and they esteemed it a subject of too much delicacy to be caviled about in a council. They did not wish to say anything further about it, anyway.

17) One of the Isht ahullos, who was an old man, and who had long been a secret teacher, among the women and children, on the nature and wants of the spirit world and the causes that made it necessary to pack the bones of the dead, arose from this seat and said:

18) “Some people can make very light talk about the bones of our deceased friends and relatives. Those sacred relics of our loved ones, who have passed away from our sight are to be irreverently stigmatized by the name of ‘oppressive burthen,’ ‘useless incumbrance,’ and the like. Awful! And it was our chief who could dare to apply the uncivil epithet to the precious and far-fetched treasures. From all these things, I am forced, unwillingly, to infer that the next thing the chief has to propose for your consideration will be for you to cast away this ‘oppressive burthen.’ Shameful! (Hofahya.) This thing must not be. This people must not cast away the precious remains of the fathers and mothers of this nation. They are charged by the spirits, who are hovering thick around us now, to take care of them; and carry them whithersoever the nation moves. And this we must not, we dare not fall to do. Were we to cast away the bones of our fathers, mothers, brothers, sisters, for the wild dogs to gnaw in the wilderness, our hunters could kill no more meat; hunger and disease would follow; then confusion and death would come; and the wild dogs would become fat on the unscaffolded carcasses of this unfeeling nation of forgetful people. The vengeance of the offended spirits would be poured out upon this foolish nation.”

19) The council before which the Isht ahullo made this appeal to the religious sentiment of the tribe was only an assembly of the leaders of the iksas. The people were not present, and did not hear it. The chief, however, was fully apprised of the secret action of these bad men; and to counteract their dark and mischievous influence on the minds of the people, he dismissed the Isht ahullos, and leaders of the iksas, with a severe reprimand, telling them plainly that he had no further business for them to attend to. Then turning to the Isht ahullo, and at the same time pointing at him with an arrow, [he] said:

20) “When you again get in council with the lazy, bad hearted men to which you belong, tell them that the time has come when you must be cautious how you meddle with the affairs of the nation. Hear my words.”

21) The minko, returning to this tent, sent for Long Arrow, to whom he communicated his designs as to the disposition of the dry bones; after which he directed him to send the tool carriers to the iksas, and instruct them to summon every man, woman and child, except 225 the leaders of the clans and the conjurers of all grades. The minko said:

22) “Tell the people to assemble at the dance ground early in the day, tomorrow. I wish to consult them on important national business. Let the people, except those I have named, all know it before they sleep.”

23) In accordance with the notice sent by the chief, the entire tribe, male and female, old and young, except the yushpakammi and the leaders, came. These were not found in the great assembly. But the healthy, clean washed, bright, cheerful people were all present, and seated at the time the minko came to his place on the council ground.

24) The minko looked around on the multitude, and very calmly speaking, addressed them as follows:

25) “It is to you my brothers, my sisters, my countrymen, that I wish to declare my thoughts this day. I look around upon the bright, cheerful countenances of the multitude and I feel assured that you will hear my words; and that you will hearken to my counsels. You are a great people, a wonderful people, a people of strength, of unparalleled courage and untiring, patient industry. Your goodness of heart has caused you to work and hunt, far beyond the needs of your families, to gain a surplus, to feed a lazy, gluttonous set of hangers on, whose aim it is to misdirect you, whose counsels are all false, and whose greatest desire is confusion and discord amongst this peaceful, happy people. I know the meaning of my words. I speak them boldly and intentionally, I do not catch you in a corner, one at a time, and secretly communicate to you messages from the spirit land; packing you with enormous and insupportable burthens, to gratify wicked and discontented spirits, who are, as you are told, hovering about the camps, threatening mischief. But I call you all in general council and standing in this bright sunlight, with every eye upon me, and declare in language that cannot be mistaken, words of wisdom and truth. I bring no message from the spirit land. I declare to you the needs and interests of the living. I have no visions of the night; no communications from the discontented spirits, who it is said are hovering around our camps, threatening disaster and death to the living, out of spite for having been rejected from the good hunting ground, to tell you of; but openly, in this bright day, I communicate to you, in deepest solicitude, the long cherished thoughts of a live man; which, when fully carried out, cannot fail to establish peace, harmony, concord and much gladness to this great live nation. I speak not to the dead; for they cannot hear my words. I speak not to please or benefit the dead; there is nought I can do or say, that can by any possibility reach their conditions. I speak to the living for the advancement and well being of this great, vigorous, live multitude. Hear my words.

26) “From new motions and indications made by the sacred pole, which I have never witnessed before, I was led to conclude that our forty-three years’ journey in an unknown country had come to its termination. And to avoid hindering and annoying the whole people with what I had on my mind to be considered, I called yesterday (pilashash) a 226 council of the leaders of the iksas, and all the conjurers, for the purpose of examining and deciding on the most prudent course to pursue, in case it should be finally ascertained, that the leader’s pole had settled permanently.

27) “They all came, and after hearing my propositions, they put on wise faces, talked a great deal of the unhappy spirits of our dead friends, of their wants and desires, and of the great dangers that would befall the people, if the failed to obey the unreasonable demands made by the spirits, through the lazy Isht ahullos, conjurers and dreamers, who, according to their own words, are the only men through which the spirits can make manifest to the nation their burthensome and hurtful desires. Finding that they had nothing to say, nor did they even surmise anything of the subject of the affairs and interests of the living, I dismissed them as ignorant of, and enemies to, the rights of the people, and, therefore, improper agents for the transaction of their business. They were dismissed on account of their secret, malicious designs on the people, and their inefficiency in the councils of the nation. I immediately sent out runners to convene the people in general council to-day. You are all here, except the secret mongers, and the leaders of the clans, whose mouths and tongues have been tied up by the Isht ahullo and yushpakammi. The nation is present to hear my words; in them there is no secret of hidden meaning. You will all hear them, and let everyone, who is a man, open his mouth this clear day, and openly and fearlessly pour out his full and undisguised feelings on the topics which will be presented.

28) “From signs which I have just named, I conclude, and I find it the prevailing impression of this multitude of self-sustaining people, that our long journey of privations and dangers in the pathless wilderness has ended. We are no in the land of tall trees and running waters, or fruit, game of many kinds and fish and fowl, which was spoken of by our good chief, who is missing, in the far off country towards the setting sun. his words have come to pass. Our journey is at end, and we shall grow to be a nation of happy people in this fruitful land.

29) “Let us now, like a sensible people, put the nation in a suitable condition for the free enjoyment of the inexhaustible bounties that have been so lavishly spread in this vast country for the use and benefit of this multitude. Let us lay aside all useless encumbrance, that we may freely circulate, with our families in this widely extended land, with no burthen to pack, but such as are necessary to sustain life and comfort to you wives and little ones. Let us call this place; this, Nunih Waya encampment, our home; and it shall be so that when a man, at his hunting camp, in the distant forests, shall be asked for his home place, his answer will be, ‘Nunih Waya.’ And to establish Nunih Waya more especially as our permanent home, the place to which when we are far away, our thoughts may return with feelings os delight and respectful pleasure, I propose that we shall by general consent and mutual good feelings select an eligible location within the limits of the encampment and there, in the most respectful manner, bring together and pile up in beautiful and tasteful style the vast amount of bones we have packed so far and with which many of the people have been so grievously oppressed. Let each set of bones remain in its sack, and after the sacks are closely and neatly pile up, let them 227 be thickly covered over with cypress bark. After this, to appease and satisfy the spirits of our deceased relatives, our blood kin, let all persons, old and young, great and small, manifest their respect for the dead, by their energy and industry in carrying dirt to cover them up, and let the work of carrying and piling earth upon them be continued until every heart is satisfied. These bones, as we all know, are of the same iksa, the same kindred. They were all the same flesh and blood; and for us to pile their bones all in the same heap and securely cover them up will be more pleasing to the spirits, than it will be to let them remain amongst the people, to be scattered over the plains, when the sacks wear out in the hands of another generation who will know but little and care less about them. 30) “You have heard my talk. I have delivered to you the true sentiments of my heart. When it comes to my time to depart for the spirit land, I shall be proud to know that my bones had been respectfully deposited in the great mound with those of my kindred. What says the nation?”

31) Some little time elapsed; and there was no move among them. The multitude seemed to reflect. At length, a good looking man of about sixty winters, arose in a dignified manner, from his seat, and gravely said:

32) “It was in my boyhood, and on the little river where we had the great fish feast, that my much respected father died. His family remained and mourned a whole moon, and when the cry-poles were pulled down, and the feast and dance had ended, my mother having a young child to carry, it fell to my lot, being the next largest member of the family, to pack on the long journey, the bones of my father. I have carefully carried them over hardships and difficulties, from that little rocky river to the present encampment. Such has been my love and respect for these sacred relics, that I was ready at any time to have sacrificed my life sooner than I would have left them, or given them up to another. I am now growing old; and with my declining years come new thoughts. Not long hence, I too must die. I ask myself, who in the coming generations will remember and respect the bones of my father? Will they not be forgotten and scattered to bleach and moulder on the carelessly trodden plain? I have sought with a heart full of anxious sorrow, for a decent and satisfactory resting place, in which to deposit the bones of my long lost father. I could think of none. And I dare assert, that there are thousands in hearing of my voice, at this very moment, whose faithful hearts have asked the same embarrassing questions. I am happy in the acknowledgment, and I trust with much confidence, that the whole people will view this important matter in the same satisfactory light. The wise propositions of our worthy chief have answered perplexing questions and have fully relieved the unsettled workings of many anxious hearts.

33) “It is true, as our wise chief has already suggested, that we can now witness the wonderful and never before heard of sight of a live nation packing on their backs an entire dead nation, our dead outnumbering the living. It is a pleasure to me, not that my eyes have been opened by the chief’s proposition to the propriety of placing these relics of the dead nation to themselves, that we have power and time to do as he suggests, and most reverently to secure them from being tumbled among our greasy packs, and from 228 the occasional dropping of the precious bones, through the holes in the worn out sacks to be lost forever. Let us, in accordance with the wise and reasonable proposition of our minko, fetch all the sacred relics to one place; pile them up in a comely heap; and construct a mound of earth upon them, that shall protect them from all harm forever.”

34) And the people rose up and with one voice, said, “It is well; we are content.”

35) The minko stood up again and said that in that great multitude there might be some whose feelings in regard to the disposition of the bones of their dead friends would not permit them to pile them with the dead nation. Then they all shouted aloud, “It is good, it is satisfactory.” 36) Men were then appointed to select an appropriate place for the mound to be erected on, and to direct the work while in progress. They selected a level piece of sandy land, not far from the middle creek; laid it off in an oblong square and raised the foundation by piling up the earth which they dug up some distance to the north of the foundation. It was raised and made level as high as a man’s head and beat down very hard. It was the floored with cypress bark before the work of placing the sacks of bones commenced. The people gladly brought forward and deposited their bones until there were none left. The bones, of themselves, had built up an immense mound. The brought the cypress bark, which was neatly placed on, till the bone sacks were all closely covered in, and dry as a tent. While the tool carriers were working with the bark, women and children and all the men, except the hunters, carried earth continually, until the bark was all covered from sight, constituting a mound half as high as the tallest forest tree.

37) The minko kindled the council fire, and, calling an assembly of the people, told them that the work on the great monumental grave had been prosecuted with skill and wonderful industry. He said that the respect which they had already manifested for the deceased relatives was very great; that notwithstanding the bones were already deeply and securely covered up; the work was not yet completed. Yet it was sufficiently so to allow them to suspend operations for a season. Winter was drawing near; the acorns and nuts were beginning to fall and were wasting. The people must now scatter into the forests and collect the rich autumnal fruits which were showering down from every tree. That done, the people must return to the encampment; and as the tool carrier had produced seed corn enough for all to have a little field, each family must prepare ground for that purpose. Then, after the corn was grown and the new corn feast and dance celebrated and over, the nation could again prosecute the work on the mound, and so on, from year to year, until the top of the great grave of the dead nation should be as high as the tallest forest tree. And it should be made level on the top as much as sixty steps (habil) in length, and thirty steps in width, all beat down hard, and planted thick with acorns, nuts and pine seeds. “Remember my words,” said the chief, “and finish the work accordingly. Now go and prepare for winter.”

38) And the people gladly dispersed into the distant forests. Fruit was found in great quantities and was collected and brought into camp in very large amounts – acorns, 229 hickory nuts, and most and best of all, the otupi (chestnuts), all of which was secured from the worms by the process of drying them by smoke and incasing them in small quantities in airtight mud cells, in the same manner, that the mud daubers (lukchuk chanuskik) preserve their spiders. Their hunters were very successful; and at midwinter, when all the clans had returned to their camps, they found themselves rich in their supplies of so many things that were good for food, they concluded that ast the best way of expressing their unfeigned gratitude (yokoke ahni) to the great sun they would celebrate a grand, gland feast, and joyous dance, before they commenced the work of clearing and breaking ground for their cornfields. So the cleaned out the dance ground, and planted the pole with the golden sun in the center of it. The people collected and, with much joy and gladness of heart, feasted and danced five days.

39) The amount of ground necessary to plan what corn they had was small, and was soon planted. Then having nothing else to be working at, a thoughtful old man, pointing to the great unfinished mound (yokni chishinto) said, “the weather is cool and pleasant, and the grave of your dead kindred is only half as high as a tall tree.” Taking the timely suggestion of the man, thousands went to work, carrying dirt to the great mound. Afterwards, it became an honorable thing to carry and deposit earth on the mound at any time they were not engaged at work in their domestic vocations.

40) The winter over, spring with its green foliage and singing birds and its grand flourish of gobbling turkeys came slowly on. Corn was planted and the companies of hunters went forth. The camps were healthy. Those who were planting soon finished it, and engaged actively forthwith in throwing earth upon the already huge mound. Their corn flourished well, producing enough, after preserving a portion of their fields for seed, to supply a full feast for the green corn dance.

41) At the Nunih Waya encampment, everything went well and there were no complaints. Their hunters made wide excursions, acquainting themselves with the geography of the country to the extent of many days’ journey around. But, as yet, they had discovered no signs of the enemy, or of any other people. In this happy condition of health and plenty – for they had enlarged their fields and were harvesting abundant crops of corn – years rolled around; the work on the mound was regularly prosecuted; and at the eight green corn dance celebrated at Nunih Waya, the committee who had been appointed at the commencement, reported to the assembled multitude that the work was completed and the mound planted with the seeds of the forest trees in accordance with the plan and direction of the minko, at the beginning of the work.

42) The minko then instructed the good old Lopina, who had carried it so many years, to take the golden sun to the top of the great mound and plant it in the center of the level top.

43) When the people beheld the golden emblem of the sun glittering on the top of the great work which, by the united labor of their own hands, had just been accomplished, they were filled with joy and much gladness. And in their songs at the feast, which was then 230 going on, they would sing:

44) “Behold the wonderful work of our hands; and let us be glad. Look upon the great mound; its top is above the trees, and its black shadow lies on the ground, a bowshot. It is surmounted by the golden emblem of the sun; its glitter (tohpakali) dazzles the eyes of the multitude. It inhumes the bones of fathers and relatives; they died on our sojourn in the wilderness. They died in a far off wild country; they rest at Nunih Waya. Our journey lasted many winters; it ends at Nunih Waya.”

45) The feast and dance, as was the custom, continued five days. After this, in place of the long feast, the minko directed that, as a mark of respect due to the fathers and mothers and brothers and sisters, for whom they had with so much labor prepared such a beautiful and wonderfully high monumental grave, each iksa should come to the mound and, setting up an ornamental pole for each clan, hold a solemn cry a whole moon. Then, to appease the restless spirits of the deceased nation and satisfy all the men and women with what they had done with the sacred relics of their dead, the Choctaws held a grand and joyous national dance and feast of two days. And returning to their tents, they remembered their grief no more.

46) All the people said that their great chief was full of wisdom; that his heart was with the people; and that his counsels had led them in the clean and white paths of safety and peace. Each of the iksas selected very tall pine poles, which they peeled and made white and ornamented with festoons of evergreens and flowers. Then in most solemn form, they performed the cry three times every day, during one whole moon. Then at the great national pole pulling, they celebrated a grand feast and dance of two days. The rejoicing of the nation was very great, and they returned to their camps with glad hearts, remembering their sorrows no more.

47) Afterward, when a death occurred, and the bones had been properly cleansed, they were deposited in a great cavity which had been constructed for that purpose, as the work of the mound was progressing. It was the national sepulchral vault; and thither the bones of all the people that died at Nunih Waya were carried and neatly stowed away in dressed leather sacks. Thus arose the custom of burying the dead in the great monumental sepulchre. And when a member of a hunting party of more than two men or a family died, too far out in the forest to pack home the bones, which could not be cleaned in the woods – for the bone pickers never went hunting – it was deemed sufficient to appease the wandering spirit to place all his hunting implements close to the dead body, just a death had left it. In such cases it was not lawful to touch the dead, and they were covered with a mound of earth thirty steps in circumference and a high as a man’s head. If death occurred at the camp of an individual family in the far off hunt, the survivors would, during the cry moon, carry, in cane baskets and [on] the blade bones of the buffalo, a sufficient amount of earth to construct a mound of the above dimensions. If there should be but two men at a camp, or a lone man and his wife, and one should die, the survivor had to carry the dead body home. Life for life, was the law; and every life had to be 231 accounted for in a satisfactory manner. It would not answer for a man to turn home and report that his hunting companion or his wife had been lost or drowned, devoured by wild beasts or died a natural death. He must show the body. There are occasionally found among the great number of tumuli scattered over the land, mounds of larger dimensions than ordinary ones. These mounds were constructed by females. Upon the death in camp of a man who had an affectionate wife, his mourning tekchi (wife), regardless of the customary time to cry, would throw down her hair and with all her strength and that of her children would carry earth, and build upon the mound as long as they could find food of any kind that would sustain life. They would then return to camp, worn out skeletons.

48) Now, my white friend, I have explained to you the origin, and who it was that built the great number of mounds that are found scattered over this wide land. The circular, conic mounds are all graves, and mark the spot where the persons, for whom they were built, breathed their last breath. There being no bone pickers at the hunting camps to handle the dead, the body was never touched, or moved from the death posture. Just as it lay, or sat, as the case might be, it was covered up, first with either stones, pebbles, or sand, and finished off with earth. In this way the custom of mound graves originated from the great mound grave, Nunih Waya, and it prevailed with the Choctaw people until the white man came with his destructive, sense-killing “fire water,” and made the people all drunk.

49) After getting in possession of this information, in regard to the origin and make of the mounds, I took pains to excavate quite a number of them, which were found on the “second flat” along the Tombecbee river. They continued invariably a single human skeleton. The bones generally, except the skull, were decomposed. The crania of most of them would bear handling, when first taken out, but when exposed to the air they soon fell to ashes. Along with the ashes of the bones, in most cases, would be found five or six arrow points, a stone ax, and not infrequently a stone skin-dresser. In all cases, the bones would be found enveloped, sometimes lying on the side, feet drawn up; at other times in a sitting posture, either in sand, pebbles, or small stones. In one or two cases, the coals and the charred ends of the pine knots that lighted up the last sad night of the deceased, lay in front and near the bones, under the sand.

50) As soon as the national cry was over, the poles pulled down, and the great dance celebrated, the families dispersed into the far off hunting grounds where they enjoyed the game and fruits, until midwinter; when they returned to their homes to prepare and put their fields in order for the coming planting time. The seasons at Nunih Waya were good every year; and they had on hand corn in abundance. Their mode of putting it way, in small lots, in air tight earthen cells, preserved it, from year to year, for an indefinite period, as sound and fresh as new corn. To keep it dry and entirely excluded from air, was all that was necessary, to preserve it for any lenght of time in the same condition in which it was when put up.

51) Feeling themselves permanently settled after the mound was completed, they planted larger crops and were beginning to construct good, dry houses in which to dwell. The 232 next year afer the mound was finished, having a very large crop of corn, they celebrated the green corn dance, eating nothing besides the corn. On the first day of feast, and at the time the people had assembled to receive instructions in regard to the manner of conducting the ceremonies, the minko came upon the dance ground, and calling the attention of the multitude said:

52) “We are a brave and exceedingly prosperous people. We are an industrious people. We till the ground in large fields, thereby producing sustenance for this great nation. We are a faithful and dutiful people. We packed the bones of our ancestors on our backs, in the wilderness, forty-three winters, and at the end of our long journey piled up their memory a monument that overshadows the land like a great mountain. We are a strong, hardy, and very shifty [!] people. When we set out from the land of our fathers, the Chata tribe numbered a little less than nineteen thousand. We have traveled over a pathless wilderness, beset with rocks, high mountains, sun-scorched plains, with dried up rivers of bitter waters; timbered land, full of lakes and ferocious wild beasts. Bravely we have battled and triumphed over all. We have not failed, but are safely located in the rich and fruitful land of tall trees and running brooks see in a vision of the night, and described by our good chief who is missing. And we number now a little more than twenty-one thousand. Assuredly we are a wonderful people. A people of great power. A united, friendly people. We are irresistibly strong (hlampko).”