New interpretations on late prehistoric and protohistoric occupation in the interior of Florida's Central Gulf Coast

New interpretations on late prehistoric and protohistoric occupation in the interior of Florida's Central Gulf Coast

Material Information

New interpretations on late prehistoric and protohistoric occupation in the interior of Florida's Central Gulf Coast
Clagett, Heather Lea
Place of Publication:
Tampa, Florida
University of South Florida
Publication Date:
Physical Description:
vi, 101 leaves : ill., col. maps ; 29 cm.


Subjects / Keywords:
Indians of North America -- Antiquities -- Florida ( lcsh )
Antiquities -- Pinellas County (Fla.) ( lcsh )
Dissertations, Academic -- Applied Anthropology -- Masters -- USF ( FTS )


General Note:
Thesis (M.A.)--University of South Florida, 1995. Includes bibliographical references (leaves 92-101).

Record Information

Source Institution:
University of South Florida
Holding Location:
Universtity of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
021581593 ( ALEPH )
33986067 ( OCLC )
F51-00023 ( USFLDC DOI )
f51.23 ( USFLDC Handle )

Postcard Information



This item is only available as the following downloads:

Full Text


Graduate School University of South Florida Tampa Florida CERTIFICATE OF APPROVAL Master's Thesis This is to certify that the Master's Thesis of HEATHER LEA CLAGETT with a major in Applied Anthropology has been approved by the Examining Committee on August 11, 1995 as satisfactory for the thesis requirement for the Master of Arts degree Examining Committee: Major Professor : J Raymond Williams, Ph. D Member: Roger T. Grange Ph. D Member : Robert Brinkmann Ph. D


NEW INTERPRETATIONS ON LATE PREHISTORIC AND PROTOHISTORIC OCCUPATION IN THE INTERIOR OF FLORIDA'S CENTRAL GULF COAST by HEATHER LEA CLAGETT A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Anthropology University of South Florida August1995 Major Professor: J Raymond Williams Ph.D.


DEDICATION To mom and dad, for all your faith, love and encouragement, and for instilling in me a belief that the only limitations are the ones you place on yourself. And to Rick, whose endless hours of patient reading and rereading, valuable advice, love and support helped to see me through that journey up the mountain.


ACKNOWLEDGEMENTS I wish to extend thanks to Dr. J. Raymond Williams for the ideas upon which this thesis is founded, as well as his advice and encouragement. Great thanks also go out to Dr. Roger Grange and Dr. Bob Brinkmann, whose comments and suggestions really helped me focus. My special gratitude goes out to Joan Deming, without whose knowledge and priceless library this thesis would not have developed. Appreciation is also extended to the many friends and colleagues who have both knowingly and unknowingly helped me along the way. And to Shelly Happel, thank you for your hours of patient instruction, advice, positive reinforcement, and friendship. I would never have made it past the research design without your help.


LIST OF TABLES LIST OF FIGURES ABSTRACT TABLE OF CONTENTS CHAPTER 1. INTRODUCTION Development of the Study Statement of the Problem Limitations of the Study CHAPTER 2. REGIONAL PREHISTORY CHAPTER 3. REVIEW OF PREVIOUS RESEARCH CHAPTER 4. THE STUDY AREA AND ITS ENVIRONMENT Environment The Study Area Defined Lithic Resources Implications of Environment on the Study Area CHAPTER 5. THE RESEARCH DESIGN Methodology The Nearest Neighbor Statistic The Data Set Defined Polk County The Lake Marion I Site (8-Po-2) The Frostproof Mound (8-Po-7) The Singletary Site (8-Po-13) The Nalcrest Site (8 Po 15) The Raulerson Mound (8-Po-123) The Philip Mound (8-Po-446) The Haines City River Site (8-Po-1 036) Hardee County The Davis Burial Mound (8 Hr-1) Carlton Ranch No 1 Site (8-Hr 5) The Orchard-Fenceline Site (8-Hr-11) iii iv v 1 1 3 3 5 18 24 24 27 30 31 33 33 37 45 46 46 48 48 49 49 50 52 52 52 53 54


CHAPTER 5. (Continued) Hardee County (continued) The Cowboy Mound (8-Hr-15) 54 The Welch Mound (8-Hr-16) 56 The Ona Road Mound (8-Hr-17) 58 The Breton Mound (8-Hr -18) 60 The Little Mound (8-Hr-19) 61 The Bostwick Mound (8-Hr-52) 63 Keen Mound Complex (8-Hr-149) 64 Manatee County 66 Sugarbowl Mound (8-Ma-64) 66 8-Ma-65 67 The Airstrip Village Sites (8-Ma-181) APLS #4a and #4b 68 DeSoto County 70 The Pine Level Site (Keen Mound Site) (8-De-2) 70 The Pine Level 2 Site (8-De-3) 72 The Brandy Branch Village Site APLS #5 (8-De-4) 72 The Mizell Mound "A" (8-De-31) 73 The Mizell Mound "B" (8-De-32) 75 The Mizell Mound "C" (8-De-33) 76 The Cunningham Mound (8-De-34) 77 CHAPTER 6 RESULTS AND INTERPRETATIONS 79 Results of Nearest Neighbor Analysis 79 A Question of Method 84 Interpretations for the Region 88 REFERENCES CITED 92 II


LIST OF TABLES Table 1 Micro-Environments and Resources within the Study Area 29 Table 2. Site Inventory of the Data Set 47 Ill


LIST OF FIGURES Figure 1. Locational Map of General Study Area 28 Figure 2. Map Depicting Location and Type of Sites within Study Area 35 Figure 3 Quadrats Subjected to Nearest Neighbor Analysis 41 Figure 4. The Effect of Area on Rn Values for a Constant Regular Point Pattern 43 Figure 5 Sites in Northern Quadrat 80 Figure 6. Significance Graph for Testing Rn Values 82 Figure 7 Sites in Southern Quadrat 83 Figure 8. Distribution of Sites with Respect to Cultural Affiliation 86 iv


NEW INTERPRETATIONS ON LATE PREHISTORIC AND PROTOHISTORIC OCCUPATION IN THE INTERIOR OF FLORIDA'S CENTRAL GULF COAST by HEATHER LEA CLAGETT An Abstract Of a thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Anthropology University of South Florida August 1995 Major Professor : J. Raymond Williams, Ph.D. v


Once thought of as a "hinterland" exploited by people of the Safety Harbor Culture (A.D 800A.D 1000), and later, the "heartland" of a related culture, the interior of Florida's Central Gulf Coast Archaeological Region now appears far less homogeneous than once believed Milanich's (1978) Cades Pond Occupational Nexus Model, W'hich defined six separate nexuses via the nearest neighbor statistic, is tested utilizing data from 27 pre-contact and post contact archaeological sites in the interior Central Gulf Coast region. The results reveal site clusters in a region of numerous diverse micro-environments. The inhabitants of this region were influenced not only by Safety Harbor, but the Glades, and Kissimmee River Valley cultures as well. Evidence also suggests this region was occupied continually for several thousand years. Conclusions addressing the archaeological evidence exhibited by sites located in the inland region and how they relate to the concept of archaeological culture will also be discussed. It is hoped the results of this study will lead to both a better understanding of this portion of the Central Gulf Coast region of Florida during late prehistoric and protohistoric times, as well as encourage a resurgence of spatial pattern studies throughout Florida. Abstract Approved: ------------------Major Professor: J. Raymond Williams, Ph.D Professor, Department of Anthropology Date Approved: tc.< l \&\"'-,-I VI


Development of the Study CHAPTER 1 INTRODUCTION 1 In the spring of 1994, Phase II excavations were conducted on CF Industries Property in Hardee County, Florida, in a cooperative project between Archaeological Consultants Inc. and the University of South Florida, Department of Anthropology. A thesis began to grow from the results of these excavations and originally attempted to focus on defining the Interior variant of the Safety Harbor culture based on Mitchem's 1989 dissertation. As data collection concerning all Safety Harbor period sites in the interior region (Polk, Hardee, and DeSoto Counties) progressed, it became apparent that there were many factors influencing site location and that not all of the sites in this region which corresponded to Safety Harbor temporally were actually Safety Harbor sites. Many of the sites seem to have been influenced by other surrounding regions, such as the Kissimmee River Valley, Glades and Caloosahatchee regions. Deming and Williams (1994) noted the obvious visual settlement pattern existing for the five CF Industries mounds in Hardee County. The mounds were spaced over the landscape in the form of two triangles They also noted the same configuration existing for three mounds in southern DeSoto County (Deming and Williams 1994; Willis and Johnson 1980). As a result of these observations, Deming and Williams ("1994) re introduced a settlement model which Jerald Milanich described in 1978. Milanich's Occupational Nexus model was developed utilizing five types of


2 Cades Pond archaeological sites: 1) mound with semicircular earthwork and adjacent village; 2) mound with adjacent village located within 0.5 kilometers; 3) mound with associated village within 0.5 to 1.5 kilometers; 4) specialized site with dense but localized midden accumulations not indicative of the full range of site activities; and 5) camp site represented by a diffuse midden (Milanich 1978:166). Milanich (1978) noted that the geographical locations of Cades Pond sites appeared to lie in clusters or groups, each of which contain one or more type #2 or #3 sites. The earlier clusters also contained at least one type #1 site, and type #5 sites occured in all clusters. It was suggested that nearest neighbor analysis would sufficiently quantify these clusters (Milanich 1978 : 167). The pattern which resulted from these clusters is one of related or connected sites occupying a restricted geographic locale which is relevant to both the social and natural environments (Milanich 1978:167). For example, villages always bordered lakes, ponds, swamps or marshes, which enabled maximum accessibility to aquatic flora and fauna; a situation almost identical to that which exists in the interior Central Gulf Coast region Milanich s model resulted in the definition of six Cades Pond occupational nexuses, each of which he believed functioned as separate political units, but which were bound together by social organization, shared beliefs and economic ties (Milanich 1978 : 170). Milanich believed the nexus model could be quantitatively tested for the Cades Pond region and applied to other areas of the state, such as the Lake Okeechobee Basin between ca. 500 B C and A.D 1000 or later It is on this model and Deming and Williams' (1994) work at the CF Industries mounds, that this thesis is based.


3 Statement of the Problem The focus of the study is twofold. First, it must be determined whether sites in the interior Central Gulf Coast region coeval with the Safety Harbor period, A. D. 800 1725, cluster to form nexuses. Second, some conclusions concerning what was occurring culturally, in the interior Central Gulf Coast region during the period from A.D 800 to 1725 must be addressed In other words, are all the sites from this period and in this region actually Safety Harbor sites, or do some of them exhibit characteristics of other cultures? Two working hypotheses were constructed: Hypothesis 1) Sites in the interior Central Gulf Coast region dating to the period A. D. 800 to 1725 form settlement clusters on the landscape; Hypothesis 2) Sites in the interior Cental Gulf Coast region dating to the period A.D. 800 to 1725 cannot exclusively be attributed to the Safety Harbor culture. These hypotheses and explanations of why this temporal period was chosen will be discussed in greater detail in Chapter 5. Limitations of the Study A number of factors affected data collection, analysis, and the results of this study. The greatest limitation on this study is that the interior of the Central Gulf Coast region is one of the least investigated areas in the state In fact, much of what is known about the area stems from cultural resource management surveys and excavations over the past 20 years. As is common in archaeology, the majority of the early work in the Central Gulf Coast region was confined to the data rich temple mounds, burial mounds, and shell middens of the coast.


With less investigation comes a smaller database with which to work, and an even smaller sample to test statistically 4 A related problem exists in variations in site recording over the past fifty years. Early Florida Site File Forms consisted of small index cards, upon which location and a brief description of the site was to be filed. In many cases, exact location was not recorded and very limited and/or vague information concerning the site was included. Thus, a handful! of possibly eligible sites had to be dropped from consideration in the sample due to a lack of needed information. Cultural Resource Management Reports yielded the best site data. This type of literature always included detailed artifact assemblage information in the form of types, counts and weights, in addition to precise site locations in both Universal Transverse Mercator and longitude and latitude coordinates. The chosen method of quantitative analysis, nearest neighbor analysis, by its nature, introduced a number of limitations to the study. These shall be discussed in detail in Chapter 5 when the pitfalls and considerations of the statistic are defined One requirement for using nearest neighbor analysis, temporal control, is also a factor which affects the study as a whole As is the nature of archaeology, temporality of a site is often determined by diagnostic artifacts within the site assemblage. But these chronologies are ever changing, and as more is learned about the archaeology of the interior region, greater chronological control will be achieved. As conditions exist now however, the rather broad temporal period of A.D. 800 to 1725, though on the large end of the scale for nearest neighbor analysis, is acceptable.


CHAPTER 2 REGIONAL PREHISTORY 5 In order to fully understand aboriginal occupation in the interior Central Gulf Coast, it is necessary first to comprehend the prehistory of the entire region. Defined by Milanich and Fairbanks (1980) as the Central Peninsular Gulf Coast this region encompasses a large area from just north of Tampa Bay southward to the northern portion of Charlotte Harbor. The eastern boundary however, is debateable. Determining a regional border is made difficult by the fact that the archaeology of the interior portion of this region is poorly understood Later prehistoric sites located in Hardee, Polk and DeSoto counties appear to contain cultural materials which either come from or have been influenced by the Kissimmee River, Caloosahatchee and (or) the Okeechobee Basin regions (Milanich 1994). Thus, as the development of reg i onal cultures becomes evident in the archaeological record, it becomes more difficult to assign cultural affiliation to sites in this interior region. This will be addressed in detail in Chapter 7 A cultural chronology of Florida has been outlined by Milanich and Fairbanks (1980), and most recently by Milanich (1994), consisting of the Paleo Indian, Archaic, Manasota, Weeden Island, Safety Harbor and Historic periods. Milanich and Fairbanks (1980) included a Transitional period between the Archaic and Manasota periods Recently, Milanich (1994) has redefined the culture periods so as to exclude the Transitional period as it is part of a cultural continuum This development in the cultural chronology of the C en tral Peninsula Gulf Coast region will be discussed in more detail below


6 Paleo-Indian peoples entered what is now Florida some time during the close of the Pleistocene. While the exact date of their arrival is under debate, most archaeologists agree people had reached this part of North America by at least 12,000 years ago (Milanich 1994). Paleo-Indian peoples were descendants of those who crossed the massive land bridge, which extended from Asia over what is now the Bering Straits and into North America. It is estimated that this "land bridge" was at least 1,000 miles wide and it is likely people not only migrated across it but lived on it for many generations (Milanich 1994; Hoffecker, Powers, and Goebel 1993). At the end of the Pleistocene, the Florida peninsula was vastly different than today. With millions of gallons of water locked in glacial ice, sea levels worldwide were as much as 430-550 meters lower than at present, placing Gulf coast shorelines approximately 75 kilometers west of where they are presently located (Milanich 1994 ). Milanich ( 1994) has noted that about half of the land exposed 12,000 years ago is now inundated. Therefore, it is highly likely that Paleo-Indian sites exist below both the waters of the Gulf of Mexico and the Atlantic coast (Deming and Williams 1994; Dunbar, Webb and Faught 1991; Ruppe 1980; Clausen et al. 1979). Florida's inland environments then consisted of arid uplands characterized by xerophytic species such as scrub oaks, palmetto, and sand pines, and occasional river drainage systems which Paleo-Indian peoples certainly utilized. As the water table was much lower, many of the springs which feed rivers such as the lchetucknee, did not exist 12,000 years ago, and marshes and wet prairies were virtually nonexistent (Milanich 1994; Daniel and Wisenbaker 1987). Potable water in the interior was obtained from limestone bottomed catchments lined with marl type deposits and fed by rainfall, as well as in deep sinkholes which received water via rainfall and occasional groundwater


7 from deep springs (Milanich 1994 ). Sinkholes of this type are characteristic of Florida's karst topography formed in Tertiary age limestone deposits. It is around spring fed sinks such as these that Paleo-Indian peoples camped and gathered resources (Deming and Williams 1994; Clausen et al. 1979). The oasis model, proposed by James S Dunbar and S David Webb (1983, 1991) suggests that watering holes such as these sinks, were crucial to Paleo-Indian settlement patterning and subsistence. Watering holes were undoubtedly gathering places for animals which in tum provided good hunting for carnivores, including Paleo-Indian populations. If this hypothesis is correct, the presence of Paleo-Indians identifiable in the form of artifacts, and perhaps the remains of butchered animals, should be evident at these watering holes Such evidence has been recovered from Little Salt and Warm Mineral Springs in Sarasota County Similar discoveries have also come from rivers such as the Wacissa, which yielded a Bison antiquus skull with a broken stone projectile point still lodged within it (Milanich 1994; Webb et al. 1984) Further research reveals that the presence of Paleo-Indian materials was almost entirely restricted to regions where karst topography is evident on the surface in the form of sinkholes In fact ninety-two percent of Clovis and Suwannee projectile points, diagnostic points of the Paleo-Indian period, come from these geographic regions {Milanich 1994; Dunbar 1991 ). The Paleo-Indian tool kit consists of the large fluted Clovis and Suwannee projectile points, unifacial scrapers adzes, various blade tools, spokeshaves and hafted bifacial knives Research at sites such as Harney Flats in the Hillsborough River floodplain, has demonstrated that Paleo-Indian peoples possessed a specialized tool kit which was an indispensable aspect of their subsistance strategies (Milanich 1994; Daniel and Wisenbaker 1987)


8 By around 7,500 B.C. climatic changes associated with the end of the Pleistocene were evident in the Central Gulf Coast region. The Pleistocene megafauna were virtually extinct by this time, and as sea levels gradually rose with the melting of the glaciers, the climate began to resemble that of today Pine forests slowly replaced the arid uplands of the interior, and marshes and wet prairies developed along the coasts. These environmental changes led to gradual changes in culture toward increasingly intensive exploitation of localized food resources. Smaller game such as deer, raccoon and opossum as well as shellfish and a wide variety of wild plants became the focus for Archaic populations (Deming and Williams 1994; Milanich 1994). Traditionally, the Archaic period has been divided into: Early (6,5006 000 B.C.) Middle (5,000 3,000 B.C.), and Late (3,000-1,200 B.C.) periods (Milanich 1994; Miianich and Fairbanks 1980). But in light of new research, archaeologists have split the Archaic period into a pre-ceramic and a ceramic sub-period Within this chronology, diagnostic projectile points are specific to each sub-period and correspond temporally with the advent of fiber tempered Orange Period pottery As noted above, the environmental changes occurring around 7,500 B.C. had an effect on prehistoric populations as well. Archaeological evidence indicates that early preceramic archaic peoples were exploiting interior and riverine environments as well as coastal areas of the peninsula. This change in subsistence strategies is apparent in the artifact assemblage The large Paleo Indian projectile poinUknives were replaced with smaller stemmed tools, such as the Kirk, Hamilton, and Wacissa types (Milanich 1994:63) Undoubtedly, tools made from bone, wood and antler and other subsistence procurement items such as baskets and nets, were manufactured and utilized by both Paleo-Indian and early pre-ceramic Archaic period peoples Although preservation of these


9 items is rare, bone and wood tools, as well as reed mats and other evidence of cordage, have been excavated from the peat bog layers at the bottom of Windover Pond in Brevard County, a pre-ceramic Archaic mortuary pond (Milanich 1994; Deming and Williams 1994). This anaerobic environment also contributed to the excellent preservation of human remains, including brain tissue and stomach contents (Milanich 1994 ; Dickel and Doran 1989). The analysis of the materials from Windover greatly contributed to our knowledge of the early Archaic period Most early pre-ceramic Archaic sites are small, seasonal campsites which suggest that small bands moved around seasonally in search of food and other resources (Deming and Williams 1994). The Little Payne Mining Tract Site #1 in Hardee County may be one such campsite as evidenced by the recovery of a Bolen type projectile point dating to the early pre-ceramic Archaic (Deming and Williams 1994; Batcho 1978 : 11-14 ). Between 5,000 and 3,000 B.C. the environment and climate ameliorated to almost modern conditions (Milanich 1994 ) A shift in settlement patterns from the dispersed pattern of early pre-ceramic Archaic populations to a system of base and resource extraction camps has been hypothesized for the middle and late pre-ceramic Archaic periods. This settlement pattern shift resulted in a maximizing of resources which suggests that larger bands of people may have lived together at least part of the year (Deming and Williams 1994) Subsistence was assisted by the use of stone tools such as the broad-bladed, stemmed Newnan, Marion and Putnam type projectile points, specialized tools such as microliths, burins, and large chopping tools, and an array of expedient tools in the form of utilized flakes and scrapers. The cessation of pond burials, such as those at Tick Island in Eastern Florida, is also a hallmark of this period The


10 Tick Island burials, dated to 3,400 B.C. are the most recent of the pond burials, which are replaced by a new burial pattern (Milanich 1994:84 ). Several campsites from the middle and late pre-ceramic Archaic period have been excavated in the Central Gulf Coast region, many as a part of the Interstate 75 archaeological project in the late 1970s and early 1980s. These include; the Deerstand (Daniel 1982), Wetherington Island (Chance 1982) and Muck Pond East (ACI/Janus Research 1994) sites in Hillsborough County, and Republic Groves (Wharton, Ballo and Hope 1981 ), and four lithic scatter sites (Wood and Williams 1976) in Hardee county. These four lithic scatter sites, 8Hr8, -9, -12, and -14, excavated by Wood and Williams (1976) and located on CF Industries property, yielded projectile points diagnostic to both the pre ceramic and ceramic Archaic periods (5,000 1,000 B. C). The late pre-ceramic Archaic period (roughly 3,000 to 2,000 B.C.) is characterized by increased populations and regionalization of cultures as people continued to adapt to more specific and diverse environments. With the arrival of essentially modern environmental and climatic conditions, shellfish, fish and other resources were available in greater numbers along the marshes, interior rivers, and coastal wetlands (Milanich 1994). Evidence that late pre-ceramic Archaic populations were utilizing these resources can be seen in the extensive shell middens along the east and west coast of Florida as well as the St. Johns river and the Ten Thousand Islands area in southwest Florida (Milanich 1994:85). The Palmer site in Sarasota County, a horseshoe shaped shell midden encircling a freshwater spring adjacent to Sarasota Bay, is perhaps the most well preserved example of this site type in the Central Peninsula Gulf Coast region (Deming and Williams 1994: 13). The broad bladed, stemmed projectile points of the preceding period continue to be utilized. However, a host of other tools associated with a greater reliance on marine resources, such as


11 fish hooks, and net weights, appear during this period. Analysis of shell midden materials has contributed a wealth of information on late pre-ceramic Archaic lifeways including subsistence data in the form of floral and faunal remains. The first fired clay pottery appears by around 2,000 B.C This technology was either invented in Florida or diffused into Florida from coastal Georgia and South Carolina, where early dates for pottery have also been obtained (Milanich 1994:86). The Orange culture of the St. Johns region in northeast Florida was thought to be the first to manufacture pottery in Florida. However, recent research has shown that fiber tempered pottery appears just as early in southwest Florida (Milanich 1994; Cockrell 1970; Widmer 197 4; McMichael 1982; Russo 1991 ). The earliest ceramics were believed to be fiber tempered and plain, but ceramics with sand, plant fibers and quartz have also been found in Orange Period contexts (Milanich 1994). After 1650 B.C fiber tempered ceramics were decorated with punctations and incisions These decorations are an essential characteristic used in dating early and late fiber tempered ceramics during their period of manufacture, 2,000 to 1,000 B.C. (Milanich 1994 : 86) The artifact assemblage at the Canton Street site in Pinellas County consists of sand tempered, sand and limestone tempered, and temperless chalky ceramics in association with what Purdy (1981) has identified as ceramic Archaic period projectile points as well as pre ceramic Archaic period points. Bullen ( et al. 1978) suggested that the admixture of basally notched, side and corner notched and Archaic stemmed projectile points, and what he saw as three different ceramic traditions, marked a dynamic period he termed the Transitional Period However, Milanich (1994) asserts that changes in ceramic technology alone do not necessitate cultural transition Rather, these changes in technology should be viewed in terms of their position on a technological


continuum. Milanich (1994) therefore suggests the use of the Transitional period be discarded. 12 By around 500 B. C. settlement patterns in the Central Peninsula Gulf Coast Region consisted of permanent residence on the coast for most of the year with seasonal forays into the interior to gather resources (Deming and Williams 1994). Luer and Almy (1982) have defined the Manasota culture, a regional culture which lasted from 500 B.C. to A.D. 800 in the Central Peninsula Gulf Coast Region. Manasota defines both the culture and the geographic region consisting of Manatee and Sarasota counties, the area in which the cultural complex was first recognized (Luer and Almy 1982). The Manasota culture is characterized by a fishing, hunting and shellfish gathering economy in which both marine and "hinterland" environments were exploited. The major villages were located on or near the mainland and were probably inhabited most of the year. Inland sites were located in the pine flatwoods, on high ground, and near a source of freshwater. From these locations multiple environments were accessible; freshwater swamp, hardwood hammocks and pine-palmetto flatwoods (Luer and Almy 1982). These inland sites were probably occupied only for brief periods during the year. There are several sites in the interior region which may be contemporaneous with coastal Manasota sites. The Orchard-Fenceline site (8Hr-11) in northwest Hardee County has been dated at 185 B.C. to A.D. 875, and may represent an interior Manasota site (Deming and Williams 1994; Wharton 1977:48). Artifacts recovered from Manasota contexts include; undecorated, sand tempered, thick-walled, large ceramic vessels such as globular bowls, an array of bone tools such as shark tooth drills and stingray spine pins, and stone tools such as knives, scrapers, drills, sandstone abraders, chert cores, and blanks, although this later stone technology was poorly developed. Shell tool


13 manufacture however, was highly developed among Manasota populations. Strombus shell hammers, Mercenaria anvils and scrapers, Busycon shell spoons, pounders, celts, columellae, spokeshaves and hammers, and Noetia net sinkers have all been recovered from Manasota sites, reinforcing the notion of a strong reliance on the sea (Luer and Almy 1982:40-44). Burial practices changed throughout the 1,300 year period from primary flexed burials in midden debris to primary flexed burials in burial mounds and finally secondary burials in burial mounds (Luer and Almy 1982 : 40-44). This method of burial in conjunction with the recovery of Weeden Island complicated stamped pottery from many Manasota burial mounds, may represent an early manifestation of Weeden Island ceremonialism (Luer and Almy 1982; Milanich 1994). Changing burial practices are one of the many influences from the Weeden-lsland culture to the north, that transformed Manasota to a Weeden lsland related culture (Deming and Williams 1994; Milanich 1994) Although larger populations and increased sedentism are inferred for this period, the subsistence pattern continued to consist of exploiting marine and interior areas for resources (Deming and Williams 1994) Sites such as the Branch Mound, Thomas Mound (Bullen 1952), South Prong I Mound (Martin 1976), Parris Mound 5 (Willey 1949), and the Stanley Mound (Deming 1976), are good examples of the site type for this period These sites usually consist of shell middens or habitation areas in association with a sand burial mound (Deming and Williams 1994) Some burial mounds might have been shared by one or more villages (Willey 1949). Burial practices, elaborate trade networks, and settlement patterns all suggest a complex socio religious system (Deming and Williams 1994 : 14)


14 The final cultural period in the Central Peninsula Gulf Coast region is Safety Harbor, named after the type site for the Safety Harbor culture in Pinellas County. It is probable that the Weeden-lsland related cultures developed into the Safety Harbor culture as many Safety Harbor sites contain evidence of previous occupations of Weeden-lsland/ Manasota related populations (Deming and Williams 1994: 15). The Bayshore Homes site and Weeden-lsland, both excavated by Sears (1971 ), are two examples of this evolutionary relationship (Milanich 1994:228). Mitchem (1989) has divided the Safety Harbor period into four phases: the Englewood Phase (A.D. 800-1000), the Pinellas Phase (A.D. 1 000-1500), the Tatham Phase (A.D. 1500-1567), and the Bayview Phase (A.D. 1567-1625) Mitchem (1989) has also identified five regional variants. The Northern Variant is bounded by the Withlacoochee River on the north and northeast and extends to south Pasco County. Hillsborough, Pinellas and part of southern Pasco counties compose the Circum-Tampa Bay Variant. The Manasota area which has previously been defined (Luer and Almy 1982) is the third regional variant. The least understood of Mitchem's five variants is the Interior Variant, which is comprised of Polk and Hardee counties, part of eastern DeSoto County, and possibly portions of Highlands County. The fifth and final regional variant is the South Florida Variant, which is composed of Lee, Collier, Glades and Hendry counties The exact boundary is under contention as it lies within Calusa territory (Mitchem 1989) The major Safety Harbor villages were located along the coasts with small, dispersed settlements in the interior (Mitchem 1989:586). These large villages consisted of a temple mound, (sometimes with a ramp extending off the mound to the west), plaza, midden, and burial mound (Deming and Williams 1994; Mitchem 1989). Large population Safety Harbor centers include Safety


15 Harbor (Sears 1958; Griffin and Bullen 1950), Maximo Point (Bushnell 1962; Sears 1958), the Narvaez Midden (Bushnell 1966), and Tierra Verde (Sears 1967) in Pinellas County; Parrish Mounds 1, 2 and 3 in Manatee County (Willey 1949); the Davis Mound (Bullen 1954) and Bostwick Mound in Hardee County; and the Keen Mound (Willis and Johnson 1980) in DeSoto county (Deming and Williams 1994: 15). It is suggested that limited maize agriculture was practiced by Safety Harbor peoples (Luer and Almy 1981 ), but Mitchem (1989) asserts that there has been no solid evidence to support this. Indeed, the only surviving floral remains from the Tatham Mound in the Cove of the With lacoochee area, consisted of a cast of a charred Cucurbita seed which was not a domestic species (Mitchem 1989). The Englewood and Pinellas phases exhibit ceramics styles and techniques and treatment of the dead similar to earlier Weeden Island cultures. Sand tempered plain pottery, usually decorated with serrated rims, continued to be the primary type and is considered utilitarian As towns continued to develop, charnel structures for preparing the dead and Southeastern Ceremonial Complex motifs appeared on Safety Harbor mortuary and ceremonial ceramics Artifacts such as copper breast plates, a copper plume head ornament ear spools and other items of personal adornment have been excavated from Tatham Mound, which also supports contact with more southern variants of Mississippian cultures (Milanich 1994; Mitchem 1989) However, Safety Harbor leaders did not hold the ascribed elite status typical of leaders among the Mississippian peoples. Additionally, it is evident that Safety Harbor populations did not practice large scale agriculture, an integral part of Mississippian economies (Milanich 1994 : 412).


16 The Timucuan Indians are recognized as the population which comprised the Safety Harbor culture. In the Tampa Bay area they were known as the T ocobaga (Deming and Williams 1994 ). These were the people whom Spanish explorers Panfilo de Narvaez and Hernando de Soto encountered when they landed on the shores of Tampa Bay. In 1513 the Spanish entrada into Florida began. The Soto expedition visited the Safety Harbor town of Uzita and Narvaez's expedition entered a town in Pinellas County, known today as the Narvaez Midden (Mitchem 1989; Milanich and Fairbanks 1980) Evidence of Spanish contact exists in the form of iron artifacts and a multitude of glass beads at major mound sites such as Tatham Mound, Narvaez Midden, and Safety Harbor (Mitchem 1989). During the Bayview Phase many Safety Harbor groups suffered from extreme stress related to European contact. Disease was rampant and the Tocobaga and Calusa were enemies most likely at war at the time of Menendez's arrival in 1567 (Mitchem 1989 : 575; Solis de Meras 1964:224-225) Reports from the Menendez expedition indicate that these factors had significantly weakened some Safety Harbor groups and had allowed the Calusa to expand northward, increasing their power (Mitchem 1989). The Calusa were not immune to the ravaging effects of the Spanish entrada and by the first half of the eighteenth century, native populations in the Central Peninsula Gulf Coast region as well as in other portions of Florida, had disappeared (Deming and Williams 1994 : 15) Soon thereafter, groups of Creek Indians were driven into Florida in an attempt to avoid annihilation from American settlement in their homelands. These people became the Seminole a Spanish word meaning wild or untamed (Hudson 1976) Their struggle to survive continued during the First Seminole War of 1818 After the Second Seminole War of 1840 many Seminole people


17 settled in the Tampa Bay area (ACI/Janus Research 1994; Milanich and Fairbanks 1980) Although temporally recent, very little is known about Seminole sites in this region. One of the few known sites in the Tampa Bay area is the Quad Block site in downtown Tampa, where hundreds of Seminole burials were encountered and recovered from part of the old Fort Brooke cemetery (Piper and Piper 1982). In response to the passage of the Homestead Act in 1847, American settlers continued to move south. Seminoles were captured and sent to Oklahoma in exile. However, several small bands of Seminole people managed to escape into the Everglades region where they were eventually locked i nto reservation lands which continue to shrink today


18 CHAPTER 3 REVIEW OF PREVIOUS RESEARCH Understanding a region demands more than just knowledge of the prehistory of that region. Equally important is the history of the archaeological research conducted in a region. This information allows one to view the breadth and scope of investigation in the region and acts as an commencement point for new research. The earliest documented archaeological investigations in the Tampa Bay area were conducted during the mid to late nineteenth century. Early investigators such as Daniel Brinton (1859, 1867), Jeffries Wyman (1870), R.E.C. Stearns (1870, 1872), Frank Cushing (1897), James Shepard (1886), W.W. Calkins (1878), and S.T. Walker (1880a, 1880b, 1885) excavated and recorded a number of sites in the Tampa Bay region. Although, their "excavations" were both destructive and stratigraphically uncontrolled, and their interpretations speculative at best, these early researchers recorded many sites which would soon be destroyed by construction activities. In the early twentieth century, Clarence B. Moore (1900, 1903, 1905) conducted surveys and excavations at numerous sites in the Central Gulf Coast Region. He kept detailed notes concerning the sites he excavated and the materials recovered from them. Although not particularly interested in culture history, Moore's work was incorporated into a subsequent synthesis of Florida archaeology.


19 The majority of the archaeological investigations in the Central Gulf Coast Region prior to the1930s, were concentrated along the coast. With the development of the WPA (Works Progress Administration) in the mid-1930s, sites in central and eastern Hillsborough, Sarasota and Manatee counties were investigated. Members from the Smithsonian Institution also explored this part of Florida. Sterling (1935) conducted excavations on five mounds near the town of Parrish in eastern Manatee County. Data from these excavations allowed Padgett ( 197 4) to draw some conclusions about the inland Manatee region. Only one of the dateable sites contained materials which were culturally affiliated to an archaeological period other than the Late Archaic or Safety Harbor period, and with the exception of the Parrish Mound 5 Site, none of the inland sites in the Manatee Region were culturally affiliated with the Perico Island or Weeden Island periods Based on Sterling's (1935) data, Padgett (1974 : 30) hypothesized a 1000 year cultural hiatus for the interior region of the Central Gulf Coast. The interior region thus became a "hinterland" into which aboriginal people ventured on short forays to obtain resources (Padgett 197 4 ). It was not until the late 1960s that counties in the interior region, (Polk, Hardee, and DeSoto) began to be investigated. In Polk County, Benson (1967) excavated the Philip Mound, a post-contact Safety Harbor period burial mound A 12 to 15 foot wide earthen ramp extended off the northern portion of the mound eastward approximately 200 feet to a borrow pit and then continued westward to within five feet of the southern portion of the mound Fragments of human bone were found with a multitude of European glass beads, including Nueva Cadiz beads which are chronologically sensitive and date between A .D. 1500 and 1550 (Deagan 1987) Other evidence of European contact includes a wide variety of metal beads, metal pendants of various sizes, and iron tools such as scissors and an axe head. Polished stone celts and pendants Pinellas and


20 pre-ceramic Archaic projectile points and drills, and aboriginal ceramics typical of the Safety Harbor period, i.e. Sarasota Incised, Pinellas Incised, Englewood Incised, Pinellas Plain, and Belle Glade Plain, were also recovered (Benson 1967). Benson (1967) believed the Philip Mound to be comparable temporally and in material association with the nearby Goodnow Mound (Griffin and Smith 1948) in Highlands County. Griffin and Smith (1948) related the Goodnow Mound to the Glades area and Benson (1967 : 131) concluded the Philip site had strong characteristics of both late Glades and late Gulf Coast influences: an important observation for sites in this interior region. Yet the idea of the "hinterland," a region which was utilized for short hunting forays and the collection of limited resources, persisted. The interior region was seen as a buffer zone between the Gulf coast cultures and those of the Okeechobee region. Milanich and Martinez (1975) excavated four sites on the Carlton Ranch Property in western Hardee County. Although three of the sites were Archaic period lithic scatters, the remaining site consisted of an aboriginal sand mound. Milanich and Martinez (1975) believed the sand tempered plain and laminated, contorted paste sherds recovered from this site, the Carlton Ranch Number 1 site (8 Hr-5), to be identical to those recovered from the Fort Center site. The site was therefore interpreted as a house site occupied by a northern extension of the Okeechobee basin peoples dated to about A.D. 200 (Milanich and Martinez 1975) In 1977, Willis and Milanich concluded, of the interior region that ... at no time during the prehistoric period was the region ever an important culture area" ( 1977:217). Some archaeologists, however, did not agree with the "hinterland" hypothesis for this region Wood and Williams (1976) presented the results of initial survey of the CF Industries Property. Five sand mounds, one lithic scatter,


21 and two habitation sites were recorded. Wood and Williams (1976) assigned these sites to the Glades periods A.D. 500-1500 and concluded that the interior region possessed potential to yield valuable archaeological information. As a part of the AMAX. Pine Level Survey in Manatee and DeSoto counties, Willis and Johnson (1980) identified 14 aboriginal sites, eight of which were sand mounds, with temporal spans from the mid to late Archaic to the Safety Harbor period. They concluded that the sites contained "important and irreplaceable elements of Florida's prehistory" and that more research was needed in the interior region; quite a different view from 1977 (Willis and Johnson 1980:121 ). In a further attempt to discount the "hinterland" hypothesis, Wharton and Williams (1980) synthesized the research conducted in Hardee County and concluded the area was not a hinterland but rather a "heartland." They cited a number of sites including the Davis Mound complex, a truncated platform mound, three village sites, and at least three burial mounds, all dating to the Safety Harbor period in eastern Hardee County (Wharton and Williams 1980:217). Another Hardee County site of great significance is the Republic Groves Site, a middle to late Archaic period site discovered during a channelization operation when human burials were exposed by dragline. Mitchell Hope conducted controlled excavations at the site to salvage as much data as possible before the site was destroyed (Wharton, Ballo and Hope 1981 ). In addition to the large number of burials (possibly over one thousand) preserved in the muck, wooden stakes, incised bone, antler and wood and most notably, the remains of human brain tissue were recovered (Wharton, Ballo and Hope 1981 ).


22 Another important site is the Keene Mound Complex located in southwestern Hardee County. The site consists of a large, rectangular platform mound, a smaller, conical, burial mound surrounded by a horseshoe shaped embankment, at least one associated borrow pit, and a small low sand mound. A rectangular shaped earthen enclosure is also present and may be of prehistoric origin (Janus Research 1994 ). The earthwork surrounding the Keene Burial mound is similar to those appearing at Glades culture sites in the Okeechobee basin (Wharton and Williams 1980:218). While the cultural affiliation of the Keene Mound has not been established, the recovery of three sand tempered plain and one Belle Glade Plain ceramic sherd(s) suggest a date of AD. 500 to 1500. Furthermore, the presence of this complex in the interior region cannot be overlooked. The Little Payne Creek #7 Site, surveyed in 1978 by Batcho and located in northwest Hardee County, yielded both Dunn's Creek Red and Englewood Plain sherds. A number of other ceramic bearing sites possibly associated with this burial mound, were located during this mining land survey (Wharton and Williams 1980:218). Piper and Piper ( 1981) carried out test excavations on seven sites, previously recorded by Willis and Johnson (1980) and located on AMAX property in DeSoto and Manatee counties. These sites include; 8-Ma-181 a large sand mound containing late Weeden Island/early Safety Harbor period ceramics and projectile points, large amounts of animal bone and some shell; the associated village site to 8-Ma-181, which was off AMAX lands and therefore not surveyed; the Brandy Branch Village site (8-De-4) which has been tentatively assigned to the Safety Harbor period; 8-De-8 and 8-De-9, the Horse Creek #2 and #3 sites respectively, characterized as late Archaic period lithic and ceramic scatters; and the Safety Harbor period habitation mounds--Mizzel Mounds A and


23 B, and the Cunningham Mound, a Safety Harbor period burial mound (Piper and Piper 1981 ) In the spring of 1994, Deming and Williams conducted phase two excavations at the five Hardee County mounds and the lithic scatter site previously investigated by Wood and Williams (1976). From the analysis of the artifact assemblages, Deming and Williams ( 1994) concluded that the CF Industries Property mounds dated from the late Weeden-lsland/ early Safety Harbor (Englewood and Pinellas phases) period and that they were probably short use domiciliary mounds. The majority of archaeological research of the interior region of the Central Gulf Coast has been conducted as part of cultural resource management projects. Undoubtedly as development in this region continues, many more archaeological sites will be encountered and evaluated. It is important to understand that the cultural picture of this region is still developing and hypotheses are constantly being reevaluated. It is only through the continuation of such research that we can more fully comprehend the prehistory of this region


24 CHAPTER 4 THE STUDY AREA AND ITS ENVIRONMENT In order to develop a more complete view of why archaeological sites exist at specific loci, it is essential that a whole range of environmental variables be considered; such as climate, geology, topography, relative elevation, soil composition, and proximity to water and lithic resources. The study area, the interior region of Florida's Central Gulf Coast, is characterized by considerable variable environmental microzonation, and an abundance of floral, faunal, and mineral resources (Wharton and Williams 1980:216). All of the aforementioned factors influence the availability of resources in the study area, which in turn, is reflected in the settlement and prehistoric land use patterns which exist today (Deming and Williams 1994). Environment The environment as it appears today is quite different from that which existed for Paleo-Indian populations in Florida. Paleobotanical and zoological studies have revealed an environmental mosaic of microhabitats which allowed for the existence of a wide array of species (Carbone 1983). Palynological studies have indicated the presence of mixed boreal and temperate, coniferous and deciduous forests within a broader matrix of grasses, sedges, and both hydrophytic and xerophytic species. Thus open and closed, wet and dry environments coexisted in close proximity with each other (Carbone 1983). Faunal analysis conducted at many Paleo-Indian sites in the southeast reveals


25 the presence of large numbers of Pleistocene megafauna such as: the woodland musk ox, mammoth, mastodon, giant moose, bison, horse, peccary, whitetail deer, caribou, elk, giant beaver, and ground sloth. Faunal remains from sites in the Vera Beach area, central and northern Florida, Alabama, and Georgia all support the scenario of an environmentally diverse mosaic confronting Paleo Indian populations (Carbone 1983). Deposits from Lake Annie, located in southern Highlands County, contain three pollen zones dating from 37,000 to 4,715 B.P. which correspond to other estimates of Quarternary vegetation for the state (Watts 1975). The Pleistocene seems to have been marked by decreased seasonality: milder winters, cooler summers, and increased dryness initiated by the presence of the Laurentide Ice Sheet. Along the Florida peninsula, this meant that rainfall patterns increased during the winter over the northern portion of the peninsula, while extended drought conditions persisted in the southern portion (Carbone 1983). Thus, the flora in north Florida consisted of open pine forest with some populations of broad leaved trees. The central and lower portion of the state was characterized by xeric scrub species along active sand dunes, with a coastal savannah strip projected for the Gulf Coast (Carbone 1983; Watts and Stuiver 1980). Pollen counts from Zone 3 (37,000 to 13,010 B.P) in Lake Annie indicate that rosemary scrub on dunes were the dominant plant cover and that pines were rare or absent in the region. Some oaks co-existed with prairie-like vegetation, all of which suggests a much drier climate than today (Watts 1975). By 14,000 B.P. the glacial climate had begun to ameliorate. The mixed conifer-hardwood forests expanded northward but the southeastern oak-hickory pine forests remained relatively stable. The Florida peninsula however, was losing land mass as a result of sea level rise and open xeric habitats were


26 considerably reduced (Carbone 1983; Delcourt and Delcourt 1981 ). Northern Florida around 14,000 B.P. was dominated by oak, hickory and hackberry, and with increased precipitation, beech became a major component of these forest stands (Carbone 1983). Watts and Stuiver (1980) interpret the vegetation in northern Florida during this time as consisting of dry oak hickory stands with local prairie in a warm, dry climate. Between 14,000 and 10,000 B.P. forests from northern Georgia to northern Florida gradually became more dense and pine and oak began to colonize previously unforested areas, which Carbone (1983) believes indicates a geologically brief return to drier conditions. Between 10,000 and 5,000 years B.P., oaks and pines were still the dominant vegetation in northern Florida, but the scrub vegetation in the peninsula was gradually replaced with oak savannah and a small strip of prairie persisting along the Gulf coast (Carbone 1983). Dry dune species are rare in the pollen counts from Zone 2 ( 13,01 0 +I-165 B. P.) at Lake Annie. These species are replaced by oaks and grasses with some pine. As with the previous period, there is no modern analog for this vegetation in Florida (Watts 1975). By the mid-Holocene (5,000 years B.P.) the climate again changed to favor more open vegetation. The southern pine forest replaced oak savannah in the Florida peninsula, extensive swamps and marshes developed along the coasts, and a subtropical hardwood forest established itself along Florida's southernmost tip (Delcourt and Delcourt 1981 ). Pollen Zone 1 (4,715 +/-95) from Lake Annie is characterized by large amounts of pine pollen and small amounts of oak pollen. These data suggest forests dominated by long-leaf pine in conjunction with cypress swamps and bayheads (Watts 1975). Sheelar Lake in northern Florida experienced an increase in oak and herb species between 9,500 and 7,200 B.P. and a decrease in pine. This area was characterized by a


mixture of oak stands and intervening prairie. By the late Holocene, pine had regained its dominance and modern conditions, including the development of cypress swamps, were established (Carbone 1983). 27 In terms of cultural adaptation, coarse grained (homogeneous) environments require greater specialization. In contrast, fine grained (heterogeneous) environments require more generalized adaptation. The late glacial and postglacial environments in the southeast were highly fine grained and demanded a rather generalized adaptive strategy for early populations which is in stark contrast to the specialized adaptive strategies present in the western half of the continent (Carbone 1983). When all of this is considered, a generalized hunting/foraging strategy with sites tied to mesic habitats can be inferred for the glacial and postglacial periods. However by A.D. 700 1500, the approximate temporal period of the sites included in this survey, the environment can be considered essentially modern. The Study Area Defined The definition of the study area is derived from Mitchem's (1989) description of the Interior Variant of Safety Harbor Mitchem originally described the Inland area as consisting of Polk, and Hardee counties, the eastern half of DeSoto County and possibly portions of Highlands County. As shown in Figure 1, the study area for this project will include all of Polk, Hardee and DeSoto counties and a small sliver of eastern Manatee county, which is not illustrated This small section of Manatee County was included because the sites in this area resembled other sites in the interior more closely than contemporaneous sites on the coast. Highlands County was not included for a number of reasons


28 Figure 1 : Location a I Map of General Study Area. c..-0" 0 150 300 Kilometers


which will be discussed later in the research design. Due to the size of the study area, a general environmental description is given. Each of the generalized micro-environments presented below lie within all four counties which make up the study area. Inhabitants in this interior region would have access to all of these resources. 29 Table 1 Micro-Environments and Resources within the Study Area (USDA, Soil Conservation Service 1983a, 1983b, 1984, 1989). Micro-Environment Uplands Flatwoods Freshwater Marsh/River Cypress Swamp Floral, Faunal, and Potable Water Resources Turkey oak, longleaf pine, slash pine, live oak saw palmetto, prickly pear, indiangrass. White tailed deer, squirrel, turkey, feral hogs, bobwhite quail, raccoon, rabbit, armadillo, opossum, skunk, bobcat, gray and red foxes. Slash pine, longleaf pine, saw palmetto, water oak, waxmyrtle, bay, cypress, maple and gum, with sawgrass and fern as ground cover. Same faunal inhabitants as in the uplands with the addition of wading birds in low areas. Potable water obtained from small streams passing through these areas. Water oak, cypress, cabbage palm, sweetgum, hickory, red maple, sawgrass, sedges and other water tolerant plants. Waterfowl such as cranes, herons, ducks, reptiles, amphibians, otter. Potable water year round. Sweetbay, sweetgum, cypress, various pines, hickory, water oak, magnolia, cedar, and understory of cattail, royal and cinnamon fern, and variety of aquatic plants. Numerous species of waterfowl, woodpecker, otter, alligator and other reptiles, amphibians. Potable water year round.


30 Lithic Resources Chert outcrops within the study area are relatively rare, as most deposits are deeply buried Ho-Never, lithic resources were available at Peace River J Hillsborough River and Brooksville outcrops. Peace River chert, although closer to the study area, was utilized much less, probably as a result of its poor flaking ability due to numerous phosphatic pellets imbedded within the rock. Quartz and limestone deposits dating from the Oligocene (38 to 22 5 million years before present) to the Holocene (10,000 years ago to the present) are present at and just beneath the surface in DeSoto County The Oligocene Series in the county consists of Suwannee Limestone which is creamy white to light yellowish gray in color It ranges in texture from packstone to grainstone and is variably crystallized (USDA Soil Conservat ion Service 1989) Suwannee Limestone contains many index fossils including a wide range of Miliolid foraminifera such as Rotalia and Elphidium which contribute to the grainstone or packstone rock fabric (Upchurch, Strom, and Nuckels 1982) In the northern half of DeSoto county Suwannee Limestone is located approximately one hundred meters below sea level and dips towards the south to a depth of two hundred and sixty meters below mean sea level. Due to the depth of this formation, surface chert outcrops of this type are not present in DeSoto County However the Peace River Formation Limestone, composed in part of Hawthorn Group cherts, can be found within the Peace R i ver Valley floodplain The Hawthorn Group is a Miocene Series group which was recently raised from Formation to Group status and includes sediments which were previously grouped with the Tampa, Hawthorn, and Bone Valley Formations In DeSoto County the Hawthorn Group is composed of the Arcadia Formation and the Peace River Formation. The Arcadia Formation is composed of the


31 Nocatee, Tampa and unnamed members. The Arcadia Formation lies over the Suwannee Limestone and generally consists of carbonate-rich quartz sand and thin clay beds (USDA, Soil Conservation Service 1989). The Peace River Formation lies at or near mean sea level throughout DeSoto county and is sixteen meters above sea level in the northwestern part of the county (USDA, Soil Conservation Service 1989). Cherts from this formation are medium dark gray to very light gray, grayish orange to very pale orange in color. Chert fabric is described as micromosaic quartz in replacement cherts, and pelletal opal cemented by later opal and chalcedony in opaline cherts Host rock fabric is either packstone or wackestone and quartz sand inclusions and silt grains are common Detrital phosphate pellets are diagnostic of the formation. Index fossils are relatively rare in Peace River Formation cherts, although replacement cherts occassionally contain Miliolid fragments (Upchurch, Strom, and Nuckels 1981 ) Implications of Environment on the Study As Table 1. indicates numerous micro-environments exist within the boundaries of the study area. Aboriginal populations had a diverse region from which to collect resources. Upland areas would have provided deer, raccoon, rabbit, armadillo, opossum skunk, bobcat, gray and red foxes and small rodents Wild herbaceous plants, grasses and legumes, and hardW'Ood trees which yield fruits, nuts, buds, bark, and foliage provided floral resources Rivers, swamp and wetland areas provided a wealth of wading birds, reptiles, otter, and freshwater fishes, as well as sedges, rushes, and reeds. Lithic resources were obtained from the Hillsborough River and Peace River outcrops Indeed many of


32 the sites in the sample are located along convergences of at least two different micro-environments, and their placements were surely intentional.


33 CHAPTER 5 THE RESEARCH DESIGN Methodology The data set of 28 sites was compiled by research of extant literature as well as the Florida Site File for sites meeting two specific criteria: 1) sites must be located in either Hardee, Polk, DeSoto or a defined eastern edge of Manatee counties; and 2) sites must exhibit an occupation which dates somewhere between A.D. 800 and A.D. 1725. The counties chosen as the spatial unit for the data set were selected on the basis of Mitchem's ( 1989) definition of the interior variant of the Safety Harbor culture. As mentioned previously, Mitchem (1989) defined the interior variant as consisting of at least Hardee and Polk counties, most of eastern DeSoto County and possibly portions of Highlands County. The chosen statistic for cluster analysis, (nearest neighbor), necessitates the use of as square a study area as possible. Reasons for this assertion are addressed on pages 37 to 45. Upon examination of site file data, it was determined that a small sliver of eastern Manatee County would be included because: 1) three sites located in this small section of Manatee County were good candidates for the data set, as they resembled sites of the interior more so than sites along the coast; and 2) the inclusion of these sites helped to satisfy the need for a square study area. These three sites as will be illustrated in a later section, occur in a triangular pattern which was also observed among five mounds in Hardee County.


34 Highlands County was excluded from the study area because it was determined that with the exception of the Goodnow Mound, the sites in this county are related to Central Florida cultures as opposed to Gulf Coast cultures. In addition, inclusion of Highlands County would have jeopardized the need for a square study area. The temporal period, A.D. 800 to 1725, was selected for a number of reasons. Principally, this time span is indicated by Mitchem (1989) as that which delimits the four phases of the Safety Harbor period: Englewood Phase (A.D. 800-1000); Pinellas Phase (A.D 1000-1500); Tatham Phase (A.D 1500-1567); and Bayview Phase (A.D 1567 -1725). Site selection was based on whether a site lie within the defined study area and fell into the chosen temporal period. Temporality was determined by diagnostic artifacts and site type. In some cases, the archaeologist interpreting the site defined its temporal period and these dates were used. As it became clear that sites in this interior region exhibited characteristics of cultures other than Safety Harbor it was realized that some degree of control was needed to assure statistical confidence. As necessitated by the nearest neighbor statistic, temporality was chosen as the control variable (Lorena Madrigal, personal communication 1994). Although this time period, A.D. 800-1725, seems quite long, it is specific to the culture chronology of the region and acts as an acceptable control variable in the cluster analysis The Universe Transversal Mercator coordinates for each site were plotted into the Geographic Information System, Maplnfo Once plotted, maps of the study area were produced. See Figure 2. Euclidean distances from each site to its nearest neighbor were obtained employing Map Info tools Utilizing the nearest neighbor method, the distances obtained via Maplnfo were subjected to


Fi gure 2 Map D e pict i ng Lo catio n and T y pe of Sit es wi thin S tudy Area. t:: z a u "' l: .. .. '--' :--i.. ,. it' b-15 * 8 H r 1 POLK COUNTY HARDEE COUNTY 8 Hr 1 8 co _. e 8-Hr-15 "C?l 8 -f=lr-16 0 1 0 co 8-H '-52 e 8-Hr-5 08-Hr-1 49 8-Hr-1 Sit e Types Loc at e d i n Study Area e Dom i c ili a ry Mounds e 8 -0e-3 -DESOTO COUN TY Sand Bur i a l Mounds T e mple Mounds .. A Village Sites 0 Site s w i t h Earthwo rks Lin ea r Sand Moun d s . *Campsit e '-* Sand Burial Site !"' Midden il .!:::!' ''. 35


36 cluster analysis. Nearest neighbor analysis, including the development of and considerations in using the statistic, are discussed later in this chapter. Results from the analysis are presented in Chapter 6. T'NO working hypotheses were presented in Chapter 1: Hypothesis 1) Sites in the interior Central Gulf Coast region dating to the period A.D. 800 to 1725, form settlement clusters on the landscape; and Hypothesis 2) Sites in the interior Central Gulf Coast region dating to the period A.D. 800 to 1725 cannot exclusively be attributed to the Safety Harbor culture. It will be illustrated that the many of the sites in this region do not reflect Safety Harbor type sites. The artifact assemblages and in some cases site type, resemble those of either the Glades or Kissimmee River Valley cultures. However, none of the sites were classified as "classic" Glades, Kissimmee River or Safety Harbor sites by the archaeologists researching them. Although Hypothesis 1 states the expected outcome of quantitative analysis, it became necessary for it to be modified and restated in compliance with the requirements of nearest neighbor analysis. Thus, the null hypothesis for the data set is that the points (sites) occur on the landscape in a totally random pattern. These terms are defined in the next section in a discussion of the nearest neighbor statistic. The second hypothesis can only be addressed qualitatively, the results of which are discussed in Chapter 6. The majority of sites in the data set were surveyed and excavated as a result of cultural resource management projects. The AMAX Chemical Corporation Pine Level Mine Project, phases one and two, and the CF Industries Property Projects, phases one and two, accounted for the excavations of many of the sites in the data set. A large number of sites dating from the Paleo-Indian to Weeden-lsland period lie within the study area. Because these sites do not exhibit occupations


37 between A D. 800-1725, they were not included in the data set, and thus do not appear on the map of the study area. Schiffer (1987:341) writes," There can be no magic number of discovered sites that will guarantee an accurate occupational history for a study area". Thus, archaeologists are perpetually working with what data are presently available. Although not as intensively surveyed as the coast, large expanses of land within the study area boundaries, i.e. the corporate property mentioned above, have been subjected to pedestrian survey; phase one, two, and three subsurface survey; and excavations. It is believed that a sufficient percentage of land has been investigated to mitigate any biases related to area and sample size which may impact this study and reasoning for this assertion will be discussed below. The Nearest Neighbor Statistic Nearest neighbor analysis was originally developed by ecologists to describe the distributions exhibited by populations of living organisms in their natural environments which occur in an infinite variety of patterns (Clark and Evans 1954:445). Once the widespread existence of non-randomly distributed populations was recognized, the search for a method to objectively assess contrasts between actual patterns and their theoretical random counterparts began (Clark and Evans 1954:445). A number of mathematical models attempted to address this question (Viktorov 1947; Cottam and Curtis 1949; Goodall 1952). Dice (1952) however, was the first to use distance between nearest neighbors to measure departure from randomness (Clark and Evans 1954:446). His method consisted of measuring the distance from a randomly selected individual to its nearest neighbor in each of six sextants of a defined circular "center of origin" (Dice 1952). This procedure proved to be laborious


38 and Clark and Evans (1954) later developed and published their own version of the formula. The nearest neighbor statistic gained widespread acceptance in ecology and gradually its usage spread to other disciplines. Geographers applied the technique to urban and rural settlement patterns and hailed it as a simple technique which encourages students to think spatially (Pinder and Witherick 1972:277). Archaeology, like geography, is concerned with objective description and analysis of the spatial corollaries of human behavior, and thus responded to the new technique with vigor (Pinder et al. 1979). Nearest neighbor analysis was used for a variety of purposes: to describe site distributions (Piog 1974; Washburn 197 4; Zubrow 1971 ), as a means of quantifying artifact distributions on living floors (Whallon 197 4 ), and as a means of evaluating burial patterns within an archaeological site (Stickel 1968; Pinder et al. 1979). In a study similar to this one, Upham (1982) utilized the technique in his study of regional organization of 14th century western pueblo societies. As previously discussed, Milanich (1978) suggested employing nearest neighbor analysis to quantitatively test his Cades Pond Occupation Nexus model. Explained in terms of two dimensional space, Clark and Evans (1954) proposed a measure of space which describes how individuals in a given population and area depart from a random distribution. Before proceeding, it is important to understand the concept of random distribution. Clark and Evans ( 1954) define random distribution as that which occurs in a given set of points and in a given area in which any and every point has the same chance of occurring on any sub-area, and that the placement of each point is not influenced by that of any other point. As defined here, randomness is a spatial concept and is completely dependent upon the spatial boundaries defined by the researcher (Clark and Evans 1954:446).


39 The procedure for nearest neighbor analysis is as follows. The distance, irrespective of direction, from each individual to its nearest neighbor is measured for a given population. Next, the mean distance to nearest neighbors which would be expected in a random distribution is calculated. The resulting ratio of observed mean distance to expected mean distance equals the measure of departure from randomness, the Rn statistic (Clark and Evans 1954:447). The Rn statistic (nearest neighbor), will fall within a range from zero to 2.1491. In a completely random distribution, R = 1 Ideally, values between 1 and 0 indicate a clustered distribution, and values ranging from 1 to 2.1491 denote a uniform distribution. The highest degree of departure from random for the Rn statistic is derived from the fact that under conditions of maximum spacing, each individual will be equidistant from six other individuals, thus forming a hexagonal distribution. In such a situation, the mean distance to the nearest neighbor is maximized and the resulting value is 2.1491 (Clark and Evans 1954:447). The nearest neighbor statistic can be calculated for two or more populations for comparative purposes, and can be evaluated for reliability via significance tests such as chi-square, the F distribution, and the Student's t distribution (Clark and Evans 1954:448). However, the application of the nearest neighbor statistic assumes that all relevant sites have been identified, which is unlikely for this lesser known interior region Therefore, before an accurate test of site clustering can be attempted, the researcher must have some level of confidence that the study are has been adequately surveyed both in method and in square kilometers. It is true that large tracts of land were surveyed within the study area boundaries: those on AMAX property, CF Industries property and Agrico property. However, in determining how


40 adequately an area has been covered, it is important to consider not only where sites are, but where they are not (Schiffer 1982). To begin addressing the notion of coverage, a map of sites (dating to the research designated temporal period) in the study area was created. Again it must be stressed that Paleo-Indian, Archaic, and Woodland period sites do exist in this region but are not included on the map as they do not meet the criteria for the study. Thus, many "blank" areas on the map actually contain sites from these periods. To eliminate as much "blank" space as possible, it was determined that two square regions, each equal to 70 square kilometers, would more accurately cover the study area and that the results would be compared to each other as opposed to the null hypothesis value of 1 (See Figure 3). At first glance, nearest neighbor analysis seems relatively simple and straightforward, but there are many pitfalls to the employment of this type of analysis which warrant some discussion As stated above, if the Rn value is equal to 1, then the observed and expected distances are identical and thus the pattern is the result of forces working at random. However, the forces which control the location of points are highly unlikely to operate at random, but are certainly capable of distorting either of the extreme conditions (regularity or clustering) to such an extent that the observed pattern could possibly be matched by a random distribution (Pinder and Witherick 1972:281-282). Therefore, an Rn value of 1 cannot be passed off simply as a product of random forces. What is more likely is that a variety of factors have influenced the position of each point (Pinder and Witherick 1972). This can be illustrated in archaeology, as site location does not occur in a vacuum and is entirely dependent upon any number of factors whose influences should be considered.


Figure 3 Quadrats Subjected to Nearest Neighbor Analysis . ... . ... ; .. .... :' A jt 8-Po-1 :3 POLK COUNTY HARDEE COUNTY r-18 'tf _, 8-Hr-15 "(?) -<-" 'tJ e 8-Hr-5 08-Hr-149 Site Types Located in Study Area e Domiciliary Mounds Sand Burial Mounds Temple Mounds Village Sites 0 Sites with Earthworks T Linear Sand Mounds *Campsite Sand Burial Site Midden 4 1


42 Thus, the research who produces a seemingly randomresult should critically inspect the data for possible cultural, environmental, socio-religious, etc., factors which might account for this distribution. Study area size is also an important consideration as each Rn value is controlled partly by this factor To avoid skewed interpretations, it is crucial to understand that the more extensive the area around a given distribution, the lower the Rn value will be. Figure 4, taken from Pinder and Witherick (1972:284), depicts the effect of area on the Rn value of a distribution of ten regularly spaced points in a one square kilometer area at the center of a 40 square kilometer region. It is clear that as areal reduction continues, the Rn value rises increasingly and does not arrive at regularity for the distribution until the area size is reduced to less than 4 square kilometers Clark and Evans (1954:450) suggested a solution to this problem, by stating that the study area should lie ... well within the total area covered by the population." Utilizing this strategy would eliminate the danger of incorporating a false element of clustering in the calculation of the Rn value (Pinder et al. 1979). It is important to note however, that in many cases, such as the present study, researchers need to widely demarcate their areas Thus, in these situations, a built in level of clustering must be accepted (Pinder et al. 1979) Another complication surrounding the nearest neighbor statistic is the range of random matching Pinder and Witherick (1972) found that the expected mean distances should never be regarded as exact definitions of randomness Instead they assert that this distance should be viewed as the center of a zone, whose width represents the range of average distances likely to be produced by a large number of random patterns utilizing the same number of points and the same area (Pinder and Witherick 1972:285) Clark and Evans (1954) included a formula for computing the standard error for the expected mean distances. This


43 2 1 r--C'O ::J E 1 0 .... 0 -----------------------LL c:: Q: 0 5 40 30 20 10 0 Area ( km2 ) Figure 4 The Effect of Area on Rn Values for a Constant Regular Point Pattern ( from Pinder et a1. 1979:4 34) formula incorporated area size and number of po i nts to define :he upper ar.:: lower limits of the range of randomness at the 05 confidence level. An Rn value which falls within the range of random matcn ing canner os interpreted as indicating the presence of signif i cant regularity or clustering due to the fact that if a large number of truly random patterns were analyzed 90% of their Rn values would lie within this range. However, outside the range increasingly strong significance can be assumed since the chances of Rn matching a random pattern fall from 5% to 2.5% and finally to 0.5% as the more stringent confidence levels are encountered (Pinder et ai. 1 979 : 438) This concept will be addressed in more detail in Chapter 6


44 A fourth consideration in computing the Rn value is that of the "boundary effect," touched on briefly by Clark and Evans (1954) and further addressed by Pinder et al. ( 1979) The boundary effect can be explained as follows. Consider a boundless space in which points are randomly distributed. If a finite area is superimposed upon this space, as is necessitated by nearest neighbor analysis, this boundary will ultimately sever those connections with nearest neighbors for the points lying on the perimeter of the boundary Thus, when nearest neighbor distances inside the boundary are calculated, some of the measurements for points near the periphery will be greater than if distances to real nearest neighbors outside the boundary had not been severed The result is that the mean observed distances will be higher than that predicted for a random po i nt pattern More importantly, this effect is greatest when the point population is small, since a higher proportion of severed distances can be expected (Pinder et al. 1979) Pinder et al. (1979) developed a corrected formula based on work by Ebdon (1976), to negate the boundary effect. Related to the boundary effect is study area size From the description above it is obvious that certain shapes chosen for study area will be more likely to sever connections between nearest neighbors. Linear rectangular and irregularly shaped study areas greatly increase this possibility The circle is the most efficient perimeter possible, but its boundary is only 11. 38% smaller than that of a square. Thus for archaeological applications it is suggested that mean predicted differences can only be considered accurate when study areas are square (Pinder et al. 1979) In archaeology, the nearest-neighbor statistic has been applied to measure the distributions of presumably contemporaneous phenomena. However, unlike geographers and ecologists, each of whom deal with modem phenomena, archaeologists lack firm temporal control over their data. Most


45 often artifacts corresponding to a specific temporal period which may span over several hundred years, are utilized as dating elements to secure contemporaneity. Additional problems of reuse and growth are introduced when site inventories are stratified for analysis based on site size or function (Pinder et al. 1979:442). Thus, it must be realized that precise description of distribution of points alone does not guarantee accurate characterizations of synchronic settlement patterns (Pinder et al. 1979:442). This is an important consideration in the present study, as all sites in the data set have been tentatively dated via their diagnostic artifact assemblages and because there is evidence of reuse at many of the sites in the sample. All of these characteristics of nearest neighbor analysis will be considered in the interpretation of findings. Nearest neighbor analysis is a powerful tool for addressing spatial patterning of archaeological phenomena. Ultimately however, it must be remembered that this method produces a relatively simple characterization of spatial patterning and should serve only as an initial step in interpreting and describing the spatial patterning of such phenomena (Pinder et al. 1979:443). The Data Set Defined The data set is composed of twenty-eight sites dating from or containing components dating from A.D. 800 to A.D. 1725. To reiterate, these dates represent the time span for the late-prehistoric and protohistoric period in the cultural chronology for the Central Gulf Coast region as defined by Willey (1949) and corresponding to the Safety Harbor culture as defined by Mitchem (1989). Every recorded site within the study area, which has been dated from A.D. 800 to the contact period, A.D. 1725, is included in the sample in order to increase statistical confidence. Many of the sites included in the data set also contain


46 evidence for earlier occupations and thus may have earlier inception dates. See Table 2 for an inventory of these sites. It is important to stress that although there are Archaic period sites in this interior region, only those sites which contain evidence for late occupation, A.D. 800-1725, were included in this sample. Each site is defined below. Polk County The Lake Marion I Site (8-Po-2). This sand burial mound is located in Township 28 South, Range 28 East, Section 5, (USGS Dundee Quad}, in the eastern portion of the county near the western shore of Lake Marion, Polk County, Florida. The mound, which measures 1.8 m high, had been dug into by many individuals by the time it was recorded by Goggin in 1947. The site lies approximately 600 m west of the lake in a forest of longleaf pine and xerophytic oaks The site form indicates that numerous burials were removed from the mound and that a gold pendant was recovered at some point. However, this pendant may possibly have come from the Lake Marion II Site, (8-Po-11 }, another burial mound lying along the shores of the lake. Goggin was shown a collection from this sand burial mound which consisted of two silver objects, at least one silver ceremonial tablet, a copper object, globular light blue glass beads, a large chevron bead, some large seed beads, and a Florida cut crystal bead (Mitchem 1989: 138) The silver tablet, recovered with a single burial, is described as measuring 65 mm by 37 mm and decorated with half moon, tear drop and linear incisions (Allerton, Luer, and Carr 1984). The sheet copper object is described as a rectangular pendant with embossed dots along the edges and four semicircular embossed shapes along


Table 2 Site Inventory of the Data Set. EUROPEAN SITE NAME DATE SITE TYPE COUNTY ARTIFACTS Bostwick Mound 5,000 B.C.-A.D. 1600 temple mound w/asso village Hardee no Airstrip Village Site 4a 5,000 B C.-A.D. 1500 village site Manatee no Airstrip Village Site 4b 5,000 B C -A.D. 1500 v i llage site Manatee no Keen Mound Complex 1 000 B C -A D 1600 m ound complex and earthwork Hardee no Mizell Mound "A" 1,000 B C -A.D 1500 sand domiciliary m o un d DeSoto n o Mizell Mound "B" 1,000 B C.-A D 1500 sand domiciliary mound DeSoto no Mizell Mound "C" 1,000 B C.-A.D 1500 linear sand domiciliary mound DeSoto n o Sugarbowl Mound 1,000 B.C -A.D. 1500 linear sand mound Manatee no Cunningham Mound 1,000 B.C.-A.D. 1500 sand burial/domiciliary mound DeSoto no Raulerson Mound 500 B C.-A.D 1500 sand mound Polk no Haines City River Site 500 B .C.A .D. 1500 m i dden Polk n o Orchard-Fenceline Site 500 B C -A.D. 1500 campsite H ardee no 8-Ma-65 100 B.C.-A.D. 1700 2 sand burial mounds Sarasota no Brandy Branch Village Site 100 B.C -A.D 1500 village site DeSoto no Carlton Ranch No. 1 Site A.D. 200-1500 mound Hardee no Frostproof Mound A .D. 500-1567 sand burial mound Polk yes Cowboy Mound A.D. 800-1500 sand domicil i ary mound Hardee no Ona Road Mound A.D. 800-1500 sand domiciliary mound Hardee n o Breton Mound A.D. 800-1500 sand domiciliary mound Hardee n o Philip Mound A.D 800-1600 sand mound w/ramp Polk yes Little Mound A.D. 800-1600 sand domiciliary mound Hardee no Nalcrest Site A.D 800-1725 campsite Polk yes Welch Mound A.D. 900-1500 sand domiciliary mound Hardee no Davis Burial Mound A.D 1400 sand burial mound Hardee no Pine Level Site A.D. 1450-1700 san d burial mound DeSoto n o Pine Level 2 Site A.D. 1500 san d burial mound DeSoto yes Singletary Site A.D. 1500-1567 sand burial site Polk yes Lake Marion I A.D 1500-1567 sand burial mound Polk yes


48 its central axis (Allerton, Luer and Carr 1984:32). If this site is to be classified as Safety Harbor, it falls into the Tatham Phase, A.D. 1500-1567. The Frostproof Mound (8-Po-7). Recorded by Chance in 1977, the Frostproof Mound is located in Township 31 South, Range 28 East, Section 28, (USGS Babson Park Quad), Polk County, Florida. The site lies 200 m from Reedy Lake in a pine flatwoods environment. Greatly disturbed, the Frostproof Mound is estimated to have originally measured 30.5 m in diameter and at least 1.2 m high. Local people had apparently unearthed extended burials, small glass beads, and a triangular silver pendant (Mitchem 1989: 139). The collection from the Florida Museum of Natural History includes: 13 St. Johns Plain ceramic sherds (2 rims), 8 sand tempered plain ceramic sherds (2 rims), 4 Belle Glade Plain ceramic sherds, 2 Dunns Creek Red ceramic sherds, and a marine shell fragment (Mitchem 1989:139). Dunns Creek Red pottery is most commonly associated with early Weeden-lsland related sites, but may possibly occur later (Mitchem 1989:139; Goggin 1948:7-8). The Florida Site File Form indicates a few other artifacts were recovered from the mound: 1 patinated blue glass bead, fragments of mammal bone, and Biscayne Plain, Red, and Cord marked pottery. Totals for the latter are not included on the site file form. The Dunns Creek Red sherds and marine shell fragment indicate a coastal connection, but the Biscayne pottery seems to indicate Glades influences. The Singletary Site (8-Po-13). The Singletary Site, which Mitchem (1989:137) believes may represent a post-contact Safety Harbor site, is located in Township 31 South, Range 28 East, Section 3, (USGS Homeland Quad), Polk County, Florida. The site lies in hardwood forest within the Peace River Drainage. The Florida Site File Form indicates that an Olive Jar was plowed up from the site in


. 1915 and that human remains were recovered at some point. The site is described as a mission but there is no evidence to support this (Mitchem 1989:139). 49 The Nalcrest Site (8-Po-15) The Nalcrest Site has recently been renamed the Indian Mounds, Lake Weohyakapka Site This site is one of many along Lake Weohyakapka in southeastern Polk County. The site is located within Township 30 South, Range 29 East, Sections 29 and 32 (USGS Lake Weohyakapka Quad). The site lies in a sand pine and scrub forest. This multicomponent site yielded numerous Pinellas type projectile points, several unidentified lead objects, two possible beads, and several flat, perforated, check stamped lead discs, which may have been suspended (Mitchem 1989 : 140). Sparse, non-diagnostic pottery was recovered from the site but was not listed as to type on the Florida Site File Form. The Pinellas type projectile points may indicate a Safety Harbor association The Raulerson Mound (8-Po-123). Also known as the District Site No. 1, the Raulerson Mound is located in Township 25 South, Range 24 East, Section 8 (USGS Rock Ridge Quad), Polk County, Florida The site lies in an oak hammock adjacent to a river flood plain containing swamp forest vegetation. The mound was partially excavated by a group of amateur archaeologists who unearthed a number of lithic and ceramic artifacts dating from the Paleo-Indian to late Weeden Island/ early Safety Harbor periods (Mitchem 1989: 140) Measurements for the mound were found to be 60 m East-West by 26 m North South by 1.65 min height. A number of Pinellas projectile points as well as ceramics ranging from early fiber tempered wares to St. Johns plain and Weeden Island incised types were recovered (Mitchem 1989) Numbers of


50 artifacts are not given on the site file form. The Raulerson Mound probably contains no later than an Englewood Phase occupation dating from A.D. 800 to 1000. The Philip Mound (8-Po-446) Also in the vicinity of Lake Marion, the Philip Mound is located in Township 27 South, Range 28 East, Section 33 (USGS Dundee Quad), in Polk County Florida. The site consists of a large circular mound measuring approximately four feet high with an attached ramp extending eastward off the northern portion of the mound towards a borrow pit directly east of the mound and then back westward to within 1 5 m of the mound. A large and diverse assemblage of both aboriginal and European artifacts \Nere recovered from the site Over sixteen thousand glass beads including some possible Nueva Cadiz beads (1500-1550), at least 12 silver and copper beads, and five Florida cut crystal beads were recovered. A large number of metal objects, such as a silver thimble, iron scissors, an iron knife blade, an iron eyed axe, an iron celt, and several miscellaneous metal objects complete the European artifact assemblage Aboriginal lithics encountered include : Pinellas points, a drill, Archaic points, and chipped and polished stone A small proportion of incised ceramics were recovered: one Sarasota, one Pinellas, and one Englewood Incised; one Pinellas Plain, one sand tempered plain, and two Gordon's Pass Incised or variant; as well as 32 Biscayne Check Stamped sherds Undecorated utilitarian wares were represented by 253 Belle Glade Plain sherds and 37 Biscayne Plain sherds. Vessel sizes approximate 10-30 em in diameter and vessel shapes include simple flattened globular and collared globular bowls. Although the incised ceramic assemblage indicates classic Safety Harbor, the largest portion of the ceramic assemblage is represented by types identified in the Glades region. Their overwhelming occurrence seems to


51 indicate either a great deal of influence from that region or that the Philip Mound was occupied by people from the Glades region. In addition to the artifact assemblage described thus far, four Busycon shell dippers, all with kill holes, were recovered from the 6-12" level. Three of these Busycon dippers were inverted and one covered sherds from an Englewood Incised pot. Ceremonial use of Busycon cups in this manner is very typical of Safety Harbor. A pair of scissors and a copper coin bead were also found in association with this feature, and their presence indicates a post A.D. 1500 occupation. Disturbed burials were encountered, but only small fragments of bone remained (Benson 1967). Benson (1967) believed the Philip Mound to be similar to the Goodnow Mound located just south of Philip in Highlands County. However, the Goodnow Mound lacked Biscayne Check Stamped and the Pinellas and Englewood components which the Philip Mound exhibits Due to the presence of Englewood ceramics and Spanish period artifacts, Benson (1967) dated the Philip Mound from the Englewood Phase to Bayview or Tatham Phase. However, it is important to note that the Englewood ceramics could have been curated items carried with the individuals occupying this site. In 197 4, Karklins located and excavated a pottery cache near the western edge of the mound which consisted of Belle Glade Plain, St. Johns Plain, St. Johns Check Stamped, and Pinellas Incised sherds and vessels, all of which had been broken and/or perforated (1974). A number of other ceramic types were recovered from the remaining portions of the mound and include: Pasco Complicated Stamped, Ocklawaha Incised, Papys Bayou Punctated, Indian Pass Incised, and three miniature vessels of Belle Glade Plain, St. Johns Plain, and Sarasota Incised (Karklins 1974:2). Karklins (1974) also recovered a perforated


52 iron fragment, 35 glass beads, two Florida cut crystal beads (A.D. 1550-1600), a lead bead, several shell beads and pendants, a coral bead, and red ochre. While Mitchem (1989) places the Philip Mound in the Safety Harbor period with a possible late Weeden Island component, the majority of the ceramic types recovered from the mound indicate some sort of interaction with other cultures; i.e. Belle Glade Plain ceramics as evidence for contact with the Glades region. The Philip Mound represents an extremely important component in attempting to understand the prehistory of the interior region The Haines City River Site (8-Po-1036) This midden, located in Township 27 South, Range 28 East, Section 20, (USGS Dundee Quad), Polk County, Florida, is also near Lake Marion. Lying on Lake Marion Creek, survey at this site yielded St. Johns Plain, Belle Glade Plain and sand tempered plain ceramics. Very little information exists on the Florida Site File form, but sand tempered plain ceramics have a rather long temporal span of 500 B C to A.D 1500. Their presence in conjunction with an absence of European artifacts indicates a pre contact occupation Hardee County The Davis Burial Mound (8-Hr-1). The Davis Burial Mound, which was completely destroyed by bulldozing activity in 1976, was located in Township 35 South, Range 27 East, Section 5 (USGS Sweetwater Quad), Hardee County, Florida. This sand mound lie in a pine flatwoods_environment 200 m north of Charlie Creek. The site was located near the Peace River. Originally 13 m in diameter and 1 3 m high, approximately three fifths of the mound was removed by bulldozer activity by the time Bullen recorded the site in 1953. A borrow pit


53 lie just to the northwest of the mound, and it was attempts to fill this pit that led to the destruction of the mound. Although a large portion of the mound had been destroyed, Bullen was still able to gather stratigraphic information. The mound consisted of a light brown sand overlain by brown sand, a dark homogeneous layer containing burials and finally capped with a layer of light tan sand which also contained burials. While the remains of 12 individuals were recovered as a result of Bullen's excavations, he estimated that the mound likely contained as many as 20 individuals (1954). Charcoal and red ochre were found in association with many of the burials, and in very few cases, a ceramic sherd. Shell beads, a total of eight Belle Glade Plain ceramic sherds, and the tip of a projectile point were also recovered. Bullen (1954) places the Davis Mound temporally at A.D. 1400, and believes the Davis Mound, like the Philip Mound, to be similar to the Goodnow Mound, located 15 miles to the southeast in Highlands County. Carlton Ranch No. 1 Site 8-Hr-5. The Carlton Ranch Property was excavated by Milanich and Martinez in 1975. The property covers a large area consisting of Township 34 South, Range 23 East, Sections 22, 23, 24, 25, 26, 27, 28, and 36, and Range 24 East, Sections 15, 16, 17, 18, 19, 20, 21, 22, 23, 26, 27, 28, 29, 30, 31, 32, and 33, and Township 35 South, Range 24 East, Sections 4 and 5. The Carlton Ranch site is located in former oak hammock on the western edge of a large swamp in west-central Hardee County, Florida. The mound is rectilinear to oval in shape and measures 54 by 17 m. Artifacts recovered from the mound include an unlisted amount of lithic debitage and a tools, fourteen sand tempered plain ceramic sherds, and two laminated and contorted paste sherds. Milanich and Martinez (1975) believed the later to be identical to those from Ft. Center, which dates from approximately A.D. 200. The researchers also


. asserted that the Carlton Ranch No. 1 Site is similar to houses occupied by a northern extension of the Okeechobee Basin people, thus dating the site between A.D. 200-1500. 54 The Orchard-Fenceline Site (8-Hr-11 ). The Orchard-Fenceline Site straddles the property line of CF Industries and Agrico in Township 33 South, Range 24 East, Section 6, Hardee County, Florida. As of 1979, undisturbed portions of the site still existed on CF property, but portions of the site on Agrico property had been destroyed by orchard operat i on. The site lies in pine flatwoods 1 00 m west of a swamp and 850 m east of Payne Creek The Peace River is also nearby, but the closest sources of water are the Hickey Branch which flows one mile southwest of the site and a spring located 50 m east of the site. A number of surface and subsurface artifacts were recovered from the Orchard-Fenceline site Ten sand tempered plain ceramic sherds, one broken projectile point, six retouched chert flakes, and 34 lithic debitage flakes were recovered from the surface of the site The subsurface assemblage consisted of 11 sand tempered plain ceramic sherds 1 chert waste flake, and 2 coral waste flakes. Based on the artifact assemblage, Wood classified the site as Glades I, II, and Ill, A.D 500 B C -A.D. 1500 However, there are no diagnostic artifacts listed which would allow cultural affiliation with the Glade Region The Cowboy Mound (8-Hr-15). The Cowboy Mound is an aboriginal sand mound located in Township 33 South, Range 24 East, Section 17, on CF Industries property in Hardee County, Florida Wood and Williams (1976) described this mound as situated in a flat field of palmettos, scrub vegetation, and scattered slash pines The mound lies fifty meters west of a dense hardwood hammock and includes areas of standing water Wood and Williams


55 (1976) estimated the mound to measure 25 by 15 m and to be approximately 46 em high. At the time of Phase II excavations in early 1994, the Cowboy Mound was covered largely in grass with a small fringe of saw palmetto along the outer margins of the site (Deming and Williams 1994) Phase I investigations into the mound produced: 25 Glades Plain ceramic sherds, three Belle Glade Plain ceramic sherds, and 37 lithic waste flakes (Wood and Williams 1976) A Hernando type projectile point was recovered from a depth of 45 to 60 em and 82 % of the pottery was recovered from levels six and seven (Deming and Williams 1994) Wood and Williams (1976) estimated the mound to date bet\veen 500 B .C. and A.D 1000. Phase II excavations conducted in 1994 found the mound to be roughly circular in configuration and to measure 20 m north-south by 19 m east-west by 0 5 to 0.6 m in height. No apparent borrow areas for mound fill could be discerned in the immediate area However, it is possible that mound fill was obtained from the swampy area 50 m to the east. A large number of artifacts were recovered from Phase II investigations The ceramic assemblage consisted of : 452 sand tempered plain ceramic sherds, 72 Pinellas Plain ceramic sherds, 36 St. Johns Plain ceramic sherds, and 14 Belle Glade Plain ceramic sherds. Sixty eight tool forms were recovered and include: 28 bifaces and biface fragments {one Florida Archaic Stemmed, one Levy projectile point, three Florida Archaic Stemmed Diminished projectile points, one Clay/Lafayette type projectile point, three Duval projectile points 17 Pinellas projectile points, one Lafayette projectile point, one Florida Morrow Mountain type projectile point, and one Bradford/Taylor type projectile point), five biface preforms, 31 flake tools, a core, and two abraders (Deming and Williams 1994:44) In addition to tool forms, 902 lithic waste flakes, two


56 sandstone abraders, one cube of galena, and one eroded piece of marine shell were also recovered (Deming and Williams 1994). Lithic analysis of the assemblage at the Cowboy Mound indicates that both tool manufacture and maintenance were occurring. Analysis of the debitage indicates that lithic materials were brought to the site in the form of early stage preforms Cherts from the Hillsborough River and Peace River quarry clusters were almost equally represented (Deming and Williams 1994:59). The Cowboy Mound is believed to be a domiciliary mound dating from the Englewood and Pinellas phases of the Safety Harbor period, A.D 800-1500 (Deming and Williams 1994:59). Deming and Williams (1994) also believe there is evidence of Archaic occupation, suggested by the presence of the Florida Morrow Mountain type projectile point recovered one meter below surface, (0.6 m below mound base). Thus, like numerous other sites in the region it was probably utilized over many thousands of years. The_Welch MouodJ_S-Hr-16} This aboriginal sand mound lies within Township 33 South, Range 24 East, Section 34, on the edge of a hardwood hammock on CF Industries property in Hardee County. Wood and Williams (1976) estimated the mound to be approximately 25m in diameter and 0.5 to 0.6 meters in height. During Phase I testing, 19 ceramic sherds, including : 11 Glades Plain, two St. Johns Plain, and six Belle Glade Plain, two projectile point fragments, and four waste flakes were recovered. Based on these data, Wood and Williams (1976) estimated the mound to date to A.D. 500-1500. Phase II archaeological investigations revealed that the mound actually measured 23m north-south by 16m east-west by 0.6 meters in height. No borrow areas near the mound were apparent, but it is possible that mound fill


57 came from the nearby swamp. Evidence for this can be explained in the composition of the mound fill itself, which consisted of 0.6 meters of homogenous blackish-gray sand (Deming and Williams 1994) The Soil Survey of Hardee County (1984) describes the site as situated on Pomona fine sand a nearly level, poorly drained soil. A single control shovel test excavated to the north of the mound revealed that the natural stratigraphy at the site does not correspond to that described for Pomona fine sand (Deming and Williams 1994 ). The mound fill does not conform to either the description of Pomona fine sand, or the natural stratigraphy at the site Further, dark soil such as that present in the mound fill are characteristic of low, depressional areas which are underwater frequently enough to produce organically rich black soil. Upon removal from these areas, as in the case of use as mound fill the soil gradually leaches and becomes lighter in color (Robert Brinkmann, personal communication 1994). This would account for the blackish-gray color of the mound fill. However, investigation of low areas in the environs surrounding the mound were not undertaken during Phase II investigations Artifacts recovered during Phase II excavations include : 96 sand tempered plain ceramic sherds, 20 Belle Glade Plain ceramic sherds, 10 St. Johns Plain ceramic sherds, one Pinellas Plain ceramic sherd, 295 lithic waste flakes, seven lithic tool forms, two flake tools, one reworked Archaic stemmed projectile point, and four biface fragments (Deming and Williams 1994) Although, 100% of the tool forms were manufactured from Hillsborough River cherts, Peace River Quarry Cluster cherts accounted for 34 % of the lithic debitage. Temporally the ceramic assemblage (excluding the sand tempered plain sherds which date from 1000 B.C to A.D 1600) indicates late Weeden lsland related through the Englewood and Pinellas phases of the Safety Harbor period, circa A.D 800-1500 (Deming and Williams 1994:73).


58 Deming and Williams (1994:73-74), interpret the Welch Mound as a single stage construction, short-term, domiciliary mound. Later stage stone tool reduction, maintenance, and modification of previously manufactured tools, and cooking, food preparation, and storage are all indicated by the artifact assemblage at the Welch Mound. Deming and Williams (1994:74) conclude by placing the mound construction during the precolumbian phases (Englewuod and Pinellas) of the Safety Harbor period, circa A.D 9001500. Intrinsic to this temporal placement, is that the Archaic Stemmed projectile point recovered from the 0.3 to 0.4 meter level, was likely a curated item (Mitchem 1994; Deming and Williams 1994). The Ona Road Mound (8-Hr-17). The Ona Road Mound is an aboriginal sand mound located in Township 33 South, Range 24 East, Section 29, in an area of palmettos, scrub vegetation and slash pines. Wood and Williams (1976) described the mound as roughly oblong in shape, measuring 15 by 10 by 0.56 meters in height. Phase I testing of the Ona Road Mound yielded one Lafayette type projectile point and one lithic waste flake. Wood and Williams (1976) assigned the mound to the Glades I-III period, ca. A.D. 500-1500. However, they believed the presence of the Lafayette point found below the mound proper, could push the date back to the Late Preceramic Archaic, ca. 3000-500 B.C., indicating reuse of this site over time. During Phase II investigations the mound was surveyed and found to measure 19 m north-south by 17 m east-west by 0.5 m in height. A much greater number of artifacts were recovered from Phase II excavations than from Phase 1 testing. A total of 25 ceramic sherds, including: 18 sand tempered plain, two Belle Glade Plain and five Pinellas Plain, were recovered during I


59 excavations. The lithic assemblage from the site consisted of 10 biface and biface fragments (five Pinellas projectile points, and two lchetucknee type projectile points), five flake tools, and a probable sandstone abrader. Of the 15 tool forms recovered, 33% were of Hillsborough River chert, 33% were of Peace River chert, 27% were of coral, and 7% were of chert from an unidentified source. In addition to the tool forms, 117 lithic waste flakes were recovered from the mound. Ninety-one percent of the debitage were of chert and 69 % of these flakes were derived from the Peace River Quarry Cluster (Deming and Williams 1994). Pinellas and lchetucknee projectile points were dated to the Safety Harbor period, A.D. 800-1600. (Bullen 1975:6). The ceramic assemblage also indicates a late Weeden-lsland related through early (Englewood and Pinellas) Safety Harbor period, A.D. 800-1500 (Mitchem 1989; Deming and Williams 1994). Deming and Williams (1994) interpret the Ona Road Mound to be a single construction, short term, domiciliary mound. As with the Cowboy and Welch mounds, the absence of human remains, mortuary artifacts, and/or cultural features such as hearths, post molds or midden lenses, precludes the possibility of the Ona Road Mound serving a burial or long term habitation purpose (Deming and Williams 1994:87). The artifact assemblage indicates a somewhat limited range of site activities: later stage reduction and maintenance of stone tools, and general camp activities, i.e., cooking, food preparation and food storage. Deming and Williams (1994) believe the mound was constructed during late Weeden-lsland or early Safety Harbor periods A.D. 8001500 (Mitchem 1989). A pre-mound occupation is still indicated by the Lafayette projectile point recovered during Phase I testing (Wood and Williams 1976). However, 9 % of the ceramic assemblage is from the lower levels of excavation,


60 the possible pre-mound occupation is thought to date to the post-Archaic period, no earlier than 1000 to 500 B.C. (Deming and Williams 1994:88) The Breton Mound (8-Hr-18). This aboriginal sand mound is located in Township 33 South, Range 24 East, Section 30, in a field of dense palmetto and myrtle, and bordered on the east side by six live oaks Wood and Williams (1976) originally described the mound as 10 m in diameter and 0 5 meters in height. During Phase I investigations the mound yielded only one lithic waste flake. Wood and Williams (1976) dated the mound to A.D. 500-1500, (Glades IIII) based on its similarity to other mounds in the region. The Breton Mound was surveyed as a part of Phase II excavations and found to be roughly circular measuring 24 m in diameter by 0 5 m in height (Deming and Williams 1994). A slight depression to the west of the mound may represent the original borrow pit. The mound lies on what the Soil Survey of Hardee County ( 1984) describes as Pomona fine sand and a control shovel test excavated north of the mound confirms this description (Deming and Williams 1994). A total of 41 ceramic sherds, including: 32 sand tempered plain six Pinellas Plain, and one St. Johns Plain sherd(s), were recovered during Phase II excavations The lithic assemblage consists of: four bifaces and biface fragments (two Pinellas type projectile points and a Pinellas preform), one biface preform, eight flake tools, a broken small hammerstone which exhibits minor retouch, and 139 waste flakes (Deming and Williams 1994 : 93) Both the tool assemblage and the debitage consist of both Hillsborough River and Peace River cherts. Peace River chert comprises 50 % of the tool form assemblage but only 48 % of the debitage assemblage (Deming and Williams 1994:1 00)


61 The Breton Mound is interpreted as a single construction, short-term, domiciliary mound As with the Cowboy Ona Road, and Welch mounds, human remains, mortuary ceramics, and cultural features indicative of long term habitation are not present at this site The ceramic assemblage and the presence of Pinellas type projectile points indicate a late Weeden-lsland/early Safety Harbor period occupation A.D 800-1500 (Deming and Williams 1994) The Little Mound (8-Hr-19). The Little Mound is an aboriginal sand mound located in Township 33 South Range 24 East, Section 30. Wood and Williams (1976) described the mound as situated in a field of palmettos, wiregrass, and low scrub vegetation 150 m to the east of Shirttail Creek The mound was originally estimated to be 10 m in diameter and 0 5 to 0 6 m in height (Wood and Williams 1994 ) As a result of Phase I testing nine Glades Plain ceramic sherds and one lithic retouch flake were recovered. On the basis of these data, Wood and Williams (1976) estimated the mound to date to the Glades I-III periods A.D 500 1500. During Phase II investigations, Deming and Williams (1994) found the mound to measure 22m north south by 14m east-west by 0 4 meters in height. Possible borrow areas were noted to the east and west of the mound (Deming and Williams 1994 ). As a result of Phase II investigations, the Little Mound yielded: three sand tempered plain ceramic sherds; three biface fragments, including the basal half of a Hernando point; a biface preform; five flake tools; and 61 waste flakes. Peace River chert accounts for 64.6% of the debitage assemblage and the majority of the tool form assemblage (Deming and Williams 1994).


62 Although the sand tempered plain ceramics are not chronologically sensitive ( 1 000 B. C. to A D. 1600), the recovery of a Hernando projectile point suggests "Transitional" period occupation, (500 B.C. to A.D. 200). However, the Little Mound's similarity to the other mounds in the area would suggest otherwise. Further, Mitchem (1994) has noted numerous instances of curated artifacts, including biface types traditionally assi gned to earlier periods at Safety Harbor sites. Thus, the Little Mound is characterized as a single construction, short term, domiciliary mound in which evidence for burial or long term habitation was recovered. The lithic assemblage from the mound suggests that stone tool production did occur However, there is no evidence for the reduction of large cores. Deming and Williams ( 1994) tentatively date the Little Mound to A D. 800-1600. There are a few characteristics which all five CF Industries property mounds exhibit. All five mounds represent examples of domiciliary mounds. All five mounds are located on the edge of a marsh, but not too far distant from pine-palmetto flatwoods or hardwood hammock which would have provided aboriginal people occupying these mounds access to aquatic and dry land resources. Interestingly, site clustering is visually apparent. The Ona Road, Breton, and Little mounds form a rough triangle in which all three mounds are situated within 1.6 kilometers of each other. The centrally located Ona Road Mound is also a connection point for another triangle formed with the Cowboy and Welch mounds, all of which are within 4.8 kilometers of each other (Deming and Williams 1994: 112). A similar arrangement will be described for the Mizell mounds in DeSoto County. It is this visual site clustering which prompted the need for a cluster analysis using the nearest neighbor statistic and formed the basis for this thesis.


63 The Bostwick Mound (8-Hr-52). The Bostwick Mound site, which consists of a sand temple mound, an associated borrow pit, and a suspected off-mound residential midden area, covers 88 acres in Township 35 South, Range 27 East, Section 6 (USGS Sweetwater Quad), in Hardee County, Florida Recorded by General Land Office surveyors in 1859, the Bostwick Mound was measured to be 6.6 m high, 50 m long and 20 m wide. In 1980, the mound measured 5.6 m high, 120m long and 76.6 m wide (Henefield and Wharton 1984). The site is located between two converging branches of Charlie Creek at the northern edge of an extensive, forested upland region. The vegetation in this upland consists of longleaf pine, and xerophytic oak forests. A bayhead, which drains into Charlie Creek, lies to the north. North of this bayhead are pine uplands and to the east lie swamps and stands of cypress (Henefield and Wharton 1984). Clearly, the Bostwick site lies on the boundaries of several micro-environments. An interesting assemblage of ceramic and lithic artifacts were recovered and/or observed from the Bostwick site. Lithic artifacts included: one Newnan projectile point, one Sarasota projectile point, one Westo type projectile point, one Florida Archaic stemmed projectile point, and numerous lithic debitage flakes. The ceramic assemblage consisted of both plain and decorated wares. Sand tempered plain, Belle Glade plain, St. Johns plain, laminated-contortedpaste plain, Pasco or Perico plain, Norwood or semi-fiber tempered plain, grog tempered plain, and Orange plain sherds comprise the plainware ceramic assemblage. The laminated-contorted-paste plain sherds most likely represent the type Pinellas Plain which is associated with the Safety Harbor culture. The decorated ceramic assemblage consisted of: one Weeden Island Red, five Dunns Creek Red, two sand tempered plain, three sand tempered linear check


64 stamped, five St. Johns Linear Check Stamped, six St. Johns Simple Stamped, three St. Johns Check Stamped, three St. Johns Incised or scored, one Belle Glade Simple Stamped, three unidentified incised, three unidentified incised/punctated, and six unidentified punctated. According to Cordell (1992), the category sand tempered plain now subsumes Glades Plain and Glades Gritty ware, Weeden Island Plain and Residual Plain, as well as incised and punctated varieties of sand tempered plain. Thus, the latter three types in the aforementioned list can be included in this category. The artifact assemblage from the Bostwick Mound site suggests repeated use of the site since the Middle Archaic period. Further, the diversity of the ceramic assemblage signifies interaction between other regional cultures. Although, Wharton and Henefield (1984) dated the Bostwick site to 5,000 B.C. to A.D. 1600, ceramics recovered from the mound itself indicate a pre-contact occupation during the Englewood and Pinellas phases of the Safety Harbor period, (AD. 800-1500). At the time of its 1980 National Register of Historic Places nomination, the Bostwick site represented the only known temple mound existing in interior central Florida. However, the Keen Mound Complex, described below, may be another example. Keen Mound Complex (8-Hr-149). The Keen Mound Complex is located within Township 35 South, Range 23 East, Section 23, in Hardee County, Florida. Janus Research (1994:38) describes the site as consisting of a large, rectangular platform mound; a smaller, conical, burial mound surrounded by a horseshoe shaped earthwork; at least one associated borrow pit; a small low sand mound; and a rectangular shaped earthen enclosure which may also be prehistoric in origin. The area surrounding the mound complex yielded lithic evidence of earlier occupation. The site lies in an upland area on the west side


65 of Horse Creek and approximately 150 m northeast of Goose Pond Road. Most of the s1te is covered with oak, pine, and palmetto, but a small portion of the site lies in an open sandy area (Janus Research 1994). The primary soil type listed for the site is Pomona fine sand (Janus Research 1994 ). However, soil profiles encountered in the test pits were generally dissimilar to the published soil profile descriptions. ''The soil conditions and the natural vegetation observed indicates that the area is better drained than the description of the soil type 'NOuld indicate" (Janus Research 1994:38). This same situation occurs at four of the CF Industries property mounds in Hardee County and indicates either a problem with the soil survey or some sort of environmental change since the time of the survey The former is much more probable. Archaeological investigations of the Keen Mound Complex were limited to shovel tests in the burial mound, the horseshoe shaped embankment, and the large platform mound. The ceramic assemblage recovered from testing consists of three sand tempered plain sherds, one possible Belle Glade Plain, and three sand tempered plain sherds with burnished surfaces The lithic assemblage includes : two tools-a Newnan projectile point, a bipointed biface with steep edges, and 55 lithic debitage flakes, the majority of which are small thinning and shaping flakes resulting from tool manufacture and/or repair (Janus Research 1994:41 ). In addition, 12 fragments of human bone and two shell fragments were recovered from shovel test investigations The horseshoe shaped earth'NOrk at the Keen Mound Complex is considered to be very similar to structures in the Lake Okeechobee and Kissimmee River areas, as well as some other sites on the southwest Gulf Coast (Janus Research 1994 : 42) A temporal period is not offered for the Keen Mound Complex. However, the presence of highly burnished sand tempered plain


66 ceramic sherds, which may be from a Weeden Island series vessel; the Newnan projectile point (4000 B.C.); and the sites' similarity to others in regions to the south and east, could possibly date the site to the same temporal range as many sites in the data set, 1000 B.C. to A.D. 1600. Manatee County Sugarbowl Mound (8-Ma--64) Originally recorded by Montague Tallant, the Sugarbowl Mound is located in Township 37 South, Range 22 East, Section 29 (USGS Murdock NE), only 20 m west of Sugarbowl Road, on AMAX Property in Manatee County, Florida The site lies within extensive palmetto flatwoods adjacent to prairie swampland (Willis and Johnson 1985:40) This swamp area is wet through much of the year and it is likely that prior to the construction of drainage canals, these areas provided a more stable and extensive source of both water and aquatic flora and fauna (Willis and Johnson 1985). The Sugarbowl Mound is a linear shaped sand mound aligned 60 m north-south and 20 m east-west. It rises approximately one m above the surrounding terrain Permanent pasture and isolated stands of palmetto and cabbage palm cover the mound today. Excavation within the mound revealed artificial construction of the mound in two stages evidenced by white sand constructed on original ground surface overlain by a cap or secondary mound of gray sand (Willis and Johnson 1985). Two sand tempered plain sherds were recovered from the gray sand cap The remaining assemblage from the mound consists of lithic artifacts including: three utilized flakes, three scrapers, one expended core, and one Archaic point tip All lithic artifacts, with the exception of the Archaic projectile point tip, were recovered from the primary mound construction which also contained charcoal


67 features. The Archaic projectile point tip was recovered from the confluence of the primary mound base and the old ground surface (Willis and Johnson 1985) Willis thought the primary mound to be Archaic and that the two sand tempered plain ceramic sherds indicated a later use of the mound (Willis and Johnson 1985:50). It is important to include here that Willis and Johnson (1985) note the possible remains of a simple log tomb in the form of charred stains connected to each other in north-west, south-east alignment. However, no further information concerning this feature is included. The Sugarbowl Mound is thought to be a multi-occupation site which dates from 100 B.C. to A.D. 1500. Although, a rather broad temporal range, the inception date marks the beginning of mound burials in Florida and the termination date marks the end of manufacture of sand tempered plain ceramics. There are no diagnostic artifacts, aboriginal or European, to allow for firmer temporal control or indicate cultural affiliation. However, if the log tomb Willis and Johnson believed they detected was in fact present, then the site may be associated with the St. Johns culture which utilized log tombs circa A.D 100 (Milanich 1994). 8-Ma-65. The Florida Site File form for this site listed no name, and it was later discovered that the site lie not in Manatee County, but in Sarasota County and has been renamed 8-So-396. Although the site is not in Manatee County, it is included in the data set because it lies within the defined square boundary necessary for the chosen statistic. Site 8-So-396 consists of two low sand burial mounds located in Township 38 South, Range 22 East, Section 5, 100m south of the AMAX property boundary in Manatee County, Florida. This site was originally recorded by Montague Tall ant but is not located on AMAX property and therefore was not


68 tested by Willis and Johnson (1980,1985). Although definitive dates have not been secured for these mounds, the site is included in this study due to: its similarity to other sand mounds in the area, and its apparent association with 8Ma-181 a probable village site which is described below. Thus, a date of 1 00 B.C. to A.D. 1700 is hypothesized. As with the Sugarbowl Road Mound, 100 B.C marks the inception of mound burials in Florida. The date of A.D. 1700 corresponds to the closing date for the associated village site 8-Ma-181 (Willis and Johnson 1985). The Airstrip Village Sites (8-Ma-181) APLS #4a and #4b. Located in Township 37 South, Range 22 East, Sections 32, and 33, 8-Ma-181 consists of two subsurface clusterings of lithic and ceramic artifacts Willis and Johnson (1979) believed these two sites, which have been grouped under one site number although they are 140 m apart, to be the village site associated with Piper and Piper (1981) believe the two clusterings should be designated as two distinct sites as Willis and Johnson did not recover artifacts in the 140 m area between the artifact clusters. Willis and Johnson (1980, 1985) describe 8-Ma-181 as measuring 60 m east-west by 100m north-south. Site 8-Ma-181 surrounds the southern tip of a bayhead, which would have allowed immediate access to fresh water and aquatic flora and fauna. Artifacts recovered from the initial Phase I survey in 1980 by Willis and Johnson consist of thirteen sand tempered plain ceramic sherds, one utilized flake, four non utilized flakes, and one bifacially worked drill Based on the artifact assemblage, the site was estimated to have been occupied between 500 B.C. to A.D 1700 (Willis and Johnson 1985). Phase 11 investigations by Piper and Piper (1981) consisted of examining #4a only, as #4b lies off AMAX property Site #4a was found to measure 60


69 meters East-West by 85 meters North-South A sample of charred midden bone was collected, and a midden stain at a depth of approximately 12-35 cmbs., associated with faunal remains and aboriginal pottery was recorded Ceramics recovered from this feature were "charred and frequently heat potlatched" (Piper and Piper 1981 : 7) A large number of artifacts were recovered during Phase II investigations. A total of 563 ceramic sherds were recovered from #4a and include: 535 sand tempered plain sherds (27 rims), 28 St. Johns Plain sherds (four rims), and one Belle Glade Plain sherd Rim sherds made up 12 % of the total ceramic assemblage Vessel forms represented are flattened globular bowls, pots with slightly converged orifices, and pots with straight rims Over 200 + fragments of faunal remains were recovered from #4a. These data included : one saltwater bivalve fragment; and numerous bones representing small mammals, fish, snakes and other reptiles The site yielded 195 lithic artifacts consisting of: three shatter, 80 thinning flakes, 99 pressure flakes seven primary flakes, three secondary flakes one utilized flake two drills one Hernando type projectile point, one Morrow Mountain type projectile point, one Pinellas type projectile point, and one miscellaneous Archaic stemmed projectile point. The small flake sizes, lack of decortication flakes and low tool to flake ratio indicate that tool maintenance I rather than manufacture was occurring at this site The dominant occupation zone at #4a was found to lie between 15-60 cmbs This zone produced the Archaic stemmed point (5000-1000 B.C .), Hernando Point (500 B .C. to A.D 200), Pinellas Point (A.D 900-1700), and the pottery dating from 200 B .C. to A.D 700. On the basis of the charcoal stains, midden material, quantity and type of pottery and tool maintenance area, it was concluded that the #4a portion of 8-Ma-181 was a village site utilized


70 intermittently over time from 5000 B.C. to A.D. 1500. Site location, vessel forms, and the presence of saltwater bivalves and the Pinellas point suggest a coastal cultural affiliation including possibly Manasota and Safety Harbor. DeSoto County The Pine Level Site (Keen Mound Site) (8-De-2) The Pine Level Site, also known as the Keen Mound or Old Whidden Cemetery, is a sand burial mound located in Township 37 South, Range 23 East, Section 14, (USGS Limestone Quad), off AMAX property in DeSoto County, Florida. The land owner, Mr. Harvey Keen, granted permission to conduct archaeological investigations in 1980 (Willis and Johnson 1980) Although, Horse Creek lies approximately 700 m to the east, the mound is located approximately 300 m west of a bayhead projection of Horse Creek's mesidhydric vegetation band. This bayhead was likely an aboriginal water source. The mound measures 60 m north-south by 20 m east-west and rises 2m above the surrounding terrain (Willis and Johnson 1985) At the time of Phase I investigations in 1980, the surface of the mound was covered by a mix of oak and hickory with an understory of pasture grass. The surrounding terrain was utilized as permanent pasture to support livestock. Willis and Johnson (1980: 11-47) reported the mound to be relatively intact. However, the mound was disturbed once by bulldozing activities which, according to the land owner, exposed burials accompanied by hematite. The exposed burials were apparently reinterred and there was no further disturbance to the mound. The excavation profiles revealed that the mound was constructed in tvvo stages. The ground surface, which overlays fine white sand, was prepared to


71 receive the primary mound of yellow sand. A white-gray sand secondary mound was later placed on original ground surface and only part of the primary mound. The final construction phase consisted of capping both mounds with white sand. Although no human bone was encountered in the excavations, a number of mortuary type artifacts and their placements within the mound indicate a burial. This burial consists of at least four whole ceramic vessels which had been broken in place. Ten em below the cached vessels, a 1 2 m long segment of carbonized wood aligned east-west was revealed. This is believed to be the base of the burial and was carbon dated to A.D 1037-1228 (Willis and Johnson 1980:K71 ). In this same level, just to the southeast of the carbonized wood stain, lie a cache of five saltwater conchs The ceramic assemblage from the Pine Level site consists of: one aboriginal, sand tempered sherd (applique adorned with incised weeping eye); 18 zoned punctate sherds (three rims); 17 4 sand tempered plain sherds ( 12 rims); five burnished chalky ware sherds (three rims); two diagonal zoned, incised sherds; and two check stamped sherds There was a low frequency of lithic artifacts in the area excavated One whole Pinellas point, one utilized chert flake, and three non-utilized flakes were the only contributors to the assemblage. Also recovered were the remaining fragments of the five conch shells accompanying the burial (Willis and Johnson 1985). The artifact assemblage from the Pine Level site places it temporally between A.D. 1450 and 1700, the late Pinellas, Tatham, and early Bayview phases of the Safety Harbor period. Willis and Johnson (1985) believe the mound due to its date the ceramic vessels and their placement over the burial, I I to have been constructed and utilized by people from the Safety Harbor culture. Further, they believe the mound is a ceremonial religious focal point at which the inhabitants of scattered small villages or a group of such villages converged


72 periodically to bury their dead and perhaps perform other religious ceremonies (Willis and Johnson 1985 : 11-51). The presence of wood in association with bone within the mound may be indicative of log tomb burial. It is interesting to note however that log tomb burial has long been thought to be a practice of the St. Johns culture to the east. This could be one other indication of influences from surrounding regions. The Pine Level 2 S i te (8-De-3) The Pine Level2 Site is a sand burial mound located in Township 36 South, Range 24 East, Section 30 (USGS Limestone Quad), DeSoto County Florida Originally recorded by Goggin, the mound was plowed down almost level with the surrounding terrain. What remains of the mound lies in a pine flatwoods environment which borders a freshwater swamp. Horse Creek flows 800 m to the east. While very little information is offered on the Florida Site File Form it is known that human bone and numerous glass beads were recovered from the mound before its destruction. Dependent upon type, the glass beads would date the mound to at least A.D 1500 However analysis of diagnostic aboriginal artifacts should they exist, will be the only way to arrive at a more secure date. This limited data also hinders assigning a cultural affiliation to the site. Attempts at addressing this problem will be presented in Chapter 6 The Brandy Branch Village Site APLS #5 (8 De-4) The Brandy Branch Village Site is located on AMAX property in DeSoto County Florida in Township 37 South, Range 23 East, Section 11, on the 50-55 foot contour within an orange grove bordering a mesic hammock which grades into hydric vegetation along Brandy Branch creek 100m east of the site (Willis and Johnson 1985). The dimensions of the site were found to be approximately 140 m north-south by 60


73 m east-west. Artifacts recovered during Phase 1 survey include: two chalky sand tempered plain ceramic sherds (similar to St. Johns plain) located on the surface, nine sand tempered plain ceramic sherds, two non-utilized lithic flakes, one utilized lithic flake, and one worked conch columella (Willis and Johnson 1980). When the Phase I survey was concluded, the site was temporally placed between 100 B.C A.D 1500. Phase II investigations were conducted by Piper and Piper in 1981. The ceramic assemblage recovered during these excavations included: 43 sand tempered plain ceramic sherds, two of which exhibit incurvate rims and one of which was decorated with a diagonal zone in relief. A total of 341ithic flakes and portions of two projectile points, as well as a small amount of unidentified bone, two unidentified shell fragments, and three pieces of modem glass were also recovered (Piper and Piper 1981 ) Phase II excavations contributed further evidence for the 100 B .C. to A.D. 1500 temporal placement and the likelihood of some sort of contact with the coast due to the presence of the conch columella recovered during Phase I survey and the other two shell fragments recovered during the second phase of investigation Cultural affiliation, however, is unclear. The presence of St. Johns like sherds and the conch columella possibly indicate a west coast contact. The Mizell Mound "A" (8-De-31 ). The Mizell Mound "A" is located in Township 38 South, Range 23 East, Section 5, in permanent pasture, 80 m east of a 30 acre swamp This site and the Mizell Mound "B" and "C" sites are situated on AMAX property in western DeSoto County, Florida. The Mizell Mound "A" is an aboriginal sand mound which Willis and Johnson (1985:11-52) believe served as burial mound, temporary domiciliary


74 mound, a meeting place, or periodic hunting/gathering camp. However, no evidence to support its use as a burial mound was recovered from either Willis and Johnson's excavations (1980) or those of Piper and Piper (1981 ). The mound itself measures 40 m by 30 m by 1 m in height (Piper and Piper 1981 ). The mound is constructed primarily with yellow sand and at the time of both Phase I and II investigations the surface was covered with pasture grass and several depressions thought to be cattle wallows. Immediately west and east of the mound are two borrow areas from which the mound fill was acquired. Phase I subsurface testing yielded six sand tempered plain ceramic sherds, three utilized chert flakes, one non-utilized chert flake, and one thumbnail chert scraper (Willis and Johnson 1980) Due to the lack of diagnostic artifacts, the site was dated on the basis of the sand tempered ceramics which exhibit the broad time span of 100 B .C. to 1600 (Willis and Johnson 1980). Phase II investigations yielded a wider range of artifacts : 28 thin sand tempered plain ceramic sherds, 10 St. Johns Plain ceramic sherds, 232 lithic flakes, one hafted side scraper, one utilized end scraper three Archaic stemmed points, one utilized biface fragment, one thermally altered cracked core fragment, and one double winged bannerstone recovered sub-mound. The aboriginal ceramics dominated the upper levels of the excavation and occurred only within the mound The St. Johns pottery, which dates to 500 B C A.D 1565 (Milanich 1994), was underlain by the sand tempered plain variety (1000 B.C. -A.D. 1600) (Cordell1992). The lower levels were dominated by lithic artifacts, which occurred sub-mound, and indicate all stages of tool manufacture. The bannerstone is very interesting in that it is composed of a metamorphic rock found in the Piedmont Region of the Carolina's and thus indicates the existence


of a trade network. The Mizell Mound "A" site was apparently utilized on numerous occasions over a period from 1000 B.C. to A.D. 1600. 75 The Mizell Mound "B" (8-De-32). Also located on AMAX property in DeSoto County, Florida, and approximately one mile southeast of Mizell Mound "A" lies Mizell Mound "B". This mound is located in Township 38 South, Range 23 East, Section 9, in permanent pasture 200m west of a large swamp system. Based on Willis and Johnson (1980:83), the mound is roughly circular in shape and measures 25m north-south by 30m east-west by 0.62 min height. As with the Mizell Mound "A" this mound is constructed primarily of yellow sand and exhibits cattle wallows on its surface. Mound "B" is capped with a grey sand layer. Artifacts recovered in Phase I investigations include: 55 sand tempered plain ceramic sherds, one chert drill, one utilized chert flake, and three unidentified mammal bone fragments (Willis and Johnson 1980). The majority of artifacts were recovered from the grey sand mound cap (Willis and Johnson 1980). Like mound "A" Willis and Johnson (1980) initially believed mound "B" to serve both burial and domiciliary functions. However, no evidence was recovered to support the burial hypothesis. At the conclusion of Phase I investigations, Willis and Johnson (1980) dated the mound to A.D. 500 to 1600. However, sand tempered plain ceramics, as indicated above, have a temporal range of 1000 B.C. to A.D 1600 (Cordell1992). Piper and Piper (1981) conducted Phase II excavations and recovered: eight St. Johns Plain ceramic sherds and 108 sand tempered plain ceramic sherds, including one flat folded rim, one single line incised rim, and three incised rim sherds. At least seven of the sand tempered plain sherds exhibit the contorted paste typical of Pinellas Plain which has a temporal range of A.D 1 000 to 1500 (Mitchem 1989). Some of the sand tempered plain sherds


76 appeared to be '\veil made" (Piper and Piper 1981) and may be the remains of Weeden Island ceramics. In addition to the ceramic assemblage, 40 lithic waste flakes, one utilized flake, one utilized blade, and one Lafayette Point, were recovered from disturbed plow zone (Piper and Piper 1981 ). Although Lafayette type projectile points date to the late preceramic Archaic period (ca. 3000 -2000 B.C .), the location of this projectile point in disturbed context has eliminated its usage in dating the occupation of the mound (Willis and Johnson 1980). One sand tempered plain ceramic sherd and nine lithic flakes were recovered from sub-mound levels. Although the sub-mound occupation at mound "B" is not as substantial as that of mound "A", there does seem to be evidence that the area was visited more than once. The presence of Pinellas Plain type ceramics could date the mound to the Pinellas phase of the Safety Harbor period, A.D 10001500. The Mizell Mound "C" (8-De-33). Approximately 1 6 km to the west of mound "B" and one km to the southwest of mound "A," lies Mizell Mound "C. The mound is located in Township 38 South Range 23 East, Section 8, on AMAX property in DeSoto County, Florida. Dense palmetto flatwoods surround the mound up to one half of a mile distant. Swamp areas near mound "C" are located at further distances to the north south and west than are those within the environs of mounds "A" and "B." Mound "C" is described as linear in shape, aligned north-south and measuring approximately 70 m by 35 m by 80 em in height. The mound appears to have been constructed rapidly of medium-grey sand (Willis and Johnson 1980). Borrow areas, from which mound fill was acquired, are located on the northwest and northeast ends of the mound. Upon surface investigation of the mound, nine sand tempered plain ceramic sherds were recovered.


77 Subsurface Phase I investigations yielded 251 sand tempered plain ceramic sherds, one chalky ware plain ceramic sherd (similar to St. Johns plain), three non-utilized chert flakes, one utilized flake, and one flake drill. Although artifacts were scattered throughout the mound, ninety-eight percent of the ceramics were recovered from the upper levels of the excavations (Willis and Johnson 1985). Willis and Johnson (1985) define Mizell Mound "C" as an aboriginal sand domiciliary mound The temporal period listed in the report is A.D. 500-1600, however, sand tempered plain ceramics in Florida do occur as early as 1000 B.C (Cordell 1992) All three Mizell Mounds are located on higher ground adjacent to or near a large swamp. This location would have allowed access to both flatwoods and aquatic floral and faunal species. It is also interesting to note that the Mizell Mounds form a triangle in their placement on the landscape and all three are located within 2.4 kilometers of each other This type of patterning was also observed among the five domiciliary mounds located on CF Industries property in Hardee County If this patterning were intentional and occupation of the three can be considered coeval the Mizell cluster can be considered to date to the Pinellas I phase of the Safety Harbor period The Cunningham Mound (8-De-34). The Cunningham Mound is located on AMAX property in DeSoto County, Florida. The site lies in Township 37 South, Range 24 East, Section 7, at the edge of a mesic mixed hardwood band which separates a large pasture to the north from a small marshy pond to the south. Located approximately 900 m east of Horse Creek, the mound is roughly circular in shape and measures 50 m by 45 m by 1.4 m in height (Willis and Johnson


78 1980; 1985). No cultural material was recovered from the mound surface which is now dotted with low grasses and scattered palmettos. Three borrow areas lie to the south-southeast, northeast, and east of the mound proper. Initial investigations at the Cunningham Mound yielded: 17 sand tempered plain ceramic sherds, including one rim, one chalky ware plain ceramic sherd, three non utilized chert flakes, three unidentified bone fragments, and three conch shell fragments (Willis and Johnson 1980; 1985) Willis and Johnson ( 1980:67 ; 1985: 11-68) believe the mound may have served as a bur ial mound as well as a seasonal domiciliary purpose. Evidence for this supposition comes from Mr. George Cunningham the former property owner, who reported the mound had been dug into several times and that human skeletal material had been encountered However, Phase I archaeological investigations failed to produce any such ev i dence. The presence of conch shell in the mound, while indicating trade with or trips to the coast, may be indicative of a burial as was seen in the Keen Mound (8-De 2) located approximately three kilometers southwest of the Cunningham Mound (Willis and Johnson 1980 ; 1985) Phase II excavations of the mound produced a sparse addition to the known artifact assemblage. Ten lithic chert flakes 12 sand tempered plain ceramic sherds (including one rounded rim) and one St. Johns plain ceramic sherd were recovered from the mound Eleven fragments of possible human bone were encountered but the absence of soil stains, mortuary type ceramics, I or other grave goods, as well as the extensive bioturbation and acidic nature of the mound soil, indicate that the possibility of the mound functioning in a burial capacity is probably slim Willis and Johnson (1980, 1985) believe the ceramic assemblage excavated from the Cunningham mound suggests an occupation between 1000 B C and A.D 1500


79 CHAPTER 6 RESULTS AND INTERPRETATIONS Results of N earest N e1ghbor Analysis As was discussed 1n Chapter 5 it was determined that two equilateral quadrats measunng 4 900 square kilometers would increase statistical confidence by assunng more accurate coverage of the study area than would one large square wh1ch woul d have Incorporated unsurveyed lands Utilizing th1s method only one t1e to nearest neighbor was severed between quadrats: 8Po-13 1n the northern quadrat and 8-Hr -11 in the southern quadrat. Both quadrats were tested agamst the null hypothesis that sites occur in a random d1str1bution and the results are presented below. The northern quad rat cons1sts of seven sites lying w1thin Po lk County Defmed 1n Chapt er 5 the s1tes 1nclude the Lake Marion I Site ( 8-Po-2). a sand bur1al mound ; the Frostproof Mound ( 8-Po-7), also a sand burial mound ; the S1ngletary S1te (8-Po 13). a sand burial site : the Nalcrest site (8-Po-15) a short term camps1te : the Raulerson Mound (8-Po-123 ), a sand mound ; the Philip Mound (8-Po-446) a ramped sand mound and the Haines City River site (8-Po1 036), a midden (See Figure 5) The computed Rn value for the northern quadrat of 1.08 falls within the range of random matching To readdress this concept, the range of random matching is that range 'Nhich hovers around 1 0 and decreases inversely with sample size An Rn value 'Nhich falls within this range cannot be interpreted as indicating


Figure 5 Sites in Northern Quadrat. 8 Po-123 ... ""'\ \ \ I I \ 8P )( 0 ... ,_ .. ' '. x 8-Po-1 3 ..Po -15 ". f.' 8-Po7 POLK COUNTY Site Types Located in Study Area l e Domiciliary Mounds Sand Burial Mounds o Sites with Earthworks *Campsite )I( Sand Burial Site )I( Midden ..; :' ..:;.:_ :=-=:r. .... ..; 80


81 significant clustering or regularity because, if large numbers of truly random distributions were analyzed, 90 % of their Rn values 'N'Ould lie within this range. Thus, an Rn value of 1 08 for a sample size of 7 'NDuld seem to indicate a random distribution (See Figure 6) However it is very important to consider the fact that any random location pattern is the product of a complex interaction of a variety of forces. The combination of these forces has arranged the points in such a way as to match those on a randomly generated map To assume that the location of these sites is entirely random is to reject the idea that there are more subtle explanations (Pinder et al. 1979) In the case of the northern quadrat, it is apparent that all but one of the seven sites are located within close proximity to water Further, these sites lie at the intersection of at least t'NO different micro-environments These microenvironments, the uplands, pine flat'N'Oods, cypress swamps and freshwater marshes, were discussed in Chapter 4 and the reader is referred back to Table 1 for a description These locations would have provided access to a wide variety of both terrestrial and aquatic flora and fauna. A similar situation exists within the southern quadrat, with the exception of the results of nearest neighbor analysis. The southern quadrat consists of 21 sites in Hardee, DeSoto and Manatee counties. As Figure 7 indicates, site types represented within this quadrat include : eight domiciliary mounds, five sand burial mounds, one temple mound, one site with erected earth'NDrks, t'NO village sites, two linear sand mounds, and one short term campsite The Rn value for the southern quadrat differed considerably from that of the northern quadrat. An Rn value of 0.4 significant at the 99.5 % confidence interval was computed, indicating significant clustering As with the sites within the northern quadrat,


Figure 6 Significance Graph for Testing Rn Values (from Pinder et al. 1979:439) 82 Range of r andom matching


Figure 7 Sites in Southern Quadrat ; (. POLK CoUNTY HARDEE COUNTY 8-Hr-18 ... 8-Hr-15 ::J ("). I 8 8-t-tr-16 I C(f-<- 8-Hr-5 o 8-Hr-149 - 8-De-3 8-t-52 -_ r 8-Hr-1 ... DESOTO COUNTY .. y .. -.; . .. .; ; ,;. Site Types Located in Study Area e Domiciliary Mounds Sand Burial Mounds Temple Mound s A Villag e Sites T Linear Sand Mounds Campsite 83


84 sites within the southern quadrat are all located within close proximity to a water source and straddle the intersections of at least two of the previously mentioned micro-environments. We are reminded however, that environment and resource availability are not the only forces driving site location. Any number and combination of social, spiritual, mythological, economic, and functional factors, to name a few, contribute to where people choose to camp and/or live. The difference in sample size between the two quadrats may account for some of the differences between Rn values, as it was expected that some measure of clustering would be evident in the northern quadrat. Undoubtedly as more sites are discovered and perhaps more importantly, recorded in this interior region, distance between sites will change, prompting a shift in Rn values. It is important to remember that the nearest neighbor statistic represents an initial step from which cautious interpretations can be made (Pinder et al. 1979). A Question of Method At the inception of this research, it was believed that sites in the interior region represented a culture which was influenced by the surrounding Glades, Kissimmee River Valley and Safety Harbor cultures; that the region existed within a zone of overlap between at least three cultural spheres of influence. At least nineteen of the 28 sites represented Safety Harbor type sites, two represented Glades type sites, and two represented Okeechobee or Kissimmee River Valley type sites. The remaining five sites were not specifically attributable to any one culture, as no diagnostic artifacts were present in either their lithic or ceramic assemblages


85 There has long been an attempt in Florida archaeology to draw boundary lines marking the extent of a particular culture. Most often, ceramic typologies have been the tools utilized to demarcate these cultural boundaries. It is obvious, by looking at the site distribution in the interior that the drawing of such lines is not possible (Figure 8). Although the majority of the sites within the sample are considered Safety Harbor type sites, sites that can be affiliated with other cultures share this same geographic area. For example, it appears that the northernmost of the five site clusters is composed of two Safety Harbor type sites and one possible Glades type site. Futhermore, none of the sites within the sample can be considered "classic" Safety Harbor, Glades or Kissimmee River Valley sites Even though many sites exhibit a large percentage of traits which could link them to a specific culture, there is still evidence for influence from surrounding regions within their artifact assemblages. For instance, the Bostwick site, the temple mound and associated village site described in Chapter 5, undoubtedly contains some Safety Harbor ceramics within its assemblage However, the ceramic assemblage at Bostwick is also composed of types representative of the Glades and St. Johns regions. The Keen Mound Complex, 8-Hr-149, also described in Chapter 5, seems to be consistent in structure with sites in the Lake Okeechobee and Kissimmee River areas as well as some sites along the southwest coast. The Keen Mound Complex lies betvveen two Safety Harbor type sites, another Okeechobee type site and a site of unknown cultural affiliation. At this point, it should be apparent that the second phase of Milanich's occupational nexus model is not applicable here. Milanich was able to make statements about the clusters within his study area because each site within each cluster represented the same culture. Obviously, this is not the case in the


Figure 8. Distribution of Sites with Respect to Cultural Affiliation Unk V' ... \ .. . .. -\ ,. ... . . u .J Itt , ...... r. . .. . . "' . ___, ... ...... \ . \ . ;'.. 'lo ). jC S .H. PoLK C ouNTY HARDEE COUNTY a Ok/Kis S.H. a o Ok/Kis -DESOTO COUNTY ... . .. . }' :.. .. I' : I. H. ades Key G l ades G l ad e s cultu r e S H Safety Harbor Ok/Kis Okeechobee / Kissimmee Unk Unknown cultural affiliation 86


87 interior region of the Central Gulf Coast. Indeed, within the study area there are five clusters, (a cluster is defined as three or more sites spaced in close proximity to each other over the landscape), and many of the sites within these clusters have been identified as Safety Harbor sites. Questions, however still remain as to the cultural affiliation of at least four sites associated with specific clusters One of the major problems in identifying cultural affiliation is that in most cases, ceramic types are utilized as the only measure of relatedness and/or interaction between groups Archaeologists typically assume that the similarity of two artifact assemblages provides an accurate measure of the degree of interaction between the producers of those assemblages Johnson (1977) asserts that while this may be the case style should not be used to signal exclusive group membership and maintain social boundaries between interacting populations since it is possible that stylistic sim ilarit y of artifact types between groups may be low while the actual group interaction is relatively high Arnold (1978 1985) delves even deeper into ceramic variab ili ty how i t relates to culture and environment and the role it has played in archaeology Archaeologists, he writes, tend to believe that ceramics reflect the culture of a people to such an extent that the main forces of cultural change are expressed in their ceramics. Further stylistic attributes and technological similarities are considered to indicate trade exchange, migration and/or conquest and differences in these traits are indicative of the lack of cultural contact or diffusion (Arnold 1985) Traditionally archaeologists have argued that ceramics are purely a product of culture with no environmental relationship, save the use of different tempering agents. However, temper type can be viewed as reflecting a 'concept in the mind of the potter' more than the availability of resources (Arnold 1978 : 37) As mentioned previously, stylistic variability has generally been the


88 criterion used to demarcate cultural boundaries in the Central Gulf Coast region. Arnold stresses that as ceramics are a part of the techno-economic system of culture they are closely tied to environment and that unfortunately, ... some archaeologists focus on the articulation of ceramics with spatial variables such as social structure or cultural boundaries, with little, if any, recognition of the possible importance of environ mental factors" (1978:41 ). Arnold cites examples from the work of MacNeish (1970), Deetz (1965) and Longacre (1970). Finally, Arnold (1985) believes that ceramic chronologies do little more than group ceramic assemblages in space and time; a category which has no connection to those which the manufacturing culture placed on them. In order to address the environmental link, Arnold (1978) suggests studying; both the presence and absence of pottery making, the numbers of vessel shapes, the kinds of decorative techniques, paste, and materials and resources. Johnson (1977) too, stresses the study of vessel shape as a more accurate cultural marker. Interpretations for the Region So what was occurring between the period of A.D. 800 to 1725 in the interior Central Gulf Coast region? This research seems to indicate that the area was extensively utilized by people whose cultural remains indicate influence from a number of what are considered regional cultures. Most of these interpretations have been based on ceramic typologies. Long term occupation is evidenced by the fact that most sites in the sample also contain components dating back to the pre-ceramic Archaic and is also evidenced in the site types themselves. Temple mounds and earthwork construction are indicative of more than just short term site usage.


89 It is possible that some of the domiciliary mounds were constructed for seasonal use, as they usually occur in areas where seasonal flooding is common. However, to my knowledge there have been no palynological studies to test seasonality at any of the sites within the sample. Burial mound location does not seem to subscribe to any detectable pattern. They are located both in association with site clusters and at some distance from any one site cluster However as more sites are discovered in this region, a detectable pattern may emerge It may be the case that specific kin groups utilized certain burial mounds. Future research questions addressing the social and spiritual aspects of the peoples inhabiting this region will begin to address these hypotheses. Sites are located within close if not immediate proximity to water and are situated at the intersection of at least two different ecological zones Thus, site location allows for the exploitation of a broad range of terrestrial as well as aquatic flora and fauna Lithic resources are also obtainable but at some distance from outcrops located in the Peace and Hillsborough River drainages and in areas to the north and west along the Ocala Uplift. Pope (1985) constructed a seasonality model defining a number of environments which sites could be associated with. The model was based on extensive research of the available flora and fauna from each microenvironment through each season for the Alafia River drainage basin. Pope (1985) asserted that sites associated with flatwood hammocks were likely to represent summer or autumn settlements and that sites associated with river swamps, marshes, ponds and streams were likely to represent spring or winter settlements. To reiterate, sites within the study area lie at the intersection of two or more microenvironments like those presented in Pope's (1985) study. These data as well as evidence of long term habitation in site types, i.e. the construction of large earthen structures, indicate a year round habitation. That is not to say

PAGE 100

90 however, that people were completely sedentary within this interior region. On the contrary, it is more likely that while some structures were permanent many people moved around the landscape in response to any number of social, ideological, economic and/or resource procurement needs. The results of this study have not ruled out the hypothesis that Safety Harbor peoples were the primary inhabitants of this region. Indeed, if each site in the data set which cannot be attributed to Safety Harbor were excluded, the results of nearest neighbor analysis in each quadrat would be quite different. Therefore, it may be possible that this interior region is a variant of Safety Harbor, as proposed by Mitchem (1989). The presence of sites attributed to other regional cultures could then be explained in politico-social, and/or economic terms. Once again, this is a question for future research and can only be addressed by continued investigation and inquiry. It is unrealistic to believe that inhabitants of this region did not share economic and perhaps social links. One direction of future research might be to examine trade relations between peoples in this area and how they are related to those of. the surrounding regions. The incorporation of methods of ceramic analysis utilizing vessel shape and ethnographic analogy, such as those outlined by Arnold (1985}, may be a good place to start. From what we know so far, no grand cultural scheme can be constructed for the interior region of the Central Gulf coast. However, it is apparent that cultural remains representing influences from more than one regional culture are present in the archaeological record. It is clear, as the data base exists today, that an interior variant of the Safety Harbor culture cannot truly be defined. Moreover, the archaeology of this region should serve as an impetus for rethinking how boundaries of archaeological cultures are addressed. Only further investigation in this region can begin to bring the picture into focus.

PAGE 101

91 There may exist unrecorded burial and temple mounds in the region, as well as a host of village and camp sites, the addition of which will undoubtedly alter the results of the research presented here. It is hoped that this study will ignite a resurgence in the use of settlement pattern studies in Florida archaeology, as well as new methods of addressing questions about the past.

PAGE 102

REFERENCES CITED ACI/Janus Research 1994 Phase Ill Mitigative Salvage Excavation at the Muck Pond East Site (8-Hi-515) Located Within the Proposed SR 574 Improvement Area Hillsborough County, Florida. Archaeological Consultants Inc. Sarasota. Janus Research, St. Petersburg. Allerton, D., G. M. Luer and R S Carr 1984 Ceremonial Tablets and Related Objects from Florida. The Florida Anthropologist 37(1 ):5-54 Arnold, D. E. 92 1978 Ceramic Variability Environment and Culture History among the Pokom in the Valley of Guatemala, in The Spatial Organization of Culture. I. Hodder ed University of Pittsburgh Press:PA. 1985 Ceramic Theory and Cultural Process. Cambridge University Press : Cambridge Batcho, D .G. 1978 Archaeological and Historical Resources Within the Uttle Payne Creek Mining Tract Polk and Hardee Counties, Florida, 1978. Miscellaneous Project Report Series No. 11. Department of Social Sciences, Florida State Museum, Gainesville Benson, C .A. 1967 The Philip Mound: A Historic Site The Florida Anthropologist 20: 118-132 Brinkmann, R 1995 Personal Communication. Brinton, D G. 1859 Florida Peninsula, Its uterary History, Indian Tribes, and Antiquities. J Sabin, Philadelphia. 1867 Artificial Shell Deposits of the United States Smithsonian Institution Annual Report for 1866. pp. 356 358. Washington D C

PAGE 103

Bullen, R.P. 1952 Eleven Archaeological Sites in Hillsborough County, Florida. Florida Geological Survey, Report of Investigations 8. Tallahassee. 1954 The Davis Mound, Hardee County, Florida The Florida Anthropologist 7 : 97-102. Bullen, R.P., W. Askew, L.M. Feder, and R. McDonnell 1978 The Canton Street Site, St. Petersburg, Florida. Florida Anthropological Society Publications 9. Gainesville. Bushnell, F. 1962 The Maximo Point Site 1962. The Florida Anthropologist 15:89-101. 1966 A Preliminary Excavation of the Narvaez Midden, St. Petersburg, Florida. The Florida Anthropologist 19:115-124. Calkins, W W 1878 Notes of Persona/Investigations Among the Shell Heaps of Florida. Proceedings of the Davenport Academy of Natural Sciences 1876-1878 Volume 2. Davenport, Iowa. Carbone, Victor A. 1983 Late Quarternary Environments in Florida and the Southeast. The Florida Anthropologist 36(1-2):3-17 Chance, M.A. 1982 Phase II Investigations at Wetherington Island: A Uthic Procurement Site in Hillsborough County, Florida. Interstate 75 Highway Phase II Archaeological Reports 3. Bureau of Historic Sites and Properties Clark, P.J and F .C. Evans 93 1954 Distance to Nearest Neighbor as a Measure of Spatial Relationships in Populations. Ecology 45(4):445-452. Clausen, C.J., A.D. Cohen, C Emiliani, J.A. Holman, and J.J. Stipp 1979 Little Salt Spring, Florida: A Unique Underwater Site. Science 203:609-614. Cockrell, W.A. 1970 Glades 1 and Pre-Glades Settlement and Subsistence Patterns on Marco Island (Collier County, Florida). Master's thesis, Florida State University Tallahassee. Cordell, A. 1992 Technological Investigation of Pottery Variability in Southwest Florida In Culture and Environment in the Domain of the Cal usa. Edited by William H. Marquardt. Institute of Archaeology and Paleoenvironmental Studies, University of Florida, Gainesville. pp. 105-189.

PAGE 104

94 Cottam, G and J.T. Curtis 1949 A Method for Making Rapid Surveys of Woodlands by Means of Pairs of Randomly Selected Trees Ecology 30:101-104. Cushing, F .H. 1897 Exploration of Ancient Key-Dweller Remains on the Gulf Coast of Florida. Proceedings of the American Philosophical Society 25(153):329-448. Daniel, I.R. Jr 1982 Test Excavations at the Deerstand Site (8-Hi-483A) in Hillsborough County, Florida. Interstate 75 Highway Phase II Archaeological Reports 2. Bureau of Historic Sites and Properties. Daniel I.R. Jr., and J Wisenbaker 1987 Harney Flats : A Florida Paleo-Indian Site. Baywood, Farmingdale. Deagan, K.A. 1987 Artifacts of the Spanish Colonies of Florida and the Caribbean, 1500-1800. Vol. 1. Smithsonian Institution Press, Washington D.C. Deetz, J. 1965 The Dynamics of Stylistic Change in Arikara Ceramics. Illinois Studies in Anthropology, No. 4. Urbana, Ill. Delcourt, P.A. and H.R. Delcourt 1981 Vegetation Maps for Eastern North America: 40,000 yr. B.P. to the Present. In Geobotany II. Edited by R.C. Romans. Plenum Publishing Corp., New York. Deming, J. 1976 An Archaeological and Historical Survey of the Lake Thonotosassa By-Pass Canal Right-of-Way in Hillsborough County, Florida Submitted to Southwest Florida Water Management District, Survey No. 139, p. 2. Deming, J and J .R. Williams 1994 Phase II Test Excavation at Six Prehistoric Archaeological Sites Located on the CF Industries Property, Hardee County, Florida. Archaeological Consultants Inc., Sarasota, University of South Florida, Tampa. Dice, L.R. 1952 Measure of the Spacing Between Individuals Within a Populataion. Contributions from the Laboratory of Vertebrate Biology, University of Michigan 55:1-23.

PAGE 105

Dickel, D.N. and G.H Doran 1989 Severe Neural Tube Defect Syndrome from the Early Archaic of Florida. American Journal of Physical Anthropology 80:325-344. Dunbar, J .S. 1983 A Model for the Predictability of Clovis/Suwannee Paleo Indian Site Clusters in Florida A Revival of W T Neill's Oasis Hypothesis. Paper presented at the 35th Annual Meeting of the Florida Anthropological Society, Tallahassee. 95 1991 Resource Orientation of Clovis and Suwannee Age Paleoindian Sites in Florida. In Clovis: Origins and Adaptations, edited by R. Bonnichsen and K. Tummier, pp. 185-213. Center for the First Americans, Oregon State University, Corvallis. Dunbar, J S S. D. Webb, and M. Faught 1991 Inundated Prehistoric Sites in Apalachee Bay, Florida, and the Search for the Clovis Shoreline. In Paleoshorelines and Prehistory: An Investigation of Method, edited by R. Bonnichsen, pp. 473-497. Center for the Study of the First Americans, University of Maine, Orono Ebdon, D. 1976 On the Underestimation Inherent in the Commonly Used Formulae. Area 8:165-169. Getis, A. 1964 Temporal Land-use Pattern Analysis with the Use of Nearest Neighbor and Quadrat Methods. Annals of the Association of American Geographers 54:391-399 Goggin, John M. 1948 Some Pottery Types from Central Florida. Bulletin No. 1. Gainesville Anthropological Association, Gainesville. Goodall, D.W. 1952 Quantitative Aspects of Plant Distribution. Biology Review 27:194-245 Griffin, J.W. and R.P. Bullen 1950 The Safety Harbor Site, Pinellas County, Florida. Florida Anthropological Society Publications 2. Gainesville . Griffin, J.W. and H. G. Smith 1954 The Cotton Site: an Archaeological Site of Early Ceramic Times in Volusia County, Florida. Florida State University Studies 16. Tallahassee

PAGE 106

96 Hoffecker, J F., W .R. Powers, and T. Goebel 1993 The Colonization of Beringia and the Peopling of the New World Science 259:46-52. Hudson, C 1976 The Southeastern Indians University of Tennessee Press, Knoxville Janus Research/Piper Archaeology 1994 Cultural Resource Assessment SuNey for Goose Pond Road (County Road 663 A) Bridge Replacement over Horse Creek, Hardee County, Florida. Janus Research St. Petersburg Johnson, G. A. 1977 Aspects of Regional Analysis in Archaeology, in Annual Review of Anthropology. B. J Siegel, A. R. Beals, and S. A. Tyler, eds Annual Reviews Inc.: California Karklins K. 197 4 Additional Notes on the Philip Mound, Polk County, Florida The Florida Anthropologist 27: 1-8 Longacre, W.A. 1970 Reconstructing Prehistoric Pueblo Societies. Albuquerque Luer G.M and M.M Almy 1981 Temple Mounds of the Tampa Bay Area. The Florida Anthropologist 34 : 127-155 1982 A Definition of the Manasota Culture. The Florida Anthropologist 35 : 34-58 MacNeish, R.S., F.A. Peterson, and K.V. Flannery 1970 The Prehistory of the Tehuacan Valley: Vol. 3, Ceramics Austin Madrigal, L. 1995 Personal Communication Martin, J D 1976 An Archaeological and Historical SuNey of the Borden Big Four Mine Properties in Southwestern Hillsborough County, Florida. Archaeological Report No. 3 Department of Anthropology, University of South Florida, Tampa. McMichael, A.E. 1982 A Cultural Resource Assessment of Horr's Island, Collier County, Florida Master's thesis, Florida Atlantic University, Boca Raton.

PAGE 107

Milanich, J. T. 1978 Two Cades Pond Sites in North-central Florida: The Occupational Nexus as a Model of Settlement. The Florida Anthropologist 31:151-173. 1994 Archaeology of Precolumbian Florida. University Press of Florida, Gainesville. Milanich, J.T. and C.H. Fairbanks 1980 Florida Archaeology. Academic Press, New York. Milanich, J .T. and C.A. Martinez 97 1975 Archaeological and Historical Resources of the Carlton Ranch Property, Hardee County, Florida. Miscellaneous Project Report Series No 1. Department of Social Sciences, Florida State Museum, Gainesville. Mitchem, J. M. 1989 Redefining Safety Harbor: Late Prehistoric/ Protohistoric Archaeology in West Peninsular Florida. Ph .D. dissertation, University of Florida, Gainesville. University Microfilms, Ann Arbor. Moore, C.B. 1900 Certain Antiquities of the Florida West Coast. Journal of the Academy of Natural Sciences of Philadelphia 11:349-394. 1903 Certain Aboriginal Mounds of the Florida Central West Coast. Part II. Journal of the Academy of Natural Sciences of Philadelphia 12:361-492. 1905 Miscellaneous Investigations in Florida. Journal of the Academy of Natural Sciences of Philadelphia 13:298-325. Padgett, T. J. 197 4 An Archaeological and Historical Survey of the WR. Grace Property in Hillsborough and Manatee Counties, Florida. Miscellaneous Report Series No. 17. Florida Division of Archives, History and Records Management, Bureau of Historic Sites and Properties, Tallahassee. 1976 Hinderland Exploitation in the Central Gulf Coast-Manatee Region during the Safety Harbor Period. The Florida Anthropologist 29:39-48. Pinder, D.A. and M.E. Witherick 1972 The Principles, Practice and Pitfalls of Nearest-Neighbor Analysis. Geography 57:277-288. Pinder, D.A., I. Shimada, and D. Gregory 1979 The Nearest-Neighbor Statistic: Archaeological Application and New Developments. American Antiquity 44(3):430445 .

PAGE 108

Piper, H.M and J.G. Piper 1981 Archaeological Testing and Evaluation of Seven Sites Located on AMAX Properties, Manatee and DeSoto Counties, Florida. Piper Archaeological Research, St. Petersburg. Submitted to Environmental Science and Engineering, Inc., Gainesville. 1982 Archaeological Excavations at the Quad Block Site, 8-Hi-998, Located at the Site of the Old Fort Brooke Municipal Parf
PAGE 109

Solis de Meras, G. 1964 Pedro Menedez de Aviles, Adelantado, Governor and Captain General of Florida. Memorial. Translated by Jeannette Thurber Connor. Reprinted. University of Florida Press, Gainesville. Originally published 1923, Florida State Historical Society, Deland. Stearns, R.E.C 1870 Rambles in Florida. American Naturalist 3:349-360. 1872 Remarks on Mounds and Shell Heaps of Tampa Bay, Florida. Proceedings of the California Academy of Science Series 1 4 : 214-215. Stickel, E.G. 1968 Status Differentiation at the Rincon Site In University of California Archaeological Survey, Annual Report, 1968, pp 209-261 Stirling, M W. 1935 Smithsonian Archeological Projects Conducted under the Federal Emergency Relief Administration, 1933-1934. Annual Report of the Smithsonian Institution 1934 : 371-400 United States Department of Agriculture, Soil Conservation Service 1983a Soil Survey of Manatee County, Florida. Department of Agriculture, Washington D C 1983b Soil Survey of Polk County Florida Department of Agriculture, Washington D .C. 1984 Soil Survey of Hardee County, Florida Department of Agriculture, Washington D.C. 1989 Soil Survey of DeSoto County, Florida Department of Agriculture Washington D.C Upchurch S B., R.N. Strom and M .G. Nuckels 99 1982 Methods of Provenance Determination of Florida Cherts Department of Geology, University of South Florida, Tampa Upham, S. 1982 Polities and Power. Academic Press, New York Walker, S.T. 1880a Report on the Shell Heaps of Tampa Bay, Florida. Annual Report of the Smithsonian Institution for 1879. pp. 413-422 Washington D.C 1880b Preliminary Explorations Among the Indian Mounds in Southern Florida. Annual Report of the Smithsonian Institution for 1879. pp 854-868. Washington D C 1885 Mounds and Shell Heaps on the West Coast of Florida. Annual Report of the Smithsonian Institution for 1883. pp. 854-868. Washington D C

PAGE 110

Washburn, O.K. 197 4 Nearest Neighbor Analysis of Pueblo 1-111 Settlement Patterns Along the Rio Puerco of the East, New Mexico. American Antiquity 39:315-335 Watts, W.A. 1975 A Late Quarternary Record of Vegetation from Lake Annie, South Central Florida. Geology 3:344-346. Watts, W.A and M. Stuiver 1980 Late Wisconsin Climate of Northern Florida and the Origin of the Species Rich Deciduous Forest. Science 21"0:325-327. Webb, S.D., J .T. Milanich, R. Alexan, and J.S Dunbar 1984 A Bison Antiquus Kill Site, Wacissa River, Jefferson County, Florida. American Antiquity 49:384-392. Whallon, R. 197 4 Spatial Analysis of Occupation Floors II: the Application of Nearest Neighbor Analysis. American Antiquity 39:16-34 Wharton, B.R. 1977 Proposed Research Design for Safety Harbor Sociopolitical Organization. Ms. on file, Department of Anthropology, University of South Florida, Tampa. Wharton, B., and J.R. Williams 1980 An Appraisal of Hardee County Archaeology: Hinterland or Heart land? Florida Scientist 43:215-220. Wharton, B., G. Ballo, and M. Hope 1981 The Republic Groves Site, Hardee County, Florida. The Florida Anthropologist 34:59-80. Widmer, R.E. 197 4 A Survey and Assessment of Archaeological Resource on Marco Island, Collier County, Florida. Miscellaneous Project Report Series 19. Bureau of Historic Sites and Properties. Willey, G.R. 1949 Archaeology of the Florida Gulf Coast Smithsonian Miscellaneous Collections 113. Smithsonian Institution, Washington D.C. 100

PAGE 111

Willis, R.F. and RE. Johnson 1980 AMAX Pine Level Survey: An Archaeological and Historical Survey of Properties in Manatee and DeSoto Counties, Florida. Department of Anthropology, University of Florida, Gainesville. Submitted to Environmental Science and Engineering, Inc., Gainesville. 1985 Draft Environmental Impact Statement U.S. Environmental Protection Agency Region IV--Supplementallnformation Document Section 11. Archaeological/ Historical Resources. Submitted to AMAX Chemical Corporation Pine Level Mine Manatee and DeSoto Counties, Florida. Prepared by Environmental Service and Engineering, Inc., Gainesville. Willis, R.F., and J.T. Milanich 1977 Archaeological and Historical Resources of the Farmland Industries, Inc., Properties, Hardee County, Florida. Miscellaneous Project Report Series No. 10. Florida State Museum, Gainesville. Wood, LN. and J.R Williams 1976 An Archaeological and Historical Survey of the CF Mining Corporation Property in Northwestern Hardee County, Florida. University of South Florida, Tampa. Submitted to CF Mining Corporation, Wauchula, Florida. Wyman, J. 101 1870 Explorations in Florida. Third Annual Report to the Trustees, Peabody Museum, 8-9. Harvard University, Cambridge. Zubrow, E. 1971 Carrying Capacity and Dynamic Equilibrium in the Prehistoric Southwest. American Antiquity 36:127-138.


Download Options

No images are available for this item.
Cite this item close


Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.


Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.


Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.


Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.