MODELING VARIABILITY IN PRE-COLUMBIAN WOODLAND HABITATION

AND SOCIAL ORGANIZATION: THE BRICKHILL BLUFF SITE,

CUMBERLAND ISLAND,

by

Stephen Andrew Wise

A Thesis Submitted to the Faculty of

Dorothy F. Schmidt College of Arts and Letters

In Partial Fulfillment of the Requirements for the Degree of

Master of Arts

Florida Atlantic University

Boca Raton, FL

August 2017

Copyright by Stephen Andrew Wise 2017

ii iii ACKNOWLEDGEMENTS

I would first like to thank my advisor Dr. Arlene Fradkin for her continued encouragement and patience. This work would not have been possible without her persistence and guidance. I am especially indebted to Dr. Mike Russo, Archaeologist for the National Park Service, for his reassurance and help. I also would like to thank

Valentina Martinez who taught me how to excavate features and the proper way to hold a trowel. Her comments and suggestions were invaluable to the organization and construction of this thesis. Thank you for being supportive of my goals.

I am grateful to everyone whom I have had the pleasure to work with on this and other related projects. Whether in the field or in the lab, the members of my excavation team and the many supportive voices at Atlantic University and the National Park

Service have contributed to my growth as an archeologist and as a person. I would especially like to thank Dr. David Morgan, Dr. Michael Harris, Richard Vernon, Hank

Kratt, Charlie Sproul, and John Cornelison who have guided my academics and career.

They have served as mentors and shown me how to be a professional archaeologist and friend. I am thankful for our shared love of the past, people and culture. They have taught me more than I could ever give them credit for here.

My family has been an important part of this project. My mother encouraged me and provided financial and emotional support. Her love and guidance are the reason I am free to pursue my interest in archaeology. My father’s good humor and his reminders not to take life too seriously shaped who I am and helped ground me during the most difficult

iv times. And finally, I would like to thank my wife, Brittany. Her support and reassurance guided this research, my presentations, and finally this thesis. She is the person I turn to during times of joy and times of sorrow. Thank you for reading and editing countless drafts and for your love. I look forward to telling our twins how their mother read each chapter between bouts of morning sickness. Finally, I would like to acknowledge our unborn twins. If you ever read this, remember that you can accomplish those things you think you cannot do.

v ABSTRACT

Author: Stephen Andrew Wise

Title: Modeling Variability in Pre-Columbian Woodland Habitation and Social Organization: The Brickhill Bluff Site, Cumberland Island, Georgia.

Institution: Florida Atlantic University

Thesis Advisor: Dr. Arlene Fradkin

Degree: Master of Arts

Year: 2017

This thesis examines Woodland settlement patterns at the Brickhill Bluff site on

Cumberland Island, Georgia. Aspects of Woodland habitation and social organization are not well understood along the Georgia coast. Using shell and artifact distribution data from excavations at Brickhill Bluff, this thesis attempts to discern how Woodland populations, specifically Deptford and St. Johns cultures, utilized the site between 1000

B.C. and A.D. 1000.

This study also examines the efficacy of the midden typology already established for the South Atlantic Coastal Plain by statistically comparing the artifact assemblage from Brickhill Bluff to samples from sites used to develop this midden typology. The aims of this research are to identify past cultural activities at Brickhill Bluff - specifically seasonal oyster collecting, general hunting and gathering strategies, and residential density. These criteria are compared with an established matrix designed to discern how past groups utilized southeastern coastal shell midden sites.

vi This thesis confirms that Brickhill Bluff most closely resembles oystering station sites with short durations of limited occupations. Brickhill Bluff, like other oystering stations sites, contained no evidence of structural features and very little evidence of domestic activities. However, because of the variability among archaeological sites and the different degrees of artifact preservation, it is essential that this midden typology is used cautiously. This thesis demonstrates that the identification of features provides the biggest clues to site function and formation and should be used in tandem with the benchmarks spelled out in the established midden typology to identify Woodland middens.

vii MODELING VARIABILITY IN PRE-COLUMBIAN WOODLAND HABITATION

AND SOCIAL ORGANIZATION: THE BRICKHILL BLUFF SITE,

CUMBERLAND ISLAND, GEORGIA

List of Figures……………………………………………………………………………..x

List of Tables…………………………………………………………………………..…xi

Chapter 1: Introduction……………………………………………………………………1

Problem Orientation………………………………………………...………...…...1

Theoretical Orientation……………………………………………………..…..…6

Objective and Organization of Thesis…………………………………………..…7

Chapter 2: Environmental Setting and Cultural History…………………………..…..…10

Environmental Setting………………………………………………………...…10

Cultural Overview…………………………………………………………..……16

Chapter 3: Previous Archaeology…………………………………………………….….29

History of Previous Archaeology on Cumberland Island………………………..29

History of Previous Archaeology at Brickhill Bluff……………………………..33

Chapter 4: Methodology…………………………………………………………...…….36

Phase I……………………………………………………………………………36

Phase II……...…………………………………………………………………....37

Laboratory Procedures, Analysis and Curation………………………………….38

Chapter 5: Brickhill Bluff Material Culture Results and Analysis……………………....42

Datable Evidence…...... …………………....………………………………….....42

Prehistoric and Protohistoric Artifacts…………………………………………...43

viii Historic Artifact………………………………………………………………….48

Chapter 6: Description of Woodland Assemblages at Comparative Sites……………….51

Multi Family Residential Base…..……….………………………………………………53

Bass Pond Dam (38CH124)………………………………………………...…53

Stafford North (9CM10)………………………………………………………..55

Table Point (9CM12)…………………………………………………………...56

Single Family Shell Midden……………………………………………………..58

Minim Island (38GE46)……………………………………………………..…58

38BU2…………………………………………………………………………….61

Single Family Limited Shell Midden…………………………………………….64

30CH779………………………………………………………………………....64

Fish Haul (38BU805)…………………………………………………………..65

Oystering Station………………………………………………………………...67

38BU833…………………………………………………………………………67

38BU2…………………………………………………………………………….68

Chapter 7: Comparative Sites and Middle Typology Model…………………………….70

References…………………………………………………………………………….….86

ix FIGURES

Figure 1. Location of the Brickhill Bluff site on Cumberland Island, Georgia...... 2

Figure 2. Evidence of erosion at Brickhill Bluff...... 15

Figure 3. Table Point Excavation showing ring and house site in the upper right...... 31

Figure 4. Brickhill Bluff. Station 2, established by Robert Wilson and David Brewer in 1987……………………………………………………….…………...……34

Figure 5. Shell midden eroding from Brickhill Bluff...... 37

Figure 6. Screening soil from an excavation unit...... 39

Figure 7. Shell Tools……………………………………………………………………..47

x TABLES

Table 1. Expectation for Various Site Types…………….………………………………..4

Table 2. Native American ………………………………………………………..43

Table 3. Sites sampled……………………………………………………………...……52

Table 4. Relative Frequency of Artifacts from Sampled Sites……………………..……74

Table 5. Comparative Frequencies of fauna, oyster shell, and minority shell found at the sampled sites ..……………………………………………………………..75

Table 6. Diversity of artifact types in each site sampled………………………………...76

Table 7. The Presence or Absence of Structural Features and Human Remains………...77

Table 8. Artifact visibility ranked low, medium, and high………………………………78

xi CHAPTER 1: INTRODUCTION

Problem Orientation

Woodland period habitation and social organization are not well understood along the Georgia coast. While shell midden sites in the region have yielded important information (Caldwell 1952:315–16; DePratter 1991:122–56; Espenshade 1993;

Espenshade and Brockington 1989; Espenshade et al. 1994; Jefferies 1994; Milanich

1971; Trinkley 1990, 1991; Trinkley and Adams 1994; Waring and Holder 1968), there remains a lack of understanding regarding cultural behaviors as they correlate to

Woodland shell midden formation. Archaeologists continue to place these sites under the umbrella term “shell midden.” This classification ignores the variety of characteristics found within and between sites and fails to differentiate divergent social and economic interpretations. A proven typology is needed that recognizes that not all shell middens resulted from the same formation processes.

In this thesis, I examine the efficacy of Espenshade’s (Espenshade et al. 1994) midden typology for the South Atlantic Coastal Plain by statistically comparing the artifacts from the Brickhill Bluff site on Cumberland Island, Georgia (Figure 1) with a sample of sites used by Espenshade to develop his four midden categories. Specifically, I analyze the shell and artifact data collected from the Brickhill Bluff site and compare it with the assemblages Espenshade used to create his midden typology. Then I explore how close-interval shell mapping can provide an additional variable to determine site structure. The aims of this research are to identify past cultural activities at Brickhill

1 Figure 1. Location of the Brickhill Bluff site on Cumberland Island, Georgia.

2 Bluff and to examine Espenshade’s midden typology as a tool to discern variability in settlement patterns during the . Cultural activities ultimately suggest; and in some regards, dictate how archaeological deposits are organized. A full detailed analysis of site formation encompassing paleoenvironmental and geomorphological data is beyond the scope of this thesis. However, determining specific cultural behaviors is the first step to an encompassed interpretation that acknowledges each variable contributing to the creation of the Brickhill Bluff site.

Shell Midden Formation and Typology

Espenshade’s model for prehistoric shell middens incorporated the variations and frequencies found in the zooarchaeological and artifact assemblages from five shell middens located on Spring Island, Beaufort County, (Espenshade et al.

1994:1). He later expanded this initial sample to include artifact assemblages from 17

Woodland period sites previously excavated on the South Carolina and Georgia coast. An analysis of the different assemblage traits found within these shell middens revealed patterns related to distinct cultural behaviors. Espenshade argues that a site’s “assemblage signature” can determine specific site types. Shell midden sites therefore represent a discernible range of behaviors neglected when sites are broadly labeled as shell middens.

According to Espenshade et al. (1994:34), the term shell midden is virtually meaningless and does not describe the manner in which past groups utilized these sites.

Espenshade addressed 12 variables found within the artifact assemblage of coastal midden sites: oyster contribution, non-oyster molluscan contribution, vertebrate faunal contribution, lithic density, lithic tool diversity, bone tool frequency, shell tool frequency, sherd tools/abraders frequency, presence of ideotechnic items, sherd density, presence of human burials, and presence of structural features (Espenshade et al. 1994:27). Using

3 these markers, one can measure Woodland shell midden variability, which in turn aids in determining site function. Espenshade describes four specific midden site types: Multi-

Family Residential Base, Single-Family Shell Midden, Single-Family Limited Shell Site, and Oystering Station (Espenshade et al. 1994:177). These site types are predicated on the composition of their artifact assemblages and the presence or absence of burial and structural features (Table 1).

Table 1. Expectation for Various Site Types (from Espenshade et al. 1994, Table 49, page 178).

Attribute Multi-Family Single-Family Single-Family Oystering Resident Base Shell Midden Limited Shell Station

Oyster Contribution High High Moderate to Low Very High

Non-Oyster Moderate Moderate Low Very Low Molluscan Contribution

Vertebrate Faunal Moderate Moderate Moderate to Very Low Contribution High

Lithic Density Moderate to High Moderate Moderate to Very Low High Lithic Tool Diversity Moderate to High Moderate to High Moderate Very Low

Bone Tool Frequency Relatively High Moderate Moderate Low

Shell Tool Frequency Relatively High Moderate Moderate Low

Sherd Tool Relatively High Moderate Moderate Low Frequency

Ideotechnic Items Relatively High Moderate to High Moderate Low

Sherd Density Relatively High Relatively High Moderate Low

Human Remains Relatively Common Occasional Occasional to Very Rare Rare Structural Relatively Common Common Common Very Rare Features

4 Multi-Family Residential Base

The multi-family residential base represents the largest number of inhabitants at a single locality. These sites are occupied for at least two seasons or potentially year-round.

They function as a residential center strategically located in areas that provide easy access to multiple environmental zones for hunting and gathering. These sites contain evidence of a wide range of domestic activities and often yield ideotechnic items. Shell deposits containing oysters are dominant; however, the midden also contains a significant number of minority shellfish species and a large number of vertebrate fauna. These sites contain high sherd density with a range of shell and bone tools. Multi-Family Residential

Base sites also typically contain features that define activity areas as well as human burials (Espenshade et al. 1994:177-178).

Single-Family Shell Midden

The single-family shell midden sites are simply a smaller version of the multi- family residential base. Like the multi-family sites, these sites are also evidenced by a wide variety of domestic activities that result in a large number of vertebrate fauna, sherds, and bone and shell tools. Often situated in estuarine settings, these sites provide access to shellfish and fish resources. Typically these sites are occupied for one season, but possibly year-round (Espenshade et al. 1994:179).

Single-Family Limited Shell Site

The single-family limited shell site represents a seasonal occupation by a small group or single family. These sites are characterized by limited amounts of shell generally located in small pits. Unlike the single-family shell midden and multi-family sites, the limited shell sites are found on upland tracts ideal for collecting nuts and hunting terrestrial species such as deer. This activity might indicate a late fall or winter

5 single-family habitation. The artifact assemblage at these sites is limited and there is a reduced emphasis on oyster. The increased amount of vertebrate faunal remains indicate a greater dependency on terrestrial species (Espenshade et al. 1994:179)

Oystering Station

The oystering station site is the final shell midden type defined by Espenshade.

Oystering stations include an abundance of oyster shell and a virtual lack of vertebrate remains. Occupied by small work teams for short durations, this type of site was used to procure and process oysters. The site contains a low frequency of sherds and generally few shell, bone, or sherd tools. Oystering stations typically lack any defined structural features and very little evidence of any domestic activities. This type of site is similar to what Trinkley (1992:39) described as an “encampment for collection activities.”

Lawrence (1992) described this type of site as a “fishing camp.” Previous work by

Trinkley (1992) shows no demonstrable evidence for intensive fishing activities

(Espenshade et al. 1994:179-180).

Theoretical Orientation

In the past, archaeologists have used site size, features (e.g. Trinkley 1990a) or types of shellfish species present (e.g., Claassen 1989) to classify shell middens

(Espenshade et al. 1994:26). This thesis takes a different approach. A clearer understanding of archaeological patterns emerges with an examination of midden form and assemblage characteristics. These patterns then correspond to past activities and behaviors. Using the theoretic framework of Processional Archaeology and Formation

Theory created by Michael Schiffer (1987, 1995) and Lewis Binford (1962, 1977, 1979,

1980, 1982), combined with the midden classification system developed by Espenshade et al. (1994), we can explore these patterns.

6 Binford and Schiffer, while different in their particular theoretical approaches, were each interested in the validity of the inferences we make from the material record and how that record formed. Formation theory governs this domain and shapes our understanding of how the archaeological record was created. The examination of formation processes permits us to make reasonable observations from the artifact assemblage. These observations can then be translated into an operative understanding of behavior (Shott 1998: 300-311).

Processual archaeology focuses on the variability found in the archaeological record and builds arguments to explain this variability and the cultural systems that produced it. Binford concluded that the archaeological record was organized by a unique cultural system. The artifacts and features found at a site contain patterns that can reveal information about the system that produced them (Binford 1962:220-224; Johnson

2004:13-14). An archaeological site is thus created by numerous cultural and natural phenomena operating at different spatial and temporal scales. Cultural dynamics as well as physical, chemical, and organic processes all contribute to the creation and modification of a site (Johnson 2004:15). The archaeologists must develop strategies, observe phenomena, and finally define the data matching that interpretation to an operative paradigm.

Objectives and Organization of Thesis

The objective of this study is to determine the efficacy of Espenshade’s midden typology for the South Atlantic Coastal Plain by statistically comparing the artifacts from

Brickhill Bluff with a sample of sties used by Espenshade to develop his four midden categories. I hypothesize that Espenshade’s theory is correct, and that an examination of artifact assemblages will reveal an “assemblage signature” that represents a discernible

7 range of behaviors. It is hoped that by examining artifact assemblages from a variety of coastal middens, observable patterns will emerge that will provide insight into how past groups utilized a particular site. Building upon Espenshade’s model, I also hope to show how shell mapping can provide an additional tool to determine site structure and consequently facilitate the ability to resolve activity areas within a site. I believe that a positive assessment of Espenshade’s model will provide additional data to advance his midden typology and aid in the identification of site function and settlement behavior between 1000 B.C. and A.D. 1000 at Brickhill Bluff on Cumberland Island, Georgia.

In Chapter 2, I discuss the physiographic and environmental setting of the

Atlantic Coastal Plain and Cumberland Island. I review the Pleistocene formation of

Cumberland Island and describe the island’s geology and varied ecosystems. In addition,

I provide an overview of the island’s flora and fauna followed by a discussion of the region’s rich cultural history. Chapter 3 presents a summary of over 100 years of archaeological investigations on Cumberland Island and then focuses more specifically on the previous archaeology conducted at Brickhill Bluff. This chapter ends with a discussion of the previous archaeology conducted at the nine additional Woodland sites sampled for this study. These sites were chosen to provide a representative sample of the four site types created by Espenshade et al. (1994).

Chapter 4 details the materials and methods used for data collection and analysis.

Chapter 5 presents an analysis of the material culture recovered from Brickhill Bluff during my fieldwork at the site. Chapter 6 provides an overview of the artifact assemblages from the nine comparative sites used for this thesis. Chapter 7 presents a comparative analysis of the material remains at Brickhill Bluff and at the nine sampled

8 shell midden sites in terms of artifact assemblages and site structure. This chapter concludes with a discussion of site variability and avenues for additional research.

9 CHAPTER 2: ENVIRONMENTAL SETTING AND CULTURAL HISTORY

Environmental Setting

Cumberland Island is the largest barrier island on the Atlantic Coastal Plain of

Georgia (see Figure 1). Commonly known as Sea Islands, Cumberland and its neighboring islands protect the mainland from the waters of the South Atlantic Bight.

This section of the Atlantic seaboard has a concave-shaped coast with relatively shallow waters and semidiurnal tidal fluctuations of 2 m during neap tides (when the difference between low and high tide is the least) and 3 m during spring tides (Hillestad et al.

1975:23). Cumberland Island’s large size is attributed to the relatively shallow waters surrounding the island and large tidal fluctuations. Islands with greater tidal ranges have greater beach accumulation and larger dunes, while the low wave energy found in the shallower waters far from the continental shelf help minimize beach erosion (Kaplan

1988:13; Schoettle 1993:17-18). Cumberland Island is approximately 29.6 km long and ranges from 2 to 9 km wide. The closest upland area on the mainland is approximately

3.6 km away across wide marshes and a network of rivers and creeks (Hillestad et al.

1975:23).

The southern extremities of Cumberland Island at Raccoon Key are low lying, rising only about 1.5 m above mean sea level. The bluffs at the island’s northern end at

Terrapin Point reach up to 18 m. Geologically, this compound barrier island was formed during two distinct time periods. The core (90 percent) of the island was formed sometime during the terminal Pleistocene, 25,000 to 50,000 years ago. The formation of

10 the remaining beach dune zone occurred during the Holocene in a process of erosion and accretion that continues today. In geological terms, this makes Cumberland Island an ex- tremely young and unstable island (Hails and Hoyt 1969:54; Hillestad et al. 1975:25-26).

The physiography includes freshwater lakes, ponds, streams, salt marshes, brackish ponds, beaches, and several different ecological communities.

The geological stratigraphy of Cumberland Island is divided into six basic zones: a Miocene limestone layer, Pliocene deposits, Pleistocene marsh-lagoon deposits,

Pleistocene barrier island sands, Holocene beach-dune sands, and Holocene marsh deposits (Griffin 1991). Most of the soils from these upper zones are derived from quartz sands that are highly resistant to decomposition. The main characteristic of these soils is their high permeability. Consequently, they have a very low water-holding capacity and experience rapid leaching. These characteristics, coupled with the island’s acidic soils, make for a highly nutrient-competitive vegetative community (Hillestad et al. 1975:32-

33).

The freshwater resources of the island include permanent, temporary, and artificial ponds. The majority of the groundwater found on Cumberland Island comes from two water-bearing strata. A deep-lying limestone artesian aquifer known as the

Principal Artesian Aquifer, or the Coastal Plain Aquifer, lies approximately 150 m below the surface and is recharged by rainfall flowing underground from inland Georgia. This aquifer surfaces as submarine springs offshore near the continental shelf (Hillestad et al.

1975:45-47). A much shallower aquifer collects water that permeates from the surface

(Hillestad et al. 1975:49). The standing surface water on the island is much less stable, with only a small number of permanent ponds. These ponds experience a dramatic reduction during times of extreme drought. Generally, these ponds average only 1 m in

11 depth. The one exception is Lake Whitney which is the largest of the permanent ponds, averaging 2 m deep and covering approximately 33 ha with 16 ha of open water.

Nineteen other natural and artificial ponds, outflows, and drainage fields were listed on the revised 1965 Cumberland Island plat map prepared for the Lucy Carnegie estate

(Hillestad et al. 1975:53).

The large size of Cumberland Island allows for the creation of several distinct ecosystems. These include the active beach, dune areas, swales, beach meadows, maritime forest, and saltwater marshes. The vegetation within these communities, although often overlapping, is used to distinguish each of the zones (Hellmann 2005;

Hillestad et al. 1975:75; Schoettle 1993:19).

The active beach zone occupies the eastern seaside portion of the island between the low- and high-tide zones. This area has virtually no sustainable vegetation. Adjacent to the beach zone is a very sensitive dune system that runs along the eastern shoreline

(Hellmann 2005; Hillestad et al. 1975:83). Dune vegetation includes sea oats, sandspur, saltmeadow cordgrass, Russian thistle, sea rocket, orach, sea purslane, and beach croton.

Areas of the dunes facing the ocean receive high amounts of salt-spray. These zones are populated by salt-tolerant plant species like marsh elder, Caribbean sea spurge, inkberry, dune panic grass, and prickly-pear cactus (Kaplan 1988:32-34; Schoettle 1993:23).

Grasses and plants such as camphorweed, wild bean, butterfly pea, pennywort, dune primrose, spurge-nettle, muhley grass, and fireweed occupy the areas behind the dunes known as swales. These pioneer plants create beach meadows (Schoettle 1993:24).

Over time, salt-tolerant shrubs and trees replace these plants and eventually provide a rich humic layer (Kaplan 1988:20). This rich soil allows species such as wax myrtle, red

12 cedar, Hercules’ club, pine, yaupon holly, groundsel-tree, redbay, and oaks to prosper and create a shrub forest within the swale (Schoettle 1993:25).

Usually located on the oldest areas of the island, that is, interior or uplands from the coast, the maritime forest is characterized by a community of live oaks, southern magnolias, pines, and cabbage palms. It is considered a climax community of the

Southeastern barrier islands. This means that the maritime forest is capable of self- propagation and remains relatively unchanged over time. However, when portions of the maritime forest are cleared for agriculture or destroyed by natural phenomena, it takes centuries for the soil to regain the nutrients needed to regrow. Most often the cleared maritime forest is repopulated by loblolly pine and slash pine. Eventually these pine forests revert to hardwood stands (Schoettle 1993:33-34).

The western shoreline of the island is dominated by an almost monospecific saltwater marsh (Hellmaan 2005; Schoettle 1993:26). The cordgrass that thrives in this marsh system covers over one-third of the island’s total acreage. In addition to cordgrass, the salt-marsh community consists of grass marsh, shrub border, and oak-juniper-palm forest (Kaplan 1988:15).

The human-altered systems on Cumberland Island vary in location. They include areas maintained directly by human activity or indirectly by feral animals. Pastures, lawns, cultivated fields, and pine plantations all make up these miscellaneous human- altered areas (Hellmann 2005).

The diversity of species on Cumberland Island is a reflection of the wide range of habitats and resources available for use. The island supports approximately 575 documented terrestrial vertebrate species (Echternacht and Harris 1993:103) and is home to the American alligator, sea turtles, migratory waterfowl, and a large number of

13 mollusks, crustaceans, and fish (Hellman 2005). The most common types of mollusks are ribbed mussels, marsh periwinkles, and oysters (Schoettle 1993:112). The estuaries provide a rich habitat for a variety of fish species including catfish, drum, mullet, and sharks. Dolphins and occasionally manatee populate the waters around Cumberland

Island (Zoodsma 1991:7, 60).

Terrestrial animals found on the island include deer, opossum, raccoon, gray squirrel, rabbit, and field mice. Armadillos are a recent invasive species found throughout the island (Stalter and Odum 1993:136-137). Reptiles found on the island include a variety of venomous and nonvenomous snakes. Cottonmouth moccasins, rattlesnakes, and colubrid snakes such as the southern black racer and the red ratsnake call

Cumberland Island home (Stalter and Odum 1993:146). The freshwater sloughs and ponds provide habitats for freshwater fish, frogs, turtles, and alligator (Schoettle 1993:35-

36).

Over 300 species of resident and migratory bird species are found on Cumberland

Island. The ringbill gull, sanderlings, whistling swan, snow goose, Canada goose, gadwall, and mallard represent migratory bird species on the island. The brown pelican, ibis, egret, and heron are examples of birds that reside on the island year-round. Along the beach and dune areas, nesting shorebirds like the Wilson’s plover, American oystercatcher, black-necked stilt, laughing gull, terns, and black skimmers share a nesting ground with loggerhead and hawksbill sea turtles. The island is also home to a community of raptors such as the peregrine falcon, merlin, osprey, kestrel, and bald eagle

(Stalter and Odum 1993: 141-145).

The Brickhill Bluff site is situated on the north end of Cumberland Island within a maritime forest ecosystem. Live oaks, pines, and cabbage palms cover a majority of the

14 site. The Brickhill River snakes along the western boundary of the site and provides deep-water access to the island for boats and barges. However, there is no protective salt- marsh zone, leaving the shoreline directly exposed to waves, tides, storm surges, and the river current. Over time, these forces have caused significant shoreline erosion resulting in substantial undercutting and slumping of the bluff (Figure 2). This erosion frequently exposes artifacts and archaeological features. A backcountry campground and several hiking trails bisect the site, leading to further degradation. This continual deterioration of the historic fabric at Brickhill Bluff highlights the importance of investigating and recording the archaeology at the site before it is lost forever.

Figure 2. Evidence of erosion at Brickhill Bluff

15 Cultural Overview

Cumberland Island is located in the northern portion of an area that has been archaeologically designated as the St. Mary’s Region. This region is a coastal area that extends from the St. Johns River in Florida north to the Satilla River in Georgia. This region has a cultural history distinct from the central and northern Georgia coast and the

St. Johns area in Florida (Russo 1992:107). Historically, we know there were two indigenous linguistic groups living around Cumberland Island. The French identified the first group as the Quade. The Spanish and English called them the and the Wallie, respectively. The second group was known as the Timucuan. Spanish records from the sixteenth century show that the Timucuan language was distinct from the Guale’s

Muskogean dialect. Priests who ministered to the Timucuan groups along Florida’s eastern Atlantic coast found it impossible to communicate with the Guale who inhabited the central and northern Georgia coast (Crook 1978:35-36).

Three anthropological perspectives are used to interpret the archaeological record in the St. Mary’s region. Prehistorically, the indigenous groups living in this area may be viewed as a border area caught between two larger cultural spheres of influence. In this scenario, competing groups from the east Florida coast and the Georgia coast struggled to occupy contested territory creating a cultural sequence where an outside group abruptly replaces an established tradition. Another theory regards the St. Mary’s region as an area of transition. In this scenario, cultural attributes from the southern groups melded with cultural attributes from the northern groups. A third view operates on the belief that the cultures in the region are unique from those surrounding them. In this third scenario, the peoples living in this area developed unique separate identities and simply practiced more than one ceramic tradition (Russo 1992:120). Current data from this region is insufficient

16 to definitively determine which of these three perspectives best characterized the prehistoric inhabitants living on Cumberland Island and in the broader St. Mary’s region

(Ashley 2008; Russo 1992)

Although all the sites sampled for this thesis date to the Woodland Period, it is useful to understand the surrounding archaeological periods to place the assemblages and sites in context. Consideration of the prehistoric settlement, subsistence, and social organization strategies and of how these strategies changed over time is especially pertinent to discerning how people formed and used their environment. In order to make an accurate hypothesis regarding how Woodland assemblages and structures correspond to specific past activities, one must consider the organization and values of the occupant’s culture.

Before the Woodland Period: Archaic Period

The Archaic period dates from about 9500 to 1000 B.C. and is characterized as a stage when populations of hunter and gatherer bands grew and became more sedentary

(Griffin 1967:178; Milanich 1994:85). Archaic people are believed to have been organized into semi-nomadic bands that occupied different territories. These groups moved seasonally to take advantage of different natural resources. For example, a group might move to a coastal site to gather oysters and later move inland to collect acorns and hunt. This mobility, while less frequent than their fully nomadic ancestors, allowed groups to fully exploit resources as they became ripened or more easily available within the environment of southeastern .

Three subperiods are generally recognized: Early Archaic (9500–6900 B.C.),

Middle Archaic (6900–3600 B.C.), and Late Archaic (3600–1000 B.C.). These subperiods are based primarily on changing climactic conditions and changes in

17 styles. The Early Archaic Period is marked by the initial 2000 years of the

Holocene and the occurrence of side– and corner-notched and bifurcated-based bifaces like Dalton and Hardaway-Dalton types. The end of the Early Archaic coincides with the onset of a warming episode known as the Hypsithermal. The corner-notched and bifurcated bifaces were replaced by square and contracting stemmed Kitck Stemmed,

Stanley Stemmed and Morrow Mountain (Anderson et al. 1996:15). The Middle Archaic has traditionally been regarded as a period of adaptation to a changing environment leading to increased settlement permanence, population growth, and subsistence intensification of the Late Archaic. The appearance of Savannah River Stemmed projectile points marks the beginning of the Late Archaic. The Late Archaic is also recognized by the development of fiber-tempered ceramic vessel technology (Sassaman et al. 1990:10-11).

Woodland Period

The Woodland period in the Southeast is characterized by the following: the establishment of semipermanent or permanent villages occupied much of the year, the widespread adoption of pottery use, the routine construction of mounds, and the elaboration of an incipient system of horticulture (Smith 1986, 1989). Moreover, the

Woodland period is divided into three subperiods – Early, Middle, and Late – that correspond to the development of these attributes. Traditionally, the Early Woodland is represented by the first widespread use of pottery. The Middle Woodland is characterized by the rise and fall of a “panregional ceremonially based interaction network.” Finally, the Late Woodland is a period of cultivation intensification, increasing population, and political fragmentation (Anderson and Mainfort 2002:1).

18 The established chronology for the three subperiods placed the Early Woodland between 1000 B.C. and 200 B.C., the Middle Woodland subperiod between 200 B.C. and

A.D. 400, and the Late Woodland subperiod between A.D. 400 and A.D. 1000. Changes in pottery types and projectile point forms during this 2,000-year period provide the clearest chronological markers for each of these subperiods (Anderson and Mainfort

2002).

Early Woodland projectile points in eastern Georgia initially included a number of stemmed types, followed by larger triangular points, and finally medium to small triangular forms, which spanned the entire Late Woodland period into the Mississippian period. Diagnostic Woodland points are typically smaller than Archaic forms. A major reduction in point size takes place at the end of the Woodland period with the advent of the bow and arrow in the Southeast (Anderson and Mainfort 2002:2; Justice 1987:229).

However, changing pottery styles – especially on the coast where lithic resources were scarce – played a more important role in defining the temporal markers of the Woodland chronology. Archaeologists also use these changes in pottery types to indicate transformations in cultural traditions.

The advent of widespread pottery use throughout the Southeast is the conventional taxonomic divider between the Archaic and Woodland periods. After approximately 1000 B.C., practically all indigenous groups within the Southeast utilized some form of ceramic technology. Pottery was a revolutionary new cooking and storage technology that created significant social change (Sassaman 2002:398-399). In the area around Cumberland Island, the Woodland ceramic chronology begins with a transitional ceramic type similar to early fiber-tempered Orange wares.

19 Refuge Culture

Antonio Waring first identified the Refuge phase in 1947 at the Refuge site on the

Savannah River. This phase marks a transition from the Late Archaic to the Early

Woodland. Within the St. Mary’s region, the Refuge culture begins approximately 1000

B.C. and ends around 500 B.C. with the advent of the (Russo 1992:113).

Refuge pottery is identified by incised, punctuated, stamped, and plain pottery with a coarse, gritty paste (Anderson 1996; DePratter 1976:6; Sassaman 1993). They are similar to Deptford ceramics in form, paste, and stamping; however, Refuge pottery shows evidence of having interior decoration as well as being abraded (Anderson 1996;

DePratter 1979:116; Sassaman 1993). Refuge wares also share elements with Late

Archaic fiber-tempered wares and are often termed “semi-fiber tempered.” The Refuge ceramic tradition is believed to be a hybrid that separates the earlier Orange pottery types and the late Deptford and St. Johns I wares (Russo 1992:113).

During the Refuge phase, sea levels were much lower than present-day levels

(Colquhoun et al. 1980; DePratter and Howard 1980, 1981). As a result, many coastal

Refuge sites are submerged or hidden under the marshes and sediments that separate the

Sea Islands from the mainland. The inaccessibility of these sites makes interpretation of the Refuge culture difficult. Some have argued that the lower sea levels resulted in a change in subsistence strategies. In Georgia, once the estuaries were drained, groups abandoned their coastal sites and moved to the interior. Here they developed a greater dependency on hunting and gathering (Colquhoun et al 1980:152; DePratter 1979; Saffer

1979). However, a change in sea level does not necessary predicate a mass exodus to the interior. The exploitation of coastal resources likely continued as the estuaries shifted farther east. The paucity of Refuge sites on today’s coastline is likely a result of an

20 unproductive environment. Today’s teeming estuaries would have been less productive as sea levels dropped and freshwater overtook the rich marine environment. Since these areas were not as desirable, the more extensive Refuge occupations would have followed the retreating estuaries in places now covered by water and marsh sediments (Russo

1992:114).

Deptford Culture

In the area along the Atlantic coast from South Carolina to northeast Florida and along the Gulf coast from the Florida Panhandle to Tampa Bay, the appearance of the

Deptford culture begins in the Early and Middle Woodland periods. On the Atlantic coast, the florescence of the Deptford culture begins between 800 and 500 B.C. Deptford emerges later in Florida, with Deptford stamped ceramic vessels appearing around 500

B.C. (Stephenson et al. 2002:318). In both Florida and along the Atlantic coast, the

Deptford ceramic tradition consisted of sand and grit-tempered, coil-made pottery fashioned by impacting the surfaces of the wet clay vessels with cylindrical sticks, split canes, or paddles. This new method of pottery production improved the fusing of the clay coils and created unique decorative motifs not found in early pottery types (Milanich

1994:112–114; Tesar 2011:1).

The Deptford pottery tradition is divided into early (500–100 B.C.) and late (100

B.C. –A.D. 200) (Tesar 2011:1). The earliest Deptford pottery types—Deptford Simple

Stamped and Deptford Cross Stamped—are distinguished from later Deptford wares by their particular stamping patterns. Deptford Cross Stamped, also referred to as Deptford

Cross Simple Stamped, is distinguished from Deptford Simple Stamped by the overlapping simple stamping that crosses at a 30-degree angle.

21 Later during the Deptford culture, paddles were introduced and became the principal tool for welding the clay coils together to form deep, cylindrical pots with rounded or conoidal bottoms (Milanich 1994:129). Plain paddles yielded Deptford Plain pottery with undecorated exterior surfaces. Wrapping paddles with twisted fiber cordage produced the Deptford Cord Marked type. Cutting parallel and transverse linear grooves of different widths into paddle surfaces produced an even wider assortment of Deptford pottery types commonly recognized today as Deptford Check Stamped and Deptford

Linear Check Stamped (Anderson 1996; DePratter 1979:111–112; Tesar 2011; Williams and Thompson 1999:36–38).

Milanich describes three Deptford subregional adaptations: Atlantic, Gulf, and

Inland Interior Riverine. The groups who inhabited these subregions practiced similar subsistence economies and ceramic traditions; however, there are differences in the developmental sequences and, of course, geographic and environmental variations

(Milanich 1994:113-115). For this thesis, we will focus on the Atlantic subregion, which extends from central South Carolina to the mouth of the St. Johns River near

Jacksonville, Florida.

Four decades ago, Milanich (1971, 1973) proposed his first model of Deptford settlement organization. While slight modifications to this model have occurred over the years, much of his original scheme remains intact. The Deptford culture is viewed as a coastal oriented band level social group subsisting on hunting, gathering, and fishing activities. Excavations of Deptford burial mounds on St. Catherines Island, Georgia, recovered modest grave goods associated with bundled burials. Most of these burials were female which led Thomas and Larsen (1979:149-150) to propose a matrilineal- based society.

22 Deptford sites within the Atlantic subregion were situated within the forest bordering large salt marshes and on the offshore islands. Camps and villages were spaced every 13 to 16 km along the coast, and a wide range of aquatic and terrestrial species were exploited for food. According to Milanich, Deptford groups lived along the coast for most of the year. However, during certain seasons small expeditions or entire groups would move inland along rivers to collect ripened acorns and hickory nuts. All of these inland sites could be reached within a few days of walking, or faster by river from the coast (Milanich 1994:112-114; Stephenson et al. 2002:328)

Archaeological survey and excavations on the Atlantic Coastal Plain have provided a substantial amount of data to test Milanich’s model of coastal settlement with seasonal inland migration. While Milanich’s theory that Deptford groups traveled between coastal and inland environments appears appropriate, most of the Deptford shell middens excavated in the past two decades do not represent the substantial camps or villages described by Milanich. Instead, most Deptford sites are formed as a series of small shell heaps that gradually accumulated. This hypothesis of gradual accumulation combined with the paucity of structural features and low artifact density reflect short periods of site use (Stephenson et al. 2002:329-330). This data illustrates the great variability in Woodland shell middens (Espenshade 1993; Espenshade and Brockington

1989; Espenshade et al. 1994).

Swift Creek Culture

The , recognized almost exclusively by its elaborately stamped pottery, occurred along the southern Georgia coast and throughout northern Florida during the late Middle Woodland and Late Woodland periods (A.D. 100–700) (Williams and Elliott 1998:1). It is believed that Swift Creek pottery is the oldest example of

23 complicated stamped vessels in the Southeast (Stephenson et al. 2002:333). In the St.

Mary’s region, it appears to converge with Deptford and St. Johns 1 and prospered for roughly 600 years (Russo 1992:116; Stephenson et al. 2002:342). Still, the arrival of

Swift Creek ceramics and the degree of Swift Creek cultural influence in the St. Mary’s region is open to interpretation (Ashley 1998:220).

Like many Deptford sites, a majority of Swift Creek sites are located adjacent to large salt marshes with smaller special purpose camps located inland. Unlike Deptford sites, large Swift Creek village sites appear inland indicating a significant movement of indigenous groups from the coast to the interior (Milanich 1994:142-143). Evidence suggests that, like the coastal Swift Creek sites, these larger interior villages were complemented by smaller special-purpose camps established with the purpose of exploiting specific environmental resources not found near the larger village (Milanich

1994:145; Stephenson et al. 2002:345).

Traditionally, archaeologists viewed Swift Creek pottery in the St. Mary’s regions as a nonlocal trade ware (Goggin 1952). However, while the St. Mary’s Region is on the periphery of the Swift Creek sphere of influence, the abundance of pottery types at habitation sites suggests a certain degree of local production of Swift Creek wares

(Ashley 1998:218-220).

St. Johns Culture

The St. Johns culture began around 500 B.C., near the end of the Early Woodland period, and continued through the Mississippian period until European intrusion significantly altered traditional Native American lifestyles. Milanich (1994:247) described six divisions in St. Johns ceramics: St. Johns I (500 B.C. – A.D. 100), St. Johns

Ia (A.D. 100–500), St. Johns Ib (A.D. 500–750), St. Johns IIa (A.D. 750–1050), St. Johns

24 IIb (A.D. 1050–1513), and St. Johns IIc (A.D. 1513–1565). A refined ceramic chronology for Northeast Florida proposed by Keith Ashley removes St. Johns I from the

Woodland period of Northeast Florida and restricts St. Johns II phase to the Mississippian period (A.D. 900-1250/1300) (Ashley 2008:125).

This modified ceramic chronology is similar to what William Sears formulated in

1957 when he worked along the lower St. Johns River. His excavations found a low occurrence of St. Johns Plain sherds and a predominance of sand-tempered plainwares.

His Woodland period ceramic chronology opened with the Deptford complex, followed by sand-tempered plain wares, and finally ending with a grog-tempered ware known as

Colorinda. According to Sears, St. Johns II replaced Colorinda and marked the beginning of the Mississippian period (Sears 1957:33). Additional testing is necessary to determine if this modified St. Johns ceramic chronology is applicable to the Georgia coast. For the purpose of this thesis, we will continue to use Milanich and Russo’s chronology, which places St. Johns I and early St. Johns II in the Woodland period.

The St. Johns wares are distinguished from other pottery types by the use of diatomaceous earth-containing sponge spicules, which give surfaces a chalky feel (Russo

1992:116). St. Johns Plain pottery predominates artifact assemblages from the St. Johns I,

Ia, and Ib sites. St. Johns Check Stamped pottery, absent at St. Johns I sites, is commonly found in assemblages from the St. Johns IIa, IIb, and IIc sites. The appearance of Dunns

Creek Red, a minority ware, indicates St. Johns Ia or early St. Johns Ib components (Mi- lanich 1994:247). Swift Creek trade wares are linked with St. Johns Ia.

Evidence suggests that early St. Johns ware and other non-fiber-tempered pottery were coeval with the transitional Orange-pottery tradition, eventually superseding it in the St. Johns region of northeast and east-central Florida. In northeast Florida, Orange

25 Incised pottery containing sponge spicules has been dated to about 2000 B.C. (Russo

1992:113). Likewise, south of the St. Johns River valley, chalky, textured, sand-tempered pottery has been recovered from the Joseph Reed Shell Ring (8MT13) near Hobe Sound.

This shell ring, which did not contain any fiber-tempered pottery, was constructed between 3,100 and 3,500 years ago, toward the end of the Late Archaic period (Russo and Heide 2000:53). More study, however, is needed to further elaborate on the transition from Orange to St. Johns wares.

Russo (1992) has suggested that construction of burial mounds, increased sedentism, horticultural activities, and the presence of a more stable environment distinguish the St. Johns culture from those of the preceding Late Archaic period. The extent cultivated crops were adopted along the Atlantic coast is unknown. Widespread cultivation seems unlikely because of the abundant natural resources afforded by the maritime environment. A hunting-fishing-gathering lifestyle, prevalent during the previ- ous Archaic and transitional periods, likely continued to provide a major portion of the diet of the St. Johns peoples.

Wilmington / St. Catherines

Wilmington sites are associated with the Late Woodland period and defined by heavy cord-marking on clay-tempered pottery (also sometimes referred to as grog- tempered). This type is commonly found on the coast between South Georgia and North

Carolina. Wilmington pottery differs from other Woodland varieties by the inclusion of clay with some variability in the composition of the paste and tempering materials (Crook

1978:72). The continued use of sand tempers and decorations consistent with previous traditions makes differentiation of early Wilmington pottery from other late Woodland ceramic types difficult (DePratter 1979:119).

26 St. Catherines pottery replaces Wilmington II phase around A.D. 1000. While St.

Catherines pottery types are generally accepted to be a Late Woodland variety, some researchers claim that grog-tempered St. Catherines wares may have persisted in some areas of Georgia beyond the Late Woodland period. This ceramic tradition may well have continued until perhaps as late as A.D. 1600 (Anderson 1996:245). St. Catherines pottery is characterized by having a clay or grog temper with smoothed and burnished interiors and exteriors. The three varieties of St. Catherines ware include St Catherines Burnished,

St. Catherines Plain, and St. Catherines Fine Cord Marked. A variety called St.

Catherines Net Marked has also been recovered; however, this type is rare at most sites

(DePratter 1979:119).

The Wilmington and St. Catherines pottery tradition marks a time characterized by the establishment of permanent villages and the beginning of horticulture in most of the Southeast (Smith 1986, 1989). It is important to note that despite changes in subsistence and settlement patterns from the end of the Late Archaic throughout the entire Woodland period, evidence suggests that people along the Atlantic Coastal Plain continued to rely heavily on the abundant natural resources traditionally gathered from the surrounding environment.

After the Woodland Period: Mississippian Period

In the development of southeastern native cultures, the Mississippian period is generally characterized as a time of great socioeconomic complexity (Griffin 1967, 1985;

Muller 1983; Peebles and Kus 1977). Along the Georgia coast, this change began with the end of band level societies and the beginning of a more complex socio-political system. This increased complexity is evidenced by the introduction of platform mounds, large burial mound construction, wall-trenched houses, an increased reliance on

27 horticulture, and nucleated settlements centered around plazas and mound structures

(Crook 1978:81; Steponaitis 1986:387-388).

Inhabitants of the Georgia coast were conservative in their adoption of

Mississippian traits and sociopolitical organization. This hesitancy to adopt the same lifeway as the indigenous groups living along the inland rivers is perhaps a function of living in a unique environment with alternative subsistence strategies (Crook 1986;

Hellman 2005). Unable or resistant to rely solely upon farming for subsistence, the indigenous groups along the coast continued to harvest wild foods and maintain some of their pre-horticultural traditions. For this reason, the areas along the coast are described as being “geographically peripheral and culturally marginal to the mainstream of

Mississippian culture development” (Cook and Pearson 1989:14).

28 CHAPTER 3: PREVIOUS ARCHAEOLOGY

History of Previous Archaeology on Cumberland Island

Clarence B. Moore conducted the first recorded archaeological excavations on the

Georgia coast in 1897. Near Cumberland Island, he excavated the Owen’s Ferry and

Woodbine mounds along the Satilla River. Moore also excavated the Fairview and Kings

Bay mounds in Camden County, recording and publishing his research findings in

Certain Aboriginal Mounds of the Georgia Coast (Moore 1897). Frank M. Setzler of the

Smithsonian Institution conducted the first archaeological project on Cumberland Island in February 1932 when he investigated the remains of a dugout canoe, which Thomas M.

Carnegie discovered in a small creek. A 2.5 m section of the canoe was removed, strapped to a wooden plank, cleaned, and sent to the Washington Navy Yard (Setzler

1932:60).

Lewis Larson conducted the first archaeological survey of Cumberland Island in

1953. This survey involved only surface collections. Unfortunately, site records from this survey are lost. The first recorded survey of Cumberland Island took place in 1969 under the direction of Charles Fairbanks. Fairbanks investigated a number of sites in Camden

County, Georgia, including sites on Cumberland Island. Fairbanks received a list of seven sites compiled by Joseph Caldwell who briefly visited the island earlier in 1969. Included in Caldwell’s list of seven sites were three mound sites, two historic sites, and two small scatters of sherds located on the beach. There is no evidence that Caldwell tested any of these sites beyond a cursory surface collection. Fairbanks invested most of his energy on

29 the island excavating two historic sites: Rayfield and Stafford slave quarters (Ascher and

Fairbanks 1971; Duetschle and Wilson 1975:2).

In July and August 1970, Jerald Milanich of the University of Florida excavated two Deptford phase sites on Cumberland Island: Stafford North and Table Point. Stafford

North contained multiple “shell piles” extending 398 m along the marsh bank (Milanich

1971:30). Seven 1.5 x 3 m excavation units revealed features associated with a living area or structure (Milanich 1971:32). This structure was comprised of crushed shell, three postholes, and a pit feature all situated within a shallow depression. The three postholes extended 27 cm beneath the shell zone into sterile sand. The artifacts recovered included mostly Deptford Check and Simple Stamped wares. A minority of cord-marked ceramics and red-slipped wares were found with various plain sherds with grit, sand, grog, and fiber paste. Faunal remains included deer, red-breasted merganser, diamondback terrapin, marine catfish, sheepshead, eagle ray, bull shark, raccoon and large amounts of shell of which 90 to 95 percent were oysters (Milanich 1971: 41-43).

The Table Point site contained shell deposits concentrated along the marsh edge running north to south for 137 m. The most eastern edge of the site extended 100 m from the high tide line (Milanich 1971:46). The site was composed of numerous piles of shell that measured 15 to 30 cm high and 6 to 9 m in diameter (Milanich 1971:26). Within the complex was a large shell ring with an oval midden in the center. The shell ring was 67 m in diameter and 6 to 7.6 m wide with a depth of 20 to 60 cm (Figure 3). Two gaps in the shell ring were attributed to shell mining for tabby construction and road paving in the nineteenth and twentieth centuries. Within the shell ring was another ring measuring 45.7 m in diameter and 18 m wide (Milanich 1971:49). The ring fill contained St. Johns ceramics as well as sherds tempered with sand, grog, and fiber (Milanich 1971:51).

30 Excavations on the northwest edge of the shell ring at Table Point identified a

Deptford period house. The oval-shaped house measured 9.7 x 6.7 m. Associated features included postholes, wall trenches, a fire pit, and shell refuse related to cooking (Milanich

1971:67). A radiocarbon date of A.D. 55 +/- 95 years was obtained from a shell pick recovered from the fire pit (Milanich 1971:69). Excavation units in the center of the shell ring and in the central midden north of the shell ring contained Deptford ceramics with the remains of turtle, catfish, raccoon, and porpoise (Milanich 1971:58-59). Probe tests

12 m northeast of the Deptford house structure identified a small buried shell midden.

This midden was 15 cm deep and contained only a trace amount of ceramic sherds and food bone (Milanich 1971:61).

Figure 3. Table Point Excavation showing ring and house site in the upper right (from Milanich 1971, Figure 5).

31 On October 23, 1972, the United States Senate and House of Representatives passed a law designating Cumberland Island as a National Seashore. Various federally funded archaeological surveys in the following years provided a broader picture of the cultural resources within the park’s boundaries. Donald L. Crusoe conducted a limited initial survey of the island in 1973 (Crusoe 1973). This survey was followed by a more expansive survey conducted by John Ehrenhard in 1975. Ehrenhard’s full-scale Cultural

Resource Inventory of Cumberland Island sought to locate and record as much of the island’s archaeological and cultural features as possible. Another goal of this project was to provide a preliminary evaluation of National Register eligibility for sites previously listed and newly discovered (Ehrenhard 1975:1). The survey consisted primarily of a pedestrian survey and surface collection from identified sites. The project identified 49 prehistoric and historic sites (Ehrenhard 1976:104).

Currently there are 90 identified archaeological sites within the boundaries of the

National Seashore. Seventeen of these sites are listed on the National Register of Historic

Places. Five sites are considered potentially eligible for nomination to the National

Register, including Brickhill Bluff (Ehrenhard 1976:104). Of the 90 known archaeological sites, 25 sites lack information recording their exact temporal position and cultural affiliations. Of the 65 sites with data showing their cultural affiliations, 53 of these sites have a prehistoric component and 27 have a confirmed Woodland cultural occupation. Additional sites on the island likely date to the Woodland period, but further archaeological testing is necessary to evaluate the cultural history of all properties. Given the available data, over half of the recorded prehistoric sites on Cumberland Island contain components that date to the Woodland period.

32 History of Previous Archaeology at Brickhill Bluff

Brickhill Bluff was first recorded in 1975 by John Ehrenhard (1976). The site was assigned the National Park Service site designation CUIS-24 and given the Georgia state site file number 9CM85. Ehrenhard estimated that Brickhill Bluff extended about 600 m north-south and had a width of approximately 80 m. He noted several shell deposits within the site that averaged 7 m in diameter. The Brickhill Bluff site is located in an area composed of Chipley sand, with considerable palmetto, oak, and pine growth limiting visual survey. Ehrenhard (1976:56) mentioned that while some of the middens appear to have been mined historically, large portions of the site appeared undisturbed. He also commented that the bluffs overlooking the Brickhill River may have once been the location of the Spanish mission San Pedro y San Pablo de Puturiba, based on the presence of several Spanish olive jar fragments (Ehrenhard 1976).

In October 1987, an archaeological site monitoring program was established for

Cumberland Island. This program evolved around increasing concerns for the stability of sites on the island's western inland shoreline and the perception that erosion was being exacerbated by the dredging of the Cumberland River and the St. Mary’s Channel by the

Army Corps of Engineers for the Navy submarine base at Kings Bay. The site monitoring program included limited surface collecting of artifacts exposed along eroding shorelines

(Faust 1987:6–7).

Archaeologists Robert Wilson and David Brewer established monitoring stations at Brickhill Bluff. Shoreline measurements were taken using these stations, and each station’s position was then referenced with known USGS benchmarks. Two stations were established at the Brickhill Bluff site. Station 1, the northernmost, was placed near the

Brickhill Bluff campground sign, while Station 2 was at the south end of the site, just

33 north of a water pump (Figure 4). Wilson and Brewer made five more trips to record changes in the shoreline and conduct surface collections of exposed artifacts. The

Brickhill Bluff erosion was so severe that two shell-filled pits had become visible as portions of the bluff sloughed away into the river. The first feature was near Station 1. It measured about 60 cm in diameter and extended 80 cm below ground surface. The second feature, measuring 95 cm in diameter and 110 cm below ground surface, was located an unknown distance north of the first feature. (Thorne 1988; Wilson 1988).

Brewer recommended site testing to gather as much information as possible about the remaining archaeological resources. Brewer also recommended continued monitoring of the erosion using video recordings to augment the documentation (Brewer 1993 and

1994). Unfortunately, these recommendations were never carried out, so erosion data is unavailable.

Figure 4. Brickhill Bluff. Station 2, established by Robert Wilson and David Brewer in 1987.

34 In 1999, archaeologists Dean Wood and John Doolin from Southern Research, a cultural resources management firm, conducted a series of six shovel tests at Brickhill

Bluff to evaluate the feasibility of locating a dock at the site. Their tests ran along 125 m of shoreline in the southern portion of the site. Based on their findings, which included the identification of dense shell midden within the project area, they concluded that the local archaeological resources could be considered eligible for listing on the National

Register of Historic Places. Wood recommended that the dock should not be constructed in this area (Wood 1999).

Until now, archaeologists speculated about the exact boundaries of Brickhill

Bluff. Additionally, there was a poor understanding of the precise cultural sequence of the site. Prior to my research, the occupational sequence of the site was based primarily on pottery types recovered along the beach. To build upon the previous research and increase our understanding of the Brickhill Bluff site, my research includes a testing strategy to determine site boundaries and provide information regarding site occupation.

Also included is the location and mapping of areas of high artifact and shell concentration and measuring shoreline accretion and attrition. Additionally, I will compare the data from Brickhill Bluff to Espenshade’s 1994 midden study. This comparison, combined with the artifact and stratigraphic evidence, should increase our understanding of

Brickhill Bluff and its prehistoric occupants.

35 CHAPTER 4: METHODOLOGY

The initial fieldwork for this project took place in July and August of 2007.

Personnel from the Southeast Archeological Center (SEAC), National Park Service,

Tallahassee, Florida, conducted the initial excavations for this project under my field supervision. The principal goal of this investigation was to ascertain accurate site boundaries and assess the impact of shoreline erosion. Additionally, we sought to gather preliminary information regarding site structure and sequence of occupation. The excavations followed a research design that entailed archaeological testing of the site using shovel tests, unit excavations, and surface collections (Kidd and Wise 2007).

Phase I

Work began with a detailed examination of the shoreline, coupled with a beach surface collection of artifacts. As a result of storms, tides, and wave action, artifacts are frequently washed out of the bank and left exposed on the beach (Figure 6). All artifacts recovered along the shoreline were given a specific field specimen (FS) number, and the location of each artifact was identified in relation to the site grid established for the shovel-testing phase of the project.

Shovel tests were excavated at 20 m intervals. All soil from the tests was screened through 6.4 mm (1/4-inch) mesh hardware cloth. A grid system was established at the site with a primary north–south baseline aligned along the bluff. The shovel tests were assigned a specific number based on coordinates linked to the grid. Shovel tests were dug

1 m deep and 30 cm in diameter.

36 Figure 5. Shell midden eroding from Brickhill Bluff

Figure 6. Shell midden eroding from Brickhill Bluff.

Shovel testing continued north-south and east-west along the grid until two consecutive negative shovel tests were encountered. Due to the large amount of shell scattered over such a vast area, I decided that shovel tests were deemed negative if only a trace amount of shell (less than 10 g) was found. A total of 183 shovel tests were excavated at the Brickhill Bluff site. Of these, 93 (51 %) contained cultural materials.

Phase II

Immediately following surface collection and shovel test excavations, we began opening excavation units. Areas of highest artifact concentration deemed to have the best potential to provide data regarding site structure and variability were flagged for Phase II testing. The placement of excavation units was also predicated by a desire to recover information in areas immediately threatened by shoreline erosion. Of the three 1 x 1 m

37

units, two were close to the bluff in areas with exposed midden and eroding artifacts. A third unit was placed in an area with a high concentration of historic material.

Units were excavated in natural stratigraphic zones designated as levels. Vertical control was maintained using a line level attached to the southwest corner of each unit.

As with the shovel tests, all soil from the excavation units was screened through 6.4 mm

(1/4-inch) mesh hardware cloth (Figure 6). Using a plastic bucket labeled in liters, the volume of shell encountered was recorded for every level of each unit.

All recovered artifacts were placed in plastic bags and labeled. Shell was discarded in the field. A 30 x 30 cm flotation sample was removed from the southwest quadrant of excavation unit 3 and taken to SEAC for processing. Plan-view maps of recorded features and variations in the soil matrix were drawn for each unit. A profile map was drawn for at least one wall of each unit. A photographic record was kept of all identified features and unit profiles. To ensure consistency, standardized task-specific forms (for example, excavation-level forms, field specimen forms, photo logs) were used to record data in the field. Forms electronically generated in the field were later printed on acid-free archival-quality paper.

Laboratory Procedures, Analysis, and Curation

All recovered artifacts were returned to SEAC for cleaning and cataloging. Most were cleaned by hand-brushing with water before being air-dried. Metal artifacts were carefully dry-brushed. All flotation samples were processed in-house using a Flote-Tech

Model A flotation tank to separate the heavy fraction from the light fraction.

Of the artifacts to be curated, those that were stable and large enough had the park acronym (CUIS) and catalog number written on them in indelible ink in an inconspicuous

38

Figure 6. Screening soil from an excavation unit.

area without decoration. Small, delicate materials were put in a bag or vial, with the information recorded directly on the container. The information for each artifact was also recorded on an acid-free paper tag placed with the artifact in its respective container.

Classification and cataloging of the artifacts followed the guidelines set forth in the Museum Handbook, Museum Records, Part II (NPS 2000) and in SEAC's 1990

Cataloging Manual for Archaeological Objects. The artifact analysis and computer data entry were guided by the Southeast Archaeological Catalog System (SACS). The database was converted into the Automated National Catalog System (ANCS+) and entered into the ReDiscovery curation program.

Following basic cataloguing procedure, the materials were sorted into three categories: mineral, vegetal, and animal. The mineral category included materials

39 manufactured from stone, clay, and glass. The vegetal and animal categories included floral and faunal remains, respectively. The materials were further divided into several categories based on material type, including bone, ceramic, glass, metal, shell, soil, stone, wood, other mineral materials, and other plant materials. Each artifact was also classified according to function. The categories, adapted from the revised version of Chenhall’s

(1988) system for classifying human-made objects, included building components, tools and equipment, masonry and stoneworking materials, personal gear, and food-processing items.

The prehistoric pottery analysis was conducted using established methods for pottery classification (Rice 1987; Shepard 1965). The pottery analysis identified patterns of technical and formal variability. Technical variability results from environmental factors, such as the types of locally available clays that affect paste composition and includes the identification of paste compositions. Formal variability results from cultural factors that include the preference of decoration and surface treatment, vessel form, and rim and lip treatments and forms. The typological classification and descriptions of vessel variability were based on formally recognized types and vessel forms established for coastal Georgia, as well as greater Georgia, South Carolina, and Florida (Espenshade et al. 1994; William and Thompson 1999).

Shell and Artifact Distribution

Using the program Surfer, I created a detailed map of the site using shell density as a Z value. Shell density was recorded on the shovel test forms for each of the 183 shovel test excavations. Using this data, I created a map in Surfer to show the location of shell deposits. Artifact density was also mapped using Surfer with the Z value representing the number of artifacts present in each shovel test excavation. Combined

40 with the GPS data, the shell and artifact distribution maps provided a comprehensive illustration of subsurface features. This data was then incorporated into GIS to create a site map of the entire project area.

Intersite Comparison

The artifact assemblages from Brickhill Bluff are compared to the assemblages from nine Woodland sites on the Georgia and South Carolina coast used by Espenshade to formulate his midden typology. For each assemblage, I examine the raw and relative frequency of five variables used by Espenshade et al. (1994). The variables include shell tools, bone tools, ideotechnic items, lithics, and pottery. Additionally, I examine the relative frequency of vertebrate faunal remains and shell among sites. Finally, I record the diversity of pottery, vertebrate faunal remains, and lithic artifacts as well as the presence or absence of structural features and human burials. This analysis allowed me to investigate Espenshade’s model for shell midden development by completing a detailed comparison of site assemblages. By examining the frequency and diversity within and between collections, I was able to demonstrate artifact homogeneity and heterogeneity, ultimately allowing a review of the expectations defined for each midden type. This comparison demonstrates the significance of Brickhill Bluff. Using this comparative methodology, along with data from mapping and excavations, a clearer determination can be made regarding the use of Brickhill Bluff and its importance among the archaeological sites along the South Atlantic Bight.

41 CHAPTER 5: BRICKHILL BLUFF MATERIAL CULTURE

RESULTS AND ANALYSIS

Testing at Brickhill Bluff resulted in the recovery of 1,817 individual artifacts and

49 small bags of cultural remains for a total of 1,899 cataloged objects. The enumerated material remains include Native American pottery, fired clay, a chert core, a chert flake, faunal remains, shell tools, historic ceramics, brick fragments, cut nails, tabby fragments, a gunflint, a buckle, metal fragments, and glass fragments. Bags of uncounted artifact recorded by weight include charcoal and diminutive shell fragments. The entire collection reflects both prehistoric and historic use of the Brickhill Bluff site spanning some 2,500 years of human occupation. For the purpose of this thesis, however, I will focus primary on artifacts attributed to the Woodland occupation.

Datable Evidence

The artifacts from Brickhill Bluff have associated dates ranging from 800 B.C. to

A.D. 1870. The datable artifacts consist primarily of Native American pottery and historic ceramics. Although radiocarbon samples were collected, the pottery analysis provides an adequate chronology of human occupation at the site. One radiocarbon sample produced inconclusive results. Additional radiocarbon samples are available for future analysis if funding becomes available.

42 Prehistoric and Protohistoric Artifacts

Native American Pottery

In using pottery to date the site, we identified several significant periods of human activity at Brickhill Bluff. Native American pottery found during the 2007 excavations ranges from the Early Woodland period (1000–500 B.C.) through the late seventeenth century (circa A.D. 1680) (Table 2). The assemblage includes several types of Deptford and St. Johns wares, one St. Catherines Cord Marked sherd, and one Swift Creek

Complicated Stamped sherd. A small number of St. Marys Cord Marked and San Pedro wares were likely used after European contact. Another 105 unclassified vessel fragments could not be assigned a specific pottery type due to their level of conservation, diminutive size, and/or lack of diagnostic features.

Table 2. Native American pottery.

Ceramic Type Count Count % Weight (g) Weight % Date Saint Johns Plain 7 3.7 39.8 4.1 Woodland to Early Mississippian Saint Johns Ware 33 17.3 47 4.9 Woodland to Early Mississippian Saint Johns Simple Stamped 1 0.5 16.7 1.7 Woodland to Early Mississippian Saint Johns Checked Stamped 5 2.6 18.9 2.0 Woodland to Early Mississippian Deptford Ware 2 1.0 6 0.6 Middle Woodland Deptford Cord Marked 12 6.3 61.1 6.4 Middle Woodland Deptford Check Stamped 10 5.2 92.8 9.7 Middle Woodland Deptford Simple Stamped 1 0.5 3.8 0.4 Middle Woodland Deptford Cross Stamped 2 1.0 20.8 2.2 Early Woodland St. Catherines Cord Marked 1 0.5 15.5 1.6 Late Woodland to Early Mississippian Saint Marys Cord Marked 7 3.7 66 6.9 Middle to Late Mississippian San Pedro Plain 4 2.1 90.6 9.4 16th to early 17th Century Swift Creek Complicated Stamped 1 0.5 1.8 0.2 Middle Woodland Untyped 105 55.0 479.4 49.9 Unknown

Total 191 100 960.2 100

Several different kinds of Deptford ware were recovered at Brickhill Bluff.

Deptford wares indicate an Early and Middle Woodland component at Brickhill Bluff. I recovered 2 plain Deptford sherds and 25 diagnostic Deptford sherds representing 4 ceramic types. As a group, the Deptford assemblage makes up 13 percent of the Native

43 American pottery assemblage. Of the four diagnostic Deptford types, Deptford Check

Stamped (10) and Deptford Cord Marked (12) were by far the most common (11.5 %).

Two Deptford Cross Stamped and a single Deptford Simple Stamped sherd comprise the remaining 1.5 percent of the Deptford assemblage.

Deptford pottery is tempered with fine to medium size sand or grit. Deptford

Check Stamped pottery is recognized by the pronounced check stamping motif covering the entire exterior of the vessel. The check design can be squares, rectangles, rhomboids, or triangles and are made using a wooden paddle (William and Thompson 1999). The overwhelming majority of check stamped sherds from Brickhill Bluff contained simple squares stamped on sand-tempered clay. The exterior of the sherds ranged from orange buff to dark gray and black in color, with black and dark gray being the most prominent.

The interior of the sherds are gritty and have visible marks from smoothing. Deptford rims are commonly straight or slightly flared. The single checked stamped rim sherd from

Brickhill Bluff had a straight rim.

Deptford Cord Marked ceramics have fine to medium grit temper and exteriors that exhibit cord marked designs. These designs are usually located just below the rim.

Like Deptford Check Marked sherds, the sherd exteriors are orange buff, or dark gray to black in color with visible marks of smoothing on a gritty interior. The sherds recovered from Brickhill Bluff contain mostly a fine grit temper with one sherd having a slightly heavier grit temper than the others. These sherds, like the Deptford Check Stamped sherds, were dark gray to black in color. Deptford Cross Stamped sherds are related to

Refuge Simple Stamped and consists of 30-degree angle simple stamping on grit- tempered pottery. The beach surface collection produced two Deptford Cross Stamped

44 body sherds. The single Deptford Simple Stamped body sherd consists of parallel linear grooves stamped on sand-tempered pottery color.

At Brickhill Bluff, I recovered a number of St. Johns sherds representing St. Johns

Plain, St. Johns Check Stamped, and St. Johns Simple Stamped varieties. Disregarding the 33 unspecified St. Johns sherds too eroded to type with indeterminate decorative techniques, St. Johns Plain (n=7; 3.7 %) is the most common type in the assemblage, followed by St. Johns Check Stamped (n=5; 2.6 %) and St. Johns Simple Stamped (n=1;

0.5 %).

As previously mentioned in Chapter 2, St. Johns ceramics have a long sequence spanning the Early Woodland to the early historic period. St. Johns ceramics are tempered with diatomaceous earth consisting primarily of the skeletons of microscopic sponge spicules giving them a chalky feel. The St. Johns Check Stamped ceramics recovered at Brickhill Bluff consists of a check stamp motif made with a wooden paddle on this specific temper. The simple stamped designs contain broad parallel or slightly crossed decoration.

One Swift Creek Complicated Stamped Sherd was recovered from Brickhill

Bluff. This sherd represents a Middle to Late Woodland period of occupation and is contemporaneous with Deptford and St. Johns I. The stamping consists of a curvilinear design stamped with a carved wooden paddle on grit tempered clay.

One St. Catherines Cord Marked sherd was recovered from all the excavations and thus represents only 0.5 percent of the total assemblage. St. Catherines Cord Marked is tentatively considered to be indicative of a Late Woodland (A.D. 500–1150) occupation at Brickhill Bluff (Williams and Thompson 1999); however, it likely persisted as a minor pottery type into the early Mississippian period.

45 Seven St. Marys Cord Marked vessel fragments were recovered from the Brickhill

Bluff excavations. This type tied with St. Johns Plain as the third most common diagnostic pottery type in the assemblage. St. Marys Cord Marked pottery dates from

A.D. 1100 and continued until the sixteenth century, when San Pedro pottery became more common (Ashley and Rolland 2002).

Three San Pedro Plain sherds and one San Pedro pottery fragment with indeterminate surface treatment were recovered from the 2007 excavations. San Pedro ware, which dates to the late sixteenth and early seventeenth century Spanish mission period (Ashley and Rolland 2002), possibly represents the last Native American use of the Brickhill Bluff site.

A total of 105 Native American vessel fragments could not be typed. By both count and weight, they represent 55 percent of the entire pottery assemblage. The surface treatment for most of the untyped fragments was either indeterminate or plain. A few fragments exhibited one of the following surface treatments: check stamping, complicated stamping, cord impressing, or fabric impressing. Still, these sherds were too diminutive to accurately type. The untyped fragments exhibited a variety of tempering agents, though the most common tempers were grit and sand. The tempering agents and surface treatments of the untyped fragments were comparable to the typed sherds from the site, with most displaying a grit or sand temper.

Shell and Lithic Tools

Four shell tools and three worked shells were recovered during the Brickhill Bluff excavations. All the shells in this category were manufactured from whelk (Buccinidae).

One shell tool was likely used as a hammer. The use of the other three tools could not be determined. Native Americans in the region used whelks for a variety of purposes. They

46 were the most commonly used shells for creating digging, pounding, or puncturing instruments (Figure 7). The shells were also used to make shell beads (Marquardt 1992).

A number of the shell tools from the site are incomplete, which is a hindrance to accurate identification. A future microscopic analysis of the shell tool assemblage may shed light on the function of this class of artifacts.

The lithic artifacts included a chert core, a non-cortical chert flake, and four bags of manuports. Manuports are unmodified stones carried to a site from another location by human intervention. Cumberland Island contains no natural stone deposits. It is therefore assumed that all stone we encountered during the excavations were brought from other locations. The prehistoric or historic purpose of these manuports is indeterminate.

Figure7. Shell Tools

47 Historic Artifacts

A brief description of the historic artifacts found at the site gives context to the reuse of Brickhill Bluff by numerous cultures. Historic American/European ceramics from the excavations at Brickhill Bluff included a fragment of a Spanish olive jar, tin- enameled ware, creamware, and pearlware.

A single olive jar fragment from the Brickhill Bluff site provides evidence of an early Spanish presence at the site. The earliest evidence for a likely English occupation at the site is a single sherd of tin-enameled earthenware. Identified as delftware, this artifact is an early English ceramic type produced between 1749 and 1800. Creamware is the most common European ceramic type from the Brickhill Bluff. Of the nine creamware sherds recovered, one had a painted annular design; the other eight were undecorated.

The pearlware from Brickhill Bluff includes one blue hand-painted sherd and one green shell-edged sherd. Originally known as China glaze, pearlware was a British attempt to compete with Chinese export porcelain. Produced from 1779 to 1840, pearlware was dropped from general production when it was replaced by whiteware. Pearlware was developed from creamware by increasing the flint content in the paste and adding a small amount of cobalt to the glaze. As a result, pearlware is lighter in color and has a bluish tint in the lead glaze. This tint is most noticeable where the glaze pooled on the vessel before firing (Miller 1991:5).

Only a small number of metal artifacts were recovered and these consist primarily of building components. Two spikes collected from the beach may have been deposited there as wreckage and, thus, may or may not have originated from the site. The nails, which were recovered from the shovel tests and excavation units, are most likely authentic to the site.

48 Because the nails are all machine cut, they were likely used sometime between

1790 and 1890. Machine-cut nails were mechanically cut from sheet metal. This production technique was an American innovation that began around 1790. Before 1790, only hand-wrought nails were available. The new, cheaper machine-cut nails were widely used until the 1890s, when machinery for the production of construction-grade wire nails was perfected (Nelson 1968). The few other metal artifacts from the project could not be accurately dated. These included the spikes from the beach, a metal buckle, and inde- terminate iron and lead fragments.

The two glass fragments from Brickhill Bluff could not be thoroughly typed and classified because of their diminutive size. Analysis was also limited by the small count.

The one blue-green glass vessel fragment (1 g) was of insufficient size to determine vessel form. The only other glass artifact was a piece of slag (1.4 g).

One of the two personal artifacts from Brickhill Bluff was a metal buckle. The iron buckle found at Brickhill Bluff is most likely a harness buckle or a modest shoe buckle. The other personal artifact was a small porcelain tube of indeterminate function.

It is likely a piece of pipestem from the eighteenth or nineteenth century, even though tobacco pipes from this time period were more commonly made from kaolinite clay. Sold in large quantities, tobacco-smoking pipes were commonplace from the late seventeenth until the beginning of the twentieth century (Noël Hume 1969:302–303).

A gunflint, or more technically a gun spall, found at the site originated from the

European practice of manufacturing firearm components from the flint stone deposits found in European chalk beds. Gunflints were mass-produced by cutting previously prepared, uniform-sized blades of flint. Gun spalls, on the other hand, were made by striking them from an irregular nodule of chert. Chert is similar to flint but originates in

49 limestone rather than chalk. Therefore, the more variable shape of the gun spall is largely a result of the way it was struck off the nodule. Other historic artifacts found at the site include building materials, such as brick fragments, concrete fragments, and tabby mortar. The excavations also uncovered iron soil concretions which occur naturally and are common in the types of soils found at Brickhill Bluff.

Artifacts came almost exclusively from a single stratigraphic layer composed of oyster shell and Woodland ceramics. At this time, it is not possible to determine the precise sequence of habitation during the 2000 year Woodland period because we see no discernable changes in pottery types or the introduction of significant features within this single stratigraphic level. Additional radiocarbon dates may help solve this problem.

Currently the large shell midden level is thought to be an accumulation taking place primarily between 1000 B.C. and 1000 A.D. This level contains a consistent range of artifacts dating from early, middle and late woodland periods. A minority of

Mississippian wares indicate a transition period around 1000 A.D. However, based on the scarcity off Mississippian ceramics, limited use of the site is assumed until the protohistoric and historic periods.

50 CHAPTER 6: DESCRIPTION OF WOODLAND ASSEMBLAGES AT

COMPARATIVE SITES

I chose nine sites to sample and compare with the Brickhill Bluff assemblage

(Table 3). Espenshade et al. (1994) includes each of these sites in his prehistoric midden study. While Espenshade ‘s studyincludes additional sites, I chose these nine sites as a representative sample. In that study, the sites are assigned to one of four specific

Woodland midden categories. These categories are based primarily on the artifact assemblage and a few additional attributes such as the presence or absence of structural features and burials.

Of the nine sites, those classified as Multi-Family Residential Base Middens are

Bass Pond Dam (38CH124), Stafford North (9CM 10; CUIS-10), and Table Point

(9CM12; CUIS-21). Minim Island (38GE46) and Block B of Site 38BU2 are Single-

Family Shell sites. Single-Family Limited Shell Middens include Site 38CH779 and Fish

Haul (38BU805). Oystering Stations consist of Site 38BU833 and Blocks A, C, D, and E of Site 38BU2 (Espenshade et al. 1994:177-180). Geographically, the chosen sites are within the Atlantic Coastal Plain of Georgia and South Carolina. The sites encompass an area that extends from Cumberland Island, Georgia, north to Georgetown, South

Carolina.

The sites sampled span the Woodland Period (1000 B.C. to A.D. 1000) with the addition of relevant Late Archaic / Transition materials. The earliest of these sites contain

Early Deptford and Thom’s Creek wares. Site 38CH779 and Fish Haul contain an Early

51 Woodland period of occupation. Sites 38BU2 and Minim Island have Early and Middle

Woodland components. Bass Pond, Stafford North, Table Point, and Site 38BU833 have evidence of a continual cultural occupation spanning the entire Woodland Period.

Table 3. Sites sampled.

Site Types Site Name Bass Pond Dam (38CH124) Multi-Family Residential Base Stafford North (9CM10) Table Point (9CM12) Ring and House Minim Island (38GE46) Deptford and Thom’s Single-Family Shell Midden Creek Block B of 38BU2 38CH779 Single-Family Limited Shell Site Fish Haul (38BU805) 38BU833 Oystering Stations Blocks A, C, D, and E of 38BU2

It is worth noting that the Thom’s Creek type ceramics begin in the Late Archaic

(2000 B.C. – 1000 B.C.) and terminate around 500 B.C. Thom’s Creek wares, like

Refuge wares, are considered a transitional ceramic type; however, the Thom’s Creek variety predates both Refuge and Deptford (Trinkley 1980:17). The Thom’s Creek pottery found at 38BU2, Bass Pond, and at 38CH779 likely indicate a Late Archaic and

Early Woodland transition period of occupation along the South Carolina coast.

52 Multiple-Family Residential Base

The Bass Pond Dam Site (38CH124)

Located on Kiawah Island, South Carolina, the Bass Pond Dam site is situated on a small rise along a salt marsh and saltwater creek. A small portion of the site is visible where it is bisected by a road revealing scattered shell, pottery sherds, and historic ceramics (Michie 1979:24). The site was recorded by John Combes in his 1975 assessment of archaeological sites on Kiawah Island. Combes remarked that the Bass

Pond Dam site is significant for its deeply stratified midden and the presence of

Awendaw pottery (Combes 1975: A15). In 1978, 25 1 x 1 m excavation units were placed across the site. Soil was removed in 15 cm levels and screened through 6.4 mm

(1/4 inch) wire mesh with the help of a gasoline powered mechanical sifter (Michie

1979:33).

The excavation revealed a homogeneous midden approximately 75 cm thick. The lower levels of this midden contained Awendaw and Thom’s Creek pottery which originated around 2000 B.C. and terminated in the early Woodland period. The upper levels of the midden contained Cape Fear pottery, which dates from A.D. 200 to 700

(Michie 1979:34). The pottery types at the Bass Pond Dam site suggests that the site was abandoned at the beginning of the Early Woodland period and reoccupied 1,200 years later in the Middle Woodland period. Unfortunately, no radiocarbon dates are available

(Michie 1979:39-40).

The artifact assemblage includes ideotechnic items (n=7), lithics (n=218), pottery

(n=2179), human burials, and structural features. Pottery at the site contributes a little over 90% of the measured artifacts in the assemblage. Lithic items account for 9% with all other items totaling less than 1%.

53 No shell or bone tools were found at the site. As would be expected at a Multi-

Family Residential base, several ideotechnic items and lithic artifacts were recovered.

These items consist of seven bone pins and 218 lithic artifacts. Lithics include of flakes, bifacially flaked projectile points, flake tools, a hammerstone, and fragments of steatite

(Michie 1979:53, 63). Pottery was by far the most abundant artifact class. The 2,179 pottery sherds recovered from the Bass Pond Dam site include Awendaw, Thom’s Creek, and Cape Fear type wares. The earliest inhabitants at the site utilized Awendaw Plain,

Awendaw Finger-pinched, Awendaw Nail-punctate, Awendaw Nail-gouged, Thom’s

Creek Periwinkle-puncate, Thom’s Creek Stick-punctate, Thom’s Creek Reed Punctate,

Thom’s Creek Periwinkle-Impressed and Thom’s Creek Simple Stamped. Later inhabitants developed Cape Fear pottery which slowly replaced the Awendaw/Thom’s

Creek pottery (Michie 1979:44).

Among the vertebrate faunal material, six species were identified. Vertebrate species include white-tailed deer, raccoon, rabbit, diamond-back terrapin, drumfish, and sea catfish. Also included in the assemblage are unidentified fish and bird species

(Michie 1979:76-77). Eastern Oysters are the most common invertebrate species. Oyster shells were found in all levels of the shell midden. The shell composition for the site is on average 79% oyster with 21% minority species such as periwinkle, quahog, whelk, anglewing, moon snail, and razor clam (Michie 1979:72).

Archaeologists uncovered eight disarticulated human skeletal fragments at Bass

Pond Dam (Michie 1979:76-77). Structural features at the site include a living floor and a single posthole (Michie 1979:39-40).

54 Stafford North (9CM10)

The Stafford North site stretches approximately 398 m along a tidal marsh on the western side of Cumberland Island, Georgia. The site is bounded to the north and south by small freshwater streams and is composed entirely of scattered piles of shell (Milanich

1971:29). In July and August 1970, Milanich excavated the Stafford North site to answer questions regarding the relationship between the Deptford Phase and other coastal traditional phases as well as examine the settlement patterns, subsistence strategies and other cultural processes of the Deptford Phase (Milanich 1971:25).

At Stafford North, Milanich established a north-south and east-west grid. He sampled the site by placing 1.5 x 3 m units along this grid. In total, eight units were excavated at Stafford North. Five units were placed at the northern end of the site where a living floor was encountered. Three additional units were placed along the north-south axis. Excavation units were excavated by natural stratigraphy and soil was screen through wire mesh screens (Milanich 1971:31-32). Based on the ceramic assemblage, the Stafford

North site dates from around 500 B.C. to A.D. 200 (Milanich 1994:112–114; Tesar

2011:1). No radiocarbon dates are available for this site.

The artifact assemblage from Stafford North is comprised exclusively of pottery

(n=427) with limited shell and faunal remains. Structural features at the site include a living floor and postholes possibly belonging to a thatched chickee-like structure

(Milanich 1971:45-46).

The pottery sherds recovered from the site included Deptford Cord Marked,

Deptford Check Stamped, Deptford Linear Check, Deptford Complicated Stamped,

Deptford Simple Stamped, Deptford Plain, and St. Johns Plain sherds (Milanich 1971:29-

42). Shell at the site was composed primarily of eastern oyster (90% to 95%). Other

55 univalves and bivalves present included clam, whelk, ribbed mussel, periwinkle, and giant Atlantic cockle (Milanich 1971:35).

The site did not contain human burials, but excavations did reveal a concave oval depression with a living floor and three postholes. This structure is somewhat similar to the concave floor found at the Table Point site (Milanich 1971:45).

Table Point (9CM12) Ring and House

Table Point, like Stafford North, is located on Cumberland Island. This site is also composed primarily of scattered piles of shell. However, this site also contains a shell and dirt ring with a large oval midden in the center. The site is approximately 137 m north to south and is adjacent to a tidal marsh. The most eastern edge of the site extends 100 m from the high tide line (Milanich 1971:46). Piles of shell measured 0.15 to 0.3 m high and 6 x 9 m in diameter (Milanich 1971:26). Within the complex, the shell ring is 67 m in diameter and 6 to 8 m wide with a depth of 0.2 to 0.6 m. Two gaps in the shell ring were attributed to shell mining for tabby construction and road paving in the nineteenth and twentieth centuries. Within the shell ring is another ring measuring 46 m in diameter and

18 m wide (Milanich 1971:49). Table Point was excavated by Milanich during the same field season as the Stafford North Site in July and August 1970 (Milanich 1971:25).

Excavation units were either 3 x 3 m, 1.5 x 3 m, or 1.5 x 1.5 m. All unit excavations were organized according to a grid established at the site (Milanich 1971:47).

The site contained two assemblages. The first artifact assemblage comes from a shell ring measuring 67 m in diameter and between 6 m and 7.5 m wide (Milanich

1971:49). The second artifact assemblage is from a Deptford period house located on the northwest edge of the shell ring. This oval-shaped house measures 10 m by 7m and is similar to the oval living floor area found at the Stafford North site (Milanich 1971:62-

56 67). The artifact assemblage from the ring feature contained exclusively pottery (n=622) and a small number of shell tools (n=3). Pottery made up 99.52% of the measured artifacts from the ring feature.

The assemblage from the Deptford house included shell tools (n=4), a bone tool fragment (n=1), a single lithic (n=1), pottery (n=559), shell, and faunal materials. The pottery from the Deptford House feature accounted for 98.93% of the measured artifact assemblage with all other artifacts accounting for less than 1%. Structural features were encountered in both locations. Human burials were absent.

Taken separately, the Table Point ring and Deptford house excavations revealed two distinct assemblages. The Table Point ring assemblage included three Busycon perversum shell picks. These three shell tools all had a single hafting hole. The ring contained 622 sherds including Deptford Simple Stamped, Deptford Cross Simple

Stamped, Deptford Complicated Stamped, St. Johns Plain, and a number of unidentified types. The vertebrate faunal material represented seven identified species, including tiger shark, white-tailed deer, sea catfish, diamond-back terrapin, raccoon, pocket gopher, and red drum (Milanich 1971:51-52). Shell at the site was composed primarily of 90% to

95% oyster (Milanich 1971:49).

Structural features at the Table Point ring included the ring itself along with a living floor. The living floor contained several features including a circular area of charcoal and burnt sand representing a possible hearth and postholes (Milanich 1971:53-

56). The Table Point house artifact assemblage included a few shell tool varieties. Four

Busycon perversum picks, a square coquina hone, and two Busycon columella awls were recovered. One broken and polished one pin tip came from the house and a single chert chip was recovered from the house wall trench. The Table Point house assemblage

57 contained 559 sherds. Pottery types included Deptford Simple Stamped, Deptford Check

Stamped, Deptford Cross Simple Stamped, Deptford Complicated Stamped, and St. Johns

Plain (Milanich 1971:70-71).

Faunal species identified from the house excavation included raccoon, cf. West

Indian seal, whale or porpoise, white-tailed deer, box turtle, salt marsh terrapin, chicken turtle, gopher tortoise, sea turtle, requiem shark, hammerhead shark, sting ray, eagle ray, sea catfish, red drum, channel bass, and sheepshead (Milanich 1971:73-74). Like the

Table Point ring, the shell midden areas at the Deptford house were comprised of 90% to

95% eastern oyster.

Structural features included wall trenches with post holes and anchored with a mixture of shell and dirt fill. A small trench inside the house may have separated the cooking and sleeping areas. A fire pit was found near the center of the house. A large posthole nearby may have helped support the roof (Milanich 1971-66-69).

Single-Family Shell Midden

Minim Island Site (38GE46) Deptford and Thom’s Creek

The Minim Island Site (38GE46) is located on Minim Island, Georgetown

County, South Carolina. Situated north of the North Santee River along the Intracoastal

Waterway, the site is in an estuarine ecosystem with tidal ranges of 2 to 4 m (Espenshade and Brockington 1989:59). The site was first recorded in 1974, although the site was known locally for many years. It was nominated for inclusion to the National Register of

Historic Places in 1981 and listed in 1982. The excavations at Minim Island site were sponsored by the United States Corps of Engineers to mitigate the adverse effects of dredging and operating the Intracoastal Waterway (Espenshade and Brockington 1989:1).

58 In 1988, a 3 x 9 m block was excavated at the site to a depth of 2.2 m

(Espenshade and Brockington 1989:17). Deptford, Thom’s Creek, and Refuge wares composed a majority of the ceramic assemblage. Radiocarbon dates from the site were consistent with a Deptford and Thom’s Creek occupation. Radiocarbon dates from the

Thom’s Creek/Refuge stratum provided a calendar data of 1170 B.C. The radiocarbon dates for the Deptford period of occupation fall between 679 B.C. and A.D. 220

(Espenshade and Brockington 1989:125-126).

The assemblage for the Thom’s Creek component of the Minim Island site included shell tools (n=13), bone tools (n=17), lithics (n=96), pottery (n=2833), faunal remains and structural features. The measured assemblage is made up of 95.74% pottery and 3.24% lithics. The remaining artifact comprise less than 1% each of the assemblage count. The artifacts recovered from the Deptford shell midden at Minim Island included shell tools (n=78), bone tools (n=31), ideotechnic items (n=6), lithics (n=61), pottery

(n=2974), and faunal remains (8150.9 g). Over 94% of the collection is made up of pottery. Lithics account for 1.94%, and 2.48% are shell tools. Bone tools and ideotechnic items are each less than 1%. Structural features and human burials are present.

The Thom’s Creek assemblage from the Minim Island site contained 13 shell tools. Shell tool types included scoop/dippers, hammers and picks. Awls, gouges, pins and a spatula comprised the bone tools found in this assemblage. While there were no ideotechnic items recovered in the Thom’s Creek levels, Minim Island did contain 96 lithic artifacts. Lithic tool types contained both expedient and formal tools: bifaces, flake tools, unifaces, a projectile point, a gorget, and abraders. The site also produced a large number of flakes (Espenshade and Brockington 1989:170-198).

59 Excavations uncovered 2,959 pottery sherds. Thom’s Creek Incised, Thom’s

Creek Plain, Thom’s Creek Shell Scraped, Thom’s Creek Separated Punctate, Thom’s

Creek Jab and Drag Punctate, Thom’s Creek Finger Grooved, and Thom’s Creek Cockle

Impressed composed the ceramic types in this assemblage. The Thom’s Creek provenance also contained a number of Refuge Plain, Refuge Dentate Stamped, Refuge

Incised, and Refuge Puncated sherds (Espenshade and Brockington 1989:151).

The total vertebrate faunal material included 18 identified species. Maximum meat weight data indicates a reliance on vertebrate species as part of the Thom’s Creek diet (Esenshade and Brockington 1989:199). Opossum, raccoon, and white-tailed deer comprise the identified mammals at the site. Turtles included cooter, diamondback terrapin, and spiny softshell. Wide varieties of fish were recovered. Included in this list are sturgeon, gar, American eel, gafftopsail catfish, freshwater catfish, largemouth bass,

Atlantic croaker, drums, and stringray (Espenshade and Brockington 1989:203). The shell composition for the Thom’s Creek cultural level was 88% oyster with 12% minority species such as mussels, cockles, and quahogs. Several structural features were found during the excavation. They include probable posts and pit features (Espenshade and

Brockington 1989:105).

The Deptford shell midden contained a moderate number of shell tools (n=78) including adze, hammers, picks, and chisels. Thirty-one bone tools were recovered from the Deptford levels at the site. Bone tools included gouges, awls, pins, and a socketed antler point. The Deptford cultural layers at the Minim site revealed six beads. Five of these beads were made of bone and one was made of shell. These beads are classified as ideotechnic items. The excavations also revealed 61 lithic artifacts from the Deptford shell midden. Lithic artifacts included flakes, bifaces, projectile points, a uniface, a core

60 fragment, worked cobbles, a cobble grinding stone, and an abrader (Espenshade and

Brockington 1989: 170-198).

The Deptford assemblage contained 2,974 sherds. The major pottery types included Deptford Red Painted, Deptford Plain, Deptford Incised / Punctated, Deptford

Check Stamped, Deptford Linear Check Stamped, Oemler Complicated Stamped, and

Deptford Simple Stamped (Espenshade and Brockington 1989: 151). The total vertebrate faunal material consisted of 30 identified species. Mammal species included rabbit, raccoon, and white-tailed deer. Reptile and fish species include spotted turtle, diamondback terrapin, spiny softshell, cooter, sturgeon, gar, American eel, shad, gafftopsail catfish, hardhead catfish, Atlantic croaker, star drum, mullet, perch, flounder, bowfin, ladyfish, American silver perch, weakfish, black drum, stringray, and cownose ray (Espenshade and Brockington 1989: 206-210). The shell composition for the

Deptford cultural level was 95% oyster with 5% minority species such as whelk, clams, mussels, cockles, and quahogs.

The Deptford level contained a primary burial and secondary/disturbed burials.

The primary flexed burial of an adult male between 30 to 35 years of age was found with six Deptford check stamped sherds and had damage to the cranium. The damaged cranium appeared to be the result of an intrusive prehistoric refuse pit into the burial

(Espenshade and Brockington 1989:120). The Deptford level contained numerous features including pit and post features and a possible house basin (Espenshade and

Brockington 1989:105).

38BU2, Block B

Site 38BU2 is located on Spring Island, Beaufort County, South Carolina. The island is within the Port Royal Sound estuary and is considered a smaller sea island. The

61 initial site survey was conducted in 1990 by Chicora Foundation with additional work by

Christopher Espenshade of Brockington and Associates in 1993. The 1993 excavations mitigated the potential adverse effects of a new golf course being constructed on the island. Limited documentation from the 1990 Chicora excavation required Espenshade to resurvey the site to locate the midden loci. The entire site was subject to shovel testing at

20 m intervals. The shovel test survey identified 20 midden loci along the marsh and inland from a former stream bed. Close interval shovel testing and block excavation sampled five shell middens. Machine stripping followed hand excavations as a method to identify subsurface features. A bulldozer was used to remove 30 cm of soil from portions of the site. The area was then shovel scrapped to identify potential structures associated with the midden (Espenshade et al. 1994:34-38). Cultural material remains indicated

Deptford phase middens (500 B.C. – A.D. 200) and a Stallings/Thom’s Creek phase midden (2000 B.C. – 500 B.C.). No radiocarbon dates are available for this site

(Espenshade et al. 1994:163-164).

The artifact assemblage for Block B included a single shell tool (n=1), a bone tool

(n=1), ideotechnic items (n-12), lithics (n-158), and a large amount of pottery (n=2820).

Pottery represents 94.25% with the remaining 5.28% of artifacts being lithics and less than 1% each of bone tools, shell tools, and ideotechnic items. Other cultural materials consisted of a diverse faunal assemblage and a single human burial. Block B was excavated by 10 cm arbitrary levels. Block B contained 8 total levels with a dense oyster shell midden identified in levels 2 through 5 (Espenshade et al. 1994: 65-70).

Block B produced one shell tool, identified as a hammer. A single bone awl from

Unit 11 comprises the entirety of the bone tool assemblage from the site. Twelve bone pin fragments found in units 12 (n=11) and 13 (n=1) represent at least seven pins and are

62 believed to have an ideotechnic function due to their proximity to a human burial. The general lack of bone tools found in Block B suggests that bone tools were used infrequently for utilitarian purposes at the site (Espenshade et al. 1994:168).

Block B produced 158 lithic artifacts. Lithic artifacts included a drill, one hammerstone, a single blade, four bifaces, one core, an adze fragment, 19 projectile points (all Savannah River Stemmed), and 130 flakes (Espenshade et al. 1994:163-164).

Block B also contained 2,820 sherds. The major pottery types were Stallings Jab and

Drag Punctate, Stallings Plain, Thom’s Creek Jab and Drag, Thom’s Creek Separate

Reed Punctate, and Deptford Check Stamped.

The total vertebrate faunal material consists of 17 identified species. White-tailed deer was the most abundant mammal based on MNI and biomass percentage. Additional mammals included opossum, raccoon, and a single mole. The only identifiable bird was a wild turkey. Reptiles included alligator, snake, and four turtle species. Catfish comprised half of the fish remains found in Block B. Black drum and flounder are also represented among the fish species. The shell composition at the site was 98% oyster with 2% minority species such as periwinkle (Espenshade et al. 1994: 174).

Block B contained human remains. Human bone recovered from Block B consists of fragments from a single mandible and maxilla scattered in units 10 and 13. These units contained no discernable pit feature suggesting a secondary burial context. Further examination of the remains revealed that the skeletal elements are from a 20 – to – 25 – year old individual. The site contains no structural features (Espenshade et al.1994:65-

70).

63 Single-Family Limited Shell Site

38CH779

Site 38CH779 is located on Sol Legare Island near Charleston, South Carolina. It was discovered when a developer cut a road through the site to build a subdivision. In

1983, a concerned citizen contacted archaeologist Michael Trinkley about the construction and looting at the site (Trinkley 1984:1). After contacting the developer,

Trinkley gained permission to excavate on the lots still owned by the developer. He proceeded to excavate the lots which had the least amount of disturbance from looters and development. Four 1.5 x 1.5 m units were excavated and a controlled surface collection of the site was made (Trinkley 1984:3-4).

Ceramics recovered from the site indicate a temporal assignment at the end of the

Thom’s Creek phase during the Early Woodland Period. Thom’s Creek Finger Smoothed and Thom’s Creek Simple Stamped sherds suggest a date around 900 B.C. A few early variety Deptford specimens are also present confirming that 38CH779 was occupied during the Early Woodland Period. No radiocarbon dates are available (Trinkley

1984:22).

The artifact assemblage included lithics (n=5), pottery (n=2533), structural features, and human burials (Trinkle 1984:3-4). While the site produced no shell tools, bone tools, or ideotechnic items, a small number of lithic types were recovered. Four primary flakes of local coastal plan chert and a single Morrow Mountain II projectile point make up the total lithic assemblage from the site (Trinkley 1984:28-29). Thom’s

Creek wares made up the majority (95.7%) of the pottery found at the site. Pottery from the Thom’s Creek series included Thom’s Creek Finger Pinched, Thom’s Creek Reed

Punctated, Thom’s Creek Drag and Jab, Thom’s Creek Finger Smoothed, Thom’s Creek

64 Shell Punctate, and Thom’s Creek Simple Stamped. The remainder of the pottery found at site 38CH779 consisted of Deptford Check Stamped and unidentified sherds too small to type (Trinkley 1984:22-25).

The few faunal remains found at 38CH779 provide an idea of the subsistence patterns of the prehistoric inhabitants. Recovered remains were well preserved and included white-tailed deer, mammals, birds, fish, and turtles (Trinkley 1984:33). Shellfish remains were also sparse at this site. Shell is represented primarily by oyster with some other tidal specimens like quahog, whelks, and periwinkle. Shell was found primarily scattered through the plowzone and in small clusters around the site. This shellfish assemblage is not dissimilar to other single family limited shell sites (Trinkley 1984:34).

Human cranial fragments appear to represent a single individual, possibly 25 to

35 years of age. The bones were found scattered in an area disturbed by the construction of a new road (Trinkle 1984:42). A number of postholes arranged in a linear order indicates the presence of some type of aboriginal structure at the site. It is hypothesized that the arrangement of these postholes is similar to structures from the Deptford site near

Savannah, Georgia (Trinkley 1984:18).

Fish Haul Site (38BU805)

Located on Hilton Head Island in Beaufort County, South Carolina, this multicomponent site consists of a Late Archaic Stallings phase, followed by a Woodland occupation, and finally an historic stage. The site was first examined in 1982 and again in

1986. These investigations revealed well-preserved features and a small non-shell

Stallings midden. The 1986 research at the site conducted systematic auger tests at 15 m intervals and block excavations. The auger tests revealed the site consisted of a number of “discrete occupation loci” (Trinkley 1986:5). The block excavations exposed 158 sq m

65 of the site consisting primarily of Stallings phase remains with a light mix of other prehistoric remains (Trinkley 1986:15). Three radiocarbon dates fall within the established temporal boundaries of the Stalling phase: 1170+/- 90 B.C., 1730 +/- 60 B.C., and 1330 +/- 80 B.C. (Trinkley 1986:6). The site is listed on the National Register of

Historic Places.

The assemblage included ideotechnic items (n=2) lithics (n=644), and pottery

(n=3541). Pottery comprised 84.57% of the collection. Lithics represented 15.38% with ideotechnic items amounting to less than 1%. Only two shell artifacts were found at the site. These are heavily eroded oyster and clam shells with a hole drilled in the hinge area presumable so they could be worn. These items are classified as ideotechnic items. The site is absent of shell tools and worked bone (Trinkley 1986: 208). Lithic artifacts include

610 flakes, 21 projectile points, 8 blade fragments, 4 hammerstones, and 1 flake tool

(Trinkley 1986: 183-195). The pottery recovered totals 3541 sherds. Pottery types included Stallings Shell Punctate, Stallings Reed Punctate, Stallings Drag and Jab,

Stallings Incised, Stallings Cord Impressed, Deptford Check Stamped, Deptford Simple

Stamped, Deptford Cord Marked, Thom’s Creek Shell Punctate, Thom’s Creek Simple

Stamped, Refuge Random Punctate, Mount Pleasant Cord Marked, St. Catherines, and

Irene (Trinkley 1986: 158).

The vertebrate faunal collection consisted of more than 4,471 bones (Trinkley

1986: 282). These bones constitute 10 species: white-tailed deer, diamondback terrapin, raccoon, turkey, common loon, sheepshead, rabbit, and mud snake (Trinkley 1986: 296).

Shell composition at the site was primarily oyster (Trinkley 1986: 335). The site contained structural features but no burials.

66 Oystering Stations

38BU833

Site 38BU833 is located on Hilton Head Island, South Carolina, along Skull

Creek. The Chicora Foundation conducted an initial reconnaissance level survey in 1987

(Trinkley 1987). An additional survey in 1992 determined that the site, while impacted by the construction of a condominium complex and parking lot, contained intact midden deposits. The site was recommended eligible for inclusion on the National Register

(Trinkley 1992:1). The 1992 excavation involved 32 shovel tests to delineate the site followed by nine excavation units in areas believed to contain cultural materials. The profiles from the units revealed that a significant portion of the site had eroded into the creek. The midden appeared to be originally about 30 cm with profiles indicating the site consisted of piles rather than a continuous sheet midden. No features or other evidence of occupation were encountered (Trinkley 1992:19). A single radiocarbon date was obtained from an oyster shell. This sample yielded a date of 820 + 60: A.D. (Trinkley 1992: 26-

27).

The artifact assemblage contained pottery (n=423), a single lithic flake (n=1), and faunal remains. Shell tools, bone tools, and ideotechnic items are completely absent at

38BU833. One small non-cortical flake of Coastal plain chert comprised the entire lithic assemblage. Three pottery types were found at 38BU833: Deptford, St. Catherines, and

Savannah. Specific diagnostic pottery include a variety of Deptford wares (simple stamped, dentate stamped, plain, fabric impressed, and check stamped) and St. Catherines wares (cord marked, plain, net impressed, and check stamped). The Savannah wares recovered included plain, cord marked, check stamped, complicated stamped, and fabric impressed (Trinkley 1992:27).

67 Very few faunal remains were recovered. Many of the areas investigated for this project yielded no faunal materials. The bulk of the remains appear to be mammalian species such as white-tailed deer. Only one provenience yielded fish remains. A small number of bird bones were found associated with two small pit features. Excavations at

38BU833 yielded shell deposits composed of 99% oyster (Trinkley 1992:34-35).

38BU2, Blocks A, C, D and E

While Block B of 38BU2 is believed to represent a single-family shell midden, excavation blocks A, C, D, and E of Site 38BU2 appear to represent components of an oystering station. As stated previously, 38BU2 is located on Spring Island in Beaufort

County, South Carolina. The excavation revealed two middens: a Deptford midden

(Block A, C, D, and E) and a Thom’s Creek midden (Block B). Blocks A, C, D, and E yielded shell tools (n=6), bone tools (n=2), lithics (n=531), pottery (n=4,278), and faunal material. Over 88% of the measured assemblage is made up of pottery. Lithics make up

11.02% with shell tools and bone tools representing less than 1% each.

Shell tools associated with the Deptford occupation of the site include shaped columellae, an adze, and hammers. The single bone awl made from split long bones from a large mammal was found in midden Block A along with an incised bone pin

(Espenshade et al. 1994:114-119). This bone pin does not appear to be associated with a burial and therefore is not deemed an ideotechnic item. Lithic artifacts from the site are made up of flakes, a flake took, a uniface, bifaces, an adze, projectile points, a drill, and an abrader (Espenshade et al. 1994:163). In total, the lithic assemblage for blocks A, C,

D, and E contains 531 items, mostly flakes.

Pottery sherds make up a large portion of the artifacts recovered from 38BU2. Out of the 4,278 sherds found, three primary pottery types were recognized: Deptford,

68 Wilmington, and Savannah. Deptford wares recovered from the site include Deptford

Cord Marked, Deptford Check Stamped, Deptford Simple Stamped, Deptford Plain, and

Deptford Fabric Impressed. Wilmington wares include Wilmington Cord Marked,

Wilmington Plain, Wilmington Fabric Impressed, and Wilmington Check Stamped.

Savannah Check Stamped and Savannah Cord Marked were also included in the ceramic assemblage (Espenshade et al. 1994:46).

The faunal assemblage contained a variety of mammals, birds, reptiles and fish.

The Deptford fauna includes species such as opossum, rabbit, squirrel, raccoon, deer, turkey, alligator, turtle, diamondback terrapin turtle, hardhead catfish, gafftopsail catfish, red drum. black drum, flounder, and pufferfish (Espenshade et al. 1994:136-138). Oyster shell made up 99% of the midden with the occasional inclusion of clams (Espenshade et al. 1994: 174). No structural features or human remains were found in this section of

38BU2.

69 CHAPTER 7: COMPARATIVE SITES AND MIDDEN TYPOLOGY MODEL

The research objective of this thesis is to determine the efficacy of Espenshade’s midden typology and its application to understanding Woodland period habitation and social organization along the Georgia coast. This proposition is based on the belief that behaviors associated with the construction of Woodland middens are discernable by examining and measuring a site’s unique artifact assemblage and archaeological features.

A secondary consideration for writing this thesis is to better understand the Brickhill

Bluff site on Cumberland Island, Georgia. What was the purpose of this site and what type of behavior contributed to its construction? The relationships identified and substantiated in this thesis provide archaeologists an alternative to “lumping” all sites into a generic shell midden classification and provides additional clarity to the poorly understood Atlantic Woodland period. This research affords archaeologists the ability to more confidently and cautiously use Espenshade’s midden typology and facilitate site identification both for research and in cultural resource management. Finally, this thesis reports on the findings at the Brickhill Bluff site and contributes to our understanding of

Woodland site structure and function on the South Atlantic Coast.

To determine the efficacy of Espenshade’s midden typology, I first examine the characteristics laid out in his model and compare those with the raw and relative frequencies from our four midden types. Espenshade looks at each characteristic and assigns them a value based on their unique frequency and diversity within the artifact assemblage. The value for the frequency of shell tools, bone tools, ideotechnic items,

70 lithics, and pottery is expressed as either very low, low, moderate to low, moderate, moderate to high, relatively high, or high. These terms are also used to describe the amount of diversity found in pottery, lithics, fauna, and shell in each assemblage. The presence or absence of human remains and structural features is characterized as either very rare, occasional to rare, occasional, relatively common, or common.

I ascribe a percentage value to the first five artifact categories. Within individual assemblages, the percentage of each artifact type works well to describe the frequency of the material recovered. Raw frequencies of artifacts within an assemblage therefore give us a good idea of what type of items are found at the site and how these items compare to each other in the overall makeup of an assemblage. I use raw numbers to show the diversity of faunal remains in grams and by the number of unique species identified. I also record the diversity of pottery and lithic artifact types by documenting the presence of each unique type and totaling them into a raw number. This data provides a method to keep track of diversity within the assemblage in a way that will later be used to identify inter-assemblage differences. A single yes or no marks the presence or absence of human remains and structural features.

Finding a pattern of inter-assemblage variation is harder to quantify than finding a pattern of artifacts within an assemblage. Preservation and sampling factors make a comparison between archaeological sites inherently difficult. Different sites will yield artifacts in different frequencies, and it can be difficult to tell if the frequency variation is a reflection of behavior or external forces. For example, several large block units may yield only a few sherds at one site while a single row of shovel tests may produce the same number of sherds at another site. The number of sherds would not necessary reflect the frequency of pottery at the site due to the divergent testing methods. Additionally,

71 some sites maintain better preservation of bone and shell materials than others depending on environmental conditions. However, the assemblages, taken as a whole, should provide a relatively constant artifact ratio expressed as a percentage, which then provides useful clues to determine behavioral patterns.

Attempting to quantify the differences between very low, low, moderate, high, and very high is also a challenge. These are subjective classifications and Espenshade does not give a numerical value. However, a few things stand out when we look at the data. Taking the artifact percentages available from the sites sampled we can make a determination on what is low, medium, and high by comparing the percentages of artifact types across all of the sites examined. Shell tools, bone tools, and ideotechnic items have frequency values between 0 and 2.48%. The median value is 0.31%. Consequently, I classify 0 – 0.05% as low, 0.12% – 0.57% as moderate, and 0.98% – 2.48% as high.

Lithic artifact frequency ranges from 0 to 14.69% with a median value of 3.24 %. Values between 0 and 0.24% are low, 1.93% and 5.28% are moderate, and 9.06 and 14.69% are high. Pottery frequency ranges between 85.26 % to 100% with a median value of

95.74%. Values between 85.26% and 90.64% are low, 94.25% and 99.52% are moderate, and 99.76% and 100% are high.

All of the sites sampled had very high concentrations of pottery relative to all other artifacts. In most instances, all other artifact types are minor, often composing less than 5% of a site’s total artifact assemblage. Using Espenshade’s model, it should be possible to match the appropriate artifact frequencies and diversities to discern a site’s cultural function. The observed pattern of inter-assemblage variation for the sites included in this study show some inconsistencies with Espenshade’s midden typology.

72 The relative frequency of artifacts (shell tools, bone tools, ideotechnic items, lithics, and pottery) among sites show some interesting results (Table 4). As previously stated, all site assemblages are dominated by pottery. In fact, no site has an assemblage of less than 84% pottery. The Stafford North site contains 100% pottery. Fish Haul has the smallest percentage of pottery at 84.57%. Lithics are variable across sites. Fish Haul has the highest concentration of lithic artifacts at 15.38%. Stafford North and the Table Point

Ring have none. Ideotechnic items do not numerically contribute in a substantial way to the composition of any artifact assemblage. Ideotechnic items from Bass Pond Dam account for 0.29% of the total assemblage from the sites. All other sites either have no ideotechnic items, or the ideotechnic items contribute less than 1% of artifacts recovered.

None of the sites included in this study yield a large number of bone tools. Bone tools within each assemblage account for less than 1% of the total artifacts recovered. The

Deptford levels of Minim Island yield the most shell tools making up 2.48% of artifacts.

The Table Point House site has an assemblage with 1.23% shell tools; however, all other sites contain less than 1% shell tools.

The vertebrate faunal remains are measured by weight (Table 5). Unfortunately, due to the variable nature of the excavation methodology at each site, determining a good relative frequency is problematic. Vertebrate faunal values range from 88.10g to

9,788.30g. There is variable shell composition across the sites; however, a majority are composed primarily of oyster. The two oystering station sites, 30BU833 and 38BU2

Blocks A, C, D, and E, contain the highest concentrations of oyster shell at 99%. The least amounts of oyster shell are found at Bass Pond Dam (79%) and the Thom’s Creek level of Minim Island (88%).

73

. Relative Frequency. Relative of Artifacts Sampledfrom Sites 4

Table

74 The diversity of pottery, vertebrate faunal remains, and lithic items provide additional data (Table 6). Site 38BU833 has the most identified pottery types. Blocks A,

C, D, and E of 38BU2 and the Thom’s Creek Minim Island midden has the second most diverse pottery assemblage. The Table Point Ring has the least unique pottery types.

Faunal diversity is highest in the Deptford midden at Minim Island and in Blocks A, C,

D, and E of site 38BU2. Of the sites with faunal remains, 30CH779 and 38BU833 have the smallest assemblages with only one identified species. Lithic types are most diverse in the Minim Island Deptford midden, Block B of 38BU2, and Blocks A, C, D, and E of

38BU2. The sites with the least diverse lithic assemblage are 38BU833 and the Table

Point House site

Table 5. Comparative Frequencies of fauna, oyster shell, and minority shell found at the sampled sites.

Table 96Table 97. Comparative Frequencies ofFauna fauna, oysterOyster shell, andMinorty minority shell found at the sampledSite sites. (g) Shell % Shell % Multi-Family Residential Base

Bass Pond Dam (38CH124) 872 79 21 Table 98. Relative Frequency of Artifacts from Sampled Sites Stafford North (9CM12) 178 95 5 Table Point (9CM12) Ring 522 95 5 TableTable 99 PointTable (9CM12) 100. Comparative House Frequencies of fauna,984 oyster shell,95 and 5 minority shell found at the sampled sites.

Single Family Shell Midden TableMinim 101 IslandTable Deptford 102. Comparative Frequencies of8151 fauna, oyster shell,95 and 5 minority shell found at the sampled sites. Minim Island Thoms Creek 9788 88 12 38BU2: Block B 1288 98 2 Table 103. Relative Frequency of Artifacts from Sampled Sites Single Family Limited Shell Site

Table30CH779 104. Relative Frequency of Artifacts from Sampled132 Sites 95 5 Fish Haul (38BU805)

Table 105. Relative Frequency of Artifacts from Sampled Sites Oystering Station 38BU833 88 99 1 Table38BU2: 106 Blocks. Relative A,C,D, Frequency and E of Artifacts from Sampled1364 SitesTable99 107. 1 Comparative Frequencies of fauna, oyster shell, and minority shell found at the sampled sites.

75 Table 108Table 109. Comparative Frequencies of fauna, oyster shell, and minority shell found at the sampled sites.

Table 110. Relative Frequency of Artifacts from Sampled Sites

Table 6. Diversity of artifact types in each site sampled. Diversity includes all types of artifacts within a particular artifact class.

Table 223. The diversity of artifact types in each site sampled. Diversity includes all types of artifacts within a particular artifact class. Site Pottery Fauna Lithics Multi-Family Residential Base TableBass 224Pond. The Dam diversity (38CH124) of artifact types in each site10 sampled. Diversity6 5 includesStafford all North types (9CM12) of artifacts within a particular artifact7 class. 0 0 Table Point (9CM12) Ring 4 7 0 TableTable 225 Point. The (9CM12) diversity House of artifact types in each site sampled.5 Diversity17 1 includes all types of artifacts within a particular artifact class. Single Family Shell Midden TableMinim 2 26Island. Comparative Deptford Frequencies of fauna, oyster 7shell, and minority30 shell 8 foundMinim at Island the sampled Thoms sites. CreekTable 227. The diversity of11 artifact types18 in each site7 sampled.38BU2: DiversityBlock B includes all types of artifacts within5 a particular17 artifact class.8

TableSingle 228 Family. The diversity Limited of Shell artifact Site types in each site sampled. Diversity includes30CH779 all types of artifacts within a particular artifact6 class. 1 2 Fish Haul (38BU805) 14 10 5 Table 229. The diversity of artifact types in each site sampled. Diversity includesOystering all types Station of artifacts within a particular artifact class. 38BU833 14 1 1

Table38BU2: 230 Blocks. The diversity A,C,D, of and artifact E types in each site11 sampled. Diversity19 8 includes all types of artifacts within a particular artifact class.

Table 7. The Presence or Absence of Structural Features and Human Remains Thefound presence at the orSampled absence Sites. of structural features and human remains provides an additional descriptiveTable 231Table element 232. The (Table Presence 7). orSite Absences with of human Structural remains Features include and Minim Island Human Remains found at the Sampled Sites. Deptford Midden, 38BU2 Block B, 30CH779, and Bass Pond Dam. These burials are

Table 233. The diversity of artifact types in each site sampled. Diversity mostly secondaryincludes burialsall types with of artifacts the exception within a particular of the Deptfordartifact class. burial at the Minim Island site. Structural features are entirely absent in the Oystering Station sites. All of the Multi- Table 234Table 235. The Presence or Absence of Structural Features and Family ResidentialHuman Remains Base sites found and at the most Sampled of the Sites. Single -Family sites have some type of structural feature.Table 236 Table 237. The Presence or Absence of Structural Features and Human Remains found at the Sampled Sites.

Table 238. The diversity of artifact types in each site sampled. Diversity includes all types of artifacts within a particular artifact class. 76

Table 239. The diversity of artifact types in each site sampled. Diversity includes all types of artifacts within a particular artifact class.

Table 240. The diversity of artifact types in each site sampled. Diversity

Table 7. The Presence or Absence of Structural Features and Human Remains found at the Sampled Sites.

Table 356Table 357. The Presence or AbsenceStructural of Structural FeaturesHuman and Human RemainsSite found at the Sampled Sites.Features Remains Multi-Family Residential Base Bass Pond Dam (38CH124) Yes Yes Table 358. The diversity of artifact types in each site sampled. DiversityStafford includes North (9CM12) all types of artifacts withinYes a particular artifactNo class. Table Point (9CM12) Ring Yes No Table Point (9CM12) House Yes No Table 359Table 360. The Presence or Absence of Structural Features

and Human Remains found at the Sampled Sites. Single Family Shell Midden Minim Island Deptford Yes Yes TableMinim 361 IslandTable Thoms362. The C Presencereek or AbsenceYes of Structural NoFeatures and Human Remains found at the Sampled Sites. 38BU2: Block B No Yes

TableSingle 363 Family. The diversity Limited of Shellartifact Site types in each site sampled. Diversity30CH779 includes all types of artifacts withinYes a particular artifactYes class. Fish Haul (38BU805) No No Table 364. The diversity of artifact types in each site sampled. DiversityOystering includes Station all types of artifacts within a particular artif act class. 38BU833 No No Table38BU2: 365 .Blocks The diversity A,C,D, of and artifact E types in eachNo site sampled.No Diversity includes all types of artifacts within a particular artifact class.

Table 366. The diversity of artifact types in each site sampled. Diversity includes all types of artifacts within a particular artifact The modelclass.Table struggles 367 . toThe meet Presence anticipated or Absence results of Structural when artifactsFeatures and are compared Human Remains found at the Sampled Sites. against expectations (Table 8). Multi-Family Residential Base sites should have a consistently highTable level 368 ofTable all 369artifact. The Presencetypes with or Absence moderately of Structural high amounts Features of lithics. The and Human Remains found at the Sampled Sites. data taken from Bass Pond Dam, Stafford North, and two loci at Table Point do not Table 370. The diversity of artifact types in each site sampled. substantiate thisDiversity claim. includesAs expected all types at ofa artifaMulticts-Family within a Residentialparticular artifact Base class. site , sherd frequency is relativelyTable 371 highTable at 372 Stafford. The Presence North or and Absence the Ring of Structural at Table Features Point . However, for and Human Remains found at the Sampled Sites. all other sites in this group, pottery as a percentage of the assemblage ranges from low to moderate. ExpectedTable 373lithicTable frequencies 374. The Presence are found or Absence only at of the Structural Bass PondFeatures Dam site, but none and Human Remains found at the Sampled Sites.

Table 8. Artifact visibility ranked low, medium, and high by comparing the percentages of artifact types across all sites examined. 77 Table 375Table 376. Artifact visibility ranked low, medium, and high by comparing the percentages of artifact types across all sites examined.

Results

Pottery

Low

High

Low

High

Moderate

Moderate

Moderate

Moderate

High

High Low

Pottery

Expectation

Low

Low

Moderate

Moderate

Relatively High High Relatively

Relatively High High Relatively

Relatively High High Relatively

Relatively High Relatively

Relatively High Relatively

Relatively High Relatively Relatively High Relatively

Lithic

Results

High

Low

High

Low

Moderate

Moderate

Moderate

Low

None None High

Lithic

Expectation

Very Low

Very Low

Moderate to High Moderate

Moderate to High Moderate

Moderate

Moderate

Moderate

Moderate to High Moderate

Moderate to High Moderate

Moderate to High Moderate Moderate to High Moderate

Ideotechnic

Item Results

None

None

Low

None

None

Low

Moderate

None

None

None Moderate

Item

Ideotechnic

Expectations

Low

Low

Moderate

Moderate

Moderate to High Moderate

Moderate to High Moderate

Moderate to High Moderate

Relatively High Relatively

Relatively High Relatively Relatively High Relatively High Relatively

Results

Bone Tool

Low

None

None

None

Low

Moderate

High

Moderate

None None None

Bone Tool

Expectation

Low

Low

Moderate

Moderate

Moderate

Moderate

Moderate

Relatively High Relatively

Relatively High Relatively

Relatively High Relatively Relatively High Relatively

Results

Shell Tool

Moderate

None

None

None

Low

Moderate

High

High

Moderate

None None

Shell Tool

Expectation

Low

Low

Moderate

Moderate

Moderate

Moderate

Moderate

Relatively High Relatively

Relatively High Relatively Relatively High Relatively High Relatively

Site

ranked. visibility Artifact low, medium, and high by the comparing percentagesof types artifact across all sites examined.

8

Table

38BU2: Blocks38BU2: A,C,D, E and

38BU833

Oystering StationOystering

Fish Haul Fish(38BU805) Haul

30CH779

Single Family Limited Shell Site

38BU2: Block38BU2: B

Minim Island Thoms Creek MinimThoms Island

Minim Island MinimDeptford Island

Single Family Shell Midden

Table Point (9CM12) House Point (9CM12) Table

Table Point (9CM12) Ring Point (9CM12) Table

Stafford North (9CM12) North Stafford

Bass (38CH124) Dam Pond Multi-Family Residential Base

78 of the sites contain relatively high concentrations of bone tools or ideotechnic items. In the shell tool category, the Table Point House site is the only site in our sample to meet expectations. In summary, while some of the sites contained artifacts within the expected frequencies, none of the so called Multi-Family Residential Base sites matched the model on more than one category.

According to the model, Multi-Family Residential Base sites should contain at least moderate amounts of lithic, pottery, and faunal diversity. None of the supposed

Multi-Family Residential sites met this expectation. Bass Pond Dam and Stafford North contains at least a moderate number of diverse pottery types. The two Table Point areas contains a moderate amount of faunal diversity, and only the Bass Pond Dam site has more than one type of lithic. Only Bass Pond Dam achieves more than one diversity target: pottery and lithics.

As predicted, oyster contribution was high; however, only the Bass Pond Dam site contains the minimum anticipated amount of minority shellfish. The model predicts that human remains and structural features are relatively common at Multi-Family sites.

Interestingly, all predicted Multi-Family Residential Base sites have structural features, while Bass Pond Dam yields the only definitive human remains. Therefore, it appears that the sampled sites only match the amount of oyster contribution and the presence of structural features in the proposed typology.

Single Family Shell Midden sites contain a moderate amount of lithics, as predicted. Most other categories are mismatched. The Thoms Creek component of the

Minim Island Site fits expectations with moderate amounts of shell and bone tools, but only the Deptford component of the Minim Island Site has the expected moderate amount

79 of ideotechnic items. All three sites have moderate amounts of pottery, when the relatively high amounts are anticipated.

While the Minim Island Thom Creek component accurately matches three of the five categories, there is still uncertainty that the proposed artifact frequencies are useful for identifying these types of sites. Lithic diversity is high when it should be moderate.

The vertebrate faunal diversity for the three sites is moderate to high which also matches the model. The number of different pottery types is mixed with high numbers for Minim

Island Thoms Creek, medium numbers for Minim Island Deptford, and low numbers for

Block B of 38BU2. Contrary to the model, the Minim Island Deptford site and Block B of 38BU2 each has a moderate amount of non-oyster molluscan in its assemblage, but

Minim Island Thoms Creek is true and accurate to expectations with moderate percentages. The model suggests that structures are common and human remains occur occasionally. Each Single Family Shell Midden site has either a structural feature, burial, or both.

Single Family Limited Shell sites fail to match all but one of the artifact frequency expectations. The Fish Haul site has a high lithic frequency that fits the hypothesized expectation. All other categories are mismatched. We then turn to artifact diversity and find that all of the Single Family Shell Middens sites have a moderate number of pottery types compared to other sites. This is slightly lower than the prediction of relatively high pottery diversity. However, single family shell sites have between seven and eight types of lithic artifacts which match the model’s assumption of a moderate to high lithic diversity. Faunal diversity at single-family shell middens should be moderate, but the data show a high number of individual vertebrate species. Oyster

80 contribution at the site was high at 95%, as predicted, but minority shellfish contribution was lower than expected at 5%.

The model also states that structural features at single-family shell sites are common and occasionally burials are found. True to the model, the Minim Island

Deptford and Thom’s Creek middens had structural features. Block B of 38BU2 did not have structural features but did contain human remains. Human remains were also found at the Minims Island Deptford midden.

Unlike Single-Family Limited Shell sites, Oystering Stations should have low or very low frequency for all artifacts. Site 38BU833 meets all expectations having either none or low amounts of every artifact class except pottery. Site 38BU833 has a high amount of pottery compared to all other sites in the sample. Blocks A, C, D, and E of

38BU2 have low or no artifacts in three out of the five categories. Compared to all other sites sampled, blocks A, C, D, and E of 38BU2 have a moderate number of shell tools and a high number of lithics.

In terms of artifact diversity, the two oystering sites have the highest diversity of pottery types. Blocks A, C, and D of 38BU2 also have extremely high faunal and lithic diversity. This diversity within the artifact assemblage is contrary to the low and very low expectations of the model. However, as expected, the two oystering stations sites do have the highest concentrations of oyster shell and the lowest concentrations of minority shellfish. Additionally, these sites have no structural features or human burials. The absence of structural features and burials is one of the biggest indicators that the site was used for short durations to harvest oysters.

Per the model, Brickhill Bluff most closely resembles the Oystering Station sites.

The Brickhill Bluff site matches four of the frequency benchmarks and contains a low

81 diversity of lithic artifacts and fauna. The artifact assemblage is made up largely of pottery with very few other items. A low frequency of lithic artifacts (1%) and no ideotechnic items or bone tools were recovered. Contrary to the model, Brickhill Bluff has a moderate amount of shell tools; however, the site conforms to the model in being comprised primarily of oyster shell with very low concentrations of minority shellfish species. While Brickhill Bluff contains a diverse array of pottery types, lithic types and vertebrate faunal diversity are also low. The site also contains no structural features or human burials. Examining all four of the site types available, it is evident that prehistoric

Native Americans used this site primarily to procure and process oysters. The site was occupied for short durations and lacks any defining evidence of structural features and very little evidence of domestic activities.

Although there are divergences between the hypothesized artifact pattern and the actual percentage of artifacts within each assemblage, certain patterns are visible that support Espenshade’s model. For instance, the presence or absence of burials and structural features are the most significant criteria for determining the category of the sites within this model. Additionally, a site might not match every benchmark but must be examined as a whole and may in fact match a number of important benchmarks that will provide clues to site formation and utility. It becomes problematic when individual artifact criteria are selected in isolation and forced to germinate larger conclusions regarding cultural behavior and thus site function and formation. The weakness of this model therefore is the tendency to overemphasis specific traits found within an artifact assemblage. These traits, while helpful, fail to provide the evidence to soundly determine site function and formation. Understanding how people and communities in the past utilized a site is dependent on a synthesis of all available data with an understanding that

82 irregularities within and between sites can, and often do, challenge our assumptions.

Ultimately, site, like human behavior, are variable and unwilling to be forced into sterile categories. However, when we look at all of the variables measured, archaeologists can find unique and helpful patterns.

Espenshade’s shell midden model demonstrates that shell middens can, with caution, be placed into exclusive categories based on matching a number of variables.

None of the sites studied for this thesis fit perfectly into a particular site type. All sites contained variability that, unfortunately, did not correspond with the model perfectly; however, there were general themes that allowed me to ascribe sites to a particular type.

The model appears to work best for identifying oystering stations. The two oystering station sites (38BU833 and Blocks A, C, D, and E of 38BU2) matched four and three of the frequency benchmarks. While, the diversity of pottery was higher than expected,

38BU833 has a low diversity of fauna material and lithic types as predicted by the model.

Nevertheless, the absence of structural features and human burials are likely the biggest clues that 38BU833 and Blocks A, C D, and E of 38BU2 are Oystering Stations.

The multi-family residential base and single-family limited shell sites are most contrary to the model. Still, these sites all contained structural features and/or burials indicating a more permeant habitation than the oystering stations. Without the structural features and burials, it might be possible to use artifact frequencies to classify these sites as oystering stations or possibly single family shell middens.

While the data gathered from the artifact frequency benchmarks was sometimes contrary to the model, the presence of structural features and burials helped justify the model. Artifact frequency worked well for sites placed within the Oystering Station category but frequency failed when trying to categorize a Multi-Family Residential Base

83 or a Single Family Limited Shell site. The results were mixed when trying to determine if a site was a Single Family shell midden.

The strongest criteria of the model, by far, is the presence or absence of human burials and structures. The presence of structural features clearly indicates some type of domestic function beyond oyster harvesting. All of the site not listed as an Oystering

Station had type of feature or burial. Many contemporaneous structures most likely indicate a more intensive use of the site by multiple family units. The use of shell probing is a useful tool for discerning these features at shell sites. The shell data at Brickhill Bluff provided valuable data regarding site limits and areas with heavier shell concentrations.

The data collected with the shell probe also confirmed that the site is devoid of features, which adds evidence to the hypothesis that Brickhill Bluff functioned as an Oystering

Station created by the slow accumulation of oyster shells during episodes of harvesting and processing. It is worth noting that mapping shell density is a useful visual that can guide further investigations and help determine site function. In addition to features and structures, burials may indicate some form of semi-permanent occupation, and multiple burials suggests either a long history of site use or concentrated habitation by multiple families. The second most helpful criteria is perhaps the diversity of artifacts found within a site followed by the amount of oyster compared to other shellfish species within the midden. Additionally, I would argue that site location, size, and type of structures found at a site are also necessary factors for determining site function.

The variability of archaeological sites makes it necessary to use Espenshade’s midden typology cautiously. This variability appears highest in sites classified as Multi-

Family Residential base and Single-Family Limited Shell sites. Fortunately, Multi-

Family Residential base sites are most likely to contain significant structural features to

84 separate them from Oystering Station sites and small Single-Family Limited Shell sites.

Limited Shell sites should be located on upland tracks and will also have features that separate them from Oystering Stations. I argue that the benchmarks for each site type in the shell midden matrix should be examined as a whole unit. Certain sites may not meet all, or even most, frequency and diversity standards, but sites may match important benchmarks that provide clues to site formation and function, the identification of features being one of the most important and telling benchmarks.

There are a number of limitations with this model. First, it is important to realize that all of the sites examined for this thesis are, to some degree, Oystering Stations.

Woodland people harvested and processed oysters for subsistence at all of these sites.

Variation among the artifact assemblages does lend clues to site function beyond oyster harvesting. Second, differences in preservation and excavation methodologies can make accurate intersite comparisons difficult. In addition, inconsistences in artifact identification also make it difficult to obtain good frequency and diversity data. This is especially true with the identification of faunal material. Finally, archaeological sites are not always homogenous. Assemblages from discrete levels and features may represent different behaviors over time or contemporaneously at the same locale. These limitations should be considered prior to future application of this midden typology.

85 REFERENCES

Anderson, David G. 1985 Middle Woodland Societies on the Lower South Atlantic Slope: A View from Georgia and South Carolina. Early Georgia 13 (1&2):29-66.

1996 Models of Paleoindian and Early Archaic Settlement in the Lower Southeast. In The Paleoindian and Early Archaic Southeast, edited by David G. Anderson and Kenneth E. Sassaman, pp. 29-57. University of Press, Tuscaloosa.

Anderson, David G. and Robert C. Mainfort, Jr. 2002 An Introduction to Woodland Archaeology in the Southeast. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 1-19. University of Alabama Press, Tuscaloosa.

Ashley, Keith H. 1992 Swift Creek Manifestations Along the Lower St. Johns River. The Florida Anthropologist 45 (2):127-138.

1998 Swift Creek Traits in Northeastern Florida: Ceramics, Mounds, and Middens. In A World Engraved:” Archaeology of the Swift Creek Culture, edited by Mark Williams and Daniel T. Elliott, pp. 197-221. University of Alabama Press, Tuscaloosa.

2008 Refining the Ceramic Chronology of Northeastern Florida. The Florida Anthropologist 61 (3-4):123-131.

Ashley, Keith H. and Vicki Rolland 2002 St. Marys Cordmarked Pottery (Formally Savannah Fine Cord Marked of Northeast Florida and Southeastern Georgia): A Type Description. The Florida Anthropologist 55(1):25-36.

Austin, John C. 1994 British Delft at Williamsburg. The Colonial Williamsburg Foundation, Williamsburg, Virginia.

Binford, Louis 1962 Archaeology as Anthropology. American Antiquity 28(2): 217–225.

1977 General Introduction. In For Theory Building in Archaeology, edited by L. R. Binford, pp. 1-13. Academic Press, New York.

1979 Organization and Formation Process: Looking at Curated Technologies. Journal of Anthropological Research 32:255-273.

1982 The Archaeology of Place. Journal of Anthropological Archaeology 1(1): 5-31.

86 Blitz, John H. 1988 Adoption of the Bow in Prehistoric North America. North American Archaeologist 9 (2):123-145.

Brewer David M. 1993 Trip Report on the Relocation of Monitoring Stations, Cumberland Island national Seashore, June 17-18, 1993; SEAC Accession 1096. Memorandum on file, National Park Service, Southeast Archeological Center, Tallahassee.

1994 Trip Report, Site Visits to Brickhill Bluff (9CM85) and River Trail (9CM14), Cumberland Island National Seashore, 12/8/94, SEAC Acc. 1156. Memorandum on file, National Park Service, Southeast Archeological Center, Tallahassee.

Caldwell, Joseph R. 1952 The Archaeology of Eastern Georgia to South Carolina. In Archeology of Eastern United States, edited by James B. Griffin, pp. 312–21. Chicago: University of Chicago Press.

Chenhall, Robert G. 1988 The Revised Nomenclature for Museum Cataloging. American Association for State and Local History (AASLH) Press, Nashville, .

Claassen, Cheryl 1998 Shells. Cambridge University Press, November 28.

Colquhound, Donald J., Mark J. Brooks, William H. Abbott, Frank W. Stapor, Walter S. Newman, and Richard R. Pardi 1980 Principles and Problems in Establishing a Holocene Sea-level Curve for South Carolina. In Excursions in Southeastern Geology, the Archaeology-Geology of the Georgia Coast, edited by James D. Howard, Chester B. DePratter, and Robert W. Frey, pp. 143-159. Guidebook 20, Department of Natural Resources, Georgia.

Combes, John D. 1975 The Archeology of Kiawah Island. In Environmental Inventory of Kiawah Island, Environmental Research Center, Inc., Columbia, South Carolina.

Cook, Fred 1977 The lower Georgia coast as a cultural sub-region. In Early Georgia 5 (1&2).

Cook, Fred C. and Charles E. Pearson 1989 The Southeastern Ceremonial Complex on the Georgia Coast. In The Southeastern Ceremonial Complex: Artifacts and Analysis, edited by Patricia Galloway, pp. 145- 165. University of Nebraska Press, Lincoln and London.

Crook, Morgan Ray 1978 Mississippian Period Community Organizations on the Georgia Coast. Ph.D Dissertation University of Florida, Gainesville.

1986 Mississippi Period Archaeology of the Georgia Coastal Zone. University of Georgia Laboratory of Archaeology Series Report No. 23, Georgia Archaeological Research Design Papers No. 1, University of Georgia, Athens.

87 Crusoe, Donald L. 1973a Cumberland Island National Seashore. A Research Prospectus. Manuscript on file, National Park Service, Southeast Archeological Center, Tallahassee.

1973b Cumberland Island Trip Report. Memorandum on file, National Park Service, Southeast Archeological Center, Tallahassee.

DePratter, Chester B. 1976 The Refuge Phase on the Coastal Plain of Georgia. Early Georgia 4 (1 & 2):1-13.

1979 Ceramics. In The Anthropology of St. Catherines Island. The Refuge-Deptford Mortuary Complex, edited by D.H. Thomas and C.S. Larsen, pp. 109-132. Anthropological Papers No. 56, American Museum of Natural History.

1991 W.P.A. Archaeological Excavations in Chatham County, Georgia: 1937–1942. Laboratory of Archaeology Series 29, Department of Anthropology, University of Georgia, Athens.

DePratter, Chester B., and James D. Howard 1981 Evidence for a Sea Level Lowstand between 4500 and 2400 Years B.C. on the Southeast Coast of the United States. Journal of Sedimentary Petrology 51: 1287-1295.

Des Jean, Thomas, Irvy R. Quitmeyer, and Karen Jo Walker 1985 A Coastal Swift Creek Community at King’s Bay, Georgia. In Indians, Colonists, and Slaves: Essays in Memory of Charles H. Fairbanks, edited by Kenneth W. Johnson, Jonathan M. Leader, and Robert C. Wilson, pp. 155-177. Special Publication No.4, Florida Journal of Anthropology, University of Florida Student Anthropology Association, Gainesville.

Deutschle, Stephen A., and Robert C. Wilson 1975a Assessment of Archeological and Historical Resources of the Georgia Coast Adjacent to Cumberland Island National Seashore. Ms. on file, National Park Service, Southeast Archeological Center, Tallahassee.

1975b Known Prehistoric and Historic Resources of Cumberland Island National Seashore. Ms. on file, National Park Service, Southeast Archeological Center, Tallahassee.

Echternacht, Arthur C. and Larry D. Harris 1993 The Fauna and Wildlife of the Southeastern United States. In Biodiversity of the Southeastern United States. Lowland Terrestrial Communities, edited by William H. Martin, Stephen C. Boyce, and Arthur C. Echternacht, pp. 81-116. John Wiley & Sons, Inc., New York.

Ehrenhard, John E. 1975 A Preliminary Evaluation of National Register Site Eligibility, Cumberland Island National Seashore Area. Manuscript. on file, National Park Service, Southeast Archeological Center, Tallahassee.

1976 Cumberland Island National Seashore. Assessment of Archaeological and Historic Resources. Manuscript. on file, National Park Service, Southeast Archeological Center, Tallahassee.

88 1988 Archaeological Site Stabilization at Cumberland Island National Seashore. Manuscript on file, National Park Service, Southeast Archeological Center, Tallahassee.

Ehrenhard, John E., and Robert M. Thorne 1991 An Experiment in Archeological Site Stabilization. Cumberland Island National Seashore. CRM 14 (2):13-16,19. U.S. Department of Interior, National Park Service, Washington, D.C.

1993 An Experiment in Archeological Site Stabilization—Part II. Cumberland Island National Seashore. CRM 14 (2): 3-4,16. U.S. Department of Interior, National Park Service, Washington, D.C.

Elliott, Daniel T. 1998 The Northern and Eastern Expression of Swift Creek Culture: Settlement in the Tennessee and Savannah River Valley. In A World Engraved: Archaeology of the Swift Creek Culture, edited by Mark Williams and Daniel T. Elliott, pp. 19-35. University of Alabama Press, Tuscaloosa.

Espenshade, Christopher T. 1993 A Few Visits in Prehistory: Data Recovery Excavation at 9Rh18, Randolph County, Georgia. Submitted to the Georgia Department of Transportation, Atlanta. Occasional Papers in Cultural Resource Management No. 5, Brockington and Associates, Atlanta and Charleston.

Espenshade, Christopher T., and Paul E. Brockington 1989 An Archaeological Study of the Minim Island Site: Early Woodland Dynamics in Coastal South Carolina. Submitted to U.S. Army Corps of Engineers by Brockington and Associates, Atlanta.

Espenshade, Christopher T., Linda Kennedy, and Bobby G. Southerlin 1994 What Is a Shell Midden? Data Recovery Excavations of Thom’s Creek and Deptford Shell Middens, 38BU2, Spring Island, South Carolina. Submitted to Spring Island Plantation by Brockington and Associates, Atlanta.

Faust, Richard D. 1987 Archeological Site Monitoring Program. Memorandum on file, National Park Service, Southeast Archeological Center, Tallahassee.

Ford, Richard I. 1979 Paleoethnobotany in American Archaeology. Advances in Archaeological Method And Theory 2:285–336.

Goggin, John M. 1952 Space and Time Perspective in Northern St. Johns Archaeology, Florida. Yale University Publications in Anthropology 47. New Haven, Connecticut.

Griffin, James B. 1967 Eastern North American Archaeology: A Summary. Science 156: 175-191.

89 1985 Changing Concepts of the Prehistoric Mississippian Cultures of the Eastern United States. In Alabama and the Borderlands; From Prehistory to Statehood, edited by R. Reid Badger and Lawrence A. Clayton, pp. 40-63, University of Alabama Press, Tuscaloosa.

Griffin, Martha M. 1991 Geological Guide to Cumberland Island National Seashore. Department of Natural Resources, Environmental Protection Division, Georgia Geologic Survey, Georgia.

Hails, J.R., and J.H. Hoyt 1969 An Appraisal of the Evolution of the Lower Atlantic Coastal Plain of Georgia, U.S.A. Transactions of the Institute of British Geographers Vol. 0, Issue 46:53-68.

Hellmann, Robert 2005 Archeological Overview and Assessment of Cumberland Island National Seashore, Camden County. MANUSCRIPT on file, National Park Service, Southeast Archeological Center, Tallahassee, Florida

Hillestad, Hilburn O., John R. Bozeman, A. Sydney Johnson, C. Wayne Berisford, and James I. Richardson 1975 The Ecology of Cumberland Island National Seashore, Camden County, Georgia. Georgia Marine Science Center, University System of Georgia, Skidaway Island, Georgia.

Jefferies, Richard W. 1994 The Swift Creek Site and Woodland Platform Mounds in the Southeastern United States. In Ocmulgee Archaeology, 1936–1986, edited by D. J. Hally, pp. 71–83. Athens: University of Georgia Press.

Jennings, Justin 2011 Globalizations and the ancient world. Cambridge University Press, Cambridge, New York.

Johnson, Amber L. (editor). 2004 Processual archaeology: exploring analytical strategies, frames of reference, and culture process. Praeger, Westport, Conn.

Justice, Noel D. 1987 Stone Age Spear and Arrow Points of the Midcontinental and Eastern United States. University Press, Bloomington.

Kaplan, Eugene H. 1988 Southeastern and Caribbean Seashores. Peterson Field Guide Series, Houghton Mifflin Company, Boston and New York.

Kidd, Steven R. and Stephen Andrew Wise 2007 Research Design for Survey and limited Testing at the Brickhill Bluff Site (9CM85), Cumberland Island National Seashore. Manuscript on file, National Park Service, Southeast Archeological Center, Tallahassee.

90 Lawrence, David R. 1992 Archaeological oysters from sites 38BU132, 38BU372, and 38BU1241, Colleton River Plantation, Beaufort County, South Carolina. In Data Recovery Investigations of Four Wilmington Phase Sites (38BU132, 38BU372, 38BU1236, and 38BU1241), Beaufort County, South Carolina: A Study in Middle Woodland Subsistence Strategy, Brockington and Associates, Atlanta, Georgia.

Marquardt, William H. 1992 Shell Artifacts from the Caloosahatchee Area. In Culture and Environment in the Domain of the , edited by William H. Marquardt and Claudine Payne, pp. 191– 221. Institute of Archaeology and Paleoenvironmental Studies, Monograph No. 1. Florida Museum of Natural History, Gainesville.

Michie, James R. 1979 The Bass Pond Dam Site: Intensive Archaeological Testing at a Formative Period Base Camp on Kiawah Island, South Carolina. Research Manuscript Series No. 154, South Carolina Institute of Archaeology and Anthropology, Columbia.

Milanich, Jerald T. 1973 Review of The Pre-Columbian Mind. The Florida Historical Quarterly 51(4): 446–447.

1971 The Deptford Phase: An Archeological Reconstruction. Ph.D. Dissertation, University of Florida, Gainesville.

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

Miller, George L. 1991 A Revised Set of CC Index Values for Classification and Economic Scaling of English Ceramics from 1787 to 1880. Historical Archaeology 25(1): 1-25.

Moore, Clarence B. 1897 Certain Mounds of the Georgia Coast. Journal of the Academy of Natural Sciences of Philadelphia 11.

Muller, Jon 1983 The Southeast. In Ancient North Americans, edited by Jesse D. Jennings, pp. 373-419. W. H. Freeman and Co., New York.

National Park Service (NPS) 2010 Museum Handbook, Museum Records, Part II. U.S. Department of the Interior, Washington, D.C.

1990 Cataloging Manual for Archeological Objects, Vols. I, II, III. Southeast Archeological Center, Tallahassee.

Nelson, Lee H. 1968 Nail Chronology as an Aid to Dating Old Buildings. History News, Vol. 24, No. 11. Technical Leaflet 48.

Noël Hume, Ivor 1969 A Guide to Artifacts of Colonial America. University of Pennsylvania Press, Philadelphia. 91 Peebles, Christopher, and Susan Kus 1977 Some Archaeological Correlates of Ranked Societies. American Antiquity 42 (3):421- 48.

Rice, Prudence M. 1987 Pottery Analysis: A Sourcebook. University of Chicago Press.

Russo, Michael 1992 Chronologies and Cultures of the St. Marys Region of Northeast Florida and Southeast Georgia. The Florida Anthropologist 45:107-126.

Russo, Michael and Gregory Heide 2002 The Joseph Reed Shell Ring. The Florida Anthropologist 55(2):67-87.

Saffer, Marian E. 1979 Aboriginal Clay Resource Utilization on the Georgia Coast. Master’s thesis, Department of Anthropology, University of Florida, Gainesville.

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

2002 Woodland Ceramic Beginnings. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp. 398 -420. University of Alabama Press, Tuscaloosa.

Sassaman, Kenneth E., Mark J. Brooks, Glen T. Hanson, and David G. Anderson 1990 Native American Prehistory of the Middle Savannah River Valley: A Synthesis of Archaeological Investigations on the Savannah River Site, Aiken and Barnwell Counties, South Carolina. Occasional Papers of the Savannah River Archaeological Research Program, Columbia.

Schiffer, Michael B. 1987 Formation Processes of the Archaeological Record. University of New Mexico Press, Albuquerque.

1995 Behavioral Archaeology. University of Utah Press, Salt Lake City.

Schoettle, Taylor 1993 A Naturalist’s Guide to St. Simons Island. Darien News, Inc., Darien, Ga.

Sears, William H. 1957 Excavations on Lower St. Johns River, Florida. Contributions of the Florida State Museum 2, Gainesville.

Setzler, Frank M. 1933 Salvaging an Aboriginal Dug-Out Canoe, Cumberland Island, Georgia. In Explorations and Field-Work of the Smithsonian Institution in 1932, pp. 57-60. Smithsonian Institution, Washington, D.C.

Shepard, Anna Osler 1965 Ceramics for the Archaeologist. Carnegie Institution of Washington, Washington.

92 Shott, Michael J. 1998 Status and role of formation theory in contemporary archaeological practice. Journal of Archaeological Research 6(4): 299–329.

Smith, Bruce 1986 Archaeology of the Southeastern United States: From Dalton to deSoto, 10,500-500 B.P. In Advances in World Archaeology, Vol. 5, edited by Fred Wendorf and A. Close, pp. 1-92. Academic Press, Inc, New York.

1989 Origins of Agriculture in Eastern North America. Science 246:1566-1571.

Stalter, Richard and William E. Odum 1993 Maritime Communities. In Biodiversity of the Southeastern United States. Lowland Terrestrial Communities, edited by William H. Martin, Stephen C. Boyce, and Arthur C. Echternacht, pp. 81-116. John Wiley & Sons, Inc., New York.

Stephenson, Keith, Judith A. Bense, and Frankie Snow 2002 Aspects of Deptford and Swift Creek of the South Atlantic and Gulf Coast Plains. In The Woodland Southeast, edited by David G. Anderson and Robert C. Mainfort, Jr., pp.318-351. University of Alabama Press, Tuscaloosa.

Steponaitis, Vincas P. 1986 Prehistoric Archaeology in the Southeastern United States, 1970-1985. Annual Review of Anthropology 15:363-404.

Tesar, Louis D. 2011 Ceramic Tales: Artifacts From the Mounds Mapping Project, Jefferson and Taylor Counties, Florida. Ms. on file, National Park Service, Southeast Archeological Center, Tallahassee.

Thomas, David H. and Clark S. Larsen 1979 The Anthropology of St. Catherines Island. The Refuge-Deptford Mortuary Complex, edited by D.H. Thomas and C.S. Larsen, pp. 109-132. Anthropological Papers No. 56, American Museum of Natural History, New York.

Thorne, Robert 1988 Filter Fabric: A Technique for Short-Term Site Stabilization. Technical Brief 1, U.S. Department of Interior, National Park Service Archeology and Ethnography Program, Washington, D.C.

Trinkley, Michael B. 1980 Investigation of the Woodland Period along the South Carolina Coast. Ph.D. Dissertation,University of , Chapel Hill.

1984 The Archaeology of Sol Legare Island, Charleston County, South Carolina. Chicora Foundation, Columbia, South Carolina.

1986 Indian and Freeman Occupation at the Fish Haul Site (38BU805), Beaufort County South Carolina. Chicora Foundation, Columbia, South Carolina.

1987 Archaeological Survey of Hilton Head Island, Beaufort County, South Carolina. Chicora Foundation, Columbia, South Carolina.

93 1991 Further Investigations of Prehistoric and Historic Lifeways on Callawassie and Spring Islands, Beaufort County, South Carolina. Chicora Foundation Research Series 23, Columbia, South Carolina.

1992 Archaeological Data Recovery at 38BU833, at St. Catherines and Savannah Shell Midden Site, Hilton Head Island, Beaufort County, South Carolina. Chicora Foundation, Columbia, S.C.

Trinkley, Michael (editor) 1990 The Second Phase of Archaeological Survey on Spring Island, Beaufort County, South Carolina, Investigation of Prehistoric and Historic Settlement Patterns on an Isolated Sea Island. Chicora Foundation Research Series 20, Columbia, S.C.

Trinkley, Michael, and Natalie Adams 1994 Middle and Late Woodland Life at Old House Creek, Hilton Head Island, South Carolina. Chicora Foundation Research Series 42, Columbia, S.C.

Waring, Antonio J., Jr., and Preston Holder 1968 The Deptford Ceramic Complex. In The Waring Papers: The Collected Works of Antonio J. Waring, Jr., edited by Stephen Williams, pp. 135–51. Papers of the Peabody Museum of Archaeology and Ethnology, Vol. 58, Harvard University, Cambridge.

Williams, Mark and Daniel T. Elliott 1998 Swift Creek Research: History and Observations. In A World Engraved:” Archaeology of the Swift Creek Culture, edited by Mark Williams and Daniel T. Elliott, pp. 1-11. University of Alabama Press, Tuscaloosa.

Williams, Mark and Victor Thompson 1999 A Guide to Georgia Indian Pottery Types. Early Georgia 27:1-167. Smith, Bruce D. (editor)

Wilson, Robert C. 1988 Archeological Site Monitoring on CUIS, 2/8-10/88. Memorandum on file, National Park Service, Southeast Archeological Center, Tallahassee.

Wise, S. Andrew 2012 Archaeological Investigations at Brickhill Bluff, Cumberland Island National Seashore. Manuscript. on file, National Park Service, Southeast Archeological Center, Tallahassee.

Wood, Dean 1999 “A Reconnaissance Survey of Two Alternative Locations for a Dock or Water Taxi on Cumberland Island National Seashore, Camden County, Georgia.” Report submitted to Pond and Co. Atlanta, Georgia.

Zoodsma, Barbara Jo 1991 Distribution and Behavioral Ecology of Manatees in Southeastern Georgia. Research/Resources Management Report SER-91/03 Kings Bay Environmental Monitoring Program Report. Ms. on file, National Park Service, Atlanta, Ga.

94