CERAMICS, MOLLUSCS, AND SEDENTISH:
THE LATE ARCHAIC PERIOD ON THE GEORGIA COAST
Rochelle A. Marrinan
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Although some research efforts may be individual undertakings, archaeo-
logical investigations characteristically are not. Many factors--the site,
quantities of material to be processed, techniques used to gather informa-
tion, and the usually public nature of excavations--preclude the possibility
that a single person can accomplish the field research singlehandedly.
During the four field sessions spent gathering data which make this disser-
tation possible, over 30 students were provided field research experience.
Their participation in the University of Florida Department of Anthropology
Archaeology Field Schools made possible the excavation of materials pre-
sented herein. It is with gratitude that I acknowledge the following field
Summer, 1973. Nain Anderson, Roberta Lemlich, Leonard Roberts, Sandra
Sampson, and Mary Turpen.
Spring, 1974. Chris Birdsall, Mark Brooks, Marion Drescher, Leslie
Gresh, Steve Hamburg, Lynne Jackson, Tim Kohler, and Hugh Prine.
Summer, 1974. Richard Atwood, John Barksdale, Debbi Baukney, Bill
Christofferson, Robin Futch, Jean Gearing, Elizabeth Hill, Brenda Lavelle,
David Lawrence, Maureen Lineaweaver, Karen Malesky, Harvey McKenzie, Janet
McPhail, and Rebeca Quintana-Garcia, Connie Welsch, and Joy Willard.
Winter, 1975. Nain Anderson, George Edwardson, Sandra Forney, Tim
Kohler, and Mary Turpen. Robin Futch, Elizabeth Hill, Roberta Lemlich,
Janet McPhail, and Rebeca Quintana-Garcia volunteered many weekends of
assistance which added needed time to field work.
Analysis of excavated materials involved the assistance of many
specialists. I should like to thank Dr. Fred Thompson, Malacologist,
Florida State Museum, for aid in identification of the molluscs; Dr.
Pierce Brodkorb of the University of Florida Department of Zoology for
assistance with the intricacies of avian osteology, and Dr. David
Hubbel, Institute of Food and Agricultural Sciences, University of
Florida, who made possible the processing of soil samples. Dr. William
R. Maples, Chairman of the Department of SociaLSciences, Florida State
Museum, aged the few human'skeletal elements recovered. Dr. Robert L.
Stephenson, Director of the Institute of Archaeology and Anthropology,
University of South Carolina, made copies of preliminary work on South
Carolina ring sites available. Dr. E. Thomas Hemmings, West Virginia
Geological Survey, made available copies of manuscripts in preparation
and preliminary field analysis.
Special thanks is extended to Chester DePratter, University of
Georgia, for his continued interest. His knowledge of the Georgia
coastal area provided numerous insights. Charles Pearson, University
of Georgia, shared his knowledge of sites within the project area. Mr.
Pearson first reported the Marsh Ring site in 1972. Dr. Donald L.
Crusoe, Southeastern Archaeological Center, made available copies of
manuscripts in preparation and discussed his research findings.
Preliminary paleobotanical identification was made by Timothy A.
Kohler using the facilities of the University of Florida Herbarium
and his own collections. Mr. Kohler was twice a field crew member and
any expression of gratitude of his contributions would be insufficient.
George Edwardson and Lynn Cunningham (University of New Mexico)
spent considerable time with illustrations presented. Roberta Lemlich
provided the photographs of lithic and shell artifacts. Dr. William
R. Maples kindly made the photographic equipment of the Department of
Social Sciences, Florida State Museum, available.
Sea Island Properties, Inc., made this project possible and it is
impossible to thank them sufficiently. However, I would specifically
like to thank Mr. William A. Jones, Sr., Mr. William A. Jones, Jr.,
and Mr. J.D. Benefield, Jr., for their kindness and cooperation. I am
very grateful that the opportunity to conduct research in such an
interesting and extremely beautiful area was mine.
To the members of my committee, I extend thanks for their continued
academic and personal support. Dr. Dickinson spent considerable time
editing this manuscript. Dr. Patton generously provided information
and encouragement. Dr. Milanich provided advice, field assistance, and
editorial time. Dr. Wing has been a great help to me since I first
came to her for assistance with a field school project as an undergraduate.
Since that time and particularly since 1973, she has made space available
for work, comparative collections, and invaluable information on almost
a daily basis. I have greatly profited from this contact.
Dr. Fairbanks initially interested me in this research topic and
continued financial support for field and analytical work. Four field
sessions constitute a major outlay of funds. I am particularly grateful
for the confidence, the personal contact, and the direction he has will-
ingly given. In the final analysis, Dr. Fairbanks is responsible for pro-
viding the opportunity to undertake this research.
There are probably others to whom an expression of gratitude is
due. I apologize for any oversight and wish to point out that while
the above-named individuals have provided information from their
special competencies, I am responsible for its presentation in this
This dissertation is a study in the application of hypothesis for-
mulation and testing to archaeological investigation of an early prehis-
toric aboriginal culture. It is presented chronologically in order to
document the considerations and observations in perspective. To do other-
wise would create an artificial impression of smoothness and correctness
that is not real. Further,'this study is seen as an exploration of the
application of more systematic methodological techniques in field exca-
vation. Analysis of data thus collected to an long-standing problem in
southeastern United States prehistory generates very interesting results.
Finally, it is intended to be a contribution to the status of knowledge
on the subject of southeastern shell ring sites. To this point, it is
a compilation of data presently available in the archaeological litera-
ture as well as a presentation of data resulting from field excavations
on Cannon's Point, St. Simon's Island, Glynn County, Georgia. It is
believed that such data be used for comparative purposes by investigators
considering the same or similar problems.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . . . . . ii
PREFACE . . . . . . . .... . . . . . vi
LIST OF TABLES . . . . . . . . . . x
LIST OF FIGURES. . . . . . . . ... . .. xi
ABSTRACT . . . . . . . . . . . . . xiii
CHAPTER 1: SHELL RINGS IN SOUTHEASTERN UNITED STATES
ARCHAEOLOGY . . . . . . . . . . 1
Introduction . . . . . . .. . . . . . I
The University of Florida St. Simon's Island Archaeological
Project . . . . . . . . . . . 2
Southeastern Shell Rings . . . . . . . . . 3
Previous Investigations and Interpretations . . . . 9
Chronological Assessment . . . . . . . . .. 13
The Cultural Phenomena . . . . . . . . .. 15
CHAPTER 2: EXCAVATIONS ON CANNON'S POINT PART 1:
SUMMER 1973, SPRING 1974 . . . . . . .. 19
Present Environment . . . . . . . . 19
Excavation Strategy: Objective and Techniques . . .. 24
Summer 1973: Excavation Summary . . . . . . .. 25
Spring 1974: Excavation Summary . . . . . .... 28
Preliminary Observations . . . . . . . .... 32
CHAPTER 3: EXCAVATIONS ON CANNON'S POINT PART 2:
SUMMER 1974, WINTER 1975 . . . . . . 37
Formulation of Hypotheses and Test Implications . . . 37
Summer 19749 Excavation Summary . . . . . . 39
Winter 1975: Excavation Summary . . . . . 41
Summary of Excavation Findings . . . . . . .
Evaluation of Hypotheses and Test Implications . . .. 45
CHAPTER 4: ANALYSIS OF EXCAVATED MATERIALS . . . . .. 48
Radiocarbon Dates . . . . . . . . . 48
Cultural Remains . . . . . . . . .. . 51
Faunal Materials . . . . . . . .. . . 67
Floral Remains . . . . . . . . . . 78
Other Analyses . . . . . . . .. .. . 80
Consideration of Methods and Techniques . . . . 82
CHAPTER 5: THE LATE ARCHAIC OCCUPATION OF CANNON'S POINT . 89
Prehistoric Environment . . . . . . . .... 89
Techno-environmental Adaptation .. . . . . . . 91
Seasonality and Sedentism. ... . . . . ... 96
Extra-areal Relationships . . . . . . . . 102
CHAPTER 6: POST-EXCAVATION ASSESSMENTS OF THE LATE ARCHAIC . 105
The Cultural Phenomena . . . ... . . . .. 105
Relationships to Other New World Sites . . . . .. 109
Summary and Conclusions . . . . . . .... .116
APPENDIX 1 . . . . . . . . . . . ... 119
APPENDIX 2 . . . . . . . .... ... . 130
APPENDIX 3 . . . . . . . .... . . . . 131
APPENDIX 4 . . . . . . . . ... . . . . 134
APPENDIX 5 . . . . . . . . . . . . . 140
APPENDIX 6. .. . . . . . .... . . . . 201
APPENDIX 7. .. . . . . . .... . . . . . 206
APPENDIX 8 . . . . . . . . . . . . . 217
APPENDIX 9 . . . . . . . . . . . . . 220
APPENDIX 10 . . . . . . . ....... ......... .225
APPENDIX 11 . . . . . . . . . . . . . 232
APPENDIX 12 . . . . . . . . . . . . . 239
APPENDIX 13 . . . . . . . . . . . 248
APPENDIX 14 . . . . . . . . . . . . . 250
REFERENCES CITED . . . . . . . . ... ...... 253
BIBLIOGRAPHY . . . . . . . . . . . . . 264
LIST OF TABLES
Table 1. Site Distribution . . . . . . . . . . 6
Table 2. Ring Dimensions in Meters . . . . . . . . . 8
Table 3. Chronological Summary . . . . . . . .... .16
Table 4. Radiocarbon Dates for Ring Sites . . . . . ... 35
Table 5. Radiocarbon Dates for Other Fiber-Tempered Sites. . ... 50
Table 6. Ceramic Distribution for the Cannon's Point Ring Sites . 56
Table 7. Ceramic Distribution for the Marsh Cultural Level . . .. .57
Table 8. Total Recovered Ceramic Material . . . . . .. 58
Table 9. Composite Molluscan Species List 9GN57, 9GN76 . . ... 69
Table 10. Faunal Lists: 9GN57, 9GN76, and Marsh Cultural Level . . 71
Table 11. Dietary Contributions of Selected Species . . . ... 75
Table 12. Seasonal Availability of Identified Flora . . . ... 79
Table 13. Human Skeletal Material . . . . . . . .... 83
Table 14. Contribution of Screening by Weight of Processed
Material 6S, 15E . . . . . . . .... . 85
LIST OF FIGURES
Figure 22 . . .
Figure 4 .
Figure 6 .
Figure 11 .
List of Figures continued
Figure 23 ... . . . . . . . . . .. 245
Figure 24 . . . . . . .. . . . . .. . 246
Figure 25 . . . . . . . . . . . . . 247
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
CERAMICS, MOLLUSCS, AND SEDENTISM:
THE LATE ARCHAIC PERIOD ON THE GEORGIA COAST
Rochelle A. Marrinan
Chairman: Charles H. Fairbanks
Major Department: Anthropology
Previous researchers have contended that during the Late Archaic Period
(circa 2500-1000 BC), human groups on the present-day coast of South Caro-
lina and Georgia practiced a mollusc-centered subsistence strategy. The
abundance of molluscan resources has been viewed as evidence of a rela-
tively permanent residence pattern. This stable pattern of residence was
thought to have been instrumental in the innovation of fired clay pottery.
Today these former habitation sites exist as annular rings and amorphous
middens of molluscan debris.
Recent archaeological investigations on Cannon's Point, St. Simon's
Island, Georgia, in such habitation sites indicate that although molluscan
resources were obviously an important contributor to prehistoric subsistence,
crustaceans and vertebrate fauna (particularly fishes) also were critical
Fiber-tempered ceramics, to date the earliest type of earthenware
reported in the United States, are present in all of the Late Archaic
Period sites excavated. Affiliations between Georgia fiber-tempered
ceramic decorative motifs and motifs associated with Orange pottery from
the St. John's River drainage in Florida are noted. This finding supports
the previous observation of Holder and others for the St. Simon's Island
area. The Georgia coastal ceramic tradition is considered influential
in the development of Orange ceramics.
Reconsideration of the generally accepted belief that decorated fiber-
tempered ceramics are later than plain ceramics is urged based on the findings
from the excavation of two shell ring sites (dated by radiocarbon at 2240-
1815 BC and 1910-1665 BC). Decorated ceramics occur in these sites at all
levels, dating coevally with plain specimens. Reconsideration is also urged
for Waring's contention that the Bilbo-style engraved bone pins are a stylis-
tic development. Both of the two bone pin fragments recovered in the Cannon's
Point excavations were from a sub-shell stratum.
In the marsh surrounding the largest ring site, a cultural deposit was
exposed 0.5-1.0 meter beneath the present surface. The association of cul-
tural materials (ceramics, worked bone, lithics, flora, and fauna) with tree
stumps suggested that environmental change is evidenced in the immediate
vicinity of the ring sites. When deposited, the Marsh Ring accumulated in
a forest environment. Ceramics from this cultural deposit are both fiber-
tempered and grit-tempered; however, the manner in which deposition occurred
is unclear. Radiocarbon dates indicate that environmental change in this
area occurred after 835-820 BC.
Permanence of residential pattern cannot be fully evaluated at present.
Evidence suggesting a Spring-Fall occupation of the Cannon's Point sites
is considered tenuous at this time. Investigation of inland riverine sites
having fiber-tempered ceramics is needed to evaluate the relationship of
these biotopes to the coast.
Excavations in the Cannon's Point shell ring sites have demonstrated
that more systematic methods and techniques must be applied to data reco-
very. Perceptions of coastal cultural developments have been framed in
terms of cultural inventory and artifact attributes. Development of a
research strategy to investigate the articulation between human culture and
physical environment is urged to provide an understanding of cultural events
during the Late Archaic Period.
SHELL RINGS in SOUTHEASTERN UNITED STATES ARCHAEOLOGY
For thousands of years prior to the founding of European colonies
in the New World, native American populations experienced remarkable
variety in cultural complexity, geographical distribution, and in their
use of environmental resources. Today archaeological sites are monuments
to this diversity of prehistoric lifestyles. Just as the strangeness
of such people and places promoted early Europeans to attribute unusual
origins and abilities to native peoples, the intricacies and apparent
inexplicabilities of human cultural remains still encourage some to
speculate novel derivations of human artifacts. It has been the task
of archaeology to attempt to demonstrate causality in the adaptability and
flexibility of mankind. To the student of prehistory change is perhaps
the most obvious fact and the one most challenging to explain. Plog
(1974:8) considers explanation of change to be the primary undertaking
of productive archaeological research. The methodology employed in
explanation is a critical issue in archaeology today for it is not
sufficient to simply ... tie data to cultural phenomena ..." (Plog
1974:4). Recovering significant archaeological data from field research
strategies having the objectives of hypothesis formulation and testing
affords the opportunity to investigate parameters germain to explication and
explanation of cultural processes (Binford 1962; Watson, LeBlanc, and Redman
1971). As a result of these objectives, contribution to the corpus of
anthropological knowledge may be achieved.
The present study demonstrates the use of archaeological data to
explain cultural phenomena. It proceeds from a review of the pertinent
archaeological literature to recognition of specific needs and problems
requiring further study within the research area. Hypotheses are generated
for field testing using excavation techniques selected to maximize data
recovery. Hypotheses, it should be noted, do not spring forth at the need
of the investigator. Rather, they are usually the result of continuing
research. In this study, it should be made clear that the status of
information regarding the research topic was not sufficient enough to
support more than very general hypotheses at the outset. As excavation and
more importantly analysis progressed, data suggested direction for
inquiry. Hypotheses were generated to structure excavation strategy and
future analysis. Working hypotheses will be explicitly stated in Chapter
2 and 3.
The University of Florida St. Simon's Island Archaeological Project
Dr. Charles H. Fairbanks and Dr. Jerald T. Milanich of the University
of Florida Department of Anthropology were approached in 1972 by Sea
Island Properties, Inc., regarding a proposed archaeological survey of
Cannon's Point, St. Simon's Island, Glynn County, Georgia. Several sites of
known archaeological interest were in areas scheduled for development
promoting the landowners' concern for recovery and preservation of
prehistoric and historic materials. Ruins of the Couper Plantation were
listed in the National Register of Historic Palces. Excavations in the
extensive shell heaps on the property had been made by Preston A. Holder
during the late 1930's. Human skeletal remains were frequently exposed
in plowed fields. A shell ring site was identified by a site survey
team in 1972.
A grant (#GB-37889) to support excavations was awarded by the
National Science Foundation (Fairbanks and Milanich) in 1973. Field
research began in that year. Investigation of the changing relationships
between man and his environment during the period 2000 BC and AD 1865
was the organizational theme of the project. The presentation and
discussion which follows concerns the Late Archaic aspect of the project
(circa 2000-1000 BC). Field research and analysis were also supported
by funds from the Sea Island Foundation.
Southeastern Shell Ring Sites
Structures of unusual size, location, composition, or configuration
have been popular subjects for study and speculation: pyramids, henges,
and earthen effigy mounds are but a few examples. Accounts of early
travelers, proceedings of learned academies, and personal letters
document the interest in such sites. William McKinley of Milledgeville,
Georgia, wrote of circular shell heaps on Sapelo Island, Georgia, in 1872.
His letter to the Smithsonian Institution is included in the Annual Report
for that year (McKinley 1873: 422-428). Apparently this is the first
mention of a shell ring site. The first excavations at Sapelo were some
twenty years later. Clarence B. Moore and field party conducted excavations
within the largest shell enclosure (1897: 71-73). As a result of Moore's
coastal investigations, two other shell ring sites were reported (1899).
However, because large quantities of artifacts were not forthcoming, Moore's
group shied away from further excavations in ring sites. Fifty years
later, the first scientific investigation of a shell ring was headed by
Antonio J. Waring, Jr., in the largest Sapelo Island ring (Waring and
Larson 1968: 263-278).
Circular accumulations of shell dot the barrier islands along the
coast of South Carolina and Georgia. Shell ring sites are located on the
first barrier island, or as is more commonly the case, on islands in the
lagoon and estuaries between the initial barrier and the mainland.
Ring sites are frequently located in salt marsh environments. At this time,
known distribution is limited to the states of South Carolina and Georgia.
South Carolina reports twenty-two; Georgia has at least half that many.
Table 1 summarizes published distributional information. It is possible
that two additional shell ring sites exist in Florida (Talbot Island,
Jupiter Island). The cultural inventories of these sites do not exhibit
any similarity to Georgia and South Carolina sites. Oysters (Crassostrea
virginica) are the most abundant constituents, but other species are
present, particularly clam (Mercenaria sp.) whelk (Busycon carica), and
ribbed mussel (Geukeysia demissa). While most species are estuarine,
several terrestrial forms are observed. In addition to molluscs, sediment,
vertebrate remains and human cultural materials are present.
Dimensions of the shell rings are extremely variable (Table 2).
In general, diameter ranges from 37 to 46 meters. Height also varies
considerably, a result to some extent of time and weathering. All
measurements are given in metric. While a few rings have a complete
circular contour, most are incomplete in some aspect. Erosion, shell
borrowing, and initial construction design may account for this.
Speculations regarding ring function began with McKinley's (1873)
suggestion of use for games, ceremonials, or torture. Since that
time, additional hypotheses have been offered. Moore (1897), as
another possible explanation, considered fortifications. Waring and
Larson concluded that the Sapelo Island rings were public in nature and
FIGURE 1. Shell Ring Site Distribution.
Table 1: Site Distribution, South Carolina and Georgia
Site Designation Location Source
Yough Hall (Auld)
Fig Island 1
Fig Island 2
Fig Island 3
Leeward of Capers Island
Leeward of Capers Island
Leeward of Capers Island
Leeward of Dewees Island
Leeward of Seabrook Island
Port Royal Island
Guerard Point 38BU21
Hemmings and Waddell n.d.
Hemmings and Waddell n.d.
Hemmings Field Map
Crane and Griffin 1964:9-10
Hemmings and Waddell n.d.
Hemmings and Waddell nd.
Hemmings and Waddell nd.
Hemmings and Waddell n.d.
Hemmings Field Map
Hemmings 1970a, 1970b
Hemmings Field Map
Hemmings and Waddell n.d.
Table 1 Continued.
Site Designation Location Source
Large Ford Ring
Small Ford Ring
Sea Pines Ring
Ring near Skidaway
Sapelo Ring 1
Sapelo Ring 2
Sapelo Ring 3
A. Busch Krick
Ring at Cannon's Pt.
Hilton Head Island, north
Hilton Head Island, north
Hilton Head Island, south
St. Simon's Island
St. Simon's Island
St. Simon's Island
Waring and Larson 1955-60
Crusoe and DePratter n.d.
Table 2: Shell Ring Dimensions (in meters)
Site Diameter Height Base Width Area Excavated Source
A. Busch Krick
Ring near Skidaway
represented some ... ceremonial or social arrangement ... (1955-60:
273). After excavations in the Sewee Mound ring in South Carolina,
Edwards concluded that the site represented the remains of an aboriginal
fishtrap (1965). Refuse debris from internal pile dwellings is another
suggestion (Caldwell as quoted in Crusoe 1973). Unpublished sources
have contended that these sites were supports-for structures built
directly on the midden accumulation or protection fot dwellings
It should be mentioned at the outset that some argument exists
regarding ring origin as being of natural construction rather than
human. Proponents of the natural origin view contend that redeposition of
naturally occurring oyster bars and human cultural debris by storms and
currents is responsible. Clarence B. Moore was convinced of their
human origin but interested himself in determining them to be aboriginal
rather than English. This study suggests human construction responsible
for ring sites. Data supporting this contention will be presented
in Chapter 3.
Previous Investigations and Interpretations
An examination of the results of four shell ring site excavations is
made at this point since information from these sites comprises the foundation
for existing interpretations of prehistory during the Late Archaic.
The relevant sites are Sapelo Island Ring Number 1 (Waring and Larson 1955-60
263-278), Sewee Mound (Edwards 1965), Hilton Head Island Rings (Calmes
1968), and Fig Island (Hemmings 1970a, 1970b). Because the following
comments are general considerations, the reader is referred to Appendix 1
for a detailed summary of information available for these sites.
Previous excavation strategy followed two basic approaches. Waring and
Larson and Hemmings excavated 10-foot wide trenches from approximate
ring center through the shell deposit to the exterior edge. Edwards and
Calmes selected a smaller sampling size (5-foot square units) and
tested more locations within the site.
Excavation techniques varied significantly. No use of screens is
reported by Waring and Larson or Calmes. Hemmings used 1/4-inch
screen as a standard at Fig Island and relied on window screen for separating
material recovered from features. After observing the quantities of
faunal material lost through 1/4-inch screen, Edwards changed to 1/8
inch. Additionally, he used water separation in the screening process.
Location of sites presented some logistical problems for investigators.
Hemmings and Edwards experienced tidal innundation during excavation
because of low elevation above sea level. The higher elevation of the
Sapelo Island sites precluded such problems with tidal water or elevated
ground water levels during excavation.
Reporting of information is sketchy. Waring and Larson's (1955-60: 263-278).
report appears in The Waring Papers, published in 1968. Until a few
years ago, this report was considered a rather thorough job. In comparison
with more recent archaeological research, it lacks information on many
relevant questions. Calmes (1968) brief paper comprises the total
published information on the Hilton Head Island sites. Oriented
toward establishing chronological placement of the sites on the basis of
ceramics, it offers very little else. Edwards' (1965) report is
entitled "preliminary." No further analysis has been made public.
Information on Hemmings' Fig Island excavations are contained in two
papers dealing very generally with that site. Faunal information for
this site is available from the Zooarchaeology Laboratory, Florida State
Museum. The Institute of Archaeology and Anthropology of the University of
South Carolina has graciously made their copies of Hemmings' artifact
Data are variously interpreted. The following discussion summarizes
interpretations of excavators and other southeastern prehistorians;. The
origins of human inhabitants of the coastal area during this time period
(circa 2500-1000 BC) are obscured by the fact that there is no solid
evidence for an inland-coastal cultural continuum. Sites dating within
the time span circa 4000-2000 BC are absent in the coastal plain.
Waring (1940:191) posits an inland cultural development for coastal
peoples based on lithic technology. Since no lithic material is
naturally available in the coastal strand, Waring and lately Crusoe and
DePratter (n.d.:15), agree that previous cultural development in an inland
area is probable. The extensive shell heaps in the states of Tennessee,
Kentucky, and Nothern Alabama are offered as examples of possible
sources of a cultural tradition that later adapted to the lower
Atlantic coast (Stoltman 1972:53). Basic similarities in riverine and
coastal subsistence are also offered as further substantiation of the
riverine origins of these peoples. Riverine mollusc gatherers are
believed to have made the adjustment of estuarine mollusc procurement while
fishing and hunting techniques required only slight reorientation to.the
Analysis of the cultural remains rests on three important artifact
classes: pottery, projectile points, and worked bone. Shell tools, a
common component of cultural inventories in coastal sites are not
distinctive enough for the purposes of tracing cultural affiliations.
Ceramics appear early in this geographical area and are very distinctive
In the United States, ceramics initially occur in the southeast and are
dated at circa 2500 BC (Stoltman 1966:872). These ceramics are fiber-
tempered; a typological distinction based on the inclusion of vegetal
material as a tempering agent. A plain ceramic horizon preceded decorated
ceramic types (Waring and Larson 1955460; Hemmings 1970b; Milanich 1971).
Evidence for this progression is observed at Bilbo, Sapelo Island, and
Stallings Island (located at the Fall Line on the Savannah River near
Augusta, Georgia). Ceramics are variously termed: Stallings, Bilbo, St.
Simons. It should be noted here that the type site for fiber-tempered
ceramics is the Stallings Island site (Claflin 1931). It is riverine
in orientation. Subsequent reports of fiber-tempered ceramics in Georgia
and South Carolina were lumped under the Stallings nomen. Decorative
types included incision, punctation, combinations of these techniques, and
drag and jab (frequently called linear punctation). Ceramics are
apparently hand-molded as no evidence of coil fracturing is observed.
In South Carolina, ceramics are primarily sand-tempered and typed
Awendaw, Horse Island, or Thoms Creek (DePratter, Jefferies, and Pearson
1974). Radiocarbon dates indicate that sand-tempered and fiber-tempered
ceramics are at least partially contemporary in occurrence with the latter
occurring first. It is possible that Awendaw and Horse Island are coastal
variants of the Thoms Creek type which is identified for inland
contexts. Ceramic types overlap geographically. Fiber-tempered types
are reported from as far north as the Edisto River drainage in South
Carolina. Sand-tempered types are reported along the Savannah River
drainage from the coast inland.
Lithic artifacts associated with fiber-tempered ceramic sites (rings
and middens) are large, triangular stemmed projectile points usually
termed Savannah River Stemmed. It has been noted that recovery of any
quantity of lithic material from ring sites is not usual. These sites are
especially non-productive of lithic debris and worked examples. Bilbo,
a non-ring midden site located near Savannah, Georgia, yielded large
quantities of lithic artifacts (Waring 1940: 152-197). It is contended
that this feature is unusual in the coastal area (Crusoe and DePratter
n.d.; Marrinan 1973).
Bone tools include both crude and sophisticated workmanship. Some
forms are consistent and stylized enough to be described and given type
names. The long bones of deer were fashioned into several types of
artifacts, but the most distinctive one is the Bilbo-style bone pin des-
cribed by Waring (1940: 165-171). This artifact is made from a split
section of the deer metapodial and frequently retains a portion of the
proximal articular surface. Incised designs are occasionally present on
the surface. Geometric motifs are the usual decoration; naturalistic
designs are unknown. Waring contends that the style changes from a
"nail-head type" to an "expanded-headed type" (1940-42: 254). Thus the
latter pin style indicated a later chronological placement to Waring
Radiocarbon dates for fiber-tempered sites (both ring and midden)
fall within a 2500-1000 BC time range. This period has been termed
Stallings, St. Simon's, Bilbo, Sapelo-St. Simon's, and Fiber-tempered.
Chronological designations are categorical conventions used to organize
information, artifacts, and sometimes intuition. Using such a structure,
prehistory is examined and interpreted. At times, these exercises
become particularly cumbersome as the chronological dilemma of fiber-
tempered sites aptly exemplifies.
Understanding Late Archaic cultural developments on the Georgia
coast is complicated by the assignment of several typological names
derived from distinct geographical locales (e.g. Stallings, Bilbo, St.
Simon's). There is a lack of consensus regarding just what criteria
should be used in developing these designations: artifacts or adaptations.
There is also a tendency to use attributes of a single artifact as the
basis for developing typological levels.
Waring (1940: 160) observed that Stallings nomen cannot be applied to
the coastal sites since there were too many dissimilarities. These observed
differences were ceramic: vessel shape and thickness, execution of a
punctated design, presence of crude simple stamping on vessel bottoms
(1940: 160). He proposed a bipartite Bilbo nomen: Bilbo I with plain fiber-
tempered ceramics and Bilbo II with decorated fiber-tempered ceramics and
elaborately decorated bone pins. Williams (1968: 320-321) presented a
Stallings Island Culture composed of a Stallings Island Phase, a Bilbo
Phase, and a St. Simon's Phase. This structuring encompasses the geographical
aspect but considers little else.
Milanich (1971: 119-128) expressed dissatisfaction with the equating
of Stallings Island and the coastal cultural traditions. He developed a
bipartite division for the coast: Sapelo Phase and St. Simon's Phase.
The Sapelo Phase is associated with shell ring sites but temporally overlaps
the midden-related St. Simon's Phase. These phases mark the beginning of what
Milanich called the "Coastal Tradition" and are based on radiocarbon
dates, differences in subsistence orientation, distinct settlement types, and
artifact inventories. Crusoe (n.d.: 4-10) presented a four-part
division (Stallings 1-4) articulating the Georgia area with Bullen's
(1954: 47) Orange chronology for Florida. The criteria for these divi-
sions are primarily changes in manufacture and decoration of fiber-tempered
ceramics. It is rather obvious that no consensus exists. Table 3 summa-
rizes the different chronologies.
The Cultural Phenomena
At this point, it is pertinent to summarize the cultural phenomena
peculiar to the coast as exemplified by fiber-tempered ceramic sites in general
and shell ring sites in particular. The explication and explanation of these
cultural developments are the tasks to which this study is addressed.
A. Reliance on molluscs is the most obvious characteristic. This
subsistence strategy is held responsible for influencing settlement pattern
and allowing more sedentary existence. Lengthy settlements are inferred
from the extensive and deep deposits. Habitation areas during the time
period circa 2500-1000 BC are of two types: ring and amorphous midden.
Both are primarily composed of shellfish refuse. Some researchers consider
coastal inhabitants derived from inland riverine shellfish gatherers (e.g.
Tennessee River Valley Archaic people).
B. Ceramics first appear in the United States at this time. The
inclusion of ceramics in a site inventory once lent credence to claims
of sedentism (as well as horticulture). A nomadic existence was said to
be ruled out by the problems of transporting ceramics. In Georgia
and extending to the Edisto River drainage of South Carolina, fiber-tempered
ceramic types are recovered. For the remainder of the area of distribution
of shell ring sites (to the Santee River in South Carolina) and with some over-
lapping at the coast and inland in the Savannah River drainage, sand-tempered
C. Cultural inventories are sparse and quantities of recovered arti-
facts vary with location. There is also some suggestion that inventories
Table 3: Chronological Summary.
Waring Williams Milanlch Crusoe
APPLICATION Inland and
CRITERIA Bilbo Period
plain or simply
head bone pins
baked clay objs.
Stallings Is. Phase
St. Simon's Phase
Stallings Is. Phase
Stallings Is. and
St. Simon's Phase
Mclntosh and Glynn
St. Simon's Phase
Georgia and South
St: Simon's Phase
coastal and inland sites for
duration of fiber-tempered
I -absence of decoration
II -punctated and incised
Ill-grooved designs and
IV -pottery began to be made
with other tempering
Table 3 Continued.
Waring Williams MTlanich Crusoe
CRITERIA Bilbo II
expanded head elab-
DATES 2000-1000BC Sapelo I 2500-1800 BC
1000-500 BC St. 11 1800-1500 BC
Simon's Phase III 1500-1200 BC
IV 1200-100 BC
SOURCE Waring 1968:254 Williams 1968:320-1 Milanich 1973:51-53 Crusoe n.d.:4-10
vary with type of site (i.e. middens may be more productive than ring
sites). Sites nearer the Savannah River contain more lithic material.
Bone pins are a very important diagnostic artifact but their function
is uncertain. Cultural inventories are composed of subsistence-
oriented items made of locally available materials.
D. At approximately 1000 BC there is an end to this lifestyle.
Molluscs decline in subsistence importance, new ceramic types
utilizing different construction and tempering techniques appear, and
cultural inventories change.
This then was the cultural problem to be archaeologically
investigated and interpreted. A better explanation of the particular
adaptation made by the occupants of the coastal area was obviously a very
real need considering both the data base available for archaeological
interpretation as well as the interpretations generated.
EXCAVATIONS ON CANNON'S POINT PART 1: SUMMER 1973, SPRING 1974
The Atlantic mainland is protected by barrier islands along the
majority of its extent. These islands are characteristically elongated in
contour and result from deposition of water- and wind-borne sediments.
Width is usually limited to several kilometers, but length may exceed 160
kilometers. St. Simon's Island is one of the southern barrier islands
along the Georgia coast. Located between Sapelo and Jekyll Islands, it is
approximately 57 kilometers from the Georgia-Florida border (Fig. 2).
The northern end of St. Simon's Island is divided into two points of
land: Butler Point on the west and Cannon's Point on the east (Fig. 3).
These points are separated by an expanse of salt marsh through which Jones
Creek meanders. The archaeological project area comprises the entire
eastern point. To the north, the area is bounded by the Hampton River;
on the south by a property line fronting on Lawrence Road. Little
St. Simon's Island forms a protective barrier to the north against the
Altamaha River and against the Atlantic Ocean on the east. Consequently,
the Cannon's Point area is sufficiently distant (3.8 km) from the first
barrier that vegetational wind forms produced by increased salt content in the at-
mosphere are observed only in a very marginal fringe along the marsh edge.
Extensive changes in the floral cover of Cannon's Point have been
affected since the Plantation Period by land clearing for cultivation of
commercial and staple crops and by free ranging cattle and pigs. In a
FIGURE 2. St. Simon's Island and Adjacent Islands.
FIGURE 3. The Cannon's Point Project Area.
very limited way, this practice is continued today. During the Plantation
Period, a lane of oaks was planted along the main access to the planter's
home. John Hamilton Couper whose family owned the Cannon's Point Plantation
from 1794-1861 was an ardent amateur naturalist who introduced many
different plant species (Lyell 1849). Land clearance probably left few
areas of Cannon's Point untouched. In some-areas, live oaks (Quercus
virglnlana) are all the same size and pines are heavily concentrated in
other areas (it Is not clear if this condition represents planting of pine
or regrowth of a previously cleared area).
Floral cover of the Marsh Ring site (9GN57) is presently dominated by
cedars (Juniperus virginiana) which evidence wind forms and stunting. A
few oaks are observed. The understory is sparse but yaupon (Ilex!vomitori )
and marsh (Iva sp.) elder are present. Yucca (Yucca alterniflora) grows
on the highest elevation of the ring. Because the ring is open on its
northeast arc, a marsh vegetation composed primarily of Spartina
alterniflora grows within. Outside the ring stretches an extensive salt
marsh dominated by Spartina and Juncus romerianus, interrupted only by
meandering tidal creeks. Presently, the closest tidal creek is located
at approximately 160 meters distance. Nearest high ground is 35 meters
southwest, the edge of which is fringed by palmetto (Serenoa reopens ,
yucca, oak, and cedar. The West Ring (9GN76) Is located some 85 meters
southwest of the marsh ring on high ground. On its south arc, this ring
erodes into the marsh. Oaks, magnolia (Magnolia grandiflora), cedar,
hickory (Carya), and hackberry (Celtis) are common in the floral cover of
this area. Yucca, palmetto, Smilax sp., yaupon, and grapes (Vitis)
contribute to a rather dense understory in the immediate ring environment.
Spanish moss is found within the forest margin.
In the Hampton River, the average annual tidal factor is guaged at 2.1
meters (United States Commerce Department 1954). At periods of high tide,
the marsh ring is completely surrounded by water and the center innundated.
Fiddler crabs (Uca spp.) abound in and around the marsh ring. A species
of land crab Is frequently observed in the trees atop the marsh ring but
incidence drops as one progresses Into the forest. Salicornia virginica
a common marsh plant is frequently observed in the moist areas near the
land edge (Carlton 1975:13). A small freshwater stream possibly the result
of ditches constructed during the plantation period empties into the marsh
west of the ring sites. Water is caught in a shallow depression inland
and flows easterly into the marsh during wet periods. Flow is minimal
during dry periods. Although topographic maps show several artesian
wells in the area of Cannon's Point, none are presently free-flowing.
Terrestrial animals seen in the vicinity of the rings include birds,
mammals, and reptiles; both native and introduced forms. Deer, raccoons,
and opossums have been observed during daylight hours and after dark.
Feral pigs, cows, and horses are presently free in the area. Marsh
hawks (Circus cyaneus) and red-tailed hawks (Buteo Jamaicensis), are the
largest predatory birds seen. Turkey vultures (Cathartes aura) are
frequently observed. Wood ibis (Mycteria americana) and herons (Ardea
herodlas and Hydranassa tricolor) are common shore birds. Red-winged
blackbirds (Agelaius phoeniceus), mourning doves (Zenaidura macroura),
pleated woodpeckers (Hylatomus pileatus), painted buntings (Passerina
ciris), and blue jays (Cyanocitta cristata) are frequently observed in
the immediate surroundings. Diamond-back terrapins (Malaclemys terrapin)
and musk turtles (Kinosternon sp.) were observed on the ring sites.
Several snakes were present: black racer (Coluber constrictor), indigo
(Drymarchon corals), and corn snake (Elaphe guttata).
Excavation Strategy: Objective and Techniques
A review of available shell ring literature initially suggested the
following problems for investigation:
a. ring function
b. subsistence reconstruction
c. radiocarbon dating for southern distributional extent
Determining the manner by which the site was deposited offered the
most plausible means of approaching the question of function. Utilization
of large, adjacent excavation units was thought to provide the most
efficient means of examining stratigraphy, changes in contour, or features
related to structures on or in the ring. A 3 x 3 meter excavation unit
size was selected. Additionally, systematic mapping of cultural and
subsistence material was planned to overcome the difficulties of excavating
a shell site.
Screening was considered imperative to reconstructing aboriginal
subsistence strategies. Past reports had all too casually dispensed with
this aspect. One-fourth inch screen was selected for standard screening
with additional reliance on flotation of material from features. Quantities
of faunal material were anticipated, but recovery of floral remains was
considered particularly important. Flotation was intended to provide the
means of recovering floral material.
Radicarbon dates were similarly considered important. The Cannon's
Point excavations represent a southern extreme of shell ring distribution.
On the basis of the nearest similar site, a relative date of circa 1500 BC
was selected as a working date. The Sapelo Island Number 1 Ring had an
average radiocarbon date of 1750 BC and primarily plain ceramics. The
marsh ring was known to have decorated sherds from a surface collection made
in 1972 by the author.
Summer 1973: Excavation Summary
Field crew members numbered five persons for the ten week session.
A section of the south arc of the marsh ring was selected for excavation
because of the absence of tree cover. A transit station (ON,OE) was
established on the southwest arc and tied to a benchmark on the nearest
high ground (Fig. 4). A surveyor's stake from the 1970 property survey
was located and a large cedar next to this stake became the benchmark.
An east grid line was set out with stakes at 3-meter intervals. From
this line, a north-south stake line was set out at 15E and 18E. The area
included a 3-meter wide cut across the ring from south to north. Originally,
excavating a north-south trench across the ring was planned.
The matrix encountered in the south arc was a sticky, chocolate
colored clay. It was impossible to separate material in the screen.
Midway through the field session, a pump was secured allowing water
separation of material. This procedure facilitated processing of excavated
midden material and undoubtedly increased recovery. Water was obtained
from two reservoir sumps dug in the marsh floor south of the ring. While
digging the second of these pits, considerable aboriginal cultural
materials were exposed at a depth of 0.6-0.7 meters below the present
marsh surface. Additionally, the stumps of two trees were observed but
no shell deposit was seen.
FIGURE 4. The Marsh Ring (9GN57) and Marsh Excavation Units.
In the midden excavation units (a6S, 15E and 09S, 15E), removal of
material proceeded in 15 cm levels; actual midden removal was accomplished
by 5 cm layers. Initially three and four prong rakes were used but
these were discarded since it became apparent that they posed problems for
vertical control as well as the trauma to midden inclusions. Trowelling
was a slower but more precise means of removing material.
Tests were begun in the ring center (OON, 15E and 03N, 15E). Stripping
of superficial marsh grasses was done by hand. Initially, screening was
attempted but this proved productive only in areas where midden was
projecting into the excavation units (in DON, 15E only). Finally, a
I x 1 meter sondage was made prior to leaving the field. This test
revealed a sandy deposit below the marsh grass and was virtually sterile.
Excavation units in the ring midden provided quantities of faunal
material. In addition, a sizeable ceramic sample was obtained. Sherds
were all fiber-tempered. Plain ceramics were more frequent than decorated.
Motifs included punctation, incision, and combinations of these. Several
sherds suggestive of Orange Series motifs were recovered. Shell tools,
basally blunted and having holes near the spire, were infrequent midden
inclusions. Several examples of worked bone were recovered but none were
decorated beyond circumferentially incised lines. Neither of the units
opened in the ring midden were completed.
Excavations below the marsh surface (027S, 18E and 0275, 21E) exposed
fiber-tempered ceramics in association with grit-tempered sherds. Faunal
remains were present but minimal. A few nutshells (Carya spp.) were
recovered. These findings suggested a now-drowned cultural deposit made
at a time when the environment in the vicinity of the marsh ring differed
Testing in the ring center was not extensive and revealed no features
suggesting either structures or disturbance. In the excavation of oON,15E,
the extent of midden deposit contracted perceptibly toward the ring
evidencing slump. This indicated that the ring was higher and basally
thinner than its present dimensions.
Spring, 1974: Excavation-Summary
The field crew numbered nine people for the ten week session. Of
initial concern was the completion of the two opened units on the south arc
of the ring. it soon became evident, however, that this pursuit was an
impossibility given the tidal conditions. Ground water levels were too
high and continuous pooling of muddy water forced abandonment of this
area. Attention was turned to the area of highest elevation on the north
side of the ring. The grid system was extended northward and two adjacent
units (ol8N,OE and I18N,3E) were opened in the area of maximum height.
These 3x3-meter units extended along the apex of the ring.
As a direct result of our 1973 findings, changes were made in both
excavation techniques and In processing. Excavation preceded by 5 cm.
levels to allow observation of changes in cultural material, flora, and
fauna. A 30 cm.2 column sample area was reserved in each square to allow
qualification and quantification of midden constituents. Screen size was
reduced to 1/8-inch. This change was motivated chiefly by dissatisfaction
with previous floral recovery. Two standing screens having 3/4x3/8-inch
expanded mesh upper screens and 1/8-inch detachable underscreens had been
constructed. Use of flotation sampling was again planned for features and
areas of high humic content.
A sump was dug several meters off the north edge of the ring. Again,
ceramics were exposed associated with the remains of trees. No aboriginal
flora or faunal materials were recovered. The grid was extended to enclose
this area (041N,12E) and the finds were mapped in place. The original
intention in digging the sump was to allow processing of excavated
materials on the north side of the ring. Wheelbarrows were used to
transport excavated midden to the screens on the marsh floor for processing.
This practice eliminated problems from backdirt.
In addition to the sump excavation unit (ahlN,12E), two additional
units were opened In the marsh north of the ring (o54N,12E and c80N,12E).
A small quantity of ceramics and one historic or modern fencepost were
recovered. The post was cedar and had bark adhering to the exterior.
As the distance of excavation units increased from the vicinity of the
Marsh Ring, the incidence of cultural material decreased.
A superficial scattering (0-10 cm.) of grit-tempered sherds was
observed in the north ring excavation units. No decorated ceramics were
recovered during this field session. One-half meter of midden material
was removed and processed from each square. One shell disc bead and one
bone point were recovered in the screens. Faunal recovery was heavy;
flora, particularly Carya and Celtis were frequently observed in the matrix
and in the screens.
Because the excavations in the marsh north of the ring had shown no
concentration of cultural material, attention was turned toward the south
marsh. Two adjacent 3x3-meter units were opened (033S,15E and 033S,12E).
Both squares proved productive far in excess of our expectations.
Excavation proceeded by shovel until cultural material was encountered.
Usually fragments of carbonized wood were exposed Immediately above the
cultural deposit. Once cultural material had been exposed, trowels were
used. Cultural materials and features were mapped in place for future
study of distribution and associations. A Igrge stump, apparently burned
in place was an interesting feature. Adjacent to it was a large cedar
root burned only on the side next to the stump_(west).
Attempts to return to the two south excavation units in the ring
proper continued to be frustrated by high tides. A new unit (030N,3E)
opened on the north arc of the ring at the intersection of midden and
marsh met with the same problems. It was possible, however, to note that
slump in this area had been more gradual than that noted in the south ring
A second ring site was identified by testing during the Christmas
holiday in 1973 (Fig. 5). The West Ring (9GN76) was located on land
southwest of the marsh ring. It was incomplete on the south side where
intersection with the marsh occurs. Presently the site is located beneath
a substantial humic layer from hardwoods and thick understory. A 2x2-meter
unit (Test 1) had been opened on its north arc and screened using 3/4x3/8-inch
expanded metal screen. Fifteen centimeter levels were used but time did
not allow completion of the test. After the unit had been excavated to
45 cm. below the surface, only the southern 1x2 half of the square was
excavated to sterile (65 cm.). Sand-tempered sherds and one possible
complicated stamped sherd were recovered in the first level; below this
FIGURE 5. The West Ring (9GN76) Diagrammatic Representation.
I 1 ' '' '.'r ?
'0 i 7 .*Is.L9 .'..cf1
'-~.4 ' *
'J*& r'. 'C'
..r ~~~ I *, :;I:-1
-I: *. -i "& .4'$.t *! I* ,
'I..,~ 'I ** *' 71
It .' I:''".
decorated and plain fiber-tempered sherds were collected. Harsh
periwinkle (Littorina irrorata) shells were recovered In greater frequencies
In this test than in the marsh ring. Their use as food items was indicated
by the consistent removal of the spires of each shell.
During the Spring 1974 Field Session, the unexcavated northern 1x2
meter half of Test I was excavated and processed using water separation
over 1/8-inch screen. The results presented in Appendix 2 offer a strong
case of the value, in terms of recovery, for use of fine screens in shell
sites. A second 2x2 meter test (o5S,30E) was opened on the east side of
the ring. Material was removed in 5 cm. levels. The unit was not completed
prior to the close of this field Session.
After the first two field sessions, numerous observations were
stated preliminary to formulating hypotheses by which to structure the
following Summer Field Session excavations. There were many interesting
aspects of excavation findings to consider for more vigorous testing. An
assessment of these observations follows.
A. Environmental change: The presence of a cultural deposit below
the present marsh surface suggested an instability in the marsh in this
area. Evidence of the continuing changing relationship of marsh and high
ground on the east side of Cannon's Point is observed by the marginal
area of dead and dying trees along the marsh edge. The marsh was thus
perceived as a dynamic factor in the Cannon's Point environment. The
fact that the ring, at deposition had stood in a forested area indicated
that the "fishtrap" hypothesis was not valid.
B. Ceramics: A plain to decorated ceramic continuum seemed to be
indicated by preliminary findings from the south ring at excavation units.
Decorated ceramics were confined to the upper three levels (0-45 cm.).
Below that, to the depth that excavation had progressed (90 cm.), only
plain ceramics were recovered. A cultural level having fiber-tempered
ceramics and grit-tempered ceramics located 0.5-1.0 meters below the
present marsh surface suggested an occupation following the ring-
building inhabitants and not associated with mollusc midden debris. While
flora, fauna, and cultural material were recovered, no mollusc concentration
was exposed beneath the marsh surface. Some contact with the St. Johns
River area of present day Florida was indicated by the presence of Orange
Series ceramic motifs. These are illustrated in Appendix 3. Evidence
of this contact had been initially observed by Holder (1938) and noted
again by DePratter (n.d.) at the Bony Hammock site on St. Simon's Island.
C. Lithics: An increased amount of lithic material, slight though
it was, in the marsh cultural level and the West Ring indicated an
availability and use not demonstrated in the Marsh Ring. This suggested
a change related to the increased amounts of deer present.
D. Subsistence: Marked differences in faunal recovery were obvious.
For example, fish remains were numerous in the Marsh Ring faunal assemblage
but the lower levels of the West Ring provided evidence that use of fish
had far greater importance (both by number of species identified and number
of individuals) to the Inhabitants of that site. Marsh periwinkles,
small marsh-dwelling gastropods were found in the West Ring midden used as
food items, suggesting a need to utilize less productive resources.
Periwinkles have far less food value to offer for the time expended in
gathering and preparation. The midden deposit was much thinner (0.65 cm)
in the West Ring than in the Marsh Ring (average 1.3 m). Appendix 5 clearly
shows that faunal elements present in the marsh cultural level were higher
in numbers of deer and small mammals. Reptiles were noticeably more
frequent. Fish, while present, were not numerous.
E. Dates: The working date, selected from a comparison of the Sapelo
Island Ring (Number 1) suggested a 1000 year-time span (marsh ring, West
Ring, and marsh cultural deposit). Because grit-tempered ceramics
included no check-stamping, a later decorative trait, a terminal date of
750 BC was selected. The working time range was then 1750-750 BC. Table 4
lists radiocarbon dates for shell ring sites.
After considering these points, an hypothesis based on marsh change
to account for the observed differences in faunal procurement offered a
point from which future excavations could be directed and analysis structured.
Test implications supporting such an hypothesis basically depended on
continuing support for the observations noted above. The proposed 750 year
time span did not seem unreasonable given the three cultural entities.
Length of debris accumulation was estimated from Hilton Head Island sites.
Calmes (1968) reported that the Sea Pines ring had two feet of deposit and
totalled approximately 300 years. On Hilton Head Island, Calmes (1968: 45-58)
obtained radiocarbon dates for the Ford Rings (Large and Small) representing
a total deposition time in excess of 800 years (eight feet and two feet of
midden deposit respectively). Test implications would be examined through
extending excavation in the West Ring and marsh cultural deposit.
Table 4: Radiocarbon Dates for Shell Ring Sites.
Site Date Material/Significance Source
Large Ford Ring
Small Ford Ring
Sea Pines Ring
A. Busch Krick
1820 BC:3770 + 130 BP (M-1209)
1635 BC:3585 + 115 BP (1-2850)
1170 BC:3120 + 110 BP (1-2849)
1940 BC:3890 + 110 BP (1-3047)
1450 BC:3400 + 110 BP (1-2848)
1160 BC:3110 7 110 BP (1-2847)
1650 BC:3600 + 350 BP (M-39)
1850 BC:3800 350 BP (M-39)
1265 BC:3215 + 80 BP (UGA-226)
1520 BC:3470 T 85 BP (UGA-227)
1815BC:3765 + 90BP (UM-521)
2240 BC:4190 + 90 BP (UM-520)
1655 BC:3605 + 110 BP (UM-523)
1910 BC:3860 + 90 BP (UM-522)
Oyster: Upper level
Charcoal: 56-57" (Level 9)
Oyster: 27" (Level 4)
Charcoal: Basal level
Clam: 20-26" (Basal level)
Conch: 0-6" (Upper level)
Oyster: Dates early ceramics
Conch: 5.7-6.0 ft.
Charcoal: 4.6 ft.
Oyster: Dates upper level
Oyster: Dates lower levels
Oyster: Dates upper levels
Oyster: Dates lower levels
Crusoe and DePratter
Completion of all opened excavation units in the marsh ring would end
excavation at that site. Analysis would be oriented at reconstruction of
the aboriginal environment.
EXCAVATIONS ON CANNON'S POINT PART II: SUMMER 1974, WINTER 1975
Formulation of Hypotheses and Test Implications
Information from analysis was at this point adequate for the evalu-
ation of research directions. Each of the previously stated observations
was carefully considered. Overall, the most concrete and interesting
aspect of the excavation was the indication that a change in resource
usage occurred. Using the marsh ring as a base, change is reflected in
the West Ring and the marsh cultural level. The question of cause was
posed. The approach then taken considered the possibilities that such
a finding could have cultural cause, environmental cause, or a combination
of these factors could be responsible. It was decided to pursue causal
statements within the ecological realm since the cultural status of these
populations was not sufficiently developed (i.e. social organization, popu-
lation density, population size, etc.).
Several hypotheses were generated and suitable test implications
Ho: No perceptible change in resource usage is evident
among the Late Archaic sites in the Cannon's Point
HI: A perceptible change in resource usage is evident
among the Late Archaic sites in the Cannon's Point
11: Inter-site vertebrate inventories are comparable.
12: Inter-site molluscan usage is comparable.
This hypothesis offered the ability to objectively test the observation
that a change in resource usage was in fact observable and repeatable
through new excavations.
H : During the period of occupation (ca. 1500-750 BC) no
changes in the marsh occurred.
Hi: During the period of occupation, marsh encroachment
and silting occurred.
I1: Submarsh shell midden having fiber-tempered ceramics
is not found.
12: Changes in the marsh occurred subsequent to the
This hypothesis was offered both to suggest causality and establish the
time of the proposed ecological change.
Ho: Marsh silting and land encroachment would have no effect
on molluscan and vertebrate fauna availability.
HI: Marsh silting and land encroachment would have an effect
on molluscan and vertebrate fauna availability.
IY: No effect can be documented in the literature.
12: No effect can be demonstrated in excavation findings.
These hypotheses offered a framework within which to structure the
Summer and Winter Field Sessions. Specific excavation goals were devised.
Extended excavations of the marsh were planned. Excavations in the marsh
north of the marsh ring indicated that the cultural concentration evident
on the south side of the ring was not present. Four units had been opened
in the marsh south of the ring and at least six more were planned to allow
testing of areas between the marsh ring and land. Determination of the
limits and focal areas of the cultural deposit was a goal. Testing east
of the ring was planned. If time allowed, additional testing of the West
Ring was planned. However, the investigation had been initially committed
to a strategy of collection of deep stratigraphic samples of floral,
faunal, and cultural materials from the marsh ring and no reason to aban-
don these priorities was seen.
Recovery techniques would continue in the same manner. Trowelling
of exposed material in the marsh excavation units was believed to be com-
parable to removing the material in a level. Mapping of in-place finds
was particularly important in determining the depositional nature of the
marsh cultural level.
Summer 1974: Excavation Summary
Only six weeks were given to excavations in the Late Archaic component
of the project locale during this field session. Work proceeded on the
north marsh ring units. Three additional units were opened in the marsh
south of the ring (l15S, 6W; 030S, 3W; 030S, 12W). Several levels of the
remaining uncompleted West Ring unit (1SS, 30E) were processed. Crew mem-
bers numbered between 12 and 15 persons daily.
Approximately one meter's worth of midden material was processed from
the north ring units (Gl8N, OE; 018N, 3E). Of special interest were two
features: Feature #18 (l18N, 3E) and Feature #21 (318N, OE). Feature #18
had initially been defined during the Spring 1974 session as an area of
increased incidence of floral and faunal material with a high humic content.
The feature was sampled for flotation on several occasions. A large crab
concentration (Callinectes sp., approximately 100 individuals represented)
was exposed during the Summer Field Session. Feature #21, also initially
defined during the Spring Field Session, was observed to be an area of
dark brown soil within the shell deposit. As excavation progressed, a
level of cemented shell was encountered. The most visible aspects of
this feature were the numerous hackberry seeds and the ashy condition
of the shells. This feature was assumed to represent a hearth. No decora-
tion was obvious on any of the ceramics recovered from the north excava-
tion units to this point.
One south ring unit (a65, 15E) was reopened but tidal problems con-
tinued. A well point was positioned in the southern (and lowest area)
portion of the square against the baulk, but problems with pump availa-
bility prevented continuous use of this means of water evacuation. The
unit was again closed after the identification of Feature #23, an area of
blue-grey sand believed to represent submidden soil mixed with obviously
weathered oyster and clam shells.
The marsh excavations were notable. Each unit contained several
lithic fragments, most of which were worked. Additionally, some worked
bone was recovered. Deer antler was a more common item than previously
observed. Ceramics continued to be exposed in the same associations.
While carbonized material was recovered in each unit, no large concentra-
tions of tree stumps or roots were encountered.
A I x 1-meter test (o6N, 8E) was opened in the shell projection in
the center of the marsh ring. The object of this test was determination
of ceramic contemporaneity of this deposit. Ceramics were all plain fiber-
tempered sherds. Sterile was reached at 0.75 meter below the present sur-
face. Material was removed by trowel and hand rake and processed through
3/4 x 3/8 inch mesh using water.
Winter 1975: Excavation Summary
With the exception of a brief look at submidden sterile in a 1 x 1-
meter test (06N, 8E), commitment to a rather laborious processing of
midden material resulted in uncompleted excavation units. An overriding
objective of the Winter Field Session was the completion of excavation
units. Crew members numbered six for the 10-week session. Weather was
expected to be problematical but fortunately was mild.
The first good look at submidden sterile-came one month into this
field session when O18N, OE was completed. Submidden was reached at 1.65
meters below the surface. Two features (#26, #27) were defined in the
interface. Both were pottery concentrations having plain fiber-tempered
sherds with the exception of one curvilinear incised and punctated fiber-
tempered sherd. A human cranium was recovered at 1.3 meters below the
surface, well within the midden deposit. A human pelvis was found at
approximately the same level but was three meters distant from the cranium.
Another remarkable feature (#24) was identified within the shell deposit.
This feature had a very high faunal content and was reminiscent of the
remains of a dumped cookpot.
Elevation of sterile was higher than anticipated. The midden deposi-
tion was apparently made in an area having a slight sandy knoll. It is
interesting to note that decorated ceramics were encountered below one
meter in depth. Decorative motifs were combinations of punctation and
Considering the time required to process midden material, it was
decided to excavate only one half of the remaining north excavation unit
(018N, 3E) to sterile. The southern half of this square was selected
because it contained Feature #18, an area of increased fauna and flora.
Sterile was reached at 1.65 meters. Decorated ceramics were encountered
at I meter and below. A projectile point and an elaborately engraved
bone pin fragment were recovered from the interface.
On the south side of the ring,Q6S, 15E was finally completed. Again
the interface was a highly productive area. A plain bone pin, a worked
lithic specimen, and a quantity of ceramics were recovered. Decorated
ceramics were recovered in this unit in Levels 1-3 (0-45 cm) and Levels
7-9 (90-135 cm). Sterile was reached at 1.3 meters below the present
surface. A well point was successfully used to evacuate water from this
unit. A human molar was recovered from the interface.
When the human cranium was recovered inDlSN, OE, no additional human
skeletal material could be observed. Since very little skeletal material
had been recovered from shell ring sites, examination of the area behind
(west) the baulk was indicated. A 1.5 x 2-meter excavation unit was
opened (019.5N, 2W) to expose the adjacent area. After removing 1.75
meters of shell overburden, no additional material was observed. No
screening was used, but a visual collection of flora, fauna, and cultural
material was made. Excavation of this area added a 2-meter expanse for
observation of the interface. More ceramics were recovered from this
area than in the midden. All ceramics were plain fiber-tempered.
Four excavation units were opened in the marsh: three south of
the ring (0455, 15E; 024S, 24E; o42S, 24W) and one east of the ring (oON, 58E).
In the east excavation unit, only three raccoon (Procyon lotor) teeth were
recovered after 0.75 meter of excavation. Instability of the floor (spon-
giness) forced a premature closure. The physical conditions encountered
in the unit were unlike any of the south units.
Each of the south marsh excavation units was remarkable in the quan-
tity of cultural material recovered. The southernmost, and least produc-
tive, square excavated in the marsh was 045S, 15E. The most westerly
and moderately productive was a42S, 24W. Additionally, below the cul-
tural level, white sand was exposed. This unit is closest to present
high ground. In the vicinity of the original marsh sump excavations of
1973, 124S, 24E was opened. A fragment of human cranium was recovered
from this unit along with the cultural and subsistence assemblage. It
should be noted that during this field session, an attempt was made to
screen material excavated from the marsh units. This proved unfeasible
because of the added water problem created by the water-screening.
There was too much additional seepage generated. If the screening could
have been done at some distance from the opened excavation unit, it would
prove more acceptable.
The West Ring excavation unit (a5S, 30E) was completed. In the basal
midden deposit, numerous potsherds were recovered. Sherds were plain and
decorated fiber-tempered ceramics. The decorative motifs were unusual
in comparison with those previously recovered. Combinations of puncta-
tion and incision in curvilinear and rectilinear designs were unlike the
primarily horizontal application usually observed. Quantities of fish
otoliths were recovered in the lowest levels of this unit in the same
manner as Test 1.
Finally, the last excavation unit to be opened in the vicinity of
the marsh ring wasdl2N, 15E, positioned in the center of the ring. The
location of this unit offered the ability to check the extent of the pro-
jection of shell in the center of the marsh ring. Previous testing had
revealed that it was contemporaneous to the surrounding midden on the
basis of ceramic constituents. However, as the overlying marsh accumu-
lations were removed, the projection was seen to extend into the center
much farther than indicated on the surface. Probing of this area revealed
that the situation with regard to extent of the projection is unclear.
Overlying marsh would need to be removed to make an assertion of extent
or absence in an area.
Summary of Excavation Findings
In terms of material excavated and information gathered, the Winter
Field Session had been one of the most profitable sessions. Excavation
units in the marsh ring had been completed providing an observation of
the condition of the submidden deposit. A second unit in the West Ring
and five additional marsh excavation units had been completed. All of the
previous conceptions of the progression of cultural events on the coast
were questionable in light of the information gathered. The working date
no longer inspired any confidence. Chronological placement by radiocarbon
dating could conceivably assign these sites early or late dates. Some
findings worthy of discussion follow.
The interface proved to be a particularly productive area. More
cultural material was concentrated in this area than any level of the
midden deposit or any feature. Elaborately engraved bone pin fragments,
supposedly a late cultural feature were recovered from the base of the
midden deposit. Decorated ceramics and two examples of lithic artifacts
came from this area of the site.
Plain and decorated ceramics occur in these ring sites without
stratification. Thus, a plain to decorated ceramic continuum could not
be supported by the findings. The fiber-tempered ceramics from the
north excavation units contained a perceptible amount of grit-tempering,
not observed in the ceramic samples from the ring sites. This suggested
a situation of transition in ceramic manufacturing techniques to grit-
Human skeletal material, though scanty, was not recovered in the con-
text of an articulated burial, but scattered throughout the midden deposit.
Further evidence of the forested nature of the environment prior to
ring deposition suggested that the site was constructed by human inhabi-
tants rather than natural forces. The fact that the interface represented
such a heavy concentration of human cultural remains was considered evi-
dence that the midden had begun to accumulate as the result of human
subsistence practices or intentional placement.
Evaluation of Hypotheses and Test Implications
As the Summer and Winter Field Sessions progressed, it became evident
that one hypothesis was unprovable given the extent, direction, and capa-
bilities of excavations and a second unsupported by excavation findings.
An examination of these hypotheses is made to point out the problems.
Hypothesis 3: Null Hypothesis Marsh silting and land
encroachment have no effect on molluscan and verte-
brate faunal availability.
While an effect upon molluscan and some vertebrate fauna by increased
amounts of sediment could be documented as possible (Buck 1956: 249-261;
Galtsoff 1956: 408-419), our excavation findings did not demonstrate such
evidence. No relict oyster beds were defined. While a probable margin
of inundated land can be observed from aerial photographs of the area,
no direct correlation between silting and encroachment could be demonstrated
for local molluscan fauna. The Null hypothesis, while not acceptable, is
similarly not rejectable either. This is a situation in which reformula-
tion and a new excavation strategy are required prior to acceptance or
Hypothesis 2: Null hypothesis During the period of
occupation (circa 1500-750 BC) no changed in the
Discernible changes in the marsh in the vicinity of the ring sites
occurred subsequent to the fiber-tempered occupation. Radiocarbon dates
indicate that inundation of the south marsh cultural level occurred after
835-820 BC. No submarsh midden deposit having fiber-tempered ceramics
was exposed by excavation supporting Test Implication 1. No molluscan
fauna was exposed in association with any material from the marsh cultural
level. Argument for accepting the Null hypothesis is felt to reflect the
findings at the site.
Hypothesis 1: Null hypothesis No perceptible change
in resource usage is evident among the Late Archaic
sites in the Cannon's Point area.
In the analysis of faunal material from the three cultural deposits
in the Cannon's Point area, consistent differences in faunal inventories
are observed. This involves molluscan and vertebrate faunal components.
Appendix 5 lists the results of the faunal analysis from the Late Archaic
occupations. Intersite vertebrate inventories are not similar. The West
Ring demonstrates a higher reliance upon fish. Analysis of fauna from
the marsh cultural level reveals an increased use of mammals and reptiles.
Intersite use of molluscan resources is not similar. The West Ring indi-
cates a subsistence use of periwinkle, an incidental mollusc in the
marsh ring. The marsh cultural level demonstFates no use of molluscs.
Therefore, neither of the test implications are supported. The Null hypo-
thesis must be rejected for lack of supporting evidence and the Test hypo-
ANALYSIS OF EXCAVATED MATERIALS
Radiocarbon dates were not available untjl May, 1975, well after
the close of the Winter Field Session. Material for dating was sub-
mitted to the University of Miami Radiocarbon Dating Lab.
Apparent age was calculated relative to 0.95X NBS
oxalic acid radiocarbon dating standard. Quoted
precisions are one standard deviation and include
only the counting precisions of the unknown sample,
background, and modern standard. (Eldridge 1975)
A discussion of the six samples, their proveniences, and significance
The Ring at Cannon's Point (Marsh Ring): 9GN57
Two samples were submitted from the north excavation units. Since
these units were located in the area of maximum depth of deposit, material
for dating from this area was desirable.
1815 BC Sample taken from (18N, 3E at approximately 13 cm
3765+90 BP below the present surface (1.13 m below datum).
UM-521 Sample consisted solely of oyster shells (Crassos-
trea virginica). This sample dates the upper levels
of this deposit and was removed just below the humic
zone. Sample is from Shell Sample #118, Field Speci-
2240 BC Sample taken from al8N, OE west baulk at a depth of
4190+90 BP 1.46-1.58 m below the surface (2.09-2.21 m below
UM-520 datum). Sample taken from the immediate vicinity
of human cranium. Sample of oyster shell exclusively.
Sample dates the lower midden levels and the human
cranium. Sample is from Field Specimen #663.
The West Ring: 9GN76
Two samples were submitted from the east baulk of Test 1 with the
intention of dating the upper levels and the basal levels of the deposit.
1655 BC Sample consisting solely of oyster shells (Crassos-
3605+110 BP trea virginica) taken from approximately 19 cm be-
UM-523 low the surface (1.86-1.97 m below datum). Sample
dates upper levels of midden. Removed just below
humic zone. From Field Specimen #83, Shell Sample
1910 BC Sample taken from the lowest level of midden deposit
3860+90 BP at approximately 46 cm below the surface (2.24 m
UM-5T2 below datum). Shell Sample #16 comprised only of
oyster shells (Crassostrea virginica). Field Speci-
Table 5 provides radiocarbon dates for non-ring fiber-tempered sites.
Marsh Cultural Level
Two samples were submitted from separate proveniences on the south
side of the marsh ring (9GN57). The first sample was collected during
the summer of 1973; the second in the spring of 1974. Dates from these
proveniences were felt to be requisite for determining the age of the
fiber-tempered/grit-tempered ceramic associations. Additionally, some
indication of the length of time represented by the marsh build-up would
820 BC Sample collected from o27S, 18E and consisted of wood
2770+95 BP taken from a tree stump associated with floral, faunal,
UM-5T9 and cultural materials. Ceramics were fiber-tempered
and grit-tempered. Depth of the sample was approxi-
mately 61 cm below the present marsh surface (3.21 m
below datum). Field Specimen #40. Dates demise of
tree--probably at time of marsh encroachment.
835 BC Sample taken from 033S, 12E at a depth of 76-87 cm
2785+80 BP below the surface (3.36-3.47 m below datum). Field
UM-5T8 Specimen #279 consisted of carbonized material from
a carbon concentration associated with faunal remains.
Sample dates the fiber-tempered/grit-tempered cultural
Table 5: Radiocarbon Dates for Fiber-tempered and Related Sites.
Site Dates Materlal/Slgniflcance Source
2515 BC:4465 + 95 BP (GXO-345)
2500 BC:4450 + 135 BP (GXO-343)
920 BC:2870 + 110 CGX01752)
1780 BC:3730 + 150 BP (M-1278)
2500 BC:4450 + 150 BP (M-1277)
2165 BC:4125 + 115 BP (0-1047)
1870 BC:3820 + 125 BP (M-11ll)
1780 BC:3730 + 125 BP (M-1112)
1750 BC:3700 + 125 BP (M-1109)
1820 BC:3770 + 200 BP (M-236)
1445 BC:3395 + 100 BP (GX-2281)
970 BC:2920 + 200 BP (M-267)
820 BC:2770 + 95 BP (UM-519)
835 BC:2785 80 BP (UM-518)
Charcoal: 18/B11 1.2-1.8 ft.
Charcoal: 15/Bl1 1.3-2.4 ft.
Charcoal: Pit 4, base of pottery
Charcoal: Pit 2, below pottery
Charcoal: 5.5-6.0 ft. (dates
Charcoal: 5.5-6.0 ft. (dates
Bone: 5.5-6.0 ft. (dates
Charcoal: 3.0-3.5 ft.
Oyster: Dates pottery
Oyster: Dates pottery,burlal
Shell: Dates refuge ceramics
Wood: Dates cultural level
Carbonized wood: Dates cultural
Bullen and Greene
Correction factors based on dendrochronological analysis is pre-
sented here to provide an indication of realistic dates. All dates are
given as date range from Ralph, Michael, and Han 1973: 1-20.
Radiocarbon Date Dendrochronological Correction Range
1815 BC+90 years 2180-2070 BC
2240 BC+90 years 2910-2580 BC
1655 BC+110 years 2110-(1750-1710) BC
1910 BC+90 years (2410-2340)-2140 BC
820 BC+95 years 1010-800 BC
835 BC+80 years 1010-810 BC
Cultural inventories reconstructed by archaeological excavations in
shell ring sites are not elaborate. Based on the raw material utilized
in construction, four categories are recognized: ceramic, bone/antler,
lithic, and shell. It is pertinent to remember that cultural items made
from plant resources constitute a potentially diverse category no longer
recoverable. In sites where conditions of preservation have been favor-
able (e.g. caves or bogs), the contribution of plant resources is impres-
sive. The absence of cultural items constructed from plant materials
generally creates a rather lopsided view of aboriginal technology.
Although occupational debris in the marsh cultural level was covered
by a considerable deposit of muddy silt, no examples of culturally modi-
fied wood or fiber were recovered from aboriginal contexts. This would
seem to indicate that inundation occurred so slowly that deterioration
of any such material was completed prior to submergence. Similarly, no
impressions of basketry, matting, or twining were observed on fiber-tem-
Ceramic typology mirrors the lack of consensus previously noted in
consideration of cultural chronology. Claflin's report (1931) of the
Stalling's Island excavations introduced the terms Stalling's Island
culture and Stalling's Island pottery. Unfortunately, Claflin did not
recognize the ceramics as being fiber-tempered or propose any relation-
ship to known vegetal-tempered ceramics from Florida (Wyman 1875). Holmes
(1894) had more succinctly described vegetal-tempering in aboriginal cera-
mics. Holder (1938) noted vegetal tempering in the ceramics in his
collections from St. Simon's Island, Georgia. As a result of his fin-
dings, a St. Simon's Fiber-Tempered type was recognized. This type,
however, was never formally described, and Holder's work has been pub-
lished only recently (Chance 1974).
That this ceramic type existed and was used is documented by Caldwell
and Waring's application (1939) of the nomen in their description of
Chatham County, Georgia, pottery types. The Bilbo site report, written
in 1940 but not published until 1968, uses the terminology. Griffin (1943)
defined the Stalling's Island types and applied the Stalling's nomen to
fiber-tempered ceramics from the Beaufort, South Carolina, area estab-
lishing a relatedness, almost an equivalence between the ceramics of the
Stalling's Island and coastal Chester Field sites. He reiterated this
point in a subsequent paper defining the Orange ceramic types of Florida
The attempts to check unwarranted creation of ceramic types resulted
in the inclusion of all fiber-tempered ceramics in one of three regional
categories. This scheme was formalized by Sears and Griffin in 1950.
Regardless of any consideration of attributes other than ceramic ones,
ceramics were identified on the basis of the geographical location of
the find. That is, if the ceramics were excavated in Florida, they be-
longed to the Orange Series; in Alabama and Tennessee to the Wheeler
Series; in Georgia to the Stalling's Island Series.
It was not until the Bilbo site excavations that Waring began to
have second thoughts (ca. 1940). He then devised a Bilbo period chronol-
ogy which he introduced in a paper written around 1952 (1955-60: 253-255).
Williams (1968) lists the report of the Sapelo Island shell ring excavation
as being written between 1955 and 1960. In this report (1955-60: 265-280),
Waring adhered to the St. Simon's type with no mention of Bilbo classifi-
cation. More recently, Crusoe and DePratter (n.d.) have redefined the Bilbo
chronology and urged its usage. Crusoe (n.d.) in a paper written subsequently
proposes a four-fold classification returning to the Stalling's nomen and
articulating with Bullen's (1954) five-fold Orange period classification.
Since 1940, the status of excavations and particularly analysis of
sites on the Georgia coast has been that of very little public communication.
This is especially true for fiber-tempered sites. Excavations have consisted
of small tests with interest continuing in a ceramics-to-chronology orienta-
tion. No adequate site survey of the Georgia coast has been done during
this time. The situation in 1975 is not appreciably improved.
Proposal of a new typology, adherence to an old one, or consolidation
of several is not intended here. Coastal sites are significant in more
meaningful ways than ceramic motifs or attributes of ceramic manufacture.
Milanich (1971) takes a refreshing direction in proposing a Coastal
Tradition based on subsistence strategy, total artifact inventories, and
settlement type within a frame of radiocarbon dates. This is one of the
first attempts to understand cultural developments on the coast in terms
of a lifestyle and not in terms of viewing the people as possessors of
The greatest problem to researchers is the absence of a systematic
body of data from excavated sites in either riverine or coastal environ-
ments. As many times as the Stalling's Island-rite has been excavated
or tested, interest in this site is primarily that of its unique ceramics
or ceramic change (Claflin 1931, Bullen and Greene 1970, Crusoe and
DePratter n.d.). The notable exception is Fairbanks' (1942) assessment
of the cultural traits represented by excavation findings. Too often
discussions of subsistence strategy in archaeological reports are obvi-
ously based on impression and not the result of analysis of food remains.
Riverine information is augmented somewhat by Miller's (1949) Lake Spring
site report and the paper by Phelps and Burgess (1964) on the White Mound.
The state of information reflects insufficient analysis and publi-
cation upon which to build a coherent structure of chronology or ceramic
typology. Until this data base is available, a moratorium on further
classification and categorization is urged to allow objective and impar-
tial collection of information. This should not be construed to propose
information gathering today and theorizing tomorrow. Rather it is recog-
nition that the framework within which the Late Archaic situation in
Georgia, as now perceived, is insufficient support. It is additionally a
recognition that the quality of the Cannon's Point collections precludes
their comparison with any extant collection. For these reasons, no new
scheme will be offered.
Before interpretive presentations of human development in the
Georgia coastal area are made, prehistorians must be willing to inves-
tigate the adaptation as a totality, not one or two traits.
A portion of this study is a contribution to such a data base.
Ceramics have been a major concern for archaeologists since their unique
ability to reflect cultural development was recognized. To this end,
a ceramic analysis is provided for each excavation unit giving level
distribution and totals. Many recovered decorated sherds are illustrated
with their respective proveniences. Illustrations of decorated ceramics
and rim forms may be found in Appendix 3 sherdss having Orange series
motifs) and Appendix 6. Additionally, distribution by level is provided
in Appendix 7. Illustrations of grit-tempered ceramics are included as
Ceramic fragments recovered from excavation units in the ring sites
were overwhelmingly fiber-tempered. Table 6 lists the distribution of
fiber-tempered and grit-tempered sherds for both ring sites. Table 7
lists ceramic distribution for each excavation unit in the marsh. Table 8
presents totals of ceramic material recovered from the three contexts.
It should be mentioned that fiber-tempered sherds tend to fragment upon
removal from the midden and as a result of excavation trauma. A sherd
is here defined as a ceramic fragment having both an interior and an
exterior surface intact. A paste fragment is defined as any ceramic frag-
ment lacking either one or both surfaces. Such a distinction reduces
sherd counts appreciably but is felt to be more representative of the
material reocvered. Weights are included to provide an indication of
paste fragment contribution to the total ceramic sample.
Table 6: Ceramic Distribution for Ring Sites (9GN57, 9GN76).
Provenience Fiber-tempered Sherds Grit-tempered Sherds
Plain Decorated Body Rim Total Weight Plain Decorated Body Rim Total Weight
o6S,15E 169 25 156 38 194 3696.2 g 1 1 1 2.0 g
a9S,15E 64 4 59 9 68 948.3 g
al8N,OE 231 8 225 14 239 3274.9 g 2 2 2 11.5 g
a18N,3E 79 4 67 16 83 1831.0 g 3 3 3 8.1 g
DON,15E 2 2 2 67.3 g 3 2 1 3 20.0 g
030N,15E 4 3 1 4 106,5 g
o6N,8E 28 28 28 184.1 g
al9.5N,2W 21 20 1 21 333.3 g
9 Units 598 41 560 79 639 10441,6 g 6 3 8 1 9 41,6 g
Test 1 45 2 44 3 47 808.9 g 17 5 22 22 69.9 g
35S,30E 33 12 34 11 45 687.0 g 2 1 3 3 3.5 g
2 Units 78 14 78 14 92 1495.9 g 19 6 25 25 73.4 g
Table 7: Marsh Ceramic Distribution.
Provenience Fiber-tempered Sherds Grit-tempered Sherds
Plain Decorated Body Rim Total Weight Plain Decorated Body Rim Total Weight
13 Units 289 27 277 39 316 5293.1 g 234 80 275 39 314 2143.4 g
Table 8: 9GN57 and 9GN76 Total Recovered Ceramic Material.
Total Total Paste Paste Paste Total Total
Provenience Sherds Weight Fragments(ft) Weight Fragments(gt) Weight Fragments Weight Ceramic Weight
16 24.9 g 71
Table 8 Continued.
Total Total Paste Paste Paste Total Total
Provenience Sherds Weight Fragments(ft) Weight Fragments(gt) Weight Fragments Weight Ceramic Weight
Ol8N,OE 241 3286.4 g 970 407.9 g 970 407.9 g 3694.3 g
1a8N,3E 86 1831.9 g 621 234.3 g 621 234.3 g 2066.2 g
tON,15E 5 87.3 g 8 22.7 g 8 22.7 g 110.0 g
a3N,15E 1 2.0 g 1 2.0 g 2.0 g
a30N,3E 4 106.5 g 78 41.6 g 1 2.2 g 79 43.8 g 149.8 g
06N,8E 28 184.1 g 36 33.8 g 36 33.8 g 217.9 g
al9.5N,2W 21 333.3 g 30 70.1 g 30 70.1 g 403,4 g
Test 1 69 878.8 g 98 102.2 g 9 8.6 g 107 t 110.8 g 981.0 g
05S,30E 48 690.5 g 363 194.4 g 363 194.4 g 884.9 g
24 Units 1395 19864.0 g 3576 2532.9 g 50 57.0 g 3626 2597,9 g 21820.6 g
Several observations regarding the ceramic collection from the
Cannon's Point Late Archaic sites are pertinent. The term fiber-tempered
is used to designate those sherds in which vegetal paste inclusions have
resulted in post-firing vermiculation and channeling. Grit-tempering
(grit in the form of visible quartz inclusions) is present in all ceramics
that are not fiber-tempered. While fiber-tempered sherds excavated in
Georgia should be typed Stallings, the situation for grit-tempered sherds
is either simple stamped or cross-simple stamped (Appendix 8). Radio-
carbon dates (820-835 BC) place these ceramics within the Refuge Phase
(ca. 1100-600 BC) as defined by Peterson (1970: 76-81).
Waring (1968: 198-208) first described this phase after excavating
the Refuge site, a shell midden located on a tributary of the Savannah
River in South Carolina. Distinction between Refuge and the later Dept-
ford phase was based on ceramic attributes. A coarser, thicker paste is
reported for Refuge sherds. This phase is best known from the Savannah
River drainage and its coastal affiliations are not clear. Peterson
(1970: 76-81) notes a distinct change in environmental usage in terms of
exploited resources and area of habitation. This information comes from
the excavations headed by Stoltman at Groton Plantation, Allendale County,
South Carolina. Peterson contends that Refuge sites are situated in the
uplands. When the swamps are used, occupation is not associated with
mollusc gathering since Refuge phase sites on Groton Plantation yield no
In the Cannon's Point project area, the marsh cultural level simi-
larly reflects no mollusc association. It is true that in this cultural
level, the majority of grit-tempered sherds were recovered in association
with fiber-tempered sherds, but it is in the latter sherds that an interesting
change is noted. In the fiber-tempered sherds from the ring sites, only
an occasional grit inclusion is observed. Fiber-tempered sherds from
the marsh cultural level, however, evidence a very high frequency of
grit inclusions. The tempering of the grit-tempered sherds from this
provenience were admittedly coarse, but assignment to a Refuge type may
be premature. As Stoltman observes (1974: 22) Waring's Refuge typology
was subjectively developed and may not be accurately reproduced by
another investigator. Perhaps the event occurring in the marsh cultural
level is one of development from predominantly fiber-tempered to grit-
tempered types--a pre-check stamped Deptford situation. Unlike fiber-
tempered ceramics, coil fracturing is frequently observed on these grit-
tempered sherds. Ceramic decorations continue to be distinct, however.
No incised or punctated motifs appear on grit-tempered sherds: no stamping
is present on fiber-tempered sherds. Decorative motifs on fiber-tempered
ceramics from St. Simon's Island shell ring sites are highly variable.
The typical Stallings motif, linear punctation, does not occur in this
sample and is generally not a common coastal decoration.
Waring (1955: 219) notes that Orange Incised, a Florida type, appears
to extend as far north as St. Simon's Island, Georgia. Excavations in
both ring sites substantiate this point (Appendix 3). Of particular
interest are sherds from the West Ring, particularly proveniences 59,
62, and 67. These sherds are nearly identical to Griffin's (1945) examples
of Tick Island Incised (see Griffin's Figures D-1). The sherds are from
basal levels of the east arc of the midden (the basal levels of the west
arc are dated at ca. 1910 BC). In the Marsh Ring, sherds having Orange
motifs are from the south arc excavations units and were recovered in the
first three levels.
As the ceramic tabulation shows, no support for a plain fiber-
tempered ceramic horizon is provided. Decorated ceramics are found be-
neath the midden as well as within it. It is interesting to note that
no decorated sherds were recovered on the north ring arc until almost
one meter of midden had been removed. On the south arc, decorated cera-
mics were very superficial.
Cultural items made of bone were far more common than those of antler.
A sizeable sample of worked bone was recovered, the majority small frag-
ments. Only three examples of utilized antler were identified. Much of
the antler recovered from marsh proveniences could not be verified as
utilized because of its poor state of preservation. All examples of
worked bone appear to be made of fragments of larger bones. Some may
be specifically identified as constructed from the metapodials of white-
tailed deer. A catalog of worked bone and antler is provided in Appendix 9;
illustrations of selected examples follow in Appendix 10.
Three basic categories of worked bone are exemplified: awls, points,
and pins. Awls are defined as pointed, tapered implements having an
expanded base. Frequently a constriction or grooving is observed at the
base. Seven examples were recovered. Only two points were identified,
one of bone and the other of antler. Points are characterized by short
length and are cylindrical but tapered in cross-section having a con-
stricted base. Bone pins have been one of the most characteristic arti-
facts from fiber-tempered sites. Pins are constructed of split fragments
of deer metapodial. Specimens are limited in length to the length of the
metapodial. In the Cannon's Point collections, length from base (or head)
to point is 12 to 16 cm. Fragments were rounded, apparently by abrasion,
and a high polish is usual. It is uncertain if the polish is a result
of abrasion in manufacture or from usage. Nine examples were recovered.
In addition, many of the worked bone fragments are parts of pin shafts
and points. The application of decorative motifs was limited to this
category. The most common design (#342) observed was a zoned triangular
or diamond incised motif. Two examples of the scroll designs of the
Bilbo type bone pins described by Waring (1940: 170-171) are included
in this collection. One is rectilinear (#776 and the other gives a hint
of a curvilinear design (#817a).
Worked bone was recovered as a midden inclusion with more frequency
than in features. In Features #19 and #21 (a highly cemented hearth
extending through the shared baulk), one awl, a broken pin tip, and one
decorated shaft fragment were retrieved. In Feature #20, another probable
hearth which lacked the consolidation of shells observed in Feature #19
and #21, most of a pin was recovered. In contrast, Feature #18 which
consistently yielded large quantities of fauna, produced only a fragment
of utilized turtle.
The elaborately engraved bone pin fragment (#776) was exposed below
the midden and its impression was visible in the submidden soil. Proven-
ience and Waring's criteria suggest a later date for these sites. The
radiocarbon dates, instead of supporting elaborately decorated bone pins
as a later development, suggest that in the Cannon's Point area this
artifact has a much earlier date.
One additional type of worked bone was frequently observed. In the
catalog, it comprises over 50% of the recovered examples. The carapace
of specific turtles evidences internal scraping which removed the verte-
bral attachments on the neurals and exposed the cancellous structure of
the pleurals and marginals. The carapace is cut from the plastron through
the bridge marginals apparently to allow removal of edible meat. This
cut area usually shows considerable polish. Such treatment of turtle
carapace is consistently seen on Deirochelys reticularia, Chrysemys cf.
floridana, and Malaclemys terrapin. These species have frenulated exteriors,
a fact that apparently was culturally valued. That they are the largest
turtles locally available is another observation which must be included.
Because scraping increases the porosity of the shell, it would seem that
items intended to be contained within (if in fact the shell was used as
a container) were dry instead of liquid.
Culturally altered lithic material recovered from the marsh ring
consisted of three examples. The location of each item is particularly
interesting. A granitic hone was recovered in Feature #19 associated
with a bone awl. In the submidden soil of the southern half of al8N, 3E,
a chert projectile point was recovered. This point, illustrated in Appen-
dix 11, is not a typical Savannah River Stemmed point, rather it is
reminiscent of a Florida Arredondo point typologically (Bullen 1975).
The third specimen is a stemmed, bifacially worked object excavated from
a pit adjacent to Feature #12. The specimen is of chert and appears to
have been repeatedly retouched. Its function may have been as a knife
During this time period, coastal sites have consistently produced
few lithic artifacts. The small inventory discussed above was expected.
Debris from tool-making or retouching was similarly meager, suggesting
that lithic items were brought to the coastal area as finished products.
No cores or reforms were recovered. A small number of chert flakes
were retrieved from the West Ring but no lithic artifacts. Quartz
pebbles were more frequently recovered, particularly in the basal levels.
The marsh cultural level presented a very different picture. Six
lithic artifacts were excavated in units south of the marsh ring. Flint
was black or bluish in color in contrast to the white specimens from
the marsh ring. Debitage was present in greater frequency than the few
chips recovered in the ring sites. It is probable that screening would
have produced additional specimens.
From the scarcity of lithic material in the ring sites and the
greater frequencies in the marsh cultural level, it seems apparent that
lithic availability in the coastal area increased with time. The change
in color is interesting also. Black flint outcrops have been reported
from the Fall Line but not from the coastal plain.
Definition of this class of artifacts if particularly problematical.
Certainly any shell is a potential tool but conditions of preservation
may obscure evidence of use. Clarification of criteria used for classi-
fication of shell tools was required as field work progressed. A mollusc
shell was considered utilized if non-naturally occurring alterations were
observed. These alterations could be evidence of use as food rather than
tool use. Additionally, it was anticipated that some specimens would
evidence both food and tool use modification.
Gastropods were the most commonly excavated mollusc class exhibiting
consistent alterations. Such molluscs can be removed from their shells
by pulling the animal out by its foot. Perforation of the shell provides
access to the muscle attachments, which when cut makes removal possible.
Holes in the shoulder of the shell or absent spires were considered alter-
ations for food use. The criteria for tool use were based on alterations
of the umbilical area (base) and lip edge (at the shell opening). Blunting
of the base, angular wear resulting in basal pointing, and lip edge wear
were the most consistently observed alterations.
Two gastropods were commonly recovered from the midden with such modi-
fications: Busycon carica and Busycon carica-eliceans. The former seemed
to repeatedly exhibit modification for food use, the latter for tool and
food use. Busycon carica is a thin-shelled species; Busycon carica eliceans
is a more robust species. Some of the archaeological specimens had segments
of shell removed exposing the columella. Such detachment produces sections
of shell which can be used for manufacture of other itesm (e.g. beads).
Bivalves were the most numerous midden inclusion, but evidence of food
or tool alteration was not as obvious. A single example of utilized oyster,
modified by the removal of a circular section from the valve body was re-
covered. Use is not known. In the West Ring, a clam shell (Mercenaria sp.)
with the umbo (hinge) removed and exhibiting considerable edge wear of this
area was retrieved. The specimen may have served as a dipper or scoop.
Single shells were generally encountered during excavation but occa-
sionally clumps of shells were observed, but these plus crustaceans (usually bar-
nacles) were not sizeable nor was their location repetitive. Many Busycons
were excavated with internally adhering oyster growth indicating that the
initial reason for collecting this specimen was not for food. Since these
specimens usually exhibited no modification from tool use, the choice of
these shells remains a question. Such gathering practices seem to suggest
a lack of discrimination in the selection of potential foodstuffs. Per-
haps the location of the habitation area was near enough to the source to
compensate for the waste of energy represented by transportation of ined-
ible mollusc shells. Another suggestion of the proximity of mollusc
sources to the habitation area is offered by one specimen which was
initially a tool then discarded in water andsubsequently regathered with
adhering immature oyster growth (Appendix 12, #568). This specimen demon-
strates immature oyster growth over edge wearand removal of the lip.
On the other hand, the functions of shell tools may have included acti-
vities done on or near the water with loss of this specimen being a pre-
dictable consequence of use.
Clams frequently exhibited exterior shell removal near the umbo.
Perforation did not occur but an inner layer of shell was exposed. A blow
to loosen the hinge and facilitate extraction of the mollusc is a probable
cause. Small concave areas of shell are frequently observed to be absent
from the lip of the clam suggesting that prying may have been attempted.
Most clams evidences neither lip nor umbo alteration. Appendix 12 catalogs
the utilized shell from both ring sites and includes photographs of selected
Molluscan resources represent such an obvious contribution to the
subsistence base of Late Archaic peoples that the perception of coastal-
dwelling groups as "oyster eaters" inevitably developed. Unfortunately,
little else was considered. Identification and enumeration of faunal
remains from the Cannon's Point sites indicates that it is more realistic
to perceive coastal inhabitants during this time period as participants
in a procurement system in which the tidal creek and the surrounding marsh
were critical sources of food. However, a discussion of molluscs is appro-
priate to initiate any consideration of Late Archaic subsistence.
While oysters are obviously the dominant species present, many other
molluscs occur as midden inclusions in significant quantities. All too
frequently, this fact has been overlooked. As noted previously, some
species combined subsistence and technological functions. A list of the
molluscs excavated from the ring sites is presented in Table 9 (identification
is based on Abbott 1974 and Burch 1962). It is apparent that mollusc habitat
varies greatly. Oysters, clams, whelks, stout tagelus (Tagelus plebius),
and mussels (Geukensia demissa) made the greatest food contribution. Marsh
periwinkles seem to have been more important in the West Ring while the subsis-
tence role of marsh snails (Ilyanassa obsoleta, Melampus bidentatus),
oyster drill (Urosalpinx cinerea), and others is less certain. These species
may have been gathered with other resources representing incidental inclusion
in the midden. Carnivorous terrestrial snails seem to have been drawn in
very large numbers to the decaying midden debris. There is no evidence that
these may have been utilized.
Table 10 presents a composite faunal list for each ring site and the marsh
cultural level. Taxonomy is based on the following sources: fishes Bailey
et al. 1970; amphibians and reptiles Conant 1975; birds Peterson 1947;
Robbins et al. 1966; mammals Burt and Grossenheider 1964, Golley 1962. The
diversity of faunal species identified is remarkable. Species lists for each
excavation unit and a level-by-level assessment of the faunal material is
available in Appendix 5. From these tabulations, it is evident that verte-
brate fauna is critically important to aboriginal subsistence. This point
is clearly supported when a dietary assessment of available data for molluscs
Table 9: Composite Molluscan Species List, 9GN57 and 9GN76
Species Common Name Habitat
Stenotrema fraterna (Pillsbry)
Triodopsis hopetownensis (Shuttleworth)
Polygyra cereolus (Muller)
Euglandina rosea (Ferrusac)
Haplotrema concava (Say)
Mesomphix vulgatus Baker
Melampus bidentatus (Say)
Detracea floridana (Gmelin)
Nerltina reclivata (Say)
Polynlces duplicate (Linn.)
Littorina Irrorata (Linn.)
Busycon carica (Gmelin)
Busycon carica eliceans (Montfort)
Busycon canallculatum (Linn.)
Ilyanassa obsoleta (Say)
eastern mud nassa
floodplains, forested habitats
warmer subtropical and temperate climate,
lowland areas, damp habitats
lowland floodplains, damp vegetation
floodplains, lowlands, forests
shallow-water sand bottoms
brackish water marshes on weeds
shore to 30 feet
shore to 30 feet
in sand; 3-50 feet
in muddy intertidal flats, salt marshes
Table 9 Continued.
Species Common Name Habitat
Urosalpinx clnerea (Say)
Eupleura caudata (Say)
Anadara ovalis (Bruguiere)
Tagelus plebius (Lightfoot)
Mercenaria mercenaria (Linn.)
Mercenaria campechlensis (Gmelin)
Dinocardium robustum (Lightfoot)
Crassostrea virginica (Gmelln)
Geukensia demissa (Dillwyn)
Crytopleura costata (Linn.)
Atlantic oyster drill
northern quahob clam
southern quahob clam
giant Atlantic cockle
Atlantic ribbed mussel
oyster bars; low tide mark to 25 feet
muddy and sandy bottoms
In sandy mud, intertidal
in sand 3-50 feet
estuaries, tidal creeks
colonies In mud and clay
Table 10: Faunal List.
Fauna 9GN57 9GN76 Level
Decapoda x x
Callinectes sp. x x
Galeocerdo cuvieri x
Sphyrna sp. x x
Rajiformes x x
Dasyatis sp. x
Aetobates narinarl x
Lepisosteus sp. x x x
Amia calva x x
Elops saurus x x
Clupeidael x x
BrevoortiQ sp. x
Ictalurus sp. x
Ariidael x x x
Arius felis x x
Bagre marinus x x x
Micropterus sp. x
Opsanus sp. x x
Pomatomus saltatrix x
Carangidae' x x
Lagodon sp. x
Archosargus sp. x x x
Sciaenidae' x x x
Bairdiella chrysura x x
Cynoscion sp. x x
Leiostomus xanthurus x x
Menticirrhus americanus x
Table 10: Faunal List. (Continued)
Fauna 9GN57 9GN76 Level
Micropogon undulatus x x x
Pogonias cromis x x x
Sciaenops ocellata x x x
Stellifer lanceolatus x x
Mugil sp. x x x
Prionotus sp. x
Paralichthys sp. x x x
Siren sp. x
Chelydra serpentina x
Kinosternon sp. x x x
Terrapene carolina x
Malaclemys terrapin x x x
Chrysemys sp. x x x
Chrysemys cf. floridana x
Deirochelys reticularia x x
Alligator mississipiensis x
Anolis carolinensis x
Natrix sp. x
Coluber constrictor x
Lampropeltis sp. x
Phalacocorax auritus auritus x
Ardea herodias wardi x
Bucephala clangula x
Buteo lineatus alleni x
Table 10: Faunal List (Continued)
Fauna 9GN57 9GN76 Level
Didelphis virginiana x x x
Soricidae x x
Cryptotis parva x
Blarina brevicauda x x
Scalopus aquaticus x
Sylvilagus sp. x x x
Rodentla x x x
Glaucomys volans x
Sciurus carolinensis x
Microtinae cf. Neofiber x
Neofiber alleni x
cf. Neotoma x
cf. Peromyscus x x
Oryzomys sp. x
Peromyscus sp. x
Sigmodon hispidus x x
Lutra canadensis x
Mustela vision x
Canis familiaris x x x
Procyon lotor x x x
Odocoileus virginianus x x x
Homo sapiens x x
Indicates elements only identified to this level.
and several vertebrate species is made. Table 11 documents the fact that
crustaceans and vertebrate fauna comprise a more significant contribution
by weight than molluscan fauna.
With a few notable exceptions, all fishes identified are marine and
inhabit the tidal creek biotope. The exceptions are freshwater catfish
(Ictalurus sp.), bowfin or mudfish (Amia calva), and bass (Micropterus sp.).
All usually inhabit freshwater biotopes but may enter brackish waters. They
are represented in the faunal collection by a very small number of fragments.
Their presence in these sites may indicate a change in the local habitat.
It is possible that these species were resident in semi-permanent ponds on
the island. Such standing water as is present on the island today depends
upon rainfall and is generally scarce. Freshwater catfish are taken today
in marine habitats when unusually high freshwater run-offs occur. These
run-offs may be caused by increased rainfall over a period of time or by
storms. A similar situation occurring prehistorically may account for the
presence of these species. Additionally, freshwater turtles were recovered
from most excavation units.
The tidal creek biotope constantly fluctuates with tidal condition.
At periods of high tide, the tidal creek expands vastly. Conversely, at
periods of low tide, the marshes extend for miles. Many mammalian species
other than man depend upon the marshes and tidal creeks for their maintenance.
Crabs (particularly of the genus Uca fiddlers) constitute a staple item in
the diet of island raccoons. Deer and rabbits (Sylvilagus sp.) include marsh
grasses in their diets. Otter and mink, both semi-aquatic mammals, prey on
numerous species present in these biotopes. Many avian species are perma-
nent residents of the area, others are seasonal. While the majority of avian
Table 11: Dietary Contributions of Selected Species (Watt and Merrll 1963)
(100 grams of edible portion)
4A E E E U
E m E 0 C <
SI 0 < E 1 I
01 2) 5 2 Z O I E O r 1
01 -. In 2 0 2 a ro u2
Iv E L 2 E E > E u
E1 2 -C E -a E I E C -0
,0 4J n0 0 I) c In C I o o U 0
S O 4 L C L 0 0 4j U
M to I I W -C L 0 0 C In
ap L) LU O LLa < U IL 0. 1 > I- Z <
Deer venison raw
84.6 66 8.4
83.8 66 9.6 1,4 3.1
86.2 49 6.5 .4 4,2
67,9 172 18.7 10.2
79.2 96 17.8 2.2
71.3 133 24.3 3.2
80.2 80 18.0 0.4
65.3 219 17.6 15.9
78,5 93 17.3 1.9 .5
73.0 135 21.0 5.0
74.0 126 21.0 4.0
1,8 3.4 1.8 94 143 5,5 121 73 310 ,14 .18 2.5
1.8 43 175 0.8
1.0 10 249
60 ,12 .08 5.5
70 .13 .10 6.5
.15 .05 3.5
2,170 .16 .08 2.8 2
.23 .48 .63
Table 11 Continued.
0 E E
M E c -0
u E E E U
E m 01 A E6 01 C E <
n .- 1A E 3 0 < E 0-
S-o E 3 E > E U
L ) c 1 >. 01 E O 0 *- C C a *-
w) G- C E E l E *- C -0
4- 0 O I- cn 0 3 E E C L.
M O a-) .0 0 A C ( *- ( o 0 U 0
S 0 4-' L .f 0 0 4-1 -0 4- .0 U
S 1- (0 *- (A ( -C 1- 0 0 C -
da o L.. U C Lw < UL a a- > i- z <
Opossum roasted 57.3 221 30.2 10.2 2.3 .12 .38
species identified from the collections are permanent residents, one migra-
tory species is present (American golden-eye).
Two species identified are not present-day residents of the area.
Aetobates narinari (spotted eagle ray) is not reported for the Georgia
coast (Dahlberg 1975: 31). Similarly, the present distribution of Neofiber
alleni (Florida water rat) does not include the Georgia coast (Burt and
Grossenheider 1964: 200-202).
Several species are particularly important in the faunal sample: blue
crab (Callinectes sp.), menhaden (Clupeidae cf. Brevoortia), marine catfish
(Ariidae), members of the drum family (Sciaenidae), and mullet (Mugi1 sp.).
The majority of individuals in the sample are represented by these species.
Mammalian species contribute a greater amount of edible meat by weight than
fish or crabs but are not present in large numbers in the collections. It
is obvious from the areas tested that mollusc gathering was augmented largely
Amphibian fauna is represented by only one genus, Siren. In general,
amphibian remains were so tiny that adequate collections were not available
for positive identification. Consequently most of the recovered amphibian
fauna is listed as unidentified. Reptilian fauna, particularly snake verte-
brae, presented a comparable problem. The condition of erosion and frag-
mentation of the elements makes the identification of snakes very tentative.
Turtles were well represented. It is interesting to note that turtles in
the marsh cultural level included two species not present in the ring sites:
Chelydra serpentina and Terrapene carolina. Chelydra serpentina was present
in considerable quantity. One fragment of alligator was recovered from the
marsh cultural level. This was the only example of this reptile retrieved.
Many species may represent incidental inclusions rather than food
remains. Some of these are flying squirrel, mole, shrews, green anole
(chameleon), and several species of rats. Regardless of their function,
the presence of these small animals add considerable information about
the prehistoric environment.
The role of the dog is uncertain. Dog remains occurred in the midden
as any other food remains might have been observed. One partial maxilla
was recovered in Feature #19. Its cracked and burned condition suggests
that whatever the premortem function of the dog, its postmortem role in-
cluded human consumption. It is interesting to note that no teeth were
recovered with the maxilla or in the feature. While dog teeth are frequently
recovered drilled for stringing, no evidence of this practice was observed.
Dog remains were recovered at each of the locations.
The marsh cultural level evidences an increased incidence of use of
turtle and deer. As previously mentioned, no molluscan remains were asso-
ciated with any of the faunal or cultural material exposed beneath the pre-
sent marsh surface. This absence suggests several hypothetical answers:
a decreased abundance of molluscs, a culturally motivated subsistence
change, differential use of the area (i.e. not for mollusc harvesting),
or differential preservation of bone and shell beneath the marsh. At this
time, adequate support for any of these hypotheses is lacking.
Regardless of the reports of previous investigators, floral material
is a midden inclusion and may be retrieved using screened samples. Floral
remains were recovered as seeds, nutshells, and carbonized nuts (particu-
larly acorn). A tentative floral list is provided in Table 12 with addi-
tional plotting of the seasonal availability of the edible parts of the
Table 12: Seasonal Availability of Identified Flora.
Species Month Comments Common Name
J F M A M J J A S 0 N D
__ mature acorns
berries may per-
S fruit available
Pinus is not present in the sample in any quantity. Juniperus spp.
is presently the dominant cover of the marsh ring and present on the
West Ring. In the marsh cultural level, juniper roots were exposed.
Identification was based on the characteristically red inner coloration
and odor. Quercus spp. was moderately abundant. Carya spp. was the most
numerically abundant nutshell. Considerable quantities were recovered
throughout the midden with a significant increase in the submidden soil.
It was not possible to determine whether these'fragments were burned or
carbonized. An increased frequency of Carya was noted in Feature #18,
#24, and #25. Celtis spp. is second to Carya in abundance. Its frequency
is particularly high in Features #19, #20, and #21. There seems to be
some correlation between the occurrence of Celtis and areas identified as
probable hearths. Celtis is also abundant in the general levels throughout
the midden. Prunus serotina, Ilex vomitoria, Bumelia lyciodes, and Fores-
teiera sp. are not present in any quantity. It should be remembered that
the leaves of Ilex vomitoria are known to have been used aboriginally for
ceremonial drinks (Swanton 1946: 699).
In general, the floral species identified still grow in the area.
Brassica spp. has been tentatively identified but is not common. This
species represents an introduction from Eurasia or Africa (Small 1933: 562).
Level-by-level species identification is provided in Appendix 13 for the West Ring.
Soil samples were taken whenever a feature was designated or an area
of unusual color or texture appeared. Additionally, a stratified sample
was taken from marsh unit 045S, 15E for comparative analysis. The question
of amount of carbonate resorption taking place was felt to be critical to
problems of differential preservation of shell and bone. Because marsh
periwinkle and Atlantic ribbed mussels were noted in marsh excavations
in a state of decay, comparative analyses were required to check the
carbonate levels. The samples were submitted to the Institute of Food
and Agricultural Sciences Soil Sciences Laboratory at the University of
Florida. Unfortunately, the tabulations were misplaced, and the only
results which are available concern pH values. These values are listed
below and indicate an alkaline environment, a condition already suspected
in view of the excellent bone preservation.
Soil Depth below
Sample # Surface pH value Associations
59 0-18 cm 7.4 Spartina growth heavy
60 31-43 cm 7.6 Spartina rootlets present
but growth not thick
61 70-80 cm 7.3 Corresponds to level of
cultural and subsistence
62 120 cm 7.5 Below the cultural level
Samples for pollen analysis were taken, particularly from the marsh
excavation units. Initial processing of these samples in the Florida State
Museum Paleoecology Lab revealed extensive Spartina contamination. Conse-
quently, the samples were of little value. One would question the value
of pollen samples in an area at the mouth of a river in any event. Much
of the pollen deposition may include water-borne pollens from non-local
Although fpur of six excavation units in the marsh ring produced some
human skeletal material, the total inventory is small. Table 13 summarizes
this material. Skeletal material recovered was fragmentary but generally
gave the impression of being crushed in place. Because cannibalism had been
hypothesized to explain the fragmentary nature-of previous finds of human
skeletal material in shell ring sites, much care was taken to observe support-
ing information. However, no solid evidence of cannibalism is offered by
this collection. Admittedly, some of the bones have articular surfaces
missing but this could be caused by the manner in which the skeletal element
became a part of the midden.
The inclusion of non-articulated skeletal material in the midden pre-
sents a puzzle. It Is possible that the lower levels of the ring were
borrowed from shell heaps in which burials had been made. It is also pos-
sible that considerable shifting of the midden overburden through the years
caused relocation of material and disarticulation. It may be that the burial
practices of these people included exposure and that subsequent burial of
the remaining bones was haphazard. No cultural items have been excavated in
association with human skeletal remains.
The cranium and pelvis are probably from the same individual, a male
aged 22-28 years, although their locations in the midden are some three meters
diagonally distant. They are, however, from approximately the same level.
The remaining material is enigmatic, particularly the right leg and foot
from the south arc excavation units.
Consideration of Methods and Techniques
At this point, it seems appropriate to consider the methods and techniques
applied during excavation since analysis points out the inconsistencies, pro-
Table 13: Human Skeletal Material
Provenience Level Element Condition Sex/Age
East Baulk pelvis
ilium and pubic symphysis present.
root absent; extensive attrition
u24S,24E 75-85 cm bs parietals
fragmented; distal articular surface in-
cluded in sample but not attached. Proxi-
mal articular surface absent from below
lesser trochanter. Sizeable portion of
proximal epiphysis present In sample but
not attached; sizeable portion of diaphysis
intact. Distal articular surface absent.
medially split; articular surfaces for
split with roots missing; extensive attri-
frontal, parietals, temporals, small amount
of occiptal present. Basal portions of
occlptal absent and left side of maxilla,
Right side of maxilla has one remaining
molar (M ) which evidences extensive attri-
tion. M Is erupted. Abcess noted in pre-
molar. Supraorbital torus present but not
or contributions of these aspects of excavation strategy. The methodology
and techniques applied are responsible for generating the raw data which
when analyzed, supports the observations made.
When an excavator makes a commitment to screening, several problems
immediately arise. Time required for field processing is greatly increased.
Material recovered similarly increases (depending on the site). The area
that can be excavated with the time allowed maybe lessened unless the
crew is large. Time required for post-excavation analysis is lengthened.
In a shell site, the amount of material to be processed when screening is
used is phenomenal. Most of the excavator's time is not spent in the exca-
vation unit but at the screen clearing shell debris and picking out the
archaeologically significant material. Davidson (1964) notes many of the
same problems and states:
In almost all cases, the critical division seemed to lie
between the 1/4 inch and 1/8 inch. Material retained by
an 1/8 inch sieve was identifiable without difficulty.
Objection to it lies in the great increase in time necessary
for a very small increase in exactness. The material re-
tained by the sieve was always less than 10% of the total
weight of the sample and usually considerably less.
(Davidson 1964: 155).
A very interesting comment on the amount of archaeologically significant
material recovered from a shell ring excavation level is offered in Table 14.
Calculations for this table were made possible by the fact that 06S, 15E was
excavated during the last few days of the Winter Field Session. Time did
not allow field processing of the material in the lower screen. Consequently,
significant material was separated in the top screen and bagged (note first
column); all material caught in the lower screen was bagged and processed at
Table 14: Contribution of Screening by Weight of Processed Material,n6S,15El
Total weight lower screen
% of lower screen weight
Total weight lower screen
% of lower screen weight
25.5 g 42.0%
22.1 g 78.9%
1.2 g 25.5%
615.9 g 96.92%
63.5 g 36.7%
130.2 g 96.3%
3.05 g 52.6%
975.6 g 99.92
35.5 g 58.0%
5.9 g 21.1%
3.5 g 74.5%
19.5 g 3.1%
5.05 g 3.7%
2.75 g 47.4%
0.8 g 0.1%
Weights given represent the amount of material considered archaeologically
significant. Shell debris represents discarded material. Weights given as
"shell" here refer to terrestrial and marine molluscs recovered.
Includes weight of potsherds excavated and mapped in place in the excavation