MISSISSIPPIAN PERIOD COMMUNITY ORGANIZATIONS
ON THE GEORGIA COAST
MORGAN RAY CROOK, JR.
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
Many individuals and several institutions were involved in the
making of this dissertation and the absence of any one of these
would have changed its form. The Kenan Field research presented in
the following pages is but a small part of a broader research objective
designed to record and assess the archaeological resources of Sapelo
Island,Georgia. Towards the fulfillment of this broad research
goal, I am pleased to acknowledge the support of Governor George
Busby and the State of Georgia, along with Commissioner Joe Tanner
and Dr. Elizabeth Lyons of the Georgia Department of Natural Resources.
I also gratefully acknowledge the National Science Foundation for
the Dissertation Improvement Grant (#77-07565) that funded the
zooarchaeological analysis for the Kenan Field Project.
Archaeological research on Sapelo Island has benefited greatly
by its association with C. V. Waters of the Georgia Game and Fish
Division and his wife Mary Lou. The assistance and courtesies
received from the Waters and their staff are accepted with gratitude
and are due repeated thanks.
Several individuals may be singled out in recognition of their
positive impact upon me through their roles as teachers, advisers,
and friends. I sincerely thank Lewis H. Larson, Jr., for his patience
and his continued interest in my growth as an archaeologist, for his
wise counsel, and for the many research opportunities which he
encouraged. Whatever achievement is expressed by this dissertation
is largely the effect of my association with Dr. Larson. I also owe
a large debt to the distinguished members of my Doctoral Committee.
I thank Dr. Fairbanks for his many excellent graduate courses and for
his open willingness to consider new points of view and to discuss
old problems. I consider myself privia d to have been taught by i'
him. I also thank Dr. Rice for her teaching ability, for he6n k
for providing criticism at the right moment, and for her alertness
Logical and stylistic flaws. My courses with Dr. Margolis provided
a firm base for my theoretical orientation and although she would
argue with particulars of my framework, it is nevertheless a result
of her influences. I appreciate Dr. Proctor for his interest in the
dissertation and for his propensity to focus upon those simple questions
which have such complex anthropological explanations. Finally in the
repetoire of prominent individuals, there is the position of Dr. Jerald
T. Milanich, the chairman of my Doctoral Committee. His impact upon
me and the course of this dissertation has been considerable. I deeply
thank him for some of the interest he was able to impart to me for
coastal archaeology, for his skepticism, and for his numerous
There are a host of other individuals whose efforts are appreciated
and who positively affected the research. The field work during the
summer of 1977 was completed with a crew composed of from ten to
fifteen high-school students and other residents of Sapelo Island.
Their efforts and attentiveness are acknowledged with graditude, as is
the able service of three field assistants. Ms. Patricia O'Grady
supervised the shell-midden sampling program and the in-field analysis
of the molluscs, Mr. Jeff Mitchem supervised much of the general
excavation, and Mr. Ken Johnson supervised the mechanical-transect
excavations. Their personal involvement in the success of the
project is to be credited.
The laboratory-analysis portion of the research fell into the
competent hands of five individuals. Ms. Mimi Saffer completed the
descriptive analysis of pottery, Ms. Pamela Stern analyzed the clam
shells, Ms. Erika Simons and Ms. Kathie Johnson completed the faunal
analysis under the supervision of Dr. Elizabeth Wing, and Mr. David
Hall was responsible for the floral analysis. The goals I set for
analysis often required the modification of standard analysis
procedures and the flexibility of the analysts towards the achievement
of these goals was admirable.
The influence of fellow graduate students in the formation of my
concepts in anthropology and archaeology, and towards the solution
of my personal research problems, cannot be over-estimated. I benefited
from numerous discussions with Ann Cordell, Nick Honerkamp, Bob Johnson,
Tim Kohler, Jill Loucks, Alan McMichael, Mimi Saffer, Robin Smith, and
Secretarial and administrative services are an important portion
of any research and I fortunately have been associated with some of
the best. I thank Ms. Mavis Riding and Ms. Linda Timmons at West
Georgia College along with Ms. Annette Fanus of the Florida State
Museum for their kind assistance. I also thank Ms. Nancy Steinen for
her patience in typing the dissertation.
Last, but not least, I would like to acknowledge my personal
appreciations. I thank a few close friends for remaining friends
while I was consumed with the dissertation. These include Ann Cordell,
Tom Eubanks, Nick Honerkamp, Steve Johnson, Leslie Lieberman, Jill
Loucks, Alan McMichael, Mimi Saffer, Dan Simpkins, Robin Smith, and
Pam Stern. A special place is reserved here for Patricia O'Grady, for
her love and support. Finally, I thank my mother and father for their
long support of me throughout my education and for their solid
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . . . . . . .. . .. ii
ABSTRACT . . . . . . .. .... . . .. vii
CHAPTER I: INTRODUCTION ..... . . . . . . 1
Concepts and Theory .... . . . .. . . . 1
Problems and Hypotheses .... . . . . . . 4
CHAPTER II: METHODS . . . . . ... . . 10
Mapping Procedures ...... . . . . .. . 10
Sampling Procedures . . . . ... . . .. 12
Faunal and Floral Analysis .. . . . . . 19
Pottery Analysis .... . . . . . . .... 20
Lithics and Other Artifacts .. . . ... . . .. 23
CHAPTER III: PERSPECTIVES ON MISSISSIPPIAN
PERIOD COASTAL ARCHAEOLOGY . . . ... .. . 24
Environmental Summary .... . . . . . . 24
Coastal Georgia during the Sixteenth Century ...... 35
Archaeological Assessments . . . ...... .. . 71
CHAPTER IV: INVESTIGATIONS AT KENAN FIELD . . ... . 97
Structure #1 .. ... ........ ........ ... .103
Structure 2 . . . . . . . . . . . 150
Structure #3 . . . .... ...... .... 173
Structure 5 . . . ..... .... . . . .. 176
Structure #4 ... . ..... . . ..... 137
Structure #6 . . . . . . . . . . . 203
Structure #7 . ... . . . . . . . 209
Mechanical Transect Test Excavations . . . .. .. .217
Shell Midden Tests .. . . . . ....... .. 225
Analysis of Clam and Oyster Shell . . . . ... ... .23A
CHAPTER V: CONCLUSIONS . . . .. . . . . . 255
Tests of the Hypotheses . ..... . . . . 255
Consideration of the Sixteenth-Century
Model and Cultural Change ... . . . ...... .265
APPENDIX: POTTERY TYPES RECOGNIZED AT KENAN FIELD,
INCLUDING POSSIBLE ALTAMAHA VARIANTS ...... .272
BIBLIOGRAPHY ........... . . . . . . ... 279
BIOGRAPHICAL SKETCH .. .. ........ . . . . .290
Abstract of a Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
MISSISSIPPIAN PERIOD COMMUNITY ORGANIZATIONS
ON THE GEORGIA COAST
Morgan Ray Crook, Jr.
Chairman: Jerald T. Milanich
Major Department: Anthropology
This dissertation examines Mississippian Period cultural adapta-
tions in a coastal environment as indicated by certain social, spatial,
and temporal dimensions of community organization. Hypotheses con-
cerning basic elements of adaptation are posited and an annual model
of the aboriginal cultural system is constructed through an examination
of sixteenth-century accounts and data from modern ecological sources.
Preliminary investigations at the Kenan Field site on Sapelo Island,
Georgia, are discussed in detail and the archaeological data are
analyzed for evidence required to test the stated hypotheses and systems
model. Analysis focuses upon subsistence information, architectural
details, and the spatial distribution of material-culture elements.
Two primary conclusions are drawn from the research. The first
is that the Savannah Phase, a Mississippian Period manifestation on the
Georgia coast, was defined by a chiefdom-level cultural system that was
distinctively adapted to the coastal environment. The second basic
conclusion is that the Savannah Phase was followed by a period of
cultural change that appears to be associated with late sixteenth
century Spanish activities on the Georgia coast.
This is an archaeological study of cultural adaptations on the
Georgia coast during a latter portion of the Mississippian Period
(ca. A.D. 1000 1540). Primary data are from preliminary investi-
gations at the Kenan Field site on Sapelo Island, Georgia, and consist
of settlement and subsistence details along with other information
concerning material culture. Ethnohistoric accounts of the coastal
inhabitants during the sixteenth century and published archaeological
data provide additional evidence for analysis and interpretation.
Concepts and Theory
The study of settlement and subsistence from materials preserved
in the archaeological record has proven to be a productive research
orientation for the archaeologist. This mode of inquiry has its
roots firmly set in the regional analysis and cultural ecology of
Julian Steward (1938, 1955), with its archaeological application in
the Viru Valley of Peru (Willey 1953).
A general definition of settlement pattern is "the way in which
man disposed himself over the landscape on which he lived" (Willey
1953:101). As Willey later points out,
settlements are a more direct reflection of social and economic
activities than are most other aspects of material culture
available to the archaeologist. . [but] like most other facts,
those of settlement are robbed of much of their importance when
considered in isolation [Willey 1956:1].
Borrowing the ideas of social "microstructure" and "macrostructure"
(Chang 1968:7), Bruce Trigger refined settlement pattern categories
with the concepts of "microsettlement" and "macrosettlement" (Trigger
1968:54-55). Macrosettlement pattern analysis is the distributional
study of habitation areas over the environment, while microsettlement
analysis examines habitation components in relation to one another
within a community or focuses on details of*individual structures.
Macroanalysis is concerned with intersite relationships and nicro-
analysis with intrasite details.
Quantitative techniques, locational models, and ecological
theory have provided a major impetus for development in settlement
subsistence studies since the 1960's. It is very clear that the
application of advanced analytic techniques, or even the study of
settlement and subsistence pattern, is of limited value as a goal in
itself. Rather, techniques and theories offer a key to the under-
standing and explanation of cultural systems. Central to this under-
standing is the concept of adaptation. This concept is a fundamental
assumption in all ecological studies, including archaeological ones
dealing with extinct cultural systems through analysis of settlement
and subsistence data. Adaptation is, in very simple terms, the process
of successful adjustment to and articulation with the environment.
In the case of a cultural system, this success is demonstrated by
the continued existence, or survival, of the system.
Structural elements of the materialist-hierarchic model of
culture, explicitly used by Leslie White (e.g. 1969) and more implicitly
by Julian Steward (e.g. 1955:134-185), consist of ideology, social
and political organization, and technology. Technology provides a
base for the cultural pyramid and is the primary means of adaptation,
implemented through productive arrangements ordered by social and
political organization, and supported by ideological constructs.
The model is systemic and dynamic. It is dynamic in that
cultural change is affected by alterations in the structural com-
ponents, basically through thermodynamic increase as proposed by
Thite (1969:363-393) and by modification in technology and productive
arrangements as suggested by Steward (e.g. 1955:37). It is dynamic
also in the sense that the elements are functionally related and
interact with one another in the form of a cultural system. That
is, culture is an interrelated whole.
Structured within a materialist framework, cultural ecology seeks
to "explain the origin of particular cultural features and patterns
which characterize different areas" (Steward 1955:36). However,
the concept is "less concerned with the origin and diffusion of
technologies than with the fact they may be used differently and
entail different social arrangement in each environment"(Steward
Cultural ecology focuses on cultural adjustments to local
environment by discerning those cultural features "which empirical
analysis shows to be most closely involved in utilization of environ-
ment in culturally prescribed ways"(Steward 1955:37). Those features
of culture basic to environmental utilization, the "cultural core,"
are those "most closely related to subsistence activities and economic
arrangements." The core includes such social, political, and religious
patterns as are empirically determined to be closely connected with
these arrangements" (Steward 1955:37).
Perhaps the most important deficiency within the cultural
ecology method is that the role of dynamic interplay between culture
and environment is minimized. The method is constructed to discern
cultural features caused by adjustment to environmental factors. A
feedback concept is missing; the treatment of culture as an interacting
element within an ecosystem. Attention to feedback loops and
cybernetic theory has proven important in archaeological research (see
Plog 1975). However, the original formulation of cultural ecology
remains a valuable basic research tool, and is used here with an
awareness of the importance of system interactions within the environment.
Cultural ecology, with the added feedback dimension, is
employed in this study as a data organizing tool and as a heuristic
device. Essentially, the concept provides a general problem orientation
for research and analysis. Ethnohistoric and archaeological data are
analyzed for details of technology, subsistence activities, productive
arrangements, social and political organization, and ideology, with
reference to aspects of the local environment and material exchanges.
Problems and Hypotheses
The explication of prehistoric cultural adaptations along the
southeastern Atlantic coast has received the increasing attention of
archaeologists in recent years. This research is discussed in later
chapters. It is sufficient to note here that major gaps exist in
our knowledge. One such gap involves Mississippian Period adaptations
in the region. Although early ethnohistoric accounts indicate a
marked degree of social complexity, there is little understanding of
the range and function of the social components in terms of adjustment
to environmental factors and dealings with other cultural systems.
Initial investigations during the summer of 1976 at the Kenan
Field site on Sapelo Island suggested that more intensive, problem
oriented research would yield substantial information on the adaptations
of a late prehistoric coastal community. Kenan Field is located on an
extension of land that juts out briefly about midway along the western
side of Sapelo Island. The site is rather rectangular in shape,
corresponding to the form of the land upon which it is situated.
Salt marsh bounds Kenan on all but its eastern side. The Duplin
River approaches the northwest corner of the field, and a portion of
the south end is accessible through a small tidal creek (Figure 1).
The site area, defined by the shell midden distribution, measures
60 hectares and contains 589 discrete shell middens.
Research efforts during 1976 concentrated on mapping and limited
test excavation. Surface debris in the planted-pine stand that covers
Kenan Field had been destroyed by fire, exposing the shell deposits,
which were mapped. Pottery from the test excavations was predominately
cord-marked and check-stamped Savannah-Phase wares, and much less
frequently, Irene complicated-stamped pottery. Post holes marking
construction activity were exposed beneath the 25 cm. plow zone.
Since excavation was restricted to a small area, the only conclusions
supported at this point were that at least a portion of the site
SAPELO ISLAND, GEORGIA
dated to the Mississippian Period and that detectable structural
features were present.
The shell middens and other topographic features, including
two earthen mounds and a long earthen embankment, were mapped to
provide a model of the location and surface distribution of cultural
features at the site. The map (Figure 2) shows rather ambiguous
patterns of refuse deposition, probably because of sequential re-
arrangement of settlement components.
However, certain general patterns are evident. The map shows
intricate linear and aggregate arrangements of midden deposits
which suggest a complex village plan. At least four distinct areas
may be defined:
(1) a separate clustering of shell middens in the northeastern
portion of the site;
(2) linear arrangements in the central area;
(3) an arc-shaped pattern of large shell middens defining an area
to the west of Mound 'A';
(4) an area more or less void of midden refuse in the southern
portion of the site, south of the linear embankment and near
These spatial data are impossible to interpret because cultural
associations are undemonstrated. Knowledge of the cultural pro-
venience of the shell middens and the relationship between midden
location and structural components are specifically required. Once
these data are established, structural location should be predictable
from refuse location. The hypothesis is that refuse was deposited
by definable social units within the village, and that this de-
position was in reference to structures that housed these units.
ii o ,
S ". "- ,,-- ' "o "
r,.... 1 : .. |
,. -.f I
..'= =-c =..t
-.IF y']- ,*-
Formally, the purpose of this dissertation is to test by advanced
field methods the following hypotheses and their implications:
Hypothesis 'A': The Mississippian Period community at the Kenan
Field site was characterized by a formal village plan which
reflects the adaptations of a ranked society to environmental
factors and relations with other cultural systems.
1. The spatial arrangement of structural remains is patterned.
2. Socially and/or functionally distinct structures are associated
with particular areas of the village, and may show repeated
construction in the same area.
3. Socially and/or functionally distinct structures are
a. Structural details;
b. Associated subsistence refuse;
c. Associated material culture.
4. Adaptations to the natural environment reflected by the
village plan are discernible in:
a. Form and distribution of structural remains;
b. Communal food-storage facilities.
5. Adaptations to the social environment reflected by the
village plan are discernible in:
a. Defensive fortifications;
b. Exotic materials.
Hypothesis 'B': Social organization of the Mississippian Period
community at Kenan Field was an adaptation for procurement of
1. Social organization is discernible in:
a. The village plan;
b. Structural form and function;
c. Distribution of contemporary and discrete material culture
d. Redistributive activities.
2. Productive arrangements are discernible in:
a. Subsistence technology;
b. Form of social units engaged in subsistence activities;
c. Settlement location in relation to strategic resources;
d. Seasonal and spatial occurrence of strategic subsistence
The map of Kenan Field provides a surface-distribution model that
is the basis for sampling and ultimately the basis for conclusions
drawn from these data. The procedures used in mapping determine the
accuracy of this base model and the amount of confidence that can be
placed in it. Thus, the mapping technique is discussed in some detail.
Kenan Field was mapped using controlled transit-stadia transects.
A base line of stations was set at 50 m. intervals along the western
side of the site, between two permanent bench marks. Corridors, 50 m.
to 75 m. wide, were defined extending east and west from these stations,
from the western marsh edge to High Point Road.
The transit was positioned within a corridor in reference to the
associated station. The rod-man then walked from one side of the
corridor to the other in a north to south and south to north fashion,
using the 2.5 m. space between the rows of planted pine trees as a guide.
Within each corridor, every other tree row was walked and the interven-
ing space was inspected. The location of shell middens was recorded
with the transit, and shell-midden size and form were estimated.
Dimensions of small middens were visually estimated and larger middens
were paced. Shell midden center-point elevations, immediate surrounding
surface elevations, and other surface elevations at approximately 15 m.
intervals were recorded from a metric stadia rod. The mounds and other
earthworks were surveyed at smaller surface intervals, generally from
1 m. to 3 n. apart.
Survey thus proceeded down a corridor, the northern extent of
reconnaissance being marked with flagging and the transit being
moved at about 150 m. intervals. When the eastern end of a corridor
was reached, the next to the north was surveyed back towards the western
base line. The formerly flagged line now marked the southern extent
of the new corridor and the northern side was flagged for future
reference. Instrument position was re-established at the western
end of the corridor in reference to the base-line station, and
the accumulated transit error averaged over the number of set-up
points used in the pair of corridors. Transit readings were then
plotted on metric graph paper at a scale of 1:2000.
Crucial to the survey was the definition of shell middens. Shell
middens were defined by noticeable concentrations of shell debris
usually accompanied by a slight rise above the surrounding surface.
Kenan Field was used for agriculture beginning perhaps earlier than
1800, and continued to be cultivated at intervals until 1951 when
the pines were planted. This agricultural activity dispersed the
shell deposits and makes definition of the smaller middens difficult.
However the results of the map are accurate and replicable, as
individual shell middens were re-located during the summer of 1977
using a transit and following the declinations and distances presented
on the map. The only inaccuracies noted were the erroneous definition
of two small shell middens and a single large shell midden. These
were originally defined more by slight increases in surface elevation
than bv noticeable shell concentrations.
The shell middens encountered at Kenan Field were divided into
five classes based on size differences (Figure 3). The three larger
diameter classes (A,B,C) were defined by natural breaks in frequency
distribution and the smaller two (D,E) by a marked decrease in
frequency. Measurement of those few middens which are elongated
rather than annular were converted to averaged diameters for class-
The two classes of shell middens with the largest diameters were
used to define sampling strata and substrata. The two earthen
mounds were included within the largest diameter class for strata
Strata were formed by constructing Theissen polygons (see Kopec
1963, Haggett 1965:247-248) with the shell middens as referents.
Each polygon defines that area which is closer to the designating
shell midden than to any other midden of the same or larger class.
Polygons were initially constructed with equal weight being placed
on both Class A and Class B middens. This defined the sample sub-
strata. Eleven strata were then defined by drawing polygons in
reference to only Class A middens. The boundaries of these strata
were then adjusted to accommodate the substrata within each polygon
One of the polygon strata, Stratum F, was chosen for sampling.
Stratum F confines an area of 51,033 square meters, contains 91
shell middens, and is divided into six substrata. Selection of this
stratum for testing was non-random. It was chosen because of its
Class A Class 3 3lass C Class I 2las I
SIZE-FREQUENCY DISTRIBUTION OF SHELL MIDDENS AT KENAN FIELD
/ 1 z z
/ -- r,
/ 7 ^ i-
/ ~ i: d4\a
/ \ ) \
/ / / ^
-------- ^ ----- \ /
\ / z
central position within the site, and because it encloses variable
areas of shell midden arrangements, from the arch-shaped pattern west
of mound A to the linear arrangement area in the central part of the
Initially, a random sample of each substratum within Stratum
F was planned. Because of several unpredicted factors that will be
discussed later, only three were investigated.
Each of these substrata was tested with a randomly selected
transect. Sample transects were aligned between 110 and 150 east
of north, determined by the arrangement of the planted-pine rows at
the site. It is assumed that placement is unbiased in reference to
the prehistoric component. The plow zone along the 1.9 m. transect
lines was removed by a tractor with a rear-mounted dozer blade, and
the transect surface was leveled by hand with shovels. All observed
features were then mapped with an alidade, and artifacts were recovered
and their provenience recorded.
The sole purpose of the sample transects was to expose structural
locations. More extensive and formal excavation was required for
definition of structural form and cultural associations. Far more
structures were located by the transects that could be exposed by
extensive excavation during the 14-week field period during the
summer of 1977. Selection of structures to be more completely
defined was non-random. It was decided that each structure within a
transect required definition and that second level random sampling
was inappropriate. Excavation was expanded in those transect areas
which contained the most structural activity. The sample is probably
biased at this point towards the archaeologically more complex
Structural areas were excavated in 2 m. x 2 m. and in 1 m.
x 4 m. units. Each unit was excavated in levels to the base of the
plow zone. The first excavation level extended 15 cm. beneath the
surface and defined a zone of maximum agricultural disturbance. This
zone in approximately every other excavation unit was screened through
1/4 in. or 1/4 in. x 3/4 in. mesh. Depth of the second excavation
level was more variable but never extended more than 30 cm. beneath the
surface. This level defined a less disturbed portion of the plow
zone and cultural material from each excavation unit was recovered
by screening, usually with 1/4 in. mesh hand screens. Excavation
was terminated in most areas at the base of the plow zone and features
which were intrusive into the underlying subsoil were mapped with
an alidade or in reference to established grid stations.
Two structural areas were investigated outside of the sample
design. Excavations were extended in the 1976 test area in an
attempt to define that structure, and excavations were initiated
in an area just west of mound A after a trial transect was cut
that exposed structural elements.
In addition to structural area excavations, random samples were
taken of each shell midden class within a substratum. The purpose
of this was to secure a quantifiable sample of subsistence remains
and determine the cultural association of the shell middens. A 25%
sample of each shell midden class within each substratum was randomly
selected for testing. The actual sample fraction varies due to the
number of shell middens per class per substratum. For example, by
definition each substratum contains only one Class A or Class B shell
midden. Thus, 100% of this class per substratum was tested.
Each selected shell midden was tested by placing a 1.5 m. x 2 m.
test pit within the deposit. This test pit was divided into four
unequal, horizontal sections (see Figure 5). Section A was excavated
in 15 cm. levels to the base of the midden and the matrix was screened
through 1/4 in. mesh. Excavation then continued until sterile subsoil
was reached. Cultural material from each level was retained and the
shell material was identified and weighed in the field.
Profiles of the eastern wall of each test pit were recorded,
and Sections B, C, and D were excavated by natural stratigraphic levels.
A total sample of approximately 1 liter from each level was recovered
in Section C. These samples are reserved for future analysis. Sections
B and D were independently screened through 1/4 in. mesh and processed
the same as Section A. In addition, midden that passed through the
screen was collected and processed by screening with 1/16 in. mesh in
water to remove the soil matrix. Recovered material was then spread
on plastic sheets to dry. This material was manually sorted in the
field and the faunal remains were separated from the shell debris and
artifacts. The debris was retained and samples were later processed
by chemical flotation (ZnC 2) to recover plant and animal remains that
were overlooked in the field.
The field methods used at Kenan Field were designed to provide
data for testing the stated hypotheses. Execution of these methods
STANDARD FORMAT OF SHELL-MIDDEN TEST PITS
| Section A __
I I * ,
STANDARD FOREIAT OF SHELL-MIDDEN TEST PITS
required more field time than was originally predicted. The main
reason for this was the unanticipated number and large size of
structures that were encountered.
A significant amount of data was gathered, but execution of the
sampling design is incomplete. For this reason the final statistical
treatment and their conclusions must await additional field work.
The study presented here is preliminary in this respect; however,
enough information is available at this point for partial resolution
of the hypotheses.
Faunal and Floral Analyses
Fauna from the Kenan Field excavations were identified using the
comparative collections of the Zooarchaeology Laboratory, Florida State
Museum. Skeletal elements recovered from each provenience unit were
identified to the lowest reliable taxon and elements of each taxon were
weighed. Minimum number of individuals were determined for species
identified from shell-midden test pits and undisturbed features. Faunal
material recovered from general excavations in the structural areas was
fragmented due to agricultural disturbance and required a different
format for the presentation of analysis results. This faunal material
was summarized by the relative skeletonal-weight contribution of each
identified species within specific contexts.
Floral remains encountered at Kenan Field were identified to the
lowest reliable taxon using the comparative collections and references
of the Herbarium at the University of Florida.
Pottery recovered from the Kenan Field excavations was analyzed
to provide evidence of both diachronic and synchronic variability.
Analysis focused on identification of surface treatment and design
elements, tempering components, and rim forms. Tempering, as used
here, refers to aplastic inclusions with the pottery paste.
Classification of tempering material conforms to categories
used in previous type descriptions of Southeastern pottery (see March
1934, Southeastern Archaeological Conference 1938). It should be noted
that although these aplastic categories are widely used and are
apparently reliable, their simple definitions have remained implicit.
Sand tempering is self-explanatory and refers to fine grains of sand,
sherd or grog tempering denotes crushed fragments of potsherds included
within the paste, and grit tempering refers to coarse granules of sand.
Sherd tempering is a nominal category, while the sand and grit categories
are ordinal. Analysis in the present study considers the presence and
relative proportions of tempering materials within each potsherd, and
thus is basically an ordinal quantification. A sample of the analysis
forms used for the pottery analysis is shown in Figure 6 and a flow
chart of the analysis procedure is presented in Figure 7.
The screening techniques employed in the field recovered many
very small sherds. These sherds were usually impossible to identify
with any confidence and they were considered to contain little infor-
mation beyond that presented in the larger sherds. Therefore, a
sampling system was devised to remove small sherds from detailed
analysis. Separation by size was accomplished by placing pottery
" Fi-l) "_ 7 F1 rter-" '- "-" l .7.777c -
.- C I I I I I I |
::-,~ ... C .~ s . t I I I I j I I I i
II 7 s I ~ 1 1C7
c | i i I I i I
Iin-! ---. I I I ______
iIii ~i i
POTTERY ANALYSIS FORM
_______________________ _______________________ (i) ______________
- -\> [l
FLOW CHART OF PROCEDURES USED IN POTTERY ANALYSIS
from each provenience unit in a Clorox bottle neck with an inside
diameter of 21 mm. Those sherds passing through the bottle neck were
counted and weighed, and small rim sherds were removed for additional
analysis. Those sherds larger than 21 mm. remained within the analysis
These larger sherds from each provenience were divided into groups
with common tempering and surface treatment. The number of sherds
within each group was recorded and notes were made of other traits
such as shell-smoothed interiors and hone marks. All rims and a large
sample collection were set aside for additional analysis and typology.
Typologies were based on published type descriptions and comparative
collections in the Florida State Museum, University of Florida Depart-
ment of Anthropology, and the West Georgia College Archaeological
Lithics and Other Artifacts
Stone material was rare at Kenan Field. Those lithic artifacts
which were encountered were identified by material and form, and when
possible by function. Another rare category was pottery artifacts,
i.e. re-utilized potsherds, consisting of a few sherd hones and one
possible spindle whorl. Neither bone nor shell artifacts were re-
Spanish ceramics were infrequent and were not temporally distinc-
tive. These consisted of three body fragments of olive jar that could
have been manufactured as early as 1500 until later than 1800
(see Goggin 1960).
PERSPECTIVES ON MISSISSIPPIAN PERIOD COASTAL ARCHAEOLOGY
Aspects of the Georgia coast have captured the attention of
scientists from several disciplines for many years. A great amount
of published work reporting coastal research is scattered through
the journals of each speciality. A useful summary of some of this
research recently has been published (Johnson et al. 1974) and a study
of the coastal environment from an anthropological point of view has
been presented by Lewis Larson (1970). These two major works are
used rather freely in the following summary description.
The coastal environment has been altered since the late pre-
historic period by agricultural and industrial activities. Thousands
of acres of delta-swamp forest around the mouths of freshwater rivers
were cut during the late eighteenth and nineteenth centuries. These
areas were dyked and became productive rice fields. Highland forest
areas were also cut and the land planted during the plantation era,
and low wet lands were drained to provide even more agricultural
land. Clearing and draining continued well into the twentieth century.
Even today some low mainland areas are being drained to provide pulp-
wood acreage (e.g. Gray 1933, Bonner 1964).
These changes in the landscape must have displaced and in many
cases destroyed large segments of the biotic community. Certainly,
repercussions were felt throughout the coastal environment. The late
perhistoric environment can, nevertheless, be approximated with modern
ecological data. This information must be used critically and
assessed with a knowledge of historic alterations. For example, vege-
tation and animal communities documented within a drained slough of
a barrier island are of limited value as a base for prehistoric
Climate along the Georgia coast is rather moderate with warm to
hot summers and cool winters. Annual rainfall averages around 50
inches and temperature averages just below 70 F. There are about
300 freeze-free days a year in the Brunswick area of the middle
Georgia coast. Temperatures exceeding 900 F. begin to occur in May
and continue into September while freezing temperatures usually begin
during the last part of November and end during March. Precipitation
is from frontal activities during the late fall and winter, and thunder-
showers during the spring and summer. These thundershowers are fre-
quently localized and may inundate small areas while leaving nearby
locales completely dry. Normally about half of the annual rainfall
occurs between June and September, but tropical storms often account
for heavy rains in August and September. Table 1 shows monthly
rainfall and temperature readings for Sapelo Island during 1960,
1965, and 1970, along with average statistics from Brunswick.
The Georgia coast defines the eastern edge of the Southern
Temperate Deciduous Forest Biome (Shelford 1963:56-88). This biome
contains three major communities or faciations: the oak-hickory
forest which extends from around the Fall Line into the Piedmont and
upper Coastal Plain; pine lands or barrens which extend from the
northern edge of the Coastal Plain to within several miles of the
Ic rc 1o
c c 0 00 cc C3 C
'.r '. 'zr
0 0 -
DaG C4 co
so s r N cc \O
CO r? ~T
AM In 0 c o cOm
1'0 \0 '.0 33 -Z '0 -'
o oo o U
'OMN 1No N.
N r '
Ut C rU
Z- cT cr' N
oN 1 o Ut o O
ccJLC '0 N a, '0 i N'01 0N 0 N
X N i >- 2
O-*iin N '. IN O- N N Ut Ut N
r^ D *-
cc 0 N 0 C
ION '.0 lr 1C ION -r N -
3Mff t C ON N 0 t O
'0 A. Ut.c c m r^c r ri=o-rc ^
.-~ \ o r" -
'0' '. N '. O '
=^ r- f- in
'0. I0 IN. '0 .
Ut Ut.- C'0U '.0Ut ON
cc oC co cc
'0 r\n CT '0 ^ 1'0 '0 r4
s - N
<^ LO Ln *< f
SC o Ic N N
0 i CN N0
^ CC C CC rUO C -,
'0 '. N
N UtV7 |C N N N Ut L-
I-Cjv^~ '^ 'Ut
Ut 'C t Utr,- Ut =
S- 01 N S
C 0 4.
'r *i GT ~T. <^' 'a' u" -. c "!!- in c
^1 c i ri ~'p _5 : E: 5 -^ ~ ^ -1 *r
coast: and the magnolia and maritime forest which defines an irregular
strip containing the coastal islands and adjacent mainland.
Progressively younger geological deposits form the Coastal Plain
from the Piedmont to che Atlantic Ocean. The Fall Line represents a
late Mesozoic shore line. Relatively thin Cretaceous sediments occur
at the edge of the Fall Line and as sea level dropped, younger and
thicker surface deposits were formed approaching the present coast
line. The youngest, Holocene, formation occurs along the ocean sides
of the barrier islands.
Most of the coastal area is the result of Pleistocene formations.
As one crosses the coastal plain towards the Atlantic, relic coastal
features such as beach ridges, islands, hammocks and former marshes
may be observed. The most obvious are within several miles of the
Atlantic, just east of the Miocene Coastal Plain formation. Although
deposits thicken towards the present coast, surface elevation gently
drops from around 100 meters above sea level in the upper Coastal Plain.
Rivers that empty into the Atlantic along the Georgia coast have
their headwaters in three physiographic provinces. The Savannah
River originates in the Blue Ridge, the Altamaha in the Piedmont through
the Ocmulgee and Oconee Rivers, and the other rivers are Coastal Plain
in origin (see Figure8 ). These river systems provided the most
practical link between the coast and inland areas during the aboriginal
period. It has been argued (Larson 1970:73-117) that the pine barrens
had a low exploitation potential given late prehistoric subsistence
technology. The pine barrens was also an effective cultural barrier,
although permeable through the Altamaha and Savannah Rivers.
LCn.%i LrAfJnCZTrIZ A
0 5C 100
MAJOR PHYSIOGRAPHIC PROVINCES OF GEORGIA
SHOWING LOCATIONS OF THE TIDEWATER BIOME AND THE PINE BARRENS
GrCEirGI .iRiiNcA G iZ i I _.M eAsE Ma P
The Georgia coast is defined by plant and animal communities that
are distinctive from the adjoining pine barrens. Topography of the
coast is also distinctive. The western edge of the coastal section
is delineated by the marsh system that accompanied the Wicomico barrier
island formations. Eastward of this Pleistocene geological formation
are remanent features of later beach ridges, barrier islands, hammocks,
and marshes that extend parallel to the present coast line. Today
these mainland Pleistocene features form a system of highland hammocks
surrounded by freshwater swamp. The most recent Pleistocene formation,
Silver Bluff, defines the active system of barrier islands and salt
marsh that separates the mainland from the open ocean. Tidewater
actions are most pervasive in the most recent formations; however,
the tides also cause fluctuating water levels in the freshwater
rivers beyond the impact of salinity. Because of the influence of
tides throughout the coastal section, hereafter this area is referred
to as the Tidewater Biome. In summary, the Tidewater Biome includes
the present barrier islands and extends to the western edge of the
Wicomico formation (see Hoyt and Hails 1967, Hoyt 1968).
The Tidewater Biome may be divided into three environmental areas
based on biotic and abiotic differences. These areas Larson (1970)
defined as the Strand Section, the Marsh and Lagoon Section, and the
While use of these environmental divisions is maintained in this
review, exception is taken to the previous areal definitions of the
Marsh and Laeoon Section. The western extent of this section has
been defined by the impact of salinity and the pine barrens are
shown bordering this salinity zone (see Larson 1970:8). The Tidewater
Biome extends a considerable distance to the west of this former
boundary, to an irregular line reaching approximately 75 kilometers
inland. Pine barrens begin to occur at this inland point. The present
areal definition of the Tidewater Biome and the pine barrens agrees
well with the observations of William Bartram in 1773 (Harper 1958:19).
The Strand Section faces the open Atlantic and is composed of
the beach with its.dunes, shore and offshore areas. Most of the plant
and animal resources of aboriginal importance in the strand occur
more regularly and abundantly elsewhere on the coast. The exceptions
are sea turtle, which is a seasonal visitor that nests on the beaches,
and coquina, a small bivalve that occurs in some abundance along the
shore. The strand may have been visited to harvest sea turtles, their
eggs, or coquina, but there is little evidence of aboriginal occupation
in the area. Only two sites have been reported on the southern Atlantic
coastal strand. One is located on Cumberland Island (Milanich, personal
communication) and the other is in Brunswick County, North Carolina
(South 1976:10). The strand appears to have been of limited, if
seasonal, economic importance during aboriginal times.
The Marsh and Lagoon Section is separated from the ocean by the
strand, and is composed of high ground, marshes, tidal streams and
lagoons. This section is the largest, the most physiographically
diverse, and contains the greatest number of species in the Tidewater
Biome. Archaeological sites occur most frequently on the high-
ground areas within this section.
The barrier islands, hammocks, and mainland high-ground areas
are examples of what Shelford (1963:67-73) identified as Magnolia
Forest, although Live Oak (Quercus Virginiana Miller) was probably
the climax, or near-climax, dominant (e.g. Johnson ec al. 1974:
44-55, Wharton 1977:185-188). The climax questions aside, this forest
community includes live oak, laurel oak, water oak, pignut hickory,
red cedar, southern magnolia, red bay, American holly, and cabbage
palm as major overstory species. Understory species include wax
myrtle, saw palmetto, and yaupon along with many herbs and vines.
Floral species of the marsh are less diverse and are frequently
monospecific. Vegetation occurs in zones dependent on salinity and
inundation factors. The most extensive of these salt marshes are
composed of smooth cordgrass, needlerush, and giant cutgrass. Those
areas along the landward edge of the marsh which are flooded for only
a short time each day contain grasswort, saltgrass, sea oxeye, and
Although salt-marsh plants had no direct value as an aboriginal
subsistence resource, the marsh was essential in the food chain of
species which were of economic importance. The three primary producers
in the salt marsh are smooth cordgrass, mud algae, and phytoplankton.
Tidal flow brings essential nutrients into the marshes and carries
enriched nutrients and detritus back into the estuary. The energy
stored in the producers flows through the ecosystem through a grazer
food chain and a detritus food chain. Only about 5% of marsh pro-
duction is consumed in the grazer food chain. The remainder is
available to detritus and suspended-algae consumers. Primary detritus
and algae consumers include species such as fiddler crabs and molluscs.
Most, if not all, of the estuarine fish species either feed on marsh
detritus and suspensions, eat species that are detritus consumers,
That portion of the salt marsh which is next to high ground
provides an important feeding habitat for raccoons seeking high-marsh
crabs and the eggs of diamondback terrapins. The high-marsh plant
cover evidently also supplies an important food source for marsh
rabbits. Both the low aquatic marsh and the high marsh are feeding
grounds for marsh mink, which search for fish, mussels, crabs, and
eggs. Fish in the low marsh and estuaries are taken by otters. Both
otter and marsh mink use the high ground adjacent to the marsh as a
nesting area. Various wading birds such as white ibis and little blue
heron also feed in the salt marshes.
Mammals inhabiting the high-ground areas in the oak forest include
white-tailed deer, opossum, raccoon, cottontail rabbit, and bobcat.
The white-tailed deer population presently on the coast is genetically
diverse, due to the introduction of deer from various areas of the east
in game management activities. Aboriginal coastal deer may well have
been a smaller subspecies (Odocoileus virginianus nigribaris Goldman
and Kellog) which has been identified on Blackbeard and Sapelo
Islands (Golley 1962:204). The average weight of the Blackbeard deer
is around 60 pounds (Lund et al. 1962 in Johnson et al. 1974:59). This
weight is about 40 pounds less than the average weight of other
Georgia deer. A metric study of archaeological specimens could
establish whether or not this subspecies was prevalent on the coast
during prehistoric times.
A fourth and rather distinct environmental zone occurs within the
Marsh and Lagoon Section. This zone is defined by low areas of
freshwater swamp that are beyond tidal influence. The water supply
of this swamp is predominately rain water. Freshwater swamp occurs
in relic marshes along the mainland in areas which are away from the
river deltas. A similar phenomena exists within the low interiors or
sloughs of the barrier islands. Little is known about the ecology
of these freshwater swamps. Vegetation appears similar to that of the
upper reaches of the freshwater river, with cypress, tupelo, and ash
as possible dominates. The barrier island slough systems are slightly
brackish, especially where they empty into the salt marsh. Wading
birds form nesting colonies in the sloughs during the spring and,
lacking predatory fish, sloughs also provide important breeding
grounds for reptiles and amphibians. Food for the reptiles is
provided by the nesting birds.
One aspect of the mainland freshwater swamps that certainly
affected aboriginal occupation was that they effectively dispersed
highland areas and their resources. Highland oak forest occurs in
patches throughout the swamp and movement between these forest areas
was to some degree impeded by the swamp lands. Since the swamps appear
to have been of limited exploitive value, the forest resources
which were of aboriginal importance occurred in dispersed sections.
Larson (1970:292-297) has pointed out that the dispersed condition of
highlands combined with limited pockets of relatively fertile soil
restricted the size and distribution of Indian agricultural plots.
Lacking natural soil renewal systems, like alluvial deposition, even
the more fertile soils are rapidly exhausted in the high-ground areas.
Based on ethnohistoric evidence, the late aboriginal inhabitants of
the coast responded to these fertility restrictions with a shifting
agricultural regime. Maize, beans, and squash were cultivated in
these swidden plots by small social units scattered over the coastal
area. Aboriginal agricultural on the coast is discussed further in the
ethnohistoric section of this chapter.
The final environmental section in the Tidewater Biome is the
delta area. The Delta Section is defined by the course of freshwater
rivers. These rivers become increasingly deltaicand brackish as they
approach saltwater. The Delta Section is low and frequently inundated,
especially during spring floods. Vegetation in the delta is water
tolerant, grading towards increasing salt tolerance as the deltas
approach the salt marsh. Although subject to the tides, deltas are
composed primarily of fresh water. Much of the area consists of fresh
to slightly brackish swamps containing cypress and gum as dominates,
with increased vegetation in the higher areas (Wharton 1977:60-62).
As with the freshwater swamps, very little is known about the ecology
of delta swamps. Brackish species, such as fiddler crab, inhabit some
delta-swamp areas. Deer, otter, and raccoons also occur in portions of
the swamp; however, it appears that each of these mammals is far more
abundant and accessible in the Marsh and Lagoon Section.
Anadromous fish are the most distinctive inhabitants of the delta
rivers, and were probably the resources of aboriginal importance.
Species such as American shad, glut herring, striped bass, and sturgeon
enter the rivers during the spring to spawn. While the abundance of
anadromous fish would have made aboriginal exploitation profitable for
short periods of time, there is little in the Delta Section to sustain
a more permanent population.
In summary, the tidal dreams and oak forest are the richest biotic
environments on the coast. As Larson (1970:33) concludes,
The Lagoon and Marsh Section with its diversity of ecology, with
its variety and abundance of resources, was potentially and actually
an area of considerable aboriginal importance. The Coastal Sector
populations resided almost exclusively within this section, and
concentrated almost entirely upon its resources.
Coastal Georgia during the Sixteenth Century
That portion of La Florida now known as the Georgia Coast was
occupied by two aboriginal linguistic groups during the sixteenth
and seventeenth centuries. The Timucua occupied the area from
around Cumberland Island and the Satilla River in Georgia south into
northeast Florida. A group known to the French as Ouade, the Spanish
as Guale, and the English as Wallie extended north from the Timucuan
area to around St. Catherines Island. Guale was actually the name
of a single town and its chief on St. Catherines Island. The name
was also used by the Spanish to refer to the entire area from
the Timucua to St. Catherines.
There is little ethnohistoric information about the area be-
tween St. Catherines Island and the Savannah River. This portion
of the coast perhaps encompassed the northern limits of Guale and
the southern limits of Cusabo. While the northern and southern
extent of the groups called Cusabo are unclear, the area from the
Savannah River north along the South Carolina coast to around
Charleston inlet seems to have been almost exclusively the land of
Cusabo (Swanton 1922:16-17, 1946:128).
The distinction between Guale and Cusabo was at least partially
due to Spanish divisions of the coast into administrative areas.
There may have been some cultural differences as well, but at least
the Cusabo just north of the Savannah River and the more southern
Guale were far more similar than they were different. The main
distinction may have been a degree of political integration. While
there is evidence chat the entire Guale province was under some form
of control of a single head-chief, no such arrangement is indicated
among the Cusabo. However, it appears that the head-chieftainship
was a rather ephemeral office. The position may have developed as
a result of Spanish definition of the province and a requirement for
a single aboriginal representative.
The Guale and Cusabo spoke a common Muskogean language. When the
French under Jean Ribault visited the Guale in 1562, a Cusabo guide
had no difficulty communicating with the Guale. In addition, a
grammar composed by Spanish missionaries among the Guale was evidently
used by missionaries among the Cusabo. The most convincing evidence
of a common language is a statement by Governor Pedro Menendez Marques
in 1580 that the Indians of Santa Elena, i.e. the southern Cusabo,
were of the same linguistic province as the Guale. There are several
linguistic traits which indicate with little doubt that the Guale
were Muskogean speakers (see Swanton 1922:18 passim).
Timucuan has been classified recently as a "Language Isolate"
that was distinct from Muskogean (Crawford 1975:65-66). Timucuan
was divergent from Muskogean to the degree that priests who spoke
Timucuan found it necessary to employ interpreters when communicating
with the Guale (Swanton 1922:15).
Most of the subsequent discussion is focused on available infor-
mation concerning the Spanish-defined Guale. This information is
directly applicable to at least the southern-most Cusabo. Likewise,
I have occasion to employ accounts from the southern Cusabo for
additional information about the Guale.
Ethnohistoric information pertaining to the Guale is limited,
however certain basic structural elements of this coastal culture may
be defined. The Guale were swidden agriculturalists with settlements
organized into towns; they had a well-developed political structure
composed of micos and several other offices; and their kinship net-
works probably had a matrilineal structure and post-marital residence
was probably matrilocal Many details of this general structure
were shared by other Muskogean groups in the southeast, although
the Guale were distinctive in several respects.
The Guale resided in towns, each with a mico as political head
and representative. Groups of towns were united with allegiance to
a mico in one of the towns. There seem to have been three such
regional town groups and three regional micos. When Governor
Pedro de Ibarra visited the Guale in November of 1604, he met in
council with the micos and other officials from each region at or
near St. Simons Island, Sapelo Island and St. Catherines Island
(Swanton 1922:81,89). The location of these meetings may be used
approximately to divide the Guale area into northern, middle and
southern town regions. Additional evidence for these regional
town groups is that following the 1597 massacre of Franciscan
missionaries, the mico of Asao is spoken of as the head of the southern
group of Guale towns. There was also a head mico for the entire
Guale province. This individual is said to have exacted tribute and
was feasted upon his visits to various towns in the province
There is some evidence that indicates succession to the town-mico
office was structured within a defined kingroup. Don Juanillo of
St. Catherines Island is spoken of in 1597 as the "eldest son and
heir of the cacique of the island of Guale" (Barcia 1951:181). Given
matrilineal kinship organization, this individual was probably the
eldest nephew of the Guale mico and this relationship was expressed
by a father-son terminology. It'is unclear whether the offices of
regional and provincial mico were ascribed or achieved. Don Juanillo
is also referred to as the one, "whose turn it was to be head mico
of that province [Guale]" (Swanton 1922:34). Distinctions between
the use of the terms cacique and mico are ambiguous. Actual dif-
ferences between offices ray be implied, however Spanish use of the
terms is inconsistent.
Guale political structure was composed of several officials other
than the mico. The Spanish called these individuals mandadors,
aliaguitas, and other principal Indians (Serrano y Sans 1912 in
Larson 1978:124). These officers were certainly an integral part of
the Guale town councils. The councils that met with Governor Ibarra
probably contained the political nexus of each Guale town that was
Guale political structure appears similar to that of inland
Muskogean groups. Micos were also the leaders of Creek towns. The
role of the Creek nico is well illustrated by Speck (1907:113) as,
to receive all embassies from other tribes, to direct the
decisions of the town council according to his judgment, and
finally to stand as a representative of the town in foreign
The Creek council contained various officials and advisors of the mico,
including most frequently a heniha with peace functions and a tastanagi
with war'functions (Swanton 1928:276-334).
The Guale micos had some control over the goods of production. A
Guale mico called Oade gave Ribault's men food supplies in 1562. "This
good Indian was ready to do the favor as they were to ask it, and he
commanded his subjects to fill our boat with corn and beans"
(Laudonniere 1975:43). A short time later the French returned to
Oade for additional supplies. Oade sent word to Covecxis, another
Guale mico and referred to as his brother, requesting corn and beans
for the French. The next morning supplies arrived from Covecxis
Guale councils met in large council houses which were functionally
equivalent to Creek square grounds. There are only a few accounts
of these structures for the Guale, however each town probably had a
council house. These building were circular in shape and were usually
quite large. Individual apartments or cabins raised above the floor
lined the walls along the inside of the building, and in the center
of the structure was an open space for a fire and activities.
Two accounts of council houses are given in San Miguel's record
bf his 1595 visit among the southern Guale. At the town of Asao,
San Miguel witnessed a chunky game which was followed by a black
drink ceremony in the council house. "The Spaniards, caciques and
important Indians sat down, each on a bed which was supported by
poles from the floor" (Garcia 1902 in Larson 1978:129). San Miguel
(Garcia 1902 in Larson 1979:131) describes another Guale council house
circular in shape, made of entire pines from which the limbs and
bark had been removed, set up with their lower ends in the earth
and the tops all brought together above like a pavillion or like
the ribs of a parasol. Three hundred men might be able to live
in one: it had within around the entire circumference a con-
tinuous bed or bed stead, each well fitted for the repose and
sleep of many men, and because there was no bed-clothing other
than some straw, the door of the cabin was so small that it was
necessary to bend in order to enter: and due to the cold although
it was spring when we arrived; and so that one may not feel the
cold at night and may sweat without clothing it is sufficient
to cover the doorway at night with a door made of palmetto.
The best descriptions of a Guale council house are given by
Jonathan Dickinson in 1699, more than 100 years after San Miguel.
The Guale had become dispersed by this time and their middle Georgia
coast territory was mostly abandoned. Some groups had moved to mission
villages closer to St. Augustine and others had fled to their
Carolinean neighbors and the English (see Swanton 1922:90-92). The
description given by Dickinson applies to those Guale who had moved
to mission villages along the south Georgia and northeast Florida
coast. Just north of St. Augustine he visited the town of Santa
Cruz, which contained a large council house. Dickinson and his
were directed to the Indian war-house: it is built round having
sixteen squares; on each square is a cabin built and painted which
will hold two people; the house being about fifty foot diameter.
In the middle of the top is a square opening about fifteen foot.
This house was very clean, and fires being ready made nigh our
cabins [Andrews and Andrews 1945:87-88].
A little later the Quaker party visited a town called St. Mary's
where they were
conducted to the war house, as the custom is, for every town hath
a war-house. Or as we understood these houses were for their
[the Indians] times of mirth and dancing, and to lodge and
entertain strangers. This house is about 81 foot diameter
built round, with 32 squares, in each square a cabin about 8
foot long of good height being painted and well matted. The
center of this building is a quadrangle of 20 foot being open at
the top of the house, against which the house is built thus. In
this quadrangle is the place they dance having a great fire in
the middle. One of the squares of this building is the gateway
or passage in [Andrews and Andrews 1945:89].
These accounts indicate Guale council houses were remarkably
similar to Creek Tcokofas, also called round houses, rotundas,
sweat houses, or hot houses. Tcokofas were part of the Creek
ceremonial structure complex which also included a square ground,
defined by three or four opposing rectangular sheds, and a ball
ground. Tcokofas were winter council houses and the square ground
construction served this purpose during the summers. Guale council
houses may have combined the functions of the Creek Tcokofa and
square ground. At least some Guale towns also contained ball grounds,
as evidenced by the San Miguel account. The similarity between
Guale council houses and Creek Tcokofas is apparent in the descrip-
tion of a Tukabachee round house.
The main structure is supported upon twelve [emphasis in original]
pillars, one end sunk in the ground. They are disposed in a circle
about 9 or 10 ft. apart, making a space within of about 120 ft.
circumference, in the centre of which, upon the ground, is the
sacred fire. The roof over this circle is a cone terminating
in a point over the fire some 20 odd feet high. The rafters extend
down from the apex of the cone beyond the twelve pillars, which
are about 8 ft. high, to within 4 or 5 ft. of the ground, which
space, of 4 or 5 ft., is closed entirely with earth. Between
the pillars and the extreme exterior, a space of several feet,
are seats of mats, like those of the sheds [in the square
ground] [Hitchcock Ms. notes in Swanton 1928:179-180].
There is much less ethnohistoric evidence concerning domestic
structures and storage facilities in the Guale area. All that is
said of domestic structures is that, "all of the houses are small,
because, as they have little reason to keep in them, they make them
only for shelter" (Garcia 1902 in Larson 1978:131). Domestic
structures shown in the De Bry engravings that relate to the
general Timucua area are all small and round, and are possibly
applicable to the Guale (see Lorant 1946:39-115). Father Or6 notes
that granaries were common throughout La Florida and that "in them
the Indians place the maize they keep for their sustenance; it is
a type of barn supported by four posts, high and bulky, raised from
the earth"(Ore 1936:24). It is not mentioned whether these store-
houses were domestic or communal. It assumed that both types
existed. Considering Guale political organization, the mico's
granary was also probably the community storehouse.
Some Guale towns were palisaded but once again there is little
information on this point. A town called Yfusinique, in the northern
portion of the Guale province, was stockaded and provided a defensible
location for the perpetrators of the 1597 Guale revolt (Swanton 1922:
88). Palisaded towns were certainly common elsewhere along the coast,
as shown in the De Bry engravings of the Virginia area and northeast
Florida towns (Lorant 1946).
Much more conclusive evidence is needed about Cuale social organ-
ization, but it appears that kinship and post-marital residence
followed a Creek pattern. Polygamy was an aboriginal condition
that the priests were determined to abolish. The Guale were equally
resolved to maintain their marriage form. When confronted with
the Christian demand, the Indians replied: "If I leave her, I will
not have anyone to give me to eat and if I do not enter the house
where my children are, and if I do not bring them food and wood, they
will perish" (Ore 1936:101). This statement strongly indicates that
a house was the property of the wife. When a man had more than one
wife, he alternated his residence between the houses and provided
each household with food and services. Thus post-marital residence
appears to have been matrilocal. Since children lived in the house
and locality of the mother, and assuming that the place of social
orientation and pre-marital residence were the same, then a matri-
lineal kinship organization is certainly suggested.
Marriage to a mico may have resulted in modification of the
matrilocal residence rule. The relations of a mico are cited by
the priests to illustrate their missionizing difficulties, and they
urged the mico to set a "good" example for his people. Referring to
the "principal cacique" or head mico of the Guale province, Father Ore
(1936:101-102) states that,
During the time of his [the head mice's] apostacy he took into
his house as a concubine and mistress one of his sisters-in-law,
the sister of his own wife, with whom he lived all that time. By
her he had three children, and by his own wife four children ..
The fathers said the reformation of morals should start with him.
All they accomplished with him was that he put her in a separate
house, which was an ancient custom of the chiefs who placed in
a separate house each one of the women or lovers they had. Even
then the Indians complained: "Until now the cacique had in one
house two women and children; now he has two houses and in each
house he has a woman as if he were a pagan." The Indians
urged him to marry her. Neither did he nor she wish, nor did
anyone dare to marry her, for it was a custom that no one should
marry or speak to the wives or the lovers of the caciques.
These statements indicate that upon marriage the wives of a mico
were imported to his locality. The mico, his wives, and children
evidently resided in the area of his consanguineous kin group, his
matrilineage. Sororal polygyny is indicated as the marital form
in this case, but this may not have been the exclusive form. It
may well be that when wives were sisters, they resided in the same
house and when from different matrilineages, they lived in separate
houses (cf. Larson 1978:126).
These inferences have important implications for the Guale
social and settlement system. Marriage to a mico probably provided
a mechanism for social mobility in a ranked society. Upon marriage
the woman moved from her matrilineage to the locality of the mico
and there began a descent group spatially separate from that of
her own orientation, though still connected through consanguineous
ties. The children of the mico would have been members of their
mother's descent group rather than that of the mico, but they pro-
bably held a degree of prestige higher than that associated with
their matrilineage alone through their relationship with the mico.
The wife's matrilineage may have accrued additional prestige through
affinal ties with the mico and his lineage. Social taboos sur-
rounding the wives would have served to solidify the position of
the wife and her offspring in the residential area of the mico.
The taboos also would have made the rank of this new matrilineage
segment more secure.
Guale lineages were probably arranged into clans or sibs, although
there is no direct evidence to support this claim. There is a vague
reference to the Timucua that may apply to the Indians of La Florida
in general. Father Ore (1936:107) states that the Indians "consider
themselves related, provided they have the same names or lineages
even if there is a difference of a hundred degrees."
Analogy with Creek social organization may be used to supplement
our information about Guale social organization. Descent group member-
ship for the Creek was reckoned through the female line. Given a male
ego, members of his descent group included his mother, mother's brothers,
mother's sisters and their children, mother's mother and her sisters
and their children, ego's sisters and thier children, and ego's
brothers. Each domicile was owned by the wife. The household was
principally composed of a husband, wife, and their unmarried children.
Older sons and daughters whose spouses had died, plus the offspring
of the widow, and occasionally an orphan and war captives were in-
cluded in the household. The nucleus of the domestic unit was the
nuclear family. Households of the same matrilineage commonly re-
sided in the same area of a town, the husbands being imported from
other descent groups and the sons leaving upon marriage to reside
with their wives and their lineages (see Swanton 1928:79-97, 170-171).
Different matrilineages were united through mythical ancestry
to form exogamous sibs. It should be noted that the term "clan" is
used by Swanton (1928:114) to refer to these matrilineal groups which
acknowledge common descent. However according to Murdock (1949:41-78),
they properly define sibs. Sibs are distinguished from clans in
the Creek case because husbands retain their own lineage and sib
Analogy with the Creek pattern becomes less secure past this point.
Creek sibs were organized into phratries which were in turn divided
into moieties. Creek towns were designated either Red or White
depending upon their moiety affiliation and ball games were played
between towns of opposing colors. San Miguel's account of a
Guale chunky game indicates that the teams were from different towns,
possibly suggesting town moieties, but this is the only hint we have
of a dual division of Guale towns.
Certainly, the Guale political structure indicates an inte-
gration beyond the individual town level. The dynamics of this are
unclear,'but some Guale towns were surely allied through kinship
ties. The account previously cited from Laudonniere refers to Oade
and Covecxis as brothers. A literal interpretation is unwarranted,
but kinship between the two micos is definitely suggested. The
micos were probably either members of the same lineage or sib and
shared reciprocal responsibilities through this relationship.
A Systems Model of the Sixteenth Century Guale
What we know of Guale social and political organization is made
more intelligible with an examination of the resource base which
supported the cultural system. It is possible to construct an annual
model of Guale social, subsistence, and settlement systems based
on a small amount of ethnohistoric evidence and a heavy reliance
upon modern ecological data. A graphic presentation of the systems
model is provided in Figure 9 for reference in the following
discussion. Essentially the model provides a testable hypothesis
for archaeological investigation. Since it is predominately
constructed with evidence contained in accounts of early historic
period (pre-1600), elements of a purely aboriginal form should be
represented. The most intensive acculturation of the Guale
accompanied the renewed mission efforts that followed the 1597
The Guale planted corn, beans, and squash. Agricultural fields
were small and scattered throughout the highland areas of the coast.
Within the highland areas fertile soils occur in small pockets,
ANNUAL SYSTEMS MODEL FOR THE SIXTEENTH-CENTURY GUALE
presumably limiting the size of fields within an already restricted
area. The sandy coastal soil is marginally fertile at best, re-
quiring fallow periods between plantings for renewal (see Larson
1970:292-297). Discussing agriculture of the Guale around St.
Catherines Island in 1570, Father Sedeno states that,
the few Indians that are there are so scattered: because as they
do not have that with which to clear the trees for their fields
they go where they find a small amount of land without forest
in order to plant their maize: and as the land is so miserable
they move with their households [sus ranchos] from time to time
to seek other lands that they can bring to productivity
[Zubillaga 1946 in Larson 1970:295-296].
As Larson points out, the "small amount of land without forest"
probably refers to fallowing fields.
Accounts from the area immediately north of Guale also serve
to illustrate the scattered nature of the Indian fields and in
addition supply information about the social units involved in cult-
ivating the swidden plots. Father Juan Rogel, a Jesuit missionary
among the Orista, states that settlements are dispersed because,
the land will not support it nucleatedd settlement], because
it is very quickly weakened and miserable and exhausted. And
thus the same ones say that because of this they move around
so spread out and shift so regularly [Zubillaga 1946 in Larson
If my analogy with Creek social organization is correct, it appears
that the swidden plots were cultivated by related households of a
matrilineage, most usually two nuclear families. Again from Orista,
Father Rogel says that in the early spring,
those members of those twenty households [casas] distributed
themselves on twelve or thirteen farms [estancias] that were
some twenty leagues, some ten, some six, some four from
one another: and there were only two inhabitants that planted
[maize] around there [around the mission] [Zubillaga 19L6 in
Crops from the swidden fields were harvested in mid-summer.
Father Rogel indicates that planting occurred in the early spring,
probably just after the period of freezing temperatures which continue
into March. Harvest occurred in late June or early July. This harvest
was accompanied by a feast through which the scattered swidden families
were brought together in a single location. Produce from the scattered
fields supplied the feast. The town larder, the mico's granary, was
probably replenished at this time, grain from the preceding year being
now depleted. Support of these inferences is found in a statement by
Father Rogel that,
it happened that the alferez, Juan de la Vandera, deputy governor
of your grace in Santa Elena, was at a Escamacu feast [in late
June or early July 1570] and forced by necessity ordered three or
four caciques, there were Escamacu and Orista and Ahoya, that
they bring him certain [numbers of] canoes of maize by a certain
day to Santa Elena [Zubillaga 1946 in Larson 1970:195].
Guale populations were nucleated and sedentary for a time following
the harvest. The next subsistence phase indicated in the ethnohistoric
accounts involved a population dispersal to gather acorns. Father Rogel
arrived at Orista in 1569 during the harvest period. He says that,
this was the time that they were together, which was two and one-
half months; and the acorn harvest arrived, all left me alone, and
they were in these forests, each one in his area [cada uno por su
cabo], and they do not assemble except for certain feasts that they
hold twice in two months,and it is not always in one area [cabo]
but one time here and another in another place [Zubillaga 1946
in Larson 1970:281].
It would appear from this statement that the Guale resided in
towns from the first part of July until the middle of September.
The dates given by Rogel and his statement about gathering acorns
agree with contemporary information about oak trees and their fruit.
Acorns, as well as hickorynuts, begin to fall from their trees in
late August and continue until early December. Acorns, especially
those from white oaks, germinate soon after dispersal, requiring
immediate collection to retard spoilage (Olson 1974:695,698.)
Oaks'within the Tidewater Biome usually occur in stands
covering several acres. Precise data about the size and composition
of these stands are lacking; however, groves of 20 large live oaks
(Quercus virginiana Mill) are common and much larger stands exist.
On Sapelo Island during December of 1977, the ground beneath live
oaks was covered with acorns. This was perhaps the result of a good
year for acorn production, associated with the preceding unusually
cold winter on the coast. An estimated bushel of acorns was spread
beneath each large live oak.
Considering these estimates, the yield per season from a single
large live oak would be 120 pounds of acorns containing 50 pounds of
meat. The stand of 20 trees would produce about 1000 pounds of
acorn meat per season (see Olson 1974:Table 3). Divided over a
15-week acorn dispersal period, the stand would yield on the average
of 65 pounds of acorn meat every week. This converts to 180,000
calories per week, capable of supporting about 13 individuals for
7 days, considering a per capital per day intake of 2000 calories.
Presumably more than one oak grove would have been visited per
week, increasing the size of the group that could have been sustained
by the acorn harvest. The caloric value of acorns is a low estimate
of 600 calories per 100 grams, based on pecans, hickorynuts, and
walnuts (Watt and Merril 1963:44,34,65).
The point of this acorn assessment is to demonstrate that the
food energy available from the acorn harvest was substantial, and
capable of supporting larger groups in one area than the one or two
nuclear families that were involved in the swidden cultivation.
Of course the acorn harvest was accompanied by other subsistence
activities, the most important being deer hunting. White-tailed
deer are also drawn to oak groves during the fall to feed on the
acorn crop (see Larson 1970:247). Deer hunting and acorn gathering
were complementary subsistence activities. Deer feed in the early
morning and late afternoon, and are far less active during the
remainder of the day. Acorn gathering could have taken place through-
out the daylight hours. The acorn season is also the only period
of the year when deer regularly occur in groups, making a communal
The social organization and procurement model for the fall sub-
sistence phase suggested by the ethnohistoric and ecological data
is one of population dispersal from large towns sometime in September,
primarily to gather acorns and hunt white-tailed deer. The seasonally
abundant resources connected with oak groves were capable of supporting
several families in a single location. Acorns were locally abundant,
but perished soon after they fell to the ground. This type of
resource is most effectively harvested by many persons over a short
period of time.
Oak stands are scattered over the sections of highland along the
coast. The oak groves defined within a highland section may have
been revisited after new acorns had dropped, perhaps by the same
group in a cyclical pattern. Following a Creek social organization
pattern, the subsistence group probably defined a single matrilineage
with four or five nuclear families forming the social core. The
entire matrilineage may have been employed in communal deer hunts
using a surround or similar technique, or the older males in the
group may have ambushed the deer at their feeding grounds. The deer
hunt would have occurred in the twilight hours, and either technique
would have been productive.
Father Rogel also says that the acorn gathering groups came
together twice in two months at different locations for feasts.
This suggests a settlement component defined by towns composed of
temporary and changing populations, as opposed to the seasonally
stable population of the summer towns. The sites of summer towns
and the periodic towns may have been the same. A mico, his wives
and children, and members of his lineage were probably permanent
occupants of the town site, and exploited nearby oak groves during
the acorn season.
Town sites would have served as storage areas for surplus
produce acquired during the acorn season, possibly in the form of
tribute to the mico. The feasts would have been an occasion to settle
disputes, debate political matters, and an opportunity to develop
social relations with individuals outside of the matrilineage.
The produce brought in for the feast probably included not only
white-oak acorns and venison for immediate consumption, but also
less perishable commodities. Acorns from red-oak species (e.g.
Quercus laurifolia Michaux, Q. shumardii Buckley) and hickorynuts
(e.g. Carva tomentosa Nutall, C. glabra Sweet) could have been
successfully stored until the spring (see Olson 1974:699; Bonner
and Maisenhelder 1974:271). Dried venison and deer skins were
other storable items. Red-oak acorns require processing to leach
out bitter tannin before they are edible (see Larson 1970:269,
281-282). Since they may be stored, the leaching process could
have been completed after the acorn season was over.
A semi-permanent settlement pattern is suggested for these
lineage subsistence groups. Settlements were probably located in
relation to acorn resources. Since dispersement of oak groves is
restricted to circumscribed highland areas, the resources of several
oak stands could have been exploited from a single settlement
location. When these resource areas were exploited beyond the point
of supporting the lineage, the settlement would have shifted to
The next subsistence phase in the model relies heavily upon
ecological inference. Subsistence is hypothesized to have shifted to
a reliance on estuarine fish and shellfish following the acorn season
and continuing until the March agricultural activities. White
tailed deer probably continued to be exploited, but by the individual
hunter through stalking because deer were now much more solitary.
It is also likely that some stored nuts were processed and eaten
during this period. A matrilineage form of social grouping probably
remained the basic settlement and subsistence unit; however, settlements
were now dispersed within a more restricted environmental area.
Settlement probably shifted from scattered locations over the highland
oak-grove areas to those highland areas adjacent to tidal streams
which permitted access to the estuarine system.
The seasonal abundance of five families common in the tidal
streams of the Georgia coast, and which are also commonly represented
in the archaeological record, is presented in Figure 10. The fishery
.'.. 3- ..a =|g
: / -- 1
I -X -"
- -I "
K \\ i ,.
^ i I >i >
^t"~~ ^ -.v^
S" ^ ^ G a
data are from trawl catches in tidal streams from around the
Satilla River up to the Savannah River. The information was col-
lected over a three-year period, from October 1970 until September
1973 (Mahood et al. 1974b). Only those streams large enough to
admit a trawl vessel, measuring approximately 0S feet in length,
were sampled in this way. Furthermore, since trawling was restricted
to the deeper portions of the streams, those peripheral areas along
the banks are not represented. However as trawl sampling was in-
tensive and the sample size was quite large, the results are con-
sidered to be a reasonable indication of the seasonal variation of
fishes in the tidal streams.
Species within certain families of fish were becoming more abundant
in the tidal streams between January and March of the hypothesized
subsistence phase. A large portion of the increased abundance of
Sciaenidae is due to the occurrence of spot (Leiostomus xanthurus
Lacepede) in tidal streams near Sapelo Island. Several other
Sciaenidae are also present, including spotted sea trout (Cynoscion
nebulus Cuvier), Atlantic croaker (Micropogon undulatus Linnaeus),
and star drum (Stellifer lanceolatus Holbrook) (see Mahood et. al
197'b:31-32). Sciaenidae as a family actually occurs year-round
on the Georgia coast, but are most common during two seasonal periods,
from January through March and from June through August. The seasonal
abundance of Sciaenidae also varies depending upon their location on
the coast. For example, during the winter months Sciaenidae are
more common in the warmer waters of the south of Sapelo Island
than in the cooler estuarine waters of the middle and northern
Georgia coast (see Mahood et al. 197Lb:Table 4).
Sciaenidae occurring near Sapelo Island during the winter are
represented by small to medium-sized species. Spot reach a weight of
around 340 gm. and a length of about 40 cm. Star drum are also small,
reaching a length of about 25 cm. and weighing perhaps 900 gm.
Atlantic croaker are quite stocky, weighing as much as 2.3 kg. and
attaining a length of 45 cm. Sea trout are a larger species, reaching
a length of 65 cm. and a weight of 3.2 kg. (Mahood et. al 1974b:
32-33, Breder 1948:192-195).
Members of the Clupeidae family frequently occur within the
tidal streams of the Georgia coast from January through March.
Schooling species of Clupeidae which are present in tidal streams
near Sapelo Island during this period consist of blueback herring
(Alosa aestivalis Mitchill), menhaden (Brevoortia sp.). Atlantic
herring (Clupea harengus harengus Linnaeus), and gizzard shad
(Dorosoma cepedianum LeSueur). However, only the Atlantic herring
is found in the Sapelo Island tidal streams exclusively during this
winter period, and as with Sciaenidae; the Clupeidae family is more
abundant in the southern coastal waters during the winter (Mahood et
Shads and herrings of this family are anadromous. They inhabit
the tidal streams during the winter, prior to spawning in freshwater
rivers during the spring. Thus, the winter occupants are mostly
mature individuals. The species represented around Sapelo Island
are composed of small mature individuals, ranging in length from
20 cm. to 45 cm. and perhaps weighing between 250 gm. and 900 gm.
(Mahood et al. 1974b:23-21).
Sturgeons (Acipenser sp.) are a large anadromous fish and are
available in the tidal streams only during the winter and early
spring months. The winter sturgeon population consists of mature
individuals, as they spawn in freshwater rivers during the spring.
Atlantic sturgeon (Acipenser ozyrhynchus Mitchell) range from
around 30 cm. to 90 cm. in length and mature individuals perhaps
weigh around 6 kg., although individuals reaching a length of more
than 5 m. have been recorded (Mahood et al. 1974b:23, Breder 1948:
The American oyster (Crassostrea virginica Gmelin), and other
marine molluscs were probably exploited during the winter period.
Subtidal oysters are rare in the coastal estuaries. The Georgia
coastal oysters are intertidal and form beds on the firmer parts of
the tidal stream banks. These oysters spawn throughout the warmer
months of the year, continuing well into the fall. During the spawning
period, oysters are in poor condition and are diseased. A parasitic
fungus, Labyrinthomyxa sp., infects the intertidal oysters at a
high rate from June through December. The oysters are predominately
free from infection between January and May (Hoese ?1968).
Factors associated with this disease are poorly understood, but
the oysters appear most susceptible during the warmer months when
they are in poor condition. The fat and carbohydrate components
of oysters are lowest during the spawning period and markedly
higher from October to around April. Protein is highest during
August, but also increases from December to February (Lee and Pepper
1956). Present inhabitants of the Georgia coast refer to oysters of
the warmer months as being "thin and milky," and they are seldom
The most productive time, in terms of energy return, to harvest
intertidal oysters is during the winter and early spring months.
There is a 33% reduction in the meat weight of heavily diseased
oysters, diminishing the energy return from this species during the
warmer months (see Ray et al. 1953). Even undiseased oysters yield
lower meat weight in response to their summer spawning condition.
Referring to commercial oysters of the North Carolina coast and the
south in general, Chestnut (1951:159) states that,
Oysters may develop their sex products to spawn in [as late as]
the fall months, with the consequence that they are in poor
condition immediately after spawning, and yield a low volume
in meat content. A month or longer may be required to recover
from spawning. ..
While the late aboriginal occupants of the coast may have occasionally
gathered oysters during other periods, this was probably due to
subsistence stress or failures in other subsistence activities.
The common link between the winter resources is that most occur
as groups within the tidal streams. Sturgeon are probably the excep-
tion. The shad, herring and Sciaenidae tend to move in schools or at
least groups, and oysters occur in discrete beds. Thus, these winter
resources occur in clusters that are dispersed within the tidal
streams and are spatially analogous to the fall acorn and white-tailed
deer resources. The difference is that the winter species, with the
exception of oysters, are more mobile. However in aboriginal times,
as today, there were probably certain favored fishing places and to
this extent the location of fish in the tidal streams was predictable.
The same section of tidal stream could have been repeatedly fished.
Owing to this, settlements were presumably much more stable during
the winter months.
Social-group composition for the winter subsistence activities
was probably the matrilineage as in the fall. Since subsistence
resources occurred in clusters within the tidal streams, these could
have been most effectively exploited by several individuals at one
The particular technology employed was perhaps influenced by
factors beyond those imposed by the nature of the resources. Water
temperatures during the winter range around 100 C. (500 F.) to 150 C.
(59 F.) (Mahood et al 1974b:Figure 12). Air temperature during the
winter months is frequently around or below freezing. These uncomfort-
able temperatures certainly restricted procurement to a technique
that minimized water contact. All the fish species discussed above
could have been caught with hook and line; however, more than one
fishing technique may have been employed. Bottom-feeding species of
Sciaenidae may have been captured with basket traps, and cast nets
may have been used to catch members of the Clupeidae family. Sturgeon
were most probably procured exclusively with hook and line, as these
individuals are too large to effectively catch with cast nets or
basket traps. Procurement of oysters would have been little trouble,
as they are easily harvested at low tide by dislodging them from
their bed with a stick.
The winter procurement season was followed by spring planting,
thus closing the annual subsistence cycle. The March planting period
has already been discussed in respect to swidden activities and the
social dispersion that it entailed. The matrilineage social group
temporarily fissioned into nuclear family groups. The primary
settlement and subsistence unit was now composed of one or two
The period defined by the growth of crops was certainly a time
of subsistence stress on the coast. With few exceptions, potential
resources were neither abundant nor clustered within the tidewater
environment. For example in May of 1565 while among the Timucua,
Laudonniere speaks of a famine time when there is little to eat
except a few acorns, certainly from red oak, and fish (Laudonniere
Food supplies stored in the granaries of Guale towns probably
postponed the shortage for a short time. A few festive occasions
may have served to redistribute this stored food. For example
early in the spring in April 1566, Pedro Menendez returned two
slaves captured by the Guale (St. Catherines) to Orista. This,
along with the Governor's visit, was cause for festivities. On
this occasion, "many Indian women [came],carrying maize, fish
boiled and roasted, oysters and many acorns.. ." (Meras 1964:175).
The maize and acorns were certainly stored foods from the preceding
year. The oysters were probably fresh, but were surely the last
of the season. The fish may have been anadromous, taken from fresh-
water spawning runs.
Several anadromous species occur on the Georgia coast. Shads,
herrings and sturgeons enter the freshwater rivers to spawn and are
available in quantities for a short period of time. The period of
migration varies somewhat with the species, but spawning is generally
a spring activity. Blueback herring (A. aestivalis Mitchill)
ascends the rivers in the spring, hickory shad (A. mediocris Mitchill)
in the late winter and spring, American shad (A. sapidissima Wilson)
from January to March, and the Atlantic sturgeon (A. oxvrhvnchus
Mitchill) during the spring and summer (Dahlberg 1975:37, Larson 1970:
177). Juvenile American shad and blueback herring have been caught
as far inland as the Oconee and Ocmulgee Rivers, over 200 km. from
salt water (Smith 1968). This probably indicates that some mature
individuals spawn well upstream during the spring.
The temporary abundance of anadromous species was certainly
noticed by the aboriginal occupants in a period which was otherwise
defined by scarcity. These fish may have been exploited by groups
of families during the spring after the fields were planted. This
subsistence trek possibly carried the Guale families well upstream,
beyond the Tidewater Biome.
The period following the late June harvest, as I have discussed,
was accompanied by a population aggregation in the towns. This large,
seasonally sedentary settlement continued until the fall acorn harvest.
Although the cultivated foods were certainly important during this
time, there was probably a renewed interest in estuarine resources.
This period corresponds with the second seasonal abundance peak in
the tidal stream species. The Sciaendae family and sea catfish are
prominent summer inhabitants.
Many species of Scianenidae are abundant in the tidal streams
near Sapelo Island during the summer months. Most notably these
species include red drnu (Sciaenops ocellata Linnaeus), black drum
(Pogonias cromis Linnaeus), Atlantic croaker (Microgogon undulatus
Linneaus), and star drum (Stellifer lanceolatus Holbrook). Atlantic
croaker and star drum were discussed as a winter resource, but are
more common in the tidal streams during the summer. The red and black
drums can be very large fish. Black drum reach a weight of about
34 kg. and a length of around 1.5 m. However the maximum length
reported in trawl catches on the Georgia coast is about 50 cm;, sug-
gesting a more usual weight of about 10 kg. Red drum are generally a
larger species, reaching a length of 1.1 m. and a weight of 66 kg.
(Mahood et. al. 1974b:31-33, Breder 1948:194-197).
Two species of saltwater catfish are abundant in the tidal
streams during the summer. Sea catfish (Arius felis Linnaeus) are
most common. This species reaches a length of about 45 cm. and a
weight of about 500 gm. is usual. Gafftopsail catfish (Bagre marinus
Mitchill) are less common but slightly larger, reaching a length of
55 cm. and weighing as much as 1.8 kg. (Mahood et al. 1974b:25,
Zinmand Shoemaker 1955:63).
Procurement technology at this time perhaps shifted to tech-
niques that employed larger numbers of people. A wide range or
fishing techniques could have been used in the tidal streams during
the warm summer months, including hook and line, basket traps, and
any of several netting techniques. Larson (1970:184-191) has argued
against the use of weirs along the Georgia coast. The basis for his
argument centers around the exceptionally high tides and unstable
bottom conditions in this area. These conditions would have made
constructions unstable and the weirs would have required continual
This hypothesis about the use of weirs on the Georgia coast is
carefully thought out and cannot be reproached based on present in-
formation. The hypothesis is, however, in dire need of testing with
data from experimental archaeology. At this point an alternative
hypothesis may be proposed for testing. While conditions that appear
prohibitive to permanent weir constructions exist along the beach, the
mouths of rivers, and in most other estuarine areas, small tidal traps
could have been constructed at certain locations within the estuary.
Only those areas immediately adjacent to large intertidal oyster
beds have a stable substrata. Weirs may have been constructed in
these areas, as they appear to lack the factors that mitigate against
construction in other tidal waters.
Whether weirs or some other subsistence technology was employed,
the oyster beds were probably at least one focal point for aboriginal
subsistence activities, as several species apparently feed here at
high tide. The most important to the aboriginals may have been the
Sciaenids (see Dahlberg 1972:351). Catfish and other species are
probably attracted to oyster beds by the presence of many associated
small invertebrates. Wells (1961) recorded an average of 43 species
associated with oyster beds in North Carolina. Durant (?1968).
recovered from 11 to 21 species from intertidal oyster beds along
the Georgia coast. Most of these recorded species are small organisms
such as annelids and boring sponges, and predators such as conch and
oyster drills. These species, along with oysters, provide a concentrated
food source for fish that is available when the oyster bed is covered
by the tide. Oyster beds have never been sampled in a way to provide
firm information on this point, but it is entirely logical that the
beds provide a localized feeding habitat for many fishes at high tide.
It should be stressed that the subsistence stages presented
in the above model deal with what are considered to be the primary
resources available within a given season, those which supplied
concentrated sources of exploitable energy. Resources other than
those discussed were certainly exploited and were certainly important,
such as wild grapes in the fall and several types of berries during
the spring and early summer. Deer would have provided an important
food source throughout the year, along with smaller mammals like
raccoons and rabbits. With a few important exceptions, it is the
abundance rather than the presence of particular resources that
varies with the season.
In summary, it is hypothesized that the Guale annual cycle
was composed of four distinct settlement and subsistence systems.
Each seasonal subsistence activity was executed by particular social
units. The form of each social group was a response to the nature
of the resources and the technology used for exploitation.
Although the model portrays the seasonal activities as static,
in reality, the seasonal boundaries were certainly more flexible.
Seasonal divisions would have shifted somewhat due to yearly fluc-
tuations in resources. For example, the acorn harvest is variable
from year to year and from grove to grove (Larson 1970:280). Poor
harvest years would have prolonged the summer subsistence period and
caused the winter subsistence phase to begin somewhat earlier. Some
overlap exists between the seasonal resources.
Four basic levels of settlement are indicated in the model.
The smallest settlement was composed of one or two nuclear families
during the agricultural season in the spring. This family settlement
pattern was quite dispersed throughout the coastal area. The second
type of settlement unit was composed of a matrilineage comprised
of 20 to 25 individuals. This lineage settlement had dispersed
locations during the acorn harvest in the fall. Settlement location
became more concentrated during the winter near the estuary.
It is likely that town sites were permanently occupied by a mico,
his wives and children, and perhaps his lineage. These residents would
have exploited areas adjacent to the settlement throughout the year.
The town sites were probably located in an area which could support
a small sedentary population. This location would have been in an
area which maximized access to productive resource zones. That is,
town sites were in a location which provided direct access to productive
estuarine areas, oak forest, and agricultural land. These settlements
were the location of periodic feasts held during the fall, spring
and probably winter subsistence seasons. During these feasts the
population level at the town site increased sharply for short
periods of time.
Town sites were also the centers for population aggregation
during the summer months and contained a relatively large and
sedentary population. The regional micos and provincial mico
resided in certain towns. These towns probably formed the apex of
the settlement structure and contained the largest summer population.
The political and social structure suggested by the ethnohistoric
accounts and the model indicate that Guale socio-political organization
was a form of chiefdom. A segmented and ranked system is clear,
although the exact form and components remain to be adequately
demonstrated. It would seem safe at this point to assume that the
micos and their lineages formed the apex of the socio-political
hierarchy. Micos were certainly at the top of the power structure,
but the mode of status and power acquisition is unclear. At least
the position of town mico appears to have been ascribed rather than
achieved, as the ehtnohistoric information suggests a line of suc-
cession to this office. Further, since another account indicates
some micos were related, it is probable that micos came from certain
lineages or sibs. In sum, aside from all the ambiguities that exist,
the Guale social system was segmented and these segments were ranked.
The function of the chiefdom organization may be understood in
terms of the annual model. Like the model, the interpretation should
be considered tentative and subject to verification and revision.
But even if certain details of the model prove to be inaccurate, the
basic nature of the interpretation should be valid if the general
form of the model is correct.
The Guale chiefdom may be understood in terms of information
processing and to a lesser extent the redistribution of goods.
Elman Service (1962:143) has pointed out that one characteristic
of a chiefdom is "the presence of centers which coordinate economic,
social and religious activities." Guale towns or town sites
functioned in this respect during the various subsistence phases.
The seasonal nature of the resource base, along with the dispersed
distribution of seasonal subsistence items, required an organization
and scheduling mechanism for the maintenance of the cultural system.
The construction and execution of this mechanism lay beyond the
capacity of individual subsistence groups.
The variant information held by each subsistence group was
channeled into town sites during periodic feasts that were held
during the fall, winter and spring subsistence phases. This in-
formation was processed, probably by councils, at the feast times
and decisions were made about future subsistence activities. The
office of mico probably functioned to voice and legitimize this
The chiefdom was essentially a socio-political mechanism
by which diverse information was acquired, processed, and then re-
distributed as a summary decision. Knowledge held by individual
groups about resources that were abundant in certain areas and
scarce in others was funneled into single locations. A coordinated
dispersal of the subsistence groups into mutually exclusive resource
areas followed decisions based on these data.
The chiefdom organization would have been perhaps most im-
portant towards the end of a subsistence season, when timing decisions
had to be made about subsistence and settlement shifts. Information
available to each group about decline of continued abundance of
current food resources and about the availability of new resources
was processed and a determination made about the appropriate course
Chiefdoms and their potential for information processing is
well illustrated in a recent discussion by Peebles and Kus (1977).
The Guale case supports their arguments. Speaking of hierarchical
social segments they point out that,
information is filtered at the lower levels and passed on to
the higher level regulator in summary form. The higher level
regulator can then deal with a variety of events that cannot
be simultaneously handled by the lower level units. For cultural
systems, hierarchical arrangements not only increase the system's
ability to process energy and information, but facilitate greater
internal complexity and external variability [Peebles and Kus
The authors emphasize the importance of ritual in this information
transfer and decision process and argue that redistribution and
ecological specialization should be abandoned as requisites for
Earlier assessments of ranked social organization stressed the
importance of ecological differentiation and the redistributive
function of the chief. For example, Service (1962:143-148) emphasized
ecological differentiation and sedentariness as major factors in the
development of chiefdoms. Goods from distinctive environmental zones
were channeled to the chief and were then redistributed so that
diverse food sources moved to the people rather than the contrary.
Overall production was increased, creating a surplus which supported
the chief and perhaps craft specialists. The surplus could be drawn
upon by the population during the times of scarcity.
This surplus and redistributive argument has been made of followed
by others (e.g. Fried 1967, Adams 1975), and remains a useful
base for archaeological explanation although it is usually difficult
to detect in the preserved elements of material culture. It seems
fairly obvious that both information processing and redistribution
are important functions of a chiefdom. The importance of one or
the other probably increases depending upon the particular environ-
mental context and subsistence technology. In areas with marked
zonation of contemporaneous subsistence resources, the redistributive
function becomes most important. The function of information
processing becomes more crucial to the maintenance of the system
in environments containing seasonally homogeneous but spatially
dispersed food resources. Redistribution actually remains a
primary characteristic in either case. What changes is the material
reallocated: goods in the former and information in the latter
case. Any particular chiefdom is likely to include both aspects.
Redistribution of stored foods was an important, though secondary
function of the Georgia coast chiefdoms. The creation of seasonal
surpluses, housed in the mico's granary, provided some support for
the population during the period of spring scarcity. The town
granary was replenished following the swidden harvest period. This
summer larder probably supported community activities such as council
meetings, public building construction and repair, large ceremonial
and festive events, and the entertainment of foreign and neighboring
dignitaries. Warfare might also be included, but the densely pop-
ulated towns would have been at their defensive peak and more vulnerable
during other seasons of the year when the resident population was
The nature and diversity of information processed by the system
would have changed dramatically during the summer months. Ecological
information was more homogeneous since the population was nucleated
and exploited a more restricted estuarine environment. Attention
at this tine may have been focused more on inter-town relationships;
their maintenance, termination, or realignment.
An excursion into the impact of western contact and mission
activities on this complex cultural system would require a separate
treatise., Certain elements of the system were surely already affected
by European influences, since the model is constructed with data from
between 1560 and 1600, more than 40 years after Allyon's first contact
with the coastal groups and in the midst of deliberate acculturation
attempts by the early missionaries. The earliest Jesuit efforts ended
in failure and the Franciscans were temporarily thwarted until after
the Guale revolt of 1597. An example of the cultural disruption that
accompanied slightly later mission activities is provided in an
account by Father Ore that describes Spanish retaliation immediately
following the 1597 revolt.
Since all the Indians were hidden in the woods, the governor
could neither punish them nor get in touch with them. They
burned the foodstuffs of the Indians: the Indians themselves
already burned their houses when they left. On this account
and due to what followed, during the subsequent years they had
no maize harvest. Moreover since they were removed from the
sea, they could neither fish nor gather shellfish, with the
result that they suffered great hunger. Though the Indians
sowed, it was little, while the Spaniards destroyed it every
year [Ore 1936:95].
Two of the main areas to be changed by the missionaries were
aspects of social organization and subsistence. The priests were
committed to the elimination of Guale polygyny and in the process
destroyed social networks which the marital form maintained.
Missionaries also required a sedentary population to whom they could
adminster Christian doctrines. The Guale finally became sedentary
and agricultural, although population size decreased. According to
my interpretation of the Guale Chiefdom, this settlement and subsistence
change would have destroyed the primary function of the aboriginal
The Mississippian Period is a temporal segment of Southeastern
prehistory which begins roughly around 700 A.D. with developments
at the Cahokia site and continues until the middle of the sixteenth
century. The De Soto expedition of 1539 is used to mark the end
of this period, as the disruptions and diseases which followed De
Soto did much to alter aboriginal culture in most of the Southeast
(e.g. Griffin 1967:189-191, Larson 1970:2, Willey 1966:292-310).
The term "Mississippian" refers to a number of related cultural
forms which developed during this period of time. The nuclear area
and nominator for this "Formative-level culture" (Willey and Phillips
1958:163-170) is the central Mississippi River Valley. Regional
variants of the culture developed in other areas of the Southeast,
each defining distinctive adaptations to particular environmental
and historical contexts (e.g. Caldwell 1958). Some common features
of this adaptation are a central position of agriculture in the
economy, fully sedentary villages, relatively large populations,
pronounced social segmentation, platform mound construction, and
elaborate ceremonialism. I am concerned here with developments on
the Georgia coast during the Mississippian Period and their relation
to Mississippian events and processes occurring elsewhere in
the Southeast, particularly in Georgia.
This period on the Georgia coast is defined by two cultural
phases called Wilmington and Savannah. Some would also include the
Irene "hase within the terminal portion of this period, however
indications are that Irene is an early historic aboriginal phase.
I briefly review in this section the available evidence for each of
these phases and provide an assessment of the associated cultural
forms and the changes that occurred on the Georgia coast during the
Mississippian Period. Important locations discussed in the text are
shown in Figure II.
Wilmington Phase (ca. 700 1000 A.D.)
Wilmington-Phase sites, defined in terms of a distinctive pottery
type, have an extensive coastal distribution. Wilmington heavy
cord-marked, sherd-tempered pottery (Caldwell and Waring 1939a:
113-116, Caldwell 1952:316-317) is commonly found at coastal sites
from around the St. Marys River at the Georgia-Florida border north
into North Carolina.
Wilmington pottery along the Georgia coast is most frequently
heavy linear cord marked and tempered with large pieces of ground
sherd; however, tempering materials and paste characteristics can
be quite variable. While sherd tempering is most common, non-plastic
inclusions may frequently range from sand to larger quartz grit.
About 14% of the Wilmington Cord-Marked sherds identified from St.
Simons Island had sand and grit inclusions in the paste to the ex-
clusion of ground-sherd fragments (Martinez 1975:70). Wilmington
pottery identified at the Groton Plantation was exclusively sand
tempered (Stoltman 1974). Variability in the composition of paste
and tempering materials is expectable, and probably results from
variation in the physical properties of exploited clay sources. It
may well be that riverine clays used in the Groton-Plantation area
were suitable for pottery manufacture without inclusion of additional
GEORGIA CRINAGE SYSTEM BASE MAP SCALE
L E G E N D 0 50 100
A SAVANNAH RIVER I WILMIG TC. ISLAND 8 GROTCN PLANTATION
9 YEDWAY RIVER 2 CSSABA'.V ISLAND 9 EBENEZER C.TEEK
C ALTAMAHA RIVER 3 ST CAT:--iR!NES ISLAND 10 IRENE MCUND SITE
-6 ( ,: I ,.--. i-_
D SATILLA RIVER 4 SAPELO ISLAND !I SEPTFORD BLUFR
E ST. MARS RiVER 5 ST S;.V.C.)S iSLAND 12 -CRMA, M .-UND
F OGEECHEE RIVER 6 CUMBERLAND ISLAND 13 SiG MCRTAR-SNUFcBOX
IMPORTANT LOCATIONS DISCUSSED IN THE TEXT
2 :' -
/ -.:, .. ,i-.... .
. ., I .
GECRGIA DRAINAGE SYSTEM SASE MAP SC A L E
L E G E N D o so 100
A SAVANNAH RIVER I WILMINGTCN ISLAND a GROTCH PLANTATION
8 MEDWAY RIVER 2 CSS-ABA, ISLAND 9 EEENEZER CREE<
C ALTAMAHA RIVER 3 ST. CAThERINES ISLAND I0 IRENE MCUND SITE
O SATILLA RIVER 4 SAPELD ISLAND I PEPTFOP.C BLPF;
E ST. MARYS RIVER 5 S. S;.,!C;S iSLAND 12 "CR;:AI4 I.f4UNC
F OGEECHEE RIVER 6 CUMDERLAND ISLAND 2BiG MCRTAR-SNIUF'BOX
G CC,"J.LC-EE RIVER 7 AMELIA ISLAND I OCMIJLGEE siC 8END
15 MACON PLATEAU
IMPORTANT LOCATIONS DISCUSSED IN THE TEXT
Wilmington pottery occurs infrequently south of the St. Marys
River, for example at sites on Amelia Island (Bullen and Griffen 1952,
Hemmings and Deagan 1973). This northeastern Florida area was the
home of a different and longer-lasting cultural phase known as St.
Johns (Goggin 1952). The northern boundary of Wilmington is more
poorly defined, but South (1976) reports Wilmington pottery along
the southern portion of the North Carolina coast. Distribution
probably ends somewhere in North Carolina. The sherd-tempered ware
is very rare in Virginia, although cord marking remains a popular
decorative technique for a long period of time (Evans 1955).
There is reason to believe that the Wilmington pottery distri-
bution is associated with variant cultural adaptations. Anderson (1975)
has shown that north of the Santee River in South Carolina the pottery
is distributed along inland rivers up to the Fall Line as well as
along the coast. This possibly suggests a coastal and interior
riverine adaptation, and perhaps seasonal settlement shifts. South of
the Santee River, Wilmington sites are most commonly located along
the coast and are rarely encountered in inland, riverine contexts
beyond the Tidewater Biome. The southern Wilmington sites are located
either adjacent to tidal streams and salt marsh or in upland, oak-forest
areas in shell-free contexts. A tidewater settlement and subsistence
system is indicated for the southern Wilmington sites, however ad-
ditional cultural similarities and differences remain to be demonstrated.
Arguments about the origin of Wilmington are speculative at this
point. Caldwell (1958:33-34) regards Wilmington as a coastal intrusion
with marked differences, primarily in pottery, from the preceding
Deptford Phase. The tetrapodal, check-stamped, sand-tempered Deptford
pottery was replaced by linear cord-marked, sherd-tempered, cylindrical
jar forms of Wilmington. He suggests coastal North or South Carolina
as the origin of this intrusion, but also points out that burial
mounds, defined as Wilmington-Phase trait, are absent in the northern
areas. This unshared trait led him to conclude that "the original
Wilmington homeland may not be easy to discover" (Caldwell 1958:34).
That Wilmington burial mounds exist at all is a debatable point.
As Stoltman (1974:24-27) has argued, none of the supposed Wilmington
mounds can definitely be attributed to the Wilmington Phase. Certainly
none of these mounds have burials associated with Wilmington pottery
vessels, nor have any of the burials been dated to the Wilmington
Period with Carbon-14 estimates.
New evidence presented by South (1976) gives some support to the
northern-origin hypothesis. South suggests that Wilmington Cord
Marked pottery is a late component and Hanover Fabric-Marked pottery
the earlier component of what he calls the Wilmington Ware Group.
While both components of this group extend from North Carolina
south to around the Charleston area, only the Wilmington component
extends further south. As I understand South, this indicates that
Wilmington developed out of Hanover and spread south along the
Atlantic coast into Georgia after the Deptford Phase. South stops
short of actually saying this, but it is the logical conclusion of
his chronology and distribution argument.
Others have suggested that Wilmington was affiliated with interior
Georgia. Waring (1968) proposed that the coastal Wilmington intrusion
was contemporary and somehow associated with the Mississippian movement
into the Macon area. Recent Carbon-14 dates suggest that Wilmington
precedes the Macon Plateau period and that his theory is probably
an error. Milanich and others (1976) have suggested that some
interior coastal plain riverine sites with cord-marked pottery are
simply interior Wilmington sites. Recent investigations along the
Ocmulgee Big Bend area (Snow 1977) have shown an extensive distribution
of small sites with cord-marked pottery. This pottery series, called
Ocmulgee Cord Marked, has broad folded rims, cross and linear cord
marking, and usually sandy to temperless paste. While the variant
paste could result from differences in clay sources, the other
characteristics seem to distinguish this series from Wilmington
pottery. Firm temporal determinations are needed for the Ocmulgee
Cord-Marked pottery series before any adequate assessment of coastal
interior relationships can be made.
Turning away from questions of origins and interior affinities,
an examination of the Wilmington Phase in its secure Georgia coastal
setting is more important for our present purposes. As previously
mentioned there appear to be two types of Wilmington sites in the
Tidewater Biome: marsh-edge, shell-deposit sites and highland oak
forest sites. More is known about the marsh-edge sites than the less
visible shell-free, oak-forest sites.
The Walthour site, located on Wilmington Island at the mouth
of the Savannah River, is a Wilmington-Phase marsh-edge site. This
site consists of intermittent deposits of shell midden extending
parallel to the marsh. Caldwell (1958:34) says that at the Walthour
site, "shell heaps were several feet high and a solid bank of oyster
shell lined the edge of the tidewater marsh from which shell fish were
secured." This linear arrangement also characterized the Deptford
site, located on a bluff along the Savannah River (Caldwell and
McCann?1940). The stratigraphic position of Wilmington-Phase pottery
over that of the Deptford Phase was clearly demonstrated from tests
along the bluff. A number of features and burials were encountered
at the Deptford Bluff but their cultural proveniences were undem-
onstrated. They were assumed to belong to either the Deptford
or Wilmington Phases. Features included wall trenches and numerous
pits filled with midden debris, but no house patterns were recognized.
Forty-two burials in flexed or extended positions were also excavated
along the bluff, but these lacked accompaniments that would indicate
their associated phase.
Additional information concerning Wilmington-Phase shell midden
sites comes from barrier islands along the central Georgia coast.
Carbon-14 estimates from two Wilmington-Phase shell middens on St.
Catherines Island indicate dates of A.D. 735 + 115 and A.D. 905- 200
(Caldwell 1970). This temporal position is supported by a Carbon
14 date of A.D. 820 70 from a Wilmington shell midden on St.
Simons Island (Milanich 1977). The Wilmington midden tested on St.
Simons was one of many shell deposits scattered along the marsh edge
at Cannon's Point. Another shell midden in this area contained mixed
components ranging from the fiber-tempered pottery of the St. Simons
Phase to Wilmington Phase. Refuse appears to have been deposited
along the marsh edge for a considerable period of time, at least
until the Wilmington Period (Martinez 1975:60-66).
The Wilmington marsh-edge middens suggest a settlement and sub-
sistence pattern with an estuarine focus. There is little information
available about subsistence other than the obvious exploitation of
marine molluscs, primarily oysters. The only analysis of Wilmington
faunal components is a small sample gathered from the midden on St.
Simons Island (Martinez 1975:60-63, 90-95). Fauna other than molluscs
included white-tailed deer, raccoon, mink, pine snake, mud turtle,
diamondback terrapin, eagle ray, drum, sheepshead, sea catfish, herring,
and mullet. Remains of at least two deer and two diamondback terrapin
were identified; however,the remaining species could have been re-
presented by only one individual each. The small vertebrate sample
prohibits reliable seasonal determinations, but subsistence evidently
involved exploitation of a variety of habitats including freshwater,
estuarine, and terrestrial areas. Based on earlier arguments presented
about oysters, it would seem that a major portion of the midden was
deposited during the winter or early spring. The turtles, pine snake,
and catfish possibly suggest that deposition was in the early spring
period, as these species are generally inaccessible during the cool
Wilmington pottery was found scattered behind the midden deposits
at Deptford Bluff and St. Simons Island. This may indicate that
residences were located adjacent to the linear area of refuse deposition,
but again no house patterns of this phase have been identified.
The presence of Wilmington-Phase sites in interior areas of the
Tidewater Biome is documented by three recent surveys. The first is
a survey of the Big Mortar-Snuffbox Swamp in Long and McIntosh
Counties, Georgia (Zurel, Gresham and Hally 1975). The next survey
is of the Ebenezer Creek watershed, located between the Ogeechee
and Savannah Rivers (Fish 1976, 1978). The third is of the Groton
Plantation along the northern side of the Savannah River (Stoltman
1974). The Big Mortar-Snuffbox Swamp is located on the mainland
opposite Sapelo Island and the other two survey areas are along the
western edge of the Tidewater Biome (cf. Fish 1976:2, 1978:334).
The Groton Plantation appears to be located just beyond the Tidewater
Biome, but the Wicomico Formation briefly turns at the Savannah River
and extends far enough northwest to include the surveyed area (see
Cooke 1936). Groton Plantation is, nevertheless, in a marginal
These interior coastal sites with Wilmington components are
located on river bluffs and on pockets of well-drained soil surrounded
by freshwater swamp. The upland locations of these sites indicate a
dispersed settlement pattern that probably results from seasonal
exploitation of oak-forest resources such as acorns and deer (cf.
Stoltman 1974:216). The bluff locations would have provided access
to the large rivers and these sites may have been utilized for oak
forest as well as riverine exploitation.
Present evidence allows only a tentative assessment of the
'ilmington subsistence system. It appears that small groups exploited
oak-forest resources during the fall and a wide range of terrestrial
and estuarine food sources during the winter and early spring. At
some part of the year, perhaps during the spring, they may have relied
on riverine resources. It is clear that Wilmington-Phase settlements
were small and scattered, and never large and nucleated. This
suggests a rather low level of socio-cultural integration. We may
expect that the Wilmington Phase was associated with small, semi-nomadic
hunting and gathering groups organized at a band level of socio-cultural
integration (see Service 1962:107-108).
This band-level organization adapted to the Tidewater Biome
evidently had a substantial history. The same general dispersed
settlement pattern in the oak forests and along the estuaries long
preceded the Wilmington Phase. Although pottery styles changed through
time, this basic cultural form applies to populations who lived on the
coast since the appearance of pottery at approximately 2500 B.C.,
and perhaps even earlier. A single exception to this organizational
level may have been the Late Swift Creek coastal settlement, as evi-
denced by the large mound and village complex along the lower Altamaha
River at the Evelyn Site (Waring and Holder 1968:1l0). Little is
known about this intrusion, but it seems to have a short life and
was possibly an unsuccessful colonization attempt by a more culturally
advanced interior group (see Cook 1978 for other speculations about
Late Swift Creek and Kelvin on the Georgia coast).
Evolutionary changes certainly occurred during this 3500-year
period of coastal prehistory, such as slowly increasing sedentariness
and gradual population growth. However subsistence technology and
socio-political integration appears to have remained remarkably stable.
Milanich (1971)has called this conservative adaptation the "Coastal
Tradition." Whether the cultural phases of this tradition were
intrusions along the coast or in situ developments, as Milanich
argues, the coast appears to have been occupied by band-level cultures
for a long period of time.
Savannah Phase (ca. 1000 1540 A.D.)
The Savannah Phase marks the end of the band level of socio-
cultural integration and the beginning of more complex socio-political
arrangements on the Georgia coast. This change evidently resulted from
modifications in existing Wilmington-Phase adaptations rather than
purely technological advancements or invasion and displacement by some
exotic group. The Savannah Phase is characterized by nucleated settle-
ments, platform mounds, and the extensive use of single locations as
cemeteries resulting in burial mound constructions. Cord-marked pottery
decoration persisted but now was defined by finer marking and crossed
as well as linear designs. Check-stamped, complicated-stamped, and
burnished-plain pottery were added to the Savannah complex, as well as
more specifically Mississippian shapes such as cazuelas.
Settlement locations occupied during the Wilmington Phase continued
to be used during the Savannah Phase. The dispersed Wilmington-Phase
settlement pattern applies to the Savannah Phase as well, but was now
part of a more complex cultural system. This was also the time of
population growth on the coast as evidenced by the greatly increased
number of Savannah-Phase archaeological sites. For example in the Big
Mortar-Snuffbox Swamp (Zurel. Gresham, and Hally 1975), there are
!.5 times as many Savannah-component sites (n=18) as Wilmington-component
Most of our information about the Savannah Phase comes from
excavations around the mouth of the Savannah River. The Irene
Mound site, located on a bluff overlooking the river about five miles
above the city of Savannah, is considered the classic example of
a Savannah-Phase ceremonial center (Caldwell and McCann 1941). This
site is quite small, covering approximately 2 hectares (5 acres), and
its Savannah-Phase features are a platform mound with seven construction
levels and ascending ramps, most of the interments in an adjacent
burial mound, numerous long wall trenches that divide the site into
segments, and three small domestic-like structures (Features 26,
53, 56) located within the enclosed segments. Each construction
stage of the platform mound was associated with rectangular buildings
and the final platform stages were also surrounded by enclosing walls.
Several important conclusions may be drawn from internal evidence
from the Irene Mound site. The energy expended in construction of
the platform mound is greater than that which can be explained by the
size of the population resident at the site. Labor forces beyond
the site must have been employed for construction. The enclosed
portions of the site indicate internal spatial organization, and
again this plan probably functioned to organize activities beyond
those of the small resident population. One of the enclosed areas
extends south and west of the platform mound and defines a large,
open plaza-like area. This area and the platform-mound buildings
were probably associated with activities involving a larger, but
less visible segment of the Savannah-Phase population.
The Savannah-Phase portion of the Irene site certainly reflects
only a segment of a functioning cultural system. The social function
of this segment may be hypothesized based on the construction features
represented at the site. The platform mounds and enclosures indicate
socio-political authority in terms of organization of a sufficient
labor force for their construction. The domicile-like structures
suggest that the site was also a small residential area. The size
of these houses suggests that they were primarily nuclear family
residences. While there is little else that would distinguish
these structures as high-status residences, their presence at a site
which is otherwise defined by communal features indicates that they
denote an integral part of the socio-political structure. It appears
that the site was occupied by a central political figure, and probably
his immediate relatives. The division into two enclosed areas may
have separated the lineage of the chief from his wife or wives and
his children, although much more evidence is needed to demonstrate
The social position of the Irene site needs to be considered
when comparing the Savannah-Phase material culture represented at
the site with other Savannah sites. It should be remembered that
the Irene site was a Savannah-Phase political and ceremonial center
rather than a purely residential area. Chronologies developed within
the Savannah Period have ignored the possibility of social distinctions
so clearly suggested at the Irene site.
Ideal Savannah-Phase divisions consist of an early ceramic
complex, Savannah I, defined by fine cord-marked and burnished-plain
pottery. The later pottery complex, Savannah II, is defined by the
addition or at least increased abundance of check-stamped pottery,
the addition of complicated-stamped designs, and the continuance of
cord marking and burnishing. The cazuela burnished-plain forms
continued to be made but vessels with other surface treatments con-
sisted of jars with flaring rims and small bowls (see Caldwell
1952, Caldwell and Waring 1939a, cf. Caldwell and Waring 1939b).
The Savannah II complex was characteristic of pottery recovered
from the Irene site. The Savannah I complex was identified from
mound contexts such as the Deptford burial mound (Caldwell and McCann
nd.), Haven Home or Indian King's Tomb (Waring 1968), and several
other mound and village areas around the mouth of the Savannah River.
Caldwell (1952:318) suggested that several mounds in the central
Georgia coastarea may also be of the Savannah I period.
That Savannah II is later than Savannah I is, with a single
exception, totally a logical chronology and without stratigraphic
support. The exception is the Norman Mound in McIntosh County,
Georgia (Larson 1957). Here the sub-mound levels contained high
frequencies of Savannah Fine Cord-Marked pottery at the expense of
Savannah Check-Stamped pottery. The central shell core and the
mound fill contained increased amounts of the check-stamped variety.
Check-stamped and burnished-plain pottery accompanied the Savannah
Phase burials. While the stratigraphy supports Savannah II as being
later than Savannah I at the Norman Mound, it is important that pre-
mound and mortuary strata are being compared. The pottery differences
could still be explained in terms of a mortuary versus domestic
association, rather than by temporal differences.
It is probably significant that on St. Simons Island, Savannah
Cord-Marked pottery appears to have increased through time at the
expense of check-stamped decorations, suggesting that Savannah I
is later than Savannah II. This may indicate that the two complexes
are actually contemporary: however, based on present evidence
conclusive arguments can be made neither for chronological nor social
differences in the Savannah pottery complexes. It is likely that
both processes were operating in the manufacture, use, and deposition
of pottery and we must be aware of the two possibilities.
A certain amount of regionalization occurred during the Savannah
Phase that appears to be directly associated with the two major
freshwater river systems. The two Savannah-Phase regions may be
defined as the Savannah Region and the Altamaha Region. Their
boundaries are somewhat amorphous, probably because of sparse infor-
mation, but can be tentatively defined. The Savannah Region extends
from around the Coosa River in South Carolina to the \Medway River
in Georgia and the Altamaha Region includes that area of the Georgia
coast from the Medway River to around the Satilla River. A marginal
extension of the Altamaha Region continues as far south as Amelia
Island, however occupation appears to have been concentrated to
the north. Each region is confined to the Tidewater Biome.
Both regions share the main diagnostic traits of the Savannah
Phase such as extensive use of burial mounds, nucleated settlement,
and fine cord-marked, check-stamped, and burnished-plain pottery.
The distinctive features of the Savannah Region include platform
mounds and a relative abundance of complicated stamped potter'-; however,
these two elements are uncommon even in the Savannah Region. Only
one platform mound other than that at the Irene site is documented
in the region, and it is assumed to be a Savannah-Phase construction
without direct evidence. This mound, Indian Hill, was investigated
by C.B. Moore and is located on St. Helena Island, South Carolina.
This truncated mound was about 13 ft. high, its base measured 138
ft. by 129 ft., and its level summit was 62 ft. across. Moore
(1898:16L) says that, "a number of post-holes from which the wood
had rotted, still unfilled, were found in four distinct levels ..
No burials were met with and we must regard the mound at Indian Hill
as erected for domiciliary purposes."
Savannah Complicated-Stamped pottery is a clear indication of
interior contact with the coast. The coastal pottery is decorated
with figure-eight, various concentric circles, and nested-diamond
stamped designs. The similarity with Wilbanks stamped pottery in
northern Georgia is so striking that for several years Wilbanks was
referred to as Savannah Complicated-Stamped pottery (see Fairbanks
1950: Wauchope 1966). Sears (1958) named the northern Georgia pottery
complex "Wilbanks." Etowah Complicated-Stamped designs, also a
Mississippian Period northern Georgia type, also are found on Savannah
Complicated-Stamped pottery. Etowah designs include nested diamonds
and other rectilinear elements in addition to the later Wilbanks
designs such as figure eights, concentric circles and scrolls. The
infrequency of complicated-stamped pottery with these designs in the
Savannah Region may indicate that many are actually trade wares.
Savannah-Phase cazuela bowls are another indication of interior
Mississippian Period influences, but unlike the complicated-stamped
pottery, this form has a widespread Savannah-Phase distribution.
The Altamaha Region lacks many of the outward signs of interior
influences. The negative evidence includes the absence of platform
mounds and much less Savannah Complicated-Stamped pottery. The
Savannah-Phase pottery complex in the Altamaha Region is at variance
with the pottery types defined for the Savannah Region. While details
of these differences are presently undemonstrated, many archaeologists
working on the coast have noted that the Savannah River type de-
scriptions are only partially applicable to the central Georgia
coast during Savannah or later Periods (Caldwell 1970: Larson 1955,
1958; Milanich 1977). Definition of these differences is an important
problem for current research on the coast. A conclusion that can be
drawn at this point is that this variation seems to be associated with
the Altamaha Region.
Arguments could be made that the Savannah Phase in the Altamaha
Region is simply the result of spreading influences from the Savannah
Region. However, certain developments just prior to the Savannah
Phase along the Ocmulgee River near the Fall Line suggest that this
may have been an additional direction of Altamaha Region influences.
The Macon Plateau Phase at Ocmulgee clearly represents the intrusion
of a fully agricultural, stratified population with fortified, planned
villages containing temple mounds and buildings with politico-religious
functions. The estimated temporal range of the phase extends from
around 900 co 1100 A.D., making it contemporary with the beginning
of the Savannah Phase (see Fairbanks 1956, Wilson 1964).
The Macon Plateau Period was rather short and transportable material
culture elements such as pottery and religious paraphernalia were
sytlistically simple. These factors may explain the sparsity of
observable Macon elements during the Savannah Phase. The multitude
of negative evidence proposed for the Altamaha Region is insecure
grounds for assessment, however if one considers that social changes
were responsible for developments in the Savannah Phase, then the
possibility of Macon Plateau influences becomes more credible.
That the Savannah Phase was built directly upon a Wilmington
socio-cultural system and included descendents of the Wilmington
population is a long-standing hypothesis. Cord-marked pottery and
sherd tempering are only two indications of Wilmington affiliations.
Based on his Savannah I distinction, Caldwell (1952:317) says that
this period "follows Wilmington, apparently without a sharp break
and might easily have been named 'Late Wilmington.'" When discussing
the Savannah II Period he concludes that it may be "regarded as a
fusion of the old coastal culture with Middle Mississippian influences
from the interior" (Caldwell 1952:318).
Savannah-Phase settlement pattern also shows a continuation of
parts of the Wilmington system. Sites with Savannah-Phase components
occur in the same locations as Wilmington and earlier phase sites
(see Stoltman 1974; Fish 1976; Zurel, Gresham and Hally 1975).
Savannah settlements are located throughout the Tidewater Biome in
dispersed highland areas surrounded by swamp, on river bluffs, and
on well-drained land along or near the estuaries. Important Savannah
Phase additions were nucleated settlements near estuaries and occasional
small settlements along the freshwater rivers beyond the Tidewater
Biome. One of these interior riverine settlements is represented by
a site (9AplO) located along the upper Altamaha River in Appling
County, Georgia. Savannah Check-Stamped and Fine Cord-Marked pottery
was identified from surface collections at this site (Snow 1977:L2).
Nucleated settlements are rather infrequent and rarely have been
investigated. Extensive excavations have been carried out only at
the Irene site. Major nucleated settlements are found on Ossabaw Island
at Middle Settlement (9Chl58) and Bluff Field (9Chl60)(Pearson 1977),
on St. Simons Island at Cooper's Field and Indian Field (Wallace 1975),
on Cumberland Island at High Point (NPS 9Cam35) (Ehrenhard 1976),
and on Sapelo Island at Kenan Field and Bourbon Field. Only the
Cooper's Field and Kenan Field sites can be assigned to the Savannah
Phase with confidence. Indian Field and probably Bourbon Field are
Irene and Sutherland Bluff-Phase sites. Pearson (1977) argues, based
on surface collections and limited test excavations, that both
large sites on Ossabaw Island are Irene Phase. While the pottery
recovered from Middle Settlement suggests an Irene-Phase provenience,
assignment of Bluff Field is based on 1l sherds and is hardly secure.
Further investigations will probably show that nucleated sites occur
on other barrier islands and perhaps the adjacent mainland, particularly
near the junction of major freshwater rivers with estuaries as at the
Irene site. It also appears that this settlement form continues into
later periods, but usually in different locations.
Savannah-Phase subsistence probably also followed the Wilmington
pattern. Considering interior influences, it is probable that agricul-
ture was practiced but there is no direct evidence. There are no
analyses of faunal materials associated with closed Savannah-Phase
contexts, but the numerous shell midden deposits attest an estuarine
focus. The high-ground swamp sites were probably, as during the
Wilmington Phase, occupied for the exploitation of oak-forest resources.
The interior riverine locations were possibly bases for the exploitation
of anadromous fish. The nucleated estuarine settlements possibly
indicate that subsistence technology involved a larger-group effort
that that represented by the scattered settlement pattern.
It is obvious that much more information is needed before secure
conclusions can be drawn concerning many parameters of this prehistoric
adaptation. However, it is important that for the first time the
available evidence suggests a cultural system much like the one
presented in the ethnohistoric model. The Savannah-Phase system may
have assumed this general form.
There is some evidence that the Savannah Phase extended into
the historic period, but a terminal date is uncertain. One of two
dogs buried in a Savannah-Phase cemetery at Cooper's Field was
accompanied by the musket ball that was evidently the cause of death.
Wallace (1975:121-141) concludes, through rather complicated and
debatable means, that the burial complex was associated with a
mortuary building and perhaps a mound. Pottery in the disturbed soil
over the burials indicates extensive use during the Savannah Phase,
and the dog was probably killed and buried late in that period. This
evidence suggests that the Savannah Phase survived into the period of
European contact. We tentatively mark the end of the Savannah Period
at 1550 A.D., however it could have ended any time between 1521 when
Allyon's men visited the coast and the last decade of the sixteenth
century. There would have been little evidence of Spanish contact
during this time since military and missionary activities were sporadic
and generally unsuccessful.
In conclusion, the Savannah Phase provides an example of a
Mississippian Period acculturation process. Details of this process
will undoubtedly become more clear as research progresses; however,
some general suggestions can be offered at this time. As Ralph
Linton has concluded, there are two common features of culture contact
and the transfer of culture elements.
First, borrowing [of culture elements] is normally a reciprocal
process, and second, its logical, although by no means always
its actual, end product is the amalgamation of the two cultures
involved, resulting in a new culture differing in certain respects
from either of its parent cultures [Linton 1940:491].
What we know of the Savannah Phase demonstrates the latter point.
It was a blend of elements from the Wilmington Phase and those from
a more complex cultural configuration, i.e. an interior, chiefdom-level
cultural system. The result was a distinct cultural system adapted
to the coastal environment. The central role of agriculture was
certainly diminished in the infertile coastal soils, and hunting and
gathering remained of primary importance. The "culture elements"
transferred to the coast appear to have been primarily social and
political. This resulted in the reorganization of Wilmington settle-
ment and subsistence systems towards greater productivity.
Invasion and annihilation, transformation, and incorporation
are three broad categories that define the nature of cultural contacts
between socieites of unequal socio-political complexity (Fried 1952).
Savannah-Phase developments may be understood in terms of a mixture
of incorporation and transformation, but possibly without the violent
military action that Fried includes in these processes. For the
Savannah-Phase case, the process appears to have involved incorporation
and alteration of the existing cultural form under an exotic social
and political superstructure. This process commonly applies to
marginal areas such as the Georgia coast, in which the natural resources
are of limited importance to the more complex cultural system. To
the degree that the contact area is isolated, the foreign superstructure
is precluded from continual contact with and support from the parent
culture and becomes an integral part of the contact-cultural system.
Irene-Sutherland Bluff Phases (ca. 1540-1680)
This period on the Georgia coast is composed of Irene and
Sutherland Bluff Phases. The Irene Phase is defined by a Pine Harbor
variant that is restricted to the Altamaha Region and an Irene
variant that is restricted to the Savannah Region. The Irene Phase
is followed by the Sutherland Bluff Phase in the Altamaha Region.
Since this period is properly beyond the scope of the present inquiry,
discussion is limited to a few comments and certain factors which
have direct bearing on the Savannah Phase. The Altamaha and Savannah
Regions continued to define areas of distinctive but related material
culture assemblages and were probably associated with culturally
similar populations. This Savannah-Phase pattern was intensified and
may be recognized with greater clarity during the Irene Phase.