Front Cover
 Table of Contents
 From the editors
 The owl totem
 The mounds themselves might be...
 A reevaluation of the Gainesville,...
 Geoarchaeological investigations...
 Petrographic evaluation of Belle...
 Wacissa boat
 Osceola's garter
 Book reviews
 About the authors

Group Title: Florida anthropologist
Title: The Florida anthropologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00027829/00201
 Material Information
Title: The Florida anthropologist
Abbreviated Title: Fla. anthropol.
Physical Description: v. : ill. ; 24 cm.
Language: English
Creator: Florida Anthropological Society
Conference on Historic Site Archaeology
Publisher: Florida Anthropological Society
Florida Anthropological Society.
Place of Publication: Gainesville
Publication Date: June-September 2007
Copyright Date: 2007
Frequency: quarterly[]
two no. a year[ former 1948-]
Subject: Indians of North America -- Antiquities -- Periodicals -- Florida   ( lcsh )
Antiquities -- Periodicals -- Florida   ( lcsh )
Genre: periodical   ( marcgt )
Summary: Contains papers of the Annual Conference on Historic Site Archeology.
Dates or Sequential Designation: v. 1- May 1948-
 Record Information
Bibliographic ID: UF00027829
Volume ID: VID00201
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltqf - AAA9403
oclc - 01569447
issn - 00153893
lccn - 56028409
issn - 0015-3893

Table of Contents
    Front Cover
        Front Cover 1
        Front Cover 2
    Table of Contents
        Page 57
        Page 58
    From the editors
        Page 59
        Page 60
    The owl totem
        Page 61
        Page 62
    The mounds themselves might be perfectly happy in their surroundings
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
    A reevaluation of the Gainesville, Ocala, and Lake Panasoffkee quarry clusters
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
    Geoarchaeological investigations in the flats of the Osceola Plain, Highlands and Polk counties
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
    Petrographic evaluation of Belle Glade and Sandy St. Johns pastes
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
    Wacissa boat
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
    Osceola's garter
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
    Book reviews
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
    About the authors
        Page 159
        Page 160
Full Text



VOLUME 60, NUMBER 2-3 June-September 2007

-_ I -., ,

Al .


THE FLORIDA ANTHROPOLOGIST is published by the Florida Anthropological Society, Inc., P.O. Box 357605, Gainesville, FL 32635.
Subscription is by membership in the Society. Membership is NOT restricted to residents of the State of Florida nor to the United
States of America. Membership may be initiated at any time during the year, and covers the ensuing twelve month period. Dues shall
be payable on the anniversary of the initial dues payment. Members shall receive copies of all publications distributed by the Society
during the 12 months of their membership year. Annual dues are as follows: student $15, individual $30, family $35, institutional $30,
sustaining $40 or more, patron $100 or more, life S500, and benefactor S2500. Foreign subscriptions are an additional $5 U.S. to cover
added postage and handling costs for individual, family, or institutional membership categories. Copies of the journal will only be sent
to members with current paid dues. Please contact the Editors for information on recent back issues.

Requests for information on the Society, membership application forms, and notifications of changes of address should be sent to the
Membership Secretary. Donations should be sent to the Treasurer or may be routed through the Editors to facilitate acknowledgment in
subsequent issues of thejournal (unless anonymity is requested). Submissions of manuscripts should be sent to the Editors. Publications
for review should be submitted to the Book Review Editor. Authors please follow The Florida Anthropologist style guide (on-line at
www.fasweb.org) in preparing manuscrpts for submission to the journal and contact the Editors with specific questions. Submit four (4)
copies for use in peer review. Only one set of onginal graphics need be submitted. The journal is formatted using Adobe In Design. All
manuscripts must be submitted in final form on CD in Microsoft format. Address changes should be made AT LEAST 30 DAYS prior to
the mailing of the next issue. The post office will not forward bulk mail nor retain such mail when "temporary hold" orders exist. Such
mail is returned to the Society postage due. The journal is published quarterly in March, June, September, and December of each year.


President: Patty Flynn, 160 W. Camino Real #190. Boca Raton. FL (pflynn@pbmnh.org)
First Vice President: Robert J. Austin, P.O. Box 28 I. Riverview. FL 33568-2818 (bob@searchinc.com)
Second Vice President: Steven Martin, (martinswa earthlink.net)
Corresponding Secretary: Antoinette Wallace. (ab_ allace@bellsouth.net I
Membership Secretary: Kay Gautier, P.O Box 13191, Pensacola, Fl 325911 (kaygautier aol.com)
Treasurer and Registered Agent: Joanne Talley, P.O.Box 788, Hobe Sound. FL 33475 (jo@ whiticar com)
Directors at Large: Gloria Fike, 2815 SE St. Lucie Blvd.. Stuart, FL 34997 (GloriaFike@prodigy.net); Bill Lucus, 333 Persimmon
St., Freeport, FL 32439; Christine Newman. 504 17th St., St. Augustine, FL 32084 (aci_staugustine@bellsouth.net)
Immediate Past President: Sheila Stewart. 2130 Burlington Avenue North, St. Petersburg. FL 33713 (sjstew@gte.net)
Newsletter Editor: David Burns, 15128 Spnngview St., Tampa. FL 33624 (daveburns(@prodigy.net)


Co-Editors: Deborah Mullins. P.O. Box 357605. Gainesville. FL 32635-7605 (dmullins.fl.anthropoiogist@gmail.com)
Andrea White, P.O. Box 6356. Tallahassee, Fl 32314-6356 (awhite.fl.anthropologist@gmail.com)
Book Review Editor: Dan Hughes. 2301 8th Ave. North, St. Petersburg, Florida 33713 (hughesarc juno.com)
EditorialAssistant: George M. Luer, 3222 Old Oak Drive, Sarasota, FL 34239-5019 (gluer@grove.ufl.edu)
Printer: Durra-Print, 717 South Woodward Ave.. Tallahassee, FL 32304
Bulk Mail: Capital City Mailing, 4013 Wood\ille Hwy, Tallahassee. FL 32311
Back Issue Sales: Palm Beach Museum of Natural History (wwwv.pbmnh.org)


Albert C. Goodyear, Institute of Archaeology and Anthropology, University of South Carolina. Columbia, SC 29208
Jerald T. Milanich, Florida Museum of Natural H:story. University of Florida, Gainesville. FL 32611 (jtm@(flmnh.ufl.edu)
Jeffrey M. Mitchem, Arkansas Archeological SurveN. P.O. Box 241. Parkin, AR 72373 (jeffmitchem@juno.com)
Nancy Marie White, Department of Anthropology. L university of South Florida. Tampa. FL 33620-3100
(nwhite@chuma 1 .cas. usf.edu
Robert J. Austin. P.O. Box 2818. Riverviev. FL 3356.s-28 18 (bobla searchinc.com)

NOTE: In addition to the above Editorial Review Board members, the review comments of others knowledgeable in a manuscript's
subject matter are solicited as par: of our peer re\ iew process.

VISIT FAS ON THE WEB: www.fasweb.org



U OF F .. R R
Volume 60 Number 2-3
June-September 2007 C


From the Editors 59

The Owl Totem. Barbara A. Purdy 61

"The Mounds Themselves Might Be Perfectly Happy in Their Surroundings": The "Kolomoki Problem" in 63
Notes and Letters. Thomas J. Pluckhahn

A Reevaluation of the Gainesville, Ocala, and Lake Panasoffkee Quarry Clusters. 77
Jon C. Endonino

Geoarchaeological Investigations in the Flats of the Osceola Plain, Highlands and Polk Counties, Florida. 97
Michael Wilder, Charles D. Frederick, Mark D. Bateman, and Duane E. Peter

Petrographic Evaluation of Belle Glade and Sandy St. Johns Pastes. 117
Ann S. Cordell

Wacissa Boat 1 (8JE1604):Example of a Plantation Flat in a North Florida River. 127
Jeffrey T. Moates

Osceola's Garter: An Analysis of a Nineteenth Century Native American Textile. 139
Mary Spanos, Virginia Wimberley, and Amanda Thompson


Mould: Choctaw Prophecy: A Legacy of the Future. Anne McCudden 153

Crooks: Jacksonville: The Consolidation Story, From Civil Rights to the Jaguars, Georgia. Hope Black 154

Pluckhahn and Ethridge: Light on the Path. Skye Wheeler Hughes 155

Eidse: Voices of the Apalachicola. EJ Ford 157

About the Authors 159

Cover: (Left) View to the east of Mounds D (lower left foreground) and A (center background) at Kolomoki,
(Right Top) Osceola's garter, (Right Center) Belle Glade Plain sherd showing large scraped facet and characteristic
drag/scratches (Photo by Mr. Pat Payne), (Right Bottom) illustration of the fastening processes of a plank-
built flat featuring the use of iron drift pins (Drawing by William Judd). Please see articles for more information.

Published by the
ISSN 0015-3893


Here we span time, topic, and even geography in
this something-for-everyone double issue of The Florida
Anthropologist. In looking for commonalities among the
subjects considered in this issue, we were struck by the concern
from a majority of our contributing authors to keep their focus
on people as their special object, regardless of the fact that
these investigations focus on a single artifact, the formation of
specific site types, particular artifact attributes, and an analysis
of a special class of artifacts (in the form of written letters).
While a focus on people may seem like a given in archaeology,
it definitely is not; especially when one considers how easy it
is to become distracted from that focus by the very things we
consider to get at them: material remains. And it is an entirely
understandable predicament. Who can refute how interesting
and consuming methodological development or the close study
of artifacts of every sort can be-or how central a nuanced
understanding of these artifacts is in getting at our main object?
In the pages of this issue we find fallibility and pride lurking
behind the written word, we find traditional craftsmanship and
the adaptability of those skills in the face of social pressures
and changes, and we catch a glimpse of the ancient cultural
landscapes so well understood and purposefully shaped for
use and gain by past Floridians.
An opening note by Barbara A. Purdy explains and
celebrates the Owl Totem that graced the cover of the last issue
of The Florida Anthropologist (Vol. 60, No. 1) and serves as
the 60th anniversary emblem of the FAS. The level of artistic
achievement embodied in the owl totem commemorates
the people and culture that created it and is a true national
treasure. The Society would like to once again thank artist
and archaeologist James W. Hunter III for his contribution of
the owl drawing, and Barbara A. Purdy for her willingness to
enlighten the readership on the significance of our anniversary
Our first article by Thomas Pluckhahn visits the human
side of the relationships among key founders of Florida
and Georgia archaeology and expands on the silence over
interpretive problems related to the Kolomoki site cultural
chronologies-a problematic silence that ultimately lessened
the importance of the site in regional interpretations. By
focusing on the personal correspondence between William
Sears and his contemporaries, including Caldwell, Fairbanks,
Griffin, Kelly, Waring, Willey, and others, Pluckhahn relates
Sears's refusal to revisit his misinterpretations of the Kolomoki
ceramic chronology even in confidence to his closest
colleagues and long past the point where those same associates
acknowledged among themselves the "Kolomoki problem"
via a steady stream of private discussion. The editors are very
pleased to bring you this important contribution on the history

of southeastern archaeology. Readers interested in learning
more about the Kolomoki debate should turn to Vernon J.
Knight and Frank T. Schnell's 2004 article "Silence Over
Kolomoki: A Curious Episode in the History of Southeastern
Archaeology" (Southeastern Archaeology 23(1): 1-11).
Lithic material is a common artifact found on
archaeological sites, and on precermaics sites is often the only
material evidence of past human activity. The result of several
years of research, Jon Endonino's publication takes a look at
Florida chert quarry clusters and the implications of sourcing
lithic material recovered in an archaeological context. Building
on the 1982 work of Upchurch, Richards, and Nuckels, as
well as other archaeologists, Endonino addresses deficiencies
in distinguishing between different Ocala Limestone cherts
and refines the diagnostic criteria for sourcing these cherts,
primarily by looking at the size and abundance of fossils
found within the stone. From a methodological standpoint, the
author's work is especially exciting because he has developed
a set of analytical criteria concerning samples, techniques, and
procedures that will allow other archaeologists to replicate and
refine this type of provenance sourcing for lithic material. A
detailed knowledge of chert resources on a landscape allows
archaeologists to address issues of social complexity on
sites where little other data is preserved-as is the case for
thousands of archeological sites across Florida.
The authors of our third paper, Michael Wilder, Charles
Frederick, Mark Bateman and Duane Peters, advance our
understanding of archaeological and geoarchaeological issues
in an understudied region of Florida through their application
of both C-14 and Optically Stimulated Luminescence (OSL)
dating of deep sand sites in the Kissimmee River Basin. Sure
to stimulate interesting discussion and comparative testing is
the contention of the authors that sites in sandy environments
are moving down en masse due largely to processes of
pedoturbation as opposed to being buried by the combined
effects of sedimentation and pedoturbation over time.
In a related article, Ann Cordell uses pottery samples from
the sites discussed in Wilder et al. and contributes an important
scholarly analysis on the sometimes frustrating and subjective
study of plainwares. Cordell attempts to clarify some of the
typological uncertainty related to the identification of Belle
Glade and sandy St. Johns pottery based on the measurement
of sponge spicules from a sample of thin sections. The paper
is straightforward, well written, and introduces a more precise
definition of these two pottery types that will undoubtedly
be incorporated into the working terminology of Florida's
archaeological community.
The historic vernacular watercraft of Florida is both an
understudied and dwindling resource. Archaeologist Jeff


VOL. 60(2-3)



Moates contributes to our understanding of the traditions of
knowledge reflected in the construction methods and daily use
of plantation craft, and his article documents the geographic
and temporal expansion of a vessel type previously unknown
outside of the southern Tidewater region and largely associated
with slaves and freed workers. Important for future research and
vessel documentation, Moates also establishes an appropriate
set of terms for use in the recording of specific construction
elements of vernacular craft.
In our last article, Mary Spanos, Virginia Wimberley and
Amanda Thompson present an in-depth textile analysis of a
garment owned and worn by Osceola, a unique individual who
made a permanent impact on his contemporaries as well as
future generations of Native Americans. The technical analysis
of Osceola's garter, a rare example of a Southeast Indian
textile, provides insight into continuities of form, construction
technique, and materials in textile traditions that would have
been passed from generation to generation. This last piece is
something of a unique submission to the FA and will hopefully
motivate other non-archaeologist friends of FAS to contribute
interesting research to this journal.
Taken together, these articles take very different paths to
that same intersection of people, action, materials, and time,
and we hope all of our readers can find something to inspire
them. Please also read the four book reviews contributed
by Hope Black, EJ Ford, Skye Wheeler Hughes, and Anne
McCudden-their efforts are much appreciated. The editors
would particularly like to thank the reviewers of the articles
featured in this issue of the journal. Their gifts of time and
thoughtful critique ensure a continued tradition of quality for
The Florida Anthropologist.

Deborah R. Mullins
Andrea P. White

2007 VOL. 60(2-3)



The owl totem, uncovered by a dragline operator, was pulled from the muck of the St. Johns River
i 'near the famous Thursby Midden (8-VO-35) in late June 1955 (Bullen 1955:61). Made from the heartwood
of a magnificent Pinus species, the owl is the largest prehistoric wood effigy in the Western Hemisphere
'. -. with a length of 12 feet, 2 inches. The carved portion is 6 feet, 3 inches. The feet are perched on a prepared
S pedestal that is more than four feet in length. Two additional feet at the bottom of this extension had partially
'.71 ""52 deteriorated when recovered. Because of its large size, the totem certainly was not portable and must have
fallen into the river as a result of a storm or from slope erosion.
The function of the owl totem may never be known. It could have been a marker for the village of the owl
S clan, or for a cemetery. The owl figured prominently with death in the belief systems of some Native American
S---- groups. Perhaps the Spanish toppled the figure some time in the late 1500s in their efforts to destroy images
worshiped by the "pagan" Indians.
i Shell, stone, and shark's teeth were employed to create the specimen using a combination of natural
and conventional art. It is thought to represent the great hored owl. The presence of five talons instead of
I four also suggests anthropomorphism. The sides and back of the owl are richly engraved, and were carefully
and symmetrically planned to represent stylized feathers. Through distinct incisions, the talented artist
S differentiated between wing and tail feathers, radial feathers typical of owls, and still other special feathers
S around the beak.
Bullen (1955) estimated that the totem dated to about A.D. 1400, the St. Johns IIb subperiod. In 1958, a
portion of the rotted base was radiocarbon dated (University of Michigan M-464) at 650+200 (ca 1100-1500)
or about A.D. 1300 (Bullen 1958). Radiocarbon dating was still in its developmental stages and no calibration
techniques were available; nevertheless Bullen's original assessment is probably accurate.
The owl was on exhibit at the Florida State Museum located in the Seagel Building in downtown
Gainesville from 1955 to 1971. When a new museum (Dickinson Hall) was built in the shape of a temple mound on the University
of Florida campus, the owl reposed on its back on the floor in the anthropology range until Congressman Charles E. Bennett (U.S.
House of Representatives) was instrumental in having it transferred to Fort Caroline National Memorial. At Fort Caroline the owl
was moved through the years from place to place in the exhibit hall. At present, it is displayed only from the talons upward, and is
against a wall so that the beautifully engraved back is not visible.
Other than periodically shellacking the outer surfaces, the owl totem has not undergone preservation treatment; i.e., it has not
been subjected to increasing concentrations of polyethylene (PEG). The carving appears amazingly stable; nevertheless, it should
be monitored regularly by a conservator knowledgeable about wood degradation.
During drought conditions in 1977, two smaller specimens, a pelican and an otter(?), were recovered from approximately the
same area as the owl. The carved portion of each is about 2.5 feet. Based on the craftsmanship, there seems little doubt that the
pelican and otter were manufactured by the same artist who produced the owl, or an apprentice. To my knowledge, the pelican and
otter carvings have not been dated. They are in Tallahassee at the Museum of Florida History.
The owl totem is not just a state treasure, it is a national treasure. I am happy to have the opportunity to put together this
information so that others may know of its significance.
Barbara A. Purdy

Further Reading

Bullen, Ripley P.
1955 Carved Owl Totem, De Land, Florida. The Florida Anthropologist 8(3):61-73.

1958 More Radiocarbon Dates and Their Significance. The Florida Anthropologist 11(4):97-109.

Purdy, Barbara A.
1987 Investigations at Hontoon Island (8VO202) An Archaeological Wetsite in Volusia County, Florida. The Florida
Anthropologist 40(1): Special Issue.

1991 Hontoon Island. In The Art andArchaeology of Florida Wetlands, 102-138. CRC Press, Boca Raton.

Schwehm, Alice Gates
1983 The Carved Wood Effigies of Fort Center: A Glimpse of South Florida's Prehistoric Art. M.A. Thesis, Department of Fine
Arts, University of Florida, Gainesville.

VOL. 60(2-3)





Department ofAnthropology, University of South Florida, 4202 E. Fowler Ave, SOC107, Tampa, FL 33620
Email: tpluckha@cas.usf.edu

Frankly, I am not at all pleased to find Kolomoki
Complicated Stamp [sic] associated with
"Mississippian" Pottery, temple mounds or not...So
in answer to your question, I am uncomfortable, but
the mounds themselves might be perfectly happy in
their surroundings.

Joseph Caldwell to Gordon Willey, letter, July 8, 1954,
National Anthropological Archives, Smithsonian

Kolomoki (9ER1), in the lower Chattahoochee Valley
of southwestern Georgia (Figure 1), is one of the most
impressive archaeological sites in the southeastern United
States. The site includes at least nine mounds, the largest of
which (Mound A) rises approximately 17 m from the broad
terrace to its west (Figure 2). The number and magnitude of
its earthworks make Kolomoki impressive, but these have also
been a source of confusion. William Sears, who conducted
intensive excavations at the site between 1948 and 1952,
correctly identified the dominant ceramics from Kolomoki as
local varieties of Woodland period types recently identified
by other researchers working in the region (Kelly 1938;
Willey 1945, 1949; Willey and Woodbury 1942). However,
distracted by the presence of the large flat-topped mound,
Sears (1950, 1951a, 1956) inverted and misplaced the ceramic
chronology to force the dominant occupation of the site into
the Mississippian period (ca. A.D. 1200), when agriculturally-
based chiefdoms constructed such earthworks throughout the
Southeast (Williams 1958). It is now widely accepted that
Kolomoki dates to the Middle and Late Woodland periods,
from around A.D. 350 to 750 (Jenkins 1978; Knight and
Mistovich 1984; Pluckhahn 2003; Schnell 1981; Sears 1992;
Trowell 1998). Sears's mistake has come to be referred to as
the "Kolomoki problem" (Trowell 1998).
Knight and Schnell (2004) have recently considered the
reaction to Sears's mistaken chronology. They identify nine
explicit critiques that were written between 1953 and 1961.
However, as Knight and Schnell (2004) also note, most of
these criticisms were never published, and the most common
reaction to the "Kolomoki problem" was silence, in accord
with the professional etiquette of the day (see also Knight and
Mistovich 1984:23).
While a polite silence may have prevailed among
the pages of the leading professional journals, personal
correspondence from the period indicates that there was a great

deal of discussion and debate among many of the principal
archaeologists of the day. Here, I draw together some of this
correspondence to provide additional insight on the "Kolomoki
problem." In particular, I rely on letters between Sears and
James Griffin (his academic advisor at the University of
Michigan) housed at the University of Michigan Museum of
Anthropology (UMMA). I also quote extensively from letters
written by Sears, Antonio Waring, Gordon Willey (Figure 3),
Joseph Caldwell, Ripley Bullen, and Arthur Kelly, archived
at the Anthropology Division of the Florida Museum of
Natural History (ADFMNH) in Gainesville, the University of
Georgia Laboratory of Archaeology (UGALA) in Athens, the
Southeast Archeological Center of the National Park Service
(SEACNPS) in Tallahassee, and the National Anthropological
Archives of the Smithsonian Institution (NAASI) in Suitland,
Maryland. These documents provide new insights into
the historical development of the Kolomoki controversy,
specifically the factors that led Sears astray, the rationale for
his errant chronology, and the reaction his mistake engendered
among his contemporaries.

"A Relatively Early Manifestation":
Background to the Kolomoki Problem

Perhaps the earliest statement concerning the temporal
placement of Kolomoki came from archaeologist Arthur
R. Kelly in 1937, more than a decade before Sears began
his first excavation. Kelly wrote to Matthew W. Stirling,
Chief of Bureau of American Ethnology at the Smithsonian
Institution, to correct an earlier statement assigning the site to
the Mississippian period:

...we have now made a study of the Kolomoki
collections in Dr. Harrold's basement and the results
will necessitate a change in the textual reference to
Kolomoki. Specifically it was found that Kolomoki
in light of present knowledge is definitely Late Swift
Creek with a few specialized incised and punctate
sherds which Ford thinks look very much like Cole's
Creek horizon. There is no suggestion of the Bull
Creek focus (Columbus) as I had supposed from my
recollection of the collections. You must remember
that I had not seen the material for nearly two years
and that in the meantime Swift Creek exploration and
analysis have completely changed the picture. It is


VOL. 60(2-3)




0 250
s itr Meters

contour interval = 1 rn

Figure 1. The Kolomoki site.

still true that Lamar is a widespread focus or aspect in
the Southeast...But Kolomoki must be expunged from
that part of the record which has to do with Lamar
or the Bull Creek focus thereof [Arthur Kelly to
Matthew Stirling, letter, July 14, 1937, SEACNPS].'

Nevertheless, other prominent archaeologists held out for a
Mississippian association for Mound A, even if the main period
ofoccupation was assumed to date earlier. Jesse Jennings, while
in general agreement with Kelley's characterization, offered
this opinion in a brief report outlining the archaeological
significance of Kolomoki:

On the basis of surface collections, from which
survey studies are made, Kolomoki would see to be
a relatively early manifestation. From our present
knowledge, the Swift Creek pottery type, which at
Kolomoki comprises 90% of the surface wares,
is one of the earlier cultures of the Southeast. It is
considered a Woodland culture...It is obvious that
Kolomoki reflects the influence of both basic cultures
[i.e., Woodland and Mississippian], because we

have here the truncated mound, while represented
by the surface collections is pottery with Woodland
affinities [Jennings 1940:1-2].2

Around the same time, Charles Fairbanks summarized the
analysis of sherds from surface collections and test excavations
at Kolomoki. He clearly identified them as Swift Creek and
Weeden Island, but drew a connection between the latter and
Mississippian cultures:

Kolomoki Mound Group would seem to have been
occupied first by Swift Creek...Later occupation by
Weeden Island complex is indicated together with
certain lower Mississippi Valley influences...Weeden
Island is indicated to be fully agricultural and may be
responsible for the large pyramidal mound [Fairbanks

Fairbanks developed this theory in greater detail in brief
reports summarizing his excavation of two small mounds at


2007 VOL. 60(2-3)


+' :f'J
-~ .4

Figure 2. View to the east of Mounds D (lower left fore-
ground) and A (center background) at Kolomoki.

At least three Indian groups have occupied the
Kolomoki Site. First, the Swift Creek group had
a very extensive village there. The large amounts
of Swift Creek Complicated Stamp indicates one
of the larger settlements of these people...Whether
or not any of the mounds can be ascribed to this
group is uncertain. Swift Creek mounds seem to be
of the domed residence type of other sites. None of
this kind are visible at Kolomoki at present...It is
probable that the large mounds are not of Swift Creek
construction...Weeden Island seems to represent the
result of a migration of culture from the Lower Valley
into Florida. This later came in contact with a Middle
Mississippi tradition, the fusion being Weeden Island.
This culture seems to have spread up into Georgia
an uncertain distance. It is obviously represented at
Kolomoki by typical sherds...In Florida the Weeden
Island sites, in general, do not have large temple
mounds. Some large mound groups are known,
however, and the large mounds at Kolomoki may
belong to this level...A very few sherds of Lamar
types were found in the surface collections from the
village. It does not seem possible that the mounds
can be assigned to Lamar on the basis of such small
collections [Fairbanks 1941a:3-4].

Figure 3. An informal conference at Kolomoki in 1949.
From right: William Sears, Gordon Willey, Antonio War-
ing, unidentified male (possibly Henry Brett), and Molly
Allee. Reproduced courtesy of the University of Georgia
Laboratory of Archaeology.

A letter written to Griffin the following year indicates that
Fairbanks's reservations concerning the possible association of
Swift Creek pottery with a large platform mound had grown:

Kolomoki is a mess from the stratigraphic point.
There is a large Swift Creek occupation...There is
a large amount of Weeden Island that may be later
than Swift Creek or combined with it as on the N.W.
Florida Coast. Lamar is very thin and does not seem
to account for the large mound. The smaller mounds
are stone and dirt conicals and could be Swift Creek
or Weeden Island. The large one I think should be
Lamar. Waring and Holder say it could be Swift
Creek but that is not very reasonable...Late Weeden
island in its northern range could run to a good deal
of Middle Mississippi traits [Charles Fairbanks to
James Griffin, letter, February 5, 1942, UMMA].

As Fairbanks's appraisals suggest, a complicating factor in
these early attempts to situate Kolomoki in the prehistory
of the Southeast was the commonly held assumption that
Weeden Island was contemporaneous with fully developed
Mississippian cultures in the interior. Indeed, as Knight
and Schnell (2004:3-4) have pointed out, in the 1940s the
Woodland and Mississippian sequence for the Gulf Coast was
compressed into the interval after A.D. 500: Fort Walton was
posited to have been essentially protohistoric in age (ca. A.D.
1500 to 1650), late Weeden Island (or Weeden Island II) was
assumed to date as late as A.D. 1500, early Weeden Island
(Weeden Island I) was thought to post-date A.D. 1000, and
Swift Creek culture was assumed to have a long duration that
began around A.D. 500 and extended as late as A.D. 1250 or
Thus, before Sears began his fieldwork at Kolomoki,
the basic components at the site and their order had been




accurately surmised by Kelly, Jennings, and Fairbanks, but the
precise temporal positioning of the major occupations was still
very unclear. All three archaeologists recognized that the first
major occupation at Kolomoki came during the Woodland
period, and was marked by Swift Creek Complicated Stamped
pottery. However, both Jennings and Fairbanks believed
that this occupation predated most of the mounds at the
site, and certainly Mound A. They suggested that most of
the mounds were constructed instead during a subsequent
occupation, marked by Weeden Island pottery and perhaps
dating to the Mississippian period. The limited amount of
Late Mississippian, Lamar pottery was correctly thought to be
associated with an occupation that postdated the major mound

"The Site Really is Something": The Problem Begins

Sears began work at Kolomoki in 1948. An early letter to
Griffin reveals his excitement at the prospects:

The site really is something. The mounds alone are
impressive, but there is a lot of village site material,
as I am sure you know. Ground surface on one low
rise a few hundred yards S.W. of the main mound
is littered with the Swift Creek and Weeden Island
material. There are a number of smaller mounds,
including several which do not appear in the maps
and descriptions. One low one, which may be Swift
Creek Burial mound, or may have something to do
with the big mound, is close enough to the proposed
road so that I'll be getting into it in the next week or
so [William Sears to James Griffin, letter, June 27,
1948, UMMA].3

This letter also reveals what may have been the seed of the
"Kolomoki problem": while Weeden Island sherds were
plentiful at Kolomoki, they were matched or even exceeded
by those of the Swift Creek type. Further, contrary to the
assumptions of Fairbanks and Jennings, there appeared to be
an association between Swift Creek pottery and at least one
of the mounds (and perhaps even Mound A). The unstated
implication would be that the mound construction (and
specifically platform mound construction) predated Weeden
Island, and thus also the Mississippian period.
Within the last twenty years, the existence of pre-
Mississippian platform mound construction has become
widely accepted (e.g., Jefferies 1994; Knight 1990). In the
1940s, however, the idea that "temple" mounds preceded the
Temple Mound period would have been considered heretical.4
Sears's dissertation (1950:52-53) provides a glimpse of the
importance of the settlement plan at Kolomoki to his (and
probably also Griffin's) conviction that the site must date to
the Mississippian period:

...I stated that the Temple Mound-Plaza-Village
Complex at Kolomoki was a Middle Mississippi
feature. The cultural relationships and relative
chronological position within Middle Mississippi

of such arrangements as this have been treated by
Ford and Willey, by Phillips, and in the report of the
Central Mississippi Valley Archaeological Survey, so
that they need not be treated at length here. It seems
quite certain that even if temple mounds do appear
with such abstract entities as Troyville, an idea
which still needs more convincing demonstration,
such arrangements as that at Kolomoki of the temple
mound, the plaza and a heavily occupied village area
are not found in early Mississippi times or Temple
Mound I but only with later Mississippi or Temple
Mound II. As pointed out earlier, this indicates that
one of the cultures at Kolomoki...must have been in
contact with peoples of Middle Mississippi culture.
The village plan at Kolomoki indicates that such
contact took place at the late end of our continuum,
during that period when the occupants of the site
were using pottery of the Kolomoki complex.5

If, as Sears was apparently beginning to surmise, the
complicated stamped ceramic component at Kolomoki was
the most substantial, and also the one responsible for much
of the mound construction, how could this be reconciled with
a Mississippian date for the mounds? We can imagine that
to Sears's thinking, one possibility was that the complicated
stamped pottery at Kolomoki included a later variety that
developed from late Swift Creek, and which could thus date to
the Mississippian period (at least according to the compressed
chronology of the day).
Sears was already beginning to entertain the notion of
a late (post Swift Creek and Weeden Island) complicated
stamped horizon at Kolomoki by the fall of 1948. He looked
to similarities between the pottery at Kolomoki and better
established Mississippian types like Savannah Complicated
Stamped and Coles Creek Incised to bolster his case:

As I told you before, there are three occupations at
the site. The Lamar like material seems finally to
have been only a small campsite. There are about
1500 sherds however, including a check stamp, quite
common, definitely associated with the Lamar rim.
Weeden Island is represented in pure form
also. Good deal of material from the plow zone,
and two pits with 2-3,000 sherds between them.
All of the Florida types of the early period, and a
few complicated stamped sherds, including a fair
proportion of Napier.
The main occupation however is even more
interesting. Besides scattered material from the
plowzone and a few small refuse pits, I have large
samples from three midden deposits. One of them,
which may show some percentage changes from
bottom to top, yielded 8,000 sherds, the other two a
few thousand apiece. Most important form is the jar,
sand tempered, complicated stamps in zones. Rim is
usually plain, rarely shows any signs of folding, may
flare slightly, a [sic] quite often has a small regular,
extrusion of paste at the very top edge (outer)...

2007 VOL. 60(2-3)


Commonest stamped design is the multi-line scroll,
as Savannah. There are also Savannah-like figure
8's. Some bullseyes appear, as well as bullseyes set
in diamonds. Swift Creek is still strong however,
probably predominant in overall effect.
I don't think Kelly agrees with me yet, and I am
quite sure Tono [Antonio Waring] won't, but it looks
to me like a case of Late Swift Creek participation
in the early stages of Temple Mound I growth. Pot
bases sound like Coles Creek, Stamps have a decided
Savannah element and there is more than a trace of
Weeden Island, particularly in flattened bowl forms,
sort of seed bowls. That queer rim brings the Weeden
Island rims to mind too [William Sears to James
Griffin, letter, October 17, 1948, UMMA].

Griffin apparently consulted with Fairbanks regarding
Sears's initial observations regarding the major occupations
at Kolomoki, finding agreement that the Late Swift Creek
occupation at Kolomoki was the most substantial. Griffin
also encouraged Sears's comparisons to better-established
Mississippian pottery types from the Mississippi Valley:

Fairbanks says that he thinks also that Kolomoki
was probably occupied and built during Late Swift
Creek. I think you will find the same types of bases
in Weeden Island as in Coles Creek [James Griffin to
William Sears, letter, October 21, 1948, UMMA].

As these letters indicate, Sears initially focused on
perceived Mississippian attributes in the rim forms, stamping,
and bases he was finding at Kolomoki. He would later use these
same attributes to formally define a new pottery type, Kolomoki
Complicated Stamped (Sears 1950:28-35, 1951a:9-16). In his
definition of this type, Sears (1951a:15-16) described it as
a development from late Swift Creek.6 However, consistent
with his comments to Griffin, he argued that the flat-bottomed
bases were more akin to Coles Creek than Swift Creek or
Weeden Island, thus finding justification for its placement in
the Mississippian period (Sears 195 la:14).
A potential flaw in the hypothesis that the principal
occupation at Kolomoki developed late in the Swift Creek/
Weeden Island sequence was the almost complete absence
of Wakulla Check Stamped, which Willey's investigations
along the Gulf Coast had established as a good marker of
the Weeden Island II period (Willey 1945, 1949; Willey and
Woodbury 1942). Kelly's (1950, 1960) work along the Flint
and Chattahoochee Rivers to the south of Kolomoki was
producing similar results. Sears commented on this problem in
a letter to Griffin before the start of his second field season:

I have been in contact with Willey right along. He
typed a batch of Weeden Island for me last fall,
and has been very co-operative ever since. Haven't
written Goggin yet, but I will....
I am still not so sure about check stamp in
Weeden Island I, largely because I have a good many
thousand Weeden Island sherds from Kolomoki,

including several large batches from pits where the
only non-Weeden Island materials were complicated
stamped sherds of non-Kolomoki types. There is not
a single check stamped sherd however. This is in
spite of the fact that Moore dug mounds down river
a bit and found Kolomoki types with Weeden Island
and Check stamp.
Kelley's [sic] sites which have only Wakulla
check, seem to represent some sort of intrusion.
Weeden Island mounds in the area have a fair
percentage of Late Swift Creek sherds, although
not of the Kolomoki types. As soon as the check
stamp shows up, it seems to replace the Swift Creek
completely [William Sears to James Griffin, letter,
February 5, 1949, UMMA].

In the summer of 1949, Sears began excavation of
Mound E. He recognized that the pottery from the mound was
consistent with Willey's Weeden Island I period. However, he
was also clearly concerned with validating his hypothesis of a
later complicated stamped horizon (his Kolomoki Complicated
Stamped), as is evident in a letter to Griffin:

Dig here at Kolomoki is turning out fine. Mound, 10-
12 ft high and 80 across, has turned out to be a Weeden
Island I burial mound, which is what I'd hoped...Some
complicated stamped jobs showing up. Not the stamp
which appeared with the Weeden Island stuff in the
village refuse deposits, nor Kolomoki complicated,
although closer to the latter. Sort of thing it could
have developed from, and probably did if my ideas
on sequence in the area are correct [William Sears to
James Griffin, letter, June 15, 1949, UMMA].

Griffin responded a few days later with some puzzlement:

I am glad to know that the burial mound at Kolomoki
has turned out to be so productive. You ought to make
a pretty good killing with this excavation of a Weeden
Island I Mound. I am not sure whether you mean the
complicated stamped in the mound developed from
Kolomoki Complicated or vice versa [James Griffin
to William Sears, letter, June 20, 1949, UMMA].

Sears wrote back to clarify his position that the complicated
stamped pottery in Mound E was ancestral to his newly defined
Kolomoki type:

The stamped ware presumably is ancestral to
Kolomoki. Same vessel forms, with some change
in the percentage make up of the complex perhaps,
same rims, and the same size and complexity "feel"
to the stamps. However, specific stamp elements
are different. Fair number of snow shoes, e.g., but
as parts of more complex elements, never as the
sole motif and so far never as fills in figure eights.
In any case, if my dope last year on who followed
which [sic] is correct, has to be earlier. Can't let both




Weeden Island I and Kolomoki groups onto the site
at the same time...
ARK [A.R. Kelly] of course isn't too happy
about any of this. He can see that the stuff is different
when I pile a batch of each in front of him, and he can
see that temporal positions must be different. But all
I hear is "We'll have to justify these fine divisions,
Bill." Of course we will, no damned point making
them without [sic] reasons.
I feel, but can't prove yet, that there are two main
streams of complicated stamped wares in S. Georgia,
not all 1 [sic] coming from Macon. The trade ware
in the Weeden Island village, the burial, Kellys
[sic] stuff at Bainbridge, and other sherds I've seen
represent one. The other is represented by Kolomoki,
this only slightly different than the present mound,
and Deptford complicated stamp. Macon seems to
have some of both, the Kolomoki-Deptford hitting it
late and developing what may be Late Swift Creek
if ARK will ever describe it. Or maybe I'm nuts.
The only thing I am sure of is that there are a lot of
complicated stamped potsherds, which cover a long
time and space range, and that pure samples vary...
Don't expect to see Tono [Antonio Waring] till
late Aug., but Willey should pull into sight next week.
I'll have somebody to fight with then [William Sears
to James Griffin, letter, June 22, 1949, UMMA].

As this letter indicates, by 1949 Sears was convinced that
there was a "Kolomoki period" occupation at the site that was
both distinct from, and later than, the Weeden Island horizon.
Despite Sears's efforts to clarify the situation, however, Griffin
continued to express some befuddlement at Sears emerging
chronology more than a year later:

Lewis [Larson] has been in, but he does not seem to
know too much about the details of the interpretation
of the digging and I must confess that in some ways
I am in the same kind of a fog that Lew is about the
detailed interpretation of the Kolomoki material and
the way it fits into the Florida West Coast sequence
[James Griffin to William Sears, letter, September
22, 1950, UMMA].

Soon, more archaeologists would express similar confusion.

"I Am Glad That I Do Not Have to Worry about That":
Initial Reactions

Sears's dissertation, the first printed account of his still
ongoing work at Kolomoki, was completed in 1950. It appears
to have engendered considerable discussion among his
colleagues. Waring, who worked with Sears at Kolomoki on
occasion, must have expressed his reservations about Sears's
chronology to Charles Fairbanks. In his response, Fairbanks
also expressed concerns, but was receptive to the possibility
that the main occupation at Kolomoki dated late in the Weeden
Island sequence, and thus could be contemporaneous with

Mississippian cultures. The presence of a red painted ware,
referred to by Sears (1950:42-44) in his dissertation as Mercier
Red on Buff, seems to have been particularly persuasive in this

If your profile of the Kolomoki situation is right
then Sears would appear to be a bit off. However,
his pottery from last year does show Southeastern
Woodland and Early Middle Mississippi mixed. I
can't see Kolomoki as a local response in view of the
red paint and Mississippian vessel shapes. Griffin has
found somewhat similar stuff at Cahokia, plus simple
cult stuff at what he says is an early level. [Charles
Fairbanks to Antonio Waring, letter, September 15,
1950, NAASI].

Fairbanks was not alone in lending credence to the possibility
that Sears's chronology was correct, and that a major revision
to the Gulf Coast sequence might be called for. Willey himself
admitted his openness to the possibility in a letter to Waring:

Heard from Sears afew days ago. I am still unconvinced
(but willing to change) on his interpretations. I am
anxious to examine the evidence in detail. I do not
hedge on some sort of Mississippian contact (one
way or the other) with Weeden Island I; but I do not
think there is any chronological overlap that amounts
to anything between Weeden Island and Fort Walton.
I am not yet clear on his dating of Kolomoki Stamped,
Late Swift Creek, and Weeden Island. As I understand
it, he tends to see these as three separate entities,
probably separate in time [Gordon Willey to Antonio
Waring, letter, September 8, 1950, NAASI].

Griffin (1984:136) later recalled some of the confusion that
Sears's dissertation created:

...A major problem of his thesis...was whether the
burial activities and associated material belonged to
the earlier or later portions of Weeden Island. While
recognizing early in the excavation that there was
a strong relationship to earlier Hopewellian burial
procedures, a number of features of the ceramic
complex seemed to have a significant relationship
to Mississippian vessels. Gordon Willey read the
thesis and regarded Kolomoki as belonging to the
early Weeden Island period with the ceramic features
suggestive of Mississippi styles being prototypes
instead of isochronological as, alas, Sears and his
thesis advisor tended to think.

Griffin's recollections indicate that Willey believed that the
Mercier Red on Buff type could have been a predecessor
to vessels of similar decoration that appeared later, at
Mississippian sites such as Etowah and Cahokia. Sears, on the
other hand, argued against this interpretation:

2007 VOL. 60(2-3)


By now Gordon's copy of his rebuttal on my
dissertation will have reached you. Some justice in a
lot of it, but I sure do bog down on his development
of Mercier Red on Buff and use of it as a donor
type to Mid Mississippi. That is carrying things too
far. Vessel shape, painting style, and several very
specialized rim forms developing out of a zoned
style at one site out of hundreds, and then acting as
a donor to thousands of groups over a wide area is
just too much to ask methinks. It certainly reverses
everyone's theories of diffusion from Boas and
Wissler to Leslie White.
Another red and white vessel has turned up too.
Large white areas apparently, red stripes, curvilinear,
plus large incised triangles pendant from the rim.
Haven't had time too [sic] do much with it, as is a
very late jar in many sherds, but appears to be the
same late Mid Miss vessel form as the Mercier Red
on Buff vessels, including the flattened lip [William
Sears to James Griffin, letter, December 8, 1950,

By 1950, Griffin was growing increasingly concerned
about the disagreement between Sears's chronology and that
of Goggin and Willey:

...I hope that you and Goggin and some of the others
can come to some agreement as to what kind of
material you are getting there. It certainly doesn't
make sense to have some people regard it as Weeden
Island I, and others as late Weeden Island (James
Griffin to William Sears, letter, November 21, 1950,
...Yes, I did get Gordon's comments on your
thesis and also his reply to your letter. I am not
particularly concerned with differences of opinion as
to the interpretation of the material, except that there is
certainly going to be a revolution if some of Gordon's
Weeden Island I is actually in your collection of
Kolomoki material, which seems to date from quite a
bit later. This means that perhaps a considerable part
of Gordon's classification of the Florida Gulf Coast
material will need a lot of redoing. I am glad that I
do not have to worry about that [James Griffin to
William Sears, letter, December 15, 1950, UMMA].

Although he expressed such reservations to Sears, Griffin
defended his student to other archaeologists, as indicated in
the following letter to Waring:

The materials that Sears excavated...look to me to fit
in pretty well with the Weeden Island sites. I went
through Willey's report with some degree of care and
went through Moore: there is no question in my mind
but that Willey's alignment of Weeden Island I and
Weeden Island II is not at all satisfactory, and that a
considerable amount of work will need to be done
in Weeden Island sites to straighten out the current

mess. I am convinced that Sears' [sic] interpretation
of the cultural alignment of the material at Kolomoki
and the related Weeden Island forms is correct,
irrespective of when this may be chronologically
[James Griffin to Antonio Waring, letter, January 13,
1950, NAASI].

Sears also found some validation from one of Griffin's other
students, Charles Fairbanks. In his review of Sears's first
two excavation reports, Fairbanks (1952) expressed tentative
appreciation for Sears's position:

....From the association of Weeden Island and
Kolomoki pottery types in all the mound pottery
deposits he concludes that the burial mound [Mound
E] belongs to the Kolomoki complex which is in turn
descended from the Weeden Island complex. The
Weeden Island pottery types present in the burial
mound would thus be regarded as the archaic hold-
overs reserved, and made, solely for funerary use
after they had gone out as domestic types...
Sears goes on to discuss his own and the Willey-
Goggin views of Weeden Island. Briefly, Willey and
Goggin hold that the Kolomoki Weeden Island is
Weeden Island I because Wakulla Check Stamp or a
comparable type is absent. Sears holds that the absence
of Wakulla Check Stamp does not of necessity mean
that his Weeden Island component is early...He cites
the presence of the zoned variety of Weeden Island
Incised as an evidence of its placement in Weeden I
but holds that the cursive, free-hand incising in the
burial mound indicates a Weeden Island II time for
that component and thus for the associated Kolomoki
complex. He adds the undoubted mature Middle
Mississippian elements on Mercier Red on Buff as
a clincher. Willey and Goggin have presented the
evidence that Wakulla Check Stamp is the marker
for Weeden Island II in the northwest Florida area.
It would take more evidence than here presented to
invalidate their assertions. However, it does seem
that Sears marshals both significant facts and valid
assumptions to bolster his revision.

Perhaps bolstered by the tacit approval ofFairbanks and Griffin,
Sears held firm to his convictions that the Kolomoki period
occupation post-dated Weeden Island. The fact that Wakulla
Check Stamped was not well-represented at Kolomoki clearly
indicated to him that Willey's chronology-rather than his
own-was in error:

Have finished most of the work in the Kolomoki
village units. No doubt at all that we have an unbroken
continuum extending from the classic Weeden Island
up to the latest. Mid-Troyville to Early Plaquemine....
Apparently, the whole period from definitely Weeden
Island to definitely Kolomoki is about equivalent to
Coles Creek and definite Early Mississippi. We quit at
the Grant-Gahagan-Alto level, with the cult probably




showing up in the next generation...Think that ties
in reasonably well with your Central Valley data on
such stuff, although it will not please Goggin.
Have no doubt at all that Gordons [sic] seriation,
through trusting the check stamp and not separating
mound and village asemblages [sic], loses the entire
climax period for at least the central part of the Fla.
N.W. Coast, a 300 year long segment. Pure Stamped
middens, basically he threw back at the bottom of his
seriation charts, as Mound Field. All of which shows
what happens to a trusting young man who puts his
faith in the nasty check stamp and very few strata
cuts [William Sears to James Griffin, letter, January
29, 1953, UMMA].

"The Sequence by the Wrong End": The Reaction Grows

While the issue may have been settled in Sears mind, his
Kolomoki chronology began to unsettle more of his colleagues.
In 1952, Joseph Caldwell began excavations of stratified
deposits at the Fairchild's Landing and Hare's Landing sites
to the south of Kolomoki. This work would clearly show
Sears's chronology was in error, and in turn provide additional
validation for Willey's sequence. Although a complete report
of these excavations would not be published for more than
thirty years (Caldwell 1978), the results began to disseminate
informally by 1954, as indicated in this letter from Kelly to

Joe Caldwell is completing his write up on Fairchilds
[sic] Landing. Shows a good Swift Creek at base,
probably just a little later than G. Willey's Santa Rosa-
Swift Creek intergrading into an early Weeden Island
I, then a Weeden Island II with Wakulla check-stamp
playing pretty much the role that Willey assigned to
it! "Kolomoki Stamped" begins in the definite Swift
Creek shell layers in the bottom of the mound, and
plays out in the upper layers. Over 50,000 sherds
in good sealed context. [Arthur Kelly to Antonio
Waring, letter, April 17, 1954, NAASI]

In a letter the following year, Kelly reiterated these

J. Caldwell's pottery design studies on Fairchild's
Landing and Hare's Landing, lower Chattahoochee,
give a complete detailed style evolution of the
Kolomoki stamping from a mid-Swift Creek point
of departure thru a Kolomoki horizon, thru [sic]
Weeden Island, into a new plain pottery-check
stamped phase which Joe and Gordon Willey
are calling Cummings Period-into a final check
stamped horizon just precedent to Fort Walton.
The amazing thing, according to Joe, is that
Sears correctly described the actual stylistic and
morphological changes in the development of
the pottery, in Kolomoki context, and in Weeden
Island, but just exactly reversed each sequential

development from the true one! Apparently, the
only thing Sears got right at Kolomoki was the
chronological position of the attenuated Lamar
material at the site. Mound D and Mound E are
perfectly good Weeden Island Mounds, and the
premound is Kolomoki of the kind Joe has beautifully
sandwiched in good shell stratigraphy in two good
sites down the river from Kolomoki [Arthur Kelly
to Antonio Waring, letter, March 7, 1955, NAASI].

Willey weighed in on the matter in two letters to Waring in
July of 1954:

Recently, I have read the able (if somewhat over-
impassioned) final Kolomoki report ofoperative Sears.
I was just about convinced of Bill's reorganization of
sequences down there, but Joe seems to have found
evidence to the contrary [Gordon Willey to Antonio
Waring, letter, July 6, 1954, NAASI].

The more I think about it, particularly after what you
have written, it looks as though Sears had the sequence
by the wrong end. I was never too convinced of the
way I had it lined up, and he had just about swayed
me over; but it looks as though Joe [Caldwell] has the
real dope [Gordon Willey to Antonio Waring, letter,
July 15, 1954, NAASI].

Nevertheless, the association of elaborate vessel forms and
"temple" mounds with Swift Creek and early Weeden Island
pottery was troublesome to both Caldwell and Willey, as
another correspondence from the same time suggests:

Frankly, I am not at all pleased to find Kolomoki
Complicated Stamp associated with "Mississippian"
Pottery, temple mounds or not. But I don't know when
the first Mississippian came in. I am not as sure as Sears
as to the direction some ceramic features might have
traveled. Your Florida cultures of Sta. Rosa times on
are in no sense backword [sic] and might have even
traded ceramic features on an even-steven basis with
whatever was becoming Mississippian elsewhere.
So in answer to your question, I am uncomfortable,
but the mounds themselves might be perfectly happy
in their surroundings [Joseph Caldwell to Gordon
Willey, letter, July 8, 1954, NAASI].

Likewise, a letter from Ripley Bullen to Caldwell indicates
that the development of Swift Creek and the chronological
placement of Kolomoki continued to be a topic of debate:

Seems to me you can't have the burial mounds
at Kolomoki Weeden Island I as long as they
have Mississippi traits in them (assuming that the
Mississippi complex is o.k. and relatively late).
Reason, these influences could only have operated on
late Weeden Island people...


2007 VOL. 60(2-3)


The Kolomoki midden below Mound D plainly
shows, as Sears wrote, the mound could be as old
as or younger than the midden (not older than). A
situation where the burial mound pottery differs from
that in the village is fairly common here along the
Florida Gulf Coast. The lack of Kolomoki pots in the
mound is a negative trait.
The zoned Kolomoki pot looks like a late or
Weeden IslandII thingto me. Hence, contemporaneous
but different burial furniture does not seem too odd
[Ripley Bullen to Joseph Caldwell, letter, April 15,
1954, UGALA].

Still, by 1955 the weight of professional opinion had
clearly turned against Sears. Fairbanks, who had given his
tacit approval to Sears in a review three years earlier, was far
more critical in his review of Sears's next excavation report
for American Antiquity:

...The underlying midden below Mound D yielded
almost solely Kolomoki Complicated Stamped, yet
the mass pottery deposit comprised about 92 percent
Weeden Island types. Thus we find the Weeden Island
types in use in the burial ritual after they had gone
out of use as a domestic ware. Sears again makes the
point that he believes the Weeden Island material
present is Weeden Island II in spite of the absence of
Wakulla Check Stamped. An alternative explanation
would be to place Kolomoki before Weeden Island
and to identify the Weeden Island as Weeden Island
I. The Kolomoki pots in the mass deposits would
then be survivals. In Sears' explanation the Weeden
Island pots are the survivals. The simpler explanation
would be to place Kolomoki before the Weeden
Island occupation...[Fairbanks 1955].

James Ford (1956) provided a similar, if slightly more
circumspect, review of Sears's first three reports the following
year for American Anthropologist.
The final report on the excavations at Kolomoki was
published in 1956 (Sears 1956). Two years later, the report
was reviewed by Stephen Williams (1958), who provided
the first-and for many years the only-extended critique of
Sears's mistaken chronology:

...The over-all picture as seen by Sears is that of 2
successive Weeden Island components, followed by
the Kolomoki culture, when all the elaborate burials
were made. Sears sees connections of this Kolomoki
culture with other Temple Mound cultures of the
Southeast...The Lamar component is seen as a later
reoccupation of the site...
In making this alignment Sears is in direct conflict
with the views of Gordon R. Willey...Sears sees the
stratigraphy at Mound Field...to be a case of inverted
stratigraphy, and thinks that Wakulla Check Stamp
can not be used in this area as a good temporal

indicator although he admits its utility in other parts
of Florida...
The suggestion that Sears's basic chart may
be inverted is a view which Sears anticipates...But
the problem must be settled on its own merits, and
the data appear to be at hand for such a solution.
The closest approach to real stratigraphy at the site
was in Mound D, where an almost pure midden of
complicated stamped ceramics...was found under a
burial mound which produced over 40 vessels of the
Weeden Island series...
What is the interpretation available from the
comparative data in the Florida area? This mound
would be dated as late Weeden Island I...with the
midden below indicating a late Santa Rosa-Swift
Creek-Early Weeden Island I period...Thus, rather
than Willey having inverted stratigraphy at Mound
Field, it appears to this reviewer that Sears has
unfortunately reversed the Kolomoki sequence...
The confusion arising from an equation of
Temple Mound to Mississippian is obvious...To
think of Kolomoki as "Mature Mississippi" because
of its temple mound was the false step that led Sears's
chronology astray...

Even before the review saw print, Sears had complained to the
editor of American Antiquity, Raymond H. Thompson:

If I may I would like to make a comment, perhaps
unnecessary on the review of my final Kolomoki
report. As you may know, Steve Williams sent me his
first rough draft for comment, in the hopes that this
would eliminate any necessity for letters to the editor,
etc. While I naturally disagreed with some points, I
will happily accept a review by Steve or anyone else,
which does the job I consider essential to a review.
After talking to Steve at Madison and telling him
this again I discovered that two things may happen.
One, he will be carrying even further, a tendency
I disapprove of, to use the review to air his own
views, reconstructions, and so on, of the history of
the area. Two, he has been justifiably corresponding
with John Goggin on this and they may now submit
a joint review. I will not complain about this nor
ask for rebuttal space if it is a review, however, it
is possible that this will turn out to be their version
of the pre-history of the area using a fair amount of
new data which was not available to me [William
Sears to Raymond H. Thompson, letter, May 8, 1957,

Although there are hints that Sears may have realized
his error by this point (e.g., the reference to "new data" in
the above letter), he was not yet ready to admit this to his
colleagues. He continued to defend his temporal placement
in letter to James Ford in 1959 (William Sears to James
Ford, letter, December 8, 1959, ADFMNH). In 1962, Sears
wrote to Olaf Prufer concerning the latter's article (Prufer




1962) drawing an association between meteoric iron and
the Hopewell culture. Sears claimed an exemption for iron
cymbals from Kolomoki:

I know of only one exception concerning the
association of meteoric iron with Hopewell and
this is the occurrence of a number of meteoric iron
earspools or cymbal-shaped ornaments in Mound D
at Kolomoki...Actually, I don't think that this effects
[sic] your thesis at all since in my opinion classic
Weeden Island culture and particularly mortuary
ceremonialism are direct lineal descendants of
Hopewell culture and ceremonialism [William Sears
to Olaf Prufer, letter, March 29, 1962, ADFMNH].

Prufer wrote back only two days later to ask some of the
questions that no doubt many other archaeologists were also
curious about:

Since 1953, have you had any new ideas regarding
the site's chronological position (specifically as it
relates to Mound D)? Any radiocarbon dates? Did
Mandeville affect your thinking on Kolomoki at all?
If so, in which way? If these questions sound naive,
the fault is entirely mine. I am not familiar with
Kolomoki beyond the broadest generalities. What I
am after, of course, is to see whether by any chance
your Mound D can be brought in any chronological
proximity of our Hopewell [Olaf Prufer to William
Sears, letter, March 31, 1962, ADFMNH].

I have found no record of a response from Sears.
No doubt much of Sears's inability to admit his mistake
was due to personal conflicts with Kelly and Joseph Caldwell.
The fact that it was Caldwell's work which effectively
disproved the errant Kolomoki chronology must have been a
bitter pill for Sears. Correspondence suggests that things were
not always so rocky between these principals. In 1948, Sears
wrote Griffin:

I think Kelley [sic] and I will get along fine. He seems
to be O.K. in all respects. Met his wife and daughters
too, who are also nice people [William Sears to James
Griffin, letter, June 27, 1948, UMMA].

Within a few years, however, the tensions and professional
jealousies ran high. In 1951, Sears wrote to Griffin decrying
the field methods employed by Kelly and Caldwell:

Wouldn't mind too much except that I saw a couple
of his [Caldwell's] digs in Buford, and his field
technique stinks, sloppy as can be. Interpretation
almost as bad. If Kelly or Carl Miller are worst, he
is running for third. [William Sears to James Griffin,
letter, July 21, 1951, UMMA].

Sears's animosity also extended to Sheila Caldwell
(wife of Joseph Caldwell and daughter of A.R. Kelly), who

he attempted to remove from her position as director of the
excavations at the site of Fort King George, on the Georgia

The clan here is being mighty secretive about Sheila.
Gather that they suspect me. However, she is still
digging, so must be tearing up developmental and
interpretive material now, out of the survey stage...
Sheila is an archaeologist now. If your friend
Lawrence can do anything, he had better get at it. If
we don't get this straightened out now, she will be
permanently intrenched [sic] in Georgia Archaeology.
At that point, I'll go to Maine if I have to. Enough is
enough. [William Sears to Antonio Waring, letter,
September 17, 1952, NAASI].7

The conflict between Sears and Kelly reached a climax of
sorts in 1953. Sears's position at the University of Georgia-
always tenuous-was coming to an end. Correspondence
suggests that he was wrangling for a more secure appointment
by deriding Kelly's abilities as Chair of the Anthropology
Department. Sears had apparently gone above Kelly to
complain about the workings of the department to Walter
Martin, Dean of the College of Arts and Sciences, as well as
several other members of the university administration (Arthur
Kelly to Antonio Waring, letter, October 21, 1953, NAASI;
William Sears to James Griffin, letter, March 10, 1953,
UMMA). Despite the apparent subterfuge, Kelly defended
Sears. Informed by the Dean of Sears's machinations, Kelly
nevertheless wrote a letter of support, trying to explain the
personal stress (principally, the pressure of a growing family
and lack of a tenured position) that had made Sears "haggard,
occasionally overwrought, grouchy, stonily aloof' (Arthur
Kelly to Dean S.W. Martin, letter, April 7, 1953, UGALA).
Kelly described Sears as one of the top archaeologists of his
generation, and one that would be difficult to replace. "I like
Bill Sears" he wrote to Dean Martin.8
Both Kelly and Sears enlisted their friends and colleagues
to choose sides in the conflict. To their credit, most appear
to have remained neutral. Waring wrote Kelly to express his
regret over the disagreement:

I had rather hoped that you and Bill had simmered
down to an amicable agreement but such is apparently
not the case. Bill has many admirable qualities in spite
of that element in his personality that goes to make a
Marine, and I do feel that it will be a loss to Georgia
archaeology when he leaves the State, as leave he
probably must if the situation has come to the boil
you indicate [Antonio Waring to Arthur Kelly, letter,
October 23, 1953, NAASI].

Willey also wrote Kelly to express his hope that the conflict
could be diffused:

It is distressing to know that there has been conflict,
overt or covert, between you and Bill. He is a
very capable man and has done a lot for Georgia


2007 VOL. 60(2-3)


archaeological research. I hope you can see your
way clear to pull in double harness rather than go in
opposite directions [Gordon Willey to Arthur Kelly,
letter, September 24, 1953, UGALA].

In a letter to Waring, Willey succinctly summarizes the personal
issues that must have formed the core of the disagreement:

For over a year I have received letters from both
Kelly and Sears about the trouble down there
between them, and I have tried to play it down the
middle. From what you say it does appear as if Bill
had been a little too vigorous in pursuing his causes
with no regard for anything else. It is interesting
that Fairbanks fits into Bill's side. Griffin must
impart some peculiar indoctrination to the boys up
there...What archaeology needs are some reasonably
generous and tolerant practitioners whose neuroses
are more or less under control. Kelly is generous
but inept and hopelessly lazy; Sears is too severe,
intolerant and uncompromising [Gordon Willey to
Antonio Waring, letter, July 14, 1954, NAASI].

As Willey observed, in addition to the personality conflicts,
there also appears to have been an underlying tension between
archaeologists from the University of Michigan on the one
hand, and those from the University of Chicago and Harvard
University on the other. Consistent with Willey's observation,
in his correspondence with Sears, Griffin spoke derisively of
the "new archaeology at Chicago" (James Griffin to William
Sears, letter, April 6, 1953, UMMA). Conversely, Kelly
complained about the number of Michigan graduates in
Georgia and a perceived attempt by Griffin to dominate the
field of archaeology (Arthur Kelly to Antonio Waring, letter,
October 21, 1953, NAASI).9
By 1960, Sears had become increasingly embittered, as
evidenced by a letter he wrote to Charles Fairbanks concerning
a submission to Current Anthropology:

I am afraid I honestly don't know what you are
supposed to do about the Current Anthropology
article...I certainly wouldn't worry about any
disagreement you might have with the article,
including one based on Lamar. I will have some
opportunity for reply I suppose. I guess that I cannot
publish anything which everyone agrees with unless
it is simply a sherd count. I doubt that there would
be general agreement with even this [William Sears
to Charles Fairbanks, letter, December 19, 1960,

Whatever objections Fairbanks may have raised, the article
was published the following year with no mention ofKolomoki
(Sears 1961).
Likewise, Sears's correspondence from the 1960s and
later contain almost no references to Kolomoki. As Knight and
Schnell (2004:9) have noted, by this time, Sears's conclusions
had proven untenable for most students of southeastern

prehistory, and there was no longer any need for debate.
Sears continued to correspond with archaeologists working
in southern Georgia, including even Kelly (William Sears
to Arthur Kelly, letter, September 21, 1960, ADFMNH). In
comparison with the chattiness of Sears's earlier exchanges,
however, the language is terse and occasionally slightly
acerbic even when showing flashes of humor, as in this letter
to Harold Huscher:

...I can't imagine that it is hot in the Chattahoochee
Valley. It never got over 125 dg. when I worked
there. Perhaps the archaeologists are getting
weaker...[William Sears to Harold Huscher, letter,
August 18, 1961, ADFMNH].

"Mea Culpa": The Aftermath

Knight and Schnell (2004:9) have discussed the problems
that resulted from the lack of any clear, well-documented,
published correction of Sears's error. These include the
perpetuation of the mistaken chronology among both scholars
and the lay public for more than half a century, as well as a
diminishment of the importance of Kolomoki to archaeology.
The lessons to be learned from personal correspondence
related to the "Kolomoki problem" are perhaps more subtle,
but nevertheless important.
The perils of hubris are evident in these letters. There
can be little doubt that his personal conflicts with Kelly and
Caldwell made it more difficult for Sears to admit that his
chronology was in error. Even Sears's later mea culpa lays
much of the blame for his mistake with the failure of his
colleagues to publish their data contradicting his assumptions
(Sears 1992:66-67). However, as the letters cited above clearly
indicate, it is inconceivable that Sears was unfamiliar with the
new data proving his chronology was in error even before his
final report was published.
A related lesson that I have taken from reading Sears's
letters and attempting to reconstruct his thinking is the
importance of entertaining alternative interpretations and
engaging with contradictory evidence. The possibility of a
later complicated stamped horizon at Kolomoki was phrased
as a working hypothesis in Sears's first excavation report
(Sears 1951a:38-42) and a subsequent summary for American
Antiquity (Sears 1953:227). Yet, in his final report (Sears
1956), no alternatives were discussed and conflicting evidence
was either casually dismissed or ignored. Had Sears maintained
the more moderate tone of his earlier work, it would have been
easier to admit his error and save face.
Personal correspondence relating to the "Kolomoki
problem" also reveals the injustices of the culture of academia.
Throughout much of the 1950s, Sears was balancing a rapidly
growing family with a series of impermanent and poorly
paying positions. An honest admission of a fundamental flaw
in what was-at least at that juncture-Sears's life's work
would surely have impeded his struggle to land a more secure
academic position, just as it would for aspiring academicians



Sears's relationship with his advisor also bears scrutiny.
Although alternative readings of these letters are possible, my
own analysis suggests that Griffin strongly encouraged Sears's
thinking that Kolomoki must date to the Mississippian period
(see note 5), while at the same time cautioning him against
directly contradicting Willey's chronology. We can imagine
that Sears, as a former Marine, was anxious to please his
mentor. Given the circumstances, he had little choice but to
argue that the pottery at Kolomoki represented something new
and different that could more easily be accommodated into the
existing chronologies for the Southeast.
Finally, these letters serve as a reminder of the challenges
faced by an earlier generation of archaeologists, struggling
to construct cultural historical sequences in the absence of
dependable radiocarbon dating or-at least in most cases-
stratified deposits. Today, we take these sequences for granted,
but these letters reveal something of the personalities, sweat,
and debate that were a part of their creation.


1. Dr. C.C. Harrold was a prominent physician from Macon
who was instrumental in the acquisition of Kolomoki for a
park. For more details, see Trowell (1998:28-35).
Kelly was working at Macon Plateau at the time this letter
was written. Elsewhere in this letter, he describes his belief
that Swift creek culture was interrupted by the Mississippian
development there, reflecting the compressed chronology that
was accepted at the time (see Knight and Schnell 2004:3-4).

2. Jennings appraisal, as well as the early work by Fairbanks
and Wauchope, was intended to assess the suitability of
Kolomoki for a National Park.

3. It is not entirely clear which mound Sears is referring to
in this letter. Mound B is the closest small mound to both
Mound A and the present-day road across the site, and thus
would seem the most likely candidate. Sears may have
conducted excavations in this mound in 1948, although in his
report he appears to mistakenly describe it as Mound C (Sears

4. As Knight (1990:170) has demonstrated, however, most
of the pre-Mississippian platform mounds-including those at
Kolomoki-lack convincing evidence for summit structures.
They thus should probably not be considered "temple mounds"
per se.

5. Sears's citation of the Lower Mississippi Valley survey
(Phillips et al. 1951), which was still in preparation at the time,
is illuminating. The wording in this section of his dissertation
closely parallels that of some parts of the report (see Phillips
et al. 1951:441).
Phillips and colleagues note the presence of pre-
Mississippian platform mounds at several sites in the region.
However, a footnote (Phillips et al. 1951:441, note 35)
indicates that Griffin dissented on this point, preferring a later
date for these features. The influence of Griffin on Sears's

thinking regarding the temporal placement of Kolomoki can
only be guessed, but must have been significant.

6. Sears (1992:66) would later admit that many of the sherds
he classified as Kolomoki Complicated Stamped "...could just
as easily belong to any nondescript 'Swift Creek' category."
Today, the Kolomoki Complicated Stamped type designation
is infrequently employed by archaeologists working in the

7. Out of respect for the Caldwell's descendants, I have
omitted from the transcription of this letter a derogatory
comment regarding Sheila Caldwell and her pregnancy.
The occasional androcentrism and misogyny apparent in the
personal correspondence of some archaeologists of this era
would seem a potentially fruitful topic for critical analysis.

8. Later, Sears lobbied Robert L. Stephenson of the
Smithsonian's River Basin project to prevent Kelly from being
awarded a contract for more work in southwestern Georgia
(William Sears to Robert L. Stephenson, letter, October 7,
1958, ADFMNH).

9. Kelly received his Ph.D. from Harvard, while Joseph
Caldwell earned his from the University of Chicago. Sears
earned an M.A. at Chicago before transferring to Michigan
for his Ph.D.
In a letter to Waring, Kelly observed that Griffin was
favorably inclined to graduate students who had left Chicago
for Michigan. He went on to described these as "...a group
animated by something of the spirit of the Old Bolsheviki who
had come out of the Siberian wilderness" (Arthur Kelly to
Antonio Waring, letter, October 21, 1953, NAASI).


Although a number of individuals and organizations
facilitated this research, I accept sole responsibility for the
content of this article. Institutional support was provided by
the University of Georgia, the University of Oklahoma, the
University of South Florida, and the Georgia Department of
Natural Resources. I thank Chris Trowell for providing copies
of documents on file at the University of Michigan Museum of
Anthropology. At the Southeast Archeological Center, I was
assisted by Bridget Beers. Donna Ruhl and Jerald Milanich
expedited my research at the Florida Museum of Natural
History. In addition, Jeannie Skylar guided me in the use
of the collections at the National Anthropological Archives.
Jim Knight, Rochelle Marrinan, and Frank Schnell provided
comments that greatly improved the quality of the manuscript.
Finally, but perhaps most importantly, I thank William Sears
and his contemporaries, whose passion for southeastern
archaeology was strong enough to induce long, thoughtful


2007 VOL. 60(2-3)


References Cited

Caldwell, Joseph R. (edited by Betty A. Smith)
1978 Report of the Excavations at Fairchild's Landing and
Hare's Landing, Seminole County, Georgia. Report
prepared for the Nation Park Service, Purchase Order

Fairbanks, Charles H.
1940 Analysis of Sherd Collections form Kolomoki State
Park. Report on file, Southeast Archeological Center,
National Park Service, Tallahassee.

1941 Archaeological Site Survey of the Kolomoki Mound
Group. Report on file, Southeast Archeological
Center, National Park Service, Tallahassee.

1952 Review of Excavations at Kolomoki, Season I, 1948
and Excavations at Kolomoki, Season II, 1950 by
William H. Sears. American Antiquity 1:81-83.

1955 Review of Excavations at Kolomoki, Season III
and IV, Mound D, by William H. Sears. American
Antiquity 20(3):289-91.

Ford, James A.
1956 Review of Excavations at Kolomoki, Season I, 1948;
Excavations at Kolomoki, Season II, 1950; and
Excavations at Kolomoki, Season III and IV, Mound
D, by William H. Sears. American Anthropologist

Griffin, James B.
1984 William H. Sears, Southeastern Archaeologist: A
Tribute. Southeastern Archaeology 3(2):134-138.

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

Jenkins, Ned J.
1978 Prehistoric Chronology of the Lower Chattahoochee
Valley. Journal of Alabama Archaeology 24(2):73-

Jennings, Jesse D.
1938 The Archaeological Significance of Kolomoki.
Report on file, Southeast Archeological Center,
National Park Service, Tallahassee.

Kelly, A.R.
1938 A PreliminaryReport on Archaeological Explorations
at Macon, Georgia. Bureau of Ethnology Bulletin
119, Smithsonian Institution, Washington, DC.

Knight, Vernon J., Jr.
1990 Excavation of the Truncated Mound at the Walling
Site: Middle Woodland Culture and Copena in the
Tennessee Valley. Report of Investigations 56.
Division of Archaeology, Alabama State Museum of
Natural History, University of Alabama, Tuscaloosa.

Knight, Vernon J., Jr., and Tim S. Mistovich
1984 Walter F. George Lake: Archaeological Survey of
Fee Owned Lands, Alabama and Georgia. Report of
Investigations 42, Office of Archaeological Research,
the University of Alabama, Tuscaloosa.

Knight, Vernon James, Jr., and Frank T. Schnell
2004 Silence Over Kolomoki: A Curious Episode in the
History of Southeastern Archaeology. Southeastern
Archaeology 23(1):1-11.

Phillips, Phillip, James A. Ford, and James B. Griffin
1951 Archaeological Survey in the Lower Mississippi
Alluvial Valley, 1940-1947. Papers of the Peabody
Museum of American Archaeology and Ethnology,
Harvard University, Volume 25, Cambridge,

Pluckhahn, Thomas J.
2003 Kolomoki: Settlement, Ceremony, and Status in the
Deep South, A.D. 350 to 750. University of Alabama
Press, Tuscaloosa.

Prufer, Olaf H.
1962 Prehistoric Hopewell Meteorite Collecting: Further
Evidence. Ohio Journal of Science 62(6):314-16.

Schnell, Frank T.
1981 Late Prehistoric Ceramic Chronologies in the Lower
Chattahoochee Valley. Southeastern Archaeological
Conference Bulletin 24:21-23.

Sears, William H.
1950 The Prehistoric Cultural Position in the Southeast of
Kolomoki, Early County, Georgia. Ph.D. dissertation,
University of Michigan, Ann Arbor. University
Microfilms, Ann Arbor.

1951 a Excavations atKolomoki: SeasonI-1948. University
of Georgia Press, Athens.

1951b Excavations atKolomoki: SeasonII, 1950. University
of Georgia Press, Athens.

1953 Excavations at Kolomoki: Seasons III and IV, Mound
D. University of Georgia Press, Athens.

1956 Excavations at Kolomoki: Final Report. University
of Georgia Press, Athens.




1961 The Study of Social and Religious Systems in North
American Archaeology. Current Anthropology

1992 Mea Culpa. Southeastern Archaeology 11(1):66-71.
Trowell, Christopher T.

1998 A Kolomoki Chronicle: The History of a Plantation,
a State Park, and the Archaeological Search for
Kolomoki's Prehistory. Early Georgia 26(1):12-81.

Willey, Gordon R.
1945 The Weeden Island Culture: A Preliminary Definition.
American Antiquity 10(3):225-254.

1949 Archeology of the Florida Gulf Coast. Smithsonian
Miscellaneous Collections 113, Smithsonian
Institution, Washington, DC.

Willey, Gordon R., and Richard B. Woodbury
1942 A Chronological Outline for the Northwest Florida
Coast. American Antiquity 7:232-254.

Williams, Stephen
1958 Review of Excavations at Kolomoki, Final Report,
by William H. Sears. American Antiquity 23(3):321-


2007 VOL. 60(2-3)



Department ofAnthropology, University ofFlorida, Gainesville, Florida 32611
Email: endonino@ufl.edu

Determining the source of lithic material recovered from
archaeological contexts has the potential to provide a number
of insights into the lives of prehistoric peoples. Sometimes
data derived from lithic tools and waste flakes are all that
archaeologists have to work with, especially at sites dating
to preceramic times and in areas where organic preservation
is poor due to acidic soil conditions. Lithic sourcing data
can also be used toward similar ends among more socially
complex societies and has been used to investigate the
political economy of Mississippian hoe production (Cobb
2000). Other questions that source attribution allows us to
address are the movement (or lack thereof) of groups across
the landscape (Austin 1996, 1997; Binford 1979; Daniel 1998,
2001; Gramley 1980; Sassaman et al. 1988), the organization
of technology (Andrefsky 1994; Cobb 2000; Daniel 1998), as
well as exchange and interaction between groups (Carr and
Steward 2004; Endonino 2003; Odess 1998).
For sourcing efforts to be successful it is vital to know the
lithicc landscape," the distribution of available chert resources
within the natural environment, and it must also be possible for
the sources of stone to be accurately and reliably determined
(Endonino 2002, 2003; Meltzer 1989; Odess 1998). A fairly
well developed body of knowledge has already been amassed
for lithic types and source areas in Florida (Austin 1997;
Austin and Estabrook 2000; Goodyear et al. 1983; Upchurch
et al. 1982). Even with all of the work that has been done there
are still some deficiencies that need to be addressed. A major
problem facing lithic provenance studies, and the one under
consideration here, is the inability to distinguish between
different chert sources derived from the Ocala Limestone in
a quantified and replicable fashion (Austin 1997; Austin and
Estabrook 2000). Numerous large quarry and production sites
in north-central Florida amply demonstrate the importance
of Ocala Limestone chert to prehistoric peoples (Bullen and
Dolan 1959; Clausen 1964; Hemmings and Kohler 1974; Purdy
1975, 1980, 1981). Unfortunately, Ocala Limestone cherts can
appear nearly identical and the different quarry clusters that
produce this material are not easily distinguished from one
another (Upchurch et al. 1982:120). Given the significance
of these quarry clusters to north-central Florida's prehistoric
inhabitants it is all the more important to be able to accurately
differentiate between them. Upchurch et al. (1982:121)
recognized this difficulty 25 years ago and anticipated that
revisions would be made to the extent and diagnostic criteria
of the quarry clusters they proposed.
The goal of this paper is straightforward to create
a stronger foundation for the investigation of prehistoric

interaction, exchange, and mobility in Florida by developing
replicable and quantifiable criteria for sourcing cherts from the
Ocala Limestone. By developing these criteria, the accuracy and
reliability of lithic raw material source determinations will be
increased. Upchurch et al. (1982) suggest that variations in the
abundance and size ofOrbitoid foraminifera (Figure 1), a group
of fossils round in plan view with a cross section characterized
by a central bulge and tapering edges (resembling a flying
saucer), are a primary criterion used in distinguishing between
the Gainesville, Ocala, and Lake Panasoffkee quarry clusters.
Although many foraminifera occur in Florida cherts, only a
few are used by paleontologists, geologists, and archaeologists
to identify geologic formations. For Ocala Limestone cherts,
Orbitoid foraminifera are the primary diagnostic fossil type
and the one considered in this study. These will be generally
referred to as "fossils."
Refining Upchurch et al.'s (1982) original quarry clusters
using fossil abundance and size criteria requires more
extensive sampling than the original study (only two samples
from each). Toward this end chert samples were collected
from Upchurch et al.'s (1982) Gainesville, Ocala, and Lake
Panasoffkee quarry clusters. These samples were then analyzed
to determine the number and size of fossils present in them.
It was anticipated that if observed patterns in the abundance
and size of fossils are valid criteria for characterizing specific
quarry clusters, then they should be differentially distributed
across the landscape and display a discernible degree of
geographic clustering. Demonstrating this to be the case, the
attribution of lithic artifacts to a particular quarry cluster is
more reliable because it is empirically based. A brief example
of the application of these refined criteria to a Mount Taylor
period lithic assemblage from the St. Johns River Valley is
presented later in the paper to demonstrate its usefulness.
First, however, it is necessary to know the distribution of lithic
resources across the landscape.

Chert Resources and Quarry Clusters

The presence of chert resources, how they are formed, and
the geologic processes acting to expose them are important in
understanding where chert occurs and why. Chert distribution
is heavily influenced by the region's geology and is generally
"restricted to the flanks of areas of uplift such as arches and
domes" (Upchurch et al 1982:12). The distribution of stone
outcrops is not uniform and is dependent to a large degree on
uplift and erosion to expose them. Frequently they are found
where overlying sand and clay have been removed; in and



VOL. 60(2-3)



Figure 1. Lepidocyclina spp. fossils in sample from the F 2 P
Figure 2. Pecten molds on the surface of a large Gaines-
Gainesville Quarry Cluster. Numerous examples are pres- v Q C c
ville Quarry Cluster chert boulder.
ent and arrows indicate some of the more typical specimens.
Circled specimen "I" is a classic example ofa Lepidocvclina

spp. fossil and "II" is a Pecten. Inset: a) Lepidocyclina spp.
fossil, b) Operculinoides spp. fossil (adapted from Austin
1997: Figure 18).

along rivers, around lakes and streams, and on the crests and
slopes of hills. In north-central peninsular Florida most of the
Ocala Limestone chert is exposed along the flanks of the Ocala
Uplift which is located on the western side of the Peninsular
Arch, a Cretaceous period (70-135 million years ago) landform
that forms the axis of the Florida peninsula and is characterized
by highlands running north-northwest to south-southeast from
southeastern Georgia into central Florida (Schmidt 1997;
Upchurch et al. 1982:12). Other residual material (including
chert) from younger deposits may also be present if they are
resistant to erosion. However, it is the Ocala Limestone cherts
that are of concern here.
The Ocala Limestone is composed of upper and lower
members (Scott 2001). It was deposited during the Eocene
Epoch some 40 to 60 million years ago (Upchurch et al.
1982:13). Lower Ocala Limestone is light colored, granular,
and dolomitic with fewer Orbitoid foraminifera. The upper
Ocala Limestone is extensively silicified and contains
abundant chert (Scott 2001; Upchurch et al. 1982: 17). Ocala
Limestone cherts are a foraminiferal grainstone to packstone'
with abundant fossils (Randazzo 1997:50). Diagnostic fossils
include several families of foraminifera and mollusks. As a
group, foraminifera consist of single-celled marine organisms
belonging to the phylum Protozoa and their shells make up
a significant portion of the ancient sediments that became
the Ocala Limestone. Three genera of foraminifera occur
frequently: Lepidocyclina, Operculinoides, and Nummulites.
Lepidocyclina spp. fossils are abundant in upper Ocala
Limestone and limited in the lower Ocala Limestone, making
the upper member quite distinctive (Scott 2001). The former
two are the most important for the purposes of the current
discussion (see Figure 1). Mollusks are also common and
Pecten (scallop-like shells) molds and casts are frequently
observed in both (Figure 2). Ocala Limestone and chert
crop out over much of central Florida, particularly along the

Ocala Uplift, but it also occurs in northern Florida along the
Chattahoochee Anticline (Upchurch et al. 1982:17). Upchurch
et al. (1982) divided the known chert-bearing exposures of
Florida, including those of the Ocala Limestone, into groups
characterized by material of the same geologic formation that
are more or less spatially isolated and distinct, pioneering the
quarry cluster method of provenance determination.

Quarry Clusters

A quarry cluster is defined as "an area known to contain
numerous exposures of chert, some of which must have been
used by early man, and in which the chert is expected to be
relatively uniform in fabric, composition, and fossil content."
Moreover, they are "usually within the same exposure belt or
a single formation" (Upchurch et al. 1982:9). Recognition of
the geologic formation that a particular specimen comes from
is a major step in provenance determination, especially if it is
spatially distinct. Of all the geologic strata in Florida few are
silicified and even fewer were available to prehistoric people
(Upchurch et al. 1982). Few geologic formations in Florida
contain significantly silicified deposits that would have
been available to prehistoric peoples: the Ocala Limestone,
the Suwannee Limestone, the St. Marks Formation, and the
Hawthorn Group (Austin 1997; Scott 2001; Upchurch et al.
1982:23). Within the Hawthorn Group, the Tampa Member of
the Arcadia Formation and the Peace River Formation contain
chert utilized by prehistoric peoples. The Avon Park Formation
also contains chert but does not have any significant surface
exposures and thus was not likely a significant source of lithic
material for prehistoric peoples.
Nineteen quarry clusters were originally proposed by
Upchurch et al. (1982:93) (Figure 3). Additional sampling was
expected to alter their findings and such has been the case.
Austin (1997:215) modified Upchurch et al.'s original clusters
because additional and more representative sampling lead to
a recognition that the criteria previously used to differentiate
between the Lower Suwannee, Gainesville, Ocala, and Lake
Panasoffkee quarry clusters were not sufficient to allow for an


2007 VOL. 60(2-3)



WC=Wright's Creek
US=Upper Suwannee
A=Alapaha River
SCS=Swift Creek Swamp
WS=White Springs
LS=Lower Suwannee
SF=Santa Fe
LP=Lake Panasoffkee
UW=Upper Withlacoochee
HR=Hillsborough River
TP=Turtlecrawl Point
PR=Peace River

50 Miles

50 Kilometers

Figure 3. Quarry clusters proposed by Upchurch et al. (1982).

accurate source attribution (Figure 4). He proposed to combine
those quarry clusters where cherts are derived from a single
geologic formation and unambiguous criteria for distinguishing
between them are lacking. The result was a decrease in the
total number of quarry clusters from 19 to 16 by dividing the
quarry clusters of central peninsular Florida into eastern and
western megaclusters. The eastern group retained the Ocala
Quarry Cluster name and consists of the Ocala, Gainesville,
and the lower portion of the Lake Panasoffkee quarry clusters.
The western group is made up of the Lower Suwannee, Santa
Fe, and northern Lake Panasoffkee quarry clusters and is
referred to as the Lower Suwannee/Lake Panasoffkee Quarry
Cluster. The Inverness Quarry Cluster was subsumed within
the Lower Suwannee/Lake Panasoffkee quarry cluster because
it was proposed by Upchurch et al. (1982) based on presumed
chert exposures and was never actually sampled. No outcrops
of chert have so far been identified in this area. Based on my
examination of a much larger sample of chert from outcrops
in north-central Florida, I believe that real differences do exist
between the Gainesville, Ocala, and Lake Panasoffkee quarry
clusters. For this study Upchurch et al.'s (1982) original
nomenclature is retained and I maintain that fossil size and
abundance criteria used for separating them are valid but are
in need of refinement and modification.

Quarry Clusters and Ocala Limestone Chert

Several quarry clusters contain Ocala Limestone chert but
only the Gainesville, Ocala, and Lake Panasoffkee clusters are
considered here. Diagnostic criteria for these quarry clusters
are described by Upchurch et al. (1982) and are summarized
in Table 1. Large and abundant Orbitoid foraminifera in a
grainstone or packstone fabric are typical of the Gainesville
cluster(Upchurch etal. 1982:122). The Ocala material generally
contains fewer fossils in a packstone fabric and these occur in
large homogeneous masses. Orbitoids such as Lepidocyclina
spp. are common as are Pectens. Lake Panasoffkee cherts
contain abundant large Orbitoids and scattered Pecten molds
in a grainstone fabric and, less frequently, in packstone fabrics.
Crystal-lined voids are another distinguishing characteristic of
the Lake Panasoffkee Quarry Cluster. While the Ocala and
Gainesville quarry clusters were heavily utilized by prehistoric
groups, the Lake Panasoffkee Quarry Cluster appears not to
have been used as intensively and this is attributed to the limited
nature of the exposures (Upchurch et al. 1982:121). Although
Upchurch et al.'s observations regarding the differences in
fossil size and abundance between these three quarry clusters
are still valid, they do not capture the range of difference that
exists among them.


2, j01 s





WC=Wright's Creek
US=Upper Suwannee uw
A=Alapaha River C
SCS=Swift Creek Swamp HR DPR
WS=White Springs TP V
SF=Santa Fe
LS/LP=Lower Suwannee/Lake Panasoffkee
UW=Upper Withlacoochee
HR=Hillsborough River
TP=Turtlecrawl Point
PR=Peace River

50 Miles

50 Kilometers

Figure 4. Revised quarry clusters proposed by Austin (1997).

Table 1. Upchurch et al.'s (1982) diagnostic criteria for quarry clusters characterized by Ocala Limestone chert in penin-
sular Florida.
Quarry Cluster Host Rock Fabric Diagnostic Criteria
Gainesville packstone, grainstone large abundant Orbitoids (especially
Lepidocyclina spp.) with common miliolids and pectens
Ocala packstone large Orbitoids, Miliolids, and
Pectens common, Orbitoid fossils less common than Gainesville, homogenous masses
Lake Panasoffkee grainstone, minor Orbitoids (Lepidocyclina spp.) large
and packstone abundant, Miliolids common, scattered Pectens, quartz-lined cavities

2007 VOL. 60(2-3)



Other quarry clusters characterized by Ocala Limestone
cherts not sampled or analyzed in this study include the Santa
Fe, Wright's Creek, Marianna, and the Upper Withlachoochee
River clusters. It is worth noting that there also are some
exposures in the vicinity of Dothan, Alabama (Claude
Van Order, personal communication, 2000) and these are
probably associated with the Marianna Quarry Cluster. While
the quarry clusters listed above have not been sampled, the
Santa Fe Quarry Cluster merits some comment relative to its
characterization and relationship to the Gainesville Quarry
Cluster and is presented later.


Increasing the number of samples used to re-characterize
the Gainesville, Ocala, and Lake Panasoffkee clusters was
critical to this research. Numerous samples from a variety of
geographically dispersed locations throughout north-central
Florida were sought and obtained (Table 2). Many of the chert
sources were known prior to sampling and the majority of them
come from the Ocala and Gainesville quarry clusters. Several
from the Lake Panasoffkee area in Sumter County were located
by driving throughout the area and looking for exposures in
road cuts, fields, rivers, and creeks. Additional locations were
sampled during the course of archaeological reconnaissance
surveys (Mitchell 1997a, 1997b; Stokes 2000; Stokes et al.
2001). UTM coordinates were taken for all samples with a
handheld Garmin IV GPS unit and each location was referenced
using its county, nearby municipalities, and major roads. In all,
47 separate outcrops were sampled (Figure 5). Of these 16 are
from the Gainesville Quarry Cluster, 19 from the Ocala, and
9 from the Lake Panasoffkee cluster. Two samples (S-25 and
S-26) near the southern end of the Lake Panasoffkee Quarry
Cluster are characterized by Suwannee Limestone material
and likely belong to the Upper Withlacoochee River Quarry
Cluster. These samples are not included in this analysis. One
locality in northern Alachua County (S-36) is a silicified coral
outcrop and does not contain any Ocala Limestone chert.
Samples were analyzed with the methods developed
by Upchurch et al. (1982). A binocular microscope with an
independent fiber-optic light source was used to visually
inspect each sample. Magnification ranged from 10-x to 70-x.
Lower magnifications (20-x and under) were the most useful.
Occasionally higher magnification was necessary to determine
the nature of certain aspects of fossil content, rock fabric, and
secondary inclusions. Moistening the samples is useful in
aiding the identification of the host fabric and fossil content
(Austin 1997; Austin and Estabrook 2000; Upchurch et al.
One sample was selected from each outcrop location for
the Gainesville (16 samples) and Ocala (19 samples) quarry
clusters. Due to the paucity of outcrops sampled in the Lake
Panasoffkee Quarry Cluster several samples were taken for
four of the sample locations from this area. This was done
in order to make up for deficiencies in the total number of
sample locations and all were selected from spatially disparate
locations within the same exposure/sample area. Three
quantitative samples each were taken from locations S-21,

S-22, and S-23 on the eastern side of the Lake Panasoffkee
Quarry Cluster, totaling nine quantitative samples in all. An
additional three samples were taken from S-24 on the western
side. The remaining five locations each contributed a single
quantitative sample, bringing the number of samples to 17
from the Lake Panasoffkee area.
Samples consist of an area 2-x-2 cm (4 cm2) in size placed
arbitrarily on a flat surface of the sample specimen. Within
each sample square, data on fossil size and abundance were
recorded. Whole and fragmented Orbitoids contained entirely
or partially within each sample square were counted and their
diameters measured. No differentiation of Orbitoid species
was made although those present were noted. All of the fossils
present within the quantitative sample square were measured
with digital calipers and the minimum and maximum fossil
sizes were recorded in millimeters.

Analysis Results

Table 3 summarizes the data on fossil abundance and
minimum and maximum fossil size. Maximum fossil size and
fossil abundance prove to be key criteria for distinguishing
between the Gainesville, Ocala, and Lake Panasoffkee quarry
clusters. Minimum fossil size was recorded during the original
analysis, but it did not prove useful in discerning among the
three quarry clusters. These data are presented for each of the
quarry clusters in order to provide a more complete dataset,
but they are not considered in depth. Fossil abundance is
expressed and discussed as a density value, in this case the
number of fossils per square cm (pcm2), rather than the total
number of fossils contained within the sample unit. Presenting
abundance data in this way provides a baseline against which
other samples can be compared.
Gainesville Quarry Cluster chert is defined by a packstone
to grainstone fabric and a fossil density ranging from 7.0-
13.0 fossils pcm2, frequently more, with an average density
of 9.55 pcm2. Maximum fossil size ranges from 6.0-10.0 mm
and average 10.5 mm. Pecten molds are common but may be
abundant in some locations. Crystal-lined voids were also
observed but are not frequent. Maximum fossil size varies
from 3.6 29.7 mm with most measuring 6.3-9.5 mm. The
majority have greater than 7.5 fossils pcm2 with most of these
possessing 7.25-12.5 pcm2. Six of the 17 samples possess
greater than 12.5 Orbitoids pcm2. Large and abundant fossils
are a key criterion in identifying this quarry cluster.
Ocala Quarry Cluster chert is defined by a homogenous
packstone fabric with a fossil density ranging from 1.0-3.0
fossils pcm2 and an average of 2.88 pcm2. Maximum fossil
size ranges from 4.0-8.5 mm with an average of 7.6 mm.
Pecten molds are common and are generally small in size
though large examples also have been observed. No crystal
lined cavities were noted and materials typically occur in
homogenous masses. Maximum fossil size varies from 2.6-
22.9 mm with the majority measuring 4.3-8.4 mm. In terms
of abundance, Ocala materials group tightly and, with the
exception of two outliers, generally have four or fewer fossils
pcm2 with most of the samples containing 0.75-2.5 fossils
pcm2. The rarity and small size of the fossils and homogenous



Table 2. Sample Locations.

Sample County
Upchurch et al.
S-1 Marion Ocala
S-2 Marion Ocala
S-3 Marion Ocala
S-4 Alachua Gainesville
S-5 Alachua Gainesville
S-6 Alachua Gainesville
S-7 Marion Ocala
S-8 Alachua Gainesville
S-9 Marion Ocala
S-10 Marion Ocala
S-11 Levy Ocala
S-12 Marion Gainesville
S-13 Levy Gainesville
S-14 Levy Ocala
S-15 Marion Ocala
S-16 Marion Ocala
S-17 Marion Ocala
S-18 Marion Ocala
S-19 Marion Ocala
S-20 Marion Ocala
S-21 Sumter Lake Panasoffkee
S-21-a Sumter Lake Panasoffkee
S-21-b Sumter Lake Panasoftkee
S-22 Sumter Lake Panasoffkee
S-22-a Sumter Lake Panasoffkee
S-22-b Sumter Lake Panasoffkee
S-23 Sumter Lake Panasoffkee
S-23-a Sumter Lake Panasoffkee
S-23-b Sumter Lake Panasoffkee
S-24 Sumter Lake Panasoffkee
S-24-a Sumter Lake Panasoffkee
S-24-b Sumter Lake Panasoffkee
S-25* Sumter Upper Withlacoochee
S-26* Sumter Upper Withlacoochee
S-27 Marion Ocala
S-28 Sumter Lake Panasoffkee
S-29 Sumter Lake Panasoffkee
S-30 Sumter Lake Panasoffkee
S-31 Sumter Lake Panasoffkee
S-32 Marion Ocala
S-33 Alachua Gainesville
S-34 Alachua Gainesville
S-35 Alachua Gainesville
S-36* Alachua Gainesville
S-37 Alachua Gainesville
S-38 Alachua Gainesville
S-39 Marion Ocala
S-40 Marion Ocala
S-41 Marion Ocala
S-42 Alachua Gainesville
S-43 Alachua Gainesville
S-44 Alachua Gainesville
S-45 Sumter Lake Panasoffkee
S-46 Marion Ocala
S-47 Marion Ocala
* Non-Ocala Limestone, not analyzed

Quarry Cluster
Lower Suwannee/Lake Panasoffkee
Lower Suwannee/Lake Panasoffkee
Lower Suwannee/Lake Panasoffkee
Upper Withlacoochee
Upper Withlacoochee
Lower Suwannee/Lake Panasoffkee
Lower Suwannee/Lake Panasoffkee
Lower Suwannee/Lake Panasoffkee

This Paper
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee East
Panasoffkee West
Panasoffkee West
Panasoffkee West
Upper Withlacoochee
Upper Withlacoochee
Panasoffkee West
Panasoffkee West
Panasoffkee West
Panasoffkee East
Santa Fe
Panasoffkee East


2007 VOL. 60(2-3)






11 S-2

S-10 S-27
* S-40 S-3

*S **

S-4* S-47

S-20 S-18




* S-28

Figure 5. Sample locations.


I 1





Table 3. Fossil abundance and size data for the Gainesville, Ocala, Lake Panasoffkee East and Lake Panasoffkee West
quarry clusters.
Quarry Cluster Abundance Min. Size (mm) Max. Size (mm)
mmin. max. avg. min. max. avg. min. max. avg.
Gainesville 3.50 20.00 9.55 0.6 3.2 1.4 3.6 29.7 10.5
Ocala 0.75 6.25 2.88 0.9 6.8 1.7 2.6 22.9 7.6
Lake Panasoffkee East 0.75 3.75 2.20 0.1 3.4 1.2 1.4 14.6 6.1
Lake Panasoffkee West 5.25 15.00 9.20 0.8 1.2 0.8 3 13.7 7.6

nature of raw material packages distinguish the Ocala Quarry
Cluster from the Gainesville and Lake Panasoffkee sources.
The two samples (S-3 and S-19) deviate from the overall
pattern observed in the Ocala sample group. The first (S-3)
was exposed through commercial limestone mining operations
and may or may not have been available to prehistoric hunter-
gatherers but is still included in the analysis. The second outlier
(S-19) is from an exposure that would have been available to
prehistoric groups and is thus the only significant deviation
from the observed pattern. Because of its location near the
southern end of the Ocala Quarry Cluster area and near the
northern end of the Lake Panasoffkee cluster, this sample may
represent a transition between the Ocala and Lake Panasoffkee
quarry clusters. Additional sampling in this area may help to
clarify the matter. Both outliers represent variation within a
group that is otherwise very consistent.
The Lake Panasoffkee Quarry Cluster merits special
consideration due to the clear differences in fossil abundance
within this source area. Samples are essentially identical
in terms of size (Figure 6). However, differences in fossil

abundance make it possible to distinguish between material
from the eastern side of the quarry cluster and that from the
west and south. Based on these differences I propose that the
Lake Panasoffkee Quarry Cluster be divided into eastern and
western sub-clusters, hereafter referred to as Lake Panasoffkee
East and Lake Panasoffkee West quarry clusters.
The Lake Panasoffkee East Quarry Cluster is defined by a
grainstone fabric with less than 3.75 fossils pcm2 and an average
of 2.2 fossils pcm2. Maximum fossil size ranges from 2.0-8.0
mm with an average of 6.1 mm. Pecten molds are common
and crystal-lined voids are common to frequent. Maximum
fossil size ranges from 1.4-14.6 mm with the majority falling
between 4.5-5.6 mm. Abundance varies from 0.75-3.75 fossils
pcm2 with most having 0.75-3.0 pcm2.
Lake Panasoffkee West cherts are characterized by a
grainstone and occasionally packstone fabric with 5.0-15
fossils pcm2 that measure 3.0-14 mm in size. Pecten molds and
crystal-lined voids are occasionally present. Maximum fossil
size varies between 3.0 mm and 14 mm with most falling
between 5.0-8.1. Samples contained a minimum of 5.25 fossils

Figure 6. Scatterplot showing maximum fossil size and abundance for Lake Panasoffkee East and West quarry cluster

kee E.

fkee W.

a Panasofl


a Panasoff


- ~ ~ ~ ~ ~ L -- - I- -I --- --l- --

max. size (mm)


2007 VOL. 60(2-3)


pcm2 and a maximum of 15 pcm2 with most having 5.25-6.25


In comparing the Gainesville, Ocala, and Lake
Panasoffkee East and Lake Panasoffkee West quarry clusters,
overlap between them is apparent (Figure 7). The Gainesville
and Ocala quarry clusters overlap in both size and abundance
(Figure 8). In terms of size, the lower end of the maximum size
distribution of the Gainesville Quarry Cluster overlaps with the
upper limit of the Ocala Quarry Cluster. Regarding abundance,
the Gainesville materials have more fossils on average with
over half falling outside the uppermost distribution of the Ocala
cluster, outliers aside. It would appear, then, that abundance is
the best indicator of difference between the Gainesville and
Ocala clusters. The best approach to distinguishing among
the Gainesville and Ocala quarry clusters is to consider both
fossil size and abundance. This also applies when comparing
Gainesville and Lake Panasoffkee quarry clusters. Gainesville
and Lake Panasoffkee West chert are similar in abundance and
both stand in contrast to the Lake Panasoffkee East Quarry
Cluster which contains notably fewer fossils (Figure 9).
Upchurch et al. (1982:126) noted the difficulty in discerning
between these two source areas but suggest that quartz crystal-
lined voids in Lake Panasoffkee cherts, both East and West,
provide one means of differentiating them. Though similar in
abundance, fossils in the Gainesville materials are still larger
than those observed for Panasoffkee West. Lake Panasoffkee
East materials are easily distinguished from Gainesville chert

by the rarity and small size of their fossils as well as the
presence of crystal-lined voids.
Ocala and Lake Panasoffkee East materials are very
similar and with a small-sized archaeological assemblage,
both in terms of the number of artifacts and their physical size,
would be virtually indistinguishable (Figure 10). Some of the
sampled Ocala Quarry Cluster cherts have more fossils than
observed for Lake Panasoffkee East, but not many. While their
maximum fossil sizes are similar, there are two samples from
Lake Panasoffkee East that have fossils significantly larger
than those observed for most of the Ocala Quarry Cluster
samples. Differences between the Ocala and Lake Panasoffkee
West samples mirror those observed for Lake Panasoffkee East
and Lake Panasoffkee West and for Gainesville and Ocala.
Lake Panasoffkee West materials are not easily differentiated
from Ocala chert due to their similarity in fossil size. There
are, however, marked differences in fossil abundance and this
criterion can be used to differentiate between them. These
differences, between the Gainesville and Ocala clusters,
between Lake Panasoffkee East and West, and between Ocala
and Lake Panasoffkee West, are in line with the differences
observed by Austin (1997) that lead to his revision of these
quarry clusters and the establishment of eastern and western
megaclusters. Chert samples from the western side of the
Ocala Uplift and those from the eastern side display fossil
assemblages with clear differences in abundance and moderate
differences in maximum size. Differing degrees ofsilicification
and texture are also useful in differentiating between Lake
Panasoffkee East and Lake Panasoffkee West. Cherts from
the west and south are generally more homogenous, better




S0A 0 130 0

10 0A0 0


a 4b


0 5 10 15 20 25 30 35
max. size (mm)

Figure 7. Scatterplot showing maximum fossil size and abundance for the Ocala, Gainesville and Lake Panasoffkee East
and West quarry clusters.

p U

* Ocala

o Gainesville
a Panasoffkee E.

A Panasoffkee W.



25 Ocala

0 13 130
0 0
3B 13

5. -- 0 --------

o. 0

0 10 20 30 40

max. size (mm)

Figure 8. Scatterplot comparing maximum fossil size and abundance for the Ocala and Gainesville quarry clusters.

25 o Gainesville
SPanasoffkee E.
20 n APanasoffkee W.

S 15 a
3 0 n D 0 0 0
o10 aAn

5 B ,, -A
00 0
0 ,

0 5 10 15 20 25 30 35

max. size (mm)

Figure 9. Scatterplot comparing maximum fossil size and abundance for the Gainesville, Lake Panasoffkee East., and Lake
Panasoffkee West quarry clusters.


2007 VOL. 60(2-3)

16 Ocala

14 Panasoffkee E.

12 A A Panasoffkee W.


2 .)Jf A B-^ <---------
6 A

a A
2 A

0 5 10 15 20 25

max. size (mm)

Figure 10. Scatterplot comparing maximum fossil size and abundance for the Ocala, Lake Panasoffkee East, and Lake
Panasoffkee West quarry clusters.

silicified, and have a smoother texture than those from the
east. Caution should be exercised when using these criteria,
however, since the degree of silicification within a single
exposure can vary widely.

Santa Fe Quarry Cluster

As noted earlier, the Santa Fe Quarry Cluster was once
thought to be characterized by Ocala Limestone chert and
is located at the northern end of the current study area.
Samples collected during this work, as well as others prior
and subsequent, have different diagnostic criteria than
originally proposed by Upchurch et al. (1982). The Santa
Fe Quarry Cluster is characterized primarily by Suwannee
Limestone chert and occasionally silicified coral. Outcrops
of Ocala Limestone material are present in northern Alachua
County south of High Springs. As one moves north from the
Gainesville Quarry Cluster and approaches the Santa Fe River,
Suwannee Formation chert appears to the exclusion of Ocala
Limestone chert. Samples collected in the vicinity of High
Springs, O'Leno State Park, and the Santa Fe River confirms
this. A single sample from an archaeological site near the
confluence of the Santa Fe and Ichetucknee rivers in Columbia
County was analyzed by Upchurch et al. (1982) and lead to the
designation of this quarry cluster. It may, however, be best to
include this location in the Lower Suwannee Quarry Cluster.
The Lower Suwannee Quarry Cluster also is characterized by
the single occurrence of Ocala Limestone chert at Fanning
Springs in Levy County. The isolation of this exposure and
the absence in general of silicified exposures in the area make

this source somewhat anomalous. Materials from this source
most closely resemble those in western Alachua County.
Further sampling is needed in this area in order to clarify the
relationship between this Ocala Limestone exposure and those
of the Gainesville Quarry Cluster to the east.

Residua: Non-Ocala Limestone Chert and Corals

During the collection of chert samples for this study
a few specimens that were clearly dissimilar to the usual
Ocala Limestone material from the Ocala Quarry Cluster
were encountered. Additionally, several locales that produced
silicified coral were brought to the author's attention by Claude
Van Order, expert flintknapper and prehistoric technologist
from Lakeland, Florida. The presence of "atypical" materials
is not insignificant and demonstrates the variability that can
be present within a single quarry cluster, a phenomenon
observed by Upchurch et al. (1982) but one that was not
pursued further. The masking of variation in the definitional
criteria of quarry clusters has recently been brought into
focus by Estabrook (2005) who notes that quarry clusters
have the potential to contain a greater diversity of material
than previously recognized. The failure of archaeologists to
address this variation can and will lead to misattribution of
residual material to more distant sources when in fact they
may be quite local. To help lessen the risk of such mistakes it
is appropriate to discuss cherts that are "out of place" within
the Gainesville, Ocala, and Lake Panasoffkee East and Lake
Panasoffkee West quarry clusters.




Figure 11. Non-Ocala Limestone chert from the Ocala
Quarry Cluster.

Two chert samples resembling Hawthorn Group material
were collected by the author from locations within the Ocala
Quarry Cluster (Figure 11). The first atypical specimen
(Figure 11, right) comes from an area near the junction of
U.S. 301 and U.S. 441 in Marion County. Poorly silicified
limestone was frequent at this location and good quality
chert was not abundant though better material is likely in the
vicinity. A medium-sized nodule collected from this location
and has a relatively fine texture and a slight luster. Its fabric
is reminiscent of Hillsborough River Quarry Cluster material
and it appears to have a mudstone or wackestone fabric.
Fossils observed consist of Rotalids (a family of small, single-
celled marine foraminifera). Abundant fine sand was present
as a secondary inclusion. Given these inclusions it would be
easy to misidentify these materials as originating from the
Hillsborough River Quarry Cluster (Upchurch et al. 1982:139-
The second atypical sample came from S-40 (Figure 11,
left); a prehistoric quarry site characterized by abundant Ocala
Limestone material. Several large nodules of this material
were present at this location and were collected by the author.
Though each nodule contained much high quality, fine-grained
material, it is typically interspersed within a poorly silicified
chert matrix. It is typically dark brown in color, lustrous, and has
a splotchy, brecciated appearance. No diagnostic fossils were
observed though a moderate amount of fine sand was present
in the poorly silicified areas. There are similarities between this
specimen and Hillsborough River Baybottom (Type 5) chert
and Caladesi Quarry Cluster chert as described by Goodyear
et al. (1983). Material bearing a strong resemblance to this
has been observed by the author from a prehistoric context
near Micanopy (Austin 2001). If nothing else, the preceding
examples demonstrate the need to develop an appreciation for
the amount of variability within quarry clusters rather than
assuming that they are characterized exclusively by one type
of material. Variation exists and the failure to recognize it can
and will lead to misidentification and misinterpretation.
In addition to the two chert samples, several locations
within the Gainesville, Ocala, and Lake Panasoffkee quarry
clusters are known to have produced nodules of silicified

coral. Silicified coral, because of its lack of diagnostic fossil
content, cannot be attributed with any certainty to a particular
source area though it is generally associated with both the
Suwannee Limestone and Tampa member of the Hawthorn
Group. No coral sources were noted in the Gainesville, Ocala,
or Lake Panasoffkee quarry clusters by Upchurch et al. (1982).
Several locations where coral has been found were brought
to the author's attention by Claude Van Order. Mr. Van Order
(personal communication, 2000) provided me with samples
and location information for several coral sources. These
include localities near Center Hill and Coleman in Sumter
County, south of Summerfield in southern Marion County,
and near Micanopy in Alachua County. Another source of
coral was discovered during an archaeological survey near
the town of Alachua in Alachua County (Stokes et al. 2001).
With the exception of the source near Alachua, none of these
coral outcrops appear to have been extensively exploited
prehistorically and in fact, only a small amount of material
suitable for the production of stone tools was present.
The presence of these atypical chert and coral deposits
support Estabrook's (2005) contention that greater variation
exists within quarry clusters and underscores the need to
account for this variation. It does not undermine the utility
of the quarry cluster approach, the general criteria used to
assign samples to these clusters, or reduce the significance
of the dominant types. By and large the central peninsular
outcrops of Ocala Limestone chert are characterized by
Orbitoids and have a grainstone or packstone fabric. It is,
however, a cautionary tale for those who use lithic sourcing to
draw conclusions about prehistoric mobility, settlement, and
exchange. Accounting for variability within quarry clusters is
yet another issue in desperate need of attention and tackling it
will be another step forward for provenance studies.

Spatial Distribution

Upchurch et al. (1982) proposed the boundaries of their
quarry clusters based to a significant degree on the extent of the
limestone formations containing chert and a limited number of
samples. Here, the distribution of chert sampled and analyzed
is used to redraw the quarry cluster boundaries. Based on
differences in the size and abundance of Orbitoid foraminifera
presented above, the boundaries for the Gainesville, Ocala,
and the Lake Panasofkee (East and West) quarry clusters
have been revised (Figure 12) and are more similar to those
proposed by Upchurch et al. (1982) and effectively reverse
changes made by Austin (1997).
The boundaries of the Gainesville Quarry Cluster extend
from an area sorth of the Alachua-Marion County line to just
south of Alachua, and west from the area around Newnan's
Lake to the Newberry area in western Alachua County and
Williston in eastern Levy County. The extent of this quarry
cluster is similar to that illustrated by Upchurch et al. (1982)
though here its eastern edge has been shifted to the west and
a portion of northwest Marion County and northeast Levy
County have been included in this quarry cluster.
The Ocala Quarry Cluster stretches from just south of
Orange Lake to southern Marion and northern Sumter county
and from the just east of Silver Springs westward to S.R. 41 in

2007 VOL. 60(2-3)



Gainesville Quarry Cluster (top)
Ocala Quarry Cluster (middle)
Lake Panasoffkee East Quarry Cluster (bottom top/right)
Lake Panasoffkee West QuarryCluster (bottom/left)



Figure 12. Revised boundaries for the Gainesville, Ocala, and Lake Panasoffkee East and Lake Panasoffkee West quarry

I 1




western Marion County. This configuration again is similar to
Upchurch et al.'s (1982) Ocala Quarry Cluster except for the
loss of some area in northwest Marion County and northeastern
Levy County mentioned earlier. A further modification made
as a result of this work is the separation of the southern end
of the Ocala Quarry Cluster and the northern end of the Lake
Panasoffkee Quarry Cluster. Both Upchurch et al. (1982) and
Austin (1997) show the Ocala and Lake Panasoffkee sources
as being contiguous. A break in the distribution of outcrops
between these two sources is the basis for this division.
Additional sampling at the southern end of the Ocala and the
northern end of the Lake Panasoffkee quarry clusters may
eventually lead to further revisions to these boundaries.
Lake Panasoffkee Quarry Cluster chert is distributed across
an area in northern Sumter County in the vicinity of the town
of Wildwood south to Bushnell and from the Withlacoohee
River at the Sumter-Citrus County line east to around U.S.
301. The East/West division of the Lake Panasoffkee Quarry
Cluster roughly corresponds to the eastern and western sides
of 1-75. Compared to both Upchurch et al. (1982) and Austin,
(1997) the Lake Panasoffkee Quarry Cluster has been reduced
as a result of this research. The northern boundary has been
retracted toward the south and the western extent has been
shifted eastward. The most significant modification (discussed
above), is the separation of this quarry cluster into eastern and
western sub-clusters.
A revised map reflecting the changes made to the extent of
the quarry clusters considered in this research as well as those
not considered is presented in Figure 13. The Lower Suwannee
Quarry Cluster is quite large in both Upchurch et al.'s (1982)
and Austin's (1997) representations, especially the latter (see
Figures 3 and 4). Based on reconnaissance in the vicinity of
the Lower Suwannee Quarry Cluster during this research and
the failure to locate any additional sources of chert suitable
for the production of stone tools in this area, I recommend
that this quarry cluster be reduced in size and encompass the
source at Fanning Springs and the surrounding environment.
The modified representation of the Lower Suwannee Quarry
Cluster can be seen in Figure 13. Changes to the extent of the
Santa Fe Quarry Cluster also have been made. Both Upchurch
et al. (1982) and Austin (1997) show the Santa Fe Quarry
Cluster in contact with the Lower Suwannee Quarry Cluster.
As a result of the reduced area of the Lower Suwannee Quarry
Cluster, the Santa Fe Quarry Cluster is here considered not to
be contiguous with it. Additional sampling is needed in order
to more accurately characterize the extent of this quarry cluster.
Provisionally, the Santa Fe Quarry Cluster can be considered
to extend from an area just south of High Springs and the
Santa Fe River north into southern Columbia County around
O'Leno State Park, and east from near the confluence of the
Ichetucknee and Santa Fe rivers up to, but no further than,
the border of Columbia and Union counties. Having presented
the fossil size and abundance criteria for distinguishing among
the quarry clusters characterized by Ocala Limestone chert,
its application to an archaeological assemblage is in order to
demonstrate its usefulness.

Archaeological Application

My own interest in interaction and exchange among
the Middle to Late Archaic Mount Taylor peoples of the St.
Johns River Valley (SJRV) stimulated this research (Endonino
2003). The absence of naturally occurring lithic raw material
suitable for making chipped stone tools in the SJRV rules out
the possibility of local procurement. The production of lithic
tools and the byproducts of their manufacture in this region
must, therefore, have come from somewhere else. Given the
proximity of the Gainesville, Ocala, and Lake Panasoffkee
quarry clusters to the St. Johns River, I anticipated that
much of the material found in the SJRV would be from these
sources. When applied to an archaeological assemblage, the
criteria for discerning between the Gainesville, Ocala, and
Lake Panasoffkee quarry clusters outlined earlier successfully
facilitated the attribution of lithic tools and debitage to each of
these sources with greater confidence. A successful application
of these criteria is demonstrated below through a comparison
of the results of two sourcing efforts on the same assemblage:
the first of these following criteria for the identification of
Ocala Limestone chert outlined by Upchurch et al. (1982) and
the second using the revised criteria presented in this paper.
The assemblage considered comes from 8V053, the Lake
Monroe Outlet Midden (LMOM), located on the western
shore of Lake Monroe north of its juncture with the St. Johns
River (Figure 14). A series of standard radiometric and AMS
dates place this site's occupation between 4040-3090 B.C.
(two sigma calibration) with most of the dates falling between
3660 B.C. and 3340 B.C. (ACI 2001:9-1). A lithic reduction
area spatially segregated from the midden deposits produced
an impressive assemblage of bifaces, microliths, flake tools,
and debitage; arguably the largest Mount Taylor period lithic
assemblage excavated to date. Both analyses ofthis assemblage
were performed on materials from the midden and the lithic
Previous work by Archaeological Consultants, Inc. (ACI
2001:5-6) determined that at least five different quarry clusters
are represented in the assemblage. Much of the material is
believed to be chert from Ocala Limestone sources, and the
Ocala Quarry Cluster in particular, as it is indicated as being
the nearest source of chert to the project area. Other sources
identified include the Peace River, Upper Withlacoochee,
Brooksville, and Hillsborough River quarry clusters.
However, source identifications are not indicated for debitage,
microliths, or other tools and were only separated according
to raw material type (chert vs. coral). Some of the bifaces
were attributed to specific source areas "when possible"
(ACI 2001:2-5). The inconsistent attribution of tools and the
apparent lack of attributions for other tools and debitage make
direct comparisons of sourcing efforts impossible. Likely
the statement that Ocala Quarry Cluster materials account
for the majority of the chert from this assemblage is based
on impressions formed through observation during sorting
and analysis. The attribution of a significant portion of the
chert in the assemblage to the Ocala Quarry Cluster, without
applying well-defined criteria, masks a great deal of variability
in the sources present in the overall assemblage, especially


2007 VOL. 60(2-3)


Figure 13. Quarry clusters reflecting changes based on this research.

considering that more than one quarry cluster characterized by
Ocala Limestone chert are just as close if not closer than the
Ocala Quarry Cluster.
A sample of the LMOM assemblage consisting of
1583 lithic artifacts was identified to source by the author
(Endonino 2003). Tools came from all excavated proveniences
at the site but only debitage from Test Units A (3x3-m) and
C (4x4-m) within the midden deposits were analyzed. Test
Unit B (4x4-m), placed at the location of the lithic workshop
and located away from the midden, produced an abundance
of debitage and microlithic tools as well as haftable bifaces

and biface fragments. Debitage from this location has not yet
been analyzed in detail, but during preliminary sorting it was
observed that somewhat less than half of the material was
chert, and of that, most appeared to be derived either from
Suwannee or Ocala Limestone sources; this is in accord with
the findings presented below. Among the artifacts analyzed are
36 haftable bifaces and biface fragments, 266 "other" tools
including a number of flake tools and microliths, and 1281
pieces of debitage.
The frequency and percent of each source area identified
is presented in Table 4. Twelve distinctive source areas




Figure 14. Location of 8VO53, the Lake Monroe Outlet Midden.

were identified and of these, five are characterized by Ocala
Limestone materials. Overall, Ocala Limestone materials
account for a little over 30 percent of the assemblage. All of
the quarry clusters characterized by Ocala Limestone chert
considered in this paper were present: Gainesville, Ocala,
Lake Panasoffkee East, and Lake Panasoffkee West. The Lake
Panasoffkee chert, especially Lake Panasoffkee West, proved
to have accounted for most of the chert present at the site. Chert
from the Green Swamp and Rock Ridge areas within the Upper
Withlacoochee River Quarry Cluster provided a fair amount
of the lithic material, about seven percent of the site total.

The Green Swamp/Rock Ridge localities are characterized
by Ocala Limestone chert but their case is unique in that it
also contains fossils diagnostic of the Suwannee Limestone
and sand as a secondary inclusion (Upchurch et al. 1982:132).
These characteristics in combination make this source area
unique and readily identifiable. Other quarry clusters are
represented as well. Hillsborough River Quarry Cluster
material identified comes from two sub-areas: the Upper
Hillsborough River and Cowhouse Creek. The presence of
frequent gastropods, sometimes called "drills" or "aguers," are
an indicator of origins in the upper reaches of the Hillsborough

2007 VOL. 60(2-3)


Table 4. Quarry cluster determinations for Lake Monroe Outlet Midden (8VO53) lithic materials.

Material Quarry Cluster Bifaces OtherDebitage Total
N et. N Pt. N Peb t. N PTt.
N Pct. N Pct. N Pct. N Pct.

Lake Panasoffkee East
Lake Panasoffkee West
Lake Panasoffkee Indeterminate
Ocala Limestone Indeterminate
Upper Withlacoochee River, GS/RR
Upper Withlacoochee River
Suwannee Limestone Indeterminate
Hillsborough River
Hillsborough River, CHC
Upper Hillsborough River

GS/RR=Green Swamp/Rock Ridge locality within the Upper Withlacoochee River Quarry Cluster, CHC=Cow House Creek locality within the Hillsborough River Quarry Cluster.

River. Charophyte oogonia, the reproductive apparatus of a
freshwater plant, are diagnostic of the Cowhouse Creek and
Harney Flats areas in Hillsborough County (Upchurch et al.
1982:74). Caladesi Quarry Cluster chert is represented by a
few of the "other tools" and debitage and these may have been
the Peace River materials identified during the ACI (2001)
analysis. The particular source within the Peace River Quarry
Cluster they cite are characterized by fossils (Soritesspp.) and
sand inclusions similar to those characteristic of the Caladesi
materials. The findings presented here, as well as those ofACI
(2001), are in agreement regarding the presence of Suwannee
Limestone chert sources, namely the Upper Withlacoochee
River (exclusive of the Green Swamp and Rock Ridge areas)
and Brooksville quarry clusters.
These source identifications diverge from those of ACI
(2001) in that more sources have been identified and, most
importantly, materials attributable to the Gainesville, Ocala,
and Lake Panasoffkee East and West quarry clusters were
discernible using size and abundance criteria. Whereas the
original analysis by ACI (2001) identified only a single quarry
cluster characterized by Ocala Limestone chert, the Ocala
Quarry Cluster following Upchurch et al. (1982), the revised
criteria presented in this paper allowed for all four sources
considered here as well as the related, though distinct, Green
Swamp/Rock Ridge area within the Upper Withlacoochee River
Quarry Cluster, to be discerned. The ability to more accurately
source lithic artifacts from 8V053, or any other assemblage,
allows issues such as mobility, band range, and exchange to be
addressed in a more precise and nuanced fashion. While a more
thorough consideration of the mechanisms for bringing lithic
raw materials into the SJRV is beyond the scope of this paper,
the criteria developed here for discerning among Gainesville,
Ocala, and Lake Panasoffkee East and West cherts have been
shown to be applicable to archaeological problems.


These results are encouraging. There are real differences
in the size and abundance of fossils in cherts from Upchurch et
al.'s (1982) Gainesville, Ocala, and Lake Panasoffkee quarry
clusters. These differences are observable, quantifiable, and can
be used to differentiate between the clusters. Table 5 presents
these revised criteria. Based on differences in the abundance and
size of fossil content of the samples studied, alterations to the
geographic extent of the Gainesville and Ocala quarry clusters
also have been made and the Lake Panasoffkee Quarry Cluster
has been divided into eastern and western clusters. Additional
sampling will, as predicted by Upchurch et al. (1982) 25 years
ago, likely result in additional changes and refinements. Future
work should focus on collecting more samples from Ocala
Limestone-derived cherts, especially at the northern end of the
Gainesville Quarry Cluster, the southern end of the Ocala and
northern end of the Lake Panasoffkee East and West quarry
clusters, and within the Lake Panasoffkee East and West
quarry clusters. Sampling within the Lower Suwannee Quarry
Cluster in order to characterize and determine its extent is yet
another important area for future work. Finally, efforts need to
be made to search for "anomalies" within quarry clusters that
might be mistaken for materials from other source areas.
In all, the results have been positive and I believe they will
be useful to others interested in the provenance determination
of cherts used by prehistoric groups. This paper has largely
been methodological in its orientation and most of it has been
of a geological nature with its relationship to archaeology
only minimally developed. Hopefully the brief example
of their application to an archaeological assemblage will
demonstrate their usefulness. Knowing the sources of stone,
their distribution, and being able to distinguish between them
using empirical and quantifiable criteria is necessary for laying
a solid foundation for provenance studies and gives further

Silicified Coral

11 0.69
38 2.40
104 6.57
161 10.17
2 0.13
52 3.28
92 5.81
69 4.36
39 2.46
3 0.20
8 0.51
2 0.13
1 0.06
14 0.88
19 1.20
666 42.07
IoR1l in 0o'

Iv~u ~ Ilel 2001 1J

5 1.00
5 3.54
4 7.21
7 14.97
! 3.92
1 7.01
S 4.99
5 2.90
t 0.77
1 1.08
0 1.26
3 50.72
13 100


y^1- .-

Table 5. Revised diagnostic criteria for the Gainesville, Ocala, Lake Panasoffkee East and Lake Panasoffkee West quarry
Quarry Cluster Host Rock Fabric Diagnostic Criteria Avg. Abund.* Avg. Size*

Gainesville packstone, grainstone abundant Orbitoids, tend to be larger 9.55 10.5
than other quarry clusters, Pectens
frequent, fairly well-silicified, some
quartz-lined voids

Ocala packstone small Orbitoids that are few in 2.88 7.6
number, Pectens common, chert
comes in homogenous and variably
silicified masses

Lake Panasoffkee E. grainstone, minor Orbitoids few in number and tend to 2.20 6.1
packstone be small, silicification is variable but
tends to be less than Ocala or
Panasoffkee W, scattered Pectens,
quartz-lined cavities

Lake Panasoffkee W. grainstone, minor abundant Orbitoids variable in size 9.20 7.6
packstone but not as large as Gainesville,
scattered Pectens, quartz-lined cavities,
better silicification that Panasoffkee East

Per square centimeter.

strength to the to the conclusions arrived at by archaeologists
using this method of inquiry.


1. Upchurch et al. (1982) define packstones a grain supported
stone where the large grains are in contact and the pores
between the grains are filled with mud. Grainstones are grain-
supported rocks with minimal mud in the pore spaces resulting
in a highly permeable and porous material.


This work was initiated as part of my master's research
at the University of Florida and I would like to thank my
committee chair Ken Sassaman and committee member Steve
Brandt for their guidance. Bob Austin and Rich Estabrook
provided valuable information, ideas, and insights as well as
thoughtful comments on earlier drafts of this paper. Marion
Almy and Joan Deming of Archaeological Consultants,
Inc. facilitated the loan of the Lake Monroe Outlet Midden
materials with permission of the Florida Department of
Transportation District 5. Bryan Harrell, Luther Quinn, and
Kevin Stofan provided technical assistance with the GIS maps
and the resulting figures used in this paper. Helpful comments
and suggestions from Rich Estabrook and two anonymous
reviewers, nearly all of which were taken into consideration
and implemented, greatly improved this paper. My thanks
also go to Debby Mullins and Andrea White for helping
me navigate the editorial process, their assistance is much
appreciated. Lastly, and by no means least, I owe a great deal
of thanks to Claude Van Order for sharing his knowledge and
chert samples with me as well as his generosity and hospitality

and that of his family. All errors and oversights are entirely
the responsibility of the author.

References Cited

Archaeological Consultants, Inc.
2001 Phase III Mitigative Excavation at the Lake Monroe
Outlet Midden (8V053), Volusia County, Florida.
Prepared for the U.S. Department of Transportation,
Federal Highway Administration and the Florida
Department of Transportation, District 5. Report
on File, Florida Division of Historical Resources,

Andrefsky, William Jr.
1994 Raw-Material Availability and the Organization of
Technology. American Antiquity 59(1):21-34.

Austin, Robert J.
1996 Prehistoric Chert Procurement Strategies and
Settlement Mobility on the Lake Wales Ridge.
Florida Anthropologist 49(4):211-237.

1997 The Economics of Lithic-Resource Use in South-
Central Florida. Ph.D. dissertation, Department of
Anthropology, University of Florida, Gainesville.

2001 Phase II Excavations at the Cameron Site (8AL45),
Tuscawilla Hills Development Site, Alachua County,
Florida. Report prepared for Tuscawilla Hills
Development by Southeastern Archaeological
Research, Inc.


2007 VOL. 60(2-3)


Austin, Robert J. and Richard W. Estabrook
2000 Chert Distribution and Exploitation in Peninsular
Florida. Florida Anthropologist 53(2-3):116-130.

Bamforth, Douglas B.
1986 Technological Efficiency and Tool Curation.
American Antiquity 51(1):38-50.

Binford, Lewis R.
1979 Organization and Formation Processes: Looking at
Curated Technologies. Journal of Anthropological
Research 35(3):255-273.

1980 Willow Smoke and Dog's Tails: Hunter-Gatherer
Settlement Systems and Archaeological Site
Formation. American Antiquity 45(1):4-20.

Bullen, Ripley P. and Edward M. Dolan
1958 The Johnson Lake Site, Marion County, Florida.
Florida Anthropologist. 12(4):77-94.

Carr, Philip J. and Lee H. Steward
2004 Poverty Point Chipped Stone Tool Raw Materials:
Inferring Social and Economic Strategies. In Signs
of Power: The Rise of Cultural Complexity in the
Southeast, Jon L. Gibson and Philip J. Carr, eds., pp.
129-145. University of Alabama Press, Tuscaloosa.

Clausen, Carl. J.
1964 The A-356 Site and the Florida Archaic. MAThesis,
Department of Anthropology, University of Florida,

Cobb, Charles R.
2000 From Quarry to Cornfield. The Political Economy
of Mississippian Hoe Production. University of
Alabama Press, Tuscaloosa.

Daniel, I. Randolph Jr.
1998 Hardaway Revisited: Early Archaic Settlement
in the Southeast. University of Alabama Press,

2001 Stone Raw Material Availability and Early Archaic
Settlement in the Southeastern United States.
American Antiquity 66(2):237-265.

Endonino, Jon C.
2002 Examining Social Landscapes Using Lithic
Provenance Determination: Exchange and Social
Interaction During the Pre-Ceramic Middle and
Late Archaic, St. Johns River Valley, Florida.
Paper Presented at the 59th Annual Meeting of the
Southeastern Archaeological Conference, Biloxi,

2003 Hunter-Gatherer Interaction and Exchange During
the Middle and Late Preceramic Archaic, St. Johns

River Valley, Florida. Non-Thesis Master's Project,
Department of Anthropology, University of Florida,

Endonino, Jon C., Robert J. Austin, Jane-Ann Blakney-Bailey,
Meggan Blessing, Carol Colannino, and Jon-Simon Suarez
2005 Preliminary Results of Phase III Mitigation
Excavation at the Monteverde Site (8LA243),
Lake County, Florida. Paper Presented at the 62nd
Annual Meeting of the Southeastern Archaeological
Conference, Columbia, SC.

Estabrook, Richard W.
2005 Mapping the Residuals: Chert Outcrop Outliers
and Lithic Reuse in Peninsular Florida. Paper
Presented at the 2005 annual meeting of the Florida
Anthropological Society, Gainesville.

Goodyear, Albert C., Sam B. Upchurch, Mark J. Brooks, and
Nancy N. Goodyear
1983 Paleo-Indian Manifestations in the Upper Tampa Bay
Region, Florida. The Florida Anthropologist 1-2:

Gramley, R. M.
1980 Raw Material Source Areas and "Curated" Tool
Assemblages. American Antiquity 45(823-833).

Hemmings, E. Thomas and Timothy A. Kohler
1974 The Lake Kanapaha Site in North Central Florida.
Bureau of Historic Sites and Properties Bulletin
No. 4, pp. 45-64. Division of Archives, History,
and Records Management, Department of State,

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

Mitchell, Scott E.
1997a An Archaeological Assessment Survey of SW 20't
Avenue from West of Tower Road to the Intersection
of 34'h Street, Alachua County, Florida. Report
prepared for FDOT District 2 by Southeastern
Archaeological Research, Inc.

1997b Archaeological Investigations at the Keeler Quarry
Site (8AL2331): A Prehistoric Quarry in Alachua
County, Florida. Report prepared for FDOT District
2 by Southeastern Archaeological Research, Inc.

Odess, Daniel
1998 The Archaeology of Interaction: Views from Artifact
Style and Material Exchange in Dorset Society.
American Antiquity 63(3):417-435.

Purdy, Barbara A.
1975 The Senator Edwards Chipped Stone Workshop Site
(8MR122), Marion County, Florida: A Preliminary



Report of Investigations. Florida Anthropologist

1980 Floridas Prehistoric Stone Technology: A Study of
the Flintworking Techniques of Early Florida Stone
Implement Makers. University of Florida Press.

1981 Investigations into the Use of Chert Outcrops by
Prehistoric Floridians: The Container Corporation of
America Site. Florida Anthropologist 34(2):90-108.

Randazzo, Anthony F.
1997 The Sedimentary Platform of Florida: Mesozoic to
Cenozoic. In The Geology of Florida, Anthony F.
Randazzo and Douglas S. Jones, eds., pp. 39-56.
University Press of Florida, Gainesville.

Sassaman, Kenneth E., Glen T. Hanson, and Tommy Charles
1988 Raw Material Procurement and the Reduction of
Hunter-Gatherer Range in the Savannah River Valley.
Southeastern Archaeology 7(2):79-94.

Schmidt, Walter
1997 Geomorphology and Physiography of Florida. In The
Geology ofFlorida, Anthony Randazzo and Douglas
S. Jones, eds., pp. 1-12. University Press of Florida,

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

Stokes, Anne V.
2000 A Cultural Resource Assessment Survey of the
Alachua West DRI, Alachua County, Florida. Report
prepared for WACO Properties, Inc. by Southeastern
Archaeological Research, Inc.

Stokes, Anne, V., Robert J. Austin, Jerald T. Milanich, and Jon
C. Endonino
2001 Cultural Resource Assessment Survey of Mill
Creek Estates, Alachua County, Florida. Report
prepared for Alan Cain Realty, Inc. by Southeastern
Archaeological Research, Inc.

Upchurch, Sam B., Richard N. Strom, and Mark G. Nuckels
1982 Methods of Chert Provenance Determination of
Florida Cherts. Report prepared for the Florida
Division of Archives, History, and Records
Management, Bureau of Historic Sites and Properties,

2007 VOL. 60(2-3)




S567 Laurel Cherry Lane, Venice, FL 34293
Email: mojowilder@gmail.com

2Department of Geography and the Environment, University of Texas at Austin, Austin, Texas 78712
Email: chasuz@toast.net

3Sheffield Centre for International Drylands Research, Department of Geography, University of Sheffield, Winter Street, Sheffield
S10 2TN UK
Email: MD.Bateman@sheffield.ac.uk

4Geo-Marine Inc., 2201 K Avenue, Suite A2, Plano, Texas 75074
Email: dpeter@geo-marine.com


This paper presents the significant findings of recent test
excavations of selected archaeological sites on Avon Park
Air Force Range (APAFR), Highlands and Polk counties,
Florida. These investigations were sponsored by the United
States Air Force, HQ Air Combat Command, APAFR, and the
U.S. Army Corps of Engineers, Fort Worth District, in order
to assess several sites' potential eligibility for inclusion in the
National Register of Historic Places. Although excavations
at any individual site were not extensive and were designed
primarily to satisfy management needs, a suite of geophysical
analyses was conducted concurrently with the archaeological
excavations. The archaeological and geophysical data, coupled
with accelerator mass spectrometer (AMS) carbon dating and
ages obtained via optically stimulated luminescence (OSL) of
quartz sand grains, have provided a better understanding of
the position of these selected sites within the archaeological
and geomorphic framework for the Osceola Plain region
of Florida. This work has implications regarding the local
cultural chronology and has also generated new information
concerning the contextual integrity of deeply buried sites in
sandy soil matrices.

The Setting

APAFR occupies 106,110 acres in southcentral Florida
and lies in portions of Polk and Highlands counties (Figure
1). Bombing Range Ridge is the principle topographic feature
in this area and runs roughly north-south through the center of
APAFR. Elevations range from a mean of about 70 feet above
mean sea level (amsl) in the flatlands to a peak of about 140 feet
amsl on the crest of Bombing Range Ridge. The ridge is the
dividing line between the Kissimmee River drainage system
to the east and Arbuckle Creek to the west. Between these
two major alluvial corridors and the ridge itself, the landscape
consists of extremely flat terrain with a pine flatwoods
vegetative complex interspersed with grassy ponds and cypress

domes. In addition to such isolated karstic depressions, there
are also widespread, cascading karstic depressions situated on
these surfaces. The hydrological connectivity of these features
appears to vary directly with fluctuations in the elevation
of the water table, with many of these features acting like
isolated closed depressions throughout most years and only
being connected when the local water table is very close to
the ground surface (cf. Choi et al. 2003; Harvey et al. 2003).
When the latter occurs, many of the karstic depressions on the
Osceola surface appear to exhibit elements of flow that are
more stream-like than the ponds/marshes that they appear to
be normally. The predominant vegetative community consists
of pine flatwoods, but hardwood hammocks are interspersed
along the margins of these streams and marshes. This habitat
supports an abundance of wildlife, including wild turkey, deer,
river otter, alligator, several species of tortoise, a variety of
snakes, and a plethora of avian species.
Eighteen archaeological sites were investigated during
this project, but the historic and historic-military-related
aspects of some of these sites are not considered in this article.'
Prehistoric sites tested included five that can be characterized
as swamp-marsh habitation middens dating primarily to post-
2,000 B.P. Four sites consisted predominantly of upland lithic
scatters that yielded little or no dateable organic materials.
Many of these sites have a demonstrably Archaic (mid-to-
late) association based on recovered point types, but other
occupations cannot be ruled out. Two sites contained elements
of both of these gross landscape/site-type categories, and one
site was situated in a lacustrine setting on the eastern shore of
Lake Arbuckle.


Fieldwork procedures followed conventional methods:
test units were excavated in 10-centimeter (cm) levels; the
removed sediment was passed through a 6.35-millimeter
(mm; 0.25-inch) hardware cloth; and all artifacts retained
on the screen were collected. The excavator drew one or


VOL. 60(2-3)




Figure 1. General location of the Avon Park Air Force Range, Highlands and Polk Counties, Florida.

more walls of each test unit. The geomorphic setting of each
site was evaluated by means of aerial photographs, which
complemented surface examination of the geomorphology and
inspection of test unit profiles. Column samples of sediment
were removed in 5-cm increments from selected test units on
each site for subsequent physical characterization. Charcoal
samples used for radiocarbon dating were collected in the field
and from flotation following the completion offieldwork.

In orderto better understand the formation processes at each
site, the column samples were subjected to a variety of physical
and chemical analyses that characterized the composition of
the deposits present. Analytical procedures employed on the
sandy sites were designed in part from comments by Leigh
(2001), and previous experience with sandy sites in Texas
was employed to facilitate the discrimination of potential
eolian deposits from sites buried by pedoturbation. Methods
employed included detailed field examination and subsequent
analytical characterization of the profile using granulometry,
magnetic susceptibility, and organic carbon content, as well as
comparison of the depth distribution of size-graded artifacts,
heavy mineral analysis, and detailed dating using radiocarbon
as well as single grained and standard aliquot OSL dating.

Some of these methods proved to be more useful than others.
On the wetland sites, the analysis of the calcium carbonate
content by Chittick was also performed to confirm the presence
of marls.


Fine resolution particle size analysis may detect subtle
depositional attributes that could be critical in understanding
the formation processes of a site (e.g., Frederick et al. 2002).
For this work, the < 2-mm particle size fraction was analyzed
for samples collected at 5-cm increments from the wall of
one or more test units at each site. These samples were taken
down to the maximum depth possible for each site, which was
generally the water table. The small subsample used in this
analysis was first digested in concentrated approximately 30
percent) hydrogen peroxide on a hot plate for approximately
20 minutes in order to remove organic material that might bind
sedimentary particles together. The sample was subsequently
analyzed on a Cilas 940 Laser Particle Sizer, and the resulting
cumulative curve exported as an ASCII text file. Various
descriptive statistical parameters were subsequently calculated
from these data using Microsoft Excel.

* .. N *

S Polk County



* UI~


2007 VOL. 60(2-3)


Heavy Mineral

One of the ways Leigh (2001) advocated distinguishing
eolian from subaqueous sands was by means of heavy mineral
analysis that compared the ratio of the weight percent of light
minerals and the weight percent of heavy minerals larger
than 0.25 mm in diameter. The rationale is that the ratio of
heavy minerals to light minerals for the >0.25-mm fraction is
significantly different for wind- and water-transported sands
owing to the different density of the minerals. To perform this
analysis, 40-50 grams of dried sediment from each 5-cm level
were weighed and subsequently mixed with hydrogen peroxide
on a hot plate for 30 minutes to remove organic material, and
then passed through a 63-micron sieve to remove the silt and
clay particles present. The samples were then dried, and the
heavy and light minerals separated in sodium polytungstate at
a density of 2.89 g/cm3. The resulting heavy and light mineral
fractions were subsequently rinsed, dried, sieved through a
0.25-mm sieve, and weighed. Unfortunately, the impoverished
heavy mineral assemblage from these deposits is completely
incomparable to Leigh's (1998, 2001) results. The absence of
any significant and clear stratigraphic trend among the samples
suggests that the depositional process responsible for these
deposits was the same throughout the sampled sections.

Artifact Distribution

The depth distribution of size-graded artifacts is one means
of examining post-depositional movement of an archaeological
assemblage. Ifthe occupation materials are undisturbed, thenthe
coarse and fine (macroartifacts and microartifacts, specifically
debitage and ceramics) should track together and have the
same depth profiles. If the cultural deposit has been disturbed,
the differences in which these two size classes are distributed
may provide clues to the agent responsible. For instance,
small artifacts may be moved by insects that would ignore or
avoid larger artifacts, thereby altering the depth distribution
of the two groups. Samples collected for this analysis were
bulk samples cut in a roughly 15-x-10-cm column in 5-cm
levels from the side of the test unit. Unfortunately, the low
artifact density of the sites resulted in few microartifacts being
recovered from most units; therefore, it is clear that the entire
unit would need to be fine-screened, not just a subsample, if
artifact distributions are to be informative.

Optically Stimulated Luminescence Dating

The most useful method employed was OSL dating. In
general terms, OSL dating measures the amount oftime that has
passed since quartz sand grains have been exposed to sunlight.
For sediments deposited by normal geological processes,
OSL ages typically represent the time that has passed since
the sediment was last transported. However, any process that
results in sand grains being exposed to sunlight can reset the
OSL clock; post-depositional disturbance is one such process
that must be considered here (discussed in detail below).

OSL dating relies on the premise that individual sand
grains of quartz or feldspar act as micro-dosimeters that store
a trapped charge in proportion to the amount of background
ionizing radiation they have been subject to since burial. Every
time grains are exposed to sunlight or high temperatures, this
trapped charge is released, thus resetting the "time clock." In
order to derive an OSL age, the total radiation dose accumulated
since last exposure to sunlight (known as the paleodose or
De) and the annual background radiation dose rate have to
be determined (see Aitken 1998 for detailed explanation).
Luminescence measurements are used to quantify the paleodose
in a sample, and through analysis of the concentrations of
uranium, thorium, and potassium in sediments, the annual
background radioactivity can be calculated. Thus, to derive an
OSL age, the paleodose (expressed in Gray units) is divided by
the annual dose rate (Grays/year).
The applicability of OSL dating to the investigation of
the formation processes of sandy sites has been discussed
in theory by Bateman et al. (2003), Frederick and Bateman
(1998), and Leigh (2001:284-286). This technique was
recently used in sandy sites in eastcentral Texas (Bateman et
al. 2003; Bateman et al. 2007a, 2007b; Frederick et al. 2001;
Frederick et al. 2002), where the issues of formation processes
are very similar to the current study area.
OSL can provide help when considering formation
processes because the reproducibility of multiple ages from
a single sample will be affected. To explain, in the case of
a sediment that was exposed to sunlight (thereby resetting
the luminescence signal) during transport, buried and
subsequently left undisturbed, sand grains buried at the
same time in the same depth/location should have stored the
same paleodose (De) from which an age can be calculated.
If multiple measurements are made of such a sediment, high
reproducibility is expected with the ages forming a single
normal (highly peaked) distribution around the mean (Figure
2-a). If not all sand grains were exposed to sunlight prior to
burial, or subsurface mixing of grains due to bioturbation has
occurred, then such a distribution is unlikely. In such cases,
multiple measurements of a single sample would lead to ages
forming a distribution with an old age tail, or a multimodal
age distribution as a proportion of the sand grains would have
retained an antecedent De signal unrelated to the final burial
(Figure 2-b; e.g., Forrest et al. 2003; Sanderson et al. 2001).
Additionally, bioturbation could also affect De distributions
by exhuming some grains and causing them to be exposed to
sunlight and reset, or by moving down profile younger grains
producing an age distribution with a young age tail (Figure 2-c).
In the latter scenario, an OSL-derived chronology could show
a deceptively reassuring increasing age with depth (reflecting
primarily the decreasing intensity of bioturbation with depth).
In extreme circumstances, rapid sediment turnover may lead to
exhumation, exposure to sunlight, and resetting of sand grains
followed by reburial leading to zero aged grains being found
at depth. For a more detailed explanation of the methodology
employed and discussion of the underlying principles of using
OSL to detect bioturbation, readers are referred to Bateman et
al. (2007a; 2007b).




2007 VOL. 60(2-3)

Age (years)

- C COAge (years) O o
Age (years)





Age (years)


Age (years)

V V V to t







AgeN N N N N (years)
Age (years)

Age (year
Age (years)

Figure 2. The age distributions generated by running multiple OSL replicates of individual samples at both single grain
and standard aliquot levels. (a) Agreement between single grain and standard aliquot measurements combined with a tight
normal distribution indicative of sediment fully exposed to sunlight prior to burial and undisturbed since burial; (b) Single
grain measurements showing much younger ages than standard aliquot measurements and a distribution with significant
tail of much older ages indicative of either incomplete sunlight exposure prior to burial or sub-surface mixing of older sand
grains into younger sediment. (c) Single grain measurements showing much younger ages than standard aliquot measure-
ments and a distribution with significant number of modern grains as well as some older grains indicative of modern pedo-
genic disturbance as well as either incomplete sunlight exposure prior to burial or sub-surface mixing of older sand grains
into younger sediment.


Table 1. Radiocarbon Dates.
Site and Sample # Material
"Ebersbach Midden North"



Conventional Radiocarbon

TU2 10-15 cmbs

TU2 15-20 cmbs

TU2 55-60 cmbs

TU2 75-80 cmbs

bulk humic TU2 55-60

bulk humic 80-85 cmbs

TU2 15-20 cmbs

TU2 35-40 cmbs

TU2 55-60 cmbs

bulk humic TU1 30-32 cmbs

bone collagen TU2 20-30cmbs

TU2 30-40 cmbs

TU8 40 cmbs





TU14 10-20 cmbs upper midden

TU14 30-40 cmbs paleosol

TU14 60-70 cmbs sub-paleosol

In addition to adopting multiple replicate measurements
for each sample, OSL measurements for this project were also
conducted at two levels. Firstly, the OSL signal was measured
at the standard aliquot level in which the signal is collected
simultaneously from approximately 2,000 grains of sand.
While this allows for better measurement precision, even if
many replicate aliquots are measured, the true distribution of
ages within a sample may be masked, for only a single average
OSL signal from all 2,000 grains is recorded on each aliquot.
This is particularly significant in samples of heterogeneous
age (i.e., sands that have been bioturbated or only partially
exposed to sunlight prior to burial). This problem can be
obviated by measurement of the accumulated dose from
individual grains, as advocated by Leigh (2001) for the study
of sandy site formation processes. Thus, for this project, each
sample also underwent multiple replicate OSL measurements
at the single grain level.
Four sandy sites (8HG941, 8HG887, 8P06085, and
8P06091) were chosen for detailed OSL dating. OSL samples
were also collected from an additional site (8P06094) on the
margin of Lake Arbuckle owing to clear evidence of dynamic
sedimentation and depositional stratification that could serve
as a control sample for purposes of evaluating the results from
the sandy sites. In all, 26 OSL samples were dated from these
five sites.

Wooded Swamp-Marsh Sites

Riparian Wetlands

Two of these wetlands sites (8P05310 and 8P05311, the
East and West Lavender sites, respectively) were situated on
either side of a channel within a cascading karstic depression
system, and three (8HG937, 8HG938, and 8P05307) were
situated within small stream valleys that exhibited channelized
alluvial flow. The Holocene sediments associated with both of
these systems are similar (probably because they share flow
histories) and consist of an early phase of marl deposition
followed by a later period when siliceous, organic-rich
sediment was deposited. This basic stratigraphic sequence
was encountered at each site of this type investigated. For
purposes of this paper, discussion will focus on one site,
the West Lavender site (8P05311), which contained the
clearest expression regarding the cultural relationship to this
stratigraphic sequence. Fortuitously at this site, a paleosol was
encountered between two phases of marl accumulation. This
buried soil was associated with a stratigraphically discrete
prehistoric component dominated by sand-tempered plain
pottery. A series of three radiocarbon samples bracketing
this surface was taken.2 All radiocarbon dates discussed in
this report are presented in Table 1. The paleosol was dated

Beta 194938

Beta 194939

Beta 194940

Beta 194941

Beta 194942


midden deposit

midden deposit

banded deposits
below midden

water table

banded deposits
below midden

eastern beach ridge

grey sand

grey sand

light grey sand

sediment below
lithified carbonate

grey sand

soot from BG sherd

Fea. 2

Beta 194943

"Burnt Hammock"

Beta 194944

Beta 194945

Beta 194946

Beta 194953

Beta 199258

Beta 199261

"West Lavender"

Beta 195378

Beta 195379

Beta 195380

Beta 195381




1,010 50 B.P.

1,030 50 B.P.

2,000 50 B.P.

2,620 40 B.P.

2,200 40 B.P.

5,570 90 B.P.

200 50 B.P.

1,630 + 40 B.P.

3,310 50 B.P.

990 40 B.P.

90 30 B.P.

1,960 +40 B.P.

1,16040 B.P.

1,000 40 B.P.

1,840 40 B.P.

2,150 40 B.P.

A.D. 960-1180

A.D. 900-1050

110 B.C.-A.D. 100

830-780 B.C.

380-160 B.C.

4580-4240 B.C.

A.D. 1640-1710
A.D. 1720-1880
A.D. 1910-1950

A.D. 350-530

1700-1490 B.C.

A.D. 990-1160

A.D. 1680-1740
AD 1810-1930
40 B.C.-A.D.120

A.D. 780-980

A.D. 980-1060
A.D. 1080-1150

A.D. 80-250

360-280 B.C.
240-60 B.C.









Inidcations of Occupation Activity
West Lavender Site (8PO5311) TU3

Bone (g)
0 5 10 15 20 25 30 0

Magnetic Susceptibility
Total Ceramics (n) 10 'm'kg-' Calcium Carbonate (%)
5 10 15 20 0 5 10 15 20 25 30 0 10 20 30 40 50

Figure 3. Indications of cultural occupations, Test Unit 3, site 8P05311.

by charcoal obtained through flotation and returned a date
of 1,840+40 years B.P. (cal A.D. 80-250 Beta-195380). The
lowermost date, corresponding to the lower marl formation,
yielded a date of 2,15040 years B.P. (cal 360-280 B.C. or 240-
60 B.c.; Beta-195381). The uppermost component at the site
yielded an associated radiocarbon date of 1,000+40 years B.P.
(cal A.D. 980-1060 or cal A.D. 1080-1150; Beta 195379). This
upper date was associated primarily with Sandy St. Johns and
Belle Glade plain ceramics, a dense amount of faunal material,
and almost no lithic artifacts. Additionally, a charcoal sample
taken from a pit feature associated with this upper component
yielded a date of 1,16040 years B.P. (cal A.D. 780-980; Beta
195378). Thus, it appears that at this site there is an upper
component corresponding to Belle Glade Period III (Austin
1996; Sears 1982) and a lower component corresponding to
Belle Glade Period I, separated by a culturally sterile period
represented (at least in the southern part of the site) by the
upper marl stratum. Toward the northern (downstream)
portions of the site, this upper marl appeared to become fused
to the younger, organic-rich siliceous sediments, but a lower
mode consisting of almost exclusively sand-tempered plain
ceramics was still evident. Figure 3 depicts the relationship of
these soil strata, the carbon dates obtained, and the distribution
of artifacts within Test Unit 3 at site 8P053 11.3
The presence of marl or freshwater calcium carbonate mud
has been the subject of considerable discussion in the geologic
literature of South Florida (cf. Browder et al. 1994; Gaiser et
al. 2005; Gleason 1972; Gleason and Stone 1994; Swift 1981),
but beyond the Everglades, its existence is poorly recognized
even today. In the Everglades, Gleason and Stone (1994:187)
identified one prominent period of regional marl deposition

where a marl bracketed by peat deposition was dated to the
interval between approximately 3,200 and 2,000 radiocarbon
years B.P. They also note that modern marls occur in well-
lighted, sparsely vegetated areas with flooding of 6-9 months
per year. If the depositional settings reported for marls in the
Everglades can be applied to the current study area, perhaps
the most important implication for past environments is that
these small stream valleys in the flats of the Osceola Plain
may have been shallow water, grassy, wet prairies prior to ca.
2,000 radiocarbon years B.P. and that a forest cover developed
in the last two millennia. Pollen data from Lake Annie (Watts
1975) records the first appearance of bald/pond cypress at ca.
3,000 radiocarbon years B.P., and this has been interpreted as
representing the northward migration of hardwood swamp
environments from the Everglades of south Florida (Griffin
2002:44). It is the establishment of this environment that
appears to usher in a significant change in cultural adaptations
in this region. Prior to this time, and based partially on the
underlying marl encountered in all the swamp margin sites
investigated, the Kissimmee region was likely characterized
by largely open prairies that were inundated 6-9 months out of
the year and which is represented geomorphically by the marl
stratum. Evidence from the current sample of sites indicates a
shift from marl deposition to siliceous sediment at ca. 2,000
radiocarbon years B.P. The timing of this shift appears to
correlate well with an increase in the overall number of sites
and a concomitant shift in their locations across the landscape.
Figure 4 depicts the chronology of local marl deposition based
on radiocarbon dates and diagnostic artifacts recovered from
the recent investigations.



2007 VOL. 60(2-3)



West Lavender (8P05311, Tu3)

Psychotria Nervosa (8HG938, Tul)

Small Psychotria (8HG937) &

Burnt Hammock (8HG887)



j 1- I I7 I I I I I I I I I

-I -


Black, organic-rich, siliceous sediment

E Marl

Figure 4. Chronology of marl deposition is based on radiocarbon dates and diagnostic artifacts. The best dates available
are from the West Lavender site, where three radiocarbon dates bracket the period of marl deposition, and two discrete
marl beds separated by a paleosol are present. The chronology from Small Psychotria and the Burnt Hammock Site are
predicated on the fact that marl deposition precedes Belle Glade Plain ceramics, which are thought to appear in this region
around A.D. 1000. At several sites marl deposition appears to be temporally discrete (e.g., West Lavender and Psychotria
Nervosa) whereas at others the frequency of marl in the deposits gradually tapers out towards the modern surface (e.g.,
Small Psychotria and Eight Mile Hammock #2). It is not clear if this is an artifact of marl deposition or post-depositional
disturbance of the deposits. The deposition of black siliceous sediment that follows marl in these streams is assumed to con-
tinue to the present.

Lake Margin Beach Sites

It is difficult to envision significant sediment movement
occurring today within or along the margin of Lake Arbuckle
with its prominent, fringing, bald cypress swamp, but
examination of the deposits around the eastern shore of this
lake at the Ebersbach Midden North site (8P06094) identified
two lacustrine beach ridges that were created by wave action
and subsequently became the loci of prehistoric habitation.
These sandy deposits are encased within low-energy palustrine-
lacustrine muds and mucks, and all of these deposits are inset
below the upland margin (Figure 5).
The Older Beach: The higher and most discrete beach
is situated at the rear of the bald cypress swamp where the
vegetation changes to a hardwood swamp and is about 10
m wide, 1 m tall, and covered with a prominent sawtooth
palmetto understory (the "older beach ridge" on Figure 5).
Coring across this narrow ridge with a GeoProbe corer
revealed that a discrete body of coarse sand is the source of
the ridge. An OSL sample from the sand at its base yielded
an age of 4,270+170 years (Shfd04018), which suggests that
this beach began to form sometime around 2436-2096 B.C.
Table II presents a summary of OSL dating results for sites
discussed in this text. This beach sand rests on an organic-
rich marsh mud, the top of which was radiocarbon dated to
5,57090 years B.P. (cal 4580-4240 B.c.; Beta-194943). The
presence of this organic-rich lake margin deposit suggests that
Lake Arbuckle had risen to this elevation by that time.
The Younger Beach: A younger, discontinuous, and less
prominent beach ridge was observed to the west of the older
beach (toward the lake). This subtle rise consists of alternating
strata of sand and silty and/or more-organic-rich sediment that

reflect a complex depositional history that occurred between
approximately 3,000-1,000 radiocarbon years B.P. The age
of this beach is constrained by three radiocarbon dates and a
single OSL age obtained from Test Unit 2 at this site. A piece
of charcoal collected from a mucky deposit at a depth of 80
cm within this beach yielded a radiocarbon age of 2,620+40
years B.P. (cal 830-780 B.C.; Beta-194941), a bulk sediment
sample from a mucky deposit at 55-60 cm in depth yielded
a radiocarbon age of 2,200+40 years B.P. (cal 380-160 B.C.;
Beta-194942), and a piece of charcoal collected from the Belle
Glade midden at the top of the ridge (10-15 cm below the
surface) yielded a radiocarbon age of 1,01050 years B.P. (cal
A.D. 960-1180; Beta-194938). A standard aliquot OSL age of
2,830+120 years (B.C. 946-706; Shfd04019) was obtained from
a clean sand between the lower two radiocarbon dates, and
this age overlaps with the lower radiocarbon date. Prehistoric
occupation of this beach occurred late in its history, after it had
emerged as a topographic rise.
Previous test excavations at Avon Park Air Force Range
have examined sites in similar geologic context, specifically
the Barker site (8P01007) and Ebersbach Midden (8P01008)
(Austin 1996). Both of these sites are located on landforms
Austin calls "lake levees." In a landscape ofcascading wetlands,
lake levees are likely to be encountered on the downstream
side of lakes where the lake spills over or cascades into a marsh
or stream. However, the ridges encountered at this site and
those described by Austin (1996) are not lake levees because
they are located midway down the east side of the lake, and
well above the "cascade point" for Lake Arbuckle, which is
where it spills into Arbuckle Creek and Arbuckle marsh at the
southern end of the lake.






SArbuckle Younger Beach

110 B.C. A.D. 100 A.D. 960-1180
(Beta 194940) (Beta 194938)

946-376 B.C. A.D. 900-1050
(Shfd04019) TU2 (Beta 194939)
830-780 B.C.
o, I83 -70,, 1, ,9 l i "" ."


Older Beach


2436-1636 B.C.

l Beach Sands
* Lacustrine mud and muck

* Bhs Horizons

Figure 5. Schematic illustration of the geologic deposits observed along the eastern shore of Lake Arbuckle at the Ebersbach
Midden North site (8P06094). The older beach ridge was comprised of a single massive, coarse sand bed that was deposited
around 4000 years, on top of a marsh mud that dated to 5570 years BP. The younger beach is comprised of interbedded sand
and muddy deposits which formed in the 2000 year period between ca. 3000 to 1000 years BP. Horizontal is not to scale,
and vertical scale of the different surfaces is estimated. The thickness of stratigraphic units is to scale based on relation-
ships observed in cores and test units. The dates shown are 2 sigma calibrated radiocarbon ages, and the OSL dates are the
minimum and maximum of the age defined by the one sigma single grain and standard aliquot OSL ages. The numbered
vertical bars on the right side of the figure are the Geoprobe cores.

Table 2. Summary of OSL dating results for site discussed in the text.

Depth Standard Single
Aliquot De* Grain De**
(cm) (Gy) (Gy)

Dose rate'

Aliquot Age

Standard Aliquot
Age (BC/AD)

"Ebersbach Midden North"

1.37 0.03 1.19 0.1 322 11 4270 170
0.74 0.01 0.70 0.06 365 10 2830 120

BC 2436-2096 3850 210 BC 2056 1636
BC 946-706 2560 180 BC 736 376

" Burnt Hammock"

Albritton Quarry, Avon Park

0.48 0.02
2.03 0.04
8.68 0.191


0.17 0.04 498 17
0.98 0.07 519 18
4.150.2 459 16

960 50

AD 994-1094

3910 150 BC 2056-1756
18920 770 BC 17,686-16,146

340 80
1890 150
9050 720

- 65337 211,17015,580 BC 224,746-193,586
- 53927 266,81018,510 BC 283,316-246,296

Mean De weighted by variance after outlier De values removed.
SMean De derived from common age model after outlier De values removed.
STotal Dose is attenuated for both grain size, density and moisture.
Note: Ages in bold thought to be closest to true depositional ages out of SA and SG measurements

Lab Code

Single Grain

Single Grain

AD 1584- 1784
BC 264 AD 36
BC 7766 6326


2007 VOL. 60(2-3)


The lithologic sequence and cultural use history
documented at the Ebersbach Midden North site correlate
remarkably well with the stratigraphy reported by Austin
(1996) at the Barker site (8P01007), which is situated on a
similar ridge to the south. There, Austin radiocarbon-dated a
sand between two mucky deposits at depth to 2,430+60 years
B.P. (cal 759-683 B.c. and cal 670-400 B.C.; Beta-76960) and
obtained a radiocarbon age of 85090 years B.P. (cal 1017-
1291 A.D.; Beta 76959) from a Belle Glade occupation near the
top of the profile.
Implications of Beach Ridges: In general terms, the
deposits revealed by these profiles suggest that the lakeshore
plain deposits of Lake Arbuckle appear to hold a distinct
sedimentary record that probably ties closely with periods of
extreme rainfall and/or wind events. Three basic deposit types
were noted in the profiles: (1) sands, (2) mucks and mucky
sands, and (3) anthrogenic sediments.
The sand deposits are likely to be storm deposits because
only during storms will there be sufficient energy to transport
sediment of this size. Some sand mobilization will undoubtedly
occur in proximity to stream mouths in association with major
precipitation events. Nevertheless, the interbedded sand-
muck deposition observed here is likely, at least in part, the
result of longshore transport of sandy sediments by wave
action. Beach ridge construction presumably occurred during
periods of extreme wind events such as hurricanes. The mucks
are most likely the product of normal conditions at the lake,
and represent periods of low depositional energy and gentle
accumulation of organic detritus or gyttja-like sediments.
The interbedded muck and sand deposits present here are
interpreted as brief periods of high-energy conditions and sand
deposition separated by long periods of low-energy conditions
and predominantly organic sedimentation. Specifically, these
deposits indicate that Lake Arbuckle has experienced at least
six major periods of high-energy conditions in the last 4,500
years, and this is not likely to be a complete record given
that it is based on just two exposures. The oldest phase of
high-energy conditions is the eastern beach ridge that has
been OSL-dated to cal B.c. 2436-2096. The muck beneath
it presumably represents organic accumulation at the lake
margin during the early phase of lake level rise following the
establishment of nearly modern sea level and the concomitant
rise of the water table across the Florida peninsula. The second
phase postdates cal B.c. 2436-2096 (Shfd04018) and predates
cal 830-780 B.C. (Beta-194941) and is probably closer to the
latter. There may be more than one high-energy event in this
time period given that we did not reach the base of the younger
beach ridge deposits in Test Unit 2. The third probably closely
follows cal 830-780 B.C., given it is OSL-dated to 946-706
B.C. (Shfd04019), which completely overlaps the underlying
age at 2-standard deviations. The fourth high-energy event is
loosely placed between cal 110 B.C.-A.D. 100 (Beta-194940)
and cal A.D. 900-1050 (Beta-194939), but the presence of a
weak soil formed in this sand favors a deposition event closer
to the older age. The fifth phase of sand deposition appears to
either predate or be concomitant with occupation of the site,
around A.D. 960-1180 (Beta-194938), and a final sixth phase
postdates occupation of the site.

General Observations on Wetland Sites

Although site 8P06097 was unique in our sample due to
its location along a lake margin, it shares many general traits
with the "swamp-midden" site types. At all of these sites,
the faunal assemblage was dominated by aquatic species,
and fresh-water turtle was one of the most important staples,
far more prominent than deer or any other mammal. While
there were important exceptions (including a large amount of
mussel from site 8PO6097, indicating a specialized adaptation
to the lake margin environment), the overall faunal assemblage
consisted primarily of highly fragmented pieces of bone.
Furthermore, although the frequency of burned bone was not
high (and was at least partially attributable to tool production),
the overwhelming majority of bone was heated, most likely
through boiling. Nearly all the faunal assemblage displayed
an almost complete lack of collagen (which was discovered
during submittal of bone samples for collagen dating). Boiling
is perhaps the most effective method of removing collagen
from bone (Darden Hood, personal communication 2004). At
site 8HG887, a large fragment of a Belle Glade plain vessel
was recovered. Based on curvature measurements of the rim
and interior, the vessel appears to have been a large, globular
bowl that would have measured 42.52 cm (16.74 inches) in
diameter and have been capable of containing a volume of
20.12 liters (5.32 U.S. gallons). Given the faunal assemblage,
the soot encrustation along the rim, and the form of the vessel,
it appears reasonable to suggest that vessels such as this were
used to cook "gumbo-" like meals composed primarily of
aquatic species obtained from the nearby swamps and wetlands.
This cooking method would account for the characteristics of
the faunal assemblage, in that the bone was heated but not
directly burned. A sample of the soot on this particular vessel
fragment was submitted for radiometric assay and yielded
a date of cal 1,96040 years B.P. (40 B.C.-A.D. 120; Beta-
199261). This date corresponds to a late Belle Glade Period I
timeframe and indicates an earlier introduction of this ceramic
type into the region than has been previously documented (cf.
Austin 1996). This date is also intriguingly close to the general
timeframe posited here for the establishment of hardwood
swamp environments in this region. Additional dates from
soot-encrusted sherds could provide clues to the introduction
and expansion of ceramic technology into the Kissimmee
region and its potential relationship to these environmental
Another pattern that was noted during excavations at these
sites was the near total lack of lithic materials. Furthermore,
when such materials were encountered, they occurred almost
invariably near the base of the midden deposits. Since the
geomorphic work has indicated relatively stable depositional
environments at these settings, it is not believed that this pattern
is the result ofpostdepositional, differential settling of artifacts.
Rather, this pattern is tentatively interpreted as further evidence
of a wetland-oriented subsistence regime. These environments
would have supplied not only an easily exploitable source
of protein, but also readily obtainable raw materials for tool
manufacture-namely, the hardwoods themselves. Prior
to the establishment of this environment, lithic materials




were obtained primarily from the Tampa Bay region based
on provenance determination (Wilder and Frederick 2006).
However, once the hardwood swamps became established, a
locally available tool source emerged, obviating the need to
travel long distances to acquire suitable raw materials for tool
production. If our chronological interpretations regarding the
cessation of marl formation in this region and its relationship to
paleoenvironmental changes are at all accurate, then both the
faunal and hardwood resources associated with these swamp-
margin habitats would have become available only sometime
after ca. 2,000 years B.P.
It should be recognized that the timeline presented in
Figure 4 is based on our limited sample of sites. Additional
data are necessary in order to refine, confirm, or refute this
sequence. Particularly important would be additional dates
from soot-encrusted pottery vessels as well as more pollen
data. If the inferences presented here are correct, then the
majority of ceramic-bearing archaeological sites should
postdate the cessation of marl formation in the region and
conform chronologically to the establishment of hardwood
swamp habitats.

Sites in Sandy Upland Settings

The other major site type investigated during the recent
testing can be called the "sandy site." These sites consist
primarily (but not exclusively) of lithic scatters situated in
sandy upland deposits. Sites such as these are common in the
southeastern U.S. from theAtlantic coast to East Texas, and they
tend to be problematic because they typically have poor organic
matter preservation, lack discernible depositional stratigraphy,
and what strata are present are typically postdepositional
pedogenic features (soil horizons). At APAFR, sites of this
type often exhibit apparent cultural stratigraphy in the form
of reasonably discrete artifact concentrations at depth that
could often be separated into lithic and ceramic components at
different depths (Figure 6). This phenomenon is not restricted
to APAFR, but is a common attribute of sandy sites across the
peninsula. Gagel (1981:100) has noted that it is often assumed
that artifact groups at similar depths were deposited at the
same time throughout a site even though independent data to
support this assumption are lacking.
Evaluating the formation processes responsible for
burying archaeological materials in this landscape context was
one of our major project goals. Many of the sites occupying the
flat geomorphic surfaces that dominate this region of Florida
are buried but lack any observable depositional stratigraphy,
cultural features, or other evidence ofstratigraphic or contextual
integrity. The interpretation of such site assemblages is often
highly contingent upon the presumed burial mechanism,
and for flat surfaces two processes are widely cited: eolian
deposition and pedoturbation.

Eolian Deposition

In a landscape mantled by fine sand, the possibility
of eolian deposition is often assumed, especially in the
archaeological literature where sites are typically found buried

and not resting on the surface. Although there are a number
of clear areas of inland eolian dunes (e.g., Scott et al. 2001;
White 1958:38-43, 1970:151) many of which are on Lake
Wales Ridge, these features are currently thought to be relict
Pleistocene landforms and not active during the Holocene
(for dated examples see Ivester and Leigh 2003; Ivester et al.
2001; Otvos 2003; Otvos and Price 2001). We collected two
OSL samples from the core of Lake Wales Ridge at Albritton
Quarry, on the northeast side of Avon Park; these samples
returned single aliquot ages of 211,000 16,000 (Shfd04048)
years and 267,000 19,000 years (Shfd04049) and support
geologic inferences concerning the age of this structure.
Away from these relict eolian landforms, the geologic
origin of the near-surface sands is the subject of considerable
debate among archaeologists. Numerous archaeological studies
in Florida have inferred site burial by eolian sedimentation for
Archaic and Paleoindian occupations (e.g., West Williams site
[Austin et al. 2004]; Harney Flats [Daniel and Wisenbaker
1983]; Silver Springs [Bullen 1975; Hemmings 1975]; and the
Colorado site [Horvath 2000]), a point also argued in general
terms by Bullen (1975:81) and in specific by Hemmings and
Kohler (1974) who suggest that eolian sand was mobilized
between 7,000 and 4,000 radiocarbon years B.P. Geological
studies map them as "undifferentiated surficial sands";
furthermore, Campbell (1986), Scott (2001:23), and Scott
et al. (2001) carefully avoided any processual connotation
with respect to the sands when describing this map unit as a
"varying thickness of undifferentiated sediments consisting of
siliciclastics, organic and freshwater carbonates." Hence, the
geologic origin of the sands is unknown, and the potential for
Holocene mobilization by wind is a possible but unconfirmed


An alternative mechanism of site burial is pedoturbation,
or the mixing of a sediment after deposition. The exploration
and documentation of the mechanics and effects of disturbance
processes on archaeological assemblages has been a common
theme in the archaeological literature since a seminal work by
Wood and Johnson (1978) that summarized a broad suite of
processes applicable to archaeological site formation. Since
then, formation process studies have appeared routinely,
examining pedoturbation by various animals (e.g., large
animals [Araugo and Marcelino 2003; Butler 1995; Voslamber
and Veen 1985], pocket gophers [Bocek 1986, 1992; Erlandson
1984; Gabet 2000; Kalisz and Stone 1984, Pierce 1992], insects
[Brussaard and Runia 1984; de Bruyn and Conacher 1994;
Eldridge and Pickard 1994; Kalisz and Stone 1984; McBrearty
1990; Nkem et al. 2000], and earthworms [Atkinson 1957;
Langmaid 1964; Stein 1983; van Nest 2002]) and plants (e.g.,
treethrows, Schaetzl and Folmer 1990). More comprehensive
summaries by Balek (2002), Johnson and Wattson-Stegner
(1987), and Schiffer (1987) cover a wide range of processes and
their archaeological implications, whereas recent summaries
by Johnson (2002) and Johnson et al. (2005a, 2005b) examine
how these processes may influence the development of soils
over long periods of time.

2007 VOL. 60(2-3)



Sandy Point Hammock

(8HG941) TU2

Arbuckle Terrace


Artifacts (count)
0 10 20 30 40 50 60 70


20 -

40 -

60 -







Lithics (count)
0 2 4 6 8

Lithics (count)
0 2 4 6 8

Figure 6. Examples of artifact counts from various test pits excavated on sandy sites.

In the Southeast, two authors have examined the issue
of site burial by pedoturbation in detail. Michie (1990:44),
summarizing excavation results from sites on the South Carolina
coastal plain, argued that burial by pedoturbation was common
and resulted in fairly consistent age-depth relationships. The
specific issue of how archaeological sites may be buried in
sandy sediments has been examined in detail by Leigh (1998,
2001). In these papers, Leigh notes that the interpretation of
artifacts buried within sandy deposits suffers from problems
of equifinality because pedoturbation and sedimentation can
both result in buried cultural deposits, and that distinguishing
which process is responsible is difficult on the basis of field
evidence alone. Demonstrating that pedoturbation rather than
sedimentation has buried an assemblage of artifacts requires

careful consideration of the geomorphic setting of the site,
the sedimentary structures/stratigraphy/soils, the particle size
distribution, distribution and integrity of cultural material and
features, soil micromorphology and dating techniques.
Key attributes of pedoturbation identified by Leigh
include the lack of sedimentary structures or other stratigraphic
features indicative of burial (e.g., buried soils or lithologic
discontinuities). The destruction of sedimentary structures
by postdepositional processes has been widely examined by
geologists working in near-shore marine environments, and
the factors involved in determining whether sedimentary
structures are preserved has been likened to a race between the
sedimentation rate and the bioturbation rate, conditioned by
the thickness of the sedimentary beds and the thickness of the




8HG887 Burnt Hammock Site

Test Unit 2

OSL Dates Stratigraphy
SA=Standard Aliquot Zone
SG=Single Grain

SA 960+50
SG 340+80

SA 3,910150
SG 1,890+150

SA18,920770 O
SG 9,050+720

Ages (years BP)

20050 BP
(sela 19444)
90-+30 BP
- 196040 BP
(810- 199259)
163040 BP
(B-t 194945)
331050 BP
(Beta 14946)

Particle Size Distribution (%) Mean/Sorting (phi units)
0 20 40 60 80 100 0 1 2 3 4 5 6
10 .Mean
--- Sorting

8P06091 Arbuckle Terrace Site

Test Unit 1

OSL Dates
SA=Standard Aliquot
SG=Single Grain
SA 3,610+260
SG 1,320+180
SA 8,720+500
SG 4,300360
SA 11,870790
SG 5,930470
SA 17,700940
SG 7,390500
SA 27,290+1500
SG 6,610+430

SA 49,400+2300
SG 10,920840

SA 145,700+6700
SG 18,900+2100


O Shfd04031 2

O Shfd04030 3
O Shfd04029

0 Shfd04028

0 Shfd04027

0 Shfd04026

0 Shfd04025 4

Particle Size Distribution (%)

Mean/Sorting (phi units) Organic Carbon (%)
0 1 2 3 4 5 6 00 0.2 0.4 0.6 0.8 10

O OSL Dating Sample
+ Radiocarbon Dating Sherd Soot Sample
I Radiocarbon Dating Charcoal Sample
* Radiocarbon Dating Bone Sample

Figure 7. Plot of the stratigraphy, granulometric data, and organic carbon content for two of the four sandy sites examined
during this study. Note the lack of variation in the particle size with depth and the subtle difference in mean particle size
between the two sites. It is also important to note that the scales for the organic carbon content are different for each site.

Organic Carbon (%)
0.0 1.0 2.0 3.0 4.0


2007 VOL. 60(2-3)

3.00 4 Sandy Sites
S 0 Willingham Branch TU2
.2 2.50 E Arbuckle Terrace TU1
m5 Arbuckle Terrace TU10
0 2.00 x Sandy Point Hammock TU1
2 x Burnt Hammock TU2
M 1.50 +
C Late Holocene Littoral Sediments
u +Ebersbach Midden TU2
u 1 Ebersbach Midden Beach

o 0.50 4

0.00 1.00 2.00 3.00 4.00 5.00 6.00
graphic mean (phi)

Figure 8. Particle Size Data for the sandy sites vs. Ebersbach midden north.

mixing layer (Wheatcroft and Drake 2003:123). They note that
these processes can best be resolved in areas that experience
episodic sedimentation

General Observations

Four sites of this type were studied in detail: the Burnt
Hammock site (8HG887), the Sandy Point Hammock site
(8HG941), the Willingham Branch site (8P06085), and the
Arbuckle Terrace site (8PO6091). Only one of these sites is
discussed in detail here. For a more comprehensive discussion
of the results, the reader is referred to Wilder and Frederick
None of the deposits at the sandy sites exhibited evidence
of depositional stratification, but many exhibited apparent
archaeological stratification. However, where temporally
diagnostic artifacts were present, there was some overlap
of different-aged artifacts, as exhibited most clearly at site
The granulometric data demonstrated that the sandy sites
are texturally homogeneous and exhibit little down profile
variation in particle size except within spodic (Bhs) horizons,
which were slightly finer-textured owing to postdepositional
alteration (Figure 7). It was noted that the mean particle size
varied slightly from one sandy site to the next. All of the
remaining sites examined that had evidence of late Holocene
sedimentation exhibited a much wider range of particle size
variation (Figure 8).

Several trends emerged from the OSL dating results that
hold implications for the burial processes operative at the
sandy sites. The most prominent trend was an age discrepancy
that occurred between the standard aliquot and the single
grain ages that increases in magnitude with depth (see OSL
ages on Figure 9). This is caused by a progressive increase
in the number of ancient sand grains in the dated population
(scenario shown in Figure 2-b). Moreover, not only does the
number of these grains increase with depth, but the age of
these grains increases as well. As a result, the standard aliquot
OSL dating method, which averages the luminescence signals
from ca. 2,000 grains, yields spurious ages that do not reflect
the time of transportation or burial. Interestingly, however, the
single grain ages from these samples, where the ages could
be directly compared with other dating results (either relative
ages of diagnostic artifacts or radiocarbon dates) often were
statistically similar to or only slightly older than the dated
material. From this, we infer that the single grain ages are close
to or slightly older than the cultural material present. Finally,
at sites with clear evidence of geological deposition (like the
younger beach at the Ebersbach Midden North site), the single
grain and standard aliquot ages were nearly identical and did
not exhibit the age discrepancy observed in the sandy upland
sites. Hence, we conclude that normal depositional processes
yield reasonable OSL ages, and that the sandy upland sites are
not the result of normal deposition, but rather are artifacts of




Ceramic Types by Level

Burnt Hammock Site,TU2

Age (years B.P.)

0 z ~ C oGo0 0
C~G b (0 0 @~ O0




* Radiocarbon date, calibrated
* Single Grain OSL date -i Pairs from a
* Standard Aliquot OSL date single sample

Figure 9. Plot of diagnostic artifact depth distribution and results of dating from Test Unit 2 at the Burnt Hammock Site.

The Burnt Hammock Site (8HG887)

The Burnt Hammock site proved particularly informative
owing to the presence of a fairly recent historic component, a
prehistoric ceramic and bone component, and a lower mode
of artifacts consisting solely of lithics and bone. This situation
allowed us to compare artifact distributions and depths from
(presumably) discrete time periods and to obtain independent
ages from these components as well as the matrix within
which they occurred. In order to date the sand matrix, trace
charcoal obtained through flotation samples excavated in 5-
cm levels was used as well as dating with the OSL method.
Attempts to determine the chronology of the prehistoric
assemblage by radiocarbon dating were in part thwarted due
to a lack of datable collagen in the bone assemblage, but the
resulting additional radiocarbon dates demonstrate that the
chronological stratification may be more apparent than real.
Three OSL samples and five radiocarbon dates were
obtained from Test Unit 2 at site 8HG887. The OSL samples
were collected at 20-, 35-, and 80-cm depths, while the
radiocarbon samples were collected at 15-20 cm, 35-40 cm,
55-60 cm, and from soot scraped off a Belle Glade plain
rim sherd recovered at 30-40 cm below surface. The fifth
radiocarbon date was derived from bone collagen of historic-
era domesticated pig remains recovered between 10-20 cm

below surface (the historic fauna was unburned and was
the only bone retaining sufficient collagen to date from this
assemblage). Figure 9 displays the depth distribution of
diagnostic artifacts from this test unit with respect to the dating
results. The top two radiocarbon samples both yielded ages
consistent with the Historic period (20050 years B.P. [cal A.D.
1640-1710, and/or cal A.D. 1720-1880, and/or cal A.D. 1910-
1950; Beta-194944]; and 9030 years B.P. [A.D. 1680-1740,
and/or A.D. 1810-1930; Beta-199258]) and are at or slightly
below the mode of the historic artifact distribution, which was
in Level 2. The OSL sample from 20 cm (Shfd04016) yielded
a single grain date age of 34080 years (A.D. 1584-1744) and
a standard aliquot age of 96050 years (A.D. 944-1094). These
ages must be considered in light of the fact that the mode of the
Historic artifact distribution was at a depth of only 15 cm. The
fact that about 28 percent of the grains measured at the single
grain level were zero-age grains (i.e., grains that had recently
been exposed to sunlight and reburied at this depth) suggests
that the Historic artifacts are still, at this depth, being affected
by pedoturbation. The shift in OSL age when using the single
grain rather than the standard aliquot data, combined with data
showing many of the grains dated by OSL are older than the
artifacts, indicates that pedoturbation has buried the historic
artifacts by moving up older-aged sand grains to the surface.





5. 40







2007 VOL. 60(2-3)


The next lower OSL age at 35 cm (Shfd04015) yielded
a single grain age of 1,890+150 years (264 B.C.-A.D. 36) and
a standard aliquot age of 3,910+150 years (B.c. 2056-1756).
This is directly comparable to a radiocarbon date on a Belle
Glade Plain sherd collected from 30-40 cm that returned
an age of 1,96040 years B.P. (cal 40 B.C.-A.D. 120; Beta-
199261) and a charcoal sample from 35-40 cm that dated
to 1,63040 years B.P. (cal A.D. 350-530; Beta-194945). The
date obtained from the sherd is probably the most reliable in
terms of dating the actual cultural assemblage at this level,
despite being a bit older than previously documented Belle
Glade plain vessels in this region. The fact that the single grain
OSL age and the sherd soot radiocarbon date overlap well at
2 standard deviations indicates that the majority of the sand
surrounding the cultural material was reset about the time the
site was occupied; however, the much older standard aliquot
age indicates that part of the sand in the matrix is considerably
older than the occupation, which is due to the postdepositional
intermixing of sand grains of different age. If the assumption
that the radiocarbon-dated sherd was part of a cluster of sherds
belonging to one vessel is correct and these items are moving
down profile but are intermixing with older sand, then they
seem to be moving downward more or less en masse. The lack
of zero-age grains indicates that at this depth, the sediments
and artifacts are no longer being significantly affected by
present-day pedoturbation.
The charcoal sample collected from 55-60 cm yielded an
age of 3,31050 years B.P. (cal 1700-1490 B.c.; Beta-194946)
and was located in the level immediately below that which
yielded sand and fiber tempered sherds. This age is younger
than the standard aliquot age at 35 cm, but much younger than
the OSL ages from 90 cm depth (Shfd04014), which yielded
a single grain age of 9,050720 years (7766-6326 B.c.) and
a standard aliquot age of 18,920770 years (17,686-16,146
The plotting of diagnostic artifacts from Test Unit 2 as a
percentage with respect to depth indicates four different modes
in the following order: Historic artifacts in Level 2, Sandy
St. Johns Plain in Level 3, Belle Glade Plain in Level 4, and
sand and fiber tempered in Level 5 (see Figure 9). This pattern
has the appearance of four different age groups of artifacts
arranged from youngest at the top to oldest at depth, although
it is not clear whether some of these ceramic wares are actually
contemporaneous. What is clear is that Historic artifacts are
found in the same levels as late Belle Glade artifacts and
that the burial and dispersion of the Historic assemblage has
occurred in less than a century. The site area consists of old-
growth hardwoods with no evidence of plowing present, so
some other mechanism is responsible for this pattern of artifact

Inference ofBurial Processes

The results of the OSL dating at the Burnt Hammock site
and other sandy upland sites clearly point toward pedoturbation
as the burial process in upland settings. If these sites were
buried by normal sedimentation, then the two forms of OSL
dating would consistently yield similar results regardless

of the depth below surface, as is the case at the Ebersbach
Midden North site. Instead, they show a progressive age
discrepancy that increases with depth below surface. If eolian
deposition was responsible for burial of these sites, then we
would expect the single grain ages of the sand to match the
age of the artifacts it surrounds, but, in the case of the Burnt
Hammock site, the Historic artifacts (estimated to be about
100 years old) are surrounded by sand that yielded a OSL 1
sigma age of 340 years. Furthermore, if there has been 20 cm
of Historic-age eolian sedimentation, we would expect to find
it in the immediately adjacent lowland segment of this site
as well, but examination of those deposits did not reveal any
deposit of this type.
Considered together, the results from the sandy site
investigations suggest that these types of sites on the flats
at APAFR are being buried by postdepositional bioturbation
rather than geologic sedimentation and that the soil within
which the archaeological remains are found is effectively a
biomantle (Johnson et al. 2005, in press). The dating results and
the artifact distributions suggest two processes are operative:
(1) the exhumation of sand (but not artifacts) from depth, most
likely by insects, and (2) mixing and dispersion of the artifacts
in the near-surface by the burrowing of small animals. If the
sites were being buried by bioexhumation of sediment alone,
then this would result in stratigraphic separation of different-
age cultural components and, although not "normal geological
sedimentation," it would effectively result in discrete buried
artifact assemblages. However, as was clear at the Burnt
Hammock site, when the occupations are close together in time,
there is considerable mixing of different-age components, and
this can only occur if something is moving sand and artifacts in
the near-surface environment (approximately the top 30 cm).
The extent to which these processes adversely affect
archaeological assemblages will depend upon how much
time separates different occupation events and the degree of
disturbance in the near-surface environment. Several of the
sites examined exhibited a deep, Archaic-age lithic component
that was stratigraphically discrete and separate from a
much shallower and younger ceramic period occupation.
At such sites, the bioexhumation process appears to favor
the stratigraphic separation of disparate age archaeological
components. However, what is clear from the Burnt Hammock
site is that what may appear to be single-component ancient
occupations may in fact be multiple different-age occupations
that have been mixed together by near-surface disturbance and
then progressively buried by bioexhumation. The fact that such
sites have been buried by bioturbation does not necessarily
mean that they have lost sufficient integrity to contribute new
and useful information, but rather that the limitations imposed
by the burial process should be carefully considered when
interpreting such assemblages.
Archaeological sites situated in deep, sandy deposits occur
throughout the eastern Gulf Coastal Plain, and archaeologists
have long wrestled with issues of contextual integrity and
research significance for these sites (cf. Austin 2002; Gagel
1981). Previous studies on APAFR (e.g., Headquarters Air
Combat Command 2001a, 2001b) have interpreted vertical
artifact distributions at sites in sandy contexts in terms of




an aggrading, depositional environment. Although this is
appropriate at some sites, the geoarchaeological investigations
reported herein have demonstrated that at least in some cases
stratigraphic positioning may be misleading. In order to draw
implications concerning past cultural behavior from these
material remains, issues of site formation/transformation and
the contextual integrity of the deposits must be addressed.
Without such fundamental knowledge, any statements on past
human behavior and culture derived from the excavation of
these materials are questionable at best.


Clearly, archaeological testing does not allow for the
attainment of major regional research goals; however, the
data gathered during these investigations have provided some
intriguing information that may prove useful in future regional
syntheses efforts. The results of this work identified three
geomorphic settings: dynamic lake margin environments,
Holocene alluvial/palustrine settings, and upland sandy sites.
Sites in the first two settings appear to have been buried by late
Holocene sedimentation, whereas site burial in the uplands
appears to be the result of pedoturbation. Sites in the first
two settings were generally wooded swamp-marsh habitation
middens that date to a postceramic timeframe, whereas the
sandy upland sites appear to date largely (but not exclusively)
to a preceramic timeframe.
Researchers are in general agreement that the Late
Pleistocene/Early Holocene was considerably more arid than
the present (c.f. Faught and Carter 1998; Watts et al. 1996)
and that sea level began to approach modern level around ca.
6,000 B.P. The close temporal correspondence between sea
level rise and changes in vegetation from open oak woodland
to a pine flatwood forest makes it difficult to disentangle the
effects of climate vs. sea level, but it is commonly agreed that
the period from about 6,000-3,000 B.P. is one characterized
by progressively higher water tables and more humid climatic
conditions (cf. Balsille and Donoghue 2004; Gleason and
Stone 1994; Watts et al. 1996:36; Winkler et al. 2002). If the
association postulated by Griffin (2002) between the influx
of bald/pond cypress pollens in the cores from Lake Annie
and the northern migration of hardwood swamps is accurate,
then the Osceola Plain may have been characterized by
open, seasonally wet prairies that favored the formation of
marl from about 6,000-2,000 B.P. If such were the case, then
upland (drier) areas would be the favored places for human
occupation in this region until the establishment of hardwood
swamps offered a new source of subsistence and sustenance.
At the present time, the above scenario can only be considered
a testable hypothesis. Further data are necessary in order to
support or refute this hypothesis, and it is hoped that future
studies in the region will shed further light on the nature and
timing of this environmental shift and cultural responses to it.
An additional important finding of these investigations
concerns the contextual integrity of deeply buried sites
within sandy soil matrices. Our studies have indicated that
the burial of cultural materials in such settings is largely
the result of postdepositional pedoturbation. While many of

these assemblages appear to be moving downward more or
less en masse, any associations of artifactual relationships
based purely on relative vertical positioning in these settings
should be made cautiously, if at all. These findings may well
be applicable to sites in similar contexts throughout the Gulf
and Atlantic Coastal plains.


1. See Wilder and Frederick 2006 for information on historic
and historic-military sites investigated.

2. All radiocarbon dates discussed in the text are reported
as calibrated years before present (cal B.P.) in order to permit
direct comparison between the radiocarbon and OSL ages.
Where more than one calibrated age range exists, the age
range cited in the text is the maximum and minimum values
of the calibrated age ranges. All radiocarbon dates, along with
other relevant data are presented on Table 1. All radiocarbon
ages were processed by Beta-Analytic using the accelerator
mass spectrometer (AMS) method, and calibrations were
performed using CALIB version 5 and the IntCal04 data set
of Reimer et al. (2004). Calibrated ages are presented at the
2-sigma confidence interval, which indicates a 95 percent
probability that the age of the material falls within the date
range presented.

3. Test Unit 3 was inspected and profiled in the field by the
project geomorphologist, while the carbon dates presented
were obtained from the same strata in a different test unit
(Test Unit 14). Both units contained an upper peak of artifacts
dominated by Sandy St. Johns plain ceramics and a lower
mode dominated by sand-tempered plain ceramics. This lower
mode was associated with a paleosol in both units and was
underlain by a marl stratum. The major difference in terms of
stratigraphy between the two is that the upper marl appears to
pinch out and become fused with the lower midden deposits
as one progresses northward (downstream) on the landform
that the site occupies. Due to the more detailed geomorphic
information that was recorded for Test Unit 3, the carbon dates
have been correlated and represented in relationship to the
artifact distribution and stratigraphy of Test Unit 3.


The authors would like to thank Bob Austin (SEARCH,
Inc.) and Rich Estabrook (FPAN) and Sylvia Scudder (Florida
Museum of Natural History) for sharing their time and
comments. We would also like to thank Dr. Andy Carr for
putting in long hours and generating reams of data that proved
crucial to understanding how these sandy sites have formed.

References Cited

Aitken, M. J.
1998 An Introduction to Optical Dating: The Dating of
Quaternary Sediments by the Use ofPhoto-Stimulated
Luminescence. Oxford Science Publication, Oxford.

2007 VOL. 60(2-3)



Araugo, A. G. M., and J. C. Marcelino
2003 The Role of Armadillos in the Movement of
Archaeological Materials: An Experimental
Approach. Geoarchaeology 18:433-460.

Atkinson, R. J. C.
1957 Worms and Weathering. Antiquity 31:219-233.

Austin, R. J.
1996 Ceramic Seriation, Radiocarbon Dates, and
Subsistence Data from the Kissimmee River Valley:
Archaeological Evidence for Belle Glade Occupation.
The Florida Anthropologist 49:65-87.

2002 Beyond Technology and Function: Evaluating the
Research Significance of Lithic Scatter Sites. In
Thinking About Significance, edited by R. J. Austin,
K. S. Hoffman, and G. R. Ballo, pp. 153-186. Special
Publication Series No. 1, Florida Archaeological
Council, Inc., Riverview, Florida.

Austin, R. J., B. E. Ensor, L. Carlson, J. Endonino, J. E. Foss,
D. H. Phillips, F. J. Rich, Y. Roh, D. L. Ruhl, and S. Scudder
2004 Multidisciplinary Investigations at West Williams,
8HI509: An Archaic Period Archaeological Site
Located Within Florida Gas Transmission Company s
Bayside Lateral Pipeline Corridor, Hillsborough
County, Florida. Southeastern Archaeological
Research, Inc., Jonesville, Florida.

Balek, C. L.
2002 Buried Artifacts in Stable Upland Sites and the Role
of Bioturbation: A Review. Geoarchaeology 17:41-

Balsillie, J. H., and J. F. Donoghue
2004 High Resolution Sea-Level History for the Gulf of
Mexico Since the Last Glacial Maximum. Report of
Investigations No. 103, Florida Geological Survey,

Bateman M. D., C. D. Frederick, M. K. Jaiswal, A. K.
2003 Investigations into the Potential Effects of
Pedoturbation on Luminescence Dating. Quaternary
Science Reviews 22:1169-1176.

Bateman, M. D., C. H. Boulter, A. S. Carr, C. D. Frederick, D.
Peter, and M. Wilder
2007a Preserving the Palaeoenvironmental Record in
Drylands: Bioturbation and Its Significance for
Luminescence Derived Chronologies. Sedimentary
Geology 195:5-19.

2007b Detecting Post-Depositional Sediment Disturbance
in Sandy Deposits Using Optical Luminescence.
Quaternary Geochronology 2:57-64.

Bocek, B.
1986 Rodent Ecology and Burrowing Behavior: Predicted
Effects on Archaeological Site Formation. American
Antiquity 51:589-603.

1992 The Jasper Ridge Reexcavation Experiment: Rates
of Artifact Mixing by Rodents. American Antiquity

Browder, J. A., P. J. Gleason, and D. R. Swift
1994 Periphyton in the Everglades: Spatial Variation,
Environmental Correlates and Ecological
Implications. In Everglades: The Ecosystem and its
Restoration, edited by S. M. Davis and J. C. Ogden,
pp. 379-418. St. Lucie Press, Delray Beach, Florida.

Brussaard, L., and L. T. Runia
1984 Recent and Ancient Traces of Scarab Beetle Activity
in Sandy Soils of the Netherlands. Geoderma 34:229-

Bullen, R. P.
1975 Implications from Some Florida Deposits and Their
Archaeological Contents. The Florida Anthropologist

Butler, D. R.
1995 Zoogeomorphology. Cambridge University Press,
New York.

Campbell, K. M.
1986 Geology of Polk County, Florida. Florida Geological
Survey. Florida Department of Natural Resources,

1990 Geology. In Soil Survey of Polk County, Florida, by
Richard D. Ford, J. M. Robbins Jr., J. T. Werner, Dean
Cowherd, Charles N. Gordon, W. B. Warmacck, M.
M. Brown, Kenneth Monroe, Walter G. George,
Therman Sanders, and Phyllis M. Basch, pp. 3-7.
Soil Conservation Service. United States Department
of Agriculture, Washington, D.C.

Choi, J., J. W. Harvey, and J. T. Newlin
2003 Significance of Microtopography as a Control on
Surface-Water Flow in Wetlands. Abstract of a poster
presented at the Joint Conference on the Science and
Restoration of the Greater Everglades and Florida
Bay Ecosystem. Available online at http://conference.

Daniel, R., and M. Wisenbaker
1983 A Preliminary Report on the Excavations at
Harney Flats, Hillsborough County. The Florida
Anthropologist 36(1-2):67-79.

de Bruyn, L. A. L., and A. J. Conacher
1994 The Bioturbation Activity of Ants in Agricultural




and Naturally Vegetated Habitats in Semiarid
Environments. Australian Journal of Soil Research

Eldridge, D. J., and J. Pickard
1994 Effects of Ants on Sandy Soils in Semiarid Eastern
Australia 2: Relocation of Nest Entrances and
Consequences for Bioturbation. Australian Journal
of Soil Research 32:323-333.

Erlandson, J. M.
1984 A Case Study in Faunalturbation: Delineating the
Effects of the Burrowing Pocket Gopher on the
Distribution of Archaeological Materials. American
Antiquity 49:785-790.

Faught, M., and B. Carter
1998 Early Human Occupation and Environmental Change
in Northwestern Florida. Quaternar, International

Frederick, C. D., and M. D. Bateman
1998 The Potential Application of Optical Dating to the
Sandy Uplands of East Texas and Northwestern
Louisiana. Journal of Northeast Texas Archeology

Frederick, C. D., M. D. Bateman, and P. Lehman
2001 Geoarchaeological Investigations. In National
Register Eligibility Testing at 41LE177, Alcoa
Sandow Mine, Lee County, Texas: Archaeological,
Geoarchaeological and Paleoenvironmental
Assessment of an Upland Sandy Mantle Site, edited
by R. Ricklis, pp. 83-123. Coastal Archaeological
Research, Inc., Corpus Christi, Texas.

Frederick, C. D., M. D. Bateman, and R. Rogers
2002 Evidence for Eolian Deposition in the Sandy Uplands
of East Texas and the Implications for Archaeological
Site Integrity. Geoarchaeology, 17:191-217.

Forrest, B., W. J. Rink, N. Bicho, and C. R. Ferring
2003 OSL Ages and Possible Bioturbation Signals at the
Upper Paleolithic Site of Lagoa do Bordoal, Algarve,
Portugal. Quaternary Science Reviews 22:1279-

Gabet, E. J.
2000 Gopher Bioturbation: Field Evidence for Non-Linear
Hillslope Diffusion. Earth Surface Processes and
Landforms 25:1419-1428.

Gagel, K.
1981 Archaeological Excavations at Site 8Hi483B:
An Archaic Habitation Site in Hillsborough
County, Florida. Interstate 75 Highway Phase II
Archaeological Reports, Number 6. Bureau of
Historic Sites and Properties, Florida Division of

Archives, History and Records Management, Florida
Department of State.

Galbraith, R. F, R. G. Roberts, G. M. Laslett, H. Yoshida, J.
M. Olley
1999 Optical Dating of Single and Multiple Grains of
Quartz from Jinmium Rock Shelter, Northern
Australia. Part I, Experimental Design and Statistical
Models. Archaeometry 41:339-364.

Griffin, J. W.
2002 Archaeology of the Everglades. University Press of
Florida, Gainesville.

Gaiser, E., A. Wachnicka, P. Ruiz, F. Tobias, and M.
2005 Diatom Indicators ofEcosystem Change in Subtropical
Coastal Wetlands. In Estuarine Indicators, edited
by S. A. Bortone, pp. 127-144. CRC Press, Boca

Gleason, P. J.
1972 The Origin, Sedimentation and Stratigraphy of a
Calcitic Mud Located in the Southern Freshwater
Everglades. Ph.D. dissertation. Pennsylvania State
University, University Park.

Gleason, P. J., and P. A. Stone
1994 Age, Origin and Landscape Evolution of the
Everglades Peatland. In Everglades: The Ecosystem
and Its Restoration, edited by S. M. Davis and J. C.
Ogden, pp. 149-198. St. Lucie Press, Delray Beach,

Harvey, J. W., J. T. Newlin, and J. Choi
2003 Characterization of Microtopography in the
Everglades. Abstract of a poster presented at the
Joint Conference on the Science and Restoration of
the Greater Everglades and Florida Bay Ecosystem.
Available online at http://conference.ifaas.ufl.edu/jc/
papers/ geer/4postergr.htm.

Headquarters Air Combat Command
2001a Management Summary/Analyses Report, Phase I
Archaeological Survey Management Units 7, 8,
9, 10, 11, 12, and Main Base, Avon Park Air Force
Range, Avon Park, Florida. Report on file, MacDill
Air Force Base, Florida.

2001b Phase II Archaeological Site Evaluation: Sites
8HG880, 8HG881, and 8HG882. Report on file,
MacDill Air Force Base, Florida.

Hemmings, E. T.
1975 The Silver Springs Site, Prehistory in the Silver
Springs Valley, Florida. The Florida Anthropologist


2007 VOL. 60(2-3)


Hemmings, E. T., and T. A. Kohler
1974 The Lake Kanapaha Site in North Central Florida.
Bureau of Historic Sites and Properties Bulletin

Horvath, E.
2000 Archaeological Investigations at the Colorado Site
(8He241 -A Lithic Workshop in Hernando County,
Florida. The Florida Anthropologist 53(2-3):82-97.

Ivester, A. H., and D. S. Leigh
2003 Riverine Dunes on the Coastal Plain of Georgia,
USA. Geomorphology 51:289-311.

Ivester, A. H., D. S. Leigh, and D. I. Godfrey-Smith
2001 Chronology of Inland Eolian Dunes on the Coastal
Plain of Georgia, USA. Quaternary Research

Johnson, D. L.
2002 Darwin Would Be Proud: Bioturbation, Dynamic
Denudation, and the Power of Theory Science.
Geoarchaeology 17:7-40.

Johnson, D. L., and D. Wattson-Stegner
1987 Proisotropic and Proanisotropic Processes of
Pedoturbation. Soil Science 143:278-292.

Johnson, D. L., J. E. J. Domier, and D. N. Johnson
2005a Reflections on the Nature of Soil and Its Biomantle.
Annals of the Association of American Geographers

2005b Animating the Biodynamics of Soil Thickness Using
Process Vector Analysis: A Dynamic Denudation
Approach to Soil Formation. Geomorphology.

Kalisz, P. J., and E. L. Stone
1984 Soil Mixing by Scarab Beetles and Pocket Gophers
in North-Central Florida. Soil Science Society of
America Journal 48:169-172.

Langmaid, K. K.
1964 Some Effects of Earthworms' Invasion in Virgin
Podzols. Canadian Journal of Soil Science 44:34-

Leigh, D. S.
1998 Evaluating Artifact Burial by Eolian Versus
Bioturbation Processes, South Carolina Sandhills,
USA. Geoarchaeology 13:309-330.

2001 Buried Artifacts in Sandy Soil: Techniques for
Evaluating Pedoturbation Versus Sedimentation.
In Earth Sciences and Archaeology, edited by P.
Goldberg, V. T. Holliday, and C. R. Ferring, pp. 269-
296. Kluwer Academic, New York.

McBrearty, S.
1990 Consider the Humble Termite: Termites as Agents
of Post-Depositional Disturbance at African
Archaeological Sites. Journal of Archaeological
Science 17:111-143.

Michie, J.
1987 Bioturbation and Gravity as a Potential Site
Formation Process: The Open Area Site 38GE261,
Georgetown County, South Carolina. South Carolina
Institute of Anthropology and Archaeology, The
University of South Carolina, Columbia.

1990 Bioturbation and Gravity as a Potential Site Formation
Process: The Open Area Site 38GE261, Georgetown
County, South Carolina. South Carolina Antiquities

Nkem, J. N., L. A. L. de Bruyn, C. D. Grant, and N. R.
2000 The Impact of Ant Bioturbation and Foraging
Activities on Surrounding Soil Properties.
Pedobiologia 44:609-621.

Otvos, E. G.
2003 Prospects for Interregional Correlations Using
Wisconsin and Holocene Aridity Episodes, Northern
Gulf of Mexico Coastal Plain. Quaternary Research

Otvos, E. G., and D. M. Price
2001 Quaternary Inland Dunes of Southern Louisiana and
Arid Climate Phases in the Gulf Coastal Region.
Quaternary Research 55:150-158.

Pierce, C.
1992 Effects of Pocket Gopher Burrowing on
Archaeological Deposits: A Simulation Approach.
Geoarchaeology 7:185-208.

Reimer, P. J, M. G. L. Baillie, E. Bard, A. Bayliss, J. W. Beck,
C. J. H. Bertrand, P. G. Blackwell, C. E. Buck, G. S. Burr,
K. B. Cutler, P. E. Damon, R. L. Edwards, R. G. Fairbanks,
M. Friedrich, T. P. Guilderson, A. G. Hogg, K. A. Hughen,
B. Kromer, G. McCormac, S. Manning, C. B. Ramsey, R. W.
Reimer, S. Remmele, J. R. Southon, M. Stuiver, S. Talamo, F.
W. Taylor, J. van der Plicht, and C. E. Weyhenmeyer
2004 lntCal04 Terrestrial Radiocarbon Age Calibration,
0-26 Cal KYR BP. Radiocarbon 46:1029-1058.

Sanderson D. C. W, P. Bishop, I. Houston, and M. Boonsener
2001 Luminescence Characterization of Quartz-Rich
Cover Sands from NE Thailand. Quaternary Science
Reviews 20:893-900.




Schaetzl, R. J., and L. R. Folmer
1990 Longevity of Treethrow Microtopography:
Implications for Mass Wasting. Geomorphology

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

Scott, T. M.
2001 Text to Accompany the Geologic Map of Florida.
Open-file report 80, Florida Geological Survey. Florida
Dept of Environmental Protection, Tallahassee.

Scott, T. M., K. M. Campbell, F. R. Rupert, J. D. Arthur, T. M.
Missimer, J. M. Lloyd, J. W. Yon, and J. G. Duncan
2001 Geologic Map of the State of Florida, Scale
1:750,000. Florida Geological Survey. Florida Dept
of Environmental Protection, Tallahassee.

Sears, W. H.
1982 Fort Center: An Archaeological Site in the Lake
Okeechobee Basin. University Press of Florida,

Stein, J. K.
1983 Earthworm Activity: A Source of Potential
Disturbance of Archaeological Sediments. American
Antiquity 48:277-289.

Talma, A. S., and J. C. Vogel
1993 A Simplified Approach to Calibrating 14C Dates.
Radiocarbon 35:317-322

Swift, D. R.
1981 Preliminary Investigations of Periphyton and Water
Quality Relationships in the Everglades Water
Conservation Areas. Technical Publication #81-
5, DRE-13. Environmental Sciences Division,
Resource Planning Department, South Florida Water
Management District.

van Nest, J.
2002 The Good Earthworm: How Natural Processes
Preserve Upland Archaic Archaeological Sites of
Western Illinois, U.S.A. Geoarchaeology 17:53-90.

Voslamber, B., and A. W. L. Veen
1985 Digging by Badgers and Rabbits on Some Wooded
Slopes in Belgium. Earth Surface Processes and
Landforms 10:79-82.

Watts, W. A.
1975 A Late Quaternary Record of Vegetation from Lake
Annie, South-Central Florida. Geology 3:344-346.

Watts, W. A., E. C. Grim, and T. C. Hussey
1996 Mid-Holocene Forest History of Florida and the

Coastal Plain of Georgia and South Carolina. In
Archaeology of the Mid-Holocene Southeast, edited
by K. E. Sassaman and D. G. Anderson, pp. 28-40.
University Press of Florida, Gainesville.

Wheatcroft, R. A., and D. E. Drake
2003 Post-Depositional Alteration and Preservation of
Sedimentary Structures on Continental Margins, I.
The Role of Episodic Sedimentation. Marine Geology

White, W. A.
1958 Some Geomorphic Features of Central Peninsular
Florida. Geological Bulletin No. 41. The Florida
Geological Survey, Tallahassee.

1970 The Geomorphology of the Florida Peninsula.
Geological Bulletin No. 51. Bureau of Geology,
Division of Interior Resources. Florida Department
of Natural Resources, Tallahassee.

Wilder, M., and C. D. Frederick (with contributions by R.
Austin, M. Bateman, E. Carlson, A. Cordell, A. Dase, K.
Mickelson, and D. Peter)
2006 ArchaeologicalandGeomorphological investigations
at the Avon Park Air Force Range, Highlands and
Polk Counties, Florida. U.S. Air Force Air Combat
Command Series Reports of Investigations Number
32. Geo-Marine, Inc., Plano, Texas.

Winkler, M. G., P. R. Sanford, and S. W. Kaplan
2002 Hydrology, Vegetation, and Climate Changes in the
Southern Everglades During the Holocene. Bulletins
ofAmerican Paleontology 361(5):57-99.

Wood, R. W., and D. L. Johnson
1978 A Survey of Disturbance Processes in Archaeological
Site Formation. In Advances in Archaeological
Method and Theory, vol. 1, edited by M. B. Schiffer,
pp. 315-381. Academic Press, New York.

2007 VOL. 60(2-3)




Florida Museum ofNatural History, Dickinson Hall, University ofFlorida, Museum Road, Gainesville, FL 32611
Email: cordell@flmnh.ulf edu


This paper presents the results of petrographic analysis
of a small sample of Belle Glade Plain and sherds with sandy
St. Johns paste. The impetus for this paper developed out
of a petrographic analysis of pottery thin sections from the
Avon Park Air Force Range (APAFR), near Sebring, Florida
conducted for Geo-Marine, Inc. (Cordell 2006; Wilder and
Frederick 2006). The Avon Park analysis was carried out to
characterize and evaluate several sandy and spiculate paste
categories that were defined during Geo-Marine's preliminary
sorting of pottery from several sites on the APAFR (Austin
2004; Wilder and Frederick 2006:39-40). The present study
focuses on two of the categories, Belle Glade and sandy St.
Johns, owing to overlapping criteria for-and persistent
inconsistencies in-their recognition (Austin 1996; Cordell
1992, 2006).

Belle Glade and Sandy St. Johns Paste Descriptions:
Sorting Criteria and Sources of Confusion

Early descriptions of Belle Glade Plain are thoroughly
recapitulated by Robert Austin (1996:75). Austin further
provides the currently accepted working definition of Belle
Glade Plain as "... a hard, spiculite [sic] paste with common
to abundant quartz sand as inclusions.... The sherds may or
may not possess a chalky feel. If the sherd is large enough it
will generally display drag marks, facets, and extruded sand
grains that are the result of dragging a tool across a nearly dry
surface" (Austin 2004:1; Wilder and Frederick 2006:40). In
contrast, sandy St. Johns is "harder and contains more sand
than typical examples of St. Johns Plain...often confused with
Belle Glade Plain but should be identifiable by the absence
of drag marks, facets, and extruded sand grains...may or may
not have a chalky feel" (Austin 2004:1; Wilder and Frederick
2006:40; for other definitions see Cordell and Koski 2003:120-
121 and Russo et al. 1989:65).
These definitions evolved out of a key developed by
Austin to reduce the subjectivity inherent in the sorting of
undecorated sand-tempered and spiculate paste pottery from
the Kissimmee River/Lake Okeechobee Region (Figure 1)
(Austin 2004;Wilder and Frederick 2006:40). The sorting
criteria include relative hardness, paste constituents or temper,
and surface finish. This key is very helpful when coupled with
the use of a microscope (70x magnification) to examine freshly
broken edges, as recommended by Austin. But there is still
room for improvement. Overlapping and/or conflicting criteria
point to obvious difficulties in distinguishing Belle Glade and
sandy St. Johns pastes. By this key, Belle Glade Plain and

sandy St. Johns Plain differ only in terms of surface finish.
As Austin alludes to, consistency in recognizing the scraped
faceted surface finish (Figure 2), especially on smaller sherds,
may be difficult to achieve.
The key also allows for the occurrence of spiculate and
non-spiculate Belle Glade categories, designated BGP-A and
BGP-B, respectively (Wilder and Frederick 2006:40). I have
elsewhere suggested that spiculate paste should be the primary
criterion of Belle Glade paste (Cordell 1992:111), which
would preclude a Belle Glade Plain designation for a sand-
tempered plain sherd with faceted or scraped surface finish.
Given the use of 70x magnification in the preliminary sorting
of APAFR pottery, I suspect that the apparent lack of sponge
spicules in some sherds that are otherwise Belle Glade-like
in surface finish, may be attributed to examination of a fresh
break made perpendicular to the preferred orientation of the
sponge spicules. The preferred orientation of sponge spicules
in Belle Glade (and St. Johns) pottery is parallel to horizontal
planes of a pottery vessel (rims, coil fractures or junctures,
etc.). Fresh breaks made parallel to this preferred orientation
expose more easily recognized longitudinal sections of
sponge spicules (Figure 3a,b). In contrast, fresh breaks made
perpendicular to this preferred orientation expose mainly
circular cross-sections of sponge spicules that are not as easily
recognized as sponge spicules, except in petrographic thin
section (Figure 4). This scenario might account for Austin's
BGP-B category, sherds with characteristic "Belle Glade"
surface finish but no apparent sponge spicules. Orientation of
fresh breaks would likewise affect the success in recognizing
the spiculate character of sandy St. Johns sherds.

The Present Study

In the present study, I will quantify particle size and
frequency descriptions of paste/temper constituents for Belle
Glade and sandy St. Johns pastes based on petrographic point-
count analysis. My goal is to attempt to eliminate the potential
for error in distinguishing between Belle Glade and sandy St.
Johns pastes.

The Pottery Sample

The Belle Glade sample consists of thin sections of 13
sherds (Table 1). Three sherds are from the Avon Park project
(Polk and Highlands counties; Cordell 2006). Four other
sherds are from sites in Brevard (n=l), Lee (n=2), and Collier
(n= 1) counties, and six are from sites in Palm Beach County,
courtesy of Florida Atlantic University (FAU)' (Table 1).


VOL. 60(2-3)




2007 VOL. 60(2-3)

Hard, Compact Paste?


Sand in Paste?





Spicules in Paste?
/ \



Sand in Paste?


Spicules in Paste?
/ \


Spicules in Paste?
/ \


Belle Glade Plain with sponge spicules
Belle Glade Plain without sponge spicules
sandy St. Johns Plain
sand-tempered plain
St. Johns Plain

Figure 1. Austin's ceramic sorting key for the Kissimmee region (source: Austin 2004 and Figure 8 in Wilder and Frederick

Figure 2. Belle Glade Plain sherd showing large scraped facet and characteristic drag/scratches. FLMNH 91-31-368 (8CR1,
Chokoloskee Key), FLMNH-CTL Type Collection. Photo by Mr. Pat Payne (Phead262@aol.com).

icules in Paste?
Spicules in Paste?


/ \







Figure 3a. Figure 3b.

Figure 3c. Figure 3d.

Figure 3. (a) Photomicrograph of sponge spicules (longitudinal sections) showing preferred orientation (sandy St. Johns
paste thin section sample UPSTJ-3, 100x magnification; approximate width of image 0.7mm). (b) Close-up of sponge spic-
ules in sandy St. Johns thin section (UPSTJ-3, 250x magnification; approximate width of image 0.3mm). (c) Photomicro-
graph of Belle Glade Plain thin section showing sponge spicules and quartz grains (mostly fine) (Belle Glade GMI-8, 100x
magnification; approximate width of image 0.7mm). (d) Close-up of Belle Glade Plain thin section showing sponge spicules
(GMI-8, 250x magnification; approximate width of image 0.3mm).

The sample for sandy St. Johns paste consists of six thin
sections (Table 1). One sherd is from the Avon Park project
(GMI-4) and one is from St. Lucie County (10Mile-15),
courtesy of New South Associates, Inc (Loubser et al. 2005).
The other four sherds include one from Brevard County and
three from St. Johns County. Although this sample of six
sherds is very small, it represents 100% of the sandy St. Johns
paste sherds that have been thin sectioned and curated at the
Florida Museum of Natural History (FLMNH)2.

Methods ofAnalysis

The petrographic analysis was conducted to evaluate
compositional and particle size homogeneity and differences
within and between Belle Glade and sandy St. Johns samples.
Point counts were made for quantifying relative abundance

of inclusions within a range of particles sizes. This procedure
involved using the petrographic microscope with a mechanical
stage and generally followed recommendations by Stoltman
(1989, 1991, 2000). Counting intervals of 1 mm by Imm and
1mm by .5mm were used as a compromise between the small
particle sizes of matrix constituents (primarily sponge spicules)
and larger sizes of the quartz inclusions. Each point or stop of
the stage was assigned to one of the following categories: clay
matrix, voids (channel voids, closed pores, and micropores
[Rice 1987:350]), silt particles, sponge spicules, and very
fine through very coarse quartz and other aplastics of varying
compositions. Most of the counts were made using the 10X
objective (magnification of 100x), but the 25X (250x) objective
(with plane-polarized light) was used when necessary to double
check the occurrence of sponge spicules. Size of aplastics
was estimated with reference to the Wentworth Scale (Rice






Figure 4a. Figure 4b.

Figure 4c. Figure 4d.
Figure 4. (a) Photomicrograph of sponge spicules in cross-section, perpendicular to preferred orientation (sandy St. Johns
thin section GMI-4, 100x magnification; approximate width of image 0.7mm). (b) Close-up of sponge spicules in cross-sec-
tion in sandy St. Johns thin section (GMI-4, 250x magnification; approximate width of image 0.3mm). (c) Photomicrograph
of sponge spicules in cross-section in Belle Glade thin section (GMI-7, 100x magnification; approximate width of image
0.7mm). (d) Close-up of sponge spicules in cross-section (and large [medium] quartz grain) in Belle Glade thin section (GMI-
7, 250x magnification; approximate width of image 0.3mm).

1987:38). Sponge spicule particle size was estimated with an
eyepiece micrometer during petrographic analysis (counts of
sponge spicules were not made within particle size categories
during the point-counting procedure). For cases in which fewer
than 200 points were counted (n=4), the thin sections were
rotated 180 on the mechanical stage and counted a second
time (after Stoltman 2000:306). Analyses were carried out
in the FLMNH Ceramic Technology Laboratory (FLMNH-
CTL) and FAU Geosciences Department. Data for the six
Belle Glade Plain thin sections from FAU collections include
point counts made by the author and Carmen Giomar Sanchez,
graduate student in anthropology at FAU.
Raw point-count data and percentages are listed in Table
2. By convention, the total counts exclude the number of
counted voids (Stoltman 1991:107). Point-count data were
used to calculate a "sand size" index for each sample following
Stoltman (2000:314). Statistical comparisons of petrographic

data were calculated using SAS/STAT Guide for Personal
Computers (SAS 1988).

Results: Evaluation of Belle Glade and Sandy St. Johns

Paste/Temper Constituents

Quartz sand and sponge spicules were the predominant
constituents in both Belle Glade and sandy St. Johns paste
samples. Polycrystalline quartz was observed in very small
amounts in most cases and counts were included with counts of
quartz sand. Other accessory constituents that were observed
petrographically include microcline (K-feldspar), unidentified
(UID) feldspar (no twinning), epidote, green pleochroic
minerals such as amphibole and tourmaline, kyanite, hematite
or ferric concretions, and UID heavy minerals. These latter

2007 VOL. 60(2-3)

Table 1. The Pottery Samples: Belle Glade paste (BGP) and Sandy St. Johns paste (SSJ).
Sample #" Cat. e Site or County Name Paste Project Reference
Number ber
Wilder and Frederick 2006;
GMI-3 pending 8HG938 Highlands County BGP APAFR/Geo-Marine, Inc. Cordell 2006
Wilder and Frederick 2006
GMI-7 pending 8PO6094 Polk County BGP APAFR/Geo-Marine, Inc. Cordell 2006
Wilder and Frederick 2006;
GMI-8 pending 8PO5306 Polk County BGP APAFR/Geo-Marine, Inc. Cordell 2006

UPSTJ-4 A21737 8BR193 Gauthier site BGP FLMNH Upper St. JohnsSigler-Eisenberg 1985
FL- A27550 8LL722 Buck Key Shell BGP FLMNH Southwest Florida
SWFL-1 A27550 8LL722 MBGP Marquardt 1992; Cordell 1992
Midden Project
FLMNH Southwest Florida
SWFL-17 A20382 8LL51 Collier Inn BGP FLNH Southwest Florida Marquardt 1992; Cordell 1992
Horrs-10 89-22-16 8CR201 Horrs Island BGP FLMNH Horr's Island Project Russo 1991; Cordell 1991
FAU-I A1900 8PB56 Boca Weir BGP Florida Atlantic University Sanchez and Cordell in progress
FAU-3 A1995-1 8PB56 Boca Weir BGP Florida Atlantic University Sanchez and Cordell in progress
FAU-4 A1998 8PB56 Boca Weir BGP Florida Atlantic University Sanchez and Cordell in progress
FAU-6 A2126 8PB30 Riviera Beach Mound BGP Florida Atlantic University Sanchez and Cordell in progress
FAU-8 A2832 unknownb BGP Florida Atlantic University Sanchez and Cordell in progress
FAU-9 A2837 8PB34 Jupiter Inlet I BGP Florida Atlantic University Sanchez and Cordell in progress
UPSTJ-3 A21691 8BR193 Gauthier site SSJ FLMNH Upper St. JohnsSigler-Eisenberg 1985
Wilder and Frederick 2006;
GMI-4 pending 8P05306 Polk County SSJ APAFR/Geo-Marine, Inc. Cordell 2006
Ten Mile Creek Project/New Loubser et al. 2005; Cordell
10Mile-15 03-66-15 8SL1181 Ten Mile Creek site SSJ
South Associates 2005
Anderson 2001; Deagan 2004;
FOY-C40 01-95-2233 8SJ31 Fountain of Youth Park SSJ FLMNH FOY project A Cden 200; D n 2004
Cordell in progress2004
FOY-C63 87-44-431 8SJ31 Fountain of Youth Park SSJ FLMNH FOY project Anderson 2001; Deagan 2004;
Cordell in progress
FOY-C65 87-44-431 8SJ31 Fountain of Youth Park SSJ FLMNH FOY project Anderson 2001; Deagan 2004
Cordell in progress
aSample/project prefixes: APAFR= Avon Park Air Force Range; GMI= Geo-Marine, Inc.; PSTJ= Upper St. Johns Project; SWFL= Southwest
Florida Project; Horrs= Horr's Island; 10Mile= Ten Mile Creek Project/New South Associates; FLMNH= Florida Museum of Natural History;
FOY= Fountain of Youth Park site; FAU= Florida Atlantic University.
site probably from Palm Beach County, Florida.

constituents were observed in very small amounts (less than
1%) and with no discernable pattern among the samples (data
on file, FLMNH-CTL). Silt particles, composed of quartz,
appeared to be under-represented in the point count data. A
comparison chart of estimated percent particle abundance
(Rice 1987:349 [Figure 12.2]) was used for estimating relative
abundance of silt (Table 2) and accessory constituents.

Belle Glade Point Count Data

The Belle Glade paste sample is characterized by means
of 14% sponge spicules and 22% quartz sand, primarily fine
and medium in size (Table 2). The mean sand size index
is 1.44. The point count percentages show a great deal of
variability in sponge spicule and quartz frequency within
the Belle Glade sample. The data likely include end and mid
points in a continuum of sponge spicule and quartz frequency
within Belle Glade paste. Mean and range percentages for
sponge spicules translate roughly to "frequent to common,"
while mean and range percentages for quartz sand is "frequent
to abundant."
The sponge spicules in Belle Glade paste sherds are
generally very small, ranging primarily 0.625-0.125mm in

length (very fine on the Wentworth Scale) and 0.008-0.024mm
in diameter (within silt size range on the Wentworth Scale)
(Table 2). Maximum length of whole spicules ranges up to
about .25mm (within range of fine sizes on the Wentworth
Scale), but most spicules are fragmented.

Sandy St. Johns Point Count Data

The sandy St. Johns sample is characterized by means of
22% sponge spicules and 22% quartz sand, primarily fine in
size (Table 3). The mean sand size index is 1.21. For sandy
St. Johns paste, mean and range percentages for sponge
spicules translate roughly to "common", while mean and
range percentages for quartz sand is "common to abundant."
Sponge spicules in sandy St. Johns sherds (and St. Johns
paste) are relatively big. Unbroken sponge spicules in sandy
St. Johns (and St. Johns) sherds can range up to 0.50 mm in
length (within range of medium sizes on the Wentworth Scale)
and 0.008-0.04mm in diameter (silt Wentworth sizes). Most
spicules are fragmented and lengths fall mainly within very
fine and fine Wentworth sizes (0.0625-0.25mm).



Table 2. Raw Point Count and Percentage Data for Belle Glade Samples.
Count, Count, Count, Count,
Total ount, Count, Count, Count Count, % V Count % % Sand Size
Sample # % % Clay % Total % % Fine % Sand Size
Count % % l Sponge % % Silt Fine Medium Coarse Index'
Count Voidsf Matrix Aplastics Spicules and Quartz Quartz Quartz
Spicules Quartz Quartz Quartz
S27 299 156 64 89 3 13 36 30 10
GMI-3" 455 1.56
6% 66% 34% 14% 20% 1% 3% 8% 7% 2%
GM b 289 8 209 80 50 28 2 6 11 10 1 1.43
3% 72% 28% 17% 10% 1% 2% 4% 4% (1%)
GM8b 255 29 134 121 21 95 5 11 43 35 71.52
10% 53% 47% 8% 39% 2% 4% 17% 14% 3%
UPST 246 11 116 130 38 89 3 5 48 28 8
4% 47% 53% 15% 37% 1% 2% 20% 11% 3%
S11 113 92 26 60 6 9 23 26 2
5% 55% 45% 13% 29% 2% 4% 11% 13% 1% 1.0
SWFL 40 190 134 50 82 2 10 37 30 5
SWFL-17E 324 1.49
11% 59% 41% 15% 25% 1% 3% 11% 9% 2%
Horrs-1b 2 47 153 85 23 53 9 12 26 14 1 130
16% 64% 36% 10% 22% 4% 5% 11% 6% (1%)
FAUd 691 71 455 236 111 119 6 18 60 37 4 1.38
-9% 66% 34% 16% 17% 1% 3% 9% 5% 1%
FAU-3d 624 66 445 179 96 80 3 12 39 26 2 136
10% 71% 29% 15% 13% (1-3%) 2% 6% 4% (1%) _
86 285 201 46 146 9 30 54 48 9
FAU-4e 486 1.47
15% 59% 41% 10% 30% 2% 6% 11% 10% 2%
e 51 410 159 89 66 4 8 36 21 1
FAU-6e 569 1.35
8% 72% 28% 16% 12% 1% 1% 6% 4% <1%
FAU8b 56 22 163 93 41 51 1 13 22 10 2 1.31
F 8 26 8% 64% 36% 16% 20% (1-3%) 5% 9% 4% 1% __
S49 362 203 134 66 2 12 24 17 6
FAU-9 0 565 1.56
8% 64% 36% 24% 12% (1-3%) 2% 4% 3% 1%
mean 9% 62% 38% 14% 22% 2% 3% 10% 7% 2% 1 44
counting interval lx.5mm, counted once.
counting interval lx 1mm, counted once.
'counting interval lx1mm, counted twice.
counting interval lx.5mm, counted twice.
'counted twice, once at interval lx.5mm and once at interval xlm m.
'excluded from total count.
sponge spicules range primarily 0.625-0.125mm (very fine on Wentworth Scale) in length and 0.008-0.024mm (silt on Wentworth Scale) in diameter.
h"total sand" is the sum of counts of very fine through coarse quartz, plus counts of polycrystalline quartz and feldspar grains.
'sand size index calculated following Stoltman (2000:13); counts of silt, sponge spicules, and ferric concretions excluded from calculations.

Comparison ofBelle Glade and sandy St. Johns Data

The comparisons of Belle Glade and sandy St. Johns
data in Table 3 show that there is overlap in most of the
compositional and particle size categories, which is not
surprising given the above descriptions as well as traditional
difficulties in distinguishing Belle Glade and sandy St. Johns
pastes. There are, however, statistically significant differences
in mean percentages of matrix, total aplastics, sponge spicules,
and mean sand size index (Table 4), despite the small sample
sizes. Mean percentages of total aplastics and sponge spicules
are higher in sandy St. Johns than Belle Glade paste (47%
aplastics and 22% sponge spicules versus 39% aplastics and
14% sponge spicules, respectively), while mean percent
matrix and mean sand size index are higher in Belle Glade
than sandy St. Johns paste (62% matrix and mean sand size
index of 1.44 versus 53% matrix and mean sand size index
of 1.21, respectively). Estimated percentages of silt (primarily
composed of quartz) are higher for sandy St. Johns sherds
(Table 4), but this distinction is likely to be obvious only in
thin section.

Instead of distinguishing Belle Glade and sandy St.
Johns pastes, the point count data corroborate the similarities
between the two categories, providing quantification for the
acknowledged difficulties in making the paste distinctions. The
translated verbal descriptions for the two paste categories are
very close to one another. Differences in relative abundance
of sponge spicules or clay matrix versus aplastics, while
statistically significant, might not be easily or consistently
detected using a standard microscope with incident fiber-optic
Although not precisely quantified, sponge spicule sizes
may offer a reliable way to distinguish Belle Glade and sandy
St. John pastes. In the present sample, sponge spicules in sandy
St. Johns sherds (and St. Johns paste) are generally larger
in length and diameter than those in Belle Glade paste (see
Figures 3 and 4), although there is overlap in both dimensions. I
have observed this same spicule size distinction in Belle Glade
and St. Johns pottery from southwest Florida sites (Cordell
1992:112) and in examples of Belle Glade Plain and St. Johns
Plain in the type collection maintained in the FLMNH-CTL. It
is uncertain whether the length difference can be attributed to


2007 VOL. 60(2-3)

Table 3. Raw Point Count and Percentage Data for Sandy St. Johns Samples.
Count! Count! Count/ Count!
T l Count, Count/ o / Count Countount/ Count ount/ ount/
Sample # Total % Count/ % Very % Fine % % Sand Size
Sample # Co % % Clay % Total % % Fine
unt Voidsd Matrix Aplastics Sponge Sand % Silt Fine Quartz Medium Coarse Index
SpiculesQuartz Quartz Quartz Quartz
UPSTJ-3a 351 65 125 226 72 141 13 41 80 16 4 117
16% 36% 64% 20% 40% 4% 12% 23% 5% 1%
GMI4 218 25 129 89 54 34 1 5 21 8 (<%) 1.24
10% 59% 41% 24% 16% <1% 2% 10% 4%
10Mile- 18 127 127 66 57 4 17 26 13
15b 254 7% 50% 50% 26% 22% 2% 7% 11% 5% (<1%) 1.23
FOY-C40' 292 24 173 119 47 62 8 5 25 26 1.45
8% 59% 41% 16% 21% 3% 2% 9% 9%
FOY-C63c 297 26 175 122 51 57 13 13 39 4 1.07
8% 59% 41% 17% 19% 4% 4% 13% 1%
FOY-C65a 237 15 127 110 65 40 5 9 26 4 1.10
6% 54% 46% 27% 17% 2% 4% 11% 2%
mean 9% 53% 47% 22% 22% 3% 5% 13% 4% <1% 1.21
counting interval Ix 1mm, counted once.
counting interval xlm m, counted twice.
Counting interval lx.5mm, counted once.
excluded from total count.
'sponge spicules range primarily 0.0625-0.25mm (very fine to fine/medium on Wentworth Scale) in length and 0.008-0.04mm (silt on Wentworth Scale)
in diameter.
f"total sand" is the sum of counts of very fine through coarse quartz, plus counts of polycrystalline quartz, feldspar, and other crystalling grains.
sand size index calculated following Stoltman (2000:13); counts of silt, sponge spicules, and ferric concretions excluded from calculations.

Table 4. Comparison of Belle Glade and Sandy St. Johns Point Count (percentages) Data.

% % % % Sponge Sponge % % Silt Very % % % Sand
Paste Voids Matrix Aplastics ponge Spicule Spicule Sand (estimated) Fine Fine Medium Coarse Size
Spicules Length Diameter Quartz Quartz Quartz Quartz Index
Belle 9% 62% 38% 14% very fine 0.008- 22% 2% 3% 10% 7% 2% 1.44
Glade (3- (47- 0.0625- (10- 3 (4- (<1 1.30-
lade ) 7(3 ) ( (28-53%) (8-24%) 0.024mm ( (1-3%) (1-6%) 0 ) (3-14%)
o 13 16%) 720) 0.125mm 39%) 20% 3%) 1.56)
sandy 9 53% very fine to 22% 5% 13% 1.21
S. 9% 53% 47% 22% fine/medium 0.008- % 3% 5% 13% 4% <1% 1.21
Johns 16%) 5(- (41-64%) (16-27%) 0.0625- 0.04mm 4 (3-10%) 12) 23) (1-9%) (0-1%) 1.07-
n 16%) 59%) .240%) 12%) 23%) 1.45)
n=6 0.25rmm

t-test comparisons between pastes for mean %matrix, mean %aplastics, %sp0
%matrix: --2.42 df-17 p<.0269
%aplastics: t=-2.41 df=17 p<.0275
%sponge spicules: t=-3.42 df-17 p<.0031
%silt: t-2.19 df-17 p<.0430
sand size index: t=4.39 df-17 p<.0004

differential fragmentation (greater in Belle Glade paste?) or to
a sponge species difference for sponge spicules in Belle Glade
and sandy St. Johns pastes. A sponge species difference would
indicate a significant difference in spiculate resources used by
Belle Glade and St. Johns potters, regardless of whether the
sponge spicules were added as temper or naturally present in
the clays (see Borremans and Shaak, 1986; Cordell and Koski
2003; Rolland and Bond 2003). Differential growth stages
at time of incorporation into clayey sediments (if natural to
clay deposit), or at time of collection (if added as temper)
might also account for the sponge spicule size variability. The
variation in spicule diameters could be attributable to species
differences or to differential growth stages at time of death or
collection, or to geographic variation in growth patterns. It is
very difficult to identify species from spicules alone (Johnson
1945:14). However, species differences may be the most likely

onge spicules, mean %silt, and mean sand size index:

explanation given the differences in gross regional distributions
of Belle Glade (south Florida) and St. Johns (east Florida)
pottery (Dr. J. Keith Rigby, personal communication3).
With respect to the Avon Park samples in particular, the
results of the petrographic analysis corroborated the Belle
Glade Plain and sandy St. Johns designations of three sherds
that were so-designated during the APAFR preliminary
analysis by Geo-Marine Inc. researchers. A fourth Avon Park
sherd (GMI#3-Table 1) had been designated sandy St. Johns
Plain during the preliminary analysis but was reassigned to
the Belle Glade paste category after petrographic analysis
(Cordell 2006:260) on the basis of sponge spicule size range
and percentage of sponge spicules.




Conclusions and Recommendations

The difficulty in distinguishing sandy St. Johns and
Belle Glade paste sherds might be most easily resolved by
considering sponge spicule size. Belle Glade paste sherds
have sponge spicules that range in length primarily within
very fine particles sizes, while sandy St. Johns sherds have
sponge spicules that range in length primarily within very
fine and fine particles sizes. Although the sample size for this
study is small, especially for sandy St. Johns paste, the sponge
spicule size distinction has been observed in hundreds of
sherds from southwest Florida and in type specimens housed
at the FLMNH. This size distinction might therefore extend
to the larger populations of Belle Glade and St. Johns pottery.
Testing the broader utility of the relationship between spicule
size and pottery type will, of course, require some effort to
estimate sponge spicule particle size ranges.
Making this particle size distinction would not require
petrography. Sponge spicule sizes can be estimated using a
standard microscope with incident fiber optic light equipped
with an eyepiece micrometer. In the absence of an eyepiece
micrometer, particle sizes can be estimated by determining
the width of the field of view at specified magnifications.
Another way of considering this is if the sponge spicules can
be seen easily with low (e.g. 30x) magnification, then the
sherd probably has St. Johns or sandy St. Johns paste. If the
sponge spicules can only be distinguished only under higher
magnification (e.g. 70x), then the sherd probably has Belle
Glade paste. In my experience, sponge spicules can be detected
in St. Johns and sandy St. Johns sherds even on non-eroded
and non-encrusted surfaces as well as in freshly broken edges
at 30x magnification. But the same can not be said for Belle
Glade pottery, which almost always requires examination of a
fresh break with 70x magnification. Even in a fresh break, the
sponge spicules in Belle Glade pottery may appear somewhat
indistinct, as the paste texture tends to resemble gritty cotton
wool. Greater success in making the distinction may be
achieved when the fresh break examined is made parallel to
the preferred orientation of the sponge spicules, i.e., parallel to
coil fractures or the rim if present. With these considerations
in mind, I offer the following descriptions of Belle Glade and
sandy St. Johns pastes:
Belle Glade paste is characterized by frequent to common
sponge spicules and frequent to abundant quartz sand very
fine to coarse in size. Sponge spicules are fragmented and
generally very fine in length; they will not be easily detected
below 70x magnification. Scraped/scratched, faceted surface
finish is characteristic of the type.
Sandy St. Johns paste is characterized by common to
abundant sponge spicules and common to abundant quartz sand
very fine to medium in size. Sponge spicules are fragmented
and generally very fine and fine in length; they should be easily
detected with 30x magnification. Scraped/scratched, faceted
surface finish is not characteristic of the type.
Scraped/scratched, faceted surface finish should still
be considered a very useful, but not primary, criterion for
corroborating Belle Glade paste. Perhaps archaeologists will
find this sponge spicule size difference helpful in making Belle

Glade/sandy St. Johns paste distinctions in cases in which
surface finish is eroded or otherwise ambiguous.


1. When this manuscript was first submitted for publication, the
sample for Belle Glade Plain consisted of seven thin sections.
About the same time, a set of thin sections including 28 of Belle
Glade Plain, was rediscovered in the archaeology laboratory at
Florida Atlantic University by FAU graduate student Carmen
Sanchez. Six of the thin sections were subsequently analyzed.
After consultation with the editors, I included the data in this
study. The 22 other Belle Glade thin sections were excluded
on the basis of poor quality thin sectioning.

2. When this manuscript was first submitted for publication,
the sample for sandy St. Johns paste consisted of seven thin
sections, four of which were actually chalky fiber-tempered
sherds (Cordell 2004) serving as proxies for sandy St. Johns
paste, with fiber-void counts deleted from point count data.
One reviewer rightly questioned the validity of making this
substitution. During the review process, I had completed
point count analysis of three sandy St. Johns thin sections
from the Fountain of Youth site (8SJ31, St. Augustine, FL).
After consultation with the editors, I deleted the chalky fiber-
tempered sherds from the study, replacing them with the three
newly acquired sandy St. Johns samples.

3. Dr. J. Keith Rigby, Professor emeritus at Brigham Young
University Department of Geological Sciences, is considered
"Mr. Sponge," a paleontologist who specializes in fossil


The thin sections were obtained through projects sponsored
by Geo-Marine, Inc., New South Associates, Florida Atlantic
University, and the Florida Museum of Natural History.
Most The thin sections are curated in the FLMNH Ceramic
Technology Laboratory (http://www.flmnh.ufl.edu/anthro/
ceramic_lab.htm) and the FAU Anthropology Department'. I
am grateful to Michael Wilder for encouraging me to write
this paper; to Carmen Sanchez for the loan of the FAU thin
sections and for her contributions to the point count data; and
to Dr. J. Keith Rigby for commenting on the photomicrographs
of sponge spicules in the pottery. Thanks also go to the United
States Air Force Air Combat Command for permission to adapt
APAFR Figure 8 for use in this paper and to Robert Austin for
a file of the original figure. Pat Payne (Phead262@aol.com)
expertly photographed the sherd used in Figure 2. Helpful
critiques were provided by FLMNH colleagues Gifford Waters
and Sylvia Scudder and by the reviewers and editors of The
Florida Anthropologist.

References Cited

Anderson, Jamie L.
2001 2001 Excavations at 8SJ31 The Fountain of Youth


2007 VOL. 60(2-3)


Park Site 1565 Spanish Campsite. Florida Museum
of Natural History Miscellaneous Project Reports in
Archaeology #54. Gainesville, Florida.

Austin, Robert J.
1996 Ceramic Seriation, Radiocarbon Dates, and
Subsistence Data from the Kissimmee River Valley:
Archaeological Evidence for Belle Glade Occupation.
The Florida Anthropologist 49(2):65-87.

2004 Ceramic Descriptions for the Avon Park Air Force
Range Project. MS on file, Avon Park Air Force

Borremans, Nina T. and Graig D. Shaak
1986 A Preliminary Report on investigation of Sponge
Spicules in Florida "Chalky" Paste Pottery. Ceramic
Notes 3, pp. 125-131.

Cordell, Ann S.
1991 Paste Variability in Horr's Island Pottery. In Final
Report on Horr 's Island: the Archaeology ofArchaic
and Glades Settlements and Subsistence Patterns,
by Michael Russo, pp. 556-590. Report prepared for
Key Marco Developments, Marco Island, Florida.

1992 Technological Investigation of Pottery Variability
in Southwest Florida. In Culture and Environment
in the Domain of the Calusa, edited by William H.
Marquardt. Monograph Number 1 of the Institute
of Archaeology and Paleoenvironmental Studies,
pp. 105-189. University of Florida, Gainesville.

2004 Paste Variability and Possible Manufacturing
Origins of Late Archaic Fiber-tempered Pottery
from Selected Sites in Peninsular Florida. In Early
Pottery: Technology, Style, and Interaction in the
Lower Southeast, edited by Rebecca Saunders and
Christopher Hayes, pp. 63-104. The University of
Alabama Press, Tuscaloosa.

2006 Petrographic Evaluation of Paste Categories from
the Avon Park Air Force Range Project, Highlands
and Polk Counties, Florida. In "Archaeological and
Geomorphological Investigations at the Avon ParkAir
Force Range, Highlands and Polk Counties, Florida"
by Michael Wilder and Charles D. Frederick, Chapter
7 Specialized Analyses, Part 1, pp. 257-261. United
States Air Force Combat Command Series, Report of
Investigations Number 32. Report prepared for U.S.
Army Corps of Engineers by Geo-Marine, Inc. Piano,

Cordell, Ann S. and Steven H. Koski
2003 Analysis of a Spiculate Clay from Lake Monroe,
Volusia County, Florida. The Florida Anthropologist

Deagan, Kathleen
2004 Summary Interpretation of Archaeological Field
Work at the Fountain of Youth Park Site (8-SJ-31)
1951-2002. Florida Museum of Natural History
Miscellaneous Project Reports in Archaeology #56.
Gainesville, Florida.

Johnson, Margaret Crile
1945 The Freshwater Sponges of Alachua County, with a
Summary of the Known Florida Forms. MA Thesis,
Department of Zoology, University of Florida,

Loubser, J, A. Cordell, L. Raymer, H. Mattemes and P.
2005 Phase III Data Recovery of 8SL1181 at Ten Mile
Creek, St. Lucie County, Florida. Report prepared
for U.S. Army Corps of Engineers by New South
Associates, Stone Mountain, GA. New South
Associates Technical Report 1184.

Rolland, Vicki L. and Paulette Bond
2003 The Search for Spiculate Clays Near Aboriginal Sites
in the Lower St. Johns River Region, Florida. The
Florida Anthropologist 56(2):91-111.

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

Russo, Michael
1991 Final Report on Horr's Island: the Archaeology of
Archaic and Glades Settlements and Subsistence
Patterns. Report prepared for Key Marco
Developments, Marco Island, Florida.


Michael, Ann Cordell, Lee Newsom, and Robert

Ceramic Analysis. In Phase III Archaeological
Excavations at Edgewater Landing, Volusia County,
Florida, by Michael Russo,Ann Cordell, Lee Newsom
and Robert Austin. Report prepared for Radnor/
Edgewater Landing, Inc. by Piper Archaeological
Research, Inc., St. Petersburg.

SAS Institute Inc.
1988 SAS/STAT Guide for Personal Computers, Version 6
Edition. SAS Institute Inc., Cary, North Carolina.

Sigler-Eisenberg, Brenda
1985 Archaeological Site Types, Distribution, and
Preservation within the Upper St. Johns River Basin,
Florida. Florida State Museum Miscellaneous Project
and Report Series Number 27, Gainesville.
Stoltman, B. James
1989 A Quantitative Approach to the Petrographic Analysis
of Ceramic Thin Sections. American Antiquity



1991 Ceramic Petrography as a Technique for Documenting
Cultural Interaction: An Example from the Upper
Mississippi Valley. American Antiquity 56 (1):103-

2000 The Role of Petrography in the Study of
Archaeological Ceramics. In Earth Sciences and
Archaeology, edited by Paul Goldberg, Vance T.
Holiday, and C. Reid Ferring, pp.297-326. Kluwer
Academic/Plenum Publishers.

Wilder, Michael and Charles D. Frederick
2006 "ArchaeologicalandGeomorphologicalInvestigations
at the Avon Park Air Force Range, Highlands and
Polk Counties, Florida" with contributions by Mark
Bateman, Robert Austin, Ann Cordell, Lisabeth
Carlson, Katherine Mickelson, and Duane Peter.
United States Air Force Combat Command Series,
Report of Investigations Number 32. Report prepared
for U.S. Army Corps of Engineers by Geo-Marine,
Inc. Plano, TX.

2007 VOL. 60(2-3)




Florida Maritime Museum at Cortez, PO Box 100, Cortez, FL 34215
Email: jeff.moates@manateeclerk.com

In response to an informant report, archaeologists from
the Florida Bureau of Archaeological Research (BAR)
recorded the submerged remains of a flat-bottomed watercraft
in the Wacissa River (Jefferson County, Florida) in April of
2006 (Figure 1) and returned for further documentation in
March of 2007. Wacissa Boat 1 (8JE1604) appears to have
been used in the latter half of the nineteenth and possibly
early twentieth centuries to ferry people and/or cargo through
one of the numerous shallow canals of the Wacissa River
drainage. The vessel represents an important relic of local
history and also reveals information on a particular type of
flat-bottomed watercraft, the plantation flat. The discovery and
documentation ofWacissa Boat 1 in a remote region of northern
Florida signifies a geographical and temporal expansion of this
specific vessel type. While Wacissa Boat 1 represents a later
version of a flat-bottomed watercraft principally suited for use
in shallow water environments, it also offers a glimpse into
the methods of waterbore traffic behind the North Florida
coastline, as well as the practicality and adaptive strategies of
the people building and utilizing these vessels.
Wacissa Boat 1 shares characteristics of two methods
of barge construction, split-log and plank-built, and shows a
close resemblance to these types that are known to have been
associated with 18t" and 19"' century coastal plantations in
South Carolina and Georgia (Harris 1992; Leech et al. 1994;
Newell 1991, 1997; see Table 1 for terminology definitions).
Split-log barges are distinguished by the beam of solid carved
wood (often cypress) shaped to create one-piece vessel
sides. The one-piece sides of split-log craft form the vessel's
angular junctions, provide attachment points for transverse
decking and bottom planking, and constitute the majority of
the vessel's longitudinal support. Conversely, vessel sides of
plank-built flats are comprised with multiple planks placed
on edge. The transverse bottom planking of plank-built flats
fasten directly to the bottom of the side planks instead of into
carved-out shelves and rabbets that are found on split-log
forms. Builders of plank-sided craft often used conventional
knees, drift pins, and framing to replace the missing structural
support. Examples of split-log and plank-built barges and
flats have been documented throughout portions of the South
where slaves, freed men and women, workers, farmers, land
and plantation owners utilized the watercraft in the shallow,
tidally-influenced coastal areas principally to cultivate and
harvest rice. Wacissa Boat 1 is a combination of the two
construction methods as techniques for both are present. This
modified form is further represented by the vessel's carved-out
one-piece side, technically known as an ile.

Research conducted for this project revealed that an
appropriate term to use when describing the carved-out,
one-piece side identified on Wacissa Boat 1 has yet to be
widely established within the archaeological community.
In comparison to similar hull elements located on the Ben
Sherrod barge (Krivor and James 1999), the Wooden Spud
Barge (Leech et al. 1994; Tidewater Atlantic Research 1992),
and the Laurel Hill Barge No. 2 (Harris 1991) for example,
described below, the one-piece side used in the construction
of Wacissa Boat 1 takes on a different form. The sidepiece on
Wacissa Boat 1 is more akin to a plank than a log or timber.
Incidentally, upon our initial visit and documentation of the
site we used, for lack of knowledge of a better term, hull
plank with integrated chine as the description for the plank-
like structural element. This description is apparent in Figure
2. However, the term chine-girder which has been used in
descriptions of previously documented split-log craft does not
apply to the particular sidepiece recorded on Wacissa Boat 1.
Technically, the term does not distinguish the log form from
the built-up form. The traditional French term ile more or less
describes the cross-sectional shape and is the more appropriate
term for describing the sidepiece documented on Wacissa Boat
1 (Mike Alford, personal communication, May 2007). The
carpenters of Wacissa Boat 1, therefore, employed a modified
version of the ile by incorporating a shelf or molded bracket
and adding a sheerstrake. The addition of the sheerstrake to
create useable freeboard was possibly necessitated by lumber
size constraints. Thus, the term ile is used when describing the
extant plank-like sidepiece documented as part of the remains
of Wacissa Boat 1. Equally, the term chine-girder is used when
describing the log form evident on other larger types of split-
log built, flat-bottomed vessels.

Environmental Setting

The crystal-clear headwaters of the Wacissa River surface
approximately 20 miles southeast of Tallahassee in central
Jefferson County from a group of large limestone springs just
below the red clay hills of the Cody Scarp, a relic marine terrace.
From its headsprings, the Wacissa travels approximately 12
miles south through a shallow and broad cypress swamp
before breaking into numerous braided channels whichjoin the
Aucilla River a few miles further downstream. Sandy bottom
covers most of the shallow Wacissa while some exposed
limestone areas and many limestone boulder outcrops occur as
well in the river and swamp.


VOL. 60(2-3)




IT '---c.

j~Ci- e 5

~*A ,

0.4 0 0.4 0.8 Mi

Wacissa Boat 1 (8JE1604) and
Slave Canal (8JE1554) (USGS
Nutall Rise, FL 1955)

Figure 1. Site Location.

2007 VOL. 60(2-3)


Table 1: Glossary of Boat Building Terminology.
barge A long narrow flat-bottomed boat used for transporting freight on rivers or canals.
bitt A strong upright post used for securing lines and cables
bottom planks Planking on the bottom, exterior of the vessel
ceiling planks Loose fitted or fastened interior planking lying on top the stringers or attached to the blocking floor timbers.
chine The angular junction of the bottom and side of a vessel; usually found on flat-bottomed hulls, or those with little deadrise.
(chine log) Can also refer to a longitudinal timber (also referred to as a chine log) located just inside the junction, to which athwartships
(transverse or cross) bottom planks are sometimes fastened. In heavily constructed barge craft, chine logs can also be
constructed with notches to accept the ends of cross ties.
chine-built Type of barge or flat constructed with carved-out, usually one piece sides.
chine girder Beam of solid carved wood shaped to create one-piece sides which form the vessel's angular junction, provide attachment
points for transverse decking and bottom planking, and constitute the vessel's longitudinal structural support. Refers to the
log form of chine-built vessels. Chine-girders are similar to chine-logs but have a carved-out rabbet to accept the ends of
transverse bottom planking. Usually are comprised of several timbers.
chine shelf The shelf created during the carving "out" process of the ile. The chine shelf acts as the point upon which transverse bottom
planking and decking are seated.
chine stringer A stringer which lies alongside the chine of a planked flat.
deck beam Transverse timber running from side to side of the flat to support a decking level.
drift pin A cylindrical bolt, headed on one end, that is slightly larger in diameter that the hole into which it is driven.
ferry craft A boat used to transport passengers, vehicles, or goods across water, especially one operating regularly across a river or
narrow channel.
frame A transverse timber, or line or assembly of timbers, that described the body shape of a vessel and to which the planking and
ceiling were fastened.
freeboard The distance between the waterline and the uppermost portion of the vessel sides.
futtock Upright timbers which provide an attachment area for outer hull side planking and lateral structural support on a planked
barge or chine log barge which has more that a single splashboard strake above the chine log.
general Usually large, heavily-constructed barges that may contain but are not limited to diagnostic features and/or structural
industrial materials such as heavily cast fastenings, bitts, industrial bilge pumps, spud boxes, and wire nails.
barge craft
ile Beam of solid carved wood shaped to create one-piece sides which form the vessel's angular junction, provide attachment
points for transverse decking and bottom planking, and constitute the vessel's longitudinal structural support. Refers to the
built-up plank form of chine-built vessels. French word for the letter "L" which describes the cross-sectional shape of the
structural element.
knee L-shaped timbers which provide structural strengthening and attachment surfaces for outer planks.
lighter A flat-bottomed open cargo boat or barge, used especially for taking goods to or from a larger vessel when it is being loaded
or unloaded.
molded depth The various dimensions of timbers as seen from the sheer and body views of construction plans; the dimensions determined
by the molds. Thus, the vertical surfaces (the sides) of keels, the fore-and-aft sides of the posts, the vertical or athwartships
surfaces of frames, etc.
phosphate Large, heavily-constructed barges related to the phosphate industry that may contain but are not limited to diagnostic
industry barge features and/or structural materials such as heavily cast fastenings, bitts, industrial bilge pumps, spud boxes, and wire nails.
plank-built Type of barge or flat constructed with planked-up sides.
planking The outer lining, or shell, of a hull.
plantation flat A long, narrow flat-bottomed open barge, used for transporting goods on shallow waterways and assisting in the
maintenance and construction of plantation ditches and canals in Tidewater Atlantic region of the U.S.
rabbet A groove or cut made in a piece of timber in such a way that the edges of another piece could be fit into it to make a tight
joint. In regards to a chine-girder barge, the term refers to the grooves cut into the bottom into which the ends of the bottom
planking were seated.
ramp The rising bottom ends found on most barge craft.
rub rail A rail (rather than a board) used on the exterior sides or ends of a vessel to prevent rubbing or bumping up against a dock or
another vessel.
scarf/scarph A joint made by chamfering, halving or notching two corresponding planks or timbers.
scow A freight barge used mainly for transporting bulk materials.
sided The distance of an unmolded surface: the distance across an outer frame surface, the forward or after surface of the stem or
dimension sterpost, or the upper surface of a keel or keelson.
sheer strake The uppermost strake.
spud box Through-hull boxes where spuds (large poles) could be lowered for use in vessel stability
stanchion An upright supporting post.
strake A continuous line of planks, running from bow to stem.
stretcher Transverse inboard floor timber running from side to side of barge providing structural strength and rigidity
stringer A general term describing the longitudinal timbers fixed to the inside surfaces of the frames.
transverse Lying or going crosswise or at right angles to something.
Glossary compiled from two sources: SCIAA, Underwater Archaeology Division, 1992 and Steffy 1994.




Figure 2. Site Plan of Wacissa Boat 1 (8JE1604).

Wacissa Boat 1 is situated on the sandy bottom of one of
the many clear watercourses that make up the Wacissa River
drainage. Located within the Aucilla Wildlife Management
Area in the Big Bend region of Florida, the submerged remains
have slowly decayed to a discreet collection of planks of wood
and rusting iron nails. Many of the recreational enthusiasts
who travel the waterways of the Wacissa River via canoe,
kayak, or small boat have most likely passed over the site,
not recognizing it amidst the relatively undisturbed natural
community of hardwood-dominated bluff, bottomland and
swamp forest associations.

Historical Background

As noted on Figure 1, the vessel remains of Wacissa
Boat 1 are located at the head of an improved portion of the
Wacissa River known locally as the Slave Canal (8JE1554).
A recognizable feature of the Lower Wacissa River, the Slave
Canal is a narrow, shallow canal spanning several miles of
swampy terrain before emptying into the Aucilla River south
of Nutall Rise. Although the canal's construction is attributed
to the work of Jefferson County slaves, use of the waterway

during the mission era in Florida indicates that the shallow
channel was flowing and navigable prior to improvements in
the Territorial and early statehood periods (1830s-1850s). Four
missions existed in the vicinity of the Wacissa headsprings in
the seventeenth and early eighteenth centuries, the Ivitachuco
(8JE100), Ayubale (8JE2), Oconi (8JE131), and the Aspalaga
(8JE1), and all were connected via surrounding waterways
that included the Wacissa River.
Having coalesced throughout the southeast and the
Florida peninsula in the previous centuries, many historic
period Native American tribes relocated to the Apalachee
region during the latter half of the seventeenth century. These
groups sought refuge from political persecution and slave
raids in Spanish held La Florida as colonial powers continued
their struggle for control of what is today the Southeastern
US (Tesar 1979). One of these groups, the Tocobaga, whose
original territory encompassed the area around Tampa Bay,
had by 1678 amassed a village of 350 persons along a channel
of the Wacissa River. In the same year a Spanish official visited
the village and delivered an order from Governor Pablo Hita
Salazar for the Tocobaga to temporarily close the channel of
the Wacissa and withdraw further inland. A pirate attack at

Wacissa Boat 1(8JE1604)
Site Plan 6 April 2006
BAR/ J Moates, R Smith

[iillpl a le
I ~\ Pl -ki.
_/ 1na+-1+ /




2007 VOL. 60(2-3)


the nearby port of San Marcos hastened the order (Bushnell
Despite the pervasive threat of such attacks, the population
of the area continued to grow and a successful trade network
developed between San Marcos, Havana, St. Augustine, and
Vera Cruz. Near the Wacissa headsprings (and up Welaunee
Creek to the Ivitachuco mission), local missions produced
mostly agricultural goods that were sent seaward while needed
daily supplies and Spanish trade goods such as bells, knives,
beads, sheet brass, cloth, razors, and scissors were transported
inland (Bushnell 1990). Trade activities gave rise to the need for
local transportation services to and from the missions and the
Tocobaga settlements. The Tocobaga were already renowned
as coastal pilots and often served as ferrymen for the materials
and people moving up and down the Wacissa, Aucilla, and
Suwannee Rivers (Memory 2000). With activities associated
with missions and Native American settlements dotting the
region, it is likely that the navigable course providing access
to the Wacissa headsprings and today known as the "Slave
Canal" continued to see heavy use.
The activities of Jefferson County plantation owners and
their slaves during the Territorial and Early Statehood periods
in Florida further document the fact that the "Slave Canal"
was actually an enlargement and improvement of the extant
stream used previously to supply nearby mission communities.
Achille Murat and James Gadsden became two of the most
prominent early landholders in Jefferson County after the
new capital was established at Tallahassee in 1824. Gadsden
named his plantation "Wacissa," and soon partnered with
Murat on his nearby plantation, "Lipona" (Memory 2000).
Together with two other early settlers, Robert Gamble and
William Nutall, the plantation owners readied their workers
and developed land for planting crops such as cotton and
sugar cane and establishing saw and grist mills, with trade to
and from these plantations conducted via road to Magnolia, a
small port established upriver from St. Marks (Ibid.).
By 1830, however, these planters wished for a more
feasible and efficient trade route with Gamble presenting
a report to a Federal Internal Improvement Committee
necessitating the establishment of a trade route via the Wacissa
and Aucilla Rivers (Carter 1959). In the report, Gamble writes
that deepening and widening the channel would erase the
difficulty in obtaining boat navigation and prompt a town to
grow up at the head of the Wacissa (Carter 1959). With the
formation of The Wacissa and Aucilla Navigation Company
by 1836, the first attempt to channelize the river proceeded for
several miles before funds ran out (Shofner 1976; Simpson
et al 1934). The name "Slave Canal" has been popularized
locally and is attributed to the remnants of that project-the
results of which are visible today (Figure 3). Improvement
efforts languished for a few years, however by late 1840s
the Second Seminole War (1835-1842) had passed and The
Wacissa and Aucilla Navigation Company was chartered
anew (Memory 2000). Together with the small community
of Wacissa, at the head of navigation, local planters made
plans to consider further enhancement of their long-navigated
channel turned canal. Indeed, a Fourth of July celebration in

Figure 3. Entrance to Slave Canal (8JE1554).

1849 found a Jefferson County official rallying local residents
with his carefully worded prose:

Ere this epoch shall again dawn upon the American
horizon, may this limpid stream bear upon its bosom
a steamer to facilitate our enterprise while in infancy,
and turn upon us navigation's richest blessing
[Shofner 1976].

Although several thousand dollars in stock was sold
for the company and more work on that "limpid stream"
continued into the 1850s, cutting through the limestone ridges
of the Aucilla River was revealed as too difficult and the idea
of further improvements to the canal were finally laid to rest
(Ibid). Use of the canal continued without pause, however, and
near the turn of the twentieth century and into the following
decades the canal was used by everyone from naturalists
to loggers to Prohibition-era whiskey bootleggers with
occasional descriptions or reports of its use finding their way
into the historical record (Stoddard 1969). In 1895, a noted
ornithologist mentioned the canal and the history of its origin
in a report of his travels:

The Wacissa empties into a canal which is five miles
long, which was dug before the war with a view of
deepening the river and making it navigable. This
object, however, was never realized, as the river is
very shallow being, on an average, three feet deep.
The canal described above joins the Aucilla River
and conveys the water into the Aucilla River, and
thence into the Gulf of Mexico [Wayne 1895].

Wacissa Boat #1

In April of 2006, BAR archaeologists investigated the
submerged remains of a flat-bottomed historic watercraft
located in the Wacissa River. Recording efforts began by




establishing a baseline along the site's long axis. Overall the
site measured approximately 20 feet east/west by 10 feet north/
south and was delineated on the southern edge of the remains
by a large structural component of the vessel known as an ile.
The plank-like hull element measured 18 feet in length and the
team captured digital photographs and sketched the site in plan
view. Later, archaeologists conducted a follow-up visit to the
site in March of 2007 to document the ile in detail.
The site of Wacissa Boat 1 is comprised of hull elements
including an ile, an eroded sheerstrake affixed on edge to the ile,
a broken, disarticulated portion of the sheerstrake, transverse
bottom planking, and stringers (Figure 3). Remaining hardware
included several cut iron nails, evidence of two iron spikes,
and a single iron drift pin. Spaced evenly in pairs, the cut nails
remained fastened in the chine shelf portion of the ile. The
iron spikes and the iron drift pin fixed the ile and sheerstrake
together on edge. A square nut is threaded onto either end of
the drift pin.
Split-log construction on antebellum plantations utilized
a single large plank for the side and chine of a flat (Newell
1991). The chine describes the point where hull shape
changes from the bottom to the side of a vessel. As noted
above, some authors today refer to these as iles and chine-
girders. However, the term "split log" was used historically
to describe the principle structural elements of vessels of this
type (Newell 1997). Traditionally, the longitudinal elements
were split from a single log of large diameter and length and
carved to form the vessel sides, including a chine-shelf along
the inside bottom edge. Once completed, carpenters would flip
the structural elements over and fasten the ends of transverse
bottom planks into a rabbet carved into the bottom edge of the
chine-shelf thus forming the shape of the hull.
The remains of the ile at Wacissa Boat 1 measured 18 feet
in length and approximately 8 to 9 inches in width at its widest
point (Figure 4). Made of cypress and badly eroded or worn,
the ile also varied in thickness from less than one inch to no
greater than 2 inches where the iron drift pin remained. The ile,
disarticulated from the rest of the hull remains, did not appear
any thicker than the sheer strake except along the chine shelf.
The eroded chine shelf of the ile also varied in form due to
natural erosion. The sided dimension of (width) the chine shelf
ranged from less than one inch to no greater than 2 inches. Its
molded depth of 2 inches appeared more consistent likely due
to the preserving quality of the cut nails that remained.
Other recorded hull elements include the remains of a
sheerstrake, three stringers, and transverse bottom planking.
The remaining portion of the sheerstrake, attached to expand
the vessel's freeboard, is affixed to the top of the ile. The
badly eroded strake measured 1 to 2 inches in thickness. A
broken portion of the plank remained mostly buried adjacent
to the ile. Three stringers measured 2 inches and 4 inches in
sided dimension and 2 inches molded and remained fastened
to a single bottom plank in a visible portion of the remains.
Recorded bottom planks measured 7 to 8 inches in width and
had a thickness of approximately 1 inch. The bottom planks
were disjoined from the ile.
The remaining hardware includes iron cut nails that appear
to be uniform in size and shape, evidence of two iron spikes,

and a large iron drift pin. The cut nails remain fastened to the
chine shelf and suggest that the construction of Wacissa Boat
1 occurred sometime after 1840 (Nelson 1968). Measuring
no more than 5 inches in length and exhibiting similar square
heads however slightly misshapen due from oxidation, the
nails are exposed to a length of no greater than 3 inches below
the chine shelf where a rabbet has been carved into the ile. Tips
of nail points were also exposed on the upper side of the chine
shelf where a few were driven clear through. Paired together
along the shelf, the cut nails were spaced approximately 7
inches apart with an average of 2 inches between the individual
nails within any given pair. Broken tips of the two iron spikes
are exposed on the eroded upper surface of the sheer strake.
The iron spikes, along with the large iron drift pin, fasten the
sheer strake and ile together on edge. The large iron drift pin
measures 1 inch in diameter and approximately 20 inches in
length. Two square nuts are threaded onto either end of the
Wacissa Boat 1 appears to be the remains of a barge or flat,
most likely of local construction, employed in the movement
of people, cargo, and/or materials through one of the shallow
canals that helps to make up the Wacissa River drainage.
However small, approximately 20 feet in length and 10 feet
in width, the vessel remains indicate construction methods
associated with two barge forms, split-log and plank-built,
concomitant in use and construction in the eighteenth and
nineteenth centuries. Similar examples of both construction
methods have been documented throughout the Southeastern
U.S. However, a barge type which Wacissa Boat 1 more closely
resembles, the plantation flat, has been found to occur in the
low country of South Carolina and Georgia.


The earliest use of split-log barges and flats in American
contexts dates to the early eighteenth century. However, historic
accounts from the later antebellum period detail that era as the
period of their greatest use-utilized especially as plantation
owners shifted rice-producing operations toward the coast and
developed river-fed canal irrigation systems (Newell 1997).
To date, these accounts provide most of the current knowledge
on the use of split-log and plank-built plantation flats and are
acknowledged in "The Historic Small Craft in South Carolina:
a General Typology with a Study of Adaptation of Flatboat
Design." (Ibid.). Wacissa Boat 1, with evidence of transverse
planking and use of an ile, closely resembles the plantation
barge craft documented in South Carolina and Georgia.
Newell mentions that another parallel found for
this construction technique, one which utilized split-log
construction and transverse planking in barge craft, is in Poland
on a tributary of the Vistula River linking this type to early
European history (Litwin 1988). As a construction technique
however, the split-log method appears to vanish from use during
medieval times only to reappear in Colonial America (Newell
1997). The construction techniques visible for Wacissa Boat
1 represent a geographical and temporal expansion as well as
potentially reflecting the latest variation in design of split-log
plantation flats. Importantly, the expansion and augmentation


2007 VOL. 60(2-3)

Inboard Sketch of Remains of lie and Sheer Strake

C.dU -, P i&r S4-.. A..
Icfi5 -3plt fj

Figure 4. Inboard Sketch of Remains of Ile and Sheerstrake (8JE1604).

of these techniques also represent the transference of learned
skills and the adaptable practicality of the builders and users
of these craft.
Several historic barge forms and flat-bottomed vessels
have been recorded by researchers throughout the Southeastern
United States. Referred to as barges, flats, lighters, and scows by
those who built and used them, a few of the types include ferry
craft, plantation flats, coal barges, phosphate industry barges,
and general industrial barge craft (Watts and Hall 1986; Newell
1991, 1997; Harris 1992; Leech et al. 1994; Krivor and James
1999). The variety of names given to barges often reflected the
primary function of the vessel, though in reality these working
platforms would have served an assortment of overlapping
duties and the terminology associated with their construction
and use would often have been used interchangeably.
Accordingly, antebellum boat builders adapted ferry craft
design for use in plantation contexts (Newell 1997).
The architectural features recorded at Wacissa Boat 1 bear
a close resemblance to the overall design of ferry craft rather
than the designs of other types of historic barge forms located
within the same region and those that have been documented
elsewhere. The structural remains of Wacissa Boat 1 such as
the remains of an ile with a planked-up sheer strake to form the
vessel side and its small size overall show characteristics that
are not seen on many of the other documented barge types.
These traits link Wacissa Boat 1 to the construction methods
known to have been associated with antebellum ferry craft
and the types, later known as plantation flats, adapted for use
in low-country rice agriculture in the Tidewater region of the
Atlantic South.
Although similarities in construction and fastening methods
do exist among most of the barge types mentioned above,
Wacissa Boat 1 does not share many of the diagnostic features,
structural materials, or characteristics that are found on or are
related to phosphate industry barges, general industrial barge
craft, and coal barges. For example, some of these diagnostic
features not evident on Wacissa Boat 1 include heavily cast

fastenings, bitts, industrial bilge pumps, spud boxes, and wire
nails. While a review of the archaeological record of some of
these flat-bottomed vessels provides a comparative context
for Wacissa Boat 1, it also reveals similar construction and
fastening methods regardless of barge type or function.
For example, archaeologists with Panamerican Maritime
documented the eroded remains of a large, heavily built coal
barge (Ben Sherrod barge) located in Wilkinson County,
Mississippi in January 1999 (Krivor and James 1999). Most
likely associated in the transportation of coal from Pennsylvania
mines to port towns along the Ohio and Mississippi Rivers in
the mid to late nineteenth century, this barge (111 feet in length
and 22 feet in width) was heavily constructed employing the
use of tightly spaced cross ties, heavy side timbers including a
chine log, and large iron drift pins and spikes. Utilizing a timber-
sided method of construction, the vessel sides were fastened
together using iron drifts to secure the large, rectilinear side
timbers to the top of the vessel's chine log. Although the Ben
Sherrod barge is much larger than Wacissa Boat 1, the use of
iron drift pins to fasten the vessel's sides reveals a fundamental
fastening technique for similarly designed watercraft apart
from vessel size.
Conceptually, the timber-sided method utilized on
Ben Sherrod is comparable to Wacissa Boat 1. Differences,
however, are apparent upon examination of the extant side
of Wacissa Boat 1 that consists of the remains of a single
plank (a sheer strake) and a carved-out ile. In comparison to
the ile, the dimensions of the chine log recorded on the Ben
Sherrod barge measured 6 inches sided (width) and ranged in
molded depth from 7 to 17 inches. These measurements are
considerably larger than the plank-like ile of Wacissa Boat
1-which measured 2 inches in thickness by 8 to 9 inches in
width'. Additionally, carpenters of Wacissa Boat 1 carved a
rabbet into the bottom of the ile to accept and fasten the ends
of transverse planking. Archaeologists at the Ben Sherrod
barge, however, determined that the chine logs were notched to
accept large cross ties. Also contrary to Wacissa Boat 1, outer

\rL- .-- -;-.iI~-~A it ~- it. A 7. a tt- j 7c-~ tiC

Wacissa Boat 1 (8JE1604)
BAR/ JMoates, KPorter
9 March 2007


Inbar Sktc of Rean f i n herSr



tI 116C~bb au~


,kc- ill I P __T
A A r_-e5k~l



hull (bottom) planking was fastened directly to the bottoms of
the chine logs and cross ties thus forming the outer shell of the
hull. In summary, though much different in size, use of similar
techniques and materials associate the Ben Sherrod barge
and Wacissa Boat 1. The barges are set apart typologically,
however, by the fact that carpenters of Wacissa Boat 1 utilized
a relatively intensive construction method of carving out the
entire vessel sides.
A study conducted by the U.S. Army Corps of
Engineers, Savannah District, to determine the significance
of archaeological resources associated with Savannah River
rice plantations assesses the remains of a wooden spud
barge, 82 feet in length and 32 feet in width (Watts 1992).
Archaeologists concluded that this massively constructed,
industrial work barge dates to the late nineteenth or early
twentieth century and may have been associated with the
construction of two bridges which span the study area rather
than daily plantation activities (Leech et al 1994). Construction
of this vessel employed massive timbers including chine-
girders that framed the vessel floor, frames, futtocks, hanging
and standard knees, deck stanchions, floor and deck stringers,
and deck beams. Also, the documented remains include
spud boxes where spuds could be lowered through the hull
for use in vessel stability, tie-rods, and machinery supports.
Together, the remains clearly demonstrate that this vessel was
purposefully constructed for the hauling of heavy machinery
and for the assistance in construction over water. The vessel's
chine-girders measure 14 inches sided (width) and 19 inches
in molded depth and are comprised of more than one timber
as a scarph along the exposed chine is apparent (Ibid). Similar
to the ile documented on Wacissa Boat 1 however, builders of
this craft carved rabbets into the bottom of the chine-girder
to accept and fasten the ends of transverse bottom planking.
Although similarities in vessel side construction exist, the
wooden spud barge displays characteristics that are not related
to the remains of Wacissa Boat 1.
In 1991, archaeologists and sport divers conducted
investigations to record Laurel Hill Barge No.2 located nearby
Laurel Hill, South Carolina along the Waccamaw River (Harris
1991). Associated with nineteenth century rice plantation
activities, the vessel remains measured approximately 56 feet
in length and 15 feet in width and consisted of chine-girders,
several keelsons (stringers), transverse bottom planking, and
ship-like features such as small framing members, knees, and
ceiling planking. But unlike Wacissa Boat 1, treenails were
used to fasten just about everything together. Archaeologists
did record, however, a few cut nails that fastened the
gunwale strake to the upper side strake and a small rubrail.
Even though this barge is relatively large and was most
likely used to carry heavier than usual loads of agricultural
produce than most plantation flats (Ibid; 49), archaeologists
recorded characteristics that are similar to those documented
on the Wacissa boat. Examples of these similarities include
the combination of the chine-girder with planked-up sides and
the use of iron fasteners to secure upper side planks together.
However, the relevance of the ship-like features and use of
treenails for fasteners along the chine and elsewhere within
the hull likely pre-dates the entirely cut nail fastened Wacissa

Boat 1 (Newell 1997). Unlike the two previous barges, Laurel
Hill Barge more closely represents the barge type adapted for
use in plantation contexts.
Similar to Laurel Hill Barge No.2 and Wacissa Boat
1 several barges or flats, with date ranges all through the
nineteenth century, have been documented as falling into one
of three categories; plank-built, split-log, or a combination of
the two (Ibid. 1991, 1997; Leech et al 1994). A clearly written
explanation of the three methods is offered by Leech et al.
and is rephrased here. In the first method, a plank-built vessel
shape is formed by the fastening of the bottom planks to the
side planks. This intersection is known as the chine of the
vessel and is usually reinforced on the inside by a chine log.
Secondly, split-log built vessels involve the carving of a split
log to form an entire vessel side. The log is carved to create the
chine and vessel side and on its bottom inboard side a rabbet is
also carved to accept the ends of bottom planks for fastening.
The third method is a combination of the above two with the
carved split-log serving as the chine and the lowest "plank"
on a plank-sided vessel. Archaeological documentation of
Wacissa Boat 1 illustrates construction methods attributed
to the combination of split-log and plank-built craft by the
presence of the ile, as the lowest "plank" on the planked-up
The sides of plank-built flats were constructed of multiple
thick planks placed on edge, one on top of the other (Newell
1997; Newell in Tidecraft 1995). Builders carefully drilled
holes down and through 2 or 3 strakes with a ship auger.
Iron drift pins could then be driven through and secured.
An illustration of the finishing processes of a plank-built flat
features the installation of iron drift pins (Figure 5). Newell
states that both construction methods (split-log and plank-
built) were utilized at the same time but exclusively split-log
plantation flats seem to disappear in South Carolina as a result
of the Civil War. Plank-built flats and combinations of the
two construction methods, however, appear to have been built
throughout the eighteenth and nineteenth centuries. Like the
carved-out log, an ile was also a "carved-out" timber. Its form
bears a resemblance to a plank, much wider than it is thick
with nearly flat sides. Similar to the split-log cross section
recorded by Newell (Figure 6), the ile of Wacissa Boat 1 could
possibly be the result of a hewn plank than a "split log". Thus,
the presence of the ile, planked-up sheer strake, use of iron
drift pins, and transverse planking suggests that the builders of
Wacissa Boat 1 were familiar with and utilized aspects of both
split-log and plank-built construction methods.
Tidewater Atlantic Research archaeologists documented
several barges during a systematic survey of the Back River
(Watts 1992), where archaeologists found that one of the eight
documented was the remains of a small flat (Leech et al. 1994).
This leftover from Pennyworth Plantation included remnants
of exposed bottom planks, longitudinal stringers, vessel sides,
and a stretcher. Importantly, the vessel sides were constructed
of "two planks that are edge-joined with irregularly spaced
vertical throughbolts" (Ibid.; 199). Measurements of the
planks varied with the upper planks being 1 foot 6 inches and
9 inches wide and the lower "planks" measuring 8 '/ and 10
inches in width. Evident in the small flat cross section, the


2007 VOL. 60(2-3)


Figure 5. Illustration of the Fastening Processes of a Plank-Built Flat Featuring the Use of Iron Drift Pins (as presented in
Fleetwood 1995) (Drawing by William Judd).

Figure 6. Cross Section of Ile of Quarter Ditch Barge at
Conway, South Carolina (as presented in Newell 1991).

lower side "planks" appear "L" shaped (Figure 7). These were
recorded to be a part of the outer longitudinal stringers made
from two separate planks, one of which was noted as a chine
log (Ibid.; 198). Unfortunately archaeologists were unable to
determine how the bottom planking was fastened underneath.
Nevertheless, based on historical records addressed by Newell

and archaeological evidence it can be assumed that the bottom
planks were fit into carved out rabbets on the underside of
the chine shelf and secured with iron nails. In addition, the
fact that the side planks were secured with "irregularly spaced
vertical throughbolts" supports this interpretation (Ibid.; 199).
Bottom and ceiling planking was also noted to run laterally
(transverse) across the vessel. The four longitudinal stringers
are slightly different in size based upon location but their
measurements compare very closely to the stringers located
on the Wacissa boat. All of the bolts recorded on the small flat
measured 7/8 inches in diameter with 1'%-inch square nuts.
Since the fasteners were not treenails, the vessel appears to
date to the last quarter of the 19th century.
Similar to the above described Back River flat, construction
techniques and hardware also date Wacissa Boat 1 to the
latter half of the 191h century. Contrary to the construction
and use of split-log plantation flats in the tidewater regions
of South Carolina and Georgia which are speculated to have a
specific temporal range linked to the period of the plantation
system, evidence suggests that a post-Civil War construction
of Wacissa Boat 1 is likely. Modern machine cut nails began
being manufactured in the late 1830s (Nelson 1968) and were
probably available in North Florida shortly thereafter. The
presence of cut nails on Wacissa Boat 1, as also seen with the
iron drift pins recorded on the small flat in the Back River,
signifies a departure from earlier fastening methods on barge
craft. Treenails prevailed as the fastening method of choice
for most of the documented plantation flats and include those
with possible construction dates determined to have taken
place later in the antebellum period, or closer to mid-century
(Newell 1997).

treena i I
f overlap

.CALE: 15




Figure 7. Cross Section of Small Flat on Back River, Georgia (as presented in Leech et al. 1994).

Additional data on split-log and plank-built plantation
flat construction methods also point to a later date range for
construction and use of Wacissa Boat 1. Newell writes that
extensive adze and axe tool-marks most often appeared
associated with earlier craft along with ship's timbers (standard
and lodging knees) and an absence of iron fastenings (Ibid.;
51). He continues that, "smaller chine-built craft with planked-
up sheer strakes may be pre-dated by larger chine-built craft
that achieved the same hull depth without additional planking"
(Ibid.). All together, vessel location, the use of iron fasteners,
and use of an additional sheer strake support a late nineteenth
century date range for construction and use of Wacissa Boat 1.
Furthermore, Newell's research shows that plank-built barges
can possibly be viewed historically as a response to factors
which affected the use of carved-out timbers in split-log barges.
He writes that the slowed production of new rice fields late
in the antebellum period combined with the labor intensity of
carving split logs, skill of the individual plantation carpenters,
availability of large lumber sizes, and cost may all have taken
a part in the decision to utilize one type of method over the
other. It appears with the combination of the two construction
methods as documented on Wacissa Boat 1, builders continued
to employ traditional split-log techniques while utilizing new
materials, cut nails and possibly hewn planks, and combining
fastening procedures of plank-built and split-log construction
to maximize lateral strength of the planked-up sides.
log and plank-built, are known to have been interchangeable
in serving many of the same functions during the plantation
era in the South. Whether or not plantation workers or
ferrymen favored a construction method over another to serve
a specific function is unknown. It is known, however, that the
two construction methods were used concurrently (Leech et

al. 1994; Newell 1997). As ferry craft, these vessels assisted
operators in a variety of purposes linked to the transportation
of people and cargo from one point of a river, canal, or
intertidal zone to a particular destination-usually the other
side of the river, canal, or zone. Possibly beginning as early
as the 1730s, plantation workers in colonial America adapted
ferry craft to efforts in the construction and maintenance of
tidally-influenced irrigation ditches and canals associated with
rice plantations (Newell 1997). Results of this adaptation are
possibly evidenced by an absence of deck stanchions and a
rise in vessel ramp angles. These vessels became known as
plantation flats. The scant remains of Wacissa Boat 1 had
no indication of stanchions nor could vessel ramp angles be
determined. Even though these structural characteristics are
not apparent, the vessel's location and possible date range
for construction and use place this vessel in the context of
the trade and supply activities that may have surrounded the
nearby plantations and towns of mid to late nineteenth century
Jefferson County, Florida.
Many factors such as function, environment, knowledge
of traditional construction practices, and local economic and
social climates acted in accord to influence historic watercraft.
Small in size, Wacissa Boat 1 is an example of several factors
dictating vessel construction and design. In all likelihood, this
flat-bottomed vessel was built nearby, used atop, and eventually
abandoned within the very same waters. Built sometime near
the Civil War and probably with locally-milled lumber. Wacissa
Boat 1 cannot be positively identified in association with either
episode of the construction or dredging of the "Slave Canal."
However, the discovery and archaeological documentation of
Wacissa Boat 1 indicates that use of split-log plantation craft
was carried on in various forms and locations even after the
plantation system languished across the Southeast.


2007 VOL. 60(2-3)



Barge craft built in the southern Tidewater regions during
the antebellum period were plantation expedient. Construction
practices apparently varied as to if barge craft were constructed
on the plantation, purchased, or outside carpenters were called
in to do the work. Even so, hard working watercraft like barges
are in constant need of repair and maintenance attention.
Construction methods therefore would have been acquired and
passed between extended kin groups and close associates-
members of which could easily have migrated across the
Southeast and into Florida over time. Though producers of
these craft tended to be slaves, there is no direct line of evidence
that knowledge of the split-log construction method and
design were brought with them from Africa or other European
or New World locales with long traditions of slavery (Newell
1997). However, Newell points out that the construction
method might have been more suited to the traditional West
African skill sets and we know that people from specific areas
of West Africa were pressed into slavery specifically because
of their specialized knowledge in rice production, the working
conditions associated with water crops, and their familiarity
with a sometimes oppressive climate duplicated across the
colonial and early American Southeast (Hall 1992).
Geographical and ethno-historical analysis may reveal
further similarities between craft building traditions that
developed among skilled slaves and free craftsmen along the
tidewater Atlantic, the Big Bend Region of Northwest Florida,
and the Gulf Coast-and there are certainly enough parallels
in economic, social, and human conditions to warrant further
comparative research. Plantation flats in Georgia and South
Carolina, for example, were constructed to perform in coastal
plantation environments where irrigation canals and ditches
were prevalent and demanded size and draft restrictions. The
Wacissa River in Northwest Florida is a shallow, spring-fed
river (1 /2 feet deep during investigations). Shallow-drafted
craft are necessary for passage in such an environment, where
fallen trees and shoals present frequent obstacles. Similar to the
barges recorded in plantation contexts in South Carolina and
Georgia, Wacissa Boat 1 is a perfect example of vessel design
dictated by both environmental restrictions and a practiced
knowledge of what works in a particular type of environment.
In the case of Wacissa Boat 1, specific primary vessel function
is unknown. However, initial function may not be as important
as knowing the daily use of the vessel. Flats and barges were
and still are everyday working watercraft and used as ferries,
lighters for transport, and general work platforms to assist in
construction and maintenance of docks, bridges, and ramps.
Historic barges in Florida waters and elsewhere serve as
a physical reminder of the link between historic knowledge
and the flexibility of daily practices. Although these vessels
have had limited archaeological scrutiny, those located
within survey areas or those known locally have not escaped
preliminary archaeological and historical documentation.
These important efforts reveal the frequency of the resource
across Florida's submerged lands and coastlines (see Carr
1974; Curren, Newby, Mikell & Smith 1998; Denson 1992;
Dunbar 1990, 1991; Ellis & Denson 1997; Franklin, Morris

& Smith 1992; Mathewson 1992; Morris & Burns 2001;
Morris, Moore & Eslinger 2004). Attention to this resource
sharpens an otherwise indistinct image of the people who
boarded Wacissa Boat 1 and other similarly utilized vessels in
the progress of everyday life. Considered in context, Wacissa
Boat 1, lying at the head of an improved portion of the Wacissa
River known locally as the "Slave Canal," links a natural body
of water first navigated by native Floridians, later missionized
Indians, and colonists of La Florida, to a waterway sculpted
and channelized by the unnamed slaves and workers of early
statehood and post Civil War Florida and into a landscape that
many modern Floridians recognize today as something of
value and worth revisiting.


1. As expressed by Steffy, protocols for recording
measurements differ between planks and timbers: "Normally,
timbers are expressed in sided and molded dimensions, while
planks and wales are listed in thicknesses and widths. Molded
and sided dimensions are used because of the changing
orientation of timbers, such as frames, where "thick" and
"wide" or "height" and "depth" become confusing" (Steffy
1994: 276).


I would like to thank Robert Daniels, retired Florida Fish
and Wildlife Commission officer, for informing the Florida BAR
about the vessel remains and escorting us through the Wacissa
narrows. Roger Smith and Kevin Porter ably assisted with the
site documentation. A thanks also goes to Mike Alford and his
technical explanations. I was inspired by the investigations of
Mark Newell into historic watercraft traditions of the southeast
and thank him for his comments on my own research. Others
to whom I am thankful for reading this manuscript and offering
guidance, comments, corrections, and comparisons are Lewis
Tesar, Jim Dunbar, Jason Burns, Michael Krivor, and Roger
Allen, as well as Deborah Mullins.

References Cited

Bushnell, Amy T.
1990 "How to Fight a Pirate: Provincials, Royalists,
and the Defense of Minor Ports During the Age of
Buccaneers." Gulf Coast Historical Review. Vol. 5,

Carter, Clarence E., compiler and editor
1959 The Territorial Papers of the United States, The
Territory of Florida 1828 1834. Vol. 24. The
National Archives and Records Service. Washington,

Fleetwood, William C.
1995 Tidecraft: the Boats of South Carolina, Georgia,
and Northeastern Florida, 1550-1950. WBG Marine
Press. Tybee Island, GA.




Hall, Gwendolyn Midlo
1992 Africans in Colonial Louisiana: the Development
of Afro-Creole Culture in the Eighteenth Century
(L.S.U. Press: Baton Rouge)

Harris, Lynn B.
1992 "The Waccamaw-Richmond Hill Waterfront Project
1991: Laurel Hill Barge No. 2 (38GE420)." Research
Manuscript Series 214. South Carolina Institute of
Archaeology and Anthropology, Columbia.

Jones, B. Calvin
1987 Letter to T Buckingham Birdregarding the importance
of the San Lorenzo de Ivatchuco Mission Site. On
file with site form 8JE100, with Florida Bureau of
Archaeological Research, Division of Historical
Resources, Tallahassee.

Krivor, Michael C. and Stephen R. James Jr.
1999 "Recordation and Assessment of a Historic
Watercraft at the Ben Sherrod Site, Wilkinson
County, Mississippi." Panamerican Maritime, L.L.C.,
Memphis, Tennessee.

Leech, Richard W., Jr, Judy Wood, Gregory D. Cook, Kyra
Bowling, & Taras Pevny
1994 Archival Research, Archaeological Survey, and Site
Monitoring, Back River, Chatham County, Georgia
and Jasper County, South Carolina. Savannah District,
U.S. Army Corps of Engineers, Savannah, Georgia.

Litwin, Jerzy.
1988 Changes in Boatbuilding Techniques along Two
Tributaries of the Vistula: The Dunejic and the
Wisxova." 4'" International Symposium on Boat &
Ship Archaeology Porto, 1985. BAR Series 438(i).
Octavio Lixa Filgueras Ed. British Archaeological
Society, London, UK.

Memory, Melissa, Jonathan Lammers, and William Stanton.
2000 "A Second Progress Report (1998-1999) on the
Wacissa River Survey, Including Historical Notes,
Aucilla Wildlife Management Area Jefferson County,
Florida." C.A.R.L. Archaeological Survey, Bureau of
Archaeological Research, March.

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

Newell, Mark
1991 "Preliminary Documentation of a Chine Girder Barge
at Conway, South Carolina." Research Manuscript
Series, Draft. On file at South Carolina Institute of
Archaeology and Anthropology, Columbia.

1997 The Historic Small Craft in South Carolina: a General
Typology with a Study of Adaptation of Flatboat
Design. University of St. Andrews. Dissertation.

Olds, Doris L.
1962 History and Archaeology of Fort Saint Marks in
Apalachee. FSU Masters Thesis on file with Florida
Bureau of Archaeological Research, Division of
Historical Resources, Tallahassee.

Shofner, Jerrell H.
1976 History of Jefferson County. Sentry Press.

Simpson, R.C., Mary McRory, Edith Barrows, Lena Clements,
and Mary Bowen
1934 History of Jefferson County, Florida A Project of
the Federal Emergency Relief Agency. Published
under the Auspices of the Kiwanis Club. Monticello,

South Carolina Institute of Archaeology and Anthropology
1992 "Rice Flat/Barge Glossary." Underwater Archaeology
Division. Columbia, South Carolina.

Steffy, J. Richard
1994 Wooden Ship Building and the Interpretation of
Shipwrecks. Texas A&M University Press. College

Stodard, Herbert L., Sr.
1969 Memoirs of a Naturalist. University of Oklahoma
Press. Norman.

Tesar, Louis D.
1979 The Leon County Bicentennial Survey Report:
An Archaeological Survey of Selected Portions of
Leon County, Florida. Performed for the Florida
Bicentennial Commission, city of Tallahassee, and
National Park Service by the Bureau of Historic Sites
and Properties, Division of Archives, History and
Records Management. Florida Department of State.

Watts, Gordon P., Jr.
1992 "Remote Sensing and Low Water Survey, Back River
and New Cut, Chatham County, Georgia and Jasper
County, South Carolina." Draft Report on file at
Savannah District, U.S. Army Corps of Engineers,
Savannah, Georgia. Tidewater Atlantic Research,
Inc., Washington, North Carolina.

Wayne, Arthur T.
1895 "Notes on the Birds of the Wacissa and Aucilla River
Regions of Florida." The Auk: A Quarterly Journal of
Ornithology. Vol.12, No.4. pp. 362-367.


2007 VOL. 60(2-3)



'Department ofAnthropology, University ofAlabama, Tuscaloosa, Alabama 35497
Email: mary.spanos@ua.edu

Department of Clothing, Textiles, and Interior Design, University ofAlabama, Tuscaloosa, Alabama 35497
2Email: vwimberl@ches.ua.edu, 3Email: athompso@ches.ua.edu


The Osceola Garter is an intact example of a garment
that can usually only be seen in Native American portraiture
from the nineteenth century. Through the events associated
with Billy Powell, also known as Osceola, one of these garters
is intact and also available for study. Other researchers have
documented Osceola's life, as well as the events surrounding
his untimely death and the removal of the Seminole Indians
from Florida. Curious readers could consult two old, but
fascinating issues of the Florida Historical Quarterly
(volumes 33 and 34 published in 1955 and 1956), which were
devoted to the research of Osceola, and also see Foreman
(1989) and Wickman (1991). Therefore, this research will only
briefly summarize the history of Osceola along with the period
when this garter entered the historical arena and will focus on
a technical analysis of this rare Southeast Indian textile, which
provides new insight into the traditional garments and textile
techniques of historic Southeastern Native Americans.
Billy Powell was bor near present-day Tuskegee,
Alabama, around 1804 and took the name asT.yaholi, or
Osceola, as a teenager (Wickman 1991:xx). Although he was
not a hereditary chief among the Indians in the areas where
he lived, southern Alabama and northern Florida, he became
a charismatic leader who fought for Indian heritage, identity,
and territorial rights. In November 1837, Osceola, along
with a group of Seminole and Mikasuki chiefs and warriors,
attended a meeting with United State's General Thomas Jessup
under a flag of truce. They met to discuss a peaceful end to
the hostilities, which had resulted from the Indian's desire to
remain in Florida and the United States' desire to supplant these
indigenous people in favor of European American settlers. In
lieu of peace negotiations, General Jessup captured the chiefs
and warriors and held them in the fort at St. Augustine. After
eighteen people escaped in a prison break, the remaining
captives were transferred along side more than one hundred
warriors and over eighty women and children to Fort Moultrie
on Sullivan's Island in South Carolina. Sadly, Osceola died as
a captive at Fort Moultrie on January 30, 1838, from what was
reported as "an affection of the throat" (Foreman 1989:358,
Wickman 1991:xxv).
The object of this research, a textile artifact known as
Osceola's garter, is an accessory that was worn tied around
the calf just below the knee (Figure 1). The garter's recorded

history begins with Dr. Fredrick Weedon, a colonel in the U.S.
army (Ward 1955:193) and the attending physician to Osceola
at his death. At that time Dr. Weedon came into possession
of a group of Osceola's belongings that included the garter.
The garter was passed on to Dr. Weedon's son, Hamilton, who
gave it to his son, Hamilton, Jr., who gave it to his daughter.
The daughter, Mrs. Robert Blount, donated it to the Alabama
Department of Archives and History in Montgomery, Alabama,
where it is on display today (Wickman 1991:125-128).
Since more than a few faked artifacts have claimed
association with Osceola (Milanich 2004:53), it is
advantageous that the garter's provenance is short and
uncomplicated; it passed from its famous owner into one
family which retained possession until it was placed in the
trust of a public institution. Unfortunately this history fails
to provide information about when, where, and by whom the
garter was made. It was produced sometime before 1838 since
it was in Osceola's possession at his death early that year. Was
it made by a Creek Indian where Osceola was bor and raised,
or by a Seminole craftsman where Osceola lived as a grown
man, or by another Indian passing through Fort Moultrie who
gave it as a gift to an admired leader? Any of theses scenarios
could be correct as evidenced by one of three coats owned by
Osceola originating from the Great Lakes region, and the other
two being of Seminole origin (Sturtevant 1956:316). Many
researchers assume that textile production was performed by
women, however there is no documented historical evidence
of Creek, Seminole, or Mikasuki gender roles in this type of
accessory production.
Setting aside the question of the origin of Osceola's
garter, this paper presents a technical analysis of this historic
textile that focuses on the information that is discernable
from the fiber content, yarn structure, the fabric construction
methods, and its incorporated beads. Although the data that
can be obtained through non-destructive measurement and
testing leave some questions unanswered, structural analysis
can provide insight into the traditional garments and textile
techniques of a culture that was in transition as their homeland
was exchanged between the Spanish, British and the United
States. After decades of contact with Europeans and European
textiles, traditional methods of costume preparation were
undoubtedly modified, changed, or altogether replaced. The
garter is an example of the finger weaving technique which was
well established in Native American traditions as evidenced by


VOL. 60(2-3)




Figure 1. The Osceola Garter on display at the Alabama Department of Archives and History in Montgomery, Alabama
(accession number 85.17.1).

examples from various archaeological sites, such as Wickliffe,
Kentucky (Blackard 1990; Kuttruff 1990: 3-7; Drooker 1992:
29-31; Downs 1995: 121).

Native American Garters and Textiles

Textiles embody information concerning their composition
and the technology needed in their preparation. Fibers, dyes,
treatments, yar production, and fabric formation processes all
can be determined from physical and chemical examination of
the textile product (Emery 1995; Sibley et al. 1991; Drooker
1992; Kuttruff 1993; Sibley et al. 1996; Song et al. 1996).
Information gleaned from textiles has been used to understand
cultural manifestations, such as status (Schreffler 1988;
Cassman 2000; Kuttruff 1993), boundaries between cultural
groups (Church 1984; Petersen and Wolford 2000), socio-
technical change (Hyland and Adovasio 2000) and specific
cultural activities, which include labor expenditure, craft
specialization, and burial rituals (Schreffler 1988; Sibley et
al. 1991; Thompson and Jakes 2005).
In addition to the artifacts recovered from archaeological
contexts, historic and ethnographic reports also provide
information about textiles, including their production and
utilization. Many early travelers to the North American
continent recorded their observations ofNative American tribes
and their life ways (Kalm 1937, Le Page 1972, Williams 1930),
and archaeologists and ethnographers have added to this body
of information (e.g. Swanton 1969, Densmore 1974, Erichsen-
Brown 1979). Descriptions of the attire of Southeastern
Indians by early European explorers and settlers include the
use of an accessory tied around the upper calves or the use
of body paint to decorate this area of the leg (Hulton 1984:
Plates 47 and 48). The use of the word garter likely reflects
these Europeans' own familiarity with the accessory item that
men used to hold up their hose (Wilcox 1958). Specifically, in

English terminology, a garter is defined as "items of clothing,
usually ties or band, used to hold up stockings. Garters have
been in use since the middle ages. In the seventeenth and
eighteenth centuries, garters were generally long, narrow
bands" (Baclawski 1995:114).
Martin and Mauldin provide multiple Creek or Muskogee
terms for the English word 'garter': "est-ele-svwvnakv, est-
ele-sohwnvakv" (Martin and Maudlin 2000:243). Although
Sturtevant's much earlier description of Seminole men's
clothing including photographs and drawings that showed
garters, his text discusses moccasins and leggings, but not
garters. Sturtevant adds that the Seminole moccasins and
leggings are "typologically indistinguishable from moccasins
and leggings of the Creek, Cherokee, and perhaps Choctaw...
they represent a persistence of pre-Seminole Creek patterns"
(Sturtevant 1987:161,170). For simplicity, this article will
continue to use the term garter for the object under study.
Le Page Du Pratz, the French entrepreneur who lived in
Louisiana in the early eighteenth century, reported that native
babies wore garters made of very soft buffalo's' wool that
covered three to four inches of their legs, from just under the
knee to above the ankle bone (Le Page 1972:308). Le Page
also records that men's garters were decorated with small
kernelstones that resembled porcelain beads. Figure 2 is a
portrait of Osceola wearing this type of garment over cloth or
leather leggings. In this example Osceola is shown wearing
the Native American garter over buttoned leggings (more in
the European manner of the garter as a functional accessory
that covered the gap between the breeches/trousers and the
leggings), a Native American sash at his waist, silver gorgets
hanging from his neck, and a European coat. It was fashionable
for young Native American men across the South during this
time period to wear a combination of European and Native
American clothes (Perdue 2004:714).


2007 VOL. 61)(2-3)


Figure 2: Osceola by George Catlin. Courtesy of the American Museum of Natural History (image number 909).



F 2-

Figure 3: William McIntosh, Creek, 1820's, oil painting
by Nathan and Joseph Negus. Courtesy of the Alabama
Department of Archives and History, Montgomery, Ala-
It is evident from a survey of Native American portraits
from the nineteenth century that the garter was a common
accessory of dress used by warriors and chiefs from the
Cherokee, Creeks, Choctaw, and Seminole Indians. Portraits
collected of Fundaburk's Native Americans in Southeastern
Indians Life Portraits: A Catalogue of Pictures 1564-1860
show diamond patterned garters worn just below the knee,
tied either directly in front of the leg or just to the outside
front of the leg. A portrait of McIntosh in Figure 3 shows
this type of garter. Works published by McKenney and Hall
show both Osceola and Tuko-see-mathla wearing diamond
patterned garters tied over cloth or leather leggings just below
the knee. Figure 4 is one such portrait. Bushnell documented
two Southeast Indian bags that are held in two British museum
collections that appear similar to the Osceola Garter, with
chevron patterns outlined by white beads that are curated
in museums in England (Bushnell 1906:177-180). The yar
in the bags was determined to be buffalo fiber and Bushnell
believed they were the only existing examples of buffalo fiber

Figure 4: Tuko-see-mathla, a Seminole chief. Courtesy of
the State Library and Archives of Florida.

textiles from Indian groups east of the Mississippi River. The
garter's chevron pattern may also denote cultural traditions in
its design. This pattern is thought to represent the markings of
a rattlesnake (Goggin 1955:182) and one Seminole informant
reported that Osceola belonged to the rattle snake sib or clan
(Sturtevant 1955:206-207).
Today, if one looks past the deterioration to imagine
this garter as it was when Osceola wore it, then one sees a
well-made and cleverly constructed accessory for the well-
appointed 19th century Native American man. Osceola's garter
has deteriorated quite a bit since he owned it; there are large
holes across the center where the thin yam appears to have
failed. Some years ago the curator at the Alabama Department
of Archives and History tacked the garter to an unbleached
muslin ground and then affixed the whole to an archival board
to stabilize the remaining pieces and prevent further stress on
the frail yarns (Bob Carson, personal communication 2005).
The oblique interlaced fabric structure is still visible in much
of the remaining fabric and it appears that beads outlined the
chevron pattern created by the use of two colored yars. Each
end of the garter is essentially the same: two rows of twining
that provide stability and a transition between the oblique
interlaced fabric and the four-strand braids that create the
fringed edge.


2007 VOL. 64)(2-3)

r I

Figure 5. A map of measurements and locations where fabric and yarn were measured and fibers were gathered for the
analysis of Osceola's Garter. Circles labeled 1D and 1L, etc., indicate where dark and light yarn diameters and yarn-per-cm
measurements were taken. Circles labeled LFS and DFS show the location of light and dark fiber samples. Lines with ar-
rows and centimeter labels provide width and length measurements.

Method of Analysis

The analysis of the artifact was conducted at the Alabama
Department ofArchives and History in Montgomery, Alabama.
With approval, a few fibers were removed from the outer
most fringe areas so as to not disturb the fabric construction
and were taken to the University of Alabama in Tuscaloosa,
Alabama, for analysis by scanning electron microscopy.
Established test methods exist for textile related attributes
such as fiber types, yam size, yam twist direction, and density
of fabric construction through the work of the American
Society for Testing and Materials (ASTM 2006)." Ideally,
measurements should be taken randomly over the complete
object in a minimum of five different locations and using
the 10% rule for avoiding selvedges (the finished edge of
the fabric) where distortions can occur in fabric construction
(ASTM 2003). For the garter, the measurement sites were
selected by visually choosing five undamaged sites spaced
as far from each other as possible. The data collected for
the Osceola Garter included: fiber analysis, yam types and
attributes, fabric attributes and the various length and width
measurements of the garter in its present state.
All garter measurements were taken on site digital
calipers and a Bausch & Lomb lx-2x stereo-microscope with
10xW.F. lenses. Fiber analysis included light microscopy with
a Leica DMLS microscope with an attached Sony CCD-IRIS
video camera at 20x magnification and a FEI Quanta 3D dual
electron ion beam instrument environmental scanning electron
microscope (ESEM) to view surface structure and measure
fiber diameter. Yam diameter was measured using the FEI XT
Docu analysis software.

The yam attributes observed included the yarn diameter
(to indicate the thickness of the yam), the number of plies
comprising the yam, the initial and final twist direction of
the yam, and the final twist angle of the yam. To create yam
similar to that used in Osceola's Garter, a hand-spinner creates
single strands of yam and then twists those strands together to
produce multi-ply yams. A two-ply yar is made by twisting
two single strands of yam together and a three-ply yam is
made by twisting three singles strands together, and so forth.
The process of spinning a single strand of yar (often referred
to as "singles") or plying a multi-ply yam requires the hand-
spinner to twist the yar in one of two directions. The common
textile terms for these two directions are S (counter clockwise)
and Z (clockwise) where the slant of the diagonal portion of
each letter indicates the slant of the visible twist in the spun
fibers or plied strands. The twist angle indicates the amount of
twist inserted in the yam during the spinning and the plying
processes. The twist angle was measured with a protractor
reticle installed in the microscope ocular. This thin lens with a
protractor printed on it allows the protractor to appear overlaid
on the artifact making precise measurements possible.
Fabric attributes analyzed included both the inter-working
elements per centimeter in vertical and horizontal directions
and the fabric construction techniques employed. Surface
embellishment and selvage construction were also recorded.
Wear patterns and subsequent repairs were noted.
A life-size photograph of the garter was used during the
site visit to record the location of all of the analysis locations
(Figure 5). The labeled arrows show the measurements of the
length and width of the garter. The circles that are labeled 1
through 5, L and D, show the five areas where the light (L) and




dark (D) yars and fabric were measured. The ovals on the
right side of the map show where the light fiber sample (LFS)
and the dark fiber sample (DFS) were removed. The fabric
and yar attributes were measured at five different locations
on the garter in order to produce a mean value to provide a
summary description of the fabric and yet be representative of
the variation found within this textile.
While deterioration of the historic textile artifact is
unfortunate, the holes and frayed areas provide visual access
to the interior of the textile, which is often hidden and
inaccessible in a more compact, intact textile. Upon close
examination of the artifact contemporary threads may be seen;
these are visible in some of the photographs presented here.
These are the threads used to secure the fragile garter to the
undyed and desized cotton cloth stabilizer.

Textile Analysis Results

Artifact Description

The garter is approximately 27.5 cm long and 18.5 cm
wide in the center at its widest point. The width at the end
of the fabric, 10 cm, may be a better estimate of the actual
width of the garter when in use, because the oblique interlaced
structure creates a fabric that is similar to standard interlaced
fabric used at an angle ("on the bias" is the textile term). When
interlaced fabric is used at such an angle it provides some
stretch and body-hugging capabilities (Humphries 2004:253).
Such a malleable structure is more difficult to measure in a
way that reflects the actual length or width of the item during
use. The length and width measurements provided here can
only reflect the current state of the garter, at rest and secured to
another cloth. The width and length measurements are likely
affected by the damage to the artifact and its attachment to the

Figure 6. During the process of weaving oblique interlaced
fabric, such as the Osceola Garter, when an active yarn
reaches the selvedge edge, it is bent back toward the fabric
to become one of the passive yarn elements. Note the finer
commercial thread securing the garter to the cloth stabi-
lizer. The small diagram at right shows the source location
of the close-up image.

The construction of this textile was quite ingenious and
efficient (Figure 6). The yars begin on one end in braids, they
pass through a twined section2 that secures them and arranges
them for the transition into the fabric body of the garter where
the light and dark yars form a chevron pattern (Figures 7 and
8). At the end of this fabric the yars pass through twining
again and end once more in braids.

Figure 7. At each end of the oblique interlaced fabric, two
rows of twining provide an effective transition between
fabric and fringe. The small diagram at right shows the
source location of the close-up image.

Figure 8. The two colors of yarn were manipulated to cre-
ate a chevron design. The small diagram at right shows the
source location of the close-up image.

One area, in the lower left of the body of the garter as
photographed here, appears to have been repaired with
commercial sewing thread (Figure 9). There is no record
at the Archives of this repair being performed there, so it is
assumed that the repair was performed prior to the garter being
given to the Archives. In Figure 9, two commercial threads
are visible, one lighter and one darker. The lighter thread is
the repair thread and the darker thread is securing the garter
to its stabilizer. The repair appears to be random stitches that
repaired a hole or tear at the corer of the garter. The braids in
this area appear to have been re-braided since they are the only
3-strand braids on the garter.


2007 VOL. 60(2-3)


Figure 9. At some point, commercial thread was used to
repair one end of the Osceola Garter. The repair thread is
the thick tan commercial thread. The fine dark commer-
cial thread is the new thread used to attach the garter to
the stabilizing fabric. The small diagram at right shows the
source location of the close-up image.

Figure 10. Magnified images of fibers removed from the
light (top) and dark (bottom) yarns suggest that the fibers
used in the Garter are protein in composition.

- "



Figure 11. ESEM micrograph of modern buffalo fiber. The
scales are well defined and intact.

Fiber Evidence

Optical microscopy revealed that both the light and dark
fibers have scales and medulla indicating the fiber to be wool
of some type (Figure 10). Although the general configurations
of both the light and dark fibers appear similar, the scales of
the dark fiber were less distinctive. The scales on the dark fiber
were primarily visible where they appeared to be sloughing off
due to deterioration.
Since the ethnographic record (Le Page 1972:308) states
that buffalo fiber was used to make garters, a variety of buffalo
fibers were obtained and compared to the fibers collected from
the garter. ESEM Micrographs of modem buffalo fiber (Figure
11) and fiber from a 19th century buffalo robe from a private
collection (Figure 12) were made for comparison. The ESEM
images of the Osceola Garter fibers were also compared to
published images of other fibers (Hatch 1993: 143), including
sheep's wool (Figure 13).
The modem buffalo fiber has well defined scales. The
sample from the 19th century shows some degradation of the
scale edges. Comparison of the fibers from the garter revealed
that the light garter fiber had similarities with the buffalo fiber
(Figure 14). The dark fiber micrographs (Figure 15) revealed
that the scale pattern is still visible, even with the loss of the
scales due to age or possible damage during the dyeing process.
The dark fiber sample also shares more similarities with the
modem buffalo fiber than with images of any other fiber.
The sample fibers from the light yam in the Osceola
Garter showed a mean diameter of 36.9 micrometers, fibers
from the dark yam averaged 39.1 micrometers, and fibers
from the 19th-century robe averaged 39.4 micrometers (Table
1). While the fibers in the garter appeared to be consistent
over the entire textile, only one light sample and one dark
sample could be removed, so the mean diameter provides a
fiber diameter estimate of the fiber in the area from which the

THE FOI A O 20 VL 60(2-3)
e93:8m 2O x,0Ok 1.m 10 16 16 AMf 21 .m

Figure 12. ESEM micrograph of 19th century buffalo fi-
ber. The scales are well defined and intact despite the age
of the fiber.

Figure 13: ESEM micrograph of 2006 sheep's wool fiber.

Figure 14. ESEM micrograph of light fiber from Osceola's
Garter. The scale pattern is still discernible and similar to
the buffalo fiber shown in Figures 11 and 12.

Figure 15. ESEM micrograph of dark fiber from Osceola's
Garter. The scale pattern in this sample is less apparent
than in the previous three micrograph images.

Table 1. Fiber diameter comparison between thel9th-century buffalo-fiber robe and the light and dark fibers from Osceo-
la's Garter.
19th century Osceola light Osceola dark
Robe fiber fiber
Mean diameter (micrometers) 38.4 36.9 39.1
Standard deviation .4 2.0 .9
Measurements averaged 9 14 13


2007 VOL. 60(2-3)


sample was taken, rather than being representative of all fibers
in the garter.
While one does not typically associate buffalo with the
American Southeast, there are historical records of buffalo in
"central Alabama, Florida, and southern Georgia" (Rostlund
1960:395). It is tempting to assume that buffalo fiber would
have been more highly valued that plant fiber, based upon
various aspects of acquisition and performance properties of
protein fibers such as resilience, warmth and dyeability; but
this may not have been the case. If value were determined by
the amount of labor necessary to prepare the fiber and produce
the yam, then more common Native American yarns made
from plant fibers would be more costly since these required
more labor to extract from plant material (Spanos 2006:19-

Yarn Structure

The yarns used throughout this textile are two-ply yars,
Z-spun as single strands and twisted in the S- direction when
those single strands were twisted together to make what is
commonly referred to as a plied yam (Figure 16). Looking at
the yars within the fabric microscopically, some Z-twist in
the singles was still visible in the plied yams, indicating that
the plying did not completely balance the original twist in the
Light and dark yars were measured in five different areas
(locations 1L and ID through 5L and 5D indicated on Figure
5). All yam measurements were taken from the woven fabric
area and not from frayed or damaged areas. On average, the
width of the two light yarns used together in the weaving of
the garter were 1.2 mm wide, while the two dark yarns were

Figure 16. The two-ply yarns used throughout the Osceola
garter were Z-twist singles plied in the S direction. The
small diagram at right shows the source location of the
close-up image.

1.4 mm wide (Table 2). Although the light and dark yarns
measured slightly different widths, all of the yar within
the garter, including those threaded through beads, appeared
visually to be two colors of the same yarn.
When yam is spun by hand, the hand-spinner chooses
how much twist to allow in the yam as he or she is spinning.
Loosely spun yars have a low angle of twist, 10 degrees or
less, a medium twist angle is between 10 and 25 degrees and
tightly spun yarns have a high angle of twist of 25 degrees
or more (Figure 17). The light yams in the Osceola Garter
averaged a twist angle of approximately 11 degrees; while the
dark yarns averaged approximately 27 degrees.
The yarns have the appearance of being soft yam created
through the use of low twist single strands and low twist
plying. In other words, they were not firm and dense like
commercial cord; rather they were soft and lofty. This type of
yam is often more difficult for a handspinner to produce, as it
requires both greater attention to consistently inserting a small
amount of twist and greater care in handling the spindle in
order that the low twist singles are not broken before they can
be strengthened by plying (Spanos 2003).
The disadvantage of low-twist soft yams is that they do
not stand up to abrasion as well as firmly spun denser yams.
The choice of a soft yam for this garter may have created a
softer and more flexible textile, but ultimately these low-twist
years may have contributed to its current condition, as the
glass beads might have more easily torn through this yar as
the garter was repeatedly worn.
These yarns may have been produced in a variety of
methods: they may have been hand-spun with a spindle

Figure 17. Low, moderate, and high twist angles (adapted
from Emery 1996:12).

Table 2. Yarn measurements collected from Osceola's garter (1L through 5D indicate locations shown in Figure 5).
Mean Mean
IL ID 2L 2D 3L 3D 4L 4D 5L 5D Man a
Light Dark
Yarn Diameter (mm) .7 .6 .6 .8 .6 .7 .6 .6 .6 .8 .6 .7




or a spinning wheel, or they may have been produced in a
commercial mill. The visible scale structure and mean fiber
diameter of the garter's fibers are consistent with a sample
of 19'h century buffalo fiber. If the fibers are buffalo fiber, as
the authors argue, then the yams were probably hand-spun
either with a spindle or a spinning wheel and would not have
been mill spun (there are no records of buffalo fiber being
mill spun during this period). Hand spindles with clay whorls
were documented tools of the Southeast Indians (Williams
1930:453) and the accounts of property lost during the Indian
removal in the early 1800s mentioned spinning wheels owned
by Indians (Foreman 1989:261). In fact, as early as 1817 the
Choctaw built and used 2,000 spinning wheels for spinning
cotton (Carson 1999:79). Both spinning wheels and spindles
were readily available and a skilled handspinner could have
created the yams used in Osceola's garter with either tool. If
the fibers were sheep's wool, then in addition to these hand-
spinning sources must be added the possibility that the yams
could have been spun in mills in Europe or New England
and traded to the Indians. Research into identifying spinning
technology in archaeological textiles has provided tools for
identifying spindle-spun and thigh-spun yams when bast fibers
are used (Tiedemann 2001:171-175). Unfortunately, a scientific
method for distinguishing between spindle-spun, wheel-spun,
and mill-spun animal fibers is not currently available.

Fabric Structure

The fabric structure of Osceola's garter is made by
oblique interlacing (Figure 8). Oblique interlacing is a flat
form of braiding related to the three-strand braid commonly
used to braid hair today. Starting from a common point yam
elements pass over and under each other as they cross from
side to side at oblique, or inclined, angles to the selvages.
It is this non-perpendicular angle of the intersecting yam
elements that distinguishes oblique interlacing from weaving
(Emery 1995:62). Warp and weft are the weaving terms for
the vertical and horizontal yam elements in woven fabric
and are not typically used to describe the yam elements in
oblique interlaced structures. The reason that these terms
are less applicable for oblique interlaced structures is that in
this braiding technique, every yar performs both functions;
sometimes a yar is the warp and then later it will become the
While the oblique interlacing technique is a braiding
technique, it is such a wide braid that it resembles a woven
fabric. The essential difference in braiding and weaving is
that in braiding there is one set of yams that are interlaced to
create the fabric, while in weaving there are two sets of yams
(the yams mounted on the loom are called the warp and the
yam that is woven over and under the warp yams is called
the weft). The particular form of oblique interlacing used in
the Osceola Garter is further complicated by the method used
to produce the chevron pattern. The interlacing was done so
that some of the yars were hidden by other yams; this gives
the appearance of warp-faced fabric, which is produced when
the weaving is done tightly so that the warp yams completely
hide the weft yars. In the Southeast Indian culture, the skill

that produced this garter may have been widely held; not
discounting its complexity, but recognizing the level of skill
common throughout a society that had produced all garments
and household textiles by hand prior to European contact.
At each end of the woven fabric section of the garter
there are two rows of compact twining (Figure 7), which
may have been used to provide stability and yar spacing
as the interlacing began and then used again at the finish to
provide a consistent appearance and a transition from the
interlaced fabric to the braids. The twining rows were made
using groups of three or four of the light and dark yams. In
its use in Osceola's garter, twining is accomplished using two
bundles of yams that are twisted together to enclose a bundle
of vertical yams (Figure 18). The pair of twining yam bundles
was twisted in an S twist direction. Since both ends of the
fabric are similarly constructed, there is no indication of which
end was the starting point and which was the finishing point.
The yars that appear in the braided fringe are the actual
yams used to construct the body of the artifact. There are
thirteen braids on the end of the garter to the left of Figure 1
(the end that was repaired, see below) and fifteen braids on the
end that appears on the right. The braids in the repaired area are
3-strand braids, while the rest of the braids are 4-strand braids.
The flat half overhand knots (Gilmurray 1981:148) encompass
all of the yarns in each braid but they were made using only a
portion of the total yarns in each braid bundle (Figure 19 and
20). Two to six cm of yam extend beyond the knots. Some of
the braids contain yams of both colors and thus, appear to be
striped. Looking at the braided fringe under the microscope,
many fibers are broken at points in the braids and knots where
they were bent at the greatest angle; a deterioration probably
due to the age of this garter.

Figure 18. The process of twining.


2007 VOL. 60(2-3)


Figure 19. The four-strand braids were secured with flat
half overhand knots. The small diagram at right shows the
source location of the close-up image.


Figure 20. A flat half overhand knot.

A weaver makes many choices when designing cloth,
such as the diameter of the yam to uses and the number of
yarns per centimeter. One thick yam of a particular diameter
can be replaced by two thinner yams that together provide
the same width in the finished fabric. The technique of using
two yars in place of one typically creates a thinner and more
flexible fabric. Threads per centimeter is a measurement that
is used to describe the density of the yarns within the fabric or
how closely adjacent parallel yars lie to each other (Seiler-
Baldinger 1994:7). With the damage to the Osceola Garter and
the securement of the artifact to the stabilizer, the width of the
garter appears to be extended more than at the time of original
construction and use. Given the inelasticity of the twined rows,
it may be that the thread arrangements measured close to the
twined section are a more accurate reflection of the garter as
it was produced than those in the center of the fabric. Table 3
provides measurements taken from the garter. As mentioned
previously, the 1L and ID through 5L and 5D labels indicated
where on the garter the measurement were taken (Figure 5).
The weft per cm and warp per cm record the number of two-
ply weft (the hidden yams) and the two-ply warp (the only
visible yars) and provide an estimation of the density of the
interlaced fabric. The mean weft yars per centimeter was 7.3
and the mean warp yars per centimeter was 10 weft yars per

Bead Structure and Usage

The white beads in the garter appeared to be opaque glass
beads (Figure 8) of the class II and type IIA1, a trade bead
common in the late eighteenth century and early nineteenth
century (Brain 1979:101). Although these beads are shiny,
there is no visibly thick clear coating. By definition, type IIAl
beads can range from 1 to 19 millimeters in length and 1 to 10
millimeters in diameter with holes from 0.5 to 3 millimeters
(Brain 1979:101). While this type of bead supports the dating
of this textile to a period prior to Osceola's death in 1838, it
is too ubiquitous over too broad a period to further refine the
production date of the garter.
Of the five beads included in this analysis, every
measurement was not available due to the position of the
remaining beads in the fabric. The mean bead diameter was
3.2 millimeters, the mean length was 2.1 millimeters, and the
mean diameter of the perforation was 1.1 millimeters (Table
4). The beads were threaded on to the yars prior to the garter's
construction and placed in their final locations as the braiding
progressed in order to outline the chevron pattern.


This garter is a valuable artifact of a culture at the point
when it was undergoing tremendous change. European
clothing was accepted by Native Americans in the early
nineteenth century but some traditional garments continued to
be produced and worn. The quality of the work in Osceola's
garter is evident even in its current state of deterioration, both
the design and the execution show evidence of technical skill
and knowledge.



Table 3. Fabric measurements collected from Osceola's Garter (1D through 5D indicate locations shown in Figure 5).

Mean Mean
ID 2L 3D 4L 5D Man a
Light Dark

Weft per cm 8 6 8 8 6 7 7.3

Warp per cm 12 12 8 10 10 11 10

Table 4. Bead measurements collected from Osceola's Garter.

Bead 1 Bead 2 Bead 3 Bead 4 Bead 5 Mean

Diameter (mm) 3.5 3.7 3.4 3.7 3.4 3.5

Length (mm) 2.5 3.0 2.5 n/a 2.6 2.6

Perforation (mm) n/a n/a n/a 1.0 1.1 1.1

As an artifact of a historic culture, the Osceola Garter is
considered rare today, but when it was made and worn it would
not have been unusual, as can be seen from the numerous Indian
portraits from the period. The use of the oblique interlacing
construction technique, twining, and the small white beads,
were all common in the 1830s.
The visible scale structure and mean diameter of the
garter's fibers are consistent with a sample of nineteenth
century buffalo fiber. While it would be ideal to have a readily
available definitive test for specific fiber type, beyond visual
identification there is not one currently available (Kathryn
A. Jakes, personal communication 2006). There are historic
accounts of buffalo as far south as Tampa Bay as late as 1772
(Rostlund 1960:397) and there are ethnographic accounts that
Southeast Indians produced buffalo fiber garters (Le Page
1972:308). While the physical evidence of the scale pattern
on the sample fibers cannot absolutely confirm that the fiber
is buffalo, the authors argue that the Osceola Garter is a
buffalo fiber garter. If the fiber is buffalo, then it would have
been hand-spun by Native Americans, using either a hand-
spindle or a spinning wheel. If the fiber is sheep's wool, then
determining whether the yar was spun by Indians or traded
from Europeans or Americans may not be possible.
Although some questions remain unanswered, this
research has explained much about the Osceola Garter. The
data provided by this oblique interlaced garment contribute to
the documentation regarding early nineteenth century Native
American textiles and provides researchers with information
on aspects of its finger-woven construction. Given the
perishable nature of textiles, and the fact that few specimens
from this period remain as intact and complete as this artifact,
the Osceola Garter provides a rare complete example that can
provide comparative data for future analysis of textiles from
the early nineteenth century.


1. Although "bison" is the more accurate term for the North
America species, historically writers of the "old west" have
and still use the term buffalo.
2. While interlacing involves one yar element weaving
over and under a perpendicular yar element, twining uses a
pair of yam elements that twist around each other, enclosing
the perpendicular yar element. In the Osceola Garter, the
twining yar elements are a bundle of three or four light and
dark yars. Pairs of these bundles were twisted around each
other in a half-turn to enclose a bundle of the yar from the
fabric, twisted again to enclose the next bundle of fabric yam,
and so on. When the edge of the fabric was reached, the two
twining bundles of yar were twisted again and a second row
of twining was created by working back across the fabric,
twining the same bundles of fabric yar, until the opposite
edge of the fabric was reached.


The authors would like to thank lan W. Brown for offering
helpful comments on the manuscript and curator Robert
Cason of the Alabama Department of Archives and History
for providing access to Osceola's garter and workspace for the


2007 VOL. 60(2-3)


References Cited


Annual Book ofASTMStandards 2003: Section Seven
Textiles, Vol. 07.01. American Society for Testing and
Materials, Conshohocken, Pennsylvania.

2006 Annual Book ofASTMStandards 2006: Section Seven
Textiles, Vol. 07.01. American Society for Testing and
Materials, Conshohocken, Pennsylvania.

Baclawski, Karen
1995 The Guide to Historic Costume. Drama Book
Publishers, New York.

Blackard, David M.
1990 Patchwork and Palmettos: Seminole and Miccosukee
Folk Art Since 1820. Ft. Lauderdale Historical
Society, Ft. Lauderdale, Florida.

Brain, Jeffrey P.
1979 Tunica Treasure. Papers of the Peabody Museum
of Archaeology and Ethnology, Vol. 71. Harvard
University, Cambridge.

Bushnell, David I., Jr.
1906 The Use of Buffalo Hair by the North American
Indians. Man 6:177-180.

Carson, James T.
1999 Searching for the Bright Path: The Mississippi
Choctawsfrom Prehistory to Removal. University of
Nebraska Press, Lincoln.

Cassman, Vicki
2000 Prehistoric Andean Ethnicity and Status: The
Textile Evidence. In Beyond Cloth and Cordage:
Archaeological Textile Research in the Americas,
edited by Penelope B. Drooker and Laurie D.
Webster, pp. 253-266. University of Utah Press, Salt
Lake City.

Church, Flora
1984 Textiles as Markers of Ohio Hopewell Social
Identities. Midcontinental Journal of Archaeology,

Densmore, F.
1974 [1928] How Indians Use Wild Plants for Food, Medicine
and Crafts. Dover, New York.

Downs, Dorothy
1995 Art of the Florida Seminole and Miccosukee Indians.
University Press of Florida, Gainesville, Florida.

Drooker, Penelope Ballard
1992 Mississippian Village Textiles at Wickliffe. University
of Alabama Press, Tuscaloosa.

Emery, Irene
1995 The Primary Structure of Fabrics. Watson-Guptil,
New York.

Erichsen-Brown, C.
1979 Medicinal and Other Uses of North American Plants.
Dover Publications, New York.
Foreman, Grant
1989 Indian Removal: The Emigration of the Five Civilized
Tribes of Indians. University of Oklahoma Press,

Gilmurray, Susan
1981 Weaving Tricks. Von Nostrand Reinhold Limited,
New York.

Goggin, John M.
1955 Osceola: Portraits, Features, and Dress. Florida
Historical Quarterly, 33:161-191.

Hatch, Kathryn
1993 Textile Science. West Publishing,

Humphries, Mary
2004 Fabric Glossary. Pearson Prentice
Saddle River, New Jersey.

Saint Paul,

Hall, Upper

Hulton, Paul
1984 America 1585: The Complete Drawings of John
White. University of North Carolina Press, Chapel

Hyland, D. C. and J. M. Adovasio
2000 The Mexican Connection: A Study of Sociotechnical
Change in Perishable Manufacture and Food
Production in Prehistoric New Mexico. In Beyond
Cloth and Cordage: Archaeological Textile Research
in the Americas, edited by Penelope B. Drooker and
Laurie D. Webster, pp. 141-160. University of Utah
Press, Salt Lake City.

Kalm, Peter
1937 [1770] The America of 1750: Peter Kalm's Travels in
North America, Vol. I & II. Translated by Adolph B.
Benson. Wilsen-Erikson, New York.

Kuttruff, Jenna T.
1990 Mississippian Textile Production Complexity as
Represented in Pottery Impressions. Paper Presented
at the 89th Annual Meeting of the American
Anthropological Association, New Orleans.

1993 Mississippian Period Status Differentiation through
Textile Analysis: A Caddoan Example. American
Antiquity 58:125-145.




Le Page du Pratz
1972 [1774] The History ofLouisiana, or of the Western Parts
of Virginia and Carolina. Claitor's, Baton Rouge,

Martin, Jack B., and Margaret McKane Mauldin
2000 A Dictionary of Creek/Muskogee: with Notes of the
Florida and Oklahoma Seminole dialects of Creek.
University of Nebraska Press, Lincoln.

Milanich, Jerald T.
2004 Osceola's Head: Close Encounters with a Famed
Seminole Chief. Archaeology 57(1):48-53.

Petersen, James B. and Jack A. Wolford
2000 Spin and twist as Cultural Markers: A New England
Perspective on Native Fiber Industries. In Beyond
Cloth and Cordage: Archaeological Textile Research
in the Americas, edited by Penelope B. Drooker and
Laurie D. Webster, pp. 101-117. University of Utah
Press, Salt Lake City.

Perdue, Theda
2004 Race and Culture: Writing the Ethnohistory of the
Early South. Ethnohistory 51:701-723.

Rostlund, Erhard
1960 The Geographic Range of the Historic Bison in the
Southeast. Annals of the Association of American
Geographers 50(4):395-407.

Schreffler, Virginia L.
1988 Burial Status Differentiation as Evidenced by Fabrics
from Etowah Mound C, Georgia. Ph.D. dissertation,
Ohio State University, Columbus. University
Microfilms, Ann Arbor.

Seiler-Baldinger, Annemarie
1994 Textiles: A classification of Techniques. Smithsonian
Institution Press, Washington D.C.

Sibley, L. R., M. E. Swinker, and K. A. Jakes
1991 The Use of Pattern Reproduction in Reconstructing
Etowah TextileRemains. ARS Textrina 15:179-202.

Sibley, Lucy R., Kathryn A. Jakes, and Lewis H. Larson
1996 Inferring Behavior and Function from an
Etowah Fabric Incorporating Feathers. In A Most
Indispensable Art, edited by J. B. Petersen, pp.73-87.
University Tennessee Press, Knoxville.

Song, C. A., K. A. Jakes, and R. W. Yerkes
1996 Seip Hopewell Textiles Analysis and Cultural
Implications. Midcontinental Journal ofArchaeology,

Spanos, Mary
2003 Spinning Soft Yarn with a Handspindle. Spin-Off

2006 Mississippian Textile at Beckum Village (ICk24),
Clarke County, Alabama. M.A. thesis, Department of
Anthropology, University of Alabama, Tuscaloosa.

Sturtevant, William C.
1955 Notes on Modern Seminole Traditions of Osceola.
Florida Historical Quarterly, 33:206-217.

1956 Osceola's Coats? Florida Historical Quarterly,

1987 Seminole Men's Clothing. In A Seminole Source
Book, edited by William C. Sturtevant, pp. 160-174.
Garland Publishing, New York.

Swanton, John .R.
1969[1946]The Indians of the Southeastern United States.
Smithsonian Institution Press, Washington, D.C.

Thompson, Amanda J., and Kathryn A. Jakes
2005 Textile Evidence for Ohio Hopewell Burial Practices.
Southeastern Archaeology, 24:137-141.

Tiedemann, Erica J.
2001 Characterization ofPrehistoric Spinning Technology:
Toward the Determination of Spinning Practices
Employedin Mississippian Textiles. Ph.D. dissertation,
Department of Textiles and Clothing, Ohio State
University, Columbus. University Microfilms, Ann

Ward, May M.
1955 The Disappearance of the Head of Osceola. The
Florida Historical Quarterly 33:193-201.

Wickman, Patricia
1991 Osceola's Legacy. University of Alabama Press,

Wilcox, Ruth Turner
1958 The Mode in Costume. Charles Scribner's Sons, New

Williams, Samuel C. (editor)
1930 Adair s History of the American Indians. Promontory
Press, New York.

2007 VOL. 60(2-3)



Choctaw Prophecy: A Legacy of the Future. Tom Mould.
2003, University of Alabama Press, Tuscaloosa. XVIII+263
pp., photos, references, notes, appendix, index. $34.95

8231 NW 45 Court, Lauderhill, FL 33351

I imagine that writing a book about oral histories and the
people who perpetuate them must have been a daunting task.
For instance, the time frame alone is one that covers three
simple time periods, the past, present, and future. The content
of these prophecies also focus around all the major topics that
any group of people face; birth, death, work, family, and war.
I must confess I am a bit biased in my review of this particular
work because of the fact that I interact daily with Native
Americans and their concept of prophecy and prediction. With
that in mind, I found this work to be very informative in regards
to how one can dissect all aspects of oral recitations, from the
narrator's physical mannerisms to the tone in their voice. In
fact, I found the passages pertaining to how Choctaws (and
more generally, Native Americans) interacted and recounted
their experiences with oral histories, to be quite accurate and
As I began to read this work, I noticed that both the
preface and the introduction were structured in such a way
as to succinctly explain to the reader why the author tackled
this subject and how his research was conducted. In giving the
appropriate amount of background information to the reader,
Mould allows one to understand what was lacking in previous
works (anthropological scholarship that focused only on the
major prophetic movements) and why he chose to study this
topic (he wanted to ensure a written legacy for the Choctaw of
I found this book to be well written, interesting, and
enlightening in its content. Mould organized his work into
five chapters focusing around Choctaw verbal communication
as well as the history of prophecy amongst Native Americans.
Mould (who obviously spent considerable time interviewing
Native American participants) accurately relates how these
interviews took place, from the physical surroundings to
the mannerisms of the interviewees. I felt that Mould did an
excellent job of providing the reader with a comprehensive
history of the early prophetic movements of the nineteenth
century. He also explains how many of the ethnocentric views
of early researchers caused them to look at these movements
more for their participants and the events that were held, than
at the actual content of the movement. In this work, Mould
attempts to provide a more enlightened twentieth century view

(albeit from an outsider) of why these prophecies took place,
what these predictions meant to those relaying them, and why
the Native Americans followed them. Mould utilizes a mix of
first person narratives and historical prophetic occurrences to
form a historical context that explains the reasoning behind
why Native Americans and specifically the Choctaw practice
this form of oral history.
As stated earlier, Mould provides the reader with very
a wonderful contrast of the ways in which these prophecies
are delivered to the masses. In chapter 2 "The Genre and
Performance of Prophecy," he delineates the difference
between when a speaker "reports" as to when they "perform"
for their constituents. He also explains (in chapter 3
"Interpreting Prophecy") that as each Choctaw speaker either
performed or reported their prophetic visions, they were at
times "interpreting" and at other times "recounting". As one
might suspect, the lines between when a speaker is reporting
and recounting become some blurred at times, which is why
a researcher such as Mould is able to discern the difference
between these skills. One of Mould's main points is that there
were different purposes or goals of these prophets. At times,
they were supposed to be telling stories to their people so that
the stories and events of the past would be kept alive. At other
times, these movements were used so that stories and events of
the past could be interpreted for the population. By interpreting
these stories for the entire populations, the prophets helped
build the basic social, religious, and spiritual framework for
the community. As with all historical recounting, we (the
listeners) hope that what we hear and what we are shown,
teaches us how to act and how to not repeat events that cause
such things as death, destruction, or loss of beliefs.
It is not until chapter 5 "The Future in Prophecy" that
Mould attempts to outline some of the other areas of research
that are being performed within Native American communities.
It is also within this chapter that the author provides the reader
with some insight into how the Choctaws live, what their home
life is like, and how they have acculturated the oral histories of
their ancestors with religion and modern technology. As with
many Native American groups, the Choctaws have a substantial
percent of their population that practice religion in a truly
acculturated manner. This mixture is generally a combination
of traditional ways and Christianity. Various denominations
come into play, but Baptist and Catholic are probably the
two most prominent. One of the main reasons these prophetic
movements incorporate Christianity into their stories and the
interpretations of their stories, is to help explain how and why
something has occurred. For example, the Choctaws believe
that the bible and their prophetic descriptions of total world
annihilation are not meant as doomsday predictions, but rather


VOL. 60(2-3)



they are written/spoken to give people hope. In this sense,
prophecy and biblical stories strike a balance between what
has happened, what will happen, and how we are to react to
these changes. As Mould points out, Choctaw men and women
quote the bible and prophecy in equal proportions.
Although Mould points out that many of the Choctaw
and other Native American prophecies superficially focus on
death, destruction, and other bleak happenings, they actually
serve a more positive and hopeful purpose. He makes it a
point to stress that although the prophecies seem pessimistic,
the people are not. They use these stories to balance their
lives and to restore their faith in each other. At the same time
though, there are some members of these communities who
turn to the bible because they feel that they have no other way
to learn about their past. This last point brings up what I think
is the most important point of this book, and that is cultural
Here again, I am showing a slight bias on my part. As
I observe some of the research being done to preserve the
languages and oral histories of Native Americans, I took from
this work that much more research needs to be done about these
storytellers and prophets. Although Mould does not overtly
stress that these languages and stories are dying out, he does
imply it when he includes passages from Choctaw regarding
future generations. As the elders pass away, subsequent
generations are taking up the job of telling these stories but the
mantra of "keep your language and you will be saved" is still
heard throughout these communities.
In writing this book, Tom Mould has provided some
insight and perspective into an area that few delve into. There
have been countless works on the Native Americans as entire
cultural groups, but very few works exist that cover the use of
language and prophecy. I thought this book was well written,
succinct, and thorough in its coverage of how these prophets
operate, what their purpose is, and where their future lies.
As someone who answers questions almost daily regarding
Native American languages and culture, I would recommend
this work as an example of how people conduct research
within a Native population and what sorts of results can some
from that research.

Jacksonville: The Consolidation Story, From Civil Rights to
the Jaguars. James B. Crooks. University Press of Florida,
Gainesville. 2004. XVIII+296 pp., photos, references, notes,
appendix, index. $27.95 (cloth).

7902 Sloane Gardens Court, University Park, Florida 34201

James B. Crooks considers the coming of the Jacksonville
Jaguars National Football League team to Jacksonville at the
end November 1993 as "the most exciting event of the decade."
Not even the most ardent booster could have imagined such an
event twenty-five years ago when the city faced spiraling crises
in educational and civic services. This is an interesting story
to read. It is not just a documentary of a decaying city that

like the frog that turns into a prince, has a sudden metamorphic
change into a vital metropolis. Crooks records the painstaking
process by tracing the evolution of the city from the beginning
of Hayden Bums' mayoral administration in 1949 through
the process of rejuvenation of its downtown, providing an
overview of the multiple problems and ramifications of the
threatened environment, the increasing suburban sprawl, and
catastrophic racial tensions.
Crooks was the historian-in-residence during Mayor
Tommy Hazouri's administration in Jacksonville from 1987 to
1991. He gained access to an enormous number of documents
packed into forty-five file cabinets discovered in the old city
hall parking garage, and with administrative permission was
able to read and make copies of the materials. Because of the
sheer quantity and scope of the documents, Crooks made the
decision to narrow the emphasis of the book to three salient
issues: race relations, environmental factors, and downtown
abandonment. He graphically documents the people and events
of the turbulent civil rights movement as it played out on the
Jacksonville stage with the consequences of the intransigent
southern white mind-set. Crooks moves the book forward
with an orderly and sequential presentation about the decision
to consolidate Jacksonville and Duval County government
agencies and discuss the ramifications of this decision to the
Beginning with a backward look at Jacksonville in the
years following World War II, Crooks describes the city center
with its contaminated and unsightly waterfront, its "rat-infested
dilapidated warehouses and wharves," and the subsequent
revitalization of the area during Burns' administration.
Burns saw himself as a city builder with a mission to renew
a decaying downtown. The story is one of determination and
shrewdness. Mayor Bums was able to raise funds to revamp
the waterfront, build a new courthouse, and begin construction
of a new sanitary sewage system. Most dramatically, he set
the wheels in motion for the development of the Jacksonville
Expressway System and the growth of the insurance industry.
The establishment of the Jacksonville Chamber of
Commerce was critical to the success of the venture. The
Chamber comprised more than five hundred business and
professional residents. In a prescient move in the early 1950s,
the Chamber created an effective planning board, which
crossed city-county lines, annexing some adjacent suburbs
to increase the tax base. By the early 1960s, however, when
the Chamber realized the importance of further annexations
in order to provide better schools, protection services, and
utilities, residents disagreed and voted down a referendum
for annexation. Those residents who had escaped to the
suburbs wanted no part of the complex problems of the urban
population and infrastructure. It would take more time.
By mid-decade, forty-two percent of Jacksonville's
population was African- American. Following consolidation,
the percentage of the African-American populace in the
greater county area diminished to twenty-three percent. The
quality of life for minorities continued to deteriorate: there
was substandard housing, slumlords, inadequate retail stores,
poor schools, scanty police and fire protection, and vast
unemployment. Unavoidably, civil unrest turned into mass


2007 VOL. 60(2-3)

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs