Front Cover
 Membership Information
 Table of Contents
 Editor's Page
 The Stratigraphic Sequence at Rollins...
 Florida Anthropological Society...
 Perpetuating Tradition on the Lower...
 The Honey Dripper Site (8NA910):...
 Book Reviews
 About the Authors

Group Title: Florida anthropologist
Title: The Florida anthropologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00027829/00184
 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.
Place of Publication: Gainesville
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: VID00184
Source Institution: University of Florida
Holding Location: Department of Special Collections and Area Studies, George A. Smathers Libraries, University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 01569447
lccn - 56028409
issn - 0015-3893

Table of Contents
    Front Cover
        Front Cover
    Membership Information
        Unnumbered ( 3 )
    Table of Contents
        Table of Contents 1
        Table of Contents 2
    Editor's Page
        Page 247
        Page 248
    The Stratigraphic Sequence at Rollins Shell Ring: Implications for Ring Function
        Page 249
        Page 250
        Page 251
        Page 252
        Page 253
        Page 254
        Page 255
        Page 256
        Page 257
        Page 258
        Page 259
        Page 260
        Page 261
        Page 262
        Page 263
        Page 264
        Page 265
        Page 266
        Page 267
        Page 268
    Florida Anthropological Society Chapters
        Page 269
        Page 270
    Perpetuating Tradition on the Lower St. Johns: Pottery Technology and Function at the Mayport Mound (8DU96)
        Page 271
        Page 272
        Page 273
        Page 274
        Page 275
        Page 276
        Page 277
        Page 278
        Page 279
        Page 280
        Page 281
        Page 282
        Page 283
        Page 284
        Page 285
        Page 286
        Page 287
        Page 288
        Page 289
        Page 290
        Page 291
        Page 292
        Page 293
        Page 294
        Page 295
        Page 296
        Page 297
        Page 298
    The Honey Dripper Site (8NA910): A Late Swift Creek Encampment in Northeastern Florida
        Page 299
        Page 300
        Page 301
        Page 302
        Page 303
        Page 304
        Page 305
        Page 306
        Page 307
        Page 308
        Page 309
        Page 310
    Book Reviews
        Page 311
        Page 312
        Page 313
    About the Authors
        Page 314
        Page 315
Full Text


2000 Florida Anthropological Society Inc.

The Florida Anthropological Society Inc. holds
source text of the Florida Anthropologist
considered the copyright holder for the text
these publications.

all rights to the
and shall be
and images of

The Florida Anthropological Society has made this publication
available to the University of Florida, for purposes of
digitization and Internet distribution.

The Florida Anthropological Society reserves all rights to this
publication. All uses, excluding those made under "fair use"
provisions of U.S. Code, Title 17, Section 107 are restricted.

Contact the Florida Anthropological Society for additional
information and permissions.







8 -1




- 4


Episode 8

Episode 7

Episode 6

Episode 5

Episode 4

Episode 3

Episode 2

^ -"^^"^F^^Sa^-^

Episode 1



F 63-4

THE FLORIDA ANTHROPOLOGIST is published by the Florida Anthropological Society, Inc., P.O. Box 6356, Tallahassee, Florida 32314.
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: individual $25, family $35, institutional $25, sustaining $35 or
more, patron $100 or more, and life $500. 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 Editor 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 Editor to facilitate acknowledgment in
subsequent issues of the journal (unless anonymity is requested). Submissions of manuscripts should be sent to the Editor. Publications
for review should be submitted to the Book Review Editor. Authors please follow The Florida Anthropologist style guide (49[1], 1996,
pp. 37-43) in preparing manuscripts for submission to the journal and contact the Editor with specific questions. Submit four (4) copies
for use in peer review. Only one set of original graphics need be submitted. The journal is formatted using Word Perfect 8.0. All
manuscripts must be submitted in final form on DOS-formatted 3.5 inch diskettes or CDs in Word Perfect, Word, or ASCII 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: David Burns, 15128 Springview St., Tampa, FL 33624 (daveburns@prodigy.net)
First Vice President: Scott Mitchell, Florida Museum of Natural History, University of Florida, Gainesville, FL 32611-7800
Second Vice President: Phyllis Kolianos, 880 Seminole Blvd., Tarpon Springs, FL 34689 (koliano@attglobal.net)
Corresponding Secretary: Gordon Davis, 2200 N. Oleander Dr., Avon Park, FL 33825 (gordondavis@hotmail.com)
Membership Secretary: Terry Simpson, 9907 High Meadow Ave., Thonotosassa, FL 33592 (tsimpson@luna.cas.usf.edu)
Treasurer and Registered Agent: Kelly Driscoll, 5313 Johns Road, Suite 205, Tampa, FL 33634 (k.driscoll@mindspring.com)
Directors at Large: Gloria Fike, 2815 SE St. Lucie Blvd., Stuart, FL 34997 (GloriaFike@prodigy.net)(One Year): Linda Allred.
5460 43rd Ave. N., St. Petersburg, FL 33709 (llrred@juno.com)(Two Years): Chuck Wilde, 195 Huntley Oaks Blvd.. Lake Placid.
FL 33852 (pillpot@strato.net)(Three Years)
Immediate Past President: Sheila Stewart, 2130 Burlington Avenue North, St. Petersburg, FL 33713 (sjstew(a@gte.nct)
Newsletter Editor: Sheila Stewart, 2130 Burlington Avenue North, St. Petersburg, FL 33713 (sjstew@gtc.net)


Editor: Ryan J. Wheeler, P.O. Box 6356, Tallahassee, FL 32314-6356 (rwheeler@nettally.com)
Book Review Editor: Keith Ashley, P.O. Box 350225, Jacksonville, FL 32225-0225 (akashlcy(@fdn.com)
Technical Editor: vacant
Editorial Assistant: George M. Luer, 3222 Old Oak Drive, Sarasota, FL 34239-5019 (gluer(rgrove.ufl.edu)
Printer: Durra-Print, 717 South Woodward Ave., Tallahassee, FL 32304
Bulk Mail: Capital City Mailing, 4013 Woodville Hwy, Tallahassee, FL 32311
BackIssue Sales: Graves Museum of Archaeology and Natural History. 481 S. Federal Highway. Dania, FL 33004. (954) 925-7770.
FAX (954) 925-7064


Albert C. Goodyear, Institute of Archaeology and Anthropology, University of South Carolina. Columbia, SC 29208 (goodyear a sc.edu)
Jerald T. Milanich, Florida Museum of Natural History, University of Florida, Gainesville, FL 32611 (jtmi(flmnh.ufl.edu)
Jeffrey M. Mitchem, Arkansas Archeological Survey, P.O. Box 241, Parkin, AR 72373 (jeffmitchem@(juno.con)
Nancy Marie White, Department of Anthropology, University of South Florida, Tampa, FL 33620-8100 (nwhite@(achuma 1.cas.usf.edu)
Robert J. Austin, P.O. Box 2818, Riverview, FL 33568-2818 (bob@asearchinc.com)

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

VISIT FAS ON THE WEB: www.fasweb.org



Volume 57 Number 4
December 2004

'NCE 19A1


Editor's Page

The Stratigraphic Sequence at Rollins Shell Ring: Implications for Ring Function.
Rebecca Saunders

Perpetuating Tradition on the Lower St Johns: Pottery Technology and Function
at the Mayport Mound (8DU96). Neill Wallis

The Honey Dripper Site (8NA910): A Late Swift Creek Encampment
in Northeastern Florida. Greg S. Hendryx


Little: Public Benefits of Archaeology. Lee Hutchinson

Phillips, Ford and Griffin: Archaeological Survey in the Lower Mississippi Alluvial Valley,
1940-1947. James C. Waggoner, Jr.

About the Authors U OF F BRAR


Cover: Reconstruction of depositional episodes at Rollins Shell Ring, see Rebecca Saunder's article on page 249.

Published by the
ISSN 0015-3893


This issue includes three articles that deal with the archae-
ology of northeastern Florida. The first paper, by Rebecca
Saunders, reports on her work at the Rollins Shell Ring-one
of several publically-owned shell ring sites in Florida (other
examples include the Guana Shell Ring and the Joseph Reed
ShellRing). Saunders' research supplements existing work on
shell ring sites, providing further radiocarbon evidence for the
great antiquity of these sites, as well as evidence that shell
rings represent an early form of monumental architecture and
a possible use as feasting sites by people of the late Archaic.
The second article, by University of Florida graduate
student Neill Wallis, presents an important analysis of
ceramics from the Mayport Mound. Many Florida
archaeologists, especially those who have worked in the St.
Johns River basin, are familiar with Rex Wilson's 1965
publication on his work at the site. Wallis' article provides a
fascinating look at the Swift Creek ceramic assemblage from
Mayport, making clear the value of re-evaluating existing
museum collections. Wallis was even able to document a
Swift Creek paddle stamp design that was used to decorate one
of the Mayport vessels as well as a vessel from the Altamaha
River area, over 100 km from Jacksonville.

The third and final article, by Greg Hendryx, is a site
report on another Swift Creek site in northeastern Florida-the
Honey Dripper site. I can not imagine why archaeologists
would name a site after a liquor store, but I am sure everyone
can ask Greg at the next FAS annual meeting. Hendryx's
article is a good complement to the proceeding paper and adds
to the increasing body of literature on Swift Creek in
northeastern Florida.
One nice thing about this issue is the diversity of the
contributors, including a professional archaeologist working
at a museum, a graduate student, and an archaeological
consultant. I am particularly happy to have the recent partici-
pation of so many talented student authors, and hope this trend
continues. This issue also contains two book reviews by Lee
Hutchinson and James C. Waggoner, Jr. I hope that every
reader finds something they enjoy.

December, 2004


VOL. 57(4)



A new video on florida's native peoples

Flonrdas Lost People"
Produced by the Florida
Funded by the
Florida Department
of State

Produced and Directed by Chaos Productions
Executive Producer: Brent Weisman
Written by Marshall Riggan
Artwork by Theodore Morris

1998 Florida Anthropological Society and the Florida Department of State

To obtain copies send $23.62 [$18.81 plus $1.31 (sales tax) and $3.50 (S&H)] to:
Terry Simpson, 9907 High Meadow Ave., Thonotosassa, FL 33592



Museum of Natural Science, Louisiana State University, Baton Rouge, LA 70803
E-mail: rsaunde@lsu.edu

Shell rings are poorly understood features of the Middle
and Late Archaic landscape in the lower Southeast. Con-
structed between ca. 4500 and 3000 years ago, these circular-
to arc-shaped deposits of shell range in diameter from 30 to
250 m and can reach over 6 m in height. Shell rings are
distributed along the lower Atlantic coast from South Carolina
to southern Florida; somewhat similar structures are found
along the northern Gulf coast as far west as the Pearl River at
the Mississippi/Louisiana border as well as along the south-
western Gulf coast (Figure 1). Thus, rings are known from all
of the coastal manifestations of the various fiber-tempered
pottery-producing cultures. However, rings are not restricted
to those cultures. Several preceramic-age rings are known and
a ring containing a spiculate ware (St. Johns?) and sand
tempered sherds, dated to ca. 3300 B.P., is present on Jupiter
Island, Florida (Russo and Heide 2002). As noted above, rings
also extend to central South Carolina, where they are associ-
ated with Thoms Creek pottery, a ware with a sandy paste and
few or no fiber inclusions.

Previous Considerations of Shell Ring Function

Since they were first described, by William McKinley in
1873, shell rings have captured the imagination of avocational
and professional archaeologists alike. Against a backdrop of
contemporaneous amorphous shell middens and artifact
scatters (DePratter 1979; Waring 1968a), rings are impressive
structures. Writing of the three rings on Sapelo Island,
McKinley (1873) was taken by the height of the largest of the
three, over six meters of near vertical wall (three meters of
shell on three meters of sandy bluff), the apparent sterile
interior of each ring, and the relative symmetry of the shell
structures. Though overstated-modern maps (e.g., Russo and
Heide 2001) have shown that many rings have irregularities in
ring height and shape and interiors sometimes do have
features (the Ford Shell Rings, for example; Calmes
1967)-these characteristics are often cited by researchers in
consideration of function (e.g., Waring and Larson 1968:273).
McKinley (1873:422-428) proposed that the rings were
"doubtless for council and games" with the largest operating
as a "house of state" and "torture chamber" and the lesser
circles as places for dance, sports, and games. Others have
speculated that rings functioned as fish weirs (Edwards 1965;
Waring 1968b: 182), or were midden accumulations around the
edges of raised structures that served as primary fishing
stations (Waring 1968b: 183). Ultimately, Waring (1968a) and

Waring and Larson (1968) concluded that rings were the
earliest monumental architecture in the Americas; Michie
(1979) and Sassaman (1993) also attributed a ceremonial
rather than a simple settlement function to shell rings.
Marrinan (1975:117) was ultimately noncommittal as to ring
function, but she did note (1975:95) that "the attention to
symmetry and the recognition of a collective desire for this sort
of edifice is as real as any Midwestern earthen effigy mound or
British long barrow." DePratter and Howard (1980:8) also
leave the question of ring function open, though DePratter
(1979:50) identified a possible domestic structure (based on a
circular area of crushed shell and a possible hearth located in
the ring) at the A. Busch Krick ring site.
During the heyday ofdiffusionist explanations, Ford (1966,
1969) noted the similarity of southeastern rings to shell rings
in coastal Columbia and proposed (Ford 1969:185) that the
appearance of rings and fiber-tempered pottery in the South-
east could be attributed to direct colonization of the Savannah
River valley and adjacent coast by peoples from the area of the
Isthmus of Panama. Thus, to Ford, rings on the Georgia coast
were no more than the transplantation of a South American
domestic settlement pattern to the Southeast. More recently,
Trinkley's (1985; see also Trinkley 1997) influential paper on
shell ring function echoed Ford's assessment. Based on a
review of past investigations, in particular the (unpublished)
work at Chester Field undertaken by Ritter and Moorehead in
1932 and 1933, respectively, and his fieldwork at Lighthouse
Point and Stratton Place in Charleston County, South
Carolina, Trinkley proposed that rings were secular habitation
sites. He (Trinkley 1985:118) argued "that the circular shape
was related to the egalitarian nature of Early Woodland
societies, where a circular clustering of habitations would
promote communication and social interaction." Trinkley
(1985:117; 1997) used the "kitchen refuse" nature of the
midden remains, the pronounced banding and crushing of
midden and artifacts reported for Chester Field and observed
at the base of the Charleston County rings, and the presence of
numerous postholes and steaming and roasting pits at the base
of many shell rings, as support for a domestic as opposed to a
ceremonial function for the sites. Cable (1997), however,
found Trinkley's "gradual accumulation theory" of ring
formation untenable'. Reviewing the stratigraphic sequence of
the Sea Pines and Skull Creek shell rings, Cable argued that
the defining feature of these deposits was the presence of large
piles of unconsolidated shell capped by thin lenses of crushed
shell. Cable hypothesized that these deposits were intention-


VOL. 57(4)




Figure 1. Location of Rollins Shell Ring and other southeastern shell rings.

ally mounded remains from feasts occurring every 10 to 20
years. Numerous contemporaneous habitation sites around the
Sewee shell ring indicated to Cable that the Sewee ring also
had a special function2. On the basis of the nature of the
deposits and zooarchaeological remains, Russo and Heide
(2002) argued for feasting at the Joseph Reed shell ring
(8MT13) near Cape Canaveral; Russo et al. (2002) took the
ceremonial function of rings for granted in their discussion of
the Guana Shell Ring.
In 1998, work was undertaken at the Rollins Shell Ring in
northeastern Florida as part of a larger project (Russo and
Saunders 1999) to explore the evolution and function-secular
or ceremonial--of several shell rings on the northeastern coast
of Florida. This initial effort at Rollins was limited in scope.
The research was designed to: 1) produce a detailed topo-
graphic map of the site; 2) use soil chemistry to determine
whether some portions of the Rollins ring may have been
borrowed for shell; 3) secure good radiocarbon dates for the
initial and final construction stages of the ring; 4) record
microstrata to understand how the Rollins ring was con-

structed; 5) develop seasonal data to determine seasons of site
deposition; and 6) to recover a good sample of artifacts that
might indicate activities at the site (Russo and Saunders 1999).
Results from this research provide additional hard data on the
characteristics of shell rings: data upon which testable hypoth-
eses about shell ring function can be based. Here the empha-
sis is on the implications of the stratigraphic sequence and the
composition of the strata at the ring. A more complete account
of results of these excavations can be found in Saunders

Rollins Shell Ring

Rollins Shell Ring is a topographically complex site
(Figures 2 and 3). It is one of the larger of the known ring
sites at 250 m in diameter (150 m across the interior) and over
4 m in height above the surrounding terrain3. The main ring
is generally horseshoe-shaped. Conventional radiocarbon
dates from basal midden deposits from the trench excavated on
the west side of the ring and basal midden deposits from 1 x 2

2004 VOL. 57(4)



F = ringlets

4900- = excavation

= excavation

4850- Note: units and trench
not to scale

4800 6.6

200 250 300 350 400 450 500 2.6

Figure 2. Rollins topographic map with 1998 excavation units (units not to scale).


Figure 3. Rollins topographic map, wire contour.

m Unit 3197 on the east side of the ring are, for all practical
purposes, contemporaneous at about 3600 cal B.P. (Figure 3;
Table 1), or in the Orange III period. The contemporaneity
between the arms of the ring indicates that the horseshoe shape
and overall site size was inherent in the original plan of the
site and that it was maintained throughout the occupation of
the site-probably some two to three hundred years (further
dating of the site is discussed below).
The Rollins ring is decidedly asymmetrical (Figures 2 and
3). The western portion of the ring is appreciably broader and
higher than the eastern arm. In addition, at least nine smaller
enclosures or "ringlets" are arranged around the western and
northern sides but are absent from the eastern arm4. Because
the highest and broadest portions of the ring are associated
with the presence of the well-defined smaller enclosures, the
ringlets may be at least partially responsible for the difference
in ring height and width from west to east. There has been
insufficient excavation in these areas to determine how the
enclosures articulated with the main ring; that is, whether they
were contemporaneous with, preceded, or postdated initial ring
construction. However, complementary excavations at the Fig
Island shell ring site (38CH42) in South Carolina indicate they
may be younger. At Fig Island, one smaller enclosure attached
to the enormous Fig Island 1 ring was radiocarbon dated ca.
200 years younger than the main ring deposits (Saunders

Trench Excavations, 1998

Ten 1 x 2 m units were excavated throughout the site, one
on the east arm of the ring (Unit 3197) and the nine others in
non-shell main ring or ringlet centers. Results of these
excavations are reported in Saunders (2003). Within the ring
itself, a 1 x 16 m trench was excavated on the western side of
the ring (Figure 3). The trench was laid out northeast to
southwest between two of the highest peaks of shell on the site.
In addition to providing material cultural and other evidence
of site activities, the trench was intended to bisect the ring
feature as a whole and to provide stratigraphic evidence of the
depositional events that made up the ring. Most important was
to determine whether the ring was composed of lenses of
compacted, accretional midden resulting from daily refuse
discard or of large deposits with little postdepositional distur-
bance indicative of feasting (or both).
Trench profiles (Figure 4) revealed that the bulk of the ring
was composed of a massive shell pile, designated Feature 1
during excavation (all features are listed in Table 2)5. Though
this was disturbed by intrusions (in Figure 4, these intrusions
are the vertical lenses of what is keyed as "Al" in Excavation
Units 1 and 2, north and south profiles; and "C4" and "E3" in
the north wall profile at the boundary of Units 2 and 4), and
stratigraphy is not as straightforward as one would like6, both
the north and south wall profiles indicated three episodes of
rapid shell deposition creating the core of the ring (Figure 4,

2004 VOL. 57(4)


Table 1. Radiocarbon dates from Rollins Shell Ring.'

FS # Description Sample Measured Conven- 133CPDB 2/1 cal 1/2 cal. 2 /1 cal (intercept) 1/2cal Lab #
B.P. tional B.P. delta R -5 20 delta R 36 14

85 Trench 1, Unit 2, Oyster 3300 70 3670 -2.5%o 3767/3658 3477/3398 3692/3617 (3537) 3449/3363 Beta-
Feature 1, bottom 119816
deposit, 90-100 cm
Unit 1097, Ringlet J, Oyster 2100 70 2460 -3.0%o 2297/2200 -1999/1941 2266/2142 (2054) 1961/1892 GX-
pit feature. 29516

459 Trench 1, Feature 11, Bulk Carbon 3741 80 3730 -25.6%o 4300/4157 (4088) 3975/38412 GX-
base (initial 25750
occupation); 200 cm

467 3197, 10-20 cm bs, Oyster 2340 60 2690 -3.7%o 2655/2470 2309/2259 2518/2437 (2329) 2275/2149 WK-
midden. 7433

488 Trench 1, Unit 1, 3229 60 3600 -2.4%o 3638/3570 3419/3353 3601/3518 (3445) 3375/3324 WK-
Feature 1, top Oyster 7438
deposit, 33 cm bs.

508 3197, 80-90 cm bs, Oyster 3300 70 3710 -0.3%o 3818/3720 3528/3455 3761/3653 (3576) 3476/3399 Beta-
midden. 119817

Timu 4850N/250E, 60-65 Oyster 3360 60 3760 %0 3860/3776 3620/3533 3805/3715 (3634) 3554/3479 Beta-
cmbs (Russo 1993).____________ 50155
1. I have included two different reservoir correction values, -5 20, which has been the most commonly used value, and is the value still used by Beta Analytic (Darden Hood, personal
communication, 2004), and 36 14, which is the value currently posted for southern Florida and the Bahamas at http://radiocarbon.pa.qub.ac.uk/marine/. As can be seen, the newer value
does narrow age ranges significantly. I have also included calibrated intercepts, calibrated with Calib 4.3. While these are just a midpoint within the range and do not represent the most
likely date, they are a useful shorthand for comparison. All calibrated ranges were calibrated with Calib 4.4.
2. Reservoir corrections do not apply to this bulk carbon date.


South Wall

\ ~ E5 EjIS F4G7 E2
East F3
wall of North Wall Reverse View 3 West
Stretch F7 wall of
S t04 trench

5F1 E

FS J 22HI H3 B2 H3
ii G9H2

0 2 4

A. Uundarnt oyster with iOYR 212 to 4/2
MM 1.(Zmw 1) sMhishr 1NOYR3/I1 W4 4(1uriaid
j=:= 2 a"o I "g Womw)a Shdsrt AO Wr3i IOYR W NuM c
& (low 1) shr3110ys 3141YR l2wDh nusofldA1
Ld 4 OUIM on""trh
B. Abundant Oyster with little to no soll
I. rue-I ) AwdotahMoleGm o id W i =W W bwc mw s hid Wue Wb w1 W
2.)Ueas "ag)"tm4"sat Oxio &Wim Moko W wki waN uiauslof IOYR 4.2 Ismdc
3. )Fsamk Is) 10Al 3/1 sai wii oxsisa stin oyisr aid WWi 1sid iabudwA bam
4. T*An 1) &Wadoi oftb oyem di us" I YR 4Q sai
5 Mokjm 3. esW & w4m kdi) Vuida *07 shd hw idst-M We13 b c no A n
6' (F#s 3 aii ad wmind IOYR V1I osass sad)

C. common to scattered oyster In 1lOYR 211 to 2W sand
1, OAm 4s) 10YR 2/1 b"s amid dii ansirai %h*t oqst
Z (Fuan 4) lOYR W1 b"l said miii ausu Wid sh*t vysid
I l3 IOYR 2M b"si sWi miA onsas Wokn~a sWi
4. IGYR 2/1 loam mid dii saisis Mde oysW mid OW hashi
5. (Ams 15) lOYR 21 W"is said %0 usia misam *mmd badin oysl
6. (Aims 4b) IOYR W/1 dii oxisi oys6/, ilidwida4d miihi b"s aid si hash

D. common oyster In ICYR 3/1 to 64 sand
I. (Zon M42 =w=su b1 scorsidmyoingk IOYR W/ nOWce wlliliYR 411 usid
S2. )Fasiam U) lO YR / saidv silmislo aps mid slid hulh
3 (Fl'siitsi) GR (4mid64 said mmd onsiorI Id/siim pm
4, coRYon to rumnd optmo In 1a4421d to 2n send

4.10YROI WO 4/I id dii de hint Ih
O 10YR4Imiy~dmgi~asnoyw hblrdH

E scattered tonooyster llYR31Ztod2 umnd
I. (Fues 9) 1M 0/1 lust mid, ss OYM *Ndsa
xmimihbml Ns aid anPwismsbdwmd
2 (JFus 13) 103R l2 t sad V& oWauBMWiYO
&. 0IYR 412 MW dii asahd Sto ysts WW *W hah
4 107 4M sWa WO nSMOdM sldeV No m hasi
5.10711312 s tWm WW

F. mottldor mixed IOYR411 to 714 sand, scattered tono
1. (Zone 2M3) 11-d IR W1b514 said dmosI Csler
Z( 'Fme 7) 101 503 mid 714 sftad made bidh a1anla
I )Psijm 8) IOYR Y1I & 514 uvid. badlasi od as WSAkbjd
4. (Feshi U) 1IM1 6/4 & I/lla sfaid urftwi 7135 ii.: ban
cIA as poW* idimOO
5 10711411,412, sad /IN niidsad
s. 1014M 4/3 aid5 513 moWed anw ixeid $WWdI
WMi dIWdi PK5C & %&*V/i
7.10711413 aid WrO nwi 6/ Wii Aid
8. 1071YR6/33dW 4/3sdasdsaid

G. 10yr 2/1 to 5/5 sand with scattered to no shell
I. (Zm 1 ns.) 10YR 311 to hunmc muls
2. (Zor 11.) 10IOR 141 4 hic mnd wi mwy scttMd
3. (Zon 1 a.i darkr) 10YR 21 humis nd
4. (Frnm 10) 10YR 412 wid dih ducol hidci
5. 1(0Y1/2ciw nd
7. IOYR 412 sWad
S10YR 13 and
7.n10YRl 2 id
9. 10YR 3,45snud

H. sterile snd 10yr 6/4 to Oi1 with no shell
1.(. 3)O10YR6/4 to 711 snd
2 (Zoe 4)1WR711 sand
3.10YR /1 sad

I. Roots and charcoal
2. mol mmodndl diu bne
3L. Dwe rid i dw

J. cadcmalinarms

n.s. a shui
Is. WeI sho

Figure 4. Rollins trench profiles, south and north walls.











Table 2. Trench Features at Rollins Shell Ring.'

Feature# Location Top Base (mbs) Description
1 Units 1, 2; expands to .20 1.35 Includes three distinct deposits of abundant large, whole, clean oyster, abundant, fine bone, almost no soil. Relative to Zone
Unit 4 in Level 7, Unit 3 A, more crab claws, burned and unburned, very little periwinkle.
in Level 8
2 Unit 2 .5 n/a Abundant oyster with small amount of sandy, brown (10YR 5/3) matrix at top-which prompted segregation from Feature
1. This sand disappeared by base of Level 6, and the Feature could not be distinguished from Feature 1. Combined with
Feature 1 after segregating for one level
3 East Unit 4 1.10 1.30 Deposit of abundant large, whole, clean oyster similar to Feature 1, but some dark gray (10YR4/1) humic sand matrix, fine
3 West Units 3, 5 shell hash. Present on east and west sides of Feature 1.
4 Units 3, 5 1.60 1.70 Semi-oval area of abundant coquina shell in dark gray to black humic sand; admixed with Zone 1 (10YR 3/1-4/1) matrix.
5 Units 4, 6 1.60 1.92 Amorphous area heavily mottled very dark grayish brown, black (10YR 3/2, 10YR 2/1) humic sand, scattered shell,
charcoal, calcium concretions. May be combination of low spot in earth midden (Zone 1 no shell) underlying shell and root
6 Unit 4 1.60 1.80 Subrectangular area of very dark brown brown (10YR 2/2) humic sand, scattered shell, charcoal, calcium concretions
6a 1.80 adjacent to Feature 5. 6a had common shell in similar matrix; north wall profile indicates a separate feature from 6.
7 Unit 4 1.60 1.80 Subrectangular area of brown and very pale brown (10YR 4/3 -5/3, 10YR 7/3 7/4) mixed sands; bottoms out onto
circular area ofparticulate charcoal mixed with sand.
8 Units 1, 3 1.60 1.88 Semi-oval area of brown and very pale brown mixed sands. Profiled in south wall. Bottoms out onto particulate charcoal.
9 Unit 2 1.60 2.20 Large, semi-oval area of brown and very pale brown mixed sands, some mottling with more humic sands, pockets of oyster.
9a Unit 2 1.85 9a has more humic sand and shell; also contained more bone that 9.
10 Unit 2 1.60 1.88 Semi-subrectangular area of mixed brown and very pale brown sands, profiled in south wall. Bottoms out onto lens of
particulate charcoal mixed with sand.
11 Unit 1 1.60 1.80 Amorphous area of mixed brown and very pale brown sands; bottoms out onto particulate charcoal mixed with sand.
12 Unit 3 1.60 1.88 Amorphous area of mixed brown and and very pale brown sands, pockets of oyster shell adjacent to Feature 8. Profiled
in south wall. Bottoms out onto particulate charcoal mixed with sand. Though consistently lighter than Feature 8, profile
suggests same episode.
13 Units 3, 5, 7 1.80 1.90 Trench bisected possible circular area of very dark grayish-brown (10YR 3/2) humic sand with very occasional shell,
scattered crushed shell along western margin. Apparently a low area of the earth midden underlying shell.
14 Unit 7 1.90 2.04 Subcircular area of darker, humic sand (10YR 2/2, 10YR 3/2) below Feature 13.
'Areas' of probable cultural origin:
Area 4, 4a Units 1, 3 .50 1.15 Linear area of very dark gray to black (10YR3/1, 10YR 2/1) humic sand extending between Zone 1 and Feature 1 on west
Area 4b 1.10 side oftrench. 4b contained moderate broken and whole shell.
Area 15 Unit 4 1.60 1.80 Semi-circular area of mixed brown and very pale brown sands, profiled in south wall. Bottoms out onto lens of particulate
charcoal mixed with sand.
1. Elevations are with reference to a datum at ground surface at the highest point along the trench, the southeast comer of Unit 1.


Bl-B4). Each of the discrete episodes was separated by thin
lenses of sand or clayey sand. According to our soil scientist,
Sylvia Scudder (personal communication, 1998), these were
not aeolian deposits and so must have been anthropogenically
introduced. Because these sand lenses were not visible during
excavation, all three episodes were excavated as Feature 1.
Each of the three deposits of Feature 1 was composed
primarily of large, whole, clean oyster shells with virtually no
soil incorporated into the deposit. There was little breakage or
other compaction of the shell that might result from living on,
or any other intense use of, the surface of each of these
deposits. Lacking any soil or sediment matrix, and lacking
compaction, the shell was extremely loose. Shell orientation
was variable, ranging from horizontal to vertical. Small fish
bone was abundant in these deposits; most upturned oyster
shells cupped scores of tiny bones. The large, clean oyster, the
jumbled appearance of the shell, and the lack of soil might
suggest that Feature 1 was a secondary deposit transported
from a nearby area to the ridge by means of loosely woven
baskets that permitted smaller shells and soil to filter out. The
presence of the innumerable small fish bones, however, argued
against this conclusion.
Deposits similar in composition to Feature 1 appeared on
either side of the feature. These deposits (Feature 3 East and
Feature 3 West, keyed as B5 in Figure 4) had the same loose,
large, clean oyster shell but contained a small amount of sand
and abundant shell hash. These physically separate deposits
were so similar that, initially at least. Feature 1 could have
been considered an intrusion into this already-extent deposit;
using the same designation for both was intended to highlight
this fact. In profile, it is clear that these were deposited after
Feature 1. On the western half of the south wall, Feature 3
West was separated from Feature 1 by two lenses of very dark,
organically enriched sand with only amounts of whole oyster
(Figure 4, South Wall, Cl and C6). These were excavated as
Area 4a and Area 4b. This is lacking in the south wall on the
east side, but may be visible on the east side of the north wall
where disturbances C4 and E3 are present. On the western
portion of the north wall, Al between B 1 and B5 appears to be
the equivalent. Feature 1, Feature 3, and these lenses comprise
the principal deposits of the core of the ring. There were no
other deposits in the core of the ring; nothing-no postholes,
pits, hearths, etc.-indicative of habitation on the ring during
or immediately after the deposition any of these proveniences.
All of these deposits were overlain with a dense shell
midden, referred to as Zone 1. Zone 1 was composed of
abundant whole and crushed oyster in a dark grayish-brown,
organically enriched sand: very similar to midden descriptions
from sites of many time periods along the coast. In the south
wall profile, the contrast between this stratum and Features 1
and 3 and Area 4 is striking (Figure 5).
Both the north and south wall evidence a large intrusion(s)
of the overlying Zone 1 (Al) deposits into Feature 1, in Unit
1 on the south wall and in Unit 2 on the north wall. These are
considered intrusions because, unlike the consistent appear-
ance of Area 4 and other slope deposits, which receded
regularly to the interior or exterior of the ring with depth, plan

maps and photographs indicate irregular, amorphous areas of
shelly midden that appeared and disappeared throughout the
excavation of the relevant levels. Patches of crushed burned
shell, some of which had a "weird smell" (field notes, on file.
LSUMNS) were also associated with this intrusion. All
aspects of this intrusion were segregated during excavation as
Zone 1 and a series of Areas.
Two coquina "pot dumps"' were found within Zone 1
One, designated Feature 4. w as on the x western side of the ring.
and was initially defined spreading into both Units 3 and 5. A
small portion of this feature can be seen in the South Wall
profile in Unit 3 (A4). A second such deposit, on the opposite
side of the ring, can also be seen in the profile of Unit 6 (also
One more lens is worthy of note. Another shell midden
lens, Zone 1 Strong Brown (Figure 4, A2). appeared on either
side of the ring, at or near the base of Zone 1. The shell and
other inclusions in this deposit were the same as in Zone 1. but
the soil matrix was distinctly browner. The presence of this
lens on either side of the ring. as well as that of Feature 3.
gives an impression of s-ymmetry to the nng deposits.
The striking difference in the characteristics of Zone 1 and
the features in the core of the nng suggests different
depositional processes. Feature 1. and the later Feature 3.
suggest intentional mounding of shell with little subsequent
disturbance. The overlying Zone 1 may indicate the more
common, accretional processes that produced manN. if not
most of the shell middens along the lower Atlantic coast: that
is. daily discard of shell and bone, common trampling to both
crush and homogenize many of the deposits. and the inclusion
(intentional and unintentional) of soils to the shell matrix. On
the other hand, it may be the result of millennia of humus and
aeolian sand deposition accompanied by compaction bx
postdepositional prehistoric and historic activity.
Another type of shell midden underlax all of the aforemen-
tioned features. Feature 1. the core of the nng. overlay a thin
(< 10 cm) shell midden (Zone 1BS [below shell]) very similar
to Zone 1. This stratum was not visible in profile beneath
Feature 1 but is apparent on the western side of the ring in
both profiles. In plan. Zone IBS was distinguished from
Feature 1 material by the presence of more humic sand and
both whole and crushed shell, but the stratum was irregular in
area and in depth. It is unclear whether Zone IBS represents
mixing of Feature 1 matenals with the underlying earth
midden stratum (by trampling of initial Feature 1 deposits?) or
whether this is a distinct deposit between the underlying earth
midden and the feature.
The earth midden, which appears in profile in a gently
sloping configuration (Figure 4. G1. G2. G4). was a dark gray
to very dark gray, fine humic sand between 10 and 30 cm
deep. This is a B horizon, not a buried A horizon (Scudder.
personal communication, 1999). Sherds were common in this
stratum; there was less small bone. but bone was still present.
Shell was scattered to absent. This stratum could represent a
period of site occupation prior to exploitation of estuanne
resources, but it was co-extensive with the ring, suggesting
that it is related.

2004 VOL. 57(4)


Figure 5. Rollins trench profile photo, vertical height exaggerated.


Originating within this stratum (not in the shell above) and
extending into sterile tan sand was a series of subcircular
features (Figures 6 and 7). Each of these features was photo-
graphed, mapped, bisected, and profiled at the top of sterile
sand, at ca. 1.70 mbs. Once profiled, the features resolved into
two groups. Features 7, 8, 10, 11, 12, and Area 15 contained
a similar fill of mixed brown to grayish brown sand. Features
8 and 12 had small amounts of shell at their surfaces but the
shell did not extend deeply into the feature fill. This shell may
have been introduced from the base of Feature 1 by root or
rodent disturbance. At the base each of these features, circular
to subcircular areas of sand heavily mixed with particulate
charcoal were encountered. These were about 10 cm in
diameter and between two and five cm deep. These shallow
pits did not extend far below the base of the earth midden
(Figure 7), with the exception of Feature 12, which also had
the least charcoal at the base. Feature function is unknown.
They were probably not postholes with post bases burned in
situ, because no wood grain was evident in the charcoal. They
also were not pits used for cooking with baked clay objects, as
not a single baked clay object or object fragment has been
recovered from Rollins. These features may have been some
other kind of roasting or steaming pits, smudge pits, or they
could have been used to burn some kind of organic offeringss.
The second set of features, Features 5, 6, and 9, also had
predominantly brown sand fills. Features 5 and 6 were
shallow; Feature 5 had a concave base with some evidence of
root disturbance while Feature 6 had a flat base just 10 cm
below the base of the earth midden and may reflect a dip in the
original ground surface. Shell at the top of Feature 6 did
extend into the feature fill and was given a separate feature
number, Feature 6a. It was profiled in the trench wall (D3),
and may be root or rodent disturbance. In contrast, Feature 9
was a deep pit, extending 60 cm below the base of the earth
midden. There was no evidence of burning associated with
this feature. It may have been a storage pit. A separate,
smaller shell-filled pit intruded into Feature 9 and indicates
subsequent, possibly unrelated activity in the same area.
Finally, after removal of the earth midden throughout the
trench-when most of the trench floor was in yellowish brown
(10YR5/6) sterile sand-an apparently circular area of dark
grayish brown earth midden remained in the western units
(Units 3, 5, and 7) (Figure 6). The size and shape of this
feature (Feature 13) suggested that it might be a house floor,
though no postholes were visible. In order to more fully
explore the feature configuration, a 1 x 1 m unit, Unit 9, was
opened on the north wall of the trench (Unit 5, east V2). When
the unit was brought level with the trench, it was clear that
Feature 13 did not extend north of the trench. Rather, at the
base of Level 18 (1.80 mbs; below Feature 3), soils in Unit 9
still contained shell and bone. Soil texture was "sticky" (field
forms, on file, LSUMNS) and a small area along the north
wall may have contained ash. Sterile sand appeared in the
next level. While this unit revealed no new information on
Feature 13, the area certainly should be explored in the future.
Within the trench, Feature 13 bottomed out within 10 cm of its
identification. The wall profiles indicate that it was simply a

distractingly regular, low area of the earth midden which
followed the slope of the underlying C horizon sands.
Given the modest size of our window into the feature
assemblage, activities prior to ring construction remain
unclear. At present, it can only be affirmed that the ring was
built on a site previously used by Orange peoples in a way that
resulted in negligible shell deposition, at least in the area
under the subsequent ring. It is unclear whether these activi-
ties were related to site clearing or other site preparation
activities that just preceded ring construction or whether they
were much earlier-radiocarbon and other dating and an
attempt at cross-mending sherds between proveniences cannot
completely resolve this question (see below\). However. similar
features were found nowhere else on site except immediately
beneath the ring. In fact, these features do not cover the same
area as the latest ring stratum. Zone 1. but appear only in the
area below Features 1 and 3. This areal congruence between
the underlying features and the ring core strongly suggests that
they are related to subsequent ring deposition.
A summary of site formation processes as seen in the south
trench profile is presented in Figure 8'. Note that for this
figure. I have excluded the last two meters on either end of the
trench. which pertain to the sequential relationship of A2
(Zone 1SB) to the other features in the trench: this remains
unclear. In addition. I have "smoothed" the conjunctions
between some deposits and removed disturbances until the
final depositional episode.
Initial activity in the location of the ring included the
deposition of an earth midden containing potter and bone.
and only minor amounts of shell, on a pre-existing sand ridge.
During the accretion of this earth midden, numerous pits were
dug. Shortly thereafter (based on pottery crossmending-see
below), shell deposition began with a thin lens in the center of
the ring area. This deposit consisted of whole and crushed
oyster and may have been trampled into the pre-existing dark,
organically-enriched sand. Because it was hard to distinguish
from some subsequent deposits, the horizontal extent of this
lens is unclear and so is marked with dashed lines. Immedi-
ately thereafter, the first deposit of Feature 1 was made. The
second and third deposits (Episodes 4 and 5) of the same
whole shell, numerous small fish bone. and virtually no soil.
separated by thin lenses of sand or sand and clay. followed
shortly. Once this initial pile was completed, shell and earth
middens (Episodes 6 and 7: Episode 6 and 1 were difficult to
distinguish and may be conflated) were deposited on the
interior and exterior slopes of the ring. perhaps to stabilize the
extremely loose shell in Feature 1. (This stratum is missing
from the ring interior of the south wall. where Feature 3 and
Feature 1 abut.) Feature 3 East and West were then deposited
on the interior and exterior of the ring0. Finally. Zone 1
overlay all these features (Episode 9). As noted, Zone 1 has a
very different character than Features 1 and 3. and may have
resulted from different activities than those that produced
Features 1 and 3.
Feature 1 and Feature 3 probably represent feasting
remains. Hayden (2001) noted that archaeological evidence
for feasting could be distinguished by "feasting middens" in

2004 VOL. 57(4)

Figure 6. Rollins trench, floor plan at top of sterile sands, elevations as indicated.

Examples of features with particulate charcoal bases

.4 ,

Feature 7

Examples of other features

grading to gray

root disturbance
Feature 5

F Mixed and/or mottled (la) grays, browns, tan sands
B Black (10YR 2/2) sand with particulate charcoal
D Brown to dark grayish brown sands
SSterile tan (10YR 4/3-10YR 5/3) sand

calcium concretions she
S. 3 3

Feature 6 and 6a (shell)


4 a t 12

Feature 12

Figure 7. Selected feature profiles from the Rollins trench. Profiled at top of sterile sand, ca. 1.70 mbs.


Table 3. OCR' dates from Rollins Shell Ring.

FS # Description Date
456 Trench 1, Unit 1, F12. 1.90-1.93 cm bs 3839 +/- 115
459 Trench 1, Unit 1, Fll. 1.97-2.00 cmbs 3830 +/- 114
506MS21 Trench 1, Unit 2. Top of earth midden below Fl, 1.30-1.33 cmbs 3888+/-116
507MS22 Trench 1, Unit 5. Top of earth midden below shell midden, 1.60-1.63 cm bs 3855 +/- 115
1. OCR (oxydizable carbon ratio) dates are based on the fact that, unlike much organic matter, charcoal and the humic material in soil degrade
only very slowly through time (Frink 1995, 1997). As these substances degrade, the relative percentage of easily oxidizable carbon increases.
An OCR date is based on the ratio of the relatively inert versus the readily oxidizable carbons present in soils or charcoal. The advantage of
OCR dates is that most soils can be dated; the procedure is not dependent on amounts of organic matter that would be necessary for
conventional radiocarbon dating. Another advantage is cost, which is $50/sample. Though despised by the operators of radiocarbon labs, the
procedure has been tested against hundreds of radiocarbon dates. Where a number of 1 would equal complete agreement and 0 no agreement
between the two procedures, these tests indicate agreement at a value of 0.98 (Frink 1995). For more information or to submit samples, go
to http:// members.aol.com/dsfrink/ocr/ocrpage.htm.

which single deposits contain massive amounts of food
remains. While the horizontal extent of Feature 1 and related
deposits is unknown, it is telling that similar deposits were
found in the Russo's 1992 Unit 4850N250E, about 15 m north
and west of the trench.

Radiocarbon and OCR dating

Radiocarbon and OCR dates are presented in Tables 1 and
3 and Figure 9. The oldest date on the site is from Feature 11,
which had a calibrated intercept of 4089 cal B.P. (see Table 1
for calibrated ranges and other information). The bulk carbon
radiocarbon date from Feature 11 is older than the OCR date
on the feature (Table 3), but single sigmas overlap". Indeed,
all four OCR dates from the subring features and the top of the
earth midden are essentially contemporaneous, and suggest
that initial use of the ridge dates to ca. 3850 B.P.12.
Shell from the base of Feature 1 was radiocarbon dated to
3617 3449 cal B.P. (1 sigma) This suggests that the earth
midden might be somewhat older than the initial shell
deposits-that there may have been a hiatus in site use
between the two proveniences. However, the dates from the
earth midden and the base of Feature 1 overlap at two sigma,
so site use may have been continuous. Certainly pottery from
all three strata, Zone 1, Feature 1, and the earth midden,
appear similar, so occupations did not straddle Orange phases.
Pottery cross-mending was undertaken to answer questions
about the temporal relationship of proveniences on the site.
For pottery over 3 cm (n = 801), only six crossmends between
proveniences were found. One of these, between Feature 1 in
Level 12 and Zone 2/1 (one of a series of transitional zones
near the base of the earth midden) in Level 14, suggests
contemporaneity between the earth midden and Feature 1, but
more crossmends would be more persuasive. A corrected,
calibrated shell sample from the top of Feature 1 dated to cal
3518 3375 cal B.P. (1 sigma), nearly contemporaneous with
the basal deposit. This, along with the stratigraphic implica-

tions for dumping of massive piles of food remains, indicates
a rapid build-up of this part of the main ridge.

Was Rollins Shell Ring a Feasting Site?

Rollins Shell Ring is a unique site for the Orange III period
in the area between the Nassau and St. Johns Rivers. The ring
is a large, topographically complex construction that cannot be
explained with reference to simple egalitarian village plans.
Radiocarbon dating suggests that the ring was constructed
quickly; top and bottom dates for the core of the ring were
separated only by some 50-100 years (but could also be
considered contemporaneous).
Is the ring the result of ritual involving feasting, or
everyday discard? I argue for the former. The evidence for
this includes: site context (it is the only Orange III ring in the
lower St. Johns drainage); intrasite organization suggesting
purposeful maintenance of a ring structure throughout the
occupation of the site; and deposits indicating purposeful
mounding and little evidence of post-depositional crushing.
On the basis of this and other evidence, it appears that Rollins
was a special purpose site where Orange III culture popula-
tions of the area aggregated seasonally for feasting and other
activities. The ring itself was probably constructed from the
remains of these feasts, piled up as a display of the success of
the corporate group (cf Russo et al. 2002). Elsewhere, I have
presented evidence indicating strong seasonality in the Rollins
feasting deposits (Saunders 2003; see also Russo 2002 for
seasonality at the Fig Island Ring Complex) and demonstrated
that the frequency of decorated pottery is much higher at
Rollins than at contemporaneous, sheet midden sites (Saunders
2003, 2004). Taken together, these data strongly suggest a
special purpose site.
That the ring constituted a separate facility for feasting and
other macroband activities is consistent with cross-cultural
comparative studies that demonstrate an association between
feasting and spatial differentiation; in other words, feasting is



0- Episode 8

o-- o -1

0- Episode 8


o- Episode 7


0- Episode b


Episode 2


0- Episode 1


0- Undsturbed


Figure 8. Reconstruction of depositional episodes at Rollins Shell Ring.

2004 VOL. 57(4)


3670 70 3600 60


-2- I
3888 116 3839 115
OCR 3830' 114 OCR 3855 115
3730 + 80
0 2 4 6 8 10 12 14 16

Figure 9. Radiocarbon and Oxydizable Carbon Ratio (OCR) dates from Rollins Trench 1 (RC dates are corrected, not calibrated; see Table 1 for calibrated


often done in special structures spatially separated from
domestic or village life (Adler and Wilshusen 1990; Deitler
and Hayden 2001: Hayden 2001). Indeed, "it is important to
emphasize that not only were these structures specially
constructed sites for feasting, but they were most probably
constructed through feasting. That is, the more marked the
architectonic elaboration, the more such features represent the
congealed labor of work feasts and are, in effect, an advertise-
ment of the feasts that went into their construction" (Dietler
and Hayden 2001:9). This labor, over and above that neces-
sary for simple function, qualifies shell rings as monumental
architecture. According to Trigger (1990), monumental
architecture is conspicuous consumption: at rings, consump-
tion is made conspicuous.
This has been a subject of debate for some time (e.g.. Cable
1997: Trinkley 1985). Rejection of the idea that shell rings
were a form of monumental architecture is based on both
theoretical preconceptions and the general unavailability of
data on the rings themselves. Each of these is examined
briefly below.
Until recently, unilineal theories of cultural evolution have
reserved monumental architecture for hierarchical societies
with an ascribed elite (agriculture was slowly withdrawn from
this equation over the last twenty years or so). This elite class
was needed (theoretically) to control (or cajole) labor for these
projects. However, an infusion of theory from feasting studies
is changing conceptions of what kinds of social structures are
necessary to produce monumental architecture. With respect
to ring construction, perhaps the most useful element of the
anthropological description and interpretation of feasting is its
use as a means to mobilize labor in egalitarian and
transegalitarian societies. In fact, feasting events appear to be
the principal means by which labor is recruited in non-
capitalist societies. While it is important to acknowledge that
feasting is also used to mobilize labor in societies with
ascribed positions of power (or, more properly, feasting
justifies the use of labor for the gain of a single individual or
an elite class of people [e.g., Dietler and Herbich 2001]), the
recognition that large groups of people can be mobilized to
perform tasks on a more corporate, cooperative level (Blanton
et al. 1996) has promulgated a re-evaluation of how some
monumental architecture was produced and used in the
southeastern United States (e.g., Knight 2001). Certainly, "a
fuller awareness of the range and operation of such practices
exposes the inadequacies of assumptions. ..such as simplistic
correlations between the existence of large-scale earthworks
and the necessity of centralized political organization" (Dietler
and Herbich 2001:257). In other words, mounds don't equal
chiefs (Gibson 2001).
There is no question that feasting was an integral part of
southeastern Native American Indian ceremony. There is
ample ethnohistoric information suggesting seasonal feasting
(e.g., Le Page Du Pratz 1972: Hudson 1976: Laudonniere
1975; Swanton 1979); the Green Corn ceremony is still
performed today. In addition, archaeological evidence
suggests deep antiquity for feasting at special purpose sites
(e.g., Blitz 1993; Kelly 2001; Knight 2001; Milanich et al.

1997: Smith and Williams 1994: VanDenrarker 1999).
Recently Knight (2001) has re-interpreted Middle Woodland
platform mounds as fundamental to ceremonialism centering
on world renewal and feasting. Shell rings, constructed
through feasting, may well be antecedent to these more widely
recognized monuments.
The monumentality of rings may also have been over-
looked because of the paucity of readily available information
on site structure, stratification, and other important informa-
tion on rings. Shell rings were excluded by the adjective
"earthen" in a 1998 volume on ancient enclosures-sacred and
secular-of the eastern Woodlands (Mainfort and Sullivan
1998). In that volume, Poverty Point is cited as the earliest
enclosure (Mainfort and Sullivan 1998:2). though dates from
shell rings are up to 500 years earlier". With the exception of
Waring and Larson's (1968) work on the Sapelo Shell Ring
complex, and Trinkley's (1985) influential article on rings as
egalitarian village sites, most information on ring excavations
is either unpublished or buried in reports w ith limited distribu-
tions (see Saunders 2002a for a revieww). In addition, until
very recently, research on Atlantic coastal rings had stalled,
with no new, extensive excavations conducted between 1979
and 1998 (though, as noted, Rollins was tested by Russo in
1991 [Russo 1993]).
A review of previous excavations and comparison of results
with more recently excavated material from Rollins (Russo and
Saunders 1999; Saunders 1999) and the Fig Island Shell Ring
complex (Saunders 2002b) indicated a fairly diverse assem-
blage of ring sizes and shapes (Russo and Heide 2001), but
also some commonalities of ring composition (Saunders
2002a). Most important for considerations of ring function
was the frequent description of "loose, clean, whole oyster" as
the principal ring fill in at least six ring sites. At Rollins and
at Fig Island, enormous deposits of loose whole oyster were
mapped. The shell in these deposits was oriented every which
way, indicating dumping: this dumping, along with the height
of Rollins and many other rings (Fig Island 1 is almost 7 m
high) must indicate that mounding was deliberate. Further,
there was no indication at either Rollins or Fig Island of the
crushing and lensing that would occur if the surfaces of the
rings accreted gradually over time and were inhabited on a
daily basis.
The huge deposits of shell contained predominantly oyster
and small net-able fishes-at Rollins, scianids, and at Fig
Island, catfish. This conforms to the menu for cooperative
feasts prepared by Hayden (2001). While feasting menus for
societies higher up the food chain, where feasts are competitive
and promotional, are likely to involve unusual or scarce foods
from high trophic levels, more egalitarian, solidarity feasts are
likely to involve an abundance of common foods from lower
trophic levels that are resistant to overexploitation (Hayden
2001). Faunal assemblages from Rollins and Fig Island fit
that bill of fare exactly. And the bulk of the stratigraphic and
zooarchaeological data from what are presumed to be succes-
sive feasting episodes indicate highly seasonal deposits (Russo
2002:Figure 39; Saunders 2003).
Hayden (2001; among others) also suggested that serving


2004 VOL. 57(4)


vessels at feasting sites might be of unusual quality or size.
This has been demonstrated at many sites of many different
time periods in the Southeast. Blitz (1993) demonstrated that
Mississippian "big shots" had big pots at the Lubbock Creek
site in Alabama; Knight (2001) cites examples of special
purpose wares at several Middle Woodland mound sites. As
noted above, the Rollins assemblage is distinctive in its high
frequency of decorated wares. The vessel assemblage, with its
concentration on shallow bowls, may indicate the predomi-
nance of serving bowls.


The Rollins Shell Ring site is most likely an example of
early monumental architecture, one of over 30 known exam-
ples, along the lower Atlantic coast. Itjoins a growing number
of other recently recognized examples of Middle and Late
Archaic monumental architecture in the Southeast. Over 20
Archaic earthen mound sites are known in the Lower Missis-
sippi River Valley and there are numerous Archaic shell and
earthen mounds in Florida as well (Russo 1996). Just how
labor was mobilized to construct these monuments, not to
mention why they were constructed, remains a mystery (see
Gibson and Carr 2004 for a range of opinions). However,
there does seem to be an emerging consensus that at least some
Archaic mound (including ring) construction could have been
a cooperative effort requiring little in the way of status
hierarchy to mobilize the labor force. At smaller, unelaborated
rings, construction was probably achieved through combining
the labor with feasting, religious ceremony, music, dancing,
mate selection, and gossip (information exchange). Access to
ring activities appears to have been more or less unrestricted
and the entire affair probably had an incorporative function.
However, the variability in ring configuration and size
suggests that some societies may have upped the ante. In
particular, the addition of ringlets, which would admit only a
fraction of the population that could participate in activities in
the main ring, may indicate increasingly5 restrictive entry into
some areas and, by extension, increasing social differentiation.
The configuration of rings may be some of our earliest
evidence for social stratification, though this hypothesis will
take much additional work to confirm.


For other criticism of Trinkley's hypothesis, see Russo and Heide

2 Though Cable believed the sites around Sewee to be Late Archaic,
survey reports suggest only Mississippi Period occupations around
Sewee (Michael Russo, personal communication, 2003).

3 Sterile sand exposed in the base of the trench was higher under the
shell than on either the interior or the exterior of the ring. This may
indicate that a portion of the ring was built on a naturally occurring
ridge. Alternatively, the shell may have protected the original C-
horizon sands from erosion by wind or water (Michael Russo,
personal communication, 2003) or from being swept away or
compacted by cultural activities. In either event, the "ridge" would

then be a remnant of the original C-horizon elevation.

4 These ringlets were so unusual with respect to known ring configu-
rations at the time Rollins was excavated, that plantation period or
modern shell borrowing were suspected to have produced those
irregular shapes. The soils analysis was designed to address this, as
well as whether or not shell once extended across the ring opening.
Results indicated that no shell borrowing had taken place in the areas
tested. Subsequent to our investigations, two other sites (Fig Island
Ring Complex [38CH42] and Sewee Shell Ring [38CH45] have been
mapped with ringlets.

5 In the nomenclature used here, "Features" are the result of discrete
episodes of human behavior. "Areas" are less surely so, and could be
the result of human or natural processes. "Zones" are the general soil
matrix. All Features, Areas, and Zones were excavated separately.

6 Profiles were drawn according to microstrata observed in the walls;
we intentionally did not impose features or areas as mapped in plan
onto the walls. Similarly, we explicitly avoided imposing a
depositional sequence on the deposits observed in profile.

7 Pot dumps are small (ca. 20-30 cm diameter) areas of shell, usually
coquina or periwinkle, that appear to have been dumped from a
cooking pot after broth has been made.

8 Features with evidence of heating or burning have been found at the
bases of Archaic mounds in Louisiana (Saunders 1994).

9 Russo and Heide (2003) should be credited for creating the first
"sequence profiles" for Rollins. My interpretation ofthe stratigraphic
sequence, which is based on field forms as well as the final profiles,
differs from theirs (Russo and Heide 2003:Figure 20) and one they
attributed to me (Russo and Heide 2003:Figure 21).

" As mapped, Feature 3 E on the south wall appears to have been
deposited before or at the same time as the last episode of Feature 1.
The appearance of Feature 3 E on the north wall, and Feature 3 W
and W on the north and south wall, argues against this conclusion.

" These OCR dates have relatively large sigmas because the samples
were 2-3 cm deep (Douglas Frink, personal communication, 1999).
Ideally, samples should be ca. 1 cm thick. Note also that OCR dates
have the highest correlation with calibrated intercept ofa radiocarbon
date (Douglas Frink, personal communication, 2003).

12 Two dates, one from the top of Unit 3197 (FS #467) and one from
a feature in Ringlet J (FS #281) indicate late Orange period activity
at portions of the site. This may account for some of the shell
crushing observed in Unit 3197.

13 Watson Brake (160U175), with calibrated intercept dates from
mound bases as early as 5600 B.P. (Saunders et al. 1994), currently
has the distinction as the earliest enclosure; Middle Archaic mounds
in the lower Mississippi River Valley may date as early as 7000 B.P.
(Russo 1996).

14 This is more of that gray literature, but a CD of the report is
available from the author.

"s The two proveniences dated from a ringlet associated with the Fig
Island 1 shell ring (38CH42) were ca. 200 years younger than the
dates from the main ring (Saunders 2002b:Table 7).





The research presented here was funded through the National
Geographic Society Research and Exploration program, Grant #
6018-97. In addition, we had logistical support from Talbot Island
State Park and the participation of numerous volunteers. Thanks to
everyone involved.

References Cited

Adler, M.A. and R.H. Wilshusen
1990 Large-scale Integrative Facilities in Tribal Societies: Cross-
cultural and Southwestern US examples. WorldArchaeol-
ogy 22(2):133-146.

Blanton, Richard E, Gary M. Feinman, Stephen A. Kowaleski, and
Peter N. Peregrine
1996 A Dual-Processual Theory for the Evolution of
Mesoamerican Civilization. Current Anthropology

Blitz, John IH.
1993 Big Pots for Big Shots: Feasting and Storage in a Missis-
sippian Community. American Antiquity 58(1):80-95.

Cable, John S.
1997 The Ceremonial Mound Theory: New Evidence for the
Possible Ceremonial Function of Shell Rings. South
Carolina Archaeology Week Poster, Shell Rings of the Late
Archaic. South Carolina Institute of Archaeology and
Anthropology, University of South Carolina, Columbia.

Calmes, Alan
1967 Test Excavations at Two Late Archaic Sites on Hilton Head
Island. Ms. On file, South Carolina Institute of Archaeol-
ogy and Anthropology, University of South Carolina,

DePratter, C.B.
1979 Shellmound Archaic on the Georgia Coast. South Carolina
Antiquities 11:1-69.

DePratter, Chester B., and James D. Howard
1980 Indian Occupation and Geologic History of the Georgia
Coast: a 5,000 Year Summary. In Excursions in Southeast-
ern Geology: The Archaeology -Geology of the Georgia
Coast, by James D. Howard, Chester B. DePratter, and
Robert W. Frey, pp. 1-65. Geological Society of America,
Guidebook 20. Georgia Department of Natural Resources,

Dietler, Michael, and Brian Hayden
2001 Digesting the Feast-Good to East, Good to Drink, Good
to Think: An Introduction. In Feasts: Archaeological and
Ethnographic Perspectives on Food, Politics, and Power,
edited by Michael Dietler and Brian Hayden, pp. 1-22.
Smithsonian Institution Press, Washington.

Dietler, Michael, and Ingrid Herbich
2001 Feasts and Labor Mobilization: Dissecting a Fundamental
Economic Practice. In Feasts: Archaeological and
Ethnographic Perspectives on Food, Politics, and Power,
edited by Michael Dietler and Brian Hayden, pp. 240-264.

Smithsonian Institution Press, Washington.

Edwards, William E.
1965 A Preliminary Report on the Sewee Mound Shell Ring,
Charleston County, South Carolina. Manuscript on file,
South Carolina Institute ofArchaeology and Anthropology,
University of South Carolina, Columbia.

Frink, Douglas S.
1995 Application of the Oxidizable Carbon Ratio Dating Proce-
dure and Its Implications for Pedogenic Research In
Pedologicl Perspectives in Archaeological Research, Soil
Science Society of America, Special Publication 44, pp. 95-

1997 OCR Carbon Dating of the Watson Brake Mound Complex.
Paper presented at the 53rd Annual Meeting of the South-
eastern Archaeological Conference. Bimningham, Alabama

Ford, James B.
1966 Early Formative Cultures in Georgia and Florida. Amer-
ican Antiquity 31:781-99

1969 A Comparison of Formative Cultures in the Americas.
Smithsonian Contributions to Anthropology 11 Smithso-
nian Institution, Washington, D.C.

Gibson, Jon
2001 The Ancient Mlounds ofPoverty Point: Place of'Rings.
University Press of Florida, Gaiesville.

Gibson, Jon, and Phillip Carr, eds.
2004 Signs of Power: The Rise of Cultural Complexity m the
Southeast, University of Alabama Press, Tuscaloosa (in
press). University of Alabama Press. Tuscaloosa.

Hayden, Brian
2001 Fabulous Feasts. A Prolegomenon to the Importance of
Feasting. In Archaeological and Ethnographic Perspec-
tives on Food, Politics, and Power, edited by Michael
Dietler and Brian Hayden, pp. 23-64 Smithsonian Institu-
tion Press, Washington, D.C.

Hudson, Charles
1976 The Southeastern Indians. University of Tennessee Press,

Kelly, Lucretia S.
2001 A Case of Ritual Feasting at the Cahokia Site. Archaeo-
logical and Ethnographic Perspectives on Food, Politics,
and Power, edited by Michael Dietler and Brian Hayden,
pp.334-367. Smithsonian Institution Press, Washington,

Knight, Vernon James
2001 Feasting and the Emergence of Platform Mound
Ceremonialism in Eastern North America. InArchaeologi-
cal and Ethnographic Perspectives on Food, Politics, and
Power, edited by Michael Dietler and Brian Hayden, pp.
311-333. Smithsonian Institution Press, Washington, D.C.

Laudonnierre, Rene
1975 [1586] Three Ioyages, translated by Charles Bennett.


2004 VOL. 57(4)


University Presses of Florida, Gainesville.

Le Page Du Pratz, Antoine S.
1972 [1758] The History of Louisiana. Claitor's Publishing
Division, Baton Rouge, Louisiana.

Mainfort, Robert C. and Lynne P. Sullivan
1998 Explaining Earthen Enclosures. In Ancient Earthen
Enclosures of the Eastern Woodlands, edited by R.C.
Mamfort and L.P. Sullivan, pp. 1-16. University Press of
Florida, Gainesville.

Marrinan, Rochelle A.
1975 Ceramics, Molluscs, and Sedentism: The Late Archaic
Period on the Georgia Coast. Ph.D. dissertation, Depart-
ment of Anthropology, University of Florida, Gainesville.

McKinley, William
1873 Mounds in Georgia. Annual Report to the Board of
Regents, Smithsonian Institution 22:422-428.

Milanich, Jerald T., Ann S. Cordell, Vernon J. Knight, Jr., Timothy
A. Kohler, and Brenda J. Sigler-Lavelle
1997 Archaeology of Northern Florida, A.D. 200-900: The
McKeithen Weeden Island Culture. University Press of
Florida, Gainesville.

Michie, James L.
1979 The Bass Pond Sites: Intensive Archaeological Testing at
a Formative Period Base Camp on Kiawah Island, South
Carolina, Columbia. Research Manuscript Series 154,
South Carolia Institute of Archaeology and Anthropology,
University of South Carolina, Columbia.

Russo, M.A.
1993 The Timucuan Ecological and Historic Preserve Phase III
Final Report. SEAC Accession Number 899. Florida
Museum of Natural History, Gainesville.

1996 Southeastern Archaic Mounds. In Archaeology ofthe Mid-
Holocene Southeast, edited by Kenneth E. Sassaman and
David G. Anderson, pp. 259-287. University Press of
Florida, Gainesville.

2002 Faunal Analysis at Fig Island. In The Fig IslandRing
Complex (38CH42): Coastal Adaptation and the Question
of Ring Function in the Late Archaic, edited by Rebecca
Saunders, pp. 141-154. Report prepared for the South
Carolina Department of Archives and History, Grant #45-

Russo, M. A., and G. Heide
2001 Shell Rings of the Southeast US. Antiquity 75(289):491-

2002 The Joseph Reed Shell Ring. The Florida Anthropologist

2003 Mapping the Sewee Shell Ring. Report submitted to the
Francis Marion and Sumter National Forests, South

Russo, Michael A., Gregory Heide, and Vicki Rolland

2002 The Guana Shell Ring. Report Submitted to The Florida
Department of State, Division of Historical Resources,
Historic Preservation Grant No. F0126.

Russo, Michael A. and Rebecca Saunders
1999 Identifying the Early Use of Coastal Fisheries and the Rise
of Social Complexity in Shell Rings and Arcuate Middens
on Florida's Northeast Coast. Report submitted the
National Geographic Society under Grant #6018-97,
Washington, D.C.

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

Saunders, Joe W., Thurman Alien, and Roger T. Saucier
1994 Four Archaic? Mound Complexes in Northeast Louisiana.
Southeastern Archaeology 13(2):134-153.

Saunders, Rebecca
1994 The Case for Archaic Period Mound Sites in Southeastern
Louisiana. Southeastern Archaeology 13(2):118-134.

1999 Feast or Quotidian Fare?: Rollins Shell Ring and the
Question of Site Function. Paper Presented at the 56h
Annual Meeting of the Southeastern Archaeological
Society, Pensacola, Florida.

2002a Previous Archaeological Research. In The Fig Island Ring
Complex (38CH42): Coastal Adaptation and the Question
of Ring Function in the Late Archaic, edited by Rebecca
Saunders, pp. 43-62. Report prepared for the South
Carolina Department of Archives and History, Grant #45-

2002b Field Excavation : Methods and Results. In The Fig Island
Ring Complex (38CH42): Coastal Adaptation and the
Question of Ring Function in the Late Archaic, edited by
Rebecca Saunders, pp. 98-140. Report prepared for the
South Carolina Department of Archives and History, Grant

2003 Feast or Quotidian Fare: Rollins Shell Ring and the
Question of Ring Function. Report prepared for the
Florida Department Historical Resources, under permit #
1A-32 9697.04.

2004 Spatial Variation in Orange Period Pottery: Spatial Varia-
tion and Function. In EarlyPottery: Technology, Function,
Style, and Interaction in the Lower Southeast, edited by
Rebecca Saunders and Chritopher T. Hays. In press,
University of Alabama Press, Tuscaloosa.

Smith, Marvin T. and Mark Williams
1994 Mississippian Mound Refuse Disposal Patterns and
Implications for Archaeological Research. Southeastern
Archaeology 13(1):27-35.

Swanton, John R.
1979 [1946] The Indians of the Southeastern United States.
Classics of Smithsonian Anthropology, Smithsonian
Institution Press, Washington, D.C.




Trigger, Bruce G.
1990 Monumental Architecture: A Thermodynamic Explanation
of Symbolic Behavior. WorldArchaeology 22(2); 121-132.

Trinkley, Michael B.
1985 The Form and Function of South Carolina's Early Wood-
land Shell Rings. In Structure andProcess in Southeastern
Archaeology, edited by R.S. Dickens and H.T. Ward, pp.
102-118. University of Alabama Press, Tuscaloosa.

1997 The Lighthouse Pont and Stratton Place Shell Rings. South
Carolina Archaeology Week Poster, Shell Rings of the Late
Archaic. South Carolina Institute of Archaeology and
Anthropology, University of South Carolina, Columbia.

VanDerwarker, Amber M.
1999 Feasting at Status at the Toqua Site. Southeastern Archae-
ology 18(1):24-34.

Waring, Antonio J., Jr.
1968a The Archaic and Some Shell Rings. In The Waring
Papers. The Collected Works of Antonio J. Waring, Jr.,
edited by Stephen Williams. Papers of the Peabody
Museum of Archaeology and Ethnology 58:243-246.

1968b The Bilbo Site, Chatham County, Georgia. In The Waring
Papers: The Collected Works of Antonio J. Waring, Jr.,
edited by Stephen Williams. Papers of the Peabody
Museum of Archaeology and Ethnology 58:152-197.

Waring, Antonio J., Jr., and Lewis H. Larson
1968 The Shell Ring on Sapelo Island. In The Waring Papers:
The Collected Works of Antonio J. Waring, Jr., edited by
Stephen Williams. Papers of the Peabody Museum of
Archaeology and Ethnology 58:263-278.


2004 VOL. 57(4)

Florida Anthropological Society Chapters

9 7~

1) Archaeological Society of Southern Florida
2495 NW 35th Ave., Miami, FL 33142

2) Broward County Archaeological Society
481 S. Federal Highway, Dania Beach, FL 33004

3) Central Florida Anthropological Society
P.O. Box 947544, Maitland, FL 32794-7544

4) Central Gulf Coast Archaeological Society
P.O. Box 82255, St. Petersburg, FL 33682

5) Indian River Anthropological Society
3705 S. Tropical Trail, Merritt Island, FL 32952



6) Kissimmee Valley Archaeological and Historical Conservancy
195 Huntley Oaks Blvd., Lake Placid, FL 33852

7) Emerald Coast Archaeological Society
333 Persimmon Street, Freeport, FL 32435

8) Panhandle Archaeological Society at Tallahassee
c/o The Tallahassee Trust for Historic Preservation
423 E. Virginia Street, Tallahassee, FL 32301

9) Pensacola Archaeological Society
P.O. Box 13251, Pensacola, FL 32591

10) St. Augustine Archaeological Association
P.O. Box 1301, St. Augustine, FL 32085

11) Southeast Florida Archaeological Society
P.O. Box 2875, Stuart, FL 34995

12) Southwest Florida Archaeological Society
P.O. Box 9965, Naples, FL 34101


13) Time Sifters Archaeology Society
P.O. Box 25883, Sarasota, FL 34277-2883

14) Volusia Anthropological Society
P.O. Box 1881, Ormond Beach, FL 32175

15) Warm Mineral Springs Archaeological Society
P.O. Box 7797, North Port, FL 34287

Florida's Lost Tribes
Theodore Morris
Commentary by Jerald T. Milanich
"Theodore Morris's well-
researched illustrations of
Florida's lost tribes provide an
extraordinary glimpse of a past
that has forever disappeared. This
book is the next best thing to time
travel!"-Mallory O'Connor,
author of Lost Cities of the Ancient
Cloth $29.95

Order through full-service
booksellers, our website at
www.upf.com, or with VISA,
American Express, or M/C toll
free: 1-800-226-3822

The Calusa and
Their Legacy
South Florida People and Their
Darcie A. MacMahon
and William H. Marquardt
"Finally, a well-researched and
entertaining look at the grand
procession of life that has been
flourishing in South Florida's
estuaries for thousands of years.
This book masterfully describes
the wondrous and little-known
stories of its inhabitants-from
plankton to mangroves to the
ancient Calusa Indians to modern-
day people."-Carol Newcomb-
Jones, Florida Gulf Coast
Cloth $39.95

The Seminole Wars
America's Longest Indian
John and Mary Lou Missall
"In this insightful book the con-
flicts known as the Seminole Wars
are placed in the larger context of
American history. Twenty-first-
century Seminole Indians and all
other Floridians have been shaped
in part by those nineteenth-
century events."-Jerald T.
Milanich, Florida Museum of
Natural History
Cloth $29.95

Indians of Central

John H. Hann
" The first survey of Florida's
natives that includes some of the
richest cultural history in the
nation."-American Archeology
* From their earliest contact with
Europeans to their disappearance
in the 18th century, the story of the
Calusa of the southwest coast, the
people of the Tampa Bay region,
the Surruque and Ais and their kin
of the east coast from Turtle
Mound southward through the
Keys, as well as their hinterland
kin from the St. Johns to
Kissimmee valleys.
Cloth $39.95


Native Georgia
Max E. White
"A concise history of Native
Americans in Georgia.... White
draws on an extensive body of
archaeological and historical data
to tell the story of these people
and their impressive accomplish-
ment ... a very readable account
of this story."-American
MARCH. $24.95


Games ile TaIllaassee Tampa Boc Raton
Pensacoia Orlando Miami Jacksonville Fort Myers


The Florida Journals

of Frank Hamilton
Edited by Phyllis E. Kolianos
and Brent R. Weisman
"In addition to a wealth of
archaeological evidence, Frank
Hamilton Cushing left a treasure
trove of fascinating images of
Florida's Gulf Coast as it appeared
to him in the late 19th century. I
wish I could visit these places and
see what his eyes saw more than
100 years ago."-Barbara A.
Purdy, professor emerita,
University of Florida
FEBRUARY. Cloth $49.95

The Lost Florida

Manuscript of Frank

Hamilton Cushing
Edited by Phyllis E. Kolianos
and Brent R. Weisman
"Frank Cushing's long-lost
archaeological manuscript adds
important details on the Hope and
Safford mounds as well as a host
of other coastal sites in southwest
Florida."-William H. Marquardt,
curator in archaeology, Florida
Museum of Natural History
FEBRUARY. Cloth $59.95

Gulf Coast
The Southeastern United States
and Mexico
Edited by Nancy Marie White
"Puts the topic of Mesoamerican
contact back on the table. ... At
last, it's okay to look to Mexican
cultures for instruction and inter-
pretation of a Pan American belief
system."-Cheryl Claassen,
Appalachian State University
MAY. Cloth $65.00

and South Florida, The Archaeology

1513-1763 and History of the




Department ofAnthropology, University ofFlorida, Turlington Hall, Room 1112, P.O. Box 117305, Gainesville, FL 32611-7305
E-mail: njwallis@ufl.edu

From ca. A.D. 300 to 900, the Mayport Mound (8DU96)
was a site of repeated mortuary ritual, where pottery was
broken and deposited and bodies were interred. Although the
burial rituals themselves appear unaltered and marked by
fairly constant practices of interment over a period of several
centuries, the pottery traditions represented within the mound
assemblage are wildly diverse and characteristic of the local
Woodland period. Similar in both morphology and assem-
blage to at least 14 other sand mounds located between
Jacksonville and the Atlantic Ocean (Ashley 1998:209; Ashley
and Wallis 2004), the Mayport Mound might represent just
one ceremonial site in a constellation of other meaningful
places across the landscape. Examination of the pottery
assemblages within these mounds has the potential to reveal
changing cultural affiliations for participants at mound events.
The typological complexity of pottery assemblages in the
lower St. Johns area has been variously interpreted by archae-
ologists as evidence of imported trade wares (Goggin
1952:49), periodic influxes of different culturally affiliated
groups (Wilson 1965:30), or mixture of St. Johns and coastal
Georgia wares due to its border location (Sears 1957:35).
While the importance of external influence has long been
acknowledged, simplistic explanations involving migration,
trade, or diffusion have done little to illuminate the mecha-
nisms driving these dynamic patterns of social integration. In
light of a recently developed Woodland period chronology
(Ashley 2003:75), however, a more nuanced explanation of
past social processes can be derived using provenienced
ceramic assemblages to monitor technological changes in
pottery production through time. In the following analysis of
the Mayport Mound pottery assemblage, I monitor variability
in technological style-the technological choices made during
pottery production-in order to infer continuity and change in
social influences that affected the way pots were produced and
used for mound events.

Woodland Period Chronology of Northeastern Florida

As a prologue to the following discussion, a brief explana-
tion of the Woodland period culture history within the lower
St. Johns is necessary. Following a half century of taxonomic
development and chronology building, Ashley (2003:75)
produced the most locally relevant and useful chronology yet
presented. According to Goggin's (1952:47) initial chronol-
ogy for the lower St. Johns River, ceramic periods largely
conformed to the culture history of the upper southern reaches

of the river, with Woodland period assemblages dominated by
plain chalky St. Johns wares. The presence of exotic minority
wares such as Deptford, Swift Creek, and Weeden Island, were
used as temporal markers to designate St. Johns I, Ia, and Tb,
respectively. Bullen and Griffin (1952) and Sears (1957) were
somewhat dissatisfied with Goggin's (1952) chronology, since
it was based mostly on site data from the St. Johns heartland
to the south and failed to address the lack of sites with pre-
dominantly St. Johns wares within the northern lower St.
Johns. Recognizing the preponderance of sand-tempered
sherds without decoration, Sears (1957) developed a chronol-
ogy that bracketed this "sand-tempered plain complex"
between Deptford and Colorinda, respectively. Based solely on
seriations and in the absence of radiocarbon assays, Sears
assigned the sand-tempered ceramic complex to a time slot of
A.D. 1 to 700. Milanich and Fairbanks' (1980) subsequent
use of Goggin's (1952) chronology with minor alterations,
however, led archaeologists to typically designate Woodland
period sites in the area as St. Johns I, despite the fact that
many Woodland period sites yield a majority of sand-tempered
sherds (Ashley 2003:68). Alternatively, Russo (1992) focused
attention on the distinctive complexity of local Woodland
assemblages and defined a unique "St. Marys" region, but it
was not until recently that Ashley (2003:75), with the aid of
new archaeological data, attempted to bring temporal order to
the minority wares that accompany the ubiquitous sand-
tempered plain complex in northeastern Florida.
Building on Sears' (1957) early chronology, Ashley
(2003:74-76) suggests that the production of Deptford pottery
in northeastern Florida gave way around A.D. 1 to a roughly
800-year period in which sand-tempered plainwares were the
dominant pottery manufactured in the region. During this
lengthy span of time, however, checked stamped and compli-
cated stamped pottery underwent periodic fluorescence. Most
notably, Early Swift Creek wares, both charcoal and sand
tempered versions, are found in mounds and middens assumed
to date to ca. A.D. 300 to 500. After A.D. 500, Late Swift
Creek sherds are the predominant minority ware, ending in
popularity with the emergence of the brief Colorinda phase
(A.D. 850-900). This revised chronology enables finer
distinctions to be made regarding changes in the cultural
affinities that persisted through time.

New Dates

While date ranges for Ashley's (2003:74-76) Woodland


VOL. 57(4)




Mayport Mound
Mayport Mo
Dent Mounnd

Dent Mound

I I I I I I l I
A D 100 200 300 400 500 600 700

I = 1 sigma calibrated
= 2 sigma calibrated

Figure 1. AMS dates for charcoal-tempered vessels.

chronology were calculated based on a series of radiocarbon
dates, some types are represented by more assays than others.
For instance, the temporal range assigned for the production
of Late Swift Creek pottery is becoming firmly established
through a battery of radiocarbon dates (Stephenson 2002;
Stephenson et al. 2002), while the designations for Deptford
and Early Swift Creek vessel production can be only tenta-
tively suggested due to a conspicuous lack of radiocarbon
assays (Ashley 2003:81, Ashley and Wallis 2004). As a step
toward better chronological placement, an AMS date was
obtained for a charcoal-tempered vessel using soot from an
exterior portion of the rim (v. 14). This most recent assay
represents one of only three compelling radiometric dates for
charcoal-tempered pottery. Three outlying and likely errone-
ous dates deserve mention. The earliest date for charcoal-
tempered pottery, yielding a 2 sigma range of 30 B.C. to A.D.
135 from a vessel at the Dent Mound, fails to overlap the next
earliest date range by more than a century, and should be
considered with caution. The early assay may in fact be a
product of the old-wood problem, perhaps dating the early age
of the fuel that was burned rather than the cooking activity
itself (Ashley and Wallis 2004). In addition, two late seventh

century dates, taken from oyster shell in mixed features at
McArthur Estates (8NA32) that contained a scant amount
of charcoal-tempered pottery, are likely to reflect the age of
the predominant Late Swift Creek occupation at that site
(Ashley and Wallis 2004, Handley et al. 2004). In light of
these considerations, the more secure AMS dates from soot
must suffice for dating the production of charcoal-tempered
The newest assay, 1510 +/- 40 (Beta-190255: soot: 13C
= -25.3%o). with a calibrated two sigma range of A.D. 440
to 640 and an intercept of A.D. 560, largely substantiates
Ashley's (2003:74) estimation for charcoal-tempered
pottery, but extends its range by at least several decades
(Table 1, Figure 1)'. The three contextually secure assays
tentatively suggest a 300 year production period for charcoal
tempered pottery, from ca. A.D. 300 to 600 (Figure 2).
Given several radiometric assays from Late Swift Creek
contexts that date to the 6th century as well, the years
between ca. A.D. 500 to 600 were likely a transitional period
in pottery production, with the predominance of charcoal-
tempered early forms gradually giving way to the production
of mostly sand-tempered late forms (Ashley and Wallis 2004,
Stephenson 2002). More dates from other sites beyond the
Dent and Mayport Mounds, especially habitation sites, will
greatly increase our temporal understanding of charcoal-
tempered pottery production.

Regional Considerations

In northeastern Florida, Early Swift Creek pottery, in both
rim forms and surface designs, closely resembles vessels from
northwestern Florida sites (Ashley 1998:203). While some of
these vessels could be explained as trade wares, the local
charcoal-tempered variant is unknown outside of the lower St.
Johns River and was almost certainly locally produced (Ashley
1998:203). In contrast, Late Swift Creek vessels, with folded
rims and often zoned designs, are more similar to those of the
southeastern Georgia coast. In fact, direct evidence of interac-
tion between northeastern Florida and southeastern Georgia
can be found in complicated stamped paddle matches. Since

Table 1. AMS dates from soot on Charcoal-Tempered Pottery.

Site # Beta # C13/C12 Conventional Calibrated 2 1 Sigma and Reference
ratio C14 age (BP) Sigma intercept
(o/oo) Range*

8DU68 182333 -24.2 1940+40 30 BC-AD 30 BC (65) 95 Ashley 2003

8DU68 182332 -24.7 169040 AD 250-430 AD 330 (385) 410 Ashley 2003

8DU96 168177 -24.5 157040 AD 410-580 AD 430 (460, 480, Ashley 2003
520) 540

8DU96 190255 -25.3 1510+40 AD 440-640 AD 530 (560) 610 Wallis 2004
* INTCAL98 Radiocarbon Age Calibration.

2004 VOL. 57(4)



Figure 2. Northeastern Florida Woodland period chronology (

decoration was made by impressing a carved wooden paddle
into the wet clay surface of a vessel before firing, cracks in the
wood or artisan error make it possible to identify the unique
signatures of particular wooden paddles used to decorate
multiple pots. Based on these paddle signatures, Frankie
Snow was able to determine that two vessels from two mounds
on the lower St. Johns River were stamped with identical
paddles as pots found on a site along the Altamaha River,
Georgia (Ashley 1995:28-32).
Thus, while charcoal-tempered wares eventually gave way
to the predominance of sand-tempered wares, Swift Creek
iconography seems to have been applied in varying degrees to
both early and late local pottery traditions. Once considered
anomalies and exotic items (Goggin 1952:49), more recent
excavations that produced large quantities of Swift Creek
Complicated Stamped sherds suggest local production (Ashley
1992:130-133). Typically, however, complicated stamped
sherds tend to make up a low percentage of total assemblages,
up to 15 percent in middens and perhaps slightly higher in
some mounds, as elaborated below (Ashley 1998:216, Russo
1992:115). A conspicuous preponderance of plain sand-
tempered pottery throughout shifts in regional influence
directly relating to the production of Swift Creek minority
wares, suggests that rather than wholesale migration into the
area, the lower St. Johns was connected to outside areas via
more complicated and interactive relationships. In addition,
the continuous use of mounds without any conspicuous change
in burial practice for centuries denotes resolute cultural
persistence. If local groups, through changing patterns of

trade networks and alliances, in fact
maintained a semblance of cultural
continuity for nearly a millennium,
what role did external influences play
in this persistence and its reproduc-
tion through time?

Previous Swift Creek Research:
Mounds, Identity, and Interaction

Before proceeding, it is important
to situate the Mayport Mound within
the Swift Creek world of which it was
a part. Importantly, the term "Swift
Creek", as I use it, does not denote a
recognizable cultural entity with
boundaries, but rather signifies a
pottery style that saw various degrees
of popularity throughout much of the
Southeastern U.S. (Anderson 1998;
Ashley 1998). Rather than represent-
ing a single "culture," the production
and exchange of Swift Creek iconog-
raphy on a regional scale signifies a
d fm A y tradition that brought together hetero-
Sgeneous groups with unique histories.
The character and degree of this so-
cial unity, however, was dependent on
the circumstances attending local social organizations. The
heterogeneity of practice among groups making and using
Swift Creek pottery and the importance of local participation
in its proliferation is especially evident on a regional scale,
thus substantiating a larger picture of the Woodland period as
multifaceted and dynamic (Anderson 1998, Anderson and
Mainfort 2002).
Many Woodland Period traditions that show expansive
spatial distribution, such as platform mound building, served
varying functions across space and through time. Of the many
recorded Woodland-period platform mounds, the occurrence
of structures, mortuary activity, or adjacent villages is variable,
suggesting that different and locally relevant practices oc-
curred at each center (Jefferies 1993; Lindauer and Blitz
1997). Yet the structural similarity of geographically dis-
persed flat-topped mounds, all beginning with initial small
platforms and proceeding with subsequent additions of
planned and uniform extent, begs comparison and points to
unifying traditions (Jefferies 1993). The differences between
mounding centers within the Woodland Southeast, therefore,
cannot be understood without considering local contexts of
practice, situated in a particular place within broader forces of
political economy, community interaction, and ideological
While in northwestern Florida, Middle Woodland mound
sites were frequently in close proximity to villages, many
others in the Southeast, including those associated with Swift
Creek pottery, generally seem to be places of periodic congre-
gation for disparate small groups, since few show evidence of







Figure 3. Distribution of northeastern Florida mounds containing Swift Creek pottery.
Figure 3. Distribution of northeastern Florida mounds containing Swift Creek pottery.

associated habitation (Anderson 1998: 283; Williams and
Harris 1998). Many of these may have served as gateways to
the goods and information networks of Hopewellian or later
Woodland period interaction, perhaps some even functioning
as outposts for channeling Gulf Coast shell into the interior
(Anderson 1998: 278, Caldwell 1964); however, each center's
role was inevitably negotiated through time. With no central
authority or orchestration, and with periodic meetings of
disparate groups at various mounds, mound locations became
arenas for establishing relationships of alliance and for
negotiations of group identity and political clout. Like
Crumley's (1995) "heterarchy" and Johnson's (1982) "sequen-
tial hierarchy", this kind of loosely organized socio-political
organization, perpetually fluid and negotiated through interac-
tion, discourages any individual's attempt to acquire power
toward a vertical hierarchy through inclusive networks of
alliance building (Cobb and Nassaney 1995:214).
Distinctive goals and actions abutted specific but varying
historical circumstances, so that tactics and their conse-
quences, intended and otherwise, produced heterogeneous
archaeological signatures across space. The variability in
social action across space was ultimately a result of each

group's changing position within economic and political
networks inherited and negotiated by each subsequent genera-
tion. Although many Woodland mound sites appear to have
been congregation centers for disparate groups, few studies
have attempted to define the groups involved in these periodic
meetings, although the data and methods for doing so might
be available (e.g., Snow and Stephenson 1998). As a premise
for the analysis presented below, I assert that scrupulous study
of pottery assemblages from mounds holds the potential to
illuminate nuanced patterns of interaction at Woodland period
gathering centers such as the Mayport Mound.
Previous research has suggested the use of paddle stamped
design data to address questions relating to exchange and
interaction networks (Broyles 1968, Snow 1975; Stephenson
and Snow 1998; Stoltman and Snow 1998), as well as to infer
social structure and social identities (Saunders 1986, 1998).
This unique opportunity is provided by the Swift Creek
complicated stamping tradition, in which a wooden paddle was
carved and subsequently impressed into the wet clay surface of
a vessel prior to drying and firing. Due to artisan error or
cracks in the wooden paddle used to stamp each vessel, unique
design flaws or signatures appear on various vessels stamped


2004 VOL. 57(4)


with an identical paddle. Thus, not only can design similari-
ties be noted within and across assemblages, but sometimes the
use of an identical paddle can be discerned among different
pots (Broyles 1968, Snow 1975, Snow and Stephenson 1998).
Saunders (1986, 1998) used the frequency of designs across
two sites in southeastern Georgia to infer an association
between family or village affiliations and particular design
elements. Substantiated by technological attribute data, such
as vessel form and paste composition, the vessels represented
at Kings Bay and Mallard Creek denote the work of two
analytical individuals, likely the result of two distinct learning
communities (lineages) (Saunders 1998). Saunders (1986)
ultimately concluded that the bimodal distribution of some
designs across the Kings Bay site represented separate lin-
eages, whereas the more general similarity of designs between
Kings Bay and Mallard Creek suggested habitation by the
same population (Saunders 1998). Conveniently, evidence of
contemporaneity came from several paddle matches between
sites (Saunders 1998).
Using a more diachronic perspective, Snow and
Stephenson (1998) traced the distribution of paddle matches
between the central-Georgia Hartford site and surrounding
sites. Their analysis showed that the direction of design
contacts with the Hartford site shifted from predominantly
upstream to mostly downstream along drainages through time.
Proceeding from the premise that the higher instances of a
design occur at the site where the paddle owner resided, Snow
and Stephenson (1998:103) postulated that more designs
moved to Hartford from outlying areas rather than vice versa,
substantiating the idea that Hartford was likely a ceremonial
center positioned on an important trail juncture. Furthermore,
Stoltman and Snow (1998) employed petrography to analyze
paddle matches between sites, hoping to discern whether
paddles or pots moved across the landscape. The results of
their analysis revealed a complex pattern of interaction
between sites. Common designs with different vessel paste
between sites suggested that paddles traveled, perhaps with the
potter, while common vessel paste between sites suggested the
movement of pots (Stoltman and Snow 1998:151).
Stoltman and Snow's (1998) regional scale analysis of the
frequency of locally produced versus imported wares is an
effective step in understanding the social processes that define
this interaction. Be it down-the-line exchange of pots,
exchange of paddles, or movement of paddles with female
potters through marriage alliances and patrilocal residence, the
precise social mechanisms responsible for design movements
must be understood from a multiscalar perspective centered on
local contexts of exchange (e.g., Nassaney and Sassaman
1995). To begin to understand the social mechanisms driving
these interactions, we must undertake detailed diachronic
studies of large collections of provenienced vessels from
radiometrically dated sites. Analysis of collections from
accretional mounds containing Early and Late Swift Creek
wares, such as the Mayport Mound, thus holds the potential to
reveal changes in the social implications of interactions
through time among groups in northeastern Florida.

Considering Technological Choices

As with the study of all material culture, a consideration of
style is fundamental to the comparison of pottery within and
between assemblages. Why do artifacts appear similar to one
another, and how can we explain changes and variation
through time and across space? In studying pottery, functional
considerations are certainly germane, since changes in vessel
attributes may reflect changes in subsistence and cooking
practices (e.g., Schiffer and Skibo 1987). In addition to these
purely functional parameters, style corresponds in various
ways to patterns in the organization of production within
groups and patterns of interaction between groups. That is,
the conformity of pottery styles to archaeological "types"
across spatial and temporal dimensions corresponds to
particular structural elements of culture, particularly the way
that pottery production is learned.
My consideration of technological style is essentially
holistic, synthesizing both functional and normative explana-
tions. Purely functional explanations, which posit practical
purposes for cultural traits, and traditional normative views,
which relegate cultural patterns to the constant duplication of
mental templates, both relegate change to external forces
(natural or cultural) (Hawkes 1954, Trigger 1989:264-270,
Watson 1995:683). Instead, a study of technological style
potentially considers all of the structural forces, both "inter-
nal" and "external," that shape uniformity and difference.
Invoking 'choice' as a variable in technological style avoids
deterministic functionalist and nonnative explanations for
material culture, and instead presents the reproduction of
material culture as an ongoing and dynamic process where
change is always a possibility, even without external agents.
Utilizing a dynamic view of structure that ultimately draws on
Bourdieu's (1977, 1984) habitus, where structure is repro-
duced over time through deep-rooted bodily dispositions, this
study assumes that pottery is produced according to the
learned, culturally relevant techniques of particular communi-
ties. At the domestic level of production, specific and distinc-
tive potting techniques are acquired primarily through emula-
tion and repetition. Thus, pottery assemblages may show
regularities at least in part because individual artisans learned
to perform their craft in the same traditional learning environ-
ment that socialized them according to like technical processes
and steps. While the social function of style may have been
significant in the past, especially the highly visible aspects of
design such as Swift Creek complicated stamping, technologi-
cal style is more relevant to the current study in its conserva-
tism and tendency to reveal cultural affiliation (Carr 1995a,
1995b; Lemonnier 1986).
Technological style (Lemonnier 1986), which refers to the
technological choices made during production, conforms to
learned techniques in predictable ways. Last order steps in the
production sequence that are highly visible, such as surface
decoration, are often subject to alteration in situations of
intercultural contact, but initial steps with lower visibility tend
to be more conservative (Gosselain 1998, Hodder 1982:202-
207). First order steps, especially those consisting of routin-




ized actions, such as clay processing and vessel forming
techniques, have tended to be culturally conservative in
ethnographic examples (Gosselain 1994, 1998; Stark et al.
2000). The organization of learning in household level
production is thus a force of stylistic conservatism, so that
particular attributes of a vessel might denote the signatures of
particular communities of practice (Costin 1998, Sassaman
and Rudolphi 2001).
Given the culturally conservative nature of the production
process, with initial manufacturing steps of a vessel likely
signifying the particular learning environment of the artisan,
an examination of technological style among vessels through
time on the lower St. Johns could reveal important changes in
tradition triggered by external influence as well as continuities
in technology that transcend the social function of style. More
specifically, are the influxes of Early and Late Swift Creek
surface treatments in northeastern Florida accompanied by
changes in other vessel attributes? The current study begins to
address this question using an assemblage derived from the
Mayport Mound.

The Mayport Mound

Before its obliteration in the 1970s, the Mayport Mound
(8DU96) was situated approximately one mile south of the St.
Johns River and two miles west of the Atlantic Ocean (Figure
3). Its location marked the easternmost mound in a series of
15 similar, small Woodland period sand monuments along the
lower St. Johns River. All but three were excavated solely by
C.B. Moore (1894, 1895) during the late nineteenth-century,
and each typically contained a suite of burial accoutrements,
including plain and complicated stamped pottery, modified
and unmodified marine shell, pebbles, pebble hammers,
projectile points, celts, sandstone, hematite, chert flake
concentrations, mica, and copper (Ashley 1998:212). The
artifact assemblage from the Mayport Mound appears to
conform to this inventory.
The excavations from which the collection derives were
carried out in June, 1964, under the direction of Rex Wilson
(1965) of the National Park Service. A salvage job from its
inception, Wilson attempted to gain as much information as
possible with limited budget. As such, Wilson (1965:7)
considered the summit and northern flank of the mound to be
"lost to archaeology" due to heavy looting activity, and
preferred to excavate only the southern and eastern portions
that appeared relatively intact (Wilson 1965:7). Even after
Wilson's excavations, however, much of the undisturbed
portion of the mound remained unexcavated (Wilson 1965:10).
Although harshly criticized by Sears (1967), who advocated a
complete excavation that included the heavily damaged
summit, Wilson's decision to end fieldwork was dictated by
the project's patron, U.S. Representative Charles Bennett, who
sought primarily to find evidence of French and native
interaction relating to nearby Fort Caroline, and was thus
disappointed with the results (Rex Wilson, personal communi-
cation, 2003). Bennett's disappointment therefore led to an
early end to Wilson's excavations. At the request of the

landowner, all trenches dug were left open for interested
"amateur archaeologists" whose field specimens have never
been located (Wilson 1965:11). Receiving no further system-
atic excavation, the former Mayport Mound location is now
the site of a small housing development.
At the time of Wilson's excavation, the Mayport Mound
was a low earthen structure that measured approximately 23m
(N-S) by 20 meters (E-W), with a height of about one meter
above the surrounding terrain (Wilson 1965). Based on
excavation results, it appears that mound construction initiated
with meticulous clearing of vegetation and subsequent overlay
of a white sand layer impregnated with charcoal. After this
initial preparation, the mound grew by accretion in both
vertical and horizontal directions. This accretion occurred
through a continuous practice of placing burial remains on the
mound surface and covering with earth, although a few burials
were placed in pits (Wilson 1965). Toward the end of its use-
life, the eastern flank of the mound was partially covered by a
5-8 cm thick shell midden. Termed "the village midden" by
Wilson (1965:5), the dimensions and temporal range of this
deposit are not entirely clear, as it remained mostly
unexcavated. Rather than representing refuse associated with
residential activity, this peripheral midden may represent the
product of repeated ritual gatherings that were accompanied by
feasting. In any case, the midden's direct association with the
mound occurred at a late date, judging by the limited number
of burials that were found in association with shell.
Wilson's excavation, which conservatively included some
30 to 40 percent of the mound area, yielded 46 burials, 6
complete projectile points and one base, a steatite disk frag-
ment, 2 tubular stone beads, one celt, 2 cut mica sheets, a
copper ear spool casing, numerous shell beads and pendants,
a ceramic pipe bowl, 3433 sherds, and 25 restorable vessels.
Units were dug in 6-inch levels or according to natural
stratigraphy, and all soil was passed through a /4-in screen.
Although careful provenience information was recorded for
most artifacts, all of the original data not recorded in Wilson's
(1965) published report have been lost. According to Wilson
(1965), however, neither pottery nor burials associated with
diagnostic artifacts could be satisfactorily grouped according
to vertical provenience. Instead, pottery seemed to be well
mixed both vertically and horizontally, likely a result of
discrete burial episodes across the mound through time. While
some provenience information for individual vessels can be
gleaned from a plan view map that lists burials and associated
materials, these associations do little to suggest a temporal
pattern of accretion, and merely suggest tentative associations
of 2 non-diagnostic vessels with diagnostic ones (see below).
In the end, stratigraphy was too varied to give Wilson (1965)
any impressions of chronological sequence.
As part of his research, Wilson (1965:31) obtained a
radiocarbon date from charcoal taken from a fire pit on the
northeastern fringes of the mound, 2 feet below the surface.
Wilson's (1965) conventional date, 1865 95, yields a
calibrated 2 sigma range of 50 B.C. to A.D. 395 and a cali-
brated intercept of A.D. 1302. In light of the pottery assem-
blage at the Mayport Mound and three other radiometric

2004 VOL. 57(4)



assays from soot, the later end of this calibrated range may
date initial stages of mound construction. The charcoal
sample's vertical provenience, near the subsoil stratum upon
which initial mound building began, suggests that this
radiometric assay may date activity at the site just prior to the
incipient stages of mound construction.
A fourth century date for initial stages of mound construc-
tion maybe substantiated through Wilson's (1965) radiometric
date, while three other assays from soot adhering to vessels
suggests continued use through the seventh century. A
terminal date for mound use, however, can only be inferred
through the pottery types within the assemblage. In the
absence of comprehensible stratigraphy, pottery is useful in
understanding a relative chronology for the mounds' construc-
tion. The earliest diagnostic wares are Deptford (mostly pre-
A.D. 300), followed by charcoal-tempered Early Swift Creek
(A.D. 300-600).
The third diagnostic element, Late Swift Creek, dates to
circa A.D. 500 to as late as A.D. 850. One vessel with fine
spiculate-paste grog, while not typical Colorinda, may have
been produced during late stages of mound use, since grog-
tempered Swift Creek pottery tends to appear late at both
northeastern Florida and southeastern Georgia (Kelvin phase)
sites (Cook 1979). No diagnostic pottery that might post-date
Colorinda is represented in the curated assemblage, although
the excavation produced 18 St. Johns Check Stamped sherds
that probably came from non-mound test units (Wilson
1965:21). Thus, the Mayport Mound was likely used over a
period of up to 600 years, from ca. A.D. 300 to 900.

Methods: Looking for Technological Tradition

The Pottery Sample

An exhaustive search was conducted for the entire exca-
vated assemblage from the Mayport Mound, yielding varying
degrees of success. The Florida Museum of Natural History
(FLMNH) in Gainesville, curates the majority of the collec-
tion, including 756 sherds from a minimum of 38 vessels, four
projectile points, one stone drill, one fragment of a steatite
disk, an unfinished stone pendant, and numerous oyster shell
pendants. In addition, FLMNH holds the unexamined burial
remains of some three dozen individuals taken from the
Mayport Mound and Rex Wilson's color profile and plan view
drawings of the site.
After conducting an extensive inventory of the FLMNH
collection, I learned that the Timucuan Ecological and Historic
Preserve held additional artifacts from the Mayport Mound at
the Fort Caroline National Memorial in Jacksonville. The Fort
Caroline museum owns 85 sherds from a minimum of five
vessels, two projectile points, two cut mica sheets, two tubular
stone beads, one earthenware pipe bowl, one sandstone celt,
179 small disk shell beads, 22 tubular shell beads, and one
copper ear spool casing. All artifacts from the Fort Caroline
museum, excluding the copper ear spool that is currently on
display, were borrowed for study and temporarily housed at

With the combined collections, all 25 restorable vessels
listed and illustrated in Wilson's (1965) report were accounted
for, as well as 147 sherds from 16 additional vessels, some of
which are near complete or almost fully restorable. Wilson
(1965) explicitly states that the 25 restorable vessels illustrated
in the report are not included in his final sherd counts. Of the
3,433 sherds reported by Wilson (1965), then, 3,286 sherds
were missing from the original excavated assemblage. An
exhaustive search for these sherds, as well as any original field
notes, was unsuccessful. Rex Wilson had only a copy of his
published report, and had dutifully delivered all artifacts and
field notes to landowner Harold Clarke in Jacksonville in 1964
(Rex Wilson, personal communication, 2003). There is no
documentation explaining why many artifacts were sent to
FLMNH, but the Fort Caroline museum acquisitions were
certainly selected by Representative Charles Bennett, who
financed the excavation, and sought material for a planned
Fort Caroline museum. Conversations with Dave Dickel,
Collections Manager with the Florida Bureau of Archaeologi-
cal Research, and Richard Vernon, Supervisory Museum
Specialist at NPS, Southeast Archeological Center, confirmed
that neither institution housed the missing sherds.
Although the location of the missing sherds remains a
mystery, some information was recorded about the assemblage
in Wilson's (1965) report, including sherd counts divided by
typological units. Without any further information pertaining
to the missing material, especially its potential division into
vessel lots, only the pottery from FLMNH and the Fort
Caroline museum was used in the present study.

Pottery Analysis

The goal of this investigation was to measure variation
within the assemblage, and to delineate trends of variability
through time. The variability that accompanies the appear-
ance of new and different decoration might be explained by the
influx of new potters, exchange of extralocal vessels, acquisi-
tion of foreign carved paddles, or a combination of all three.
While the highly visible attributes of form and decoration are
often the most prone to change, technological attributes such
as paste composition, firing and finishing techniques, and
cooking practices tend to be more conservative and reflective
of the potter's original learning community. Thus, the present
study attempts to assess continuity and change in the techno-
logical tradition represented by the Mayport Mound assem-
The entire collection of 879 sherds was analyzed for the
purpose of assigning vessel lots, or groups of sherds that were
originally part of the same vessel. Surface decoration, rim and
lip treatment, wall thickness, and general paste composition
(including level of mica inclusions) were all considered in this
initial sorting. Many of the vessels were already organized
into lots by virtue of them having been reassembled with Duco
glue by Rex Wilson in 1964. Since that time, however, the
majority of mends had failed and these vessels were once again
reduced to boxes of sherds. Analysis resulted in the identifica-
tion of 41 separate vessel lots.




Table 2. Vessel lots and sherd counts by type.

Vessel Type Vessels Sherds
Deptford Simple Stamped 3 32
Charcoal-Tempered Plain 4 56
Charcoal-Tempered Swift Creek (Early) 2 67
Mayport Dentate Stamped 4 47
Late Swift Creek 8 242
Weeden Island 1 12
Colorinda 1 15
St. Johns Plain 4 152
Dunns Creek Red 2 24
Sand-Tempered Plain 11 230
Unclassified 1 2
Total 41 879

Fresh breaks were made on sherds from each vessel to use
under a low power microscope for paste analysis3. Using
between 40x and 70x magnification, paste was characterized
according to nine attributes: quartz grain size, grain angular-
ity, degree of sorting, inclusion of non-quartz tempering
material, and frequencies of temper, mica inclusions, clay
lumps, iron concretions, and mafic minerals. These attributes
potentially assess degrees of similarity in both clay sources and
added temper among vessels. Clay and temper selection, both
first order steps in the production sequence with low social
visibility in the finished product, are presumed to correlate
well with the potter's original learning community (Carr
1995b; Gosselain 1998). While variation in clay and temper
within the assemblage could indicate either a multicomponent
technological style or signify the presence of disparate tradi-
tions, continuity could indicate perpetuation of a local potting
In order to evaluate variation in vessel form throughout the
assemblage, vessel reconstruction was undertaken in a major-
ity of cases. This tedious step was deemed necessary because
more than half of the vessels had fully reconstructable walls
from near the base to rim, yielding unusually precise estima-
tions of vessel dimensions. Using metric calipers, a standard
diameter template, and a form gauge, interior orifice diameter,
rim thickness, lip form, rim form, and rim fold depth were
measured for each vessel when possible, and wall profiles were
drawn. While vessel dimensions might fail to reveal the
learning environment of a potter, being highly malleable to
new social conditions (Carr 1995a, 1995b; Gosselain 1998),
correlations of form and decoration could yield better under-
standings of intended vessel uses, perhaps changing through

Surface treatments and finishing tech-
niques, including interior and exterior scrap-
ing, smoothing, slipping, and stamping, were
recorded for each of the vessel lots. Inevita-
bly, some of these techniques were used
during the forming process, such as scraping,
and were subsequently obliterated before
drying and firing, and thus evade classifica-
tion. Using fresh breaks, coring and degree
and orientation of oxidation were assessed in
order to estimate variations in firing and
cooking environments. Using this informa-
tion in concert with assessments of the pres-
ence and location of fire clouding and
sooting on the surface of each vessel allows
determination of which vessels were defini-
tively used over or within fire (Hally 1983,
1986). Using Munsell color charts, the range
of surface color was recorded for all vessels.
Although this information is best used in
retiring experiments to assess the original
firing temperature of a vessel, its variation in
the assemblage might reveal patterns worthy
of future focus (Cordell 1984:40-55; Rice


The 41 vessel lots identified in the analysis represent 12
typological groups: Sand-Tempered Plain (n= 1), Late Swift
Creek Complicated Stamped (n=8), charcoal-tempered plain
(n=4) and complicated stamped (n=2), St. Johns Plain (n=4),
Mayport Dentate Stamped (n=4), Deptford Simple Stamped
(n=3), Dunns Creek Red (n=2), Weeden Island plain (n=l),
and fine grog tempered plain (n=l) (Table 2).
These results differ somewhat from Wilson's (1965)
examination. While analysis of the distribution of sherds by
type is admittedly problematic, with some vessels being
represented by over one hundred sherds and others by only
one, some important differences deserve mention (see Appen-
dix 1). Most important are some key typological additions and
revisions to Wilson's (1965) original report. First, Wilson
(1965) observed no charcoal-tempered pottery but rather
identified sherds with holes as limestone-tempered. Since no
pottery in the analyzed collection was tempered with lime-
stone, it is likely that many of these sherds containing pitted
surfaces were in fact tempered with charcoal that had "ashed
away" during firing and cooking (Ashley 1998:203). Interest-
ingly, both Wilson's (1965) count of limestone-tempered and
the analyzed charcoal-tempered sherds made up about ten
percent of their respective collections (see Appendix 2).
Second, Wilson (1965) failed to recognize any grog-
tempered sherds, because the fine-sized particles of grog were
difficult to identify. The finely crushed spiculate-paste grog
found in v. 3 during the present study differs from that of
Colorinda pottery, which normally is tempered with large

2004 VOL. 57(4)



pieces of crushed St. Johns pottery. In petrographic thin
section, in fact, a red slip was clearly visible on some of the
tiny grog particles, demonstrating that pieces of a Dunns
Creek Red vessel was used as temper (Ann Cordell, personal
communication 2004). Whether Colorinda or not, the crushed
St. Johns grog in this vessel (v. 3) was hardly visible without
magnification, and had in fact been classified as sand-tem-
pered plain in Wilson's report (Wilson 1965:Plate VI, A).
Thus, many grog-tempered sherds may in fact have been
wrongly classified as sand-tempered plain. Fine grog temper-
ing has been recognized at several Late Swift Creek sites in
northeastern Florida and is typical of terminal Late Swift
Creek in southeastern Georgia (Kelvin phase) (Ashley and
Wallis 2004). Radiometric dates for grog-tempered Swift
Creek contexts are lacking, but they may represent a precursor
to the easily recognizable large St. Johns particles typically
found in Colorinda pottery, and may therefore signify a late
eighth or ninth century date of production (Ashley and Wallis
Third, Late Swift Creek Complicated Stamped wares make
up a much larger percentage of the analyzed assemblage than
in Wilson's (1965) sherd count. Although Wilson's (1965)
criteria for distinguishing between types of complicated
stamped pottery are flawed in light of more recent research
(Ashley 1992, 1998, 2003), as a whole, complicated-stamped
wares are notably less well represented in Wilson's (1965)
report (8%) than in the analyzed collection (28%). Perhaps
this discrepancy could be explained by complicated stamped
vessels being primarily interred whole rather than smashed
and spread across the surface of the mound, thus yielding a
greater percentage of whole vessels. In fact, some Late Swift
Creek vessels may represent a distinctive ceremonial ware
intended for burial, in contrast to cooking vessels that were
used extensively before being broken and scattered across the
mound surface (see below). In addition, it is possible that
some complicated stamped wares in Wilson's (1965) analysis
were in fact charcoal-tempered Early Swift Creek, since these
were not designated in his report.
Fourth, Wilson's (1965) "Savannah" types, once consid-
ered as evidence of interaction with Georgia (Sears 1957), are
more recently considered part of a local sand-tempering
tradition (Ashley 2003:74-76), and therefore this terminology
is not employed in the analyzed collection. Fifth, as might be
expected, some of the least represented minority wares were
not present in the analyzed collection at all, including
Carrabelle Incised, Crooked River Complicated Stamped, and
several Deptford varieties. The analyzed collection, then,
makes up an admittedly biased sample, mostly including only
those sherds reassembled into substantial portions of vessels.
As a sub-sample of Wilson's original sample, the results of the
foregoing analysis must be considered as a first step toward
analyses of more collections from other lower St. Johns
Woodland sites.
One last typological point deserves mention. Mayport
Dentate Stamped was defined by Wilson (1965:22-23) as a
new pottery type found at the Mayport Mound, based primarily
on unusual stamping decoration. This unique decoration

consists of parallel, linear grooves applied in rocker stamping
fashion, placed both vertically and horizontally. The Mayport
Dentate Stamped classification has seldom been used by
archaeologists in northeastern Florida, although there are
some exceptions (e.g. Johnson and Basinet 1997). The paste
characteristics and vessel forms associated with the type are
similar to the majority of early vessels within the Mayport
Mound assemblage, suggesting local production. While this
decorative style makes up a small percentage of the analyzed
assemblage (-5%), the utility of designating it a separate type
is presently unclear. The paucity of Mayport Dentate Stamped
sherds recorded at other Woodland period sites in northeastern
Florida may be due to archaeologist's general unfamiliarity
with the type, an actual dearth of similar sherds, or both. This
particular decoration may merely be an idiosyncratic part of
the larger check and simple stamping fluorescence that
occurred locally in northeastern Florida beginning ca. A.D. 1
and continuing at least into the fourth century. In the absence
of comparative data, however, Mayport Dentate Stamped was
used as a type in the present analysis.


Of the nine paste attributes recorded, the size, angularity,
sorting, and frequency ofquartz grain inclusions demonstrated
the most notable consistency across the assemblage. Using
Wentworth's size classification and the Mathew et al. (1991)
percentage inclusion chart, vessels almost always contain a
very fine to fine-sized quartz sand particle that makes up more
than ten percent of the paste. While determining whether this
temper was added to the paste intentionally or was included
within the clays chosen by the potters is difficult to ascertain,
its commonly sub-angular shape suggests the possibility that
it existed in the clays. Typically, large quartz grains within
natural clays are rounded, while smaller ones are predomi-
nantly angular (Cordell 2004, Cordell and Koski 2003, Rice
1987:411). Two vessels of primarily sponge spiculate paste
that also contain very fine sand, lend support to the hypothesis
that this sand was not intentionally added. Nonetheless,
producing "temperless" or naturally tempered wares is a
cultural choice that seems to have been perpetuated throughout
the potting tradition represented at the Mayport Mound.
Thus, when Early and Late Swift Creek surface decorations
gained popularity among groups using the Mayport Mound,
their adoption was not accompanied by the coarse-grained
tempering characteristic at many western Florida and southern
Georgia Swift Creek locales (Ashley and Wallis 2003; Willey
1949:378-383). Wilson (1965:22), in fact, notes that many of
the Deptford and Swift Creek wares fit accepted typological
definitions on all accounts except for their inclusion of temper
much finer than normal.
This acknowledgment is particularly important when
examining typological definitions, for while "Swift Creek" can
be defined in northeastern Florida on the basis of hallmark
complicated stamped designs and paddle matches with other
distant Swift Creek sites, generic simple stamping on so-called
"Deptford" vessels may in fact be part of a local stamping




Figure 4. Absolute frequency of vessel types by quartz grain size.

fluorescence in fine sand tempered wares prior to the adoption
of complicated stamping. Deptford Simple Stamped, which
was presumably made from ca. A.D. 1-300 (Ashley 2003),
largely outnumbered check stamped varieties in Wilson's
(1965:24) analysis, and was the only Deptford type represented
in the present study. Based on paste characteristics, which are
substantially similar to most of the assemblage, these simple
stamped vessels were produced from local clays and conform
to the "temperless" variety of vessel common to the lower St.
Johns during Woodland times.
The presence of typical Deptford pottery would not be
surprising at the Mayport Mound, given its close proximity to
recorded Deptford occupations. Excavations at 8DU5541,
roughly two miles west of the Mayport Mound, revealed a
substantial Deptford component of similar size and density to
the type site near Savannah (Kirkland and Johnson 2000).
While the Deptford Linear Check Stamped and Deptford
Check Stamped sherds at 8DU5541 seemed indistinguishable
from the type site examples, Deptford Simple Stamped
appeared to be rather different, with bolder and larger stamps
than typical for Deptford series, perhaps created using an
individual flat stick rather than a paddle bearing that design
(Kirkland and Johnson 2000). Mayport Mound simple
stamped sherds conform to this description, with larger and
bolder stamps than normally associated with Deptford series.
Its association with the classic Deptford wares of the Southeast
would seem doubtful; however, its persistent presence in small
quantities in most test units at 8DU5541 substantiates its
identification as Deptford Simple Stamped at the Mayport
Charcoal-tempered wares at the Mayport Mound present
another conspicuous local version of a regional pottery type.
Early Swift Creek wares in northeastern Florida are character-
ized by simple round and flat lips as well as hallmark

crenulated and notched forms (Ashley and Wallis 2004).
While wares of this phase are predominantly plain, also
common are complicated stamping designs that are similar to
northwestern Florida and cover the entire vessel body. Fine
sand and charcoal temper are both common in northeastern
Florida versions. In fact, charcoal tempering represents an
enigma of local technologies, not documented on sites beyond
a 30 mile radius of the Mayport Mound. As with most midden
and mound sites, at the Mayport Mound undecorated charcoal
tempered vessels (n=4) outnumbered complicated stamped
ones (n=2), and were predominantly used for cooking to some
degree before their burial, as indicated by soot on exterior
rims. All six vessels contain both very fine to fine sized quartz
sand as well as charcoal that constitutes five to ten percent of
the paste. Charcoal particle size varied from fine flecks to
large pieces over 5mm in diameter. It appears that small chips
of wood were added to vessels in the initial stages of produc-
tion, and were subsequently burned or "ashed away" during
firing and subsequent cooking (Ashley and Wallis 2004).
Since charcoal-tempered vessels were used for cooking in a
majority of cases, an understanding of the technological
performance of this temper could prove insightful. Wood
temper may have improved workability of clay or allowed
lighter weight in the finished product, but with few archaeo-
logical or ethnographic examples of wood or charcoal being
used as temper worldwide, experimental studies may be
required to determine its practical utility (e.g., Skibo et al.
1989, 1997). In light of its chronological context, however,
the possible social reasons for its adoption and abandonment
will be considered briefly in the concluding section of this
With an overwhelming majority of very fine or fine sand
quartz "temper" among vessels (85%), the selection of clay
sources with these inclusions seems to be part of a long-lived


2004 VOL. 57(4)


Figure 5. Unique vessel forms.

pottery tradition The only vessels that contain quartz temper-
ing larger than "fine" grains are Late Swift Creek and sand-
tempered plain (Figure 4). Using Wilson's (1965:12-13)
provenience and burial information, two nondiagnostic sand-
tempered plain vessels (v. 7 and v. 11) were associated with
Late Swift Creek vessels, while two others (v. 16 and v. 20)
were found together in direct association. This limited sample
demonstrates that vessels with coarse-grained tempering most
likely became part of the mound assemblage after Late Swift
Creek decorations became popular, with plain wares of coarse
temper being predominantly contemporary with Late Swift
Creek (Figure 4). In contrast, diagnostic early wares, includ-
ing Deptford and Early Swift Creek, invariably have only fine
tempering. The later coarse-temperedvessels may be imported
exotic vessels from southeastern Georgia, a hypothesis that
seems likely in the context of confirmed paddle matches
between the two areas (Ashley 1995), but the continuation of
fine tempering demonstrated by vessels with mixed sizes of
temper could signify local production. Definitive answers to
these questions can be provided only by petrographic analysis.
In any case, the insertion of southeastern Georgia technology
into northeastern Florida pottery traditions during the Late

Swift Creek phase appears much more conspicuous than the
design influences that seem to have derived from northwestern
Florida. The addition of new temper, which has low social
visibility compared to form and decoration, may signify the
integration of southeastern Georgia people with northeastern
Florida groups in some fashion, perhaps in the form of
marriage alliances (see below) (Carr 1995b).
The remaining paste characteristics displayed no discern-
ible patterning according to type. Frequency of mica inclu-
sions, clay lumps, iron concretions, and mafic material were
distributed randomly across the assemblage. The occurrence
of mafic material and mica, however, which likely denote
natural mineral inclusions within the clay (Rice 1987:34), are
highly correlated and could indicate common clay sources
through time (see Appendix 4). The possibility of continuous
use of particular raw clay sources, however, cannot be substan-
tiated without a larger collection of vessels and clay sampling
in the area.
In contrast to the paste characteristics described above, one
Late Swift Creek vessel (v. 35) represents a complete diver-
gence from the rest of the assemblage, having tempering
ingredients of mostly mafic mineral and crushed bone. Bone




-CDO Or-: O CD 0M N b(0 00 0 0C4 (
i co r- 0) m- Lo IN- o co
o? D CIN N t 04 cD N

orifice diameter
total n=35

Figure 6. Absolute frequency of vessels by orifice diameter.

temper, finely crushed pieces present in enough frequency to
suggest intentional addition rather than accidental inclusion,
has also been recorded in limited quantity at the McArthur
Estates Site (8NA32) on Amelia Island (Handley et al. 2004).
While bone temper has not been identified at sites outside of
northeastern Florida, it is possible that some early identifica-
tions of shell temper within Swift Creek wares in northwestern
Florida may in fact be bone (e.g., Phelps 1969:18). Bone at
this minuscule size is difficult to distinguish from shell
without testing with hydrochloric acid under a micro-
scope-the chemical reaction will cause the acid to bubble on
the surface of shell, while bone produces no reaction. Perhaps
misidentified in the past, bone temper may be part of a
regional Swift Creek technology, rather than a purely idiosyn-
cratic northeastern Florida tradition.

Vessel Form

Myriad vessel forms in a variety of sizes are represented in

. ... ... ... ... .. ... ... ... ... .. ... ... ... ... .. ... ... ... ... ..





Ovaloid Spherical


Spherical Straight
Bowl Conical/Rounded

Figure 7. Vessel form typological paradigm.



2004 VOL. 57(4)


the Mayport Mound assemblage. Several forms,
.... perhaps some designed for special ceremonial
essels purposes, were entirely unique within the assem-
blage. These included a tiny boat-shaped vessel
(v. 6) and a shallow, wide-mouthed bowl (v. 21),
both typed as Dunns Creek Red; a grenade-
shaped incised vessel (v. 36); a triple-orificed
Weeden Island style vessel (v. 19); a St. Johns
gourd-shaped bottle (v. 24); a diminutive com-
pound jar with Swift Creek complicated stamp-
ing (v. 34); and a long collared jar with zoned
co C Swift Creek complicated stamping (v. 5) (Figure
o) 5). While some of these forms, such as the
Co cro Weeden Island compartment vessel (Moore
1895:Plates LXXII and LXXIX), the long col-
lared Swift Creek jar (Willey 1949:374), and the
St. Johns gourd-shaped bottle represent common
forms of their respective types and time periods,
the others may represent unique forms not re-
corded elsewhere.
The remaining vessels have forms more easily identifiable
as utilitarian in function. Indeed, the evidence of soot on
exterior rims of many of these vessels substantiates the
inference that they were used for cooking over fire (see below).
The range of variation in interior orifice diameter within the
assemblage shows abroad distribution of vessel sizes, with the
most common diameter (around 25 cm), representing a
medium-sized cooking pot (Figure 6). In addition, most vessel
forms are distributed in a variety of sizes, so that no particular
shape corresponds to a limited range of orifice diameters.
Since Swift Creek vessel form in general and lower St. Johns
versions in particular have received only perfunctory attention,
I have created form categories based on consistencies within
the Mayport Mound assemblage (see Broyles 1968, Giles 2001,
Kelly and Smith 1975:55, Willey 1949:431) (Figure7).
Accordingly, these vessels can be divided into two main
categories of unrestricted and restricted vessels. These
categories can be further divided into six groups: unrestricted
vessels include 1) straight conical or rounded pots, 2) ellipsoid


Table 3. Number of vessels, relative chronology, and percent sooted of each vessel form.

Form Number of Chronological Placement % sooted
Straight Conical/Rounded pot 3 early and late 100
Ellipsoid bowl 2 undetermined 50
Spherical bowl 2 undetermined 0
Ovaloid pot 5 early and late 80
Spherical jar 4 late 25
Deep Globular pot 4 early 50

bowls, and 3) spherical bowls, whereas restricted forms
include 4) ovaloid pots, 5) spherical jars, and 6) deep globular
pots. Most of the vessel forms, save for the spherical jars and
deep globular pots, appear to have been produced throughout
the Mayport Mound's entire period of use. The spherical jars
and deep globular pots, alternatively, may be temporally
sensitive forms.
The straight conical or rounded pot form is represented by
three vessels of three different series: Deptford, Mayport, and
Late Swift Creek (Figure 8). Thus, this form was likely used
throughout the entire time represented at the Mayport Mound,
and is indeed a common form for Swift Creek (Caldwell
1958:38-40) and the Woodland period in general (Linton
1944). Conoidal or slightly rounded pots with unrestricted
orifices are well suited to cooking in fire, with ample surface
area exposed to heat and no sharp angles that might encourage
thermal gradients and vessel fracture (Linton 1944, Rice
1987:226). Additionally, the unrestricted orifice provides easy
access to vessel contents and facilitates easy pouring (Rice
1987:235). As all three vessels of this form are sooted on the
exterior rim (Table 3), it is likely that this form was used
consistently for wet cooking with fire.
The ellipsoid bowl form is represented by two vessels, one
St. Johns (v. 27) and one sand-tempered (v. 16), neither of
which can be given a precise date (Figure 8). These pots
would have been better suited for cooking over fire than in
fire, in contrast to the conoidal pots discussed above (Linton
1944), but with even easier access to vessel contents. This
vessel form may in fact have been used communally, with food
distributed to many people-the St. Johns vessel (v. 27) was
peppered with mend holes, demonstrating its particular
importance to the people who used it. Notably, however, due
to these holes, this vessel could not have been used to hold or
cook liquid contents. Whether used in mound ceremonies or
elsewhere, these wide orificed vessels likely were used to serve
larger groups.
The spherical bowl, roughly spherical in shape with a
direct rim, is represented by two vessels (v. 9 and v. 10), both
sand-tempered and non-diagnostic. It represents a minority

form, most likely not used for cooking food in any quantity, as
indicated by the absence of soot, small capacity, and shallow
depth (Figure 9). This small vessel form may have been used
as a serving vessel, with the direct rim providing easy access
to contents.
The ovaloid cooking pot, a restricted form seen in both
early and late diagnostic wares, is represented by at least seven
vessels (vessels 3, 7, 8, 11, 18, 26, and 30) (Figure 9, some not
pictured). This slightly restricted form is characterized by
predominantly rounded bases and high perpendicular walls
typically with a slight inflection toward the top of the vessel.
The inflected wall allows for the technological advantage of
slightly less heat loss and spillage but maintains a rim shape
that facilitates easy pouring. This form shares similar features
to Swift Creek vessels found in northwestern Florida and
throughout Georgia (Broyles 1968:Plates 5, 6, 10, 12-14,
Willey 1949:379). Perhaps variations of this vessel form were
used for cooking throughout the Woodland Period, as is
suggested by the limited assemblage from the MayportMound.
All but one of these vessels had soot on the exterior surface
rim (Table 3).
The spherical jar form is represented by four vessels,
including two Late Swift Creek (v. 22 and v. 25) and two
sand-tempered plain (v. 16 and v. 20) (Figure 10). The two
sand-tempered vessels represented, as described above, were
found in direct association with Late Swift Creek vessels and
can therefore be considered contemporaneous. Thus, this
spherical jar form appears late in the assemblage, and is not
found in earlier vessels. There appear to be a variety of sizes
associated with this general form, with presumably only the
largest (v. 16) being used for cooking, as evidenced by soot
around its exterior rim and the conspicuous lack of soot on the
smaller vessels (Table 2). With restricted orifices and
incurvate rims that deter access to contents, the small spherical
jars likely had no practical serving function. As Swift Creek
iconography united locally distinct cultural groups, these late
period vessels may have in fact served primarily ritual func-
tions at mound ceremonies. A systematic comparison of
mound vessel forms to village forms is necessary to understand




Figure 8. Unrestricted cooking forms.

the spherical jar's possible mortuary-specific function. A
notable dearth of form data from habitation sites with substan-
tial amounts of Swift Creek pottery, however, precludes such
a comparison at this time.
Another temporally sensitive vessel form, the deep globular
pot, is more ovaloid in profile, with higher incurvate walls,
probably representing an early form of cooking vessel. This
form is represented by four vessels, including charcoal-
tempered (v. 14), Mayport Dentate Stamped (v. 38), St. Johns
Plain (v. 17), and sand-tempered plain (v. 4), with half of the
exterior rims sooted (Figure 10, Table 2). These vessels tend
to have large capacities and seem well suited to holding and
transporting liquids without spilling, although pouring is ill-
suited to this form and access to contents is limited. A
variation of this form, with rims less incurvate, is represented
by some Georgia examples (Broyles 1968:Plate 11, 16). As
evidenced by soot on some rims, these globular vessels were
used in cooking in some cases.
With such a small sample and without comparative data,
the observed temporal restrictions of spherical jars and deep
globular pots must be considered tentative. Furthermore, the
functional reasons for abandonment or addition of these
particular forms to the potting tradition have not been consid-
ered in any depth. The abandonment of the deep globular pot,
which was certainly used for cooking in some cases, could
reflect changes in subsistence, cooking practices, or both. In

contrast, the addition of Late Swift Creek jars may reflect the
advent of new ceremonial paraphernalia that has no counter-
part in earlier vessels. Future research of small jar forms in
southeastern Georgia, from which influences are presumed to
have come into northeastern Florida, is necessary to begin to
understand the distribution and use of this form.
Rim thickness, measured at three cm below the lip of each
vessel, seemed to vary randomly across the entire assemblage.
Neither vessel type nor form correlated with rim thickness, and
overall vessel size, using orifice diameter as a proxy, only
correlated weakly, with larger vessels generally tending to be
thicker. In general, rims are fairly thin, with a median of 6
mm, reflecting the overall thinness of most pots that made
them good conductors of heat for cooking over fire. Con-
versely, several of the thickest vessels, which lack soot and are
made of forms not conducive to conductive heating (e.g.,
vessels 3, 19, and 25), were thus presumably not used for
cooking, although this correlation is not statistically signifi-
cant using Chi Square Tests (Figure 11)4.

Evidence of Firing and Cooking

In general, most larger vessels display characteristics
common to pots used in direct-heat cooking (Hally 1983). Fire
cores are common on upper portions of vessels, with oxidation

Straight Conical/Rounded
.i- ,


Ellipsoid Bowl

- -,


2004 VOL. 57(4)


Ovaloid Pot

. ........

Spherical Bowl

(: ~

U m -rr~r

Figure 9. Ovaloid pot and spherical bowl forms.

more pronounced toward the exterior of the vessel, and basal
sherds typically fully oxidized. Fire clouding was observed on
the exterior of most vessels, but only those not used for
cooking seemed to retain clouding on basal portions of the
vessel (see Appendix 6). Interior fire clouds are distributed
randomly, not corresponding to evidence of direct cooking on
the exterior, consisting of oxidized bases and sooted rims.
Ultimately, sooting around the rim of many vessels is the best
evidence for their use in direct heat cooking. Interestingly,
neither typology nor vessel form corresponds to level of
exterior sooting. Rather, vessel size, as estimated by vessel
diameter, is the attribute most significantly correlated with
sooting (Figure 11). Thus, vessels of many shapes, tempers,
and decoration were used for cooking over fire, but few were
less than 15 cm in orifice diameter. Accordingly, firing and
cooking practices appear to have changed little throughout the
Mayport Mound assemblage, even in the context of diverse
and changing morphology. The use of relatively large vessels
for cooking throughout time may support the idea that they
were used for communal consumption at the site, although
form data from habitation sites is necessary for comparison.

Summary of Results

To summarize, my analysis showed tremendous continuity
in paste characteristics, particularly in the inclusion of sub-


angular very fine and fine sand. These ubiquitous quartz
grains, with their typically sub-angular shape and occurrence
in wares with other tempers, are likely a natural inclusion in
the local clays selected by potters. The temporal persistence of
fine sand thus demonstrates continuity in the initial step of
local pottery production, which is the selection of clays. The
most notable exceptions to these paste characteristics are
charcoal and grit tempering. Charcoal temper is a distinctly
local phenomenon that occurs almost exclusively in wares on
the lower St. Johns, while grit temper occurs in Swift Creek
wares throughout Georgia and northwestern Florida. Thus,
charcoal temper was an autochthonous creation while the
tradition of adding grit may have come from external sources.
Continuity in cooking vessel forms through time likely
denotes unchanging cooking practices, although the deep
globular pot may prove an exception. The advent of the Late
Swift Creek jar presents the most notable change in vessel
morphology, and may signify a mortuary form that ultimately
derives from southeastern Georgia traditions. Comparison
with vessel forms from village contexts is necessary to desig-
nate domestic cooking forms, although village sites are
conspicuously lacking in the area at this time. Whether
domestic or ceremonial, the cooking vessels represented in the
Mayport Mound demonstrate an unchanging practice of
efficient direct heat cooking with fire.




Spherical Jar

t V

Deep Globular Pot


w -

Figure 10. Spherical jars and deep globular pots, both restricted vessel forms.


As a whole, the attributes recorded for vessels across the
assemblage suggest that participants in Mayport Mound
ceremony perpetuated a pottery tradition with a distinctive
suite of technological properties throughout much of the
Woodland period. The ubiquitous fine sand-temper found
throughout the assemblage likely signifies a deep historical
tradition of exploiting particular clay sources rather than
intentionally adding temper. Because temper sizes across
Deptford and Swift Creek assemblages found outside the lower
St. Johns are variable (Milanich 1994:111, 146; Willey
1949:179) and southeastern Georgia Swift Creek assemblages
typically have "grit" tempering (Ashley and Wallis 2003), this
preponderance of fine sand, whatever the source or degree of
intentionality, suggests a persistent pottery tradition with little
changes in desired paste properties.
In contrast, the addition of charcoal tempering to otherwise
naturally tempered clays represents a distinctive 300-year
period of local production when other technological or
aesthetic properties may have been desired. The apparent
geographical circumscription of this technology, limited
almost exclusively to the lower St. Johns River, demonstrates
a degree of social insularity even in the context of design and
decoration influences from northwestern Florida. Connected

to Swift Creek pottery-making groups in northwestern Florida
via goods and information networks, these distinctive lower St.
Johns wares may have served as identity markers, especially
during interactions. Since intensive intercultural contact often
leads to the development of more conspicuous differences in
material culture between groups (e.g. Hodder 1982), Early
Swift Creek pottery with charcoal temper may have served the
dual social functions of both external affiliation (through
common design) and local group solidarity (through unique
This tenuous relationship with northwestern Florida groups
that allowed for both external connections and internal
cohesion declined around AD 600. It is probably no coinci-
dence that the rise in popularity of classic Late Swift Creek
pottery demonstrating direct connections with southeastern
Georgia sites through paddle matches, occurred at the same
time that charcoal-tempered pottery production ceased. Not
only did the direction of emphasis in regional social ties shift
during the sixth and seventh centuries, but also the character
of these relationships appears distinctly different. Coarse
quartz tempering, which only appears in limited amounts with
the advent of Late Swift Creek designs and rim forms, repre-
sents a penetration of the long-standing endogenous tradition
of fine sand tempering. As increasingly integrative social
relationships developed with southeastern Georgia groups,


2004 VOL. 57(4)


perhaps conspicuous ethnic signifiers, such as charcoal
tempered pottery, lost popularity.
Meanwhile, heightened levels of decorative and design
similarities with southeastern Georgia were encouraged
through assimilative relationships. If fine sand represents a
natural inclusion within lower St. Johns clays, then those
vessels with mixed temper may represent vessels produced by
kin-related southeastern Georgia potters using local clays and
their traditional temper. Based on ethnographic evidence from
the Southeast (Hudson 1976:264) and worldwide cross-cultural
regularities that show that women are almost exclusively the
makers of pottery in household level production (Arnold
1985:101-103: Skibo and Schiffer 1995), these were most
likely women relocated through marriage alliances and
patrilocal postmarital residence. Alternatively, the very few
vessels (n=2) containing only coarse temper could be
extralocal vessels that were trade-wares or gifts, or simply
accompanied women's changes in residence.
A more intricate understanding of the extralocal social
relationships pertinent to lower St. Johns groups becomes
increasingly more feasible with the study of paddle matches
between sites. For instance, Frankie Snow identified nearly
identical paddle impressions on two vessels found 100 km
apart, one from the Mayport Mound and the other from the
banks of Lewis Creek, a tributary of the Altamaha River
(Ashley 1995). The impressions on the Mayport Mound vessel
(v.18) were crisp enough to observe the grain in the wooden
paddle, while the Lewis Creek vessel exhibited stampings from
the same paddle, now older, more worn, and developing a
distinctive crack in the wood (Frankie Snow, personal commu-
nication, 2004). Based on these attributes, we can be certain
that the Lewis Creek vessel was crafted and stamped sometime
after the Mayport Mound vessel was created. Furthermore,
petrographic analysis suggests that the two vessels were likely
formed from the same clay source, confirming that in this
instance the vessels) moved across the landscape (Ann
Cordell, personal communication, 2004). Even more conspic-
uous is the morphological similarity of the two vessels, both
with broad rim folds and similar orifice diameters that appear
to be the distinctive product of a particular crafting community
or "analytical individual", whose vessels moved across the
landscape. In this way, paddle stamp matching studies have
the potential to add specificity to our understanding of the
social mechanisms that define Swift Creek interaction,
especially in combination with techno-functional analyses of
entire assemblages.
In this case, several cooking vessel forms were produced
and used throughout the entire period of Mayport Mound use.
Some of these forms, particularly the unrestricted conical and
rounded pots, were standard cooking ware for many communi-
ties across the Southeast (Linton 1944). Even with the
supposed temporal sensitivity of one form, the deep globular
pot. which occurs early, relative sizes of cooking vessels
appear constant throughout the assemblage regardless of form,
which may signify consistent practices of food distribution to
similar-sized groups in either domestic or ritual contexts, or
perhaps both. Certainly, vessels that demonstrate a long use-

life, such as the extensively mended bowl (v. 27), were not
created for immediate burial. Analyses of assemblages from
domestic contexts are needed in order to distinguish possible
ceremonial from domestic forms, but few substantial Swift
Creek pottery-bearing domestic sites have been recorded (see
Smith and Handley 2002 for exception). Indeed, one distinc-
tive characteristic of Woodland mound sites on the lower St.
Johns is the conspicuous lack of substantial midden deposits
that suggest more than a brief seasonal occupation, in direct
contrast to some year-round inhabited Swift Creek mound and
village complexes in northwestern Florida (Bense 1998:255-
262; Willey 1949:368-369).
Although not exclusively a mortuary ware in northeastern
Florida, Swift Creek pottery was nonetheless implicated in
ritual, as were all Woodland period types. Notably, macro-
scopic paste characteristics and generalized categories of
surface decoration, from which most Southeastern U.S.
typologies derive, seem to have had little bearing on a vessel's
ritual use, since every type found in the Mayport Mound has
been documented in Woodland period middens, usually in
similar proportions (Sears 1957). Instead, vessel form may be
a defining characteristic for some mortuary wares, although
comparative analysis awaits the review of material from many
ephemeral middens and the discovery of more habitation sites.
During late times, when vessels were being carried across
the landscape of the southern coastal plain, increasing social
integration between groups caused distinct changes in mortu-
ary ritual at the Mayport Mound. While cooking and serving
practices remained somewhat constant, the late introduction of
spherical jars, perhaps buried in whole form more often than
larger cooking vessels, represents the advent of new funerary
accoutrements whose impetus for adoption may have come
from Georgia. More than just a stochastic shift in material
culture, the adoption of these new items was likely implicated
in constructing distinctly different social memories and group
identities. At the very least, the shift from scattering largely
utilitarian wares across the surface of the mound to burying
whole vessels deliberately created for mortuary events delin-
eated a new form of ritual space that was more clearly
circumscribed from mundane activities. In the absence of
comparative data, future analysis of other collections is critical
to test this hypothesis.
Thus, as interaction networks of various forms shifted
geographical emphasis over time, from west to north, so too
did the nature of this interaction. The local distinctiveness of
early wares suggests less integrative forms of interaction with
western Florida, most likely trade and information networks,
while the slightly more conspicuous southeastern Georgia
influences in the late vessels suggest a closer assimilation of
social groups, perhaps through marriage alliances or simply
more regular interaction. The occurrence of Weeden Island
vessel forms on many lower St. Johns sites, however, generally
demonstrates that exchange networks with northwestern
Florida were not entirely abandoned during the Late Woodland
(Ashley 1998:219), and is confirmed by the Mayport Mound's
representative vessel (v. 19). The more socially penetrating
relationship between lower St. Johns and southeastern Georgia




groups deserves closer inspection using more assemblages
from both areas. If fine sand tempering represents a natural
inclusion within lower St. Johns clays, then those vessels with
mixed temper may represent vessels produced by kin-related
southeastern Georgia potters using local clays and their
traditional temper, most likely women relocated through
marriage alliances and patrilocal postmarital residence.
Alternatively, the very few vessels (n=2) containing only
coarse temper could be extralocal wares that were gifts or
trade-wares or simply accompanied these changes in residence.
Using petrographic analysis of vessels from both areas with
paddle matches, and through technofunctional analysis of
greater numbers of assemblages, these tentative trends of
social interaction may be brought to light.


'INTCAL98 Radiocarbon Age Calibration.
2Beta Analytic #: Beta-GX0315.
3 Four vessels, the property of Timucuan Ecological and Historic
Preserve at Fort Caroline, were not broken, at their request.
Although handicapped by this restriction, paste was characterized
based on older breaks. Firing characteristics could not be determined
for these vessels.
4 Using diameter groups compared to presence/absence of soot, N=35,
P=.07, x 2=22.3,df- 14.


I am indebted to Ann Cordell for her willingness to share her
knowledge of pottery analysis and restoration, and for her constant
enthusiasm and encouragement in this project. Thanks also go to
Keith Ashley, who kindled my interest in lower St. Johns Swift
Creek, divulged all of his knowledge on the subject, notified me of
the Fort Caroline collection, and read a draft of this article. I would
also like to thank Ken Sassaman, Jerald Milanich, and Jane
Southworth for their comments and criticisms on an earlier draft, and
Michelle LeFebvre for her helpful comments on important sections.
Rex Wilson was eager to share his recollections of the project, for
which I am grateful. Thanks to Frankie Snow, who shared informa-
tion about paddle matches with the Mayport Mound, as well as
relevant sherds and paddle designs. Funds for the AMS date came
from the 2004 John Griffin Student Grant, for which the Florida
Archaeological Council is thanked. Lastly, I express gratitude to
Scott Mitchell (FLMNH) and Anne Lewellen (Fort Caroline
museum) for their help in giving access to the respective collections
in their charge.

References Cited

Anderson, David G.
1998 Swift Creek in Regional Perspective. In A World Engraved:
Archaeology of the Swift Creek Culture, ed. by Mark
Williams and Daniel Elliot, pp. 274-300. University of
Alabama Press, Tuscaloosa.

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

1995 The Dent Mound: A Coastal Woodland Period Burial

Mound Near the Mouth of the St. Johns River, Florida.
The Florida Anthropologist 48:13-34.

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

2003 Interaction, Population Movement, andPolitical Economy:
The Changing Social Landscape of Northeastern Florida
(A.D. 900-1500). Ph.D. dissertation, Department of
Anthropology, University of Florida. Gainesville.

Ashley, Keith H., and Neill J. Wallis
2004 Northeastern Florida Swift Creek: Overview and Future
Research Directions. Ms. submitted to The Florida Anthro-

Bourdieu, Pierre
1977 Outline of a Theory of Practice. Cambndge University
Press, Cambridge.

1984 Distinction: a social critique of the judgement of taste.
Harvard University Press, Cambridge.

Broyles, Bettye J.
1968 Reconstructed Designs from Swift Creek Complicated
Stamped Sherds. Southeastern A.rchaeological Conference
Bulletin 8:49-74 Morgantown, West Virginia.

Bullen, Ripley P., and John W. Griffin
1952 Archaeological Survey of Amelia Island, Florida. The
Florida Anthropologist 5:37-64

Caldwell, Joseph R.
1958 Trend and Tradition in the Prelustorn of the Eastern United
States. American Anthropological Association Memoir 88.
Springfield, Illinois.

1964 Interaction Spheres in Prehistory. In Hopewellian Studies,
edited by Joseph Caldwell and Robert L. Hall, pp. 133-143.
Illinois State Museum Scientific Papers 12, Springfield.

Carr, Christopher
1995a Building a Unified Middle-Range Theory of Artifact
Design: Historical Perspectives and Tactics. In Style,
Society, and Person: A.rchaeological and Ethnological
Perspectives., ed by Christopher Carr and Jill Nietzel, pp.
151-170. Plenum Press, New York.

1995b A Unified Middle-Range Theory of Artifact Design. In
Style, Society, and Person: Archaeological and Ethnologi-
cal Perspectives., ed. by Christopher Carr and Jill Nietzel
pp. 171-258. Plenum Press, New York.

Clay, R. Berle
1998 The Essential Features of Adena Ritual and Their Implica-
tions. Southeastern Archaeology 171-21.

Cobb, Charles R., and Michael S. Nassaney
1995 Interaction and Integration in the Late Woodland South-
east. In Native American Interactions: IMultiscalar


2004 VOL. 57(4)


Analyses and Interpretations in the Eastern Woodlands,
edited by M. S. Nassaney and K. E. Sassaman, pp. 205-
226. University of Tennessee Press, Knoxville.

Cook, Fred C.
1979 Kelvin: A Late Woodland Phase on the Southern Georgia
Coast. Early Georgia 7:65-86.

Cordell, Ann S.
1984 Ceramic technology at a Weeden Island period archaeologi-
cal site in north Florida. Ceramic Notes no. 2. Florida
State Museum, Gainesville.

2004 Paste variability and possible manufacturing origins of late
archaic fiber-tempered pottery from selected sites in
peninsular Florida. Draft chapter in Early Pottery: Tech-
nology, Style, and Interaction in the Lower Southeast,
edited by Rebecca Saunders and Christopher Hayes. The
University of Alabama Press, Tuscaloosa. In press.

Cordell, Ann S., and Steven H. Koski
2003 Analysis of a spiculate clay from Lake Monroe, Volusia
County, Florida. The Florida Anthropologist 56:113-124.

Costin, Cathy Lynne
1998 Introduction: Craft and Social Identity. In Craft and Social
Identity, ed. by Cathy Lynne Costin and Rita P. Wright, pp.
3-16. American Anthropological Association, Arlington,

Crumley, Carole L.
1995 Heterarchy and the Analysis of Complex Societies. In
Heterarchy and the Analysis of Complex Societies, ed. by
R. M. Ehrenreich, C.L. Crumley, and J.E. Levy, pp. 1-6.
Archaeological Papers, No. 6. American Anthropological
Association, Arlington, VA.

Goggin, John M.
1952 Space and Time Perspective in Northern St. Johns Archae-
ology, Florida. Yale University Publications in Anthropol-
ogy 47, New Haven.

Gosselain, Olivier P.
1994 Skimming Through Potter's Agendas: An
Ethnoarchaeological Study of Clay Selection Strategies in
Cameroon. In Society, Culture, and Technology in Africa,
edited by S. Terry Childs, pp. 99-107. MASCA Research
Papers in Science and Archaeology, Supplement to Volume
11. University of Pennsylvania Museum of Archaeology
and Anthropology, Philadelphia.

1998 Social and Technical Identity in a Clay Crystal Ball. In The
Archaeology of Social Boundaries, ed. by Mariam T.
Stark, pp. 78-106. Smithsonian Institution Press, Washing-
ton D.C.

Hally, David J.
1983 Use Alteration of Pottery Surfaces: An Important Source of
Evidence for the Identification of Vessel Function. North
American Archaeologist 4:3-26.

1986 The Identification of Vessel Function: A Case Study from
Northwest Georgia. American Antiquity 51:267-295.

Hawkes, Christopher
1954 Archaeological Theory and Method: Some Suggestions
from the Old World. American Anthropologist (56):155-

Jefferies, Richard W.
1993 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.

Johnson, Gregory A.
1982 Organizational Structure and Scalar Stress. In Theory and
Explanation in Archaeology: the Southhampton Confer-
ence, ed. by Colin Renfrew, Mark Rowlands, and B.
Seagraves, pp. 389-421. Academic Press, New York.

Johnson, Robert E., and B. Alan Basinet
1997 Phase II Archaeological Investigation of Sites 8DU5544
and 8DU5545, Queen's Harbour Yacht and Country Club,
Duval County, Florida. Report on file, Division of Histori-
cal Resources, Tallahassee.

Kirkland, S. Dwight, and Robert E. Johnson
2000 Archaeological Data Recovery at Greenfield Site No. 5,
8DU5541. Report on file, Division of Historical Resources,

Lindauer, Owen, and John Blitz
1997 Higher Ground: The Archaeology of North American
Platform Mounds. Journal of Archaeological Research

Linton, Ralph
1944 North American Cooking Pots. American Antiquity 9:369-

Mathews, A. J., Woods, A. J. Oliver, C.
1991 Spots Before Your Eyes: New Comparison Charts for
Visual Percentage Estimation in Archaeological Material
in Middleton, A. P., and Freestone, I. C. In Recent Devel-
opments in Ceramic Petrology, British Museum Occasional
Paper 81, pp. 211-263. British Museum, London.

Milanich, Jerald T.
1994 Archaeology of Precolumbian Florida. University of
Florida Press, Gainesville.

Milanich, Jerald T., and Charles Fairbanks
1980 Florida Archaeology. Academic Press, New York.

Moore, Clarence B.
1894 Certain Sand Mounds of the St. Johns River, Florida, Parts
I and II. Journal of the Academy of Natural Sciences of
Philadelphia, Second Series 10:129-246.

1895 Certain Sand Mounds ofDuval County, Florida. Journal of
the Academy of Natural Sciences ofPhiladelphia, Second
Series (10):449-502.

Nassaney, Michael, and Kenneth Sassaman.
1995 Introduction: Understanding Native American Interactions.
In Native American Interactions: MultiscalarAnalyses and




Interpretations in the Eastern Woodlands, ed. by Michael
Nassaney and Kenneth Sassaman, pp. xix-xxxviii. Univer-
sity of Tennessee Press, Knoxville.

Phelps, David S.
1969 Swift Creek and Santa Rosa in Northwest Florida. Note-
book 1(6-9):14-24. South Carolina Institute of Archeology
and Anthropology, University of South Carolina, Columbia.

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

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

Rye, Owen S.
1981 Pottery Technology, Principles, and Reconstruction.
Taraxacum, Washington, D.C.

Sassaman, Kenneth E., and Wictoria Rudolphi
2001 Communities of Practice i the Early Ceramic Traditions
of the American Southeast. Journal of Anthropological
Research 57:407-425.

Saunders, Rebecca
1986 Attribute Variability in Late Swift Creek Phase Ceramics
from King's Bay, Georgia. Unpublished M.A. thesis,
Department of Anthropology, University of Florida,

1998 Swift Creek Phase Design Assemblages from Two Sites on
the Georgia Coast. In A World Engraved: Archaeology of
the Swift Creek Culture, edited by Mark Williams and
Daniel Elliot, pp. 154-180. University of Alabama Press,

Schiffer, Michael B., and James M. Skibo
1987 Theory and Experiment in the Study of Technological
Change. Current Anthropology 28:595-622.

Sears, William H.
1956 Excavations on Lower St. Johns River, Florida. Contribu-
tions of the Florida State Museum 2, Gainesville.

1967 Review: Excavations at the Mayport Mound, Florida.
American Antiquity 32:120-21.

Seeman, Mark F.
1994 When Words are Not Enough: Hopewell Interregionalism
and the Use of Material Symbols at the GE Mound. In
Native American Interactions: Multiscalar Analyses and
Interpretations in the Eastern Woodlands, edited by M.S.
Nassaney and K.E. Sassaman, pp. 122-143. Univeristy of
Tennessee Press, Knoxville.

Skibo, James M., Michael B. Schiffer, Kenneth C. Reid
1989 Organic-Tempered Pottery: An Experimental Study.
American Antiquity 54:122-146.

Skibo, James M., Tamara C. Butts, and Michael B. Schiffer

1997 Ceramic Surface Treatment and Abrasion Resistance An
Experimental Study. Journal of Archaeological Science

Smith, Greg C., and Brent Handley
2002 Addendum to. Archaeological Data Recovery and Mitiga-
tion at DU5544/5545. Queen s Harbour Yacht and C ountr
Club, Duval County, FL. Report on file, Division of
Historical Resources. Tallahassee.

Snow, Frankie
1975 Swift Creek Designs and Distnbutions. A South Georgia
Study. Early Georgia 3:38-59.

Snow, Frankie, and Keith Stephenson
1998 Swift Creek Designs: A Tool for Monitoring Interaction. In
A World Engraved Archaeology of the Sw!ft Creek Culture,
ed. by M. Williams and D. T. Elliott. pp 99-111. Univer-
sity of Alabama Press, Tuscaloosa.

Stark, Miriam T., Ronald L. Bishop, and EIlizabeth Miksa
2000 Ceramic Technology and Social Boundanes Cultural
Practices in Kalinga Clay Selection and Use. Journal of
Archaeological method and Theori 7:295-331.

Stephenson, Keith
2002 Regional Evaluation ofDeptford and SNwift Creek Chronol-
ogies from Radiometric and AMS Dates Paper presented
at the 5 5 Annual Meeting of the Southeastern Archaeolog-
ical Conterence, Biloxi, MS

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

Stoltman, James B., and Frankie Snow
1998 Cultural Interaction Within Swift Creek Society: People.
Pots, and Paddles. In A World Engraved: Archaeology of
the Swift Creek Culture, edited bv M. Williams and D. T
Elliott, pp. 130-153 Unmversitv of Alabama Press.

Trigger, Bruce G.
1989 A History ofArchaeological Thought Cambridge Univer-
sity Press, Cambndge.

Watson, Patty Jo
1995 Archaeology, Anthropology, and the Culture Concept
American Anthropologist 97:683-694.

Williams, Mark, and Freer Harris
1998 Shrines of the Prehistoric South: Patterning in Middle
Woodland Mound Distribution. In A World Engraved:
Archaeology of the Swift Creek Culture, edited by Mark
Williams and Daniel Elliot, pp. 36-47. University of
Alabama Press, Tuscaloosa.

Wilson, Rex L.
1965 Excavations at the Aavport Aound, Florida. Contributions
of the Florida State Museum 13. Gainesville


2004 VOL. 57(4)


Appendix 1. Vessel lot numbers, classification, and number of sherds per vessel.

Index # Vessel # Classification # sherds
103006 8 charcoal temp. plain 22
103012 14 charcoal temp. plain 32
103025 41 charcoal temp. plain 1
103025 37 charcoal temp. plain 1
105196 29 charcoal temp./stamped 61
105196 30 charcoal temp./stamped 6
103002 3 colorinda 15
103010 12 Deptford 30
103025 23 Deptford 1
103025 33 Deptford 1
103004 6 Dunns Creek Red 3
103018 21 Dunns Creek Red 21
103011 13 Mayport dentate stamped 33
103024 28 Mayport dentate stamped 10
105196 38 Mayport dentate stamped 2
105196 40 Mayport dentate stamped 2
105196 32 sand temp. plain 6
FOCA 428 2 sand temp. plain 18
103009 11 sand temp. plain 15
103013 15 sand temp. plain 36
103007 9 sand temp. plain 21
103008 10 sand temp. plain 11
103017 20 sand temp. plain 59
103003 4 sand temp. plain 27
103005 7 sand temp. plain 17
103014 16 sand temp. plain 6
105196 39 sand temp. plain 14
103023 27 St. Johns 70
103015 17 St. Johns 49
103020 24 St. Johns 21
FOCA 429 31 St. Johns 12
103021 25 Sw. Cr. Comp Stamp 20
103022 26 Sw. Cr. Comp Stamp 22
FOCA 435, 436 5 Sw. Cr. Comp Stamp 40
103025 34 Sw. Cr. Comp Stamp 2

103001 1 Sw. Cr. Comp Stamp 111



Index # Vessel # Classification # sherds
103016, FOCA 430, 431 18 Sw. Cr. Comp Stamp 14
103019 22 Sw. Cr. Comp Stamp 29
103025 35 Sw. Cr. Comp Stamp 4
105021 36 unique incised 2
FOCA 433 19 Weeden Island 12
lotal 879

Appendix 2. Mayport Mound sherd and vessel counts by type.

Type Original Analyzed

# % # ves- % ves- # sherds % sherds
sherds sherds sels sels
St. Johns Plain 633 18.49 4 9.76 152 17.29
Dunns Creek Red 518 15.13 2 4.88 24 2.73
St. Johns Check Stamped 18 0.52 0 0 0 0
St. Johns Simple Stamped 3 0.08 0 0 0 0
St. Johns Dentate Stamped 2 0.06 0 0 0 0
Sand Tempered Check Stamp 10 0.29 0 0 0 0
Swift Creek Complicated Stamped 278 8.12 8 19.51 242 27.53
Crooked River Complicated Stamped 2 0.06 0 0 0 0
Carabelle Incised 3 0.09 0 0 0 0
Mayport Dentate Stamped 152 4.44 4 9.76 47 5.35
"Hare Hammock Surface-Indented" 5 0.15 0 0 0 0
Deptford Simple Stamped 14 0.41 3 7.32 32 3.61
Deptford Linear Check Stamped 3 0.09 0 0 0 0
Deptford Bold Check Stamped 16 0.47 0 0 0 0
Deptford Check Stamped 9 0.26 0 0 0 0
Charcoal-Tempered Plain ("Pasco" and 330 9.64 4 9.76 56 6.37
"limestone" tempered)

Charcoal-Tempered Complicated 0 0 2 4.88 67 7.62
Plain Sand-Tempered 1424 41.60 11 26.83 230 26.17
Weeden Island 0 0 1 2.44 12 1.37
UID stamped 3 0.09 0 0 0 0
Colorinda 0 0 1 2.44 15 1.75

unique incised 0 1 2.44 2 0.23

total 3423 100 41 100 879 100.05
NOTE: Wilson's "Early," "Late," and "sandy paste" variant were all classified as "Swift Creek." Savannah Burnished Plain is included in "Plain
Sand Tempered" and "Savannah Check Stamped" are labeled "Sand Tempered Check Stamped." "Savannah Dentate Stamped" is included in
"Mayport Dentate Stamped."


2004 VOL. 57(4)


Appendix 3a. Quartz grain size, angularity, and other temper frequency.

Vessel Classification Grain Grain Angularity % other temper other %
# Size
12 Deptford v.f. sub-angular 20
23 Deptford f sub-angular 20
33 Deptford f. sub- angular 20
8 Charc. P1. f. sub- angular 10 charcoal 10
14 Charc. P1. f sub-angular 10 charcoal 5
37 Charc. P1. f. sub-rounded 20 charcoal 5

41 Charc. P1. v.f.-f sub- angular 20 charcoal 10
29 Charc. Sw. Cr. v.f.-f sub- angular 15 charcoal 10
30 Charc. Sw. Cr. f angular 20 charcoal 10
13 Mayport v.f. sub-rounded 10
28 Mayport f angular 20
38 Mayport f sub- angular 20
40 Mayport f sub- angular 20
18 Sw. Cr. f -m. mixed 15
22 Sw. Cr. m. sub-rounded 15 course sand 5
25 Sw. Cr. f. sub-rounded 20
26 Sw. Cr. f. sub- angular 20
34 Sw. Cr. f. sub-angular 20
35 Sw. Cr. v.f sub-angular >5 mafic, bone 10, 5
1 Sw. Cr. f. sub-angular 20
5 Sw. Cr. v.f.-f. sub-angular 20
36 unclass v.f.-f. sub-angular 15
2 Sand v.f.-f sub-angular 20
4 Sand f sub-angular 10 course sand 5
7 Sand f.-m. sub-angular 20 course sand 5
9 Sand v.f sub-rounded 20
10 Sand f. sub-rounded 5
11 Sand f. sub-rounded 10 coarse sand 5
15 Sand v.f.-f sub-angular 20
16 Sand c. sub-rounded 5
20 Sand f. sub-angular 10 coarse sand 5
32 Sand v.f.-f. sub-angular 20
39 Sand f. Angular 20
6 D.C.R. 20 sponge spic.
21 D.C.R. v.f.-f sub-angular 5 sponge spic.




Vessel Classification Grain Grain Angularity % other temper other %
# Size
17 St. Johns 20 sponge spic.

27 St. Johns v.f. sub-rounded 5 sponge spic.

24 St. Johns sponge spic.
31 St. Johns sponge spic.
19 W. I. f.-m. sub-angular 20

3 Colorinda f. sub-angular 5 sherd 5

Appendix 3b. Relative abundance of paste inclusions.

Vessel Classification Sorting Mafic Clay Iron Concre- Mica
# Lumps tions
12 Deptford v.g. no no no no
23 Deptford v.g. no no occ. no
33 Deptford v.g. no no occ. no

8 Charc. P1. v.g. no no occ. no

14 Charc. P1. v.g. no no occ. occ
37 Charc. P1. v.g. no no occ. no
41 Charc. P1. g. no no no no
29 Charc. Sw. Cr. no occ. no
30 Charc. Sw. Cr. v.g. no no no occ
13 Mayport v.g. no no no occ

28 Mayport g. occ no freq occ
38 Mayport v.g. no no no no
40 Mayport v.g. no no no no

18 Sw. Cr. f. occ no no occ
22 Sw. Cr. f no no no no

25 Sw. Cr. v.g. occ. no no no
26 Sw. Cr. v.g. no no no no
34 Sw. Cr. v.g. no no no no
35 Sw. Cr. v.g. occ. occ. occ. no
1 Sw. Cr. v.g. occ no occ occ

5 Sw. Cr. g. no no no occ
36 unclass g. no no occ. no

2 Sand g. freq. no occ. occ.

4 Sand p. no no no no

7 Sand f no no occ. no

9 Sand v.g. no no occ. no

10 Sand v.g. no no no occ

2004 VOL. 57(4)

Vessel Classification Sorting Mafic Clay Iron Concre- Mica
# Lumps tions
11 Sand p. no no no occ
15 Sand g. no no no no
16 Sand g. no no occ. no
20 Sand p. occ. occ. occ. occ.
32 Sand g. occ. no occ. no
39 Sand v.g. occ. no occ. occ.
6 D.C.R. no no no
21 D.C.R. g. no occ. no no
17 St. Johns no no occ. no
27 St. Johns v.g. no no occ. no
24 St. Johns no no no no
31 St. Johns no no no no
19 W. I. g. no no occ. occ.
3 Colorinda v.g. occ. occ. occ. occ.

Appendix 4. Chi Square Test of Significance for common occurrence of mica and mafic minerals.



mafic 7
no mafic 8
Probability = .0064
Degrees of freedom = 1

Appendix 5. Vessel form and finishing characteristics.

Vessel Classification interior rim rim form vessel form fold interior exterior
# orifice thick depth finish finish
(cm) (mm)
12 Deptford 14.3 6.25 plain straight conical sc/sm stamped
23 Deptford 18 5.75 scraped smoothed
33 Deptford -smoothed smoothed
8 Charc. P1. 24.6 7.2 crenulated ovaloid pot sc/sm smoothed
14 Charc. P1. 22 5.9 plain deep globular sc/sm smoothed
37 Charc. P1. -scraped scraped
41 Charc. P1. smoothed smoothed
29 Charc. Sw. Cr. 26 5.4 crenulated straight conical scraped stamped
30 Charc. Sw. Cr. 16 4.1 folded ovaloid pot 3.75 smoothed smoothed
13 Mayport 18 6.5 plain straight conical sc/sm stamp
28 Mayport 16 5.4 folded ovaloid pot 7 smoothed stamp
38 Mayport 20 4.45 deep globular smoothed stamp



Vessel Classification interior rim rim form vessel form fold interior exterior
# orifice thick depth finish finish
(cm) (mm)
40 Mayport smoothed stamp
I Sw. Cr. 24 7 folded straight conical 14.25 sc/sm stamp
5 Sw. Cr. 10 4.6 folded long collared 7.1 sc/sm stamp
18 Sw. Cr. 26 9.8 folded ovaloid pot 9 scraped sm/st
22 Sw. Cr. 8.5 6.1 folded sherical jar slip stamped
25 Sw. Cr. 13 8.1 incised sherical jar 5.9 smoothed stamped
26 Sw. Cr. 26 6.2 folded ovaloid pot 19.75 smoothed stamped
34 Sw. Cr. 3.5 5 compound 6.8 smoothed sm/st
35 Sw. Cr. 7 5.5 sherical jar smoothed stamped
36 unclass 3.5 6 grenade smoothed smoothed
S Sand 6.75 5.15 plain beaker 4 burnished smoothed
4 Sand 20.5 7.1 plain deep globular none smoothed
7 Sand 17.5 7.8 plain ovaloid pot sc/sm smoothed
9 Sand 12.5 6.1 incised sherical bowl 6.25 sc/sm smoothed
10 Sand 9.85 5.5 plain sherical bowl scraped smoothed
11 Sand 21 7.1 plain ovaloid pot scraped smoothed
15 Sand 26 6 folded ellipsoid bowl 5.9 scraped sc/sm
16 Sand 30 9.1 folded sherical jar smoothed smoothed
20 Sand 14 5.75 folded sherical jar smoothed smoothed
32 Sand 8 5.4 folded 11 smoothed smoothed
39 Sand smoothed smoothed
6 D.C.R. 5.2 plain boat slip slip
21 D.C.R. 16.5 4.5 open bowl slip slip
17 St. Johns 12.25 5.5 plain deep globular sc/sm smoothed
27 St. Johns 39 6.5 plain ellipsoid bowl scraped smoothed
24 St. Johns 4.5 4.9 plain bottle smoothed smoothed
31 St. Johns smoothed
19 Weeden Island 12 10.35 three compart- slip slip/
Plain ment burnish
3 Colorinda 11 8 plain ovaloid pot sc/sm smoothed
Charc. P1. = charcoal-tempered plain
Charc. Sw. Cr. = charcoal-tempered Swift Creek (Early)
Unclass = unclassified
D.C.R.= Dunns Creek Red
Unique vessel forms in bold.
Rim thickness taken at 3 cm below lip of each vessel


2004 VOL. 57(4)


Appendix 6. Distribution and degree of coring, clouding, and sooting.

Vessel Classification fire oxidized f.c. exte- % surface f.c. inte- % sur- sooting % sooted
# cores area rior rior face
12 Deptford 3 ext rim med base med rim high
23 Deptford 3 ext
33 Deptford 3 both no no no
8 Charc. P1. 3 int rim med rim med rim med
14 Charc. P1. 3 ext rim High base med rim high
41 Charc. P1. 3 ext
37 Charc.. 8 no
29 Charc. Sw. Cr. no rim high no rim high
30 Charc. Sw. Cr. 7 both rim high no rim high
13 Mayport 3 ext rim med no rim high
28 Mayport no rim med no rim med
38 Mayport 3 both rim med no -low
40 Mayport no
25 Sw. Cr. 3 ext no no no
26 Sw. Cr. no no no no
5 Sw. Cr. 3 ext wall/ rim med base med wall/ high

34 Sw. Cr. 3 ext no no no
1 Sw. Cr. no wall/ rim high base med rim high
18 Sw. Cr. 3 wall/ rim high base med wall/ med

22 Sw. Cr. 3 ext no no no
35 Sw. Cr. 3 both no no no
36 Unclass 3 both no no no
32 Sand 3 ext no no no
2 Sand no wall/ rim med no no
11 Sand 3 ext rim high no rim high
15 Sand 3 ext base/rim med base med rim high
9 Sand 3 ext base/rim med no no
10 Sand 3 ext rim low base low no
20 Sand 3 both wall/ rim med no no
4 Sand 3 ext wall/ rim med base med rim low
7 Sand 3 ext rim med no rim med
16 Sand 3 ext all high rim high
39 Sand no -rim
6 D.C.R. 3 both no no no


Vessel Classification fire oxidized f.c. exte- % surface f.c. inte- % sur- sooting % sooted
# cores area rior rior face

21 D.C.R. 3 both no no no

27 St. Johns 3 ext rim low no no med

17 St. Johns 3 both no no

24 St. Johns 3 ext no no no

31 St. Johns 3 both no no rim

19 Weeden Island 3 ext base low no no

3 Colorinda 3 both no base low no

Charc. Pl. = charcoal-tempered plain
Charc. Sw. Cr. = charcoal-tempered Swift Creek (Early)
D.C.R. = Dunns Creek Red
f.c. = fire clouding
med = medium
unclass = unclassified
3 = oxidized, diffuse core margins
7 = reduced, diffuse core margins
8 = reduced, no core
Fire core numbering follows Rye (1981:Figure 104).


2004 VOL. 57(4)



Environmental Services, Inc., 7220 Financial Way, Suite 100, Jacksonville, FL 32256
E-mail: ghendryx@esinc. cc


In March of 2001, Environmental Services, Inc., (ESI)
performed a cultural resource survey of a small tract on
Amelia Island in Nassau County (Hendryx et al. 2001). As a
result, a multicomponent site was discovered and named the
Honey Dripper Site (8NA910) after a nearby package store and
lounge. The site lies 1100 m west of the Atlantic Ocean on a
high, moderately well drained bluff that overlooks the
Intracoastal Waterway (Figure 1).
During initial site testing, archaeologists encountered
features, dense shell middens, and an artifact assemblage
dominated by Swift Creek pottery. More specifically, the
ceramic assemblage contained a high incidence of Crooked
River sherds, which commonly occur as minority ceramic
types at Swift Creek sites, but which are not reportedly found
as the dominant ceramic type. The site was not avoided by
proposed construction, and given the paucity of well preserved
Swift Creek sites in the region, it was recommended for
mitigation, which was completed in September and October,
2001 (Hendryx and Smith 2001).
This article provides a brief review of what is known of the
Swift Creek period along the lower Atlantic Coast and
discusses how the results of this project contribute to the
existing data. As a result of the excavations, terminal Swift
Creek dates for the Atlantic Coast were refined through
radiometric dating, site seasonality was inferred through
biometric analysis, and subsistence was addressed through
faunal analysis and shell identification. Also, ceramic analysis
provided information on vessel morphology, stamping designs,
and paste characteristics, while artifact and feature patterning
document activities associated with this discrete occupation.

Swift Creek in a Regional Context

The Swift Creek culture is believed to have originated in
southwestern Georgia and southeatern Alabama, where its
influence spread rapidly to the Florida Gulf Coast, and
eventually to northeastern Florida and coastal Georgia (Adams
1985; Ashley 1992; Blanton 1979; Cook 1977; Kelly and
Smith 1975; 1979; Kirkland 1979; Wood et al. 1986). By and
large, Swift Creek-related ceramics have been recovered from
sites within Tennessee and Savannah River valleys (Elliott
1998), and even as far away as the Mann Site in southern
Indiana (Kellar 1979).
It has been suggested that the spread of Early Swift Creek

to northeastern Florida was likely a result of contacts with
northwestern Florida people (Ashley 1998). Over time, as
Weeden Island culture emerged in north-central Florida,
northeastern Florida Swift Creek interaction shifted to south-
central and coastal Georgia. Therefore, during the Late Swift
Creek period, strong similarities between southern Georgia
and northeastern Florida ceramics are evident (Ashley
In northeastern Florida, Swift Creek pottery has tradition-
ally been viewed as nonlocal (Goggin 1952:50). However,
current understanding is that Swift Creek ceramics have been
recovered in appreciable quantities from both mound and
midden contexts near the mouth of the St. Johns River, about
12 miles south of the Honey Dripper Site, to indicate local
production (Ashley 1992, 1998). In the last decade, four Swift
Creek habitation sites were excavated in the region. These
sites include two on Greenfield peninsula, in Jacksonville
(Florida Archeological Services 1995; Smith and Handley
2002), one on nearby Crane Island (Dickinson and Wayne
1999), and one on Amelia Island, about 3 miles northeast of
the Honey Dripper Site (Johnson et al. 1997); this latter site
contained mainly curvilinear stamped sherds with minor
amounts of Crooked River, St. Andrews Complicated
Stamped, and Weeden Island pottery.
The Crooked River pottery type common at the Honey
Dripper Site exhibits a rectilinear stamped design that has
been identified in both Early and Late Swift Creek assem-
blages in the Gulf Coast region (Willey 1949:383-384, 435-
436). Pottery with similar markings has been found associated
with Swift Creek curvilinear sherds at sites throughout the
Florida Panhandle (Bense 1998; Jones and Tesar 1996),
northeastern Florida (Ashley 1998, Johnson et al. 1997),
southern Georgia (Smith 1998), and even at the Mann Site in
Indiana (Kellar 1979); however, nowhere is it reported as the
dominant ceramic type.
The curvilinear pottery at the Honey Dripper is similar to
that reported on by Cook (1979) for the lower Georgia Coast.
Specifically, Cook recovered curvilinear sherds that were
faintly stamped, overstamped, and of poor workmanship, and
noted that stratigraphically these dated later than the classic
Swift Creek. As such, he assigned these sites to a new phase
termed Kelvin, distinguished from the earlier Swift Creek
based on differences in the ceramics, stratigraphy, and
mortuary and ceremonial practices; however, this phase
designation has largely been discontinued.


VOL. 57(4)




Figure 1. Honey Dripper site location.

Survey and Site Testing

The Honey Dripper Site covered 75 m (east-west) by 65 m
(north-south). As indicated by shovel testing, subsurface
probing, excavation, and soil stripping, the area of artifact
concentration and dense shell midden was restricted to an
approximate 100 square m area in the center of the site. The
site may extend further to the north and east, but property

boundaries restricted further testing in these directions. The
area to the north is residential, yet appears to possess large
undisturbed areas of midden, while the area to the east has
been impacted by Highway A1A construction. The western
boundary was the edge of the bluff, and the southern boundary
was delineated through shovel testing.
Twenty-six shovel tests were dug at 12.5 m intervals across
the site, with 23 producing cultural material. Twenty-one tests


2004 VOL. 57(4)


r C


S 0Q.S. Grader Strip
S- Dense Shell Midden
0 1 2m

Figure 2. Honey Dripper site map and plan of excavations.

contained prehistoric artifacts and 8 yielded historic artifacts
that were temporally related to the nearby, late eighteenth
century Vaughn Plantation. Varying amounts of shell were
also encountered; 9 tests contained moderate to dense shell
midden and 15 had light shell. All shovel tests measured 50
by 50 cm, were dug to 1 m below surface, and were sifted
through %-in mesh.
The prehistoric ceramic count ranged from 1 to 17 sherds
per shovel test, with a mean frequency of 2.9 per shovel test.
Areas of artifact concentration and dense shell midden were
found near the center of the site to a depth of about 50 cm.

Swift Creek ceramics (rectilinear and curvilinear) constituted
the majority of the artifacts, although lesser amounts of
Weeden Island, Colorinda, St. Marys II, and San Pedro series
ceramics were recovered, as were Colonial period artifacts.
Seven 1 by 2-m test units were excavated during prelimi-
nary site testing. Two of the units were placed along the
western bluff in the area of historic occupation; one unit was
dug along the eastern periphery of the site where dense shell
midden was associated with sparse artifacts; and four units
were dug in the central portion of the site where shell and
ceramic artifacts were most frequent. All test units were




Feature IE

Feature 17





Table 1. Block Unit Prehistoric Ceramics.

Ceramic Count Count (%) Weight (g) Weight (%)
San Pedro 15 6.3 114.2 5
St. Marys II 16 6.7 61.9 2.7

St. Johns 6 2.5 41.9 1.8
Weeden Island 2 0.8 13.4 0.6
Crooked River 93 38.9 1223.3 53.4
Swift Creek 27 11.3 208.4 9.1
Uid, decorated 33 13.8 222 9.7
Plain 47 19.7 404.2 17.6

diminutive (omitted 476 635.4
from tabulation)
TOTAL 239 100 2289.3 99.9

excavated in 10-cm levels with attention to natural stratigra-
Excavations in the historic area revealed considerable
artifact mixing, and the unit in the eastern portion contained
few artifacts. The four units dug in the central portion of the
site contained intact deposits with Swift Creek ceramics
associated with features and dense faunal remains. Based on
these findings, a mitigation plan was devised that allowed
residential construction to continue following completion of an

Excavation and Stripping

To mitigate proposed damages to the site prior to construc-
tion an additional 20 square m was hand excavated in the
central portion of the site and 186 square meters were exposed
through grader stripping, accomplished using a Takeuchi,
which is the Japanese equivalent of a Bobcat. Two of the units
dug during the preliminary site testing were incorporated into
the 20-square m area, thus the bulk of the excavations were
within a 24-square m block (Figure 2).
Two strata were revealed in the block, including an
approximate 30-cm deep, black shell midden and an underly-
ing light brownish gray to pale brown sand. There was
considerable artifact mixing within the shell midden, but the
light sand sub-midden assemblage was dominated by Swift
Creek pottery, including Crooked River and curvilinear
stamped sherds. Most of the Swift Creek artifacts from the
block, however, came from subsurface features. In fact, only
715 (2.9 kg) prehistoric ceramics were recovered from the
general matrix of the 20 square m area, including 476 that

were smaller than 2 cm in size and not classified by type
(Table 1).


During test unit excavation and grader stripping. 17
cultural features were excavated, including 9 that dated to the
Swift Creek period, one dated to the St. Marns II period. and
seven of unknown prehistoric temporal affiliation (Table 2).
The Swift Creek features included 7 cooking pits and two
refuse pits, and some of the cooking pits appeared to have been
subsequently used for refuse. Given the dominance of Swift
Creek ceramics at the site, it is likely that many of the features
of unknown affiliation also dated to the period.
Many of the features were densely packed with artifacts.
shell, and faunal remains. Cooking pits were the most
prevalent feature type. In plan view, they ranged in size from
30 x 49 cm to 139 x 91 cm. One of the cooking pits (Feature
4) was surrounded by 12 posts that may have functioned as
supports for suspension over the pit. Two of the cooking pits
provided opportunity for radiometric dating (Features 3 and 4),
and two more (Features 4 and 18) provided a sample of
impressed odostome shells suitable for determining site
seasonality. Many of the pits revealed both curvilinear and
rectilinear sherds. Feature 16 revealed a brushed sherd along
with Crooked River pottery, and Feature 18 produced Weeden
Island Punctate along with rectilinear and curvilinear stamped
sherds. All of the cooking pits provided faunal remains
suitable for analysis.
Feature 6 was a refuse pit that yielded fragmentary remains
of a large portion of a reconstructed pot. One of the sherds


2004 VOL. 57(4)


Table 2. Feature Inventory.

that mended with the pot was recovered from Feature 3, a
cooking pit 4 m to the north, and this sooted sherd from
Feature 3 was submitted for AMS dating prior to unearthing
the remaining pieces of the vessel in Feature 6.

Ceramic Analysis

Combining the survey and excavation artifacts, Swift Creek
sherds (combining Crooked River under the rubric of Swift
Creek) comprised 82.8 percent of the diagnostic prehistoric
ceramic assemblage by count or 92.4 percent by weight (Table
3). Crooked River Complicated Stamped sherds represented
the bulk of the assemblage, and the remaining Swift Creek
wares consisted of curvilinear sherds, including a few that
were rectilinear stamped (Figure 3). Willey (1949:383-384)
described the Crooked River surface treatment as exhibiting a
rectilinear stamped pattern with zigzag lines or chevrons; this
ceramic type was encountered 2.1 times more frequently than
curvilinear sherds.
Generally, the curvilinear paddle stamped designs from the
Honey Dripper Site were less intricately carved than those of
the classic Swift Creek period, and many were also

overstamped or faintly stamped and did not provide clear
stamp recognition; these are characteristics of Kelvin phase
sherds of the lower Georgia Coast (Cook 1979:77). The small
sample of curvilinear sherds with clear impressions were
submitted to Frankie Snow for examination; however, none of
the motifs represented known paddle designs (Frankie Snow
personal communication, 2001).
Because sites with a high frequency of Crooked River
sherds are poorly represented in northeastern Florida, very
little information is available for their paste composition.
Therefore, 14 Crooked River rim sherds that appeared to
represent different vessels were selected for detailed micro-
scopic analysis using a 105x microscope with a built in
micrometer. Temper characteristics were based on measure-
ments provided in the Wentworth Size Classification Scale
(Shepard 1980:118) and the Relative Abundance Scale
established by Rice (1987:12.2). Although there was some
variability, most of the selected Crooked River rim sherds
contained abundant medium (0.25 to 0.5 mm) grit particle
inclusions. Grog is a minority tempering agent among the
Kelvin specimens described for the lower Georgia Coast (Cook
1979:77,80), but it was absent among the Honey Dripper

Size Depth
Type Provenience (cm) (cmbs) Affiliation
1 Unknown Pit Unit 4 28 x 70 45-75 Prehistoric
2 Storage Pit? Unit 5 58 x 34 30-52 Prehistoric
3 Cooking pit Units 7 and 15 104 x 67 50-94 Swift Creek
4 Cooking pit Units 6 and 16 139 x 91 17-37 Swift Creek
5 Cooking/Refuse pit U-7 and 15 105 x 90 43-63 Swift Creek
6 Refuse pit Units 12 and 13 94 x 74 23-38 Swift Creek
7 Cooking pit Units 10 and 11 52 x 70 26-37 Prehistoric
8 Cooking pit Unit 12 40 x 22 20-34 Prehistoric
10 Post Unit 14 18 x 18 25-38 Prehistoric
12 Cooking pit Units 12 and 13 66 x 26 35-45 Prehistoric
13 Refuse pit Trench 3 90 x 96 37-51 St. Marys II
14 Cooking pit Trench 4 64 x 55 43-62 Swift Creek
15 Post Trench 5 23 x 23 46-55 Prehistoric
16 Cooking pit Trench 3 30 x 49 38-74 Swift Creek
18 Cooking pit Trench 3 and 8 93 x 90 70-101 Swift Creek
19 Refuse pit Trench 4 52 x 62 49-78 Swift Creek
20 Cooking pit Trench 8 39 x 45 40-65 Swift Creek





Figure 3. Representative Swift Creek sherds. Top: Crooked River; Middle: curvilinear stamped design; Bottom: curvilinear
with rectilinear stamped design.

complicated stamped sherds.
In addition to microscopic analysis, formal attributes of the
rim sherds were recorded whenever possible. Vessel orifice
diameter estimates were obtained from 7 Crooked River rim
sherds that exhibited between 5 and 50 percent of the total
rim; the orifice diameters ranged from 15 to 25 cm. Five
sherds had a straight profile, and appeared to represent
unrestricted vessels; the rim diameter for these sherds ranged

from 17 to 25 cm, with between 5 and 15 percent of the rim
available. One sherd exhibited an inverted profile and a rim
diameter of 15 cm, with 10 percent of the rim available. A
partially reconstructed vessel exhibited a slight rim flare and
a diameter of 20 cm, with 50 percent of the rim present.
Three lip classification were observed on the selected
Crooked River rims: round, tapered, and flat. Ten sherds
exhibited rounded lips, three exhibited tapered lips, and the

2004 VOL. 57(4)



Table 3. Diagnostic prehistoric ceramics from 8NA910.

Ceramic Classification Count Count (%) Weight (g) Weight (%)
San Pedro 25 5.7 178.7 3.2
St. Marys II 34 7.8 127.1 2.3
St. Johns 6 1.4 2.5 0.1
Colorinda 2 0.4 7.9 0.1
Weeden Island 8 1.8 102 1.8
Crooked River 228 52.3 3487.7 62.9
Swift Creek Curvilinear 133 30.5 1636.4 29.5
TOTAL 436 99.9 5542.3 99.9

remaining lip was flat.

Morphological vessel information was derived from a
partially reconstructed vessel (Figure 4). A collared jar with
a slight rim flare and an elongated, globular body was recov-
ered from Feature 6. This vessel was stamped with a Crooked
River design that consisted of nested chevrons and an inter-
locking rectangular pattern, as illustrated in Figure 5.
There was some variability among the Crooked River
paddle stamped impressions, yet the number of paddles at this
site appears to have been limited. At least three different
Crooked River paddle designs were present. A few sherds
exhibited stamping that was applied with a paddle that
contained both curvilinear and rectilinear motifs, others had a

Figure 4. Reconstructed vessel (mended).

nested chevron design (Figure 3), and still others were
stamped with the nested chevron and interlocking rectangular
pattern, as depicted in Figure 5.

Site Chronology

A Swift Creek chronology was established for the Atlantic
coast based on 20 radiocarbon dates (Stephenson et al.
2002:337), and seven more have recently been added
(Dickinson and Wayne 1999; Smith and Handley 2002;
Stephenson 2002). Nineteen of the dates come from four sites
in Georgia (9CM171A, 9CM233, 9MC-10-1, and 9MC360),
and eight come from four sites in Florida (8DU68, 8DU96,
8DU5545, and 8NA709). Intercept dates range from A.D. 480
to 800. The earliest date was obtained from the
S soot off a charcoal-tempered sherd from the
Mayport Mound (8DU96) in Jacksonville, and the
most recent date was from an oyster shell from
Crane Island (8NA709), about two miles north-
west of the Honey Dripper Site.
These radiometrically dated sites are domi-
nated by pottery stamped with classic curvilinear
Swift Creek motifs. As has been noted, the poorly
executed, overstamped, and faintly stamped sherds
referred to as Kelvin along the lower Georgia coast
are thought to be later than those of the classic
Swift Creek series (Cook 1979). Although no
dates are available for sites of this nature on the
Atlantic coast, late Woodland period dates were
anticipated for radiocarbon samples from the
Honey Dripper Site, since the site wares were
similar to those found on "Kelvin phase" sites.
Two samples from the Honey Dripper Site
were submitted for radiometric dating, and the
returned dates are later than the other 27 Swift
--_ Creek dates known for the Atlantic Coast. A
sooted Crooked River sherd from Feature 3 was
submitted for AMS dating, and a sample of stout




Figure 5. Crooked River paddle designs from reconstructed vessel.

tagelus shell from Feature 4 was submitted for radiocarbon
dating. The intercept dates for the two samples were only 10
years apart and dated to A.D. 870 (Beta-159965) and 880
(Beta-159964) (Table 4). These dates suggest that the tempo-
ral span for Late Swift Creek occupation along the Atlantic
coast should be extended for another 100 years. Moreover,
there may be a need to establish a Late Swift Creek subperiod
for the region, one defined by sites with rectilinear sherds
associated with poorly executed, faint, and overstamped
curvilinear sherds. Excavations at similar site types and

Table 4. Calibrated Radiometric Dating Results.

obtaining more radiocarbon dates will assist resolution of these


Subsistence was addressed through an analysis of faunal
remains recovered from feature contexts and through field
observations regarding discarded shell. Faunal remains within
the midden and test unit levels were minimal and due to mixed
cultural assemblages within these proveniences faunal analysis

13~i prr

Lab # Material Measured Conventional 1 sigma calibrated 2 sigma calibrated
Radiocarbon Age Radiocarbon Age date range date range
Beta-159964 food residue 1150 +/-40 BP 1180 +/-40 BP Cal AD 790 Cal AD 770 to 970
to 900
Beta-159965 shell 1120+/- 60 BP 1540 +/-60 BP Cal AD 780 to 920 Cal AD 720 to
* Corrected for C13/C12 using regional estimation.


2004 VOL. 57(4)


Table 5. Faunal Analysis.

Identification Category Count (NISP)
blue crab 7
uid. Crab 1
Canid spp. 1
Crevalle Jack 1
drum 33
Gopher Tortoise 1
uid. Small mammal 5
uid. Small-medium mammal 23
uid. Medium mammal 19
uid. Medium-large mammal 11
uid. Large mammal 56
uid. Mammal 6
Sand tiger shark 1
(probable) sea bass 1
sea catfish 13
Sheepshead 1
(probable) softshell turtle 1
uid. Bird 49
uid. Fish 2564
uid. Shark 1
uid. Snake 16
uid. Turtle 36
White-tailed deer 23
uid. Bone 186
TOTAL 3056

was only conducted on the remains found within features and
from a 10-cm square column sample. A total of 3056 bone
remains was analyzed from 1' in and fine screen samples.
Although many of the remains were too small to be identified
to species level, the analysis of diagnostic remains indicated an
emphasis toward maritime aquatic exploitation (Table 5).
Some avifauna and terrestrial species were exploited, such
as deer, tortoise, turtle, snake, bird, and various mammals not
conclusively identified, yet fish bone was most prevalent
within the assemblage. One point of interest is the presence of
crevallejack, which is a very large fish commonly encountered
in Swift Creek faunal assemblages (Keith Ashley, personal

communication, 2001), yet its method of capture remains
During excavation all recovered shell was quantified by
volume in the field using buckets marked in metric incre-
ments. A total of 1,117 1 was collected from a 20-square m
sample area, excluding feature-derived shell. Oyster
(Crassostrea virginica) shell represented the dominant species,
comprising 96.4 percent of the assemblage (n= 10771), and the
remaining shell types included: quahog clam (Merceneria
mercenaria) (n=33 1), stout tagelus (Tagelusplebeius) (n=4
1), whelk (Busycon contrarium) (n=2 1), and a trace of moon
shell (Naticidae).


Biometric analysis of impressed odostome shells (Boonea
impressa) was employed to investigate site seasonality.
Impressed odostomes are small parasites that prey upon
oysters, and are inadvertently collected along with the shellfish
and incorporated into archaeological deposits. They spawn in
the spring or early summer (May and June) and have a life
cycle of approximately one year. Over their lifetime Boonea
experience measurable and predictable accretionary growth,
making them an ideal subject for seasonality studies. Length
measurement of impressed odostomes at their time of death
has been used to suggest the time of the year (season) when
oysters within a sampled deposit were harvested (Russo 1991;
Russo et al. 1993), which has allowed researchers to infer the
seasons) that a site was occupied.
All impressed odostomes were removed from the fine
screen window mesh samples of Features 4 and 18, and whole
specimens were tallied and measured using a 7x ocular
micrometer. These two features were selected because they
both contained dense shell and good evidence for a Swift
Creek affiliation. All Boonea samples were measured to the
nearest tenth of a millimeter, were placed into 1 mm incre-
ment size categories, and were compared by modal frequency.
The Feature 4 sample size consisted of 67 Boonea and the
Feature 18 sample consisted of 145 (Table 6). In order to
interpret the seasons) represented, the data were depicted
graphically (Figure 6) and compared to a master seasonal
graph developed by Russo (1991:210), based on live odostomes
collected from Crescent Beach, Florida over a 14-month
period. The two lines in Figure 6 are very similar and both
support a late autumn/early winter oyster procurement pattern.
This period of oyster harvest may suggest an awareness of a
favored exploitation period, as those collected during the
spawning summer months are more prone to disease, and
contain fewer nutrients.

Site Structure/Function

Through an examination of artifact and feature patterning,
some interpretations of site structure and function are possible.
The site was settled intermittently throughout history with
emphasis during the Late Swift Creek phase of the Woodland
period. The group that settled here at that time employed



30 --- -- Feature 4
20 --- Feature 18
10 --, ,-_- ,-,----

1 mm 2 mm 3 mm 4 mm 5mm 6mm 7 mm
(spring) (summer) (autumn) (winter)

Figure 6. Oyster seasonality, Features 4 and 18.

cooking pits for food preparation and were concerned with
maintaining an orderly living quarter, as evidenced by the
consolidated accumulation of discarded remains within refuse
pits and concentrated piles of discarded shell observed amidst
the sheet midden. Specifically, two concentrated shell
middens were exposed during block excavation and were even
more obvious during the site stripping. One midden was 2 m
wide and 8 m long and located within and around the north-
eastern corner of the block unit; a smaller (4 x 6 m) and less
dense shell midden was to the southwest of the excavation
block (see Figure 2).
Thevarying-density sheet midden was observed throughout
the entire block unit and stripped portions of the site; however,
it was not as dense as the discrete shell clusters underlying and
intermixed with it. Pottery types other than Swift Creek were
found in higher density in the sheet midden than in the
features. Therefore, it's likely that the sheet midden accumu-
lation was a result of later transient groups that harvested
shellfish and discarded the shell in a random manner across
the site, a characteristic of nonpermanent site usage.
Shovel testing, excavation, and site stripping revealed the
core of a Late Swift Creek site within an approximate 100
square m area. The duration of occupation is believed to have
been short. It's likely that a small group, perhaps 1 to 2

Table 6. Modal Size Class Data for Impressed Odostomes
(Boonea Impressa).

Size Class Feature 4 Feature 18
2 mm (n=2) 3% (n=l) 1%
3 mm (n=18) 27% (n=20) 14%
4mm (n=21) 31% (n=52) 36%
5 mm (n=25) 37% (n=56) 39%
6mm (n=l) 1.5% (n=16) 11%
Measured Total 67 145

families, occupied the site during the late autumn/early winter
over a single season or some sequential reoccupation. This
assertion is supported by the radiometric dates, the confined
cluster of archaeological deposits, and biometric analysis of

Summary and Suggestions for Future Research

Excavations at 8NA910 have provided valuable data on the
Late Swift Creek in northeastern Florida, where few such sites
have been investigated. Future research at similar sites can
continue to resolve whether the preponderance of Crooked
River Complicated Stamped and non-intricately stamped
curvilinear pottery at late sites reflects the decline of Swift
Creek pottery manufacture, or whether this site and others
with comparable ceramic assemblages should be differentiated
from Late Swift Creek.
It is also important to try to identify the relationship
between Late Swift Creek coastal occupants and that of the
Late Woodland Colorinda people, who were known to have
occupied the region between AD 850 and 900 (Ashley 2003).
Future work should also investigate cultural contacts and
influences through intersite comparative paddle stamp
analysis, which may provide insight on seasonal movements
or movements through intermarriage (Stoltman and Snow
1998:152). Finally, more radiometric dates are necessary
to continue refining the temporal placement of Swift Creek
pottery varieties.


I would like to extend thanks to the Summer Beach Develop-
ment Group, Ltd. for project sponsorship. The survey and site
testing crew did a great job revealing the site's significance; they
include Tony Kuhner, Brian Floyd, Brent Handley, and Sean
Taylor. The excavation crew of Tony Kuhner, Greg Gonzales,
and Jeff Lanham are thanked for their careful work. Brent
Handley is thanked for analyzing the faunal remains. Greg Smith
provided considerable insight on an earlier draft and served as
Principal Investigator for the project. Thanks to Keith Ashley for
sharing his knowledge of the Swift Creek and providing editorial
comment. Ryan Wheeler and an anonymous reviewer are also


2004 VOL. 57(4)


thanked for their editorial comment. Special appreciation is extended
to Brian Floyd, who analyzed the artifacts, provided the artifact
illustrations, and donated his weekend time, along with his wife
Danielle, to conduct additional excavations at the site. Finally,
Marsha Chance, and Rhodes and Sarah Robinson are thanked for
their support throughout the project.

References Cited

Adams, William Hampton (editor)
1985 Aboriginal Subsistence and Settlement Archaeology of the
Kings Bay Locality, Vol. 1. University of Florida, Depart-
ment of Anthropology, Reports of Investigations 1,

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

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

2003 Interaction, Population Movement, and Political Economy:
The Changing Social Landscape of Northeastern Florida
(AD 900-1500). Unpublished Ph D. dissertation, Depart-
ment of Anthropology, University of Florida, Gainesville.

Bense, Judith A.
1998 Santa Rosa-Swift Creek in Northwestern Florida. In A
World Engraved: Archaeology of the Swift Creek Culture,
ed. by Mark Williams and Daniel T. Elliott, pp. 247-273.
University of Alabama Press, Tuscaloosa.

Blanton, Dennis
1979 An Archaeological Survey of the Upper SatillaBasin. Early
Georgia 7(2):43-64.

Cook, Fred C.
1977 The Lower Georgia Coast as a Cultural Sub-Region. Early
Georgia 5(1-2):15-36.

1979 Kelvin: A Late Woodland Phase on the Southern Georgia
Coast. Early Georgia 7(2):65-86.

Dickinson, Martin F., and Lucy B. Wayne
1999 Island in the Marsh: An Archaeological Investigation of
8NA59 and 8NA709, The Crane Island Sites, Nassau
County, Florida. Report on file, Division of Historical
Resources, Tallahassee.

Elliott, Daniel T.
1998 The Northern and Eastern Expression of Swift Creek
Culture: Settlement in the Tennessee and Savannah River
Valleys. In A World Engraved: Archaeology of the Swift
Creek Culture, ed. by Mark Williams and Daniel T. Elliott,
pp. 197-221. University of Alabama Press, Tuscaloosa.

Florida Archeological Services
1995 Phase II Archeological Investigations at Sites 8DU5541,
8DU5542, and 8DU5543 at the Queens Harbour Yachtand

Country Club. Report on file, Division of Historical
Resources, Tallahassee, Florida.

Goggin, John M.
1952 Space and Time Perspective in Northern St. Johns Archae-
ology, Florida. Yale University Publications in Anthropol-
ogy 47. New Haven.

Hendryx, Greg S., and Greg C. Smith
2001 Archaeological Data Recovery and Mitigation at 8NA910
(The Honey Dripper Site) Nassau County, Florida. Report
on file, Division of Historical Resources, Tallahassee.

Hendryx, Gregory S., Greg C. Smith, and Keith H. Ashley
2001 A Cultural Resource Assessment Survey of the River Place
at Summer Beach and Site Testing at 8NA910, Nassau
County, Florida. Report on file, Division of Historical
Resources, Tallahassee.


Robert, Myles C.P. Bland, B. Alan Basinet, and Bob

Archaeological Investigation of the Ocean Reach Site
(8NA782), Nassau County, Florida. Ms. On file, Florida
Archeological Services, Inc., Jacksonville.

Jones, B. Calvin, and Louis D. Tesar
1995 Emergency Archaeological Salvage Excavation Within the
Swift Creek Subarea of the Block-Sters Site (8LE148),
Leon County, Florida: A Public Archaeology Project.
Report on file, Division of Historical Resources, Tallahas-

Kellar, James H.
1979 The Mann Site and "Hopewell" in the Lower Wabash-Ohio
Valley. InHopewellArchaeology: The Chillicothe Confer-
ence, ed. by D.S. Brose and N. Greber, pp. 100-107. Kent
State University Press, Kent, Ohio.

Kelly, Arthur R., and Betty Smith
1975 The Swift Creek Site, 9BI3, Macon, Georgia. Manuscript
333 on file, Laboratory of Archaeology, University of
Georgia, Athens.

Kirkland, S. Dwight
1979 Preliminary Investigations on Floyd Creek, Camden
County, Georgia. Early Georgia 7(2):1-25.

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

Russo, Michael
1991 A Method for the Measurement of Season and Duration of
Oyster Collection: Two Cases from the Prehistoric South-
east U.S. Coast. Journal ofArchaeologicalScience 18:205-

Russo, Michael, Ann S. Cordell, and Donna L. Ruhl
1993 Timucuan Ecological and Historical Preserve Phase III
FinalReport. National Park Service, Southeast Archeologi-
cal Center, National Park Service, Tallahassee.

Shepard, Anna O.
1980 Ceramicsfor the Archaeologist. Publication 609, Carnegie




Institution of Washington, Washington, D.C.

Smith, Betty A.
1998 Neutron Activation Analysis of Ceramics from Mandeville
and Swift Creek. In A WorldEngraved: Archaeology ofthe
Swift Creek Culture, ed. by Mark Williams and Daniel T.
Elliott, pp. 112-129. University of Alabama Press,

Smith, Greg C., and Brent M. Handley
2002 Addendum to: Archaeological Data Recovery and Afitiga-
tion at 8DU5544/5545 Queen's Harbour Yacht and
Country Club Duval County, Florida. Report on file,
Division of Historical Resources, Tallahassee.

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

Stoltman, James B., and Frankie Snow
1996 Cultural Interaction within Swift Creek Society: People,
Pots, and Paddles. In A World Engraved. Archaeology of
the Swift Creek Culture, ed. by Mark Williams and Daniel
T. Elliott, pp. 130-153. University of Alabama Press,

Willey, Gordon R.
1949 Archeology of the Florida Gulf Coast. Smithsonian Aiscel-
laneous Collections, Vol. 113. Washington, D.C.

Wood, W. Dean, Daniel T. Elliott, Teresa P. Rudolph, and Dennis B.
1986 Prehistory in the Richard B. Russell Reservoir: The
Archaic and Woodland Periods of the Upper Savannah
River. Archeological Services Branch, National Park
Service, Russell Papers. Atlanta.

2004 VOL. 57(4)


Public Benefits of Archaeology, edited by Barbara J. Little.
2002. University Press of Florida, Gainesville. xix+260 pages,
figures, tables, contributors, index. $39.95 (hardcover)

Archaeological Consultants, Inc. 8110 Blaikie Court, Suite A,
Sarasota, FL 34240

Public Benefits ofArchaeology is a must for any archaeolo-
gist involved with the public. It is easy to follow, being divided
into four sections: Finding Common Ground; Many Publics,
Many Benefits; Learning from an Authentic Past; and Promot-
ing the Public Benefits of Archaeology. The first section
emphasizes authenticity, a common theme occurring through-
out all the sections. The second section deals with the many
types of people (i.e., the public) different archaeological
projects affect; and how to involve the public in a project that
benefits all. Section three also deals with the general public,
but focuses on the relationship among tourism, archaeology,
and bringing archaeology into the classroom. The final section
touches briefly on archaeology in politics.
Section 1, Finding Common Ground contains two chapters:
"Archaeology as a Shared Vision" and "Public Benefits of
Archaeological Research." William Lipe focuses on how
authenticity helps the public make a connection between the
past and present. When an artifact, feature, or structure is
found, it provides a direct link between the past and present.
When the connection is made, the public becomes willing to
learn and to help in preserving archaeological sites.
Section 2, Many Publics, Many Benefits, focuses on pro-
grams where professionals have been directly involved with
the public. In chapter 3, "Heritage, History and Archaeological
Education," Francis McManamon explains the benefits of
heritage and history to archaeology, which enables a person to
feel associated with the past. A good example is St. Augustine
where visitors can walk through history. History benefits
archaeology through continued investigations as we obtain the
truth about people and events of the past. James Whittenburg
uses the Chesapeake, Virginia area to emphasize the impor-
tance of field investigations, which can change our under-
standing about sites. Leigh (Jenkins) Kuwanwisiwma discusses
the importance of archaeologists working together with the
descendants of the people who created the artifacts. Chapter 5,
discusses how descendents of two other cultural groups were
involved in archaeological projects; the immediate descendants
(from the turn of the century) of Chinese Americans and
African Americans in two separate communities. These two
projects helped show the "historical bases of modern society"
(p. 58). Rathje's garbology project examines recent garbage to

help manage our current and future resources. Having been
involved in his Naples (Florida) project, I was able to observe
the garbage being recovered from landfills which dispelled
such myths about what does and does not decompose. "Under-
water Heritage and the Diving Community" by Lynn Harris,
discusses South Carolina's program for educating divers about
the rules and methods for underwater archaeology.
Section 3, Learning from an Authentic Past, focuses on
taking the tourist beyond viewing archaeological sites and
features to the whys and hows of the discipline. This section
also emphasizes the importance of the early introduction of
archaeology to students. "Protecting the Past to Benefit the
Public" discusses the how and why of protecting the past:
legislation, funding, public education, awareness, and coopera-
tion. "Roadside Ruins" notes that museums offer authenticity
as do "along the roadside" sites such as the Alamo or Chaco
Canyon. "Archaeology and Tourism at Mt. Vernon" highlights
research and fieldwork as main tourist attractions. Similarly,
"Broadening the Interpretation of the Past at Harpers Ferry
National Historical Park," emphasizes not only excavation but
how to interpret the discovery, which appeals to the tourist.
Chapter 14, "Myths, Lies, and Videotapes," discusses how
children given incorrect information can come to believe it as
the truth. "Project Archaeology, Putting the Intrigue of the
Past in Public Education" by the Bureau of Land Manage-
ment's National Heritage Education Program prepares educa-
tors to teach archaeology in their instructional settings (p.
179). Another program for school age children that is
discussed is "Pursuing the ZiNj Strategy Like There's No
Tomorrow," aimed at making science fun.
In the final section, Promoting the Public Benefits of
Archaeology, several chapters are devoted to discussions of the
politics that involve archaeology; later chapters focus on
relating archaeology to the general public in a non-scientific
way. In Chapter 17, "Irreplaceable Heritage, Archaeology and
the National Register of Historic Places" Carol Shull discusses
the importance of the National Register of Historic Places
(NRHP) and the benefits of listing a site in the NRHP. The
following chapter "Archaeology in Santa Fe, A Public Private
Balancing Act," deals with local ordinances and why they are
necessary. More locally, similar ordinances exist in some
Florida counties, and those who have conducted work under
these ordinances realize the importance of such ordinances in
cultural resource management. In "Pot Sherds and Politics"
Terry Goddard gives a local government's perspective in which
he discusses how politics stepped in and stopped a project; but
the long term outcome was still good. He also makes a good
point: "archaeology sites need not be seen as impediments to
progress but as opportunity zones" (p. 215). In addition,



VOL. 57(4)



politics and archaeology can go together. In Chapter 20,
"Archaeology and the Tourism Train," Katherine Slick lays
out some great ideas for including archaeological sites into
tourism so that they can be further protected.
The last two chapters outline ways to promote archaeology
in a non-scientific way for the benefit of the public. For
example, we have the internet ("The Web of Archaeology, Its
Many Values and Opportunities), which provides many
opportunities to learn about current and previous research on
archeological sites. Information about archaeology can also be
communicated through stories. According to Peter Young
(editor of Archaeology Magazine), the public enjoys reading
stories about current archaeological research and archaeolo-
gists are encouraged to submit information about their work
without worrying about being good story writers. However,
when writing for the public, keep the information simple so
that the reader will be able to follow what is being written.
Thus, in the final chapter, Mitch Allen gives the scientific
community a few hints in "Reaching the Hidden Audience,
Ten Rules for the Archaeological Writer."
This book is a good tool for archaeologists seeking ideas on
ways their project can benefit and include the public; and the
book's success stories provide encouragement. Public Benefits
of Archaeology also emphasizes what most archaeologists
know: authenticity plays a big role in securing the public's
attention and aids in their appreciation of our cultural heritage.
Living in a state where heritage tourism is an important
economic factor, we have the opportunity to promote our
heritage to people from allover the world. We have authentic
remains of Florida's heritage allover the state and the artifacts
we have, from the coquina buildings in St. Augustine to the
shell mounds in Fort Myers, visually help promote the authen-
ticity of Florida's historic past.

Archaeological Survey in the Lower Mississippi Alluvial
Valley, 1940-1947. Philip Phillips, James A. Ford, and James
B. Griffin edited by Stephen Williams. 2003. University of
Alabama Press, Tuscaloosa. 457 pages. $65.00 (hardcover),
$34.95 (softcover)

Department of Anthropology, Box 117305 University of
Florida, Gainesville, FL 32611
E-mail: jwaggo@ufl. edu

Archaeological Survey in the Lower Mississippi Alluvial
Valley, 1940-1947 by Phillip Phillips, James A. Ford, and
James B. Griffin is a classic work pertaining to the archaeol-
ogy of southeastern North America. Working together, the
University of Alabama Press and Stephen Williams have made
this well-known publication available to a new generation of
students and scholars. Both should be commended for their
efforts. This volume is not only useful to archaeologists
interested in the study of the Mississippian Period, but it also
gives a glimpse into the minds of three leading figures of the
culture historical paradigm that was entering its peak of
influence at the time of its initial publication. Originally

reviewed by Williams (1952). William Haag (1953). and
highlighted by Robert Dunnell (1985). the volume was
commended for its scope and the problems the authors
attempted to address in its voluminous 457 pages. Dunnell
goes as far to comment that it is one of the more progressive.
because of its regional as opposed to site emphasis. yet
undervalued works of it's time, primarily due to its focus on
chronological concerns which were later eschewed by the
emerging new archaeology (Dunnell 1985:297).
The volume begins with a new introduction written by
Stephen Williams who had both professional and personal
relations with each one of the authors. Having had such an
intimate connection to the authors not only allows him to
present both insights into the project, and the thinking of and
interactions between Phillips. Ford. and Griffin. but also to
address recent reinterpretations of the work by culture histori-
cal revisionists. More importantly, however. he discusses at
length each of the players involved and the impetus behind the
project that began with a meeting at Louisiana State University
(LSU) in Baton Rouge, during the fall of 1939. Each member
brought with them varying background experiences and
approaches to the interpretation of the archaeological record.
The Lower Mississippi Archaeological Survey was sponsored
by the School of Geology at LSU. the Museum of Anthropol-
ogy of the University of Michigan. and the Peabody Museum
of Harvard University and encompassed a massive portion of
the Mississippi Valley from the mouth of the Ohio River to
Vicksburg, Mississippi.
Two parts comprised the heart of the project. one being the
exploration and survey of a previously understudied area and
the other a search for the origins of Middle Mississippian
culture. Fieldwork commenced in 1940 and continued in 1941
only to be interrupted by World War II. with each of the three
survey members making various contributions to the war
effort. Following the war's conclusion, the survey was
continued in 1946 and completed in 1947. A total of seven
months was spent in the field during which 382 sites were
recorded, eleven of which were excavated and produced
346,099 sherds. It was hoped that the findings could be used
to augment Ford's earlier work in the valley's lower reaches.
All in all, the project was an immense undertaking considering
its time and available resources.
Section I of the volume is an extensive overview of the
project area's geographical features that comprised roughly
350 miles of the Mississippi Valley and begins with a general
division of the local physiography. Emphasis is given to
alluvial attributes, specifically the flood plain topography that
is dominated by natural levees and back swamps. Phillips,
Ford, and Griffin generally divided the valley into two major
sections, northern and central that are highlighted on a pull
out map, and further subdivided into tributary basins, which
the authors hope will have a bearing on archaeological
findings. Each sub-area is discussed with regard to its climate.
soils, vegetation, flora, and fauna. The section concludes with
an interesting discussion of the interrelationship of culture and
environment, which the authors contend should be considered
together as opposed to singular entities in order to be truly

2004 VOL. 57(4)


The fieldwork is discussed in Section II in which the general
survey program is laid out regarding field methodologies. A
table of all the sites recorded during the survey is included at
the end of the section. Section III is devoted to the classifica-
tion of pottery collected during the project. This section is one
of the most interesting because of the varying approaches of
the authors to artifact classification, which Williams discusses.
Prior to devoting a lengthy section on the description of each
of the 42 pottery types encountered in the lower valley, the
authors discuss the concept of "type," "pottery complex," and
"series," along with the binomial aspects of classification, and
the methodology they implemented during the pottery analysis.
Two important, yet differing, points are presented in the
discussion of use of the "type" concept. One is that it has
relevance to the expression of ideas and behavior of people in
the past (culture), and the other is that it serves as a means to
address spatial and temporal concerns (scientific empiricism).
In their discussion, the authors express their belief that by
defining types at increasingly finer scales of analysis they will
eventually take on a cultural meaning rather than just a
scientific one. However, they also caution against a total
emphasis on pottery types at the expense of learning about the
people who initially made them. The discussion of "type" is
relevant because it represents a snap-shot of general thinking
at the time. Archaeologists were attempting to move beyond
the discussion of artifact types, a strong point of contention
that marked the early stages of the culture history paradigm,
and trying to ultimately present the equivalent of ethnographic
Sections IV through IX represent the nuts and bolts of the
project and are primarily relevant to the prehistory of the
Mississippi Valley. Section IV pertains to the distribution of
Mississippi vessel shapes and features in which some of the
material presented originated from Phillips' earlier disserta-
tion research in the area, while Section V is Ford's seriation
analysis of the excavated pottery; their relationship to site
stratigraphy is presented in Section VI. These findings are
further related to the drainage history of the Mississippi in
Section VII, followed by an analysis of occupation site plans,
conical mound sites versus small ceremonial centers, for
example, in Section VIII. Section IX relates to the identifica-
tion of sites from documentary sources such as early travelers
who trekked through the area such as the Hernando de Soto,
Jolliet-Marquette, and later expeditions.
The summary and conclusions are discussed in Section X in
which the authors present descriptions of time periods and the
shortcomings of their findings. Ultimately, they conclude that
the origins of the Middle Mississippian is not in the Lower
Mississippi but suggest that influence may have originated in
the Southwest. These findings illustrate the split between the
beliefs of the authors regarding culture change. While Ford
held that the culture was continual, Phillips and Griffin
believed that it was a process of starts and stops. Evidence for
this, they believed, occurred in the distinctive qualities of
Mississippian and the preceding Hopewell cultures. While
they attribute Mississippian influence to the Southwest, they

suggest that archaeologists interested in the origins of
Hopewell should confine their search to eastern North Amer-
ica. Despite these shortcomings, Phillips, Ford, and Griffin
point out the contribution of the project to a previously
understudied geographic area.
In sum, Archaeological Survey in the Lower Mississippi
Alluvial Valley 1940-1947 is valuable for two major reasons.
First, it will be a useful resource to those interested in the
archaeology of the Mississippian heartland, and secondly as an
important historical document because it was written and
published during an interesting juncture of the culture history
paradigm. It was published just prior to the advent of radio-
carbon dating which would later make methods such as
seriation obsolete but at the same time it was progressive
because the authors were attempting to establish a chronology
on a regional scale as opposed to taking a site unit of analysis.
Considering its time and scope, this is a classic work that is
relevant to anyone interested in the prehistory of southeastern
North America.

References Cited

Dunnell, R.
1985 Archaeological Survey in the Lower Mississippi Alluvial
Valley, 1940-1947: A Landmark Study in American
Archaeology. American Antiquity 50:297-300.

Haag, W.G.
1953 Review ofArchaeological Survey in the Lower Mississippi
Alluvial Valley, 1940-1947, P. Phillips, J.A. Ford, and J.B.
Griffin. American Antiquity 18:275-277.

Williams, S.
1952 Review ofArchaeological Survey in the Lower Mississippi
Alluvial Valley, 1940-1947, P. Phillips, J.A. Ford, and J.B.
Griffin. Archaeology 5(2):124.

About the Authors:

Greg Hendrvx is a senior archaeologist with Environmental Services, Inc., in Jacksonville, Florida. He has worked as an
archaeologist throughout the southeastern United States since 1990 and for the past five years his research has focused on
Florida's prehistory.

Lee Hutchinson received her M.A. in 1990 from USF in Anthropology/Public Archaeology. She is currently a senior
archaeologist with Archaeological Consultants, Inc. (ACI), where she has been employed for almost 13 years. Lee has also
worked as an archaeologist for the Tampa Bay Regional Planning Council and Sarasota County. and through USF helped
manage "Le Project du Garbage" under the direction of William Rathje.

Rebecca Saunders is Curator of Anthropology at the Museum of Natural Science at Louisiana State University and an
Associate Professor with the Department of Geography and Anthropology at LSU. She is interested in coastal adaptations
and in functional and stylistic aspects of pottery production and exchange.

James C. Waggoner, Jr. received his Master's degree in Anthropology with a specialization in archaeology from Florida State
University in 2002. He is currently pursuing a Ph.D. at the University of Florida. His interests include the social organization
and land-use strategies of Archaic period hunter-gatherers in the interior coastal plain of southeastern North America.
Geographic Information Systems, and the history of southeastern archaeological thought..

Neill J. Wallis is a graduate student in anthropology at the University of Florida. He received a M.A. from the University
of Florida and a B.A. from Duke University. Having concluded his research of the Mayport Mound. he intends to pursue a
more expansive analysis of Swift Creek interactions in the Southeast.


2004 VOL. 57(4)


Non Profit Org.
Permit No. 911
Tallahassee, FL


Editor's Page

The Stratigraphic Sequence at Rollins Shell Ring: Implications for Ring Function.
Rebecca Saunders

Perpetuating Tradition on the Lower St Johns: Pottery Technology and Function
at the Mayport Mound (8DU96). Neill Wallis

The Honey Dripper Site (8NA910): A Late Swift Creek Encampment
in Northeastern Florida. Greg S. Hendryx


Little: Public Benefits ofArchaeology. Lee Hutchinson

Phillips, Ford and Griffin: Archaeological Survey in the Lower Mississippi Alluvial Valley,
1940-1947. James C. Waggoner, Jr.

Copyright 2004 by the
ISSN 0015-3893

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