The Florida anthropologist

Material Information

The Florida anthropologist
Abbreviated Title:
Fla. anthropol.
Florida Anthropological Society
Place of Publication:
Florida Anthropological Society.
Publication Date:
Quarterly[<Mar. 1975- >]
Two no. a year[ FORMER 1948-]
Physical Description:
v. : ill. ; 24 cm.


Subjects / Keywords:
Indians of North America -- Antiquities -- Periodicals -- Florida ( lcsh )
Antiquities -- Periodicals -- Florida ( lcsh )
serial ( sobekcm )
periodical ( marcgt )


Contains papers of the Annual Conference on Historic Site Archeology.
Dates or Sequential Designation:
v. 1- May 1948-

Record Information

Source Institution:
University of Florida
Holding Location:
Department of Special Collections and Area Studies, George A. Smathers Libraries, University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
01569447 ( OCLC )
56028409 ( LCCN )
0015-3893 ( ISSN )


This item has the following downloads:

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.








THE FLORIDA ANTHROPOLOGIST is published quarterly in March, June,
September, and December by the Florida Anthropological Society, Inc., do
Room 102, Florida State Museum, University of Florida, Gainesville, FL 32611.
Subscription is by membership in the Society for individuals or institutions in-
terested in the aims of the Society. Annual dues are $6. 00; student members
$4.00. Requests for memberships and general inquiries should be addressed to
the secretary; subscriptions, dues, changes of address and orders for back issues
to the treasurer; manuscripts for publication to the editor; and newsletter
items to the president. Second class postage paid at Gainesville, Florida.





March 1974

Additional Notes on the Philip Mound, Polk County, Florida
Karlis Karklins . ... .. ... .... .. 1

A Conecuh River Site (Cv. 30), Alabama by Robert J. Fornaro 9

A Mid-Eighteenth Century Indian Village on the Chattahoochee River
C. G. Holland . . . . . 31

Oven Hill (Di-15), A Refuge Site in the Suwannee River
Stephen J. Gluckman and Christopher S. Peebles ..... 21

Clay Balls: Ceremonial or Utilitarian by Jennins W. Bunn, Jr, 47


President John W. Griffin
P. O. Box 1987, St. Augustine, FL 32084

1st Vice President Benjamin I. Waller
4911 NE 7th St., Ocala, FL 32670

2nd Vice President Wilma B. Williams
2511 McKinley St., Hollywood, FL 33020

Secretary Robert H. Steinbach
P. O. Box 1987, St. Augustine, FL 32084

Treasurer Donald L. Crusoe
P.O. Box 2416, Tallahassee, FL 32304

Editor-Resident Agent Ripley P. Bullen
102 Florida State Museum, University
of Florida, Gainesville, FL 32611

Executive Committeemen

Three years: Wesley Coleman
Miami, Florida

Two years: J. Anthony Paredes
Tallahassee, Florida

One year: Yulee Lazarus
Ft. Walton Beach, Florida

At large, for one year

Arthur F. Dreves, Orlando

E. Thomas Hemmings, Gainesville

Roger Grange, Tampa



Karlis Karklins

The Philip burial mound is situated near the eastern shore of Lake Mar-
ion in Polk County, Florida. Results of extensive salvage operations carried
out at this site were reported in The Florida Anthropologist by Carl A. Benson
(1967). The artifacts recovered were of both native and European origin and
suggested a Safety Harbor period utilization with an overlap from the earlier
Englewood period, ca. 1600-1700, or slightly earlier (Benson 1967:130-131).
The author also conducted salvage operations before the mound was totally
destroyed by pot hunters. This produced a number of artifacts which are not
mentioned in Benson's report. The following describes these artifacts and
presents some previously unreported observations.

Mass Pottery Deposit

An undisturbed area about six by four feet near the western edge of the
mound yielded a large quantity of pot sherds, as well as the lip of a large conch
shell and fragments of an iron colander. The sherds appeared at a depth of
seven inches below the surface and continued down to the least 19 inches. The
exact vertical extent of the deposit could not be determined due to the fact that
this area was leveled by treasure hunters before excavation was completed. No
skeletal remains were found in direct association with the deposit. Apparently,
this pottery concentration represented a mass ceremonial offering, a feature
common to both Englewood and Safety Harbor period burial mounds (Willey 1949:

A number of vessels were partially or totally restored from the sherds.
All of these had been "killed". Some exhibited neat, round kill-holes made by
striking the interior surface of the base. Others had had their entire bottoms
knocked out. Belle Glade Plain pottery comprised the bulk of the deposit, al-
though St. Johns Plain and St. Johns Check Stamped pottery was also well rep-
resented. Several sherds of a Pinellas Incised vessel were also found.

Belle Glade Plain vessel shapes included: simple bowls with round bot-
toms (Fig. 1, a) or flat ones; flattened globular bowls (Fig. 1, b); collared
globular bowls; and small, deep bowls or pots (Fig. 1, c). St. Johns Check
Stamped sherds bore from four to nine square checks per linear inch. Par-
tially restored vessels consisted of simple, flat-bottomed bowls.

An unusual St. Johns Plain vessel consists of a flat-bottomed container

Florida Anthropologist, vol. 27, no. 1, March 1974


with everted sides (Fig. 1, e). Its lip is round and 5.0 mm thick. The exterior
is smoothed; the interior is brushed. The juncture of the sides and base is
sharp. The entire bottom had been knocked out. This vessel is 7.0 mm high;
diameter of the orifice is 12. 5-13. 7 cm.; that of the base 7. 5 cm.

The colander fragments (Fig. 2, b) mentioned previously have linear
series of perforations. Holes are 2.0 mm in diameter and are set 2.0-5.0 mm
apart. Some fragments have unperforated sections which represent the rim
area. The metal is about 2. 0 mm thick.

Artifacts Recovered from Disturbed Fill

In addition to the artifacts described by Benson (1967), those men-
tioned below were also recovered from the disturbed portion of the mound. In
the descriptions of glass beads, colors are designated using the names and
color codes in the Color Harmony Manual (Container Corporation of America
1958), followed by the equivalent code in the Munsell color notation system
(Munsell Color Company 1960). This is done to facilitate future intersite
comparisons of bead assemblages.


Pasco Complicated Stamped (1 specimen; Fig. 1, f). Paste is heavily
tempered with crushed limestone and fine sand. Surface bears a design con-
sisting of at least five raised, concentric circles with a raised dot in the cen-
ter. The sherd is 7.0 mm thick. Ocklawaha Incised (1 specimen; Fig. 1, g).
A thick (12. Omm.) sherd of soft St. Johns type paste. The exterior surface
exhibits several bold incised lines and remnants of red paint. Papys Bayou
Punctated (1 specimen; Fig. 1, h). The paste is soft and temperless. Decora-
tion consists of a shallow, linear punctate design. Sherd thickness is 4.0 mm.
Indian Pass Incised (?) (1 specimen; Fig. 1, i). The paste is hard and tem-
pered with fine sand. The sherd is 6. 5 mm thick. Unclassified sand-tempered
incised (3 specimens; Fig. 1, j-k).

Three specimens of what appear to be miniature vessels were recovered.
One specimen, a Belle Glade Plain simple bowl (Fig. 1, d), was totally recon-
structed. The lip is 8. 0-12. 0 mm thick. It is flattened and slopes down toward
the interior. The bottom is rounded and has a neatly made, circular kill-hole.
The bowl is 3. 5 cm high and has a diameter of 6. 6 cm. An even smaller simple
bowl (Fig. 2, d) is made of St. Johns Plain paste. This vessel is fragmentary,
but appears to have been ca. 1.7 cm high and ca. 3.4 cm in diameter.

The third specimen is a rim fragment from a Sarasota Incised vessel
(Fig. 2, e). The paste is relatively soft and temperless. The lip is round and
4. 0 mm thick. A portion of a horizontal projection is present on the rim just


below the lip and may be a handle lug. Decoration consists of down-pointing,
finely incised, punctate-filled triangles which begin at the middle of the lip and
extend down onto the exterior surface of the rim. The diameter of this vessel
is estimated to have been ca. 4. 6 cm. Height could not be determined.

Drawn Glass Beads
Tubular, decorated, round cross-section (1 specimen; Fig. 2, h).
Translucent deep blue (13 1/2 pc; 5PB 3/9) core with three opaque, vertical,
shadow blue (14 ie; 2. 5PB 5/4) stripes with a redwood (6 le; 7. 5R 4/6) stripe
in the center of each. Ends of the bead are unfinished (unaltered breaks).
Length: 12. 0 mm (fragmentary); diameter: 4.0 mm; perforation: 1.5 mm.
Tubular, decorated, round cross-section (1 specimen; Fig. 2, i).
Translucent turquoise green (20 nc; 5BG 5/8) core with three opaque, vertical,
redwood (6 le; 7. 5R 5/6) stripes with a narrow white stripe running down the
center of each. Ends of the bead are unfinished. Length: 5. 7 mm (fragmentary);
diameter: 6. 0 mm; perforation: 2.0 mm.
Tubular, brite navy (13 pg; 7. 5PB 3/4), square cross-section (6 speci-
mens; Fig. 2, j ). Transparent glass. Corners on four specimens have been
ground down. Other beads have sharp corners. Ends are unfinished. Length:
4. 0-7. 0 mm; diameter: 2. 4-3. 0 mm; perforation: 1.2-1.5 mm.
Tubular, clear, round cross-section (1 specimen; Fig. 2, k). This bead
consists of a very short section of glass tubing with rounded ends. Length: 3. 4
mm; diameter: 2.2 mm; perforation: 1.0 mm.
Hexagonal, scarlet (7pa; 5R 4/14) (2 specimens; Fig. 2, 1). Trans-
parent glass. Ends are formed by unaltered breaks. The unusual thing about
these beads, considering their size, is that the corners on both specimens have
been ground. Length: 1.4-1.5 mm; diameter: 1.8-2.0 mm; perforation: 0.7
Hexagonal, brite green (22 nc; 2. 5G 5/9) (1 specimen; Fig. 2, m).
Transparent glass. The ends of the bead are unfinished and the corners have
been ground. Length: 1.5 mm; diameter: 2. 1 mm; perforation: 0.8 mm.
Oblate to ovate spheroidal, decorated (6 specimens; Fig. 2, n). Trans-
parent brite navy (13 pg; 7. 5PB 3/4) core decorated with four straight, opaque
white stripes. Length: 3.0-8.2 mm; diameter: 5.6-7.8 mm; perforation: 1.4-
1.9 mm.
Spheroidal, decorated (1 specimen; Fig. 2, o). Transparent brite navy
(13 pg; 7. 5PB 3/4) core decorated with three redwood (6 le; 7. 5R 4/6) and three
white stripes that appear alternately around the bead. Length: 4. 9 mm; dia-
meter: 5.0 mm; perforation: 1.5. mm.
Oblate to ovate spheroidal, brite navy (13 pg; 7. 5PB 3/4) (4 specimens;
Fig. 2, p). Transparent glass. Each bead exhibits several irregular, randomly
applied, cut facets. Length: 2. 8-4.0 mm; diameter: 3.3-3.6 mm; perforation:
1.0-1.3 mm.

1 CM.

1 CM.


it I


o P *q

t 'u V

r O

1 CM.
i is --

Fig. 1. Ceramics from ,he Philip lound
a-d, Belle Galde Plain; e, St. Johns Plain; f,
Pasco Complicated Stamped; E, Ocklawaha Incised;
h, Papys Bayou Punctated; i, Indian Pass Incised(?);
j-k, unclassified sand-tempered.

Fig. 2. Miniature vessels and other artifacts
a, coral bead; b, iron colander; c, imitation
shark tooth pendant; e, Sarasota Incised;, f,
shell pendant; g, quartzite; h-v, glass beads;
w, cut crystal bead.






Wound Glass Beads

Elongate, rose wine (8 le; O1RP 4/6), faceted (1 specimen; Fig. 2, q).
Transparent glass. Eight pentagonal, pressed facets cover the surface.
These were applied with a small paddle while the glass was still viscid Ends
are square. Length: 6.4 mm; diameter: 5.2 mm; perforation: 1.9 mm.
Triangular to square, rose wine (8 le; 1ORP 4/6} (4 specimens;
Fig. 2, r). Transparent glass. These specimens have a triangular (3) or a
square (1) cross-section. They were shaped by pressing the surface with a
small paddle while the glass was still soft. Length: 2. 7-3. 7 mm; diameter:
3.0-3.6 mm; perforation: 1.31, 5 mm.
Oblate spheroidal, gilded (4 specimens; Fig. 2, s). Transparent
cinnamon (3 le; 1OYR 6/5) to lite gold (2 ie; 2. 5Y 7/7) glass core covered with
a thin gilt layer which is covered by a very thin layer of clear glass. Length:
4.2-5.6 mm; diameter: 5. 2-6. 1mm; perforation: 1.2-1.6 mm.

Blown Glass Bead

Globular, very pale blue (15 ca; 10B 9/3) (1 specimen; Fig. 2, t).
Transparent glass. This bead was made by blowing a bubble in a section of
drawn glass tubing. Ends of the bead flare out very slightly and exhibit un-
altered breaks. This suggests that this specimen may be a portion of a seg-
mented bead. Length: 3.8 mm; diameter: 3.7 mm; perforation: 1.8 mm.

Glass Beads of Uncertain Manufacture

Oblate spheroidal, decorated (1 specimen; Fig. 2, u). Opaque black
bead with a bold white line encircling the equator. Outline of the bead is very
symmetrical. Specimen is wound and/or mold pressed. Length: 5.9 mm; dia-
meter; 6.6 mm; perforation: 1.4 mm.
Oblate spheroidal, decorated (1 specimen; Fig. 2, v). This bead is
barrel-shaped and very heavily patinated. The original color could not be de-
termined. Five vertical ribs, three of which exhibit parallel, diagonal grooves,
decorate the bead. The glass exhibits swirls that are at right angles to the
axis of the perforation, suggesting that the bead was wound and then pressed in
a mold. Length: 5.2 mm; diameter: 5.8 mm; perforation: 1.0 mm.

Cut Crystal Beads

Oblate spheroidal, faceted (1 specimen). There are 32 facets on the
surface. Those encircling the perforation are irregular pentagons; the 24 facets
encircling the diameter are diamond-shaped. This bead is very battered. Length:
12. 2 mm; diameter: 14.0 mm; perforation: 2.4 mm.
Oblate spheroidal, faceted (1 specimen; Fig. 2, w). This specimen is
decorated with eight ovate facets with slightly concave surfaces that span the en-


tire length of the bead. The ridges at the juncture of adjacent facets bear
two small oval facets which are set about 3.0 mm. apart and encircle the
equator of the bead. The ends of the bead are flat. This specimen also ex-
hibits signs of wear. Length: 12. 0 mm; diameter: 15. 5 mm; perforation:
2. 2 mm.

Miscellaneous Beads

Cylindrical Coral Bead (Fig. 2, a). This unusual item was made from
a section of branch coral, probably Oculina sp. (ivory coral). The surface
has been smoothed by grinding down the protruding edges of the corallites.
The perforation is hourglass-shaped. Length: 23.0 mm; diameter: 14.0 mm.
Large Shell Beads (5 specimens). Shapes include oblate-spheroidal
(length: 10.0 mm; diameter: 12.0 mm); flat disc (length: 5.0-6.0 mm; dia-
meter: 10.0-12.0 mm); barrel-shaped (length: 8. 0 mm; diameter: 9.0 mm);
and tubular with a trianguloid cross-section (length: 12.0 mm; diameter: 8. 5
mm) .
Lead Bead (1 specimen). What appears to have been a large cylindrical
lead bead is represented by a very badly decomposed fragment.


Imitation Shark Tooth Pendant (Fig. 2, c). This artifact was manu-
factured from shell. It is triangular and has a 3.0 mm wide perforation in the
center of the slightly concave base. The other edges are sharp and slightly con-
vex. Length: 27.0 mm; width: 17.0 mm; thickness: 3.0 mm.
Shell Pendant (Fig. 2, f). This specimen is elongate-ovate in outline.
The suspension end is eroded but appears to have had a transverse groove.
Length: 41.0 mm; maximum width: 9.0 mm; thickness: 7.0 mm.
Quartzite Pebbles (Fig. 2, g). Four small, smooth quartzite pebbles
may have served as polishing or smoothing stones.
Red Ocher. This material is represented by six small pieces of ocher-
cemented sand.

Discussion and Conclusions

While the artifacts described on the previous pages do not alter the
date (ca. 1600-1700, or slightly earlier) assigned to the Philip Mound by Ben-
son (1967:130-131), some of the ceramics do substantiate a pre-1600 origin for
the site.

Papys Bayou Punctated pottery is diagnostic of both Weeden Island pe-
riods, especially the second (Willey 1949:443). However, it has been found to
occur in an Englewood, as well as a Safety Harbor context (Willey 1949:132,
153). In the latter instance, it was considered an heirloom. The Indian Pass


Incised sherd, if correctly identified, can be assigned to the Weeden Island II
period (Willey 1949:427) and the Pasco Complicated Stamped sherd can be at-
tributed to a late Weeden Island II period (R. P. Bullen: personal communica-
tion). The miniature vessels may also relate to the Weeden Island II period
since they are not, to the author' s knowledge, an acknowledged trait of either
the Englewood or Safety Harbor burial complex.

The Ocklawaha Incised sherd presents a problem. Goggin (1952:103) at-
tributes this type to his St. Johns Ia, early, period which dates around 150 B. C.
to A. D. This suggests that perhaps the sherd is mis-identified or, less likely,
that this type has a much greater temporal range than was previously believed.

Since the ceramics mentioned above were not found in situ at the Philip
Mound, it is not certain whether they represent a late Weeden Island II occupa-
tion or are heirlooms. However, their extreme scarcity suggests that they fall
into the latter category. The cut crystal beads may also be heirlooms, since their
battered condition suggests a long period of use. Because there is no way to de-
termine the longevity of heirloom pieces, in this case at least, it cannot be
stated with certainty when the mound was first utilized. However, if the mid-
point in the transition from Weeden Island to Safety Harbor occurred ca. 1500,
as proposed by Sears (1967:70), then a pre-1600 origin for the site is assured
by the Weeden Island II period ceramics. It does not seem very likely that heir-
loom pieces used in Safety Harbor period mortuary rites and various other com-
munity ceremonies (Sears 1967:69) would have remained unused or unbroken
until the 17th century.

The ceremonial deposit, on the other hand, can definitely be attributed
to the Safety Harbor period. The Pinellas Incised pottery found in the deposit is
a key type for this period (Willey 1949:475). The St. Johns Plain, St. Johns
Check Stamped, and Belle Glade Plain pottery forming the remainder of the
deposit also occurs in this context.

Extensive trade with various other cultural areas is indicated by the
ceramics: St. Johns series from the Northern St. Johns Region; Papys Bayou
Punctated, Pinellas Incised, Pasco Complicated Stamped, and Sarasota Incised
from the Central Gulf Coast/Manatee Region; and Belle Glade Plain from the
Glades Area. The coral and marine shell artifacts, of course, denote trade
with inhabitants of the Atlantic and/or Gulf Coasts.

The quartzite pebbles are also trade items. Although this stone is not
considered indigenous to Florida, R. O. Vernon, Chief of the Florida Bureau
of Geology, stated (personal communication) that quartzite pebbles do occur
along the highland sections and in the sand mines just west of Lake Okeechobee.
The pebbles may have been collected from these deposits. The other nearest
source of quartzite is Georgia. Therefore, these pebbles originated either from


the Glades Area or from the area to the north of present-day Florida.

Little, if anything, is known about the artifact assemblages and burial
customs at the Philip Mound. Unfortunately, due to the destruction of the
mound, this information will never be forthcoming.

References Cited

Benson, Carl A.
1967 The Philip Mound: A Historic Site. The Florida Anthropologist,
Vol. 20, Nos. 3-4, pp. 118-132. Tallahassee.

Container Corporation of America
1958 Color Harmony Manual, 4th edition. Chicago.

Goggin, John M.
1952 Space and Time Perspective in Northern St. Johns Archeology,
Florida. Yale University Publications in Anthropology, No. 47.
New Haven.

Munsell Color Company
1960 Munsell Book of Color, Pocket Edition. Baltimore.

Sears, William H.
1967 The Tierra Verde Burial Mound. The Florida Anthropologist,
Vol. 20, Nos. 1-2, pp. 25-73. Tallahassee.

Willey, Gordon R.
1949 Archeology of the Florida Gulf Coast. Smithsonian Miscellaneous
Collections, Vol. 113. Washington.

National Historic Sites Service
Ottawa, Ontario, Canada
March 1972


Robert J. Fornaro

In January of 1971 DePauw University tested a site alongside the banks
of the Conecuh River. The site (Cv. 30) lies about one-half mile west of the city
limits of Andalusia in Covington County, Alabama, sixty miles north of the
Florida state line. Though work on the site is by no means complete, this re-
port is designed to reveal what has been accomplished.

Covington County lies on the lower coastal plains of Alabama. Its soils
are predominately Norfolk sandy loam and require substantial amounts of rain-
fall and fertilizer for cultivation. Land elevation ranges from 160 to 335 feet
above sea level, with a topography of rolling hills. The Conecuh River, a non-
navigable stream, flows diagonally across the county, northeast to southwest.
Ultimately, the Conecuh River flows into the Escambia River, which in turn
empties into Pensacola Bay, Florida.

Approximately 67 per cent of Covington County can be classified as
woodland, with about one-half of the timber land in Longleaf-Slash pine and
one-half in Loblolly-Shortleaf pine. In addition to the pines, there are oak
and hickory trees. Game in the area includes deer, quail, dove, rabbit, opos-
sum, raccoon, squirrel and wild turkey. The climatic conditions are favorable
for general agriculture with an average growing season of 255 days. Average
rainfall for the area is 60.1 inches per year with a mean temperature of 65.5

The site area is bounded on three sides by the Conecuh River (Fig. 1).
This "loop" of land is subject to periodic flooding and the spring of 1970 found
three-fourths of the area under water. About one-third of the loop area is un-
der cultivation and it was in this portion that the test excavations were made.

Tests were run on the soil from one of the test pits to determine its
composition at natural levels. The percentage of sand, silt and clay was de-
termined by decanting off the clay. Then, the silt and sand were separated by
sifting them through a screen mesh. Since the clay had already been removed
from the samples it was reasoned that anything that would pass through a
screen with an opening of 1/256 of an inch would be silt. Having first weighed
the total sample and then weighing each of the separated elements it was poss-
ible to determine percentages for each of the elements in proportion to the
total sample. Acidity was tested by placing a piece of pH paper in a sample of
the test soil that had been moistened with distilled water. Results, given in
Table 1 clearly say that this type of soil composition is not conducive to the
preservation of bone or other forms of organic material.
Florida Anthropologist, vol. 27, no. 1, March 1974


*O.#A :Cv. V i. .

. -C .

: 4) .l' : -
C ,.
0 .05 .10.

4 : '' :'sP

&mi .-


- - -



f A-0
* *T


FIG. 1.

W 4


Table 1

Depths in inches

0-9 9-21 21-32 32-48 48-48

Sand 91.3% 94.8% 95.2% 95.6 96.1%
Silt 4.5% 4.2% 3.4% 3.0% 2.5%
Clay 4.0% 1.0% 1.3% 1.3% 1.3%
pH 4 5 5 4 4


Work began with an investigation of the cultivated surface area. A
variety of ceramic and non-ceramic artifacts were found, these having been
brought to the surface by the agitation of the plow. In all, several projectile
points, some quartzite blades and about three hundred sherds were collected
by members of the Russell family, owners of the property, and the DePauw
University crew. Of the three hundred sherds recovered, 97. 67% were classi-
fied as Weeden Island period (Willey 1949:396-452). The remaining specimens
were clearly from the Fort Walton period (Willey 1949:452-470).

Digging began in the area with the heaviest concentration of surface
specimens. A datom point was established and the site area was divided into
10-foot units. In all, twelve 10-foot units were excavated starting with unit
170-180E, 200-210N thereafter referred to as pit A (Fig. 2). As each pit was
opened it was given a letter designate in addition to its co-ordinates. In all
units there were heavy concentrations of material from the surface to ten
inches; i. e., the plow zone. The next heaviest concentration came at depths
of 17, the next at 25 and the last at 35 inches. Pottery fragments still appeared
at 39, but from 40 to 44 inches only worked stone and several projectile points
were found. Total sherd counts by units are indicated in Figure 2. It is possi-
ble that, due to the absence of pottery, the 40 to 44 inch zone represents an
Archaic level. Charcoal samples were taken at several levels, but as of this
date, no C-14 runs have been made on these samples.


Excavations revealed ceramic specimens (Figs. 3-4) from several
periods; i. e., Santa Rosa-Swift Creek, Weeden Island and Fort Walton. The
vast majority of specimens, however, were of the Weeden Island period. Def-
initions of pottery types follow Willey (1949) and Bullen, Partridge, and Harris
(1970). For a tabulation of types giving their quantities and vertical distributions,
see Table 2. The following descriptions refer to the various types as found at
the Conecuh River site.


Weeden Island Plain - Tempering ranges from fine-to-medium sized grains
of sand, quartz and mica. Rims are usually thickened with one or two incised
lines below the lip. Surface color varies - light-buff, red-buff, gray and mot-
tled black. Surface areas are usually well smoothed with some fire clouding.

Weeden Island Incised - Tempered with fine sand, quartz and mica grains.
Surface color ranges like Weeden Island Plain. Decoration is a combination of
incised lines of varying depths and widths made with a pointed instrument.
Most of the specimens of this type show light incising.

Weeden Island Punctated - Tempered with both fine and medium sand and quartz
grains, Surface color ranges like Weeden Island Plain. Decoration consists of
fine round or triangular punctations made before firing.

Plain Red - Tempered with fine quartz grains. Surface and interior colored
red; color applied before firing. Specimens fairly thin.

Savannah Fine Cord Marked - Tempered with fine-to-medium sand and quartz
grains. Surface color ranges like Weeden Island Plain. Decoration: fine to
medium cord markings.

Mound Field Net Marked - Tempered with medium-to-large sand grains. Sur-
face color light buff. Decoration consists of impressions made by a net.

West Florida Cord Marked Fabric Impressed -~- Tempered with large-to-me-
dium sand and quartz grains, Surface color light buff. Decoration consists of
impressions made with some coarse fabric, impressions not very clear.

Carrabelle Incised - Tempered with medium-to-fine sand and quartz grains.
Surface color buff, or gray with some fire clouding. Decoration: medium and
fine incised lines placed vertically or diagonally on vessel beneath rim. (See
Figure 3 and 7)

Carrabelle Punctated - Tempered with medium sand and quartz grains, Sur-
face color light buff, buff and gray. Decoration consists of triangular, round
and fingernail punctations, just below rim on rim pieces,

Wakulla Check Stamped - Tempered with medium-to-large sand and quartz
grains. Surface color usually light buff, some gray. Decoration consists of
fine to medium sized checks lightly impressed right up to vessel lip.

Kieth Incised - Tempered with medium-to-fine sand quartz grains. Surface
color buff and dark gray. Decoration: incised lines diagonally crossed to
form diamond pattern,


\*4 'a~B


16'- 11"

onto-" 0 MiNE ammou


21" 20

LU 0 L"

30"- 1H'

FIG. 4.

z u



Surface A B

C D E F G H I J K L Totals


Weeden Island Plain 280 140 297 185 109 172 176 166 100 127 102 127 61 2042 all

Weeden Island Incised 7 19 15 6 7 9 5 5 2 2 8 5 90 all

W. I. Punctated 5 14 2 4 1 7 6 1 5 6 3 54 all
Plain Red 2 1 3 zone

Savannah Fine Cord
Marked 2 4 13 4 7 6 7 7 5 2 57 all

Mound Field Net
Marked 2 1 3 1 1 1 9 all

West Florida Cord
Marked Fabric Impsd. 3 3 at 15"
25 to
Carrabelle Incised 7 1 2 1 2 1 3 3 2 1 1 24 surface
13" and
Carrabelle Punctated 1 4 5 1 1 1 1 14 25"
25" to
Wakulla Check Stamped 6 11 12 7 6. 2 3 10 5 8 12 8 6 96 surface
25" to
Kieth Incised 1 2 1 1 5 surface
Indian Pass Incised 2 2 zone
16" to
St. Johns Ck. Stmpd. 2 3 1 1 1 8 surface
plow z.
Ruskin Dentate Stmpd. 1 1 1 3 & 14"

Swift Creek 18" and
Complicated Stamped 1 3 3 1 8 27"
pl.z. to
Lake Jackson Plain 7 2 1 10 surface

305 175 374 218 136 195 207 198 113 146 128 155 78

Total Ceramic Specimens





1 B 11tri


FIG. 5.


Grooved S"tr.e.

Poh\v^P, Srtouts
CPo Po v

I &1Ae



S cr iper



Indian Pass Incised Tempered with medium-to-fine sand and quartz grains.
Surface color dark gray. Decoration: fine incised lines making curvilinear
loop s.

St. Johns Check Stamped 8 sherds. Tempered with medium-to-fine sand and
quartz grains. Furface color light buff and dark gray. Decoration: deep oblong
and square checks impressed on vessel exterior.

Ruskin Dentate Stamped 3 sherds. Tempered with fine sand and quartz grains.
Surface color light buff and buff. Decoration: rows of small tooth-edge denta-
tions arranged in irregular patterns.

Swift Creek Complicated Stamped, Late Variety 8 sherds. Tempered with
medium-to-fine grains of sand, quartz and mica. Surface color buff to dark
gray. Decoration: design curvilinear made by stamping design with carved

Lake Jackson Plain 10 sherds. Tempered with medium sand grains and
crushed clay. Surface color buff. Decoration: one rim sherd with pinched
or fluted exterior. One rim appendage with three small nodes.

Projectile Points

Projectile points (Fig. 5) were found at all levels. Bullen (1968) and
Cambron and Hulse (1969) were used in classifying specimens. While worked
stone, charcoal and projectile points were discovered as far down as 43, pot-
tery was not found below 39 inches. There was a dominance of stemmed forms
plus several small excurvate trianguloid points with concave bases. The last
were associated with the Weeden Island pottery in middle zones (Fig. 5).

Other Non-Ceramic Artifacts

Besides the projectile points several other stone implements (Fig. 6)
were recovered: 8 quartzite blades from various levels, 1 chert scraper at 13,
1 quartzite scraper at 39, 2 hematite grooved stones at 10, 1 hematite nutting
stone at 14, 2 pottery smoothing stones, one round at 12, and one cylindrical
at 20 inches below the surface.


In view of the ceramic and non-ceramic artifacts recovered, there is
little doubt but that Cv. 30 is predominantly a Weeden Island period site. Many
of the diagnostic pottery types mentioned in Willey (1949:397) are represented.
The depth and consistency of ceramic remains seems to indicate a fairly long
cultural association.


From 39, the deepest level for pottery, to about 25 inches below the
surface, pottery types seem to indicate an early Weeden Island association.
The recovery of eight Swift Creek Complicated Stamped specimens and the
relative absence of Wakulla Check Stamped and absence of St. Johns Check
Stamped specimens lends support to an early Weeden Island time period. How-
ever, it should be noted that the paucity of complicated stamped sherds seems
to exclude a "very" early Weeden Island association.

Above 25 inches to the plow zone, the ceramic evidence points to a
late Weeden Island period; i. e., the increasing occurrence of Wakulla Check
Stamped plus some St. Johns Check Stamped sherds. Sherds at these levels
were larger and more numerous. It was also in these levels that our only
identifiable vessel forms were discovered. The first vessel, or rather the
rim of a small Wakulla Check Stamped vessel, occurred at 13 inches. The ma-
jor portion of this vessel was probably destroyed and scattered by the plow.
The only other identifiable vessel was half of a Carrabelle Incised flared rim
vessel found at 14 inches (Fig. 7). It should also be noted that all 10 Lake
Jackson Plain sherds were in the plow zone.

Projectile points found in conjunction with the ceramic specimens in-
clude types like Jackson, Flint River Spike, Camp Creek, Bradley Spike, Co-
taco Creek, Mclntire, and all have woodland associations (Bullen 1968, Cam-
bron and Hulse 1969). Of thirty projectile points only five were chert, the rest
being quartzite. Below a depth of 40 inches, ceramic specimens ceased and
there was a dramatic increase in the occurrence of worked stone. On the basis
of the absence of pottery and the recovery of two projectile points, apparently
Archaic in type, it is possible that Cv. 30 includes an Archaic level. The lith-
ic inventory of the 30- to 41-inch zone (Fig. 5) strongly supports the presence
of an Archaic component.

In conclusion, Cv. 30 was in all probability a seasonal habitation site
used intermittently by the Weeden Island peoples.

References Cited

Bullen, Ripley P.
1968 A Guide to the Identification of Florida Projectile Points.
Florida State Museum. Gainesville.
1965 Florida' s Prehistory, Chapter XXIII, Florida from Indian Trail
to Space Age, Vol. 1, pp. 305-316, by Tebeau, Carson; Chauvin,
Bullen and Bullen. Southern Publishing Company. Delray Beach.
1956 Some Florida Radiocarbon Dates and Their Significance.
Florida Anthropologist, Vol. IX, No. 2, pp. 31-35.


Bullen, Ripley P., William L. Partridge, and Donald A. Harris
1970 The Safford Burial Mound, Tarpon Springs, Florida. Florida
Anthropologist, Vol. 23, No. 3, pp. 81-118. Gainesville.

Cambron, James W., and David C. Hulse
1969 Handbook of Alabama Archaeology, Part 1 Point Types,
The Archaeological Research Association of Alabama, Inc.
University of Alabama. Tuscaloosa.

Griffin, John W.
1949 The Florida Indian and His Neighbors. Inter-American Center,
Rollins College. Winter Park.

Merrill, J. M.
1967 Overall Economic Development Program Covington County
Alabama. Cooperative Extension Service, Auburn University.

Sears, William H.
1958 Burial Mounds on the Gulf Coastal Plain. American Antiquity,
Vol. 23, No. 3, January, 1958, pp. 274-284.
1959 Two Weeden Island Period Burial Mounds, Florida. Contribu-
tions of the Florida State Museum, Social Sciences No. 5.
Gaine sville.

Willey, Gordon R.
1949 Archeology of the Florida Gulf Coast. Smithsonian Miscella-
neous Collections, Vol. 113. Washington.

DePaux University
Greencastle, Indiana
April 1972


Stephen J. Gluckman and Christopher S. Peebles

In 1960, John M. Goggin defined four major types of underwater sites:
"(1) refuse sites; (2) submerged sites of former human occupation; (3) shrines
or places of offerings and interment; (4) shipwrecks" (Goggin 1960:351). The
following year Donald Jewell (1961:441) defined two broad categories of sites:
"those in which cultural materials have been introduced into existing bodies of
water, and those in which cultural sites have been inundated by submergence
or damming." Two years later Stephan de Borhegyi followed Goggin' s classi-
fication and also described four major types of subsurface sites (Holmquist
and Wheeler 1964:2). Frederic Dumas (1962), in Deep-Water Archaeology rec-
ognized only one type of site, the shipwreck. In Marine Archaeology, edited
by Joan du Plat Taylor (1965), none of the articles specify a classification of
underwater sites but the book is divided into sections which indicate a recogni-
tion of such a classification. The same can be said of George Bass' s (1966)
Archaeology Under Water.

This rather truncated listing of authors who have offered classifica-
tions of underwater sites is meant to serve a dual purpose: first, to show that
a variety of underwater sites has been recognized in print for over ten years;
and second, to introduce the point that definitive studies (if that term can be
applied to any underwater work) have been attempted only for shipwrecks. To
corroborate this last statement, one needs only to check quickly through the
literature. A bibliography which has been complied by one of us over the last
few years (Gluckman 1967 ms) shows over two-thirds of the citations to be con-
cerned in part or completely with shipwrecks.

In this paper we shall make an attempt to redress this imbalance by
defining and describing a very common type of underwater site which has re-
ceived little attention in the literature. We also hope to show, by using Oven
Hill as a specific example, that this typical Florida refuse site is representa-
tive, generally, of all refuse sites, and that lessons learned at Oven Hill al-
low the formation of work methods and of theoritical concepts which are ap-
plicable to all sites of this type.

The Refuse Site: A Definition

The refuse site as defined by Goggin (1960:351) refers to any archeo-
logical site formed either by the deliberate deposition of cultural refuse into
a body of water or by the unintentional deposition of such materials through
loss as at a watering place, anchorage, or landing. Goggin recognized three
Florida Anthropologist, vol. 27, no. 1, March 1974


varieties of this type of site: (1) Sites formed in bodies of water near habita-
tion sites by the deliberate throwing of trash and garbage into the water. (2)
Sites formed primarily by loss at watering places. (3) Sites formed by car-
rying refuse some distance from a living area before depositing it in a body
of water. To these we shall add a fourth variety, sites formed in areas com-
monly used as a ship' s anchorage.

The importance of the above distinctions may seem small, but they
point out physical differences which may have cultural significance. In vari-
eties one and three we are presented with two different patterns of garbage
disposal which could indicate an archaeologically exploitable cultural differ-
ence, or at least a dissimilar or variable water-use pattern. Varieties two
and four could help to define functional and environmental categories in re-
covered artifacts. That is, pottery recovered from a watering place might
pinpoint water vessel wares and types and indicate ceramic specialization,
while artifacts from an anchorage area might highlight differences and simi-
larities between shore use and shipboard use of common cultural items.

Goggin' s examples for varieties one and three--Oven Hill (Fig. 1),
where Seminole material was thrown into the Suwannee River directly in
front of the site, and Fig Springs (Co-17), a site just off the Ichetucknee River
in north central Florida, where spanish mission material was apparently car-
ried some distance before being introduced into the spring--represent con-
trasting cultural and temporal situations. Sites roughly comparable to Oven
Hill which have been examined along the Suwannee River, such as Pine Bluff
(Di-17), and the Lang Site (Di-62), present configurations much like that of
Oven Hill (Fig. 1).

Other mission sites like that which Goggin postulated for Fig Springs
are not known, but there are Spanish sites of approximately this period which
do show refuse sites in the immediate vicinity of the living area. Fort Matan-
zas, just south of St. Augustine, Florida, on the Matanzas River, and Wright' s
Landing, a plantation site north of St. Augustine on the Inland Waterway, might
be cited as cases in point.

Goggin' s example for variety two, the site built through loss at a wa-
tering place, is the cenote (Xlakah) at Dzibilchtdn, Yucatan. To the best of
our knowledge, no comparative analysis has been made of the pottery from
the cenote in relation to the ceramics from the remainder of the site. The
fourth variety of refuse site, the anchorage, could be exemplified from a
number of sites in the Mediterranean, and, in fact, has been recognized by
Dumas (1962:3), although his interpretation is slightly different than that pre-
sented here.

A site in the harbor of Louisbourg, Nova Scotia, Canada, which its


finders rightly referred to as the "Garbage Dump" (Hansen and Bleakney
1962:27), is such a site. Located in that part of the harbor identified as the
wharf area, the site presents a typical refuse site pattern. There is a broad
shallow scattering of broken ceramics, bottles, clay pipes, and ships' metal
work and fittings spread around the bottom in this specific area. A prelimi-
nary examination done by the laboratory staff at Louisbourg shows different
pottery type percentages from those of pottery recovered at the Fortress it-

The following examination of a specific refuse site in detail should add
depth and perspective to the discussion. The site (Di-15) to be discussed was
chosen for three reasons: (1) because it is typical of such sites in the Suwan-
nee River; (2) despite its having been worked extensively, no report has been
published; and (3) it is representative of the work done by the University of
Florida under John M. Goggin.

The Oven Hill Site

Oven Hill is located 40 miles west of Gainesville, Florida, a few miles
below (south of) Old Town, on the west bank of the Suwannee River (Fig. 1).
The land portion of the site is lightly overgrown in coarse grass and low
bushes with a few tall pines. It occupies a high bluff above the river which
forms its eastern boundary while highway Alternate 27 contains it on the west.
The site has been bulldozed or graded at least twice, the last time to provide
parking space at the now defunct barbeque stand which marks its southern li-
mits. The entire area occupied by the site is about 100 yards long, that is,
north and south, and 40-50 yards wide.

The underwater portion of the site slopes gradually down from the
river' s edge to an average depth of about 12 feet. This depth varies from
about 8 feet to 20 feet depending on rainfall. The bottom, though silty near
the bank, is mainly medium coarse to very coarse brownish to black sand
overlying grey clay which rests on limestone bedrock. All of the cultural ma-
terial lies in the sand layer. This layer rarely exceeds three to four inches
in depth except where the currents have gouged gentle longitudinal gullies
and banked the sand in one to three foot depths along their edges. This sand
extends perhaps ninety feet toward the center of the river at which point the
current is sufficiently strong to scour the bottom clean.

The water site extends the length of the land site and is marked by
European and Indian material as well as modern trash, including two row-
boats. Material is found from the shore line out to a distance of 40-60 feet.
The underwater visibility when good is about 10 feet; when bad, nonexistent.
This restricts work to the drier seasons of the year when the reduced silt
load of the river allowed the best visibility. The river bottom shows evidence


Old Town

Negro Towns

BowLeg'. Town
.c. lrsl

- - - -



* Archoological Sites

m ld Field
Are.. of Selimolo Settleme.e


1 2 3 4 Mi

C.S.1eable. ,S ot... 19s5


Indion Flold.


of disturbance and rearrangement due to the periodic flooding of the river
but apparently this has had little effect on distribution of the archaeological
material. There is a sterile area of river bottom between Oven Hill and the
Lang site which is immediately upstream, and the lower end of the water site
corresponds well with the southern end of the land site. This indicates surpris-
ingly little movement of the material by water and argues against contamina-
tion from other upstream sites.

The land part of Oven Hill was excavated by Goggin and students from
the University of Florida between January and June of 1958. The work was
done primarily on weekends and comprised fourteen days of excavation. Dur-
ing this time 10 trenches from 95 to 50 feet west of the river were played out,
and 75, 5- by 5-foot sections were excavated. The cultural deposit was
shallow and the excavation cards record no depth greater than 14 inches with
most sections having been excavated only to a depth of 8 inches. The under-
water work began prior to the completion of the land excavation and continued
intermittently until the summer of 1962. During this four year period the site
was visited and dives made on most weekends during the winter months. The
heaviest concentrations of work were in 1958 and in the fall of 1960.

The underwater work began as a simple collecting exercise with artifi-
cial controls imposed in the form of a three-part division of the site into up-
stream, center, and downstream portions. The early work, that done in 1958,
maintained these controls rather strictly until it had been determined that the
underwater material confirmed the surface findings that this was a single pe-
riod, single occupation site. This, of course, precluded the recovery of hori-
zontal stratigraphy and because of this, the division of the site was dropped
and it was treated as a unified whole.

The pattern of diving established at Oven Hill, though simple, was
effective and consistent. Divers entered the water at the downstream end of
the site, usually four at a time, using standard SCUBA gear, and swam slowly
upstream collecting as they went. The site was always worked against the
current to prevent divers from being swept into potentially dangerous ob-
stacles. The divers maintained sight contact and each sweep upstream over-
lapped the outside edge of the previous one. Each sweep took about thirty min-
utes and four overlapping dives were usually sufficient to cover the site.

By late 1960, material had become quite scarce. Each series of dives
produced only a handful of sherds and most of the University of Florida' s un-
derwater activities were transferred to other sites on the Suwannee, Santa
Fe, and Ichetucknee Rivers. That Oven Hill became for all practical purposes,
worked out was to be expected. This is a characteristic of refuse sites which
are not being constantly renewed by redeposition from water destruction of ad-
joining land sites. The land deposit at Oven Hill was too shallow and too far


above the river, about 12 feet, to allow river wash to account for much of the
underwater material. Hence, Oven Hill was closed down until some months
later when it became the site for the first tests of the airlift Goggin designed.

In 1960, the Department of Anthropology at the University of Florida
purchased a set of 26-foot steel pontoons. When decked over, the resultant
craft was a shallow draft barge with an open, 24- by 8-foot deck. This diving
platform was powered by an 18 horsepower outboard motor and was designed
specifically to provide a floating work surface from which to operate a five-
inch airlift.

This rig was first tested at Oven Hill, and though the airlift worked
well in water as shallow as 8-10 feet, it proved to be of little practical use
at this site. The airlift moved sand in gratifying amounts, and it was felt that
small modifications would make it an excavating tool of some value when ex-
treme care was exercised in its use. However, the results at Oven Hill were
disappointing as little was recovered while the amount of time spent uncover-
ed far fewer specimens than the river uncovered in the week which had passed
between diving visits to the site. It soon became obvious that the material
which remained following the years of extensive collecting was both scarce
and scattered, and that it was far easier and more productive to let it be un-
covered by the river.

The experimentation with the airlift marked the termination of the
work program at Oven Hill. Dives were made periodically at the site until
1962, but these visits were only a part of the general diving program in the
Suwannee River which called for regular visits to all of the refuse sites in
the river to collect the material which had been uncovered by the river since
the last visit.

Specimens from Oven Hill

Oven Hill is the oldest and most spectacular Seminole site yet excava-
ted in Florida. Literally everything that a Seminole household might dispose of
during a relatively short period of occupancy was found. For convenience we
will deal with the river and land portions of the site as separate units; but there
is no difficulty in interdigitating them. The few broken European vessels which
were found in the river match specific sherds fromlthe land portion of the site.
The same is also true of the many aboriginal vessels which came from the riv-
er and whose restoration was completed with pieces which came from the land.

The River Section Oven Hill yielded over 35 whole and partial vessels of ab-
original manufacture, plus enough large sherds to keep a jigsaw puzzle fiend
happy for months. With the exception of a few museum pieces, these vessels
are the only whole examples of Seminole ceramic art in existence. These ves-


sels which are from a "pure" or single component site), have vastly increas-
ed the known range of avriability of Seminole ceramics.

Seminole brused ware in the past has been divided into two types on the
basis of whether the vessel was limestone or sand tempered (Goggin 1958). At
Oven Hill both "types" are present, and with the single exception of temper-
ing material, both are identical. The microstylistic elements of decoration
and manufacturing techniques are such that the conclusion that the pots are
made by the same or closely related persons is inescapable. Possibly, in this
context, preference for crushed limestone or sand as temper may be a per-
sonal matter.

European ceramics, which constitute less than 5 per cent of the ce-
ramic assemblage, consists of two small "Lambeth Delft" bowls and one
small red paste lead glazed bowl. The land portion of the site yielded some of
the remaining pieces of these vessels plus one small sherd of Ksang Hsi blue-
on-white porcelain. Other than pottery the only other material recovered from
the river was one trigger guard, one well rusted stirup and one rifle grenade.

The Land Portion supplied gun parts representing at least four weapons. Al-
though these parts are not interchangeable they are identical in form. Axes,
razors, and knives, like the guns, present an amazing degree of internal homo-
geneity. Also recovered from the site was a pair of spurs, and a snaffle bridle.
Other items recovered were worked brass wire segments, buckles, gift trunk
hinges, and several buttons. These buttons have been analyzed for us by Dr.
Lewis Binford, who has done part of the analysis of the Fort Michilimackinac
(Canada) material. One large button is a British military legging button and
the others are military coat buttons. These buttons had an official life span
among the British military in America from 1764-1768. There were also nu-
merous items of personal adornment: silver cones, a silver maltese cross,
ear rings, and a "stone ring". Also represented on the site are large num-
bers of both black and white barleycorn beads and the Oven Hill Chevron Bead.
A few pipe fragments were recovered from Oven Hill. Again, through the good
offices of Lewis Binford, these pipes have been compared to the Fort Michili-
mackinac material. With the exception of one bowl whose mark can not be
clearly read, all of the pipes are marked "T D". However, there are many
different forms of "T D" markings. Of those examined, one, according to
Binford, dates immediately post-1765, which ties in nicely with the buttons.

In addition to specific dating of single artifacts, the Oven Hill assem-
blage has been compared to the earliest gift lists kept by the British relative
to the Seminoles in Florida (Gage papers, Stuart to Gage 7-19-1764, and Cov-
ington 1960). There is an amazing correspondence between the types and des-
criptions of goods contained in these lists and the material recovered at Oven
Hill. The homogeneity of various artifact classes at Oven Hill is absent from


other British period sites in Florida. For instance, the gun parts from Oven
Hill have been compared with those recovered by Goggin (1949) at Spalding' s
Lower Store (1763-1783) on the St. Johns River. The diversity of the store
parts, while including some like those recovered at Oven Hill, is far broader
than those at Oven Hill. Both these facts argue for a short and early occupa-
tion of Oven Hill.

Probably the Oven Hill house was one from a village which Pittman wrote
about in 1767 (Gage papers, 8-5-1767). It seems certain that it is not part of
either Bowlegs or the Negro towns of ca. 1814-1818 (Young 1934-35), which,
as can be seen from the map (Fig. 1), are in the immediate area. The only
other possibility is that it may be a part of Talahosochte which the naturalist
William Bartram (1943, 1958) visited in 1744, but, at present, we think this

The dating of the site is only one answer to the many problems which
Oven Hill poses. The underwater portion of the site has given us many Seminole
ceramic vessels. But these are only a starting point toward a micro-stylistic
or structural analysis of Seminole ceramics. The many European goods pose
questions of how gifts given in St. Augustine by the British to Seminole leaders
were distributed throughout the various villages and family units in the coun-
tryside. General ecological comparisons of the villages which existed in this
area at various times between 1767 and 1840 must also be made and analyzed.
The latter two questions are at present being worked on by one of the authors
using ethnohistorical sources. But to answer the first we are going to have to
get back to the water again.

The other sites noted on the composite map of the area have been ex-
plored underwater in a preliminary way with suggestive results. Di-62, the
Lang site, has yielded two whole ceramic vessels which are copies of a Euro-
pean brass kettle and an iron tripod pot. We suspect that this is one of the
houses from the Negro town of 1814-1818. The remaining sites along the river
have each yielded small assemblages of ceramic material; and each group,
while falling into broad types, has its own distinctive internal structure--its
own micro-stylistic elements. There is enough work in this area to keep us
busy for quite some time to come.


Underwater work at Oven Hill can be both summed up and characterized
as intensive collecting carried out over a number of years. Although an airlift
was used, this was experimental and not part of the comprehensive program of
underwater investigation of the site. The use of the airlift showed that excava-
tion of this site after extensive collecting was not practical because of the mea-
ger returns from such a shallow deposit.


Work at this particular site brought into sharp focus elements which
are common to all refuse sites. These common denominators suggest proce-
dures which are basic enough to be applicable to all refuse sites:

(1) Refuse sites are formed in a manner identical with the surface midden
with one important distinction--no one ever lived on a refuse site. This im-
plies that deposition and therefore context will be accidental, fortuitous, or
random, and this distinction makes all associations suspect in the sense that
only the overall pattern and not the specific instance need have cultural signi-

(2) Most refuse sites will be related to land sites and should be worked
and interpreted with this in mind. As a rule of thumb, any surface site near
a suitable body of water may indicate the presence of a refuse site.

(3) Refuse sites are probably best worked by the application of stringent
surface collecting techniques. Exceptions to this might be any multi-compo-
nent or multi-period site where the presence of recoverable stratigraphy is

(4) Given the present state of underwater excavating equipment and con-
trols, excavation of a refuse site is not recommended if the site is allied with
a surface site. A land excavation should prove easier and more profitable of
data. This stricture must be immediately amended to admit the very real need
for the complete excavation of a number of sites of this type to check, among
other things, the validity of the hypotheses presented here.

(5) The material from a refuse site, as Goggin (1960:351) has pointed out,
may be difficult to handle. Where artifact typologies are well-known or in an
instance like that at Oven Hill where older material showed more wear, these
difficulties are only slightly greater than for equivalent surface sites. As most
refuse sites are formed in relation to land habitation sites, there should usu-
ally be no need to interpret them in an underwater vacuum, and standard ar-
cheological procedures should be applied.

The function of an archaeologist is the collection and interpretation of
data about man' s past. Underwater archaeology is a technique which adds a
dimension to the archaeologist' s ability to gather information. It is hoped that
the proceeding has contributed information about a refuse type of site and has
suggested a framework within which the refuse site can itself be used empiri-


References Cited

Bartram, William
1943 Travels in Georgia and Florida, 1773-74: a report to Dr. John
Fothergill (annotated by Francis Harper). Transactions of the
American Philosophical Society, 33:121-242.
1958 The Travels of William Bartram (edited by Francis Harper).
Yale University Press. New Haven.

Bass, George

Archaeology Under Water. Frederick A. Praeger. New York.

Covington, James
1960 English Gifts to the Indians, 1765-1766. Florida Anthropologist,

Dumas, Frederic
1962 Deep-Water Archaeology. Rutledge and Kegan Paul. London.

Gluckman, Stephen J.
1967 Underwater Archeology: Method and Theory. (unpublished M. A.
thesis) University of Florida Library. Gainesville.

Goggin, John M.
1949 A Florida Indian Trading Post, Circa 1763-1784. Southern
Indian Studies, 1:35-37. Chapel Hill.
1958 Seminole Pottery. In Prehistoric Pottery of the Eastern
United States, edited by James B. Griffin. Museum of Anthro-
pology, University of Michigan. Ann Arbor.

Hansen and Bleakney
1962 Underwater Survey of Louisburg Harbor for Relics of the
Siege of 1758. Acadia University Institute. Wolfville,
Nova Scotia.

Washington, D. C.
May 20, 1972


C. G. Holland

In June 1965 the Georgia Kraft Company appropriated money to the
Smithsonian Institution for salvage archaeology at their new plant near Cot-
tonton, Alabama, a little north of the State of Florida line. Indian sites had
previously been discovered at this location by Harold A. Huscher (1959) of
the River Basins Survey, Smithsonian Institution. It was anticipated that Hus-
cher would conduct the field activities, but after he went on sick leave Dr.
Clifford Evans of the Smithsonian Institution asked me, as a Collaborator of
the Smithsonian Institution, if I would take Huscher' s place. These matters
delayed the start of field work.

Construction in the plant area was far advanced when I first- visited
the property on September 21, 1965. The plant site, approximately a mile
west of the river had been bulldozed level, truck traffic was heavy, and the
approach railroad had been laid. A field survey of this and adjacent de-
forested areas revealed no signs of Indian occupation but it was apparent from
maps that three of Huscher' s sites would be destroyed by the construction of
settling basins.

Figure 1 presents a sketch map of the area. The region between
Route 165 and the river was featureless and covered with "dog fennel" which
grew in thick sets up to eight feet high. Several small tests were dug which
produced an occasional nail, a few sherds, some chips, and, in one instance,
a projectile point, 2 sherds and 3 chert flakes in the upper 6 inches. Finally
a careful survey was made along the river bank from the northern line of the
Georgia Kraft property to the railroad track or end of the access road. Sherds
concentrated on what appeared to be a remnant of a natural river terrace a
little north northeast of the railroad bridge (Fig. 1, site) .

In this location we excavated twenty 10- by 10-foot squares to form
three trenches as indicated in Figure 2. The upper 6 inches, containing most
specimens, was stripped off first after which excavation continued to a depth of
12 inches. In each trench, one 10-foot square was carried down to the under-
lying red clay without finding any cultural debris below a depth of 12 inches.
The proveniences of aboriginal pottery by types is given in Table 1 by trenches
and squares, numbered as in Figure 2. Material from the 6- to 12-inch levels
was so scarce that number 21 in Table 1 combines sherds from that level of
Squares 11, 12, and 17; number 22 combines from those Squares 13, 14, 15,
and 16; and 23 those from Squares 18 and 19.

Florida Anthropologist, vol. 27, no. 1, March 1974

*-, /


I r, t~

C /

I, *I -





\ -**-

Fig. 1. Site location map.

I ') I-





1129 FEET TO



4 20 3 2 1

143 FEET /

,/4_F ENGE


84 FEET -


939 FEET


12 I




15 16

Fig. 2. Excavation plan.











A total of four features were found. In Square 5, at a depth of 2 inches,
a rusty-red, hard-baked clay circle, 2 1/2 feet in diameter, was encountered
in the southwest corner (Fig. 2, circle in Sq. 5). It had two shallow, parallel
ruts across it, running with the axis of the trench and 6 inches apart (plow
lines?). Scattered on and around this feature were glass, pipe stems, aborig-
inal pottery, bone fragments, and flecks of charcoal. On excavation the surface
was 2 inches thick, beneath was a dark stained sand layer, then a second clay
layer, another sand layer, and finally a third clay layer, to a total depth of 18
inches. This feature was interpreted as a fireplace and a 10-foot lateral exten-
sion of the trench (Fig. 2, 6) was excavated with the hope of finding a house
pattern but no post molds or other suggestions of structure were encountered.

Two charcoal stains ran from the fireplace in Square 5 to the north-
east. One was fragmentary and faded out within 2 feet. The other, more signi-
ficant, was dark stained, had parallel sides 2 1/2 3 inches wide, and was dis-
cernable at a depth of 12 inches. Towards the northeast at 26 1/2 foot inter-
vals were 4 1/2-inch square post molds extending 6 inches deeper into the soil.
This line of post molds ran through Squares 7, 8, and 9 and was followed 57 1/2
feet beyond Square 9 as indicated in Figure 2 by a dashed line. It was also acci-
dentially discovered 2 inches west of Square 17. As shown in Figure 2 this fea-
ture bends towards the river and was over 1000 feet long. The fact that the fire-
place had been built into this feature indicates the feature is older than the fire-
place. Its fence-like quality and length suggests some type of fragile fortifica-
tion or palisade but its exact nature is unknown.

The third and fourth features consisted of extremely hard, flat, sandy
areas flecked with charcoal. In Squares 18 and 19, the feature did not have sharp
borders but faded out within the squares. In Squares 13-16, this compact zone
with charcoal flecks covered the entire area of that lateral trench. Only here in
the southern trench and in Square 5 was there any suggestion of house floors or
heavily travelled passageways.


Our excavation produced both aboriginal and European (Colonial) arti-
facts. These are briefly described below and the best examples illustrated in
Figures 7-8. Mr. Ivor Noel Hume of Colonial Williamsburg very kindly analyzed
the colonial pottery and Mr. Roy Huntington of Charlottesville, Virginia, the ma-
terials relating to firearms. Aboriginal pottery, except that found in part of
Square 5 where it was protected to some extent from mechanical damage by the
fireplace, was represented by rather small sherds. Most were smaller than an
inch across (2.0 by 2.0 cm) and many were in the 0. 5 by 0.5 cm range. This
suggests a concentrated occupation resulting in much trampling and resultant
miniturization of sherds.


Table 1 Distribution of aboriginal pottery at Georgia Kraft site

Proveni- Type Type Type Type Type Type Type Type Eroded Bits/ Unclass-
ence I II III IV V VI VII VIII Sherds Pieces ified

1 105 165 51 5 6 31 43
S2 140 147 44 1 1 4 134 57
3 117 152 41 2 4 126 54
S20 0-6" 88 74 25 2 7 112 16
S20 6-12" 30 28 1 146 11
4 63 74 18 1 8 68 13
10 148 80 41 4 6 28
9 247 105 94 8 1 81 105 35
S8 0-6" 297 181 94 1 20 99
S8 6-12" 44 37 3 2 18 5 5
_7 211 118 90 5 472 234 31
S6 329 63 129 14 895 157 15
5 210 84 142 11 56
F-5 77 40 91 9
11 236 239 159 15 52 5 8 215 82
$ 12 195 225 112 11 3 26 12 161 85
m17 262 336 101 18 1 2 1 16 290 81
S21a612" 184 214 97 7 11 37 1 303 23

13 309 376 148 21 1 1 14 14 314 72
14 264 296 95 21 3 1 1 5 263 73
S15 346 415 206 20 1 1 14 1 15 207 75
16 250 347 105 14 3 2 5 2 6 122 46
22b6-12" 13 4 4 4 2

18 265 212 145 14 7 1 3 258 38
o 19 277 306 116 16 2 6 262 101
2 23c6-12" 119 118 43 6 1 9 15 1 4 52 22

Totals 4826 4436 2164 219 11 32 176 17 1606 3692 1045
aLower levels of Squares 11, 12, and 17. bLower levels of Squares 13-16.
CLower levels of Squares 18-19.

Aboriginal Pottery

I have divided the aboriginal pottery into eight provisional types based
on a modal analysis which is abridged below. A more detailed account is in-
cluded in the site report on file with the Department of Anthropology, Smith-
sonian Institution. Correlated pottery types follow Willey and Sears (1952) re-
port on the Kashita site in the same region.


Type I~ Temper includes white to orange quartz sand, 1~*3 mm across, which
roughens the surface and forms elongated furrows on exteriors where dragged
by finishing techniques. Temper may constitute as high as 50% of the paste.
Many sherds sparkle from minutes particles of mica that was not an intentional
ingredient. Fire clouds are prominent and color varies from black through buff
to reddish orange. Sherds are hard, 2.5#3.5 on Moh's scale. Exteriors are
swept with a stiff brush leaving fine to 1 mm deep striations (Fig. 7, b). In-
teriors are smoother. Rims are illustrated in Figure 3. Pots have globular
bodies, 20+34 cm in diameter, and are sometimes decorated with rim strips,
4 to 9mm wide, added 4 to 10 mm below lips. These are finger pinched or ime
pressed with a blunt stick to form a wavy line (Fig. 7, a). Minor decorations
consist of square or diamondeshaped punctations on lip edges or nicking at 4 to
6 mm intervals where fillets occur. Identification: Chattahoochee Brushed
(Willey and Sears 1952:5#7; see also Bullen 1950 and Goggin 1953).

[These rim strips are similar to those found on Jefferson Ware (Smith ]1948:P1, 32)]

Type II - Paste and color are the same as for Type I. Exteriors and interiors
have been rubbed smooth, even to the point of burnishing, but both surfaces are
bumpy and speckled by the coarse sand when it is at the surface. Rims (Fig. 4)
occasionally have a stepped segment. Vessels are bowl shaped, 12-48 cm in
diameter, and decorated near the rims with incised designs formed of 1-3 mm
wide lines (Figs. 4; 7, c). Occasionally rows of stabs, linear or square in shape,
are found on rims or lips. Except for the incised decoration these sherds are

like the ''coarse plain" listed by Willey and Sears (1957:7~8)}. [The paste is the
same as that of the Lake Jackson series while the incision is reminescent of
Aucilla Incised, Ed. ]

Type Ill + Temper is either silt or very fine sand with grains less than 1 mm

in size. Clay and temper are well mixed and impossible to differentiate. Color
is black to light grey with occasionally a buff or reddish orange sherd. Vessels
are incompletely fired at relatively low temperatures but firing is more uniform
than for Types land II] Hardness is 2.5=3.5. Surfaces have a smooth, almost
slick surface without tool marks. Rims, lips, and decoration are shown in Figure
5. Sherds indicate open mouthed bowls with slightly everted, occasionally in-
verted, rims and mouth diameters clustering at 20-28 and at 38-48 cm. On
upper bodies are incised nested arcs and parallel straight lines, often made be-
fore the final smoothing process so that lines are partly obliterated. Interrupted
lines are more common than continuous ones. Occasionally there are rows of
punctations 4-5 mm apart. Included in the collection are parts of small loop
handles. These vessels (Fig. 7, a) probably are the same as Ocmulgee Fields
Incised (Willey and Sears 1957:8).

Type IV = No obvious temper is present for the majority of Type IV sherds al-
though extremely fine sand may be noted in some cases. Most sherds are tan
to buff on interior and grey to black on exterior surfaces, are incompletely


Fig. 3. Type I rim profiles
(interiors to left).

-( D lD r

Fig. 4. Type II rim profiles and decoration.





/ Fig. 5. Type III rim profiles and decoration.

oxidized, and have a hardness of 3. 0. Surfaces are uniformly smooth but not
burnished. Most lips are rounded as shown in Figure 6. Open mouth, shallow
bowls cover most containers, with diameters clustering at 4-18 and 26-30 cm,
but one sherd suggests a bottle and another a plate. All lips exhibit a red film
which, in some cases, extends downward on the inside for about 2 cm. An
occasional sherd will have a red band on both inner and outer surfaces parallel



Fir.. 6. Type IV rim profiles and decoration.

to the lip. More commonly, as shown by stippled areas in Figure 6, the red
film occurs in areas or zones which may be outlined by incision (Fig. 7, e).
Sometimes the original color has been oxidized to a dull brown. One strap
handle, 2. 2 cm wide with 3 longitudinal grooves and a loop handle, 1. 1 cm
wide, were found. These sherds can be identified with Kashita Red Filmed
(Willey and Sears 1956:7). [In Florida they would be called Mission Red Filmed,

Type V Sherds of this type have paste characteristics identical to that of Types
III and IV but have been separated out because they have a thin, fragmentary
black coating on their exteriors (Fig. 7, g), which in several cases was arrange-
ed in bands. These sherds have an almost glazed appearance. They do not seem
to fit any formally named type.

Type VI Clay has been mixed with vegetal matter which burned out during
firing to leave holes of irregular size and spacing. Texture is clayey flakey and
sherds are relatively light in weight. Exteriors are light tan to orange red, cores
and interior surfaces grey. In a few sherds, cores and inner surfaces have the
same color as outer surfaces. Oxidation is incomplete and hardness only 2.5,
the lowest of any type. Surfaces are smooth but irregular due to temper holes
and flaking of the paste (Fig. 7, k). Wall thickness vary from 4 to 10 mm.
Size and shape are unknown. Sherds must belong to the Wheeler, Stallings, or
Orange Series (Sears and Griffin 1950:1-1, 5-1, 7-1. [As the author does not
mention sand inclusions, they probably are classifiable as Orange Plain, Ed.]

Type VII These sherds, tempered with fine sand, are incompletely fired in
a reducing atmosphere and sand grains rub off easily giving the pottery a "sandy


feel. They are also easily eroded. Surface color is grey to rusty brown while
cores are grey to black. Hardness is 3. 0 and surface smooth, plain, and with-
out tool marks. Rims appear to be straight but one is thickened below the lip.
The two lips available are both tapered. Body wall are 7-8 mm thick and appear
to pertain to conoidal pots with straight sides. Loop handles are present.
Correlation with a formally defined type is uncertain but Bullen thinks these
sherds probably represent Miller Plain (Smith

Type VIII Eighteen sand-tempered sherds were paddle stamped but designs
were almost completely unrecognizable because of small size. What clues there
were suggested both linear and filfot designs (Fig. 7, f).

Other Aboriginal Artifacts

Flakes Four quartz, 1 quartzite, and 317 chert flakes were found. In Square
20, where substantial depth of artifacts was found, 24 were in the highest 6
inches, 76 in the 6- to 12-inch zone, and 14 in the 12- to 24-inch zone indicat-
ing chips to have a deeper average provenience than pottery.

Projectile Points Two projectile points, one of tan chert and one of rhyolite
were collected. The first was broken, the second 2 1/2 inches long. Both were
stemmed with parallel-sided tangs 1/2 inch long, one with a fairly straight and
the other with a rounded base.

Pottery Disc One disc, 11/16 by 13/16 inches with rounded edges and corners
was found in Square 24.

Spoon Handles Of two, one in Square 13 was of Type II pottery, the other
(Fig. 7, j) from Square 8 was of Type III ceramics and was attached to part of
a bowl.

Clay pipe A bowl fragment, 1 1/2 inches high, had a smooth exterior spoiled
by many spall scars.

Colonial Artifacts

Iron specimens, all heavily rusted, included 4 "square headed" spikes
with 1/4-inch square bodies, 4 inches long; 25 "squared headed" nails, from 2
to 3 inches in length (Fig. 8, a-d); and a 3/4 inch long tack. A fizzen spring
(Fig. 7, m) from a military musket lock (French) was associated with the pit
in Square 5. Three knife blades and 18 miscellaneous fragments complete items
made of iron.

Lead was represented by 5 musket balls: 3 flattened after firing and
weighting 17.3, 12.0, and 12.7 grams; 1 unfired and weighing 14.5, probably




Fig. 7. Specimens of clay, glass, metal, and stone.
a, Type I rim with pinched decoration; b, Chattahoochee Brushed; a, Type II
decorated rim; d, Ocmulgee Fields Incised; e, Kashita Red Filmed; f, Type
VIII or Jefferson Complicated Stamped; g, Tyle V, black filmed; h, Delftware
with cobalt decoration; i, Yorktown glazed earthernware; J, spoon-shaped
handle, Type 3 paste; k, fiber-tempered rim sherd; 1, glass bead; A, fizzen
spring; n-o, brass cones; p, white metal button; q, chewed musket ball; r-s,
gun flints.


P q



e g

c c 9
Fig. 8. Other metal specimens.
b^ a, d, iron spikes; b-c, "square-headed"
b nails; e, brass bracelet; f, silver
bandle; g, musket sling ring, brass.

55 caliber; and 1 chewed, weighing 16. 2 grams. The last (Fig. 7, q) may have
been used by a person in pain such as during surgery. Also present were 9 flat
pieces of lead weighing up to 10. 6 grams. One was square but folded. These
fragments could have been used to seat gun flints in their locks.

Present also were fragments of sheet brass, a ring 15/16 inches in
diameter (Fig. 8, g), a wire type bracelet (Fig. 8, e), a bar 3/16 inch thick
and 1/2 inch long, 4 cones 13/16 inch long (Fig. 7, n-o), the dragon side plate
of a trade musket, a spring or trigger guard fragment, a tapered fragment
broken across a hole in one end, and 12 miscellaneous fragments.

Seven buttons were found, all approximate an inch in diameter. Six
were plain, flat, and made of white metal. Only one had a mold seam. Five
had iron and the sixth a brass eye (Fig. 7, p). The seventh button had a tudor
rose face, was convex, not spun, and made of brass with a brass shank. This
button with shank and face as a single cast and a drilled shank has not been
dated. Five of the other buttons date to 1726-76 by South' s (1964) analysis.
Four of these had a collar of metal on each side of the wire eye, the result of
the molten metal conforming to the shape of the eye in casting.

Four gun flints--2 grey, 1 waxy tan, and 1 waxy greyish tan--include 1
commercially made with a French rounded heel and 1 home made but also show-
ing a rounded heel (Fig. 7, n-s).


Kaolin pipes were represented by 108 stem and 77 bowl fragments. Of
measurable holes 25 were 4/16, 73 were 5/16, and 15 were 6/16 inch in dia-
meter which by Harrington' s graph date in the 1710-50 range, by Binford' s
formula the date comes to 1743. 61 A. D.

Colonial ceramics include 8 examples of Yorktown stoneware (1730-
1750), 1 of a ribbed Staffordshire mug, (1710-30), 1 of a Staffordshire red-
bodied ware with lead glaze of Astbury or Bell type (1735-55), 1 white salt
glaze stoneware (1735-65), 1 English delftware with cobalt decoration, (poss-
ibly early 18th century), 3 from small Delftware bowls (Fig. 7, h) with cobalt
decoration (1710-40), 2 English white salt glaze (1710-40), 3 Yorktown brown
stoneware (1720-1760), 1 English delftware of uncertain date and form, 1 Vir-
ginia earthenware similar to Yorktown earthenware, and 1 Yorktown type lead-
glazed earthenware (Fig. 7, i).

Decorative items were a silver bangle (Fig. 8, f) and 5 glass beads:
1 clear, flat, 1/4 diameter; 1 white, 5/16 by 1/8 inch, 1 clear, 1/4 high by
6/16 inch diameter (Fig. 7, 1), 1 black, 1/2 by 1/8 inch, and 1 red, 5/16 long
by 1/4 inch diameter. Also present were 127 fragments of bottle glass, Most
were thin and green but 13 were thick and green, and 11 clear and thin.

Animal Remains

Excavated animal bones included 45 bird, 39 deer, 68 large mammal
fragments (probably deer), 208 small bone fragments, and 1 fish bone. Twenty-
three riverine shell fragments, 1 fossil shark tooth and 2 fossilized pecten shells
complete the inventory. It is not clear whether or not the fossilized items
were used.


The main occupation of the village is estimated to have been between
A. D. 1735 and 1750. This is based on three sets of criteria: pipe stem dating,
Harrington's graph method, 1710-50 and Binford's formula, 1743.61; colonial
ceramics manufactured between 1710 and 1765; and buttons manufactured be-
tween 1726 and 1776. Trade musket parts and gun flints are of this general
time period but are not as specifically dateable. The aboriginal ceramics--
except for the fiber-tempered sherds--are closely similar to types found in
Florida at Spanish Mission period settlements known to have been destroyed
by Gov. Moore's (of Carolina) raids of 1702 and 1704. These forms continued
to be manufactured in Alabama and western Georgia and undoubtedly were still
in use in the middle of the 18th century. Huscher, from his survey and inter-
pretation of historical data, considered this area was occupied by a historic
Creek town, Sawokli (Huscher 1959:52-3, sites Ru-20, 21). The archaeology
is in complete agreement with such a correlation.











I '



8 1

S 0 I

I r'

1 y

1%, / '

. I


4 0 3 2 I 10 9 S 7 6

SII 12 i 14 II IS I? IS IS

Fig. 9. Diagram showing association between pipe fragments, glass, and
selected aboriginal pottery (in hundreds of sherds), and lack of association
with chert chips (in tens of chips).



The 32 fiber-tempered sherds (Type VII do not fit into the above picture.
In Florida such sherds are not known to have been made after 1000 B. C. As
shown in Table 1, these sherds came from squares which also had substantial
quantities of later aboriginal pottery. However, all came from the southern
trench and 20 of the 32 sherds (62. 5%) came from the 6-12 inch level of Squares
11, 12, 17, 18, and 19, while only about 26. 6% of the other aboriginal pottery
from these same squares was in the lower level. The fiber-tempered sherds,
then, have a relatively deeper provenience which is consistent with their pre-
sumed earlier date of manufacture.

If it is assumed that the greatest amount of archaeological debris in this
village was deposited where either the greatest number of people lived or where
a small contingent lived for the longest time, this relationship may be tested by
plotting numbers of artifacts against their proveniences. This was first done for
the colonial specimens as categories (not numbers of specimens) and it was
found that these items concentrated between Squares 8 and 19 including the ex-
tension east of Square 12, with a minor concentration in Square 1. It may be
noted that, in general, all of these are east of the "fence line" of Feature 2.
This approach was continued using quantities of pipe fragments, glass, selected
aboriginal pottery, and chert chips. The results are presented in Figure 9 and
indicate a close correlation between the colonial artifacts and aboriginal pottery
with the heavier concentration in the southern part of the site (Square 6 outside
the "fence line" as might be expected is divergent).

It was assumed at first that the chert chips were also deposited by the
people of the mid-18th century village. Figure 9, however, shows a very
definite reverse relationship. Nor do they show a close correlation with the
distribution of fiber-tempered ceramics. This suggests that the site was oc-
cupied at three different times: first towards the north during the preceramic
Archaic period (stemmed projectile points and chips), second around 1000 B. C.
in the most southern area (fiber-tempered pottery), and third the mid-18th cen-
tury village in the middle and southern areas. Perhaps, the chips and fiber-
tempered pottery parts represent the same occupation with chipping activities
occurring in the northern and food preparation in the southern parts of the site.

The last village extended over 1000 feet parallel to the river and was 150
to 200 feet in width. Houses were not apparently built in the usual native fashion
by sinking posts into the ground. I can only assume they were above surface con-
structions like log cabins which floods could wash away leaving no more trace
of a village than was found. No stones large enough to have acted as foundation
or chimney were on the site. The recovered bones indicate the people's diet of
protein came from forest and river. Charcoal samples have not yet been an-
alyzed but some charred material appears to be seeds but not those of corn.

Lower Creek and other Indians, later to be known as Seminoles, drifted
southward down the Chattahoochee and Apalachicola Rivers into Florida after


1715. Excavations at the Georgia Kraft plant near Cottonton, Alabama, give
an insight into conditions in the area from which they came at a little later date.
It will be noted that no suggestion of Spanish influence was uncovered. Indians
living on the Middle Chattahoochee River were oriented towards the British
colonies during the 18th century. This is in agreement with history but it is
interesting to note products of the Virginia colony were distributed by traders
west of the Chattahoochee River at this early date.

Literature Cited

Bullen, Ripley P.
1950 An archeological survey of the Chattahoochee River Valley in
Florida. Journal of the Washington Academy of Sciences, vol.
40, no. 4, pp. 106-25. Washington.
Goggin, John M.
1953 Seminole pottery, 9-53. 'In Prehistoric Pottery of the Eastern
United States, edited by James B. Griffin. Museumof
Anthropology, University of Michigan. Ann Arbor.
Huscher, Harold A.
1959 Appraisal of the archeological resources of the Walter F.
George Reservoir area, Chattahoochee River, Alabama and
Georgia. Mimeographed report on file at River Basin Surveys,
Smithsonian Institution. Washington.

Sears, William H., and James B. Griffin
1950 Fiber-tempered pottery of the southeast. In Prehistoric
Pottery of the Eastern United States.

Smith, Hale G.
1946 Two historical archaeological periods in Florida. American
Antiquity, vol. 13, no. 4, pp. 313-19. Menasha.

South, Stanley
1964 Analysis of the buttons from Brunswick Town and Fort
Fisher. Florida Anthropologist, vol. 17, no. 2, pp. 113-133.

Willey, Gordon R., and William H. Sears
1952 The Kashita Site. Southern Indian Studies, vol. 4, pp. 3-18.
Chapel Hill.

Charlottesville, Virgina
May 1972


Jennings W. Bunn, Jr.

In October, 1968, while surface collecting on what I believe to be one
of the oldest sites in the Choctawhatchee Bay area, I was fortunate to locate
a cache of sixteen complete clay balls (Fig. 1), or baked clay objects as they
are commonly referred to, and several fragmentary ones. The site, apparently
a midden, is located on the west side of Four Mile Point on the eastern shore
of Horseshoe Bayou, and approximately one quarter mile north of the mouth
of the bayou. The state site number, (WL-36), was ascertained from Mr. L.
Ross Morrell, State Archaeologist. The site environment is composed pri-
marily of oak, hickory, pine, and dense palmetto, all situated in a low ham-
mock area. Several small shell midden areas are scattered about the approxi-
mately one acre site, none over eighteen inches in height.

The cache was located approximately twenty feet from shore in one
foot of water, and all the balls were in close proximity with each other. Five
of the specimens were completely exposed on the sand bottom, and the addi-
tional eleven were directly under these to a depth of approximately one foot.
All specimens were perfectly formed, and in an excellent state of preserva-
tion. It was impossible to ascertain whether the clay objects had originally
been in a cooking environment or hearth area as they were found in the water
and probably had been for several years. This was the first cache of baked
clay objects I had found there, but by no means the first individual finds.

On June 7th, 1969, while again surface collecting on the same site, an-
other cache of thirteen complete baked clay objects (Fig. 2) was located. This
cache was eroded from the low bank still encased in a black muck and plant fi-
ber matrix (Fig. 3). I believe this matrix, which is several inches below the
present water level, represents the oldest strata on the site. Besides producing
numerous baked clay objects, the site also yields fiber-tempered ceramics,
and a large number of Archaic-like stemmed projectile points.

A few theories as to the use of these clay balls are: cooking objects in
place of stone, which is scarce in the Choctawhatchee Bay area, hunting imple-
ments such as sling stones, bolas weights, or fishing net weights. Possibly
they were also used for ceremonial purposes as heating stones for sweat houses.
Actual use for these objects was probably in food preparation, but in a real
sense their use is, and probably will remain, problematical. It should also be
noted that these clay balls show cultural connections with Poverty Point and
other sites in Louisiana and the lower Mississippi River valley. Those from the
first cache (Fig. 1) have long shallow longitudinal grooves and are of the Clay-
borne type.

Florida Anthropologist, vol. 27, no. 1, March 1974


Fig. 1. Grooved clay balls
from Cache 1,

Cache 2.

r Fig. 3. Cache 2 balls in
black matrix.

Tallahassee, Florida
March 1973

~ "~n""4f~P~


Back issues of the FLORIDA ANTHROPOLOGIST, if available, may
be purchased from the Treasurer at $6.00 per volume ($1.50 for single and
$3.00 for double numbers). Volume 25:2 pt. 2 (Publication No. 6) is $2.00.
Volumes 1:1-2; 5:1-2; 6:1, 4; 7:3; 8:1, 2; 10:1-2; 11:1; 13:1 17:1; 18:1; 21:4;
24:1; and 26:1 are out of print.

a mongraph series, may be purchased as follows: Numbers 1, 5, and 6 at
$2.00 each, number 4 at $4.00 each. Numbers 2 and 3 are out of print.

Orders with remittances for any of the avoce should be sent to the
Treasurer whose address is given on the inside of the front cover. The
Society has no reale discount for dealers or resale agents.


THE FLORIDA ANTHROPOLOGIST publishes original papers in all
subfields of anthropology with an emphasis on archaeology. Contributions
from allied disciplines are acceptable when concerned with anthropological
problems. The journal's geographical scope, in order of importance, in-
cludes Florida, Southeastern United States, Caribbean, and related regions.

Manuscripts should be double spaced and typed on one side only of an
8 1/2 by 11 inch sheet leaving at least a 1 inch margin on all sides. Contri-
butors are referred to American Antiquity and the Chicago Manual of Style
for matters of style and reference. Textual references should be similar to
"Smith 1964: 78-80; Jones 1969: Pl. II, b) ." No footnotes. In the "Refer-
ence Cited" section, at the end of an article, do not abbreviate names of ser-
ies or journals. Books need full title, publisher's name, and place.

All illustrative material-line drawings (in India ink) add photographs-
should be included in one numbered series of "Figures" suitable for a maxi-
mum reduction to 6 1/2 by 8 1/2 inches. Parts of figures are designated by
lower case letters. Figures must be accompanied by brief descriptive cap-
tions. All tabular material must be included in separately numbered "Tables. "

Twenty-five reprints without covers of each article are furnished free
to each author or group of authors. Additional copies, withor without covers,
may be purchased but the editor must be informed of the number desired when
the article is offered for publication.

4 -r

Full Text