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.


^ *



MARCH 1976



THE FLORIDA ANTHROPOLOGIST is published quarterly in March, J
September, and December by the Florida Anthropological Society,Inc., c/o
Room 102, Florida State Museum, University of Florida, Gainesville, FL 32Z
Subscription is by membership in the Society for individuals or institutions i
terested in the aims of the Society. Annual dues are $6. 00; student members
t4. 00. Requests for memberships and general inquiries should be addressed 1
the secretary; subscriptions, dues, changes of address and orders for back iss
to the treasurer; manuscripts for publication to the editor; and newsletter
items to the president. Second class postage paid at Gainesville, Florida.




XXIX No. 1

March 1976



A New Look at the Georgia Coastal Shell Midden Archaic by
Donald L. Crusoe and Chester B. DePratter . 1

Use of a Mechanical Earth Auger at the Torreya Site, Liberty
County, Florida by George Percy . .. . 24

Some Thoughts on Florida Projectile Points by
Ripley P. Bullen .. . . .. .. 33

Hinderland Exploitation in the Central Gulf Coast-Manatee Region
during the Safety Harbor Period by Thomas J. Padgett 39


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

Ist Vice President Raymond Williams
Dept. of Anthropology, University of
South Florida, Tampa, FL 33620

2nd Vice President Jerald T. Milanich
111 SW 23rd Terrace, Gainesville,
FL 32604

Secretary George W. Percy
Div. of Archives, History, and
Records Management, 401 East Gaines
Street, Tallahassee, FL 32304

Treasurer Norcutt Henriquez
1510 Dewey St., Hollywood, FL 33020

Directors at Large

Three years: Robert E. Johnsor
4250 Melrose Avenue
Jacksonville, FL 32210

Two years: Ray C. Robinson
1020 4th Street North
St. Petersburg, FL 33701

One year: Wesley
10 NW 124 Avenue
Miami, FL 33126


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



Donald L. Crusoe and Chester B. DePratter

With the publication of the Waring Papers (Williams 1968), most ar-
chaeologists would have conceded that the general prehistoric fabric of the
Georgia coastal marsh area was known (cf. Ford 1966) Indeed, the coastal
area, aside from Ocmulgee, has received the most scrutiny in the state. The
results of these investigations have generally characterized the Georgia
coastal Archaic as an extension of the Stallings Island "Culture" of the Cen-
tral Savannah River locality. Our recent re-evaluation of the Georgia coastal
marsh (which included surveys, tests, and analyses of museum collections)
has led us to conclude, as Caldwell (1970:88-92) did for the Wilmington Phase,
that the few sites which have been excavated have produced a distorted picture
of the coastal marsh. This paper is dedicated to the presentation of some of
our findings.

We are presenting our ideas in the form of hypotheses and suggestions,
hoping to generate new interest in the archaeology of the Georgia coastal
marsh. Specifically we will propose: (1) some new interpretations, (2) a
slightly modified time-space framework which will be based upon recent geo-
environmental investigations of the coastal marsh, and (3) a ceramic typology
which, we feel, will facilitate comparative ceramic analysis.

Development of the Coastal Marsh and Islands

The coastal geo-environmental history of the Southeastern United States
is tightly tied to the Pleistocene. As each ice mass was formed and waned, a
recurrent pattern of water level recession and stabilization occurred; and, on
the Georgia Coast for example, each time the water level stabilized, an island
chain was formed. There were at least six Pleistocene and one Holocene sta-
bilization; but for our purposes we shall exclude the earlier four stabilizations,
since they have no bearing upon our topic, and concentrate on the latter three
known as the Princess Ann, Silver Bluff, and Holocene stabilizations with their
associated island formations.

At the time of the Princess Ann (ca. 50,000 B. P.), sea level was about
15 feet above the present mean sea level (Hoyt, Weimer, & Henry, 1964:1972;
Hoyt, 1968:26). The resultant land form can be equated with sections of the pre-
sent Georgia coastline. Hoyt (et. al. 1964) has found that at the time of the sec-
ond stabilization under consideration -- the Silver Bluff formation (ca. 25,000
B. P.) -- the water level was only 2 5 feet above the present mean sea level.
At this depth, major portions of Cumberland, Saint Simons, Jekyll, Sapelo, St.

Florida Anthropologist, vol. 29, no. 1, March 1976


Catherines, Ossabaw, Skidaway, and Wilmington Islands were developed.

Following the formation of the Silver Bluff islands, the sea level dropped
to approximately 330 feet below present sea level (ca. 18,000 Bo P.); with the
melting of the glacial ice, sea level again rose and probably stabilized somne-
time during the last four to five thousand years (Hoyt, 1968:23, Bloom 1971)
with only a minor rise in sea level since. Since this final stabilization, Wassaw
and Tybee Islands have been developed and additions have been made to all of
the Silver Bluff islands save Skidaway and Wilmington (Hcyt, 1968:23).

Hoyt and Hails (1967:1542) have found that during the development o!
some Pleistocene shorelines (including the Silver Bluff islands) shallow. opera
water lagoons were present in the area behind the islands. During the Wiscon-
sin glaciation, these lagoon areas were probably dry land or fresh water svan--p,
and it was only toward the end of the most recent sea level rise that the area be-
hind the islands was once again flooded and became a lagoonal-salt marsh (Ho-t
and Hails 1967:1543). This area of marsh-lagoon is an ideal area for the for:a-
tion of extensive oyster beds. According to a recent study by Linton (1370) ,
oyster beds are located predominately in the area behind the barrier islarcds
or along the creeks and rivers which drain the marsh areas. Very ev. 1 a-:,
sizeable beds are located in the foreshore area of the coastal islands. This dis-
tribution is more easily understood when the salinity (5 to 30PPTs/1000) ,
perature (340 to 86 F) free water exchange patterns, and soil base require-
ments of the oyster (Crassostrea virginica) are taken into account (Johson,
et. al. 1971:176) .

Although temperature and free water exchange requirem-ents of the : -s:er
could be met in isolated foreshore areas, other factors such as high sali.ity
turbulence, and an inadequate soil base would all probably limit the oornation
of extensive oyster beds. It therefore seems reasonable to suggest that large
oyster beds did not reform behind the Silver Bluff and Holocene islands unti
the area was reflooded (at ca. 4000 5000 B. P. ). Further, it is reasonable
to suppose that, unless the specific requirements of the oyster existec -arter
out on the continental shelf, there were probably no large oyster bes ior-e
during the most recent sea-level rise across the continental foreshore At
present, we have no reason to believe that such conditions existed farther o :
on the shelf, and therefore we would not expect to find any Shell Mound Archaic
sites on that portion of the shelf now covered by the sea. In light of the evidence
presented thus far, which, when taken in conjunction w ith the Rabbit M.V-: ev-
idence and dates, we, as Waring (Williams 19{68:191), predict that the area
slightly inland is where the earliest Georgia fiber-te -pered ptte r-.- ea :1
sites are located (Map 1).

Having dispensed with the notion that there are non--ceran~ic Shell i
Archaic sites in the Georgia Coastal Marsh area, at least to our satis-actc.






-Sapelo Island


Map 1


we would like to present an inventory of the major known ceramic-bearing
Shell Mound Archaic sites in the study area. The purpose of this inventory is
to provide a backdrop for the re-evaluation of the time-space framework which
we will shortly develop.

Inventory of Known Shell Mound Archaic Sites

Before continuing with the rationale for separating the coastal and inland
fiber-tempered pottery-making sites, we would like to provide sketches o: the
extent and content of major Shell Mound Archaic sites east of the fall-line in
Georgia. Let us say, at once, that it is not our intent to develop a complete
site report for we shall merely note the extent of the sample (excavations),
the stratigraphy, and the vertical distribution of the ceramic artifacts. The
site sketches which follow represent the result of our own research, as .well
as archaival documentation of sites excavated since the late 1930' s under
Government contract. Site locations will be found on Map 1.

Stallings Island (9Col) References: Claflin 1931; Bullen and Green 1970;
Fairbanks 1942. A "tear-drop" shaped shell midden occupies the center o3
Stallings Island in the Savannah River, about 10 miles northwest of Augusta,
Georgia (Figs. 1-2) It measures about 500 feet on its long axis and 300 fee
on its short axis. At its thickest, the midden has a depth of about 9 feet. In
general, the ceramic component is restricted to the upper 3 feet w ith the no-
ceramic component lying below and being restricted, it would appear, to the
more central area of the midden. Six test cuts \were excavated by us consist-
ing of two 9-by 15, one 10-by 10, two 3-by 15, and one 3-by 18-feet (Pit 1-
respectively). all excavated by natural levels and all except Pits 1 and 2 dug to
basal clay. Figure 1 displays the configuration of the depositional pattern at
Stallings Island. We will allow this profile section to speak for itself, for a
verbal description would be too lengthy and is extraneous for the purposes o:
this paper. Pit 2, however, has many salient features which are central to the
purpose of this paper. Accordingly, we shall outline then here.

Pit 2 was a 9-by 15-foot test cut oriented north to south on its long axis
and located between Claflin' s Trenches 2 and C4. The northern end of this pit
was fairly level before excavation, while the southern end had a ridge of earth
approximately two feet high, and this earth ridge ran diagonally- across the
southern one-quarter of the pit. We believe that this ridge represents spoil
from Claflin' s Trench 2.

First we excavated the northern seven feet of Pit 2 in 0. 3 foot levels.
When excavation to the 1. 2 foot mark was accomplished, we had completely
removed all of the spoil from Claflins excavation. With this kno. ledge, the
southern half of the cut was dug. We exposed an undisturbed compact mnidden.
We term this Zone 4 (numbered from the bottom upward) It as 1.3 3eet t:ik


and composed almost entirely of dark earth, mussel, and snail shell. Through-
out the excavation of Zone 4 there was no change either in consistency nor in
content. Having removed Zone 4 (a ceramic-bearing zone), Zone 3 was exposed.
Zone 3 was completely distinct from Zone 4 in that it contained relatively few
shells. After one foot of earth had been removed we found a red clay lens
(Zone 2) which had apparently been fired. This lens apparently extended to the
north and we opened an extension to Pit 2 in an effort to delimit this feature.
Unfortunately there was too much disturbance in that area to trace this feature.
Zone 2 had a thickness of one foot. Below it was Zone 1 which we did not exca-

The essence of the above sketch is that two cultural zones were excavated
while the third zone (Zone 5) probably represents Claflin's spoil. Zone 4 was a
humic midden with shell inclusions; Zone 3 was a clay midden with a few shells
included. This is the stratigraphic picture intimated by Claflin but, as we will
discuss later, the artifact picture is different.

Z Z -X 1.- 1

0- .1, jI "I -

00 I-

o ... :

4 r*
/ 4,,



Fig. 2. Stallings Island exca-
vations: upper left, south
face of Pit 3; lower left;
Zone 2 depth of Pit 2;upper
right, Pits 4, 5, and 6.

Charlie King (9Gn3) References: None. Preston Holder first worked the Char-
lie King site under W. P. A. funding. A few random monthly reports are on file
at the Southeast Archeological Center, Tallahassee.

Located on Saint Simons Island, Georgia, .75 miles south of the Airport
sitejHolder (1935:10-11) describes the site as "A sand burial mound and asso-
ciated shell-covered habitation area... The shell areas have been very badly
disturbed by road-building crews; while the center of the burial mound has been
explored by amateur enthusiasts..." An undetermined number of test cuts were
excavated by him. Mr. King reported that Holder excavated what remained of
the mound. We excavated one 6-by 6 foot test square but found no fiber tem-
pered pottery.

Cannon and Butler Points References: None. Holder's monthly reports re-
pose at the Southeast Archeological Center, Tallahassee.


Cannon's Point occupies the extreme northeastern point of St. Simons
Island, Georgia, while Butler Point is an extensive but thin shell midden front-
ing the Hampton River [on the northwestern end of Saint Simons Island, Georgia]"
(Holder 1935:12). Holder (1935:5) felt that Cannon's Point "... may well have
been the most extensive site on St. Simons Island, since cultural debris is
scattered over many hundred acres of the point and the shores of the adjacent
creeks". Butler Point consisted, as we noted earlier, of a thin shell midden
deposit along the bank of the Hampton River. No information is available re-
garding any excavation.

Sansavilla and Paradise Park References: None.

Sansavilla is located five miles northeast of Mt. Pleasant, Georgia, on
a high bluff overlooking the Altamaha River. Paradise Park is located five
miles northwest of Sansavilla near the fishing camp of the same name and is
also on a high sandy bluff overlooking the Altamaha River. Size is difficult to
ascertain since both middens are located in heavily wooded areas, but the dis-
tance from the edge of the bluff to the outer edge of the sites is less than 100

We excavated a 31-by 20-foot trench at Sansavilla in a northeast to south-
west direction across the presumed "center" of the site. At Paradise Park we
dug a 3-by 8-foot trench parallel to the river. No shell midden was encountered
at either site and the matrix was sand, with little or no zonal differentiation.
Numerous plain fiber tempered potsherds have been excavated and surface
collected from these sites. They are deposited at the Laboratory of Archeology,
University of Georgia.

A. Busch Krick (9McI87)- Crusoe 1972.

This horseshoe-shaped site is situated about 7 miles east of Darien,
Georgia, on Creighton Island. The midden measures slightly over 60 feet along
its north-south axis and presumed erosion has dwindled the "horseshoe" to less
than 130 feet on its east-west axis (Fig. 3). The midden appears to have a
maximum thickness of eight feet.

We excavated one 3 by 15 foot trench and one 6 x 6 foot test square as por-
trayed in Figure 3. Between the superficial humus zone and basal sand there were
four natural zones and one large lense. Above the sterile yellow sand was a
silty shell zone (Zone 1) Next was a dark crushed shell zone (Zone 2) which
forked into two sections surrounding a zone of dark crushed shell with a high
concentration of periwinkle shell (Zone 3) Above was a zone of whole and
crushed shell (Zone 4). Finally there was a zone (Zone 5) of whole shell which
was situated immediately below the humus. Table 1 presents the ceramics
(Fig. Q) excavated from A. Busch Krick.






0 10 2dI

Fig. 3. Contour map, A. Busch Krick site.

Table la. Distribution of sherds in Pit 1, A. Busch Krick site.

Levels 1-2 3-4 5-6 7-8 9-10 11-12 13-14 15-16 17-18 19-20 21 Totals

Plain 14 28 19 45 40 45 57 46 20 31 5 350
Incised 1 1 1 1 2 6
Simple stamped 1 1
Punctated 1 2 2 3 1 1 1 11
Incised & punct. 1 1
Punct. in Inc. line 2 2 4

Totals 18 20 22 52 41 47 59 47 20 32 5 373

Cane Patch Island (9Ch35) References: Crusoe 1972.

This site occupies the northern end of Cane Patch Island near Ossabaw


Island, Georgia. The midden measures about 200 feet east-west axis and perhaps
250 feet north-south, though disturbed in most areas, with a maximum thickness
of about 10 feet.

We excavated one 3, 5-by 15-foot test cut in the west, southwestern sec-
tion of the site. The stratigraphic picture at Cane Patch Island is as follows:
Zone 1, sterile yellow sand; Zone 2, black much midden, which may have been
exposed to fire; Zone 3, black soil; Zone 4, whole oyster shells; Zone 5, dark

Table lb. Distribution of pottery in Pit 2, A. Busch Krick site.

Levels 1 2 3 4 5 6 7 8 9 10 Totals

Plain 12 15 23 19 11 10 17 15 19 26 167
Incised 1 1 1 2 6 11
Simple stab & drag 1 1
Punctated 1 1 1 1 2 6
Punct. in inc. line 4 1 5
Unidentified 1 1 1 3

Totals 13 16 23 21 22 16 21 17 27 27 193

crushed shellmidden; Zone 6, humus. Table 2 presents the ceramics (Figs.
8, 10) recovered from the Cane Patch midden.

Shell "Crescent" at Skidaway (9Ch60) Reference: Beasley 1971.

This crescent shaped midden, about 100 feet north-south and scarcely
30 feet east-west, is situated on a small peninsula which extends to the south
to the mouth of a small inlet or cove. Two 5-by 5-foot test cuts were excavated.
This site has reverse stratigraphy with fiber tempered pottery appearing in
zones superior to those containing Wilmington ceramics. Upon closer examina-
tion of the area around the site, extensive earth movement scars were noted.
It is possible that the crescent shape of the site is due to the removal of shell
many years ago.

Pagan Plum Point (9Ch61) References: Beasley 1971.

This small shell midden lies on the edge of the marsh 2 miles north of
the Skidaway Crescent. It extends inward from the marsh bank some fifty feet
and has a north-south axis of about 150 feet. Two 4-by 4-foot test cuts were
placed near the bank which overlooked the marsh. They were excavated to ster-
ile sand at 4 feet. The composition of the midden was a nearly homogeneous
mixture of dirt and shell.


Table Za. Distribution of pottery in Section A, Cane Patch site.

Levels 1-3 4-6 7-9 10-12 13-15 16-18 19-20 Totals

Plain 25 33 25 52 67 74 15 291
Incised 7 10 9 16 3 4 49
Stab and drag 1 2 1 4
Punctated 9 19 12 13 19 17 2 91
Random Punctated 10 8 8 5 3 34
Inc. and punctated 1 1 6 1 9
Punct. in inc. lines 2 1 3 5 3 14
Zoned inc., punct. below 3 1 3 1 2 10
Unidentified 1 3 3
Grit tempered 3

Totals 57 62 60 99 107 100 17 511

Inland and Coastal Shell Mound Archaic Sites

Following the excavations outlined above at Stallings Island, Cane Patch
Island, A. Busch Krick (Creighton Island) and Pagan Plum Point (Skida\way
Island), we began to question current ideas on the origin of coastal fiber tem-
pered pottery--that is, its relationship to the earlier Stallings Island non-ce-
ramicramic phase. Evidence from the above-listed sites led us to consider the
hypothesis that instead of the gradual transition and cultural continuity front.
non-ceramic to the ceramic there could have been a hiatus or discontinuity i:
the sequence. Indeed there appears to be definite and distinctive differences
between these phases and there is little evidence for the transition.

Although the Stallings Island site has been excavated and re-excavated,
there are many problems with the published data from this very important site.
The material we recovered from our excavation in the ceramic levels appeared
to differ greatly from what we expected (via Claflin 1931) and so we began to
compare the ceramic and non-ceramic Stallings levels (Fig. 4).

First, we noticed that steatite "net-sinkers" were more abundant in the
non-ceramic than in the ceramic levels. Miller (1948:40) also found the same
distribution at Lake Springs. Second, we noticed that projectile points also dis-
played a change from large, rhyolite forms (non-ceramic) to smaller quartz
and chert forms (ceramic) (Bullen and Greene 1970; Miller 1948). Thirdly, no
burials were encountered. The only burial recovered by Miller at Lake Spring
was from the non-ceramic occupation. Fourth, although Claflin (1938:28) un-
earthed 60 atatl weights and weight fragments from his excavation (three asso-
ciated with burials), we recovered none in the ceramic levels, and the tvo that


Table 2b. Distribution of pottery in Section B, Cane Patch site.


1-3 4-6 7-9 10-12 13-15 16-18 19

Simple incised
Compound incised
Zoned compound incised
Curvilinear incised
Discontinuous simple grooved
Simple grooved
Zoned simple grooved
Simple punctated
Zoned simple punctated
Curvilinear punctated
Zoned curv. punctated
Random punctated
Zoned random punctated
Simple incised and punctated
Zoned simple inc. and punct.
Curv. incised and punctated
Simple punct. in inc. line
Zoned sim. punct. in inc. line
Grooved above, punct. below
Simple punct. in grooved line
Sim. punct, in inc. line & s. p.
Comp. inc. with sim. punct.
Zoned sim. punct.w. ran. punct.
2 inc. 1. followed by 2 lines sim.
punct. in inc. 1, followed by 3
lines simple punctations

1 1 1

4 3

2 10

8 2 6 8
1 2
1 2
7 10 4 10


1 2 1

2 2

50 41 42 84 134 29 9 389

Miller found were from non-ceramic levels.
grooved axes and fragments, we found none,
ceramic level (possibly out of context?) .

And fifth, Claflin also found 18
and Miller found one in his upper

The above comparisons fortified our notion of discontinuity in that there
was apparently a great deal of difference between the non-ceramic and ceramic
levels at Stallings. In addition to the "loss" of artifact types such as steatite
"net-sinkers" and atlatl weights, and the change in the form of projectile points
or knives, we also found an apparent lack of burials in the ceramic components
at Stallings Island. At present we are at a loss to explain all of these changes,
but we would like to put forth a hypothesis to explain some of them.


7 209
1 34
1 4





As has been noted, during our excavations at Stallings Island w e found a
very significant difference between the shell content (quantity) of the non-ceran-
ic and ceramic components (in Pit 2): the ceramic zone was made up almost en-
tirely of mussel shells, while the non-ceramic zone was composed of reddish
soil with a few shell inclusions. The difference in shell content was n ost strik-
ing. We posit that the change in shell content (quantity) is indicative of a differ-
ent subsistence pattern. If this is the case then, the discontinuity in the artifact
classes becomes easier to understand.

Our hypothesis that there was little continuity between the non-ceran-ic
and the ceramic levels at Stallings seemed sound, so we began a comnpariso
of the ceramic phase materials from Stallings and from sites on the Savannah
River and in the Georgia coastal marsh.

We found that steatite "net-sinkers", so abundant in the Stallings non-
ceramic levels, appear to be fading out in the ceramic. They are found at a few
sites between Stallings and the Atlantic. Eight-six fragments were found at Rab-
bit Mount (Stoltman 1972), and three were found at the coastal marsh Bilbo site
(Williams 1968). We recovered no additional "net-sinkers" in our v.ork in th
coastal marsh. Since only a few projectile points have been recovered from
Coastal Plain sites, all we can add to the previous comparison is that the pro-
jectile points which have been found are made of chert and not rhyolite. O -i.-
two complete burials (possibly associated with fiber tempered pottery) are
known from the coastal shell middens: one at Daws Island, South Carolina
(Hemmings 1969), and one in Depratter' s excavations at Pagan Plum Point.
No artifacts were found with either one, but it is assumed that both are associ-
ated with a fiber tempered ceramic occupation since neither show s evidence of
intrusion. Human bone fragments were found at Bilbo and Rabbit Mount. T7 '
atlatl weight fragments have been found at Rabbit Mount. Additional arti-
facts from the coast include perforated conch shell tools (Bilbo, Paar-
Plum Point, A. Busch Krick, and Rabbit Mount), and baked clay objects
(Dulany, Pagan Plum Point, and Sapelo).

While the differences we have noted could, as Stoltman su~eests, be
due to lack of certain resources on the coastal plain, it appears to us that
most of the lithic traits were beginning to fade toward the end of the non-
ceramic Stallings phase.

Waring made a trait list for Bilbo (Williams 1968:188) and cornpared
it to Fairbanks' (1942) list from Stallings Island. He found that oo of the
118 traits were shared between the two sites. The problem with this com-
parison, as we see it, is that the list prepared by Fairbanks included both
non-ceramic and ceramic period characteristics. Therefore Warin2's com-
parisons are useful if one wishes to compare the Shell Mound Archaic i
general; but if our hypothesis of discontinuity is valid, then comparisons
such as this are not pertinent.


Utilizing Waring's trait list we compared Bilbo to A. Busch Krick,
Cane Patch, Pagan Plum Point and Charlie King, and found that sites in our
sample are also small in size, constructed on silt and sand, displayed ex-
tensive use of shellfish, gar, and perhaps sturgeon and walnuts, had pounded
fragments of animal bone, and displayed no postmold evidence. Category II
(chipped stone complex) and category III (rubbed stone complex) are not
represented as we have already noted. The bone and ceramic complexes,
excluding specific "sub-types" of artifacts, are present in our sample of
sites, except spatulate pins, worked animal jaws, and bone chisels. Out of
Waring's 66 comparable traits (with Stallings Island) we find only 40 iden-
tities between our sample of coastal sites and the ceramic and non-ceramic
trait list from Stallings Island. This points to the uniqueness of the Bilbo
site. The inventory from Sapelo (Williams 1968:263-80) and our sample is
more typical of coastal marsh sites than are those from Bilbo. For this rea-
son we suggest that the few sites which have been excavated have produced
a distorted view of the Georgia coastal marsh area.

We, like Waring, feel that the coastal marsh sites are representative
of a group which moved to the coast from an inland location, and that the
two groups of people (inland and coastal) are related in time and, to a cer-
tain extent, in space (Williams 1968: 191). However, on present evidence,
we reject Waring's contention that "the artifact assemblage is very much like
the artifact assemblage of any Archaic site in the Southeast" (Williams-1968:
254, italics ours). Certainly, the unique absence of stone artifacts from most
coastal marsh sites is related to environmental absence of stone in the area,
but we suggest that it could also represent the distinctiveness of the cultural
adaptation to a salt-marsh-lagoon environment. At any rate, we feel that
the Bilbo phase is a viable and useful concept which allows the coastal
marsh fiber-tempered pottery-making complex to be distinguished from
similar inland cultural manifestations which are both earlier and later.

The foregoing appears sufficiently strong to restate Waring's (1968:
254) hypothesis that the Georgia coastal marsh, during the Shell Mound Ar-
chaic, represents a cultural complex which possesses traits sufficiently
characteristic to enable it to be distinguished, for pruposes of preliminary
archaeological classification, from earlier and later manifestations of cul-
tural development of which it is a part (Willey and Phillips 1958:22). Further,
we suggest that the various manifestations of the Shell Mound Archaic on the
rimland, in the marsh, and on the islands are representative of a cultural
phase distinct from the inland Stallings Island phase. We would prefer to utilize
the term the Saint Simons phase, which has precedent (Caldwell and Waring
1938) but will bow to the nomenclature already'well entrenched in the litera-
ture (Williams 1968) and utilize the term the Bilbo phase.

A Review of the Chronological Framework for the Coastal Marsh

Waring's original conception was that the Bilbo phase was comprised of


two subphases, Bilbo I and II. Our recent work has cast ne\w light on these t o
subphases and at the same time has substantiated the basic definitions of the
subphases. We, however, propose stronger spatial and temporal controls than
Waring. Spatially, the limits of the Bilbo phase extend from Cumberland Island.
Georgia, on the south, to the South Carolina marsh (ca. Hilton Head, though
only Georgia sites will be considered here) in the northeast. On the east -is te
Atlantic Ocean, while on the west only those areas immediately adjacent to t-e
marsh are included. The abruptness of the western border is related to the
brackish (versus fresh) water life zones which uere explo ted by the Bilbo peo-
ple. In terms of time, we feel that this phase lasted about 1200 years, extenc-
ing from about 2200 B.C. to 1000 B. C. : Bilbo I subphase sites cluster in the
period 2200 B.C. to 1700 B.C., while Bilbo II sites are found in the 1700 C. C.
to 1000 B. C. range.

Waring (1968:251-2; 255, passim) envisioned a hiatus between the Bilbo
subphase and the following Deptford I subphase. Some evidence iron inland
Georgia--the Waycross, Georgia area--as well as some from the coastal ar
hints at a continuity; the question of Bilbo-Deptford continuity is, how ever, ci-
vergent from the purpose of this paper. The next section of this paper ill be
devoted to an examination of the Bilbo subphases.

BILBO I (ca. 2200-1700 B. C.)

Slightly before the second millennium before Christ small settlements
began to be established on and in the developing coastal marsh. It is our vy-
pothesis, as well as Waring's (1968:191) that the Bilbo I people moved i
the coastal marsh from slightly inland and adapted their riverine subsistence
pattern to a brackish water marsh environment. The diagnostic artifacts of
Bilbo I sites are: plain fiber tempered pottery, plain and simply decorated bone
pins, and baked clay objects (cf. Williams 1968:254) This inventory is found
at the following sites: Dulany (9Ch54) Pagan Plum Point (QChol) the !o er
levels of Bilbo (9Ch4) and Sapelo (9MNcI23) The first three of these sites are
typical middens, while the latter site (Sapelo) is a shell ring. Spacially, the
Bilbo I subphase appears to have been restricted to the upper sector Do the
Georgia coastal marsh (this distribution may be more apparent than real)
Temporally, we have carbon dates from Dulany (1820+125, M-l1ll1 170i325,
M-1112), and Sapelo (1750250, M-39) (Crane and Griffin 1958:1122; Cra-e anC
Griffin 1963:239-40; and Crane 1956:665). We now propose to discus so:.e :
the salient features of the three diagnostic classes of artifacts which disti:-
guish the Bilbo I subphase.

Pottery: One of us (DLC) has studied the ceramic fabric of fiber temnperec
pottery in detail. This study revealed that pottery from the coastal area, as
compared with other fiber tempered pottery making complexes in the _New.
World, had the highest content of fiber inclusions, and the lov.est sand arai.-
content (except for those from Saint Simrons, Georgia) and the highest a:-.n



Fig. 4. Bilbo phase artifacts: a-d, clay objects from 9Ch6l; e-f, socketed
antler tools from 9M 87, bone pin and awl from 9Ch35;, projectile
point from surface 9Ch61; j-l, stone tools from 9Ch35 and 9McI87.
of inter-site variation (Crusoe 1972). Our second comment is related to the
finding of sand tempered ceramics in association with fiber-tempered ceramics
in coastal marsh middens. Sand tempered pottery does occur with fiber tem-
pered wares though in a low frequency. In passing we would like to make note
of the clay fabric study by Schneider. Schneider (personal communication)
radiated (low flux) a clay object and potsherd from Pagan Plum Point which
seemed to indicate that the clay source for both was the same. We conclude,
on the face of this and other evidence (Crusoe 1972) that the fibrous material
in fiber tempered pottery was not native to the clay as Weaver (1963) suggests,
but instead was intentionally added. Further along these lines, it is of interest
to note that the baked clay objects which have been recovered from the Shell
Mound Archaic sites are not fiber tempered but are sand tempered.


Baked Clay Objects: As mentioned above, the baked clay objects are sand
tempered rather than fiber tempered. In general, they range from one inch to
less than three inches in diameter. They are rather reddish to matt grav in
color, and the external surfaces are irregularly formed. The typical for is
more elipsoidal rather than sphereoid. The surfaces are irregular and "grooved"
perhaps to facilitate handling if used in cooking (boiling) or perhaps the grooves
were functional in increasing the surface area of the object or both. All of these
clay objects were crudely formed -- they were formed by modeling -- and none
were decorated (Williams 1968: Fig. 92 d, p- 276).

Bone Pins: The Bilbo report describes Bilbo phase bone pins in detail There-
fore we refer the reader to that report for further discussion of this class of
artifact (Williams 1968:169).

Other Bilbo I Characteristics: At the Pagan Plum Point site a burial as un-
earthed by one of us (CBDeP). The "burial" unlike the Bilbo "burials" was
represented by a complete skeleton. No pit was associated with this burial.
It appeared to have been "tossed" on the shell heap and covered with a layer of
shell, as little care was taken to preserve the remains or to place the,- in an
area of low traffic.

"Food" mammal remains at Bilbo I subphase sites are con paratively-
scarce. This scarcity should be compared to the abundance of nmamimal re~ alns
in inland sites. Judging from the remains unearthed, the primary sources of
food during the Bilbo I subphase were shellfish, crabs, and fish.

Stone is not native to the Georgia coastal marsh area and such finds must
be attributed to trade or inland excursions. Given the relative scarcity of stone
in Bilbo I subphase sites (excluding Bilbo), it is possible to postulate that
coastal-inland contact during the Bilbo I subphase was minimal.

Shell tools--Busycon adzes and grubbers and oyster shell scrapers--
form an important section of the Bilbo I tool kit (Fig. 5).

BILBO II (ca. 1700 1000 B. C )

Our survey work has produced a number of sites which fit into Warin.' s
Bilbo II subphase. These sites are: A. Busch Krick, Cane Patch, Charlie
King, Cannon's and Butler Point Midden and other small middens on Saint Si-
mons Island. Sites whose collections are no\w at the University of Georgia
which fit into the Bilbo II subphase are those from Wilmington Island such as
Oemler Marsh Midden. This site, in addition to the upper levels of Bilbo and
Sapelo, comprise the inventory of Bilbo II subphase sites. All of the sites are
shell middens, save A. Busch Krick, Oemler, and Sapelo, ,which are shell
rings. As can be seen, the distribution of sites encompasses the entire -Georia
coastal marsh area. Temporally, the only carbon dates which are applicable to
the Bilbo II subphase are ours from the A. Busch Krick shell rin0 on CreiGhton


Fig. 5. Hafted shell tools from Pagen ^ ,, "rJ
Plum Point and Wilmington Island sites.

Island. The first date is 321580 B. P. (UGa-ZZ6) or 1265 B.C. This date was
obtained from a conch shell at the base of the midden 5. 7 to 6. 0 feet, Pit 1).
The second and third dates were apparently contaminated, the one at 3' ca.
1000 B.C.; and the other at 6' was ca. 600 B.C. As can be seen, the dates
are reversed, but there was no evidence of midden disturbance.

The diagnostic characteristics of the Bilbo II subphase are: decorated
fiber tempered pottery and elaborately decorated bone pins. We again refer
the reader to the Bilbo report (Williams 1964:169) for a discussion of the bone
pins (Fig. 6 and 7). Most of the remaining pages of this section shall concen-
trate on the decorated ceramics (Fig. 8-10).

The Pottery: The most striking feature of Bilbo II subphase sites is the de-
gree of variation of surface decoration of fiber tempered ceramics. So diverse
are the design patterns that our preliminary attempts at classification resulted
in a short descriptive paragraph discussing the particular design of each pot-
sherd. It became evident to us that we were over-describing the various de-


Table 3. Patterns of decoration on Bilbo fiber tempered pottery.


Simple incised
Zoned simple incised
Compound incised
Zoned compound incised
Curvilinear incised
Zoned curvilinear incised

Simple stab and drag
Zoned simple stab and drag
Compound stab and drag
Zoned compound stab and drag
Curvilinear stab and drag
Zoned curvilinear stab and drag

Simple stamped
Zoned simple stamped
Compound stamped
Zoned compound stamped

Simple Punctated
Zoned simple punctated
Compound punctated
Zoned compound punctated
Curvilinear punctated
Zoned curvilinear punctated

Random punctated
Zoned random punctated

Simple incised and punctated
Zoned simple incised and punctated
Compound incised and punctated
Zoned compound incised and
Curvilinear incised and punctated
Zoned curvilinear incised and

Simple punctated in incised line
Zoned simple punctated in
incised line
Compound punctated in incised line
Zoned compound punctated in
incised line
Curvilinear punctated in incised line
Zoned curvilinear punctated
in incised line

signs, and that perhaps we might be able to devise a "system" which would
simplify classification and at the same time form a standard for inter-site c
prison. Table 3 is a listing of the "system" which we devised. Some expla a-
tion of our categories will be required. In our view, \e see no reason to co3-
tinue to classify fiber tempered pottery by nmicro-decorative attributes.
stead we feel that the "key" to any useable typology of fiber ten-pered potter
lies in the description of the over-all geomretrical pattern.

We found that four basic patterns can be used to describe any cesi:.
pattern found on Bilbo fiber tempered pottery. We call these: siple, co-
pound, curvilinear, and random patterns (Table 3). A simple pattern co -
sists of incisions, stab and drag punctations, stamping, or w-hat have -.
run in parallel lines to each other (Figs. a, c f, a,i- ;: 9,b-c) A c :. -
pound pattern involves a simple pattern and at least one pattern v.iiih ru- a
right angles to the simple pattern (Figs, 7, b; 8, c; 9, j). Curvilinear a:d
random patterns need no explanation (Fig. 9, e, g). One major variat-o:t
which we found is zoning (Fig. 7, h; 8, d; 9, j). Zoning can occur .with a:1.-
of the patterns. We feel that this "system" \ill aid inter-site com:pariso:-s.


) 1 2

Fig. 6. Decorated bone lines from 9Ch35 (upper two) and from 9McI87.

Other Bilbo II Characteristics: No burials have been unearthed in Bilbo II sites.

Comparatively speaking "food" mammal remains appear to be more fre-
quent in Bilbo II sites. Deer (Odocoileus virginianus) and raccoon (Procyon
lotor) occur in high frequencies. Crabs and fish appear to diminish in frequency
as does the size of the shells. In the upper levels, lenses and zones of "crushed"
shell and smaller sized bivalves are encountered. Also the frequency of peri-
winkle (Littorina sp. indef.) appears to decrease in the more superficial levels.
These changes occurred in every Bilbo II midden which we excavated except
Cane Patch. It could be that the reduction of frequency represents over-exploi-
tation of this food source or an increased utilization of bivalve shell as tools,
or as we suggested earlier that the coastal marsh environment was going
through a state of change.

During the course of this paper we have made several points which we
shall summarize here. First, we feel that the geo-environmental evidence
limits the times of possible occupation of the Georgia coastal marsh by Shell
Mound Archaic people to a beginning around 2200 B. C. Secondly, we feel
that Waring's separation of coastal and inland complexes is valid. And finally,
we are in accord with Warings two subphase division of the Bilbo Phase.

References Cited

Beasley, Dana
1971 Skidaway Island: A Research Prospectus. University of
Georgia Publications. Athens.


S1 2 3 4

Fig. 7. Decorated pottery

from Cane Patch Island.

Bloom, Arthur L.
1971 Glacial-Eustatic and Isostatic Controls of Sea Level Since
the Last Glaciation. Late Cenozoic Glacial Ages, Karl K.
Turekin, Editor. New Haven.

Bullen, Ripley P. and H. Bruce Greene
1970 Stratigraphic tests at Stallings Island, Georgia. Florida
Anthropologist, Vol. 23, No. 1, pp. 8-28.

Caldwell, Joseph R.
1970 Chronology of the Georgia Coast. Southeast Archaeological
Conference Bulletin, No. 13, pp. 89-91. Morgantown.

Claflin, William H., Jr.
1931 The Stallings Island Mound, Columbia County, Georgia. Pa-
pers of the Peabody Museum of Archaeology and Ethnology.
Harvard University, Vol. 14, No. 1. Cambridge.


pottery from A. Busch Krick.

Crane, H. R.
1956 University of Michigan, Radio Carbon Dates I. Science,
Vol. 124, pp. 664-672.

Crane, H. R., and James B. Griffin
1958 University of Michigan, Radio Carbon Dates III.
Vol. 128, pp. 1117-1123.
1963 University of Michigan, Radio Carbon Dates III,
Vol. 5, pp. ZZ8-253.



Crusoe, Donald L.
1972 Interaction Networks and New World Fiber Tempered Pottery.
A Ph.D. dissertation, University of Georgia, Athens.

Fairbanks, Charles F.
1942 The Taxonomic Position of Stallings Island, Georgia, Amer-
ican Antiquity, Vol. 7, No. 3, pp. 223-231. Menasha.


Fig. 9. Decorated pottery from Cane Patch Island.

Ford, James A.
1966 Early Formative Cultures in Georgia and Florida. American
Antiquity, Vol. 31, No. 6, 781-799. Salt Lake City.


"An Early Ceramic Site Near Beaufort, South Carolina". In-
stitute of Archeology and Anthropology Notebook, Vol. 1, Nos.
6-9, pp. 6 and 7. University of South Carolina. Columbia.

Holder, Preston
1935 W.P.A. Glenn County Project, Monthly Progress Report.
Archives, Southeast Archeological Center. Tallahassee.

Hoyt, John H.
1968 Geology of the Golden Island and Lower Georgia Coastal Plain
in The Future of the Marshlands and Sea Islands of Georgia.
Georgia Natural Areas Council and Coastal Area Planning and
Developments Commission, D.S. Maney, Editor.


Hoyt, J. H. and John Hails, Jr.
1967 Deposition and Modification of Pleistocene Shoreline Sediments
in Coastal Georgia Science, 155:1541-1543.

Hoyt, John H. Robert J. Weimer, and Vernon J. Henry
1964 Late Pleistoce and Recent Sedimentation, Central Georgia Coast,
U.S.A. Contribution No. 46, University of Georgia, Marine In-
stitute, Sapelo Island, Georgia.
1968 Age of Late Pleistocene Shoreline Deposits, Coastal Georgia.
Contribution No. 107, University of Georgia Marine Institute,
Sapelo Island, Georgia.

Johnson, A. Sydney, Hilburn O. Hillestad, Sheryl A. Fanning, and G.
Frederick Shanholtzer
1971 An Ecological Survey of the Coastal Region of Georgia, on file
National Park Service, unpublished manuscript.

Linton, T. L.
1970 Inventory of the Intertidal Oyster Resources of Georgia. In
T. L. Linton (ed.) Feasibility Study of Methods for Improving
Oyster Production In Georgia. Georgia Game and Fish Com-

Miller, Carl
1949 The Lake Springs Site, Colombia County, Georgia. American
Antiquity, Vol. 15, No. 1, pp. 38-51. Menasha.

Stoltman, James B.
1972 The Late Archaic in the Savannah River Region. In Florida An-
thropological Society Publications, No. 6. Gainesville.

Weaver, Elizabeth C.
1963 Technological Analysis of Prehistoric Lower Mississippi Ceramic
Materials: A Preliminary Report. American Antiquity, Vol. 29,
No. 1, pp. 49-56.

Williams, Stephen, editor
1968 The Waring Papers. Papers of the Peabody Museum of Archaeology
and Ethnology, Harvard University, Vol. 58. Cambridge.

August 23, 1974


George Percy

The purpose of this article is to describe the results obtained using
a mechanical earth auger for certain purposes on the Torreya site (8Li8),
a Weeden Island habitation site, located in Torreya State Park in the north-
western corner of Liberty County about in the center of the Florida Pan-
handle. The auger was extremely helpful in the preliminary exploration of
the southern end of the site, where the Florida State University Department
of Anthropology archaeological field school conducted investigations in the
Spring and summer of 1972. The northern end of the site had been pre-
viously explored by the field school through most of 1971 (Percy 1971a: 1471»:
1972a; 1972b). The auger was so helpful in the 1972 season that it was Zeit
its advantages were important enough to bring to the attention of other arch-

The Torreya site is located in the hilly uplands east of the Apalachicola
River, about one-half to three-quarters of a mile from the steep bluzis that

border the east side of the river (Fig. 1). The site is on a narrow ricge
that slopes away sharply to the east and west into the bottomlands of t4>
small creeks. The ridge continues beyond the site both to the north ané

south, and the terrain is much gentler in these directions, although the
stretch of level land is not very wide, and the ridge continues to drop 922:
sharply on the east and west beyond the limits of the site. The character
of the local topography can be easily seen from the U. S. Geological Survey
Rock Bluff Quadrangle map, where the site is mostly in the middle one-thire
of the swl/4 of Section 16, Township 2N, Range 7W. As can be seen Zrom
Figures 1-2, the site area is very heavily overgrown. Figure ! is an ob-
lique aerial photograph, looking west across the site towards the Apalachi-
cola River in the background. The bluff line shows clearly near the tap 2:
the picture. In the foreground is the main park road and the park ranger’ s
house and driveway; several trees partially obscure the house. On the far
right is the dirt road into the park campground. In right center, to the le::
(south) of the campground road is a small clear area (with open excavation
units just visible) which is part of the northern end of the site that was ex-
plored in 1971. Between this and the ranger's house is the southern portion
of the site.

Figure 2isa scaled map showing the main features of the southern ené
of the site. Concentrated refuse areas (discussed below) are outlined anc
indicated with capital letters. Also shown are the woods edge, the main
park road, the ranger's house and garage, and his dirt driveway.

Florida Anthropologist, vol. 29, no. 1, March 1976


00 R200 R300


Fig. 1. Aerial view looking
west across Torreya site to
Apalachicola River. Park
superintendent's house and
driveway are in left fore-

R400 R500 R600

Fig. 2. Map of
southern end of
Torreya site

/ Area F (') produced midden material '
and artifacts in several holes, but not all.
It may be a small separate area, or port of Area
. This will have to be determined by future excavation.


-200 -

-300 -

-400 -

-500 -

-00o -

-900 -

-900 -



The tree cover is mixed hardwood and pine forest, all secondary
growth. At all times of the year, pine needle and leaf cover on the ground
is heavy, and, of course, this makes it almost impossible to look for arch-
aeological remains on the surface and place any great degree of confidence
in negative results of surface search, The remains may be there, but it
is quite easy to walk over them without being aware of it. The situation is
obvious from the illustrations. Needless to say, this presents a major pro-
blem for archaeological investigations of the general upland area along the
east side of the Apalachicola River, since practically none of this land has
been cleared in recent years.

With regard to the Torreya site, the heavy ground cover of leaves and
pine needles presented a problem in attempting to determine the extent o:
the site through search for surface artifacts. Three areas of midden refuse
-- those labelled A,B, and C in Fig. 2 -- had been located by surface in-
spection, but these were in places that were relatively free of trees, and it
was not at all certain that they were the only refuse areas on the southern
end of the site, or that the site was coterminous with the edge of the woods.
One characteristic of all three of the areas seemed important, how ever.
This was the presence of large amounts of shell. From this it was reasoned
that, if other potential areas were characterized by shell refuse, they should
be easily recognizable, if only there were an efficient way of getting a look
beneath the leaf cover. "Efficient" meant not only being able to look under
the leaf cover in any one place, but also being able to cover a fairly large
area rapidly and systematically, since we had no idea of the overall size or
plan of the southern end of the site.

One common approach in a situation like this is to resort to extensive
exploratory trenches or test pits, the express purpose of which is to find
site boundaries. However, there was another characteristic of the site v.hih
raised the possibility that a program of exploratory excavation for this one
purpose could be tremendously time consuming. The three refuse areas,
(A, B, and C,) which were known to exist were also known to be distinct re-
fuse areas, separate from one another. They did not belong to one contin-
uous area of uniforly dense refuse deposition, although scattered artifacts
did occur between them. Consequently, it seemed quite possible that the
southern end of the site, however large it might be, was made up o: similar
small, individual, refuse areas, variously spaced and scattered over a siz-
able area of the ridge top. The amount of trenching and test pitting which
would have been necessary to determine this and, if found true, to determine
the boundaries and basic composition of each of these possible areas would d
have been an undertaking of several months.

In casting about for alternatives to excavation, the idea of the soil aug-
er was hit upon. It seemed that it might provide the ideal solution First


it offered a very easy way of penetrating the leaf and needle cover. It was
mechanically powered and would be much easier to use for extended periods
of time than manually operated augers or posthole diggers. Holes could
be drilled quickly with it, and therefore large areas could be covered in a
short period of time. Secondly, it was very inexpensive in terms of invest-
ment of time, labor, and capital, compared to the investments necessary to
perform exploratory excavation over corresponding areas, and even compar-
ed to the time involved in performing hand augering or chemical surveying.
Augering would also save survey time, since it would only require setting
occasional temporary stakes to lay out auger lines, as compared with the
amount of transit work involved in laying out many squares and trench lines
over a large area.

In general, the advantages of the auger seemed very great, provided
it would give a clear indication of the presence of refuse deposits, if and
when it penetrated them. The one disadvantage of the auger, as compared
with normal excavation methods, is that it bores only a small hole (for ex-
ample, we used a 4 inch bit) and, if the refuse is extremely low density and
does not include some easily recognizable material, such as shell or midden
soil, the results of augering might well be inconclusive. Many sites, espec-
ially towards the peripheries, are marked only by a thin scattering of pot-
sherds of stone debitage, and it is quite possible the auger would not be suit-
able for exploratory work in these circumstances.

Given our knowledge of the characteristics of refuse areas A, B, and C,
however, it was felt that the auger probably would be a reliable guide to the
presence or absence of other similar refuse deposits and a useful means of
exploring portions of the ridge surrounding these three areas.

As it turned out, these considerations were correct. The auger saved
untold excavation time and enabled us to define the general boundaries of the
southern end of the Torreya site, to determine the location of other individ-
ual refuse areas, and to obtain good approximations of the size and bound-
aries of each of these. It also indicated some variations in the composition
of these areas; some, for example, had a great deal less shell and other
kinds of midden refuse in them than others. Finally, it provided us with
basic soil profiles and miscellaneous information on the subsurface geology
of the site area. In some places, for example, perched water tables were
encountered quite close to the surface. These can be plotted on the site
map with reference to the distribution of refuse areas, and they probably ex-
plain why refuse areas (which are directly associated with house areas) are
not located in certain places.

All told in the process of augering in the summer of 1972, an area of
approximately 700 by 700 feet was covered, and six or seven additional re-


fuse areas were located. Nearly 700 holes \were bored, and the total length'
of time involved was about two weeks -work, four to six hours a day, by
three persons or, upon occasions, two persons. In the remainder o thnis
article the equipment and procedure followed in augering are described.

The particular auger used (Figs. 3 and 4) w'as a Sears' One-nan
Earth Auger, Model Number 147.142001. There are similar augers, manu-
factured by other companies, on the market; this particular one \%as used
simply because it was already owned by the Department of Anthropology,

A list of specifications for this model can be obtained from Sears. Roe-
buck and Co. It has a 3 HP Tecumseh VII engine %aith recoil starter, nan-
ually operated throttle allowing various speeds, and a centrifugal clutch
which disengages if the auger strikes an obstruction or is overloaded- Var-
ious size augers are available, ranging from 15/8 inches to 9 inches i di-
ameter; we used the 4-inch size. The standard auger length is 30 inches.
but extensions are available. The weight of the entire assembly (. without
extensions) is approximately 35 pounds. Total height (again, %without the
extension) is about 3 1/2 feet. Although it is advertised as a one-man a:uer,
if it is to be used for any length of time, t\\o people are necessary to oper-
ate it. Otherwise the weight becomes burdensome. Also, the auger occas-
ionally gets tied up in tree roots or stuck in compact soil, and two people
are necessary to pull it up. In our experience, we found \work went fastest
with three people: two to operate the auger itself, and one to record the re-
sults. This arrangement is shown in Figs. 3 and 4.

In general, the Sears auger was suitable for the ,work we were ding,
although if we were to invest in another machine, we would probably be in-
clined to buy a slightly heavier-duty model. Also, we were not impressed
with the recoil starter assembly in the present model, since it contained an
aluminum pulley wheel which broke on several occasions. Sears may have
improved on this design feature in more recent models.

The field procedure involved in the use of the auger was quite sin ple.
Basically, it consisted of drilling holes at regular intervals along selected
grid lines and recording the results of each hole, the record being made up
of certain standard observations, as well as any miscellaneous information
felt to be of possible significance.

Before each hole was drilled, the leaf cover in the immediate vicinity
was pulled away and a careful search was made for even the smallest sherd
or bone or tiniest fleck of shell. This was done for t'wo reasons: first, as
a check on the augering results; second, to compare surface indications in
any particular place with auger results in the same place as a check on the
reliability of surface observation. It was often found that surface search


and auger results tended to coincide. Where the auger produced material
there was frequently some corresponding indication on the surface under
the leaf cover, though this might be in the form of only a single tiny fleck
of shell; generally, however, the auger gave a better indication of sub-
surface density of refuse. Also, surface observations were frequently not
reliable, since many holes produced small pieces of shell or a sherd or
two, where the surface had shown nothing. This was most true where
there was a thick zone of recent humus beneath the leaf cover, or where
the refuse deposits were very light, even in the subsurface. Overall, the
auger gave a more reliable and much more easily interpretable picture of
the distribution of archaeological refuse, both in qualitative and quantitative

Following the brief surface search, the auger hole was drilled. In
practically all cases, the hole was drilled to the maximum possible depth
without the extension; this was approximately 33 inches, slightly more than
the 30 inch length of the auger itself. For each hole, the following infor-
mation was recorded: (1) presence or absence of shell or other artifactual
material, however little or small, on the surface, within about a 2 foot
radius around the hole; (2) presence or absence of shell or other artifact-
ual material sherdss, bone, stone, charcoal, midden soil, burned earth,
ash, etc.) in the auger hole; (3) depth at which artifacts occurred; (4) thick-
ness of the zone in which artifacts occurred; (5) changes in the color and
texture of soil thrown up by the auger and, also, the approximate depth at
which these were observed; (6) an estimate of the relative density of art-
ifactual material, if present; (7) miscellaneous information on cultural re-
fuse, such as whether or not animal bone or shell was burned, the type
classification of potsherds, etc.; (8) miscellaneous geological information,
such as the occurrence of a perched water table or heavy root disturbance.

By careful observation it was possible to recognize soil changes quite
reliably when they occurred. When such changes were observed, the auger
was pulled up out of the hole, and a depth measurement was taken. It was
found that it was necessary to subtract 0.15-0.20 feet from each measure-
ment to approximate the correct depth at which a soil change first occurred.
This correction factor is made necessary by the fact that the auger contin-
ues drilling, while soil it has already penetrated is working its way up the
hole. The correction factor is more critical with depth. Above approx-
imately a foot below surface, the correction was not found to be necessary.
Subsequent checking by excavation showed that the basic soil texture and
color changes recorded with the auger, and the depths at which these occurr-
ed, were quite reliable. Depth measurements were never more than 0.5
feet off. Textural and color descriptions usually had to be amended only
to account for mottling or fine lensing (and, of course, detailed bedding


Results at Torreya

The procedure for determining the location of auger holes at the
Torreya site was a combination of an attempt to explore the ridge top in
a systematic way, employing consistent intervals between auger holes and
augering lines, and knowledge gained about the site as augering progressed.
Also, there were certain areas that could be eliminated on other grounds.
These included, for example, the hardtop road that ran through the park
and also, small, steep, deeply cut ravines, especially on the east side of
the ridge (Fig. 2). Many of these were the locations of small springs and
small flowing rivulets tributary to the eastern creek and were obviously
not places where cultural activities would have been located or remains of
activities preserved in situ. They were inspected for evidence of sinrple
refuse dumping, but no indications of this were found in any o: then-. In-
spection consisted mostly of careful surface observations and very little

The augering was begun by working out in all directions from the three
known areas. Whenever refuse was encountered at points along a major
line of auger holes, short supplementary lines, perpendicular and parallel
to the main line, were drilled in order to determine the size and boundaries
of refuse deposits.

The placement of some of the major auger lines is shown in Figure 2,
relative to the refuse areas that were ultimately located and some of the
other features of the southern end of the site. The figure does not shove. all
of the auger lines; particularly, it does not show all the many short lines
drilled to determine the boundaries of individual areas, once located. -ach
line represents a set of holes, generally spaced about -en feet apart, except
where obstructions, such as large trees or a yello\ jackets' nest,
encountered. As can be seen from Figure 2,the spacing between major auger
lines was generally 50 feet on both the east and west sides of the site.
This interval was deemed satisfactory on the basis of \what %was discovered
about the minimum size of refuse areas.

No augering was done north of the area shown in Figure 2, since tne
northern part of the site was explored by excavation in the sumrner o: .
Also, very little augering -was done south of about -850, because, except i
the ranger's yard (see Fig. 1), this was a very heavily dissected area t:a:
seemed particularly unsuitable for settlement. The one exception is the lire
shown running to the southeast, from -850 to -1100, on the east side of the
road. This line is positioned so that it runs do\n the center of a narrDw.
spur of the main ridge. It is bounded north and south by ravines, and it
runs down to the eastern creek. In the creek at the end o- the ridge a nurn-
ber of sherds were found. However, augering down the ridge, and supple-


mentary augering and surface search of the banks of the creek upstream
and downstream from where the sherds were found produced no artifacts
of any kind. At present, we are at a loss to explain where the sherds
are coming from.

On the west side of the site, the augering was halted after the ridge
top began to slope sharply away. On the east, augering was stopped after
several hundred feet from the hardtop road, because nothing was being
found and because the land began to slope more sharply. In the area be-
tween -200 and -500 and west of the north-south 0 line, no augering was
done because the ground was wet in many places and the water table was
quite close to the surface. Springs broke through in the area between -200
and -300 and west of L150 (Fig. 2).

As a result of the augering, the southern end of the Torreya site, that
portion south of the -200 line, is seen to consist of nine, or perhaps ten,
refuse areas. These areas are strung out in roughly a north-south line
approximately 550 feet long and 300 feet wide. All areas are roughly oval
in shape, and they vary in size from approximately 60 feet to 160 feet in
long diameter. All of the areas are marked by shell, although other kinds
of refuse are present, including animal bone, charcoal, potsherds, and stone
debitage. This was determined by the auger without any other type of ex-
cavation. In one instance, the auger also showed the existence of a deep
midden filled pit. The augering has indicated variation in the density of re-
fuse among different areas. In general, shell and other refuse are densest
in the southernmost areas (Fig. 2, A-E) and become progressively less
dense, moving from G and H to I and J. This is also confirmed by the 1971
excavations on the northern end of the site, where refuse was uniformly light
though still patterned into areas of concentration. It is important to empha-
size that the definition of the boundaries of these areas at the southern end
of the site is based almost exclusively on the results of the augering. This
is particularly important for those areas containing lighter amounts of refuse.
The boundaries seem fairly accurate in most cases, because the augering
was carefully done, but they will ultimately have to be checked by excavation.
This has already been done for areas A and G, and the augering results have
proved to be very reliable. The other areas still remain, but in the mean-
time, the soil auger has provided us with a good working picture of the basic
size, composition, and organization of the site and has saved a great deal
of time consuming and expensive preliminary exploratory excavation.

References Cited

Percy, George W.,
1971a Current Research: Florida. Newsletter of the Southeastern
Archaeological Conference, Vol. 15, No. 1, pp. 7-8. Morgantown.


1971b Preliminary Report to the Division of Recreation and Parks,
Department of Natural Resources, State of Florida, on Arch-
aeological Work in Torreya State Park during the Year 1971
by the Department of Anthropology at Florida State University.
Tallahassee: Department of Anthropology, Florida State Un-
iversity. Mimeographed.

1972a A Preliminary Report on Recent Archaeological Investigations
in Torreya State Park, Liberty County, Florida. Paper pr-
sented at the 24th Annual Meeting of the Florida Anthropo-
logical Society, Winter Park, Florida.

1972b Current Research in the Southeast: Florida. Newsletter of
the Southeastern Archaeological Conference, Vol. 16, No. 1,
pp. 3-6. Morgantown.
Tallahassee, Florida
February 4, 1975

Fig. 3. Three-person crev
preparing to use auger.
Note ground cover and
cleared space below bit.

Fig. 4. Aager at maximum
penetration. If caught in
compact soil or roots, two
people may be necessary to
remove it.


Ripley P. Bullen

Recently, I have revised, reorganized, expanded, and brought more-
or-less up to date "A Guide to the Identification of Florida Projectile Points. "
Major changes include the addition of eleven new types, a realignment of the
pages so that similar types are more apt to be found near each other, and a
reassessment of culture periods. The last is based more on typological than
on stratigraphic data.

While doing this, I realized that projectile points have a much more in-
teresting and important story to tell than only their chronological position in
Florida. A strong case can be made that various types reflect diffusion of
ideas if not of people to Florida. Apparently migrations of small groups into
Florida has been a factor in acculturation in the state for nearly 10,000 years.

I have divided the Paleo-Indian period into an early and a late phase.
Clovis points, both fluted and unfluted, appear in the earlier and Simpson,
Suwannee and possibly Santa Fe points in the later phase. This is in agree-
ment with the situation in other parts of the country but it does not mean these
points will not in the future be found together in a closed context. Clovis is
only earlier on the average; and if makers of Clovis points arrived in Florida
after Simpson and Suwannee points had been developed elsewhere, they may
have brought all three with them. This arrangement is, of course, similar
to that presented by Hemmings (1971) at the Eastern State Archeological Fed-
eration meeting held in Gainesville in 1971.

The next or Dalton period presents intriguing possibilities. Not only is
there a distinct break typologically with the earlier Paleo-Indian period but
there is also the suggestion of two lithic traditions. One is the Dalton tradi-
tion proper with Santa Fe, Tallahassee, and Dalton points. The other, based
on relatively thick blanks, is a notched tradition with Greenbriar, Bolen, and
Hardaway points. What must be an early example is the Union Side Notched
which is suggestive of shallow notches added to the sides of Suwannee points.
However, Union Side Notched points are bifacially beveled while Suwannees
are not.

This use of beveling to form side edges on thick blanks, while present at
the Stanfield-Worley shelter in Alabama (De Jarnette, Kurjack, and Cambron
1962:49, 53), does not seem to be as common elsewhere as in Florida. Per-
haps it is an adaptation to low grade chert by chippers of points who vere used
to better materials. Relatively early side notching on the contrary is fairly


Florida Anthropologist, vol. 29, no. 1, March 1976


common to the north and northeast. This suggests diffusion of the notches
tradition southward along the Atlantic Coastal Plain or the Chattahoochee-
Apalachicola River valleys,

The Dalton complex is much better known as a result of recent or-<.
particularly that published by Don F. Morse (1973). Sone archaeolois:s i-
clude the Dalton culture--even Bolen points--under the Paleo-cIdian rbe.
With this I do not agree, as I believe the lithic industries and faunal assio-
tions are different. Santa Fe and Tallahassee points clifer fro:n Paieo-I ndan
points in having transverse parallel or collateral flakin- This nay seen like
a difference in degree and not in kind, but certainly sharp-ernir._ serration
or beveling is not a Paleo-Indian trait. It also nay be poiJntec out trar hi:s
from plowed fields and rivers suggest a closer association of ole pits
with Archaic rather than with Paleo-Indian artifacts.

There seems to be no question abour the locus o: or-:: o_ the Dai~
complex. It is clearly in the Mississippi craina probaDol in :ts t.,-st -
portion. It, undoubtedly, entered Florida alone, the Gulf Coastal Plai.
included various small tools such as extremely small enc scrapers It. _.:-
cave edges (Goodyear 1973, Bullen and Beilman 1973).

There is no persuasive evidence as to whether r te side- che a--
tion or the Dalton complex has priority in Floric a Both i-opha=i -:-a al
grinding. One thing is certain, there are n any nore _lie- ?oil-. '_- -j -'
eled and unbeveled, than all the other Florica Dalton vari-ties p-t toe h-.r
Perhaps they were used concurrently for different functions. Certa -
method of hating was different. Perhaps the Bolens, v which aret son'-:
serrated, represent an amalgamation of the tw o traditions.

The genesis of the Archaic around o000 B. C. seen s to oe a per(:-
experimentation with roughly flaked Florida Spikes, Thono'tsassa, Arr--
dondo, and, possibly, Hamilton points. It is hard to understand ch. thes-
points were made. They seem too heavv for atlatl spears. nor do they ap-
pear to be a necessary developmental statg in the invention of stcnm ecn pDints
They have been placed in an early precerarnic Archaic per-iod because c -:-:
apparent lack of association with stemmed points at the Fletcher Davis site
(Warren 1971). They do resemble the Kirk Serrated points in thickness ani
cross section but there is the possibility they w ere made to function as ha-ied
digging implements. If this should be the case, they would be more apt to ca:
around 1000 B. C. instead of 6000 B.C. Alternatively, they mav indicate an
unknown procurement of root food during the Archaic.

The Kirk Serrated point still seems to be the earliest real sten n:r e
point in Florida. It is massive and predominantly supplied w ith serrated
edges. Serration here, as for the earlier Dalton-Tallahassee-Bolen points


is a method of sharpening edges. Another apparent technological holdover is
the beveling on Hardee Beveled points. As might be expected these points
have thicker cross sections than is customary for Florida stemmed points.
The roughly chipped Hamilton points are intermediate between Arredondo and
Florida Archaic Stemmed points. Apparently there were several groups living
in Florida in Archaic times with somewhat different lithic technologies. They
accepted the presumedly introduced idea of stemmed points as a superior haft-
ing technique, but finished them in the older style--some beveled, some ser-
rated, and some coarsely flaked. This is very reminiscent of diffusion and
acculturation. This diffusion occurred, apparently, along the Atlantic Coastal

The time for these mixed or transitional specimens was very'brief and
shortly the stemmed points with which we are all familiar became the mode.
Throughout the rest of the preceramic Archaic very similar points are found
all the way from Florida to central Maine but north, south or centralAtlantic
state priority is not evident. Similar points are, of course, found west of the
Appalachian Mountains but not in the quantities found between those mountains
and the ocean. The apolgee of this period in Florida, as far as lithic work-
manship is concerned, occurred around 3400 B. C. at which time Newnan and
Hillsborough points have been radiocarbon dated at two sites.

As we approach the close of the Archaic, points which are corner or
basally notched occur. The first or Culbreath subtype 1 shows definite sim-
ilarity to the Eva I point of Tennessee and its development in that general
area seems likely. The typical Culbreath has a short stem, excurvate blades,
and drooping barbs. These points, or their concept, were certainly intro-
duced into Florida during the reign of the Archaic Stemmed points. The latter,
somewhat less well made, also hold over into the Orange period.

Culbreath points apparently were first introduced into north Florida
and the general Tampa Bay area. There progressive modification appears to
have occurred. The first modification was to form an expanding stem which,
in effect, forms a corner notched point. Judging from the Zabski site (Atkins
and MacMahan 1968) this had occurred before 960 B.C. The next change was
to extend the corners to the base to form the basally notched Hernando and
Citrus points found with Perico Linear Punctate pottery along the Gulf north-
west of Tampa Bay. This appears from present knowledge to be a local Flor-
ida development.

The next important innovation was the introduction, apparently from
southwest Georgia of Florida Copena points. These trianguloid points are
hard to place typologically but they were found by Fairbanks (1954) with Car-
terville Checked Stamped pottery at Site 9HL64 in the Buford Reservoir. In
Florida they occurred at the Sunday Bluff (Bullen 1969:33, 50, "Jackson-like")


and Zellwood sites (Dreves 1974:74, "Pinellas-like"; Bullen, John, and
Brooks 1974:16), They appear to be small copies of Adena points as found
at the Spring Creek site in Tennessee (Peterson 1973), Migration of people
with a Copena-like culture from Tennessee to western Georgia to central
Florida might be hypothesized on their distribution. If so, it took then an
appreciable length of time and they acquired check stamped pottery on the way.

A similar route may be suggested for the few Adena and Hopevellian
forms which have been added to the Guide as Florida Adena and Ocala points
There are certainly enough Adena and Hopettellian traits found in Florida so
that recognizable projectile points or ceremonial knives of those cultures
will not surprise anyone.

I am not going to say much about other projectile points found in Flor-
ida since the time of Christ. In general they are small form s--frequently
small copies of older forms--which suggest the use of the bow and arrow.
Very little is known about Broward and Gadsden points. The narrow. and
somewhat unique Duval points have their counter parts to -he west. The small
triangular Pinellas point is the latest and was still made when the Spanish ar-
rived. It represents a Middle Mississippian culture and spread from that re-
gion along the Gulf coast to be found in most of Florida. Fronm the technolo.-
ical standpoint it represents the ultimate in efficient manufacture but rarely
is it an art object.


Diffusion has been used in this paper to explain chronological chanLes
in Florida projectile points. In general these changes are similar in se-
quence to those reported for other areas of the southeast. In other v~ords
the southeast has been an interacting area for nearly 10,000 ears.

Judging from projectile points, the process of change has functioned
by a series of "fits and starts. New projectile point forns or n ltifor
complexes or old forms with new modes have been developed somv he re i:
the region--or introduced into it from further north or v.e t--and have t rad-
ually diffused throughout most of the region. In the process or in the recip-
ient areas, regional variations have occurred. After these periods of difcu-
sion or transition, there seems to have been long periods of stability duri:
which projectile point typology remained constant w ith, of course, variations
around the ideal mode. During this time certain traits or trends nma bec ome
more emphasized in one area and others more emphasized in other areas.
Old forms also tend to still be made, especially for certain functions. hen,
for some reason, the process repeats itself. The end result, nearly u:niver-
sally, seems to be small triangular points.


This process of independent invention and diffusion is not limited to
projectile points but occurs in practically all phases of culture, both mun-
dane and ceremonial. It may be easily studied with ceramics but tight typol-
ogy is necessary. There is also the complication that regional conservatism
becomes a greater factor in later as opposed to earlier times. Unless the
new idea or article has obvious functional advantages, people are less apt to
accept it.

There is one more point I wish to make. Projectile points are, of
course, only one type of artifact. To demonstrate a migration of people as
opposed to trait diffusion it is necessary to have a complex of non-ceremonial
artifacts introduced into an area where modally similar artifacts were not
previously present. It is also helpful to have a suitable donor culture some
distance removed,

The introduction of the Dalton complex into Florida may represent an
expansion of people into the state as other new tools as well as new projectile
points can be posited. However, to be sure we need a collection of various
types of Dalton-like tools from one site or part of one site without too much
extraneous material (Bullen and Beilman 1973).

There seems to be little evidence for mass immigration into Florida at
any subsequent period. The case for Copena invaders is weak. Adena and
Hopewellian influences or trade are evident but tend to be ceremonial in im-
plication and not to approach complete trait complexes. The Middle Missis-
sippian culture may have introduced a new variety of corn with attendant
ceremonies but it did not arrive as a ruthless invading force. Temple mounds
and small triangular points were present before handled, casuela-shaped ves-

References Cited

Atkins, Stephen C. and Jeannie MacMahan
1967 The Zabski Site, Merritt Island, Florida.
Florida Anthropologist, 20:3-4, 133-45.

Bullen, Ripley P.
1969 Excavations at Sunday Bluff, Florida. Contributions of the Florida
State Museum Social Sciences, No. 15. Gainesville.

1975 A Guide to the Identification of Florida Projectile Points. Revised
Edition. Kendall Books, Gainesville.


Bullen, Ripley P. and Laurence E. Beilman
1973 The Nalcrest Site, Lake Weohyakapka, Florida. Florida Anthro-
pologist, 26:1, 1-22.

Bullen, Ripley P., Otto Jahn, and Mark J. Brooks
1974 Some test at the Zellwood Site (Or-17) beside Lake Apopka, Florida.
Florida Anthropologist, 27:2, 62-66.

De Jarnette, David L. Edward B. Kurjack, and James W. Cambron
1962 Stanfield-Worley Bluff Shelter Excavations. Journal of Alabama
Archaeology, 8:1-2. University.

Dreves, Rick
1974 Archeological Investigations of 8-OR-17: An Early Aboriginal Camp-
site on the Shore of Lake Apopka, Florida. Florida Anthropologist,
27:2, 67-76.

Fairbanks, Charles H.
1954 1953 Excavations at Site 9HL64, Buford Reservoir, Georgia.
Florida State University Studies, No. 16, pp. 1-26.

Goodyear, Albert C.
1973 Archaic Hafted Spokeshaves with Graver Spurs from the Southeast.
Florida Anthropologist, 26:1, 39-44.

Hemmings, E. Thomas
1971 Early Man in the South Atlantic States. Bulletin of the Eastern
States Archeological Federation, No. 31, pp. 10-11.

Morse, Dan F.
1973 Dalton culture in Northeastern Arkansas. Florida Anthropologist,
26:1, 23-38.

Peterson, Drexel A.
1973 The Spring Creek Site, Perry County, Tennessee: Report of the
1972-1973 Excavations. Memphis State University Anthropological
Research Center, Occasional Papers, No. 7;

Warren, Lyman O.
1971 The Fletcher Davis Site. Florida Anthropologist, 24:2, 81-90.

Gainesville, Florida
February 15, 1975


Thomas J. Padgett

That portion of the Florida Gulf Coast lying between and including
the territory around Tampa Bay and Charlotte Harbor has long been rec-
ognized as a locality rich in archaeological deposits. Prehistoric shell
heaps and earthworks located there have attracted the attention of natura-
lists and archaeologists since the mid-ninteenth century and were the sub-
ject of early publications on Florida archaeology (Calkins 1877-1880; Walk-
er 1880; Moore 1900, 1905; Hrdlicka 1922). Major excavations were under-
taken in the area as federal relief projects during the 1930's. These were
reported in Gordon R. Willey' s comprehensive 1949 study while eleven were
the subject of Ripley P. Bullen's 1952 report. From the late 1940's through
the 1960's occasional excavations and surveys were carried out by the Flor-
ida Park Service and the University of Florida. In the more recent past
sites have been recorded primarily by archaeologists from the University of
Florida, University of South Florida and the Florida State Museum, often
relying upon information and assistance from local collectors.

With few exceptions all of the archaeological investigations mentioned
have been concerned with sites located along or within a few miles of the
coast. This is entirely reasonable considering the fact that the prehistoric
population was undoubtedly concentrated along the coast for at least the
last 2,000 years. The advantages that a marine adaptation offered were
great and will be discussed later in this paper. Within the last two years,
however, a number of surveys have been carried out in the interior sec-
tions of this area which can aid in the explanation of aboriginal land use
and subsistence for the entire region.

The Safety Harbor Period

The chronology of the Central Coast-Manatee Region has been well
established (Willey 1949; Goggin 1947; Bullen 1951, 1965, 1973 Griffin
1952) and shows evidence of occupation from the Paleo-Indian through
historic periods. In this paper I shall consider only the Safety Harbor
period or phase, which has a time span beginning around A.D. 1500 (poss-
ibly earlier) and lasting into the historic period, perhaps as late as
the early 1700' s.

The Safety Harbor phase has been previously defined, and many com-
ponents of the phase have been reported (Griffin and Bullen 1950; Sears
1967; Bushnell 1962; Bullen 1951). The most recent synthesis of the Saf-
ety Harbor phase was accomplished by Albert Goodyear (N.D.). Although

Florida Anthropologist, vol. 29, no. 1, March 1976


Goodyear deals only with the Tampa Bay vicinity, focusing upon the Toco-
bago Timucuans specifically, he presents a basic ethnohistorical recon-
struction of the political and religious manifestations of the Safety Harbor

Goodyear presents a model of the Safety Harbor settlement patter.
in what might be called a modified central town analysis. This is to sa,
a model is presented to show a pattern of large tons \ith temple mounds,
surrounded by smaller villages whichh depend upon the temple-ton .s or
political and religious leadership services. He places Safety Harbor sites
in six "eco-mode" categories: Bay-Beach, Bay-Island, Estuary, Inland-
River, Inland-Creek, Inland. Goodyear lists all of the temnple-to\n sites
(Safety Harbor, Hirrihigua, Navarez, Maximo, and Terra Ceia) in h:s
"Bay-Beach" ecomode and proposes a sequencial occupation of these s-tes
with periods of overlapping occupation where two temple-tow ns n av have
been operating at the same time.

Goodyear superimposes an east-west ecological dicotomy in the settle-
ment pattern, pointing out the advantages for agriculture on the eastern
side of Tampa Bay, though not dismissing the possibility that sone naize
agriculture was carried on in the western section (Pinellas Co.). As evi-
dence of this he points out the superiority of soils on the eastern since Do
the bay and the larger number of shell middens on the western side. iml-
ing a greater reliance on shellfish for the western side of Tamnpa Bay.
Goodyear emphasizes, however, that more detailed control data on so1is,
plant ecology, and oceanographic factors, as well more site data in the for
of faunal, botanical and shell analyses, will be needed to test this h-pothesis

Goodyear's model of settlement pattern and its relationship to the
socio-economic system is of definite value to the understanding of the c:lt
ural dynamics of the Safety Harbor phase. However, he limited the scope
of his analysis to the immediate Tampa Bay vicinity. While this enco: opass-
es the major portion of the Safety Harbor site distribution, a large anoun~ t
of land to the south and east is omitted. If the people represented by the
Safety Harbor phase were utilizing this territory then it should be possible
to gain additional information on their cultural system.

The eastern boundary of the Safety Harbor phase is not ell defined,
although it now appears that it lies a bit to the west of Willey's boundar-
ies for the Central Coast and Manatee Regions (Willey 1949, Iap I).
would place this boundary to include much smaller sections or Hardee, Polk.
and DeSoto counties (Fig. 1). The southern extent of the Safety Harbor phase
is Charlotte Harbor, based on the existence of Safety Harbor sites around
Boca Grande (Ch 1, Ch 2), and Peace Creek (Ch 5, De 1) (Bullen 100).






5 tO0

Fig. 1.

T. J.R T197


Land Use Model

As an alternative to Goodyear's model for the Safety Harbor phase,
I propose an ecological-geographical paradigm or model that encompasses
the entire geographic extent of the phase (Clark 1972: 7). The distribut-
ion of Safety Harbor sites can be organized into three generalized envir-
onmental zones: Coastal, Riverine, and Hinterland. Each of these zones
was utilized for different socio-economic activities and the sites located in
each zone reflect these differences in a variety of ways.

The Coastal Zone corresponds to Goodyear's Bay-Beach, Bay-Isla.d,
and Estuary eco-mode categories. This grouping seems logical since the
population in these eco-modes apparently shared a marine-based subsistence
pattern. Also included in this zone are sites not listed in Goodyear's data.
These are the southern sites near Boca Grande and sites in bay and estu-
ary locations in Sarasota County (So 2, So 4, So 5). Very few sites are
found directly on the Gulf side beaches. One coastal burial mound (He 1)
excavated by Moore in Hernando County contained a few Safety Harbor ser-
ies potsherds. Although land animal remains are found in middens in this
zone, the dependence on fish, shellfish, and turtles was great (Bushnell
1962; Griffin and Bullen 1950).

Goodyear points out that most of the temple mound sites are situated
in this zone, indicating that political and religious power was centered in
villages around Tampa Bay. However, he classifies the Parrish Mound 2,
located in the Riverine Zone along the Little Manatee River, as a burial
mound and not a temple mound (Goodyear N.D., Table 2). Willey states
that the charred remains of a four sided structure were uncovered in the
top layers of this mound, along with cremated burials (Willey 1949: 47),
suggesting that this tumulus was a temple sub-structure or supported a
charnal house.

The Riverine Zone encompasses the territory along the river systems
that penetrate the interior and drain into Tampa Bay and Charlotte Harbor.
The most important of these rivers are the Palm, Alafia, Little Manatee,
Manatee, Miakka, and Peace. There is very little data in the form of
faunal analyses, but it may be inferred that subsistence activities were bas-
ed on fresh water fishing, hunting, collecting, and agriculture. The rivers,
even today, furnish a variety of fresh water fish. The river floodplain for-
ests and moist hardwood hammocks could reasonably be expected to provide
an abundance of deer, turkey, land tortoise, and small mammals and birds,
as well as a variety of edible plants. Direct subsistence data is lacking
because, for the most part, these sites (Parrish Mounds, Picknic Mound,
Arcadia Mound) were excavated during the 1930's and emphasis as upon
mounds and not middens. Small scale agriculture, however, has been sub-


stantiated by early historical reports (Garcilaso de la Vega, Varner trans-
lation 1962:110, and the De Soto manuscripts) and can be inferred by the
finding of grinding stone implements in the mound fill at several sites.
Shell artifacts found in several of the Parrish Mounds indicate interaction
with the Coastal Zone.

The third zone, the Hinterland, can be roughly defined as the area
drained by the headwaters and tributaries of the rivers mentioned above.
This is a more restrictive usage of the term Hinterland than is usually
employed (Caldwell 1958). The land is typically flat and low. Geologically
the entire area is part of the Miocene deposited Hawthorn and Bone Valley
formations of highly phosphatic sands and clays. Mixed pine-oak forests
and palmetto flatlands dominate large tracts of this zone, but moist hard-
wood hammocks and occasional cypress and gum swamps are found along
streams. Today a considerable amount of acreage is in citrus groves, pine
plantations, and improved pastureland. Soils are poorer here than in the
Riverine Zone or the eastern Coastal Zone, and probably would not have
accomodated maize agriculture.

Hinterland Exploitation

During the past two years, five independent surveys have been
carried out in this area as part of regional planning and development
studies (Browning 1973; Hemmings 1974; McMurray 1975; Padgett 1974;
Swindell 1974). These surveys, in the Alafia, Little Manatee, and Man-
atee drainages, reported a sparce distribution of small sites, many of
which could not be classified as to culture period because of lack of
diagnostic artifacts.

As a working hypothesis, I would propose that this Hinterland Zone
was used exclusively as a hunting territory, with perhaps some very
minor collecting activities taking place. Although this land use model
may also apply to earlier culture periods or phases, I am limiting it
here to the Safety Harbor phase.

Basic here is the notion that human behavior is non-random. This
is very important in the study of cultural patterns and is reflected in
the form and distribution of human settlement and the interaction be-
weeen environment and society. Another basic assumption is that sites
representing earlier or later cultures may be found in the same terri-
tory, but will differ from the Safety Harbor phase sites in proportion
to the differences in technology and social structure. We might, there-
fore, expect a greater similarity between Safety Harbor sites and late
Weeden Island sites than between Safety Harbor sites and to other areas
or regions, and might profitably be tested as a general law, though that


is beyond the scope of the present paper.

With hypothesis and assumptions thus stated, we should test the
implications in the form of site function, content and location. All
of these are intimately interrelated but must be broken down for ana-
lysis. The implications of the hypothesis are as follows:
1) If this zone is being used only as a hunting territory, no village sites
will be found in this zone, but will be located in adjacent areas. Sites .wil be
small encampments occupied for short periods of time, perhaps having been
visited repeatedly and therefore possibly showing horizontal stratigraphy.
2) If hunting is the only subsistence activity taking place in this zone, the
the artifact assembleges will reflect this and will consist of huntirn ar--
butchering implements: projectile points, knives, scrapers, blades and
flakes. There will be conspicuous lack of ceramics and plant food prep-
aration tools such as grinding stones.
3) If hunters are moving through this territory, their camps w ill tend 3
be located near streams or other sources of fresh \water. Fresh water is
a need not only of hunters, but of the animals they hunt. The tributary
streams would also provide transportation routes to the Riverine and Coast-
al Zones.

No ecological implications can be tested by the data derived :rD:. tche
surveys mentioned above. Limiting factors include the scope of the surve-s.
which did not involve excavations beyond one or two test pits maximum oer
site, and the range of the surveys, since all of these surveys \were per-
formed in the northern half of the Hinterland Zone. Since all of the servers
were carried out independently under no one set research design, the kinds
of data reported differs slightly according to investigator. How ever, usin
the available data the following analysis may be made.

Implication 1, Priority of Hunting Camps over Villages. Bro..n
recorded 9 sites, 4 were classified as Safety Harbor, 2 as Archaic, anc
3 unclassified. One site is listed as a possible village or large can.p, all
other sites were classified as small (under 1 acre) camps. No sites sho1- ed
organic midden deposits. Hemmings reported 13 aboriginal sites, one D3
which was a possible village site for the Picknic 'Mound and might be better
placed in the Riverine Zone. All other sites were one or t.o acres in extenE
unless they represented contiguous campsites. Data from the other three
surveys indicate only small artifact scatters %with no organic hidden dep oi:s
which would be expected from village sites.

Implication 2, Artifact Assembleges will Reflect Hunting Activities.
Although quantitative data is not available for some of these surveys, over
90% of all cultural materials reported consist of flint \waste flakes. Pro-
jectile points were second highest in frequency, with a very negligible anour.t
of ceramics.


Browning found a total of 4 potsherds, Swindell found one sherd, and
no potsherds at all were found during my survey. No flint outcrops occur
in the area except in the Lake Thonotosassa area. Simpson suggest that
blanks were formed there and carried away to be finished (Simpson 1941:
32-33). Our data substantiates this.

Implication 3, Site Location. Again, information on proximity to
water sources was not available for all the surveys, but, when recorded,
it indicated site locations within 200 yards of a water source in all but
three cases. Two sites recorded by Browning were located one half mile
from a stream and one site which I recorded was a mile from a stream.


I have presented here a simplistic model of land use and settlement
in the Safety Harbor phase. There are many gaps to fill in and some prob-
lems of quantification and definition are yet to be solved. What constitutes
a village as opposed to a camp, for example? These problems must be
contended with before we can make explicit statements.

I have tried to point a direction for future field research and syn-
theses. There are some lessons which have run like undercurrents through
this paper. One may be stated simply, "All that is non-ceramic is not
Archaic." There is a tendency among some archaeologists to classify any
site that does not contain pottery as pre-ceraric. There is also a tendency
to call these sites "chipping stations" or flint working sites, when in fact
stone working was a secondary activity to support the primary site function.
Another item to be pointed out is the need for more consistency in site re-
porting. No matter what the original purpose of an archaeological survey
is, as much data and as many different kinds of data as is possible or
feasible to collect should be reported. The ideal repository for this infor-
mation in Florida is the Site Survey files of the Division of Archives, Hist-
ory and Records Management.

All these problems aside, I think that this model is a workable one,
one which deserves more attention. But if further investigation should prove
this model lacking, it should be modified accordingly or another model
formulated and tested.

References Cited

Browning, William
1973 An Archeological and Historical Survey of the Florida Power and
Light Company Manatee Power Plant. Division of Archives, History
and Records Management. Bureau of Historic Sites and Properties.


Miscellaneous Project Report Series, No. 6. Tallahassee.

Bullen, Ripley P.
1951 The Terra Ceia Site, Manatee County, Florida. Florida Anthro-
pological Society Publications, No. 3. Gainesville, Florida.

1952 Eleven Archaeological sites in Hillsborough County, Florida.
Report of Investigations No. 8. Florida Geological Survey.

1965 Florida's Prehistory. Chapter 23 in Florida from Indian Trail
to Space Age. Vol I. Tebeau et al. Delray Beach.

1969 Southern Limits of Timucua Territory. Florida Historical Quar-
terly, 47, No. 4, pp 414-19. Tampa.

1973 Introduction. In 1973 edition of Gordon R. Willey's Archeology
of the Florida Gulf Coast. American Museum Service.

Bushnell, Frank
1962 The Maximo Point Site - 1962. Florida Anthropologist. Vol. 1
No. 4, pp. 89-101.


Caldwell, Joseph R.
1958 Trend and Tradition in the Prehistory of the Eastern Unitec
States. American Anthropological Association Memoir 88.

Calkins, W.W.
1877-80 Notes on personal investigations among the shell mounds
ot florida. Proceedings ot the Davenport Academy of Natural Science,
1976-78, Vol. 2. Cited in Willey 1949.

Clark, David L.
1972 Models and Paradigms in Contemporary Archeology. In Models
in Archeology, D.L. Clark, ed. London.

De la Vega, Garcilaso
1962 The Florida of the Inca. Varner translation. University of Texas

Goggin, John M.
1947 A preliminary Definition of Archeological Areas and Periods 2:

Fiorida. American Antiquity. Vol. 13, pp. 114-127.

Goodyear, Albert
N.D. Political and Religious Change in the Tampa Bay Timucua: An


Ethnohistorical Reconstruction. Manuscript on file at Florida Division
of Archives, History and Records Management, Tallahassee; Florida
State Museum, Gainesville; and Department of Anthropology, Univer-
sity of South Florida, Tampa.

Griffin, John
1952 Prehistoric Florida: A Review. In Archeology of the Eastern
United States, J.B. Griffin, ed. University of Chicago Press. Chicago

Griffin, John, and Ripley P. Bullen
1950 The Safety Harbor Site, Pinellas County, Florida. Florida Anthro-
pological Society Publications. No. 2. Gainesville.

Hemmings, E. Thomas
1974 Archaeological and Historical Survey of the Lonesome Mine Tract
in Southeastern Hillsborough County, Florida. Preliminary Report
to Brewster Phosphates. Manuscript on file at Florida State Museum.

Hrdlicka, Ales
1922 Anthropology of Florida. Publications of the Florida Historical
Society, No. 1. Winter Park.

McMurray, Carl
1975 Personal Communication

Moore, C.B.
1900 Certain Antiquities of the Florida West Coast. Journal of the
Academy of Natural Science. Vol. II. Philadelphia.

Padgett, Thomas J.
1974 Archeological and Historical Reconnaissence of the W.R. Grace
Co. Property in Hillsborough and Manatee Counties, Florida. Division
of Archives, History and Records Management, Bureau of Historic
Sites and Properties, Miscellaneous Project Series, No. 17. Tallahassee.

Sears, William
1967 The Tierra Verde Burial Mound. Florida Anthropologist, Vol. 20,
nos. 1-2, pp. 25-74. Tallahassee.

Simpson, J. Clarence
1941 Source Materials for Florida Aboriginal Artifacts. Florida Acad-
emy of Science, Vol. 5, pp. 32-34.

Swindell, David E.
1974 An Archeological and Historical Survey of the Horatio Property,
Hillsborough County, Florida. Manuscript on file at Division of Arch-


ives, History and Records Management, Department of State. Tall-
ahassee, Florida. To be published in the Division's Miscellaneous
Project Series.

Walker, S.T.
1880 Preliminary Explorations among the Indian Mounds in Southern
Florida. Annual Report of the Smithsonian Institution for 1879.

Willey, Gordon R.
1949 Archaeology of the Florida Gulf Coast. Smithsonian Miscellaneous
Collections, Vol. 113 (Whole volume).

Tallahassee, Florida
Feburary 26, 1975


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). Volumes 25:2, pt. 2 (Publication No. 6) and 28:3,
pt. 2 (Publication No. 7) are $2. 00 each. Volumes 1; 2:3-4; 5:1-2; 6:1, 4; 7:2, 3;
8:1, 2, 4; 10; 11:1, 2;12:1; 13:1, 2;17:1; 18:1; 21:1-2; 24:1; and 26:1 are out of print.

a mongraph series, may be purchased as follows: Numbers 1, 5, 6, and 7 at
$2.00 each. Numbers 2, 3, and 4 are out of print.

Orders with remittances for any of the above 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: P1. 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.


Northwest Florida Chapter:
Mary Lou Watts
4702 19th Street
Panama City, FL 32401

Northeast Florida Chapter:
Robert E. Johnson
4250 Melrose Avenue
Jacksonville, FL 32210

Central Florida Chapter:
Arthur F. Dreves
2414 Summerfield Road
Winter Park, FL 32789

Tampa Bay Chapter:
William J. Webster
740 17th Avenue South
St. Petersburg, FL 33701

Indian River Chapter:
Cliff E. Mattox
225 Colonial Court
Satellite Beach, FL 32935

Broward County Chapter:
Norcott S. Henriquez
1510 Dewey Street
Hollywood, FL 33020

Palm Beach County Chapter:
Mrs. F. B. Morrison
936 Upland Road
West Palm Beach, FL 33401

Dade County Chapter:
Lorraine Vagner
1204 Wm. Walsh Avenue, Apt. 24D
Coral Gables, FL 33146