2000 Florida Anthropological Society Inc.
The Florida Anthropological Society Inc. holds
source text of the Florida Anthropologist
considered the copyright holder for the text
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.
PUBLISHED BY THE
FLORIDA ANTHROPOLOGICAL SOCIETY, INC.
THE FLORIDA ANTHROPOLOGIST is published quarterly in March,
June, September, and December by the Florida Anthropological Society at
3400 East Grant Street, Orlando, Florida 32806. Subscription is by mem-
bership in the Society for individuals interested in the aims of the Society.
Annual dues are $4. 00; student membership $2. 00. Requests for member-
ship and general inquiries should be addressed to the secretary; subscrip-
tions, dues, and back issue orders to the treasurer; newsletter items to
the president; and manuscripts for publication and books for review to the
editor. Second class postage paid at Orlando, Florida.
THE FLORIDA ANTHROPOLOGIST
Volume XXIII, No. 2
The Colby Site, Marion County, Florida
Stephen L. Cumbaa and Thomas H. Gouchnour
Regionalism in Florida during the Christian Era
Ripley P. Bullen . . . . . . . .
The Fish Creek Site, Hillsborough County, Florida
Karlis Karkins ... . . . . ...
OFFICERS OF THE SOCIETY
President James W. Covington
Tampa University, Tampa, Fla. 33606
1st Vice President Carl A. Benson
3400 East Grant Ave. Orlando, Fla. 32806
2nd Vice President William M. Goza
P. O. Box 246, Clearwater, Fla. 33515
Secretary-Treasurer Sarah B. Benson
3400 East Grant Ave., Orlando, Fla. 32806
Editor-Resident Agent Ripley P. Bullen
Florida State Museum, Gainesville, Fla.,
Three years: Thomas H. Gouchnour
Two years: Cliff E. Mattox
Cocoa Beach, Florida
One year: Don D. Laxson
At large, for one year:
Charles A. Hoffman, Jr.
Donald W. Sharon
Fort Walton Beach
THE COLBY SITE, MARION COUNTY, FLORIDA
Stephen L. Cumbaa and Thomas H. Gouchnour
The Colby site (Mr-57) was excavated early in 1968 by Dr. Thomas H.
Gouchnour and his associate, William J. Webster, both of Jacksonville. Three
test pits, each 5 by 5 feet in area, were dug and the materials screened, re-
moved, processed, and analyzed by arbitrary 6-inch levels. These specimens
and a report of the vertical distribution of pottery types have been donated to
the Florida State Museum, University of Florida, where they were reanalyzed
by Ripley P. Bullen, Chariman of the Department of Social Sciences of that
institution. Dr. Elizabeth S. Wing, Assistant Curator of Zoo-archaeology of
the Florida State Museum, analyzed the abundant faunal remains.
The site is at the northwestern edge of Colby' s Landing Fish Camp on
the east bank of the Oklawaha River approximately a half mile north of the
Florida Highway 40 bridge and a little east of Ocala, Florida (Fig. 1). The
property is presently owned by Francis S. Gay who lives just east of the site
and who graciously gave his consent to the excavation by Dr. Gouchnour.
The Colby site consists of a low rise or bluff on the east bank in the
middle of a bend in the river. It was originally disclosed by a bulldozer cut
at the bottom of the bluff toward the river' s edge (Fig. 1). The site has been
built up by freshwater shell midden deposits and is presently overgrown with
low grasses and dotted with cedar trees which grow nearly to the water in the
rich organic soil. The midden is separated from the river by a small swamp
which will be inundated by the rise in water resulting from the construction
of the Cross-Florida Barge Canal. The midden itself will not be covered.
Test I was dug to a depth of 5 feet in a high part of the bluff near the
western edge of the midden (Fig. 1). This pit, as were the other two tests,
was dug to the heavy clay base which seemingly underlies the whole site. It
was tested some 4 feet deeper with a posthole digger, but proved sterile as
far as artifactual material was concerned. Some bits of charcoal were noted.
After examination of the vertical distribution of pottery types from Test I
as recorded in Table 1, it seems best to discuss the ceramic sequence in terms
of three apparent time periods. As suggested in pottery tabulation, Levels 7,
8, 9, and 10, featuring predominantly Orange series sherds, seem to repre-
sent the earliest period. Of a total sherd count in these four lower levels of
118, 108 were Orange series (75 Orange Plain, 33 Orange Incised). Two
I"- RPPKGo 100'
\\\' BULLDoZER. CrTS
,/'" FaLFF oR RiSE
P DiRT ROAD
Fig. 1. Sketch map, Colby site, locating test pits.
sherds represented the Norwood series, one punctated and the other Plain.
These sherds, like the Orange series sherds, are fiber tempered but also
contain an appreciable amount of sand (Phelps 1965: 66-67). The 8 remaining
sherds from these lower levels were of the St. Johns series. Of these, five
sherds were of St. Johns paste with varying degrees of fibrous inclusions as
a tempering medium. These seem to be transitional between the Orange and
the St. Johns series. Some sherds with somewhat more fiber temper but si-
milar paste might be described as "chalky" Orange. The sixth was a St.
Johns Incised sherd.
CUMBAA AND GOUCHNOUR 45
VERTICAL DISTRIBUTION OF SHERDS, COLBY SITE, TESTS I AND II
Test 1, Levels
1 2 3 4 5 6 7
Prairie Cord Marked
Alachua Cob Marked
Deptford Check Stamped
Deptford Linear C'k St'd
Deptford Simple Stamped
Perico Punctated B
St. Johns series
Check Stamped 12
Punctated (like Perico Punc.
9 10 talks
1 2 3
13 20 4
26 42 15
4 5 82
8 10 172
1 2 3 4tals
3 1 4 1 9
Totals 90 242 150 138 85 49 64 21 14 19 872 319 173165 162 819
The second division incorporates Levels 4, 5, and 6 (Table 1). The
study of the 272 sherds from these levels indicates that the significance here
is in the growing number of St. Johns series sherds and the appearance of
St. Johns Plain without fiber tempering. There were 121 sherds of the St.
Johns series and 138 sherds of the Orange series present. The remainder,
13 sherds, included 2 of the Norwood series, 4 Pasco Plain with their char-
acteristic limestone tempering, 6 sand-tempered plain, and in the highest of
these three levels, one Prairie Cord-Marked sherd (Goggin 1952: 109).
The third division, most recent in time, includes the material from
Levels 1, 2, and 3. As might be expected, it includes the greatest variety of
sherd types. It also includes the greatest number of sherds, 482, due in part,
no doubt, to surface traffic and breakage. The significance in these levels is
in the virtual absence of Orange series sherds (8), the introduction of St. Johns
Check-Stamped (32 sherds), and the growing predominance of sherds of a sand-
tempered series (200). St. Johns Plain sherds numbered 185 and there was only
one St. Johns paste sherd with fiber tempering. A few of the St. Johns Plain
sherds showed some clay or sherd inclusions. Other sherds in these levels in-
cluded 14 of the Norwood series, 27 St. Johns Incised, one St. Johns Simple
Stamped, 14 sherds of the limestone-tempered series (Pasco Plain and Incised
and Perico Punctated B), and, included in the total of the sand-tempered series,
8 Prairie Cord-Marked and 1 Alachua Cob-Marked.
There is a problem of terminology in our discussion of sherds that are
sand-tempered "Plain". These are sometimes called Deptford Plain (Bullen
1969: 30-31). Wauchope (1966: 52-54) discusses the occurrence of Deptford
Plain in northern Georgia. The sherds from the Colby site generally fit into
his description of Deptford Plain as having dark to gray or tan paste with a
temper of fine to medium sand. He notes, as here, that some specimens
showed coarse sand tempering. However, tetrapods were not found in the
material from the Colby site and not enough rim sherds were found to make
further comment on similarities.
Test II (Table 1) will not be discussed in detail, except to say that it
generally confirmed the ceramic sequence found in Test I. Exceptions are the
increased percentage of the St. Johns and "Deptford" sherds. It will be noted
in Table I that the deposits underlying Test II were not as deep as those of
Test I. The clay base here was reached at a depth of two feet.
Test III followed much the same pattern but will not be discussed due
to possible field or laboratory error in mixing the contents of two levels.
However, it should be noted that the lowest level (Level .6) contained 12 Orange
Plain sherds to the exclusion of other types. Levels 3-5 produced about 100
sherds of the Orange series per level as well as sherds of other series. The
highest level supplied only a total of 4 Orange series sherds, with a great
many St. Johns II and Hickory Pond period sherds.
Much of the ceramic material from the Colby site seems to fall in the
Florida Transitional period described by Bullen (1959: 45-53). The site seems
to have an earlier Orange period base and to have been inhabited until the St.
Johns II A-Hickory Pond Period, thus bracketing transitional times.
CUMBAA AND GOUCHNOUR
Fig. 2. Orange and St. Johns series sherds.
a-b, d, Orange Incised; c, Orange Plain; e, Norwood Incised; f-g,
chalky Orange Plain; h, St. Johns Plain with fiber inclusions; i, St.
Johns Plain; j-k, St. Johns Incised.
,E ld ..
Fig. 3. Miscellaneous artifacts and sherds.
a, drilled shark's tooth; b, large shell bead; c, scraper
edge on blade; d, Pasco Incised; e, Pasco Plain; f, Perico
Punctated B; g, punctated, sand-tempered; h, Deptford
Plain; i, Deptford Cross-Stamped; j, Deptford Linear Check
Stamped; k, St. Johns Linear Check Stamped; 1-n, St. Johns
Check Stamped; o, shark's tooth scraped sherd, St. Johns
paste; p-q, Prairie Cord Marked; r, Alachua Cob Marked.
CUMBAA AND GOUCHNOUR
The faunal remains from the Colby site proved to be fairly consistent
through time. Dr. Wing's analysis of the material showed 24 species repre-
sented by some 326 individuals. The faunal material from the three test pits
was grouped according to the three apparent time periods represented by the
The earliest levels of Test I, Levels 7-10, showed 38 individuals of 15
species, the most predominant in number and in species being members of the
turtle family (16 individuals, 6 species). Raccoon, deer, and fish were next
in relative abundance. Also represented were opossum, cottonmouth mocca-
sin, otter, and alligator. Of special interest were two individual beaver
(Castor canadensis) fragments, found in Levels 9 and 10. Beaver, presently
extinct in Florida, have been reported from this same time period at a few
other sites in the state.
Levels 4-6 presented much the same picture. Of 55 individuals, 17
were turtle. Fish, with 21 individuals of 4 species, moved up in relative abun-
dance. Opossum, raccoon, deer (7 individuals), cottonmouth, and alligator
were also present. Of interest in Level 4 was a bone of the domestic dog
Again we have much the same situation in the three highest levels (1-3)
of Test I with 70 individuals representing 19 species. Fish were most fre-
quent, with 28 individuals of 5 species. Turtles continued to be of importance,
with 22 individuals of 6 species. Also found were opossum, deer (5 indivi-
duals), bird, cottonmouth and other snakes, and a bone of a bear.
Tests II and III did not differ significantly from Test I in faunal
content. We do find the addition of rabbit and fox to our inventory, and the
presence of another Canis bone in Level 2 of Test II. We also find 3 more
beaver remains in Tests 2 and 3, but in higher levels than in Test 1. Levels
2 and 3 of Test 2 each have one individual Castor, as does Level 2 of Test III.
This is perhaps the latest recording of beaver bones in a prehistoric site in
The forest, river, and swamps near the site are still rich in these same
species. The only one to have vanished over the years is the beaver. At pre-
sent it is not certain when the beaver left Florida, but we seem to have a re-
cord at the Colby site of this animal at least up into the first few centuries
after Christ. Not mentioned but of course present were hundreds of fresh-
water snails and mussels of several species.
Stone And Bone Artifacts
The test pits and bulldozer cut at the Colby site yielded a large quantity
of stone and bone artifacts. Some of these are pictured in Figures 4-7. For
clarity, projectile points were identified using Bullen' s A Guide to the Identi-
fication of Florida Projectile Points (1968).
In discussing these artifacts by levels, we note in Test I the paucity of
material in the lower levels, i.e. below a depth of 30 inches. Levels 7-10 pro-
duced only one bone awl and 2 utilized flint chips. Test II in its lowest level
(4) did not produce much more. Two bone awl fragments, two worked stone
fragments and a Citrus projectile point (Fig. 5, k) were the only non-ceramic
Levels 4 and 6 of Test I produced more artifacts. Here we have one bone
awl and one fragment of an awl, 4 worked chert fragments, 3 utilized chips, one
Florida Archaic Stemmed projectile point, Marion variety (Fig. 5, g) a small
bone point which appears to have been part of a composite tool (Fig. 4, a), and
a blade scraper (Fig. 3, c), which may be a broken fragment of a larger tool.
Test II at the same relative time period (Levels 2 and 3) produced consider-
ably more bone and stone artifacts. Twelve bone awl fragments, 2 bone pins,
and 4 other worked fragments comprised the total bone artifacts from these
levels. One ovate knife, 2 Citrus projectile points (Fig. 5, i-j), one ovate
scraper-knife, one concave based assymetrical knife, and 10 worked fragments
made up the stone inventory. Other artifacts included a perforated (drilled)
shark's tooth and a large bead (Fig. 3, a-b). The latter has been identified
by Dr. H. K. Brooks of the Department of Geology, University of Florida, as
being manufactured from fossil oyster shell. This fossilized shell is probably
of the geologic period known as the Cretaceous, which ended some 55 million
years ago, and is thus too early to have come from deposits in Florida.
This indicates trade with peoples probably considerably north of the present
The highest levels of Test I produced a large number on non-ceramic
artifacts, including 10 bone awl fragments, one broken bone pin, 2 fragments
of worked bone, 3 ovate flint knives, 2 broken assymetrical trianguloid knives
(Fig. 5, c), one Citrus projectile point, (Fig. 5, a), 13 worked chert fragments,
2 complete assymetrical trianguloid knives (Fig. 5, d), one stemmed drill,
(Fig. 5, b) and 2 Archaic Stemmed projectile points, one Levy variety and
one Alachua variety (Fig. 5, e-f). Test II produced only one ovate knife at
the highest level.
Erosion of the bluff from the bulldozer cuts resulted in many surface finds,
the majority of which were some 48 broken bone implements, mostly awls. Five
ovate knives were also found, as were some 20 worked chert fragments, one
stemmed drill (Fig. 5, b), 1 bone pin, 2 Pinellas projectile points, 2 hammer-
CUMBAA AND GOUCHNOUR
Fig. 4. Worked bone and beaver remains.
a, bone point from a composite tool; b-e, pins (bis
incised), f-i, awls; j, splinter awl; k-l, concave sided
awls; m, hollow bone point, rectangular at open end; n,
left premaxillary, Castor canadensis (beaver); o, left
Fig. 5. Vertical distribution of points, knives, and drills at Colby's Fish Camp.
a, i-k, Citrus; b, stemmed drill; c-d, asymmetric trianguloid knives; e, Levy'
f, Alachua; g, n-o, Marion; h, Jackson-like; 1, Pinellas; m, p, u, Hernando;
and r, Clay projectile points; q-r, t, drills.
CUMBAA AND GOUCHNOUR
I I I
I 3 II
Fig. 6. Chipped tools from bulldozer cut.
a, Jackson-like; c-d, Hernando; g, Culbreath;
and h, Lafayette projectile points; b, hafted drill;
e, asymmetric triamguloid knife; f, drill.
stones, 2 worked deer antlers, 4 rectanguloid knives, and an example each of
a flake drill point, a hollow bone point (Fig. 4, m), an incised bone pin (Fig. 4,
h), a Jackson-like Culbreath, and a Lafayette projectile point (Fig. 6, a, g-h).
The variety of bone and stone tools and ornaments would seem to indicate
that the inhabitants of the site, through time, were making wide use of the varied
resources available to them in their environment. This is again indicated in the
variety of food bone present at the site.
The frequency of deer bone awls, the most common non-ceramic artifact,
suggests a greater variety of uses than just hide-punching. The process of
making the awls seems clear from the sequence and degree of finish of some
fragments. Starting with the whole bone (usually tibia), the process involved
cutting, splitting, and finally shaping and polishing the bone.
Fig. 7. Large chipped tools.
a-d, ovate knives; e-f, scraper-knives; g, chipping
hammer; h, hammerstone.
CUMBAA AND GOUCHNOUR
The Colby site, heretofore untouched as far as can be determined, is
valuable in the added clarity it lends to our picture of the Florida Transitional
period. It is one of the few sites, along with Sunday Bluff (Bullen, 1969), the
Zabski site (Atkins 1967), and the Cato site (Bullen, Bullen, and Clausen 1968),
to name those recently reported, which seems to date from this period which
separates the Orange-St. Johns ceramic sequences. At the Colby site, the
Orange occupation seems to have been the longest, contrary to the Cato and
Zabski sites which are predominantly St. Johns. Lacking at the Colby site was
the remarkable variation in St. Johns ware decoration as noted especially at
the Zabski site. Perhaps because of its location and resultant environment,
the Colby site is most similar to nearby Sunday Bluff. Indeed, the ceramic
sequences are very similar (Bullen 1969).
The data on the faunal material, both food bone and worked bone, are
extremely important, as more than anything else they give us an idea of the
utilization of the environment by the inhabitants of the site. Enough deer bone
awls and fragments were found to piece together the process whereby these
tools were made.
The workmanship and finishing of some of the bone tools, the pins and
the incising of at least one of them, the drilled, fossilized shell bead and the
drilled shark' s tooth indicate that the inhabitants could spend at least some
time on luxury items and did not have to spend all their time and energy in the
quest for food.
The shell bead, if nothing else, indicates trade with an area or areas
considerably north of the present state of Florida. Perhaps more work at
Colby' s Landing or other Transitional period sites, utilizing carefully con-
trolled stratigraphy, can illuminate this aspect of the relationship of Florida' s
Indian population with other peoples.
Atkins, Steve, and Jeannie MacMahan
1967 The Zabski Site, Merritt Island, Florida. Florida Anthropologist,
Vol. 20, nos. 3-4, pp. 133-145.
Bullen, Adelaide K. and Ripley P.
1950 The Johns Island Site, Hernando County, Florida. American
Antiquity, Vol. 16, no. 1, pp. 23-45. Salt Lake City.
1953 The Battery Point Site, Bayport, Hernando County, Florida.
Florida Anthropologist, Vol. 6, no. 3, pp. 85-92.
Bullen, Adelaide K. Ripley P. Bullen, and Carl J. Clausen
1968 The Cato Site Near Sebastian Inlet, Florida. Florida Anthro-
pologist, Vol. 21, No. 1, pp. 14-16. Tallahassee.
Bullen, Ripley P.
1959 The Transitional Period of Florida. Southeastern Archaeo-
logical Conference Newsletter, Vol. 6, pp. 45-53. Chapel Hill.
Bullen, Ripley P.
1968 A Guide to the Identification of Florida Projectile Points.
Florida State Museum, University of Florida. Gainesville.
1969 Excavations at Sunday Bluff, Florida. Contributions of the
Florida State Museum, Social Sciences, No. 15. Gainesville.
Goggin, John M.
1952 Space and Time Perspective in Northern St. Johns Archaeology,
Florida. Yale University Publications in Anthropology, No. 47.
Phelps, David Sutton
1965 The Norwood Series of Fiber-Tempered Ceramics.
Southeastern Archaeological Conference, Bulletin.
1966 Archaeological Survey of Northern Georgia, With Some Cultural
Hypotheses. Memoirs of Society for American Archaeology,
No. 21, pp. 52-54. Salt Lake City.
Willey, Gordon R.
1949 Archaeology of the Florida Gulf Coast. Smithsonian Mis-
cellaneous Collections, Vol. 113. Washington.
REGIONALISM IN FLORIDA DURING THE CHRISTIAN ERA
Ripley P. Bullen
This paper, which neglects the Florida panhandle, is concerned with
regionalism which started in peninsular Florida before the time of Christ
and was still present when Europeans first settled there in 1564. That four
distinct cultural units could maintain their apparent separate identities for
over 1500 years in such a small area without geographical barriers to cul-
tural interchange is noteworthy.
Florida, the southeasternmost part of the United States, is a long and
narrow peninsula, 240 kilometers in maximum width, which extends some
480 kilometers south-southeasterly towards the Caribbean. In general it
enjoys a subtropical oceanic climate while its coastal areas until recently
have been extraordinarily rich in sea food both fish and shellfish. Birds,
turtles, deer, and other small animals have always been plentiful. Wild
vegetable food especially zamia was also available to the aborigines.
In fact the only drawbacks to life in Florida, then as now, were swarms
of mosquitos and other insect life.
Indians have lived in Florida for 10,000 years. During these years
many changes occurred in their tools, ornaments, and other industrial pro-
ducts. These changes have been used by students of prehistory to define
cultural groups and to arrange these groups in time and space. For the
purposes of this paper four geographical areas are recognized, three of
which touch each other and surround the fourth. They are: 1) the central
Gulf coast or greater Tampa Bay region, 2) the St. Johns River or north-
eastern region, 3) the south central region around Lake Okeechobee, and
4) the Glades region or south Florida east, southeast, and south of Char-
Regionalism started in Florida during the Florida Transitional period,
perhaps as early as 1000 B.C. certainly by 500 B.C. At the close of the
Transitional period we find limestone-tempered vessels with inturned rims
sometimes decorated with geometric designs in linear punctations north of
Tampa Bay; temperless openmouthed bowls along the St. Johns River; fine
sand-tempered, beveled rim, vessels with a peculiar dragged surface a-
round Lake Okeechobee; and simple pottery tempered with medium-sized
grains of sand in the Glades region.
These ceramic differences may seem slight, and indeed they were in
the beginning, but these slight differences except for decoration ex-
clusively maintain themselves throughout the rest of archaeological time.
The resultant ceramic traditions, which I am about to trace, together with
other differing aspects of these cultures show fundamental divergences in
At the close of the Transitional period, the Tampa Bay region received
influences from the Deptford culture to the north which introduced sand-
tempering, linear and check stamping, tetrapods, burial mounds, and pro-
bably horticulture. To the north of Tampa Bay such vessels were still
limestone-tempered. Shortly this region participated in a wide-spread cul-
tural development which extended from Charlotte Harbor west to New Or-
leans. Called in Florida the Weeden Island period, it is characterized by
specially made ceremonial vessels which artistically represent, I believe,
the best aboriginal ceramics of the United States. A cult of the dead was
very prominent with especially constructed burial mounds having prepared
bases, and dedicatory pottery caches or paths of sherds leading from the
periphery to the center of the mound. Every Indian town had its burial
mound and, based upon implications from burial arrangements, its charnel
In some central gulf coast burial mounds are found pottery from north-
west Florida, recognized by its micaceous content, as well as vessels which
were made in the St. Johns River and Lake Okeechobee regions. Local
ceramics and those from the northwest were decorated with punctation, in-
cision, modeling, and red paint. Some of the zoned-red containers are quite
decorative. Imports from the Lake Okeechobee region are not so decorated.
Around A. D. 1300 new influences entered Florida and the Tampa Bay
region from the lower Mississippi valley introducing small triangular arrow
points, handles on vessels, new vessel shapes, new pottery decoration,
temple mound ceremonialism, and probably an improved type of corn. The
resultant culture in the Tampa Bay region is referred to as Safety Harbor.
In this region settlement patterns were well established with a midden re-
sidential area, a plaza which was kept clean of debris, a temple mound whose
ramp pointed towards the residential area, and, to one side a burial mound
and charnel house. This complex, of middle Mississipian origin, apparently
did not enter the Tampa Bay region as a sudden invasion but as a series of
discrete traits. This culture was the well developed agricultural economy
encountered by Narveaz and De Soto early in the sixteenth century.
The situation in the St. Johns River region, a short distance to the east,
had a different history. After a brief initial phase, when linear and check
stamping was copied on temperless St. Johns paste; pottery in this region,
except for funeral ware, was undecorated until A. D. 850 when check stamping
was reintroduced. Burial mound pottery was plain, red painted, or supplied
with a cut, appliqued, flat collar. Occasionally small modelled animals or
other objects are found.
This is obviously quite different from the contents of the contempora-
neous Weeden Island period burial mounds of the contiguous Tampa Bay re-
gion described above. While plain, check-stamped, and red-painted St. Johns
paste pottery is found in large quantities in Tampa Bay burial mounds, the
reverse (Weeden Island pottery in St. Johns area mounds) does not occur.
While a few temple mounds are known for the St. Johns River region, the for-
mal settlement pattern of the Gulf coast has not been reported. Also unre-
corded here are vessel handles, shapes, and decoration found in the Tampa
Bay region during the Safety Harbor period. Small triangular points are,
however, sometimes found and,more rarely,copies in St. Johns paste of gulf
coast vessels, showing that communication during this time was still occur-
ring. The strange animal and inanimate forms found at Thursby's burial mound
on the Beresford peninsula and elsewhere are not duplicated on the gulf coast.
Before discussing any possible reasons for the non-acceptance of many
Weeden Island and Safety IHrbor traits by the Indians of the St. Johns River
regions, it seems best to examine the Lake Okeechobee and Glades regions.
Near Lake Okeechobee are four large prehistoric ceremonial centers each
consisting of many mounds and earthworks in a complicated arrangement.
Typically some mounds are located at one end of a roadway formed by two
long, low, ridges of dirt. Sometimes these mounds are backed by semi-cir-
cular ridges or borrow pits. Sometimes the roadways are connected to form
large semi-circles. Ditches or ponds were also dug and large flat-topped
mounds constructed. This arrangement bears no obvious relationship to the
civic plans described for the Tampa Bay region. It is, however, somewhat
suggestive of Adena earthworks found north of the Ohio River!
Complexes of the Lake Okeechobee type are not known elsewhere in
Florida or in southeastern United States. In spite of this great structural
elaboration, the local ceramics throughout the life of these sites is the simple,
fine, sand-tempered, slanted lip, dragged surface, undecorated pottery men-
tioned earlier for this region. By trade sherds we know that these sites star-
ted around 750 B. C. and, to judge from Spanish metal found in their' burial
mounds, they were still in use after the Spanish conquest of Mexico. This is
a long time for one, narrowly defined, pottery type to have persisted. As for
the St. Johns region, plain vessels of the Lake Okeechobee region are found
in Tampa Bay burial mounds but the reverse is not known to any extent. How-
ever, a few trade sherds from all three other regions are found in Lake Okee-
The Glades region remains to be mentioned. Here, along both the east
and west coasts, pottery appears at first to be undecorated. By A. D. 300 or
shortly thereafter, pottery decoration emphasized short incised lines or ticks
to form simple geometric designs or curved lines, such as might have been
made by fingernails, arranged one above another on the sides of vessels.
During the next period (A. D. 500-850) incised arches or crossed straight lines
were more popular. In the last period, cross-dated by check-stamped trade
sherds from the St. Johns River region, notched lips are prevalent. This
simple incised ceramic tradition of the Glades region has no counterpart else-
where in Florida.
As far as I am aware, there are no burial mounds in the Glades region
south of Charlotte Harbor and the Big Circle mound group. On the east coast,
burial mounds are found nearly as far south as Miami along the high (rela-
tively) ridge which divides the Glades areaproper from the Atlantic coastal
region. It is possible some of the immense shell middens of the Thousand
Island area may have functioned as temple mounds. Even if so, no civic
plan like that of the Tampa Bay region nor any complex like that of the Lake
Okeechobee region was present. Ceramics of the Weeden Island and Safety
Harbor periods of the Tampa Bay region are not found in the Glades region as
defined in this paper. On the north side of Charlotte Harbor, where these two
regions meet, pottery of both regions is found associated with each other in the
highest levels of various sites.
I have indicated, I believe, specific differences in the four regions -
differences in pottery types, ceremonial life, and settlement patterns which
imply cultural differences of some magnitude. These differences cannot be
explained on the basis of culture lag, isolation, lack of communication, or,
entirely, by geography. They appear to be tribal in nature and to correlate
really with the tribal groups ethnologically referred to as Western Timucua,
Eastern Timucua, Myamimi, Calusa, and Tequesta. Minor differences in
pottery support the Calusa-Tequesta tribal distinctions.
It is true that the best agricultural land lies in north central Florida and
that the Safety Harbor culture of the Tampa Bay region had a full-fledged agri-
cultural economy as noted by De Soto' s narrators. It was also the last com-
plex to enter Florida from the west. These facts, particularly the first, may
be used to explain the failure of this cultural manifestation to successfully
penetrate east and south Florida where the land is not suitable for the Middle
Mississippian aboriginal type of agriculture. Such reasoning, however, would
not explain differences between the other three culture areas.
Culturally speaking, the St. Johns River region was continually under
attack by strong influences from Georgia to the north, from west Florida, and,
in its southern area, from the Lake Okeechobee region. At different times it
gave up a little territory in the north to "Georgian" groups but it always main-
tained its cultural identity and individuality. It accepted some traits, such as
burial mounds, but it did not accept the elaboration which accompanied them in
other areas except near Jacksonville and to the northeast. The same state-
ments apply in general to the other two areas whose inhabitants refused to be
acculturated to any great entent.
Obviously the cultures of the three areas who refused acculturation were
extremely conservative and successfully self-perpetuating. And it must be ad-
mitted that they were shielded a little from the full effects of the influences
which acculturated the Tampa Bay region. However, a little more complete
explanation seems needed to explain these phenomena.
I believe the explanation lies in a combination of three reasons. The St.
Johns region was a well integrated area in late Archaic times. For nearly a"
thousand years (2000-1000 B. C.) they had made fiber-tempered pottery, the
first pottery in continental United States, and hence had a long tradition of con-
servatism behind them. The superiority of well made, temperless pottery
over that of fiber-tempered pottery was obvious to them but they wasted no
time on unnecessary embellishment. This early conservatism they maintained
in later times. It is also evident that the other areas had a large Archaic com-
ponent in their makeup as they were established at the end of the Archaic period
and, at first, exhibit some decorative holdovers from that period.
This conservatism was helped by Florida' s peninsula position as a cul-
de-sac into which influences were continually entering but which was not at
these times an expanding culture center. Having only one frontier exposed to
attack was a distinct advantage. It is also evident that these conservative re-
gions were fortunate in having an ample and assured food supply. They were
economically entirely self-sufficient and had no need to import new ways of
producing food. This permitted them to pick and choose in their acceptance
or rejection of new traits.
That they were basically conservative is indicated by the lack of change
in their local ceramics. Apparently, feeling no need to embellish their local
ceramics, they also felt little need to embellish their culture with exotic traits.
Certainly the inhabitants of the Lake Okeechobee region had a rich cere-
monial life as did those of the Tampa Bay and northern St. Johns regions. To
judge from carved wood from the Glades, Lake Okeechobee, and St. Johns
River regions, the lack of emphasis on the decoration of pottery may reflect
ornate wooden objects. In emphasizing the conservatism of these societies I
do not mean to indicate in any sense that they were entirely drab and without
artistic accomplishments. The important point is that they were able to main-
tain their cultural identity although in continuous communication with their
neighbors for over 1000 years. In this they were aided by their peninsular
location and adequate food supply.
THE FISH CREEK SITE, HILLSBOROUGH COUNTY, FLORIDA
The Fish Creek site (Hi-105) is located in a mangrove swamp at the
edge of Old Tampa Bay, in the extreme western portion of the city of Tampa,
Hillsborough County, Florida (Fig. 1). Fish Creek, for which the site is
named, cuts through the swamp and empties into the bay. Except for ma-
terial dredged from the creek to widen it and sand pumped into the area
north of the creek for land reclamation purposes, the swamp is completely
inundated at normal high tide.
Stone artifacts, potsherds and numerous spalls have been recovered
from the fill along the creek, as well as from the muck along the western
edge of the mangroves. The artifact bearing area along the shore begins
just south of the point where Courtney Campbell Causeway (State Road 60)
joins the mainland. From here it extends intermittently for nearly three
quarters of a mile to the south, ending abruptly at the edge of a small stream
situated about one hundred yards to the south of Fish Creek. More precisely,
the site is located in the NE 1/4 of the NE 1/4 of Sec. 13, R17E-TZ9S, and
the SW 1/4 of the NW 1/4 of Sec. 18, R18E-TZ9S.
The author has collected artifacts from this site for the past five years.
Dr. Lyman O. Warren and Al Goodyear, both of St. Petersburg, have also
accumulated sizeable quantities of material. The three collections, com-
prising nearly five hundred artifacts, will be discussed as a single unit for
the purposes of this paper. A detailed description of the cultural material
is presented to aid future comparative studies and to indicate any possible
regional variations in artifact types. The projectile points are identified
using Bullen' s (1968) taxonomic system. The measurements given for the
chipped stone tools indicate the range of length, width and thickness, fol-
lowed by the average where applicable.
A total of 183 potsherds and 304 stone tools were recovered from
Hi-105. Knives represent 52 per cent of the total stone inventory. Scrapers
and points appear in about equal quantities and are the next most common
items, comprising a combined total of 34 per cent. The majority of the stone
artifacts were fashioned from local chert. Very few were manufactured from
silicified coral. The stone recovered from the muck is heavily patinated and
dull white or grey-blue in color. Some of the material from the creek fill
bears a brown, iron stain.
FISH CREEK SITE 8 Hi 105
--- airport limits
x artifact concentration
FISH CREEK SITE
Norwood Plain (2 sherds) (Fig. 4, k). This pottery type, defined by Phelps
(1965: 66-67), has previously been referred to as semi-fiber tempered plain
(Bullen and Bullen 1953: 88) and St. Simons Plain (Willey 1949: 359). The
sherds are tempered with about equal amounts of fiber and fine sand. Paste
core is black, exteriors are reddish or black. Sherd thickness is about 1 cm.
St. Johns Incised (2 sherds) (Fig. 4, j). Both sherds are large with flat-
round lips and represent large, open bowls. The paste is temperless and
grey-black in color. The sherds are 0. 4 to 0. 7 cm. thick. Decoration con-
sists of series of parallel vertical and diagonal incised lines within triangular
St. Johns Plain (15 sherds) (Fig. 4, b). There are two varieties of this
pottery type. Six sherds have a soft, temperless paste. These are 0. 5 to
1. 1 cm. thick, the average being 0.7 cm. The remainder have a hard, gritty
paste containing a small quantity of very fine sand. The thickness of these
ranges from 0. 4 to 0. 6 cm. with an average of 0. 5 cm. One sherd also con-
tains very small, infrequent clay particles.
Pasco Plain (17 sherds) (Fig. 4, a). These sherds may be grouped into two
categories. In, eight sherds the tempering consists of fine sand and an abun-
dance of small to large pieces of crushed limestone. These are 0. 6 to 0. 9
cm. thick with an 0. 8 cm. average. The rest are sparsely tempered with
small limestone particles and probably represent an early variety of this type.
Thickness varies from 0.7 to 1.3 cm. the average being 0.9 cm. Paste
color in both types is grey to black,
Pinellas Plain (4 sherds) (Fig. 4, c). The paste is contorted and laminated.
Tempering consists of fine sand with one sherd also containing small pieces
of crushed clay. The lips on rims are rounded. One rim narrows towards
towards the lip which bears a series of notches set 1. 0 cm. apart. The other
rim has a plain lip. Rim sherds are 1. 0 to 1. 1 cm. thick; body sherds are
0. 6 to 1.0 cm. thick. Paste color is buff to black.
sand-tempered plain (142 sherds). The majority of these are tempered with
fine sand. Only a few sherds exhibit coarse particles. Thickness varies from
0. 5 to 1, 5 cm. 0. 9 cm. average. Only thirteen sherds are over one centi-
meter thick. Almost all of these sherds were found in the sand pumped out of
the bay and are quite water worn.
sand-tempered incised (1 sherd) (Fig. 4, d). One sherd bears an incised
line with a diagonal running off from it. It is tempered with fine sand and is
0. 7 cm. thick.
Suwannee (2 specimens) (Fig. 3, a, e). Large, lanceolate points with con-
cave bases and fairly prominent basal ears. The blade is waisted slightly
just above the ears. On one specimen the waiting is slight on one side and
almost non-existant on the other. Blade edges are excurvate above the waist.
The blade edges are ground for a distance of 2. 5 to 3. 0 cm. up from the cor-
ners of the ears, as is the base. One base has been thinned slightly through
the removal of several flakes from both sides. The other is thick and blunt.
Workmanship ranges from good to poor. Both appear to have been percussion
flaked from large chert flakes, and are represented by basal fragments.
Length is 11.0 cm. (estimated) 4.3 to 8.4 cm. (existing), width 3.9 to 4.0 cm,,
and thickness 0.8 to 1.0 cm.
Bolen Plain (1 specimen). A side notched point with straight blade edges and
a slightly concave base with rounded corners. Both the base and notch inter-
iors are ground. Notches are broad (0.8 cm. across) and shallow. Length
is 3.4 cm. (existing), width 2.3 cm. thickness 0.7 cm.
Florida Archaic Stemmed (10 specimens) (Figs. 2, a-d; 3, c). Triangular-
bladed points with contracting or parallel sided tangs. Two subtypes are re-
presented: Putnam (7) and Levy (3). The Putnam points (Fig. 2, d; 3, c)
have straight (4), excurvate (2), or slightly incurvate (1) blade edges. Tang
edges are recurvate and continuous. On two specimens the tang is set off
slightly to one side of the blade. Basal blade edges slope upward and outward.
Shoulders are angular and sharp.
The Levy types (Fig. 2, a-c) have contracting tangs with straight (2)
or concave (1) bases bearing angular corners. The tang with the concave
base also exhibits basal thinning. The basal blade corners droop to form
slight barbs on two specimens. The third point has basal blade edges that
slope upward and angular shoulders. Blade edges are excurvate on all three..
Length is 5. 1 to 8. 0 cm. 6. 7 cm. average; width 2. 9 to 4. 8 cm. average
3.4 cm.; thickness 0.6 to 1.6 cm. average 1.0 cm.
Florida Morrow Mountain (1 specimen) (Fig. 3, b). A large, heavy, tri-
angular-bladed point which is very crudely percussion flaked. The broad,
rounded tang is set off to one side of the blade. The cross section of the blade
is trianguloid, with the thickest portion being on the same side as the tang.
The blade tapers to a much thinner, sharp edge on the opposite side. The
blade edge on the tang side is slightly concave; the other is convex. The tip
is blunt. The asymmetric characteristics indicate that this artifact repre-
sents a knife, rather than a projectile point. Length is 8. 1 cm. width 5. 5
cm. and thickness 1.4 cm.
FISH CREEK SITE
Newnan (2 specimens) (Fig. 3, g-h). Unlike the classic Newnan points,
the examples from Fish Creek have horizontal basal blade edges, rather
than those which slope downward toward the shoulders. Both specimens are
very well made, and have contracting tangs with straight edges. One point
has a squat, triangular body with a long tang. Blade edges are excurvate,
tang base is flat, and all corners are sharp and angular. The other point is
elongate with slightly convex blade edges. Tang base is excurvate and the
corners are rounded slightly. However, upper tang corners are sharp.
Length is 4. 6 to 6.7 cm. width 3. 5 to 3. 9 cm. and thickness 0. 5 to 0. 7 cm.
Culbreath (8 specimens) (Figs. 2, e-f; 3, f, j). Medium to large sized
points with parallel sided tangs and drooping barbs. On four specimens the
barbs are small, on three they are long and prominent, and on one large
point they are very bold, almost reaching to the base of the tang. Sides are
excurvate on all but one, straight sided specimen. Tang bases are either
flat (3) or wry slightly convex (5). All appear to have been percussion flaked
and workmanship ranges from good to fair. Length is 4. 6 to 5. 9 cm. 5. 2 cm.
average; width 2. 9 to 5. 5 cm. 3. 8 cm. average; thickness 0. 6 to 1. 1 cm.,
0. 8 cm average.
Lafayette (3 specimens) (2, g). Corner notched points with drooping barbs
and expanding tangs bearing slightly convex bases. Blade edges are excur-
vate. Workmanship is poor to crude. One specimen is unifacially chipped
except for retouching along the edges on the unaltered side. Length is 5. 3 to
7. 0 cm. 6. 2 cm. average; width 3. 4 to 4. 0 cm. 3. 7 cm. average; thickness
0. 6 to 0. 9 cm. 0. 7 cm. average.
Westo (1 specimen). This type has almost straight blade edges and a poorly
shaped, rounded tang. Blade corners are slightly rounded, but distinct. The
example from Fish Creek most closely resembles Bullen' s (1968: 25) Westo
sub-type 2. The material is chert. Length is 4. 4 cm., width 3. 0 cm., and
thickness 0. 7 cm.
Citrus (9 specimens) (Figs. 2, h-j; 3, d). These are basally notched points
with broad tangs and excurvate bases. Notches vary in depth from 0. 4 to 0. 9
cm. Sides are straight (5), excurvate (3), or asymmetrical (1). The latter
probably served as a knife (Fig. 2,i). Five points are well made, the re-
mainder are poorly percussion flaked. Length is 4. 4 to 7. 4 cm. 5. 8 cm.
average; width 3. 2 to 4. 9 cm. 4. 1 cm. average; and thickness 0. 6 to 1. 0
cm. 0. 7 cm. average.
Hernando (1 specimen) Fig. 3, i). Though these points are usually well
made, this specimen is very crudely manufactured. It is basally notched;
the tang and barbs are rounded and almost equal in size. The base is convex.
One blade edge is straight, the other is excurvate. Length is 2. 9 cm. (exist-
ing), width 2. 9 cm., and thickness 0. 6 cm.
Bradford (2 specimens) (Fig. 3, k). These points have expanding tangs
with rounded corners and slightly excurvate bases. Blade edges are excur-
vate; shoulders are weak. Workmanship is poor. Length is 4. 6 to 6. 5 cm.,
width 2. 2 to 3. 1 cm. and thickness 0. 6 to 0. 8 cm.
Pinellas (1 specimen). A small, thin point with straight sides. The base is
strongly concave, leaving the basal corners projecting as very slender, sharp
barbs. Both blade faces are finely pressure flaked. Length is 4. 4 cm.,
width 2.3 cm. and thickness 0. 5 cm.
"Fish Creek" Points (2 specimens) Fig. 2, n-o). This type is distinct from
all of the corner notched points described by Bullen (1968). These are large,
ovoid specimens with excurvate blade edges, expanding tangs with strongly
convex bases bearing distinct corners, and small to medium sized, sharp,
drooping barbs. The notches are situated at the imaginary "corners" of the
base. A line drawn from barb to barb, including the tang base, is continuous
and strongly excurvate. Notches are as deep as they are wide (ca. 0. 8 cm.).
Both points are bifacially chipped and exhibit good workmanship. Length is
7.9 to 8.0 cm. width 3.9 to 4.0 cm. and thickness 0.8 to 0.9 cm.
Unattributable Fragments (4 specimens). Three tips and one body fragment
are represented in the collections.
Triangular Knives Knives in this category have a triangular outline. Bases
are flat to strongly convex. Basal blade corners are distinct and the greatest
width is across the base. Blade edges are straight to slightly convex. Tips
are usually pointed, but may be blunt. These knives may be separated into
four groups on the basis of size and shape.
Group 1. Short Narrow Triangular (11 specimens) (Fig. 2, r-s). Short
knives with slightly excurvate bases and sides. One tip is blunt, the remain-
der are pointed. All are bifacially percussion flaked from longitudinally cur-
ved flakes. One specimen is pressure flaked along one edge forming a slight
bevel and it may also have served as a scraper. Length is 5. 2 to 6. 9 cm. ,
5. 7 cm. average; width 3. 0 to 3. 9 cm. 3. 5 cm. average; and thickness 0. 6
to 1.3 cm. 0.9 cm. average.
Group 2. Long Narrow Triangular (1b specimens) (Figs. 2, x; 3, i).
Elongate knives with slightly to strongly convex bases. Basal blade corners
are distinguishable though they tend to be rounded on specimens with very
excurvate bases. Blade edges are straight for most of the length, converg-
ing near the tip to form blunt points. These are poorly made, percussion
flaked tools. One specimen is very crude and has a plano-convex cross
FISH CREEK SITE
section. Blade length is approximately equivalent to twice the width. Length
is 7. 2 to 9. 7 cm. 8.3 cm. average; width 3. 2 to 5. 2 cm. 4. 4 cm. average;
and thickness 0. 8 to 1.7 cm. 1 cm. average.
Group 3. Broad Triangular (7 specimens) (Fig. 2, w). These have wide
blades with excurvate bases. Blade edges are straight except near the tip
where they curve inward slightly, forming rounded points. Most examples
are crudely percussion flaked and the surfaces bear large flake scars. Blade
width is equivalent to at least two thirds of the length. Length is 4. 8 to 9. 7
cm. 7. 0 cm. average; width, 4. 0 to 6. 6 cm. 5. 0 cm. average; and 0. 8 to
1.5 cm, 1. 1 cm. average.
Group 4. Asymmetrical Triangular (23 specimens). This group contains
the asymmetrical varieties of the previously mentioned triangular knives.
Two types are distinguishable:
Type A. (19 specimens) (Fig. 2, u-v, z). One blade edge is straight
for most of its length and is at approximately a right angle to the base which
is almost perfectly straight on all but three specimens. The other edge is
either slightly to strongly excurvate (15) or is also straight but at much less
than a ninety degree angle to the base (4). One example ( Fig. 2, z) has an
extremely convex blade edge composed of three distinct, straight sections
connected by rounded corners, each set at about a forty-five degree angle to
the other. Due to the strong curve of the blade, the widest point is 2. 5 cm.
above the base.
On four knives the basal corner on the excurvate side is rounded. Five
specimens have the opposite corner rounded. The remainder have corners
both of which are either angular or rounded. One knife is beveled along the
excurvate blade edge. Another is unifacially flaked, one side consisting of
the original flake surface which is retouched only along the edges. Length
is 5. 6 to 8. 3 cm. 7. 0 cm. average; width is 2. 2 to 6. 3 cm. 4. 3 cm. aver-
age; and thickness 0. 5 to 1.4 cm., 0. 9 cm. average.
Type B. (4 specimens) (Fig. 2, t). This type has one blade edge
which is straight but forms an angle with the flat base that is less than ninety
degrees. The other edge is parallel to it for about half the length of the arti-
facts and then angles inward to meet it in a blunt point. These are rather
crudely made, one specimen bearing a surface consisting of the original flake
surface retouched only at the edges. Length is 5. 9 to 10. 5 cm. 8. 0 cm. aver-
age; width 3. 1 to 5.0 cm., 4.0 cm. average; and thickness 1. 1 to 1.4 cm.,
1. 2 cm. average.
Ovate Knives These knives have an ovate shape. Bases are convex and the
juncture of base and blade edges is indistinguishable. Greatest width is usu-
ally well above the base due to the lack of distinct basal blade corners and
the convexity of the base. Blade edges are excurvate for the most part, but
may be straight for short distances. Tips are pointed or blunt. Seven
groups are distinguishable on the basis of shape:
Group 1. Oblong Ovate (4 specimens) (Fig. 2y, cc). Knives in this
group are long and narrow, their length being roughly twice the width.
Greatest width is from one quarter to one third of the way up from the base.
Blade edges are straight at the middle but curve inward at the tip and base.
Tips are pointed or blunt. All specimens are crudely percussion flaked.
Length is 8. 3 to 11. 1 cm. 9.7 cm. average; width 4. 5 to 5. 5 cm. 4. 9 cm.
average; and thickness 1.2 to 1. 5 cm. 1.4 cm. average.
Group 2. Elongate Ovate (2 specimens) (Fig. 2, bb). These are nar-
row, elongate knives with convex blade edges. Greatest width is almost at
the middle of the blade. The body tapers toward the tip which is missing on
both examples but was probably pointed. The base is strongly excurvate.
Both are percussion flaked; one is very crude. Length is 7. 5 to 8. 8 cm.
(estimated), width is 3. 6 to 4. 0 cm., and thickness is 1.3 to 1.4 cm.
Group 3. Broad Ovate (2 specimens) (Fig. 2, aa). Squat knives with
the greatest width about one third of the way above the base. Width is equal
to about two thirds of the length. Sides are excurvate, tips are pointed.
Workmanship is poor. One specimen is unifacially flaked, one side being re-
touched only along the edges. Length is 8. 0 cm., width is 5. 4 cm. and thick-
ness 1.2 to 1.3 cm.
Group 4. Rounded Ovate (1 specimen). A very crude, almost circular
knife fashioned from a lamellar chert fragment. Blade edges are excurvate
and continuous. Both surfaces bear the original, rough cortex. The tool is
percussion flaked along the entire circumference. However, only one side of
the blade is chipped to a sharp edge; that opposite it is blunt. Length is 8. 0
cm., width is 6. 0 cm. and thickness 2. 0 cm.
Group 5. Bi-point Ovate (1 specimen). A large knife with excurvate
blade edges and blunt points at both ends of the blade. The shape is some-
what asymmetrical in that one blade edge is more strongly convex than the
other. Length is 8. 8 cm., width 6.3 cm. and thickness 1.2 cm.
Group 6. Leaf-shaped Ovate (1 specimen) (Fig. 2, gg). A very crude,
percussion flaked knife with one surface consisting of the original flake sur-
face retouched only along the edges. The base is very slightly excurvate;
one corner is sharp and angular, the other is rounded. Blade edges are ex-
curvate and meet in a sharp, but lopsided tip. The widest point is at the mid-
dle of the blade. Length is 8.7 cm. width 4. 5 cm. and thickness 1. 7 cm.
Group 7. Asymmetrical Ovate (11 specimens). This group contains
FISH CREEK SITE
the asymmetrical varieties of the previously mentioned knives. There are
Type A. (8 specimens) (Fig. 2, dd). One blade edge is slightly ex-
curvate, the other is strongly convex. Bases are only slightly convex, for
the most part, with the central portion being almost flat. Tips are pointed
and sharp. Workmanship ranges from good to poor. Length is 6.7 to 8.5
cm. 7. 7 cm. average; width is 4. 0 to 6. 9 cm. 4. 8 cm. average; and thick-
ness is 0.6 to 1. 6 cm. 1.3 cm. average.
Type B. (1 specimen) (Fig. 2, ff). One blade edge is almost perfectly
straight for most of its length, then curves inward at the tip. The other edge
is excurvate. The tip is blunt. This knife is finely pressure flaked. Length
is 7. 3 cm. width is 3. 6 cm. and thickness 0. 9 cm.
Type C. (2 specimens) (Fig. 2, ee). One blade edge is slightly excur-
vate. The other is strongly convex near the base, then recurves to form a
narrow, tapering tip with a blunt point. Bases are excurvate. These are
poorly percussion flaked tools. Length is 6. 0 to 6. 8 cm. width, 3. 1 to 3. 3
cm. and thickness 1.0 to 1. 1 cm.
Miscellaneous Knives This category contains the knives which are neither
triangular nor ovate in shape. Five groups are distinguishable:
Group 1. Elongate Knives (4 specimens) (Fig. 2, hh). These are long,
narrow knives with sides that are parallel for about two thirds of the total
length. The blade edges then converge to form a bluntly pointed tip. The
basal edge is straight on all specimens. On two examples the blade edges
are perpendicular to the base, on the others the base is set obliquely to the
edges (Fig. 2,hh). All specimens are crudely percussion flaked and only
one is complete. Length is 8. 1 cm. width 2. 6 to 2. 7 cm. and thickness
0. 8 to 0. 9 cm.
Group 2. Concave Base Elongate (1 specimen) (Fig. 4, 0). This
type is represented by a basal fragment. The blade is narrow and almost
parallel sided. The base is slightly concave and has been thinned by the re-
moval of several flakes. It resembles a Suwannee point except for the lack
of basal grinding and waiting. This specimen may possibly represent a
"Suwannee knife" or an unfinished Suwannee point. Length is 7. 3 cm.
(existing), width 4. 1 cm. and thickness 0. 8 cm.
Group 3. Trapezoidal Knives (2 specimens) (Fig. 4, f). Squat knives
with relatively straight edges. Blade edges are parallel for most of the length
then one side angles inward near the tip. The sides are at right angles to the
base. These are very crude, bifacially percussion flaked tools. Length is
5.2 to 5.7 cm. width 3.8 to 3.9 cm. and thickness 1.0 to 1. 1 cm.
Group 4. Flake Knife (1 specimen) (Fig. 4, e). This specimen is
roughly rectangular in outline, with fairly straight edges. The flake is re-
touched only along one side and end, their juncture being marked by a sharp
corner. The opposite corner is rounded. The longest edge is chipped on
both sides, the end is worked only on one side. The side opposite the pri-
mary cutting edge bears the original core cortex and is unaltered. Length
is 4.9 cm., width 2. 8 cm., and thickness 1.4 cm.
Group 5. Hafted Knives (2 specimens) (Fig. 2, k). These knives are
asymmetrical and have basally notched, convex, Citrus-like bases. One
specimen (Fig. 2, k) bears a blade edge which is straight for two thirds of
its length and perpendicular to the basal edge. It then curves inward, form-
ing a sharp point with the other, strongly excurvate side. Basal notches are
small, measuring only 0. 3 cm. in depth. The widest point is at the center of
the blade. Workmanship is very good. The blade is elongate.
The other example has a broad, ovate blade with a strongly convex
edge opposite one which is convex near the base and almost straight near the
tip. Only one wide notch is present. The other corner of the blade is crudely
shaped and appears to be unfinished. The greatest width is one third of the
way up from the base. Length is 5. 6 to 6. 6 cm. width is 3. 5 to 4. 4 cm. and
thickness 0. 9 to 1. 1 cm..
Unattributable Fragments (68 specimens). There are 19 basal fragments
with rounded (12) or angular (7) corners. Of 44 tips, 27 are narrow, 17 are
broad. There are five body fragments.
Other Chipped Tools
Plano-convex End Scrapers
Group 1. Irregular Flake Scrapers (32 specimens) (Fig. 4, h). These
consist of irregular flakes retouched only along the working edge. The slightly
to steeply beveled, excurvate, working end is always situated opposite the bulb
of percussion. The convex flake surface is either keeled (10) or flaked ir-
regularly (22). The working edge may be symmetrical, being confined to the
end of the tool (25), or it may be asymmetrical and also extend down one side
(7). Length is 4. 1 to 6. 5 cm. 5.2 cm. average; width is 2.4 to 5. 2 cm. 3.5
cm. average; and thickness 0. 6 to 1.5 cm. 1.0 cm. average.
Group 2. Ovate End Scrapers (3 specimens) (Fig. 4, i). These are
roughly ovate in outline and are chipped along all edges although the most
careful flaking appears on the edge opposite the bulb of percussion. The
dorsal surface is flaked over its entirety. In all cases the bulb of percussion
FISH CREEK SITE
and the striking platform have been removed by flaking these areas on the
plano side. The widest point is at the steeply beveled, working end. Length
is 4. 5 to 5. 5 cm. 5. 0 cm. average; width 4. 0 to 4. 4 cm. 4. 1 cm. average;
and thickness 1.0 to 1.8 cm. 1.5 cm. average.
Bifacially Flaked End Scrapers
Group 1. Ovate Scrapers (Z specimens). These are roughly ovate in
outline and are worked on both flake surfaces. One specimen is very thin and
flat with a wide, steeply beveled working edge. The other is thick and has
been flaked so that a central, longitudinal ridge with a triangular cross sec-
tion was formed. This artifact is especially carefully flaked along the work-
ing end. Length is 4. 3 to 5. 6 cm. width 3. 5 to 3. 7 cm. and thickness 0. 4
to 2. 4 cm.
Group 2. Hafted Scrapers (3 specimens) (Fig. 2, 1-m). All speci-
mens have slightly to strongly convex bases with Citrus-like, basal notches.
One very well made scraper (Fig. 2, m) has notches that are one centimeter
deep. The others are crudely percussion flaked and the notches are shallow.
Tangs are expanding (2) or contracting (1). The working end is gradually to
very steeply beveled. Length is 3. 6 to 4. 9 cm. 4. 2 cm. average; width is
4. 2 to 4. 8 cm. 4. 6 cm. average; thickness 0. 6 to 0. 8 cm. 0. 7 cm. average.
Plano-convex Side Scrapers
Group 1. Irregular Flake Scrapers (15 specimens). These consist of
elongate flakes chipped along one or both longitudinal edges. The remaining
surfaces are unworked. The convex surfaces are usually keeled. A number
of examples have a lunate outline. These are fairly large tools, averaging
8. 0 cm. in length, 4. 4 cm. in width, and 2. 4 cm. in thickness.
Group 2. Leaf-shaped Scraper (1 specimen) (Fig. 4, g). This type
has a flat base formed by an unaltered striking platform. Blade edges are ex-
curvate; the tip is blunt. The convex surface is flaked over its entirety. The
other side is unaltered. The edges are gradually beveled. The widest point
is one third of the way up from the base. Length 5.7 cm. width 3. 5 cm.,
and thickness 0. 9 cm.
Bifacially Flaked Side Scrapers
Group 1. Oblong Ovate (1 specimen). A crudely percussion flaked oblong
scraper with one strongly excurvate edge chipped for use. The remaining sur-
faces are less carefully worked. The ends are rounded. Length is 6. 3 cm.,
width 3.4 cm., and thickness 1. 9 cm.
Group 2. Sub-triangular (1 specimen) (Fig. 3, m). One edge is
straight, the other is straight for two thirds of its length down from the
blunt tip. It then curves inward to form an off-center, convex base. The
artifact is poorly percussion flaked. Length is 7.4 cm. width is 2.9 cm.,
and thickness 2. 0 cm.
Knife-Scrapers These are, for the most part, large and heavy, bifacially
flaked core tools. Only two specimens have areas of unworked surface due
to manufacture from lamellar chert fragments. Edges on all are sharp.
Chipping and shaping are very crude. Three groups are distinguishable on
the basis of shape:
Group 1. Triangular (4 specimens) (Fig. 4, m). Sides are slightly
excurvate, corners are distinct, tips are pointed.
Group 2. Ovate (3 specimens) (Fig. 4, 1). Sides and bases are con-
vex, basal corners are weak or indistinguishable, tips are pointed.
Group 3. Oval (6 specimens) (Fig. 4, n). Circular specimens with
continuous, excurvate edges. This group contains the smallest examples.
Lengths are 5. 8 to 13. 5 cm. 9. 2 cm. average; widths 5.3 to 8.3 cm. 6. 9
cm. average; thickness 1.6 to 3.0 cm. 2.5 cm. average. Four unattribu-
table tips from pointed specimens are also in the collection.
Stemmed Drill (1 specimen) (Fig. 2, q). This type was apparently made by
reworking an archaic stemmed point. The tang is tapered with a slightly con-
vex base and rounded corners. The tang begins just inside the shoulders of
the blade. The tip has been flaked to form a narrow, off-set point. Length
is 5.9 cm. width, 2.7 cm. and thickness 0.8 cm.
Waisted Drill (1 specimen) (Fig. 2, p). The blade edges just above the base
are constricted slightly. Above the waisted area the sides are parallel to a
point two thirds of the way up from the base. The blade edges then angle in-
ward to form a sharp point. The base is slightly convex and semi-fluted on
one side by the removal of a centrally located flake. The other side has no
flute. Length is 5.4 cm., width 2.4 cm., and thickness 1.0 cm.
Elongate Drill (1 specimen). This is represented by a medial fragment from
a narrow-bladed specimen. The blade edges are roughly parallel for the ex-
isting length of the fragment. Length is 2.5 cm. (existing), width is 1. 6 cm.,
and thickness 0. 6 cm.
Pick This is a lunate artifact with blunt ends, one of which is wider than the
other. The ends and edges exhibit battered areas. Length is 7. 3 cm. width
1. 8 to 3.8 cm. and thickness 2. 3 cm.
FISH CREEK SITE
Unfinished Tools. Unfinished knives or projectile points are represented by
one triangular and two ovate specimens. All three exhibit well shaped tips.
However, the rest of the blade is rough and unfinished. Each specimen is
missing a section of the base and was probably discarded when it broke in
the process of manufacture. Length is 6. 8 to 8. 8 cm. 7. 7 cm. average;
width is 4. 5 to 5. 1 cm. 4. 8 cm. average; and thickness 0. 9 to 1.4 cm.,
1. Z cm. average.
Miscellaneous Stone Tools
Hammerstones are represented by thirteen, small to large, rounded
chert or silicified coral cobbles exhibiting battered edges. Five flat sand-
stone fragments are present in the collection and probably represent abra-
ders. Also recovered from the site were a large quantity of bifacial and
"horse' s hoof" cores, and a small, fossilized shark tooth.
Shell tools are represented by four perforated Melongena hammers, a
possible conch shell gouge, and what appears to be a columella pendant. The
latter is very water worn but has a transverse groove near the tip of the nar-
Discussion and Conclusions
Since all of the artifacts from Fish Creek are surface finds, a complete
discussion of the cultural sequence is impossible. Much valuable information
concerning artifact associations and their chronological position is not avail-
able. Thus, no specific inferences can be made regarding occupational in-
tensity or duration since only ten per cent of the total artifact assemblage
consists of datable items with a narrow temporal range. Nevertheless, much
can be gleaned from the material collected to further the knowledge of the
archaeology of the Tampa Bay area.
On the basis of the two Suwannee points, the initial occupation occurred
during the late Paleo-Indian period, ca. 9, 000 B. C. (Bullen 1968: 48). Con-
tinued use or visit, during the Early Preceramic Archaic around 7,000 B. C.,
is suggested by the single Bolen point (Bullen 1968: 42). Then, judging from
the greatly increased quantity and variety of the projectile points, utilization
of the site became much more intensive during the latter half of the Late Pre-
ceramic Archaic, ca. 4,000 2,000 B. C. This is indicated by the Archaic
Stemmed, Morrow Mountain, Newnan, and Culbreath points (Bullen 1968: 6).
The Orange period may be represented by some of the Archaic Stemmed and
Lafayette points (Bullen 1968: 6). An Orange occupation is substantiated by
the Norwood Plain sherds (Phelps 1965: 67). The Transitional period is re-
presented by the Lafayette points, as well as the St. Johns Incised pottery.
It will be noted that the Archaic Stemmed and Lafayette points may be
ascribed to more than one archaeological period. Since no stratigraphy is
present at the site, it is uncertain whether they are associated with only one
or both of the periods involved. The ceramics, however, indicate that there
is a general continuity of occupation, regardless of actual projectile point
A Perico Island (Deptford) occupation is suggested by the Westo, Citrus,
Hernando, and Bradford points (Bullen 1968: 6), as well as the basally notched,
Citrus-like knives and scrapers. However, there are no Deptford or Perico
series sherds in the collection to substantiate this. The St. Johns Plain and
Pasco Plain sherds are poor time markers, but some of them are probably
assignable to either the Transitional or Perico Island period.
There are no artifacts specifically diagnostic of the Swift Creek and
Weeden Island periods. It is therefore assumed that the site was abandoned
during this time. The next, and last, occupation occurred during the Safety
Harbor period as evidenced by the one Pinellas point and the Pinellas Plain
The majority of the artifacts recovered from the Culbreath Bayou site,
about two miles to the south, are duplicated in the Fish Creek collection.
Culbreath Bayou is a stratified, two component site. The earliest component
consists of a Late Preceramic Archaic workshop. It is characterized by a
predominance of Culbreath points and a few Archaic Stemmed points, as well
as ovate and trianguloid knives, plano-convex end scrapers, and stone picks
(Warren 1967: 159). The knives from the workshop (Warren 1967: Fig. 7)
correspond to those from Fish Creek in Groups 2 and 4A in the triangular and
Groups 1 and 3 in the ovate categories.
The second component is represented by a Perico Island midden con-
taining a predominance of sand-tempered plain, some Pasco Plain, and minor
quantities of St. Johns Plain and Deptford series sherds (Warren 1967: 158).
Citrus and Hernando points appeared in the lowest levels. Comparing the two
sites, the above data tend to corroborate a Late Archaic and Perico Island
occupation for Fish Creek.
A lack of Orange and Transitional ceramics at Culbreath Bayou suggests
that this site was not in use during those times (Warren 1967: 162). Although
sherds distinctive of these periods are present at Fish Creek, they are few in
number and may indicate that this site was then only intermittently used with
an occupational rejuvenation during the Perico Island period. As at the Fish
Creek site, abandonment of the Culbreath Bayou site appears during the Perico
Hernando, Westo and Bradford points, bifacial end scrapers, ovate and
FISH CREEK SITE
triangular knives, and numerous spalls have also been recovered from the
southwestern section of the Tampa International Airport, about a half mile
east of the Fish Creek site (Fig. 1). Thus, it appears that this general area
was in use as a workshop during the Perico Island period, if not earlier.
The Fish Creek site yielded a distinctive corner notched projectile
point which appears to be new to Florida archaeology. The shape of the base
most closely resembles that of the Plevna points of Alabama (Cambron and
Hulse 1964). However, the lack of such Plevna characteristics as basal
grinding and beveled blade edges on the Fish Creek specimens indicates that
the two types are not equivalent. Nevertheless, these points may represent
an intrusive holdover omitting the earlier Bolen-like characteristics. As
more sites are excavated, it will be interesting to see if they appear else-
where. If they do, then their age, distribution, and relationship to the other
types may be determined. This information will further our knowledge of the
technology and possible origins of the early prehistoric cultures in Florida.
In conclusion, the Fish Creek site may be added to the ever-growing
list of early submerged sites on the central west coast of Florida. It is im-
portant as another source of Suwannee points and may be considered as one
of the earliest sites in Hillsborough County.
Special thanks are extended to Ripley P. Bullen of the Florida State
Museum for making the collections (Nos. 103189 and 103451) donated by Dr.
Lyman O. Warren to the Museum available for study and for providing photo-
graphs of this material. Al Goodyear also deserves thanks for permitting the
study of his collection. Bill Harrison' s help in photographing the author' s
artifacts is greatly appreciated.
Fig. 2. Artifacts from mud flat at the Fish Creek site.
FISH CREEK SITE
0 I 2 3
Fig. 3. Projectile points from dredged fill (Florida State
Museum collection), knife and scraper from mud flat.
Fig. 4. Ceramics and chipped stone tools from Fish
Creek (j-k, o in Florida State Museum Collection).
Bullen, Adelaide K. and Ripley P.
1953 The Battery Point Site, Bayport, Hernando County, Florida.
The Florida Anthropologist, Vol. 6, No. 3, pp. 85-92.
Bullen, Ripley P.
1968 A Guide to the Identification of Florida Projectile Points.
Florida State Museum, University of Florida. Gainesville.
Cambron, James W. and David C. Hulse
1964 Handbook of Alabama Archaeology, Part 1, Point Types.
Archaeological Research Association of Alabama, Inc.
Phelps, David S.
1965 The Norwood Series of Fiber-Tempered Ceramics.
Southeastern Archaeological Conference Bulletin, No. 2,
pp. 65-69. Cambridge.
Warren, Lyman O., William Thompson, and Ripley P. Bullen
1967 The Culbreath Bayou Site, Hillsborough County, Florida.
The Florida Anthropologist, Vol. 20, No. 3-4, pp. 146-163.
Willey, Gordon R.
1949 Archaeology of the Florida Gulf Coast. Smithsonian
Miscellaneous Collections, Vol. 113. Washington.
February 28, 1969
Copies of Volumes I through XIV may be purchased from Johnson
Reprint Corporation, 11l Fifth Avenue, New York, N. Y. 10003 at $8.00
per volume or $2.00 for single and $4.00 for double numbers if available.
Copies of Volumes XV through XXIII may be purchased from the Treasurer
of the Society at $4.00 per volume or $1.00 for single and $2.00 for double
numbers. The Society has no resale discount for book dealers.
INFORMATION FOR AUTHORS
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) and photographs--
should be included in one numbered series of "Figures" suitable for a maxi-
mum reduction to 6 1/2 by 8 1/2 inches. Parts of figures are designated by
lower case letters. Figures must be accompanied by brief descriptive cap-
tions. All tabular material must be included in separately numbered "Tables. "
Twenty-five reprints without covers of each article are furnished free
to each author or group of authors. Additional copies, withor without covers,
may be purchased but the editor must be informed of the number desired when
the article is offered for publication.