Citation
The Safety Harbor site, Pinellas County, Florida

Material Information

Title:
The Safety Harbor site, Pinellas County, Florida
Series Title:
Florida Anthropological Society. Publications - Florida Anthropological Society ;
Creator:
Griffin, John W
Bullen, Ripley P. ( author )
Place of Publication:
Gainesville
Publisher:
Pub. at the University of Florida
Manufacturer:
Pepper Print. Co.
Publication Date:
Language:
English
Edition:
Number 2, 1950
Physical Description:
42 pages : illustrations, maps ; 23 cm.

Subjects

Subjects / Keywords:
Indians of North America -- Antiquities -- Florida ( lcsh )
Antiquities ( fast )
Indians of North America -- Antiquities ( fast )
Antiquities -- Pinellas County (Fla.) ( lcsh )
Florida ( fast )
Florida -- Pinellas County ( fast )
Genre:
serial ( sobekcm )
bibliography ( marcgt )

Notes

Bibliography:
Includes bibliographical references (pages 36-37).
Statement of Responsibility:
John W. Griffin and Ripley P. Bullen.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Florida Anthropologist Society, Inc. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
01936060 ( OCLC )
ocm01936060
Classification:
GN2 .F55 no.2 ( lcc )
913.75905 ( ddc )

Full Text
FLORIDA ANTHROPOLOGICAL SOCIETY
PUBLICATIONS
Number Two
THE SAFETY HARBOR SITE
PINELLAS COUNTY, FLORIDA
John W. Griffin
and
Ripley P. Bullen
Published at the University of Florida
Gainesville
1950




FLORIDA ANTHROPOLOGICAL SOCIETY
PUBLICATIONS
Number Two
e
THE SAFETY HARBOR SITE
PINELLAS COUNTY, FLORIDA
John W. Griffin
and
Ripley P. Bullen

Published at the University of Florida
Gainesville


John M. Goggin
Editor
Printed by the Pepper Printing Company
Gainesville, Florida


ACKNOWLEDGMENT
In recognition of his cooperation and interest in
archaeology, particularly for his encouragement
and assistance with the work at the Safety Harbor
site, this publication is dedicated to Mr. Lucius S.
Ruder.




CONTENTS
Page
Introduction 7
Typology 10
Pottery 10
Stone Tools 12
Plummets 13
Shell Tools 14
Bone Tools 14
The Village Area 15
Excavations 15
Other Collections 23
The Burial Area 27
Summary of the Safety Harbor Site 29
Conclusions 33
Bibliography ; 36
ILLUSTRATIONS
Plates
I.Miscellaneous Sherds from the Safety Harbor Site 39
II.Miscellaneous Chipped Stone Tools 40
III. Plummets, Shell Tools, Daub, and Hammerstones 41
IV. Bone Tools, Pottery Hone, and Grindstone Fragments 42
Figures
1. Map of Safety Harbor Site 6
2. Profiles from Large Mound and Village 16
TABLES
1. Stratigraphic Tabulation of Pottery from Large Mound 13
2. Stratigraphic Tabulation of Pottery from Area B 20
3. Stratigraphic Tabulation of Pottery from Area C 21
4. Stratigraphic Tabulation of Points from Area C 23
5. Tabulation of Pottery from Village Area 24
6. Tabulation of Stone, Shell, and Bone Tools 25


Fig. 1. Map of Safety Harbor Site


INTRODUCTION
The Safety Harbor site is located in the northeast quarter of
Section 34 and the northwest quarter of Section 35, T. 28 S., R. 16 E.,
Pinellas County, Florida. It occupies a point of land, known as
Phillippi Point, which divides Old Tampa Bay from its northern
arm, known as Safety Harbor (Fig. 1). The map given as Figure
1 discloses that the site is on high land, at least for Florida. Bluffs
about twenty feet in height border the water on two sides of the site.
The large mound, about 150 feet across, rises over twenty feet
above the twenty-five foot ridge on which it stands. A ramp leads
down the west side of the mound, away from the water. The summit
plateau is about 100 by 50 feet in size, and quite level. The natural
steepness of the sides has been accentuated on the east by wave cut
ting during storms of years past. Reference to the map will show
that the mound is located at the east end of Phillippi Point.
From the mound the village debris runs in an L-shaped plan,
paralleling the shore to the northwest and southwest. To the noi'th
of the mound at least twenty feet, probably substantially more, of
the village area has washed away during the last forty years. There
is little or no debris back of the slight ridge paralleling the shore.
About a quarter of a mile northwest of the large mound is located
the burial mound, excavated some years ago by Stirling (Fig. 1).
The large mound itself, and the village area lying northwest of
it is in dense hammock vegetation. Most of the southwestern arm of
the village has been cleared, and behind the ridge lies an old citrus
grove. The village continues across the road shown on the map.
The site was settled by Count Odet Phillippi sometime in the early
nineteenth century, but the exact date seems to be in dispute at the
moment. Straub says that Phillippi came to Tampa in the early
30s, and settled at Phillippi Hammock in 1835.1 Newspaper ac
counts, which constitute the only other available evidence at the
moment, and which seem to be based on tradition in the area, give
the date as 1823.2 This latter date would seem to be in error how
ever for Phillippi is not listed in the U. S. census of 1830 for the
Tampa Bay region.3 Phillippi was a colorful person. The tradi
tional story of his life attributes to him the rank of chief surgeon
of the French navy under Napoleon, who was a friend and classmate
of the count. He was captured at Trafalger and thereafter did not
return to France. The story of his settlement at Tampa Bay is
entangled with encounters with pirates. Phillippi is supposed to
have introduced citrus to the area, and is credited with originating
the Parson Brown orange. It would be a major bit of historical
research to check on all of these points and produce an authentic
story of the man. This is, however, unnecessary for present purposes.
Phillippis first house stood near the shore to the south of the
large mound. The fierce hurricane of 1848 destroyed his house,
killed his grove, and cut back into the mound.4 He resettled on higher
1 Straub, 1929, p. 33.
2 See for example McKay (1946) and Sands (1948) in the Tampa Tribune.
3 Microfilm in the P. K. Yonge Library of Florida History, Gainesville.
4 An eyewitness account of this hurricane is in Bethel (1914, pp. 63-64).
7


land near our Test C (Fig. 1), and this house burned many years
later.5 Count Phillippi died in 1869 ; his grave is on the site (Fig.l).
S. T. Walker published a very brief note on the site in 1880,
describing it as one of the largest mounds on Tampa Bay and men
tioning the physical stratigraphy of the large mound as revealed by
erosion of the bay side of the mound. He conducted no excavations
at the site.8 Clarence B. Moore mentions the site only by saying
that permission to dig was refused.7
In 1929 Matthew W. Stirling visited the site and conducted some
test excavations,8 and returned to the site the following year for
more intensive work. Stirling dug the greater portion of the sand
burial mound uncovering over 100 burials with some of which historic
trade goods were found. He also did some work in the village portion
of the site. By 1936 Stirling was speaking of a Safety Harbor ware,
thus making Safety Harbor a type-site.10
In 1942 Willey and Woodbury referred to the Safety Harbor cul
ture, thus broadening the concept introduced by Stirling.11 At pres
ent a Safety Harbor Period is recognized by Florida archaeologists.12
Willey has given both a preliminary 13 and a detailed 14 account of
Stirlings work at the Safety Harbor site.
Early in 1948 Pinellas County acquired the land surrounding the
Safety Harbor site for purposes of a public park, named Phillippi
Park in honor of Count Odet Phillippi. The senior author visited
the site several times following this in order to delimit the archae
ological area for purposes of park master planning by the Florida
Park Service. In May he dug several test holes in the village.
From August 1 until August 14, 1948, excavations were conducted
by the writers. The enterprise was a joint one; the Florida Park
Service supplying the technical personnel and Pinellas County sup
plying the labor force. The many courtesies tendered by the Board
of County Commissioners of Pinellas County and the Pinellas County
Park Board, and in particular its chairman Lucius S. Ruder, made
our work most pleasant despite the late summer temperatures.
During this period of excavation the top of the large mound was
tested rather extensively, and was found to consist of super-imposed
strata with historic materials near the very top. A test in another
presumptive mound was not too productive. The third part of the
work consisted of trenching a portion of the village site. The details
of this work will be given later in the report; the location of the
excavations may be seen on Figure 1.
Subsequently collections at the Florida State Museum and those
made years ago by Messrs. Harrison, Armistead, and Wolf of Tampa
5 Information from Mr. Monroe Booth, Clearwater, descendant of Count
Phillippi.
Walker, 1880, pp. 410-411.
T Moore, 1900, p. 356.
8 Stirling, 1930, p. 186.
Stirling, 1931.
10 Stirling, 1936, p. 353.
11 Willey and Woodbury, 1942, p'p. 235 and 245.
,2 Goggin, 1947, p. 118; Willey, 1948, p. 213.
18 Willey, 1948, pp. 213-214.
14 Willey, 1949c; we wish to thank Dr. Willey for use of his manuscript prior
to its publication.
8


were examined. Dr. Gordon R. Willey supplied additional informa
tion about collections at the United States National Museum.
In August, 1949, we returned to the site once more, at the request
of the Pinellas County Park Board. Messrs. Frank E. Munley and
L. M. Bond of St. Petersburg had asked permission of the County
Commission to dig into the large mound where they believed their
electronic instruments had located a cache of metal at a considerable
depth.
Accompanied by C. W. Goyder of the Electronics Research Labor
atory of the University of Florida, whose expert opinion was greatly
appreciated, we witnessed a demonstration of the instruments on
the mound, and made additional tests with these instruments. The
conclusion was that the slope at the edge of the mound disturbed
the magnetic field of the instrument, giving an indication incorrectly
interpreted. There was not a definite indication of metal, and we
advised against permitting excavation.13 The county park board
concurred in this decision, and recommended that no digging be per
mitted. The County Commission unanimously accepted this recom
mendation.
This is not the first instance of supposed treasure in Indian
mounds, nor is it likely to be the last. In this case, the fact that
the site was under public ownership for park purposes led to a care
ful examination of the evidence before permission to excavate was
granted, and resulted in protection of the mound.
15 A full report of this investigation is on file with the Florida Park Service
and with the Pinellas County Park Board.
9


TYPOLOGY
This section of the report presents the typology of pottery and
other artifacts from the Safety Harbor site. Only the source of the
type description is given unless there is some reason to comment on
variation from the description.
Pottery
Pinellas Plain. Defined by Willey.10 Most of our specimens tend
toward the crumbly, contorted end of the paste range as defined.
Notched lips on Pinellas Plain at this site vary from mere ticks to
small triangular notches at the outer edge of the lip, (PI. I, Y).
Lips are usually flat and frequently chamfered inward. Some of
the few unique, deeply indented, rims are illustrated (PI. I, Z-BB).
Pinellas Incised. Defined by Willey.17 The paste of Pinellas
Incised at this site varies somewhat from the paste of Pinellas Plain.
It is more frequently sand-tempered, with a harder and better surface,
than is the case with Pinellas Plain. The three sub-types used in
this report are the same as those employed in the analysis of the
Lake Jackson site: A, loops, or more frequently arcades, bordered
by punctates (PI. I., G) ; B, parallel horizontal lines around the rim
(PI. I, I-K) ; and C, loops pendant from horizontal incised lines
around the rim.18
Lake Jackson Plain. Defined by Willey.19 As at the Lake Jackson
site, the paste is toward the coarse and contorted end of the range.
The four sub-types found at the Lake Jackson site 20 are all present
at Safety Harbor. Lake Jackson Plain handles are illustrated (PI.
I, C-F).
Fort Walton Incised. Defined by Willey.21 On our specimens
the design is executed solely by incision (PL I, T-U).
St. Johns Plain and St. Johns Check Stamped. Named by J. B.
Griffin.22 This is the chalky ware so characteristic of northeast
Florida. In this paper the types have been classified as St. Johns
rather than Biscayne, which seems to be a synonym. The checks
run between 4 and 8 per inch, and in almost all cases there is linearity
in the stamping; the checks being more or less rectangular with the
lands along the long side being more prominent than the cross lands
(PI. I, EE-FF).
Sarasota Incised. Defined by Willey.23 One sherd was collected
on the surface. Two other sherds were found which in motif appear
to resemble this type as defined, but they are on a sand-tempered
rather than a chalky paste (PI. I, V-W).
Willey, 1949c.
17 Ibid.
18 Griffin, N.D.6.
10 Willey, 1949c.
20 Griffin, N.D.6.
21 Willey, 1949c.
21 Willey, 1949c.
22 Griffin, 1945.
23 Willey, 1949c.
10


Other named types. The following are pottery types from the
site which seem to need no qualification: Safety Harbor Incised
(PL I, M-R) and Wakulla Check Stamped as defined by Willey 24
and Belle Glade Plain, Glades Plain and Pasco Plain as defined by
Goggin.25 Three shell-tempered sherds may probably be classified
as Pensacola Plain.20
Leon-Jefferson types. Sherds of types associated with the Leon-
Jefferson Period of northwest Florida include Leon Check Stamped,
Jefferson Stamped (PL I, B) and types 1, 3, 4 and 5 Jefferson rims 27
(PL I, A, illustrates Type 3). Annular ring bases were also found
(PL I, GG).
Spanish pottery. The majority of the Spanish sherds are from
plain olive jars. Several sherds have a green interior glaze. One
piece of earthenware has a white glazed exterior and several sherds
are yellow glazed on both sides.
Random-punctated. This is a category set up to cover sherds
whose exterior surface is covered with irregularly shaped, irregularly
placed, shallow punctations (PL I, CC-DD). Such punctations occur
only on crumbly Safety Harbor paste.
Cord-marked. In this category are included sherds which have
a surface malleated with a fine cord apparently wrapped on a paddle,
and cord-impressed sherds which have parallel lines neatly crossed
at nearly right angles to give a diamond-shaped pattern.
Unique pottery. From the mound came a sherd with three paral
lel lines of incised loops, apparently repeating, rather than pendant
from horizontal lines as in Pinellas Incised C. The remainder of
the specimens to be noted below came from the collections of Harri
son, Armistead, and Wolf. One sherd combines the sub-types Pinel
las Incised A and B, and has a strap handle with vertical incision;
the inside of the sherd has a protruding shelf about one-half of an
inch below the lip (PL I, H). Two horizontally flattened lip lugs
occur; one of them is decorated like Safety Harbor Incised (Pl. I, S).
There is also a probable tail adorno of shell-tempered pottery. One
sherd of Pinellas paste is decorated with slanting parallel incised
lines which oppose each other, giving a herringbone-like pattern.
Two sand-tempered sherds, one with punctation and bottle neck
curvature, the other with incision and punctation, do not appear to
qualify under existing type names. Two rim sherds are decorated
with appliqued discs a short distance below the lip, on one of these
the centers of the discs had been punctated (PL I, X). Another
sherd has two conical nodes; these are not podal supports. Three
sherds with a chalky paste have vertical incised lines defining zones
filled with elongated punctations. One sherd of a St. Johns paste
has parallel, interrupted, incised lines. There is also one sherd simi
lar to Safety Habor Incised but on a chalky paste. One limestone-
tempered sherd is decorated with criss-cross incision, apparently ar
ranged near the rim. Two limestone-tempered rim sherds bear verti
cal concave flutings, probably from scraping.
-4 Ibid.
"'Goggin, 1944; Goggin, 1948.
20 Willey, 1949c.
"7 Smith, 1948.
11


Stone Tools
Projectile Points: Projectile points have been classified as triangu
lar, lanceolate, corner-notched, side-notched, double basal-notched,
and stemmed. In the tables these categories have been further di
vided into small, medium-sized, and large. Points were predomi
nantly made of chert, a few of chalcedony, and only a half dozen of
fossil coral.
Small triangular points approach an equilateral triangle in shape
and are about l/2 to % of an inch in length. Medium-sized triangular
points are similar to an isosceles triangle in shape but have ex-
curvate sides (PI. II, A-B). They are about % to 1 inch in length.
Long, narrow triangular points vary from 1 to 2 inches in length
and usually have straight sides and straight bases (PI. I, E-F). They
tend to be relatively thick near the base. Triangular points were
frequently made from thin flakes with one or both of the original
flake surfaces unreduced. Occasionally, the base appears to have been
broken off in manufacture and the edges of the break slightly re
duced by flaking. Six long, narrow triangular points are serrated.
Eight have small lateral extensions of the base similar to ears.
One is both eared and serrated.
Lanceolate points are about 2 inches in length. One with a con
vex base is illustrated (PI. II, G). Those with a concave base are
much wider, about % of an inch in this respect.
Corner-notched points vary considerably in shape and form. The
one illustrated has excurvate sides and an expanded convex base
(PL II, H). Others are similar but have straight sides or have
slightly excurvate sides and deeply indented corner notches. One
has serrated edges and a concave base.
Side-notched points also have a wide range of variation. Small,
expanded base, side-notched varieties have excurvate sides, small
shallow side-notches, and a base as wide as the blade. They are
about 1(4 inches in length. Large, expanded base, side-notched
points are 2 to 3 inches in length, have wide notches, and bases as
wide as blades (PL II, I-J). Bases and notches have been smoothed,
presumedly by rubbing. These points appear to be more patinated
than others in the collections. Included under miscellaneous side-
notched are three long narrow points with serrated edges and one
short wide point with beveled edges.
Double basal notched points have a basal outline similar to the
Eva Double Basal Notched, points of Tennessee.28 They vary from
1(4 to 2% inches in length. One third have pronounced stems which,
however, do not protrude below the rounded barbs. Stems of the
balance are smaller and about the same size and shape as the barbs.
A few stemmed points are small. Most are medium-sized with
straight or contracting stems (Pl. II, K-L) or large with similar
stems (PL II, N-Q). Contracting stems may have straight sides or
the outline of the stem may be curved.
One unique point, 2% inches long, has very deep and wide ser
rations.
28 Lewis and Kneberg, 1947, p. 18.
12


Drills: Most drills are similar to long, narrow, triangular points
in outline (PI. II, R). However, they are thick, instead of thin,
throughout their length and have blunted ends. A few are crude
(PI. II, S) and one has an expanded, side-notched base.
Scrapers: Various forms of scrapers were found. These included
end, long side, elongated turtleback, trianguloid, and circular-like
scrapers. One large thumb-nail scraper has a stem. While fairly
large side scrapers predominate, no particular form seems to be
typical. A few are illustrated (PI. II, T-AA).
Knives: Knives are ovate or more or less triangular in outline,
fairly broad, with excurvate sides and convex or straight bases.
Some have asymmetric blades. One is notched on one side. A nar
row blade, 4 inches long, is also included in this category.
Small amorphous tools: This term has been used to designate
narrow flakes, V2 to 1 inch in length, with retouched sides. They
may be small drills, scrapers, or engravers. Possibly they were
hafted.
Worked fragments: Unfinished or fragmentary specimens in
sufficiently complete for classification.
Utilized flakes: Chips with retouching along one or more edges,
(PI. II, BB-CC). Retouching may have been done to form a working
edge or may be the result of use.
Hammer stones: Hammerstones have been divided into three
groups. Small tools with battered ends, 2 to 21/2 inches long and
1 to li/2 inches thick, have been called chipping hammers based on
the assumption they were used for light percussion flaking (PI. Ill, I).
They were made of sandstone or quarzite. Ball hammerstones are
those whose shape, either from manufacture or use, resembles a
sphere. At Safety Harbor they were made of chert or quartz. Other
hammerstones, of chert or limestone, are medium-sized or larger
battered rocks (PI. Ill, J). Some appear to be old cores later used
as hammerstones.
Grindstone fragments: Pieces of sandstone slabs with flat or
concave smooth surfaces (PI. IV, O-P).
Large sharpening stone: A sandstone boulder, 51/2 by 6 by 9i/£
inches, has one concave surface which had been used for polishing
or grinding. In this concave surface is a longitudinal groove about
3/16 of an inch deep, % of an inch wide, and 5% inches long.
Other stone objects: A truncated limestone cone, a limestone
bead, and a discoidal stone of pink quarzite are included here. The
cone is lVs inches high with % of an inch top and 11/4 inches bottom
diameters. The top is flat and the bottom smooth and slightly convex.
The bead is shaped like a fat doughnut, % of an inch thick, with
11/4, inches outside and about % of an inch inside diameters. The
discoidal is 2i/2 inches in diameter and one side is slightly concave.
Plummets
Plummets have been classified as knobbed, grooved, pebbled, un
classified, top-grooved columella, and centrally-grooved columella va
rieties. Sides of the knobbed variety slope inward to form a distinct
13


knob (PI. Ill, A). Sides and top of grooved plummets form a continu
ous surface or curve which, however, is broken by the groove (PI. Ill,
B). In pebble-type plummets the body is unworked or very little
altered (PI. Ill, C). Unclassified plummets are fragmentary speci
mens of one of these types. Columella plummets, as the name im
plies, are made from the central spiral of shells. The top-grooved
variety are long, narrow, and grooved near the top. In only a few
cases is a knob suggested. The sub-type name adequately describes
centrally-grooved columella plummets.
Plummets of the first four categories were made of shell, lime-
- stone, sandstone, fossil bone, slate, quartz, and an unidentified ig
neous rock. The one of slate is rectangular in cross-section.
Shell Tools
Typology of shell tools follows that used by Goggin and Sommer
for a site on Upper Matecumbe Key.29 A Busycon pick, type A, a
perforated Venus shell, and two columella chisels are illustrated
(PI. Ill, F, G, D-E).
A unique shell tool consists of a lenticular specimen, 31/2 by 11/2
inches, with rubbed edges which make only a slight angle with the
concave side. It is reminescent of pottery smoothing tools used in
the Southwest.
Bone Tools
Typology of bone pins follows that used by Willey for artifacts
from the site at Belle Glade.30 Included in the collection from Safety
Harbor are two peg-topped, one hollow shaft, two expanded head,
one simple prepared head, one unique, and three fragmentary bone
pins (PI. IV, A-H, respectively). The unique pin has a prepared
semi-annular head formed by leaving part of the wall of the bone.
This detail cannot be seen in the illustration (PI. IV, G).
Sockets for hafting tools are represented by three proximal ends
of deer cannon bones (PI. IV, I). They do not show evidence of use
and may therefore, be refuse from the manufacture of bone pins.
A gorge (PI. IV, J), two splinter awls (PI. IV, K-L), a double
pointed awl, a possible bone tube (PI. IV, M), and several worked
fragments were also found. The bone tube is about 4 inches long
and considerably eroded. One of the worked fragments has been
scraped with a sharks tooth.
31 Goggin and Sommer, 1949.
30 Willey, 1949a.
14


THE VILLAGE AREA
The village area of the site is represented by data from three
excavations conducted by the writers and by specimens from other
collections. We will first examine the results of the excavations,
and then compare the materials from the other collections.
Excavations
Tests were made at three places identified by the symbols A, B,
and C on the map (Fig. 1). A grid of five foot squares was used for
horizontal control, and material was excavated and bagged by arbi
trary six inch levels. Local datums were tied in with surveyors
markers of known elevations.
Area A. At the large mound trenches, whose locations are shown
in Figure 1, were carried down to a depth of about five feet below
the surface. From the bottom of these trenches test holes were
made with post hole diggers. Thus we have a good idea of the com
position of the upper eight feet of this mound. As it is about twenty
feet in height, our excavations did not quite penetrate the upper
half.
The trenches revealed an extremely complicated structure. A
slightly simplified profile, taken about five feet west of the major
axis of the mound, is given in Figure 2.
Toward the south, between stakes -5R3 and -5L2, what appeared
to be a dome shaped deposit of olive-green clay was located at the
lower part of our excavation. The base of this clay was not reached
in the -5 trench, but to the east, in the R1 trench, it was found to
be 23 to 29.5 inches thick, resting on a gray-white sand.
Before proceeding further it may be well to mention the results
of a hole dug in this mound in 1933, information on which we secured
while in the field. Part of this hole is shown at the southern end of
the profile (Fig. 2). It was about six by eight feet in size, and was
carried down by. steps to a depth of about 18 feet. At this level
bluish clay, charcoal, and some bones were found.81 At about a
depth of 12 to 14 feet, whole bones with clay packed around them
were noted. Whether or not these were human bones is not certain.
At lesser depths zones of sand and shell were encountered, similar
to those revealed in our excavation. Large pieces of pottery were
found, but whether they were associated with the bones in the clay
is not known.
This evidence, while sketchy, suggests the possibility that a dome
shaped burial mound made of clay and sand lies under the southern
part of the large mound. If this is correct, the large mound was
constructed by adding material to a smaller burial mound.
The portions examined by us showed that shell, sand, village
debris, and some clay was used in the construction of the mound.
The most common shells were those of the oyster (Ostrea virginica),
crown conch (Melongena corona), and the sea snail (Polinices dupli-
cata). Less common were the clam (Venus sp.) and the channeled
whelk {Busycon canaliculatus). The two deposits of oyster shells be-
31 Walker (1880, p. 411) also mentions that many bones were said to have
been washed out from the mound by a storm.
15


PROFILE OF VILLA GE
Fig. 2. Profiles from Large Mound and Village


tween stakes -5R2 and -5R1 and below -5L5 were composed of clean
shells and were practically sterile. The clay is olive-green or mixed
green and yellow in color, and was probably dug at or near the shore
of Old Tampa Bay near the site, where such clay may be found today.
It is believed that the work of construction proceeded by stages,
and that the top of the mound was occupied during pauses in build
ing. Unfortunately, evidence for such a series of stages is mostly
inferential. The more or less horizontal layers of clay may repre
sent floors; in general the tops of these floors were fairly smooth,
but ridges were present. Into the clay had been pressed sherds and
food bones. Edges of these clay zones, where we were able to examine
them, were extremely irregular. No post holes were found asso
ciated with the clay layers. In places, particularly towards the north
in square -5L3, this clay did not form a floor but consisted of a series
of lumps of clay forming a zone one to four inches thick.
The chief argument in favor of these clay zones being floors lies
in the presence of relatively large amounts of occupational debris
on and just above them. Another argument in favor of the clay
levels as floors is by analogy to the clay cappings or mantles com
monly found on pyramidal mounds throughout the Southeast.
Several post holes were found in the upper part of the mound.
These were few in number and formed no recognizable floor plan
or pattern. The apparent tops of some of these were encountered
at a depth of five to seven inches. Two small pits also led down from
this depth. Some of these features cut through zones of clayey dirt
or shell. They may indicate habitation or use at an elevation ap
proximating that of the present surface of the mound.
Another group of post holes were met with at a depth of 14 to
18 inches. Two of these are indicated on the profile (Fig. 2), one
south of stake -5L3 and the other north of stake -5L5. These suggest
an occupational zone about half-way between the present surface
and the top of the deposit of mixed shell and debris.
Considering these data, our reconstruction of the large mound
would be about as follows: A clay and sand primary mound, capped
by a heavy clay mantle and possibly used for burial, was enlarged
by the addition of sand and other materials, including many shells.
The first enlargement increased the area of the mound by adding
material northward, but did not increase it much in height. A clay
floor near the base of our excavation in squares -5 and -5L1 was
built and lived upon. Shortly thereafter another clay floor was
added, larger than and partly overlying the first floor.
After a period of occupation, the mound was enlarged and made
higher. Probably at this time the large lens of oyster shell below
stakes -5L4 and -5L-5 was deposited, on what appears to have been
a shoulder of the mound. Much clean sand was added to the top,
and possibly the clay floor between stakes -5L7 and -5L8 was built
at this time. The rather extensive and thick layer of mixed shell
and debris, extending across the mound, may mark the occupation
level atop this building stage. On the other hand it may represent
debris scraped from the village and deposited atop the mound. Post
holes leading downward from a depth of 18 inches suggest an oc-
17


TABLE 1. Stratigraphic Tabulation of Pottery from the Large Mound
Item
Depths in Inches
Totals
0-6
6-12
12-18
18-24
24-30
30-36
36-42
42-48
48-54
54-60
spoil
Spanish olive jar sherds
2
5
2
9
Jefferson ware, rim type 3
1
1
2
Jefferson ware, complicated stamped
1
1
2
Jefferson ware, annular ring base....
1
1
Lake Jackson Plain, handles
1
1
1
3
Lake Jackson Plain, knobbed
1
1
Belle Glade Plain
1
1
2
4
2
3
2
1
16
St. Johns Check Stamped
1
1
1
1
2
1
7
St. Johns Plain
3
3
1
3
2
1
13
Pinellas Incised B and C
1
1
2
1
2
7
Pinellas Incised A
1
3
1
1
6
Pinellas Plain, notched rims
4
7
7
1
1
1
1
22
Pinellas Plain, other rims
34
45
50
42
36
40
54
34
7
9
14
365
Pinellas Plain, body sherds
210
284
283
244
224
177
318
188
28
39
73
2068
Safety Harbor Incised
1
1
1
1
2
1
7
Random-punctated
1
1
1
5
4
1
1
1
15
Smooth plain
1
1
1
2
1
6
Unique incised
1
1
2
Miscellaneous incised
3
2
2
1
1
1
10
Pine criss-cross incised
1
1
Cord-marked
1
1
Glades Plain
1
1
Semi-gritty residual plain
6
22
22
17
15
16
15
8
1
4
7
133
Totals
264
371
371
318
289
244
407
241
38
54
101
2698


cupation when the surface of the mound was near the top of this
zone of mixed shell and debris.
Subsequently, more material was added and another occupation
occurred at a level equivalent to present depths of five to seven inches.
This occupation zone is suggested by an increase in the number of
sherds and by pits and post holes leading downward from that depth.
Final accretion, possibly natural, brought the mound to its present
height.
The top of the mound had been slightly disturbed by picnicers
and others, which accounts for charcoal briquettes, recent glass, and
fragments of iron found in small quantities in the upper one foot.
An animal burrow was found in square -5L2. Pieces of modern brick
had fallen or been tossed down this hole. Some decaying cedar
posts suggested the presence of a fence or small building in post-
Indian times. Otherwise the mound was undisturbed except for
the 1933 excavation mentioned earlier.
The vertical distribution of pottery in the large mound is given
in Table 1. In terms of numbers of sherds, two frequency peaks
are indicated. The one at six to eighteen inches correlates with pits
and post holes whose tops were noted at the same depths, lending
support to the upper occupation zone or zones postulated above.
The second peak, between 36 and 42 inches lies above the clay floors
and below the white sand, corresponding to the occupation postulated
at this depth. The stratigraphic implications of pottery types will
be discussed at a later point.
Only three projectile points were found in the mound. Two
medium-sized triangular points (PI. II, B-C) were found between
the surface and a depth of one foot. One medium-sized stemmed
point (PI. II, L) was found in the 24-30 inch zone.
Specimens of chipped stone from the mound include, in addition
to the projectile points mentioned above, three worked fragments,
three utilized chips and four scrapers.
A plummet type pendant of fossil bone (PI. Ill, C) was found
in the mound; there were also three pieces of fossil bone and a fossil
sharks tooth. Two sandstone grindstone fragments and the large
sandstone boulder with concave surface came from this portion of
the site, as did a quartzite pebble with a chip removed, a piece of
coral rock, and some burned sandstone fragments.
A perforated Venus shell (PL III, G), a fragment of a Busycon
pick, and a columella chisel (PI. Ill, E), fashioned of Busycon per
versa rather than the usual Fasciolaria gigantea, comprised the total
of shell artifacts from the mound.
Ten bone pins from the mound included the following types:
two peg-topped, one hollow shaft, two expanded head, one simple
prepared head, one unique and three fragmentary specimens (PL
IV, A-H, respectively). Three probable bone handles or sockets
(PL IV, I) were found, and there were a bone gorge (PL IV, J), two
splinter awls (PL IV, K-L), five fragments of worked bone, and a
possible bone tube (PL IV, M).
Area B. A small hillock, reaching a height of about four feet
above the surrounding land, is located near the parking lot. This
19


was investigated by means of an interrupted exploratory trench.
The hillock proved to be a natural feature, probably an old sand
dune, built up in pre-Indian times. The profile consisted of a su
perior zone of humic stained sand, 5 inches thick, and the following
zones beneath it: a 7 inch occupation zone, a 9 inch zone of gray-
brown sand containing only a very small amount of Indian material,
and finally sterile yellow or yellow-brown sand.
Indian material was concentrated between depths of 6 and 14
inches. In this same zone were two fragments of glass, one of brick,
and one of a slate pencil, suggesting some disturbance. The basal
fragment of a large, stemmed, stone knife was found in the top six
inches, and a Busycon pick, type A (PI. Ill, F), was found between
12 and 18 inches. A tabulation of sherds from this area will be
found in Table 2.
TABLE 2. Stratigraphic Tabulation op Pottery from Area B
Item
0-6
Dep
6-12
ths in I
12-18
nches
18-24
spoil
Totals
Spanish olive jar sherd
1
1
St. Johns Plain
1
1
Miscellaneous incised
1
1
Plain, red-painted interior
2
2
Plain smooth
1
1
Fragment of handle
1
1
Pinellas Plain, notched rims....
4
1
1
6
Pinellas Plain, other rims
3
4
9
1
17
Pinellas Plain, body sherds ....
15
45
39
3
102
Residual Plain, gritty paste ....
1
3
1
1
1
7
Totals
19
61
51
6
2
139
Area C. Area C is included in the main village area of the site.
It is believed that the contour shown for this area of the map
(Fig. 1) is the result of the accumulation of debris during Indian
occupation. A trench was excavated across this ridge with extensions
as shown on the map. A profile of the main trench is presented in
Figure 2.
This section exhibits a normal soil profile superimposed upon an
Indian site. It is similar to that found at Area B, but more com
plicated due to more intensive occupation. Half of the Indian ma
terial was found between the depths of 6 and 12 inches, and half
of the remainder in the next lower 6 inches. Shells, while present,
were not abundant west of Stake 13.
Examination of the profile will indicate a large pit between stakes
1 and 2. No function other than that of a storage pit could be sug
gested for this feature. Between stakes 3 and 11 portions of many
pits, and post holes will be noted in the profile. This area was in
tensively occupied and contained a great deal of aboriginal disturb-
20


TABLE 3. Stratigraphic Tabulation of Pottery from Area C
Depths in Inches
Pits
Spoil
Totals
Item
0-6
6-12
12-18
18-24
over
24
18-24
over
24
Spanish olive jar sherds
4
2
1
7
Jefferson rim, type 1
1
1
Jefferson Complicated Stamped, type 5 rim
1
1
Lake Jackson Plain
1
1
Safety Harbor Incised
1
2
2
1
6
Pinellas Incised, A
2
2
Pinellas Incised, B
1
1
4
1
i
1
9
Pinellas Incised, C
i
1
Unique incised
1
1
Miscellaneous incised
2
1
1
4
Punctated, tail adorr.o
1
1
Belle Glade Plain
1
3
3
3
1
11
Glades Plain
1
1
St. Johns Check Stamped
1
2
3
2
i
9
St. Johns Plain
3
2
7
3
2
3
20
Random-punctated
1
4
1
6
Plain smooth
2
1
3
Pinellas Plain, notched rims
7
40
21
3
3
3
2
79
Pinellas Plain, other rims
49
209
169
42
10
9
21
17
526
Pinellas Plain, body sherds
319
2074
989
253
114
58
90
60
3957
Residual Plain, gritty paste
8
19
15
9
2
3
1
57
Scored surface, gritty paste
1
1
1
3
Pottery disks, Pinellas paste
2
2
Totals
397
2359
1225
318
127
73
124
85
4708


ance. Our trench was widened between stakes 3 and 5 in an attempt
to secure the floor plan of a house, but the multiplicity of post holes
and small pits, and the impossibility of determining planes of origin,
made the delineation of post hole patterns impossible.
Between stakes 13 and 14 a thick layer of mixed shell and debris
begins. Three squares to the east of stake 14 were excavated to
investigate the extension of this zone. Profiles of these squares con
sisted of a deposit of shells, about two feet thick, with black dirt in
the upper half and gray dirt in the lower half. Below was a zone
of light gray sand containing only a few shells and resting on white
sand. This rather extensive shell deposit contained very few sherds
or other cultural material.
The horizontal distribution of sherds is rather interesting. The
quantity found between stakes 12 and 14 is disproportionately high,
while that for the squares to the east is very low. Presumably, the
main living area was where the pits and post holes were found.
Seemingly, the Indians living at Safety Harbor preferred to keep
their living area clear of shells, which they accomplished either by
tossing the shells toward the water, thus forming the shell deposit,
or by opening the shells in the area of the deposit and taking the
animal rather than the shells to their homes. It is also possible that
the great amount of black dirt containing charcoal found with these
shells to the east of stake 14 may indicate that shellfish were roasted
there. Any of these alternatives would explain the relative lack of
cultural debris among the shells.
The very large number of sherds found at the western edge of
this shell deposit may indicate that this was a place where pots were
frequently broken. Such might be the case if vessels were carried
to the shell heap to be filled with shucked shellfish, either cooked or
uncooked.
The village area has been under cultivation at various times dur
ing the past hundred and twenty-five years. Fragments of glass,
iron, nails (including wire nails), bricks, earthenware, and stone
ware attributable to this period were found in the upper twelve inches.
Table 3 summarizes the pottery from this portion of the site,
treating it by levels. An interesting feature of the pottery tabulation
is the presence of two pottery discs, marking the only occurrence
at this site of a trait rather common in both Ft. Walton and Leon-
Jefferson horizons farther north.
Table 4 gives the vertical distribution of projectile point types
in the village area. It should be noted from this tabulation that
stemmed points had a deeper overage promenience than did triangu
lar points. A similar suggestion of change in types of projectile
points has been mentioned earlier for the large mound.
Chipped stone, other than projectile points, included twelve worked
fragments, eleven utilized flakes, and five scrapers. There was also
a triangular-shaped drill (PI. II, R). Raw material was represented
by five broken nodules of chert. Six hammerstones (PI. Ill, J) and
three chipping hammers (PI. Ill, I) represent pounding tools. There
was a piece of fossil bone and two large water-worn pebbles. Eleven
fragments of sandstone grindstones (PI. IV, P) came from the vil
lage. Three plummet type pendants (PI. Ill, A-B), one of sandstone
22


and two of shell, were found. A fragment of a Busycon pick and a
pottery hone (PL IV, N) were recovered. One fragment of worked
bone, scraped with a sharks tooth, completes the artifact inventory.
We should also note the presence of a fragment of burned clay
daub with a wattle impression (PI. Ill, H), suggesting wattle and
daub construction at the site.
TABLE 4. Stratigraphic Tabulation of Projectile Points from Area C
Item
Depths
in Inches
Totals
0-6
6-12
At 12
12-18
18-24
24-30
Medium sized triangular ..
i
1
Long, narrow triangular ..
i
1
1
3
Medium sized lanceolate ..
1
1
Corner notched
1
1
Medium sized stemmed ....
1
1
2
Large stemmed
1
1
2
4
Totals
2
1
3
2
2
2
12
Other Collections
Three other collections of material from the village area of this
site are available for comparison. A small collection in the U. S.
National Museum is the result of Stirlings work in the area, and
has been described by Willey.32 There is a certain amount of ma
terial in the Florida State Museum, collected from the beach by T.
Van Hyning in 1914 as well as some collected by H. H. Francis in
the same year. The third group of material is found in the private
collections of John N. Harrison, Jr., W. J. Armistead, and Harold
Wolf, all of Tampa, who surface collected the area from 1917 onward.
The bulk of the collection of these three men came from the shore
a short distance northwest of the large mound.
Materials from these collections have been tabulated along with
our own excavated specimens in Tables 5 and 6. Table 5 gives the
pottery data. A final category under the pottery types reads other
sherds. In the case of the Park Service collections the contents
of this category have been presented in the sections on our excava
tions, but in the case of the other collections the contents of this
category should be noted.
The National Museum collection, as analyzed by Willey, includes
three sand-tempered sherds with broad incised lines and one red
on buff sherd with fine shell temper. The Florida State Museum
collection includes the following sherds: seven miscellaneous stamped,
two miscellaneous incised, two miscellaneous punctated, one probably
podal, support, and one sherd with double rim notching. The six-
32 Willey, 1949c.
23


teen sherds from the Harrison, Armistead and Wolf collections, here
included under other sherds, are described under the section on
typology.
In the table the following abbreviations are used: FPSFlorida
Park Service, USNMUnited States National Museum, FSM
Florida State Museum. OthersHarrison, Armistead, and Wolf col
lections.
TABLE 5. Potsherds from the Village Area
Pottery Type
FPS
1
USNM
FSM
Other
Totals
Spanish olive jar sherds
17
19
2
38
Jefferson rims
4
2
6
Jefferson Stamped
2
2
Leon Check Stamped
3
3
Ring base
1
1
2
Pinellas Plain, body
6127
182
8
6317
Pinellas Plain, rims
908
85
993
Pinellas Plain, notched rims
107
20
2
129
Pinellas Incised A
8
I
2
10
Pinellas Incised B
13
1
23
6
43
Pinellas Incised C
2
J
6
2
10
Safety Harbor Incised
12
5
13
10
40
Fort Walton Incised
2
4
1
7
Lake Jackson Plain
5
1
5
11
St. Johns Plain
34
1
1
3
39
St. Johns Check Stamped ....
16
3
6
5
30
Belle Glade Plain
27
12
39
Glades Plain
2
5
7
Wakulla Check Stamped
3
3
Sarasota Incised
2
2
Pasco Plain
3
1
4
Pensacola Plain
2
1
3
Random-punctated
21
1
1
1
24
Cord-marked
1
1
3
5
Residual plain
197
5
4
206
Other sherds
35
4
13
16
68
Totals
7543
23
398
77
8041
It will be noted that by and large the various collections give one
about the same ceramic picture for the site.
The picture presented by artifacts of stone, shell, and bone, listed
in the previous table, is basically homogeneous except for projectile
points. All four collections include an abundance of narrow triangu
lar and of medium-sized to large stemmed points. These types seem
to be typical of the site and it will be remembered that in our ex
cavations triangular points overlay stemmed points.
The large number of double basal notched points and of large,
expanded base, side-notched points in the Harrison-Armistead-Wolf
collection merit comment. That the former were known at the site
is indicated by their inclusion in the Florida State Museum collection.
It may be noted that both these collections came principally from
the beach to the north and northwest of the large mound where at
24


least twenty, and possibly, fifty feet, has eroded away (Fig. 1). They
may represent a cultural manifestation not reflected in our excava
tions. Possibly the same applies to the large, expanded-base, side-
notched points with rubbed bases. The question mark against the
four points of this type listed for the United States National Museum
collection indicates a question as to whether they duplicate those of
the Harrison-Armistead-Wolf collection.
TABLE 6. Nonceramic Artifacts from the Village Area
FPS
USNM
FSM
Others
Projectile points:
Small triangular
20
Medium-sized triangular
3
13
7
39
Long, narrow triangular
3
6
6
100
Lanceolate, convex base
1
Lanceolate, concave base
1
1
Miscellaneous corner-notched
1
4
3
Small, expanded-base side-notched....
1
4
1
Large, expanded-base side-notched....
4 ?
4
Miscellaneous side-notched
4
4
Double basal-notched
3
27
Small stemmed
8
3
Medium-sized stemmed
3
1
16
36
Large stemmed
4
(12
5
9
Unique, deeply serrated
1
Drills:
Narrow triangular
1
5
Other types
1
1
1
2
Scrapers:
Miscellaneous types
6
33
18
3
Knives:
Miscellaneous types
1
3
3
4
Narrow blade, 4 inches long
1
Small amorphous tools
1 ?
4
Worked fragments
15
60
10
Utilized flakes
14
30
Hammerstones:
Chipping hammers
o
O
1
Ball hammerstones
2
Other hammerstones
6
1
3
Grindstone fragments
13
1
Large sharpening stone
1
Truncated limestone cone
1
Limestone bead
1
Discoidal stone
1
Pottery hone
1
Plummets:
Knobbed
1
5
Grooved
1
1
2
1
Pebble
1
1
1
Unclassified
1
2
Top-grooved columella
2
3
23
Centrally-grooved columella
4
25


(Table 6 continued)
Shell tools:
Busycon picks, type A
1
4
Busycon hammers, type A
1
J 4
3
Melongena or Strombus hammers ..
X
Busycon dippers
2
1
Perforated Venus shells
1
Perforated Area shells
10
12
Celts
4
2
Gouges
2
Unique tool
1
Columella chisels
2
Bone tools:
Pins
10
Sockets ( ?)
3
Gorge
1
Splinter awls
2
Double pointed awl
1
Worked bone
7
1
Bone tube (?)
1
26


THE BURIAL AREA
The only excavations in the burial area of which we have record
are those of Stirling.33 Although Willey has summarized this ma
terial,34 and we can do no more than quote from him, the evidence
is given here for the sake of completeness of the report. The location
of the burial mound is shown in Figure 1.
The burial mound was composed of sand, about 80 feet in di
ameter and 10 to 12 feet high. The surface had been considerably
dug over before Stirlings work. To the east of the mound was a
deep circular depression representing the borrow pit from which
sand was secured to build the mound. Successive layers of sand
showed the mound to have been built in stages.
Over 100 burials were removed from the mound; most of these
were secondary in nature and were placed in no particular order.
The vast majority of the pottery was found in a deposit at the mound
base near the edge of the mound; many of the vessels had been
killed. The few European artifacts were with burials near the
top of the mound.
The pottery from the burial mound is given by Willey as follows:
Safety Harbor Incised 30
Pinellas Incised 28
Pinellas Plain 43
Biscayne Plain 2
Biscayne Check Stamped .-. 11
Glades Plain 15
Unclassified incised and/or punctated deep horizontal
incised lines, Biscayne paste 1
Total sherds 130
All of these pottery types are ones that are common in our village
site excavations. However, none of the late pottery types, associated
with the Leon-Jefferson period, are present in this collection.
Artifacts from the burial area include one stemmed projectile
point, a grooved limestone plummet, a shell hammer or pick, a shell
cup or dipper, and miscellaneous scraps of shell. The European
artifacts consist of two iron axes, a sheet silver ornament, and a
sheet silver tubular bead. One iron axe, about 6 inches in length, is
of the poll-less, tree-felling type with haft hole. It is very similar
to, but smaller than, one excavated by Hale G. Smith at the San
Francisco de Oconee mission near Tallahassee.33 The major point
of interest in the artifacts is that these European items tend to lie
high in the mound, just as the European materials lie high in the
large mound and in the village area.
Fifty-one skulls from the burial mound have been measured by
Ales Hrdlicka.30 Of these 27 are male and 24 are female. All are
above 18 years of age, but this does not necessarily mean that younger
individuals and infants were not present; they merely may not have
been treated in the anthropometric analysis.
33 Stirling, 1931.
34 Willey, 1949c.
35 Records, Florida Park Service.
30 Hrdlicka, 1940.
27


The distribution by age of these 5i skulls is as follows:
Sub-Adult
18-20
Young Adult
21-35
Middle-Aged
36-65
Old Adult
56-75
Total
Male
0
3
14
10
27
11.1%
51.7%
37.1%
Female ..
1
12
8
3
24
4.1%
50.0%
33.3%
12.5%
It is of course impossible from this data to erect a population
pyramid for the site, but there are certain suggestions which come
out of the data. It would appear that the life expectancy of the
male was greater than that of the female. The high percentage of
female deaths in the young adult age group suggests that childbirth
might be an important factor in this difference.37
An examination of Hrdlickas figures for the crania at this site
permits us to characterize the population to a certain extent. Taking
the absolute sizes of the length, breadth, and height of these skulls
we note that they are of medium length and breadth, but are high,
when compared to the standards for man in general.38 Both males
and females are similar in all indices; the only variation in means
being in the height-breadth index which shows the males to be high
headed, the females medium. The cranial index indicates a round
headed, or brachycranic, population; the males are less so than the
females, which is to be expected. Faces are of medium breadth and
height, as are the eye orbits. The nasal index discloses a medium
width nose. No evidence is available for the long bones.
37 Senyurek (1947, pp. 63-64) notes this same feature in Anatolian series and
gives other sources for the phenomena for the Danubian Bronze Age, Egyptians
of the Roman Period, and Neanderthal, Upper Palaeolithic, and Mesolithic peoples.
Casual perusal of the literature for North America shows it to be quite consistent.
38 Krogman, 1939, p. 7.
28


SUMMARY OF THE SAFETY HARBOR SITE
The preceding sections of this report have summarized the typ
ology of artifacts used in our analysis, the features of our investiga
tions, and the materials recovered by us and other investigators.
This section is an attempt to pull this material together, not as a
listing of materials, but as an interpretation of the way of life at
the site. The present attempt is felt to be conservative. One should
not seek here a full realization of the conjunctive approach advocated
by Taylor.39
The economic base of Safety Harbor culture is not certain. No
definite evidence of agriculture either in the form of charred plant
remains or of artifacts of certain agricultural association were
found. Neither the shell picks nor the fragments of grindstones
are necessarily indications of agriculture. However, the size of the
community, the lack of the tremendous shell deposits found elsewhere
in Florida, and the late time position with reflections of influence
from agricultural horizons farther north, together with historic ac
counts of agriculture in all but south Florida, combine to make a
postulation of agriculture for the site seem probable.
We are on safer ground in dealing with other pursuits concerned
with sustinence. The gathering of shell fish was an important ac
tivity. The common oyster, some of the specimens over six inches
in length, was the most frequent shellfish utilized. The crown conch
(Melongena corona) and the sea snail (Polinices duplicata) followed
in frequency. Less common were the clam (Venus sp.) and the left
handed whelk (Busycon perversa). The concentration of shell off
the actual living area of the village has been previously mentioned.
Large numbers of fish bones were found, including shark verte
brae. Methods of fishing are suggested by the probable bone gorge,
indicating line fishing, and the perforated shells believed to be net
sinkers. The possibility of spearing must not be ruled out although
no direct evidence was found. The turtle bones found appear to be
from land forms, and were probably simply gathered.
In at least the later stages of the site, where the triangular pro
jectile points are found to predominate, hunting was no doubt car
ried on with the bow and arrow. The larger points, many of them
too large for arrows, might have been used with atlatl darts or
spears, although some of them are almost certainly knives and
scrapers rather than weapons. The deer seems to have been the
animal most commonly taken, if we are to judge by the bones found,
but birds and other small game were also hunted.
As was the case with agriculture, we have no direct evidence for
the gathering of seeds, roots, nuts and the like, but such collecting
no doubt was important, as it is with most primitive peoples.
The community itself was dominated by the single large pyra
midal mound. Whether this was purely ceremonial or whether it
served as a base for the chiefs residence, and combined residential
with ceremonial use is not known. Suggestions of stages of build
ing are present in the mound, as are post holes suggesting structures
Taylor, 1948.
29


atop the mound. To either side of the large mound the village
stretched along the bay shore ridge. Post holes in our excavation
C definitely suggest a type of construction framed with fairly closely
spaced uprights, but we cannot state anything of floor plan. A single
piece of burned clay daub with wattle impressions suggests that the
common wattle and daub technique was used in building, at least
some of the time. Pits encountered in the village suggest storage
pits inasmuch as the concentration of refuse in them was rather
light and probably secondary.
The arts and crafts practiced by the inhabitants involved the
use of varied techniques on various raw materials. For chipping
the common materials were chert and silicified coral from local lime
stones. Both percussion and pressure flaking were used to work
this stone into the various types of projectile points, drills and
scrapers noted earlier. Some of the stone pendants show knowledge
of the techniques of pecking and grinding stone artifacts.
Bone and shell were cut, sawn, and ground into various artifacts.
One bone fragment appeared to have been scraped with a sharks
tooth.
The somewhat marly local clay served for the potters art in most
of the sherds found, although some showed a finer texture and prob
ably were made from other clay deposits. Pottery was made by
segmental coiling, decorated in a minority of cases with either in
cision, punctation, or stamping. The area of decoration, except for
certain stamped pieces, was confined to the upper part of the vessel
body. Design was rather crude in inspiration and execution. The
potters art was definitely more strictly utilitarian than in the earlier
Weeden Island Period of the area.
Wood working is suggested indirectly by several things at the
site. The arrow points indicate both the bow and the arrow, the
stamped pottery implies carved wooden paddles, and cruder wood
work is suggested by shell tools needing handles and post holes once
containing posts. Canoes are strongly suggested by the site location
and the use of marine life as food.
Cordage is rather directly implied by the sherds with cord marked
surfaces, and indirectly implied by the net weights.
We have, above, discussed the types of materials utilized and some
of the methods of manufacture. Now we may summarize the tools
possessed by these people in terms of the probable uses of the arti
facts found.
Hammering and pounding were carried on with hammerstones
and shell hammers. Scraping with chipped stone tools, shells and
sharks teeth. Cutting was done with chipped stone tools, and pierc
ing with chipped stone and bone artifacts. Digging was probably
done with shell picks and wooden sticks, the latter not having been
preserved. Grinding was done on standstone pieces, and some other
smoothing was probably done with sand and water abrasion.
Extra-village contacts and trade are suggested by a slate plummet
and another of an unidentified igneous rock. The sherds of the Leon-
Jefferson complex and the Spanish pottery and objects also are
evidences of trade.
30


Aside from a mention of historic notation of agriculture in this
part of Florida, we have not utilized sources for the ethnology of the
area. Without question we are here dealing with a horizon that
was at its earliest very close to historic times, and which extended
well into the historic period. It was almost certainly a Timucua
site. What may we glean from ethno-historical sources?
Unfortunately, there is little data for the particular region in
which the site is located, or for that matter for the central Gulf
coast. Swanton places the tribe or province of Tocobago, a branch
of the Timucua, in the Tampa Bay area, and the principal town at
the head of Old Tampa Bay, evidently the town site in Safety Har
bor.40 Tocobago was visited by Menendez in 1567, who left some
solders who were driven out the next year. In 1612 an expedition
went out against the Tocobago. They are seemingly not mentioned
further. There is a good possibility that the Safety Harbor site may
be the main village of the Tocobago, although no contact material
definitely attributable to the sixteenth century has been shown to
come from the site.
One direct reference to the Tocobago describes burial practices
as related to caciques. The body was cut up and cooked in large
pots for two days. The flesh was removed and the bones, arranged
in anatomical order, carried to the temple. Fasting followed for four
days. At the end of this time the bones were carried to the burial
spot and interred.41 Thus, we have an historical account of secondary
burial, a practice noted in the burial mound.
The Menendez visit, according to the Spanish sources, saw 1,500
warriors gathered near the village of Tocobago, but as Swanton
points out there is no way to judge how accurate this estimate was,
nor how many of the Indians were bonafide members of the tribe.42
The De Soto narratives contain some references to the general part
of Florida occupied by the Tocobago. Ucita is described as a village
of seven or eight timber houses covered with thatch, the chiefs
house was on a mound near the beach, and a temple was at the other
end of the town, with carved wooden fowl with gilded eyes atop it.43
Practically the only other information, also from the De Soto
expedition narratives, concerns bows and arrows. The bows were
of wood, and the wooden or cane arrows bore heads of stone or fish
bones.44 Fish bone or teeth projectile points are frequently men
tioned in early accounts and seldom found in the ground. The pres
ence of the bow and arrow, postulated from the projectile points
at the site, is substantiated by this reference.
The ethno-historical sources shed but little additional light on
the culture of the area in early historic times. We could go further
afield and include the remarks on Timucua culture as seen by the
French Huguenots in northeast Florida,45 but in so doing we would
40 Swanton, 1946, p. 196.
41 Swanton, 1922, p. 374.
42 Swanton, 1946, p. 196.
43 Swanton, 1922, p. 353.
44 Ibid, p. 357.
43 See Swanton (1922, pp. 345-387) for a summary of Timucua culture largely
from French Huguenot sources.
31


introduce the possibility of error, for there are suggestions that
Timucua culture at the time of contact differed somewhat over its
areal range. Even more remote geographically would be data from
other Southeastern tribes.46 Taking cognizance of both of these
sources of data, we are probably safe in assuming that the residents
of the Safety Harbor site utilized an extended kinship system,
probably related to a clan organization. Phratries were possibly
present, and at least a semi-aristocratic structure prevailed. Descent
was probably through the female line. Beyond this it is not safe
to speculate in any detail.
It might be suggested that certain conclusions could be drawn on
the basis of probability by reference to trends in primitive groups
of a comparable cultural complexity. The difficulty here is consider
able, however. Exceptions abound to almost every correlation which
has been advanced, and very little in the way of overall examination
of the correlation of phenomena with, say, economic base has been
attempted since the classic work of Hobhouse, Wheeler, and Gins
berg.47 According to their framework, Safety Harbor would prob
ably fall in Agriculture II.
Certain trends such as the growth of public at the expense of
private justice or the increase of slave-holding and nobility, are ob
vious in the study mentioned above, when the entire range from
Low Hunter to Agriculture III is considered. The stage with which
we are concerned, Agriculture II, lies at a point on the range where
no trait is overwhelmingly dominant in all cases, hence, no conclu
sions of value can be extended to our material.
It is a recognized axiom of archaeology that the record of life
at a site is extremely fragmentary. Even the material possessions
of the people are but partially represented.48 Fragmentary as the
artifactual record is, aspects of social organization and sacred ideology
which persist are even poorer. This we have seen in the attempted
reconstruction of the culture of Safety Harbor above. We must
leave this summary of the Safety Harbor site inadequately character
ized as a culture, but fairly well characterized as an archaeological
complex.
411 Swanton, 1946.
47 Hobhouse, Wheeler, and Ginsberg, N.D.
48 See, for example, Braidwood (1945).
32


CONCLUSIONS
Information secured by excavation should place us in a better
position to discuss the relationships of the Safety Harbor Period,
or at least the relationships of the type site of that period, than has
previously been possible.
While our evidence gives some suggestion of stratigraphic change,
the period as represented in our excavations is rather homogeneous.
One of the major stratigraphic changes is the introduction of certain
Leon-Jefferson pottery types, together with Spanish sherds, in the
upper portion of the mound, with a suggestion of the same trend
in the village. This corresponds to Stirlings observations of trade
goods with the upper burials in the burial mound.49 The vast bulk
of the pottery at higher levels, however, shows little change from
that below. There is a suggestion of projectile point change, with
the smaller, stemless triangular points tending to overlie the larger,
stemmed forms.
The evidence suggests that the site ranges from late prehistoric
times until late enough in the historic period for Leon-.Jefferson ma
terials to make their appearance. Smith has considered 1650 a rough
approximation of the beginning of Leon-Jefferson,50 and it certainly
cannot be much before that time, for Ft. Walton sites with European
trade materials are not uncommon in northwest Florida. The finding
of the Spanish sherds in the same levels as the Leon-Jefferson speci
mens, and not below, suggests that trade in the Tampa Bay area
intensive enough to make itself visible in the village debris is well
within historic times, perhaps a hundred years or more after De Soto.
This is an interesting point in and of itself, for it is usually assumed
that the absence of trade goods indicates a prehistoric date. This
is certainly not true for the levels closely underlying those with
Spanish sherds at the Safety Harbor site.
Viewing the site as a whole, we can concur with Willeys inter
pretation that Safety Harbor bears at least a partial contemporaneity
with both the Ft. Walton and Leon-Jefferson periods of the Florida
northwest coast.51 It has been possible to suggest rather significant
ceramic differences within the time range covered by the term Ft.
Walton at the Lake Jackson site, with earlier types having marked
relationships to Weeden Island, and with the later types presaging
Leon-Jefferson types in some cases.52 These later types are, many
of them, Pinellas types which occur also at the Safety Harbor site.
The problem of determining whether Pinellas Incised is a central
Gulf coast type which spread into the northwest coast area in later
Ft. Walton times, or whether on the other hand it actually originated
in Ft. Walton and spread to the central Gulf coast is really a knotty
one. If the latter is true then Safety Harbor as it is currently under
stood has a later origin than Ft. Walton. However, it may be that
we have on the central Gulf coast a situation more strictly analogous
to Ft. Walton than our excavations would suggest. Safety Harbor
"'Willey, 1949c.
M Smith, 1948, p. 316.
51 Willey, 1948, p. 217.
62 Griffin, N.D.6.
33


Incised bears much the same relation to Weeden Island Incised as
does Ft. Walton Incised. They are both somewhat sloppy and de
generate children of the classic Weeden Island Incised.
We did not recover much Safety Harbor Incised pottery from the
Safety Harbor site, but other sites, particularly burial mounds have
yielded more. It may be that these other sites represent an earlier
stage of Safety Harbor than that which we investigated. If this is
the case, both Safety Harbor and Ft. Walton, would have an earlier
stage with resemblances to Weeden Island pottery, and a later stage
characterized by the Pinellas types.
The northern origin of Pinellas Incised is a distinct possibility.
Pinellas Incised and Lake Jackson Plain have relationships to pottery
from Mississippian horizons in the broader Southeastern area. We
may postulate that the general region of northwest Florida, southern
Alabama, and adjacent areas, underwent change from Weeden Island
to the early Ft. Walton characterized by punctate filled incised areas,
and thence into a late Ft. Walton characterized by Pinellas Incised
and similar types. From the eastern part of the northwest Florida
area, where the Lake Jackson site is located, these influences made
themselves felt on the central Gulf coast.
There are several arguments which may be used to supplement
the general one advanced above. In the first place the amount of
decorated pottery is significantly less at Safety Harbor than at Lake
Jackson, as is shown below:
Site
Decorated
Plain
lii
Total
Lake Jackson
1028
13.3%
6743 86.7%
7771
Safety Harbor
216
2.8%
7308 97.2%
7524
Total
15,295
Not only is the amount of decoration less at Safety Harbor, but
only about 11.5% of the decorated sherds from Safety Harbor are
Pinellas Incised, whereas 27.8% of the decorated sherds from Lake
Jackson are of this type. This percentage becomes even higher
w'hen we consider that two stages are pooled in the Lake Jackson
figures, one of which shows a dominance of the type.
Can we explain Pinellas Incised as a central Gulf coast type which
diffused northward to the northwest coast? On the basis of these
figures we think not.
Secondly, there is the matter of the paste on which Pinellas In
cised is found. At the Lake Jackson site the type is found on a
coarse contorted grit-tempered paste exactly like that of the majority
of the sherds from the site. At Safety Harbor some sherds of Pinel
las Incised are on this same paste, while a few are on the crumbly,
laminated paste so characteristic of the site and classified as Pinellas
Plain. It seems far more probable to regard these few as a motif
34


transfer to the crumbly paste than to regard the more numerous
Lake Jackson specimens as a transfer. It would appear, then, that
the name Pinellas Incised is not too apt.
A northern origin for Pinellas Incised would fit in with other
features of the Safety Harbor Period, particularly the temple mound
and the triangular projectile points, both presumably due to Missis-
sippian influence. In other places one of the authors has expressed
the opinion that Safety Harbor and Ft. Walton both stem from
Weeden Island, which itself probably had areal variations, and are
the result of Mississippi influence through time.53 Safety Harbor,
being further from the source of this influence, would show less
effects at the same time level than would Ft. Walton. Too, all of the
influences did not penetrate simultaneously, or, at least, they were
not all adopted at the same time. Evidence suggesting that triangu
lar projectile points were adopted by Safety Harbor peoples after
the introduction of the temple mound has been cited earlier in this
paper.
J. M. Goggin has included the Safety Harbor Period in his Gulf
Tradition, rather than in his Florida Mississippian Tradition, which
includes Ft. Walton.54 Willey inclines toward retaining it in the
Gulf Tradition,55 and notes that one of the present writers stresses
the Mississipian influences. This difference of opinion is more ap
parent than real, and is primarily taxonomic.
These traditions are broad abstractions, of extreme value for
certain types of understanding and interpretation. However, there
is no absolute line dividing one from the other, and we have pre
ferred to look at the data from the point of view of culture change
and influence across areas rather than in terms of traditions within
areas. It is the transitional zone at the edges of the taxonomic boxes
in which we are interested. For the broad overall picture of culture
history the centers of the taxonomic boxes, the abstracted peaks
of the continuum, are more informative.
In conclusion we may state that the Safety Harbor site represents
a community which existed from late prehistoric times, or perhaps
early historic times in terms of dates if not in terms of trade goods,
and continued well into historic times, perhaps as late as 1700 A. D.
Thus, its time range overlaps both the Ft. Walton (the latter por
tion) and Leon-Jefferson periods of the Florida northwest coast.
No such marked change as occurred between Ft. Walton and Leon-
Jefferson occurred in this area. The isolation from direct Spanish
contact and from the aboriginal contacts to the north, which exerted
such an influence on Leon-Jefferson, were lacking in the Tampa
Bay area.
"'Griffin, 1949; N.D.a.
64 Goggin, 1949.
r Willey, 19496.
35


BIBLIOGRAPHY
Armistead, W. J.
1950. An Indian Stone Saw. The Florida Anthropologist, vol. 2, pp. 47-48.
Gainesville, Florida.
Bethell, John A.
1914. History of Pinellas Peninsula. St. Petersburg, Florida.
Braidwood, Robert J.
1945. The Order of Incompleteness of the Archaeological Record. Human
Origins: Selected Readings. Series 2, Article 9. University of Chicago,
Chicago, Ill. (Hectographed).
Goggin, John M.
1944. A Tentative Formulation of Pottery Types for the Glades Area, Flor
ida. New Haven, Conn. (Mimeographed).
1947. A Preliminary Definition of Archaeological Areas and Periods in
Florida. American Antiquity, vol. 13, no. 2, pp. 114-127. Menasha, Wise.
1948. Some Pottery Types from Central Florida. The Gainesville Anthro
pological Association, Bulletin, no. 1. Gainesville, Florida. (Mimeographed).
1949. Cultural Traditions in Florida Prehistory. In The Florida Indian
and His Neighbors, 3. W. Griffin, editor, pp. 13-44. Winter Park, Florida.
Goggin, John M., and Frank H. Sommer, III
1949. Excavations on Upper Matecumbe Key, Florida. Yale University
Publications in Anthropology, no. 41. New Haven, Conn.
Griffin, James B.
1945. The Significance of the Fiber-Tempered Pottery of the St. Johns
Area in Florida. Journal of the Washington Academy of Sciences, vol.
35, no. 7, pp. 218-227. Washington.
Griffin, John W.
1949. The Historic Archaeology of Florida. In The Florida Indian and
His Neighbors, J. W. Griffin, editor, pp. 45-54. Winter Park, Florida.
N.D.o,. Prehistoric Florida: A Review. Essays in the Archaeology of the
Eastern United States, J. B. Griffin, editor. In press.
N.D.5. Test Excavations at the Lake Jackson Site. American Antiquity.
In press.
Hobhouse, L. T., G. C. Wheeler, and M. Ginsberg
N.D. The Material Culture and Social Institutions of the Simpler People.
Monographs on Sociology, no. 3, London School of Economics and Political
Science. London.
Hrdlicka, Ales
1940. Catalog of Human Crania in the United States National Museum
Collections: Indians of the Gulf States. Proceedings of the United States
National Museum, vol. 87, pp. 315-464. Washington.
Krogman, W. M.
1939. A Guide to the Identification of Human Skeletal Material. FBI
Law Enforcement Bulletin, vol. 8, no. 8. Washington.
Lewis, Thomas M. N., and Madeline Kneberg
1947. The Archaic Horizon in Western Tennessee. Tennessee Anthro
pology Papers, no. 2. Knoxville, Tenn.
McKay, D. B.
1946. Colorful French Count Planted First Citrus Grove in this Area.
Tampa (Florida) Tribune, Dec. 29.
Moore, Clarence B.
1900. Certain Antiquities of the Florida West-Coast. Journal of ihe
Academy of Natural Sciences of Philadelphia, n.s., vol. XI, pt. 3, pp.
350-394. Philadelphia.
36


Sands, Earl
1948. Pinellas Creates Park for Phillippi, Citrus Pioneer. Tampa (Flor
ida) Morning Tribune, Feb. 4.
Senyurek, Muzaffer Suleyman
1947. A Note on the Duration of Life of the Ancient Inhabitants of
Anatolia. American Journal of Physical Anthropology, vol. 5 (n.s.),
no. 1, pp. 55-66. Philadelphia.
Smith, Hale G.
1948. Two Historic Archaeological Periods in Florida. American An
tiquity, vol. 13, no. 4, pp. 313-319. Menasha, Wise.
Stirling, M. W.
1930. Prehistoric Mounds in the Vicinity of Tampa Bay, Florida. Ex
plorations and Field Work of the Smithsonian Institution in 1929, pp.
183-186. Washington.
1931. Mounds of the Vanished Calusa Indians of Florida. Explorations
and Field Work of the Smithsonian Institution in 1930, pp. 167-172.
Washington.
1936. Florida Cultural Affiliations in Relation to Adjacent Areas. Essays
in Anthropology in Honor of Alfred Louis Kroeber, R. L. Lowie, editor,
pp. 351-357. Berkeley, Calif.
Straub, W. L<
1929. History of Pinellas County, Florida. St. Augustine, Florida.
Swanton, John R.
1922. Early History of the Creek Indians and Their Neighbors. Bureau
of American Ethnology, Smithsonian Institution, Bulletin no. 73. Wash
ington.
1946. The Indians of the Southeastern United States. Bureau of American
Ethnology, Smithsonian Institution, Bulletin no. 137. Washington.
Taylor, Walter W.
1948. A Study of Archaeology. American Anthropological Association,
Memoir, no. 69. Menasha, Wise.
Walker, S. T.
1880. Preliminary Explorations Among the Indian Mounds in Southern
Florida. Annual Report of the Smithsonian Institution for 1879, pp.
392-413. Washington.
Willey, Gordon R.
1948. Culture Sequence in the Manatee Region of West Florida. Amer
ican Antiquity, vol. 13, no. 3, pp. 209-218. Menasha, Wise.
1949a. Excavations in Southeastern Florida. Yale University Publica
tions in Anthropology, no. 42. New Haven, Conn.
19496. The Florida Indian and His Neighbors: A Summary. The Florida
Indian and His Neighbors, J. W. Griffin, editor, pp. 139-167. Winter
Park, Florida.
1949c. Archaeology of the Florida Gulf Coast. Smithsonian Miscellaneous
Collections, vol. 113. Washington.
Willey, Gordon R., and R. B. Woodbury
1942. A Chronological Outline for the Northwest Florida Coast. Amer
ican Antiquity, vol. 7, no. 3, pp. 232-254. Menasha, Wise.
37


Explanation of Plates
Plate I. Potsherds.
A, Jefferson Rim, type 3; B, Jefferson Stamped, type 5 Rim; C-F, Lake
Jackson Plain; G, Pinellas Incised, type A; H, Pinellas Incised, types A and B
combined; I-K, Pinellas Incised, type B; L, Pinellas Incised, variant of type C;
M-S, Safety Harbor Incised; T-U, Ft. Walton Incised; V-W, similar to Sarasota
Incised; X, unique; Y, Pinellas Plain with notched lip; Z-BB, unique lips;
CC-DD, Random punctated; EE-FF, St. Johns Check Stamped; GG, ring base.
H, R, S, and X from surface collections, balance excavated.
Plate II. Chipped Stone Tools.
A-F, Triangular points; G, Lanceolate point; H, corner-notched point; I-J,
side-notched point; K-Q, stemmed points; R-S, drills; T-AA, scrapers, BB-DD,
utilized flakes. D, I-J, from surface collections, balance excavated.
Plate III. Plummets, Shell Tools, Daub, and Hammerstones.
A-C, Plummets; D-E, columella chisels; F, Busycon pick, type A; G, per
forated Venus shell; H, fragment of daub; I, chipping hammer; J, hammerstone.
Plate IV. Bone Tools, Pottery Hone, and Grindstone Fragments
A-H, bone pins; I, bone handle; J, bone gorge; K-L, splinter awls; M, bone
tube; N, pottery hone; O-P, fragments of grindstones.
38


Plate I.
Potsherds
39


Plate II.
Chipped Stone Tools
40


Plate III.
Plummets, Shell Tools, Daub, and Hammerstones
41


Plate IV.
o
Bone Tools, Pottery Hone, and Grindstone Fragments
42


\






PUBLICATIONS
FLORIDA ANTHROPOLOGICAL SOCIETY
Publications of the Society appear at irregular intervals
as funds permit. These are provided by the Sustaining
Memberships and include, in some instances, special gifts
or subsidies provided by the author. Publications are dis
tributed to all members of the Society on issuance. Addi
tional copies may be secured from the treasurer, as follows:
1. Two Archeological Sites in Brevard County, Florida.
Hale G. Smith. 31 pp., 2 figs., 4 pis. Price $0.50.
2. The Safety Harbor Site, Pinellas County, Florida.
John W. Griffin and Ripley P. Bullen. 42 pp., 2 figs.,
4 pis. Price $0.50.
Address orders to the Treasurer, Florida Anthropological
Society, Florida Park Service, Seagle Building, Gainesville,
Florida.