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
M l MERN SNflf C HClHrdl
VOL. 7, f
VOL. VI f
THE FLORIDA ANTHROPOLOGIST
Vol. VI March, 1953 No. 1
Stratigraphy at a Hialeah Midden - -- -D. D. Laxson 1
The Famous Crystal River Site -- -Ripley P. Bullen 9
Editor's Notes ------------ ------- 38
Space and Time Perspective in Northern St. Johns
Archeology, Florida, by John M. Goggin --John W. Griffin 39
Contributors to this Issue - - -- - 41
Editorial Office at the Florida State University, Tallahassee
Published at the University of Florida, Gainesville, for the
Florida Anthropological Society
STRATIGRAPHY AT A HIALEAH MIDDEN
D. D. Laxson
A quarter of a mile south of the junction of Gratigny Road and Peters Pike in
the NW of Section 35, Township 52 South, Range 40 East, Dade County, Flori-
da, is a cluster of three small hammocks. They are strung out in a general WNW-
ESE direction with distances between them varying from three hundred to six
hundred feet. A narrow ridge connects two of the hammocks. A road and canal,
north of the hammocks, runs westward connecting rock pit operations in the
vicinity of Peters Pike. The area lies a mile and eight-tenths north of the Miami
The middle hammock (Site Dd-75, U. of F. records), eleven hundred feet east
of Peters Pike and fifteen hundred feet south of the rockpit road, was the scene
of test pit excavations reported in this article. Land surface near the midden is
about five feet above sea level. Rockpit operations are carried on to the north-
east and east. Small truck farms and pasture lands are to the north and south.
The soil is muck over'a calcareous marl. Underneath this is the gray quartz
sand of the Pamlico formation. Bedrock is a creamy pot-holed oolite. Shallow
water covers most of the area approaching the midden a considerable part of the
The midden is located in the center of the hammock and is roughly circular.
Dense vegetation prevents ascertaining its exact shape. Distances across the
midden are 170 feet north and south and 131 feet east and west. The slope is
gradual with the exception of a small area to the north where a rock ledge drops
The highest point in the hammock is three feet, eight inches above the terrain.
Several measurements were taken in this respect and an average height of 3.3
feet was obtained.
A base line running north and south was established parallel to a large dead
tree and all distances were measured from this line. A test or exploratory trench
of sixty feet in the form of a cross was first dug, and results warranted further
excavation. The test pits were all five feet square and only in one instance,
Pit 4, was it necessary to gradually reduce this distance on account of loose
soil and material falling in from the sides. These test pits were excavated and
the material screened by four-inch arbitrary levels except for Pit 4 where s;x-
inch levels were used because of the looseness of the debris. Relationships
between tests and the base line are shown in the excavation plan (Fig. 1).
All pits were relatively shallow, the deepest being approximately two feet
FICUS TR\E FE
Fig. 1. Excavation plan of Hialeah No. 1 midden site. Distances across mound:
N.-S., 170 feet; E.-W., 131.5 feet. Slope: N.-S., gradual, E.-W., abrupt.
Height of mound above surrounding terrain, 3.25 feet. Scale of above draw-
ing, one-eighth inch equals one foot.
before bedrock or a hard calcified deposit of bone and shell was reached.
Stratigraphic results are given in Table 1 and discussed below.
As conditions probably have not changed much since Indians inhabited this
spot, the following notes on our observations of wild life may be of interest.
Birds such as the hawk, vulture, cara cara, and whippoorwill were seen in the
hammocks. Water birds were well represented by the heron, egret, coot, and
Several small animals were observed in the late afternoons, including coon,
opossum, rabbit, and swamp rat. Only two reptiles were encountered, the alert
and speedy Florida skink and the arboreal chicken snake. Several specimens of
the latter were seen. They varied in length from three to six feet and were in
the trees at heights of from six to twenty feet.
By far the most numerous insects were the small black mosquito and the
yellow and black striped butterfly known as the "zebra".
Vegetation was profuse and varied. The list of water plants and grasses
would include arrowhead, Boston fern, swamp fern, para grass, sawgrass, and
bladderworts. The only known trees and scrubs were wax myrtle, dahoon holly,
willow, mastic, scrub oak, and strangler fig.
Exploration of the midden began in February, 1952. The excavation of
stratigraphic pits started in May, 1952 and ended the latter part of January, 1953.
Work was done on convenient weekends when weather permitted and other engage-
ments did not interfere with the presence of a full crew.
Specimens are illustrated in Figures 2 and 3 and their provenience given in
the discussion. Typology follows that used in reports listed later under
The 60-foot exploratory trench produced seven shell- or coquina-tempered,
nine St. Johns Plain, thirteen St. Johns Check Stamped, 266 Glades Plain, Seven
Glades Tooled, and three Key Largo Incised sherds. This result presented an
archaeological problem, as St. Johns Check Stamped and Glades Tooled are
recognized as Glades III B and C types dating between 1400 and 1700 A.D. and
Key Largo Incised should represent Glades II times or some time between 100
and 1100 A.D. (Goggin, 1950, a and b), while the shell- or coquina-tempered
sherds were previously unreported in the literature of the Miami area.
Shell- or coquina-tempered sherds refers to pottery fragments containing
finely crushed white material which may be either ground shells or ground rock
formed from shells. The tempering material is too fine for identification. These
sherds represent an unnamed pottery type known previously only from surface
collections. Until more information is obtained, it would not seem advisable
21 I .- 1 IN.
Fig. 2. Sherds from Hialeah midden. First row, and first two sherds of second row,
Glades Tooled variations. Last sherd, second row, Surfside Incised. Third
row, first three sherds, Key Largo Incised; last sherd, Opa Locka Incised.
Fourth row, first sherd, Opa Locka Incised; upper two, Matecumbe Incised
variant; lower two, unique incised; lower right, Miami Incised.
Fig. 3. Miscellaneous artifacts from Hialeah midden. First row, fragment of cut shell
ornament, blue glass bead, and bone bead. Second row, two fragments of
bone awls or pins, a "peg-topped" bone pin, and two bone points. Sherds
are variations of St. Johns Check Stamped pottery.
TABLE 1. VERTICAL DISTRIBUTION OF SHERD TYPES.
PIT 1 PIT 2 PIT 3 PIT 4
TYPES 0-4 4-8 8-12 0-4 4-8 8-12 12-16 0-4 4-8 8-12 12-16 0-6 6-12 12-18
Shell-coquina-tempered, plain 15 8 1 1 4 1 3
St. Johns Check Stamped 1 1 1 4 1
St. Johns Plain 1 7 3 1 4 1 1 3 2
Glades Tooled 1 1 1 3 4 1
Glades Red 1 2 1
Glades paste, black paint 1 1
Glades Plain 100 8 90 98 18 25 127 97 40 6 65 87 7
Belle Glades Plain 1 1 2 4 1
Surfside Incised 1 1 2
Key Largo Incised 3 8 1
Matecumbe Incised 2
Opa Locka Incised 1
Unique Incised 1 1
Miami Incised 1
to give them a type name, although Hialeah Plain might be suggested.
As the site seemed to represent two time periods, stratigraphic excavations
were made to determine local relationships between these pottery types and to
extend to the Miami area, if possible, the ceramic sequence determined by
Goggin for Matecumbe Key to the South (Goggin, 1949) and for the Everglades to
the west (1950a), for the later periods. It was also thought it would supplement,
very well, Goggin's work at Fisheating Creek (Goggin, 1950a), also in the Miami
area, as the Fisheating Creek site pertained chiefly to relatively early pottery
times and the one at Hialeah represented the later end of the sequence. As will
be seen below, this proved to be the case.
Pit 1, unfortunately, was located where the deposit was not very thick and
the results were not very good. Pit 2 was much more encouraging, as St. Johns
Check Stamped, Glades Tooled, and shell-coquina-tempered sherds were found
to be limited to upper zones. However, decorated pottery, such as Key Largo
Incised, was not found in lower zones and a Miami Incised sherd seemed out of
Pits 3and4 were very satisfactory stratigraphically. As shown in the Pottery
Distribution Table, St. Johns Check Stamped, shell-coquina-tempered, and Glades
Tooled sherds were well concentrated in the uppermost zone to represent Glades
III B times. Surfside Incised very nicely spanned this zone and the one below to
indicate the presence of Glades III A occupation. Glades II types, Key Largo
Incised, Matecumbe Incised, Opa Locka Incised, and a unique incised type (Fig.
2, fourteenth through seventeenth) were located only in a lower (or intermediate)
The thirteen Glades Plain sherds from the lowest zones of these pits were
found among the crushed food bones and shells (sometimes cemented into a
breccia) which formed the basal portion of the midden. They are substantially
lighter in color than sherds from higher up in these pits. Presence of these
lowest sherds is suggestive of a Glades I, predominantly undecorated, period
but the quantity is inconclusive.
Presence of Belle Glade Plain sherds in upper zones should be noted. Pre-
sumedly, these represent "trade sherds" from the Lake Okeechobee area.
Non-ceramic specimens are illustrated in Figure 3. The glass bead came
from the top zone of Pit 2. Some pieces of iron were also found in the very top
of the midden. Presumedly these indicate post-Indian use of the hammock, but
the glass bead probably indicates early European contact or Glades III C
A fragment of a cut shell ornament (Fig. 3, first) was found in the middle
zone of Pit 1. The small but very nice "peg-topped" bone pin is Glades III B
as it was in the highest zone of Pit 3. The small, short bone.points (Fig. 3,
second row, last two) are clearly Glades II as they came from the second zones
of Pit 3 and 4, as did the tip of a Busycon pick and a fragment of worked col-
umella (not illustrated). Of the fragmentary bone points or pins, one came from
Pit 1 and the other from the second zone of Pit 3. A Busycon pick, Type A, was
found in the exploratory trench.
Food remains are represented by deer, turtle, rodent, snake, fish including
shark teeth and vertebrae, and bird bones, crab claws, and alligator teeth as
well as snail, clam, oyster, Strombus, Busycon, and fresh water mussel shells.
No significant trends were noted with depth in respect to food remains. Indians,
living at Hialeah many hundreds of years ago, enjoyed a varied diet of land and
Appreciation is expressed to Mr. Jack Christiansen, owner of the land, for
permitting us to dig, to Bob Masters and Noel Hermann for helping with the dig-
ging and screening and for the map of the area and last but not least, for the
generous portions of time, patience, and advice, to Mr. Ripley P. Bullen,
Curator of Social Sciences, Florida State Museum, Gainesville.
Goggin, John M.
1949. "Excavations on Upper Matecumbe Key, Florida." Yale Univer-
sity Publications in Anthropology, No. 41. New Haven.
1950a. "Stratigraphic Tests in the Everglades National Park." American
Antiquity, Vol. 15, No. 3, pp. 228-246. Menasha.
1950b. "Florida Archaeology 1950." Florida Anthropologist, Vol. 3,
Nos. 1-2, pp. 9-20. Gainesville.
1951. "Archeological Notes on Lower Fisheating Creek." Florida
Anthropologist, Vol. 4, Nos. 3-4, pp. 50-66. Gainesville.
THE FAMOUS CRYSTAL RIVER SITE
Ripley P. Bullen
Importance of the Crystal River site is indicated by numerous references in
archaeological literature. Twenty-four titles in the following bibliography are
either chiefly concerned with this site or specifically refer to it.
Clarence B. Moore partially excavated the large burial mound (Fig. 1, E and
F) at the site in 1903, continued that work in 1906, and returned again in 1917
to uncover additional burials. His finds consisted of decorated pottery (includ-
ing three negative-painted vessels), smoking pipes, shell ornaments, copper
discs, ear plugs, a conjoined tube, and an embossed plate, a great many plummet-
like objects of shell and of stone, as well as shell and bone tools and other
artifacts which he did not illustrate (Moore, 1903; 1907; 1918). Some of these -
the copper objects, sheets of mica, rock-crystal pendants, soapstone pipes,
greenstone celts and other artifacts of igneous rock represented exotic materials
which must have been imported into Florida by Indians.
After 1917 the site had not been visited by an archaeologist until 1949, when
Gordon R. Willey, Antonio J. Waring, Jr., and Rufus Nightingale went there and
made a collection of sherds (Willey, 1950). In February, 1951, a party from the
Florida State University headed by Hale G. Smith and James B. Griffin, the lat-
ter from the University of Michigan, visited the site, made a stratigraphic test,
and collected sherds from the surface (Smith, 1951).
Smith's test, while indicating more than one ceramic period, did not uncover
any decorated pottery which could be used to correlate the local sequence with
existing chronologies. To meet this need, the Florida ParkService made two
additional stratigraphic tests and another surface collection in June, 1951.
The Crystal River Site, in Citrus County, Florida, is situated on the north
side of the river about two miles west of the town of Crystal River and about
four miles east the point at which the river finally meets the Gulf of Mexico.
Fig. 1 (after Moore, 1903, p. 380) locates the major features of the site. These
features have been described by Moore and, more recently, by Willey (1950).
Due to lack of time and an abundance of rainfall, I gave these features only a
cursory examination. It seems desirable, however, to record my impressions to
supplement earlier descriptions.
All land surrounding these features is fairly low and, as noted on the plan
(Fig. 1), standing water was found in two places. To some extent this low
elevation must reflect removal of material by Indians to construct the large
s G ;
S=E -- --
: :-,,,*,,, I',, .
, D ,
I I ll ll te r
"B. ,/I ''.r
0 100 E00
Fig. 1. Map of the Crystal River Site, Citrus County, Florida (after Moore, 1903,
Fig. 13). A, Temple mound and ramp. B and BB, Shell midden. C, Cir-
cular embankment. D, Level area. E. Prafform of sand and shells. F.
Burial mound proper. G, Shell deposit. H, Temple mound. B, Location
of Florida State University test, and I and II, Florida Park Service tests,
burial mound (Fig. 1, C-F). Shell and other material, as well as residual lime-
stone boulders, were also removed to build the temple mounds (Fig. 1, A and H).
Temple Mound A, with its ramp, remains as illustrated and described by Moore
and Willey, except for a small hole in the top and some erosion by the river at its
southern corner. An 1859 description of this mound by F. L. Dancy (Brinton,
1859, p. 179) notes:
". .it is on all sides nearly perpendicular, the faces covered with brush
and trees to which the curious have to cling to effect an ascent. It is
nearly forty feet in height, the top surface nearly level, about thirty feet
across, and covered with magnolia, live-oak, and other forest trees, some
of them four feet in diameter. Its form is that of a truncated cone, and as
far as can be judged from external appearances, it is composed exclusively
of oyster shell and vegetable mould."
At the end of the ramp there is a clearly defined ridge or walkway of shells
and midden material extending northeasterly towards the eastern end of the shell
ridge or midden (Fig. 1, BB). This feature, not previously mentioned in the
literature, seems to me to be important in chronological interpretation of the
site. It gives access to the ramp from the eastern end of the shell midden, the
only location where we found definitive pottery of the Safety Harbor period. This
is the period during which temple mounds are usually supposed to have been
built in the area.
The extensive shell midden (Fig. 1, B-BB) did not seem to me to be as wide
as Moore indicates. Natural growth of soil, muck, humus, etc., over the lower
portions of the sides of this midden, during the nearly fifty years since Moore
made his plan of the site, probably explains this discrepancy. We made two
small tests at the northwestern edge of the standing water shown north of the
walkway. Buried shells and Pasco Plain sherds were found near the edge of this
water but not, apparently, below the area covered by water.
Three houses, mentioned by Willey, two gardens, and various shade trees are
located on the eastern portion of the midden (Fig. 1, BB). Smith's test in the
western portion of the midden was located at the point marked B and ours, Tests
I and II, as indicated, slightly to the south of B.
Extending nearly 200 feet northward from the bend of the shell midden near
B, is a low, irregular ridge, wider towards the north, which ends at a low area
filled with standing water (Fig. 1). This water continued an undetermined dis-
tance to the west but only a short distance to the east. A small test at the
northern end of this ridge showed it to be a shell or shell-midden deposit cover-
ed with six to nine inches of dirt. I did not proceed further north to ascertain
whether or not this shell deposit was the same as that mentioned by Moore (Fig.
1, G). That such may be the case is suggested by the fact that I had to walk
southeasterly from this point to reach the burial mound.
The burial mound (Fig. 1, C-F), according to Moore, originally was a complex
consisting of a circular embankment (C) built of midden material which surrounded
a level area (D). Within this level area was the burial mound proper (F), made of
sand with a small amount of included shells, surrounded by an artificial platform
of sand and shells (E). Today the mound is a series of hillocks representing
Moore's spoil piles. As implied above, the burial mound may have been located
a little further south than is indicated on Moore's map.
Temple Mound H was not seen by either Willey or me. Smith, who dug a small
test there and found eight Pasco Plain sherds, tells me it was, apparently, built
of limestone boulders and dirt.
THE BURIAL MOUND
Before taking up the results of our stratigraphic tests, it seems worthwhile to
examine Moore's three reports on his work at the Crystal River burial mound to
see what implications of culture change may be found therein. In this connection
it should be remembered that Moore found copper objects (as well as other artifacts)
diagnostic of the Hopewell of Ohio (Greenman, 1938), the cultural equivalent of
which in Florida is believed to be Santa Rosa-Swift Creek; decorated pottery typical
of the Deptford, Santa Rosa-Swift Creek, and Weeden Island periods; and negative-
painted vessels which, outside of Florida, are considered to belong in a late Mid-
dle Mississippi or Temple Mound I horizon (Willey and Phillips, 1944, p. 175).
The Florida representatives of the latter horizon would be Fort Walton or its more
southern temporal equivalent, the Safety Harbor period.
If the burial mound at Crystal River was used over any such time range as is
suggested by this pottery, some evidence of the passing of time should be found
in Moore's writings, even though he was notoriously thoughtless about chronology.
As we shall see, there are some such suggestions.
Moore describes the burial mound as ten feet, eight inches in total height
(Fig. 1, F), and as surrounded by a platform or apron (Fig. 1, E). He presents
two surface profiles from which it is possible to learn that the platform had a
height of about three and a half feet while the mound eminence rose about seven
feet higher (Moore, 1903, Fig. 17). It is evident that anything found at a depth of
four feet or more must have come from the middle or lower part of the mound proper
and not from the surrounding platform. Objects below a depth of seven feet must
have come from the lower part of the mound.proper and have been deposited rela-
tively early in the history of the burial mound.
The first season, Moore excavated all the mound proper and over half of the
surrounding platform, finding over two hundred burials about equally divided be-
tween bundle, flexed, and extended interments. For the platform, Moore noted
that grave shafts frequently cut through other burials. Unfortunately, he did not
note which form of burial was found in these pits, but the passing of at least a
small amount of time is evident.
More burials came from the platform than from the mound proper, while artifacts
from the mound were, on the average, of better quality and more spectacular than
those from the platform. With only three exceptions, all burials in the platform
were below masses of shells. In the mound proper many burials were also below
masses of shells (probably in lower zones) but forty were unassociated with
shells. These differences in burial modes may be temporal in nature, as Moore
suggested (1907, p. 425).
One copper object came from the platform and twelve from the mound proper.
Of the latter -
"All came from considerable depths in the mound, from 4 to 8 feet" (Moore,
1903, p. 408).
"On the base of the mound, in the southern slope, was the skeleton of an
adult, lying full length on back. Extending across the pelvis, sagging down
somewhat, was.a row of pendants of stone, among which were three of cop-
per ." (Moore, 1903, p. 399).
"About 5 feet down, (also) in the southern side of the mound, near a skull
belonging to a bunched burial ." (Moore, 1903, p. 400).
It would appear from the above that copper objects were relatively deep and
hence early in the history of the mound proper and that, in at least one case,
prone burials were deeper than bundle burials.
On Moore's second visit, he excavated the remaining portion of the surround-
ing platform. He found 186 burials, among which were 118 in prone position,
fifty-two flexed, and only eight of the bundle type. Of these burials Moore writes:
"Bunched burials, only a few of which were found at our second visit,
invariably lay in sand, and were always at a less depth than the shell
deposits beneath which other forms of burial lay. However, these bunched
burials were found where the mound joined the sloping ground (platform)
and probably were left over from the great number'of bunched burials which
we found in the mound proper at the time of our first visit" (1907, p. 408).
The above paragraph clearly implies that bundle burials were shallower and
hence later than other forms of burials and that they were typical of the mound
proper, i.e., of the part of it which extended above the surrounding platform.
Hence the upper part of the mound proper could represent a late building stage
while the lower part of the mound should be earlier than the surrounding plat-
Moore's third trip to Crystal River was limited to excavations in the embank-
ment or circular ridge (Fig. 1, C) which encircled the burial mound area. This
embankment, which proved to be made of midden material, produced twenty-four
burials, all either flexed or lying full length on the back (Moore, 1918, p. 57Z.
Lack of bundle burials supports the hypothesis presented above that the bundle
form of burial represents a different time period than that of prone or flexed inter-
The above analysis of the sketchy data presented by Moore suggests but does
does not demonstrate three periods of use of the burial mound at Crystal River.
One would be represented by the lower portions of the mound proper, where prone
and flexed burials were found with copper objects; the second by the surrounding
platform and embankment, with prone and flexed burials; and the third by the
higher portions of the mound proper, with bundle burials.
Lower portions of the mound proper may be considered Santa Rosa-Swift
Creek in date because of the Hopewellian-like copper objects found there. The
footed or tetrapodal vessels found by Moore presumedly came from this lower
Moore's collection from his second visit, when his digging was done entirely
in the surrounding platform, includes sherds of Weeden Island I and earlier ves-
sels. The latter may have come from the underlying periphery of the mound prop-
er, or they may represent the beginnings of the platform. A Weeden Island I
date, at least in part, is confirmed by Willey's and Smith's collections (Willey,
1950, p. 44; Smith, 1951).
Willey's collection includes five Dunns Creek Red sherds. This is important,
as this type was definitely placed in upper zones in the stratigraphic tests (see
Tables 1, 2, and 3). His collection also includes Thomas Simple Stamped and
Tampa Complicated Stamped sherds. These types are sometimes considered
Weeden Island II in date (Willey, 1949a, pp. 436 and 440). Further south, across
Tampa Bay, all three of these types have been found associated with Wakulla
Check Stamped sherds in a Weeden Island II complex (Bullen, 1951a).
As Willey's collection came from Moore's spoil piles and from under an up-
rooted tree at the encircling embankment, the Thomas Simple Stamped and Tampa
Complicated Stamped sherds may have come from the latter location. At this
latter location (the encircling embankment) J. B. Griffin found two St. Johns
Check Stamped sherds (Smith, 1951). Hence, the encircling embankment must, by
definition, be assigned a Weeden Island II date, at least in part.
The bundle burials from the upper part of the mound proper I would consider
at the earliest, Weeden Island II in date. To support this allocation I offer the
All burials in the Safety Harbor, Parrish I, and Parrish II burial mounds, all
mounds of the Safety Harbor period, were of the bundle type. At the Weeden Is-
land, Thomas, and Prine burial mounds, bundle burials overlay other types and
Safety Harbor pottery is present (Willey, 1949b; Bullen, 1951a). These sites are
all to the south of Crystal River. At Buzzard's Island, two miles from the Crystal
River site, predominantly bundle burials and Safety Harbor pottery were found
To the north of Crystal River, bundle burials, while typical of Fort Walton
sites, are also clearly present in mounds from which all the available pottery is
Weeden Island. Burial habits were changing during Weeden Island times, with
the practice of bundle burials being introduced from north to south along the gulf
coast of Florida.
Ceramic analysis of Tests I and II dug by the Florida Park Service are pre-
sented in Tables 1 and 2, and Table 3 gives similar data from the test made by
Smith and the Florida State University party. Smith's data have been rearranged
under the same headings used in Tables 1 and 2.
Both in Tests I and II, the upper four feet consisted of black dirt and shells,
with concentrations of ash between depths of two and four feet. Shells represent-
ed about 40 per cent of the debris and 70 per cent of the shells were those of
oysters. Below four feet, the deposit contained relatively few shells (about 15
per cent by volume) and no concentrations of ash. In Test II crushed oyster shells
were encountered between depths of seven and eight feet.
Examination of these tables will reveal that, except for one sherd, Dunns
Creek Red pottery is limited to depths of less than thirty-four inches. In all cases
-- 0 CM. OR 3 18/16 IN.
Fig. 2. Bone Adze Socket from Crystal River Site.
TABLE 1. VERTICAL DISTRIBUTION OF SHERDS. TEST 1.
Depths 0 -a 0 I
0-4 96 1 2 1 1 6 106
4-8 218 7 3 15 243
8-12 227 19 5 1 25 277
12-16 132 6 4 2 25 1 170
16-20 103 4 4 6 117
20-24 117 7 4 2 130
24-28 137 1 5 4 1 1 1 7 1 158
28-34 112 5 5 4 126
34-40 120 S 2 1 128
40-45 119 2 1 1 2 1 126
45-48 84 1 5 1 91
48-60 131 4 14 1 150
60-66 54 1 4 1 1 2 63
66-82 10 1 11
talt mg1660 1 1 67 31 5 1 3 2 115 1 1 1 1 1 1 1 1 3 1896
TABLE 2. VERTICAL DISTRIBUTION OF SHERDS, TEST 2.
6 1 6
6 2 1 2
4 3 4
11 12 1
Totals 1758 2 2 3 1 1 116 37 1 1 1 1 67 1 3 1 1 1 1998
0 u b u
TABLE 3. VERTICAL DISTRIBUTION OF SHERDS AND TOOLS, F.S.U. TEST.
Surface a *-
in & w a E 8
W A K: IZ 0 IQ
Inches j*^ 0 h | I
0-4 59 1 262
4-8 86 3 1 2 92 2 1
8-12 49 4 7 1 6 67 1 1
12-16 36 3 3 1 43 1
16-20 58 4 2 2 66 1 1
20-24 81 5 1 2 89 1
24-28 71 1 1 1 74 2
28-32 53 2 55
32-38 71 2 3 1 77 1 1
38-42 51 1 52 1
42-48 7 7
Totals 622 22 17 3 19 1 684 3 1 i 2
there is a substantial pottery-producing zone below the lowest sherd of Dunns
Creek Red. As all three tests are in agreement on this point, there can he no
question but that at least two periods are represented in the portion of the midden
tested, one period being before and the other after the introduction of Dunns Creek
Further examination will disclose that Pasco Plain (limestone-tempered), St.
Johns Plain (chalky), and Gritty Plain (sand-tempered) sherds were found in
practically all levels of all tests. As Pasco Plain was the dominant ware, sherds
of it were carefully examined in an attempt to break it down into smaller, chrono-
logically valid categories. Segregation by the amount or fineness of temper did
not give consistent results. Sherds with well smoothed interiors and those having
vertical scraping or scoring marks a short distance below the rim (Fig 3, largest
sherd) tended to be concentrated in the upper part of the midden, a short distance
above the level of the introduction of Dunns Creek Red.
In these tests, decorated pottery was extremely rare. The lower zones of
both Tests I and II may be considered Santa Rosa-Swift Creek in time period,
as Franklin Plain rims (Fig. 3, fourth row right), Swift Creek Complicated
Stamped sherds (Fig. 3, third row right and bottom row last sherd) were found
in them. A few sherds of the Deptford complex (Fig. 3, bottom row, first two
sherds) were also found at depths below those of the first introduction of Dunns
Creek Red. This Santa Rosa-Swift Creek or pre-Dunns Creek Red period should
correlate with the lower part of the burial mound where Moore found Hopewellian
Higher levels producing Dunns Creek Red pottery should represent the suc-
ceeding or Weeden Island period as this pottery type is associated with Weeden
Island pottery elsewhere on the central gulf coast of Florida (Willey, 1949a,
p. 121; Bullen, 1951a). Two Weeden Island-like rims (Fig. 3, top row, last two
sherds), a Carrabelle-like incised sherd and a Carrabelle-like punctated sherd
(Fig. 3, second row last two sherds) would seem to substantiate this interpreta-
tion. Due toan absence of check-stamped pottery below a depth of four inches,
it would seem proper to call these levels Weeden Island I. They should correlate
with the surrounding platform of the burial mound where Moore, Willey, and Smith
found Weeden Island I pottery and where Willey, presumedly, found Dunns Creek
The top four inches of Test I, with a St. Johns Check Stamped sherd, would,
by definition, be Weeden Island II. It should correlate with part of the encircling
embankment of the burial mound complex, where Smith found two sherds of St.
Johns Check Stamped pottery.
Artifacts of shell, stone, and bone from the tests have been arranged by stra-
tigraphic levels in Tables 3, 4, and 5. Busycon gouges, it will be noticed, are
TABLE 4. VERTICAL DISTRIBUTION OF NON-CERAMIC SPECIMENS, TEST 1.
t z. IU
1 2 0
1 1- r- -
1 1 1
Totals 1 2 1 1 20 2 1 7 22 1 1 3 18 2 4 3 1 736
1 1 1 5
TABLE 5. VERTICAL DISTRIBUTION OF NON-CERAMIC SPECIMENS, TEST 2.
Depths s a
Below 0 U 4 4 l |
iU V o Se
in g3 -4 W a u 0
Inches W U QU 1 m 0 Cm W ko
~~6~ 1 av
2 1 3
3 1 8
1 2 1
1 1 114
Totals 8 1 2 3 2 43 7 7 3 81 1 2 86 2 8 1 2 1 1 2 1 1 8 1518
restricted to upper zones and appear to be a Weeden Island type of tool, although
they are found in much older horizons in East Florida. Otherwise, except for pro-
jectile points and stone knives, none of these artifacts is limited in vertical dis-
tribution or, if so limited, is present in sufficient quantities to be definitive.
Melongena hammers, the most common artifact, was popular during all periods
penetrated by tests.
Small projectile points (Fig. 4, top row) were limited to the eight to twelve
and sixteen to twenty inch levels of Test II and therefore are Weeden Island in
date. The only Weeden Island village site in Florida for which sufficient data
are available for comparison, the Butler site on the Chattahoochee River, also
produced small projectile points of various shapes (Bullen, 1950, Fig. 5).
One of these specimens (Fig. 4, top row, last point), from the eight to twelve
inch level is probably a drill, and not a projectile point, as the edges and point
appear blunted. Made of a concavo-convex chip, it is worked only on the upper
or convex side, while the base appears to have broken off during manufacture.
In these respects, albeit larger and cruder, it resembles some of the narrow tri-
angular projectile points from the Safety Harbor site (Griffin and Bullen, 1950,
Wider, heavier projectile points with slightly expanding tangs or with side
notches (Fig. 4, second row and third row, third point) came from greater depths
and pertain to the Santa Rosa-Swift Creek period. The one found by Smith at a
depth of thirty-eight to forty-two inches is of the same type. These points com-
pare closely in typology with those from the Tan Vat site in the Chattahoochee
Valley, which produced very late Santa Rosa-Swift Creek pottery (Bullen, 1950,
A fragment of a stemmed point (Fig. 4, third row, second point) came from
Test II at a depth of forty-eight to fifty-four inches. It is heavily patinated, and
may be a hold-over from an earlier period.
As is shown in Tables 4 and 5, stone knives (Fig. 4, third row first and fourth,
bottom row all) were limited to the lower or Santa Rosa-Swift Creek zone. Again,
confirmatory evidence for such dating comes from the Chattahoochee Valley, where
a similar knife was found at the Tan Vat site, a village of the late Santa Rosa-
Swift Creek period (Bullen, 1950, Fig. 4, s).
Summarizing data from the stratigraphic tests, we find the typology of both
chipped stone tools and pottery defines two major periods for the tested portion
of the Crystal River midden. Furthermore, the correlation between ceramic periods
and types of chipped stone tools agrees with data from the Chattahoochee Valley.
EXPLANATION OF ACCOMPANYING FIGURES
Fig. 3. Decorated pottery. First row: Incised, Safety Harbor period; Pinellas
Incised with Lake Jackson type lug; Weeden Island-like rims, Pasco
Plain paste. Second row: St. Johns Check Stamped; Swift Creek Com-
plicated Stamped (late variety); Carabelle-like Incised and Punctated.
Third row: Zoned textile imprinted; Pasco Plain with vertical scoring
below rim; scored or simple stamped; Swift Creek Complicated Stamped.
Fig. 4. Chipped stone artifacts. First row: Small projectile points. Second
row and middle third row: Expanding tanged and side-notched projec-
tile points. First and fourth specimens in third row, and fourth row:
Fig..5. Miscellaneous artifacts. First row: Fragments of plummets; antler
punch; polished bone. Second row: Pottery discs (?). Third row:
Grindstone fragment; mica; fragment of polished stone; chipping ham-
mer. Fourth row: Handle, Dunns Creek Red ceramic vessel; uniquely-
shaped fragment, St. Johns paste ceramic vessel.
Fig. 6. Bone artifacts. First row: Four grooved pins; four simple pins; bone
point. Second row: Flattish pieces. Tlird row: Spatulate forms;
splinter awls; finished awls.
0 1 2 3 45
Fig. 3. Decorated Pottery.
"*^ ''^ Mtak--
Fig. 4. Chipped Stone Artifacts.
o 1 23
Fig. 5. Miscellaneous Artifacts.
Fig. 6. Bone Artifacts.
Surface collections were made by us from Temple Mound A (Fig. 1, A) and
from the eastern end of the shell midden (Fig. 1, BB). Smith also collected from
Temple Mound A and from the other temple mound (Fig. 1, H). Willey's and
Smith's collections from the burial mound complex have been mentioned earlier.
At both temple mounds, Smith found Pasco Plain sherds. At Temple Mound A,
we found forty-eight Pasco Plain, one Gritty Plain, two St. Johns Plain, and two
St. Johns Check Stamped sherds (Fig. 3, second row, first sherd). Presence of the
last two sherds implies that this temple mound was built no earlier than Weeden
Island II times. These sherds are not incompatible with a Safety Harbor date, as
sherds of this type were found in the temple mound at the Safety Harbor site
(Griffin and Bullen, 1950, p. 18).
From the surface of the eastern end of the midden (Fig. 1, BB), we collected
175 Pasco Plain, eleven St. Johns Plain, one of plain, micaceous paste, seven
Gritty Plain, and two typical Safety Harbor period sherds. One of the latter (Fig.
3, top row, second) is Pinellas Incised, Sub-type B, with a Lake Jackson, Sub-
type B, lug (Griffin, 1950, p. 106). The other (Fig. 3, top tow, first), tempered
with finely ground limestone, has a well smoothed surface into which three lines
have been deeply incised parallel to the rim.
These two sherds indicate Indians were present at the Crystal River site dur-
ing Safety Harbor times. The fact they were found on the eastern portion of the
midden, towards which the ramp and walkway lead from Temple Mound A, would
seem to be significant. This mound may have been built during a Weeden Island-
Safety Harbor transitional or an early Safety Harbor period.
DESCRIPTION OF SPECIMENS
Specimens are listed in the tables and, in part, illustrated in Figs. 4 to 6.
Pottery classification follows that published by Willey (1949a). Supplementary
Two subdivisions of Dunns Creek Red, having the same vertical distribution,
were encountered. In one, the surface appeared mottled, as much of the paint had
rubbed off or weathered away. In the other, the surface, after painting, had been
burnished, with the result the paint was bright and well preserved.
One Dunns Creek Red sherd (Fig. 5, bottom row left) is shaped like an ex-
tended handle. It came from Test II between depths of eight and twelve inches,
and so is relatively late Weeden Island. Another sherd (Fig. 5, bottom row, right),
of St. Johns Plain, exhibits a flattened area adjacent to a concave area, some-
what like the form of a modern plate. It was found in Test II between depths of
thirty-two and thirty-six inches, or in the transitional zone between Santa Rosa-
Swift Creek and Weeden Island.
Two sherds have been listed as Crystal River Incised (Fig. 3, bottom row,
third and fourth sherds) although incisions are not deep nor punctations large.
The latter are hemiconical, like those of Santa Rosa Punctated. Both sherds
have well smoothed and nearly polished, exterior surfaces, fractured edges of
incisions, and contain minor inclusions of mica.
One Franklin Plain rim sherd (Fig. 3, fourth row, right) is thin, one-eighth of
an inch thick, and adorned by small grooves impressed radially into the lip. It
has about five grooves to the inch. The other sherd of this type (Fig. 3, fourth
row, second) is heavier, one-quarter of an inch thick, with a slightly inturned rim,
thinned to one eighth of an inch thick at the lip. The lip has wide notches, spaced
about three to the inch. Neither of these sherds contains any mica.
One of the micaceous-paste complicated-stamped sherds (Fig. 3, second row,
second sherd) has an undecorated area below the lip which is sharply out-turned.
It is worthy of note that this sherd, which is of the Late Variety of Swift Creek
Complicated Stamped, was the shallowest complicated stamped sherd (Test II,
twenty-four to twenty-eight inches) and the only one found in the Weeden Island
zone. This agrees closely with Willey's findings on the Florida northwest coast.
One Swift Creek Complicated Stamped sherd (Fig. 3, bottom row, last sherd)
is noteworthy because of the excellence of the stamped impression. It came from
the next to the bottom level (72-78 inches) of Test II.
Three pottery discs (Fig. 5, second row), one with a rubbed periphery, came
from Test II between depths of four and twelve inches. They are suggestive of
the Fort Walton period as found at the Lake Jackson site (Griffin, 1950, p. 103).
Being formed of the first few spiral turnings of vessels, they must be considered
doubtful artifacts, with probablyno value here as time markers.
Projectile points and stone knives have been covered in an earlier section of
this report. Worked chert refers to fragments too incomplete for classification.
The piece of polished stone (Fig. 5, third row, third) is of igneous origin and
may have come from a celt. The grindstone fragment (Fig. 5, third row, first) is a
piece of laminated limey-sandstone. One flat surface is smoothed as if it had
been used as a whetstone. The largest piece of mica is illustrated (Fig. 5, third
Typology of artifacts of shell and of bone will be found in Goggin and Sommer
(1949), Willey (1949a,b), and Rouse (1951). It should be noted that what we list
as Busycon gouges Willey terms "celts, concave surface variety." Additional
comments on shell and bone artifacts follow.
Fragments of three plummet-like pendants were found. One (Fig. 5, top row,
second), made of the columella of a shell, is not grooved. Another (Fig. 5, top
row, ftrst), also of shell, has been notched and broken off at one end. The third
(Fig. 5, top row, third) is grooved and made of fishbone. Provenience of these
three fragments suggests that plummet-like pendants may have been common in
late Santa Rosa-Swift Creek times.
Several of the objects classified as bone pins are represented only by short
lengths. Some of them, therefore, may be parts of awls or other pointed bone
Four double-pointed bone artifacts have been called grooved pins (Fig. 6,
top row, first through fourth). All have a shallow groove, presumedly for suspen-
sicn, one-half inch to three quarters of an inch from the blunter end. This end
looks as if it had been lightly pounded. The other, more gradually tapering end,
is present in only two cases and does not show signs of wear.
Two flattish pieces of worked bone (Fig. 6, middle row) are listed in Table
5. The narrower end of the shorter one broke off after manufacture. In the case
of the other, this end was grooved and intentionally broken off. The shorter is the
flatter of the two, especially at the wider end.
An interesting bone adze socket, found by Smith between depths of thirty-eight
and forty-two inches in the Florida State University test, is illustrated in Fig. 2.
The specimen has been damaged by heat and is partially calcined. It was care-
fully made from a large bone, possibly a manatee rib, and all exterior surfaces
exhibit fine striations of the type made when bone is scraped with a shark's tooth.
As will be noticed in the illustration, a deep hole has been made in the flat
end, and a shallower hole appears in the other end. The top has a raised portion
towards the curved end, which terminates in a curved edge. Unfortunately, two-
thirds of the lower surface is missing. Curvature of the remaining portion suggests
it also may have had a raised area like that of the top. If so, its edge was dis-
This specimen, undoubtedly from a composite tool, was probably part of an
adze used for woodworking. A bent handle was fitted into the larger hole and a
stone tip into the smaller. Similar artifacts have been found at the Belle Glade,
Surfside, Opa Locka I, and Key Marco sites (Willey, 1949b, Pls. 8, a, b; 15, b)
as well as in the Indian River area (Rouse, 1951, pp. 164, 255). The Crystal
River specimen is better or more carefully made than the ones just referred to.
Smith informs me that a socket nearly identical to the one from Crystal River was
found some years ago in a shell midden near Sanford by H. James Gut of that city.
Occurrence of food bones has been indicated by depths in Tables 4 and 5.
Those of deer only slightly exceeded in number those of turtle, while either were
about three times as abundant as bones of fish. Bird bones, while present, were
extremely rare. Two crab claws were found. Unidentified bones include jaws of
three small animals.
Shell fish remains from the tests are dominately those of oysters. Next in
quantity are those of the crown conch (Melongena sp.) followed by other shell
fish of the area including clams. Only one example of Polinices duplicate, Say,
COMPARISONS AND CONCLUSIONS
Stratigraphic tests at the Crystal River site have clearly defined two periods.
The earlier one, on a Santa Rosa-Swift Creek time level, is characterized by
large amounts of Pasco Plain pottery together with substantially smaller amounts
of St. Johns Plain and Gritty Plain sherds. Associated decorated sherds, possi-
bly trade sherds, include those of the Deptford and Santa Rosa-Swift Creek com-
plexes. Deep provenience of two sherds of Crystal River Incised supports Willey's
contention that this is a pre-Weeden pottery type.
Also typical of Santa Rosa-Swift Creek times at the site are relatively broad,
expanding stemmed and side-notched projectile points, and thick, lanceolate-
shaped stone knives. Late in this period are found plummets and mica. Melong-
ena hammers are common tools in both this and the succeeding period.
Levels in the tests, above a depth of three feet, we have designated as
Weeden Island I, because of the presence of appreciable quantities of Dunns
Creek Red, a few Carrabelle-like sherds, Weeden Island-like rims, and a Late
Variety, Swift Creek Complicated Stamped sherd. No Deptford sherds were found
in this zone, but Pasco Plain, St. Johns Plain, and Gritty Plain were encountered
at about the same frequencies as before. Also typical of this period are Busycon
gouges and small projectile points. The latter are variously arranged for halting
and, by their size, suggest the introduction of the bow and arrow. Mica is still
The uppermost level, with a St. Johns Check Stamped sherd, should mark a
Weeden Island II period. The final period, Safety Harbor, is definitely represented
by only two sherds from the surface of the eastern part of the shell midden. Temple
Mound A, whose ramp points towards this area, must be at least as late as Weeden
Island II in time of construction, as two St. Johns Check Stamped sherds were
found there. In agreement with other sites on the Florida central Gulf Coast, this
temple mound may have been built during the early part of the Safety Harbor period.
Similarly, we suggested the burial mound complex was used over the same
archaeological periods: Santa Rosa-Swift Creek, Weeden Island I, and Weeden
Island II, and perhaps, Safety Harbor.
Occupation of the site during the Safety Harbor period was not intensive, to
judge from the small amount of material attributable to that period. It is reason-
able to believe the Crystal River site may have functioned as a ceremonial center
for the surrounding area. Farming at or adjacent to the site would seem to be im-
possible, but farmers could have come to the site for ceremonies.
Evidence of Indians in the vicinity during Safety Harbor times has been found
at two other sites on Crystal River. Two miles to the west of the Crystal River
site, where that river meets the Salt, William Z. Harmon of the Florida Park Serv-
ice collected many Pasco Plain, one cord-marked, two St. Johns Check Stamped,
and two Lake Jackson Plain sherds. The latter are tempered with quartz sand and
have pinched rims (Lake Jackson Plain, Sub-type A, Griffin, 1950, Fig. 37, 19).
About two miles east of the Crystal River site, on Buzzard's Island, F. G.
Rainey excavated many bundle burials and pottery of the Safety Harbor period,
including Pinellas Plain and Incised (Willey, 1949a, pp. 323-4), Lake Jackson
Plain, Sub-types A and C (Griffin, 1950, p. 106), St. Johns Check Stamped, and
cob-marked sherds (Rainey, 1935, frontispiece). Like Crystal River, neither of
these sites is located in a place where agriculture would be feasible.
Negative-painted vessels, excavated by Moore from the Crystal River burial
mound, have received considerable comment in the literature (Willey and Phillips,
1944; Willey, 1948b). Our tests did not produce data to resolve this problem. It
may be pointed out, however, that painting as a trait appears in the tests at a re-
latively constant depth. It would seem likely that negative-painted vessels should
not be found until some time after the introduction of painting. Late Weeden Island
I would seem to be the earliest logical time for such vessels. There is no reason,
as far as data from the Crystal River site are concerned, why they could not be
late Weeden Island II or even early Safety Harbor in date.
Several vessels illustrated by Moore from the Crystal River burial mound were
built with podal supports. No similar supports were found in any of the tests. This
fact cannot be given undue significance due to the large size of the site. How-
ever, if such supports are typical of the Deptford and early Santa Rosa-Swift Creek
periods, and if the chronology we have presented is reasonably correct, this ab-
sence would be understandable.
Four stratigraphic sections on the central Gulf Coast of Florida are available
for comparison with the ones from Crystal River. The earliest, published by S. T.
Walker in 1883, covers his observations on a twelve and a one-half foot road-cut
through a shell midden at Cedar Keys, about thirty miles northwest of Crystal
River. Walker defined four ceramic periods. While precise canparisons are im-
possible, his sequence is similar to that accepted today and may be interpreted
as Santa Rosa-Swift Creek, Weeden Island, and Safety Harbor.
Another test was made at Shired Island, nearly fifty miles northwest of Crystal
River, by a University of Florida party under John M. Goggin. In lower levels
Goggin found fiber-tempered pottery associated with St. Johns Plain, Pasco Plain,
Gritty Plain, and Deptford complex sherds. This seems to represent a period
earlier than any encountered at Crystal River. At slightly higher levels fiber-tem-
pered pottery was not found, but Swift Creek Complicated Stamped sherds were
included (Goggin, n.d.). This would seem to be the equivalent of the lower zone
at Crystal River, with the same mixture of Pasco Plain, Swift Creek Complicated,
and Deptford sherds. In agreement with its more northern location, the Deptford
complex seems to have been stronger and earlier at Shired Island than at Crystal
Tests at Johns Island, thirteen miles south of Crystal River, differed consider-
ably from those at Crystal River in types of projectile points and in the inclusion
of much more decorated pottery (Bullen and Bullen, 1950). The lower zone (Johns
Island I) may be disregarded as representing a period earlier than any found at
Crystal River, but it should be noted that Pasco Plain, St. Johns Plain and a few
Gritty Plain (plus Pasco Incised and St. Johns Incised) were dominant. The next
zone (Johns Island II) contained some simple stamped and cord-marked sherds as
did the lower zone at Crystal River, with which it may be presumed to equate.
Swift Creek Complicated Stamped sherds were not present, but Crooked River Com-
plicated Stamped (Early Variety) sherds were.
Three Unique Punctated sherds of micaceous paste (Bullen and Bullen, 1950,
Fig. 17, s-t) from the lower part of this zone are available for comparison with the
two Crystal River Incised sherds from the tests at Crystal River. The only common
features between these two groups of sherds is that both contain mica and exhibit
punctations, and even the punctations are differently executed. Both groups, how-
ever, show that the ideas which produced the types, Weeden Island Incised and
Weeden Island Punctated, were present in this area in pre-Weeden Island times.
Slightly higher zones at both sites produced Carrabelle or Carrabelle-like In-
cised and Punctated sherds and may be considered Weeden Island I in time al-
though Dunns Creek Red was not found at Johns Island. The top twelve inches
at Johns Island contained Weeden Island II pottery types including Wakulla and
St. Johns Check Stamped pottery and so would equate with the top four inches of
Test I at Crystal River.
While these two sites differ sharply in their ceramic content, and completely
in their lithic component, there are sufficient similarities to indicate them to have
had, in a broad sense, similar ceramic histories. That decorated pottery was so
common at Johns Island and so rare at Crystal River, has cultural connotations
which are difficult to grasp.
A fourth stratigraphic test was made in the fifteen-foot thick shell midden at
Terra Ceia, ninety to 100 miles south of Crystal River and on the other side of
Tanpa Bay (Bullen, 1951a). In certain respects this test is closer to the ones
at Crystal River than are the others mentioned earlier. These agreements consist
in the presence of shell hammers in all levels, the nearly complete absence of
decorated pottery, and the sudden appearance of Dunns Creek Red (Biscayne Red)
at relatively shallow depths (thirty-four inches at Crystal River and thirty-six to
forty-two inches at Terra Ceia). At both sites, using different arguments, this
red-painted pottery has been taken to represent a Weeden Island time period.
In spite of the relatively large amount of work done recently on the central
gulf coast of Florida, the archaeological situation is far from being clear. Stra-
tigraphic tests in some sites produce a large amount of decorated pottery, while
at others the reverse is true. Shell tools seem tobe abundant in middens where
pottery is undecorated and rare when decorated pottery is abundant.
In the area around Crystal River, limestone was the preferred material for
temper from fiber-tempered to Safety Harbor times. To the north and to the south,
sand was used as temper, but the use of limestone extended southward below
Tampa Bay as a secondary tempering material in late Santa Rosa-Swift Creek
(Perico Island or possibly Weeden Island I) and in Weeden Island times. Dept-
ford pottery types are not the earliest post-fiber-tempered decorated pottery.
In spite of site to site variation, major periods which we call Santa Rosa-
Swift Creek (Perico Island to the south), Weeden Island, and Safety Harbor may
be abstracted. In this continuum Crystal River may be viewed as a site occupied
over several archaeological periods and for a long period of time, probably from
about the time of Christ to around 1600 A.D. (Goggin, 1950, p. 10).
During much of this long period of time, inhabitants of the Crystal River site
used essentially the same day to day industrial products to judge from the conti-
nuity in the manufacture of Melongena hammers and Pasco Plain vessels. Deco-
rated pottery, of which there seems to have been only a small amount, was more
or less reserved for the dead, who were well taken care of with grave goods.
With sufficient food available, they had time for ceremonial life and decora-
tive artifacts. Inhabitants of the Crystal River site seem to have been "conserva-
tive royalists" who did not adopt every passing fancy in pottery decoration. In
Safety Harbor times the site still functioned as a civil and religious center al-
though many of the people, if they practiced agriculture, must have lived else-
1859. Notes on the Florida Peninsula, its Literary History, Indian Tribes
and Antiquities. Philadelphia.
1867. "Artificial Shell Deposits of the United States." Annual Report of
the Smithsonian Institution for 1866, pp. 356-8. Washington.
Bullen, Ripley P.
1950. "An Archeological Survey of the Chattahoochee River Valley in
Florida." Journal of the Washington Academy of Sciences, Vol. 40,
No. 4, pp. 101-125, Washington.
1951a. "The Terra Ceia Site, Manatee County, Florida." Publication No. 3,
Florida Anthropological Society. Gainesville.
1951b. "The Enigmatic Crystal River Site." American Antiquity, Vol. 17,
No. 2, pp. 142-143, Salt Lake City.
Bullen, Adelaide K., and Ripley r .
1950. "The Johns Island Site, Hernando County, Florida." American
Antiquity, Vol. 16, No. 1, pp. 23-45. Menasha.
Fairbanks, Charles S.
1949. "A General Survey of Southeastern Prehistory." The Florida Indian
and His Neighbors, J. W. Griffin, editor, pp. 55-75. Winter Park.
Goggin, John M.
1947. "A Preliminary Definition of Archaeological Areas and Periods in
Florida." American Antiquity, Vol. 13, No. 2, pp. 114-127. Menasha.
1948. "A Revised Temporal Chart of Florida Archeology." The Florida
Anthropologist, Vol. 1, Nos. 3-4, pp. 57-60. Gainesville.
1949. "Cultural Traditions in Florida Prehistory." The Florida Indian and
His Neighbors, J. W. Griffin, editor, pp. 13-44, Winter Park.
1950. "Florida Archeology-1950." The Florida Anthropologist, Vol. 3,
Nos. 1-2, pp. 9-20. Gainesville.
n.d. Report on Test I, Shired Island, Dixie County. (ms. in progress).
Goggin, John M., and Frank H. Sommer, III
1949. "Excavations on Upper Matecumbe Key, Florida." Yale University
Publications in Anthropology, No. 42. New Haven.
Greenman, Emerson F.
1938. "Hopewellian Traits in Florida." American Antiquity, Vol. 3, No. 4,
pp. 327-332. Menasha.
Griffin, James B.
1946. "Cultural Change and Continuity in Eastern United States Archae-
ology." Man in Northeastern North America, Frederick Johnson,
editor. Papers of the Robert S. Peabody Foundation for Archaeology,
Vol. 3, pp. 37-95. Andover.
1949. "Meso-America and the Southeast: A Commentary." The Florida
Indian and His Neighbors, J. W. Griffin, editor, pp. 77-99. Winter
Griffin, John W.
1950. "Test Excavations at the Lake Jackson Site." American Antiquity,
Vol. 16, No. 2, pp. 99-112. Menasha.
1952. "Prehistoric Florida: A Review." Archeology of Eastern United
States, James B. Griffin, editor, pp. 322-334. Chicago.
Griffin, John W. and Ripley P. Bullen.
1950. "The Safety Harbor Site, Pinellas County, Florida." Publication
No. 2, Florida Anthropological Society. Gainesville.
Moore, Cla'ence B.
1903. "Certain Aboriginal Mounds of the Florida Central West-Coast."
Journal of the Academy of Natural Sciences of Philadelphia, Vol.
12, Pt. 2, pp. 362-438. Philadelphia.
1907. "Crystal River Revisited." Journal of the Academy of Natural
Sciences of Philadelphia, Vol. 13, Pt. 3, pp. 406-425. Philadelphia.
1918. "The Northwestern Florida Coast Revisited." Journal of the Acad-
emy of Natural Sciences of Philadelphia, Vol. 16, Pt. 4, pp. 514-518.
Rainey, F. G.
1935. "An Indian Burial Place at Crystal River Florida." The Florida
Historical Society Quarterly, Vol. 13, No. 4, pp. 185-192. Tallahassee.
1949. "The Southeast and the East Indies." The Florida Indian and His
Neighbors, J. W. Griffin, editor, pp. 117-137. Winter Park.
1951. "A Survey of Indian River Archeology, Florida." Yale University
Publications in Anthropology, No. 44. New Haven.
Shetrone, Henry C.
1931. The Mound Builders. New York.
Smith, Hale G.
1951. "Crystal River, Revisited, Revisited, Revisited." American Antiq-
uity, Vol. 17, No. 2, pp. 143-144. Salt Lake City.
Walker, S. T.
1883. "The Aborigines of Florida." Annual Report of the Smithsonian
Institution for 1881, pp. 677-680. Washington.
Willey, Gordon R.
1945. "The Weeden Island Culture: A Preliminary Definition." American
Antiquity, Vol. 10, No. 3, pp. 225-254. Menasha.
1948a. "Culture Sequence in the Manatee Region of West Florida." Ameri-
can Antiquity, Vol. 13, No. 3, pp. 210-218. Menasha.
1948b. "The Cultural Context of the Crystal River Negative-Painted Style."
American Antiquity, Vol. 13, No. 4, pp. 325-328. Menasha.
1949a. Archeology of the Floridt Gulf Coast. Smithsonian Miscellaneous
Collections, Vol. 113. Washington.
1949b. "Excavations in Southeast Florida." Yale University Publications
in Anthropology, No. 42. New Haven.
1950. "Crystal River, Florida: A 1949 Visit." The Florida Anthropologist,
Vol. 2, Nos. 3-4, pp. 41-46. Gainesville.
Willey, Gordon R. and Philip Phillips
1944. "Negative-Pained Pottery from Crystal River, Florida." American
Antiquity, Vol. 10, No. 2, pp. 173-185. Menasha.
1875. "Fresh-Water Shell Mounds of the St. Johns River in Florida."
Peabody Academy of Science, Fourth Memoir. Salem.
When the Florida Anthropological Society was organized in 1948 and this
. journal was initiated, it was the intention to publish at quarterly intervals as
soon as the membership and treasury had grown to the point where such a program
codd be supported. The Society still is not so large as it should be, and its
treasury is not so secure as it might be. However, in the opinion of the Society's
officers, we can venture now into a policy of four separate numbers yearly. Wheth-
er this will stimulate reader-membership remains to be seen. But a larger member-
ship, actually, can come only from the effort of present members, as the Presidents
have pointed out so often in our Newsletters.
To maintain under this policy the standard established by previous editors,
we must have more manuscripts from our members and friends of the Society. And
in accordance with the policy expressed at the last two annual meetings, we should
like to have papers in all of the fields of anthropology, insofar as they shall be
relevant to Florida, to the Southeastern States, and to the Caribbean, the region
which we serve.
Present members who wish to interest friends in joining the Society might
note the announcement on page 44iof this issue, which has been reprinted from
the March Newsletter.
The Sixth Annual Meeting of the Florida Anthropological Society will be held
on the campus of the University of Miami in February, 1954.
Names of officers of the Society, who were elected at the Fifth meeting in
Gainesville, appear on the inside front cover.
Space and Time Perspective in Northern St. Johns Archeology, Florida by John
M. Goggin. 148 pp., 12 pls., 9 figs. Yale University Publications in Anthropo-
logy, No. 47, $2. New Haven, 1952.
The St. Johns River may be said to be the birthplace of Florida archaeology,
and the birth may be dated from the work of Wyman between 1867 and 1871. From
this area, too, have come several large volumes by C. B. Moore. Other workers,
earlier and later, have contributed to our knowledge, but the striking fact that
emerges from a comprehensive survey, such as the one under review, is that the
area has received so little modern stratigraphic attention.
As defined by Goggin, the Northern St. Johns Area runs from Lake Harney
northward to the Florida-Georgia line and includes drainages to the west of the
river, as well as a strip along the Atlantic coast. Data from 432 sites are con-
sidered, and it should be emphasized at once that these data are unequal in
quantity and quality. This is not the fault of the author; his job was to organize
whatever information was available into a time-space framework.
Goggin recognizes four sub-areas, based on both geography and cultural
materials, and these are used in his analysis. Temporally, the materials range
from the preceramic Mt. Taylor Period to the historic Seminole occupation. This
yields six major sequential periods, three-of which are further subdivided. Goggin
(p. 79) notes the Seminole as the first demonstrable interlopers in this sequence,
which, therefore, provides a long range of indigenous cultural development under
the influence of extra-areal contacts.
The body of the report is introduced by a section which treats the geograph-
ical, ethnological and historical background of the area, and outlines the history
of archaeological research. This is followed by a discussion of the cultural
sequence, as derived from a study of the literature and available collections,
and the cultural traditions, of which there are five Paleo-Indian, Archaic, St.
Johns, Spanish-Indian, and Seminole. This is followed by a general discussion
containing an historical reconstruction, a statement of problems, and a summary.
Appendix A lists the known sites by sub-area, and indicates the periods be-
lieved to be represented at each. Appendix B contains descriptions of the pottery
types and other artifacts. As we have come to expect in Goggin's work, a gen-
erous bibliography is provided.
It is not possible to check the period placement of sites on the basis of the
published material. It would probably have been both uneconomical and impracti-
cal to include the necessary information in a report of this kind, particularly in
view of the fragmentary information available on many of the sites. Nevertheless,
some readers will find the absence of quantitative data disturbing.
Naturally, there are points at which the reviewer does not find himself in full
Naturally, there are points at which the reviewer does not find himself in
full agreement with the author. I am more inclined to question the validity of
the non-ceramic level of Nelson at Oak Hill than is Goggin (pp. 38-39). I cannot
agree that coquinas "could hardly he an important food supply" (p. 67). The
vast size of such a midden as the Cotten site seems to me to argue that coquinas
were an important food supply at certain times and in certain places. The points
of disagreement are, however, relatively minor, and many of them may be semantic.
There are several places where a definite word such as "demonstrated" is used
when I would have preferred one such as "indicated" or "postulated".
Goggin has pointed out to the reviewer that "St. Johns la" in Fig. 2 should
read "St. Johns lb". There are several, but not many, typographical errors. On
page 17, "Mosquito Inlet" should probably read "Mosquito Lagoon" (now
Indian River North). Ampullaria is consistently used in the text in place of the
more modern designation Pomacea. These errors are really inconsequential.
Space and Time Perspective in Northern St. Johns Archeology will long stand
as a highly important reference on the area. The wealth of material summarized,
and the inclusiveness of the bibliographical coverage, will assure this. More
importantly, the book provides a framework which maybe regarded as a series of
hypotheses tobe tested in the future, and suggest, both explicitly and implicitly,
problems to be investigated in the field. In this respect it is undeniably the
most significant single work which has appeared on this important area.
John W. Griffin
CONTRIBUTORS TO THIS ISSUE
D. D. Laxson, who reports here on his excavation of a Dade County midden,
was elected an executive committeeman of this Society at the Gainesville meet-
ings. He lives in Hialeah, and is associated with Eastern Air Lines in Miami.
Ripley P. Bullen is curator of social sciences of the Florida State Museum,
Gainesville, and treasurer of the Society. His work at the perennially-visited
Crystal River site was done when he was assistant archaeologist of the Florida
Board of Parks and Historic Memorials.
John W. Griffin, Gainesville, who reviews John M. Goggin's St. Johns mono-
graph, has served as president and secretary of the Florida Anthropological
Society and editor of this journal.
A LIMITED PUBLICATION OFFER TO MEMBERS
For a limited time (until July 1, 1953) it will be possible for the members of
The Florida Anthropological Society to purchase "back numbers" of The Florida
Anthropologist at a saving of fifty cents each. During this period the price of
each double number will be fifty cents per copy (exception: Vol. I, Nos. 1-2, which
will continue to sell for $1.00 to members). If you wish to take advantage of this
offer to round out your series, send your order to:
Ripley P. Bullen, Treasurer
Florida Anthropological Society
Florida State Museum
Vol. I Nos. 1-2 $1.00
Vol. I Nos. 3-4 .50
Vol. II Nos. 1-2 .50
Vol. II Nos. 3-4 .50
Vol. III Nos. 1-2 .50
Vol. III Nos. 3-4 .50
Vol. IV Nos. 1-2 .50
Vol. IV Nos. 3-4 .50
Vol. V Nos. 1-2 .50
Vol. V Nos. 3-4 .50
FLORIDA ANTHROPOLOGICAL SOCIETY
First Vice President:
Second Vice President:
Frederick W. Sleight, Mount Dora
H. James Gut, Sanford
Raymond F. Bellamy, Tallahassee
James A. Gavan, Orange Park
Ripley P. Bullen, Gainesville
Robert Anderson, Tallahassee
D. D. Laxson, Hialeah
Leigh M. Pearsall, Melrose
Membership in the Florida Anthropological Society is open to everyone
interested in its aims. Dues are $3.00 per year. Members receive the Florida
Anthropologist, the Newsletter, and other publications of the Society. Appli-
cations should be sent to the Treasurer, who should be addressed also con-
cerning receipt of publications. His address is 103 Seagle Building, Gaines-
General inquiries concerning the Society should be addressed to the
Manuscripts and publications for review should be sent to the Editor, at
the Department of Anthropology and Archaeology, Florida State University,
Address items for the Newsletter to the President, Box 94, Mount Dora.