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
MIF A I EMER I EAllN 1l llmM!
THE FLORIDA ANTHROPOLOGIST
Crystal River, Florida: A 1949 Visit
---- --- ---- --- ----------
Gordon R. Iilley 41
An Indian Stone Saw
____----------- --------------- ---- .. Armistead 47
The Woodward Site
--------------- -------------- Ripley P. Bullen 49
Cultural Occupation at Goodland Point, Florida
-___------------------------- ohn M. Goggin 65
Notes on the Archeology of Useppa Island
-------------------------ohn W. Griffin 92
Contributors to this issue------------------------- 94
Published at the University of Florida
s3Y'\ ~ 1 n~S
Fig. 12. Map of the Crystal River Site. (Reproduced from Moore, 1903, Fig. 13.)
CRYSTAL RIVER, FLORIDA: A 1949 Visit
Gordon R. Willey
The archeologically famous Crystal River site attracted the attention of
scientists at least 80 years ago. Brinton, in a Smithsonian Report published in
1867 (Brinton, 1867), remarked upon the shell piles and the sand mound located
four miles up the Crystal on its north bank. Brinton's writings were based upon
the observations of F. L. Dancy, onetime State Geologist of Florida. Shortly
after the turn of the century, Clarence B. Moore visited the Crystal River mound
site and conducted two separate seasons of excavations there (Moore, 1903,
1907). It is from Moore's work that the site and its significance in Florida pre-
history are known. Since Moore's investigations, there have been no reported
surveys or excavations at Crystal River; and, judging from the appearance of
the site as we saw it in July of 1949, it does not seem to have been seriously
disturbed by unreported diggings.
Moore's first description of the site (Moore, 1903: 379-382) is still our
best and only one. It would appear to be fairly accurate. This is important as
Moore's account of the types of sand and shellworks which he found at Crystal
River seemed surprising in view of the nature of the ceramics and other archae-
ological materials which he recovered from his excavations there. For the Moore
map (reproduced as Fig. 12) and text show and describe two large shell
mounds; and these are not shell mounds in the sense of great rubbish heaps.
They are described and drawn as purposefully constructed, flat-topped, rectan-
guloid, pyramidal-type mounds with ascending ramps. As far as Mound A is
concerned, the large impressive mound on the river bank, Moore is quite ac-
curate. He has described it as being 28 feet 8 inches in height, oblong in
horizontal section, and with basal diameters of 182 by 100 feet. The summit
plateau of the mound is given as 107 by 50 feet. The graded way or ramp he
measured as 80 feet long and from 14 to 21 feet in width. The ramp ascends
from the level ground on the northeast side of the mound to the very top
(Moore, 1903: 379).
A. J. Waring, Jr., Rufus Nightingale, and I were at the site for a period
of about three hours on the afternoon of July 22, 1949. We took no measure-
ments but our estimates of the height of Mound A ranged from 25 to 35 feet.
Our brief survey confirms the mound form as described by Moore. The summit
platform is exceedingly level although not well squared. The ramp approach
is perfectly preserved. In fact, both Waring and I felt that there was no other
temple-type mound in the Southeast, with the exception of the great mound at
Moundville, Alabama, that has retained such a well-formed ramp feature. Both
mound and ramp seem to be built of a mixture of shell and black earth midden
rather than pure shell, although this cannot be accurately determined as there
is no exposed cut in the mound. Apparently, the mound has never been exca-
vated, at least to any extent. Today, it still stands as the most imposing land-
mark on the otherwise lowlying, tropical riverbank. In July it was covered with
dense bush and grass vegetation. It also supports a stand of large trees.
Unfortunately, we were unable to verify Moore's map drawing and descrip-
tion of the other ramped mound, H. Its location is depicted as about 400 yards
from Mound A in a north-northeasterly direction, but we did not find it. The
whole site area lying back inland from the big mound, A, and the riverbank is
an extremely dense, mucky swamp. My admiration for Moore rises when I
reflect upon the difficulties he must have overcome in mapping the site if he
was confronted with anything like the growth cover which we saw there in
1949. Even a reasonably accurate sketch map under such conditions would
be a signal achievement.
After wandering through the underbrush for the better part of two hours
we finally found the sand burial mound, the centrally located and best known
feature of the site. This mound, which Moore (1903:379 ff.). lettered as F,
had a height of about 10 feet above the surrounding ground. It was circular,
conical, and 70 feet across the base. Our observations verify the Moore de-
scriptions and measurements. The mound, as the result of his excavations, is
somewhat lower now than originally, pitted and uneven on the surface, and
more spread out. Moore's account and map also list a feature, E, which he
discovered to be an annex or apron to Mound F proper. Some evidences of
this are still to be seen although the junction of mound proper and apron is
not sharply perceived. Encircling both mound and apron is a sand enclosure
which Moore called feature C. This circular embankment had a maximum
height of 6 feet and a maximum width of 75 feet according toMoore. We
were unable to trace the embankment or enclosure all the way around Mound
F owing to the dense growth. However, we did locate the embankment at a
point about 25 to 30 yards south of the mound. Arriving at a small clearing in
the trees and bush, we were first of the opinion that we had come across
another smaller sand mound which Moore had not mapped. However, the loca-
tion and distance of this clearing from the mound proper check very closely
with the Moore map. There is now little doubt in my mind but that we were
on a section of the enclosure although we were unable to follow the circle in
either direction. In this clearing which we found, a tree had been uprooted
recently, and in the grey sand of the relatively fresh hole portions of an
adult human skeleton and a number of potsherds were recovered. The presence
of burials in the sand enclosure was noted also by Moore (1907: 425).
The other features which Moore mapped and described (1903), besides the
two flat-topped ramp mounds, the burial mound and apron, and the enclosure,
were a low irregular shell mound about 100 by 150 feet in extent (feature G)
and a long shell ridge (feature B) along the river bank. We were unable to
locate feature G which lies somewhere to the northwest of Mound A, but fea-
ture B was very much in evidence. Moore mapped the shell and midden ridge
B as a long, curving affair connecting to one side of Mound A, at the river's
edge, and then turning inward and eastward to eventually follow the river -
bank. In total length he shows it as over 1000 feet, and this is prob-
ably correct. Today there is a modern home and a caretaker's home built upon
what was evidently a part of the ridge. In many places the ridge can be seen
quite clearly rising two or three feet above the surrounding ground. In most
places Moore shows it to be 100 feet or more in width, and our impressions
are that this is correct. It is composed of shells and rich black midden. It
undoubtedly represents the refuse remains of prehistoric houses or occupation.
The Crystal River site, then, is much as Moore reported it. Certainly tLe
feature of flat-topped mounds with ramps is verifiable. In view of the ceramic
periods represented in the burial mound, the presence of the substructure or
temple-type mounds here is most surprising. I have, previously, analyzed the
Moore collections from the burial mound and apron (features F and E) (Willey,
1948); and I have concluded from these analyses that the burial ceramics
from Crystal River are predominantly of the Santa Rosa-Swift Creek Period.
This period has been equated with the Marksville-Hopewellian-Early Swift
Creek horizon in the East and Southeast. A more recent and thorough study
(Willey, 1949) of the Crystal River finds corroborates this Santa Rosa-Swift
Creek Period dating for the burial mound proper (feature F) but places the
collection from the apron (feature E) as mixed Santa Rosa-Swift Creek and
Weeden Island I. The surface collection of 1949, from around the burial mound
and from the base of the uprooted tree on the enclosure, also dates as Weeden
Island I, the absence of the types Wakulla Check Stamped or St. Johns Check
Stamped precluding a Weeden Island II date (Table 1). Thus both periods,
Santa Rosa-Swift Creek and Weeden Island I seem to be represented at the site.
In either case, the general cross-dating of these periods is believed to be too
early for the temple mound trait in the Southeast. What, in view of all this, is
the explanation of the flat-topped mounds at Crystal River?
There are two interpretations. The first is that the temple mounds at the
site represent constructions of a later period, presumably of the Ft. Walton
or Safety Harbor time level. Pottery of these periods has not yet been found
at the site, although this might be due to lack of sufficient exploration. The
second interpretation would be that the substructure mounds were constructed
by the peoples who occupied the site during the Santa Rosa-Swift Creek or
Weeden Island I Periods. This may not be as much out of line as it sounds.
Krieger's East Texas work has offered evidence that the temple mound was
present in an early Caddoan-type culture in that area on a time level coeval
Table 1. Tabulation of Pottery Types from Surface Collection, Crystal
Weeden Island Series:
Weeden Island Plain ........................... 1
Weeden Island Punctated ....................... 1
Carrabelle Incised ......... ............. ... 1
Thomas Simple Stamped ........................... 6
Tampa Complicated Stamped ....................... 1
St. Johns Series:
St. Johns Plain ............................. 11
DunnsCreekRed ............................ 5
St. Johns Incised (?) ... ....................... 1
Pasco Plain ....... ......................... 8
PascoRed ................ .................. 1
Plain Red (sand-tempered) ............... ..... 2
Smooth Plain (sand-tempered) ... ........ ...... 1
Residual Plain (sand-tempered). ................... 29
Deptford Bold or Gulf Check Stamped (?) ............. 1
Total Sherds ....... .69
a. Collection made by Waring, Nightingale, and Willey, July 22, 1949. Sherds
were gathered from the top of the burial mound (feature F) and from the uprooted
tree location on the south side of the enclosure (feature C). Specimens are now
in the.U. S. National Museum, Washington, D.C.
with the Louisiana Marksville or Troyville Periods (Newell and Krieger,
1949). Temple mounds are also reported for the Troyville culture. Very recently,
W. H. Sears (personal communication, August 1949) has excavated a burial
mound at the big Kolomoki site in southwest Georgia which is pure Weeden
Island I. Sears equates this burial mound with the bulk of the occupation
at Kolomoki and also with the huge flat-topped pyramidal at that site. In this
way it is possible to follow a linkage of evidence across the lower Southeast
in which the temple mound idea could be seen as advancing from west to east,
from Texas to Florida, on a pre-Middle Mississippian time horizon. The case
for this is not demonstrated, but it is too strong to ignore. Is Crystal River an
eastern extension of such an early diffusion of the temple mound?
In connection with the spread of the temple mound complex we cannot
avoid wondering if maize horticulture was attached to it. There is no direct
evidence for maize with either Santa Rosa-Swift Creek or Weeden Island any-
where in Florida, although it has been assumed that in these periods the
Indians were maize horticulturists. After a visit to Crystal River both Waring
and I were struck by the fact that none of the land in the immediate vicinity
of the site appears to be suitable for maize cultivation. As mentioned, most
of the immediate area of the site, a-square quarter of a mile or more, was low
and mucky. In many places we sank knee-deep in black swamp soil while we
were trying to make our way across the site. The ecological set-up does not
appear to be one in which primitive horticulture would have been successful.
This, together with the bountiful supply of fish in the wide and clear Crystal
River, argues for a primary dependence upon wild foods. But this does not
jibe with our usual concepts of temple mound cultures in the Southeastern
The Crystal River site is an ideal spot for intensive research at our pres-
ent stage of knowledge in Floridian.and Southeastern archaeology. First, a
new map of the site should be made. As we have stated, the Moore map seems
to be correct as to the features we were able to check; but even if it is perfect-
ly accurate it is not sufficiently detailed. Secondly, extensive stratigraphic
pitting and trenching should be carried out in the midden ridge designated on
the Moore map as feature B. This would appear to be the occupation area, and
its contents may aid in the dating of the flat-topped Mound A. Third, a test
should be made in the shell mound or ridge listed by Moore as feature G. All
of these operations could be carried out with limited funds and small crews.
If larger scale, longtime work were feasible, the platform mounds, A and H,
should be excavated in part. Finally, I would hazard the guess that more work
-on the burial mound, apron, and enclosure complex (F, E, and C) might repay
the effort. There have been many instances where Moore has claimed to have
"completely demolished" a sand burial mound and where later investigation
has revealed considerable undisturbed portions.
Such a program of field study at Crystal River would be of great value.
The Santa Rosa-Swift Creek Period is only barely outlined; its relationships
to earlier and later periods in Florida, its connections with the archaeological
cultures of areas outside Florida, and its own internal content are all obscure.
For example, we do not know, as yet, whether the IIopewellian elements in
Santa Rosa-Swift Creek and at Crystal River preceded or followed the appear-
ance of the early Swift Creek stamped pottery in Gulf Florida. Stratigraphic
digging at Crystal River might help resolve this. Intensive work at the site
may throw more light on the puzzling early appearance of the negative painted
technique on pottery which occurs there. And it might also clear up the matter
of the presence of the temple-type mounds and their proper relative dating.
BRINTON, D. G.
1867 "Artificial Shell Deposits of the United States." Smithsonian
Institution Annual Report for 1866, pp. 356-358. Washington.
MOORE, CLARENCE B.
1903. "Certain Aboriginal Mounds of the Florida Central West Coast."
Journal, Academy Natural Sciences, Vol. 12, n.s., pp. 361-438.
1907. "Crystal River Revisited." Journal, Academy Natural Sciences,
Vol. 13, n.s., pp. 406-425. Philadelphia.
NEWELL, H. PERRY and
KRIEGER, A. D.
1949. "The George C. Davis Site, Cherokee County, Texas." Memoir,
Society for American Archeology, No. 5. Menasha, Wise.
WILLEY, GORDON R.
1948. "The Cultural Context of the Crystal River Negative painted
Style." American Antiquity, Vol. 13, pp. 325-328. Menasha, Wise.
1949. "Archaeology of the Florida Gulf Coast." Smithsonian Miscel-
laneous Collections, Vol. 113. Washington.
Bureau of American Ethnology
AN INDIAN STONE SAW
W. J. Armistead
About thirty years ago, when we were students in high school, John N.
Harrison, Jr., Harold Wolf, and myself frequently went on our bicycles from
our homes in Tampa to Phillippi Point, the location of the famous Safety
Harbor Indian site. This was a long trip in those days before the construction
of the present Courtney Campbell parkway connecting Tampa with Clearwater
and Safety Harbor.
Much of our time at the site was spent looking for Indian relics. We had
full permission to go anywhere we wished on the premises and to collect all
the artifacts we could find but we were not allowed to dig any holes. Fortun-
ately for us the land was under cultivation so that specimens were continually
being turned up. Erosion of the northern side of the point had already started
and we were able to find many artifacts on the advancing beach.
On one such expedition I was lucky to find on this beach, within an area
10 feet across, the three specimens illustrated in Figure 13. While these
relics are not particularly impressive or beautiful to look at, they tell a story
which to me is most interesting.
Specimens A and C in Figure 13 are portions of the central spiral or col-
umella of large shells. The central object (Fig. 13,B) is made of native chert
and at first glance appears to be a stone knife. Quite likely the basal projec-
tions were used to fasten this tool to a wooden handle. This specimen differs
from a knife, however, in being relatively thick with blunt edges much worn by
use. This wear extends up onto the sides as will be noticed in the illustration.
It will also be noticed that Figure 13,A has a very deep and wide but some-
what smooth groove which nearly divides it into two pieces. The blunted edge
of the chert tool fits this groove exactly. This leads me to the unescapable
conclusion that this columella was in the process of being cut or sawed with a
tool like the one illustrated. The fit between the tool and the groove is so close
that it is very probable that we have here the actual tool which some Indian
used to saw this particular columella.
As is well known, the columellas of our large shell-fish are extremely hard
and tough. They were used extensively by Indians in the fabrication of tools.
Obviously techniques would have to be developed for cutting this hard shell.
Here we have one method of reducing them, that of sawing.
A B C
Fig. 13. Specimens from the Safety Harbor Site.
The right-hand shell object (Figure 13,C) shows what may be a slightly
later stage in the operation of making tools or other objects of shell columellas.
In this case the sawing has proceeded further and the surplus material removed
Our knowledge of the industrial processes of Indians is, at best, limited. It
is interestingto be able to illustrate here an Indian stone saw, a shell colum-
ella in the process of being sawed to which the saw fits, and a third specimen
on which the sawing is completed and the columella successfully broken in two.
THE WOODWARD SITE
Ripley P. Bullen
The Woodward site, consisting of a mound and adjacent village area, is
located about nine miles south of Gainesville in the Tacoma section of Mic-
anopy, Alachua County, Florida (Fig. 14). This site was brought to the atten-
tion of the Florida Park Service by Dr. John M. Goggin of the University of
S/ LEDWITH LAKE MILES
Fig. 14. Map Showing Location of Woodward Mound.
Florida whose students located it while surveying the Gainesville area. The
northeastern portion of the mound had been removed by the Woodwards for
fill. As they planned to eventually remove the balance of the mound, excava-
tion was conducted by the Florida Park Service to conserve the remaining
data and gain some idea of aboriginal burial habits in central Florida.
We wish to express our appreciation to Mr. and Mrs. Burton M. Woodward,
owners of the site, for permission to do this work and for making our daily
visits so pleasant. We also are indebted to Mr. James I. Bryan of Micanopy,
neighbor of the Woodwards, for many hours of able assistance excavating.
The Woodward site occupies the southwestern portion of fairly high,
sandy land abutting on what was once the northeastern shore of Levy Lake.
A steep bank, about thirty-five feet high, separates it from the old lake bed.
To-day, Levy Lake is considerably smaller than shown on maps. In Indian times
it probably extended to the steep bank beside the site. In times of high water
Levy Lake, as well as Ledwith Lake to the south, may have been connected with
Payne's Prairie. The Arredondo Sheet, U.S.G.S., shows a small pond with ephe-
meral brooks suggestive of such a connection (Fig. 14, near Wacahoota).
The village site borders only a small portion of the old lake shore but extends
northeasterly a substantial distance. It occupies some ten acres of good, well
drained, sandy farm land. A small amount of occupational debris, charcoal, chips,
and sherds, were found scattered over the village area. Concentrations of char-
coal and of sherds suggested locations of houses. Tests did not disclose any In-
dian remains below the base of the plowed zone but pits are, no doubt, present.
Excavations to ascertain post hole patterns of houses were not attempted.
Because of this distribution, we believe the Woodward site represents the
remains of an agricultural community. Fish and birds from the adjacent lake system
and adimals from the surrounding area undoubtedly supplemented farm products
and gave the inhabitants a well balanced diet.
THE BURIAL MOUND
An excavation plan and two profiles of the mound will be found in Figures 15
and 16. Data regarding the burials will be found following this section describing
the mound and its construction.
The mound was comprised of three superimposed zones indicated on the pro-
files as dark gray, mottled.brown, and brown (Fig. 16). The gray zone was com-
posed of sand well mixed with fine particles of charcoal. It contained chips,
worked chert, and occasionally small pieces of charcoal but was otherwise sterile.
The mottled zone consisted of tan, brown, and dark brown sand and loam mixed,
in places with a fair amount of charcoal. It is believed to represent material taken
from the adjacent edge of the village area, partially mixed with subsoil at the
time of excavation, and deposited to form the mound. It contained chips, a few
sherds, three projectile points, a celt, and occasionally claystones.
The superior brown zone is similar in composition and content but uniformly
brown in color. Probably this zone was once mottled but leaching, oxidation,
and the partial formation of a soil profile have eliminated the variations in color.
A superficial loam or dark brown humic deposit has formed in the top few inches
of the superior brown zone.
With these zones in mind, the building of the mound may be reconstructed.
In preparation Indians removed material to a depth of two to three feet below the
EDGE OF PIT 15
EXCAVA- 8 2
LIVE OAK P
\ 0 A /o 7 5A
5W 4W BW 0 I E 3E
Fig. 15. Excavation Plan, Woodward Mound.
present elevation of the surface. They also dug pits near the periphery of this
lowered, uncovered area. (Figs. 15 and 16, Pits C, D, and a third, three feet
southwest of stake C). These pits contained charcoal and may have had a
function in the formation or preparation of the gray zone. Some of the subsoil
removed in this process is believed to be represented by the yellow deposit
shown to the right of the upper profile (Fig. 16). This yellow deposit had all
the appearances, except that of location, of undisturbed subsoil. On the low-
ered surface Indians deposited the dark gray zone shown on the profiles at the
base of the mound. This deposit also filled or lead into the pits mentioned
After the installation of this gray base, it was penetrated by a central
burial pit (Burial 15). A slight rise in the elevation of the top of the gray zone
LIGHT COLORED SAND DARK GRAY
SPROFILE-I FT. NE OF 0 LINE
Fig. 16. Details of Woodward Mound.
on each side of this burial pit suggests that it was dug before the addition of
the mottled zone. The other central burial (Burial 16) was not as deep and its pit
was only dug a short distance into the upper part of the gray deposit. It may or
may not be contemporaneous. No evidence of a burial shaft was found but it is
doubtful if one dug from an elevation within the mottled zone would be noticeable
in this deposit. The same ambiguity applies to a third burial (Burial 2) similarly
located in the upper part of the gray zone.
After the installation of Burial 15 (and possibly after the interment of Burial
16 and 2) the mound was built using materials which formed the mottled brown
zone. How high the mound was built at this time is not certain but a suggestion
of a pause in construction may be noted in the lower profile at about the elevation
of Burial 19 (Fig. 16). These suggestions consist of the burial itself, the charcoal
area and charred log to the left, the fireplace shown to the right, and the changes
in slope of the top of the mottled zone -all at about the same elevation. The fire-
place was a well defined, shallow, circular pit filled with charcoal. Other charcoal
deposits shown in the profiles were not so well defined and may or may not be
The lack of any burials in the lower part of the mottled zone suggests that it
was built up at one time. The rest of the mound, certainly the eastern part of it,
is believed to have been constructed by an accretional process. Inspection of the
excavation plan will indicate that most burials were extended. In all cases except
one, arms, hips, and feet were progressively lower than the head. It seems clear
that these bodies were laid on the sloping surface of the side of the mound, or
in shallow pits scooped out of this surface, and then covered with dirt. The con-
centration of these burials towards the east caused the mound to grow in that
direction. Consequently, the slope there was much more gradual than to the west
(Figs. 15 and 16).
Dirt to cover these burials and part, at least, of that used to form the mound
was dug from a contiguous area southeast of the mound. While a burrow pit is no
longer evident on the surface, test holes showed it to have been in this location.
On the profiles it will be noted thet the dark brown loam becomes thicker to the
left of Stakes "0" and "lW". Ten feet southeast of Stake 1W this "loam" was
24 inches thick, at 18 feet it was 20 inches thick, at 25 feet it was 11 inches, while
at 47 feet it was only 7 inches thick. The last value is about normal for the site.
These tests show the presence of a burrow pit which has gradually become filled.
After abandonment of the site by Indians, it became a farm and a citrus and
pecan grove. About seventy-five years ago the mound was used as a family bur-
ial plot by the owners. These burials were removed some years ago (Fig. 16,
Pit A). These historic burials do not seem to have disturbed any aboriginal
Previously, the mound was tested for treasure. Pit 3 (Fig. 15) shows the
outline of this test. It had nearly vertical sides, penetrated seven feet six
inches, and ended just above the base of the gray zone. This pit was very
clearly defined. A piece of wood, found sloping upward and inward from a
point two feet above the bottom suggested that the sides of this hole may have
been shored. A piece of binding wire and sixteen links of a chain, found in this
hole at a depth of four feet, suggest that dirt was hoisted up by a bucket. This
pit cut off the upper part of Burial 12. A few bones found in the pit (Fig. 15,
Burial 14) are presumed to be from this burial.
The last use of the mound, prior to its partial removal by the Woodwards,
was for the interment of a dog. As this skeleton was within the area of Pit A,
it postdates the removal of the historic burials Fig. 15). All flesh had disappeared
but dirt had not completely surrounded the dog bones and hair was still present.
These remains were in a pit whose shaft showed it to have been dug from the
present surface of the mound.
The horizontal arrangement of burials has been plotted in the excavation
plan (Fig. 15). Skeletal material was in extremely bad condition. Excepting
bundle burials, practically all bones uncovered have been indicated. Examina-
tion of this plan will show that some burials were represented by only a skull
and leg bones (for example note Burial 4). Observations and some measure-
ments couldihave been made on some of the skulls prior to removal from the
ground. Many, however, were crushed or broken by roots and all were more or
less eroded by ground acids.
All adult skulls indicated a brachycephalic (round headed) population who
did not practice cranial deformation. Brief comments on the burials and skele-
tal material follow. Depths given are from the surface to the top of skull or to
other bones in absence of a skull.
Burial 1. Middle aged adult, possibly female. Fair amount of tooth
wear, 2nd lower molar lost in life. Depth 2' 9".
Burial IA. Fragmentary skull. Probably goes with Burial 11 and dis-
placed at time of interment of Burial 1. Depth 2' 9".
Burial 2. Middle aged adult male. Very thick skull, large mastoids,
tooth wear shows poor occlusion. Red powder behind skull
and extending down to right shoulder. Depth 4' estimated,
ankles 16" deeper.
Burial 3. Upper portion removed by Woodwards. Depth, proximal ends
of femurs, 9", ankles 5" lower.
Burial 4. Young adult, sex indeterminate. Moderate tooth wear, shovel-
shaped incisors. Depth 18", ankles 4" lower than base of
Burial 5. Middle-aged adult male. Some suture closure, moderate tooth
wear, shovel-shaped incisors, additional cusp at base of
"shovel" depression of upper medial incisor. Depth 14". Leg
bones probably in unexcavated area.
Burial 5A. Late -i' dle-aged adult male. Fair amount of suture closure,
good oruw ridges, pronounced tooth wear, five lower molars
lost in life. Skull 171mm. long, 147 mm. wide, C.I. 85.9%.
Depth 18", lower leg bone 6" deeper than base of skull.
Burial 6. Young adult, sex indeterminate. Moderate tooth wear, shovel-
shaped incisors, caries in lower second molars. Depth about
30", in part under Burial 3, legs lower than skull.
Burial 7. Late middle-aged female. Moderate tooth wear, some teeth
lost in life. Depth about 2' 8", femurs lower than base of
Burial 8. Bundle or disturbed burial of an adult. Depth about 2' 4".
Burial 9. Bundle burial of middle-aged adult, probably female. First
lower molars lost in life. Long bones found on top of skull.
Bones in pit measuring 18 by 24" across and 8" vertically.
Pit shaft not noted above long bones at depth of about 2' 4".
Burial 10. Crushed skull and one long bone of young baby. No teeth or
jaw bone present but frontal fontanelle closed. Depth 18".
Burial 11. Portion of robust middle-aged adult, probably male. Depth
18". Burial lA is probably skull of this burial.
Burial 12. Portion of adult skeleton, balance cut off by Pit B. Red
powder beside left forearm. Depth 4'.
Burial 13. Skull of middle-aged adult. Some suture closure, moderate
tooth wear. Depth 20". Apparently an isolated skull.
Burial 14. Fragmentary bones in disturbed area. Depth 18 to 28". Prob-
ably part of Burial 12.
Burial 15. Old adult, sex indeterminate. Fair amount of suture closure,
pronounced tooth wear, several molars lost in life. In pit dug
through gray zone, Depth 6' 9", ankles 6" lower than base of
skull. No red powder.
Burial 16. Middle-aged female. Moderate tooth wear, shovel-shaped inci-
sors, 3rd molar and 2nd bicuspid lost in life. Red powder be-
side head, left shoulder, right hip, both thighs, and right ankle.
In pit in uppqr part of gray zone. Depth 6'.
Burial 17. Old adult male. Thick robust skull, very brachycephalic, mod-
erate brow ridges, large mastoids, all lower molars lost in life.
Depth 20". Crossed bones to southeast were 3" higher than
long bones of this burial and presumedly represent remains of
a bundle burial.
Burial 18. Middle-aged to old adult male. Large mastoids, some suture
closure, pronounced tooth wear, shovel-shaped incisors, 1st
lower right molar lost in life, caries in 2nd bicuspid and Ist
molar upper right. Depth 18", femurs lower than base of skull,
ankles substantially lower than knees.
Burial 19. Middle-aged adult male. Thick robust skull, large mastoids,
some suture closure. Red powder along left shoulder and be-
tween femurs. Depth 3' 11", lower legs samb elevation as
Burial 20. Child's skull and jaw, probably about 5 years old. Shovel-
shaped incisors. Depth 18", base of skull 4" above knees of
Burial 21. Child's skull and jaw, age about 7 years, shovel-shaped in-
cisors, some tooth wear. Probably bundle burial as jaw under
skull but face of skull upwards. Depth 7", higher than and
apparently not associated with Burial 23.
Burial 22. Isolated child's skull, age about 7 years. Some tooth wear.
Chert chip (Fig. 17, B) sticking in suture between occipital
and left parietal bones, probably cause of death. Depth 14".
Burial 23. Leg bones of adult, skull missing. Depth 14", ankles lower
Burial 24. Young adult, probably female, aged about 17 years. Small
mastoids, some tooth wear, shovel-shaped incisors, large
caries in 2nd lower left molar, 3rd molars not erupted, Inca
bone. Skeleton partially removed by Woodwards. Depth 21",
pelvis and femur lower than base of skull.
Burial 25. Child's skull, jaw, ribs, and clavicle; aged about 5 years.
Two tubular and one disc shell beads (part of a necklace?)
below jaw (Fig. 17, C and D). Depth 18".
Burial 26. Child, probably aged about 13-14 years. Shovel-shaped
incisors, some tooth wear. Below skull, fragments of what
appeared to be decayed shell. Alachaa Plain sherd in burial
fill. Red powder around head and beside right femur. Depth
3' 3", legs lower than,base of skull. Skeleton in dark deposit,
possibly a pit.
Burial 27. Old adult, probably female. Mastoids small to medium, pro-
nounced tooth wear, some molar teeth lost in life. Depth 8",
ankles 1V lower than skull.
Burial 28. Skull of young child, probably 3-4 years old. Wormian bone
at lambda. Depth 5", base of skull 6" above interfemural
area of Burial 27.
Summarizing the above data, we find remains of twenty-nine burials. Of
these eighteen were extended on the back, three were bundle burials, one con-
sisted of an isolated skull, another of only the lower extremities, and six were
burials of young.children. The latter appeared to be bundle burials but we can-
not be positive in all cases due to the poor preservation of bone in this mound.
Extended burial was the accepted mode for youths and adults from the be-
ginning to the end of the use of this mound. Bundle burials were, on the aver-
age, shallower and so probably relatively later in date than prone burials. Bur-
ial goods were limited to a few shell beads in the case of one, possibly two,
children. Red powder was present in six cases, including both males and fe-
Burials indicate a normal population with the usual distribution of old,
middle aged, and young adults, and children.
The Woodwards removed ten to fifteen skeletons in an area some six by
fifteen feet. They told us these skeletons were packed fairly close together
and near the ground level. Some were on top of others with dirt in between. Pre-
sumedly these burials were concentrated between burials 6 and 12 (Fig. 15).
They found no grave goods or pottery. Our findings agree with those noted by
Chips, sherds, and stone tools were found in the mound but were not asso-
cited with burials as grume goods. Artifacts are illustrated in Figure 17. All,
except one, are accidental inclusions transported to the moundwith the dirt
used in its construction. They are evidence that this dirt came from the adja-
cent village area.
I \ / \
S I I I
Fig. 17. Specimens from the Woodward Mound.
About a hundred and fifty chert chips were found. They were randomly
distributed throughout the mound but had a definite tendency to concentrate
in the upper few inches of the gray zone.
A utilized chip and a projectile point (Fig. 17, J and G) were found in the
gray zone. The latter was unfortunately lost and its illustration has been drawn
One of the stemmed points (Fig. 17, E) came from near the edge of the gray
zone at a depth of three feet while the other (Fig. 17, F) was in brown dirt at a
depth of only one foot.
A small scraper (Fig. 17, H) and a greenstone celt (Fig. 17, I) were found
in the mottled zone. The latter is characteristic of Weeden Island mounds of the
Florida Gulf Coast.1 The tip of a point and a fragment of a chipped knife also
came from this fill.
An interesting find was a fragment of a Folsgm-like point (Fig. 17, A). The
base and lower edges of the sides had been smoothed by rubbing. This point
was in the mottled zone, three feet northwest of the AlW-A2W line, at a depth
of slightly over two feet. Typologically this specimen should date from a time
several thousand years before the Woodward mound was built. It probably repre-
sents a pointlost by a hunter on what was later to become an Indian village and
unknowingly transported to the mound in a basketload of dirt.
Possibly the most interesting specimen is the chip illustrated as "B" in
Figure 17. As has been mentioned it was found protruding from the skull of a
child, Burial 22. This chip does not exhibit any retouching or working of any
kind. Yet we must consider it to be a tool and to have caused the death of a
child. Its upper third, which was inside the skull, has a slightly different color
than the rest.
Could this chip have been hated on an arrow for children's sport? If so,
did the parents think it harmless because it had a flat end instead of a point?
Use of such flakes as arrow points is not at all impossible. At Ventana Cave
in Arizona unworked chips were found hated to shafts.2 Finds of this nature
make us realize that many unworked flakes were undoubtedly used although the
evidence is lacking.
In the following list sherds from the mound are compared with those from
the adjacent village area. Included for the mound, as a sherd of St. Johns Check
Stamped, is a pottery hone. Half of this sherd is plain while the other half is
covered with a linear check stamp. On the undecorated area is a deep groove
and multiple striations showing use as a sharpening tool.
IWilley (1945: 231, 2"36Y.
2Personal communication, Dr. Emil W. Haury, University of Arizona.
TABLE 2. Pottery From Woodward Mound and Village Site.
From fill From surface of village
Types of mound F.P.S.a U.F.A.L.b
Alachua Cob Marked c 2 2 1
Prairie Cord Marked 16 6 2
Punctated over cord-marked 2
Gainesville Linear Punctated 1
Eroded, surface modified 1 8 2
Alachua Plain 2 13 3
St. Johns Plain 1 4
St. Johns Check Stamped 1
Weeden Island Plain 1
Plain, Weeden Island paste 1
Brushed, gritty paste 1
Plain, gritty paste 2 7
Plain, Pinellas-like paste 2
Spanish olive jar 2
Totals 24 37 23
aF.P.S. Collection, Florida Park Service
bU.F.A.L. Collection, University of Florida Anthropological Laboratory,
A48 and A83 combined.
CFor local pottery types see Goggin (1948a).
Sherds in the mound fill were concentrated in the mottled zone. A few came
from the brown zone but none from the gray basal deposit. While vertical distri-
bution probably has no special significance in this case, it may be well to mention
it. The Weeden Island Plain sherd was found between Stakes A and AIW at a
depth of only ten inches. Prairie Cord Marked sherds, while concentrating in the
mottled zone, were also found in the brown zone. The two Alachua Cob Marked
sherds, on the other hand, were relatively deep one was two and a half and the
other three feet deep and near Stakes B1W-B2W.
Sherds from the mound are similar to those from the village. The mound must
have been build after the establishment of the village. Presumedly, it was built
and used by the villages (and not by some other group) for their dead.
Based upon seriation and trade sherds, Goggin has suggested five pre-
historic ceramic periods for central Florida in which the Woodward site is
located. The last three, Cades Pond, Hickory Pond, and Alachua are believed
to be temporal equivalents of Weeden Island I, Weeden Island II, and Fort Wal-
ton on the Northwest Gulf Coast and of St. Johns IB, IIA, and IIB on the north-
east coast of Florida, respectively. Goggin writes that Hickory Pond and Ala-
chua are "characterized by an abundance of cord-marked pottery in the former
and cob-marked pottery in the latter."3
The Woodward site would thus be considered a component of Hickory Pond.
The Weeden Island Plain sherd, the St. Johns Check Stamped sherd, and the
celt support a Weeden Island II St. Johns IIA temporal correlation.
Burial mound traits at the Woodward site are both similar to and divergent
from those of Weeden Island. This is what might be expected in an area re-
moved by a substantial distance from the center of Weeden Island manifesta-
tions. The major divergences are the lack of a pottery cache4 and the predom-
inant presence of prone instead of secondary burials.5
Mound and burial traits found at the Woodward mound include a subsurface
base of dark gray sand impregnated with charcoal, central burials in pits dug
into this prepared base, red powder with burials, predominantly prone burials,
and burials concentrated towards the east. With reservations to be mentioned
shortly, these traits have been reported by Moore for mounds in the St. Johns
drainage to the east and on the Gulf Coast to the West.6 In general they occur
to the east with celts and check-stamped pottery (St. Johns II) and to the west
with celts and the contemporaneous Weeden Island pottery.
Concentration of burials in the eastern part of a mound, with or without an
accompanying pottery cache, is, to judge from the literature, limited to the
northwestern part of Florida and, with the possible exception of a site near
Fort Walton, to a Weeden Island time period.7 Prone burials while present are
not the mode for Weeden Island as most interments of that period were flexed
or bundle burials. The same seems to be true for the northern St. Johns area in
St. Johns IIA times. Prone burials are more typical of the central Atlantic and
Gulf coasts of Florida.8
4Tests beyond the periphery of the mound, not shown in Figure 16, failed to
locate a pottery cache.
5See Willey (1945) for a summary of Weeden Island traits.
6Moore (1894: 195), (1896: 464. 470, 489), (1900: 359),(1901:459), (1902: 130,
147, 229), (1903: 374, 382, 434, 435). (1918: 554, 572), (1922: 19);
See also Stirling (1935: 377, 382) for similarities with the central Atlantic
and Gulf coasts of Florida.
7Moore (1901: 437), (1902: 167, 174, 176, 183, 211, 257).
8Walker (1880: 394, 401. 403-4), Stirling (1935: 376, 386-8), Griffin and Smith
A glance at the excavation plan (Fig. 15) will show that burials were not
only concentrated in the eastern part of the mound but also that they were ar-
ranged, in general, with heads towards the center. It would seem that the plac-
ing of burials on the sloping surface of a mound would more or less automati-
cally result in such an alignment. If this procedure had been continued around
the mound, the result would have been an arrangement similar to the spokes of
a wheel. Such arrangements are reported by Stirling for three mounds on Cape
Canaveral and by Walker for a mound near the mouth of the Kootie River.1
European goods were found in two of the Cape Canaveral mounds.
Traits at the Woodward mound in north central Florida seem to reflect its
central location between various similar but also divergent cultural manifesta-
tions. These variations may reflect tribal differences.
Construction of the mound with its prepared base is similar to mounds of
northeast and western Florida as is the presence of red powder with some of
the burials location of burials in the eastern part of the mound is like some of
the Weeden Island mounds to the west but the absence of a pottery cache is
more like east Florida. The Woodward mound is similar to some on the central
Atlantic and Gulf coasts of Florida in the preponderance of prone burials and
their arrangement with heads towards the center of the mound.
The Woodward mound was built and used as a cemetery by the inhabitants
of the adjoining village. The site seems to represent an agricultural community.
Temporally the mound and village may be assigned to the Hickory Pond period
of central Florida. Probably it was flourishing fairly late in that period.
Stirling (1935: 386-8).
10Walker (1880: 394).
1883. "Mounds in Alachua County, Florida." Smithsonian Insti-
tution Annual Report for the Year 1881, pp. 635-637.
GOGGIN, JOHN M.
1948a. "Some Pottery Types from Central Florida." Bulletin, The
Gainesville Anthropological Society, No. 1. Gainesville, Fla.
1948b. "A Revised Temporal Chart of Florida Archeology." The
Florida Anthropologist, Vol. 1. pp. 57-60. n.p.
GRIFFIN, JOHN W., AND HALE G. SMITH
1948. "The Goodnow Mound, Highlands County, Florida." Contri-
butions to the Archaeology of Florida, No. 1. The Florida
Park Service, Tallahassee.
MOORE, CLARENCE B.
1894. "Certain Sand Mounds of the St. John's River, Florida."
Journal, Academy of Natural Sciences, Vol. 10, n.s., pp.
1896. "Certain River Mounds of Duval County, Florida." Journal,
Academy of Natural Sciences, Vol. 10, n.s., pp. 449-502.
1900. "Certain Antiquities of the Florida West-Coast." Journal,
Academy of Natural Sciences, Vol. 11, n.s., pp. 351-394.
1901. "Certain Aboriginal Remains of the Northwest Florida
Coast." Journal, Academy of Natural Sciences, Vol. 11,
n.s., pp. 421-497. Philadelphia.
1902. "Certain Aboriginal Remains of the Northwest Florida
Coast." Journal, Academy of Natural Sciences, Vol. 12,
n.s., pp. 127-355. Philadelphia.
1903. "Certain Aboriginal Mounds of the Florida Central West
Coast." Journal, Academy of Natural Sciences, Vol. 12,
n.s., pp. 363-438. Philadelphia.
1918. "The Northwestern Florida Coast Revisited." Journal,
Academy of Natural Sciences, Vol. 16, n.s., pp.515-577.
STIRLING, M. W.
WALKER, S. T.
"Additional Mounds of Duval and Clay Counties, Florida."
Indian Notes and Monographs, Miscellaneous Paper, No. 26,
Museum of the American Indian, Heye Foundation. New York.
"Smithsonian Archeological Projects conducted under the
Federal Emergency Relief Administration, 1933-34." Smith-
sonian Institution, Annual Report for 1934, pp. 371-400.
"Preliminary Explorations among the Indian Mounds in South-
ern Florida." Smithsonian Institution, Annual Report for
1879, pp. 392-413. Washington.
WILLEY, GORDON R.
1945. "The Weeden Island Culture: A Preliminary Definition."
American Antiquity, Vol. 10, pp. 225-254. Menasha, Wise.
Florida Park Service
CULTURAL OCCUPATION AT GOODLAND POINT, FLORIDA
John M. Goggin
In January of 1949 a visit was undertaken to the shell midden at Goodland
Point, Key Marco, Florida. From previous experience on the southwestern
Florida Gulf Coast, it was anticipated that only a limited conception could be
obtained of site details, and no more than a few artifacts collected because of
of the thick tropical hammock. However, at Goodland Point it was surprising
to find that recent clearing and grading had stripped the midden surface leav-
ing it almost completely bare.
Preliminary reconnaissance indicated that large surface areas of almost
pure shell were relatively sterile in terms of artifacts, while other exposed
areas rich in dark soil and ashes had a high artifact content. The possible
information to be derived from controlled sample collecting (Goggin, 1939)
was immediately apparent, so collections were made from a number of widely
scattered areas, sixteen in all on this day, and two others in the same and
Goodland Point Midden (Cr 45) forms the point of the same name project-
ing from the extreme southeastern corner of Key Marco, the largest of the Ten
Thousand Islands, on the Gulf Coast in Collier County, Florida.2 In most
physical respects the site is similar to other large ones in the vicinity, a
region known as the Calusa subarea of the Glades archeological area (Goggin,
1947). These sites are usually irregular deposits of shell characterized by
uneven surface contours resulting from numerous small mound and ridges on
the main deposit. Goodland Point differs from others in the region mainly in
1. This paper represents a research project of the Department of Sociology and
Anthropology, University of Florida, made possible by a grant from The Viking Fund,
Inc. The Collier Corporation, owners of the site, and in particular Miles and Sam
Collier and the late D. Graham Copeland, the former manager, have been of help in
many ways, especially in furnishing a contour map of the site which was used in the
field, and permitting its adaptation for use in this paper. Robert Reilly, of Miami, as-
sisted in the field.
In March another visit was made to the site in company with Irving Rouse and D.
Graham Copeland. Collections made at this time are in Yale Peabody Museum. They
have not been included in this study, because they were collected under somewhat
2. This site should not be confused with the "Court of the Pile Dwellers" (Cr
49) on the northern end of Key Marco which is well known because of Frank Cushings'
(1897) unusual finds there. (All site numbers used in this paper refer to the statewide
University of Florida archeological survey.)
size-it is an unusually large midden, one of the major sites in Florida in
The midden extends some 1700 feet in an East-West line and measures
about 1900 feet on North-South axis. It is difficult to arrive at an average
elevation of the site due to the extensive removal of shell in the northern
portion and the grading and subsequent smoothing of contours in the western
and southern portion. An estimate, based on the Collier Corporation map and
personal observation, is that one half of the site was from 2 to 5 feet in ele-
vation, one-third ranged from 5 to 10 feet in elevation, with a few high points
perhaps reaching 12 or 15 feet, while the remainder around the edges was
from 0 to 2 feet in elevation.
Fig. 18. Map, Goodland Point Midden. Contours are those surveyed after
clearing and grading. In most places they approximate the original
The cubic content of the site is staggering when one thinks of the labor
of collecting the shellfish. The total has not been computed but an idea may
be obtained when it is realized that an estimated 150,000 cubic yards of shell
have been removed from the borrow pit (communication, D. Graham Copeland,
September 26, 1948). An examination of Figure 18 will indicate the small size
of this pit in relation to the whole site.
The greatest part of the site, the southern and western portions (Fig. 18),
property of the Collier Corporation, has been recently cleared, graded, and
staked out for development. Much of it is now (1950) occupied by families
resettled from the old village of Caxambas on the southern end of Key Marco.
Until recently it was covered with a tropical hammock although parts were
laid out in an avocado grove. The northwest section of this property has been
used as a borrow pit, a large amount of shell being removed forming a small
lagoon. The remainder of the sites, the northeastern block, is known as the
Pettit property. Extensive, but not precisely determined amounts of shell have
been removed from here, and a small boat slip was recently cut into the site.
A number of small houses, cottages, and several trailers occupy the property.
The surface is mainly cleared, but parts of the southwest corner of the Pettit
property are still in hammock vegetation and avocado grove.
The composition of the midden is mainly shell; in fact the bulk of it is
pure shell with only narrow lenses or layers of black dirt. In all, shell con-
stitutes about 95 per cent of the total contents. Of this shell an estimated
90 per cent or more consists of small oysters (2 to 3 inches long). No other
species of mollusc is great number but the Busycon perverse is perhaps the
most numerous of the minor forms, followed by the clam, Venus campechiensis.
In spots Arca ponderosa is noticeably concentrated. Other forms less com-
monly found include Melongena corona, Strombus pugilis, Murex sp., Fascio-
laria gigantea, Fasciolaria tulipa, Macrocallista sp. and Pecten sp. A care-
ful search would undoubtedly reveal numerous other species in small quantities.
The black dirt layers appear to have a high organic content and in places
contain pockets of pure ash. Animal and fish bones, particularly the latter,
are numerous in this material. In addition potsherds and other artifacts are
very common and actually form an important proportion of the total content of
Due west of the northern part of the site, lying in the mangroves, are a
number of shell and dirt ridges (Cr 46). These are reported by Clarence B.
Moore (1900: 372) to contain burials, an observation confirmed by recent ex-
cavations of Mrs. Violet Hanley (personal communication). Due to limitations
of time these ridges were not examined although they were noted from the road.
On the March trip small collections were made here (material in Yale Peabody
PREVIOUS ARCHEOLOGICAL RESEARCH
Goodland Point Midden has been known in the archeological literature for
some time, although it has generally been overshadowed by the unique "Court
of the Pile Dwellers" at nearby Marco. Frank H. Gushing was apparently the
first archeologist to inspect the site, visiting it during the course of his work
at Marco (Cushing, 1897). Clarence B. Moore was there soon after in 1899, ex-
cavating the nearby burial areas (Moore, 1900: 371-6) and later making other
visits (Moore, 1905: 325, 1907: 460-1, 1920). Although Moore made few or no
excavations in the midden proper he obtained an unusual series of artifacts
from here, probably by purchase from the local residents as was his custom.
In 1918, Ales Hrdlicka stopped at Goodland Point in the course of his in-
tensive survey of the Ten Thousand Islands. He was apparently impressed by
the site noting that "it would seem to deserve a comprehensive survey. The
amount of shell heaps and made ground is imposing" (Hrdlicka, 1922: 24).
Henry Collins, Jr., of the Bureau of American Ethnology, made a small
surface collection in 1928. More recently (circa 1946), collections were made
and some small excavations were carried out in the nearby burial area (Cr 46)
by Mrs. Violet Hanley, a graduate student in Anthropology at the University
of Michigan. The Florida Park Service Archeological Survey also has made
surface collections which include various artifacts and potsherds.
The total cultural,sample available as a result of the efforts of these col-
lectors is fairly extensive although few ceramic specimens are represented.3
The present collection greatly supplements this earlier material both qualita-
tively and quantitatively.
The total collection made from the site in three days' work included 4615
potsherds and 351 other artifacts. Of these, 4607 potsherds and 339 artifacts
came from some one of sixteen specific collecting areas, the remainder are
specimens from the general surface other than the special areas.
An attempt was made to obtain uniform sampling from the entire surface
of the site. This was not possible because artifact concentrations were not
uniform over the site, the bulk of the material lying along the central ridge,
although important concentrations were found elsewhere. A definite attempt
was made to obtain an even distribution of samples in the southern half of
the site, and where large gaps exist without samples it is because no materi-
als. were present. However, the same cannot be said for the northern half of
3. The largest collection, other than that at the University of Florida, is in Yale
the site. There time was not available for so thorough a survey as was made
to the south and, moreover, the vegetated southwest corner of the Pettit pro-
perty made collecting there impracticable. Nevertheless, a substantial part of
the northern half was covered, only the vegetated area and the section around
the borrow pit not being examined.
In collecting every attempt was made to avoid any selection in gathering
material, especially potsherds. It cannot be said that every single sherd seen
was collected, but every time I bent over to pick up one sherd, all others
within reach were taken. Unfortunately most of the sherds are very small, aver-
aging less than a square inch in size. In great part this appears to be a direct
result of breakage after exposure, sherds having been crushed by the bulldozer.
Many broken pieces in situ could be seen to have once been much larger sherds.
In general, all fragments of such larger sherds were collected, only the small
"crumbs" being left, and each will be considered statistically as a single
sherd. Most errors resulting from this will tend to be compensatory. Although
a certain number of sherds were not collected in each area, only a large sample
being desired, all non-ceramic artifacts seen were collected. The only excep-
tion was small fragments, usually the outer whorl, of whole shell, Busycon
In going over each collecting area two factors were always kept in mind.
The first was: are we dealing with moved material? The value of any given
sample would of course be negligible if it was a mixture of in situ specimens
with material moved from elsewhere. In most instances it was easy to deter-
mine the in situ material since it was usually in a black dirt layer. A few in-
stances of possible or probable movement of material have been noted under
Secondly an attempt was made to determine the depth of the collecting
area below the original midden surface. In some cases this could not be deter-
mined, but in many instances the land clearers had left a scattering of large
trees on small pedestals, and the height of their bases above the surrounding
area was easily observed.
The size of each sample area varied. No attempt was made to have them
uniform; each area was usually exploited until its limits were reached. In
some instances the limiting factor was a sudden appearance of a new ceramic
or artifact style.
The sampling system used here is admittedly arbitrary, and is in actuality
a compromise involving a number of factors. This work was done between
academic semesters when time was limited and, personnel, moreover, was
limited to the writer, with an assistant for only one day. A more objective
method would perhaps have been an arbitrary staking out of the site in a grid
system, or else using the lots already staked out as controls. However, such
methods would have involved more time and labor than available. Moreover,
if such a purely artificial methodology had been rigidly adhered to it might per-
haps have caused confusion by lumping together nearby but culturally different
areas. It is the writer's belief that the method used is probably as effective a
sampling as was possible.
As has been indicated, collections were made in a series of areas wherever
a concentration of artifact material was found. Their location is indicated on
Figure 18. These locations and size (when indicated) are only approximate, be-
ing drawn in by eye on a base map in the field. The sherd collections from each
area are listed in Table 3 while the other artifacts are listed in Table 4.
Area 1. Judging from surface indications part of the present surface materi-
al in this area may have been moved.4 On the other hand the material collected
is, at least in terms of ceramics, a homogeneous cultural unit. Therefore, such
movenient as may have taken place might have been within an area containing
only material of the same period. Apparently the collecting surface was one to
two feet below the original midden surface.
Area 2. Some movement of surface material was suggested by surface ap-
pearances. Since ceramic material of two periods is present, the collection was
either mechanically mixed in clearing the site or the area had closely succes-
sive occupations. The former appears most probable. Present surface indica-
tions suggest thatthe former midden surface was one to two feet above the pres-
Area 3. There was perhaps some movement of material along the edges of
the area but it was probably negligible.
Area 3A. This "area" is a designation applied to material screened out
of a black dirt and ash concentration in area 3. Artifacts were very common.
This distinct refuse layer, in a general shell matrix, was only six inches deep.
4. Since the site was cleared and leveled by a bulldozer or grader movement of
material can, in most instances, be readily detected by examining the swaths cut by
the blade. In its cutting the blade picker up material, which, if it does not accumulate
to too great an extent, is merely deposited as low ridges on one or both sides of the
blade (depending on the blade angle). Several runs over an area may remove all these
side ridges, by successively moving them to one side. On the other hand, if the blade
accumulates too much in front of the grader, the operator relieves the pressure by lift-
ing the blade slightly, depositing a layer of material in the swath. Such deposits may
include material moved from some distance and are generally distinguishable in the
field by their loose packing and heterogeneous composition in contrast to surrounding
Tobl 3. Ditrlibutlon of Pottery by Type
i : i
I S "
ti I I
An c r
3 1 1010131 1 111810 8 14132 10
16 3 0 2 94 4 01 612 4 39 16 I
272 10 1 1 0 14424 0
147413 0 3 0 9183F
' ' ' "
Table 4. DIstrlbtlon of Non-Ceromic Artifacts by Type
SiS .E !
jz U, -, L
1 6 5 11
31 21 1
1 3 1
57 43 46
8 3 1 2 11 2 11 11 11 3 1 1 1 1 1
Area 4. Possibilities of movement were noted, but it probably was not
Area 5 and 6. Some movement may have taken place, but it is not certain.5
Area 7. Evidences of moved material may be seen in this area. Parts are
one to two feet below the original surface.
Area 8. It is quite probable that some movement took place here. The bulk
of the sherds were concentrated at the inner end of the area away from the
Area 9. Some movement probably took place here since the present surface
is clearly one to two feet below the original modern level.
Area 10. Movement here is uncertain; probably very little.
Area 11. There was very little sign of moved material in the collecting
area, which is now one to two feet below the original surface.
Area 12. Some material appears to have been moved in here from the north.
The general surface level is one to two feet below the old midden surface.
Area 13. Most of the area had little moved material on the surface.
Area 13A. This area consists of material from the south edge of area 13.
The sample is selective and not representative.
Area 14. Most of this area was unmoved but along the edges there were
ridges of moved material. Many of the specimens came from these sections.
The sherd collection, at least in terms of diagnostic types, reflects a mixture.
Area 15. This area is a few feet in from the water's edge and all speci-
mens were collected on the old midden surface.
Area 16. Like the former area, all specimens were collected on the orig-
inal midden surface, where a pair of low shell ridges were separated by a
muck filled depression, apparently an old canal leading in to a small basin.
Recent activity has leveled this area and filled in the basin.
As is customary in the Southeast ceramic remains are classified accord-
5. The collection from area 5 was small but unfortunately has been lost between
the field and laboratory.
ing to a binominal system. Most of the forms found here are recognized types
elsewhere. For that reason the specimens are only briefly described; and more
detailed descriptions may be found in other papers to which reference is made.
Glades Plain. This is a plain form of the basic pottery ware of the area,
the Glades series. It is characterized by a gritty texture and feeling, resulting
from abundant medium to large sized grains of quartz sand included as a tem-
pering material. In shape and surface finish it varies by period but the usual
form is an open bowl. Detailed discussions may be found in Goggin (1939,
1950), Goggin and Sommer (1949), and Willey (1949a).
Glades Red. This type is similar to Glades Plain in all respects except
for red paint being applied to either the inside or outside of the vessel or to
both sides. This coating is very thin.
Sanibel Incised. A type of the Glades series, it is decorated with designs
formed by rows of small tick marks (Goggin, 1944; Willey, 1949a, P1. 3, K).
These usually form simple chevron or running V designs (Fig. 19, L, N).
Gordons Pass Incised. This is also a Glades series type and is similar
in many respects to the previous form. It differs, though, in that the ticks al-
ways border an incised line (Goggin, 1939, Fig. 2; 1944). The geometrical
patterns are more complex (Fig. 19, A-K, P, AA).
Miami Incised. This Glades series form has a group of lines, usually
four, diagonally incised on the rim of the vessel (Goggin, 1950, Fig. 77, F.H;
Goggin and Sommer, 1949, PI. 1, A).
Matecumbe Incised. An incised cross-hatched design on the rim charact-
erizes this Glades series type (Goggin, 1944; Goggin, 1950, Fig. 77, U-W;
Goggin and Sommer, 1949, P1. 1, B, F; Willey, 1949a, P1. 3, B).
Surfside Incised. One to four, but most commonly two incised lines around
the vessel on the rim, is the typical decoration of this Glades series type
(Fig. 19, U; Goggin and Sommer, 1949, P1. 3, J). A distinctive feature of the
type is the presence of triangular rim extensions forming handles (Fig. 19, V;
Goggin and Sommer, 1949, P1. 3, A-C; Willey, 1949a, P1. 14, I).
, Unclassified incised and/or punctuated gritty ware. A number of sherds,
with a paste typical of the Glades series but decorated in a manner not recog-
nized as a type occur. Most unclassified sherds listed in Table 1 are small
and bear only a line or two of incision. Others have more details. Several
(Fig. 19, X and Z) are forms generally considered to be developments out of
Surfside Incised. A related specimen is from unit 14.
Figure 19, Q represents a sherd from an asymetrical vessel. The design
l a A\ /
Fig. 19. Potsherds, Goodland Point Midden. (See Explanation of Figures at
end of article.)
is very suggestive of Miami Incised.
Tick marked and incised sherds such as Figure 19, M and 0 are not classi-
fiable; the former is undoubtedly related to Sanibel Incised, the latter is not so
Deep punctations and incisions are combined on a hard, smooth, paste speci-
men (Fig. 19, BB). It is vaguely similar to some Fort Walton material from north-
western Florida. Figure 19, T, FF and GG are unique.
Simple incised patterns such as Figure 19, R, S, Y and Z cannot be placed
although Fig. 19, S is somewhat suggestive of Glades II material from the East
A simple punctated sherd from unit 3A may be Fort Drum Incised.
Biscayne Plain. This undecorated pottery with a fine soft paste lacks
tempering inclusions and, because of its nature, is known as chalky ware.
Similar forms occur elsewhere in Florida, the cognate type in northern Florida
being called St. Johns Plain.
Biscayne Check Stamped. Impressed check-stamped markings are typical
of this type of chalky ware (Goggin, 1940; Goggin and Sommer, 1949, P1. 3, K;
Willey, 1949a, P1. 4, A-F).
Belle Glade Plain. The paste of the Belle Glade series is intermediate in
texture between-the Glades and Biscayne series. The paste is somewhat chalky
but includes quartz sand. It is rarely decorated. The plain form has typical
wide scraping facets on the outer surface (Willey, 1949a, P1. 2, D-G).
Belle Glade Red. Red paint on the vessel surface characterizes this rare
form. In all other aspects it is similar to Belle Glade Plain.
SGoodland Plain. This is a type representing a new series. Although ex-
amples have been sporadically found elsewhere, the large number occurring
here was used as the basis of the type description. This will be given in full
in a forthcoming paper (Goggin, MS) and is only summarized here.
The paste is the most distinct feature of the ware, yet the most difficult
to characterize. It seems to be related to the chalky series in that it is soft
and smooth in texture but in contrast contains fine grit as a tempering materi-
al a feature of the Glades series. However, this tempering is much finer.
Little is known of vessel form but all sherds examined came from bowls.
Decoration is not typical.6
/ Goodland Red. This type, similar in all ware features to Goodland Plain,
is distinguished by a coat of red paint. This is quite heavy but soft and often
worn away. Paint may be found on both sides of the vessel or on only one sur-
Little Manatee Zone Stamped. A small chalky ware sherd (Fig. 19, CC)
may be this type (Willey, 1949b, Pis. 36, G-H and 37, E-F).
Spanish olive jar. Several sherds of this distinctive utility ware were
found. These typical in most respects with the usual "throwing" marks on
the interior (Fig. 19, EE) and the small rounded lip ring mouth (Fig. 19, DD).
Miscellaneous modern wares. These mainly come from areas 15 and 16 but
a few specimens were found elsewhere. Sherds from area 15 are mainly of a
heavy earthenware with a thin worn green glaze with several small sherds sug-
gestive of slip ware.
Typical slip ware sherds were obtained in unit 16. They came from a thin
large bowl decorated with green lead glaze over a design in white on the red
paste. Other sherds similar to the small ones from unit 15 were also present.
Salt-glazed "ironware" or "stoneware" sherds are probably of quite recent
deposition. They were found in units 16 and 8. .
A rather extensive series of shell tools comprise the main artifacts in this
category. These are mainly forms well known through the collecting of Clarence
B. Moore (1900, 1905, 1907). A more detailed description and comparative dis-
cussion of the types will be found in a monograph soon to be published (Goggin,
MS). The arrangement of these types is in functional order (not given here) dis-
cussed in the previously mentioned paper. Several categories, not types, are
added to include fragmentary tools.
Busycon dipper. The interior whorls of a Busycon perverse shell were re-
moved to form a receptacle with a handle (Fig. 20, A).
Busycon pick A. This is a widely distributed form of shell tool character-
ized by the grinding of the beak to a sharp cutting or picking edge, and the
distinctive preparation for a handle (Goggin, 1950, Fig. 79, X). This was done
by cutting a notch high in the lip and pecking a hole in the opposite whorl,
6. A few sherds of this ware decorated with Gordons Pass Incised designs have
been found. They are classified with that type.
Fig. 20. Artifacts, Goodland Point Midden. (See Explanation of Figuresat
end of article.)
allowing a haft to be inserted in the notch extending past the columella or
interior whorl and out the hole.7
Busycon pick B. This tool is similar to the,former,, differing only in man-
ner of hating, with two holes in the shell to hold the handle. One of these is
located in the whorl just below the shoulder and to the right of the orifice, while
the other is opposite (Fig. 21, A, a-hafted example is in Cushing, 1897, P1.
Busycon pick C. Like the previous types this is similar, but distinctive in
its manner of hating. It resembles type, A, with the notched lip and opposite
perforation, but differs in that the whorl to the right of the orifice is generally
broken away leaving a large hole.
Busycon pick D. This form is a development of type A in which a large
shell was used. The outer whorl was stripped off, exposing the long columel-
la, which is distinctive. The hating modifications are similar to type A (see
Busycon hammer D, Fig. 21, D).
Strombus celt. Sections of the lip of the large conch Strombus gigas were
given an obldng shape and ground to a rounded cutting edge at one end (Fig.
20, B). The single specimen from here is unusually well finished, as is so
often the case with those rare examples found on this coast.
Columella chisel. Sections of the long massive columella of Fasciolaria
gigantea were ground smooth, often flattened on one side, and given a chisel
like edge at one end. A questionable specimen was found here.
Busyton hammer A. Busycon hammers are similar in hating method to
Busycon picks of the same type but have a working surface of the tool that
is rounded by hammering instead of being sharpened into a cutting edge. That
this is not a simple dulling of a sharpened edge through use is indicated by
the great abundance of hammers in the Calusa subarea, where they are the
commonest artifact found. They are very rare in the Tekesta subarea, where
picks are most common.
Busycon hammer B. A hammer with hating modifications similar to Busy-
con pick B.
Busycon hammer C. This is the commonest of hammer forms and is similar
to Busycon pick C (Fig. 21, B). It is probably the most common single type of
artifact in the Calusa subarea.
Busycon hammer D. A hammer with variations similar to Busycon pick D
(Fig. 21, i).
7. Tools still containing the original handle and illustrating this and other methods
of hating were found at the nearby "Court of the Pile Dwellers" (Cushing, 1897).
Fig. 21. Cutting and Hammering Tools, Goodland Point Midden. (See Explan-
ation of Figuresat end of article.)
Busycon tool A. Fragmentary Busycon tools were found of type A. Since
the end of the columella was broken off it can not be determined whether these
were picks or hammers.
Busycon tool fragments. Unidentifiable fragments of Busycon tools were
collected. These were unassignable to either pick or hammer types.
Fasciolaria hammer, short. Hammering tools were formed from medium to
large Fasciolaria gigantea shells by removing most of the large outer whorls
and much of the columella. Two perforations were made in the remaining whorls
for hafting (Fig. 21, E, F,). The end of the columella, and often the apex, were
used for hammering.
Fasciolaria hammer, long. This type is similar to the former except for the
longer columella. As in the case of the short type the end of the columella
shows much wear from hammering (Fig. 21, C).
Columella hammer. The Fasciolaria gigantea columella was often removed,
smoothed, and used for some hammering or pounding purpose judging from the
wear on both ends. Broken ends of long Fasciolaria hammers can be distin-
guished from this type by the rough surface on one end.
Grinding stone. A flat piece of sandy limestone shows wear from grinding.
Socketed bone point. Small mammal or bird bones were shaped into points
by cutting the articular end off square and then making a diagonal cut in the
middle of the bone to form the points. This produced a sharp pointed bone tube
with a natural socket (Fig. 22, C).
Stemmed bone point. A small flat section of bone was given an expanded
pointed end with a long stem or shank (Fig. 22, G). The bone surface shows
the parallel marks of working with a shark's tooth knife. This is an unique
Perforated Arca shell. Valves of the Arca ponderosa were perforated by
knocking a hole in the shell adjacent to the hinge (Fig. 23, A-C). Judging
from Cushing's (1897: 366) finds at the "Court of the Pile Dwellers" these
were tied in clusters as net weights.
Notched shell weight. Heavy sections of clam shell, Venus sp., were
roughly given a rectangular shape and notched on the lip side (Fig. 23, D-I).
Cushing (1897: 366) found them to be used as net weights.
Oliva bead. Beads or cord knobs (Cushing, 1897: 374) were simply made
by breaking off the spire of the Oliva sp. shell and removing interior whorls
(Fig. 22, E, F).
A B c D
Fig. 22. Projectile Points and Ornaments, Goodland Point Midden. (See Ex-
planation of Figures at end of article.)
Fish vertebrae beads. Neat smooth sharks vertebrae were apparently used
as beads. Their only modification is a perforation in the center (Fig. 22, H).
Shell columella pendant. The columella of the Fasciolaria gigantea was
removed, ground down and smoothed, and prepared for suspension by means of
a single groove around one end (Fig. 22, A, B).
Bone pin fragments. Bone pins, one of the most common ornaments in the
Glades area, are slender sections of bone, variously shaped, and given a
rounded point. Fragments of these, not identifiable as to subtype, were found
(Fig. 22, D).
Perforated turtle shell. A small perforated fragment of turtle carapace was
probably part of a rattle.
Fig. 23. Net Weights, Goodland Point Midden. (See Explanation of Figures
at end of article.)
Shell filled with red ocher. A broken Macrocallista sp. shell contained
some red ocher.
Worked stone. A small unidentified fragment of stone, not native to the
area, shows wear from grinding.
TEMPORAL RELATION OF COLLECTION UNITS
With the awareness of differences in material from various sections of
the site came the idea of arranging these collections in a chronological
sequence. Previous work in this subarea had led to the recognition of the
temporal position of certain key pottery types and pottery complexes. At
Gordons Pass Midden two cultural periods were isolated (Goggin, 1939), an
early one characterized by Glades Plain ware and a later by Gordons Pass
Incised. Subsequently the temporal relationship of Gordons Pass Incised
to certain East Coast types of Glades I times was seen at Belle Glade
(Willey, 1949; 63-67), and finally various excavations in the Tekesta subarea
(Goggin and Sommer, 1949; Goggin, 1950) indicated these same types to be
earlier than Surfside Incised and Biscayne Check Stamped.
Thus, to arrange the material in a sequence we had the knowledge that a
predominately Glades Plain sample (lacking incised types) was earlier than
a sample with Gordons Pass Incised, while a sample with Surfside Incised
was later. With these clues the classified data from the collection areas were
analyzed, and the samples arranged in a chronological order (Table 3) and
grouped in a series of alphabetically, designated sequence periods.
Areas 15 and 16, although containing non-ceramic artifacts, had only pot-
tery of European origin and so could be considered late. The distinctive
material in Area 16 consisted of fragments of Spanish olive jars and Slip ware,
a type most closely associated with northern Europe and colonial United States.
The material from Area 15 appears to be of the 19th century and for that rea-
son is considered to be more modern than the ceramics of Spanish and English
colonial times. Thus Area 15 is considered most modern and Area 16 next in
age. Together they will be dated in sequence period A.
Surfside Incised and Biscayne Check Stamped are found in collection 1.
Their absence, except in association with other distinctive material, in other
collections indicated-a special position for this unit. It will be called period B.
Other Surfside Incised sherds were also found in collections 12 and 14
(one in each). Here, though, they are associated with a few sherds of Gordons
Pass Incised, a type known to be earlier. It is possible, then, that we are here
considering mixed material or material dating from a transitional period. Both
explanations may in part be correct, for disturbance and movement were ob-
served in these collecting areas. The high percentage of Glades Plain sherds
in relation to incised material is notable. These collections will be dated as
sequence period C.
Examining our material further, we found that two other units had similari-
ties to the last discussed. These are collections 13 and 2. Surfeide Incised
pottery is lacking, but a few Gordons Pass Incised sherds are included with a
high percentage of plain ware, especially Glades Plain. These collections
along with 13A, form the next sequence period, D. It may be wondered why
these units were placed here instead of proceeding (in time) the main concen-
tration of Gordons Pass Incised. This is because Sanibel Incised, believed
to be the prototype of Gordons Pass Incised, is lacking.8 If this was the
initial appearance of Gordons Pass Incised instead of the final, Sanibel In-
cised should be present in the sample. Thus, collections 2, 13, and 13A form
sequence period D.
Two collections, 3 and 3A, stand out next with a relative and numerical
abundance of Gordons Pass Incised pottery. These, together with collections
6 and 11, are dated as period E. Collection 3 is considered the later of the
two because it is in part physically underlain by collections 3A. The rela-
tively high proportion of Sanibel Incised sherds in collection 6 places it
next earlier. The sample from collection 11 is too small for close analysis
but since it is related to material in this or the next later period it is tenta-
tively placed here.
Collection 4 lacks any incised type, although it contains over 100 sherds.
This suggests a correlation with the early period reported at Gordons Pass
Midden. This collection forms sequence period F.
Four collections, from areas 4, 7, 8, and 9, contain only plain wares but
are such small units (ranging from 21 to 69 sherds) that they do not seem to
constitute a sample from which it is possible to draw any valid conclusions.
CORRELATIONS WITH THE GLADES SEQUENCE
In the previous section a series of collection units were given a relative
temporal position and placed in sequence periods. On examination of these
sequence periods it is seen that a fairly good correlation may be made between
them and the Glades sequence mentioned previously.
The Glades sequence consists of three major periods, divided into a num-
ber of subperiods (Goggin, 1947, 1950). Glades I is characterized by a lack
of incised pottery, several plain wares being found. Glades II is a period of
great exuberance in pottery incision, numerous types being present with some
serving as subperiod markers. In the latter part incision dies out. The intro-
8. Most data for this belief are unpublished as yet, but it is based on slender stra-
tigraphic suggestions from Belle Glade (Willey, 1949a) and Onion Key (unpublished
notes), seriation data, and typological inference.
duction of check stamping marks period III, as does the appearance of Surfside
Incised, and in later times historic material.
Period A can be equated with Glades IIc or later because of the presence
do Spanish olive jar and other European ceramic types.
Period B is clearly IIIa, marked by the Surfside Incised and Biscayne
Check Stamped material. Glades IIIb seems to be absent at the site at least
the collecting did not reveal any examples of the distinctive marker type,
Period C, since it contains marker types of two different periods, is either
transitional or mixed. In either case the mixture or gap spanned is between
Glades I and Ila. This is further suggested by the high proportion of plain
ware material, which is reminiscent of Glades HIc, a predominantly plain ware
period between the distinctively characterized lib and Ilia periods. The dating
of period C may be IIc-IIIa.
Period D is similar to C but, since it lacks Surfside Incised, is probably
earlier. The high percentage of plain wares suggests a dating of Ic (?).
Period E undoubtedly equates with Glades II on the basis of previous
correlation of Gordons Pass Incised at Belle Glade (Willey, 1949a) and Onion
Key (unpublished notes). Within the period it probably equates in part with
Ha and lib, as suggested by the presence of Matecumbe Incised, marker for
IIb, and Miami Incised, a marker for Ha.
If the sample called period F is representative, then that period can be
equated with Glades I. However, inasmuch as this is a small sample from
which to draw conclusions based on negative data, the dating as I is only
In summary it can be noted that if the interpretations presented here are
valid we can demonstrate almost the entire range of Glades Area occupation
at Goodland Point. Only a few periods such as mb and early II, were not pres-
ent or well represented in our sample.
UNIT DATES AND PHYSICAL LOCATION
On the basis of observations of modern shell depositing peoples (in South
Florida and the West Indies), it has been assumed that the South Florida coast-
al middens were formed by discarding material more commonly on the peripher-
ies than on top, once a certain optimum elevation was reached. Moreover it is
believed that growth took place unequally, with a greater tendency to build
outward into the water along the shore line. At Goodland Point the borrow pit
excavation reveals much material under water, either a result of such expan-
sion, or a rise in sea level, or a combination of both factors. Where specific
archeological examination has been made of this outward building phenomenon
the postulated development has been proven, as at Upper Matecumbe Key
(Goggin and Sommer, 1949). With this in mind, the dated collection units were
considered in their relative concentric relation to the center of the site.
In Figure 18 a map shows the distribution of the collection units. Those
considered to be most modern, areas 15 and 16, are directly on the shore. Area
1 next in date is also adjacent to the old shore line. Other collections near to
periphery, Units 7, 8, 9, and 10 are too small to be dated. In the center we
have units 2, 3, 3A, 12, 13, and 14, all relatively early or mixed early and
late, but among these we see the earlier 3 and 3A units occupying a central
position, to which the others have a concentric relation. The location of units
4, 5, 6, are inconclusive in relation to dating.
Although this is not a clear cut picture, there is nevertheless good evi-
dence to support in broad terms the concept of age and area in relation to the
developmental history of the site. The very latest occupations are along or
near the shore while the earlier ones are centrally located. Among these units
the earliest occupies a central position.
CHRONOLOGY AND ARTIFACT CHANGE
The segregation of temporal units by means of variations in decorated pot-
tery offers us the opportunity to consider plain pottery types and non-ceramic
artifacts in the hope of determining whether they exhibit significant changes
in frequency, presence, or absence throughout the periods established. This is
especially desirable inasmuch as such details are lacking in the subarea.
To facilitate analysis the non-ceramic artifact count was arranged by type
and collection in the same order as the ceramic units (Table 4). The plain
pottery wares are included in Table 3.
The distribution of these plain types, especially Goodland Plain, Goodland
Red, Biscayne Plain, and Belle Glade Plain shows little definite clustering.
The omnipresent Glades Plain is in all collections the most numerous type. On
the other hand, Biscayne Plain, Glades Red, Belle Glade Red, and Belle Glade
Plain are represented by so few specimens that positive conclusions are diffi-
cult to formulate.
Of these, Belle Glade Plain is most numerous and exhibits the most pro-
nounced pattern. From early to late it shows a steady increase in number and
frequency, reaching a high point in Collection 12 (Glades Ilc-ma). Glades Red
also has an increasing frequency from early to late reaching a peak in Collec-
Goodland Plain and Goodland Red are in a definite but not a clear pattern.
Both reach their peak in collections 3 and 3A (Glades IIa-Ib).
The data to be derived from the distribution of non-ceramic artifacts are a
surprise since much less variation in time is exhibited than was expected.
This could be due to a misinterpretation of the dates of the collection units
or to a truly long history for certain forms. Cross checking with other sites
suggests the latter to be most probable.
Of the 29 artifact categories listed in Table 4, the great majority, 20, are
represented by too few specimens (less than five) to be chronologically signi-
ficant. Of the remainder, three types, Oliva bead, Busycon hammer B, and
Busycon pick A, have barely enough specimens to be considered.
Busycon hammer C has a distribution in all the sequences units of the site.
It appears to be most common, though, following Glades Ia-IIb times. Busycon
hammer B also seems relatively late. Busycon hammer A is distributed through-
out the sequence with no appreciable relative concentration.
The Fasciolaria hammers do have a distinct distribution. The short form is
present throughout with no appreciable peak in relative frequency. But the
long form is clearly late, being abundant in IIIa times with a specimen from the
Net weights as a group and as types are distributed throughout the sequence,
but the points of greatest frequency differ. Notched shell weights are most com-
mon in IIa-IIb times but rarer later.9 Perforated Arca shells have an opposite
distribution, being relatively rare in Glades II times and more abundant later.
The Oliva bead is sparsely represented in IIa-IIb, and IIc-IIa units. The
only other type to be noted is the Busycon pick A with a IIa-IIb and Ilia occur-
In conclusion we can note that throughout the sequence there is little im-
portant change in frequency of the major plain pottery wares and common tools.
These seem to be, in general, stable or only changing somewhat in frequency
through the site.
SUMMARY AND CONCLUSIONS
Several thousand potsherds and several hundred other artifacts were col-
lected in sixteen distinct surface areas of Goodland Point Mlidden. All materi-
al was typical of the Calusa subarea of the Glades archeological area of South-
9. They are ilso very common at Gordons Pass Midden, a Glades I and II site.
However, Cushing (1897: 366) found them in numbers at the "Court of the Pile Dwell-
ers," a IIIb site.
On analysis these segregated samples could be arranged by ceramic seri-
ation in a series of five sequence periods. These in turn were correlated with
periods in the major Glades sequence and the site is seen to have been occu-
pied for much of the known archeological time in this area.
A check of the physical location on the midden surface of the dated col-
lection units corroborates in general a theory of midden building in Southern
Florida. This postulates a concentric refuse deposition out from the center,
following the shoreline of the midden.
Where sufficient non-ceramic artifacts were present for a valid comparison,
the analysis revealed little change throughout time in the more abundant tools.
The frequency of some varies but perhaps not significantly and only locally.
CUSHING, FRANK H.
1897. "Exploration of Ancient Key Dwellers' Remains on the Gulf
Coast of Florida." Proceedings, American Philosophical Society,
Vol. 35, pp. 329-448. Philadelphia.
GOGGIN, JOHN M.
1939. "A Ceramic Sequence in South Florida." New Mexico Anthropol-
ogist, Vol. 3, pp. 36-40. Albuquerque.
1940. "The Distribution of Pottery Wares in the Glades Archeological
Area of South Florida." New Mexico Anthropologist, Vol. 4,
pp. 22-33. Albuquerque.
1944. A Tentative Formulation of Pottery Types for the Glades Area.
1947. "A Preliminary Definition of Archeological Areas and Periods in
Florida." American Antiquity, Vol. 13, pp. 114-127. Menasha,
1950. "Stratigraphic Tests in the Everglades National Park." American
Antiquity, Vol. 15, pp. 228-246. Menasha, Wise.
MS. Archeology of the Glades Area, South Florida. (Manuscript.)
GOGGIN, JOHN M. AND FRANK SOMMER
1949. "Excavations on Upper Matecumbe Key, Florida." Yale Univer-
sity Publications in Anthropology, No. 41. New Haven.
1922. "The Anthropology of Florida." Publication, Florida State His-
torical Society, No. 1, Deland, Florida.
MOORE, CLARENCE B.
1900. "Certain Antiquities of the Florida West Coast." Journal, Acad-
emy of Natural Science, Vol. 11, n.s., pp. 349-394. Philadelphia.
1905. "Miscellaneous Investigations in Florida." Journal, Academy of
Natural Science, Vol. 13, n.s., 299-325. Philadelphia.
1907. "Notes on the Ten Thousand Islands." Journal, Academy of
Natural Science, Vol. 13, pp. 458-470. Philadelphia.
WILLEY, GORDON R.
1949a. "Excavations in Southeast Florida." Yale University Publica-
tions in Anthropology, No. 42. New Haven.
1949b. "Archeology of the Florida Gulf Coast." Smithsonian Miscel-
laneous Collections, Vol. 113. Washington.
EXPLANATION OF FIGURES
Fig. 19. Potsherds, Goodland Point Midden. A-K, Gordons Pass Incised;
L, Sanibel Incised; M, unclassified incised gritty ware; N, Sanibel Incised;
O, unclassified incised and punctated gritty ware; P, Gordons Pass Incised;
Q-S, unclassified incised gritty ware; T, unclassified incised and punctated
gritty ware; U-V, Surfside Incised; W, unclassified incised gritty ware; X, un-
classified incised and punctated gritty ware; Y, unclassified incised gritty
ware; Z, unclassified incised and punctated gritty ware; AA, Gordons Pass In-
cised; BB, unclassified incised and punctated gritty ware; CC, Little Manatee
Zone Stamped; DD-EE, Spanish olive jar; FF-GG, unclassified incised and
punctated gritty ware.
A-D, Area 3A; E, area 6; F-H, area 3A; I, area 6; J, area 3A; K, area 13;
L, area 6; M, area 11; N, area 3A; O, area 13; P-Q, area 6; R-S, area 3A; T-U,
area 14; V-X, area 12; Y, area 3; Z, area 12; AA, area 3A; BB, area 7; CC, area
14; DD-EE, area 16; FF-GG, area 3.
Fig. 20. Artifacts, Goodland Point Midden. A, Busycon dipper; B, Strom-
A, area 16; B, area 8.
Fig. 21. Cutting and Hammering Tools, Goodland Point Midden. A, Busycon
pick A; B, Busycon hammer C; C, Fasciolaria hammer, long; D, Busycon hammer
D; E-F, Fasciolaria hammer, short.
A, area 3; B, area 3A; C, area 3; D, area 12; E-F, area 1.
Fig. 22. Projectile Points and Ornaments, Goodland Point Midden. A-B,
shell columella pendant; C, socketed bone point; D, bone pin fragment; E-F,
Oliva beads; G, stemmed bone point; H, fish vertebrae bead.
A, area 3A; B, area 3; C-D, area 3A; E, area 14; F, area 12; G, area 3A;
H, area 3.
Fig. 23. Net Weights, Goodland Point Midden. A-C., perforated Arca shells;
D-l, notched shell weights.
A, area 1; B, area 3; C, area 1, D-E, area 3; F, area 12; G, area 3; H, area
12; I, area 3.
NOTES ON THE ARCHAEOLOGY OF USEPPA ISLAND
John W. Griffin
In 1947 expansion of a tennis court on Useppa Island, in Pine Island,
Sound, Lee County, Florida, uncovered human skeletal remains. At the
invitation of Mr. Sam C. Collier, the writer and Mr. Hale G. Smith visited
the island, which had previously been noted in the literature only in pass-
The north and east portions of Useppa Island are covered by an exten-
sive shell deposit, whose highest portion, on the east side of the island,
is largely covered by hotels, cottages, a golf course, and other facilities.
The exposed section, about 100 feet long, seven feet high at the south
end and two feet high at the north end, is located just south and east of
the Tarpon Club of America, on the west side of the tennis court. The
section was that of a typical shell midden, with shell making up the vast
bulk of the volume. Physical stratigraphy was poor in the section which
we saw. Sixteen species of shells were collected, with many conchs and
scallops being noted relative to clams, and with oysters being rather rare.
Portions of several human burials had been removed from the midden
before our arrival, and we found one badly disintegrated skull protruding
about one foot below the surface near the north end of the face. Burial
position could not be ascertained. Artifacts from the site include two
broken Busycon picks of either Type A or B, one Busycon hammer, one
detached columella, and one chert chip. More sherds were collected in
the shallow north end of the midden than in the deeper south end, which
may indicate association with the burials, but this is not certain.
Aboriginal sherds to the number of 127 were collected, and were typed
as follows: 2 plain fiber-tempered; 18 "transitional", with some fiber-
tempering; 86 Glades Plain; 3 Ft. Drum Punctated (?); 4 unclassified
incised Glades series sherds; 1 Glades Red; 6 Belle Glade Plain; 1 Perico
Plain (?); and 3 residual plain.
The most significant sherds found were the fiber-tempered specimens,
since with the single sherd noted by John M. Goggin from the Palm Beach,
Inlet Midden, they mark the present known southern end of fiber-tempered
pottery distribution in Florida. The sherds termed "transitional" in the
analysis appear to represent the transition from fiber tempering to grit
1M. H. Simons, "Shell-heaps of Charlotte Harbor, Florida", Annual Report of
the Smithsonian Institution for 1882, pp. 794-796.
tempering. The location of the fiber-tempered pottery on an island such
as Useppa suggests some form of water transportation for fiber-temperedt
times in Florida.
Turning to the Glades Series sherds we note that two of the incised
pieces have a single horizontal line beneath the lip on the exterior, and
another has the same treatment on the inside. These may be a typical Surf-
side Incised specimens of Glades IIIa times.2 One sherd with a horizontal
line below the lipand parallel slanting lines beneath is somewhat sugges-
tive of Miami Incised of Glades II times. The possible Ft. Drum Punctated
sherds would also be early Glades II. Glades Red appears sporadically
through the Glades sequence, and the remainder of the pottery is of little
help in dating.
The ceramics suggest an early fiber-tempered and transitional occupa-
tion, as well as occupations in either or both Glades II and III. Glades I.
a.primarily plain period, could only be established by stratigraphy. Some
Spanish olive jar sherds from the beach suggest possible contact period
occupation, but may as well relate to Cuban fishing camps. Interestingly.
sherds of blue-edged Staffordshire "china" suggest a nineteenth century
non-Spanish occupation before 1840. Later debris, of course, occurred in
Useppa Island would appear to offer good prospects for a stratigraphic
picture ranging from fiber-tempered times to the full historic period.
Florida Park Service
2The temporal placement of Glades types is from John M. Goggin, "Strat-
igraphic Tests in the Everglades National Park", American Antiquity, Vol. 15,
No. 3, pp. 228-246, 1950.
CONTRIBUTORS TO THIS ISSUE
Gordon R. Willey. Dr. Willey, well known for his extensive work on the
Florida Gulf Coast, in Louisiana and Georgia, and in Peru and Panama, is an
anthropologist on the staff of the Bureau of American Ethnology, Smithsonian
W. J. Armistead. Mr. Armistead is an active member of our society from
Ripley P. Bullen. Well known to many of our readers, Mr. Bullen is Assis-
tant Archeologist, Florida Park Service.
John M. Goggin. Dr. Goggin is Associate Professor of Sociology and An-
thropology, University of Florida.
John W. Griffin. Mr. Griffin, past President and former Editor of our
Society, is Archeologist for the Florida Park Service.