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.
PULISuHED BY THE
FLORIDA ANTHROPOLOGICAL SOCIETY, INC.
NUMBER 3 SEPTEMBER 1979
THE FLORIDA ANTHROPOLOGIST is published quarterly in March,
June, September, and December by the Florida Anthropological
Society, Inc., c/o Room 130, The Florida State Museum, The
University of Florida, Gainesville, FL 32611. Subscription
is by membership in the Society for individuals and institu-
tions interested in the aims of the Society. Annual dues are
$8.00; student members $5.00. Requests for memberships and
general inquiries should be addressed to the Secretary; dues,
changes of address, and orders for back issues to the Treas-
urer; manuscripts for publication to the Editor; and newsletter
items to the President. Address changes should be made at
least 30 days prior to the mailing of the next issue. Second
class postage paid at Gainesville, Florida 32601.
OFFICERS OF THE SOCIETY
President: Jerry Hyde
4233 Oristano Road
Jacksonville, FL 32210
1st Vice President: Thomas Watson
3705 Dilwood Drive
Panama City, FL 32407
2nd Vice President: Irving Eyster
Route 1, Box 91
Islamorada, FL 33031
Secretary: Marion M. Almy
5321 Avenida del Mare
Sarasota, FL 33581
Treasurer: Larry Hochen
P.O. Box 330754 Coconut Grove
Miami, FL 33133
Three years: Norcott Henriquez
1510 Dewey Street
Hollywood, FL 33020
Two years: Adelaide Bullen
Florida State Museum
Gainesville, FL 32611
One year: Robert Marsh
1424 Casson Street
Brooksville, FL 33512
Editor: Jerald T. Milanich
Florida State Museum
Gainesville, FL 32611
Dade County Historical Survey
Kathleen A. Deagan
Department of Anthropology
Florida State University
University of Florida
John W. Griffin
St. Augustine, Florida
George M. Luer
Div. of Archives, History, and
Records Management, Tallahassee
THE FLORIDA ANTHROPOLOGIST
VOLUME 32, NUMBER 3 SEPTEMBER 1979
Editor's Page ........................................ 84
Preliminary Archeological and Geological Evidence
for Holocene Sea Level Fluctuations in the
Lower Cooper River Valley, S.C.,
by Mark J. Brooks, D.J. Colquhoun, Richard
R. Pardi, Walter Newman, and W.H. Abbott ....... 85
An Evaluation of Wet Site Resources of Florida,
by Barbara A. Purdy ........................... 104
An Aboriginal Canoe from Lake Apopka, Florida,
by Arthur F. Dreves .......................... 114
Determination of Site Functions through the
Analysis of Modified Bone,
by Ronald L. Wallace and Susan Jacquith ....... 122
Statement of Ownership, Management, and Circulation 128
A number of publications which are of interest to Florida
archeologists have recently become available. Two deal with the
Georgia coast. They are The Anthropology of St. Catherines Island,
I. Natural and Cultural History, by D.H. Thomas, Grant Jones,
R.S. Durham, and Clark Larsen (Anthropological Papers of the
American Museum of Natural History, Vol. 55, part 2, pp. 155-248)
which includes an excellent overview of the historic Guale Indians
and The Anthropology of St. Catherines Island, 2. The Refuge-
Deptford Mortuary Complex, by D.H. Thomas and Clark Larsen (same
series, Vol. 56, part 1, pp. 1-180). They can be ordered for $6.30
and $12.80, respectively, from the Museum, Central Park West at
79th Street, New York, NY 10024.
An important article by David Brose and George Percy entitled
Fort Walton Settlement Patterns is contained in Mississippian
Settlement Patterns, a collection of papers edited by Bruce Smith.
Published last year, the 544 page volume can be ordered from
Academic Press, Order Dept., 111 Fifth Avenue, New York, NY 10003,
Wilfred T. Neill's book Archeology and a Science of Man has
been published by Columbia University Press (526 West 113th Street,
New York, NY 10025). Neill uses a large number of examples from
Florida to show how archeology can benefit from incorporating
other disciplines into research. The book, dedicated to Ripley
and Adelaide Bullen, is written in non-technical style and I
believe that readers will find it an important addition to their
libraries. It sells for $20.00
Still another collection of essays has been edited by
Charles Hudson. Black Drink--A Native American Stimulant contains
articles summarizing all aspects of the drink, which was brewed
from the leaves of the Ilex vomitoria. Chapter authors include
Charles Fairbanks (on Creek useage), William Sturtevant (on
European use from Colonial times to the present), and myself
(on the archeological evidence for prehistoric useage). The
170 page book may be ordered from the University of Georgia
Press, Athens, GA 30602 for $11.00.
PRELIMINARY ARCHEOLOGICAL AND GEOLOGICAL EVIDENCE FOR HOLOCENE
SEA LEVEL FLUCTUATIONS IN THE LOWER COOPER RIVER VALLEY, S.C.
Mark J. Brooks, D.J. Colquhoun,
Richard R. Pardi, Walter Newman,and W.H. Abbott
In this paper we present archeological and geological
arguments and data supporting the occurrence of Holocene sea
level fluctuations and relative sea level positions from 4200-
1000 B.P. in the Lower Cooper River Valley, S.C. The available
archeological data suggests relatively high sea level stands at
times during the temporal intervals from 4200-3700 B.P., 3100-
2850 B.P., 2250-1750 B.P., and 1600-1000 B.P., and the geological
data indicates lower sea level stands at 3100 years B.P. and be-
tween 2695 and 2330 B.P. with higher stands before and after
these dates. We demonstrate an association of archeological site
date clusters with rising sea level as indicated by transgressive
salt marsh clays.
The study area lies within Berkeley County S.C. in the west
central portion of a triangular-shaped region of the Lower Coastal
Plain bounded by the courses of the Santee and Cooper Rivers, and
the shore of the Atlantic Ocean (Fig. 1). The stratigraphy and
geomorphology of the region have been reported (Colquhoun, Bond
and Chappel 1973). They note that most of the surficial landforms
were formed during stages of the last Interglacial (Sangamon 80,000-
120,000 years B.P.) during transgression and regressions of the
Atlantic Ocean in response to melting of the ice caps. Three
major regions of barrier island emplacement are indicated and exist
today as relatively well drained, sandy areas extending from the
southwest toward the northeast between the Cooper and Santee Rivers.
Each of these areas is approximately three to five kilometers in
width and is bounded toward the northwest and southeast by former
marsh and deltaic plains. These latter are today poorly-drained
with erratic stream courses resembling contemporary tidal channel
drainage in existing salt marshes. Exceptions to this are the
courses of Grove Creek and the East Branch of the Cooper River.
These extend in shallow meandering floodplains almost directly
east-west from their confluences with the main branch of the
Cooper River, terminating eight or more kilometers toward the
east within the old, elevated, poorly-drained marsh plains.
Grove Creek intersects the main branch of the Cooper River,
five kilometers to the north of the central Pleistocene barrier
ridge (Cainhoy) while the East Branch of the Cooper lies about
four kilometers south of the more northerly barrier system (Bethera).
Between the floodplains themselves, as well as between their nearest
sand hills, occur partially dissected generally clayey, poorly-
drained elevated Pleistocene marsh regions, in which well-drained
sandy soils are rare.
The Florida Anthropologist, vol. 32, no. 3, September 1979
SEA LEVEL FLUCTUATIONS
Figure 1. Location maps.
The floodplain of the Cooper River within the region is broad,
being a minimum of about four kilometers wide adjacent to the Cain-
hoy old barrier system and expanding to about eight kilometers
adjacent to Grove Creek. Within the river valley, which divides
present Holocene sediments from the Late Pleistocene geomorphology,
extensive salt marshes are present with the Cooper River flowing
in a meandering pattern adjacent to the western bank. The river
is several hundred meters in width and up to ten meters in depth.
In many places it can be shown to be scouring the underlying
Tertiary sediments, and shifting large point-bar sand bodies
developed mostly downstream on the inside of meander bends.
Within the floodplain in the vicinity of Grove Creek several
small islands are present which extend landward toward a peninsula
joining the mainland a few kilometers above the confluence. The
peninsula, the islands, as well as many portions of the valley
wall of the Cooper show an undulating boundary between the present
marshes and the mainland which probably reflects a cut bank origin
formed, most probably, during lower stands of sea level developed
BROOKS ET AL.
during Wisconsin Glaciations. Relief, being nowhere conspicuous
in the entire region, is most abrupt adjacent to this valley wall.
Elevations of up to five meters are common on these elevated
Pleistocene marsh surfaces.
A recent archeological and geological survey of Amoco Realty
property, centering around Grove Creek immediately adjacent to the
Cooper River Estuary, resulted in the discovery of nineteen pre-
historic sites (thirty-four temporal components) and at least two
peat beds intercalated with marsh clays in the adjacent estuary.
The archeological sites range in time from ca. 10,000 to 1,000
B.P. and likely, though by no means proven, represent small, up-
land deer hunting and oak-hickory nut extraction camps situated
on small, dispersed patches of relatively well drained soils
(Brooks and Scurry 1978). Soils of this type are capable of
supporting high densities of nuts and deer, which could be
efficiently exploited in the fall by aboriginal human populations
when the nuts ripened and the deer aggregated to feed on them
(Quarterman and Keever 1962; Smith 1975). Unpublished soil data
provided by the Soil Conservation Service indicates that these
well and moderately well drained soils currently comprise rel-
atively small percentages of the total soils in most Lower Coastal
Plain areas. In the Amoco project area, for example, less than
40% of the 2,000 acre tract of land surveyed contained relatively
The unpublished soil data was the basis for an 11% stratified
random sampling strategy designed especially for the discovery
and definition of small, low artifact-density sites in heavily
vegetated areas through subsurface testing. Assuming that nuts and
deer were the resources emphasized (probably not the exclusive
resources) in predominantly upland, inter-riverine areas such as
Amoco, it was expected that there would be a strong positive
correlation between the better drained soils and prehistoric sites
(due largely to environmental and hence, resource variability;
this is not expected to be true in all lower coastal plain areas
[see Pearson 1977]). To substantiate this, however, it was also
necessary to demonstrate that sites tend not to occur on the poorer
drained soils. Consequently, both the better and poorer drained
soils received sampling representation in proportion to their
actual occurrence in the project area. Therefore, assuming that
soil drainage quality was a major variable determining site
location in this area throughout prehistory, it is likely that the
sites discovered represent the temporal and functional range of
It would appear then, based on the research of Brooks and
Scurry (1978), that nuts and deer were likely the target resources
of aboriginal populations in upland, inter-riverine areas through-
out most of prehistory. Since most elevations in the Amoco project
area are less than five meters above present mean sea level, fluc-
tuations in sea level would have had a major impact on the relative
SEA LEVEL FLUCTUATIONS
proportion, distribution, and nut-deer productivity of the well
to moderately well drained soils and, in turn, human adaptations,
over time. This is so because many areas with marginally adequate
drainage would become too poorly drained during higher sea level
and water table stands. Therefore the temporal variability in
prehistoric settlement patterning observed in the Amoco project area
may be attributed in large part to Holocene sea level fluctuations.
Investigations of adjacent peat and salt marsh sequences in
the Cooper River Estuary indicate that such fluctuations did occur
and suggest that the archeological hypotheses are tenable.
Holocene sea level changes within estuaries and coastal marshes
have been observed by European workers by examining the shallow
stratigraphy of those areas, and determining the nature of water
conditions present during the deposition of the beds. The following
observations and conclusions drawn from that literature are important
in interpreting the stratigraphic sequence in the Cooper River
1) Brackish water peat deposits as indicated by diatoms and
palynology have been shown to have been deposited at or
very close to mean sea level near fresh water-salt water
boundaries in parallic environments (Jelergsma 1961;
Hageman 1960; see Brand et al. 1965 for summary).
2) In Holocene sections such peats occur in two aspects:
a) as intercalated peats separated by salt marsh clays
reflecting incursion of fresh waters into formerly
brackish and marine regions which subsequently return
to marine conditions; b) as basal peats developed on
Tertiary or Pleistocene compacted sediments overlain by
salt marsh clays, deposited under water conditions re-
flecting rising water tables and subsequent marine
3) The transition from salt marsh clays to fresh and brackish
water peats and back to salt marsh clays upward in a
Holocene section indicates either transgression-regression-
transgression resulting from rise-fall and rise in sea
level, or prolonged low fainfall-high rainfall-low
rainfall in an estuary and its associated drainage basin
4) Radiocarbon dating of the tops and bottoms of blanket-
like intercalated peats have shown a remarkable coincidence
of dates for extensive areas of Northern France, Belgium,
Holland and Northwest Germany indicating processes
operating under 3 above extend over broad areas (Somme
1975; Paepe 1960; Hageman 1969; Behre et al. 1975).
5) At least eight transgressive-regressive cycles have
been identified between the present day and 8,000 B.P.
in the European Lowlands, and similar observations have
BROOKS ET AL.
been reported in northwest England (Tooley 1977), southern
Atlantic France (Ters 1973) and in southwestern Sweden
6) Actual elevation of sea level during these transgressions
and regressions is indicated by coincident radiometric
dates of associated basal peats deposited on compacted
Pleistocene cover sands; the intercalated peats possibly
Data from the Charleston area indicates similar processes
are operative as in the European Lowlands, and several of the
associated peats have been identified, at least two of which
may be contemporaneous with similar peats in Europe.
The major research goal of the Amoco archeological project
was to examine prehistoric adaptations over time, which are
directly reflected by temporal variability in subsistence -
settlement systems (Schneider 1974; Jochim 1976). To do so,
and to explore possible correlations with the sea level
fluctuations, hypothesized in light of radiometric dates reported
in the European literature, it became necessary to establish
tight chronological controls.
In the absence of radiometrically dateable material from
the archeological sites in question (soil acidity has leached
most carbonate material), relative dating was accomplished by
seriating the ceramic assemblages from the respective sites
using the ceramic sequence established for the North Georgia
Coastal Plain. The Georgia sequence was derived from radio-
metrically dating ceramic assemblages (the deposits in which
they occurred) in stratigraphic context (DePratter 1977).
By emphasizing the temporal overlap of earlier and later
artifacts, in conjunction with cross-dating artifact assemblages
of unknown age with similar assemblages that have been radio-
metrically dated, refined cultural chronologies may be established
through seriation. In line with traditional and current anthro-
pological concepts of stylistic variability and culture change,
seriation is a method that is theoretically sound. It is a
common archeological tool that has been demonstrated to be
quite accurate if properly utilized (Ford 1962; Deetz and Deth-
lefsen 1972; LeBlanc 1975).
In terms of the prehistoric sites discovered during the
Amoco survey, only twenty-one of the thirty-four temporal
components contained sufficient ceramic materials to be dated
reasonably through seriation. All of the dateable components
fall within the time range from 4200 to 1000 B.P. Within this
range the components were, in most instances, assigned to
temporal intervals of less than 200 years.
SEA LEVEL FLUCTUATIONS
DePratter's (1977) temporal units range from 50 to 600
years in duration. In the event that the units exceeded 200
years, they were further broken down for our purposes into two
or three equal intervals (depending on the duration of the period
and the existence of earlier and later ceramic assemblages). The
temporal boundaries of these smaller intervals were defined by
the presence and relative proportion of elements of earlier and
later ceramic assemblages as well as by the relative occurrence
of the types that characterize the larger units. The assemblage
assigned to each of DePratters units was viewed here as repre-
senting the mid-point of that period. Consequently, the fewer
elements of this assemblage present, the greater the likeihood
of being further in time from the mid-point. If earlier and
later ceramic assemblages exist, elements of them would be present
to an increasingly larger degree the further in time we move away
from a given mid-point.
Upon plotting the components according to their respective
temporal intervals, it was observed that the components cluster
discretely in time with intervening periods when the area was
apparently not utilized, at least in an archeologically recog-
nizeable way (see Fig. 2). Although it could be argued that the
"gaps" are a result of small sample size, radiometric dates
obtained from Cooper River-Grove Creek sediments, and elsewhere,
suggest that the gaps are real and that the temporal clustering
of components generally correlates with high sea level stands
from approximately 3100-2850 B.P. and 2250-1750 B.P. Additional
clustering between 4200-3700 B.P. and 1600-1000 B.P. is also
present, but cannot as yet be correlated with fluctutuations
in sea level recorded in the adjacent marshes.
370 350030 300 2950 280 27 260 250 2400 22 2050 190 170 160 15 14 120 110
3700 3500 3300 3100 2950 2850 2700 2600 2500 2400 2250 2050 1900 750 600 1500 1400 1250 1150 1000
TEMPORAL INTERVALS IN
YEARS BEFORE PRESENT
O COMPONENT CLUSTERS
Figure 2. Archeological site component clusters.
BROOKS ET AL.
While seriation was necessary to establish dating of
sufficient definition of less than 200 years for the archeo-
logical sites, such was not the case for adjacent estuarine
sediments. For dating in those areas the liquid scintillation
counting of benzene as developed by Noakes et al. (1967) was
used. Samples were converted to benzene by use of a synthesis
apparatus fully described by Pardi (1976).
Geological Evidence for Holocene Sea Level Fluctuations
Sixteen cores were taken from the Grove Creek-Cooper
River marshes in conjunction with numerous bore holes drilled
on the adjacent highlands. We find the highlands to be under-
lain by Oligocene, Pliocene and Pleistocene sediments, and the
region to have been most recently deeply incised by erosion
developed during low stands of sea level associated with the
Wisconsin glaciations. Holocene sediments are present deposited
within the Cooper River estuary and Grove Creek on an irregular
surface associated with this erosion interval. That surface
seems to reflect former river and stream valleys which drained
the adjacent highlands. We find basal peats overlain by salt
marsh clays (as indicated by their contained clay mineral
assemblages; see Colquhoun et al.  for further discussion)
immediately above that surface in several locations. In addition,
a buried forest composed of stumps of what appear to be cypress
trees, sharply truncated by salt marsh clays is encountered
frequently on the Pleistocene basement. In addition we find
peat beds deposited in fresh and brackish waters (as indicated
by their contained diatom assemblages), overlain and underlain
by salt marsh clay sequences (as indicated by their contained
clay mineralogy), as well as the presence of abundant Spartina
roots and stems, away from the present boundary between the up-
land and the marsh.
Dates obtained from the brackish basal peats and logs
occurring above relatively compacted Tertiary and Pleistocene
sediments have provided information indicating the elevation
of high marsh surfaces at the time measured, prior to the
invasion of marine waters as reflected in the overlying salt
marsh clays. Dates obtained from intercalated peats (with
fresh and brackish water fossil assemblages) overlain and under-
lain by salt marsh clays (as indicated by their clay mineral
assemblages) have allowed us to observe times of transition
from marine to fresh and brackish and back to marine conditions.
They do not indicate any measurement of sea level, however, since
those sections are in all cases compacted.
Elevations were obtained relative to the high marsh surface.
The various datums are strictly local and are not tied into a
universal geodetic levelling system. Results from these tidal
marsh dating compare relative changes in sea level at a station
and from station to station. A potential problem may be thought
to exist in the selection of a high marsh surface datum because
of its variability. However, this variability is not significant
SEA LEVEL FLUCTUATIONS
in that the relief of the equilibrium marsh surface covered with
a phreatophytic sere climax usually is 0.3 meters or less (Adams
1963; Redfield 1972).
Our conclusions are based in part on data illustrated
in Table 1 below.
Table 1. Cooper River-Grove Creek estuary samples.
-1.0 contact not tested
BROOKS ET AL.
Sample numbers A-8,2 and A-8, though dated only once, were
obtained from extensive peat bodies encountered widely within
the stratigraphic section tested within the area. In each
case they reflect a change from marine conditions to brackish
and fresh conditions and back to marine conditions operative
over several miles along the long axis of the estuary (in this
case from north to south). They represent therefore, either
a) a fall in sea level and subsequent rise, with consequent
seaward and then landward migration of the fresh-brackish-marine
water boundary, or b) periods of more extensive run-off within
the Cooper River drainage basin, which would tend to produce
the same effect. The Cooper River drainage basin is relatively
small, and would quickly respond to effects of drought which
would cause a decrease in run-off within the system and a migration
of the marine conditions landward thus destroying the possibility
of peat development of this type within this area. We believe
a fluctuation in sea level appears more likely, though unproved.
Examining all these occurrences, with respect to the paleo-
environmental elevation and date, we conclude that the marine to
fresh-water transition zone in theCooper River-Grove Creek area
has fluctuated with respect to both elevation and time during the
last 6000 years.
We observe that it stood relatively farther up the estuary
prior to 3100 + or 100 years B.P. when the transition zone
migrated down river to the position of sample A-8 near Grove
Creek. We note that this regression down the estuary was followed
by a transgression of the zone back up the estuary during which
time marine salt marsh clays were deposited over the 3100 year
B.P. intercalated peat. We observe that this transgression was
followed by another regression down the estuary of the fresh water-
brackish water transition zone during which a second peat was
emplaced dated at 2330 + 140. This emplacement was again followed
by a transgression of the transition zone back up the estuary
depositing marsh clays associated with the present high marsh(Fig. 3).
AA u0 14 \ A7 A3 A
COOPER RIVER "- --i _o-- o : -
fo1 TERUARY-PLEISTOCENE 3
_7_j i A14(4135 65)
PEAT WITH FRESH AND -
[UBRACKISH WATER DIATOMS ,
SALT MARSH CLAYS I
0 t 00_FEET
Figure. 3. Stratigraphic cross-section, Cooper
River-Grove Creek area.
SEA LEVEL FLUCTUATIONS
When we examine that portion of the stratigraphic section
deposited between the island underlain by Tertiary and Pleistocene
sediments and the mainland, we note that rooted trees were es-
tablished growing on that Pleistocene surface. One of these
stumps in contact with the overlying salt marsh clays 2 meters
below the present high marsh surface was dated at 2695 + 115 years
B.P. A second stump several hundred meters toward the west and
occurring 1 meter below the present high marsh (again in sharp
contact with marsh clays atthat depth), gave a date of 2035 + 105
years before present. We assume that the trees occurred growing
either on drained land, or within a fresh to brackish water
swamp, and that the transition between that environment and the
overlying marsh clays (abundant in Spartina roots) represents
transition to a marine environment. We note that the contact
between the buried forest and the overlying marsh clays occurs
uninterrupted, and gradually rises from -2 to -1 meter below
the present high marsh. It continues to rise approaching the
present island. It does not extend to the island itself however,
being interrupted by marsh clays lying directly on Pleistocene
clays near the present shoreline. We assume, therefore, that
prior to 2695 B.P. the marine zone was present farther up the
estuary above Grove Creek, as indicated by marine marsh clays
underlying the intercalated peat at 2330 B.P.; that by 2695 B.P.
either the Grove Creek area was drained with sea level lying
below -2 meters or that sea level lay near -2 meters, and that a
forest was established; that beginning in 2695 B.P. sea level
rose and lay close to -1.6 meters near 2300 B.P. (from the inter-
calated peat) and that by 2035 B.P. sea level attained -1 meter
(all measurements relative to the present high marsh surface and
all radiometri dates uncorrected). This latter regression is in
general agreement with the observations of DePratter and Howard
(1977) based on beach ridge accretion rates, in Wilmington and
Tybee Islands Georgia.
We summarize our thoughts in curve form in Figure 4. Our
higher stands are indefinite with respect to elevation. We do
not know as yet precisely how much higher sea level was before
or after the regressive intervals, for there are still many
windows in our dating in which the elevation is still not
determinable. Nevertheless these two periods of lower stands,
coupled with the near absence of archeological sites dating to
these periods, and the relative abundance of sites during other
times, indicate that a closer examination of the data is
BROOKS ET AL.
Years B.P Ix 103
7 6 5 4 3 2 1 0
i __ i --- i ------ i ------U-- 0 0U
?\ /. aQC583 1
t I I I 0
I \ I= QC 613 .2
QQ1 C 585 W
PQC588 3 o
JQC 587 I
// ,4 "
0 QC 586
Figure 4. Preliminary sea level curve for the Cooper
Archeological Evidence for Holocene Sea Level Fluctuations
Several inter-related factors suggest that the-four temporal
clusters of archeological components represent primarily high
sea level stands, with the intervening temporal spans representing
lower sea level stands. First, during higher sea level stands
there would be a general reduction in land mass. This, in
combination with an assumed trend in human population growth would
bring about a corresponding reduction in the size of band terri-
tories. Under these conditions, which would essentially produce
a "packing affect" on human populations, a more labor-intensive
economy in terms of the upland resources in question would be
probable (Binford 1968; Birdsell 1968). Possible evidence for
SEA LEVEL FLUCTUATIONS
a trend in population growth and a more labor-intensive economy
during certain periods may be indicated for the Amoco project
area (see Brooks and Scurry 1978). Generally, there is an
increase in the number and size of sites over time. This in-
crease is most notable during the woodland, especially the
Middle-Late Woodland period. It is during this period that we
see not only an increase in the number of small sites ( ca. 60-
300 square meters), but also the appearance of comparatively
larger, multicomponent sites (ca. 900-3,000 square meters).
It should be noted that the larger sites (7 sites-37%) are still
rather small and appear to be functionally equivalent to the
smaller, usually single component, sites. An assessment of
functional equivalence is based on the occurrence of very similar
artifact assemblages and the strong tendency for sites, regard-
less of size, to occur on the better drained soils. It would
appear then, that the major difference between the larger and
smaller sites is that the larger sites tend to be situated on
correspondingly larger and more productive (in terms of nuts
and deer) patches of better drained soils and were, therefore,
more intensively utilized.
Second, higher sea level stands would directly and indirectly
bring about a reduction in the amount of well to moderately well
drained soils. Correspondingly, there would be a general reduction
in nut and deer productivity, with localities optimal for these
resources becoming fewer and more dispersed. This condition would
also promote a labor-intensive economy.
Thus, with a labor-intensive economy, resulting in large
part from high sea level stands, there would be quantitatively
greater archeological outputs (artifacts) and, correspondingly,
an increase in the number of recognizable archeological sites.
Conversely, during lower sea level stands much of the "pressure"
would be off and we would expect nuts and deer to be exploited
on a more extensive, as opposed to intensive, basis. Under
these circumstances, artifacts would be less likely to accumulate
at any given spatial locus (site) and would, therefore, produce
relatively few archeologically recognizable sites.
Following the above reasoning, there should be a general
correlation between the number of sites (components) during a
given temporal span and relative sea level position, with the
higher the stand resulting in more archeological sites (components).
Thus, archeological data might be utilized not only to infer sea
level fluctuations, but it might also be utilized to indicate
relative sea level position over time. Accordingly, again using
data from Brooks and Scurry (1978), the component cluster from
3300-2850 B.P. (8 components) should generally represent a
higher sea level stand than the 4200-3700 B.P. cluster (2 compo-
nents) with the intervening temporal span from 3700-3300 B.P.
representing a relatively low sea level stand in comparison with
that of the two earliest component clusters. Similarly, the
component clusters from 2250-1750 B.P. (5 components)and 1600-
1000 B.P. (6 components)should represent relatively high sea
BROOKS ET AL.
level stands intermediate between the high sea level stands
represented by the two earliest component clusters.
Before proceeding, it should be noted that there is a
slight contradiction between the above archeological data and
the geological data obtained from the adjacent marsh. It is
maintained, however, that this discrepancy is more apparent
than real. Assuming that the 3100 B.P. peat date is correct,
and that this represents a low sea level stand, then we would
expect few archeologically recognizable sites. This is not
to say, however, that no sites (components) are expected to
occur during periods of low sea level stand. Consequently,
the single component occurring during the 3300-3100 B.P. interval
(see Fig. 2) cannot be used in itself as an argument for relative
sea level position. Nevertheless, assuming a low sea level stand
at 3100 B.P., the four archeological components in the 3100-2950
B.P. interval should represent a higher, and possibly rapidly
rising, sea level.
In support of the hypothesis that the four component
clusters generally represent relatively higher sea level stands,
and that the relative height of these stands is generally corre-
lated with the number of components in the respective clusters,
the following site-settlement locational data developed by
Brooks and Scurry (1978) is presented. First, for purposes of
the following discussion of site-settlement locational data,
it should be kept in mind that temporal components occur at
discrete spatial loci, i.e. sites. As such, site locational
attributes also apply to the contained components.
Well and moderately well drained soils in the Amoco project
area typically occur in small, dispersed patches at the lower
elevations, e.g. below the 10' contour. At the higher and more
environmentally stable elevations, however, these soils tend to
occur in larger, less dispersed patches. Consequently, such
soils at the higher elevations would have been the most productive
for nuts and deer over time. It is likely for this reason that
most of the prehistoric sites (often multicomponent) are situated
on the better drained soils in these areas, especially given
that intensive utilization occurred during higher sea level stands.
Further, given the relative environmental stability of the higher
elevations, at least in terms of sea level, it seems reasonable
to assume that existing soil drainage qualities in these areas
are much the same as they were during most of prehistory. This
is supported by the fact that even today when sea level is rela-
tively higher, most of the sites are still situated on the better
Soil data indicate that although archeological sites in
the Amoco tract tend to be located on well to moderately well-
drained soils during all temporal periods, this is variable
for each of the four component clusters (see Table 2). To
demonstrate this, soils were assigned values from 1 (poorly drained)
SEA LEVEL FLUCTUATIONS
through 5 (well drained). The resulting mean values for the
components in each of the four temporal clusters were R = 4.5
for the 4200-3700 B.P. cluster, X = 3.5 for the 3300-2850 B.P.
cluster, X = 3.8 for the 2250-1750 cluster, and X = 3.16 for
the 1600-1000 B.P. cluster. It is recognized that the use of
mean values is not entirely appropriate for ranked data.
However, it should be generally adequate for indicating vari-
ability in site-soil associations over time.
Table 2. Archeological site components
in the Cooper River-Grove Creek area.
Cluster Temporal Interval Number of X Soil Drainage X Elevatio
1 4200-3700 B.P. 2 4.5 10'
2 3300-2850 B.P. 8 3.5 18.75'
3 2250-1750 B.P. 5 3.8 15'
4 1600-1000 B.P. 6 3.16 17.91'
Given that there were fewer well and moderately well-drained
soil patches during higher sea level stands, and that it might,
therefore, often be necessary to exploit those soil patches that
were less than optimal for nuts and deer, then the above values
indicate that from 4200-3700 B.P. that more well to moderately
well drained soils were available than during the later component
clusters. This suggests that this high sea level stand was rela-
tively lower than those represented by the later component clusters.
A value of 3.5 for the 3300-2850 B.P. cluster indicates less
utilization of the well to moderately well drained soils and a
generally higher sea level stand. A value of 3.8 (between)
4.5 and 3.5) indicates a high sea level stand for the 2250-1750
interval that is intermediate between the two earlier component
clusters. Finally, a value of 3.16 for the interval from 1600-
1000 B.P. indicates a high sea level stand approximating that of
the 3300-2850 B.P. interval, or slightly higher. As can be seen
from this data, there is close agreement between the relative
position of sea level based on the number of components during a
given temporal span, and, the relative position of sea level based
on the drainage qualities of the soils selected for use by pre-
historic human populations.
Finally, an examination of mean elevation for the components
in each of the four temporal clusters produces similar results
(see Table 2). Starting_from the earliest to the latest cluster,
the values are X = 10', X = 18.75', = 15, and X = 17.91'; respec-
tively. A value of 18.75' indicates a selection by human populations
BROOKS ET AL.
during the 3300-2850 B.P. period for higher elevations than
during the earlier 4200-3700 B.P. (10') component cluster. A
generally higher sea level stand is also indicated. Values
of 15' and 17.91' for the most recent temporal site clusters
indicate high sea level stands intermediate between those of
the two earliest clusters, closest to the generally higher sea
level stand represented by the 3300-2850 B.P. interval.
It should be clear that temporal variability in the
elevations selected for site location is largely correlated
with changes in the availability of the better drained soils.
That is, with increasing sea level height there will be
greater intensification, resulting from the necessity to locate
sites at successively higher elevations in order to take advantage
of the better drained soils, which would be declining in amount
with increasing sea level.
Future Research in Archeology
In terms of general patterning we feel that the preceding
arguments and supportive data are on the correct order of
magnitude. However, many details need tobe worked out. The
site (component) sample size is one major problem area.
Although it is probably adequate for deriving general patterns,
there are two major areas where it is likely inadequate.
First, it is difficult to believe that the several hundred
year "gaps" represent times during which the area was not
utilized. Even granting that there were relatively few sites
in areas such as this during the low sea level stand intervals,
at least some should be archeologically recognizable. One of
these may be the single site (component) occurring in the 3300-
3100 B.P. interval. Also, it is possible that some of the
sites discovered during the Amoco survey were actually utilized
during these intervals although their components could not be
dated through seriation.
Second, given that the four temporal clusters represent
high sea level stands, and that the relative position of these
stands is generally correlated with the number of components,
it seems reasonable to assume that the number of components
within each of the 100-200 year intervals comprising-each of
the clusters should be indicative of short-term sea level
variability. However, given the relatively few components
in each cluster, it is difficult to determine if the resulting
component cluster curves represent intracluster sea level
variability, or, simply an inadequate component sample size.
An additional problem for interpreting this data in terms of
inter-/intracluster variability, is that the temporal units
are not of the same duration and are, therefore, not directly
comparable. Consequently, a larger component sample size and
more directly comparable temporal units are needed.
SEA LEVEL FLUCTUATIONS
In addition to obtaining a larger component sample size,
either from the Amoco area or from environmentally similar
areas nearby, coastal oyster shell middens in the Cooper River
estuary need to be examined, and dated. It is suspected that
the coastal oyster shell middens will tend to cluster in time
during the "gaps" suggested by the Amoco data. The reason
being that during lower sea level stands oyster productivity
should increase, due in part to a reduction in salinity in and
near optimal oyster bed areas that should approach the 5-30%
salinity optimum for oysters (Galtsoff 1964).
Finally, with the above archeological data in hand for the
lower Cooper River Valley, in conjunction with detailed environ-
mental reconstructions provided by geology and related disci-
plines, it will be possible to refine, or modify, our model.
It will then be useful to test the model in other nearby Lower
Coastal Plain areas. Only in this way will we be able to under-
stand the broader aspects of prehistoric behavioral systems as
representing, in large part, an adaptation(s) to variability in
the subsistence resource base over time and space.
Future Research in Geology
Our geological studies in the lower Cooper River Valley
are as yet incomplete. But we have learned enough from the
present work to realize that correlation exists between our
archeological and geological data. Our future efforts will be
directed toward additional coring in the marshes up and down
the valley in order to outline the limits of these peats. It
is hoped that as this work progresses, we shall encounter
additional basal peats, the dates of which (if confirmed to
indicate fresh to brackish water transitions through diatom
and palynological studies) will allow us to produce a sea level
rise curve for this area without lacunae. Such a curve will
permit us to interrelate many other aspects of human culture
both with respect to reconstruction of paleogeography in this
low coastal region, as well as, possibly, climate during the
last 8,000 years.
The archeological research was made possible through an
Environmental Impact Contract between the Amoco Realty Cor-
poration and the Institute of Archeology and Anthropology.
Dr. Robert Kratsas and Dr. Stephen Elbert of the Amoco Realty
Corporation are to be especially thanked for encouraging and
supporting the archeological research. Members of the Institute
of Archeology and Anthropology are also to be thanked. James
D. Scurry assisted with the field work. Dr. Paul E. Brockington,
Dr. Robert L. Stephenson and Mr. Glen T. Hanson reviewed the
paper and offered many useful comments.
BROOKS ET AL.
Newman and Pardi were supported by N.S.F. Grant No.
EAR 17-13666. Messrs. Leonard J. Cinquemani, Howard Craig
and James K. Chylinski were members of our geology field
sampling crew. Marion Newman prepared the samples for
We are indebted to Dr. Jon Sperling, Biology Dept.,
Queens University for identification of salt-brackish water
diatom assemblages in samples A-12 and A-14.
1963 Factors Influencing Vascular Plant Zonation
in North Carolina Salt Marshes: Ecology 44:445-456.
Behre, K.E., K. Brandt, J. Barckhausen, and H. Streif
1975 Guidebook to the 1975 Coastal Excursion of INQUA
Subcommission Shorelines of N.W. Europe, Niedersach-
sisches Landesame, D-3 Hannover 51, 14 pp., 18 figs.
Binford, Lewis R.
1968 Post-Pleistocene Adaptations. In New Perspectives
in Archeology edited by Sally R. Binford and Lewis
R. Binford, pp. 313-341. Aldine, Chicago.
Birdsell, Joseph B.
1968 Some Predictions for the Pleistocene Based on
Equilibrium Systems Among Recent Hunter-Gatherers.
In Man the Hunter, edited by Richard B. Lee and
Irven DeVore, pp. 229-240. Aldine, Chicago.
Brand, G., B.P. Hageman, S. Jelgersma, and K.H. Sindowski
Die Lithostratigraphische Unterteilund des Marinen
Holozans and der Nordseekuste. Geol. Jahrb. 82:
Brooks, Mark J., and James D. Scurry
1978 An Intensive Archeological Survey of Amoco Realty
Property in Berkeley County, South Carolina with a
Test of Two Subsistence-Settlement Hypotheses for
the Prehistoric Period. Institute of Archeology
and Anthropology, University of South Carolina Research
Manuscript Series, 147.
Colquhoun, D.J., T. Bond, and D. Chappel
1973 Santee Submergence. Geol. Soc. Am. Mem. 133:475-496.
Deetz, James F., and Edwin S. Dethlefsen
1972 Death's Head, Cherub, Urn and Willow. In Contemporary
Archeology, edited by Mark P. Leone, pp. 402-410.
Southern Illinois University Press, Carbondale.
SEA LEVEL FLUCTUATIONS
DePratter, Chester B.
1977 Environmental Changes on the Georgia Coast During
the Prehistoric Period. Society for Georgia
Archeology 5 (1&2), September.
DePratter, Chester B., and J.D. Howard
1977 History of Shoreline Changes Determined by Archeological
Dating; Georgia Coast, U.S.A. Transactions Gulfcoast
Assoc. Geol. Soc. 27:251-258.
Ford, James A.
1962 A Quantitative Method for Deriving Cultural Chronology,
Pan American Union, Washington, D.C.
1964 The American Oyster (Crassostrea Virginica). U.S.
Fish and Wildlife Service, Bull. 64.
1960 De Holocene Ontwikkeling van de Rijn-Maasmond.
Geol. en Mijnb. 39:661-670.
1969 Development of the Western Part of the Netherlands
During the Holocene. Geol. en Mijnb. 48:373-388.
1961 Holocene Sea Level Changes in the Netherlands.
PhD. thesis Leiden; Meded. Geol. Sticht. CVI 7.
Jochim, Michael A.
19 6 Hunter-gatherer Subsistence and Settlement: A
Predictive Model. Academic Press, Inc., New York.
LeBlanc, Steven A.
1975 Micro-seriation: A Method for Fine Chronological
Differentiation. American Antiquity 40:22-38.
1969 Eustatic and Climatic Changes During the Last 15,000
years. Geologie en Mijnbouw 48:389-399.
Noakes, L.E., S.M. Kim, and L.K. Akeis
1967 Recent Improvements in Benzene Chemistry for
Radiocarbon Dating Geoghem. et Cosmochem. Acta.
1960 La Plaine Maritime Entre Dunkerque et la Frontiere
Belge. Bull. Soc. Belg. et Geogr.
1976 Queens College Radiocarbon Measurements I.
BROOKS ET AL.
Pearson, Charles E.
1977 Analysis of Late Prehistoric Settlement on Ossabow
Island, Georgia. University of Georgia, Laboratory
of Archeology Series, 12.
Quarterman, Elsie,and Catherine Keever
1962 Southern Mixed Hardwood Forest: Climax in the
Southeastern Coastal Plain. Ecological Monographs
1972 Development of a New England Salt Marsh. Ecological
Schneider, Harold K.
1974 Economic Man: The Anthropology of Economics. The
Free Press, New York.
Smith, Bruce D.
1975 Middle Mississippi Exploitation of Animal Populations.
Museum of Anthropology, University of Michigan,
Anthropological Papers 57.
1975 Les Plaines du Nord de la France et Leur Bordure,
1973 Les Variations du Niveau Depuis 10,000 ans le Long
du Littoral Atlantique Francais, in Recherches sur le
Quaternaire Marin, Le Quaternaire, Centre National de
la Recherche Scientifique, pp. 114-135.
1974 Sea Level Changes During the Last 9,000 years in
North-west England. Geog. Jour. 140:18-42.
Columbia, S.C., and New York, N.Y.
AN EVALUATION OF WET SITE RESOURCES OF FLORIDA
Barbara A. Purdy
Organic deposits in the many and varied water-saturated
environments of Florida can provide data about past climates,
flora, fauna, and human occupations that are not preserved else-
where. Plant and animal remains from the mucks and peats of
Florida and their climatic implications have been studied for
decades, but the archaeological importance of these resources,
while recognized, has not been systematically examined.
Cultural material from wet sites can add a dimension to
prehistory that many people believed had vanished forever. It
is imperative that this aspect of Florida's heritage be investi-
gated before it is destroyed by the increased utilization of peat
deposits for agriculture and for future energy sources.
This paper reviews the present state of knowledge pertaining
to water-saturated sites in Florida and evaluates the potential
of these sites as valuable sources of prehistoric cultural material.
In addition, methods are proposed to assess the current status and
future requirements of the entire maritime heritage of Florida.
Occurrence and Age of Peat Deposits
Peat deposits are found in Florida in the following regions
and areas: (1) marshes and mangrove swamps of lowlying shores, in
lagoons and estuaries, and some depressions among dunes; (2) in
some of the large, nearly flat, poorly drained regions, especially
in the Everglades; (3) in river-valley marshes, such as the upper
St. Johns River and the Oklawaha River; (4) in some swamps particu-
larly of the tyty-bay or bay-tree types, and a few of the cypress
pond type that are common in most flatland regions; (5) in the
marshes bordering numerous lakes and ponds; (6) in some seasonally
flooded shallow depressions covered by a marsh or wet prairie type
of vegetation; (7) as deposits in the bottom of some lakes; and
(8) in layers buried under clay, sand, and rock strata (Davis
It is estimated that there are peat areas in Florida totaling
about 3,500 square miles and containing over 1,750,000,000 tons of
peat on an air dry basis (Davis 1946:2). Many peat soils have been
used for decades as agricultural lands and to a lesser extent for
horticulture, soil conditioning, and with fertilizers. Peat
deposits, not expansive enough for agricultural purposes,may be
utilized in the future for decentralized energy resources to meet
local needs or partial requirements. Utility companies are be-
coming interested in wood and other biomass products that may be
used as dilutants in conjunction with low-cost, high sulfur-
containing coal (Wayne H. Smith, personal communication, July 1979).
The Florida Anthropologist, vol. 32, no. 3, September 1979
Organic remains have been preserved, identified, and dated
to 37,000 years B.P. in Lake Annie, Highlands County (Watts 1975),
20,900 years B.P. along the Caloosahatchee River (Brooks 1968),
8,000 years B.P. in Mud Lake in north-central Florida (Watts 1969)
and in a slough adjacent to the Little Salt Spring Site near
Charlotte Harbor (Clausen et al. 1979). Early dates such as these
are limited but peat deposition on a geographically large scale in
the Holocene appears to have begun in Florida approximately 5,000
years ago (Gleason 1974:301). There is evidence that the rate of
deposition increased exponentially from 5,000 years B.P. to present,
averaging 8.4 cm/century (about 3.3 inches). It is calculated that
the initial slow peat development increased to 7.3 cm/century from
about 3,500 years B.P. to 1,200 years B.P. and to 16 cm/century from
1,200 years B.P. to present (Gleason 1974:301). The distribution
of peat deposits in Florida is extensive. Many of these deposits
contain archaeologically significant organic material.
In 1895, Frank Hamilton Cushing excavated hundreds of carved
wooden objects from an archaeological context at Key Marco, Florida
on the lower Gulf coast. The artifacts were discovered when a
resident of the area was digging garden muck from a mangrove swamp.
Much of the area of interest (The Court of the Pile Dwellers) was
underwater. Cushing relates that:
.excavation looked at first to be almost impossible.
I had a trench cut through the sea-wall to as great a depth
as possible without letting water in from the bay outside.
I then had a long trough of ship planks constructed and
placed on stakes driven deep into the muck bed, so that one
end rested over the excavation and the other, lower end, in
the mouth of the sluice-way through the sea-wall. Then
laying heavy planks over the boggy surface to furnish foot-
hold for the men, I set them at work baling. It was at
first like trying to bale out the sea itself. I soon
found that by thus proceeding each morning for a couple of
hours, the water could be kept sufficiently low to enable
us to excavate the entire place. Yet, even thus, much of
our search in the lower depths had to be made merely by
feeling with the fingers [emphasis mine] (Cushing 1897:28).
About the deposit, Cushing says:
The surface deposit throughout the entire court consisted
of a stratum of spongy black or dark brown muck, permeated
by both rotting and living rootlets. It was thin at the
margins but eighteen or twenty inches thick throughout
the middle. Below this was a somewhat thicker stratum of
brownish gray peaty marl, soft, tremulous, exceedingly
foul-smelling, and rich in the best preserved relics we
discovered. Underneath it, throughout the middle of the
court, was a less well-defined layer of the less peaty marl,
intermixed with shells and other debris, and also with
WET SITE RESOURCES
abundant ancient remains--which, indeed, we continued
to encounter even in the underlying, comparatively firm
shell and clay-marl bottom. This, however, although
nearly a foot and a half thick, we could not venture
to excavate, since the slightest opening made through it
into the sandy reef below let in a steady stream of water
from the sea (Cushing 1897:30).
Cushing's account accurately reports the depositional sequence
and reinforces the observations of Davis (1946) that stratified
peat deposits indicate changes in sea level during Recent times.
Radiocarbon analysis of a limited number of wooden artifacts
from Key Marco yielded dates from 1,000-2,000 years B.P.
It is unfortunate that the Key Marco remains did not await
the development of present-day excavation and preservation tech-
niques. Despite its uniqueness (masks, figurines, bowls, pestles,
toy canoes, paddles, float pegs, adze handles, etc.), the Key
Marco collection was largely neglected following the initial
excitement over its recovery. Cushing reports that the use of
trowels, flexible-pronged garden claws, and fingers destroyed
at least twenty-five percent of the wooden specimens. He also
describes how the objects warped, shrunk, split, or disintegrated
after being exposed to the light and air for only a few hours
(Cushing 1897:30-31). The specimens were never preserved and
today, those that remain are in an advanced state of degradation.
They could have provided an insight into the types of material
items made and used by prehistoric Floridians including selection
of raw materials and processes of manufacture. At the Ozette
site on the Olympic Peninsula, where an entire village was buried
bya mud slide, Daugherty (1974): (1) solved the problem of
excavation by pumping water away from the area, (2) averted the
destruction of the artifacts upon recovery by using a fine spray
of water to "excavate" instead of metal tools, and (3) immediately
applied preservation methods to the fragile, water-saturated
No other site in Florida has yet been discovered with as many
wooden artifacts as Key Marco, but isolated finds are fairly
frequent. In 1960, a figurine made of Brazilwood was recovered
in the Tomoka River during a dredging operation in the north end
of Tomoka State Park at Ormond Beach. Since Brazilwood trees are
found today only in Central America, the figurine had to have been
made from a piece of driftwood, imported, or the tree once grew in
Florida (Florida Conservation News, October 1974).
In 1965, an aboriginal dugout was found during dragline
operations at a phosphate mine near Zellwood, Florida. Radiocarbon
dated at 3,100 years B.P., this canoe is one of the oldest in the
Western Hemisphere (Bullen and Brooks 1967).
Wooden artifacts have been recovered on Kramer Island in
Lake Okeechobee. Several years ago the Florida State Museum
preserved a small wooden bowl from the island. This bowl is
now on display in the Belle Glade Public Library.
At the Fort Center Site (Sears 1971, 1974) two large eagle
totems have been excavated from Fisheating Creek. In the St.
Johns River near DeLand, an owl totem was recovered in the 1950s
when underwater utility lines were installed (Bullen 1955).
Other totems are described below. These may be similar to those
mentioned in historic accounts: ". .it was customary to place
carved images of birds, usually three, on the roofs of the temples.
In the town of Ucita in Tampa Bay, where De Soto landed, was a
temple upon which was a wooden bird with its eyes 'gilded'"
Skeletal material is found in an excellent state of
preservation in peat deposits. One large burial area is near the
southeastern end of Lake Okeechobee in Belle Glade. This site
was reported by Stirling (1934) and by Willey (1949) but no
further work was accomplished. The owner has now built a house
on the site thus precluding the possibility of obtaining valuable
information about the physical attributes of the Indians interred
there. Adjacent to the Little Salt Spring Site, Clausen has found
an Archaic period cemetery, more than 6,000 years old, containing
an estimated 1000 individuals that were buried in the peat. These
skeletal remains provide the best opportunity for physical anthro-
pological studies on an Archaic population since the discoveries
decades ago at Indian Knoll, Kentucky (Clausen 1979:613).
Throughout the summer and fall of 1977, the lakes and rivers
of Florida experienced drastic lowering of water levels as a result
of unprecedented drought conditions. As the water receded it
exposed strata that had entombed dugouts and other wooden items
for an unknown length of time. These objects survived only because
they had been buried in oxygen-free peat deposits. Once exposed,
they were vulnerable to attack by natural degenerating agents and
to vandalism. The well-known peat fires occur because peat be-
comes very flammable as it dries. In addition, drying destroys
the properties of peat and thus the peat itself as well as what
has been preserved in it. Once dried below 40 percent moisture,
peat does not readily reabsorb water (Davis 1946:46).
With grants from the National Endowment for the Arts and
the National Science Foundation an emergency program was initiated
to locate, retrieve, preserve, study, and display wooden artifacts
exposed because of drought conditions. Mr. Raymond F. Willis,
doctoral student in anthropology at the University of Florida,
served as graduate research assistant to the project. Four nearly
complete and five fragmentary canoes were recovered and ten
additional canoes were reported and investigated. Several privately
owned dugouts were examined. Two large carved wooden objects, a
pelican and a frog, w@re also recovered during the period of the
WET SITE RESOURCES
drought near Hontoon Island on the St. Johns River near the
area where the owl totem was found. A property owner was
running a dragline to build a marina when they were discovered.
Limited studies of the dugouts revealed information about
stylistic variations, construction techniques, and wood prefer-
ences and properties. For example, wood tissue examination of
31 canoes indicated that pine was used rather then cypress as
previously believed. Radiocarbon dates ranged from 300 A.D. to
1400 A.D. for the specimens retrieved during the project.
A large wooden bowl, two feet in diameter, was found in a
stream near Ft. Ogden in South Florida and was radiocarbon
dated at 1,000 years B.P. It is now being preserved in poly-
ethylene glycol at the Florida State Museum in Gainesville.
An important outcome of the emergency program was that it
made very clear that wet sites are valuable sources of pre-
historicartifacts. Now that we have become aware of this resource
and its potential, it is important to determine its extent and
Field Investigation An exploratory project should be initiated
to survey, inventory, and evaluate those areas of Florida that
are reputed to produce prehistoric organic cultural material
suggesting the location of wet archaeological sites in the
vicinity. Once this is accomplished an investigator can review
the wet sites resources in a region and decide if any site
provides data directly relevant to a problem of Florida pre-
history. If the site is to be sampled he can also decide how
the site could best be approached in terms of conservation and
excavation (MacDonald, personal communication, June 1979).
Wet sites have been defined as sites with a water-saturated
soil matrix below a water table or capillary fringe in which
cultural objects made of vegetal material have become waterlogged
and preserved (Croes, personal communication, 1979). Wet sites
are critical in understanding the prehistoric development of
woodworking technologies, art styles, raw material choices, etc.
This information will provide valuable data about prehistoric
complexes in Florida now known primarily from ceramic and lithic
A major justification for such an investigation is that most
wooden artifacts in Florida have been found accidentally during
dredging operations, agricultural pursuits, peat mining, or as
a result of shoreline recession. Specimens have been retrieved,
as found, with little consideration of the topographic, vege-
tational or climatic history, or the possibility that additional
objects may exist in the area. In contrast, a computer synthesis
of data would incorporate environmental characteristics useful
for site prediction plus ethnographic data and archaeological
report data on known site locations. Long range goals should
include the development of models which will enable predictions
about wet site locations and prehistoric economies.
Carbon-14 dates should be obtained from each discovered
wet site and immediate steps should be taken to protect any
organic objects encountered during the survey. Conservator
wood scientists should be hired to analyze wet wood from each
site to record the degree of degradation and determine what
techniques of conservation will be needed during any future
sampling of the site. The field of wet wood conservation is
moving away from a trial and error approach to a more analytic
and scientific one.
Preservation Water-saturated archaeological sites have produced
large quantities of well-preserved perishable materials. The
artifacts from these sites add a new dimension to our understanding
of past activities in maritime environments. Until recently we
believed that this dimension was lost forever. But with the
recovery of this new-found heritage comes a challenge and a
responsibility. Once removed from the maritime grave, objects
manufactured of organic materials will check, warp, and shrivel
if allowed to dry. Water-logged specimens must be protected.
Forty-five dugouts and numerous other cultural objects of
wood should be stabilized using the most satisfactory methods.
Many of these once waterlogged specimens are in an advanced
state of degradation. The woods from which the artifacts were
made need to be identified to determine (1) the kinds of plant
materials selected by the Indians and early settlers, (2) proper-
ties of the species that led to their selection, and (3) the best
course of action needed to preserve them.
While there is a need for a systematic investigation of wet
sites in Florida, immediate concern is the identification, preser-
vation, and restoration of what already exists before more
recoveries are made and subsequently ignored. If a program of
preservation is not initiated soon, the information to be learned
from these unique specimens will be irretrievably lost. This
situation occurred with the Key Marco specimens when they became
so degraded that molds could not be made; in addition, they were
so shrunken and warped that reproductions were impossible because
the original form was unrecognizable. When preservation and
restorationare completed, the objects can be displayed. They will
provide a wider understanding of technology and art works of the
early residents of Florida and will add appreciably to the know-
ledge now available about early people in North America.
Maritime Conference A coordinated statewide effort is needed to
outline the parameters and priorities concerning the many faceted
aspects of the maritime heritage of Florida. This course of action
can be justified by considering that Florida:
WET SITE RESOURCES
1. Has the most extensive coastline relative to land
mass of any state in the contiguous United States;
2. Was the first state explored and permanently settled
by Europeans; the intruders came by sea (Purdy 1977);
3. Has extensive river systems linking the interior to
the Gulf and Atlantic coast; virtually the entire
state of Florida is a maritime environment;
4. Is the only state with karst topography;
5. Had a native population that made and used dugouts;
Florida may well have the largest collection of pre-
historic watercraft in the world;
6. Has an Indian population today that makes and uses
dugouts (the Seminoles);
7. Has recorded numerous shipwrecks off her coast;
8. Has major present-day maritime enterprises such as
commercial and pleasure fishing, boating, etc.;
9. Is one of the fastest growing states in the nation
with population pressures encouraging extensive
development especially along the coastline;
10. Is engaged in offshore drilling activities;
11. Promotes tourism (the second largest source of revenue
in the state) and has developed elaborate recreational
and other facilities to accommodate tourists, especially
12. Has quantities of cultural remains in lakes, rivers,
sinks, and along the coast, both historic and prehistoric,
that need to be recovered by underwater archaeologists
13. Is concerned about coastal erosion, about past and
present climates, and about the effects of weather such
as storms and droughts on the coastline and water supply;
14. Has already destroyed a great deal of her maritime
heritage. We must take stock of what remains;
A three-phased project is proposed: (1) Meet with repre-
sentatives of organizations involved in a broad spectrum of
activities relating to the maritime heritage of Florida as an
initial integrated effort to address this subject and discuss
the needs for its preservation; (2) plan and hold a conference
with invited speakers from diverse disciplines asked to focus
on a variety of topics pertaining to the impact of Florida's
waterways on the history of the state; and (3) publish the
proceedings of the conference. An overview approach of this
kind may serve as a prototype for other states to follow.
A multidisciplinary orientation toward the preservation
of America's Maritime Heritage in Florida will (1) concentrate
attention on the problem and present an opportunity to explore
the many facets involved; (2) provide an efficient vehicle to
acquaint experts with the work of others; (3) eliminate or
minimize overlap of effort and lead to cooperative research;
and (4) alert the public to the endangered nature of their
Excellent preservation of perishable objects occurs only
at wet sites. These locations are recognized resources that
add new dimensions to prehistory. They provide insights about
the kinds of organic materials selected by people in the past
and about the technology used to manufacture cultural items.
One gains an appreciation of the important interpretive value
of organic materials when one considers that they represent at
least 85 percent of the cultural inventory of all human groups.
Thus a study of maritime environments provides not only a way
of learning about the influence the sea had on the daily life
of people, their ability to communicate with others, and their
link to needed resources, but it also becomes a maritime grave-
yard that preserves artifacts of the broader culture. These
"graveyards" need to be systematically explored.
In addition, all aspects of the Maritime Heritage of Florida
need to be assessed. It would be most unusual to find an indi-
vidual or organization possessing all of the skills necessary to
study and preserve that heritage. It is essential that a state-
wide effort be coordinated to preserve the entire Maritime Heritage
of Florida. A conference is recommended to initiate this effort.
It is time to turn from a hit and miss Sunday science commitment
that is often inefficient and more costly in the long run to a
well-planned, scientifically-based program. Proposals have been
submitted to the Maritime Preservation Grants Program of the
National Trust for Historic Preservation requesting funds to
proceed with the investigations as outlined in this paper.
1968 The Plio-Pleistocene of Florida, with special reference
to the strata outcropping on the Caloosahatchee River.
Miami Geol. Soc., Second Annual Field Trip Guidebook,,
WET SITE RESOURCES
Bullen, Ripley P.
1955 Carved Owl Totem, Deland, Florida. Florida
Bullen, Ripley P. and H.K. Brooks
1967 Two Ancient Florida Dugout Canoes. Quarterly Journal
of the Florida Academy of Sciences 30:97-107.
Clausen, C.J., A.D. Cohen, Cesare Emiliani, J.A. Holman, and J. J.
1979 Little Salt Spring, Florida: A Unique Underwater Site.
Croes, Dale R.
1979 Personal Communication. Dr. Dale R. Croes is at
the Washington Archaeological Research Center, Pullman,
Washington. Dr. Croes along with Dr. MacDonald
organized a symposium entitled Wet/Frozen Archaeological
Sites--The Strategies, Techniques, Research, and
Potential which was held at the Annual Meeting of the
Society for American Archaeology, April, 1979 in
Vancouver, British Columbia.
Cushing, Frank Hamilton
1897 A Preliminary Report on the Exploration of Ancient
Key-Dweller Remains on the Gulf Coast of Florida.
Proceedings of the American Philosophical Society
Daugherty, Richard D.
1974 Hunters of the Whale. William Morrow and Company,
Davis, John H.
1946 The Peat Deposits of Florida: Their Occurrence, Develop-
ment, and Uses. State of Florida, Department of
Conservation, Geological Bulletin No. 30.
Gilliland, Marion Spjut
1975 The Material Culture of Key Marco, Florida. The
University Presses of Florida, Gainesville.
Gleason, Patrick J., Editor and Compiler
1974 Environments of South Florida: Present and Past.
Miami Geological Society, Memoir 2.
Kenner, Captain Joe
1974 Rare Art from Florida's Past, Florida Conservation
News, October 1974.
MacDonald, George F.
1979 Personal Communication. Dr. George F. MacDonald
is Senior Archaeologist, National Museum of Man,
Ottawa, Canada. Dr. MacDonald along with Dr. Croes
organized a symposium entitled Wet/Frozen Archaeo-
logical Sites--The Strategies, Techniques, Research,
and Potential which was held at the Annual Meeting of
the Society for American Archaeology, April, 1979,
Vancouver, British Columbia.
1971 Food Production and Village Life in Prehistoric
Southeastern United States. Archaeology 24(4):93-102.
1974 Archaeological Perspectives on Prehistoric Environment
in the Okeechobee Basin Savannah. In Environments
of South Florida: Present and Past, edited by Patrick
J. Gleason. Miami Geological Society Memoir 2, pp.
Smith, Wayne H.
1979 Personal Communication. Dr. Wayne H. Smith is the
Director of the Environmental and Natural Resource
Center, University of Florida.
Stirling, Matthew W.
1934 Smithsonian Archaeological Project Conducted Under the
Federal Emergency Relief Administration, 1933-34. In
Annual Report of the Smithsonian Institution, 1934,
Swanton, John R.
1946 The Indians of the Southeastern United States. Bureau
of American Ethnology Bulletin 137.
1969 A Pollen Diagram from Mud Lake, Marion County,
North-Central Florida Geol. Soc. Am. Bull.
1975 A Late Quaternary Record of Vegetation from Lake Annie,
South-Central Florida. Geology 3:344-346.
Willey, Gordon R.
1949 Excavations in Southeast Florida. Yale University
Publications in Athropology, No. 42.
AN ABORIGINAL CANOE FROM LAKE APOPKA, FLORIDA
Arthur F. Dreves
Lake Apopka is one of the larger lakes in the state of
Florida, located about 20 km west-northwest of Orlando. As
the headwaters of the Oklawaha River, the lake offered an
ideal location for aboriginal settlement, as is evidenced by
the significant number of encampments that have been uncovered
by contemporary development, site surveys, and excavations
(e.g., Dreves 1974, Bullen, Jahn, and Brooks 1974).
The lake was not deep; indeed, it was a fairly simple
matter for local farmers to have the northeastern quarter of
the lake diked off and drained in the 1930s so they could farm
its rich bottomland (Dreves 1974). The U.S. Soil Conservation
Service (1960) classifies the lake bottom as Everglades Mucky
Peat, which it says is the result of the accumulation and de-
composition of "sawgrass and other sedges, lilies, myrtle
bushes, and other grasses overlying nearly neutral or alkaline
sands and sandy clays..." (USSCS 1960:14). This description
suggests that the northeastern part of the lake more nearly
resembled a shallow, grassy marsh with some open areas of water
in its original configuration.
Thus, it is not surprising that some of the farmers who
today cultivate the lake bottom have reported unearthing pieces
of aboriginal canoes, pottery, projectile points, and other
accoutrements of the early peoples who settled on the lake's
shores. The marshes at the lake's edges no doubt provided
much of the aborigines' food supply, as well as a source of
protection from the elements (especially strong winds) and a
place for recreation.
Search for Canoe
In the spring of 1976, one of the Lake Apopka farmers,
Mr. Charles Grinell, was disc-harrowing one of his fields near
the north shore of the lake [about 2 km due south of the town
of Zellwood, in the northeast 1/4 of Section 34, Township 20S,
Range 27E, in Orange County (Fig. 1)]. Each time his tractor
made a pass over one particular row, pieces of wood became
lodged in the harrowing machinery. Each piece was firm, dry
pine with a peculiarly curved shape and bore signs of having
been burned. Subsequent plowing unearthed similar pieces which,
when matched with pieces previously brought up by the harrow,
fitted together to form what appeared to be a part of an up-
swept bow from an aboriginal canoe.
Grinell, a member of the John Young Museum in Orlando,
contacted the Central Florida Anthropological Society (CFAS)
through the Museum and agreed to temporarily discontinue harrow-
ing that field while a search was made for the remainder of the
The Florida Anthropologist, vol. 32, no. 3, September 1979
scale in meters
Figure 1. USGS topographic map of area showing location
of canoe and old shoreline of Lake Apopka in relation
to the town of Zellwood. From USGS (1975).
canoe. The field in which the pieces had been unearthed
measures almost 0.8 square kilometers, and the particular
north-south row in which most pieces had appeared was 5mwide
by almost 400 long. A visual search of that row turned up
about a half-dozen smaller wood fragments roughly 195m south
of the north end of the field. As there was no other organic
matter in the peat with a mass nearly as large or as dense as
the wood fragments in question, it was assumed that all of the
wood recovered up to that point came from the canoe.
Measurements indicated that the harrowing equipment was
penetrating the peat to a depth of about 35 cm. Because only
bow sections from the canoe were being brought to the surface
by the harrow, it was assumed (correctly, it turned out) that
the major portion of the vessel lay more than 35 cm below the
surface, and parallel to it, or that the vessel had settled
into the peat at an angle, leaving only the bow close enough to
the surface to be sheared off by the harrow.
Because the harrow often dragged the wood fragments for
some distance before they would rise to the top of the disc
assembly (and thus become visible to the harrow operator or
other observer), pinpointing the location of the canoe in the
400-meter-long row became a difficult task. Excavation of the
entire row (or a significant part of it) was practically impo-
sible. Instead, a more detailed search was undertaken in an
area one row (5 m) wide by about 30 m long. This probe con-
sisted of inserting one meter metal rods into the peat at
regular intervals, to a depth of about 65-70 cm, in hopes of
striking the main body of the vessel.
Three weeks of such probing (on a weekend basis) revealed
nothing because, as became apparent later, the canoe wood was
so thoroughly water-logged that probing rods simply slid through
the wood without signaling its presence. (Because the bow
fragments were closer to the surface, where the peat periodi-
cally dried to an almost powdery consistency, they had never
become thoroughly water-saturated, and thus remained dense
enough to be caught by the harrow, leading one to believe that
probing rods would not penetrate the wood).
The search was called off pending discovery of a more
effective and yet practical means of locating the canoe, and
Grinell resumed harrowing and planting the field. In the spring,
1977, Grinell again began striking fragments of the canoe, and
because no better means of exploration had been devised, rod
probing was tried again.
This second attempt was also unsuccessful, leading to
concern among CFAS members that if the canoe was not found soon
the bow end of the vessel would be completely destroyed. That
would leave the remainder of the vessel below the cutting depth
of the harrow and thus completely undetectible by above ground
The second fruitless attempt was the subject of a small
article in the Orlando Sentinel-Star newspaper (O'Neill 1977)
which caught the attention of Ms. JoAnn Jones, an area resident
who is a full-blooded Winnebago Indian and self-proclaimed
psychic. Ms. Jones contacted the CFAS and expressed her desire
to pursue the search for the canoe through the use of psychic
abilities. On May 6, 1978, Ms. Jones and three other psychics
joined CFAS members and others at the site to renew the search.
As before, the effort was unsuccessful, but Ms. Jones felt
certain that the vessel could be located and offered to make
one last attempt.
Two weeks later, on May 21, 1978, Ms. Jones and the three
other psychics tried again. Working in a systematic fashion,
the four criscrossed the field, allegedly hoping to "sense" the
presence of the canoe. After about three hours of traversing
the field, the four converged on a particular location very
near where the initial fragments of wood had surfaced. A small
(one square meter) excavation was made, and at a depth of 30 cm
the harrow-damaged bow of the canoe was uncovered.
Excavation of Canoe
The remainder of the vessel was easily uncovered in about
an hour's time by excavating a series of 1 by 1 m units along
the almost exactly northeast-southwest axis of the canoe (Fig. 2).
The main section of the canoe was situated about 300 m due west
of the old northeastern shore line of the lake, 50 cm below the
mean surface of the peat, and practically parallel to the surface.
At this depth the main section of the vessel was entirely
within a stratum of extremely moist Everglades Mucky Peat that
began about 35 cm below the surface. The moisture content of
the peat was so high that only 30 cm2 of peat yielded almost
60 ml of water when squeezed dry. The main deck of the canoe
was almost equally as saturated.
As mentioned earlier, the bow section of the vessel was in
a far drier, almost powdery stratum of peat that extended from
the surface to about the 35 cm level. The transition zone
coincided almost exactly with the maximum depth to which the
harrowing equipment had penetrated. On a recommendation from
the Florida State Museum, the vessel was covered with a heavy
sheet of dark plastic soon after excavation to keep the wood
moist and prevent fungi from developing on the canoe's surface.
Description of Canoe
As it lay in the ground (Fig. 3) the canoe measured 4.05 m
in length, .51 m in width (at the upswept--presumed bow--end),
and tapered to about .41 m in width at the presumed stern. The
Figure 2. Excavation showing exposed Figure 3. Composite photo-
canoe. Arrow on main deck indicates north- graph of canoe in situ during
west. Note harrow-damaged bow sections in excavation.
Figure 4. Views of reassembled bow sections; left, exterior,note
darker tint on some portions, the result of burning as the
canoe was hollowed out of the tree; center, exterior, showing
prow face which is considerably lighter in color than worked
(burned) sections of bow below it; right, profile; note prow
face and severe checking and cracking of wood due to warpage
main deck of the craft averaged 8.0 cm in thickness. As
mentioned earlier, the upswept bow was first encountered about
30 cm below the surface, and the main section of the vessel
lay at 50 cm, leaving about 20 cm of the bow in situ. Whether
or not this 20 cm represents the original height of the bow is
not known for sure, as the harrow may have destroyed the upper-
most part of the bow. It seems, however, that most of the bow
has been recovered, and that its full height does not substan-
tially exceed 20 cm above the bottom of the craft.
The majority of the vessel was badly checked and cracked,
with the exception of the bow and stern sections, which had not
been totally saturated with water by virtue of their having been
in the drier stratum of peat (within 35 cm of the surface).
The bow and stern were, for the most part, intact, except where
the harrow had clipped off the uppermost pieces of each. The
canoe was constructed from what must have been a large (over
50 cm diameter) pine, perhaps of the long-needled variety still
common throughout most of central Florida. The surface of the
canoe was quite smooth except for check marks, cracks, and
some signs of the charring and burning employed in the craft's
Several pieces of pine bark were found against the vessel's
bow as well as numerous bits of charcoal, which may have been
lodged in cracks in the vessel and floated free as the vessel
sank to what was then the bottom of the lake. During the initial
excavation, one rather thick, straight, black hair was extracted
from one section of the bow. As it was solidly pinched between
two joined pieces of pine, it seems reasonable to speculate that
the hair either came from an occupant of the craft or some type
of animal that may have been transported in the craft (for in-
stance, wild game broughtback to camp from a hunting expedition).
The U.S. Geological Survey map of the area suggests that
the water depth at that part of the lake in which the canoe sank
was about 0.8 to 1.2 m (USGS 1975). This raises the obvious
questions of why the canoe sank and why it was not recovered by
its owners from such a shallow grave.
Due to a total lack of anthropomorphous remains or accouter-
ments (pottery, stone tools, etc.) associated with the vessel,
one is led to believe that the craft was unoccupied and may not
have been in use at the time of its sinking. The vessel may,
however, have gotten away from its occupants or its mooring in
a storm and could have become mired in the thick grasses which
made up the marshes at the lake's edge. No doubt it would have
been difficult for the owner of the canoe to return to the area
and raise the vessel once it became trapped in the sawgrass,
waterlogged, and finally submerged. It probably sank quickly
in what was then a thick, fluid layer of decomposing grasses and
mulch on the lake bottom.
As discovered during excavation, the main deck of the
canoe was virtually flat. Only the upswept bow and a smaller,
also slightly upswept stern protruded above the deck level.
(The protrusion of the bow and stern was obvious, because
both sections had been hit by the harrowing equipment, while
the main deck remained unscathed). Two hypotheses have been
advanced for the flatness of the main deck: one, which holds
that the craft was designed as a floating plank with upswept
bow and stern (a cross between a "conventional" canoe and
today's surfboard), and the other, which holds that the craft
did originally have small but distinct sides. The second
hypothesis presumes that the weight of the peat overburden,
combined with the water saturation and subsequent weakening
of the wood, crushed the sides, and flattened what was an
almost circular hull into a flat slab.
Figure 3 depicts the canoe in its present flattened
configuration, with splits that may be marking the places
where the sides broke loose running parallel to each other
through the center of the vessel. Firsthand observation of
the canoe, however, usually leads one to believe that the flat
("surfboard") configuration was its original form, and that
the splits are merely major stress marks.
Deposition of Canoe
After excavation and before covering the canoe with
protective plastic, all measurements were taken and photographs
made. The structurally sound, relatively dry sections of the
bow were removed in order to prevent them from becoming water-
logged under the plastic sheet and thus possibly subject to
decomposition. The bow has been reconstructed from fragments
unearthed by the harrowing equipment and from those fragments
removed during excavation (Fig. 4). The canoe was sealed
under the protective plastic and left in situ pending develop-
ment of a workable plan for removal, storage, reconstruction,
and display of the vessel.
The Florida State Museum expressed no interest in taking
possession of the craft due to its condition and the John
Young Museum in Orlando regretfully declined the CFAS' offer
of the canoe on the grounds that space for the canoe's resto-
ration was not available. Additionally, the CFAS was unable
to muster the financial resources necessary to complete the
removal and restoration of the canoe.
The 1978 summer rainy season inundated the site, leaving
the canoe in a water-filled pit, and by the fall of 1978, Charles
Grinell became anxious to resume farming of his field. The
excavation was backfilled over the protective plastic, and the
canoe remains in its original resting place for re-excavation at
some future date.
The author would like to acknowledge the assistance of
the following persons in this project: Mr. Charlie Grinell
for bringing the canoe to the attention of the CFAS and allowing
excavations to proceed on his property; Ms. JoAnn Jones for
locating the vessel after conventional methods failed; members
of the Central Florida Anthropological Society of Orlando for
assistance in excavating the canoe site; and Rick Dreves for
editorial and photographic assistance.
Bullen, Ripley P., Otto Jahn, and Mark J. Brooks
1974 Some Tests at the Zellwood Site (Or-17) beside
Lake Apopka, Florida. Florida Anthropologist
1974 Archaeological Investigation of 8-0r-17: An
Early Aboriginal Campsite on the Shore of Lake
Apopka, Florida. Florida Anthropologist
1977 Archaeologist Hoping to Ambush Indian Canoe.
Orlando Sentinel Star, September 25, 1977, p. 4B.
Sentinel-Star Company, Orlando, Florida.
U.S. Geological Survey
1975 Apopka Quadrangle topographic map, 1:24,000 scale,
7.5 minute series. 1960 edition, photorevised
1975. Washington, DC.
U.S. Soil Conservation Service.
1960 Soil Survey of Orange County, Florida. Series 1957.
Washington, DC: US. Government Printing Office.
Winter Park, Florida
DETERMINATION OF SITE FUNCTIONS
THROUGH THE ANALYSIS OF MODIFIED BONE
Ronald L. Wallace and Susan Jacquith
In recent years, archeologists have devoted considerable
attention to the discovery of animal bones that were used as
tools. The analyses have ranged from Raymond Dart's proposed
"osteodentokeratic" technology to Hind Sadek-Kooros' experi-
mental bone-fracturing investigation of the Jaguar Cave site
in Idaho (Sadek-Kooros 1972). Unfortunately, only minimal
attention has been devoted to the examination of human-modified
bones (whether tools or non-tools) as a means of determining
the activities that occurred at a site. The present paper is
a preliminary statistical investigation of a sample of modified
bone to determine which (if any) of three hypothesized activi-
ties were present at an aboriginal site in central Florida.
Hypothesis and Methodology
The region selected for the experiment was the upper St.
Johns River, directly east of the present city of Orlando,
Florida. Until the time of European contact in the sixteenth
century, the area was apparently inhabited by groups practicing
a hunter-gatherer mode of subsistence. Although maize and squash
cultivation were adopted possibly as early as 500 B.C., regional
subsistence activities were still largely devoted to the procure-
ment of wild plant and animal species until the beginning of
fully historic times. As a consequence of this long-term riverine
adaptation, one of the most frequently encountered types of sites
in the region is a low, circular shellfish midden containing
ceramic and zooarcheological materials.
Animal bones for the analysis were excavated from 8Se32, a
shell-fish midden dating from the early Archaic period (5000 B.C.)
until European contact (A.D. 1500). The total population was
16,760 animal bones which were taken to the laboratory for subse-
A stratified random proportional-allocation sample was taken
from the population of 16,760 bones. Appropriate sample sizes
for the two levels and each provenience was determined by the
N = Wi
N2D+ .Ni 3
The Florida Anthropologist, vol. 32, no. 3, September 1979
WALLACE AND JACQUITH
when N is the sample taken from each level, N. is the number
of bones removed from each provenience, 6 is the variance,
Wi is the proportion of total bones within a level (N) repre-
sented by the bones within a provenience and D is the signifi-
chance level (1.96)2. Quantities of bones selected for the
stratified random sample are presented below in Table 1.
Table 1. Provenience and sample sizes
LEVEL ONE LEVEL TWO
PROVENIENCE BONE CONTENT SAMPLE PROVENIENCE BONE CONTENT SAMPLE
479N486E 1183 40 500N497E 570 4
482N501E 803 27 509N497E 87 1
509N497E 69 3 512N497E 446 3
500N497E 213 7 497N486E 11207 76
512N497E 482 16 482N501E 1700 12
2750 93 14010 96
The sample (93+96=189) was examined in the University of
Central Florida archeological laboratory by the naked eye and
by a Bausch and Lomb binocular lens microscope with 250X
magnification. The bone attributes (discussed below) were
usually visible to the naked eye. The binocular microscope was
used for cross-verification, and for the closer examination of
weathered surfaces. In the event of disagreement between the
eye and the microscope, the evidence from the microscope was
accepted. Each bone was individually labelled with black India
ink to facilitate cross checks of the analysis. Attributes were
recorded on cards for subsequent key punching of data. Eight
binary attributes were selected for the analysis. These included:
charred, notched, incised, split, sliced, striated, edge-modified,
It was hypothesized that cooking and/or butchering and/or
tool use-manufacture activities were present at the 8Se32 site.
Bones attributes were selected with these activities in mind.
The presence of cooking would be indicated by charring; butchering
would be indicated by notching and/or incising and/or splitting
and/or slicing; tool use-manufacture would be indicated by
polishing and/or edge-modification and/or striation. Definitions
of attributes are presented below in Table 2.
Table 2. Attributes and definitions.
Any bone with microscopic evidence for
discoloration of the surface toward a more
darkened appearance as a result of direct
thermal modification of the bone rather than
from associated ash or charcoal materials
is charred. All other bones are uncharred.
Category designations: Charred: CH. Uncharred UNCH.
Any bone with microscopic evidence of a v-shaped
cut at any point on its surface is notched. All
other bones are un-notched.
Category designations: Notched: NO.
Any bone with microscopic evidence of cut lines
along its surface is incised. All other bones
are unincised. Category designations: Incised:
INCS. Unincised: UNINCS.
Any bone with microscopic evidence of cultural
cross-sectioning by cuts that are made parallel
to the disphysis is split. All other bone is
unsplit. Category Designations: Split: SPL.
Any bone with microspic evidence of a cut surface
surrounded by natural bone is sliced. All other
bones are unsliced. Category designations:
Sliced: SL. Unsliced: UNSL.
Any bone with microscopic evidence of shallow,
multiple lines on its surface. All others are
unstriated. Category designations: Striated:
STR. Unstriated: UNSTR.
Any bone with microscopic evidence of an angle
created by the joining of two cut surfaces or
by the joining of cut and natural surfaces is
edge-modified. Others are non-edge modified.
Any bone with microscopic evidence of a lustered
surface is polished. All other bone is unpolished.
Category disignations: Polished: POL.
Microscopic and naked-eye examination of the sample deter-
mined that 97 of the 189 bones were characterized by one or more
of the eight attributes. Thus, 52% of the stratified random
sample has been modified through some sort of cultural activity.
WALLACE AND JACQUITH 125
We wished to determine if these culturally-modified bones
could effectively distinguish between proveniences. This
would provide indirect statistical evidence of activity areas
located at different portions of the site. Then, through a
tabular presentation of bone attribute frequencies, we wished
to determine the relative significance of specific cultural
activities and the degree to which this had changed throughout
The technique involved in the first investigation is known
as Discriminant Analysis. It is a measure of the probability
of correctly assigning an individual to its group, when provided
with the individual's characteristics. A high probability of
correct assignment is diagnostic of a highly hetereogenous
sample consisting of very distinctive subgroups. This analysis
was performed for both levels within the site, using the
Statistical Package of the Social Sciences, SPSS-6, 1976
edition, program DISCRIMINANT. The results seem indicative of
a hetereogenous site. The probability of correct provenience
assignment of a bone from Level 1 simply on the basis of its
attribute characteristics was 54.84%, and the probability for
Level 2 was 64.58%. (It should be bonne in mind that the
probability of correct assignment on a purely random basis is
1/5 or 20%.)
The relative significance of the different cultural activi-
is reflected by a frequency tabulation. As is apparent in Table
3, the majority of culturally-modified bones were characterized
by one or more attributes of butchering. Cooking and tool use-
manufacture activities were additionally present at the 8Se32
site. These latter appear to be of lesser significance, and
the situation seems stable throughout time.
Table 3. Culturally modified bones at 8Se32.
Bones with Cooking Bones with Bones with Tool
Attribute (CH) Butchering Use/Manufacture
Attribute (s) Attribute (s)
Unit Level I Level II Level ILevel Level II Level I Level II
500N 1 2 2 3 0 2
509N 0 0 2 0 0 0
482N 501E 4 0 8 7 0 1
497N 485E 1 12 10 19 6 5
512N 497E 2 0 6 1 2 1
level 8 14 28 30 8 9
in sample 22 58 17
In a recent discussion of "primitive bone fracturing".
Sadek-Kooros has noted that
Bone is not wood and it is not horn. Better
described as an "animal stone" its properties
can be studied, like flint, by experimentation,
and by laboratory and statistical analyses....
Diagnostic criteria of bone-tool manufacture
and bone-tool use should be employed, therefore,
in studying this massive category of potential
technological information....before it is
dropped into paper bags and shipped off to
zoologists and paleontologists (Sadek-Kooros
With which notion we strongly concur. We would only add to this
view our own contention that animal bone is subjected to various
kinds of human modification in addition to "bone-tool manufacture
and bone-tool use". Thus, the amount of retrievable cultural
information potentially involved in modified-bone studies is even
greater than Sadek-Kooros has suggested.
The present preliminary study (done on an admittedly small
sample) suggests that midden depositions of the Upper St. Johns
River were areas of animal butchering, cooking, and tool-
manufacture activities, with butchering activities predominant
among the three. But in the process of eliciting this cultural
information, we discovered a number of methodological problems
which inevitable must be dealt with in further studies.
The most obvious problem of all is that of the "pseudo-
artifact" (although this is certainly not confined to studies
of bone). Effects of weathering, animal gnawing, and the pressure
of the ground can simulate the modification by human agency.
(_Miller 1969, 1975). Additionally, since animal bone can
deteriorate very quickly (the deterioration frequently beginning
at the surface), bone that was in fact subject to cultural
modification may not be recognizable as such.
The problem of recognition suggests another research priority:
the spatial distribution of animal bones. Daly has demonstrated
that the carcasses of animals are likely to be butchered closer
to the kill than to the habitation site, particularly if a
considerable traveling distance is involved or if the animal in
question is very large (Daly 1969:146). For this reason it is
conceivable that an archeological site could consist almost
completely of unmodified bones. Taphonomy (the study of the
spatial arrangement of carcass bones) could conceivably aid in
the recognition of such sites, and in distinguishing them from
areas of animal bone remains in which no sociocultural factors
were involved (Leakey and Lewin 1977:83, 1978:17).
WALLACE AND JACQUITH
Finally, a refinement of animal bone analytic categories
to a degree that is comparable to the study of lithic materials
will be crucial to the development of this technique. Given
these developments, there is a good possibility that the familiar
midden depositions of southeastern United States will be more
informative with regard to prehistoric activities than many
archeologists have been led to suspect.
1969 Approaches to Faunal Analysis in Archeology.
American Antiquity 34:146-153.
Leakey, R. and R. Lewin
1977 Origins. E.P. Dutton, New York.
1978 People of the Lake. Anchor, Garden City.
1969 A Study of Cuts, Grooves, and Other Marks on
Recent and Fossil Bone: I, Animal Tooth Marks.
Tebiwa 12(1), Journal of Idaho State University
1975 A Study of Cuts, Grooves, and Other Marks on
Recent and Fossil Bone; II, Weathering Cracks,
Fractures, Splinters, and other Natural Phenomena.
In Lithic Technology: Making and Using Stone Tools,
edited by Earl Swanson, pp. 211-226. The Hague,
1972 Primitive Bone Fracturing: A Method of Research.
American Antiquity 37:369-382.
STATEMENT OF OWNERSHIP, MANAGEMENT, AND CIRCULATION
In accordance with U.S. Postal Regulations No. 132.622, the
following statement of OWNERSHIP, MANAGEMENT, AND CIRCULATION
for THE FLORIDA ANTHROPOLOGIST is included in this issue.
The Florida Anthropological Society, Inc.
c/o Norcott Henriquez, Resident Agent
1510 Dewey Street
Hollywood, FL 33020
Officers of the Society
See inside cover for names
Total number of copies
printed varies from 1018
Sold to UF Library for
Total paid circulation
Office use, reprints,
Average number of
copies each issue
June, 1979, vol.
32, no. 2
Back issues of the FLORIDA ANTHROPOLOGIST, if available,
may be purchased from the Treasurer. Volumes 1-30 are $6.00
per volume ($1.50 for single and $3.00 for double numbers);
Volume 31 and subsequent volumes are $8.00 per volume ($2.00
for single numbers). Volumes 25:2, pt. 2 (Publication No. 6),
28:3, pt. 2 (Publication No. 7), and 31:2, pt. 2 (Publication
No. 9) are $2.00 each. Volumes 1; 2:3-4; 5:1-2; 6:1, 4; 7; 8;
9; 10; 11:1,2,4; 13:1; 17:1; 18:1; 20:1-2; 24:1; 25:3-4; 26:1: and
30:1 are out of print.
Copies of the FLORIDA ANTHROPOLOGICAL SOCIETY PUBLICATIONS,
a monograph series, may be purchased as follows: Numbers 5, 6,
7, and 9 at $2.00 each, Number 8 at $3.00. Numbers 1, 2, 3, and
4 are out of print.
Orders with remittances for any of the above should be
sent to the Treasurer whose address is given on the inside of
the front cover. The Society has no resale discount for
dealers or resale agents. Please allow four to six weeks for
delivery of back issues.
INFORMATION FOR AUTHORS
THE FLORIDA ANTHROPOLOGIST publishes original papers in
all subfields of anthropology with an emphasis on archeology.
Contributions from allied disciplines are acceptable when
concerned with anthropological problems. The journal's geo-
graphical scope is Florida and adjacent regions.
Manuscripts should be double-spaced and typed on one side
only of 8 1/2 by 11 inch paper. Authors should submit the
original and one copy of their manuscript. For matters of
style and reference see a recent issue of the journal. Text
references should be similar to "(Smith 1970:44-45; Jones 1972:
P1. II, b)." Footnotes are normally not accepted.
All illustrative material--line drawings (in India ink)
and photographs--should be included in one numbered series of
"Figures." Each figure must be accompanied by a brief descrip-
tive caption; parts of figures are designated by lower case
letters. All tabular material must be included in a separately
numbered series of "Tables." Authors should examine a recent
back issue of the journal for figure and table layouts.
Twenty-five reprints without covers of each article are
furnished each author or group of authors.