Material Information

Zooarchaeology papers to honor Elizabeth S. Wing
Series Title:
King, F. Wayne
Porter, Charlotte M., 1948-
Wing, Elizabeth S
Place of Publication:
Gainesville, Fla.
Florida Museum of Natural History
Publication Date:
Copyright Date:
Physical Description:
208 p. : ill., maps ; 28 cm.


Subjects / Keywords:
Animal remains (Archaeology) ( lcsh )
Environmental archaeology ( lcsh )
Paleozoölogie ( gtt )
City of Gainesville ( local )
Florida Museum of Natural History ( local )
Bones ( jstor )
Biomass ( jstor )
Archaeology ( jstor )
bibliography ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references.
General Note:
Cover title.
General Note:
"Publication date: 07-31-03"--P. 2 of cover.
Vol. 44, no. 1, pp. 1-208
Statement of Responsibility:
F. Wayne King and Charlotte M. Porter, editors.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright held by the Florida Museum of Natural History, University of Florida. All rights reserved. Text, images and other media are for nonprofit, educational, and personal use of students, scholars, and the public. Any commercial use or republication by printed or electronic media is strictly prohibited without written permission of the museum. For permission or additional information, please contact the current editor of the Bulletin at
Resource Identifier:
52975771 ( OCLC )
0071-6154 ; ( ISSN )


This item has the following downloads:



















































































































































































































Full Text




F. Wayne King and Charlotte M. Porter, Editors

Vol. 44, No. 1, pp. 1-208




irregular intervals. Volumes contain about 200 pages and are not necessarily completed in any one calendar year.
The end of a volume will be noted at the foot of the first page of the last issue in that volume.

MARGARET E.B. JOYNER, Managing Editor

Send communications concerning purchase or exchange of the publication and manuscripts queries to:

\lnj.m in Editorof the BL LLI ETIN
Florida Museum of Natural History
University of Florida
PO Box 117800
Gainesville. FL 32611 -'1 M, U.S.A.
Phone: 352-392-1721. x457
Fax: 352-846-0287
e-mail: mjoyner @

ISSN: 0071-6154


Publication date: 07-31-03

Price: $20.00


Enduring Foundations to a Holistic Science: Lessons in Environmental
Archaeology from Elizabeth S. Wing
K ITTY F. EMERY ................................................ ................. 3
Zooarchaeology at FLMNH in the Context of the Growth and
Development of This Discipline
ELIZABETH S. W ING ................................................... ......................... 11
Deep Sand: Soil and Landscape Relationships at the Blueberry Site
(8HG678), Highlands County, Florida
SYLVIA SCUDDER ................................................................... 17
Screen Size and the Need for Reinterpretation: A Case Study from the
Northwest Coast
KATHLYN M. STEWART and REBECCA J. WIGEN ............................................. 27
Middle Preclassic Landscapes and Aquatic Resource Use at Cuello, Belize
ARLENE FRADKIN and H. SORAYYA CARR .............................................. 35
A Preliminary Survey of the Birds from Velikent (Bronze Age,
Daghestan, Russia)
ARTURO MORALES-MUNIZ and YEKATERINA ANTIPINA ...................................... 43
Middle Bronze Age and Hellenistic Mollusk Meals from Sourpi Bay
(Thessaly, Greece)
W IETSKE PRUMMEL .......................................................... .................... 55
Resource Use through Time at Paloma, Peru
ELIZABETH J. REITZ ................................... ..................................... 65
Use of Fire in Shell Bead Manufacture at Cahokia
LAURA KOZUCH .................................. ......................................... 81
An Illustrated Guide to Trunk Vertebrae of Cottonmouth (Agkistrodon
piscivorus) and Diamondback Rattlesnake (Crotalus adamanteus) in
KAREN J. W ALKER .................................. ...................................... 91
A Millennium of Migrations: Proto-historic Mobile Pastoralism in Hungary
LASZLO BARTOSIEWICZ ..................................... ...... .. ..................... 101
Zooarchaeology of Cinnamon Bay, St. John, U.S. Virgin Islands:
Pre-Columbian Overexploitation of Animal Resources
IRVY R Q UITM YER .......................................................... .................... 131
Animal Remains from Late Medieval Capabiaccio: A Preliminary
Assessment of the Stock Economy
D IANA C C RADER .......................................................... ..................... 159
Horticultural Hunters: Seasonally Abundant Animal Resources and
Gender Roles in Late Prehistoric Iroquoian Subsistence Strategies
ELIZABETH M SCOTT ........................................... ..................... 171
Imagining Sixteenth- and Seventeenth-Century Native American
and Hispanic Transformations of the Georgia Bight Landscapes
DONNA L. RUHL .......................................... 183
Soldiers' Diet at Valley Forge: An Analysis of the Faunal Remains
from the 2000 Excavation Season
DOUGLAS V. CAMPANA and PAM J. CRABTREE ........................................... 199
A Bibliography of Literature by Elizabeth S. Wing
K ITTY F. E MERY ............................................ ...................................... 205

F. WAYNE KING and CHARLOTTE M. PORTER. eds. 2003. Zooarchaeology: Papers to Honor Elizabeth S. Wing. Bulletin of the
Florida Museum of Natural History 44:1-208. [End of volume.]

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 3-10


Kitty F. Emery'

The chapters contributed to this volume honoring
Elizabeth S. Wing's research at the Florida Museum of
Natural History (FLMNH) provide an excellent overview
of some of the most intriguing issues and controversies
in zooarchaeology and environmental archaeology today.
Analyses from nine countries spanning the New and
Old worlds illustrate the global nature of Wing's influence
on the science of environmental archaeology, while the
predominance of research from the eastern U.S.
highlights the particular importance of her work in this
geographic region. The authors tackle topics from the
effects of sieve gauge on bone recovery to the impact
of human predation on ancient environments, reflecting
the breadth of Wing's influence in these varied arenas.
Many of the contributing authors were taught in the
FLMNH Environmental Archaeology (EA) laboratory,
but whether student or colleague, all their research has
been touched in some way or another by Wing's work.
Most important to the value of this volume is the
fact that all these chapters emphasize one of the
fundamental principles of Wing's research: that our
interpretations, whether of ancient ecology, diet, or ritual,
are only as strong as the methods we use to gather our
data. This principle, a central tenet of her work, has
guided her research in all areas of zooarchaeology. Wing
is recognized worldwide for her incorporation of
innovative techniques borrowed from the biological
sciences, her interest in addressing broad theoretical
issues, and her encouragement of a holistic view of
ancient human/environment relationships in
environmental archaeology. The value of her pioneering
work rests both in her enthusiasm for all aspects of the
developing science and in the emphasis she has always
placed on rigorous methods as a foundation to any
environmental research. An understanding of the
importance of strong methodological foundations has also
been passed on to her students and to all of us who

'Assistant Curator of Environmental Archaeology, Florida
Museum of Natural History, University of Florida, Gainesville,
FL 32611, USA.

continue to learn from her example. It is one of the
strongest commendations to our science that, although
we have often moved cautiously in the study of the
cultural aspects of the human/environment relationship,
our conclusions are based on generations of intensive
evaluation of the efficacy of our analytical methods.
In the following pages I will review some of the
issues pertinent to environmental archaeology and
zooarchaeology as illustrated in these chapters. The
choice of these particular issues and themes is not
intended to provide a comprehensive review of
environmental archaeology (see Dincauze 2000, Evans
and O'Connor 1999, and Reitz and Wing 1999 for
excellent surveys) or to suggest that these are the only
areas in which Wing has been influential. I am guided
instead by the current research interests of the authors,
themselves influenced by Wing's ongoing work. My
discussion will highlight the important unifying theme that
binds the chapters of this volume together: Elizabeth
Wing's principle of methodological strength at all levels,
from initial project assessment to the middle range theory
required for accurate interpretations of cultural processes.

Analytical and interpretive methods differ among
branches of environmental archaeology, but requirements
for methodological accuracy do not. Samples must reflect
both the archaeological questions at hand and the area
under investigation. They must be sufficiently large to
overcome sample-size dependencies in statistical analysis.
Recovery methods used for their acquisition must provide
for sufficient detail in the assemblage, and subsequent
handling and preparation of the samples must follow
protocols that diminish loss and destruction of data. Once
samples have been recovered, specimens must be
accurately identified and the broadest possible range of
data must be acquired from each. Methods of
quantification must be appropriate to both the sample
and the questions under investigation. Finally, whether
the specimens are pollen grains or bivalves, interpretations
must be similarly well supported by a clear understanding

of biological and cultural contexts. These must be well
informed with regard to biological taxonomy and habitats,
the taphonomic history of the assemblage and its possible
effect on distributions, and the cultural context within
which the assemblage was recovered. The authors
contributing to this volume recognize, both implicitly and
explicitly, that these requirements are fundamental to all
investigations in environmental archaeology.

Environmental archaeology and, in particular,
zooarchaeology are still seen by some as simply an
answer to the archaeologist's "identification problem"
(Wing this volume). This is unfortunate because Wing
and others have been using direct faunal, floral, and
geomorphological data for complex environmental and
cultural research since the early 1960s, when they already
were aware of the need for methodological accuracy,
especially appropriate archaeological sampling and
collection methods to create assemblages that included
the broadest possible range of taxa or specimens.
There are many vital issues in the discussion of
accuracy in assemblage sampling. Despite years of
controversy, one of the most contentious debates today
is still the utility of fine-gauge sieving (Cannon 1999;
James 1997; Shaffer and Sanchez 1994; Vale and Gargett
2002). First discussed in the literature of the early 1970s
(Clason and Prummel 1977; Payne 1972), screening of
archaeological deposits was already seen by
environmental archaeologists as essential, not only to
increase the number of biological taxa and individuals
recovered, but also to effectively answer questions about
species composition and population dynamics (Grayson
1984; Lyman 1982). By the 1980s, it became clear that
the issue was not whether deposits should be screened,
but at what gauge. In the 1979 volume Paleonutrition,
Wing and Brown stated clearly that 5 mm (1/4") screen
was not fine enough for samples containing fish or other
species with small elements, and research in the FLMNH
laboratories helped determine that a 1.5 mm (1/16")
screen was more appropriate to environmental research
goals. Wing's insistence on fine screening has been
pivotal in many areas of the world, and all the authors in
this volume discuss its importance. "[T]he work of
Elizabeth Wing and her students, through their insistence
on fine-gauge recovery techniques, revolutionized the
picture of subsistence among prehistoric southeastern
Indian groups" (Scott this volume).
Here, Quitmyer and Stewart and Wigen tackle the
screen-size issue directly. Quitmyer's studies over the

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

last decade have shown conclusively that in fine-
screened assemblages from Caribbean and Florida
aquatic habitats the representative size-classes of fish
recovered are expanded and that the proportion of fish
in the sample increased (Reitz and Wing 1999; Wing
and Quitmyer 1985). Stewart and Wigen's work in
Canada reveals similar results. Assemblages recovered
using fine-gauge screens are significantly different from
those recovered using coarser meshes. Among sites in
this area, interpretations of diet can vary from mammal/
salmon dominated to herring dominated, depending on
the screen size used. Their work is substantiated by
similar results from other researchers (e.g., Gordon 1993)
who have found that smaller taxa and often smaller
diagnostic elements are missed when deposits are not
fine screened. Surprisingly, despite evidence from direct
analyses of the impact of differential recovery method,
archaeologists in many regions still remain unconvinced
of the value of such detailed sampling strategies.
Worldwide, the realities of budget and time limitations
discourage the use of many of the methods so necessary
for the recovery of environmental data (Vale and Gargett
While other issues related to sampling are equally
important (including sample size, variable sampling across
cultural contexts, and the effects of taphonomic variation
in archaeological deposits), the ongoing debate over fine-
mesh screening typifies the chasm that continues to exist
between archaeological method and the needs of
environmental archaeologists. It also suggests a two-
fold solution that first begins with increased participation
by environmental archaeologists in defining
archaeological hypotheses and objectives and in
designing sampling strategies to fit those objectives. The
second solution to the problem is a simple awareness of
the potential biases inherent in our samples. In this volume,
Reitz's paper on the Preceramic remains from Paloma,
Peru, and Bartosiewicz's regional analysis of the
Carpathian Basin of Hungary are both excellent
examples of this awareness and treatment of bias. Reitz
provides specific and detailed descriptions of sampling,
quantification, and secondary analysis methods and
correlates findings with a discussion of potential bias
created by each of the methods used. Bartosiewicz has
the more challenging task of coordinating data from
multiple excavations. Although it is impossible to
standardize the methods used in sample collection or
treatment in multiple datasets collected over many years,
his statistical correlations bridge the inherent biases in
the analysis. Such work also emphasizes the importance

EMERY: Enduring foundations to a holistic science

of encouraging the creation and dissemination of
"standards" for gathering environmental data from
archaeological deposits.

Issues of methodological accuracy are not confined to
the field. Many arise in our own labs during identification
and analysis. It is self-evident that one of the
fundamentals of our science is the accurate identification
of our specimens, whether they be sand grains, corn
kernels, canid phalanges, or oyster valves. Yet another
controversial issue in modem environmental archaeology
is the use of appropriate comparative collections for
specimen identification. Accurate identifications are based
on comparison with a broad modem taxonomic collection
and recognition of the essential diagnostics for each taxon
or skeletal element. However, a lack of appropriate
resources (and sometimes a lack of recognition of the
nature of appropriate resources) means that much of
the environmental archaeological research being
conducted worldwide is based on limited and often
unvouched comparative collections and untrained eyes.
Despite trends at the time emphasizing research at
a cellular level, Elizabeth Wing began her career as a
zoologist with an interest in whole animal biology. Her
earliest publications stemmed from Master's thesis
research on reproductive behavior of the Florida pocket
gopher, while her Ph.D. dissertation on fauna from the
Trinidad tar pits was an important bridge from zoology
to zooarchaeology. As a zoologist, she has continued to
conduct research on modern comparative collections as
well as environmental remains from archaeological
One important and early innovation at the FLMNH
EA laboratory was Wing's emphasis on the essential
importance of the study of whole communities.
Consequently, while many zooarchaeologists ignored
segments of the faunal taxa. she included all vertebrates
and most invertebrates in her analyses. FLMNH's
regional scope, with its emphasis on southeastern U.S.
and Caribbean coastal sites, fostered Wing's early interest
in marine fishes and the utility of fisheries and
ichthyological data for archaeological studies. As a result,
she actively encouraged the incorporation of marine
fishes into comparative collections. The EA collections
currently contain more ichthyological skeletal remains than
most other zooarchaeological labs and certainly more than
many ichthyological research centers (Poss and Collette
The EA laboratory's modern faunal comparative

collection with its 9000 specimens comes close to Wing's
initial objective of providing a representative collection
of all taxa from the southeastern U.S. and the Caribbean
(Wing this volume). The collection also contains
representative species from Central America (Mexico
and Guatemala) and northern South America (Peru,
Ecuador, and Panama), reflecting her continuing interest
in Peruvian and Mesoamerican research. The EA
collections also contain representative modern botanical
and soil reference materials and, while these collections
are small, they are growing rapidly in tandem with Wing's
focus on holistic research on environmental materials.
Over 3500 cataloged soil samples represent 34 sites, 25
in Florida and 9 in the Caribbean. The botanical collections
contain over 500 reference specimens (predominantly
from the southeastern U.S.), as well as a growing collection
of carbonized seeds for comparison with archaeologically
burned examples. Both zooarchaeological and
archaeobotanical collections are being digitized to create
virtual-image files for use by a wider audience.

Wing's work on tropical fishes is of primary
importance to coastal research, but her emphasis on
comparative work with invertebrates has been equally
influential in the expanding science of zooarchaeology.
Zooarchaeological analyses in the EA laboratory have
included mollusks and common invertebrates like crabs
and have extended to studies of shrimp mandibles and
other robust microinvertebrate remains. Work on marine
bivalves is pivotal to both cultural analyses and
environmental reconstructions in the EA laboratory
(Quitmyer this volume; Quitmyer et al. 1997). The
importance of this emphasis on mollusks is reflected in
Prummel's metric analysis (this volume) of mollusks from
Sourpi Bay, Greece, and the information this provides
on local habitat changes over time. Prummel distinguishes
molluskan morphological changes that indicate habitat
change from those attributable to human predation.
Biological information on both the modern and the
zooarchaeological specimens in this collection and an
emphasis on the habits and habitats of the species have
always been fundamental to Wing's interpretations of
animal use. Early in her career, a detailed understanding
of marine fish habitats allowed her to move beyond simple
species lists to discussions of fishing practices,
catchments, and environmental change based on the
feeding habits of marine fishes. Fradkin and Carr's
chapter on Preclassic aquatic resource use in Belize
follows the path laid down by Wing's early work on fish

behavior and habitat requirements and her later analysis
of other collections from the same site (Wing and Scudder
1991). Fradkin and Carr combine turtle habitat details
with ethnographic information on both acquisition and
taste preferences to correlate their findings with other
environmental indicators, suggesting use of the local
resources as opposed to those from a broader catchment.
The findings are important in view of current discussions
in Mesoamerican archaeology about Preclassic trade in
marine products (Powis et al. 1999; Shaw 1991, 1999;
Stanchly 1995).

In keeping with a true biological collection, all biotic
specimens collected for the EA laboratory include
information on sex, age, size, body weights, habitat, and
season of capture. The availability of such detailed
biological information has allowed those who use this
material to fine-tune the identification process. This work
has defined and in some cases increased the number of
diagnostic elements and element characteristics useful
for identification (e.g., Kozuch's use of gastropod
columellas, this volume) and delineated those similarities
that restrict accurate identifications beyond the level of
genus or, in some cases, family (Reitz and Wing
1999:154). Walker's chapter (this volume) is an excellent
example of this type of collection-based research. Based
on detailed use of modem comparative specimens from
the FLMNH, she expands the roster of diagnostic
elements useful in separating Florida cottonmouth from
diamondback rattlesnakes to include mid-precaudal
vertebrae. Wing's work on the domestic dog has also
emphasized the caution that must be taken in
identifications at the species level, but has shown that
with a substantial database it is possible to find markers
even for the identification of domestic breeds and types.
Detailed direct research on the EA comparative
collections has encouraged refinements in quantification
of ancient animal use. For decades, environmental
archaeologists have debated the means by which we
quantify the "proxy measures" found in the archaeo-
logical and sedimentological record. Pollen grains, animal
bones, and chemical soil signatures do not accurately
represent the patterns of ancient environmental variation.
No one debates this essential truth. The challenge lies in
the way these remains can be quantified to best reflect
ancient environments and use of its resources.
A central issue in zooarchaeology is the relationship
between bone fragments (NISP), individual animals
(MNI), and actual contribution to diet or other activities.

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

NISP and MNI are now well-accepted measures used
by most zooarchaeologists (and most authors in this
volume) to create a bracket of maximum and minimum
counts for each species. Wing was one of the earliest to
see both the validity of careful quantification (Wing 1963)
and the dangers of unfettered use of MNI measures,
particularly in the analysis of complex societies where
resources are shared among community members (Wing
and Brown 1979). Her early application of skeletal and
whole-body allometry formulae to bone weights, skeletal
counts, and MNIs (based on studies of modem animals)
provided impetus for more complex analyses of the
relative ancient contribution of different species in the
archaeological record (Reitz and Wing 1999:69-72; Wing
and Brown 1979). Her methods are used by Quitmyer
and Reitz (both this volume) as a basis for interpretation
of the relative contribution of species to ancient diet.
These more complex estimates allow us to move beyond
such strictly data-centered questions as "What are they
and how many did we find?" to the more intriguing
questions of ancient environments and the relationships
between these and the human communities we study as

One of our most enduring efforts in environmental
archaeology has been the reconstruction of human
impacts on their environments, from the creation of
anthrosols (soils resulting from human activities) to the
creation of new species through domestication to the
eradication of others through habitat modification and
direct predation. Two goals of environmental
archaeology are particularly relevant to modern
environmental awareness (or the lack thereof): first,
determining whether or not humanity has ever sustainably
managed an environment and, second, determining if
there are lessons that can be learned from both the
mistakes and successes of past human-environment
interactions. Accurate reconstructions of the effects of
human interaction with the environment are also
premised on accurate reconstructions of natural
environments and on the clear differentiation between
anthropogenic and non-anthropogenic processes.
Scudder's complex reconstruction of the landscape
history of the Blueberry site in Florida emphasizes the
importance of direct evidence for evaluation of processes
and perpetrators of environmental change. Using soil
micromorphology and elemental analyses, Scudder
describes the development and use of a "subsurface

EMERY: Enduring foundations to a holistic science

landscape" and ties the local landscape genesis to global
patterns of climate change. As the volume's only soil
scientist, Scudder does a superb job of highlighting the
importance of strong methodological foundations in this
branch of environmental archaeology for separating
human from non-human environmental modifications, for
defining the details of geomorphological change, and for
linking these to broader issues.
Morales-Mufiiz and Antipina clarify the methods used
for identification of the effects of human versus non-
human activity on archaeological assemblages. In their
preliminary survey of birds from Bronze Age Russia,
they provide markers to differentiate among human,
predatory birds, and other agents involved in the
deposition of bones in separate assemblages. Combined
criteria devised by detailed study of agency in
archaeological deposition allow these authors to quantify
with certainty the human activity in the accumulation of
the different assemblages.

It is not a simple process to separate the effects of
natural variation from that caused by humans. An
important issue within the broad theme of environmental
reconstruction is the controversy surrounding models of
anthropogenic faunal extinctions, particularly with first
human incursions into pristine ecosystems. The North
American Pleistocene overkill hypothesis has been an
enduring model (Alroy 2001; Martin 1967; Martin and
Steadman 1999) despite arguments that these continental
extinctions should be attributed to the environmental
conditions associated with the Pleistocene/Holocene
boundary. But over the past decades a robust
archaeological and paleontological database provides
clear evidence of a direct association between human
colonization and extinction events on many islands
(Steadman 1995; Steadman and Stokes 2002; Wing
2001). The characteristics of island biogeography make
these locales particularly susceptible to extinction, and
the effect of human colonization is most dramatic on
smaller islands.
The discussion of human impact on ancient
environments extends beyond the land. Wing's years of
research on marine fish communities have provided
extensive information on the impact over time of human
predation on marine fish populations. Her recent studies
of the ancient process of "fishing down the food web"
adapt modem trophic-level methods described by Pauly
et al. (1998) for modern aquatic populations. The method
requires appropriate samples derived from fine-gauge

sieving (1/8"or smaller), an estimate of biomass using
average body weight calculated from appropriate
allometric formulae, and the application of a mean trophic
level index for each species, as provided by modern
research (Pauly et al. 1998; Wing 2001). Using trophic
level analyses, Wing has shown the effects of ancient
human predation on reef fishes from various Caribbean
islands and has emphasized the remarkable similarity
between ancient and modern processes despite
differences in hunting or fishing techniques.
Other researchers in this volume test the specific
predictors from Wing's studies of human impact and
discuss both overall correlations and important additional
considerations. Quitmyer's analysis of faunal
assemblages from Cinnamon Bay in the U.S. Virgin
Islands expands on Wing's trophic level research by
suggesting the importance of ceremonial activities as an
additional factor in human predation. Reitz correlates
trophic level shifts on the Archaic Pacific coast of Peru
with climate fluctuations associated with the
Hypsithermal. As in any other reconstruction of human
impact on environments, the possible effects of non-
anthropogenic changes and the variations introduced by
cultural complexity must be considered.
Needless to say, the process of human impact on
plants and animals continues far beyond initial contact.
Ruhl provides a glimpse of the archaeobotanical
perspective on ancient environments, describing
environmental changes associated with secondary
human impact on natural and anthropogenic landscapes,
in this case Hispanic intrusion on the inhabited Georgia
Bight region. Combining archaeobotanical,
zooarchaeological, and documentary evidence, Ruhl
reconstructs the prehistoric environment and documents
the marked differences resulting from Old World-
introduced domesticates and land-use practices. She also
emphasizes that our increase in knowledge has come
only as a result of the use of specialized recovery
methods and an awareness of the special conditions of
wetland preservation.

While environmental archaeologists have often been
accused of methodological navel-gazing and of seeing
little beyond the biological characteristics of ancient
environments and the resources they provide for the
sustenance of life, the papers in this volume demonstrate
quite the opposite. The attention paid to methodological
rigor by environmental archaeologists has led to
substantial concentration on the accuracy of our

reconstructions of the environment as a baseline for
understanding human activity. As the papers in this
volume show, the authors have not neglected the cultural
aspects of the human/environment relationship. Even
early studies of diet and domestication of plants and
animals incorporated aspects of the symbolic nature of
landscapes and the ritual associated with domesticated
species in most societies.
It is true that environmental archaeologists have
often avoided discussions of ritual, cosmogeny, and
symbols as interpretations unreachable by current
methods (for a detailed discussion, see Albarella 2001).
Within the cultural sphere, these papers also show
attention directed at processual issues well supported
by both data and analogies as bridging arguments.

Environmental archaeological research on human
impacts extends beyond early interactions with pristine
environments to consider direct and often intentional
manipulation of species communities through the creation
of new habitats, transportation of species to new
locations, and the process of domestication. The timing
and process of early domestication were among the first
issues explored by environmental archaeology. Wing and
other zooarchaeologists have had an abiding interest in
the cooperation, necessitated by domestication, between
humans and animals, and Wing's work in the Peruvian
Andes has been important to continuing research on
New World domesticates.
Domestication involves gradual processes, and some
of our recent questions address precursor behaviors (both
plant/animal and human). Rindos's (1984) early proposal
of co-evolutionary adaptation as a mechanism for
domestication is an excellent model within which to view
the process of synanthropism. Any human manipulation
of the environment, even the establishment of temporary
encampments, provides a new niche for animal activity.
Clearly human refuse was one precursor to dog
domestication, and current research like that presented
by Morales-Mufiiz and Antipina (this volume) identifies
other species in the archaeological record likely to be
attracted to human-modified environments. Their survey
of occupation and subsistence economies in Bronze Age
Russia (Boev 1993) relies in large part on the definition
of certain bird species as seasonal synanthropes, passive
synanthropes, and synurbanists to provide a guide to the
levels of cooperative living achieved by humans and their
avian neighbors.
One intriguing consideration that is emphasized by

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

the Rindos (1984) model, but often neglected in
examinations of proto-domestication, is the human
response to synanthropizing species. Scott (this volume)
reintroduces the Linares (1976) model of tropical garden-
hunting in a North American perspective. Scott uses fine-
screened assemblages from late prehistoric New York
State to provide a detailed analysis of species use. In
the small agricultural hamlet of Spaulding Lake, her data
suggest seasonal occupation and the exploitation of fauna
attracted to gardens and clearings. Her results indicate
that traditional perspectives on Iroquois diet and
settlement may have been skewed by early
archaeological studies of hilltop villages.

Our understanding of the role of domestic plants
and animals in both ancient and historic societies is further
complicated by the nature of human society. Again, the
difficulty lies in the methodological realm. To understand
the process of domestication, the effects of
synanthropism, and the eventual dynamic of social control
over domesticated natural resources, we need accurate
counts of the remains. This simple factor is compounded
many times over by the effects of social interactions,
market economics, and complex ritual beliefs (e.g.,
Crabtree 1990; Zeder 1991). Several excellent examples
of research that directly confront these complications
are presented in this volume.
Bartosiewicz's regional review of the shift from
nomadism to sedentism in the Carpathian Basin of
Hungary reveals the effects of territorial circumscription
on generations of mobile pastoralists. Shifts in faunal
species over time mark the transition to sedentary
pastoralism, but his work also suggests that traditional
foodways are often retained even under unfavorable
conditions. While optimization models might predict the
transition to sedentism under pressure from declining
territorial sizes, more structural models of social responses
to perceived circumscription could explain retention of
ineffective subsistence systems.
Kozuch's discussion of shell bead manufacturing
methods at Cahokia reminds us of the importance of
long-distance trade of resources and the movement of
secondary products through communities. Recognition
of markers of artifact production allows us to consider
implications for both quantification of species use and
the role of environmental products in defining social
structure (in this case the evidence for economic
In this volume, two papers effectively illustrate the

EMERY: Enduring foundations to a holistic science

advantages and difficulties of working with historic
samples. Crader's evaluation of the "stock economy"
of Medieval Italy takes into account the diametrically
opposed effects of economic autonomy versus market
economies on the zooarchaeological record. Mortality
profiles provide useful information on domesticates, but
these data are also well contextualized with reference
to the varied uses and distribution of these animals.
Campana and Crabtree face similar difficulties in their
evaluation of soldiers' diets in Pennsylvania during the
Revolutionary War. Combining documentary and
zooarchaeological evidence lets them determine not only
the quantity, but also the quality, of foodstuffs available
to soldiers at the Valley Forge encampment.

I would like to suggest that attention to the details and
accuracy of scientific conclusions has become even more
important in view of our current role in providing ancient
examples for solving environmental problems (Amorosi
et al. 1996; Cannon and Lyman 2002: Lyman 1996). At
FLMNH, the EA laboratory provides a broad array of
services to academic, professional, and public audiences,
chief among these services data for governmental (local,
national, and international) and non-governmental
organizations' environmental management. Can
zooarchaeological information define natural population
dynamics and habitat extents for endangered populations?
Can archaeopedology provide substantive information
on the effects of wetland clearing? Can archaeobotanical
models be used to predict global warming? As many
academics ponder the potential utility of our science for
modem decision-making, the EA program under Elizabeth
Wing's direction has met the immediate need to contribute.
While the chapters in this volume underscore the
importance of accurate data in a purely intellectual
forum, if the future of our environment rests even in
part on our interpretations of the archaeological record,
the value of these strong methodological foundations
becomes even more relevant.
Two primary concerns in environmental archaeology
are the accuracy of our methods for providing secure
baseline data and our ability to conceptualize and describe
the full extent of the ancient relationship between people
and their environments. This volume addresses these
concerns through discussions of research from around
the world, touching on issues from allometry to women's
roles. The geographic and cultural breadth of these
chapters, in combination with their theoretical and

methodological unity, is a tribute to the strength of the
foundations upon which the science is built. The research
of these authors reflects Wing's emphasis on the
fundamental importance of methodological accuracy for
interpretation in environmental archaeology.
Wing's impact on environmental archaeology goes
far beyond the methodological foundations. Her more
important influence lies in the development of
environmental archaeology as a multidisciplinary and
holistic science. Following her lead, environmental
archaeology has come to be defined in large part by the
enthusiasm of its participants and their adventurous
exploration of the multiple ways of integrating the
biological and cultural sciences. This spirit is amply
reflected in the research presented in this volume.

Albarella, U. (ed.). 2001. Environmental Archaeology: Meaning
and Purpose. Boston: Kluwer Academic Publishers.
Alroy, J. 2001. A multispecies overkill simulation of the end-
Pleistocene megafaunal mass extinction. Science 292:
Amorosi, T., J. Woollett, S. Perdikaris, and T. H. McGovem. 1996.
Regional zooarchaeology and global change: Problems and
potentials. World Archaeology 28(1): 126-157.
Boev, Z. 1993. Archeo-orithology and the anthropisation of
birds: A case study from Bulgaria. Archaeofauna 2: 145-153.
Cannon, K., and R. L. Lyman. 2002. Zooarchaeology's
contribution to conservation biology. Session presented
at the Society for American Archaeology meetings,
Denver, Colorado, 20-24 March.
Cannon, M. D. 1999. Mathematical model of the effects of
screen size on zooarchaeological relative abundance
measures. Journal of Archaeological Science 26: 205-214.
Clason, A. T., and W. Prummel. 1977. Collecting, sieving, and
the archaeological record. Journal of Archaeological
Science: 171-175.
Crabtree, P. J. 1990. Zooarchaeology and complex societies:
Some uses of faunal analysis for the study of trade, social
status, and ethnicity. Archaeological Method and Theory
Dincauze, D. F 2000. Environmental Archaeology: Principles
and Practice. Cambridge: Cambridge University Press.
Evans, J. G., and T. P. O'Connor. 1999. Environmental
Archaeology: Principles and Methods. Stroud,
Gloucestershire, U.K.: Sutton Publishing.
Gordon, E. A. 1993. Screen size and differential faunal recovery:
A Hawaiian example. Journal of Field Archaeology 20:
Grayson, D. K. 1984. Quantitative Zooarchaeology. New York:
Academic Press.
James, S. R. 1997. Methodological issues concerning screen
size recovery rates and their effects on archaeofaunal

interpretations. Journal of Archaeological Science 24: 385-
Linares, O. E 1976. "Garden hunting" in the American tropics.
Human Ecology 4(4): 331-349.
Lyman, R. L. 1982. Archaeofaunas and subsistence studies.
Advances in Archaeological Method and Theory 5: 331-
Lyman, R. L. 1996. Applied zooarchaeology: The relevance of
faunal analysis to wildlife management. World
Archaeology 28(1): 110-125.
Martin, P. S. 1967. Prehistoric overkill.Pp. 75-120 in P. S. Martin
and J. H.E. Wright, eds. Pleistocene Extinctions: The
Search for Cause. New Haven, Connecticut: Yale
University Press.
Martin, P. S., and D. W. Steadman. 1999. Prehistoric extinctions
on islands and continents. Pp. 17-52 in R. D. E. MacPhee,
ed. Extinctions in Near Time. New York: Kluwer Publishers.
Pauly, C., V. Christensen, J. Dalsgaard, R. Froese, and E Torres,
Jr. 1998. Fishing down marine food webs. Science 279:
Payne, S. B. 1972. Partial recovery and sample bias: The results
of some sieving experiments. Pp. 49-64 in E. S. Higgs, ed.
Papers in Economic Prehistory. Cambridge: Cambridge
University Press.
Poss, S.G., and B.B. Collette. 1995. Second survey of fish
collections in the United States and Canada. Copeia
Powis, T. G, N. Stanchly, C. D. White, P. F. Healy, J. J. Awe, and
F. J. Longstaffe. 1999. A preliminary reconstruction of
middle preclassic subsistence economy at Cahal Pech,
Belize. Antiquity 73(280): 364-376.
Quitmyer, I. R., D. S. Jones, and W. S. Arnold. 1997. The
sclerochronology of hard clams, Mercenaria spp., from
the south eastern USA: A method of elucidating the
zooarchaeological records of seasonal resource
procurement and seasonality in prehistoric shell middens.
Journal of Archaeological Science 24: 825-840.
Reitz, E. J., and E. S. Wing. 1999. Zooarchaeology. Cambridge
Manuals in Archaeology. New York: Cambridge
University Press.
Rindos, D. 1984. The Origins of Agriculture: An Evolutionary
Perspective. New York: Academic Press.
Shaffer, B. S., and J. L. J. Sanchez. 1994. Comparison of 1/8-
inch and 1/4-inch mesh recovery of controlled samples of
small-to-medium sized mammals. American Antiquity 59:

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Shaw, L. C. 1991. The Articulation of Social Inequality and
Faunal Resource Use in the Preclassic Community of
Colha, Northern Belize. Ph.D. Dissertation, University of
Massachusetts, Amherst.
Shaw, L. C. 1999. Social and Ecological Aspects of Preclassic
Maya Meat Consumption at Colha, Belize. Pp. 83-102 in
C. White, ed. Reconstructing Ancient Maya Diet. Salt
Lake City: University of Utah Press.
Stanchly, N. 1995. Formative period Maya faunal utilization at
Cahal Pech, Belize: Preliminary analysis of the animal
remains from the 1994 season. Pp. 124-149 in P. F. Healy
and J. J. Awe, eds. Belize Valley Preclassic Maya Project:
Report on the 1994 Season. Peterborough, Ontario,
Canada: Trent University, Occasional Papers in
Anthropology no. 10.
Steadman, D. W. 1995. Prehistoric extinction of South Pacific
birds: Biodiversity meets zooarchaeology. Science 267:
Steadman, D. W., and A. V. Stokes. 2002. Changing exploitation
of terrestrial vertebrates during the past 3000 years on
Tobago, West Indies. Human Ecology 30(3): 339-367.
Vale, D., and R. H. Gargett. 2002. Size matters: 3-mm sieves do
not increase richness in a fishbone assemblage from
Arrawarra 1, an Aboriginal Australian shell midden on
the mid-north coast of New South Wales, Australia.
Journal of Archaeological Science 29: 57-63.
Wing, E. S. 1963. Vertebrates from the Jungerman and Goodman
sites near the East Coast of Florida. Contributions of the
Florida State Museum 10: 51-60.
Wing, E. S. 2001. The sustainability of resources used by
Native Americans on four Caribbean islands. International
Journal of Osteoarchaeology 11: 112-126.
Wing, E. S., and A. Brown. 1979. Paleonutrition: Method and
Theory in Prehistoric Foodways. New York: Academic Press.
Wing, E. S., and I. R. Quitmyer. 1985. Screen size for optimal
data recovery: A case study. Pp. 49-58 in W. H. Adams,
ed. Aboriginal Subsistence and Settlement Archaeology
of the Kings Bay Locality. Reports of Investigations no.
2. Gainesville: University of Florida, Department of
Wing, E. S., and S. J. Scudder. 1991. The exploitation of animals.
Pp. 84-97 in N. Hammond, ed. Cuello: An Early Maya
Community. Cambridge: Cambridge University Press.
Zeder, M. A. 1991. Feeding Cities: Specialized Animal Economy
in the Ancient Near East. Washington, D.C.: Smithsonian
Institution Press.

Bull. Fla. Mus. Nat. Hist. (2003)44(1): 11-16


Elizabeth S. Wing'

While zooarchaeology is not yet a household word, our
program goals demonstrate its relevance to a better
understanding of the past and present. Zooarchaeologists
examine the animal remains, often discarded refuse,
excavated from archaeological sites. These remains reveal
which animals were caught and used by humans in the
past. Habitat preferences of the animals encountered in
the refuse provide insight into the past economy at the
site. The effects of human hunting and fishing does not
stop with the meal that the catch provided, it also impacts
the subsistence enterprise on the animal populations.
Similarly, land clearing and plant cultivation can enhance
habitats for those animals that might prefer ecotonal zones
and consume crops. Land clearing might also promote
erosion, producing turbid water and less favorable
conditions for those animals that require clear waters.
While the complex of animals in archaeological contexts
reveals much about past lifeways, the effects of past
human exploitation leave their marks on the modern
landscape and contemporary animal populations.
Even though some sites, such as the big shell mounds
in the southeast U.S. and other parts of the world, are
composed overwhelmingly of animal remains,
zooarchaeology got a slow start as a sub-field of
archaeology. The molluscan remains that compose the
mountains of shells that make up some sites often did
not attract much interest for anything other than a source
of paving material. The earliest studies of animal remains
that focused on documenting the source and the history
of animal domestication were centered in the Near East
(southwest Asia). Europeans with training in veterinary
medicine or zoology undertook much of this research.
The animals of concern were sheep, goats, cattle, pigs,
and horses and how their domestication not only changed
those species from which they are derived but also
transformed the cultures that domesticated and herded
them. Some outstanding studies of the animal remains
left by people who were hunters and gatherers in North
America were conducted during the late 1800s and early
1900s. By the middle of the twentieth century,
zooarchaeology took on a more quantitative and
interpretive character.

'Professor Emerita, Florida Museum of Natural History,
University of Florida, Gainesville, FL 32611, USA.

The proceedings of a conference entitled The
Identification of Non-Artifactual Archaeological
Materials, edited by Walter W. Taylor and published by
the National Academy of Science in 1957, discussed the
"identification problem." The situation that instigated the
conference was that archaeologists would bring bags of
bones to zoologists to identify in hopes of learning about
the past uses of animals. Many of the zoologists pressed
into service had projects of their own and did little more
than provide the archaeologist with a list of taxa present.
The recommendations of the members of the conference
were to encourage specialization in zooarchaeology by
people able to accurately identify fragmentary remains
and who were also well informed about the nature of
the archaeological deposits. Archaeologists clearly needed
specialists with an interdisciplinary approach to transform
animal remains into a scientific endeavor. The
establishment of the International Congress of
Archaeozoology (ICAZ) promoted these ideas. ICAZ
as a professional organization flourishes to this day.

Zooarchaeology was initiated at the Florida Museum of
Natural History (FLMNH; formerly known as the Florida
State Museum) in this intellectual climate. The very
efforts to use zooarchaeological materials began in a
fortuitous way. Faunal remains excavated in Trinidad by
Irving Rouse of Yale University and John Goggin of the
University of Florida were given for identification to
Pierce Brodkorb in the Zoology Department at the
University of Florida. A student of Dr. Brodkorb began
work on these collections, then for personal reasons left
both the university and the bags of bones. These materials
happened to fit into my dissertation proposal to understand
changes in the mammal fauna of Trinidad as it broke
from the mainland and became an island. I was given
permission to take over the study of these materials,
then incorporate the results into my dissertation. This
study led to a Ph.D. (1962) and permission to continue
the pursuit of zooarchaeological research made possible
by funding from the National Science Foundation.
From these beginnings, a program in zooarchaeology
grew with the support of the FLMNH and the National
Science Foundation (NSF). The program developed both

according to design and serendipitously, taking advantage
of research opportunities. Basic questions needed to be
explored, including subsistence economies in the past,
environmental change through time, human impact on
island ecosystems, uses of animals by people who
cultivated domestic plants, and New World animal
domestication. Before any of these topics could be
addressed, an adequate comparative collection had to
be assembled. Growth of the comparative collection
continues to this day, but in the beginning the aim was to
get one specimen of each species that might have been
used by humans in the southeast and Caribbean. The
next step was to expand the range of sizes and to include
specimens of both males and females. The early
shoestring curation of these specimens was in baby food
jars because they were a useful size and were very
plentiful. The Wing and colleagues' families had young
children during the germination and early growth of the
zooarchaeology program. We would often buy baby food
as much for the type of jar the food came in as for the
food itself. Food in jars with screw-on lids sustained
Molly (born 1961) and Stephen Wing (born 1965).
Preparing specimens and taking field trips were parts of
their early experiences.

The southeast U.S. and the Caribbean were logical areas
in which to begin zooarchaeological studies. Many
samples from sites had been recovered and were
available for study. They were not necessarily recovered
in a systematic way and, if the material was sieved, it
would have been through coarse-gauge screen (1/4- to
1/2-inch). This early exploratory work formed the
foundation of a database of identified faunal samples
assembled in this opportunistic way. The results formed
the basis for more focused research addressing questions
of seasonal occupations of sites, fishing technology, trade
in animal resources, contribution of mollusks to the
prehistoric diet, and change in animal use through time.
The initial exploratory work was made possible by
two successive NSF grants. The first project from 1961
to 1965 was entitled "Approaches to the analysis of post-
Pleistocene environments in the southeast and Caribbean
based upon study of vertebrate remains from
archaeological sites." Co-investigators were Clayton E.
Ray, John M. Goggin, and William H. Sears, with E.A.
Wing as an associate investigator. A part of this work
included collecting modern comparative fish specimens
in the West Indies (1964) with the help of Carter Gilbert
and John Randall. This paved the way for future
zooarchaeological work in the Caribbean. The second

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

project from 1966 to 1968 was entitled "Basic economic
patterns of prehistoric fishing communities of the Gulf
of Mexico and the Caribbean." In the time between these
two NSF grants, I had a contract for zooarchaeological
work in Jamaica under an NSF grant awarded to R.R.
Howard. During the course of this research, and
particularly after the studies in Jamaica, it became
possible to ask more focused and interesting questions.
It also became clear that methods of recovery,
identification, and analysis had to be improved before
more progress could be made in this field of endeavor.

The first opportunities to work on faunal remains from
inland Mexico came in 1968 with a year-long contract
under an NSF grant awarded to Michael D. Coe for
research at the Olmec site of Tenochtitlan in Vera Cruz.
With the help of Robert R. Miller, it was possible to collect
freshwater fishes, including many confusingly similar
species of cichlids. This was followed in the period of
1972 to 1975 by research in Vera Cruz with Jeffery
Wilkerson, the Marismas Nacionales of Sinaloa and
Nayarit with Stuart Scott, and Panama with Olga
Linares. Earlier studies in Yucatan were done under the
auspices of E. Wyllys Andrews. Together these projects
allowed study of the animals associated with people who
had a fully developed agricultural system but no
domesticated animals other than the dog and turkey.
The Wing children were old enough and self-
sufficient enough to take on some of the field work in
Mexico. The most memorable trip was to the Marismas
Nacionales in 1972. Rochelle Marrinan joined the two
children and me to explore this large estuarine area and
study the faunal remains excavated from nine sites. We
lived in a small shark-fishing community. Steve took in
all the activity with relish, singing, "This is the happiest
day of my life," as he checked the catch of the day and
acted as midwife to the dead mother of 35 unborn tiger
sharks. He was allowed to keep the tiger shark jaws,
which we took back to the museum, along with other
specimens we had prepared. While our luggage with
these treasures was deemed too undesirable to allow us
to register in the hotel in Mexico City, at least we did not
have the stuffed toad playing a mandolin that Rochelle
and Steve thought I should buy for him.

Archaeological sites in the Andean highlands hold
evidence for domestication of most New World animals.
Study of the processes of domestication could be modeled

Wing: Zooarchaeology at the Florida Museum of Natural History

after the research done at other centers of animal
domestication, such as the Near East, keeping in mind
the many differences between these areas. The
opportunity to work on the animal remains excavated by
the University of Tokyo Andean Expedition became
available. Once again NSF made this possible with a
grant from 1968 to 1969 entitled "Utilization of animal
resources of the Kotosh site, Peruvian highlands." One
component of this grant was to bring a member of the
Tokyo team to the FLMNH to translate the labels and
help with identification so that upon return to Japan he
could be their faunal analyst. As it turned out, he was a
gifted artist and illustrated our report.
Even though human occupation of Kotosh was of
long duration, that single site could not yield answers to
all the questions regarding animal domestication. To
further pursue this inquiry, a second project, entitled
"Prehistoric man-animal relationships in the Central
Peruvian Andes," was funded by NSF from 1970 to 1973.
In addition, I had the opportunity to work in Ayacucho
on contract under a NSF grant awarded to Richard S.
MacNeish. This interdisciplinary research was coordinated
with Kent Flannery, who worked on the faunal samples
from the high-elevation Puna altiplano, while I worked
on the samples excavated from the valley. Seven years
later, in 1980, Elizabeth Reitz and I, with the help of two
Peruvian zooarchaeologists, Carman Rosa Cardosa and
Denise Pozzi-Escot, worked on faunal remains excavated
by Craig Morris at the Incan urban center of Huanuco
Pampa. Molly Wing came along to Huanuco and
illustrated the worked bone from the Huanuco Pampa site.
(Molly, a tall, slender, blue-eyed blonde, did not go unnoticed
in the town square of Huanuco, where the opportunities to
practice Spanish and scientific illustration were good.) Even
today, there is still much to do to fully understand the
source of animals that could be domesticated, the process
by which they were, and the spread of domestic animals
and animal husbandry knowledge.

Based on improvements in the methods of recovery, our
expanded comparative collection, and the complementary
studies of plant remains, we initiated a new project funded
by NSF from 1989 to 1992 entitled "Subsistence in
prehistoric West Indian economies." Using fine-gauge
screens, this allowed identification of many newly
recovered samples, with the plant remains being identified
by Lee Newsom. The use of fine-gauge screen was
critical for analyses of overexploitation because one of
the lines of evidence for overexploitation was the decline
in the size of organisms. When only coarse-gauge

screens are used to recover faunal samples, the size
range is biased because the small end of the size range
is missing. Improved recovery strategies allow
investigation of size declines, changes in species
composition, and corresponding changes in plant remains.
This research on both plant and animal remains is being
integrated from the perspectives of human adaptation
and island biogeography. Lee Newsom and I have
prepared a manuscript on this research that will be
published by the University of Alabama Press.

One part of the research program outlined briefly above
was the collection of comparative specimens and
zooarchaeological samples. Often the zooarchaeological
samples were on loan to the FLMNH for identification
and analysis, ultimately to be returned to the archaeologist
working on the site or to the country from which the
material was recovered. In other cases the archaeologist
excavating the site chose to deposit the faunal material
at the museum for curation and long-term care. The
comparative specimens were always collected in full
accordance with local and federal regulations governing
the capture of animals. By the year 2000, the
zooarchaeology collection included 8,700 comparative
specimens and 600 zooarchaeological samples.
The curation of the zooarchaeological and
comparative collections involves preparation, cataloging,
and organization, requiring the full-time work of several
people. A full-time faculty position, that of assistant
curator, was offered to me in 1979, which I accepted
with enthusiasm. A part of the NSF Kotosh project
supported the research and curation work that had been
done by Takeshi Ueno from 1968 to 1971. A number of
other people, including Kent Ainslie, David Dorman, Lynn
Cunningham, and Sandra Courter, contributed in
substantial ways to the curation of this growing collection
and were supported either by museum or grant funds.
Real continuity and growth came when Sylvia
Scudder began work in the zooarchaeology collection in
1980. Sylvia started work at the museum in 1972 when
collections were moved from the Seagle Building to the
new museum building now called Dickinson Hall. She
worked in the herpetology collection and then the
mammal collection before she began work in
zooarchaeology. At present, she is the collection manager
of our multi-facetted collection. In this capacity, she has
organized work in the collection and has hired work-
study students to help maintain it. In addition to her work
as a collection manager, Sylvia has done basic research
in zooarchaeology, identifying and analyzing faunal

remains from Florida, Jamaica, Middle Caicos, Belize,
and Yucatan. She has used the comparative collection
to analyze the vertebrate remains from bald eagle nests,
alligator stomach contents, and other samples brought
to the museum by wildlife biologists. Since completing
her Master's degree in soil science, she has become
one of the few authorities on anthropogenic soils and
has studied the soils from many sites in the southeast
and Caribbean to shed light on the soil development of
archaeological deposits.
The National Science Foundation supported three
initiatives to improve curation of the zooarchaeology
collection. The first, for 1979-1980, was funding to
prepare a photographic atlas of the nine most diagnostic
skull elements of each fish species in the comparative
collection. The rationale was that identification could start
with a review of the photographic atlas before specimens
were examined, thereby reducing wear and tear on the
specimens. Bonnie McEwan selected the specimens and
Donna Borne Drake photographed approximately 9,000
elements. The second grant, for 1984-1985, was to install
compact storage units for the zooarchaeological
collection. The units save space by eliminating all but
one aisle that is created by moving whole rows of cabinets
on rails. The third curation grant, 1995-1998, was to
computerize collection data. The museum as a whole
had begun computerization using the SELGEM
system, but with the incredible growth of computer
capabilities we transferred the old SELGEM files to
a relational data base called PARADOX, then once
again to ACCESS, in order to conform with other
museum collections.
During the development and growth of the
zooarchaeology program, computing systems became
transformed. Our first "calculator" was a Monroe adding
machine that would also subtract and was about the size
of an unabridged dictionary. When they first came out, I
bought a Texas Instrument hand calculator at enormous
expense, but with it we were able to get rid of the Monroe
monster. Data entry into the SELGEM system required
using a machine that recorded information on punched
tape. It did this so loudly it required the data entry person
to wear ear protectors the size of large earmuffs. In
1985, we were able to buy an Apple computer through a
gift from Bobby Dorion. We continued to use Apples for
a number of years until communications between
computers and a uniform system throughout the museum
was recommended, at which time we switched to a
Microsoft Word/PC system. With these advances and
the addition of photocopying machine, scanner, laser
printer, and fax machines, we have become quite self-

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

One of the next steps we plan is to wed the
photographic atlas with the web page and make the
photographs available on the Internet. The comparative
collection data are already on the web page. With the
addition of photographs, these data will become even
more useful for research initiated within and outside of
the museum.

An upper-division and graduate-level class in
zooarchaeology has been taught on a regular basis since
1970. Recent instructors have been Karen Walker and,
at present, Susan deFrance. Over the years, students
wanting more advanced studies in zooarchaeology do
independent studies or become involved in some funded
research project.
Graduate students have participated in projects in
the collection and some engaged in masters and doctoral
research in zooarchaeology. During the course of their
work, they have become important contributors to the
program and have enhanced the level of inquiry. Those
students who have engaged in masters or doctoral
research include Kathlene Byrd, Stephen Cumbaa,
Arlene Fradkin, Susan deFrance, Stephen Hale, Laura
Kozuch, Rochelle Marrinan, Bonnie McEwan, Elizabeth
Reitz, Elizabeth Scott, and Karen Walker. Some graduate
students from other institutions have also used the
zooarchaeology collection for their research. These
students include Helene van der Klift and Mark Nokkert
(University of Leiden), Sandrine Grouard and Nathalie
Serrand (National Museum in Paris), Nancy Hamlin
(University of Arizona), Erin Henry (Beloit), Julian Kerbis
(University of Chicago), David Morgan (Tulane), and Susan
Scott (University of Michigan).
The graduate students bring a vitality and excitement
of discovery to the research and activities in the
laboratory. New points of view and questions discussed
are refreshing for everyone. Groups of students organized
events around some projects in the laboratory. These
evolved into social gatherings that brought cohesion to
the group and things to remember and laugh about years
afterwards. For example, the students initiated a crab
feast when we needed data on the ratio of weight of
meat to shell for blue crabs. Wild-meat sampling and
bone-saving parties, with armadillo as the piece de
resistance, became the featured meat at the annual
Armadillo Roast still put on over thirty years later by
graduate students in the Anthropology Department in
honor of Dr. Charles Fairbanks. When the freezer got
too full, the Friday afternoon specimen-preparation parties
got the job done and became a social event. On the
more frivolous side, one Easter the students put on a

Wing: Zooarchaeology at the Florida Museum of Natural History

Mad Hatter competition for the most original or
extraordinary hats, which, after construction, were taken
on parade. Such gatherings did not detract from the
research mission or the sense of responsibility for the
best possible care of the collections, rather they put a
human face on the work and engendered cooperation
and mutual support.
Thus far, several books intended to guide students
in zooarchaeology have been written. The first, entitled
Paleonutrition, by Antoinette Brown and me, was
published by Academic Press in 1979. I supplied the
zooarchaeology background and Brown the physical
anthropology evidence for studies in assessing human
nutrition in the past. This book was followed twenty
years later by a book entitled Zooarchaeology, written
by Elizabeth Reitz and me and published by Cambridge
University Press in 1999. This later book attempts to
update all the methodological and theoretical advances
made in the field of zooarchaeology to that point. Its
extensive bibliography helps students explore the field
beyond the topics discussed in the book. Though the
book has worldwide coverage, its focus is on work in
coastal or riverine zones. Examples presented in the book
include invertebrates and all classes of vertebrates. The
perspective is holistic, integrating data from remains of
plants and animals, soils structure and chemistry, and
human changes to the landscape.
A third book, Case Studies in Environmental
Archaeology, is a collection of papers edited by
Elizabeth Reitz, Lee Newsom, and Sylvia Scudder and
published by Plenum in 1996. It includes papers on a
whole array of topics that relate to past environments.
This book integrates data from different types of
archaeological materials to gain a better understanding
of past human conditions and the impact of human
activity on the landscape and on other animal populations.

1990 TO 2001
With this holistic perspective in mind, we re-named the
collection "Environmental Archaeology" to better reflect
the advances made in understanding the past, of which
zooarchaeology is a part. The new name better expresses
what the collection has become and recognizes the
multifaceted nature of the research approach.
Zooarchaeology continues to be one of the main
features of the collection. Initially the focus of research
was on vertebrates. However, while invertebrates,
primarily mollusks, may be the material that produced
huge prehistoric mounds, the mollusks themselves often
were not studied. That did not make sense because much
can be learned from detailed studies of both mollusks

and vertebrates. One example is the work done by Irv
Quitmyer, who began work in zooarchaeology on a big
project at Kings Bay on the southeast coast of Georgia.
As a part of that project, he used fine-gauge screens to
recover vertebrate and invertebrate remains. After
analysis, he published the results of this major endeavor
in 1985. The technique continues to be a model for
research today. In addition to the research that was done
on that project, Irv has sectioned molluscan shells along
their growth axes and used isotope analysis to study
changes throughout the lives of organisms. This has
produced new information about seasonal occupation of
sites and past temperature regimes. Irv continues to
conduct research in the southeast and Caribbean,
integrating data from the entire range of animals that
are preserved and recovered.
Susan deFrance, assistant professor of anthropology
at the University of Florida, and Elizabeth Reitz, professor
and former director of the Museum of Natural History
at the University of Georgia, each have courtesy
appointments in the FLMNH and conduct research in
the collections as their schedules permit. Both Susan
and Betsy work on faunal material from historic and
prehistoric sites in the southeast U.S. and Caribbean
and Andean countries.
Archaeobotany, the botanical equivalent of
zooarchaeology, is an important component of
environmental archaeology, as well as any research that
attempts to understand past economies and
environmental change. Lee Newsom undertook the first
work in archaeobotany within the zooarchaeology
collection. Her research, a study of the plant remains
from the Hontoon site in Florida, became the subject of
her Master's degree. This was followed by research on
plant remains from a series of West Indian sites that
became the topic of her Ph.D. dissertation. This work
coordinated with research on animal remains, now being
prepared for publication. After graduation, Lee went to
Southern Illinois University and is now at Pennsylvania
State University. Donna Ruhl, a University of Florida
graduate student in anthropology, is continuing work in
archaeobotany. She has worked on plant remains from
both historic and prehistoric sites in the southeast. She is
also maintaining and adding to the archaeobotanical and
comparative plant collections begun by Lee.
Donna Ruhl and Karen Walker have joined forces
on a long-term research project to examine plant and
animal remains excavated from sites in the Everglades.
They are being helped by Lesley Martin and Tanya Peres.
Much of their research adds new information about how
people lived in this unique habitat in the southern-most
part of the state. They are also integrating their finds

and examining environmental changes spanning the time
that people have lived in that area. This important
research exemplifies the integrative approach of the
Environmental Archaeology unit.
Another important component of this unit is the study
of anthropogenic soils. Sylvia Scudder, who manages
the zooarchaeology collection and oversees work in the
unit, completed a Master's degree in soil science in 1993.
Her thesis was a study of the soils at the Horrs Island site
in southwest Florida. This work complemented that of Irv
Quitmyer, who studied the animal remains excavated from
this site. Sylvia has since studied soils from many sites
in the southeast and Caribbean and been able to
document the soil development and landscape changes
resulting from both natural and human forces. This work
was often done in concert with studies of plant and animal

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

remains, such as the projects at Lake Monroe, Archaic
Shell Rings, and Mosquito Lagoon.
The vitality of the Environmental Archaeology unit
owes much to the hard work of the many people
dedicated to the highest standards of research and
collection care. These include people on the museum
payroll, those supported by grants and contracts, and
those who graciously volunteer their help. This research
and curation would not be possible without the support
of the FLMNH and the many archaeologists who have
entrusted and continue to entrust to us the soils and the
plant and animal remains they have excavated for care,
preservation, and study. A special thanks is due to the
National Science Foundation for their funding over the
years that has made the establishment and growth of
this endeavor possible.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 17-26


Sylvia Scudder'

Soil chemical and physical analyses were conducted at the Blueberry site (8HG678) in Highlands County, Florida, to address two
questions: 1) the relationship between a sand-buried Belle Glade Period midden (A.D. 1410-1455) and a much older but more
superficially situated Archaic Period (4500 to 2500 B.P.) fiber-tempered locus, and 2) the nature and origin of the charcoal-laden
sand layer covering the Belle Glade midden. A significant increase in coarse sand and the condition and quantity of organic
carbon deep in the central part of the site imply that that area was formerly a stream or seep drainage separating the two older, more
stable landforms that supported the archaeological deposits. Soil morphology, color, and chemistry, and charcoal and organic
carbon distributions present evidence of past cycles of area-wide fires, episodic erosion onto the site from adjacent higher sand
ridges, and partial recovery of the vegetative cover in the intervening periods between fires.

Key words: archaeopedology, Florida Archaic, geomorphology, landscape change, soils

The Blueberry site (8HG678) in Highlands County,
Florida, is buried within the deep sands of the Lake Wales
Ridge. It contains artifacts from two cultural periods:
the Late Archaic (5000 to 2500 B.P.) and the Belle Glade
(2500 to 300 B.P.) (Milanich 1994:291-297). The two
assemblages are located within 20 m of each other on
the flank of a sand ridge bordering a large wet prairie.
The Belle Glade midden was found approximately 70
cm below the modern soil surface, associated with a
former soil surface now covered by gray, charcoal-laden
sand. The older, fiber-tempered sherds characteristic of
Late Archaic culture were in a topographically higher
position than the Belle Glade midden and closer to the
modern soil surface. Two questions regarding the spatial
and temporal relationships between these assemblages
and their surroundings are addressed in this report.
First, what is the physical relationship between the
buried Belle Glade midden and the locus that yielded
fiber-tempered sherds and second, what is the nature
and origin of the charcoal-laden sand covering the Belle
Glade midden.

The Blueberry site is located in south central Florida
(Fig. 1), at the boundary between the southern tip of the

'Collection Manager, Environmental Archaeology, Florida
Museum of Natural History, University of Florida, Gainesville.
FL 32611, USA.

Central Lakes region and the Eastern Flatwoods region
(Brooks 1981). It lies on a small sand ridge projecting
from the southeast margin of the Lake Wales Ridge (Fig.
2), a Pleistocene beach ridge that runs parallel to the
Atlantic coast (White 1970). From that vantage point, it
overlooks Indian Prairie Basin, a wide swath of wetland
that drains Lake Istokpoga into Lake Okeechobee and
was once part of the Everglades. A deep drainage ditch
at the base of the sand ridge intersects sand-covered
peat, undoubtedly once part of the wet prairie (Fig. 3).
South of the southern terminus of the site, a seep spring
issues from the lower flank of the ridge, flowing in a
small channel through fern, swamp bay, and palmetto
toward the prairie to the east. To the west, a larger sand
ridge, planted in citrus, parallels the long axis of the site.
A perched pond surrounded by palmettos is situated on
the western flank of this ridge.
The archaeological remains were unearthed in a
series of test units arrayed in a north-south line along
the ridge flank (Fig. 3). The units lie in a narrow swath
of relatively level land between the margin of the citrus
grove to the west and a mixed oak/palm woodland
descending to the seep spring to the east. The buried
Belle Glade midden was located at the south end of the
test array, the fiber-tempered material at the north end.
Belle Glade potsherds were also found close to the
modern soil surface at the south end of the site.
Local soils in the vicinity of the Blueberry site have
developed from two different parent materials: sand and

Figure 1. Approximate location of the Blueberry site in High-
lands County, Florida.

the organic matter that accumulated in the wet prairie
basin and smaller wetlands. The deep sandy soils of the
ridges and less-well drained "flatwoods soils" of the
lowlands and swamp edges all developed in marine-
derived beach-ridge sediments deposited during the
Soils evolving from organic material are commonly
known as mucks and peats. They occur on wet prairies,
lake margins, and anywhere that large quantities of plant
material accumulate and degrade under anaerobic
conditions (Soil Survey Staff 1975). Organic soils in
the vicinity of the Blueberry site once occurred as
surface soils at the base of the sand ridge on which
the site is situated, as exemplified by the ditch profile
mentioned above. Today, those soils occur in Indian
Prairie, east of the site, where in some places they are
covered by more than a meter of sand. The plant
communities supported by these wetlands soils range
from open marsh and swamp to forested wetlands and,
when combined with the upland habitats of the sand ridge
areas, would have provided aboriginal inhabitants with a
wide variety of resources.

Human effects on natural soils and landscapes can
be detected using the basic tools of soil science: soil

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

morphological descriptions and interpretations, particle-
size distribution analysis, and patterns of chemical
element accumulation. One of the most widely used
techniques to describe soils physically is particle-size
analysis, which is used to determine grain-size class
distributions with depth and to relate those distributions
to soil-forming processes and the dynamics of landscape
evolution (Farrand 1975; Hassan 1985).
Chemical analyses are used to measure the soil
content of such elements as phosphorus, calcium,
magnesium, and some trace minerals that accumulate
as a result of human habitation, then to compare those
levels with residual levels of the same elements in local,
native (non-human-impacted) soils (Conway 1983; Eidt
1985; Lillios 1992; Lippi 1988; Woods 1977). These
patterns of element accumulation can be used to
delineate site boundaries and to locate and interpret the
function of such intrasite features as hearths, storage
pits, and burials. The actual quantitative differences
between native and anthropogenic soil-element contents
are a measure of habitation duration and intensity.
Measurement of pH (acidity) of the soil can be used to
explain the physical condition or absence of bone, the
presence or absence of pollen or phytoliths, and the
preservational status of artifacts.

Soil samples were removed from the center of each
natural horizon and midden layer of the south profile of
test unit 985N/1004E, which exposed the Belle Glade
midden. Manually augered samples were taken adjacent
to the other test units, which had been back-filled prior
to general soil sampling, and at higher elevation on the
ridge at the northern end of the site (Fig. 3). Individual
samples of approximately 300 to 400 g were removed in
10-cm increments, bagged, labeled by soil test (ST)
number and depth, and transported to the Florida Museum
of Natural History in Gainesville for further preparation
and curation. Field descriptions of augered soils included
soil horizon designation, color, texture, and relative
moisture. The descriptions also included the approximate
depths at which these characteristics changed.

Soil samples were air-dried and sieved through 2 mm
screen in the Environmental Archaeology laboratory at
the Florida Museum. Catalog numbers were assigned
and two sets of 50 g subsamples were removed from

SCUDDER: Deep Sand: Soil and Landscape Relationships at the Blueberry Site

Figure 2. Regional physiographic features in the vicinity of the Blueberry site (after White, 1970).

bulk samples. One set was sent to the Environmental
Pedology laboratory in the Soil and Water Sciences
Department, University of Florida (UF), for particle-size
distribution analysis using the pipette method outlined by
Day (1965). Samples judged to have greater than 1%
organic carbon (estimated by a gray or dark gray color)
were pre-treated with hydrogen peroxide and heat to
digest the organic material before particle-size analysis
proceeded. The second set of subsamples was sent to
the Analytical Research laboratory on the UF campus
for analysis of extractable soil elements. That
procedure uses a filtrate extracted with 0.05 N
hydrochloric acid in 0.025 N sulfuric acid. The
extractant was analyzed using inductively coupled
argon plasma (ICAP) spectroscopy.

The ridge soils at the Blueberry site are from 92 to
99.7% sand, which identifies their parent material as
"coarse marine sediments" (Carter et al. 1989). The
overwhelming proportion of total sand in the soil tells us
little more than that. To understand more about internal
relationships among sub-areas of the site, the individual
sand-size classes were studied, as well as size-class
ratios that compare "fine" sizes (very fine and fine
designations) with "coarse" ones (coarse and very coarse

designations). These ratios are clarified by not including
medium-sized sand, the most abundant fraction in the
Blueberry site soils.
Table 1 summarizes particle-size distribution and
particle-size ratios among natural horizons in all soils
sampled. Medium sand dominated virtually all horizons
in all tests, with a range of 50.6 to 79.5% by weight. The
second most abundant size class, fine sand, ranged from
10.6 to 28.4% in all but four samples. The relationship
between fine- and medium-sand content changed with
depth over the site: medium-sand content generally
decreased with depth at each test, while fine-sand content
increased with depth.
Most horizons contained from 4 to 8% coarse sand.
Exceptions were found in the two deepest subhorizons
of ST #4, which contained 15.4 and 12.2% coarse sand.
Subsurface horizons of ST #5 and #6 also showed
increased coarse sand content, up to 21%.
Silt content varied considerably over the site area,
ranging from <1.0 to 6.9%. There was no discernable
pattern of accumulation or depletion with depth. In some
tests, silt content and coarse sand content varied
inversely-when one fraction was well-represented in
a particular horizon, the other was scant. Conversely, in
some tests the increase in coarse sand was accompanied
by an increase in silt. Clay content across the site was
generally less than 1%.

GROVE Fh er- S[
tempered 0 /
b1] 7 SIX locus / /
.\ / S
ST I ST 4 ST 5 ST6

Midden ST 3 Cwle


Figure 3. Blueberry site test units and soil test (ST) localities.

The highest charcoal content occurred in the A
through Ab horizons-the modem soil surface-and in
the E horizons and the buried midden level. Charcoal
content of the subhorizons of the soil buried below the
midden was minimal, except the fraction of very fine
sand in the buried E horizon.

All the soils at the Blueberry site have an A horizon
composed of a mixture of organic matter and mineral
sand. This horizon varies in darkness and thickness over
the site, reflecting differences in organic inputs from
vegetation and, in some cases, a partial mantling by
cleaner aeolian and colluvial sands. A cross-section of
the south end of the site perpendicular to the long N-S
axis is shown in Fig. 4. The buried A horizon with midden
(designated Ab/midden) can be seen in the ST #2 and
985N columns, with the suggestion that the dark A3
horizon of ST #1 may be a continuation of that former
surface horizon. Soil tests 3 through 6 showed no
evidence of buried surfaces. The soil color lightened with
depth, indicating that the densest accumulation of
organic material was on the soil surface. Soil test #7
contained a thin, intermittent Ab horizon at
approximately 130 to 140 centimeters below surface
The leached E (eluvial) horizon underlying the A
horizon varied in thickness and depth across the site. At
the south end, it was underlain by a light brown to dark
brown B horizon (zone of accumulation). At the north
end, the very deep and well-developed soil at ST #10
had two E horizons separated by B horizons that darkened

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

with depth (Fig. 5). The second E horizon began at 400
cmbs and was the last sample possible to take without
losing the auger to the forces of friction and gravity.
What underlies this remains an unanswered question. In
the central tests, no dark horizon was encountered
beneath the E horizon, although samples taken from close
to the water table in the zone of saturation contained
small black flecks of organic matter suspended in the
soil water. These flecks were not attached to the clear
quartz sand grains that comprised the pale matrix of the
soil-as they would be in a darkly colored horizon-
rather, they settled out in the liquid portion of the samples.
The B horizon also varied across the site. As
mentioned above, it was absent, or at least not
encountered at the depths augered, in the central tests.
In these areas, stripped white sands with black particles
suspended in the soil water were encountered below
the A horizon. A Bw horizon (beginning of color
development) occurred in tests 1 through 5 and again in
10. A Bh horizon (containing dark humus), articulating
with the water table, was evident at tests 1 through 4
and in non-saturated conditions at test 10.
The morphology of the soil encountered at ST #9
differed from all other areas and represented the margin
of wetland organic-soil development. This soil had a gray
sandy A horizon and an intermediate, mixed organic and
mineral AE horizon, all underlain by a blackish-red peat.

Table 2 summarizes chemical element contents and
pH measurements for all samples. Measurement of pH
indicate that the soils at the Blueberry site are mildly to
moderately acidic, becoming more acidic with depth. The
lowest surface acidity (highest pH, or alkalinity) was
generally associated with the high calcium content of
midden deposits.
Organic carbon (OC) content was generally less
than 1% and highest in the dark surface horizons, the
buried A/midden horizon, and in BwBh or Bh horizons.
Across the site, organic carbon content was lowest in
the E horizon and decreased with depth.
Highest soil calcium (Ca), magnesium (Mg), and
phosphorus (P) contents were found in the A horizon of
ST #2, 3, 4, and 5, and in the Ab/midden horizon of unit
985N. Soil Ca and Mg decreased irregularly with depth
to the deepest levels sampled, then increased again as
soil moisture increased, except in tests 8 and 10, and in
unit 985N. Soil P content was lowest in E horizon samples
and increased in subadjacent Bw and BwBh horizons.

SCUDDER: Deep Sand: Soil and Landscape Relationships at the Blueberry Site

Table 1. Particle-size distribution analysis (wt.%), Blueberry site (8HG678).

unit/test horizon (cm)

ST#1 Al 20
A3 70
E 150
BwBh 290

ST #2 A 20
El 60
ABI 80
Mid/Eb 140
Eb 160
BhBw 300

ST#3 Al

ST#4 Al 25
AE 75
E 115
EBw 155
BwBh 205

ST#5 Al 20
El 80
EBw 130
E 210

ST#6 Al 20
E 80
Ew/bl 140

ST #7 A2 20
E 100
Eb 180
E w/bl 230

ST#8 Al 20
AE 80
E 150
W w/bl 280

ST#10 Al 20
AE 80
EBw 140
Bw2 200
Bh 380
E' 400

985N/ Al 5
1004E A2 25
AE(E) 50
Abl 75
Ab2 100
AbE 115
Ebl 140
Eb2 185
Bwb 210
BwBhsb 250

sand size*

F VF sand silt

6.0 72.4 20.0 1.0
6.0 73.2 19.4 0.8
4.8 66.2 25.8 1.4
5.4 68.4 24.0 0.8

6.8 73.2 16.4 0.6
5.0 74.8 18.4 0.8
6.4 75.4 16.6 1.0
5.8 71.0 21.2 0.8
6.2 71.6 20.6 1.0
6.8 67.6 22.2 0.8

2.2 20.2 68.8 0.8
6.0 72.0 19.0 0.2
6.0 53.2 35.6 1.8
3.4 50.6 45.2 0.2

7.0 73.2 14.2 0.4
5.4 74.0 19.8 0.4
4.6 42.2 48.0 0.4
15.4 55.4 27.4 0.4
12.2 55.2 28.4 1.2

9.2 76.6 10.6 0.4
7.8 76.0 14.4 0.6
21.2 52.6 15.2 2.0
5.2 65.2 26.6 2.0

7.4 72.2 16.6 1.0
21.0 55.4 15.2 0.8
7.2 71.2 19.6 1.0

8.6 77.2 12.0 0.4
6.4 75.4 16.8 0.8
5.4 71.0 22.0 0.4
5.0 69.4 23.8 1.0

8.2 78.2 11.4 0.2
6.4 75.6 16.0 0.8
5.4 71.4 20.6 1.2
5.0 70.6 21.8 1.6

8.6 77.6 10.8 0.8
5.8 77.0 15.0 0.8
6.8 77.2 13.4 0.8
4.6 70.8 22.2 1.0
6.6 76.0 14.6 0.8
4.0 63.6 28.2 2.8

8.0 77.8 13.0 0.1
9.5 79.5 9.6 0.0

0.0 9.9 71.1 15.7 0.8

0.0 9.8 74.3 15.4 0.2

0.0 9.7 73.6 15.6 0.1

99.4 0.6
99.4 0.6
98.2 1.7
98.6 1.4

97.0 2.0
99.0 1.0
99.4 0.4
98.8 0.8
99.4 0.6
97.4 2.5

92.0 6.9
97.2 2.6
96.8 2.1
99.4 0.4

94.8 4.5
99.6 0.2
95.4 3.7
98.8 0.3
97.6 2.0

96.8 2.4
98.8 1.2
93.0 6.8
99.0 1.0

97.2 2.8
94.4 5.3
99.0 0.3

98.2 1.8
99.4 0.5
98.8 1.1
99.2 0.7

98.0 1.4
98.8 0.6
98.6 0.2
99.0 0.8

97.8 1.9
98.6 1.4
98.2 1.4
98.6 1.2
98.0 1.9
98.6 1.2

99.0 0.0
98.7 0.0

97.8 0.1

99.7 0.0

99.1 0.0

* Sand-size class designations: VC = very coarse. C = coarse. M = medium, F = fine, VF = very fine.

clay ratio

0.1 3.4

2.1 1.6

0.3 1.6

0.9 1.6

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

I 985N/ I

I ST 1

iAb-intermittert -
AE --------

Abm/idden Ab2/nidden
Ab/midden -' -
/ AEb


S A3- darkest








t saturated

Figure 4. Cross-sectional profile of south end of site.

Aluminum content was highest in horizons with the
most OC (A horizons and buried A/midden horizons)
and in the "colored" Bw and BwBh or Bh horizons. It
was lowest in the E horizon across the site, and lowest
throughout the soil column at ST #7 and #8.
Iron content was highest in the most well-developed
soil profile, ST #10, and in the Ab/midden horizon of unit
985N. Within the other tests it was generally lower in E
horizon samples and higher within B horizons.

Local variations in soil characteristics suggest a
former landscape quite different from that which
constitutes the Blueberry site today. Midden thickness
(Fig. 4) increases to the east, downslope toward the edge
of the wet prairie. The dark A3 horizon of ST #1, to the
west of unit 985N/1004E, may be the up-slope edge of
the midden itself, but is certainly contiguous with the
former soil surface. It is at the same depth as the true

midden layers in the other two tests and shows a similar
pattern of accumulation of organic C, Ca, and Mg, though
absolute contents of those constituents are not as high
as in well-defined midden levels. One can envision the
Belle Glade Period inhabitants living on the flank of the
ridge, using the resources of the seep spring, upland
sandhills and ponds, wet prairie marginal woodlands, and
the huge open marsh itself. The debris from their
habitation site and activities spread downhill to
accumulate in a wedge-shaped deposit on the sloping
foot of the ridge. Although the south end of the site is
now essentially flat, the topography of the subsurface
soil horizons underlying that area indicates that the land
sloped more abruptly toward the prairie when the midden
was forming. Because the formation of Bh horizons in
the southeastern United States is dictated by the position
of the water table (Garman et al. 1981), and the water
table under a gently rolling landscape generally parallels
the ground surface (Birkeland 1984), it can be inferred
that the subsurface horizons here formed in mounded


Soil surface >


SCUDDER: Deep Sand: Soil and Landscape Relationships at the Blueberry Site

Figure 5. Cross-sectional profile of long axis of site.

sediments and not in flat, horizontally bedded ones. The
midden itself was deposited on this sloping landform.
Other variations in the "subsurface landscape" begin
to address the relationship between the buried Belle Glade
midden and the northern end of the site with its older,
fiber-tempered sherds. The profile of the long axis of
the site indicates that, although the north and south ends
of the site are now connected by a slightly concave stretch
of ground, in the past they may have been less directly
connected and were certainly separated by more relief.
At the north end of the site, a deep sequence of gray-
to-brown soil horizons can be seen. The gray A and AE
horizons change with depth, first to the brown Bw horizon
and then to the very dark grayish-brown Bh horizon. The
Bw horizon is tinted by iron oxides, very small amounts of
which can color the stripped "white" sand grains brown or
red. These colors can develop only in well drained soils.
The humic-stained Bh horizon is composed of sand grains
coated with a combination of aluminum oxides and organic
carbon, both transported by a fluctuating water table. As

with the Bw horizon, the Bh horizon must have formed
under at least partially aerobic conditions. In addition, the
Bh horizon at ST #10 is underlain by a second E horizon 4
m below the surface, indicating a second round of horizon
differentiation under the influence of a lower water table.
This profile presents evidence of a soil that has undergone
extensive pedogenesis on a very stable land area under
well-drained conditions.
At the south end of the site the soil also shows signs
of long-term development under conditions favorable to
the formation of a Bh horizon. However, between the
north and south ends of the site, some interesting
subsurface transitions have occurred. No B horizon of
any kind was found in the central area. The deepest E
horizon in this central area consisted of stripped quartz
grains with black organic carbon flecks suspended in
the soil water. The zone of saturation itself was closer
to the surface than at either end of the site, and
terminated in the "E with black flecks" horizon at each
test. The lack ofa B horizon means that conditions were

ST 3 "Perched pond- ST10
Soil surface > 3 ST4 4 1
A ~ST7 ST 8^ -
Al ST5 A2 darker
Al A A3
31 Ab/midden Ai"- T Al Fiber- -
SBelA2 A2 Al tempered AE
Belie Glade A2 -A
- AF \v A2 4- locus _
Midden AE AE Al2 EBwl
Eb E E El AE ElBw
o E2 E2 El /
30 " E / Bw2
Bw Bw EBw
S EBwb .- ..: ST,
P-' Eb '9
S BwEb BwBh Bh w/ E w/ bl. E2 s Bw3
-h- black E w
29 BwBhb flecks Eb w/ bl. A "
Bh E w/bl. OA BhBw

S= coarse sand lens Bh



never sufficiently drained for the oxidizing or
accumulative part of the horizon-formation process to
occur. These lower levels probably were always
saturated and more related to the prairie edge than the
sand ridge. In addition, one of the most abrupt particle-
size distribution changes occurred here. Coarse and very
coarse sand content virtually tripled in a deep, centrally
located lens-shaped area.
The coarse sand lens in the subsurface horizons,
the slightly lower elevation of the ground surface in the
center of the site, and the perched pond to the west of
this area on the flank of the higher ridge suggest that the
central area may have once been an area of drainage
from the pond or a seep similar to the one now flowing
just south of the site. This drainageway, or depression,
would have been downhill from the higher, stable area
around ST #10, including the locus that produced the
fiber-tempered sherds. Across the way to the south, at
the level of the Belle Glade midden, was another slightly
elevated, stable area that would someday be populated
by the Belle Glade inhabitants. In essence, what the
subsurface soil horizons and local landscape features
reveal is a picture of relationships no longer in existence:
two relatively stable and well-drained end points
separated by a lower swale or drainage and used at
different times by different peoples-and all being slowly
buried by sand.

The two intermittent A horizons found in unit 985N
are traces of a thin surface accumulation of OC separated
by varying thicknesses of charcoal-containing light gray
sand. These partial horizons, coupled with the pale, poorly
developed Al horizon underlain by a darker, more
enriched buried A, indicate a generalized and episodic
burial of parts of the site by sand moving down from the
higher ridge areas. Particle-size analysis indicates that
most of the sand moved colluvially, or at least that sand
of the same size-classes moved locally. Although there
is slightly more silt in some surface horizons-indicating
some aeolian input-the medium-sand fraction does not
change. Consequently, there was no substantial
importation, by wind, gravity, or human endeavor, of soil
or sediments of a very different character from those
that already existed on the site.
The particle-size data, coupled with the faint build-
up of OC in the intermittent A horizons, the homogeneous
distribution of fine charcoal in the A and AE horizons in
unit 985N, and the absence of charcoal in the midden-

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

buried horizons can be used to address the origin and
nature of the charcoal-laden sand covering the Belle
Glade midden. A reasonable scenario is this: 1) the
landscape occupied during Belle Glade times (a period
believed to be cooler and drier than today, see below),
was subjected to fire and denuded of vegetation, 2) sand
eroded down the ridge flank and was briefly stabilized
by colonizing vegetation that began forming a thin A
horizon, then, 3) area-wide fires swept the landscape
again and the cycle was repeated.
Whether the fires producing the evenly distributed
charcoal in the upper horizons of the Blueberry site soil
were human- or lightning-induced is not within the scope
of this study. However, studies of sea-level changes and
their effects on climate in the Pleistocene and Holocene
epochs are pertinent. Researchers have found evidence
of low-amplitude sea-level oscillations worldwide, in beach-
ridge sets (Tanner 1992), tropical ice packs (Thompson et
al. 1988), zooarchaeological faunal samples (Walker et al.
1995), stable isotope ratios (Hodell et al. 1991), and wave-
cut beach rock (Fairbridge 1961, 1974). As the more
conservative view of a smooth sea-level rise gives way
to multiple lines of evidence for a punctuated, episodic
series of rises and falls, the timing of these episodes
becomes more refined (Walker et al. 1995). Certain dates
proposed by individual researchers for particular
transgressive or regressive events in localized geographic
areas begin to overlap with dates from other areas. One
range of dates that correlates with a recent lowstand,
the "Little Ice Age," is A.D. 1430 to 1850 [This range is
according to Gribbin (1978); other authors broaden or
contract the range according to regional data, e.g., Eddy
(1977), sets the range at A.D. 1250 to 1920.] The
calibrated radiocarbon date obtained from charcoal in
the buried Belle Glade midden at the Blueberry site-
A.D. 1410 to 1455 (Beta-83917)-falls within the Little
Ice Age. Conditions in Florida during sea-level lowstands
were generally cooler and drier. As sea level fell, fresh-
water lake levels also fell, along with inland river
discharge rates, rainfall volume from convective weather
systems, and the subsurface water table (Widmer 1988).
Plant communities adapted to more xeric conditions,
becoming more vulnerable to the effects of fire and other
disturbance. These conditions prevailed during the Belle
Glade Period of habitation of the Blueberry site, circa
A.D. 1430. On-site evidence of area-wide fires, in the
form of intermittent soil surface horizons and well-mixed
charcoal in the upper horizons, corroborates this
interpretation of climatic conditions at that time.

SCUDDER: Deep Sand: Soil and Landscape Relationships at the Blueberry Site

Table 2. Chemical analyses, Blueberry Site (8HG678).

Unit/ Depth Organic Element content (mg/kg)
Test Horizon (cm) pH carbon* Ca Mg K P Mn Zn Cu Na Al Fe

ST#1 Al 20
A3 70
E 150
BwBh 290

ST #2 A 20
El 60
AB1 80
Mid/Eb 140
Eb 160
BhBw 300

ST#3 Al

ST#4 Al 25
AE 75
E 115
EBw 155
BwBh 205

ST#5 Al 20
El 80
EBw 130
E 210

ST#6 Al 20
E 80
Ew/bl 140

ST #7 A2 20
E 100
Eb 180
Ew/bl 230

ST#8 Al 20
AE 80
E 150
Ww/bl 280

ST#10 Al 20
AE 80
EBw 140
Bw2 200
Bh 380
F 400

0.3 118.0 7.2 9.2 1.2 2.22 0.72 0.04 1.4 12.9 6.64
0.6 275.0 5.7 3.1 4.4 0.79 1.53 0.00 2.4 30.6 6.39
0.1 27.1 0.9 0.0 0.9 0.26 0.30 0.00 0.7 7.2 6.46
0.6 113.0 19.9 6.9 18.7 0.89 1.87 0.00 1.2 158.0 6.76

1.0 1370.0 102.0 200.0 141.0 24.90 5.79 0.00 8.6 23.7 1.46
0.3 196.0 12.8 3.0 4.2 0.93 0.85 0.00 0.7 5.4 0.80
0.3 223.0 13.7 4.6 8.1 1.10 0.95 0.00 0.3 5.8 0.92
0.3 405.0 11.6 1.0 108.0 4.86 3.35 0.00 3.7 24.2 5.32
0.2 81.6 3.0 0.2 15.2 1.61 1.04 0.00 0.8 5.6 3.29
0.6 60.3 13.0 6.9 77.8 1.04 0.69 0.00 1.7 447.0 5.62

0.7 2190.0 57.1 15.1 581.0 6.90 11.00 0.01 12.8 48.3 2.43
0.3 337.0 16.9 1.7 14.2 0.44 0.65 0.02 0.7 9.6 1.75
0.2 71.8 3.6 0.0 6.6 0.52 0.54 0.00 0.3 5.0 2.57
0.3 87.0 7.2 2.6 22.1 0.52 0.64 0.00 1.3 85.4 2.98

1.0 2720.0 82.9 19.6 687.0 3.96 10.90 0.00 12.9 54.5 3.24
0.3 237.0 6.7 0.1 28.9 0.78 1.04 0.00 0.7 5.5 1.30
0.2 72.4 3.0 1.5 6.3 0.48 0.84 0.00 1.0 2.1 1.14
0.2 54.8 2.1 0.0 16.1 0.20 0.47 0.00 0.3 23.1 4.62
0.3 168.0 10.9 2.9 26.5 1.21 1.27 0.00 1.4 71.1 2.47

0.9 1190.0 75.1 7.6 28.3 2.16 8.32 0.01 1.8 29.5 1.77
0.2 50.6 3.3 1.3 0.4 0.10 0.39 0.02 0.6 2.3 2.95
0.2 28.6 2.1 1.5 0.5 0.20 0.40 0.04 1.0 3.3 3.97
0.1 47.7 5.2 3.4 0.7 0.39 1.35 0.01 1.4 5.4 1.75

0.2 738.0 48.1
0.2 36.7 2.8
0.1 92.7 8.1

8.4 13.8 9.10 8.40 0.03 1.3 21.0 1.54
1.2 1.0 0.25 0.36 0.00 0.3 1.3 1.34
2.9 0.8 1.65 2.22 0.01 1.2 3.2 3.65

0.5 171.0 26.9 2.8 0.2 0.51 0.84 0.00 0.8 17.1 3.95
0.2 38.2 2.8 3.0 0.1 0.76 0.63 0.00 0.7 1.7 1.86
0.2 18.2 3.9 0.8 0.0 0.48 0.54 0.00 2.4 2.0 2.50
0.2 24.7 5.7 2.0 0.0 0.62 3.43 0.00 2.2 1.9 2.19

0.5 617.0 32.0 12.7 1.3 5.52 3.12 0.00 2.6 13.0 1.50
0.2 53.9 6.0 9.9 0.3 0.44 0.28 0.03 1.7 5.4 3.26
0.1 20.1 1.8 0.3 0.0 0.14 0.22 0.00 0.6 1.5 1.77
0.2 10.9 3.6 2.8 0.0 0.32 0.37 0.00 1.4 1.4 1.78

0.7 296.0 22.5 11.7 2.3 5.34 0.74 0.00 1.9 18.0 3.99
0.2 30.2 4.8 2.4 0.2 0.19 0.21 0.00 0.8 13.5 5.78
0.2 19.6 2.5 0.7 2.5 0.83 0.36 0.00 1.9 16.4 11.70
0.3 18.8 2.4 0.4 47.8 0.37 0.40 0.00 8.2 114.0 54.40
0.5 41.1 9.6 3.2 39.1 1.74 1.18 0.00 3.2 268.0 8.88
0.2 21.8 5.8 3.2 11.2 0.62 1.76 0.00 2.9 61.4 5.99

*Organic carbon is recorded as weight%.

I thank Robert Austin, archaeologist with Southeast
Archaeological Research, Inc., for inviting me to
participate in this project, which was part of his Ph.D.
dissertation research. For logistic support, excellent
volunteer labor, and financial support of analysis, I
thank the Kissimmee Valley Archaeological and
Historical Conservancy, and particularly KVAHC
president Anne Reynolds, who also provided lodging
by the lake. Most of all I thank Elizabeth Wing,
curator of Environmental Archaeology at the Florida
Museum of Natural History, for encouraging me to
initiate-and for enthusiastically supporting-
archaeological soils studies at the FLMNH.

Birkeland, P. W. 1984. Soils and Geomorphology. New York:
Oxford University Press.
Brooks, H. K. 1981. Physiographic Divisions of the State of
Florida (map). Gainesville: Center for Environmental and
Natural Resources. Institute of Food and Agricultural
Sciences, University of Florida.
Carter, L. J., D. Lewis, L. Crockett, andJ. Vega. 1989. Soil Survey
of Highlands County, Florida. Washington, D.C.: U.S.
Department of Agriculture, Soil Conservation Service.
Conway, J. S. 1983. An investigation of soil phosphorus
distribution within occupation deposits from a Romano-
British hut group. Journal of Archaeological Science 10:
Day, P. R. 1965. Particle fractionation and particle-size analysis.
Pp. 548-567 in C. A. Black, ed. Methods of Soil Analysis,
Part 1. Madison, Wisconsin: American Society ofAgronomy.
Eddy, J. A. 1977. The case of the missing sunspots. Scientific
American 236(5): 80-92.
Eidt, R. C. 1985. Theoretical and practical considerations in
the analysis of anthrosols. Pp. 155-191 in G. Rapp and J. A.
Gifford, eds. Archaeological Geology. New Haven,
Connecticut: Yale University Press.
Fairbridge, R. W. 1961. Eustatic change in sea level. Pp. 99-185
in L. H. Ahrens, K. Raukawa, and A. K. Runcorn, eds.
Physics and Chemistry of the Earth, vol. 4. New York:
Pergamon Press.
Fairbridge, R. W. 1974. The Holocene sea level record in south
Florida. Pp. 427-436 in P. J. Gleason, ed. Environments of
South Florida: Present and Past. Coral Gables, Florida:
Miami Geological Society.

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Farrand, W. R. 1975. Sediment analysis of a prehistoric rock
shelter: The Abri Pataud. Quaternary Research 5: 1-26.
Garman, C.R., V.W. Carlisle, L.W. Zelazny, and B.C. Beville.
1981. Aquiclude related spodic horizon development. Soil
and Crop Science Society of Florida Proceedings 40: 106-
Gribbin, J., ed. 1978. Climatic Change. New York: Cambridge
University Press, pp. 70-72.
Hassan, F. A. 1985. Paleoenvironments and contemporary
archaeology: A geoarchaeological approach. Pp. 85-101
in G Rapp and J. A. Gifford, eds. Archaeological Geology.
New Haven, Connecticut: Yale Univ. Press.
Hodell, D. A., J. H. Curtis, G A. Jones, A. Higuera-Gundy, M.
Brenner, M. W. Binford, and K. T. Dorsey. 1991.
Reconstruction of Caribbean climate change over the past
10,500 years. Nature 352: 790-793.
Lillios, K. T. 1992. Phosphate fractionation of soils at Agroal,
Portugal. American Antiquity 57(3): 495-506.
Lippi, R. 1988. Paleotopography and phosphate analysis of a
buried jungle site in Ecuador. Journal of Field Archaeology
Milanich, J. T. 1994. Archaeology of Precolumbian Florida.
Gainesville: University Press of Florida.
Soil Survey Staff. 1975. Soil Taxonomy. U.S.D.A. Soil
Conservation Service Agriculture Handbook no. 436.
Washington, D.C.: U.S. Government Printing Office.
Tanner, W. E 1992. Late Holocene sea-level changes from
grain-size data: Evidence from the Gulf of Mexico. The
Holocene 2: 249-254.
Thompson, L. G, M. E. Davis, E. Mosely-Thompson, and K-B.
Liu. 1988. Pre-Incan agricultural activity recorded in dust
layers in two tropical ice cores. Nature 336: 763-765.
Walker, K. J., F. W. Stapor, and W. H. Marquardt. 1995.
Archaeological evidence for a 1750-1450 BP higher-
than-present sea level along Florida's Gulf coast.
Holocene Cycles Climate, Sea Levels, and
Sedimentation. Fort Lauderdale, Florida: Coastal
Education and Research Foundation, Special Issue
no. 17, pp. 205-218.
White, W. A. 1970. The Geomorphology of the Florida
Peninsula. Tallahassee: Florida Dept. of Natural
Resources, Bureau of Geology, Geological Bulletin no.
Widmer, R. J. 1988. Evolution of the Calusa-ANon-Agricultural
Chiefdom on the Southwest Florida Coast. Tuscaloosa:
University of Alabama Press.
Woods, W. I. 1977. The quantitative analysis of soil phosphate.
American Antiquity 42(2): 248-252.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 27-34



Kathlyn M. Stewart' and Rebecca J. Wigen2

There has been much discussion in the archaeological literature on the utilization of different screen mesh sizes for recovering
faunal elements, with many researchers decrying the use of the 6.4 mm (1/4") screen, which allows small elements to fall through
the mesh. There has been less discussion on the merits of the data obtained through recovery of smaller elements, specifically, is
the "new data" worth the extra time and labor? This paper examines faunal material from three sites from the Northwest coast of
Canada. The material has been recovered using either 6.4 mm or 2.8 mm (1/8") mesh screens. The results suggest that elements of
herring and other small fish have been greatly underestimated in Northwest coast sites, and that these formed an important part
of the coast economy and subsistence. Where salmon, a large fish with often well preserved elements, has been seen as the
mainstay of Northwest diet and economy, excavation with small-mesh screens may indicate a much greater importance for herring
and other small fish on the coast. For accurate reconstruction of past lifeways, at least on the Northwest coast, screens with mesh
2.8 mm must be used for at least a substantial part of the matrix, in combination with the 6.8 mm mesh.

Key words: screen gauge, fish bones, Northwest coast, sampling, North American prehistory

The question of mesh size for screening archaeological
matrix has been the focus of much debate (eg., Butler
1993; Cannon 1999; Clason and Prummel 1977; Gordon
1993; Grayson 1984; James 1997; Lyman 1982; Payne
1972; Rick and Erlandson 2000; Schaffer 1992; Schaffer
and Sanchez 1994). Traditionally the 6.4 mm (1/4") and
even larger meshes have been used in archaeological
excavations, but their utilization has come under attack
because of loss of small bone elements through the mesh
and consequent loss of faunal data. But how important
are these lost data? While much debate focuses on the
faunal elements lost and optimal strategies for recovery,
there is less discussion about the value of these elements
to our overall knowledge of the cultures studied or the
reconstruction of the associated environments and animal
communities. Several studies have undertaken recovery
of fauna with <6.4 mm mesh, and have added the "new"
taxa to their lists of resources exploited, and calculated
resulting diversity indexes and/or edible meat weights
(e.g., Gordon 1993; Rick and Erlandson 2000). Is this
additional knowledge worth the considerable time, labor,
and cost needed for screening, sorting, and identifying
small faunal elements from small-mesh screens?

' Canadian Museum of Nature, PO Box 3443, Stn D, Ottawa,
2Dept of Anthropology, University of Victoria, Victoria, BC,
For communication on this paper, please contact the senior

In this paper, we analyze fauna from three
Northwest coast shell midden sites where either 6.4 mm
or 2.8 mm (1/8") mesh screens were utilized. Depending
upon which screen size was used, completely divergent
faunal results were obtained. These divergent results
mean that completely different representations of
Northwest coast economy and subsistence can be
inferred, depending upon which set of data is consulted.
Our study illustrates that, at least in Northwest coast
shell middens, accurate subsistence and economic
reconstruction is absolutely dependent on recovery of
the faunal elements through screens with <2.8 mm mesh.
We suggest that 50% of the matrix be screened through
<2.8 mm mesh.

In 1994, 1995, and 1997, the University of Victoria field
school under the direction of Don Mitchell, Quentin
Mackie, and Becky Wigen began excavating the
Kosapsom site (DcRu 4) located on southeastern
Vancouver Island along the Gorge waterway in the
present-day Victoria area (Fig. 1). The excavation was
in part a salvage recovery prior to future construction of
a sidewalk. Thirty-eight 1 x 1 m excavation units were
randomly and subjectively selected and opened in a strip
that more or less followed the shore along the Gorge.
Although the deposits were somewhat mixed, there was
evidence of three distinct cultural components, each of
which concentrated in different areas of the site: Locarno

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Figure 1. The Kasapsom site, southern Vancouver Island.

Beach Phase, dating from about 3500 to 2500 B.P., Gulf
of Georgia Phase, dating from about 1500 B.P. to contact,
and Historic, dating from 1855 A.D. to the present.
In the 1994 field season, all material recovered from
excavation by 5 cm levels was screened using a 2.8 mm
screen. Recovery of the small bones from the screens
in the field was so time-consuming that the decision was
made to bag the material from the 2.8 mm screens and
finish the sorting in the lab. In addition, a 20 x 20 x 5 cm
sample was taken from a corner of each excavation
unit. Retrieved material was taken to a laboratory at the
University of Victoria for sorting. Unfortunately, the time
required for sorting the material into the categories of
shell and vertebrates was enormous, and after four
weeks of full-time work by several students, little of the
material had been sorted. With so little material
processed, the decision was made to use a 6.4 mm screen
for recovery in the new units opened during both the 1995
and 1997 field seasons. The 2.8 mm mesh was used for
column samples.
All vertebrate material was identified by one author
(KS) and invertebrate material by the other (RW),
although the latter is not discussed in this paper.
Vertebrate material was identified using both the

collections of the Canadian Museum of Nature in
Ottawa, and the University of Victoria osteological
collections. In this paper we present the results of the
Kosapsom faunal analysis, and also compare the results
with the fauna from two nearby and reasonably
contemporaneous sites, Esquimalt Lagoon (DcRu 2)
and Fort Rodd Hill (DcRu 78). These sites are about 3
and 4.5 km, respectively, west of the Kosapsom site on
the Gorge waterway.
At Esquimalt Lagoon, fourteen 1 x 2 meter units
were excavated and fauna collected by 10 cm levels.
Matrix excavated was screened through a 6.4 mm mesh
screen. All vertebrate material was identified to element,
but bird elements were not identified to taxon (Hanson
1991). At Fort Rodd Hill, five 2 x 2 meter units were
excavated, and matrix screened through a 5 mm (1/5")
mesh screen. All vertebrate fauna was analyzed (Hanson
1991). Comparisons with the Esquimalt Lagoon and Fort
Rodd Hill faunas are based on numbers in Hanson (1991).

To date, a total of 3707 elements have been analyzed
from the Gulf of Georgia levels at Kosapsom, from three
classes (Table 1). As can be seen in the total numbers,

STEWART and WIGEN: Screen Size and the Need for Reinterpretation

fish are most abundant, with considerable diversity, while
mammals are next in abundance, but with less diversity.
Birds were poorly represented, with very little diversity.
When the faunal analysis is sorted by screen size.
however, two very different pictures of the fauna emerge
(Fig. 2). In the fauna recovered using 2.8 mm mesh, fish
is clearly the dominant class, comprising 75.8% of the
total number of elements, compared to less than half
(32%) recovered in the 6.4 mm mesh. A similar reversal
occurs with mammalian elements, comprising only 19.6%
of the 2.8 mm assemblage, but a much higher 62.1% in
the 6.4 mm assemblage. Birds were about equally, but
poorly, represented in both assemblages.
Diversity and abundance also differ between
screen-mesh sizes, particularly with fish taxa (Figs. 3,
4). Thirteen different genera/families of fish were
recovered with the 2.8 mm screen, while the assemblage
recovered from the 6.4 mm screen contains only 10
genera/families. In addition, the abundance differs
markedly between assemblages, with herring dominating
the 2.8 mm fauna at 72.4%, but comprising a mere 5.3%
of the total in the 6.4 mm mesh assemblage (Fig. 3). A
similar reversal occurs with salmon, which comprise
53.6% of the larger 6.4 mm mesh assemblage, but only
16.1% of the smaller 2.8 mm fauna. Among the less
numerous fish taxa (Fig. 4), dogfish, lingcod, and flatfish
are most common in the 2.8 mm screen, while dogfish,
surfperch, and sculpin dominate the 6.4 mm fauna.
An analysis of the estimated total lengths of the fish
recovered in each screen size shows more dissimilarities
(Fig. 5), with a clear peak for smaller-sized fish
recovered in the 2.8 mm screen, and a dominance of
larger-sized fish in the 6.4 mm assemblage.
Mammalian taxa also show dramatic differences
between the two screen mesh sizes (Fig. 6), with dogs,
rodents (including beaver), and deer predominating in
the 6.4 mm assemblage, but rodents, sea otters, and dogs/
deer predominating in the 2.8 mm assemblage. There
are three taxa that are only represented in one or the
other size group.
Comparison of the fish fauna of Kosapsom, sorted
by screen size (6.4 mm and 2.8 mm meshes), with the
fish fauna of Fort Rodd Hill, screened through 5 mm
mesh, and Esquimalt Lagoon, screened through 6.4 mm
mesh (Fig. 7), shows that salmon predominates in the
sites utilizing the larger screen meshes. In contrast,
herring overwhelmingly predominates in the Kosapsom
levels screened with 2.8 mm mesh.
Trends are not as distinct when examining fish taxa

Table I. Numbers and percentages of faunal elements recovered
from Gulf of Georgia levels at Kosapsom.

no. percent

Squalus acanthias (dogfish)
Raja sp. (ray)
Clupea harengus (herring)
Engraulis mordax (anchovy)
Oncorhynchis sp. (salmon)
Gadus macrocephalus (Pacific cod)
Gasterosteus aculeatus
(3-spine stickleback)
Embiotocidae surfperchh)
Sebastes sp. (rockfish)
Ophiodon elongatus (lingcod)
Hexagrammidae greenlingg)
Cottidae (sculpin)
Hippoglossus stenolepis
(Pacific halibut)
Pleuronectiformes (flatfish)

Rodentia (rodents)
Canidae (dogs or wolves: mainly dogs)
Enhydra hltris (sea otter)
Mustelids (otters, minks etc.)
Odocoileus henmionus (deer)
Cervidae (deer, elk, etc.)

Anatidae (ducks, geese. etc.)
Laridae (gulls)

3 42.9
4 57.1

other than salmon and herring (Fig. 8), but, in general,
the sites using the 6.4 mm mesh screen had poor or no
representation of the smaller fishes, including surfperch,
sculpins, sticklebacks, and anchovies. When we
compared mammals between the three sites (not
graphed), we found no discernable pattern in the elements
recovered. The smallest elements, those of rodents, were
equally represented in all sites, regardless of screen mesh
used. Birds were not compared among sites, as few
bird elements have been recovered at Kosapsom, and
birds were not identified beyond class in the Esquimalt
Lagoon assemblage.

The faunal data retrieved from the 6.4 mm screens
and from the 2.8 mm screens present very different
pictures of subsistence and economy at Kosapsom. If
only the results from the 6.4 mm screen are used,
mammals are twice as common in the site refuse as

H 6.4 mm/1/4"


0 20 40 60 80
% of all fish elements
Figure 2. Percentages of each class, of all elements, by screen
size at Kosapsom.









Stickleback __ _



ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

1 6.4 mm/l/4"



0 20 40 60 80
% of all fish elements
Figure 3. Percentages of salmon and herring elements, of all
fish elements, by screen size at Kosapsom.

I 2
S6.4 mm/1/4"
S 2.8mm/1/8"



0 5

Figure 4. Percentages of

10 15 20 25
% of all fish elements
"other" fish elements by screen size at Kosapsom.

STEWART and WIGEN: Screen Size and the Need for Reinterpretation

fish, and, among fish, salmon predominate, with almost
60% of the remains, followed by dogfish, lingcod, and
flatfish. The mammals represented were primarily dogs
or wolves, but size suggests mainly dogs, which
ethnohistoric accounts suggest were kept as pets, not
usually for food (e.g., Drucker 1963). Next most important
at the site were rodents, which, other than beavers, may
have been site intruders, as the Coast Salish did not
consume smaller rodents (Suttles 1974), followed in
abundance by deer. Deer are very common today on
Vancouver Island, and were likely common in Kosapsom
times as well. They were hunted for food and their hides
for clothing, and bone and antler were used for tools and
personal ornaments (e.g., Stewart 1987). Salmon, the
dominant fish represented in the 6.4 mm screenings, can
achieve a total length of 147 cm (Hart 1980) and were
probably caught on spawning runs up rivers and streams.
Dogfish, caught for food and for their skin, used for
woodworking, can grow to about 130 cm (Hart 1980).
Lingcod, a large greenling (up to 152 cm TL), is a valued
food fish caught in shallow bottom waters. Flatfish, a
diversified group, are well known food fish, ranging from
50 to 267 cm, depending on species (Hart 1980).
If only elements retrieved from the 2.8 mm screen
are examined, a completely different picture emerges
of Kosapsom subsistence and economy from that from
the 6.4 mm assemblage: fish elements outnumber
mammalian elements by almost 4 to 1; among fish, herring
elements outnumber salmon, again almost 4 to 1. Among
mammals, rodents are most numerous, again possibly
intruders into the site. Sea otter elements, the next most
abundant, surprisingly are completely absent from the
6.4 mm screen elements. Whether a rare inclusion, or
only represented by small elements (teeth), the sea otter
was valued for food, fur, and teeth for ornaments (Stewart
1987). Dog and deer are next most abundant, but are
not well represented in the 2.8 mm fauna.
Among the fish in the 2.8 mm assemblage, herring,
not salmon, is by far best represented, comprising 70.1 %
of all fish, mammal, and bird elements identifiable to order
or lower. Herring generally grow to a maximum length
of 25 cm in British Columbia (Hart 1980) and have been
much prized in historic times for food and oil (Hart 1980).
Northwest coast cultures so revered herring, along with
salmon and eulachon, that they performed annual rituals
to ensure good catches (Drucker 1963). The Gorge
waterway, where Kosapsom is located, has historically
been a herring fishery, and is fished today for herring.
Herring spawn in late winter, most heavily in March (Hart

% of all fish elements


10-30 30-50 50-100 >100
Estimated length of fish

-- 6.4mm (1.2") 2.8mm (1.8")

Figure 5. Estimated length of fish by screen size at Kosapsom.

1980). After herring and salmon, dogfish is best
represented in the 2.8 mm screening, followed by
surfperch, sticklebacks, and sculpins, the majority being
smaller fish of <30 cm total length.
Based on different screen sizes, in the 6.4 mm
assemblage, Kosapsom inhabitants were mammal hunters
and salmon eaters, while in the 2.8 mm assemblage,
inhabitants consumed mainly herring, with smaller amounts
of salmon, rarely consuming mammals. We asked, why
such discrepancies? One scenario might be that, because
the remains derive from different squares, but the same
cultural levels, the creators of the 6.4 mm faunal
assemblage were indeed primarily mammal and salmon
eaters, while the 2.8 mm fauna reflected different
preferences. This seems highly unlikely, however, because
the squares were all located in the same area, often
contiguous with each other, and such consistently different
proportions are unlikely given the contemporaneity and
proximity of the squares to each other.
A far more reasonable explanation is the obvious
one: that the larger, 6.4 mm screen allowed most of the
smaller fish bones, particularly the herring elements, to
drop through. The smaller, 2.8 mm mesh screen retrieved
a more representative proportion of the fauna in the
levels, particularly herring.

When the Kosapsom fauna are compared with the
Esquimalt Lagoon and Fort Rodd Hill site faunas, all

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing








Sea Otter

Dog or Wolf




6.4 mm/1/4"
2.8 mm//8"

0 i

0 5 10 15 20 25 3
% of elements

Figure 6. Percentages of mammalian elements by screen size at Kosapsom.

I I Kosap 6.4 mm
S Esqui 6.4 mm
Ft. Rodd 5 mm
Kosap 2.8 mm




0 20 40 60 8(
% of all fish elements

Figure 7. Percentage of salmon, herring, and other fish ele-
ments at Kosapsom, Equimalt Lagoon, and Fort Rodd Hill,
with screen size.

located less than 5 km from each other, clear differences
exist in the fish exploited. The Esquimalt Lagoon and
Fort Rodd Hill inhabitants procured salmon and herring
in similar proportions to that of the 6.4 mm fish fauna at
Kosapsom, while the smaller-mesh fauna from
Kosapsom show salmon and herring in reversed
proportions, perhaps the result of cultural differences
between the sites, but given the sites' proximity and the
similarity of results in sizes of screen meshes, it seems
that mesh size is the important variable. Further, the
coastal area, including Esquimalt Lagoon, is a well known
spawning area for herring in both historic and modern
times. It would be highly unusual if herring were not
also exploited in prehistoric times, given its value among
the historic Coast Salish (e.g., Suttles 1974) and its ease of
procurement when spawning (e.g., Drucker 1963). The
low proportions of herring at the Fort Rodd Hill site and,
especially, the Esquimalt Lagoon site seem best explained
by the use of the 6.4 and 5 mm meshes.

The primary purpose of archaeological excavation,
recovery, and analysis is to reconstruct past lifeways. A



I I :

STEWART and WIGEN: Screen Size and the Need for Reinterpretation

long-standing view of the prehistoric Northwest coast
cultures has been their overwhelming reliance on salmon,
a view derived from historic ethnographies and
archaeological reports. The often large size and
distinctiveness of salmon bones, in particular vertebrae,
have enhanced their preservation in archaeological sites,
while remains of smaller or less robust fish either have
not been preserved or are less visible. When only larger-
mesh screens are used, it is predictable that salmon
remains dominate to the near-exclusion of all but the
largest of other fish species.
While salmon was certainly a valued food fish and,
indeed, a staple among Northwest coast cultural groups,
reliance on it may not have been as complete as was
once thought, particularly in areas where other resources
were easy or easier to procure. In a point relevant to
this paper, Hart (1980:111) notes that pink salmon
(Oncorhyncus gorbuscha), the most abundant salmon
in British Columbia, spawn in a "very great number of
coastal streams and in all the major rivers with the
exception of those along the southeast part of Vancouver
Island," where Kosapsom, Esquimalt Lagoon, and Fort
Rodd Hill are located. Other less abundant salmonid
species spawn up the coastal rivers of southeastern
Vancouver Island, so alternate resources were probably
Because most archaeologists face time and labor
constraints, this study does not advocate that screens
with mesh <2.8 mm be used for all archaeological
matrices. We recommend, however, that more than a
column sample must be screened through small meshes.
For recognition of small fish, we suggest that 50% of
the matrix be screened through 6.8 mm and 50% through
<2.8 mm mesh.

The Kosapsom data highlight the differences in data from
faunal assemblages depending on screen size used in
excavations. We believe that the importance of small
fish may be highly underestimated in archaeological
reports from the Northwest coast because of faunal
recovery strategies, both in past excavations where matrix
was not screened, thus small elements were lost, or in
more recent excavations where 6.4 mm screens have
been used and, again, small elements have fallen through.
Screening matrices through <2.8 mm mesh will certainly
increase recovery and awareness of small faunal
elements. Use of still smaller meshes, e.g., 1.4 mm
(1/16") screens is even more preferable (Stewart et al.

I I Kosap 6.4 mm
I | Esqui 6.4 mm
Ft. Rodd 5 mm
Kosap 2.8 mm

0 10 20 30
% of all fish elements

40 50

Figure 8. Percentage of "other" fish elements at Kosapsom,
Equimalt Lagoon, and Fort Rodd Hill, with screen size.

in press), because considerable numbers of herring
vertebrae fall through the 2.8 mm screens. Although
increased time and labor is involved, the data support a
more accurate reconstruction of coastal subsistence and
The importance of herring to Northwest coast
peoples has been discussed in this paper, but, according
to both historic and ethnographic records, other small
fish, including eulachon, anchovy, smelt, and pilchard,
were highly valued. Being similar in size to herring, their
elements are only recovered in small-mesh screens.
Because several of these taxa are localized or only
seasonally available, certain cultural groups on the coast
had access to them, while others did not. Some of these
fish were valuable, for example, eulachon for its oil.
Those groups with access to such fish could profit
handsomely from trade. Use of fine-mesh screens and
recovery of small elements could contribute an entirely
new set of data to be used for interpretation of Northwest

coast subsistence and economy, and it is essential that
archaeologists make the effort to recover small elements.

I [KS] first met Elizabeth Wing at the ICAZ Fish Working
Group meeting in York, England, in 1988. She was and
continues to be an inspiration to my work in
zooarchaeology and to everyone else who has met her
or read her work.
We are indebted to the editors of this volume for the
invitation to contribute and for all their work in putting it
together. KS is also grateful for a Canadian Museum of
Nature research grant and to the University of Victoria
for use of their osteological collection.

Butler, V. L. 1993. Natural versus cultural salmonid remains: Origin
of The Dalles roadcut bones, Columbia River, Oregon, U.S.A.
Journal of Archaeological Science 20: 1-24.
Cannon, M.D. 1999. A mathematical model of the effects of screen
size on zooarchaeological relative abundance measures.
Journal of Archaeological Science 26(2): 205-215.
Clason, A. T., and W. Prummel. 1977. Collecting, sieving and
archaeozoological research. Journal of Archaeological
Science: 171-175.
Drucker, P. 1963. Indians of the Northwest Coast. New York:
The Natural History Press.
Gordon, E. A. 1993. Screen size and differential faunal recovery: A
Hawaiian example. Journal ofFieldArchaeology 20:453-460.
Grayson, D. K. 1984. Quantitative Zooarchaeology: Topics in the
Analysis of Archaeological Faunas. Orlando: Academic Press.
Hanson, D. K. 1991. Late prehistoric subsistence in the Strait

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

of Georgia region of the Northwest Coast. Unpublished
Ph.D. Dissertation, Simon Fraser University, Vancouver.
Hart, J. L. 1980. Pacific Fishes of Canada. Ottawa: Fisheries
Research Board of Canada.
James, S.R. 1997. Methodological issues concerning screen
size recovery rates and their effects on archaeofaunal
interpretation. Journal of Archaeological Science 24:
Lyman, L. 1982. Archaeofaunas and subsistence studies. Pp. 331-
393 in M. B. Schiffer, ed. Advances in Archaeological Method
and Theory, vol. 5. New York: Academic Press.
Payne, S. 1972. Partial recovery and sample bias: The results of
some sieving experiments. Pp. 46-64 in E. S. Higgs, ed.
Papers in Economic Prehistory. London: Cambridge
University Press.
Rick, T.C. and Erlandson, J.M. 2000. Early Holocene fishing
strategies on the California coast: Evidence from CA-SBA-
2057. Journal of Archaeological Science 27: 621-633.
Schaffer, B.S. 1992. Quarter-inch screening: Understanding
biases in recovery of vertebrate faunal remains. American
Antiquity 57:129-136.
Schaffer, B.S. and Sanchez, J. L. J. 1994. Comparison of 1/8-
inch and 1/4-inch mesh recovery of controlled samples of
small- to medium-sized mammals. American Antiquity 59:
Stewart, H. 1987. The Adventures and Sufferings of John R.
Jewitt. Vancouver: Douglas and McIntyre.
Stewart, K.M., G Coupland, and D. Naughton. (in press). Effects
of screen size on fish element recovery in northern
Northwest coast middens. In K. Stewart, ed. Transitions
in Zooarchaeology: New Methods and New Results.
Ottawa: Canadian Museum of Nature.
Suttles, W. 1974. The Economic Life of the Coast Salish of
Haro and Rosario Straits. New York: Garland Publishing.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 35-42


Arlene Fradkin' and H. Sorayya Carr2

The aquatic animals identified among the vertebrate faunal remains recovered in the 1990-1993 excavations at the Maya site of
Cuello, Belize, are examined. The detected patterns of aquatic resource use are comparable to those described by Elizabeth Wing
and Sylvia Scudder in their faunal analysis from previous excavations. These zooarchaeological findings, combined with paleoecological
data, suggest that the people of Cuello focused their aquatic resource procurement efforts primarily on local wetland habitats,
which may have formed part of a managed landscape surrounding their community in the Middle Preclassic period.

Key words: aquatic resources, Belize, Cuello, Maya, Middle Preclassic, zooarchaeology

The archaeological site of Cuello in northern Belize (Fig.
1) has yielded abundant information on many aspects of
ancient Maya life. Extensive excavations conducted over
the past three decades, under the direction of Norman
Hammond, have revealed a long occupational record,
spanning the earliest Middle Preclassic through the Early
Classic (ca. 1200 B.C.-A.D. 400). Cuello is especially
important for having a substantial amount of material
cultural remains dating to the Middle Preclassic, a time
period poorly represented at most Maya sites. Of
particular interest here are the faunal remains, which
can provide insight into Maya use of animal resources
during this period.
One of the insights to be gained from research at
Cuello is an understanding of how the people of this
community interacted with the landscapes and resources
of their natural world. An important contribution to this
understanding was provided by the work of Elizabeth S.
Wing and Sylvia J. Scudder, who analyzed the vertebrate
faunal remains from the 1976-1980 excavations at Cuello
(Wing and Scudder 1991). Their study documented the
use of aquatic as well as terrestrial resources at this
inland farming community. They noted that aquatic
animals were obtained from several kinds of habitats in
the region and that the catchment area was extensive.
Nevertheless, they felt that most of these animals were
primarily collected in the vicinity of the site (p. 96).
Furthermore, water turtles constituted a significant
portion of the aquatic animal remains represented in their
samples. Of particular interest was the abundance of

'Florida Atlantic University, Boca Raton, FL 33431, USA.
2924 Contra Costa Drive, El Cerrito, CA 94530, USA.

the small mud turtle (Kinosternon spp.) throughout the
Middle Preclassic (p. 85).
When Hammond et al. (1995) reopened excavations
at Cuello in 1990-1993, they focused primarily on Middle
Preclassic contexts, thus expanding the faunal database.
When we were given the opportunity to study this new
material, one of our primary objectives was to examine
the aquatic animal remains in light of the previous findings.
We were particularly interested in determining the
specific habitats from which these resources were
obtained. Additional paleoenvironmental data have
become available since the time of the initial study. In
this paper, we incorporate this information with our
zooarchaeological data to examine and explain the overall
pattern of aquatic resource use at Cuello.

The vertebrate faunal samples considered here were
recovered from two units within Platform 34, a locus at
the center of the site. Most of our material came from
the North Square, a 10 x 10 m area, which included the
remains of household structures and the yard area behind
them (Hammond 1991:15; Hammond et al. 1995:121).
Some additional material came from the South Trench,
a 20 x 3 m unit that primarily contained construction fill
(Hammond et al. 1991:354; Hammond, personal
communication 1993). Excavations were carried out by
natural stratigraphic levels. The faunal remains were
recovered by trowel excavation (Hammond, personal
communication 2001).
Analysis of the animal remains followed standard
zooarchaeological procedures (Reitz and Wing 1999:142-
238). Specimens were identified to the lowest taxon

Figure 1. Location of the Cuello site.

possible by direct comparison with reference collections
housed at the University of California Museum of
Vertebrate Zoology in Berkeley, the California Academy
of Sciences in San Francisco, and the Florida Museum
of Natural History in Gainesville.
Quantification of the faunal materials included a
count of the total number of fragments identified for
each taxon (NISP) and calculated estimates of the
minimum number of individual animals represented
(MNI). The MNI figures were determined separately
for each cultural context and were based on the
concept of paired elements and individual size. Bone
weights, which can provide information on the dietary
contribution of each animal, were not determined. Our
concern was the nature and location of aquatic
resource procurement. Therefore, the materials
examined here represent a subsample of the total
faunal assemblage.

A total of 6,967 bone and tooth fragments representing
85 vertebrate taxa have been identified in the faunal
assemblage. Of these totals, 2,458 fragments (35%) and
36 taxa (42%) represent aquatic animal remains. These
taxa are listed in Table 1 and, with their most common

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

habitats, in Table 2. Quantification of this subsample is
presented in Table 3.
Aquatic or semiaquatic animals identified include
turtles, fish, amphibians, several kinds of birds, and one
crocodile. Of the turtle remains, the most abundantly
represented is the small mud turtle (Kinosternon spp.),
including the red-cheeked mud turtle (Kinosternon
scorpioides). Other kinosternids present, though far
fewer in number, are the narrow-bridged musk turtle
(Claudius angustatus) and the northern giant musk turtle
(Staurotypus triporcatus). Two emydid turtles
(Emydidae), the common slider (Trachemys scripta) and
the furrowed wood turtle (Rhinoclemmys areolata), are
also well represented. A few large turtle shell fragments
not clearly identifiable to the family level may possibly
represent the Central American river turtle (Dermatemys
mawii), which was identified as a minor constituent in
two of the samples analyzed by Wing and Scudder
Fish constitute a small proportion of the total faunal
assemblage. Freshwater species predominate and
include cichlid (Cichlasoma spp.), swamp eel
(Synbranchus marmoratus), and blue catfish (Ictalurus
furcatus). We have also included in this group bigmouth
sleeper (Gobiomorus dormitor), which usually is found
in brackish waters, but in Belize occurs more often in
fresh water (Greenfield and Thomerson 1997:230). The
remaining kinds of fish, represented by only a few
specimens each, are estuarine and/or marine forms and
consist of tarpon (Megalops atlanticus), bonefish
(Albula vulpes), sea bass (Serranidae), jack
(Carangidae), snapper (Lutjanidae), mojarra (Eugerres
sp.), and possibly grunt (Haemulidae) (see Table 1).
Amphibians identified in the assemblage include
Mexican burrowing toad (Rhinophrynus dorsalis),
giant toad (Bufo marinus), other Bufo specimens not
identified to the species level, true frogs (Rana spp.),
and one possible treefrog (Hyla sp.). Amphibians are
potentially important as environmental indicators, although
their cultural significance is less certain. Their presence
in archaeological deposits may be incidental.
Several kinds of aquatic birds are represented.
These include pied-billed grebe (Podilymbus
podiceps), anhinga (Anhinga anhinga), green-
backed heron (Butorides virescens), great blue heron
(Ardea herodias), and possibly a sandpiper
(Scolopacidae) and a rail/gallinule/coot (Rallidae).
Crocodile (Crocodylus sp.) is represented by one
tooth. Although two species naturally occur in the region,

FRADKIN and CARR: Middle Preclassic Landscapes and Aquatic Resource Use

we can only identify this tooth taxonomically to the genus
Overall, our sample of aquatic animal remains is
comparable to that examined by Wing and Scudder for
the Middle Preclassic period. The turtle remains are
similar, although there is some variation in the relative
proportions of different taxa and Dermatemys has not
been positively identified in our collection. A consistent
feature is the predominance of mud turtles. Fish remains,
though admittedly a small proportion of both assemblages,
demonstrate a preponderance of freshwater specimens
with minor representation of marine forms. Birds are
minimally represented and, in our sample, only aquatic
species have been identified.

The new faunal data from Cuello lend further support to
the pattern of aquatic resource use described by Wing
and Scudder for the Middle Preclassic. Moreover,
additional paleoenvironmental information has since
become available, providing insight into the nature of the
landscape surrounding Cuello at that time and potentially
explaining the patterns observed in the zooarchaeological
The site of Cuello is situated between two rivers,
the New River and Rio Hondo (see Fig. 1), which are,
respectively, 5 and 10 kilometers away. Nevertheless,
there is evidence that other kinds of aquatic habitats
may have been in closer proximity to the site.
A study of plant and molluscan remains recovered
at Cuello by Charles Miksicek (1991) demonstrates the
development of marsh, pond, or shallow lake habitats
around the site during the course of the Middle
Preclassic. In his samples of freshwater snails, he found
large numbers of certain lake-dwelling forms, such as
Pyrgophorus, whereas snails that prefer flowing water,
such as Pachychilus (Feldman 1978:7), were rare
(Miksicek 1991:74, Table 4.4). Furthermore, he noted
an abundance of razor grass (Cladium) seeds, also
indicative of a marsh habitat (Miksicek 1991:77,83). The
composition of our faunal sample seems to indicate that
the Cuello residents focused primarily on these more
immediate wetland areas.
Turtles potentially provide the best information on
aquatic habitats exploited because of their large numbers
in our collection. Preeminent among the Cuello turtles
are small mud turtles (Kinosternon spp.). These turtles
typically live in quiet, mud-bottomed shallow water, such

Table 1. Taxonomic list of aquatic animals identified with
scientific and common names.

cf. Megalops atlanticus
Albula vulpes
Ictalurus furcatus
Synbranchus inarmoratus
cf. Serranidae
cf. Eugerres sp.
cf. Haemulidae
cf. Cichlasoma spp.
Gobiomorus dormitory
Rhinophrynus dorsalis
cf. Bufo marinus
Bufo spp.
cf. Hyla sp.
Rana spp.
Claudius angustatus
Staurotypus triporcatus
Kinosternon scorpioides
Kinosternon spp.
Rhinoclenmmvs areolata
Trachemvs script
Crocodvlus sp.
Podilvmbus podiceps
Anhinga anhinga
cf. Ardea herodias
Butorides virescens
cf. Rallidae
cf. Scolopacidae

blue catfish
swamp eel
sea bass
bigmouth sleeper
Mexican burrowing toad
giant toad
tree frog
true frog
narrow-bridged musk turtle
northern giant musk turtle
red-cheeked mud turtle
mud turtle
musk/mud turtle
furrowed wood turtle
common slider
pond turtle
pied-billed grebe
great blue heron
green-backed heron

' In the study region, the common name mojarraa" is applied to both
gerreids and cichlids.

as that found in ponds, lakes, sluggish streams, and slow-
moving backwaters of rivers. They are bottom-walkers
rather than swimmers and, unlike some other turtles, do

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 2. Habitats of aquatic animals identified.

Inland Freshwater Habitats
Taxon Lakes/Ponds Rivers/Streams Wetlands

Coastal Habitats

cf. Megalops atlanticus x
Albula vulpes x
Ictalurus furcatus x
Synbranchus marmoratus x x
cf. Serranidae x
Carangidae x
Lutjanidae x
cf. Eugerres sp. x
Gerreidae x
cf. Haemulidae x
cf. Cichlasoma spp. x x x
Cichlidae x x x
Gobiomorus dormitor x x
Rhinophrynus dorsalis x x
cf. Bufo marinus x
Bufo spp. x
cf. Hyla sp. x x
Rana spp. x x
Claudius angustatus x x
Staurotypus triporcatus x x x
Kinosternon scorpioides x x
Kinosternon spp. x x x
Rhinoclemmys areolata x
Trachemys scripta x x
Crocodylus sp. x x x x
Podilymbus podiceps x x x
Anhinga anhinga x x x
cf. Ardea herodias x x x x
Butorides virescens x x x
cf. Rallidae x x x x
cf. Scolopacidae x x x x

not bask at the surface (Campbell 1998:104). Although
they are not restricted to shallow waters, their bottom-
dwelling habits would make them more difficult to
collect in deep water. Consequently, large numbers of
mud turtles argue strongly for shallow-water
procurement. The related Claudius is similar in its
behavior and habitat preferences, tending toward even
smaller and shallower water bodies (Campbell 1998:105),
while Staurotypus is found primarily in lakes and slow-
moving parts of rivers but also enters marshes and

flooded grasslands to forage, preying upon mud turtles,
as well as other animals and some plants (Campbell
Among the emydids, the semiaquatic Rhinoclemmys
is most common in savannas and at edges of clearings
but also occurs in marshes (Campbell 1998:116; Lee
1996:165). On the other hand, the more aquatic
Trachemys is found in larger freshwater bodies; in the
tropics it is more riverine than elsewhere in its range
(Ernst and Barbour 1989:205).

FRADKIN and CARR: Middle Preclassic Landscapes and Aquatic Resource Use

Table 3. Quantification of aquatic animal remains.

Taxon NISP % MNI %
cf. Megalops atlanticus 1 0.04 1 0.27
Albula vulpes 1 0.04 1 0.27
Ictalurus furcatus 1 0.04 1 0.27
Siluriformes 1 0.04 1 0.27
Synbranchus marmoratus 30 1.22 19 5.05
cf. Serranidae 1 0.04 1 0.27
Carangidae 3 0.12 3 0.80
Lutjanidae 2 0.08 2 0.53
Lutjanidae/Haemulidae 2 0.08 1 0.27
Lutjanidae/Gerreidae 1 0.04 -
cf. Eugerres sp. 1 0.04 1 0.27
Gerreidae 3 0.12 2 0.53
Gerreidae/Haemulidae 1 0.04 1 0.27
Gerreidae/Cichlidae 1 0.04 -
cf. Cichlasoma spp. 3 0.12 3 0.80
Cichlidae 23 0.94 17 4.52
Gobiomorus dormitor 2 0.08 2 0.53
Osteichthyes 339 13.79 18 4.79
TOTAL FISH 416 16.92 74 19.68

Rhinophrynus dorsalis 22 0.90 15 3.99
cf. Bufo marinus 2 0.08 2 0.53
Bufo spp. 25 1.02 9 2.39
cf. Hyla sp. 1 0.04 1 0.27
Rana spp. 5 0.20 5 1.33
Anura 35 1.42 15 3.99
Amphibia 6 0.24 1 0.27
TOTAL AMPHIBIANS 96 3.91 48 12.77

Claudius angustatus 8 0.33 4 1.06
Staurotypus triporcatus 47 1.91 15 3.99
Kinosternon scorpioides 84 3.42 33 8.78
Kinosternon spp. 449 18.27 86 22.87
Kinosternidae 22 0.90 3 0.80
Rhinoclemmys areolata 74 3.01 33 8.78
Trachemys scripta 61 2.48 24 6.38
Emydidae 52 2.12 13 3.46
Testudines 1140 46.38 34 9.04
Crocodylus sp. 1 0.04 1 0.27
TOTAL REPTILES 1938 78.84 246 65.43

Podilymbus podiceps 2 0.08 2 0.53
Anhinga anhinga 2 0.08 2 0.53
cf. Ardea herodias 1 0.04 1 0.27
Butorides virescens 1 0.04 1 0.27
cf. Rallidae/Scolopacidae 1 0.04 1 0.27
cf. Scolopacidae 1 0.04 1 0.27
TOTAL BIRDS 8 0.33 8 2.13

2458 100.00 376



Dermatemys is conspicuously absent from our
sample and was minimally represented in the assemblage
analyzed by Wing and Scudder. This turtle is distinct from
others in the region in that it is so fully aquatic that it is
barely able to move on land, being primarily a denizen of
large rivers and lakes, spending long periods submerged
(Campbell 1998:111-112).
When compared to contemporary ethnographic
accounts, the relative abundance of the various turtles
in the Cuello zooarchaeological record is, interestingly,
almost a mirror image of their relative food value today.
The larger, meatier riverine turtles occur in fewer
numbers, whereas the very small turtles, typically found
in shallow lakes and ponds, are far more common.
The rarest turtle, Dermatemys, the largest freshwater
turtle in the region, has meat that is highly esteemed
today. A single individual provides a substantial amount
of meat. Almost as large, Trachemys is also highly sought
after and consumed. Staurotypus, slightly smaller, is
eaten, despite its musky flavor. The smaller turtles,
Rhinoclemmys and Claudius, contain much less meat
but are still eaten today. Kinosternon, though eaten by
some people, is not as widely consumed because of the
overpowering odor of musk it produces during cooking
(Campbell 1998:104-117).
Interviews conducted by Carr (1986:191-194) with
Maya archaeological workers in northern Belize
substantiate ethnographic accounts. Except for
Kinosternon, all kinds of turtles are consumed. The
most plausible explanation for the abundance of these
very small mud turtles and for the lesser
representation of larger, meatier turtles in the faunal
remains is that Cuello people obtained aquatic
resources primarily from seasonally fluctuating local
marsh and pond environments rather than from the
more distant riverine habitats.
The kinds of freshwater fish identified also suggest
that fishing focused on nearby marsh wetlands. The
swamp eel, as its common name implies, is present in
stagnant water habitats. It is well suited to seasonal
wetlands and withstands temporary dry spells by
burrowing into mud (Burgess and Franz 1989:265;
Greenfield and Thomerson 1997:138-139). Belizean
cichlid species are found in a variety of waters, ranging
from rivers to warm, murky wetlands and ponds
(Greenfield and Thomerson 1997:184-206; Konings
1989). Blue catfish, a large, meaty fish typically occurring
in large rivers in Belize (Greenfield and Thomerson
1997:78), is represented in our sample by a single

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

specimen, further indicating infrequent use of this habitat.
One other habitat occasionally exploited was quite distant
from the site. The several estuarine or marine fish most
likely were brought in from the coast, approximately
35 km away.
The presence of amphibians in fairly significant
numbers also suggests the proximity of moist habitats.
The two identified toad genera, Rhinophrynus and Bufo,
frequent open habitats and breed in temporary ponds
(Lee 1996:56, 77-80). Hylids occupy a variety of habitats
and most also breed in temporary bodies of water (Lee
1996:92-97). The more aquatic Rana frogs are most
often found near lakes, ponds, and sluggish portions of
streams (Campbell 1998:94-98; Lee 1996:123-127).
The aquatic birds in the Cuello assemblage have
broad habitat tolerances. In Belize, grebe, anhinga, blue
heron, and green-backed heron are found in or alongside
ponds and streams (Russell 1964:35-38). Similarly,
crocodile can occur in a variety of waters, including inland
lakes, rivers, and marsh wetlands as well as brackish
coastal waters (Campbell 1998:283-289).
The procurement of animal resources from local
wetlands could have been integrated into the overall
subsistence economy in several ways. One scenario is
suggested by the savanna fishing practiced today in
northern Belize. This takes place late in the dry season
when shrinking pools concentrate the fish and the water
is shallow enough to wade through (Pascual Mesh,
Johnny Montalvo, and Valeriano Tun, personal
communication 1981). As a dry-season activity, savanna
fishing complements the rainy season focus on swidden
cultivation and garden hunting (Johnny Montalvo,
personal communication 1981). Some support for dry-
season use of wetlands is indicated by the presence of
grebes and blue herons, birds that are winter (dry season)
visitors to Belize (Russell 1964:35,38).
On the other hand, mud turtles and amphibians are
most visible in the rainy season, virtually disappearing in
the dry season, when they may burrow into mud to
aestivate (Campbell 1998:46,104; Lee 1996:56,163). The
successful capture of large numbers of mud turtles thus
could argue for rainy season procurement. Alternatively,
the harvesting of mud turtles could have been facilitated
by keeping a controlled population in a limited area. These
turtles are nocturnal and can be difficult to find by day
even where they are plentiful (Campbell 1998:109).
Penning off flooded lands as the waters began to recede
would be a convenient way to keep turtles within a known
area, where they could be collected by digging in the

FRADKIN and CARR: Middle Preclassic Landscapes and Aquatic Resource Use

mud. Both Hammond (1991:239) and Scudder (personal
communication 2001) have speculated that such
management techniques could have been practiced at
Wetland margins may also have been a locus of
farming activity. To date, there is no evidence at Cuello
of raised- or drained-field farming (Hammond 1991:239).
Nevertheless, as suggested by Pohl et al. (1996:366), a
form of dry-season cultivation could have been practiced
along wetland margins without creating formal field and
canal features. Thus, crops and aquatic animals would
have been in close association so that farming and
collecting activities could have taken place
simultaneously. Indeed, the entire complex of fields and
wetlands may have been a managed landscape.
Evidence of wetland resource procurement in the
Middle Preclassic is not restricted to Cuello. At Colha,
another inland site 27 km to the southeast, Leslie Shaw
(1999) shows considerable use of aquatic animals in the
Middle Preclassic. Like the freshwater fishes at Cuello,
most fish remains were not identifiable, but they were
generally of small size (Shaw 1999:88, 91). Of the other
aquatic and semiaquatic animals that were more precisely
identified and assigned to a habitat category, wetland
animals greatly exceeded riverine animals in the Colha
samples (Shaw 1999: Fig. 4.3).

Our study of the aquatic animal remains recovered from
recent excavations at Cuello corroborates the findings
of Wing and Scudder (1991) concerning patterns of
resource use in the Middle Preclassic. The most
frequently collected animals were water turtles, in
particular mud turtles. Although some of the aquatic
resources were obtained from habitats at varying
distances from Cuello, the majority were collected in close
proximity to the site. According to paleoenvironmental
information, marsh, pond, or shallow lake habitats existed
nearby during the Middle Preclassic. The general pattern
of aquatic resource use documented at Cuello shows
similarities to that of Colha during the same period.
As part of Maya economic practices, wetland
foraging was coordinated with farming activities. The
collection of certain small animals, such as mud turtles,
could have been facilitated by the construction of
pens. We believe that this overall pattern was a
common aspect of Maya subsistence economies
wherever appropriate wetlands were part of the
surrounding landscape.

Our thanks go to Norman Hammond for the opportunity
to study this material and for providing partial funding
for this study; to Elizabeth Wing and Sylvia Scudder for
allowing us to use comparative materials at the Florida
Museum of Natural History and to incorporate their
earlier findings into this paper; to the California Academy
of Sciences and the University of California Museum of
Vertebrate Zoology for providing access to their
comparative collections; to John Iverson of Earlham
College for assistance in the identification of certain
kinosternid elements; to Sylvia Scudder for reading the
draft manuscript and providing valuable suggestions; to
Heather Norby for drawing the map; and to the people of
Chunox, Belize, for information on fishing and turtle
Most importantly, we wish to express our deepest
gratitude and appreciation to Elizabeth Wing for teaching
us the intricacies of zooarchaeology and for mentoring
us through the years. Her enthusiasm for the discipline
has been a true inspiration.

Burgess, G. H., and R. Franz. 1989. Zoogeography of the
Antillean freshwater fish fauna. Pp. 263-304 in C. A.
Woods, ed. Biogeography of the West Indies: Past,
Present, and Future. Gainesville: Sandhill Crane Press.
Campbell, J. A. 1998. Amphibians and Reptiles of Northern
Guatemala, the Yucatan, and Belize. Norman: University
of Oklahoma Press.
Carr, H.S. 1986. Faunal Utilization in a Late Preclassic Maya
Community at Cerros, Belize. Ph.D. dissertation. Ann
Arbor, Michigan: University Microfilms International.
Ernst, C. H., and R. W. Barbour. 1989. Turtles of the World.
Washington, D.C.: Smithsonian Institution Press.
Feldman, L. H. 1978. Invertebrados arqueologicos. Boletin de
la Escuela de Ciencias Antropologicas de la Universidad
de Yucatan 6(33): 2-23.
Greenfield, D. W., and J. E. Thomerson. 1997. Fishes of the
Continental Waters of Belize. Gainesville: University Press
of Florida.
Hammond, N., ed. 1991. Cuello: An Early Maya Community in
Belize. Cambridge: Cambridge University Press.
Hammond, N., A. Clarke, and C. Robin. 1991. Middle Preclassic
buildings and burials at Cuello, Belize: 1990 investigations.
Latin American Antiquity 2(4): 352-363.
Hammond, N., A. Clarke, and S. Donaghey. 1995. The long
goodbye: Middle Preclassic Maya archaeology at Cuello,
Belize. Latin American Antiquity 6(2): 120-128.
Konings, A. 1989. Cichlids from Central America. Neptune
City, New Jersey: T.F.H. Publications.

Lee, J. C. 1996. The Amphibians and Reptiles of the Yucatan
Peninsula. Ithaca, New York: Cornell University Press.
Miksicek, C. H. 1991. The ecology and economy of Cuello. The
natural and cultural landscape of Preclassic Cuello. Pp. 70-
84 in N. Hammond, ed. Cuello: An Early Maya Community in
Belize. Cambridge: Cambridge University Press.
Pohl, M. D., K. O. Pope, J. G. Jones, J. S. Jacob, D. R.
Piperno, S. D. deFrance, D. L. Lentz, J. A. Gifford, M. E.
Danforth, and J. K. Josserand. 1996. Early agriculture
in the Maya lowlands. Latin American Antiquity 7(4):
Reitz, E. J., and E. S. Wing. 1999. Zooarchaeology. Cambridge
Manuals in Archaeology. Cambridge: Cambridge

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

University Press.
Russell, S. M. 1964. A Distributional Study of the Birds of
British Honduras. Ornithological Monographs, no.1.
American Ornithologists' Union. Lawrence, Kansas: Allen
Shaw, L. C. 1999. Social and ecological aspects of Preclassic
Maya meat consumption at Colha, Belize. Pp. 83-100 in C.
D. White, ed. Reconstructing Ancient Maya Diet. Salt Lake
City: University of Utah Press.
Wing, E. S., and S. J. Scudder. 1991. The ecology and economy
of Cuello. The exploitation of animals. Pp. 84-97 in N.
Hammond, ed. Cuello: An Early Maya Community in Belize.
Cambridge: Cambridge University Press.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 43-54



Arturo Morales-Mufiiz' and Yekaterina Antipina2

This paper presents a preliminary overview of the bird remains from the early/middle Bronze Age site of Velikent, a series of
mounds situated on the Caspian coastal plain of the Russian republic of Daghestan. A total of 25 taxa, including 21 species, have
been identified thus far. They represent species from both aquatic and terrestrial biotopes, although one species, the Great
bustard, Otis tarda, constitutes the dominant element of all subsamples. Whether this was actually so or not and whether most
of the secondary patterns reported below are trustworthy is an open question partly due to the small samples thus far available
for study and partly to a manual retrieval of remains that will need to be improved in the future if patterns are to be coupled with
those available for comparison from other faunal sets, domestic mammals in particular.

Key words: birds, Bronze Age, Caspian Sea, Daghestan, hunting, Velikent

The archaeological site of Velikent is situated on the
southern edge of the contemporary village of Velikent
which is ca. 25 km northwest of Derbent and ca. 12 km
west of the present-day shore line of the Caspian Sea,
approximately in the middle of the sea's littoral plain
(Fig.l). The archaeological remains are located in five
separate natural mounds (I-V) ca. 5-7 meters high. These
natural clay mounds constitute part of an ancient terrace
formed by an earlier transgression of the Caspian.
A.A. Rusov first recognized the archaeological
significance of this site in the late nineteenth century.
Subsequently, under the direction of M.G. Gadzhiev,
the Institute of History, Archaeology, and Ethnography
of the Daghestan Scientific Center, USSR Academy
of Sciences, conducted excavations at Velikent from
1977 to 1979 and from 1982 to 1984. The Daghestan-
American Velikent Expedition (DAV), established in
1993, conducted a preliminary field season in 1994 and
two later field seasons in 1995 and 1997. From 1998
through 2000, yearly seasonal digs have been undertaken
by the Daghestanis (Gadzhiev et al. 1997, 2000).
The faunas from Velikent did not receive detailed
attention until Morales-Muiiz (in Gadzhiev et al. 1997,
2000) and Morales-Mufiiz and Antipina (2000) presented
preliminary reviews from the 1995 and 1997 field
'Laboratorio Arqueozoologia. Departamento de Biologia.
Universidad Autonoma de Madrid, E-28049 Madrid, Spain.
'Laboratory of the Natural Sciences in Archaeology. Institute
of Archaeology. Russian Academy of Sciences. Dm. Ulianova,
19. 117 036 Moscow, Russia.

seasons. The same authors have recently concluded a
detailed analysis of the mammals from the 1995-2000
campaigns from which the bird remains reported in this
paper come (Antipina and Morales-Mufiiz, in prep.).
Faunal analyses were undertaken 1) to document
whether or not the occupation at Velikent was a
continuous one throughout the chronological sequence
and 2) to define the nature of the subsistence economy
at the site, in particular domestic stocks and the relevance
of the exploitation of the Caspian Sea resources. To that
end, two mounds have been excavated. Mound II, the
older one, evidenced a chronocultural sequence that
spans from 3,300 to 2,800 B.C. The sequence of mound
I ranges from 2,700 to 1,800/1,700 B.C. (Magomedov,
pers. com.). Using a square grid, a series of excavated
trenches in both mounds uncovered a complex
stratigraphy that revealed stratified fill from domestic
areas arranged along a sequence of three building
horizons (Gadzhiev et al. 1997, 2000).

Most of the remains reported originate from two trenches,
IIC and IID, located in mound II. The sediments
consisted of stratified fill, mostly from open courtyard
areas containing various features that included hearths
and a series of pits. All remains were retrieved manually,
a method that often implies possibly important taphonomic
losses that seriously limit the inference potential of the
samples, as well as the use of abundance estimators
other than the identified number of remains (Grayson

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Figure 1. Map of Daghestan, showing the location of Velikent and the extent of the Caspian coastal plain.

1984). The skewed distribution of remains,
concentrated on the IIC '95 trench and minimal
representation of bird bones from mound I, similarly
restricts the comparative possibilities of the samples from
the various other units.
The identification of remains was carried out using
the reference collection of one of us (AMM), housed at
the Universidad Autonoma de Madrid. The western
Palearctic character of the Caspian Sea avifaunas,
together with the migratory habits of many species,
ensured that this reference collection permitted a reliable
taxonomic assignment of most remains (Dement'iev
1951; Harrison 1982). This reliability is strengthened in

the case when a particular bone morphology (i.e., a
morphotype) was only possessed by a restricted
number of species (Table 1). In some difficult cases,
such as those of certain waterfowl and corvids, use was
made of diagnostic features mentioned in the works
of Bacher (1967), Tomeck and Bochenski (2000), and
Woelfle (1967).
Estimation of the minimum number of individuals
(MNI) followed conventional procedures (e.g., Clason
1972; Grayson 1984) for the various archaeological units
(i.e., levels, pits, and so on) provided by the excavators.
Measurements will not be considered in this preliminary
overview. Recording complementary data, in particular

MORALES-MUNIZ and ANTIPINA: Birds from Bronze Age Velikent

Table 1. The Velikent bird taxa grouped according to morphotypes with an indication of osteologically similar species in the area
at present. Species codes as in Table 2.

Species in
Morphotype Velikent Osteologically Similar Species Present in the Area

16 (large)
17 (small)

Ardea purpurea (smaller), Casmerodius alba (slight osteological differences)
Ciconia nigra (smaller)
Anser fabalis. A. brachyrrhynchos (both smaller)
none of that size and morphology
Cygnus columbianus (smaller; see Fig. 2)
Tadorna ferruginea
Anas strepera, A. clypeata, A. penelope (smaller than 8# similar to 9), A. crecca
A. querquedula (smaller than 9# much smaller than 8)
Aegypius monachus (larger, slight osteological differences)
Aquila heliaca (smaller)
Aquila pomarina (?) /A. rapax (?) (missing from reference collection)
Falco tinnunculus (slight morphological differences), E vespertinus and
F naumanni (both smaller); F peregrinus and E cherrug (both larger)
none of that size and morphology
none of that size and morphology

none of those sizes and morphologies
none of that size and morphology
none of that size and morphology
Numenius phaeopus. N. tenuirostris (both smaller)
Conrus corax (larger), C. monedula (smaller), C. fugilegus (slight morphological
differences; see Fig. 3)

fractures and manipulative traces, was carried out
whenever possible. A combined use of such data together
with skeletal abundance profiles permits gross analysis
of assemblage in terms of taphonomic groups (sensu
Gautier 1987). For paleoenvironmental purposes, use was
made of the concept of the analogue as defined by Baird
(1989), with complementary biological data taken
from Boev (1993), Dement'iev (1951), Harrison
(1982), Jonsson (1992), Nikol'skii (1891/1892), and
Silant'ev (1898).

Table 2 provides a general overview of the bird
assemblage from Velikent and Table 3 provides a
distribution of the bird remains in the various trenches
by campaigns. At this gross level of analysis, one peculiar
feature is the high taxonomic diversity of the samples in
relation to their minuscule sample sizes. This ratio reflects
to no small extent an extensive taphonomic loss that
influences the comments that are to follow. Also, the
skewed distribution of remains, both taxonomically

and in terms of archaeological units, dictates that only
data from trench IIc and the Great bustard sample
(65% of the identified NISP) can be considered to
have minimal potential reliability.
Indirectly, the method of retrieval likewise might be
responsible for the generally good condition of remains
and for their rather restricted fragmentation (see below),
both contingencies aiding identification and high level of
resolution, with barely 7% of the samples remaining
unidentified. Retrieval biases in certain skewed skeletal
distributions (i.e., dominance of appendicular bones, Table
4) would require additional data in order to be
Due to the limiting factors summarized in Table 1,
the reliability of identifications is not strictly comparable
for the various taxa, with osteologically distinct species
such as coot, pheasant, and Great bustard posing no major
problem, but the various duck remains being more
questionable. During the identification process, a series
of diagnostic features were either spotted (Fig. 2) or
called into question (Fig. 3), although in most cases

46 ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 2. The Velikent bird assemblage in terms of identified number of remains (NISP) and minimum number of individuals (MNI).



Gray heron, Ardea cinerea
White stork, Ciconia ciconia
Greylag goose, Anser anser
Lesser white-fronted goose, Anser erythropus
Whooper swan, Cygnus cygnus
Mute swan, Cygnus olor
Unspecified swan, Cygnus sp.
Shelduck, Tadorna tadorna
Mallard, Anas platyrrhynchos
Shoveler, Anas clypeata
Unspecified waterfowl, Anatidae indet.
Griffon vulture, Gypsfulvus
Golden eagle, Aquila chrysietos
Spotted eagle, Aquila clanga
Hobby, Falco subbuteo
Hobby/Kestrel, Falco subbuteolE tinnunculus
Pheasant, Phasianus colchicus
Coot, Fulica atra
Great bustard, Otis tarda
Little bustard, Tetrax tetrax
Avocet, Recurvirostra avosetta
Black-winged stilt, Himantopus himantopus
Curlew, Numenius arquata
Hooded crow, Corvus corone
Hooded crow/Rook, Corvus corone/C. frugilegus
Total identified
Aves indeterminate





osteometry was needed in order to substantiate specific
assignals (Morales and Antipina, unpub. data). The case
of the Spotted eagle is special in that, despite a perfect
match with our reference specimen, the lack of
presumably similar species in our collection does not allow
one to ascertain to what extent the features recorded on
the distal tarsometatarsus are diagnostic for determining
species (Fig. 4 ).

To determine the identity of the agents involved in
the formation of the Velikent bird assemblages would
prove crucial for setting apart human behaviors from
those of other potential bone accumulators. This subject
is far from new and has given rise to an important
literature of its own (Bochenski 1997; Bochenski et
al. 1998,1999; Bramwell et al. 1987; Ericson 1987; Gautier
1987; Laroulandie 2000; Livingston 1989; Morales and
Rodriguez 1997; Mourer-Chauvir6 1983; and Serjeantson

et al. 1993). As it happens, when several agents converge
upon the same assemblage, to set them apart from one
another is far from straightforward because most
signatures are subject to no small amount of convergence
(sensu Morales and Rosello 1998).
At Velikent both the domestic nature of the deposits
and the almost total dominance of consumed domesticated
mammals point toward the human accumulation of animal
remains (Antipina and Morales, in prep.). Such an
hypothesis is reinforced by the application of the criteria
of Mourer-Chauvir6 (1983) to the samples. Thus, the
combined abundances of coracoid + humerus + femur
(59%) over the combined total represented by these three
bones, plus the radius, ulna, carpometacarpus, tibiotarsus,
and tarsometatarsus, would define Velikent as an
"anthropic" accumulation (Table 4). Preybirds as
accumulators could be ruled out on the grounds of the
low frequencies of carpometacarpus + tarsometatarsus
(14% of the previous overall total) and of the comparatively

MORALES-MUNIZ and ANTIPINA: Birds from Bronze Age Velikent

Table 3. Distribution of avian taxa, expressed as NISPs, for the different excavation units.
Taxon/Unit IC'95 IIC'97 IIC'98 IIC'00 IC IID'97 IID'98


IA'95 IA97 OC'00

Grey heron, Ardea cinerea
White stork, Ciconia ciconia
Greylag goose, Anser anser
Whitefronted goose, A.erythropus
Whooper swan, Cygnus cygnus
Mute swan, Cygnus olor
Cygnus sp.
Shelduck, Tadorna tadorna
Mallard, Anas platvrrhynchos
Shoveler, Anas clypeata
Anatidae indet.
Griffon vulture, Gypsfulvus
Golden eagle, Aquila chrysietos
Spotted eagle, Aquila clanga
Hobby, Falco subbuteo
Falco subbuteolE tinnunculus
Pheasant, Phasianus colchicus
Coot, Fulica atra
Great bustard, Otis tarda
Little bustard, Tetrax tetrax
Avocet, Recurvirrostra avosetta
Stilt, Himantopus himantopus
Curlew, Numenius arquata
Crow, Corvus corone
Corvus coronelC. frugilegus

Aves indeterminate

1 7
2 14
1 2
1 3
1 2 3
6 4 2 50

1 1

1 1

1 1

1 1

1 11 12 1 1 1 3

1 1

8 10 3 114
8 10 3 122

1 19
1 22

1 1 4 6

1 1 4 6

low frequencies of proximal humeri (24%, unpub. data)
in the humerus samples. More important, these patterns
are essentially the same as those based on the Great
bustard samples (e.g., coracoid + humerus + femur = 83%;
carpometacarpus + tarsometatarsus = 3.7%; proximal
humeri = 27%).
Other Velikent patterns appear more consistent with
remains accumulated either by preybirds or natural
deaths. In this way, the application of Ericson's index
provides a value of 68 (65 for the Great bustard), which
these authors consider indicative of "natural"
accumulations (Ericson 1987). Serjeantson et al. (1993),
on the other hand, report a high frequency of upper limb
bones in accumulations of shearwaters preyed on by
gulls and explain this in terms of the upper limb bones'
greater tendency to remain articulated for a longer period

when undisturbed (meaning absence of butchering).
Finally, Bochenski (1997) records the humerus as being
the most common bone on accumulations of Snowy
owl meal leftovers. All these data indicate that, despite
uncertainties, a high frequency of upper limb bones
does not conform, in principle, with a strictly anthropic
accumulation of bird bones
At Velikent, analysis of archaeological bird
assemblages requires consideration of many other
factors, not just skeletal profiles, in order to avoid
taphonomic convergence (e.g., Livingston 1989 working
on data from Rich 1980; see also Bochenski et al. 1998,
1999; Bramwell et al. 1987; Laroulandie 2000). When
complementary data are taken into consideration for our
assemblages, the following picture emerges:
a) Fire. Only 13 bones exhibit traces of fire in one

48 ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 4. Skeletal distributions within species. Species codes as in Table 2.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

16 17 18 19 20 21 TOTAL %



1 1 1



2 26

1 0.7
1 0.7
3 2.2
1 0.7
3 2.2
3 2.2

1 1

7 1
1 7 1 1 2

1 1

2 1 9 16 1 3 1 5 1 2 3 2 2 3 1 65 3 4 1 1 7

1 5

133 100.0

form or another. Six bones, namely one carpometacarpus
of shoveler, three humeri (one from the lesser White-
fronted goose plus two from unspecified ducks), and a
sternum and phalanx from the Greylag goose, are
charred. One unspecified splinter was calcinated and
the remaining bones in this sample (all except one
unspecified splinter from Greylag goose) were burned
in a more conventional way (i.e., no extensive surface
erosion). Overall, the dominance of waterfowl in this
sample is overwhelming (27% of this group's NISP),
especially when one considers the restricted contribution
of this group to the overall assemblage, which may point
toward a differential treatment of remains worth exploring
in the future.
b) Manipulative traces. These appear to be restricted
to a cut mark on the distal articular surface of a humerus
and a drilled proximal furcula. Both bones belong to the
Great bustard. In view of this species' abundance, large
size, and putative food value, samples of such low
frequencies are all the more remarkable and more difficult
to explain in terms of manual retrieval (see below).
c) Fracture patterns. Most, if not all, recorded

fractures at Velikent are post-depositional. Those parts
more likely to be missing are the fragile laminar or tubular
portions of bones (Fig. 5). Diagenetic factors might also
be responsible for the scarcity of other laminar/tubular
elements such as ribs, furcula, or even the skull (the
only skull bone retrieved was a pterygoid [Table 4]), but
one cannot be sure with the data at hand. The evidence
against this sample qualifying as a "natural" accumulation
involves the absence of bones below a critical threshold
of 3 cm (unpublished data), reinforcing our original
impression of a great taphonomic loss having taken place.
Under such circumstances, it would be futile to speculate
on whether the missing portions of the avian assemblage
would change the characteristics of our samples referred
to in the last three paragraphs.
d) Age data. All remains, except for a Great bustard
tibiotarsus, apparently belong to adult birds (sensu
Hargrave 1970), meaning a bird able to fly.
e) Articulated specimens. None has been recorded
at Velikent.
Put together, data from a) and b) essentially point to
human accumulation of remains despite low frequencies


1 2

11 1
1 1 1 1
3 3 1
2 1 1

MORALES-MU&IZ and ANTIPINA: Birds from Bronze Age Velikent

-J i

Figure 2. Differences in selected bones of Whooper swan (WS, on right in 1-4) and Mute swan (MS, on left in 1-4). For the furcula
(1, depicted in semilateral view), the dorsal process is long and cylindrical in WS (d) and short and flattened in MS (d'), whereas
the caudal arch is elongated and bent dorso-caudally in WS (c) but short, blunt, and directed ventro-caudally in MS (c'). The
coracoid (2, depicting the caudal view of the distal extremity) in WS exhibits a sharp uncinated process (u) and an irregular margin
for the faces articularis caudalis (a), whereas MS lacks an uncinated process and the faces articularis caudalis has a sharp and
straight margin (a'). In the tibiotarsus (3, depicting the dorsal view of the distal extremity), the lateral and medial condyles are of
similar width in WS, but in MS the lateral condyle is clearly wider than the medial condyle. In addition, MS has a comparatively
smaller muscular impression over the Retinaculum (r') than WS (r). The humerus (4, depicting the lateral/dorsal view of the distal
extremity) of WS exhibits a deep impresio musculi brachialis (b) with a sharp proximal (upper) margin, which is nowhere to be
seen in the very shallow impresio musculi brachialis of MS (b'). Finally, WS displays a very shallow (flat) muscular impression
over the medial condyle (m) that is round and bump-like in MS (m').

of burned and chopped bones. On the other hand, the
absence of anthropic fractures in c) speaks in favor of a
natural accumulation, in particular when one considers
the large size of the most frequent taxa. Conversely, d)
and e) do not support either one or the other alternative.
All this forces us to rely on context provided on
archaeological grounds and by data from other faunas
in order to decide what sort of taphocenosis the avian
one is. When this strictly circumstantial evidence is taken
into account, without ruling out the possibility that certain
taxa could have become incorporated into the
taphocenoses in a more or less "natural" manner (i.e.,
Gautier's taphonomic "intrusive" groups numbers 4 and
5), we tend to favor humans as the main accumulators
of birds at Velikent. Determination of the anthropic
assemblages as consumption refuse (i.e., Gautier's group

number 1) or manufacture refuse (group number 2) would
require larger and more trustworthy samples.

Table 5 summarizes a selection of biological features
for the different species and their zooarchaeological
record in the East European steppe sites from a
qualitative standpoint. One peculiar feature is the
discordance exhibited between aquatic and steppe taxa
where steppe birds constitute the bulk of the remains,
both in terms of NISP and MNI, but only a minimal
fraction of the diversity, a fact that relates in part to the
contribution of the Great bustard to the samples (Table
6). If partial recovery is important, a size bias toward
the largest birds should be expected, but should not, in
principle, apply to comparisons restricted to the larger

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Figure 3. Distal humeri of selected corvids. For crow (Corvus corone, A) the internal apex of the processus supracondylaris
dorsalis (1) is approximately of the same length or shorter than the external apex (2) as reported by Tomeck and Bochenski (2000)
on 84% of their specimens. That this feature might be subjected to a lot of geographic variation is indicated by the fact that close
to 50% of our reference specimens (14) exhibit the alternative conditions reported by these authors for 73% of their rooks (B,
internal apex longer) and 62% of their jays (C, internal apex developed as a spike). Thus the need to allocate one of the corvid
humeri from Velikent, with its broken internal apex apparently developed as a spike, to the category of Corvus corone/C.
frugilegus (Table 1).

fraction of the sample. Within this group of large birds
Great bustard was probably the main meat source,
with the diversity and relative frequencies of swans and
geese indicating that the cropping of aquatic birds was
secondary to that of terrestrial birds and was not
restricted or preferentially targeted to particular species.
This logic, however, partly breaks down if, for any

Figure 4. Distal tarsometatarsus of Spotted eagle, Aquila
clanga (above: dorsal view; below: ventral view).

reason, large birds cannot be considered strictly comparable
as a group in the area around Velikent. First, waterfowl are
more vulnerable in the summer, when molting. Second, the
easiest time of year to catch bustards in great numbers is
after a heavy snowstorm (Nikol'skii 1891/1892; Silant'ev
1898). Such seasonal limitations have both economic and
taphonomic implications in terms of time devoted to the
activity, processing of carcasses, bones that could end up
in the sediments, and the like. Third, waterfowl, but not the
Great bustard, provide products other than meat. Swan,
for example, has always been considered a low-quality
meat in Daghestan; what hunters were seeking from these
birds were feathers, not likely to leave traces in the
sediments (Dement'ev 1951; Nikol'skii 1891/1892; Silant'ev
1898). Like feathers, eggs leave few or no signatures at
the level of coarse analysis. Obviously, partial recovery
would compound the difficulties involved in a direct
comparison of remains, but if large-sized birds are not strictly
comparable to start with, there should be no point in trying
to specify provisioning strategies beyond a very coarse
level of analysis documenting that both terrestrial and
aquatic biotopes were cropped.
Much the same reasoning applies to the data on
seasonality. Of the three most abundant taxa, only the
lesser White-fronted goose is a wintering species (Table
5; Harrison 1982). Thus, based upon phenological
patterns exhibited by birds today (Burton 1995), the most
one can say is that some winter bird-hunting took place

MORALES-MUNIZ and ANTIPINA: Birds from Bronze Age Velikent

S"' b

] I N. -,

-I-- \

e -

(fI = I h

Figure 5. Fracture patterns for selected bone categories. Dashed lines indicate fracture lines or planes or both. Striped areas
define characteristic bone portions at Velikent (a: coracoid; b: scapula; c: carpometacarpus; d: humerus; e: furcula; f: femur;
g: ulna; h: tibiotarsus; i: sternum).

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 5. Selected biological features and zooarchaeological record of the Velikent avifaunas in neighboring areas of the East-
European Steppes. Status codes (Daghestan today): breeding (B), wintering (W), transit (T), resident (R); Abundance codes:
frequent (N), regular (R), infrequent (I); Phenology codes: January (J), February (F), March (M), April (A), May (Y), August (G),
September (S), October (0), November (N), December (D), Summer (SU); Biotope codes: Marshes/aquatic (MA), Steppe/open
(SO), Indifferent (IN); Food codes: plants (P), carrion (C), predator (D), omnivorous (0), small invertebrates (S), fish (F); Nesting
codes: ground (G), reeds (R), burrows and ruins (B), cliffs (C), trees (T), buildings (H); Limiting factors of populations' codes:
human activities (H), storms (R), brackish water (W), snow level in winter (S), winter temperature (T), nesting places (N), water
availability (A), carrion (C), prey (P); Archaeological record codes: Early Holocene (EH), Iron Age (IA), forest steppe (FS), open
steppe (S). Don steppe record comes from a single A.D. 900-1200 site; data taken from. Nikol'skii 1891/1892, Silant'ev 1898,
Dement'ev 1951, Voinstvenskii 1967, Harrison 1982, Jonsson 1992 and Boev 1993.

Code Status Abundance Phenology Biotope Food Nesting

Limiting EH IAFS IAS



+ +

+ +

+ + +


+ +


+ +


+ +
+ +
+ +

+ + +

+ +

+ + +

at Velikent. Most of the Velikent taxa qualify for more
than one phenological category, occasionally all four
(Table 5; Dement'ev 1951; Harrison 1982; Jonsson
1992). Velikent being a permanent settlement, one could
most parsimoniously expect bird hunting to have taken
place throughout the year, although perhaps shifting to
those taxa (and products) that happened to be more
accessible or sought-after at a particular time of the year.
Today quite a few of the Velikent taxa are recorded
as infrequent species (Table 5). This includes the Great
bustard, swans, and large preybirds whose hunting has
been amply documented since Paleolithic times

(Dement'ev 1951; Nikol'skii 1891/1892; Silant'ev 1898;
Voinstvenskii 1967). Still, abundances of large preybirds
might have always been low, with exceptions like the
Griffon vulture, so the decline might have more to do
with changes in stockbreeding practices and availability
of carcasses than with active human interference. There
is fairly good correspondence between abundance of a
particular taxon at Velikent and the zooarchaeological
record of species in neighboring steppe areas (Table 5).
Additionally, species previously unrecorded in Russian and
Ukrainian sites are also infrequent at Velikent.
Concerning the much debated issue of



MA S ?
IN O E ?

MORALES-MUNIZ and ANTIPINA: Birds from Bronze Age Velikent

synanthropization, our data are limited. The first
synanthopic birds in eastern Europe appeared during the
Bronze Age, becoming a typical feature of the agrarian
landscapes (Boev 1993; Iankov 1983; Doncev and
Iankov 1989). The earliest synanthropes are seasonal
(AI, at Velikent exemplified by the Mute swan and the
Grey heron) and passive (AII, represented at our site by
the coot and pheasant). Later to appear were the
synurbanists, of which at Velikent we have the crow, an
example of an initial synurbanist (BI), and the White
stork (BIII), with no representatives from the
intermediate group of advanced synurbanists (BII). If
avian colonization of human settlements took place in a
gradual way, the absence of advanced synurbanists is a
bit disturbing, although this might be resolved should the
C. corone/C. frugilegus humerus and ulna from
DAV95-IIC turn out to belong to a rook (Tables 2 and
3). Still, if aquatic birds had been transported to the site,
proof for synanthropization would vanish for the incipient
stages of colonization. As things stand now, the discussion
centers upon the largely arbitrary issue of trying to
grant an unequivocal status to a few species
represented by too few bones. From our standpoint,
Velikent's avifaunas are far too early and much too
displaced to the north to guarantee such an assignal.
Thus, our wish is to leave the issue open for the
moment until further, more reliable information
becomes available.

As stated in the introduction, the faunal analyses at
Velikent were undertaken with the specific aims first,
to prove the existence of a continuous occupation of
the settlement throughout the early/middle Bronze Age
and second, to assess the peculiarities of the
subsistence strategy, in particular in relation to the
exploitation of resources from the Caspian Sea. Bird
bones, indeed, have been documented in all the
excavation units thus far, although often in such low
numbers that, except for DAV'95IIC where they span
the whole sequence, one is not able to either prove or
disprove the first aim. As for the second, the
comparatively low NISPs of waterfowl and other
aquatic taxa contrast with the dominance of Great
bustard in all samples. Whether such a phenomenon
is a result of the manual recovery technique practiced
remains to be seen, although both the dominance of
terrestrial mammals and the small number of fish and
seal bones indicate that the Caspian Sea was not by
any means the basis of the animal economy at the

Table 6. Abundance spectra of the avifaunas from Velikent
expressed as NISPs (n = 140). MNIs (n = 101), and number of
taxa (n = 25).


% No. % No.

49 35 33 33 14 56
Steppe, Open
71 50.7 50 50 3 12
Indifferent, Other
20 14.3 18 18 8 32

site, which was one of production based preferentially
on domestic stock.
Still, even at this incipient stage, a series of hypotheses
that merit further analysis have been formulated. Thus,
the inference that the taphocenoses are mainly, but not
exclusively, a product of human activity, that there might
have been a differential treatment of groups (i.e.,
signatures of fire preferentially concentrated on the
waterfowl) and that in the east European steppe Velikent
might harbor the earliest proofs for the existence of
synanthropic taxa should be explored in detail in order to
refine aspects of this secondary resource procurement
strategy. Matters dealing with seasonality, the assignal
of taxa/remains to specific taphonomic and synanthropic
groups, and secular trends in taxonomic diversity will
require both an enlarged and more reliable database in
which patterns are not under permanent suspicion of
being distorted by partial recovery of remains.

We would like to thank Dr. Phil Kohl (Wellesley
College) and Rabadan Magomedov (Institute of
History, Archaeology, and Ethnography, Dagesthan
Scientific Center of the USSR Academy of Sciences)
not only for granting us the possibility to analyze the
Velikent faunal remains, but also for providing
information necessary for a more adequate sorting
and interpretation.

Antipina, Y.Y., and A. Morales (in preparation). Velikent: The
Bacher, A. 1967. Vergleichend Morphologische Untersuchungen
an Einzelknochen des postkranialen skeletts in Mitteleuropa
Vorkommender Schwine und Ganse. Ph.D. Dissertation.
Munich: Ludwig-Maximilians Universitat.
Baird, R.E 1989. Fossil bird faunal assemblages from Australian

caves: Precise indicators of Late Quaternary
environments? Paleogeography, Paleoclimatology,
Paleoecology 69: 241-244.
Bochenski, Z. 1997. Preliminary taphonomic studies on damage
to bird bones by Snowy owl, Nyctea scandiaca, with
comments on the survival of bones in paleontological
sites. Acta Zoologica Cracoviensia 40(2): 279-292.
Bochenski, Z., K. Huhtala, P. Jussila, E. Pulliainen, R. Tornberg,
and P. S. Tunkkari. 1998. Damage to bird bones in pellets
of Gyrfalcon, Falco rusticolus. Journal of Archaeological
Science 25:425-433.
Bochenski, Z., K. Huhtala, S. Sulkava, and R. Tornberg. 1999.
Fragmentation and preservation of bird bones in food
remains of the Golden eagle, Aquila chrysietos.
Archaeofauna 8: 31-38.
Boev, Z. 1993. Archeo-ornithology and the anthropisation of
birds: A case study from Bulgaria. Archaeofauna 2: 145-
Bramwell, D., D. W. Yalden, and PE. Yalden. 1987. Black grouse
as prey of the Golden eagle at an archaeological site.
Journal of Archaeological Science 14: 195-200.
Burton, J.F. 1995. Birds and Climate Change. London:
Christopher Helm.
Clason, A.T. 1972. Some remarks on the use and presentation
of archaeozoological data. Helinium 12(2): 139-153.
Dement'ev, G.P. 1951. Birds of the Soviet Union (5 volumes).
Moscow: Sovetskaya Nauka.
Doncev, S., and P. Iankov 1989. State and tendencies of
synanthropisation of birds in Bulgaria. Ecologia BAS 22:
35-42 (in Russian, English summary).
Ericson, P. 1987. Interpretations of archaeological bird remains:
A taphonomic approach. Journal of Archaeological
Science 14:65-75.
Gadzhiev, M. G, P. L. Kohl, R. G Magomedov, D. Stronach, A.
Morales, and A. M. Arnanz. 1997. The 1995 Daghestan-
American Velikent expedition: Excavations in Daghestan,
Russia. Eurasia Antiqua 3: 181-222.
Gadzhiev, M. G, P. L. Kohl, R. G. Magomedov, D. Stronach, S.
L. Gadzhiev, P. Lopez, B. Lopez, M. Moreno, A. Morales,
J. Moreno, A. Lopez-Saez, and M. Martin. 2000.
Daghestan-American Archaeological Investigations in
Daghestan, Russia, 1997-1999. Eurasia Antiqua 6: 47-124.
Gautier, A. 1987. Taphonomic groups: How and why?
Archaeozoologia II(a): 47-51.
Grayson, D. K. 1984. Quantitative Zooarchaeology. New York:
Academic Press.
Hargrave, L. L. 1970. Mexican macaws: Comparative osteology
and survey of remains from the Southwest. Tucson:
Anthropological Papers of the University of Arizona, no.
Harrison, C. 1982. An Atlas of the Birds of the Western
Palaearctic. London: Collins.
Iankov, P. 1983. On the stages and criteria of synanthropisation
in birds. Biologicheskie Osnovy Osvoenja Rekonstrukstii

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

i Ochrany Zhivotnogo Mira Belorusii: 133-134. Minsk:
Nauka i Teknica Publicatie House (in Russian).
Jonsson, L. 1992. Birds of Europe with North Africa and the
Middle East. London: Christopher Helm.
Laroulandie, V. 2000. Taphonomie et Archeozoologie des
Oiseaux en Grotte. Unpublished Ph.D. University de
Bordeaux I. Talence.
Livingston, S. D. 1989. The taphonomic interpretation of avian
skeletal part frequencies. Journal of Archaeological
Science 16:537-547.
Morales, A., and J. Rodriguez 1997. Black rats (Rattus rattus)
from medieval Mertola (BaixoAlentejo, Portugal). Journal
of Zoology 241: 623-642.
Morales, A., and E. Rosello 1998. Casual or intentional?
Comments on fish skeletal representation from Iberian
archaeological settlements. In P. Anreiter, L. Bartosiewicz,
E. Jerem, and W. Meid, eds. Man and the Animal World.
Budapest: Archaeolingua Alapitviny.
Morales-Mufiiz, A., and Y. Y. Antipina. 2000. Late Bronze Age
(2500-1000 B.C.) faunal exploitation of the East-European
steppe. Pp. 267-293 in Marsha Levine, ed. Late Prehistoric
Exploitation of the East-European Steppe. Cambridge:
McDonald Institute for Archaeological Research.
Mourer-Chauvir6, C. 1983. Les oiseaux dans les habitats
pal6olitiques: gibier des hommes ou proies des rapaces?
Pp. 11-124 in C. Grigson and J. Clutton-Brock, eds. Animals
and Archaeology. 2. Shell Middens, Fishes, and Birds.
Oxford: British Archaeological Reports (International
Series) no.183.
Nikol'skii, A. M. 1891/1892. Crimea Vertebrate Animals.
Moscow: Proceedings of the Imperial Academy of
Sciences, vol. 68 (in Russian).
Rich, P. V. 1980. Preliminary report on the fossil avian remains
from Late Tertiary sediments at Langebaanweg (Cape
Province) South Africa. South African Journal of Science
Serjeantson, D., B. Irving, and S. Hamilton-Dyer 1993. Bird
bone taphonomy from the inside out: The evidence of
gull predation on the Manx shearwater. Archaeofauna 3:
Silant'ev, A. A. 1898. Survey of Hunting Produce in Russia (in
Tomeck, M., and Z. Bochenski 2000. The Comparative
Osteology of European Corvids (Aves: Corvidae).
Cracow: Institute of Systematics and Evolution of
Animals, Polish Academy of Sciences.
Voinstvenskii, M. A. 1967. Fossil Ukrainian avifaunas. Pp. 3-
87 in Natural Landscapes and Fossil Faunas. Volume 3.
Kiev (in Russian).
Woelfle, E. 1967. Vergleichend Morphologische
Untersuchungen an Einzelknochen des postkranialen
Skeletts in Mitteleuropa Vorkommender Enten, Halbganse
und Sager. Ph.D. Dissertation. Ludwig-Maximilians
Universitit, Munich.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 55-64


Wietske Prummel'

Mollusk shells from three sites at Soirpi Bay in Thessaly, Greece. were studied. These are the Middle Bronze Age site of Magolila
Pavlina, the Hellenistic town of New Halos, and Hellenistic dwellings at the former Southeast Gate of this town. The mollusks of
all sites were collected in Sodrpi Bay and in the lagoon that existed just inland from the bay during the Middle Bronze Age.
Mollusks from shallow water with a soft floor prevailed at the three sites, among which the Mediterranean lagoon cockle,
Cerastoderma glaucum, was most numerous. Significant differences in valve wall thickness were found that presumably can be
explained by the change in habitat: the Middle Bronze Age cockles from the lagoon had thick valves, the cockles from the
Hellenistic period, which grew in river estuaries, had thinner walls. The inhabitants of the large town of New Halos consumed
small as well as large cockles, whereas the inhabitants of the Magodla Pavlina and the Southeast Gate could be more choosy and
consume larger cockles almost exclusively. The deep-water species Spondylus gaederopus was gathered by divers more often in
the Middle Bronze Age than in the Hellenistic period.

Key words: Cerastoderma glaucum, cultural preferences, ecology, mollusks, valve wall thickness

Marine mollusk shells have been found in large numbers
at three sites at Sodrpi Bay in Thessaly, Greece, that
date to the Middle Bronze Age (one site) and the
Hellenistic period (two sites) (Fig. 1; Table 1). Sodrpi
Bay is in the southwest corner of the Pagasitik6s Gulf.
The present town of V6los is situated in the northeast
corner of this gulf.
The western and southern shores of Soirpi Bay are
composed of gravel, sand, and mud. The floor of Soirpi
Bay falls gradually from west to east. A depth of 5 m is
reached at 150 m from the western shore, that of 10.8 m
at 550 m. At the southern end of the bay, which is even
shallower, a depth of 5 m is reached at 325 m from the
shore, that of 10.8 m at 1125 m.
The eastern shore of Soirpi Bay along the Mitz6la
peninsula is rocky and steep and has small gravel
beaches. A depth of 10.8 m is reached here about 55 m
from the coast. The deepest point of Soirpi Bay, 27.5 m,
is found at the outlet of the bay into the Pagasitik6s Gulf,
close to the eastern shore of the bay.
During the Middle Bronze Age a lagoon was situated
behind the western and southern shore of Soirpi Bay
(Fig. 1, right, shaded area). The lagoon was protected
from the bay by a spit (van Straaten 1988). The Middle
Bronze Age site that is discussed in this paper is the
Magodla Pavlina (MP), situated on the lagoon (Fig. 1,

'Groningen Institute of Archaeology, Poststraat 6, 9712 ER
Groningen, The Netherlands.

right). The human population of MP was about 200-400
(H. R. Reinders, personal communication).
In the Late Bronze Age the spit completely closed,
after which the lagoon was replaced by a salt marsh
along the open bay (van Straaten 1988). Several river
estuaries emptied into Sodrpi Bay (Fig. 1, right). The
first Hellenistic site discussed in this paper is the town
of New Halos (NH), then 1.5 km from the west shore
of Sodrpi Bay. (The west shore has moved some 10 m
since the Hellenistic period.) Founded ca. 302 B.C., this
large town of 1440 houses with an estimated population
of 8,000 to 9,000 people, mainly soldiers and their families,
was abandoned ca. 265 B.C. (Reinders 1988; Reinders and
Prummel 2002). The second Hellenistic site studied
consisted of a small number of dwellings at the former
Southeast Gate (SEG) of the town of New Halos, dated
at ca. 260-220 B.C. (Reinders and Kloosterman 1998) (Fig.
1). The population occupying these dwellings was, at most,
only several tens.
The three sites allow the study of the influence of
culture (Middle Bronze Age versus Hellenistic Period),
environment (a lagoon versus open sea and river estuaries)
(Claassen 1998: 126-129), and human population density
(low in MP and SEG, high in NH) on the marine shellfish
that was collected and consumed. Strong differences in
strength and size of the most common species, the lagoon
cockle (Cerastoderma glaucum), at the three sites
became obvious during the examination of the mollusk
materials. Studies of modem Mediterranean lagoon cockles

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Figure 1. Map of Greece with the area of the studied sites in black (left) and the current situation of the Pagasitik6s Gulf and Sourpi
Bay (right). 1) Magolla Pavlina (MP: Middle Bronze Age); 2) town of New Halos (NH: Hellenistic period); 3) dwellings at former
Southeast Gate of New Halos (SEG: Hellenistic period). The wavy shading represents the area that was a lagoon during the
Middle Bronze Age and a salt marsh during the Hellenistic period. Drawings by H. Zwier.

(Ivell 1979; Nicolaidou et al. 1988; Trotta and Cordisco
1998) were used to try to explain these differences.

During a surface survey on MP in June 1996, students
and staff of the Groningen Institute of Archaeology
collected all visible finds: pottery, flints and other worked
stone, bones, and mollusk shells. With a few later
exceptions, the pottery was all of Middle Bronze Age
date (Dijkstra et al. 1997; Reinders et al. 2001).
The almost complete homogeneity in date of the
pottery is seen as a guarantee that the bones of humans
(3), animals (124), and marine mollusk shells (285) that
were collected during the surface survey date to the Middle
Bronze Age. This was confirmed by the 14C dates of two
cattle bones (GrA-16888: 3580 40 B.P. and GrA-16889:
3675 40 B.P.), calibrated (95.4% confidence level) 2033-
1989, 1987-1873, 1843-1811, 1799-1775 and 2195-2175,
2143-1943 cal B.C., respectively. These dates place the
occupation at the middle of the Middle Bronze Age.
During excavation, Hellenistic materials NH and SEG
were collected. Animal remains were sampled carefully
by hand. Sieving was done at both sites, but no subfossil
bone or shell was recovered. The NH deposits were 0.4 m
deep at maximum. In these deposits, 899 mammal bones,
13 tortoise bone fragments, 1 fish vertebra, 504 marine
mollusk shells, 20 shells of terrestrial mollusks, half of them
the edible Helix figulina, plus one echinoderm

endoskeletal fragment were found. Although preservation
of bone and shell in the soil of the town of New Halos was
good due to the high lime content of the soil and the low
precipitation, remains of groups with more fragile skeletons,
such as echinoderms, were very rare or, in the case of
birds and crustaceans, absent (Prummel 2003).
The slightly younger SEG deposits were up to 1.5 m
in depth. The excavation of these dwellings and the
identification of the animal remains are in progress. At
present (May 2001), 2884 mammal bones, 1827
marine mollusk shells, 120 terrestrial gastropod shells
(most of them Helix figulina), 39 tortoise bone
fragments, 6 fish bones, 76 bird bones, and 7 crab
fragments (crustaceans) have been studied. The better
representation of fish and the presence of bird bones
and crustacean remains suggest slightly better
preservation conditions in the SEG than in the NH
deposits. Still, echinoderm remains have not been found
in the SEG deposits, and fish bones are rare. This could
suggest that fish was not much consumed in the two
Hellenistic sites at Sodrpi Bay.

All marine mollusk remains were identified with a
reference collection that was assembled for this purpose
from beach finds of Sotrpi Bay and the Pagasitik6s Gulf.
Identifications of the reference material were made with
Poppe and Goto (1991; 1993) and Delamotte and

PRUMMEL: Mollusk Meals from Sodrpi Bay, Thessaly, Greece

Vardala-Theodorou (1994) and checked by Rob
Moolenbeek of the Zoological Museum of the University
of Amsterdam (the Netherlands).
The numbers (N) for the bivalvia in Table 1 are the
sums of left and right valves. Minimum numbers of
individuals (MNI) are the numbers of gastropod shells
(N) and the numbers of bivalvia valve fragments divided
by two and raised to the nearest integers (Table 1).
The Cerastoderma glaucum valves were measured
to an accuracy of 0.1 mm according to standards for
bivalvia (length: from anterior to posterior margin; height:
from umbo to midline). The shells and valves of MP and
SEG were weighed to an accuracy of 0.1 g and those of
NH to an accuracy of 1 g. Graphs and statistics of
Cerastoderma glaucum valves were processed using
SPSS 8.0.

Most mollusk shells and shell fragments from the three
sites are from animals that were collected with the living
animal in it, and thus come from mollusks consumed by
humans. Few shells were dead beach finds. Damage by
opening the shells with a stone knife is visible on many
MP spiny oyster (Spondylus gaederopus) valves
(Prummel 2001). Evidence, such as large numbers of
small shell fragments of Hexaplex and Bolinus, that
would indicate purple dye production (Becker 2001) does
not exist. A Luria lurida from SEG had a hole in the
shell and obviously had been used as a pendant. A bead
made of mollusk shell was found in NH (Prummel 2002).
The Middle Bronze Age site differs from the
Hellenistic sites by the small number of recovered marine
mollusk species: 7 as opposed to 25 and 22 (Table 1).
Species with fragile shells, such as Patella sp., Mactra
stultorum, Donacilla cornea, Tapes decussatus, and
Solen marginatus, are lacking in the MP material, which
had been plowed to the surface and therefore may be
underrepresented. Excavation is needed to ascertain
whether or not fewer mollusk species were consumed
in MP than NH and SEG.
Slight differences in preservation conditions between
the two Hellenistic sites are illustrated by the absence of
Solen marginatus valves in the NH material. The much
deeper SEG deposits may have preserved the fragile
valves of this species better than the shallow NH layer.
Even if the fragile species are under-represented in
MP, the strong representation of Spondylus gaederopus
in the material of this Middle Bronze Age site (22.4% of
MNI) is obvious. In the two Hellenistic sites, the
proportion of spiny oyster valves is no higher than 1.7

and 0.9% of MNI (Table 1). The species lives cemented
to rock bottoms in 7 to 50 m of water (Poppe and Goto
1993: 68) and was procured by diving. With this species,
the deep, rocky habitat is much better represented in the
total mollusk MNI of MP than in the Hellenistic sites
NH and SEG (Table 2).
The lagoon cockle of the Mediterranean
(Cerastoderma glaucum) (Fig. 2), a species of
shallow, soft-bottomed lagoons and river estuaries, is
better represented in MP and SEG than in NH. In
NH, however, it also is the most common mollusk.
The shallow, soft habitat is best represented in each
site (Table 2). Limpets (Patella sp.), species of hard
substrate shallow-water habitats, oyster (Ostrea
edulis), a species of hard or soft substrate habitats
of shallow waters up to 90 m deep, and Noah's ark,
(Arca noae), a species of hard substrates in shallow
to 119 m water, are more common in NH than in the
two other sites (Table 1). In NH, also, hard surfaces
under shallow water (14.7% of MNI) and soft or hard
surfaces under shallow or deep water (17.7% of MNI)
are well represented (Table 2).
The MP lagoon cockle valves differ from those of
NH and SEG in two aspects. First, the MP valves are
on average much heavier than those from NH or SEG
(Table 3, Fig. 3). While the weight of mollusk valves
may change during their stay in the soil, the good
preservation conditions for bones and shells in MP soil
at the three sites would not change the diameter of the
valve wall. Nonetheless, this diameter is difficult to
measure because the shell sculpture has radiating ribs
and furrows. Moreover, cockle-valve wall thickness
decreases from the umbo to the bottom of the valve and
from the anterior to the posterior margin. The pallial line
halfway between the anterior and posterior margin was
chosen to measure the valve wall thickness (Fig. 2).
MP lagoon cockles had thicker walls than the NH
and SEG cockles. As might be expected, wall diameter
increases with valve height (Fig. 4). MP valves of 25-40
mm in height had wall thicknesses between 3 and 4 mm,
whereas NH valves of the same height had wall
thicknesses between 2 and 3 mm. Shell height and valve
wall thickness were measured on the same valves for
only part of the SEG lagoon cockles. The SEG valves of
25-40 mm height had wall thicknesses between 1.7 and
2.9 mm (Fig. 4), about the same as the NH valves.
Second, valve size, the mean of valve length and
height, differs considerably (Fig. 5, Table 3). Because
the lagoon cockle is equivalve, left and right valve
measurements have been combined. The ranges and

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 1. The numbers (N) of shell and valve fragments of marine mollusks in three sites near the coast of Solrpi Bay, their minimum
numbers of individuals (MNI) and the proportions for MNI. The sites are Magodla Pavlina (Middle Bronze Age), the Hellenistic
town of New Halos and some dwellings at the former Southeast Gate of New Halos (Fig. 1). The habitat column indicates where
the species lives; the first character refers to the substratum: H hard substrata, S soft substrata, A hard or soft substrata (all);
the second character refers to the depth of water: S: shallow water (to 6 m), D: 6 m or more deep water, A: shallow and deeper water
(habitat data: Poppe and Goto 1991; 1993).

Magodla Pavlfna New Halos

Southeast Gate
N MNI MNI% Habitat

Marine gastropoda
Patella caerulea Linnaeus, 1758 -
Patella rustica Linnaeus, 1758 -
Monodonta articulata Lamarck, 1822 -
Gibbula divaricata (Linnaeus, 1767) -
Gibbula cf. umbilicaris (Linnaeus, 1767) -
Gibbula albida (Gmelin, 1791) -
Cerithium vulgatum (Bruguibre, 1792) 13 13 6.5
Luria lurida (Linnaeus, 1758) -
Tonna galea (Linnaeus, 1758) -
Bolinus brandaris (Linnaeus, 1758)
Hexaplex trunculus (Linnaeus, 1758) 10 10 5.0
Buccinulum corneum (Linnaeus, 1758) 1 1 0.5
Fusinus syracusanus (Linnaeus, 1758) -
Conus ventricosus Gmelin, 1791 -
Unidentified marine gastropod

Marine bivalvia
Arca noae Linnaeus, 1758 8 4 2.0
Glycymeris insubrica (Brocchi, 1814) -
Mytilus galloprovincialis Lamarck, 1819 -
Pinna nobilis Linnaeus, 1758 -
Chlamys cf. glabra (Linnaeus, 1758) -
Spondylus gaederopus Linnaeus, 1758 89 45 22.4
Ostrea edulis Linnaeus, 1758 8 4 2.0
Psoedochama gryphina (Lamarck, 1819) -
Acanthocardia tuberculata
(Linnaeus, 1758) -
Cerastoderma glaucum (Poiret, 1789) 248 124 61.7
Mactra stultorum (Linnaeus, 1758) -
Donacilla cornea (Poli, 1795) -
Solen marginatus Pulteney, 1799 -
Callista chione (Linnaeus, 1758) -
Tapes decussatus (Linnaeus, 1758) -
Venus verrucosa Linnaeus, 1758 -
Unidentified marine bivalve
Unidentified marine mollusc

41 41

2 2
10 10
1 1

8 8
19 19
1 1

2 2

38 19
4 2
1 1
3 2

9 5
50 25
1 1

22 11
188 94
3 2
4 2

31 16
50 25

21 21 2.1 H/S
1 1 0.1 H/S
1 1 0.1 H/S
- H/S
2 2 0.2 S/S
43 43 4.3 S/S
1 1 0.1 H/A
75 75 7.5 S/A
33 33 3.3 S/A
- H/S
1 1 0.1 A/A
1 1 0.1 H/A

3 2 0.2 H/A
2 1 0.1 S/S
7 4 0.4 S/A
17 9 0.9 H/D
56 28 2.8 A/A
- H/D

16 8 0.8 S/A
1480 740 73.8 S/S
3 2 0.2 S/A
25 13 1.3 S/A
2 1 0.1 S/A
14 7 0.7 S/S
18 9 0.9 A/A

Total 377 201 100.0 504 293 100.0 1827 1003 100.0

histograms (Fig. 6) of valve length and height show that of the NH and SEG lagoon cockle valves strongly differ
minimum values for length and height in NH are much from the normal distribution (Fig. 6), nonparametric tests
lower than those in MP and SEG. SEG has slightly larger were used to test whether the lagoon cockles from MP,
maximum values for valve length and height than MP or NH. NH, and SEG could come from the same or identical
Because the distributions of valve length and height populations of lagoon cockles collected for consumption.

PRUMMEL: Mollusk Meals from Sodrpi Bay, Thessaly, Greece



Figure 2. Above: right valves of Cerastoderma glaucum from
the Middle Bronze Age site Magodla Pavlina. The star marks
the place where valve wall diameter was measured. Scale in
cm. Below: the measurements taken (length, height, valve wall

(It should be noted that the descriptive statistics of cockles
that were collected for consumption will differ from those
of populations not selected for consumption). The
Kruskal-Wallis H test for three independent samples (the
length and height valve measurements for MP, NH, and
SEG) gave Chi-squares for length and height of 40.992
and 71.424, respectively. The asymptotic significance
of these two Chi-square values (with df = 2) is 0.000.
Thus, MP, NH, and SEG are not significantly
homogeneous in valve length and height.
Results of Mann-Whitney-Wilcoxon tests for three
sets of two independent samples (MP-NH, MP-SEG,
NH-SEG) of lagoon-cockle valve lengths and heights
showed that the lagoon cockles from these three sites
came from three different populations of collected
cockles (Table 4). The lagoon cockles from SEG were
significantly larger in length and height than those from
NH, whereas those from MP were significantly larger
than those from either NH or SEG (Fig. 5).

Mollusks were consumed in quantity at the three sites
studied, the Middle Bronze Age site, MP, and the two
Hellenistic sites, NH and SEG. In the Hellenistic sites, a
wide variety of mollusks was consumed. Because the
MP material was collected during a surface survey, it is
not possible to state whether the choice of species was
really more restricted at this Middle Bronze Age site
than in the Hellenistic sites (Table 1). For all sites, we
cannot determine the potential consumption of
cephalopod species, which lack shells.
Mollusks living in shallow water with a soft floor
were the most consumed species in each of the three
sites (Table 2). The main species in this group was the
lagoon cockle (Cerastoderma glaucum). Shallow water
with soft bottoms was present at short distances from the
sites. The Middle Bronze Age lagoon, the west and south
shores of Sodrpi Bay, and the river estuaries all had this
type of marine bottom. The lagoon that was in existence
during the Middle Bronze age was a very good habitat
for C. glaucum (Ivell 1979; Nicolaidou et al. 1988; Trotta
and Cordisco 1998). During the Hellenistic period the
species lived in the estuaries of the rivers flowing into Sodrpi
Bay. In the bay itself, fish predation restricted the species.
The east coast of Soirpi Bay (Fig. 1) was rocky

Table 2. Magoula Pavlina, Hellenistic New Halos and Southeast
Gate of New Halos. Proportions of MNI of mollusks living on
(columns): hard habitats (H), soft habitats (S) and hard or soft
habitats (A) and below (rows): shallow water (S), shallow and
deep water (A) and deep water (D) (for the preference habitats
of the mollusk species see Table 1).

% living
in soft

% living
in soft or
on hard

% living
on hard

Shallow water(S)
Magodla Pavlina 68.2 -0.5
New Halos 42.3 -14.7
Southeast Gate 79.1 -2.3

Shallow and deep water (A)
Magodla Pavlina
New Halos
Southeast Gate

Deep water (D)
Magoila Pavlina
New Halos
Southeast Gate

0.3 0.3 2.0

Table 3. Number, range, mean, and standard deviation
(Std.Dev.) of length, height and weight of Cerastoderma
glaucum valves (left and right valves taken together) from the
Magotila Pavlina, New Halos and the Southeast Gate. The
weight was only measured for complete valves. Length and
height in mm, weight in g.
Length N Range Mean Std.Dev.
Magodla Pavlina 92 18.0-48.0 33.2 6.11
New Halos 49 12.2-48.8 26.3 7.03
Southeast Gate 323 14.4-54.4 29.9 6.22

Magoila Pavlina 144 17.2-45.5 31.6 5.30
New Halos 68 14.6-40.3 24.9 5.28
Southeast Gate 338 13.9-47.7 27.8 5.54

Magoila Pavlina 93 0.7-18.2 7.0 3.60
New Halos 54 0.9-11.0 2.6 1.86
Southeast Gate 265 0.5-16.4 3.5 2.65

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

and suitable for human collection of mollusks. For MP
and NH, mollusks collected on coastal rocks were of
some importance (Table 2). In MP, the main species
collected from rocks was the spiny oyster (Spondylus
gaederopus). In NH, limpets (Patella sp.), and Noah's
Ark (Arca noae) were the most common species
collected from rocks (Table 1).
Mollusks from shallow to deep water with soft or
hard bottoms were collected rather often by or on behalf
of the inhabitants of NH (Table 2). These species, of
which Bolinus brandaris, Hexaplex trunculus, Ostrea
edulis, Venus verrucosa, and Acanthocardia
tuberculata are the most common (Table 1), may have
been collected along the west and south shores of Solirpi
Bay. Bolinus brandaris and Hexaplex trunculus shells
are also common in SEG. Remains of Ostrea edulis,
Venus verrucosa, and Acanthocardia tuberculata in
NH suggest that, for the large human population of this
town, all types of mollusks were collected, whereas the

v 3

+ 2

0 1


Figure 3. Scatterplot of '"log of weight (in g) (LOGWEIGH) against height (in mm) of complete Cerastoderma glaucum valves from
1. Magodla Pavlina (Middle Bronze Age), 2. New Halos (Hellenistic period), 3. Southeast Gate (Hellenistic period).

PRUMMEL: Mollusk Meals from Sourpi Bay, Thessaly, Greece

10 20 30 40 50


Figure 4. Scatter plot of valve wall diameter (WALLDIAM) against height of Cerastoderma glaucum valves from the Southeast
Gate site (SEG). Scales in mm.

N= 92 144 49 68

SITE 1 2



323 338

Figure 5. Box plots of length and height of Cerastoderma glaucum valves from 1. Magotla Pavlina (Middle Bronze Age), 2. New
Halos (Hellenistic period), 3. Southeast Gate (Hellenistic period). Scale in mm.










v v v
V v v v
v v v V

Sw v V
vv ~w W Wv
vw v wv v v
v vv v v
v v wv v


ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing


St d Dev= 5 54
Mean= 27 8






,I 2LJ] 2al ml l.El 'OO u' 4'0


00 ,

Std. Dev = 5.30
Mean =31 8
N = 144.00

N5d.Om- a.

Std. Dev= 5.28
Mean = 24.0
N = 88.00

Std, D-2= 0.22

N 32300
22.0 300 38,0 460 540
18.0 260 34.0 42,0 50.0



150 200 250 300 35.0 400 450 500

Figure 6. Histograms and the normal curves of length and height of Cerastoderma glaucum valves from 1. Magodla Pavlina
(Middle Bronze Age), 2. New Halos (Hellenistic period), 3. Southeast Gate (Hellenistic period). Scale in mm.

MP and SEG inhabitants were more selective in their
choice of mollusks and hardly consumed these species
(see below). All mollusk species that were consumed in
the three sites could have been collected in Sodrpi Bay.

Spondylus gaederopus is far more common in MP
than in NH and SEG (Table 1). That the taste for this
delicious species was lost from the Middle Bronze
Age to the Hellenistic period is improbable. A second



PRUMMEL: Mollusk Meals from Sotirpi Bay, Thessaly, Greece

possible explanation is that the Hellenistic inhabitants
less often took the trouble to dive for this deep-water
species. A third explanation is that the consumption of
spiny oysters in the Pagasitik6s Gulf during the Neolithic
(Reinders et al. 1997) and the Bronze Age (Falkner 1975;
Prummel 2001) and the ring, bead, and button
production from spiny oyster valves during the
Neolithic (Tsuneki 1989:8-13) gradually decreased the
stocks of spiny oysters. It is clear that this last aspect
needs further study.
For Cerastoderma glaucum from the Middle
Bronze Age site MP, the much thicker valve walls than
those from the Hellenistic sites may be caused by
environmental factors. At Lago Lungo, a brackish lagoon
in Italy, Ivell (1979) studied the biology and ecology of
lagoon cockles (C. glaucum) from a lake and freshwater
mouth of a river and a seawater mouth that all fed Lago
Lungo. The freshwater mouth showed a much larger
community of lagoon cockles than the seawater mouth.
The same result was found by Trotta and Cordisco (1998),
who observed much larger densities (782 g/m2) of lagoon
cockles in the Lesina Lagoon on the Italian Adriatic coast
than in the nearby Fortore River estuary (417 g/m2).
The preference of the species for confined lagoon waters
that are protected from open sea was also found by
Nicolaidou et al. (1988) in the Messologhi lagoonal
system on the west coast of Greece. Predators that enter
from the open sea in more exposed areas of the lagoonal
system heavily reduce lagoon cockle densities.
Ivell (1979: 371 and his Fig. 4) found that growth in
valve height was greater in the freshwater mouth
community than in the seawater mouth community of
Lago Lungo. Freshwater-mouth cohorts (the lagoon
cockle may have three to four cohorts yearly; also see
Trotta and Cordisco 1998) increased from 7 to 25 mm in
valve height in 7 months, or from 11 to 22 mm in 5
months, while a typical seawater-mouth cohort in this
lagoon increased in valve height from 10 to 22 mm in
11 months. Correspondingly, the amount of CaCO3
and the dry organic content of the shell attained higher
absolute values in less time in the freshwater mouth
than in the seawater mouth.
The rates of shell growth and wall thickness of
species of Cerastoderma are influenced by temperature,
temperature fluctuations, trophic state, pH, oxygen content,
bottom type, and other factors (Eisma 1965). While the
influence of environmental factors on valve wall thickness
of the Mediterranean lagoon cockle has not been studied
yet, we conclude that one or several ecological factors
caused the differences in valve wall thickness between

Table 4. Mann-Whitney-Wilcoxon Test for two independent
variables of three samples of valve length and valve height of
Cerastoderma glaucum.
Mean Sum of
Length N rank ranks
Magoila Pavlina (MP) 92 85.77 7891
New Halos (NH) 49 43.27 2120
Mann-Whitney U 895
Wilcoxon W 2120
Z -5.884
Asymp. Significance
(2-tailed) .000

Magodla Pavlina (MP) 92 254.35 23400
Southeast Gate (SEG) 323 194.80 62920
Mann-Whitney U 10594
Wilcoxon W 62920
Z -4.201
Asymp. Significance
(2-tailed) .000

New Halos (NH) 49 125.20 6135
Southeast Gate (SEG) 323 195.80 63243
Mann-Whitney U 4910
Wilcoxon W 6135
Z 4.282
Asymp. Significance
(2-tailed) .000

Magoila Pavlina (MP) 144 128.38 18486
New Halos (NH) 68 60.18 4092
Mann-Whitney U 1746
Wilcoxon W 4092
Z -7.556
Asymp. Significance
(2-tailed) .000

Magoila Pavlfna (MP) 144 304.14 437%
Southeast Gate (SEG) 338 214.81 72607
Mann-Whitney U 15316
Wilcoxon W 72607
Z -6.445
Asymp. Significance (2-tailed) .000

New Halos (NH) 68 148.97 10130
Southeast Gate (SEG) 338 214.47 72491
Mann-Whitney U 7784
Wilcoxon W 10130
Z -4.200
Asymp. Significance
(2-tailed) .000

the lagoon cockles from the Middle Bronze Age, that were
collected in the lagoon, and those from the Hellenistic
period, that were collected in the river estuaries.

The differences in shell size (height and length) (Fig.
5, tables 3-4) can be explained partly by environmental
factors and partly by the strong differences in human
population size at the sites and, thus, in the degree of
exploitation of the lagoon cockle populations. The large
size of the MP lagoon cockles may be correlated to a
degree with the high shell-growth rate of the communities
that lived in the lagoon. The lagoon cockles in the less
confined, more open estuaries will have shown a lower
shell growth rate than those in the lagoon. The inhabitants
of the Hellenistic sites had to wait longer before the
lagoon cockles attained a good size.
No environmental differences in the estuaries would
have existed between the occupation periods of NH and
SEG. The size differences between the lagoon cockles
from these sites have to be explained by the differences
in human population size. Whereas the few inhabitants
of the last site could be choosy in the size of selected
lagoon cockles, the many of the first site had to accept
smaller lagoon cockles as well.
The enormous reproductive potentiality of the
Mediterranean lagoon cockle would have prevented
over-exploitation of the species, even by the inhabitants
of NH, the large town. Lagoon cockle populations have
seeds at almost all times of the year (Trotta and Cordisco
1998). Should humans harvest 90% of the lagoon cockles,
the remaining 10% are able to replenish the stock in a
short period (Ivell 1979: 378).

Becker, C. 2001. Did the people in Ayios Mamas produce purple-
dye during the Middle Bronze Age? Considerations on
the prehistoric production of purple-dye in the
Mediterranean. Pp. 122-134 in H. Buitenhuis and W.
Prummel, eds. Animals and Man in the Past. Essays in
honour of Dr. A.T. Clason, Emeritus Professor of
Archaeozoology Rijksuniversiteit Groningen. ARC-
Publication 41. Groningen: ARC.
Claassen, C. 1998. Shells. Cambridge Manuals in Archaeology.
Cambridge: Cambridge University Press.
Delamotte M., and E. Vardala-Theodorou. 1994. Kochulia apo
tis Ellinikes Thalasses (Shells from the Greek Seas).
Kifisid: Mousefo Goulandri Phusikis Istorias.
Dijkstra, Y., H. R. Reinders, V. Rondfri, and Z. Malakasi6ti.
1997. Van Duivelsberg tot Rode Rots: de survey van 1996
in de vlakte van Almiros (Griekenland). Paleo-aktueel 8:
Eisma, D., 1965. Shell-characteristics of Cardium edule L. as
indicators of salinity. Netherlands Journal of Sea Research
Falkner, G. 1975. Systematische Ubersicht uiber die
Molluskenfauna von der Magula Pevkakia. Pp. 189-190

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

in B. Jordan, ed. Tierknochenfunde aus der Magula
Pavkakia in Thessalien. Ph. D. Dissertation. Munich:
Fachbereich Tiermedizin, University of Munich.
Ivell, R. 1979. The biology and ecology of a brackish lagoon
bivalve, Cerastoderma glaucum Bruguiere, in Lago
Lungo, Italy. Journal of Molluscan Studies 45: 364-382.
Nicolaidou, A., F. Bourgoutzani, A. Zenetos, O. Guelorget,
and J.-P. Perthuisot. 1988. Distribution of molluscs and
polychaetes in coastal lagoons in Greece. Estuarine,
Coastal and Shelf Science 26: 337-350.
Poppe, G. T., and Y. Goto. 1991. European Seashells I
(Polyplacophora, Caudofoveata, Solenogastra,
Gastropoda). Wiesbaden: Christa Hemmen.
Poppe, G. T., and Y. Goto. 1993. European Seashells II
(Scaphopoda, Bivalvia, Cephalopoda). Wiesbaden:
Christa Hemmen.
Prummel, W. 2001. Spiny oyster (Spondylus gaederopus)
consumption during the Middle Bronze Age in Thessaly,
Greece. Pp. 465-471 in W. H. Metz, B. L. van Beek, and H.
Steegstra, eds. Patina: Essays Presented to Jay Jordan Butler
on the Occasion of His 80th Birthday. Groningen,
Amsterdam: Metz, van Beek & Steegstra private publishing.
Prummel, W. 2003. Animal husbandry and mollusc gathering
in Hellenistic Halos (Thessaly). Pp. 175-223 in H. R.
Reinders and W. Prummel, eds. Housing in New Halos, a
Hellenistic Town in Thessaly. Lisse, Netherlands: A. A.
Reinders, H. R. 1988. New Halos, a Hellenistic town in
Thessalia, Greece. Utrecht: Hes.
Reinders, H. R., S. Floras, E. Karimali, Z. Malakasioti, W.
Prummel, V. Rondiri, I. Sgouras, and M. Wijnen. 1997.
Karatsidhagli, a neolithic site in the Almir6s Plain
(Thessaly, Greece). Pharos 5: 85-143.
Reinders, H. R., and A. Kloosterman. 1998. Hellenistische
muntvondsten uit Nieuw Halos. Paleo-aktueel 9: 40-45.
Reinders, H. R., A. Kloosterman, and E. Schrijer. 2001. Middle
Bronze Age sites in the Almir6s and Soirpi plains
(Thessaly, Greece)-Interregional contacts? Pp. 473-479
in W. H. Metz, B. L. van Beek, and H. Steegstra, eds.
Patina: Essays Presented to Jay Jordan Butler on the
Occasion of His 80th Birthday. Groningen, Amsterdam:
Metz, van Beek & Steegstra private publishing.
Reinders, H. R., and W. Prummel, eds. 2003. Housing in New
Halos, a Hellenistic Town in Thessaly. Lisse, Netherlands:
A. A. Balkema.
Trotta, P., and C. A. Cordisco. 1998. Gonadal maturation,
conditioning, and spawning in the laboratory and
maturation cycle in the wild of Cerastoderma glaucum
Brugui6re. Journal of Shellfish Research 17: 919-923.
Tzuneki, A. 1989. The manufacture of Spondylus shell objects
at neolithic Dimini, Greece. Orient: Report of the Society
for Near Eastern Studies in Japan 25: 1-21.
van Straaten, L. M. J. U. 1988. Mollusc shell assemblages in
core samples from ancient Halos (Greece). Pp. 227-235 in
H. R. Reinders, ed. New Halos, a Hellenistic town in
Thessalia, Greece. Utrecht: Hes.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 65-80


Elizabeth J. Reitz'

Animal remains from Paloma, a Preceramic site in the Chilca Valley occupied between 7800 and 4700 B.P., indicate that marine
resources were primary to animal-based subsistence. In the Paloma material, terrestrial vertebrates and terrestrial invertebrates
contribute 10% of the estimated biomass, whereas marine vertebrates and invertebrates contribute 90% of the biomass. This
stratified site offers evidence for subsistence change, but the focus on marine resources did not change. Vertebrates of warm-
temperate waters are common in the Paloma assemblage and more common than at Peruvian coastal sites farther north at this same
time. Decline in anchovies between 5300 and 5100 B.P. coincides with the brief presence of warm-tropical animals and increase in
diversity. At approximately 4700 B.P., there is another decline in anchovies, an increase in high-trophic-level fishes, and a decrease
in diversity. Data are consistent with the end of the warmer, more humid Hypsithermal around 5000 B.P.

Key words: Archaic Period, climate change, marine resource use, mid-Holocene, Peru, zooarchaeology

Archaeologists have long debated the extent to which
humans used marine resources relative to terrestrial
animals and plants. The debate has centered on the ability
of marine resources to support large, sedentary
populations with complex social organizations and
monumental architecture, such as at the late Preceramic
site of El Paraiso, Peru (Quilter and Stocker 1983). From
a nutritional perspective, it is unlikely that people ever
concentrated exclusively upon either terrestrial or marine
resources (e.g., Weir et al. 1988; Weir and Dering 1986),
but vertebrate and invertebrate data from the middle
Preceramic village of Paloma, on the central coast of
Peru, support the argument that marine resources
provided most of the animal-based portion of the diet
during the Preceramic Period. Data from Paloma and
other Peruvian sites do not support the hypothesis that
this pattern developed out of an earlier coastal hunting
tradition focused on terrestrial resources.
The Paloma data, however, do support the hypothesis
that a change in fishing strategy occurred after 5300
B.P. when changes in subsistence strategy also occurred
at other coastal sites. At sites on the Peruvian north
coast, this change is seen as a shift in focus from warm-
tropical fishes to warm-temperate fishes (Reitz 2001)
and a decline in mean trophic level of the catch (Andrus
2000). At Paloma, it appears as a decrease in warm-
temperate fishes associated with upwelling, specifically
a decrease in anchovies, and an increase in mean trophic
level of the catch.
'Professor of Anthropology, Georgia Museum of Natural
History, University of Georgia, Athens, GA 30602-1882,

In 1975, Michael Moseley (1975) proposed the
Maritime Foundations Hypothesis. This hypothesis
proposes that the economic foundation of a civilization
need not be agricultural; a complex social organization
could be supported with maritime resources as the
subsistence base. Mosley's hypothesis applies to the
period between 3000 and 1000 B.C. It also has application
to the subsistence base at Paloma during the centuries
preceding the late Preceramic, or Formative, Period (Reitz
1988a, 1988b).
This paper examines the contributions of maritime
resources to coastal economies before 3000 B.C., rather
than implying that those resources formed the subsistence
base for complex social organization after that time. The
argument will focus on data from the Preceramic
(Archaic) village of Paloma. These data from Paloma
indicate that the maritime subsistence base of Peruvian
coastal economies in the late Preceramic was not a
recent phenomenon in Peruvian history, but appeared as
early as 7800 B.P. In fact, data from six Peruvian sites
occupied between 10,575 and 4780 B.P. support this
hypothesis (Keefer et al. 1998; Reitz 1995, 2001; Reitz
and Sandweiss 2001; Sandweiss et al. 1989; Sandweiss
et al. 1998).
The Paloma data are important in this argument
because the site is stratified, permitting an examination
of marine resource use over a 3000 year period. The
faunal remains from Paloma clearly demonstrate
continuity in the marine focus throughout this time. They
also indicate two important discontinuities. The first
discontinuity appears after 5300 B.P. as a decrease in
the use of anchovies with the brief presence of warm-

Figure 1. Location of Paloma and other sites.

tropical animals and an increase in diversity. The second
discontinuity is a further decrease in the use of anchovies,
an increase in mean trophic level of fishes represented,
and a decrease in diversity of fishes caught at the end of
the occupation, around 4700 B.P. Fishes associated with
upwelling shift in importance in the catches at Paloma
after 5300 B.P. The mean trophic level of marine
vertebrates exploited at sites further north is consistently
higher than that at Paloma. Over the period that the
mean trophic level of fishes used at Paloma increases,
the mean trophic level at northern sites declines (Andrus
Located in the Chilca Valley of Peru, about 65 km
south of Lima (Benfer 1984, 1999; Engel 1980; Fig. 1),
Paloma is an Archaic, or middle Preceramic, village with
no monumental architecture. Paloma is situated on an
alluvial plain 200-250 m above sea level on the north edge
of the lomas of Paloma. Lomas, also known as fog oases,
are plant communities found at about the 200 m line along
the desert coast of Peru. Lomas plants are generally
seasonally available, sustained by the austral winter fog
that forms at this elevation. The plain is approximately 250
x 600 m and the village occupies approximately 15 ha.
The Chilca River is dry, except for the summer when

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

some runoff from the highlands flows down its course.
The coast near the village is composed of both rocky and
sandy beaches. Paloma is located in the warm-temperate
Peru-Chilean Province.
More than 55 houses, associated with over 200 burials,
were encountered in the 2,860 square meters excavated.
The village is about 8 linear km north of the Chilca River
and approximately 4 linear km from the Pacific Ocean.
The number of levels per square varied depending
on the depth of the deposit. Four strata are represented,
designated as Levels 100 through 400. Level 100 is the
most recent stratum and Level 400 is the oldest. Earlier
strata were present at the site, but faunal remains from
them were not represented in the probability samples
reported here. Uncorrected radiocarbon dates indicate
the village was occupied between 7800 and 4700 B.P.
(Benfer 1990). In order of increasing depth, the dates
for each level are: Level 200 (4700-5100 B.P.), 300 (5100-
5300 B.P.), and 400-600 (5300-7800 B.P.). For Level
100, the occupational date is not based on radiocarbon
tests. Stratigraphically, Level 100 was deposited after
4700 B.P and the occupation probably ended before
3000 B.P. Level 100 is considered pre-Formative (before
the Cotton Preceramic) because neither ceramics nor
cotton were recovered. The end of the Hypsithermal is
roughly associated with the 5200-5000 B.P. date,
occurring at the end of the Level 300 occupation or during
the Level 200 occupation (Benfer 1999).

Two sampling strategies were followed by Benfer and
his colleagues for recovering faunal remains (Fig. 2).
During excavation of the 6 x 6 meter squares in 1976
and the 3 x 3 meter squares in 1979 some faunal materials
were collected as they were encountered, generally
without the aid of a screen. Fauna from these "grab"
samples are published elsewhere (Reitz 1988a, 1988b).
Because they provide the only evidence for marine
mammals, seasonal occupation, and long-distance trade,
the grab samples are important to the interpretation of
the site.
Fig. 2 shows the data from probability samples
collected from the walls of sixteen 3 x 3 m probability
squares, of which six are reported here. These six were
collected using a nested series of geological screens with
6 mm, 3 mm, and 1.5 mm mesh (Weir et al. 1988). The
samples were collected as standardized volumetric 5-
liter (5000 cm3) samples from each natural level. As
Fig. 2 shows, the portion of the site excavated was quite
small, but the samples are thought to represent the site

REITZ: Resource Use through Time at Paloma, Peru


1 25 50

L.j 0-0.25 m deposit depths
0.25-0.5 m deposit depths
0.5-1.0 m deposit depths

Probability Square

*10 -

Figure 2. Map of Paloma. Probability squares are shown as hatched squares and general block excavation units are darkened. The
squares from which the samples reported here were taken are designated by stars. The three strata also delimit the site.

as a whole because they are random stratified samples
of three areas of site defined by density of deposits
(Benfer 1990, 1999).
Standard zooarchaeological methods were used
during identification and analysis of the animal remains.
The identifications were made using the comparative
skeletal collection of the Environmental Archaeology
Laboratory, Florida Museum of Natural History,
University of Florida. Minimum Numbers of Individuals
(MNI) are estimated using conservative criteria. For
vertebrates these criteria include symmetry, size, and
age. For gastropods, MNI is based on the number of
apexes. In the case of bivalves, symmetry for valves
with hinge areas intact is determined, and the larger count
(left or right valves) is used as an estimate of MNI.
Each probability square is treated as a separate analytical
unit because the units were widely spaced across the
site. Each level within a square is also a separate
analytical unit. MNI is not normally estimated for high
taxonomic categories such as class. However, in the

case of Unidentified Mammal, the five unidentifiable
mammal specimens represent at least one additional
mammalian individual because they were too large to be
from the Unidentified Rodent. Hence the MNI includes
one Unidentified Mammal individual and one
Unidentified Rodent individual. It is not possible to
determine if the Unidentified Mammal was a marine
or terrestrial animal. The single bird specimen
indicates that at least one bird individual, of unknown
family, genus, or species, was present.
Although MNI is a standard zooarchaeological
quantification method, it has several problems (Reitz and
Wing 1999:194-199). A faunal assemblage may have a
higher percentage of anchovy individuals than guanaco
individuals, although it is unlikely that anchovies
contributed more meat than did guanacos. Further, MNI
is based upon the assumption that the entire animal was
utilized at the site. This ignores a basic facet of human
behavior: exchange and trade. In addition to these
problems, MNI is based upon paired elements. A large

Table 1. Allometric values used in this study.

Faunal category n

log a

Biomass from bone weight. k2


97 1.12 0.90
307 1.04 0.91
17 1.68 0.86
393 0.90 0.81

17 1.23 0.88
25 0.84 0.82
99 0.81 0.74

)le meat wgt. from shell wgt., gm
79 0.018 0.681
134 -0.162 0.918

quantity of unpaired elements, such as anchovy
vertebrae and drum teeth, generally indicates only
one individual in spite of the large number of vertebrae
or teeth present. In addition, the manner in which
data from multiple archaeological proveniences are
aggregated during analysis influences MNI results
(Grayson 1973). Some elements are more easily
identified than others. The taxa represented by such
elements may appear more significant in the species
list than they were in the daily diet.
One way to overcome these difficulties is to estimate
the amount of meat the identified animals contributed to
the diet. Estimates of meat (biomass) contributed by the
identified specimens and an animal's body size can be
made using the allometric principle that the proportions
of body mass, skeletal mass, and skeletal dimensions
change with increasing size. Meat weight, estimated from
the allometric relationship between body mass and skeletal
mass, provides information on the quantity of meat supplied
by the identified species. The relationship between body
mass and skeletal weight or a skeletal dimension is
described by the allometric equation
Y= aX"

(Simpson et al. 1960:397). In this equation, X is the
skeletal weight, Y is the quantity of meat, b is the constant
of allometry (the slope of the line), and a is the Y-intercept
for a log-log plot using the method of least squares

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

regression and the best fit line (Reitz and Wing 1999:221-
231). Values for a and b, derived from calculations based
on data at the Florida Museum of Natural History,
University of Florida, and the Zooarchaeology
Laboratory, Georgia Museum of Natural History, are
presented in Table 1.
Several factors need to be considered when
evaluating the biomass results. Biomass estimates for
invertebrates may be in error because no Pacific taxa
were included in the data from which the formulae were
obtained. Likewise, few Pacific vertebrates are included
in the vertebrate formulae. It is, however, doubtful that
the growth patterns for Pacific members of the drum
family, for example, are very different from that for
Atlantic members of that family. An additional problem
is that the allometric calculations in which specimen
weight is used to predict meat weight are influenced by
the weight of the specimens themselves, which in turn is
influenced by a variety of factors (Reitz and Wing
1999:170). The formulae for invertebrates predict soft-
tissue weight from archaeological shell weight, whereas
that for vertebrates predict specimen and soft-tissue
weight from the weight of the archaeological specimen.
The estimated biomass values for invertebrates and
vertebrates, therefore, are not as equivalent as the
treatment of them here would suggest. Improved
allometric formulae are needed to correct this problem.
The variety of exploited species and degree of
subsistence specialization of the human inhabitants can
be summarized by calculating the diversity and equitability
in the species identified. Diversity measures the number
of species used at the site. Equitability measures the
degree of dependence on the utilized resources and the
effective variety of species used at the site based on the
even, or uneven, use of individual species. These indices
enable discussion of food habits in terms of the variety of
animals used at the site (diversity) and the evenness
(equitability) with which species were utilized. Diversity is
measured using the Shannon-Weaver Index (Pielou 1966;
Shannon and Weaver 1949:14). Equitability is measured
using the Sheldon formula (Pielou 1966; Sheldon 1969).
Diversity increases as both the number of species
and the equitability of species abundance increases. A
diversity index of 4.99 is the highest possible value. A
sample with many species in which the number of
individuals slowly declines from most abundant to least
abundant will be high in diversity. Diversity is increased
if a new taxon is added to the list. Diversity is decreased
slightly if another individual of a taxon that is already
present, though scarce, is added. A low diversity can be

REITZ: Resource Use through Time at Paloma, Peru

obtained either by having few species or by having a
low equitability, where one species is considerably more
abundant than others. A low equitability value indicates
that one species was more heavily used than other species
in the sample. A high equitability index, approaching 1.0,
indicates an even distribution of species in the sample
following a normal pattern with a few abundant species,
a moderate number of common ones, and many rare
ones. It is important to remember that diversity and
equitability are dependent on sample size (Grayson
1981:82-85). They are no more reliable than the primary
and secondary data (MNI, specimen weight, number of
identified specimens [NISP], or biomass) used to
generate the indices.
Another way to assess the fishing strategy is to
estimate the mean trophic level of the catch for each
time period and to identify the trophic level from which
most of the resources were taken. Daniel Pauly and his
colleagues suggest that since the 1950s commercial fishing
has declined in response to a significant change in marine
ecosystems caused by over-fishing (Pauly and Christen-
sen 1995; Pauly et al. 1998, 2000). They term this "fishing
down the food web" and estimate the mean trophic level
of commercial fisheries using modern fisheries data.
In the study of Paloma, the trophic level exploited is
estimated using a modification of this approach,
substituting archaeological data from Paloma for modem
fisheries data. Archaeological data are assigned a trophic
level using modern trophic level data presented by
Froese, Pauly, and others (1998; Froese and Pauly 1998).
This calculation was done using only fishes in order to
be comparable to similar calculations using archaeological
data from northern Peru (Andrus 2000). In some cases
it was necessary to use higher taxonomic levels because
the taxonomic identification in the archaeological data.
the trophic level study, or both were insufficiently
precise. If the archaeological taxonomic identification
was not included by Pauly, the trophic level for the closest
taxonomic category was used. In estimating the
importance of a trophic level to the fishery of a specific
time period, the formula

TL, = I (TL,)(Biomass,) / Y Biomass,

was used; where TL, is the mean trophic level for the
time period i; TLij is the mean trophic level of the taxon
j for the time period i; and Biomassj is the estimated
biomass of the taxon j for the time period i. This
formula estimates the mean trophic level at one point
in time.
In interpreting biological materials, careful attention

Table 2. Paloma probability samples, summary, levels combined.

n %

Terrestrial mammals
Terrestrial gastropods
Marine gastropods
Other mollusca

Rocky beach mollusks
Sandy beach mollusks

gm %
84.5 8.6
8.9 0.9
746.3 75.6
5.4063 0.5
10.7499 1.1

85.9123 8.7
45.3716 4.6


96.6622 9.8
45.3716 4.6

should be paid to sample reliability. A clue to this is
sample size. Samples of fewer than 200 individuals may
be unrepresentative of the subsistence activity at a site
(Wing and Brown 1979:118-121). Specimen count and
MNI are interrelated; samples with low specimen counts
and low MNI are probably incomplete. The species list
will be too short and the relationships of the identified
taxa will be inaccurate. Derived measures, such as
diversity and equitability, are also dependent on sample
size (Grayson 1981). In the case of the Paloma data,
increased sample sizes would probably alter details of
the species list and diversity for each level. It is unlikely
that the basic emphasis upon marine resources
observed in these samples will change with more data,
but only additional work will prove or disprove this

The primary data are presented in six tables. Table 2
summarizes all the probability data for the site. Tables 3
and 4 summarize these same MNI and biomass data,
but for each level. MNI and biomass for taxa identified
in the probability samples from each of the four levels
are presented in tables 5 and 6. Table 7 presents the
diversity and equitability values. Data on taxa for which
MNI is not estimated (e.g., Unidentified Vertebrate) are
reported elsewhere (Reitz 1988b).
Marine resources were the primary source of the
diet derived from animal foods (Table 2). Less than 1%
of the individuals are birds and mammals, and these
contribute less than 9% of the biomass. Even these
animals might represent marine rather than terrestrial

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 3. Paloma probability samples, summary of MNI by level.
Level 100 Level 200
Taxa MNI % MNI %

Terrestrial gastropods
Marine gastropods
Other mollusca


13 39.4

13 39.4

35 31.3

Level 300
2 0.9
1 0.4
54 23.8
6 2.6
19 8.4

Level 400

27 49.1

2 3.6

25 45.5

1 1.8

Rocky beach mollusks 16 48.5 68 60.7 148 65.2 27 49.1
Sandy beach mollusks 3 9.1 5 4.5 13 5.7 0

Vertebrates 13 39.4 35 31.3 57 25.1 27 49.1
Invertebrates 20 60.6 77 68.8 170 74.9 28 50.9

Small fish 5 15.2 25 22.3 38 16.7 20 36.4
Large fish 8 24.2 10 8.9 16 7.0 7 12.7
Mytilidae 3 9.1 46 41.1 129 56.8 25 45.5

resource use because identification of the bird and one
of the mammals is not sufficiently precise to determine
if they were terrestrial or marine. Terrestrial
invertebrates, represented only by land snails (Scutalus
spp.), contribute 1% of the individuals and less than 1%
of the biomass. Marine fishes contribute 30% of the
individuals and 76% of the biomass. The remaining
individuals (68%) and biomass (14%) are marine
Although some change can be seen in the percentage
of individuals and biomass contributed by invertebrates
and vertebrates during the 3000 years represented by
the four levels, fish are the major contributors of biomass
in all four levels (tables 3 and 4). Fish average 36% of
the individuals and 83% of the biomass in the four levels.
Invertebrate animals average 64% of the individuals in
the four levels and 13% of the biomass. The number of
vertebrate individuals is highest in Level 400 and lowest
in Levels 200/300. The reverse is true for invertebrates.
There is a similar change between vertebrate and
invertebrate biomass in Levels 400 and 300. At the later
end of the sequence (Level 100), small fishes and mussels
(Mytilidae) are less common and large fishes are much
more abundant.
Vertebrates. Changes in the percentage of fish

individuals and biomass are found among the levels (tables
3 and 4). Use of fish declines from 49% of the individuals
and 92% of the biomass in Level 400 to 24% of the
individuals and 67% of the biomass in Level 300. Fish
percentages increase to 31% of the individuals and 82%
of the biomass in Level 200, and 39% of the individuals
and 91% of the biomass in Level 100. The decline in
fish between Levels 400 and 300 affects both large
fishes, such as cabrilla (Paralabrax spp.), jacks (cf.
Carangidae, Trachurus spp.), grunts (Haemulidae,
Anisotremus spp., Haemulon spp., Isacia spp.,
Orthopristis spp.), and drums (Cynoscion spp.,
Paralonchurus peruanus., Corvina [Sciaena]
deliciosa), as well as small fishes, such as herrings
(Clupeidae) and anchovies (En graulidae). However, the
increase in fish in Level 100 is primarily due to an
increase in MNI and biomass of grunts and drums (tables
5 and 6). Herrings are not present in Level 100 and the
percentage of anchovies declines from Level 200 to
Level 100. Anchovies are the most common individuals
in samples from Levels 200, 300, and 400. They are the
largest source of biomass in Levels 200 and 400.
Compared to Level 400, in Level 100 they contribute
about half the number of individuals and the biomass.
The decline in fish between Level 400 and Level

REITZ: Resource Use through Time at Paloma, Peru

Table 4. Paloma probability samples, summary of biomass (gm) by level.

Level 100
Taxa Biomass %

Terrestrial gastropods
Marine gastropods
Other mollusca

Rocky beach mollusks
Sandy beach mollusks


Small fish
Large fish

84.5 90.9

3.2552 3.5

2.7108 2.9
2.4571 2.6


5.966 6.4
2.4571 2.6

84.5 90.9
8.423 9.1


Level 200
Biomass %

186.7 82.5

2.8786 1.3

27.943 12.3
8.8602 3.9


30.8216 13.6
8.8602 3.9

186.7 82.5
39.682 17.5


Level 300
Biomass %


46.0884 8.3
34.0543 6.2


50.5381 9.1
34.0543 6.2

462.6 83.7
89.999 16.3


Level 400
Biomass %

105.9 91.9

0.1664 0.1

9.1701 8.0



105.9 91.9
9.337 8.1


300 is accompanied by an increase in invertebrates.
However, this decline also is due to the identification in
Level 300 of the only non-marine organisms in the
probability samples, Unidentified Mammal, Unidentified
Rodent, Unidentified Bird, and land snails (Scutalus
spp.). This might be interpreted as a major shift in the
subsistence effort were it not for the fact that the Level
300 collection is much larger than the collections from
other levels. The number of identified specimens in Level
300 is 3,318 out of a total 6,152 specimens in the
probablility samples. The two mammals and the single
bird are from the same level in the same square
(N125E145). The number of identified specimens from
Level 300 of N125E145 constitute 22% of all of the
probability specimens included in this study. Such an
increase in diversity, and the identification of unusual
taxa, is what would be expected of a larger sample. (In
the grab samples, Camelidae and Cervidae are ubiquitous
in Level 300, but very rare in Levels 100-200.)
Invertebrates. Among the marine invertebrates
some changes in exploitation through time are also evident
(tables 3 and 4). Invertebrates increase from 51% of
the individuals and 8% of the biomass in Level 400 to
75% of the individuals and 16% of the biomass in Level
300. The increase in invertebrates between Levels 400

and 300 is partly due to an increase in the use of mussels
(Mytilidae), particularly Perumytilus purpuratus (tables
5 and 6). P. purpuratus increases from 14% of the
individuals in Level 400 to 24% of the individuals in Level
300 even though it contributes only 5% of the biomass in
both levels. Much of the increase in mussels between
Level 400 and Level 300 also is due to an increase in P.
purpuratus, a rocky shore mussel, as well as the mussel
Semimytilus algosus.
Another reason invertebrate use increases in Level
300 is the identification of several new taxa. Two of
these new taxa, the wedge clam (Mesodesma
donacium) and the Venus clam (Protothaca thaca),
contribute 6% of the biomass in Level 300. The more
abundant of the new taxa, P thaca, contributing 4% of
the individuals and 5% of the biomass, is found only in
Level 300 of square N125E145, the same one with the
only evidence for terrestrial animal use in the probability
samples. Both M. donacium and the land snails Scutalus
spp. are represented in several squares. M. donacium
and P. thaca are both sandy-beach resources, a biotope
not represented in Level 400.
Invertebrate use appears stable between Levels 300
and 200 largely due to the identification of the mussel
Choromytilus chorus in Level 200 of square N120E10.

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 5. Paloma probability samples, MNI by level.


Unidentified Mammal
Unidentified Rodent
Unidentified Bird
Clupeidae (herrings)
Engraulidae (anchovies)
Paralabrax spp. cabrillaa)
cf. Carangidae (possible jacks)
Trachurus spp. (jurels)
Haemulidae (grunts)
Anisotremus spp. (grunt)
Haemulon spp. (grunt)
Orthopristis spp. (grunt)
Isacia spp. (cabinza)
Cynoscion spp. (seatrout)
Paralonchurus peruanus
Corvina (Sciaena) deliciosa (lorna)
Fissurella crassa (keyhole
Scurria parasitica (limpet)
Crepidula spp. (slippersnail)
Scutalus spp. (landsnail)
Polyplacophora chitonss)
Aulacomya water (mussel)
Choromytilus chorus (mussel
Perumytilus purpuratus (mussel)
Semimytilus algosus (mussel)
Mesodesma donacium (wedge
clam [macha])
Protothaca thaca (Venus clam)
Decapoda (crabs)


Level 100

Level 200

Level 300

Level 400

5 15.2

5 15.2

5 4.5

3 1.3

2 3.6

1 0.9

1 3.0

1 3.0

4 3.6


1 1.8



C. chorus contributes 15% of individuals and 8% of
biomass in Level 200, though identified only from this
one unit. P thaca is not present in Level 200, but M.
donacium is, contributing 5% of the individuals and 4% of
the biomass, an increase from its contribution in Level 300.
The percentage of invertebrates declines between
Levels 200 and 100, primarily because of a decrease in
mussels (Mytilidae). Aulacomya ater and S. algosus,
the only mytilids identified in Level 100, contribute only
9% of the individuals and 3% of the biomass in Level
100. Two taxa show an increase: slippersnails (Crepidula
spp.) contribute 24% of the individuals and 3% of the

biomass in Level 100; wedge clams (M. donacium)
increases to 9% of the individuals, although biomass
decreases to 3% in Level 100.
More individuals and more biomass in each of the
four levels are from rocky beaches than from sandy
beaches. M. donacium and P thaca are the only sandy
beach taxa identified. M. donacium is not present in
Level 400, but is present in small amounts in the other
levels. In combination with the lower intertidal mytilid,
Perumytilus, P. thaca are identified only in Level
300, where they are one of the largest contributors
of invertebrate biomass. In Level 300 sandy and rocky

REITZ: Resource Use through Time at Paloma, Peru

Table 6. Paloma probability samples, biomass (gm) by level.

Level 100
Taxa Biomass %

Level 200
Biomass %

Level 300
Biomass %

Unidentified Mammal
Unidentified Rodent
Unidentified Bird
Clupeidae (herring)
Engraulidae (anchovies) 23.2
Paralabrax spp. cabrillaa)
cf. Carangidae (possible jacks)
Trachurus spp. (jurels)
Haemulidae (grunts) 30.6
Anisotremus spp. (grunt)
Haemulon spp. (grunt)
Orthopristis spp. (grunt)
Isacia spp. (cabinza)
Cynoscion spp. (seatrout)
Paralonchurus peruanus
(coco) 11.8
Corvina (Sciaena) deliciosa (lorna) 18.9
Fissurella crassa (keyhole
Scurria parasitica (limpet) 0.02
Crepidula spp. (slippersnail) 3.23
Scutalus spp. (landsnail)
Polyplacophora chitonss)
Aulacomya water (mussel) 2.15
Choromytilus chorus (mussel
Perumytilus purpuratus (mussel)
Semimytilus algosus (mussel) 0.55
Mesodesma donacium (wedge
clam [macha]) 2.45
Protothaca thaca (Venus clam)
Decapoda (crab)




13.4 5.9
97.2 42.9

25.2 11.1
6.4 2.8

44.5 19.7

52 3.5

23 2.3

'85 0.6

71 2.6




8.8602 3.9


beaches contribute similar percentages of biomass,
although individuals found in rocky habitats are far
more common. As individuals from rocky habitats
decline, M. donacium individuals increase between
Levels 200 and 100.
Diversity and Equitability. Differences in diversity
and equitability among the levels are slight in most cases
(Fig. 3; Table 7). Diversity in the MNI data rises from a
low in Level 400 to a high in Level 200, and then declines
in Level 100. MNI equitability is high in Level 100,
probably because only ten taxa were identified, but four
of these dominate the collection.

Biomass diversity is generally low (Fig. 4; Table
7). Level 300 collection is the most diverse. Although
three taxa dominate Level 300, the amount of biomass
contributed by the remaining 19 taxa is variable. This
is also the largest of the collections. The biomass
diversity is similar in Levels 100, 200, and 400.
Biomass equitability is relatively unchanged
throughout the occupation.
Diversity and equitability estimates for MNI and
biomass data indicate that relatively few taxa were used.
Most of the taxa contribute individuals or biomass in
relatively equal amounts, but two or three animals dominate

Level 400
Biomass %


38.8 7.0

19.7 3.6
50.0 9.0


27.7609 5.0
18.3275 3.3

7.3590 1.3
26.6953 4.8


13.4 11.6
47.7 41.4

9.4 8.2
17.2 14.9
6.4 5.6

11.8 10.2

0.0832 0.1
0.0832 0.1

5.9113 5.1
3.2588 2.8


ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 7. Paloma probability samples, diversity and equitability.

MNI n Diversity Equitability

Biomass, gm n Diversity Equitability

100 33 10 2.0773 0.9022 92.9231 9 1.6309 0.7423
200 112 14 2.3311 0.8833 226.3818 12 1.7486 0.7037
300 227 23 2.1075 0.6721 552.5987 22 2.4071 0.7787
400 55 11 1.7231 0.7186 115.2365 10 1.7612 0.7649

the collection in each level. The abundant taxa are usually
a combination of anchovies, grunts, drums, and a species
of mussels. The dominant resource in terms of meat are:
Level 400, anchovies; Level 300, mammals, herrings, and
anchovies; Level 200, anchovies; and Level 100, anchovies,
grunts, and loma (Corvina [Sciaena] deliciosa).

The method by which vertebrates are assigned to
temperate and tropical groups is described in detail
elsewhere (Reitz and Sandweiss 2001: Table 2). Using
a number of sources (Chirichigno 1974, 1982; Fowler
1945; Hildebrand 1945; Schweigger 1964), vertebrates
found in early Peruvian and Ecuadorian archaeological
assemblages are classified as warm-temperate animals;
warm-tropical animals; mixed, unclassified marine
animals; or terrestrial animals. Warm-temperate forms
are more likely to be found in the Peru-Chilean Province
and warm-tropical ones in the Panamanian Province.
At most coastal sites marine vertebrates dominate the
collection. The most abundant warm-temperate forms
are penguins (Spheniscus humboldti), boobies (Sula
spp.), herrings (Clupeidae), anchovies (Engraulidae), and
some species of grouper (Paralabrax spp.), grunt
(Anisotremus scapularis, Isacia conceptionis), and
drum (Paralonchurus peruanus, Corvina [Sciaena]
deliciosa, Cilus [Sciaena] gilberti, as well as mackerel
(Scombridae). These contrast with the most common
warm-tropical fishes, such as bonefish (Albula vulpes),
sea catfishes (Ariidae), other species of grouper
(Epinephelus spp.), grunt (Conodon spp., Haemulon
spp., Orthopristis spp.), and drum (Larimus spp.,
Umbrina cf. xanti), as well as lisas (Mugil spp.), and
puffers (Spheroides spp.). Several other taxa are typical
of warm-temperate waters, but may also be found in
warm-tropical settings. These variable species are
unclassified and referred to as mixed. While none of
these species is confined to the condition in which it is
classified here, this approach permits a rough
classification of fishes in terms of water conditions.

The Paloma probability samples are dominated by
temperate fishes typical of the Peru Current; only 1%
of the individuals are tropical/estuarine. Tropical animals
are present in the Paloma probability samples only in
Level 300 (Table 8). This not simply a function of
recovery technique; tropical vertebrates constitute only
5% of the individuals in the grab samples (Reitz and
Sandweiss 2001: Table 3). The tropical species identified
in either the probability samples or the grab samples are
Arius spp., Caranx spp., Haemulon spp., Orthopristis
spp., and Bairdiella spp.

Using modern fisheries data, Daniel Pauly and his
colleagues demonstrate that the mean trophic level of
commercial fisheries as reported by the Food and
Agricultural Organization of the United Nations has
declined throughout the world (Pauly and Christensen
1995; Pauly et al. 1998, 2000). They report a shift away
from long-lived, piscivorous, high-trophic-level bottom
fishes to short-lived, low-trophic-level invertebrates and
small planktivorous pelagic fish in response to changes
in the relative abundance of the preferred catch. Fishing
down the food web to lower trophic levels initially led to
larger catches, then to stagnant or declining ones. They
argue that today's fishing industry is focused on small
pelagic organisms, such as anchovies. This strategy is
unsustainable in large part because it removes a food
source that is important to higher trophic level fishes.
The higher trophic level fishes are the ones that humans
have more frequently preferred. In essence, humans are
competing with their preferred prey for resources at the
lower trophic levels.
Both the theory and the method used to support it
are applicable to archaeological data. Archaeologists have
occasionally argued that over-harvesting might be an
explanation for cultural changes observed at various
times and places around the world. Such arguments are
largely qualitative; the observations upon which they are
based are difficult to quantify and test within an objective


REITZ: Resource Use through Time at Paloma, Peru

theoretical framework. In theory, such changes could
be stimulated by conquest or internal cultural dynamics
unrelated to environmental change. It is difficult to
substantiate that a change that took place in resource
use was associated with environmental change alone
and none of the other alternatives. It is particularly difficult
to verify that environmental change was caused by
human behavior. The method used by Pauly and his
colleagues (Pauly and Christensen 1995; Pauly et al.
1998, 2000) to quantify changes in trophic levels exploited
provides a way to quantify fishing strategies which can
be applied to archaeological data to demonstrate that a
change did or did not occur in fishing strategies. The
problem of causality remains unresolved, but it comes
closer to resolution as it becomes more clearly defined.
In Levels 200 to 400, the focus was on fishes from
trophic levels 2.2-2.6 for both MNI and biomass, but this
changed in Level 100 (Fig. 5; tables 9 and 10). In Level
100, the emphasis is on trophic levels 3.4-3.5. This shift is
reflected in the mean trophic level for each
archaeological level (figs. 3 and 4; Table 9). The mean
trophic level is relatively stable from Levels 400 through
200, but rises sharply in Level 100 for both MNI and
biomass. The rise in Level 100 is accompanied by a
decline in diversity. Corresponding with this shift is a
decline between Levels 200 and 100 in herrings
(Clupeidae) and anchovies (Engraulidae). Herrings are
not present at all in Level 100 and anchovies decline
from 59% of the vertebrate biomass in Level 200 to
27% of the biomass in Level 100. Temperate drums
(Paralonchurus peruanus and Corvina [Sciaena]
deliciosa) increase between Levels 200 and 100. Over
70% of the individuals and 56% of the biomass are from
trophic levels 2.2-2.6 in Levels 400-200 (Table 10). This
dominance of low trophic level resources is not present
in Level 100. In Level 100, 61% of the individuals and
72% of the biomass are from trophic levels 3.4 and 3.5.
This increase in mean trophic level in Level 100 can be
attributed to a reduction in the use of low trophic level
fishes associated with upwellings and a focus on just a
few large temperate-water fish from higher trophic levels.

1. Marine resources were the primary source of
animal protein at Paloma, with no evidence of a
developing marine focus out of a terrestrial hunting base.
2. The overall subsistence strategy at Paloma was
relatively stable throughout the 3,000-year period
represented by these data. There were changes,
however, within that strategy.







Figure 3. Paloma probability samples, trophic level, diversity,
and equitability derived from MNI. B.P. dates uncorrected.

3. The degree to which fish contributed to the diet
varied during the occupation. Fish were most commonly
used in the earliest occupation, Level 400. They were least
abundant in Level 300, with a wider range of possibly
terrestrial vertebrates and sandy beach invertebrates used.
4. Diversity in the MNI data rises from a low in
Level 400 (7800-5300 B.P.) to a high in Level 200 (5100-
4700 B.P.), and declines after 4700 B.P. (Level 100).
5. Biomass diversity is generally low. Level 300
(5300-5100 B.P.) is the most diverse collection in terms
of sources of animal protein.
6. The Paloma probability samples are dominated
by temperate fishes typical of the Peru Current. Tropical
animals are present in the Paloma probability samples
only in Level 300. This is also the level in which terrestrial
animals are most frequent.

400 300 200 100
7800-5300 5300-5100 5100-4700 4700



S &

- 1,0

Figure 4. Paloma probability samples, trophic level, diversity,
and equitability derived from biomass. B.P. dates uncorrected.

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 8. Paloma probability
and warm-tropical vertebrae
Warm-tropical fishes
Warm-temnperat fishpe

samples, MNI of warm-temperate 9. Another change occurred after 4700 B.P. There
tes by level, was greater emphasis on high-trophic-level grunts and
100 200 300 400 drums, a decline in diversity, and a decline in the use of
2 low-trophic-level fishes, specifically anchovies,
8 31 47 94 associated with upwellings.

Mixed fishes 5 4 5 3
Other vertebrates 0 0 3 0

Total 13 35 57 27

7. The most commonly utilized trophic levels from
7800 to 5100 B.P. were 2.2-2.6. This changed
dramatically in Level 100, where the emphasis was on
trophic levels 3.4-3.5. This increased use of higher trophic
levels is accompanied by a decline in diversity. The mean
trophic level throughout the Paloma occupation, however,
is much lower than the mean trophic level at sites in
northern Peru, such as Ostra Base Camp (Andrus 2000;
Reitz and Sandweiss 2001).
8. A decline in the use of anchovies between 5100
and 5300 B.P. coincides with the brief presence of warm-
tropical animals (Level 300), but the mean trophic level
does not change and most animal protein continues to
be from trophic levels 2.2-2.6.

Moseley (1975:40-43) postulates that marine products
constituted the main source of protein in a diet in which
loma resources contributed the majority of the food in
the late Archaic. Both plants and animals could be taken
from the sea as well as from the lomas, but the lomas
offered few animal protein sources. Most of the loma
resources were probably plants rather than animals. It
is not known to what extent the sea provided plant foods,
but marine animals clearly dominated that part of the
subsistence strategy directed toward animals.
The use of terrestrial or marine plants does not
necessarily correlate with the use of animals from those
same biotopes. The diet might have included primarily
terrestrial plants along with primarily marine animals.
Zooarchaeological evidence only addresses one
component of that diet, however. In order to evaluate
the way these resources were combined it is necessary
to have evidence of both plant and animal use.

90- 90-

80- 80

70- 70-

60- r 60
50- 50

30- 30-

20 20-

10- 10-

2.2 -2.6 3.1-3.3 3.4- 3.5 2.2 -2.6 3.1-3.3 3.4- 3.5
Trophic Levels Trophic Levels
% MNI % Biomass

Order of Stratigraphic Levels: 400 300 [ 200 100

Figure 5. Paloma probability samples, percentage of vertebrate MNI and biomass from three combinations of trophic levels in
archaeological levels 400 through 100.

REITZ: Resource Use through Time at Paloma, Peru

Fortunately, plant remains from Paloma have been
recovered and studied. Weir and Dering (1986;
Benfer 1990, 1999) found evidence of extensive use
of lomas plants in their study of coprolites. They
report columnar cactus (Loxanthocereus sp.), mito
(a papaya-like fruit, Carica candicans). as well as
chenopod (Chenopodiaceae) and composite
(Compositae) seeds in macrobotanical samples and
coprolites. Pearsall has also found tuberous Begonia
and amancay (non-showy spiderlilies, Hymenocallis
amancaes). Beans, gourds, and squash, too, are
present in some samples.
The Paloma data indicates heavy uses of marine
resources long before the Formative Period. Although
some changes occurred in the specific marine
resources deposited in the identified levels of Paloma,
terrestrial animals never contributed significant
percentages of biomass. The only evidence for use
of terrestrial animals, such as guanaco (Camelidae)
and deer (Cervidae), is found in the grab samples
(Reitz 1988b). The remains of an estimated three
guanaco and one deer were found (Table 11).
Primarily in Level 300, these do not provide evidence
for a gradually developing focus on marine resources
out of an earlier terrestrial hunting tradition. By the
time Paloma was occupied, use of marine animals
was already a major part of the subsistence strategy.
If a shift from terrestrial animals to marine animals
did occur, it might have happened either before Paloma
was settled or at another location. However, there is no
evidence of a subsistence strategy focused on terrestrial
animals at any of the early sites for which vertebrate
faunal data are available. In addition to dominating the
Paloma faunal assemblage, marine resources dominate
the assemblages from Quebrada Tacahuay, Quebrada
Jaguay, the Ring Site, Ostra Base Camp, and Almejas
(Fig. 1; Keefer et al. 1998; McInnis 1999; Reitz 1995,
2001; Reitz and Sandweiss 2001; Sandweiss et al. 1989:
Sandweiss et al. 1996; Sandweiss et al. 1998). These
are all early sites that either were occupied prior to
Paloma or were contemporaneous with Paloma. They
represent early subsistence from the northern to the
southern border of Peru (Fig. 1). They all support the
conclusion that use of marine resources in this area was
a phenomenon throughout the Holocene rather than a
strategy that developed out of a terrestrial hunting
The only major assemblage with terrestrial animals
is from a complex of sites located in the Chicama Valley,
representing an early hunting complex known as Paijin

Table 9. Paloma probablility samples, mean vertebrate trophic
Level MNI Biomass
100 2.99 3.13
200 2.59 2.76
300 2.63 2.82
400 2.54 2.78

(Chauchat 1988, 1992). This complex may have been in
existence from about 10,400 to 8200 B.P. during the early
Preceramic Period (Chauchat 1988:59; see also
Chauchat 1992:340). Vertebrate remains from 11 Pampa
de los F6siles samples, Quebrada de Cupisnique, and
Ascope are reported by Elizabeth S. Wing (1986,
1992). The closest of these sites to the present coastline
is 14 km and the furthest is 36 km. A fine-gauge screen
was used to recover these materials.
The abundant fish remains in the Paijin sites are
typical of warmer coastal conditions and not of
upwellings (Reitz 2001). Temperate fishes contribute
less than 1% of the individuals in the Paijin complex.
The temperate forms are herring (Clupeidae), anchovy
(Engraulidae, Anchoa sp.), and coco (Paralonchurus
sp.). The fine-gauge screen used during excavation
should have captured anchovy remains had they survived
in the archaeological record, but only two anchovy
individuals are represented in this assemblage. Most of
the Paijin vertebrates are terrestrial; 7% of the MNI
are terrestrial mammals, 74% are lizards and snakes,
3% are birds, and 16% are fishes. The terrestrial
mammals are primarily rodents; less than 1% of the MNI
were deer (Cervidae). The lizards are primarily cafiin
(Dicrodon sp.), which constitute 70% of the MNI.
Tropical fishes constitute 14% of the MNI and temperate/
Peru Current fishes constitute 0.5%. Sea catfish
(Ariidae) and lisa (Mugil sp.) are the most common
fishes, though bonefish (Albula vulpes) and croaker
(Micropogonias sp.) are present in some assemblages.

Table 10. Paloma probability samples, percentage of
vertebrate MNI and biomass from trophic level groups.
level MNI (%) Biomass (%)
100 200 300 400 100 200 300 400
2.2-2.6 38.5 71.4 70.4 74.1 27.5 59.2 60.0 57.7
3.1-3.3 3.7 3.7 2.4 8.9
3.4-3.5 61.5 28.6 25.9 22.2 72.5 40.8 37.7 33.4

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Table 11. Paloma grab samples: summary.

n %

Terrestrial mammals
Marine mammals
All vertebrates
Marine gastropods
Terrestrial gastropods
Other bivalves
All invertebrates




* Includes Ateles spp., Homo sapiens, and Felis concol

Small numbers of mojarra (Gerridae, Euc
and porgy (Sparidae) are also represent
distance of these sites from the coast, parti
the early Holocene when due to lowered
shoreline was farther west, the percent
resources is high, although the local terre
presented by the cafin lizard was clearly
the diet, but not what advocates of an
strategy had in mind as the terrestrial d
source. Amore likely Paijin "hunting com
the trapping of small animals.
It also seems unlikely that these
abandoned seasonally. Support for thi
Paloma is present in the grab samples, in w
contribute 6% of the individuals, birds 8
84% (Reitz 1988b). At least part of the
Paloma was in both the austral summ
winter. Evidence for this occupation
provided by the identification of juve
(Otariidae), an austral summer resident
and by the presence of such austral winter
as guanaco (Camelidae), deer (Cervidae
fulmar (Macronectes giganteus). Sea lio
common in the grab samples than are terr
or the fulmar, which may indicate that mo
in the village in the summer when sea lion
on the shore than in the winter when guai
formed part of the lomas community

1969:71, 72). The presence of guanacos and deer in the
grab samples, however, indicates that some people may
have remained in the village during the winter, using
kg % terrestrial vertebrates along with other lomas resources.
8.675 10.9 The prominence of sea lions in the grab samples (3% of
55.731 70.0 the vertebrate individuals and 65% of the biomass)
4.993 6.3 compared to guanaco and deer (1% of the individuals
5.361 6.7 and 11% of the biomass) suggests such seasonal shifts
0.632 0.8 in population. On the other hand, the ungulates may have
75.392 94.7 been taken from a location closer to the Chilca River
3.231 4.1 during the austral summer when the river supports
0.081 0.1 vegetation along its banks.
The Paloma grab samples also contain evidence of
0.587 0.7 long-distance contact in the form of a worked, proximal
0.319 0.4 femur from a spider monkey (Ateles spp.) in a grab
sample from the burial pit of a teenage male (Square
4.218 5.3 N120E65, Level 300, T159). Although ordinarily it would
be inadvisable to draw such a conclusion from a single
element, this may be an exception because spider
'or (Reitz 1988b). monkeys have specialized diets and it is unlikely that a
wild spider monkey population could live on the coastal
inostomus sp.) plain when Paloma was occupied. Their populations,
ed. Given the today found in the montaiia of eastern Peru (Eisenberg
cularly during and Redford 1999; Kellogg and Goldman 1944:12), also
I sea level the extend to the western side of the Andes near the equator.
age of coastal The femur from the burial pit suggests either direct or
strial resource indirect contact with one of those regions. Obsidian at
a mainstay in Paloma appears to derive from the highland area east of
early hunting Paloma, in the present-day Department of Huancavelica
dietary protein (Benfer 1999).
plex" includes The real question posed by Moseley's Maritime
Foundations Hypothesis, however, is not whether marine
sites were all resources were used, but whether they provided an
s argument at adequate nutritional base for such non-subsistence
which mammals activities as construction of monumental architecture.
8%, and fishes This cannot be demonstrated by a simple list of taxa
occupation at found at archaeological sites. To determine whether or
er and austral not maritime and wild terrestrial resources could provide
at Paloma is a base for complex social interactions without agricultural
nile sea lions input requires additional evidence, including tests of
on the coast, human skeletal materials for mineral levels, morphological
lomas visitors evidence of nutritional stress, and isotopic ratios. For
), and a giant this reason the human biological data reported by Benfer
is are far more (1984, 1990) are important.
estrial animals From Benfer's data it appears that a diet with mostly
re people lived marine-derived protein was adequate to support village
is were present life, although perhaps not overly so during the early time
nacos and deer periods (Benfer 1984, 1990). Benfer found not only that
(Grimwood, the population increased at Paloma between 7800 and

REITZ: Resource Use through Time at Paloma, Peru

4700 B.P., but that life expectancy increased for all age
levels except the oldest age categories over time. There
was no peak of childhood mortality associated with
weaning. Health also improved, as indicated by a reduction
of skeletal indicators of stress. Stature of both men and
women increased from Level 400 to Level 200. Some
men between 5100 and 4700 B.P. may have been as tall
as 170 cm. Benfer attributes this to a better diet, to
better health during childhood, or both (Benfer 1990).
There also was a strong trend for decreasing anemia.
Infection levels were stable. Data on trace elements
(zinc, strontium, and fluoride) support the hypothesis that
human consumption of animal protein was constant
throughout the occupation. Although the
zooarchaeological data indicate there was a change in
the protein sources used, there is no trend toward a more
protein-rich diet or toward increased consumption of
marine instead of terrestrial animals.
There are several indicators of activity changes over
time for men and women (Benfer 1990). Bony growths
(auditory exostoses) in the ear indicate that men may
have done most of the diving for deep-water mollusks.
A decrease in musculature of men occurred at the same
time that there was an increase in musculature of women
and a change in the humerus shape of both sexes. These
changes may reflect a change in subsistence-related
activities, especially activities which made greater use
of the upper body-such as hauling nets (Benfer 1990).

In reviewing not only the data from Paloma, but from
other sites along the Peruvian and Chilean coast, it no
longer seems appropriate to debate whether marine
resources were used. The time has come to apply the
best archaeological field and laboratory methods currently
available to explore the diversity of ways in which people
used marine resources at Paloma and other sites along
the Peruvian and Chilean coast. We need to obtain large
samples as unbiased as possible by recovery methods
and to combine evidence provided by palynology,
zooarchaeology, archaeobotany, and human biology. By
reconstructing the complex early Peruvian diet, we can
explore the diversity of subsistence strategies practiced
on the coast and also better understand the interaction
between coast and highlands.

I wish to thank Robert A. Benfer, Frederic-Andre Engel,
Glendon H. Weir, and Alice N. Benfer for the opportunity

to examine faunal materials from Paloma. Benfer and
Weir made extensive comments on the manuscript, which
are appreciated. The staff of the Centro de
Investigaciones de Zonas Aridas, particularly Carlos
L6pez Ocafia, provided invaluable assistance. I also
gratefully acknowledge the assistance of Kurt Auffenburg,
George Burgess, Diana Matthieson, Oliver P. Pearson,
Harold B. Rollins, Carlos Saavedra, Daniel H. Sandweiss,
Richard Thorington, and Jane Wheeler. All identifications
and analysis of the faunal materials were done at the
Florida Museum of Natural History, Gainesville. I
appreciate the generous assistance of Elizabeth S. Wing.
Funding was provided by National Science Foundation
grants BNS 76-12316, BNS 78-07727a/b, and BNS
81-053940, Robert A. Benfer, principal investigator.
Co-principal investigators were Frederic Andre Engel
(1976/77), Alice N. Benfer (1979), and Glendon H. Weir

Andrus, C F. T. 2000. Oxygen isotope analysis of ancient
otoliths: A mid-Holocene record of El Nifio variation and
human adaptation from Peruvian archaeological sites.
Ph.D. dissertation. University of Georgia, Athens.
Benfer, R. A. 1984. The challenges and rewards of sedentism:
The preceramic village of Paloma, Peru. Pp. 531-558 in M.
N. Cohen and G J. Armelagos, eds. Paleopathology and
the Origin of Agriculture. New York: Academic Press.
Benfer, R. A. 1990. The Preceramic Period site of Paloma, Peru:
Bioindications of improving adaptation to sedentism. Latin
American Antiquity 1(4): 284-318.
Benfer, R. A. 1999. Proyecto de excavaciones en Paloma, valle
de Chilca, Peru. Boletin de Arqueologia PUCP 3:213-237.
Chauchat, C. 1988. Early hunter-gatherers on the Peruvian
coast. Pp. 41-66 in R. Keatinge, ed. Peruvian Prehistory.
Cambridge: Cambridge University Press.
Chauchat, C. 1992. Pr6histoire de la C6te Nord du Pdrou: Le
Paijanien de Cupisnique. Paris: Centre National de la
Recherche Scientifique Cahiers du Quaternaire 18.
Chirichigno E, N. 1974. Clave para Identificar los Peces Marinos
del Peru. Callao, Peru: Instituto del Mar del Peri, Informe 44.
Chirichigno E, N. 1982. Catilogo de Especies Marinas de
Interds Econ6mico Actual o Potencial para America Latina:
Parte II, Pacifico Centro y Suroriental. Rome: FAO.
Eisenberg, J. E, and K. H. Redford. 1999. Mammals of the
Neotropics: The Central Neotropics, Vol. 3, Ecuador, Peru,
Bolivia, Brazil. Chicago: University of Chicago Press.
Engel, F A. 1980. Paloma: Village 613. Pp. 103-135 in E Engel, ed.
Prehistoric Andean Ecology. New York: Humanities Press.
Fowler, H. W. 1945. Los peces del Peru: Catilogo sistemitico
de los peces que habitan en aguas peruanas. Lima, Peru:
Museo de Historia Natural "Javier Prado," Universidad

Nacional Mayor de San Marcos.
Froese, R., and D. Pauly. 1998. FishBase. [Online] Available:
Grayson, D. K. 1973. On the methodology of faunal analysis.
American Antiquity 38(4): 432-438.
Grayson, D. K. 1981. A critical view of the use of archaeological
vertebrates in paleoenvironmental reconstruction. Journal
of Ethnobiology 1(1): 28-38.
Grimwood, I. R. 1969. Notes on the distribution and status of
some Peruvian mammals. Bronx, New York: American
Committee for International Wildlife Protection Special
Publication no. 21.
Hildebrand, S. F. 1945. A Descriptive Catalog of the Shore
Fishes of Peru. Washington, D.C.: United States National
Museum Bulletin no. 189.
Keefer, D. K., S. D. deFrance, M. E. Moseley, J. B. Richardson
II, D. R. Satterlee, andA. Day-Lewis. 1998. Early maritime
economy and El Nifio events at Quebrada Tacahuay, Peru.
Science 281: 1833-1835.
Kellogg, R., and E. A. Goldman. 1944. Review of the spider
monkeys. Proceedings of the United States National
Museum 96 (3186): 1-45.
McInnis, H. E. 1999. Subsistence and maritime adaptations at
Quebrada Jaguay, Camand, Peru: A faunal analysis. M. A.
thesis. University of Maine, Orono.
Moseley, M. E. 1975. The Maritime Foundations of Andean
Civilization. Menlo Park, California: Cummings.
Pauly, D., and V. Christensen. 1995. Primary production required
to sustain global fisheries. Nature 374: 255-257.
Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and E Torres, Jr.
1998. Fishing down marine food webs. Science 279: 860-863.
Pauly, D., V. Christensen, R. Froese, and M. L. Palomares.
2000. Fishing down aquatic food webs. American Scientist
Pielou, E. C. 1966. Species-diversity and pattern-diversity in
the study of ecological succession. Journal of Theoretical
Biology 10:370-383.
Quilter, J., and T. Stocker 1983. Subsistence economies and
the origins of Andean complex societies. American
Anthropologist 85(3): 545-562.
Reitz, E. J. 1988a. Faunal remains from Paloma, an Archaic site
in Peru. American Anthropologist 90(2): 310-322.
Reitz, E. J. 1988b. Preceramic animal use on the central coast.
Pp. 31-55 in E. S. Wing and J. C. Wheeler. Economic
Prehistory of the Central Andes. Oxford: BAR International
Series 427.
Reitz, E. J. 1995. Environmental change atAlmejas, Peru. Ms.
on file, Zooarchaeology Laboratory, Georgia Museum of
Natural History, University of Georgia, Athens, Georgia.
Reitz, E. J. 2001. Fishing in Peru between 10,000 and 3750 B.P.

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

International Journal of Osteoarchaeology 11:163-171.
Reitz, E. J., and D. H. Sandweiss. 2001. Environmental change
at Ostra Base Camp, A Peruvian Preceramic site. Journal
of Archaeological Science 28(10): 1085-1100.
Reitz, E. J., and E. S. Wing. 1999. Zooarchaeology. Cambridge:
Cambridge University Press.
Sandweiss, D. H., H. McInnis, R. L. Burger, A. Cano, B. Ojeda,
R. Paredes, M. Sandweiss, and M. Glascock. 1998.
Quebrada Jaguay: Early maritime adaptations in South
America. Science 281: 1830-1832.
Sandweiss, D. H., J. B. Richardson III, E. J. Reitz, J. T. Hsu, and
R. A. Feldman. 1989. Early maritime adaptations in the
Andes: Preliminary studies at the Ring Site, Peru. Pp. 35-
84 in D. S. Rice, C. Stanish and P. R. Scarr, eds. Ecology,
Settlement and History in the Osmore Drainage, Peru.
Oxford: BAR International Series 545(I).
Sandweiss, D. H., J. B. Richardson III, E. J. Reitz, H. B. Rollins,
and K. A. Maasch. 1996. Geoarchaeological evidence from
Peru from a 5000 years B.P. onset of El Nifo. Science 273:
Schweigger, E. 1964. El Litoral Peruano. Lima, Peru: Universidad
Nacional "Federico Villarreal."
Shannon, C. E., and W. Weaver. 1949. The Mathematical Theory
of Communication. Urbana: University of Illinois Press.
Sheldon, A. L. 1969. Equitability indices: Dependence on the
species count. Ecology 50: 466-467.
Simpson, G G, A. Roe, and R. C. Lewontin. 1960. Quantitative
Zoology. New York: Harcourt, Brace, and Co.
Weir, G. H., R. A. Benfer, and J. G. Jones. 1988. Preceramic to
early Formative subsistence on the central coast.
Preceramic animal use on the central coast. Pp. 56-94 in E.
S. Wing and J. C. Wheeler. Economic Prehistory of the
Central Andes. Oxford: BAR International Series 427.
Weir, G, and P. Dering. 1986. The lomas of Paloma: Human-
environment relations in a central Peruvian fog oasis:
Archaeobotany and palynology. Pp. 18-44 in R. Matos
M., S. A. Turpin, and H. H. Eling, eds. Andean
Archaeology. Los Angeles: UCLA Institute of
Archaeology Monographs in Archaeology no. 27.
Wing, E. S. 1986. Methods employed in the identification and
analysis of the vertebrate remains associated with sites
of the Paijdn Culture. Ms. on file, Environmental
Archaeology Laboratory, Florida Museum of Natural
History, University of Florida, Gainesville.
Wing, E. S. 1992. Les restes de vertdbr6s. Pp. 42-47 in C.
Chauchat, ed. Prdhistoire de la C6te Nord du P6rou: Le
Paijanien de Cupisnique. Paris: Centre National de la
Recherche Scientifique Cahiers du Quaternaire 18.
Wing, E. S., and A. Brown. 1979. Paleonutrition: Method and
Theory in Prehistoric Foodways. New York: Academic Press.

Bull. Fla. Mus. Nat. Hist. (2003) 44(1): 81-90


Laura Kozuch'

This paper reports on the evidence for heat-treatment of marine shell in bead manufacture at the Cahokia site in Illinois. Of burned
shell fragments, high percentages of burned columellas were found, suggesting that columellas were targeted for heat treatment.
Additionally, the presence of all shell elements reveals that whole lightning whelk shells (Busycon sinistrum) were transported to
Cahokia for artifact manufacture, probably after being de-fleshed. A columella bead-working reduction sequence is presented.

Key words: bead, Cahokia, fire, shell, trade

Shell beads from archaeological sites generally come in
three forms: (1) disk beads made from bivalves or the outer
whorl of a gastropod shell; (2) whole shell beads made
from such small gastropod shells as marginella or dwarf
olive shells that have had the apex ground off; and (3)
beads made from the columella of large marine gastropod
shells. The third form is the focus of this paper.
Beads made from the columellas of marine shells take
many forms and have been termed massive, tubular, barrel,
and a host of other cylindrical and spherical terms (Brown
1996; Holmes 1883:223; Moore 1905:154: Ottesen
1979:377). I prefer to call these types of beads columella
beads, since this term indicates the portion of the shell from
which beads were made. The drill hole in these larger beads
runs parallel to the axis of the shell columella.
Columella beads have been recovered in great
numbers from Mississippian archaeological sites
(Fig. 1). At least 43,277 columella beads made from
marine shell were found at Spiro (Brown 1996:283).
More than 30,740 columella beads were excavated from
Mound 72 at Cahokia (Fowler et al. 1999:136). Baker
(1932) and Moore (1905:154) each wrote they had found
"many" columella shell beads in Mound C at Moundville,
although I was only able to locate 23. At least 91
columella beads were recovered from the Etowah site
in Georgia (Kozuch 1998).
Most (88%) columella beads from Cahokia, Etowah,
Moundville, and Spiro were made from sinistral shells,
and the remaining 12% were made from dextral shells
(Kozuch 1998), despite the fact that the overwhelming
majority of available gastropod shells are dextral.
'Curator of Archaeology, University of Illinois, ITARP, 209
Nuclear Physics Laboratory, 23 East Stadium Drive, Champaign,
IL 61820, USA.

Additionally, the largest shell in the Atlantic Ocean and
Gulf of Mexico is the horse conch (Pleuroploca
gigantea), yet columella beads were not commonly
made from this shell. Sinistral columella beads excavated
from Cahokia, Moundville, and Spiro were far more
abundant (92-94%) than dextral columella shell beads.
The columella beads from the Etowah site were fairly
evenly divided between sinistral (51%) and dextral (49%)
shells. The Etowah ratios may be a result of the small
sample size of 23.
Drilling. The production of chert microdrills for
drilling holes to make shell beads has been most
extensively studied by Richard W. Yerkes (1983, 1991,
1993), although many others have commented on
microdrills from Mississippian sites (Koldehoff and
Kearns 1993; Prentice 1983; Trubitt 1995). Yerkes
approached the subject of craft specialization using the
production of chert microdrills as evidence for shell bead
manufacture. The process of making chert microdrills
involves breaking chert nodules into smaller pieces to be
worked into microdrills or microblades (Yerkes 1991).
Yerkes (1983) used experimental archaeology and
incident light microscopy (200x) to confirm that chert
microdrills, rather than bone or wood, were used to drill
shell. The chert microdrills Yerkes replicated and used
to drill shell had use-wear patterns that are very similar,
if not identical, to archaeological specimens.
At the Cahokia site in an area known as Ramey Field,
Mason and Perino (1961) found microdrills in association
with shell working debitage. They found "thousands of shell
beads and enormous quantities of burned conch columellas
and shell scrap" (p. 554). Microdrills have also been found
at Cahokia from Powell Mound, the Kunneman Mounds,
and the Dunham Tract (Yerkes 1991).

Figure 1. Major Mississippian sites.

Shell bead production areas, inferred by
concentrations of microdrills, have been identified at
Cahokia and surrounding sites in the American Bottom
(Koldehoff and Kearns 1993; Mason and Perino 1961;
Trubitt 1996; Yerkes 1991), an area defined as part of
the Mississippi River Valley bounded by the mouth of
the Illinois River and the mouth of the Kaskaskia River
(Fowler 1969:1-5). Other Mississippian sites that have
microdrills are Moundville (Peebles and Kus 1977:442),
sites in central Tennessee, northern Alabama, western
North Carolina, and northeastern Georgia, and the
Zebree site in northeast Arkansas (Yerkes 1993:237).
Because chert microdrills have not been found in contexts
dating to earlier time periods, they are considered a
Mississippian phenomenon (Yerkes 1993:240).
Additionally, microdrills have been found at only 10% of
the sites in the American Bottom (Yerkes 1991:58).
Sectioning. The method used to cut the thick,
durable shell columellas into sections is uncertain.
Researchers have assumed that some type of lithic tool,
such as a sandstone saw or chert cutting tool, was used.
The groove-and-snap technique, whereby the columella
was cut around the outside and then snapped, was likely
used (Pauketat 1993; Trubitt 1995).

The Cahokia site has provided the materials for a
wealth of archaeological research on the development
of complex societies, population estimates, exchange

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

of exotic goods, and craft specialization. It has at
least 120 mounds (Iseminger 1996:32), including one
that covers about 8 square km (5 square miles), the
largest mound north of the Mexican site of
Teotihuacin. The mounds at the Cahokia site are
arranged around many plazas (Fowler 1989:11).
Monks mound is about 30.5 m (100 feet) high and
covers an area of land 316.1 by 240.8 m (1037 by
790 feet) (Reed 1969). In use since Paleoindian times
(8,000 B.C.), Cahokia was continuously occupied
through the Mississippian period. Population estimates
of Cahokia at its zenith, about A.D. 1100, range from
8,000 to 40,000 (Fowler 1989:7; Iseminger 1996), the
most cited figure being that of 30,000 to 40,000
residents (Milner 1990:11). Cahokia was abandoned
about A.D. 1400.
Generic identification of columella beads. It is
often possible to identify the marine snail genus used for
the manufacture of columella beads. Holmes (1883:223)
was the first to note that the beads "often retain the
[characteristic] spiral groove as well as other portions
of the natural surface." This columella (or spiral) groove
slants toward the left or right, depending on the sinistral
or dextral shell, respectively, from which it came. The
closest readily recognizable analogy is either the "back
slash" or "forward slash" of today's computer keyboards.
Applying this analogy, sinistral shells have columella
grooves with a back slash, while dextral shells have
columella grooves with forward slash. Some columella
beads retain the columella groove (Fig. 2). The direction of
the slant, important in identification of a sinistral or dextral
shell, does not change when the bead is rotated 180.
The only sinistral shells in the Atlantic Ocean or
Gulf of Mexico from which large columella beads can
be made are from the genus Busycon. These are the
snow whelk (B. laeostomum [Kent, 1982]), lightning
whelk (B. sinistrum [Hollister, 1958]), and prickly
whelk (B. pulleyi [Hollister, 1958]). These
gastropods, capable of interbreeding, are recognized
as distinct species based on their geographic separation
(known as allopatric speciation). Changing zoological
nomenclature may have acted as an impediment to some
archaeologists trying to identify shell artifacts to the
species level. Recent efforts by a team investigating
Busycon genetics supports the position that all sinistral
Busycon be relegated to one species (Wise et al. 2002).
The sinistral Busycon shells from which beads
were made probably originated on the west coast of
Florida (Hale 1976; Kozuch 1998). Since good quality
chert to work the shells is lacking in Florida (Kozuch

KOZUCH: Use of Fire in Shell Bead Manufacture at Cahokia

1993), it would have been difficult for Florida
inhabitants to make the beads themselves. In the
Atlantic Ocean or Gulf of Mexico, the only dextral
shells large enough to be suitable for making larger
columella beads are Charonia tritonis (trumpet
triton), Busycon carica knobbedd whelk), Strombus
spp. (pink or milk conch), and Pleuroploca gigantea
(horse conch).

I examined 8,333 marine shell specimens from the
Cahokia site (Kozuch 1998) and found 14 species,
representing 12 univalve and two bivalve species.
Ninety-three percent (n = 7,737) of specimens were
small, whole shell beads made from marine snail
shells, usually olive shells (genus Oliva) or marginellid
shells. This large proportion of small, whole shell
beads suggests a bias in the data, because most disk
beads cannot be identified even to the genus level.
Tens of thousands of disk shell beads were excavated
from Cahokia, and it can only be assumed that these
were made from large marine whelk shells, probably
sinistral Busycon shells. A total of 372 specimens of
sinistral whelk shell of the genus Busycon were found.
Fragments were found from all parts of sinistral
Busycon shells, indicating that whole sinistral whelk
shells were brought to the site for bead manufacture
(Kozuch 1998).
Most columella beads had no remaining columella
grooves, so only those that have a columella groove
clearly slanting to the right or left are included in my
sample. The few specimens with the groove parallel to
the drill hole were not included. Ninety-three percent of
the finished beads from Cahokia were made from
sinistral whelk columellas, but a few were made from
dextral columellas (Fig. 3). Unfinished columella beads
that show evidence of being worked (either cut or
scored) were labeled bead blanks. Columellas without
direct evidence of being worked were classified as
debitage. No bead blanks or debitage fragments had drill
holes. I found evidence for a columella bead-reduction
sequence, starting with an unworked (sometimes burned)
columella, and ending with a finished columella bead (see
figs. 4 to 6).
Burned specimens. The total of 172 burned
specimens from Cahokia include beads, bead blanks, and
debitage fragments (Appendix 1). A majority of these
(125) were whole shell beads made from olive shells
(Oliva spp.) with burned portions. These olive shell beads
were from Powell Mound #2.

Figure 2. Columella beads from sinistral Busycon shell (after
Holmes 1883).

The remaining 47 fragments were from sinistral
whelk shells, of which 91% (43) were burned columellas,
6% (3) were inner whorl fragments, and 2% (1) an apex
fragment. These burned specimens were from three
different contexts at the Cahokia site: (1) Kunneman
Mound, (2) Ramey Field/Mound 34, and (3) Wilson
Mound. The large percentage of burned columellas at
Cahokia suggests that fire was used to treat them before
fashioning them into beads.
Archaeological contexts of burned specimens. In
1956, Gregory Perino excavated portions of an area east
of Monks Mound known as the Ramey Field and Mound
34 inside Ramey Field (Fowler 1989; Mason and Perino
1961; Perino 1959). Mound 34, a conical mound east of
Monks Mound, has been mostly destroyed (Fowler
1989:88). Perino excavated a trench through Mound 34,
as well as a refuse pit north of the mound, and found
pottery from both the Caddo area and the lower
Mississippi River Valley (Perino 1959). Lithics from
Arkansas also were recovered (Fowler 1989:88; Perino
1959). Brain and Phillips (1996:267) pointed out that all
artifacts at Cahokia that relate to the Southeastern
Ceremonial Complex are associated with Ramey Field



ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Figure 3. Finished columella beads from Mound 72, Cahokia. Top row: sinistral beads; bottom row: dextral beads.

and Mound 34. Recent work by James A. Brown and
John E. Kelly (2000) using information from engraved
motifs from marine shells and ceramics shows that
Ramey Field dates to the Moorehead phase, dated to
between A.D. 1150 and 1250.
Powell Mound, also destroyed, is 2.6 km west of
Monks Mound, about 13 m high, and 94.5 by 54.9 m at
the base (Ahler and DePuydt 1987). Powell Mound was
constructed in at least two stages. The first was a
truncated pyramid mound; the second, a modification of
the first, has the final appearance of a rounded linear
ridge-top mound (Ahler and DePuydt 1987:3-4). Two
burial pits were found between these two mound-building
stages. One burial pit had already been destroyed by
the time Kelley arrived, the other contained remains of
20 to 30 individuals accompanied by "a very large number
of disc shell or Marginella [whole shell] beads, but in no
instance were both types of beads found with the same
burial" (Titterington 1977:2). Ceramics found at Powell
Mound indicate construction between A.D. 900 and 1150
(Ahler and DePuydt 1987). In addition to much locally
produced pottery, exotic ceramics were found. Of the
latter, one sherd was tentatively identified as originating

from the Caddo area, and two more (Nodena White-
filmed and Red-and-white Painted) came from the lower
Mississippi River Valley (Ahler and DePuydt 1987:23).
Chert microdrills at the base of the mound indicate a
shell artifact workshop (Yerkes 1989:97), probably of
the Lohmann phase, about A.D. 1000-1050.
The Kunnemann Mound, 10.64 m high, is located in
the Kunnemann Tract (Fowler 1989) about 1.5 km north
of Monks Mound and across the channel of Cahokia
Creek (Pauketat 1993:8). After Moorehead's
excavations in 1921, Preston Holder partially excavated
the Kunnemann Mound (#10 and 11) in 1955 and 1956.
A publication from Holder's notes was prepared by Tim
Pauketat (Pauketat 1993). The Kunnemann Mound had
two parts, a conical mound conjoined on top of a lower
terrace. In addition to hearth features and human burials,
structural remains, such as post pits and wall trenches,
were recovered. The mound dates from A.D. 1000 to
1200 (Pauketat 1993:5). Carbonized fabric was
recovered, and the quantity of charcoal suggests that a
building on the site had burned down (Pauketat 1993:43).
Sixty-eight chert microblades, a sandstone saw, and
unfinished marine shell beads are evidence of a shell

KOZUCH: Use of Fire in Shell Bead Manufacture at Cahokia

bead workshop. The probable context of the burned shell
specimens from Kunnemann Mound dates to the
Lohmann phase, A.D. 1000 to 1050 (Pauketat 1993:56).
There is scant information on the Wilson Mound,
excavated by Preston Holder, although secondary burials
were found (Milner 1984:480). The archaeological phase
to which Wilson Mound can be assigned is uncertain,
but, based on pottery typologies, it probably dates to the
Lohmann phase (Milner 1998:130).

To understand methods of working large gastropod shells
without electric-powered metal tools, I made four shells
cups. I purchased lightning whelk shells that presumably
came from Florida's west coast. The first one I made
was from a small shell about 18 cm long. The second
was from a slightly larger shell, about 22 cm long, but
very gracile and thin-shelled. The third was from a very
large and gracile shell about 37 cm long. The last was
from a shell about 26 cm long. Because I only had modem
tools, I used a small metal ball-peen hammer to remove
unwanted portions of the outer and inner whorls. This
was a substitute for the lithic hammers available to
Mississippian peoples.
Removal of the outer and inner whorls was relatively
easy. I removed the entire columella from all but the
largest specimen mentioned above. After the whorls were
taken out, however, removal of the columella required
more precise hammering. With the smaller specimens, I
was able to remove all of the columella by hammering.
Employing hammering on the very large shell. I found I
could only remove the columella by breaking the outer
whorl and ruining the cup.
A crafts instructor at the University of Florida (Ray
Ferguson, per. comm.) suggested that I use fire to help
make the shell more friable so that it could be more
easily worked. Shell that is being exposed to heat releases
an odor not unlike burning hair. After the organic
constituents have been oxidized, the shell is more brittle
and chalky. I followed the suggestion to use a blowtorch.
Although Mississippian peoples did not have such devices,
they did have means of using directed fire, such as burning
pitchpine or small torches. The blowtorch worked well,
as long as the heat was not too intense. If the heat was
kept focused on one spot for too long, the shell began
spelling off, threatening to explode. This made it clear
that the application of low, even heat was necessary.
Consequently, I directed the small flame evenly over the
columella area that I wanted to break. After about an

.-s m
Figure 4. Unworked, sometimes burned, beads.

Figure 5. Transitional beads.

Figure 6. Finished beads.

Figure 7. Example of heat-treated Busycon sinistrum columella.

hour of heating the desired area on the shell, and after
allowing time for the shell to cool a little, I was then able
to hammer the burned columella off the rest of the shell
without breaking the shell into pieces, thereby
demonstrating that the heat treatment had made the shell
much easier to work. The resulting burned columella,
pictured in Fig. 7, is 145 mm in length.

Shell is a durable substance that lends itself well to
archaeological studies. Shells are exoskeletons of
mollusks, and are made of calcium carbonate crystals in
an organic protein matrix (Vermeij 1993:39). Claassen
comments on the structure of shell after burning: "Heating
shell physically alters the crystallography and
compromises the internal cohesion of the structure.
Burned shell fractures more easily and weighs less than
does unburned shell" (Claassen 1998:61).
Fire can be used as an aid to working shell. High
heat burns off the organic constituents of shell, leaving
behind a material with a higher percentage of inorganic
material and an altered crystallography (Claassen 1998).
The large proportion of burned columellas indicates that
Cahokians used fire to help separate the columella from

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

the rest of the shell. My own attempts to remove a
columella from a large lightning whelk shell supports the
importance of heat in shell-working technology. The
resulting burned columella from my experiment looks
remarkably similar to the archaeological specimens
(compare figs. 4 and 7). After heat treatment, the shell
material is more chalky and soft and much easier to
work into some type of artifact.
Ethnohistoric sources lend credence to the suggestion
that heat was used as a tool in shell artifact manufacture.
Garcilaso de la Vega (1988:330 [1605]) mentioned in his
account of DeSoto's expedition that, in the town of
Cofachiqui (in Georgia), pearls were pierced using hot
copper needles, supporting the notion that heat was a
known method of working shell materials in the
Southeast. Another early record (among the Natchez in
Louisiana) from Dumont de Montigny (in Swanton
1946:486), stated that shell gorgets were pierced by
means of fire. More evidence comes from the
archaeological record from the Channel Islands where
"bead-makers heated [dwarf olive shell] bead blanks"
(Arnold 1997).
Because two contexts in which burned columellas
were found have evidence of structural fires, it is not
conclusive that columellas were heat-treated. Those two
sites, Powell and Kunnemann mounds, had clear
evidence of burned buildings. There was, however, no
evidence of fire at Wilson Mound, Mound 34, or Ramey
Field where columellas were gathered. Clearly, more
precise contextual information is required.
At Cahokia, shell debitage was found, including
spire, apex, columella, and outer whorl (Kozuch 1998).
Most columella beads are from sinistral shells, probably
from lightning whelk (B. sinistrum), that were
transported minus the flesh to Cahokia for artifact
manufacture. Some mistakes made during the drilling of
columella beads were fixed by filling the faulty drill-holes
with fitted shell plugs (Milner 1998:141). These repairs
show that special care was taken when making
columella beads.
The time span during which columellas were burned
is not long, from A.D. 1000 to 1250. Additionally, burned
columellas appear during the second half of the
Mississippian period, from A.D. 800 to 1450. Considering
the time-depth of whelk shell-working practices in North
America (Carstens and Watson 1996), evidence for heat-
treatment of shell shows up very late.
There may not be any visible evidence from a
finished columella bead that the columella from which it

KOZUCH: Use of Fire in Shell Bead Manufacture at Cahokia

was fashioned had been heat-treated. The outer burned
portion was probably ground off, thereby erasing traces
of burning. Also, the organic constituents in shell beads
naturally degrade while in the ground, making it hard to
tell if it has been burned. More studies are needed to
determine if burned shell can be distinguished from shell
that has disintegrated due to acidic soil conditions or
natural oxidation.
The topic of craft specialization in marine-shell bead
production has relevance to the question of social
complexity among Mississippian cultures (Muller 1987;
Pauketat 1987; Prentice 1983), and the socio-political
aspects of Cahokia as a bead-production center are still
being explored. At Cahokia, production of shell beads
was emphasized almost to the exclusion of cups, gorgets,
and other shell artifacts (Kozuch 1998). Among
Mississippian archaeological sites, this emphasis at
Cahokia on beads stands out as somewhat anomalous.
Some researchers believe that bead making was a
centralized activity; others think that shell artifact
manufacture was not centralized and was performed at
the household level (see Milner 1998; Pauketat 1993;
Prentice 1983; Trubitt 1995; Yerkes 1983). If shell beads
at Cahokia were intensely sought after, then some type
of control may have existed over the distribution, though
perhaps not the production, of such a desirable product.
Many questions remain. For instance, how were
columella beads that reached over 2 cm in length drilled?
Some researchers have suggested that long hardwood
sticks or porcupine quills were used in combination with
sand to make the drill hole, but these items would be
very difficult, if not impossible, to detect in the
archaeological record. Columella bead replication studies
also need to be performed to compare the amount of
time it takes to manufacture a bead from burned and
from unburned columellas. Presently I am in the process
of replicating columella beads, and it is hoped that the
results of this work will shed light on the lost methods of
columella bead manufacture.

This research was done in conjunction with my doctoral
dissertation work, which was supported by the Charles
H. Fairbanks Award in Archaeology, Dissertation
Scholarship, May 1996. Photos of Cahokia artifacts are
courtesy of the Illinois State Museum. John E. Kelly
graciously provided information on the Wilson Mound.
The constructive comments of Charlotte M. Porter and
an anonymous reviewer are greatly appreciated, as is

Margaret Joyner's editing. Terrence Martin at the Illinois
State Museum greatly facilitated my researches there
during my dissertation work. My work continues on
columella bead replication, and I am indebted to Brad
Koldehoff and Larry Kinsella for their contribution of a
shell-working tool kit, including a sandstone saw, chert
cutting tools, and chert microdrills.

Ahler, Steven R., and Peter J. DePuydt. 1987. A Report on the
1931 Powell Mound Excavations, Madison County, Illinois.
Springfield: Illinois State Museum, Reports of
Investigations, no. 43.
Arnold, Jeanne. 1997. Household Archaeology on the Channel
Islands. Backdirt Spring/Summer [Online]. Available: http:/
Baker, Frank C. 1932. Molluscan shells from Etowah Mounds.
Pp. 145-149 in W. K. Moorehead, ed. Etowah Papers. New
Haven, Connecticut: Philips Academy/Yale University Press.
Brain, Jeffrey P., and Philip Phillips. 1996. Shell Gorgets: Styles
of the Late Prehistoric and Protohistoric Southeast. 2 vols.
Cambridge, Massachusetts: Peabody Museum Press.
Brown, James A. 1996. The Spiro ceremonial center: The
archaeology of Arkansas Valley Caddoan Culture in
eastern Oklahoma. Ann Arbor: University of Michigan,
Memoirs of the Museum of Anthropology no. 29.
Brown, James A., and John E. Kelly. 2000. Cahokia and the
Southeastern Ceremonial Complex. Pp. 469-510 in S. R.
Ahler, ed. Mounds, Modoc, and Mesoamerica: Papers in
Honor of Melvin L. Fowler. Springfield: Illinois State
Museum Scientific Papers 28.
Carstens, Kenneth C., and Patty Jo Watson, eds. 1996. Of
Caves and Shell Mounds. Tuscaloosa: University of
Alabama Press.
Claassen, Cheryl. 1998. Shells. Cambridge: Cambridge
University Press.
de la Vega, Garcilaso. 1988 [1605]. The Florida of the Inca.
Translated and edited by J. G. Varner and J. J. Varner,
Austin: University of Texas Press.
Fowler, Melvin L. 1969. The Cahokia Site. Pp. 1-30 in M. L.
Fowler, ed. Explorations Into CahokiaArchaeology. Urbana:
Illinois Archaeological Survey, Inc., Bulletin no. 7.
Fowler, Melvin L. 1989. The Cahokia Atlas: A Historical
Atlas of Cahokia Archaeology. Studies in Illinois
Archaeology, no. 6. Springfield: Illinois Historic
Preservation Agency.
Fowler, Melvin L. 1991. Mound 72 and Early Mississippian at
Cahokia. Pp. 1-26 in J. B. Stoltman, ed. New Perspectives
on Cahokia: Views from the Periphery. Monographs in
World Archaeology no. 2. Madison, Wisconsin:
Prehistory Press.
Fowler, Melvin L, Jerome Rose, Barbara Vander Leest, and

Steven R. Ahler. 1999. The Mound 72 Area: Dedicated
and Sacred Space in Early Cahokia. Springfield: Illinois
State Museum, Reports of Investigation, no. 54.
Hale, H. Stephen. 1976. Marine Shells in Midwestern
Archaeological Sites and the Determination of Their Most
Probable Source. Unpublished M.A. thesis. College of
Social Science, Florida Atlantic University, Boca Raton.
Hollister, S. C. 1958. A review of the genus Busycon and its
allies. Palaeontographica Americana 4: 49-126.
Holmes, William H. 1883. Art in shell of the ancient Americans.
Second Annual Report of the Bureau of Ethnology, 1880-
1881, pp. 179-305. Washington, D.C.: Bureau of
Iseminger, William R. 1996. Mighty Cahokia. Archaeology,
May/June: 30-37.
Kent, Bretton W. 1982. Patterns of coexistence in busyconine
whelks. Journal of Experimental Marine Biology and
Ecology 66: 257-283.
Koldehoff, Brad, and Ronald Kearns. 1993. Using stone tools
to detect Mississippian homesteads. Illinois Antiquity
28(3): 4-8.
Kozuch, Laura. 1993. Sharks and Shark Products in South Florida.
Monograph 2. Gainesville: University of Florida, Institute
for Archaeology and Palaeoenvironmental Studies.
Kozuch, Laura. 1998. Marine Shells from Mississippian Sites.
Unpublished Ph.D. dissertation. Gainesville: University
of Florida, Department of Anthropology.
Mason, Ronald J., and Gregory Perino. 1961. Microblades at
Cahokia, Illinois. American Antiquity 26(4): 553-557.
Milner, George R. 1984. Social and temporal implications of
variation among American bottom Mississippian
cemeteries. American Antiquity 49(3): 468-488.
Milner, George R. 1990. The Late Prehistoric Cahokia cultural
system of the Mississippi River Valley: Foundations,
Florescence, and Fragmentation. Journal of World
Prehistory 4: 1-43.
Milner, George R. 1998. The Cahokia Chiefdom: The
Archaeology of a Mississippian Society. Washington,
D.C.: Smithsonian Institution Press.
Moore, Clarence B. 1905. Certain aboriginal remains of the
Black Warrior River. Journal of the Academy of Natural
Sciences of Philadelphia 13: 124-244.
Muller, Jon. 1987. Salt, chert, and shell: Mississippian exchange
and economy. Pp. 10-21 in E. Brumfiel and T. Earle, eds.
Specialization, Exchange, and Complex Societies.
Cambridge: Cambridge University Press.
Ottesen, Ann I. 1979. A Preliminary Study of Acquisition of
Exotic Raw Materials by Late Woodland and
Mississippian Groups. Ph.D. dissertation. New York: New
York University, Department of Anthropology.
Pauketat, Timothy R. 1987. Mississippian domestic economy
and formation processes: A response to Prentice.
Midcontinental Journal of Archaeology 12(1): 77-88.
Pauketat, Timothy R. 1993. Temples for Cahokia Lords: Preston

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Holder's 1955-1956 Excavations of Kunnemann Mound.
Ann Arbor: University of Michigan, Memoirs of the
Museum of Anthropology no. 26.
Peebles, Christopher S., and Susan Kus. 1977. Some
archaeological correlates of ranked societies. American
Antiquity 42(3): 421-448.
Perino, Gregory. 1959. Recent information from Cahokia and
its satellites. Central States Archaeological Journal 6(4):
Prentice, Guy. 1983. Cottage industries: Concepts and
implications. Midcontinental Journal of Archaeology 8(1):
Reed, Nelson A. 1969. Monks and other Mississippian mounds.
Pp. 31-42 in M. L. Fowler, ed. Explorations into Cahokia
Archaeology. Urbana: Illinois Archaeological Survey, Inc.,
Bulletin no. 7.
Swanton, John R. 1946. The Indians of the Southeastern United
States. Washington, D.C.: Smithsonian Institution, Bureau
of American Ethnology, Bulletin no. 137.
Titterington, Paul F. 1977 [1938]. The Cahokia Mound Group
and Its Village Site Materials. Cahokia, Illinois: Cahokia
Mounds Museum.
Trubitt, Mary Beth D. 1995. Marine shell ornament production
at Cahokia. Paper presented at the 60th Annual Meeting,
Society for American Archaeology, Minneapolis,
Trubitt, Mary Beth D. 1996. Household Status, Marine Shell
Bead Production, and the Development of Cahokia in the
Mississippian Period. Ph.D. dissertation, Department of
Anthropology, Northwestern University, Evanston, Illinois.
Vermeij, Geerat J. 1993. A Natural History of Shells. Princeton,
New Jersey: Princeton University Press.
Wise, John, M.G. Harasewych, and Robert T. Dillon. 2002.
COI, allozyme, and morphological survey of the sinistral
Busycon of North America. Paper presented at the 68th
annual meeting of the American Malacological Society,
Charleston, South Carolina, August 3-7, 2002.
Yerkes, Richard W. 1983. Microwear, microdrills, and
Mississippian craft specialization. American Antiquity
Yerkes, Richard W. 1987. Prehistoric life on the Mississippi
flood plain: Stone tool use, settlement organization, and
subsistence practices at the Labras Lake Site, Illinois.
Chicago: University of Chicago Press.
Yerkes, Richard W. 1989. Mississippian craft specialization on the
American Bottom. SoutheasternArchaeology 8(2): 93-106.
Yerkes, Richard W. 1991. Specialization in shell artifact production
at Cahokia. Pp. 49-64 in J. B. Stoltman, ed. New Perspectives
on Cahokia. Monographs in World Archaeology no. 2.
Madison, Wisconsin: Prehistory Press.
Yerkes, Richard W. 1993. Methods of manufacturing shell
beads at prehistoric Mississippian sites in southeastern
North America.Liege, Belgium: Colloque International de
Liege, Editions ERAUL, vol. 50, pp. 235-242.


ISM Illinois State Museum, Springfield, IL
LOA/UI Lab of Anthropology, University of Illinois, Urbana-Champaign, IL
NMNH National Museum of Natural History, Washington, DC

Artifact Length
Provenience Taxon # Type (mm) Comments Elements

Powell Mound #2
Powell Mound #2
Powell Mound #2
Powell Mound #2
Powell Mound #2
Powell Mound #2
Powell Mound #2
Powell Mound #2
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Kunnemann Mound
Ramey Field, Mound 34
Ramey Field, Mound 34
Ramey Field, Mound 34
Ramey Field, Mound 34
Ramey Field
Ramey Field
Ramey Field
Ramey Field
Ramey Field
Ramey Field
Ramey Field
Ramey Field
James Ramey Mound
Powell Mound #2
James Ramey Mound
James Ramey Mound
James Ramey Mound
James Ramey Mound
Madison Co.
Madison Co.
Madison Co.

Oliva spp.
Oliva spp.
Oliva spp.
Oliva spp.
Oliva spp.
Oliva spp.
Oliva spp.
Oliva spp.
Busycon cf. sinistrumn
Busycon cf. sinistrumn
Busycon cf. sinistrunm
Busycon cf. sinistrumn
Busycon cf. sinistruun
Busycon ct. sinistrumn
Busycon cf. sinistrumn
Busycon fc. sinistruin
Busycon cf. sinistruim
Busycon sinistrium
Busvcon sinistrum
Busycon sinistrumn
Busvcon cf. sinistrumn
Busycon cf. sinistruin
Busycon cf. sinistruni
Busycon cf. sinistrum
Busycon cf. sinistrumb
Busycon cf. sinistrumn
Busycon cf. sinistrum,
Busycon cf. sinistrumn
Busycon cf. sinistrum
Busycon sinistrum
Busycon sinistrunm
Busycon sinistrumi
Busycon sinistrumn
Busycon sinistrum
Busycon sinistrum
Busycon sinistrum
Busycon sinistrum
Busycon sinistrumn

bead blank
bead blank
bead blank
bead blank
bead blank
bead blank
bead blank
bead blank
bead blank




Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
Preston Holder Collection, cut on posterior end
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
1956 Perino excavations
cut columella?

Cyrus Thomas & Powell
Cyrus Thomas & Powell
Cyrus Thomas & Powell

siphonal canal & lip
columella & siphonal canal

outer lip
post. '/2 of shell-ant. colum burned
post. /2 of shell want. columella burned
post. V2 of shell want. columella burned
partial shoulder & apex
columella only
columella only
columella only
columella only
columella only
columella only
columella only
columella only
columella only
columella only
columella only
inner whorl
inner whorl
inner whorl
columella, frag outer whorl
columella, frag outer whorl


Appendix: Cahokia burned shell artifacts (cont.)
Artifact Length
Provenience Taxon # Type (mm) Comments Elements
Madison Co. Busycon sinistrum 1 debitage 151.5 Cyrus Thomas & Powell columella, frag outer whorl
Madison Co. Busycon sinistrum 1 debitage 169.8 Cyrus Thomas & Powell columella, frag outer whorl
Wilson Mound Busycon cf. sinistrum 1 bead blank 64.2 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 88.5 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 49.6 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 71.1 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 60.1 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 49.9 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 43.5 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 79.4 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 53.0 Preston Holter collection cut on anterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 35.6 Preston Holter collection cut on posterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 78.3 Preston Holter collection cut on posterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 61.0 Preston Holter collection cut on posterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 61.3 Preston Holter collection cut on posterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 78.2 Preston Holter collection cut on posterior end
Wilson Mound Busycon cf. sinistrum 1 bead blank 38.5 Preston Holter collection cut on posterior end

Bull. Fla. Mus. Nat. Hist. (2003)44(1): 91-100

COTTONMOUTH (Agkistrodon piscivorus) AND DIAMONDBACK
RATTLESNAKE (Crotalus adamanteus) IN FLORIDA

Karen J. Walker'

The cottonmouth, or water moccasin (Agkistrodon piscivorus), and the diamondback rattlesnake (Crotalus adamanteus)
are distributed throughout Florida and their skeletal remains, usually vertebrae, are often present in zooarchaeological
assemblages. Although the two viperid snakes exhibit different habitat preferences-one a semi-aquatic snake, the other
terrestrial-their vertebrae are very similar. This illustrated guide helps to distinguish between the vertebrae of the two
taxa. A strategy of limiting species identifications to the middle trunk series of mature adults and employing multiple
characteristics is recommended to overcome the intracolumnar and individual variability that occurs in the vertebrae of
these snakes.

Key words: Agkistrodon piscivorus, Crotalus adamanteus, Florida snakes, vertebrae, viperid

In varying degrees of abundance and diversity, the skeletal
remains of snakes are a component of many
zooarchaeological assemblages collected from Florida's
pre-Columbian archaeological sites (Fradkin 1978). Their
presence in these assemblages can be attributed to
incidental inclusion (i.e., a snake was attracted to the
human habitation site and died there either naturally or
by human hands) and the snake may or may not have
been eaten by humans. Their presence also may be due
to purposeful hunting of snakes for human consumption.
If one or two individuals are identified in an assemblage,
often the snakes are considered by zooarchaeologists to
be incidental inclusions. Size may be another factor in
determining whether or not the snakes were targeted for
food with the reasoning that the meat provided by smaller
species and juvenile forms may not have been worth the
effort. If several or many individuals are represented by
the remains, then snakes are considered to have been a
food resource. Archaeological sites in the southern third
of the state in particular produce significant quantities of
snake skeletal material. There is no doubt that snakes
were an important food for American Indians who
inhabited such areas as the Big Cypress Swamp and the
Everglades (Danielson 1991; Fradkin 1978; Griffin 1988;
Hale 1984; Wing 1984).
While an impressive taxonomic array of snake taxa
has been identified in Florida's zooarchaeological
assemblages, this paper focuses on two taxa, both

'Environmental Archaeologist, Randell Research Center,
Florida Museum of Natural History, Gainesville, FL 32611, USA.

members of the Viperidae family and Crotalinae
subfamily. Within this pit-viper family, there are eleven
species and subspecies known for the southeastern U.S.
(Conant and Collins 1998). Of this total, six occur in
Florida (Ashton and Ashton 1988; Conant and Collins
1998; Tennant 1997). Of these six, one copperhead
subspecies, Agkistrodon contortrix contortrix, one
cottonmouth subspecies, Agkistrodon piscivorus
piscivorus, and one rattlesnake, Crotalus horridus, are
limited to localized areas within the state: north-central
northwest panhandle, extreme northwest panhandle, and
northeast Florida, respectively.
Only three vipers are distributed throughout the entire
state. These are the Florida cottonmouth, Agkistrodon
piscivorus conanti, often called the water moccasin,
the eastern diamondback rattlesnake, Crotalus
adamanteus, and the dusky pigmy rattlesnake, Sistrurus
miliarius barbouri. For the majority of Florida
archaeofaunal assemblages then, zooarchaeologists have
only these three viperid taxa with which to concern
themselves. Osteologically, no differences are known
among the three southeastern cottonmouth subspecies
or the three pigmy rattlesnake subspecies, so Florida
zooarchaeologists need not be concerned with the
subspecies names conanti or barbouri. Even in the
small areas of Florida and elsewhere in the southeast
where copperheads (A. contortrix) and cottonmouths
(A. piscivorus) co-occur and the diamondback (C.
adamanteus) and timber rattlesnakes (C horridus) co-
occur, these species sometimes can be separated
osteologically (Auffenberg 1963:200; Meylan 1982:57-59).

With the keen observation skills that normally are required
of the zooarchaeologist, one can readily distinguish
between most of the primary cranial and mandibular
elements of Agkistrodon piscivorus and Crotalus
adamanteus. However, the majority of skeletal elements
that are preserved and found in archaeological sites are
not from the fragile skull. Rather, the curved rib and the
compact but complex vertebrae are the most frequently
recovered elements. Ribs are diagnostic of only the Class
level, Serpentes. Vertebrae, however, can often be
assigned to family and genus levels, and sometimes to
species levels, if fragmentation is absent or minimal.
Thus, I have arrived at the purpose of this paper, which
is to provide an illustrated guide for distinguishing between
the vertebrae of A. piscivorus and C. adamanteus from
Florida's zooarchaeological assemblages. It is critical,
however, that the guide be used in concert with multiple
modern comparative specimens so that the researcher
can appreciate the existing intracolumnar and individual
Much of what is described and illustrated here is
based on the work of Auffenberg (1963) and Meylan
(1982) who studied Florida assemblages of snake bones
representing many taxa from deposits dating to the
Miocene, Pliocene, and Pleistocene epochs. The present
paper combines their observations with the present
author's into a more usable form, including a table and a
correlating series of five illustrations presenting the
vertebral diagnostic characteristics for each species side-
by-side by osteological view. It further recommends,
through the use of the table and figures, a strategy of
limiting identification to the middle precaudal, or trunk,
vertebrae of mature individuals and using multiple
characteristics to achieve an identification.

Zooarchaeologists and paleontologists alike generally
strive for a species-level identification of a given skeletal
specimen when that specimen is a diagnostic element,
either complete or fragmented, but still exhibiting
major structural characteristics. Clearly, for the
zooarchaeologist, a species-level identification
maximizes the amount of information that can be inferred
about the specimen, the source animal, and the relationship
between it and the human residents of the site from which
its remains came. From a natural history perspective,
biodiversity and biogeographic databases benefit from
archaeological identifications made to species level.
These are also issues of interest to zooarchaeologists.

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Specific-animal habitat preferences can allow inferences
about environmental conditions during the times of site
occupation. Depending on the extent of the
zooarchaeological dataset being used in a given study,
inferences can be made at multiple scales, spatially and
temporally. Concerning Agkistrodon piscivorus and
Crotalus adamanteus, zooarchaeologists studying Florida
archaeofaunal assemblages especially want to distinguish
between these two taxa because the first is a semi-
aquatic (freshwater) snake and the second almost
exclusively terrestrial. Combined with other lines ofevidence,
past fluctuating water levels might be inferred from varying
abundances of water snakes through time based on present-
day correlations between intra-annual snake abundance
and wet seasons (e.g., Dalrymple et al. 1991).

The snake vertebral column can be divided, more or
less, into four sections. The anterior, or cervical,
vertebrae, including the atlas, are located in the "neck"
region. These exhibit hypapophyses (median ventral
processes, also called haemal processes) even in taxa
whose trunk vertebrae have no hypapophyses, such as
many of the colubrid genera. In taxa whose trunk
vertebrae do have hypapophyses, the anterior vertebrae
often exhibit hypapophyses that are longer than those of
the trunk vertebrae. In the anterior region, an elongated
hypapophysis functions in concert with the muscles of
this area to facilitate the snake's lifting of the head and
neck region. The sacral vertebrae, located in what might
be thought of as a pelvic region, have multiple gracile
projections, both ventrally and dorsally, and thus, if not
completely broken and eroded, are easily recognized.
The caudal vertebrae do not have these projections, but
they are easily recognized by the presence of paired
median processes called lymphapophyses. Both sacral
and caudal vertebrae have these. Even in poorly
preserved archaeological specimens, one can almost
always recognize caudal vertebrae based on the presence
of lymphapophyses.
The last and most important category of vertebrae
is the trunk vertebrae (already mentioned above), all
those found between the anterior and sacral/caudal
vertebrae. All trunk vertebrae exhibit rib attachment
structures, called paradiapophyses, one on each side of
the centrum, adjacent to the cotyle. Auffenberg
(1963:154) and others agree that the trunk vertebrae in
adult individuals are the most consistent vertebrae in
terms of intracolumnar and individual morpologhical

WALKER: Guide to Trunk Vertebrae of Agkistrodon piscivorus and Crotalus adamanteus in Florida

variation and therefore are the most useful for
identification purposes. Even so, all researchers observe
that subtle variations do occur in these vertebrae, for
example, the more anterior of the trunk vertebrae versus
the middle trunk vertebrae versus the more posterior
trunk vertebrae. This is to be expected since the
morphological change from one end of the vertebral
column to the other end is gradual. Nonetheless, with
considerable comparative study and thus familiarity with
snake vertebrae, one can with some confidence
determine that an isolated vertebra is a middle trunk
specimen of an adult individual, the most diagnostic of
all the trunk vertebrae even though these too exhibit some
variation. If one cannot determine that a vertebra is a
middle trunk one, identification should be made only at
the family or subfamily level. Auffenberg (1963:154)
states that he used the relative large size of the neural
canal as a guide to assigning a middle trunk position to
an isolated vertebra. Additional characteristics are an
overall large size and a high neural spine. Thus, in
following the tested lead of Auffenberg and others, this
guide is focused on only these diagnostic vertebrae for
zooarchaeological genus and species identification.
Structural vertebral terminology, most abbreviations,
and vertebral measurements used in this paper follow
Auffenberg (1963:151-155), who outlined these in detail,
with the goal of future consistency among researchers.
Directional terms (anterior, posterior, lateral, dorsal, and
ventral views) also follow Auffenberg (1963), although
others may prefer "cephalad" and caudad" over
"anterior" and "posterior." Although Auffenberg's (1963)
and Meylan's (1982) foci were on pre-Holocene faunal
assemblages, these studies are highly useful to
zooarchaeologists and should be consulted for complete
osteological terminology and study of a variety of snake
taxa, certainly ones common to zooarchaeological
assemblages. In addition, Holman's (1963, 1979, 2000)
work should be consulted. The illustrations presented
here (figs. 1 through 5) are the original work of artist
Sue Ellen Hunter, executed in the Florida Museum of
Natural History's (FLMNH) Environmental Archaeology
Laboratory in consultation with the author. They
represent the most complete set of vertebrae illustrations
known for these taxa, providing five views essential to
the guide's purpose. With few exceptions, these
illustrations are not stylized, so that illustrated differences
in a morphological feature, even between one half of a
vertebra and the other, are realistic and thus are not
artistic error. As indicated in the figure legends,
FLMNH modern specimens from the Herpetology

and Zooarchaeology comparative collections were
used for illustration. In addition, other skeletons were
used for comparative study. These include:
Agkistrodon piscivorus (Zooarchaeology: Z62,
Z1703, Z2381, UF27539-S, UF27540-S; Herpetology:
UF893, UF8950, UF14435, UF676335, UF14107,
UF99084, UF115019, UF9828, UF99027, UF11834,
UF 11928, UF14330, UF37020; Crotalus adamanteus
(Zooarchaeology: Z1252, Z1260, Z2183, Z3092, Z3668,
Z3671, Z3686, Z3795, Z7726; Herpetology: UF9705,
UF14444, UF18396, UF32545, UF32557, UF35130,
UF37513, UF41510, UF53428, UF56113, UF99060); and
Sistrurus miliarius (Zooarchaeology: Z1301, Z1780;
Herpetology: UF 19078, UF 19092, UF40610).

In addition to the viperids, archaeofaunal
assemblages from sites in fresh-water environments
often contain skeletal remains of the aquatic natricines.
Like the viperids and unlike most other colubrids, the
natricines have hypapophyses along the entire vertebral
column. Depending on condition and place in the column,
archaeological natricine and viperid vertebrae can
sometimes be confused, as some of the natricine genera
have high neural spines similar to the middle trunk
vertebrae of the viperids. In particular, some of the
anterior natricine vertebrae with their elongated
hypapophyses can be similar to viperid vertebrae. But
even if the hypapophysis is broken, its cross-section
shape can help to distinguish the natricine vertebra from
a viperid one. The natricine trunk hypapophysis is thin
and bladelike with a distinctive shape in the lateral view
and, if unbroken, is still relatively short. The viperid
hypapophysis, on the other hand, is more rodlike, more
rounded in cross section, and, if unbroken, relatively long.
In addition, natricine centra are somewhat longer than
those of the viperids and natricine accessory processes
are well developed (but not in some anterior vertebrae),
unlike those of the viperids. In those anterior natricine
vertebrae that are so similar to viperids, the neural spines
are directed more posteriorly, while in the viperids the
neural spine is more erect.

As observed by Auffenberg (1963), Holman (1963,
1979), Meylan (1982), and this author, viperid middle trunk
vertebrae generally exhibit the characteristics discussed

below organized by osteological view. Many of the structural
characteristics are described in relation to snakes of other
families, especially those of Colubridae. The vertebrae
of Agkistrodon piscivorus and Crotalus adamanteus
serve to illustrate these general viperid descriptions.
Lateral view (Fig. 1). The centrum (c) is short,
giving each vertebra a compact quality. The condyle (co)
is slightly to moderately oblique, on a short neck, and
directed posteriorly. The hypapophysis (h) is long and
appears narrow, and exists on all vertebrae within the
column. The neural spine (ns) is much higher than long,
either straight at the anterior edge or overhanging and
usually overhanging at the posterior edge. The
parapophysial processes (pp) project ventrally and
anteriorly well beyond the lower lip of the cotyle (ct).
The paradiapophysial articular surfaces (pas) are
separated and distinct from one another.
Ventral view (Fig. 2). The centrum has well
developed subcentral ridges (sr), extending from the base
of the diapophysis (d, the upper articular surface of the
paradiapophysis) posteriorly to near the bottom of the

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

condyle. Accessory processes (ap) are short and not well
developed; sometimes they barely project beyond the
margins of the prezygapophysial articular surfaces (pras).
Dorsal view (Fig. 3). The prezygapophysial
articular surfaces are variably oval in shape. Accessory
processes are short. From above, the zygosphene (z) is
variable in shape.
Anterior view (Fig. 4). The paradiapophysial
articular surfaces are separated. The hypapophysis is not
compressed laterally, is almost rodlike, and is sometimes
thickened distally. The cotyle shape is oval to sometimes
rounded. The zygosphene is variable in thickness.
Posterior view (Fig. 5). Also seen in this view is
that the hypapophysis is not compressed laterally, is
almost rodlike, and is sometimes thickened distally.
The condyle is large and round. The neural arch (na)
is wide, short, and depressed. Auffenberg (1963:200)
states that the neural arch sometimes exhibits an
epizygapophysial spine (es) and, based on my
observations, this seems to be the case in older individuals
of Agkistrodon piscivorus.


Figure 1. Comparative lateral view of typical mature Agkistrodon piscivorus (Herpetology: UF893) and Crotalus adamanteus
(Zooarchaeology: Z1252) trunk vertebrae: anterior (a); posterior (p); dorsal (d); ventral (v); neural spine (ns); neural arch (na);
epizygapophysial spine (es); condyle (co); centrum length (cl); hypapophysis (h); parapophysial process (pp); paradiapophysial
articular surfaces (pas); cotyle (ct); accessory process (ap).

Agkistrodon piscivorus Crotalus adamanteus

WALKER: Guide to Trunk Vertebrae of Agkistrodon piscivorus and Crotalus adamanteus in Florida 9:

Agkistrodon piscivorus

Crotalus adamanteus


Figure 2. Comparative ventral view of typical mature Agkistrodon piscivorus (Herpetology: UF893) and Crotalus adamanteus
(Zooarchaeology: Z 1252) trunk vertebrae (anterior is up): parapophysial processes (pp); cotyle (ct); diapophysis (d); subcentrum
ridge (sr); condyle (co); neural spine (ns); hypapophysis (h); prezygapophysial articular surface (pras); accessory process (ap).

Zooarchaeologists must always consider the
possibility of the presence of S. miliarius vertebrae in
Florida's archaeofaunal assemblages. Adult vertebrae
are small (unlike those of adult Agkistrodon piscivorus
and Crotalus adamanteus), as S. miliarius in life
averages only 20" (51 cm), according to Ashton and
Ashton (1988:165). Auffenberg (1963:201) used the
following combination of vertebral characteristics to
distinguish S. miliarius specimens from young A.
piscivorus and C. adamanteus specimens: round cotyle;
absence of epizygapophysial spine; wide zygosphene,
convex in anterior view; longer and narrower centrum.
Meylan (1982:61) separated out fossil S. miliarius
vertebrae from young A. piscivorus and C. adamanteus

based on the proportionally smaller, round cotyles of S.
miliarius. Of all these characteristics, while the rounded
cotyle is the clearest and most consistent one, my
observations include some small, round cotyles within the
columns of both A. piscivorus and C. adamanteus
specimens, away from the middle trunk vertebrae. This
author's recommendation is to leave small viperid vertebrae
exhibiting round cotyles at the family/subfamily level.

Once natricines and Sistrurus miliarius are eliminated
as possible identifications for an isolated adult trunk
vertebra, the final challenge is to determine whether or
not a vertebra can be assigned an Agkistrodon
piscivorus or Crotalus adamanteus identification or


ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Agkistrodon piscivorus


Crotalus adamanteus


Figure 3. Comparative dorsal view of typical mature Agkistrodon piscivorus (left, Herpetology: UF893) and Crotalus adamanteus
(right, Zooarchaeology: Z1252) trunk vertebrae (anterior is up): prezygapophysial articular surfaces (pras); zygosphene width
(zw); neural spine (ns); neural arch width (naw); hypapophysis (h); accessory process (ap).

whether it is best left at family/subfamily level. Those
structural characteristics observed to be the most
constant within the middle trunk vertebrae series of these
two taxa are presented in Table 1 and described below.
Nonetheless, as pointed out above, some variation within
the middle trunk series does occur and for this reason,
no one characteristic can be relied upon for separating
A. piscivorus and C. adamanteus. A multicharacter
approach is required. There must be enough of a
vertebra present in order to discern a combination of
two or more primary diagnostic characteristics (those
with an asterisk in Table 1) for a confident identification.
The variation is such that some of the primary
characteristics may be evident and some not. If
observed, secondary characteristics may be helpful, but
should not be relied upon. Even if the vertebra is complete,

uneroded, and clean of encrusting sediment, the
characteristics in Table 1 may not clearly point to one
taxon or the other, in which case the vertebra should be
identified to family or subfamily only.
Lateral view (Fig. 1). Several primary
characteristics can be seen in the lateral view. The
hypapophysis of Agkistrodon piscivorus is slightly
shorter, more gracile, and projects slightly more
posteriorly than that of Crotalus adamanteus. Related
to the projection, the angle between the anterior edge of
the A. piscivorus hypapophysis and a vertical line drawn
through the center of the centrum is greater than that of
C. adamanteus. The parapophysial processes project
more anteriorly in A. piscivorus. Auffenberg (1963)
noted that the neural spine of A. piscivorus has a
"tendency" to be lower than the neural spine of C.

WALKER: Guide to Trunk Vertebrae of Agkistrodon piscivorus and Crotalus adamanteus in Florida

adamanteus. Illustrator Hunter and this author observed
a less than 50% occurrence of higher C. adamanteus
spines and for this reason consider this characteristic to
be of secondary importance. (The neural spines are not
drawn to illustrate the difference.) Another secondary
characteristic is that the neural spine of A. piscivorus is
usually thinner and shows no thickening on the upper
anterior edge. The A. piscivorus condyle is usually more
oblique and gracile. Meylan (1982:59) experimented with
several vertebral-measurement ratios in an attempt to
separate A. piscivorus and C. adamanteus. One is
centrum length (cl, Fig. 1) divided by neural arch width

Agkistrodon piscivorus

scale: 2/1 KEY
scale: 2/1

(naw, Fig. 3). Whereas Meylan's means for the two
taxa were similar, 1.25 mm and 1.24 mm, respectively,
results from a small sample combining two small
(Zooarchaeology Z3092, UF27539-S) and two large
(Herpetology UF893; Zooarchaeology UF32540)
individuals produced means of 1.10 mm for A. piscivorus
and 0.85 mm for C. adamanteus. The very different
results suggest that too much variability may exist for
this ratio to be diagnostically useful.
Ventral view (Fig. 2). Three primary characteristics
are exhibited in the ventral view. The parapophysial
processes of Agkistrodon piscivorus appear narrower

Crotalus adamanteus



Figure 4. Comparative anterior view of typical mature Agkistrodon piscivorus (Herpetology UF893) and Crotalus adamanteus
(Zooarchaeology Z1252) trunk vertebrae: neural spine (ns); zygosphene angle (z 0); neural canal (nc); cotyle (ct); hypapophysis
(h); paradiapophysial articular surfaces (pas).

ZOOARCHAEOLOGY: Papers to Honor Elizabeth S. Wing

Agkistrodon piscivorus

Crotalus adamanteus


Figure 5. Comparative posterior view of typical mature Agkistrodon piscivorus (left, Herpetology: UF893) and Crotalus
adamanteus (right, Zooarchaeology: UF1252) trunk vertebrae: neural spine (ns); neural arch (na); zygantrum (zg); condyle (co);
hypapophysis (h); epizygapophysial spine (es); neural canal (nc).

and more parallel to each other, while the Crotalus
adamanteus processes are wider and more diverging
or V-shaped. The third primary feature is that often at
the base of the C. adamanteus cotyle is a well developed
ridge, usually absent in A. piscivorus. A secondary feature
is that the condyle of C. adamanteus often is more bulbous.
Dorsal view (Fig. 3). The shapes of both the
accessory processes and the prezygapophysial articular
surfaces can be variable even between left and right
sides of a single vertebra. But a fairly constant and
therefore primary characteristic is that both are directed
more laterally in the Agkistrodon piscivorus and slightly
more anteriorily the Crotalus adamanteus. Secondary
characteristics are also seen in the dorsal view. The neural
spine of A. piscivorus is usually thinner. The distal end
of the C. adamanteus hypapophysis is more robust and

roughly grooved. The zygosphene is often more concave
in A. piscivorus and more crenate in C. adamanteus.
Meylan's (1982:59) results of calculating the zygosphene
width to neural arch width ratio (zw/naw) were the most
promising for separating the two taxa because they
resulted in a difference in the means of A. piscivorus
(0.88 mm) and C. adamanteus (0.82 mm). However, a
test of the A. piscivorus ratio mean, measuring vertebrae
(n = 25) of four recent individuals (Zooarchaeology Z62,
Z2381, UF27539-S, UF27540-S), produced a mean of
0.78 mm with a range of 0.74 to 0.80 mm. Based on this
author's admittedly limited measurements, the ratio might
be an unreliable characteristic for identifying isolated
archaeological vertebrae.
Anterior view (Fig. 4). Observed in the anterior
view, a primary characteristic is the flattened top edge


WALKER: Guide to Trunk Vertebrae of Agkistrodon piscivorus and Crotalus adamanteus in Florida

Table 1. A comparison of structural characteristics of the middle trunk vertebrae of mature Agkistrodonpiscivorus and Crotalus
adamanteus. Asterisks indicate primary characteristic; figure citations are for the present paper. "A" indicates Auffenberg 1963
as primary reference; "H/M"as Holman 1963 in Meylan 1982; "H" as Holman 1979; "M" as Meylan 1982; "W" as Walker, this

Vertebral structure

*hypapophysis (Fig. 1)

*angle between anterior edge of
hypapophysis & line drawn perpen-
dicular to center of centrum (Fig. 1)

*parapophysial processes (Fig. 1)

epizygapophysial spines (Fig. 1)

neural spine (Fig. 1)

neural spine (Figs. 1,3, 4)

condyle (Fig. 1)

cl/naw (centrum length/neural
arch width) (Figs. 1,3)

*parapophysial processes (Fig. 2)

*parapophysial processes (Fig. 2)

*base of cotyle (Fig. 2)

condyle (Fig. 2)

*accessory processes (Fig. 3)

*prezygapophysial articular surfaces
(Figs. 3,4)

*distal end of hypapophysis (Fig. 3)

zygosphene shape (Fig. 3)

zw/naw (zygosphene width/neural
arch width) (Fig. 3)

*cotyle (Fig. 4)

*indentations or pits on either side
ofcotyle (Fig. 4)

*indentations or pits on either side
ofcotyle (Fig. 4)

*position ofzygantrum within
neural arch (Fig. 5)


A. piscivorus characteristic
A: shorter

W: greater

A: more projected anteriorly

A: faint when present

A: lower

A: thinner with no thickening
on upper anterior edge

W: more oblique

M: (n = 31) mean of 1.25 mm
W: (n= 6) mean of 1.10 mm

W: narrower

A: more parallel

W: area is smooth

W: more gracile

W: more laterally directed

A: more laterally directed

W: more gracile and smooth

W: more concave

M: (n = 32) mean of 0.88 mm
W: (n = 25) mean of 0.78 mm

W: dorsal edge usually rounded

A: more deeply indented

H/M: pits are distinct and each
contains one large foramen

W: lower

C. adamanteus characteristic
A: longer

W: less

A: less projected anteriorly

A: usually absent

A: higher

A: thicker, usually with a tubercle
on anterior, upper edge

W: less oblique

M: (n = 31) mean of 1.24 mm
W: (n = 6) mean of0.85 mm

W: wider

A: more diverging

W: well developed ridge

W: more robust

W: more anteriorly directed

A: more anteriorly directed

W: more robust and grooved

W: more crenate

M: (n = 31) mean of 0.82 mm

W: dorsal edge is flattened

A: less indented

H/M and H: one or more small
foramina in indistinct pits when

W: higher