• TABLE OF CONTENTS
HIDE
 Preface
 Main














Title: Journal of Caribbean archaeology
ALL VOLUMES CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00091746/00010
 Material Information
Title: Journal of Caribbean archaeology
Series Title: Journal of Caribbean archaeology
Alternate Title: JCA
Abbreviated Title: J. Caribb. archaelo.
Physical Description: Serial
Language: English
Publisher: Christopher Ohm Clement ;
Christopher Ohm Clement
William F. Keegan
Place of Publication: Gainesville FL
Publication Date: 2008
Frequency: annual
regular
 Subjects
Subject: Archaeology -- Periodicals -- Caribbean Area   ( lcsh )
Antiquities -- Periodicals -- Caribbean Area   ( lcsh )
Genre: periodical   ( marcgt )
 Notes
System Details: Mode of access: World Wide Web.
Dates or Sequential Designation: Vol. 1 (2000)-
General Note: Title from title screen (publisher's Web site, viewed Dec. 2, 2002).
General Note: Latest issue consulted: Vol. 5 (2004).
 Record Information
Bibliographic ID: UF00091746
Volume ID: VID00010
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 003345617
oclc - 41077527
lccn - sn 99003684
issn - 1524-4776

Downloads

This item has the following downloads:

Binder2 ( PDF )


Table of Contents
    Preface
        Page i
        Page ii
        Page iii
        Page iv
    Main
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
Full Text




Journal of CaribbeanArchaeologv
Copynght 2008
ISSN 1524-4776


AN EXPLORATORY STUDY INTO THE CHEMICAL CHARACTERIZATION
OF CARIBBEAN CERAMICS:

AN INTRODUCTION TO A SPECIAL VOLUME
OF THE JOURNAL OF CARIBBEAN ARCHAEOLOGY
IN MEMORY OF JAMES B. PETERSEN

Special Editors:

Christophe Descantes
Archaeological Research Facility
University of California Berkeley

Robert J. Speakman
Museum Conservation Institute
Simithw i ,ian, Institution

Michael D. Glascock
Archaeometry Laboratory
University of Missouri Research Reactor

I A/ihe' T. Boulanger
Department ofAnthropology & Archaeometry Laboratory
University of Missouri


The papers presented in this issue of the
Journal of Caribbean Archaeology were
presented at the 2006 Society for American
Archeology (SAA) annual meetings in San
Juan, Puerto Rico. The conference papers
were part of a session entitled "An
Exploratory Study into the Chemical
Characterization of Caribbean Ceramics: In
Memory of James B. Petersen." A
participant in and strong supporter of the
session, Prof. James B. Petersen was
tragically murdered in the summer of 2005
while conducting fieldwork in Brazil. The
organizer for the session was Christophe
Descantes; the session chair was Michael D.
Glascock. Ronald Bishop (Smithsonian

Journal of Caribbean Archaeology, Special Publication


Institution) served as the discussant for the
session and offered invaluable comments on
the papers. For the journal issue, Christophe
Descantes, Robert J. Speakman, Michael D.
Glascock, and Matthew T. Boulanger are the
special editors. Descantes, Speakman, and
Glascock were involved in the instrumental
neutron activation analysis (INAA) of the
ceramic and clay specimens at the University
of Missouri Research Reactor Center
(MURR), and Boulanger helped in the
editing, typesetting, and pre-press of the
articles. Daan Iseendoorn et al.'s work was
analyzed using x-ray fluorescence (XRF) at
Vrije Universiteit Amsterdam.


2, 2008 i












Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776










































































This Page Intentionally Left Blank.



Journal of Caribbean Archaeology, Special Publication ', 2008









Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776


THE A AUTHORS DEDICATE THE PAPERS IN THIS VOL UME
TO THE MEMORY OF

JAMES B. PETERSEN
1954-2005


Among his many passions, Jim Petersen was dedicated to the thorough analysis of human
material culture, past and present. His meticulous study of Amerindian technologies including
ceramics, stone tools, and perishable textiles, led to groundbreaking insights into the material
signatures of social interaction and ethnic boundaries. During his career, Jim analyzed hundreds
of thousands of artifacts, all the time focused on what the careful study of objects could tell us
about people.
Above all other classes of artifacts, ceramics held a special place for Jim. He could squeeze
diamonds of data from the smallest of fragments. While decorated vessels certainly caught his
eye, he loved every sherd equally. He always looked beneath the surface, always went one step
further in his quest to learn more about pots and their potters. Jim was a champion of the small
and underappreciated attributes, defending them with countless hours of his free time and the
binocular microscope he lugged around in his carry-on bag. While Jim practiced his own more
"traditional" approach to ceramic analysis, he was eager to explore the ways in which INAA
could help us elevate our understanding of ceramic manufacture and exchange in the Caribbean.
He would be excited by the progress this volume represents.

John G. Crock
University of Vermont


Jim Petersen, Anguilla 1992 (photo by David R. Watters).
Journal of Caribbean Archaeology, Special Publication 2, 2008


F












Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776










































































This Page Intentionally Left Blank.



Journal of Caribbean Archaeology, Special Publication #2, 2008








Journal of CaribbeanArchaeology
Copyright 2008
ISSN 1524-4776


COMPOSITIONAL STUDIES OF CARIBBEAN CERAMICS:
AN INTRODUCTION TO INSTRUMENTAL NEUTRON ACTIVATION ANALYSIS

Christophe Descantes
Archaeological Research Facility
University of California Berkeley
2251 College Building
Berkeley, CA 94720-1076
cdescantes@berkeley. edu

Robert J. Speakman
Museum Conservation Institute
S.mithui, inu Institution
4120 Silver Hill Road
Suitland, MD 20746
speakmanj@si. edu

Michael D. Glascock
Archaeometry Laboratory
University of Missouri Research Reactor
1513 Research Park Drive
Columbia, MO 65211
glascockm@missouri. edu

The University of Missouri Research Reactor Center (MURR) initiated a research program in
2003 centered on the chemical characterizations of ceramics from the Caribbean region. This
article describes the technique and procedures involved in instrumental neutron activation
analysis (INAA), applied to all but one of the studies in this issue. These papers were presented
at the 2006 Society for American Archaeology (SAA) 71st annual meetings in San Juan, Puerto
Rico. The collaborative studies presented are preliminary in nature, but hold promise for ad-
dressing a myriad of issues regarding Caribbean prehistory, ranging from island eIleinuent to
culture change.


The Caribbean region has been one of the
most underrepresented areas of the world for
chemical characterization studies of ancient
material culture. In recent years,
compositional studies have begun to address
this deficiency (e.g., Bohus et al. 2005; Carini
1991; Crane 1993; Fitzpatrick 2000; Hofman
et al. 2005; Lundberg et al. 2002; Padilla et


al. 2003; Padilla et al. 2006; Winter and
Gilstrap 1991). This paper introduces one
such program initiated at the University of
Missouri Research Reactor Facility (MURR)
where ceramics from eight projects were
investigated to begin building a
compositional database of Caribbean
ceramics. A description of the techniques and


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Studies of Caribbean Ceramics

procedures of INAA applied to the ceramics
and clays of most of the papers, is presented
followed by an introduction of the different
papers and a discussion of the potential of the
techniques for addressing issues of Caribbean
archaeology. Isendoorn et al.'s paper
introduces the archaeometric program at
Leiden University where various
archaeometric techniques have been used for
the study of Caribbean ceramics (see also
Hofman et al. 2005).
The senior author initiated this
archaeometric project on Caribbean ceramics
in 2003 while a postdoctoral researcher at the
Research Reactor Center. Several leading
Caribbean researchers based at US academic
institutions were invited to participate in the
study and were offered 50 free analyses in
exchange for participating in the SAA
conference session. The long-term goal of the
study is to address some of the major
anthropological questions of the region by
building an elemental database of the most
common artifact type found in the historic
and prehistoric Caribbean archaeological
contexts-ceramics. The technique has been
applied successfully to numerous other
cultural regions in the Americas, notably the
American Southwest and Mesoamerica
(Glowacki and Neff 2002; Nichols et al.
2002). INAA has allowed for the exploration
of the changing roles of markets and the
development of preindustrial markets in the
Basin of Mexico. In the American Southwest,
INAA has become a useful tool in
understanding exchange, migration, social
identity, and economic organization.
Financial support for the INAA studies
presented in this issue was provided by the
National Science Foundation (BCS-0102325)
and the US Department of Energy Office of
Nuclear Energy, Science and Technology


Descantes et al.


(Award No. DE-FG07-03ID14531) to the
Midwest Nuclear Science and Engineering
Consortium under the Innovations in Nuclear
Infrastructure and Education program.
Isendoorn et al.'s X-ray fluorescence (XRF)
analyses and fieldwork were funded by the
NWO VIDI-project "Mobility and exchange:
dynamics of social, material and ideological
relations in the pre-Columbian insular
Caribbean" directed by Professor C. L.
Hofman and by the Byvanck Fonds.
The Yale University Peabody Museum of
Natural History and the late Professor Irving
Rouse graciously facilitated Christophe
Descantes' compositional studies of Indian
Creek ceramics from the West Indian island
of Antigua. The INAA study of these data
was submitted for publication in the
Proceedings of the 21st Congress of the
International Association for Caribbean
Archaeology (IACA) in Trinidad and Tobago
(Descantes et al. 2007). Antiguan data
presented in the paper above are used in this
introductory paper to illustrate the INAA
technique and the reduction of the data.
In addition to the compositional programs
at Leiden University and MURR described in
this volume, there have been other recent
compositional studies of Caribbean
ceramics-most notably the work of Roman
Padilla et al. at the Centro de Aplicaciones
Tecnol6gicas y Desarrollo Nuclear
(CEADEN) in Cuba, and of Magdalena
Mackowiak de Antczak and Andrzej Antczak
at Sim6n Bolivar University in Caracas,
Venezuela.
The INAA work of Padilla et al. (2003)
from CEADEN successfully combined INAA
and scanning electron microscopy X-ray
analysis (SEM-EDX) to study Cuban pottery
from prehistoric and historic assemblages.
Incorporating petrographic analyses in their


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Studies of Caribbean Ceramics

work, they confirmed that different clay
regions in central Cuba have different clay
compositions. When investigating the
provenance of Majolica ware recovered at
Cuban sites, Padilla et al. (2003) argue that
the pottery originates from Spain, Old
Havana in Cuba, and possibly Veracruz,
Mexico. Recently, Padilla et al. (2006)
combined two energy dispersive X-ray
fluorescence methods-improved sensitivity
polarized X-ray fluorescence analysis
(EDPXRF) and radioisotope X-ray
fluorescence (R-XRF)-for fast and non-
destructive analyses of ceramic sherds. They
found the chemical measurements were
comparable to that of INAA, provided the
surface of the sherd was flat and had no
evident signs of physical alteration, such as
cracks or cavities.
Similarly, Magdalena Mackowiak de
Antczak and Andrzej Antczak from Sim6n
Bolivar University in Caracas, Venezuela,
collaborated with the Institute of Isotope and
Surface Chemistry in Budapest to determine
specific areas of production of Los Roques
Archipelago ceramic figurines. As in other
archaeometric studies into Caribbean material
culture, various analytical techniques, such as
INAA, total reflection X-ray fluorescence
(TXRF), and prompt gamma activation
analysis (PGAA) are combined with
traditional petrographic analyses to not only
provenance the stylistically diverse ceramic
figurines, but also to investigate the social
and ideological aspects of the figurines in
their respective archaeological deposits.
Preliminary results reported indicate that
similar styled figurines from the Los Roques
Archipelago and from the Lake Valencia
region in Venezuela differ significantly in
chemical composition (Bohus et al.
2005:255).


Descantes et al.


The INAA Technique
Instrumental neutron activation analysis is a
powerful quantitative analytical technique
that has been widely applied in
archaeological studies for the last 50 years
(Glascock and Neff 2003; Neff 2000; see
papers in Speakman and Glascock 2007). The
precision and accuracy of this radiometric
technique in generating empirical data for
major, minor, and trace elements of objects of
interest makes it one of the most reliable
compositional characterization techniques,
especially useful for addressing questions of
provenance or origins.
In brief, INAA involves activating the
nuclei of objects by bombarding them with
neutrons generated from a nuclear reactor
(Figure 1). Once activated, nuclei emit
gamma rays with characteristic energies that
can be measured and attributed to specific
isotopes. Quantitative INAA data can be
generated by including known reference
standards in the analyses (see Glascock
1992).

Sample Preparation
Standard MURR procedures were used in
the INAA of the Caribbean ceramics and clay
samples (see Glascock 1992). Fragments of
about 1cm2 were removed from each sample
and abraded using a silicon carbide burr in
order to remove glaze, slip, paint, and
adhering soil, thereby reducing the risk of
measuring contamination. Specimens were
then washed with deionized water before
being crushed into fine powder in an agate
mortar. Where possible a portion of each
specimen was retained, unpowdered, for the
MURR archive of analyzed ceramic fabrics.
Powdered samples were oven-dried at 100
degrees Celsius for 24 hours. Portions of
approximately 150 mg were weighed and


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Studies of Caribbean Ceramics


beta
particle


target prompt
nucleus gamma re


incident
neutron


compound
nucleus


Figure 1. Schematic diagram of instrumental neutron activation analysis (adapted from
Glascock 1998:100).


placed in small polyvials used for short
irradiations. At the same time, 200 mg of
each sample were weighed into high-purity
quartz vials used for long irradiations. Along
with the unknown samples, reference
standards of SRM-1633a (coal fly ash) and
SRM-688 (basalt rock) were similarly
prepared, as were quality control samples
(e.g., standards treated as unknowns) of
SRM-278 (obsidian rock) and Ohio Red Clay
(an in-house standard).

Irradiation and Gamma-Ray Spectroscopy
INAA of ceramics at MURR consists of
two irradiations and three gamma counts. As
discussed in detail by Glascock (1992), a
short irradiation is carried out through the
pneumatic tube irradiation system (Figure 2).
Samples in the polyvials are sequentially
irradiated, two at a time, for five seconds at a
neutron flux of 8 x 1013 n cm-2 s-1. The 720-
second count yields gamma spectra
containing peaks for nine short-lived


elements: aluminum (Al), barium (Ba),
calcium (Ca), dysprosium (Dy), potassium
(K), manganese (Mn), sodium (Na), titanium
(Ti), and vanadium (V). The samples
encapsulated in quartz vials are subjected to a
24-hour irradiation at a neutron flux of 5 x
1013 cm-2 s-1. This long irradiation is
analogous to the single irradiation utilized at
most other laboratories. After the long
irradiation, samples decay for seven days,
and then are counted for 1800 seconds (the
"middle count") on a high-resolution
germanium detector coupled to an automatic
sample changer. The middle count yields
determinations of seven medium half-life
elements, namely: arsenic (As), lanthanum
(La), lutetium (Lu), neodymium (Nd),
samarium (Sm), uranium (U), and ytterbium
(Yb). After an additional three- or four-week
decay, a final count of 9,000 seconds is
carried out on each sample. The latter
measurement yields the following 17 long
half-life elements: cerium (Ce), cobalt (Co),


delayed
gamma ray


Journal of Caribbean Archaeology, Special Publication ', 2008


Descantes et al.









Compositional Studies of Caribbean Ceramics

chromium (Cr), cesium (Cs), europium (Eu),
iron (Fe), hafnium (Hf), nickel (Ni), rubidium
(Rb), antimony (Sb), scandium (Sc),
strontium (Sr), tantalum (Ta), terbium (Tb),
thorium (Th), zinc (Zn), and zirconium (Zr).
The element concentration data from the
three measurements are tabulated in parts per
million using Microsoft Excel.

Quantitative Analysis of the Chemical Data
The following section is a brief description
of the data reduction and analytical
procedures used in the chemical
compositional analysis. See Neff (1994,
2000) for more detailed information. As is
customary in ceramic provenance studies at
MURR (Bishop and Neff 1989), the data are
converted to base-10 logarithms of
concentrations. Use of log concentrations
rather than raw data compensates for
differences in magnitude between major
elements. Chemical data values for the 33


Descantes et al.


elements in the analyzed samples are
examined prior to identifying compositional
groups. Specimens with anomalous
concentrations are treated as outliers and
rejected from subsequent statistical
procedures. In addition, elemental
abundances that are non-existent or below
detection limits in many of the samples are
also dropped from further analyses. The
elimination of nickel measurements is not
uncommon in the chemical analysis of
ceramics in general and in Caribbean
ceramics in particular. The last procedure
involves searching for a pattern of calcium
concentrations greater than 1% or 10,000
parts per million. The high calcium
abundances are most likely a result of either
of the original clay sources or a calcium-rich
temper. A correction factor is applied to all of
the elemental data to counter the dilution
effect of calcium on the other elemental
abundances (see Cogswell et al. 1998). As in


HPGe
detector




sample


Rabbit with two samples

Figure 2. Short irradiation of a rabbit carrying two samples in a pneumatic tube system (left)
and the measurement of gamma-ray emission in front of a high-resolution germanium detec-
tor (right).


Journal of Caribbean Archaeology, Special Publication 2, 2008










Compositional Studies of Caribbean Ceramics


+. + +
F *
CM: 5
0~
o


OL 0 ++ + +
I ? "* *.+ +

+: +
+ +,
U



C-


-1,0 -0.8 -02 02 0 6
Principal Component I


I -1.0 -0.6 -0.2 0.2 0-6
Principal Component 1


Figure 3. Plot of principal components 1 and 2 displaying the ceramic INAA data from the Indian Creek Site, Anti-
gua (left) and plot of the same data with identified compositional groups (right). Unassigned specimens (+) are not
labeled.


previous treatments of ceramics with
enriched calcium concentrations, we
eliminated the elemental concentrations of
calcium and strontium. A calcium correction
factor was applied to the ceramic data from
the Dominican Republic and Puerto Rico.
The goal of quantitative analysis of the
chemical data is to recognize compositionally
homogeneous groups (that are hopefully
anthropologically meaningful) within the
analytical dataset. Based on Weigand et al.' s
(1977) "provenance postulate", such groups
are assumed to represent geographically
restricted sources or source zones. The
location of sources or source zones may be
inferred by comparing the unknown groups to
knowns (source raw materials) or by indirect
means. Such indirect means include the
"criterion of abundance" (Bishop et al. 1982)
or arguments based on geological and
sedimentological characteristics (e.g.,
Steponaitis et al. 1996).
Initial hypotheses about source-related
subgroups in the compositional data are
derived from the application of pattern-
recognition techniques to the chemical data.


Principal components analysis (PCA) is used
to recognize patterns, that is subgroups in the
compositional data (Figure 3). PCA provides
new reference axes that are arranged in
decreasing order of variance subsumed. PCA
can be used in a pure pattern-recognition
mode, i.e., to search for subgroups in an
undifferentiated data set or in a more
evaluative mode to assess the coherence of
hypothetical groups suggested by other
criteria, such as archaeological context and
decoration.
Biplots are used to display both the
variables (elements) and objects (individual
analyzed samples) on the same set of
principal component reference axes (Figure
4). Displaying objects and variables on the
same plots makes it possible to observe the
contributions of specific elements to group
separation and to the distinctive shapes of the
various groups.
The Mahalanobis distance statistic is used
to statistically test the separation between
groups or between individual points and
groups on multiple dimensions (see Bieber et
al. 1976; Bishop and Neff 1989; Harbottle


Journal of Caribbean Archaeology, Special Publication 2, 2008


S Group 1




Group 3
Group 2


1-0 1.4


Descantes et al.









Compositional Studies of Caribbean Ceramics

1976; Neff 2001; Sayre 1975. Mahalanobis
distance takes into account variances and
covariances in the multivariate group and is
analogous to expressing distance from a
univariate mean in standard deviation units.
Similar to standard deviation units,
Mahalanobis distances can be converted into
probabilities of group membership for each
individual specimen (e.g., Bieber et al. 1976;
Harbottle 1976).
Mahalanobis distance-based probabilities of
group membership for members in small
groups may fluctuate dramatically depending
on whether or not each specimen is assumed
to be a member of the group to which it is
being compared. This limitation can be
circumvented by cross-validation or
jackknifingg", that is, by removing each
specimen from its presumed group before
calculating its own probability of
membership (Baxter 1992; Leese and Main
1994). All probabilities discussed in each of
the papers in this issue were cross-validated.
Mahalanobis distances can be calculated for
the log concentrations and the principal
components extracted from the variance-
covariance or correlation matrix of the
complete data set. Calculating Mahalanobis
distances on the principal components
permits us to reduce the dimensionality of the
data set and statistically test the separation of
the chemical groups, which is crucial when
dealing with small groups. We use enough
principal components to subsume
approximately 90% of total variance in the
data set. Typically, elemental plots are
provided in the analysis to demonstrate that
the identified compositional groups are not an
artifact of the PCA algorithm (see Figures 4
and 5).


Descantes et al.


Potential of INAA studies
The compositional study of ceramic
materials from archaeological sites promises
to shed light on some of the major
archaeological questions of the region. Such
questions include ancient population
movements within the region, the
development of exchange networks both
within and among various island
archipelagoes, and local adaptations and
historical developments from ca. 4000 BC up
to and including the time of Euroamerican
contact.
Material culture in the Caribbean has long
served as a marker for identifying cultural
traditions and modeling population
movements (Keegan 1995; Rouse 1992). In
regards to the original movements of
preceramic peoples into the Antillean region,
many areas have been suggested as possible
sources (e.g., Cruxent and Rouse 1969; Veloz
and Vega 1982; Wilson et al. 1998). The
Yucatan Peninsula is favored as the most
likely source for the early people who settled
the Greater Antilles (Keegan 2000). In
contrast, the later horticultural ceramic-
producing peoples who populated the Lesser
Antilles sometime during the first millennium
BC are hypothesized to be migrants from
northern South America (near the mouth of
the Orinoco River). Compositional analyses
of the material culture, and where possible,
isotope ratios (e.g., Sr) of human bone and
teeth, will contribute to our understanding of
the complex migration movements into the
region, and serve as an important
counterpoint to previous migration
hypotheses based on artifact attributes and
historical linguistic studies. Compositional
data for ceramics from early assemblages can
provide quantifiable evidence and track
technological changes during the rapid island


Journal of Caribbean Archaeology, Special Publication ', 2008











Compositional Studies of Caribbean Ceramics


Descantes et al.


0

a Group 1

.- 1.0 -06 -. A
C Ca Sr Ti
O a wHf

0 ce
0 cCa
-- Group 4 Sb





Eu Yb Group 2

o
I -1.0 -0.6 -0.2 0.2 0.6 1.0 1.4

Principal Component 1

Figure 4. Correlation matrix PCA biplot of principal components 1 and 2 showing the
four compositional groups identified in the Indian Creek Site ceramic sample. Ellipses
represent 90% confidence level for membership in the groups. Unassigned specimens are
not shown.


Group 2

a

Group 1






Group 3

0.8 1.0 1.2 1.4 1.6 1.8 2.0
Chromium (log base-10 ppm)


0 0.9 1.0 1.1 1,2 1.3 1,4 1.5 1.5 1,7
Chromium (log base-10 ppm)


Figure 5. Left: Plot of base-10 logged chromium and iron concentrations of the four ceramic compositional groups
at the Indian Creek Site, Antigua. Ellipses represent 90% confidence level for membership in the groups. Unclassi-
fied samples (+) are not labeled. Right: bivariate plot of base-10 logged chromium and rubidium concentrations
displaying the chemical distinctiveness of compositional Groups 3 and 4. Ellipses represent 90% confidence level
for membership in the groups. Unclassified samples are not shown.


Journal of Caribbean Archaeology, Special Publication 2, 2008









Compositional Studies of Caribbean Ceramics

settlement of the Caribbean. In particular, the
acquisition of ceramic and clay compositional
data can allow one to identify the possible
interactions between people of new island
settlements with the people from whence they
came. Finally, studies using INAA can also
contribute insights into the origin of ceramic
technological behaviors of African slave
societies in the Caribbean by investigating
comparable technologies in West Africa.
Constructing past exchange models is a
major task facing Caribbean archaeologists
(e.g., see Crock and Petersen 2004; Hauser
2001; Knippenberg 2006). Exchange
networks are extensive in Caribbean
prehistory; ceramics, cherts, and other
materials found in the archaeological record
are important sources of evidence for
identifying past exchange behaviors.
Petrographic analyses of ceramics in the
region have long provided invaluable
evidence for addressing questions of ceramic
production and exchange, and still have a
vital role to play as a complimentary
technique to finer-grained analytical
techniques. Detailed studies that employ
INAA and petrographic analysis to study the
material culture of the peoples of the
Caribbean contribute to our understanding of
exchange networks, which are important for
developing and testing hypotheses about the
origins, the operation, and the transformation
of exchange systems.
Permeating all aspects of island life,
exchange initiates interaction between people
and is an important dynamic in discussions of
culture change. Exchange studies based on
compositional data can provide insights on
cultural processes of ethnogenesis,
enculturation, colonialism, and resistance in the
prehistoric and recent past. Leiden
University's Archaeology Program presently


Descantes et al.


heads several Caribbean research projects
that employ state-of-the-art archaeometric
techniques, such as thermal ionization mass
spectrometry (TIMS), to gain insights into the
movement of objects and their impacts on the
societies involved.
Compositional analyses of ceramics can
provide insights into the technologies of past
peoples (e.g., Curet 1997). Compositional
analyses of undecorated and decorated wares
aimed at understanding behavioral processes,
from procurement of clay at specific geologic
sources to the deposition of broken ceramic
vessels in the archaeological record can allow
archaeologists to explain local adaptations and
historical developments in the societies of the
region. INAA analyses of Caribbean ceramics
permit an investigation into the technological
attributes of the decoration-based typological
categories of the ceramics. A diachronic
perspective based on the composition of
particular ceramic wares allows us to
investigate issues of the so-called decrease in
ceramic quality as well as the great cultural
diversity in the West Indies around AD 800
(e.g., see Keegan 2000:155). We know that
dramatic changes in subsistence and pottery
decoration mark the boundary between
Saladoid and Ostionoid cultures of Puerto
Rico (Keegan 2000:152). Is this shift
represented in the compositional record of the
sherds? It has been argued that cultural
pluralism is a key factor in understanding the
emergence of Taino chiefdoms (Wilson
1999:2). Is this pluralism represented in their
ceramic recipes? In conjunction with the
ethnohistoric sources (see e.g., Hofman and
Bright 2004) that exist for the region, the
compositional studies we propose stand to gain
from understanding pottery manufacturing in
changing and social and cultural environments
and understanding how pottery was negotiated
in African, European, and indigenous societies.


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Studies of Caribbean Ceramics

Relying on small data sets, the papers in this
issue use compositional data from ceramics to
address some of the major questions of the
region and offer preliminary interpretations. All
of the studies presented here are works in
progress, and will undoubtedly be expanded
upon by enlarging the sample sizes of the clays
and ceramics, and applying various analytical
techniques to the research questions. The work
by Isendoorn et al. presents preliminary results
of a case study on St. Lucia to address wider
concerns of reconstructing contact and/or
distribution networks in the Lesser Antilles
during the Ceramic Age. Their micro-regional
case study involving three archaeological
sites in southern St. Lucia finds that the
majority of ceramics originate from local clays
in the region. Siegel et al.'s exploratory paper
attempts to identify compositional signatures
for ceramics from various periods in Puerto
Rican prehistory to investigate exchange. They
conclude that ceramic production and exchange
remained at the local domestic level despite the
changing social and political contexts of the
island. Crock et al. with a dataset of Late
Ceramic Age ceramic sherds from Anguilla in
the British West Indies and the Salt River
Site in St. Croix of the U.S. Virgin Islands
depend on stylistic analyses and INAA
compositional work to suggest a degree of
specialized production and the use of a
limited number of ceramic recipes for both
island communities. They conclude that
ceramic recipes were limited in both island
contexts; the ceramic sample from St. Croix
appeared to have more of a local origin than
that from Anguilla.
Conrad et al. use the compositional
heterogeneity of Taino ceramics from the
Dominican Republic to argue that the site of La
Aleta served as a ceremonial center for a
regional rather than a local population.
Fitzpatrick et al. attempt to reconcile


Descantes et al.


differences in their petrographic and chemical
characterization interpretations of ceramics on
the island of Carriacou in the southern
Grenadines. To offer preliminary insights into
the production and exchange of ceramics on the
island of Carriacou in the southern Grenadines
from ca. AD 400-1200, they suggest several
different possible exotic sources for their
ceramic sample.
Several papers in the volume address issues
in the more recent historical past. Kelly et al.
investigate low-fired earthenwares to address
issues of exchange and interaction by focusing
on the enslaved peoples of African descent in
Guadeloupe during the French Colonial Period
of the eighteenth and nineteenth centuries.
Preliminary conclusions on the heterogeneity of
ceramic recipes imply complex interisland
trade of industrial wares. Ahlman et al.
examine the wares of enslaved Africans at
Brimstone Hill Fortress National Park on St.
Kitts from contexts dating from1790-1850 for
insights in their systems of production and
exchange. Incorporating possible local clay
sources in their compositional analysis, and
using a mineralogical and chemical
compositional approach, they interpret that
most of the ceramics in their sample were
locally made. They also suggest the possibility
of socio-economic differences amongst the
enslaved Africans at Brimstone Hill, where
slaves with specialized tasks at the fortress had
the means to purchase more expensive non-
local pottery than the less skilled laborers, such
as plantation slaves. Finally, Hauser et al. focus
on the mineralogical and chemical
compositions of Yabba ware for insights into
the eighteenth century craft production of
enslaved and free Jamaicans. At this
preliminary stage, it appears that ceramic pot
recipes on the north and south coasts and the
central part of the island were similar,


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Studies of Caribbean Ceramics

indicating a larger than expected scale of
production.

Conclusion
Caribbean pottery technology has received
relatively little attention (Curet 1997:497). To
address this need, several analytical programs
have begun to use compositional data of
ancient and historic Caribbean material
culture to address the salient research
questions of the region. MURR has initiated a
new program for the chemical
characterization of ceramics using INAA, as
well as other analytical techniques at its
disposal such as inductively coupled plasma
mass spectrometry (ICP-MS), energy
dispersive X-ray fluorescence spectrometry
(EDXRF), and portable X-ray fluorescence
spectrometry (PXRF). All of the papers in
this volume were presented at the 2006
Society for American Archeology (SAA)
annual meetings in San Juan, Puerto Rico, in
a session entitled "An Exploratory Study into
the Chemical Characterization of Caribbean
Ceramics: In Memory of James B. Petersen.
The INAA compositional data for eight of the
nine papers in this volume were generated at
MURR; the XRF data for one paper was
generated at Vrije Universiteit Amsterdam.
The research reported in this conference
session and the baseline ceramic
compositional data generated for the
Caribbean will serve as a foundation for future
research into the prehistory of this region. We
are optimistic that archaeologists are
beginning to recognize the value of
integrating such studies for understanding
important aspects of prehistoric human and
social dynamics. Our hope is that we will
foster greater collaboration between
archaeology and the physical sciences in this
region of the world by demonstrating the


Descantes et al.


intrinsic value of provenance studies to
understanding past human social dynamics.

Acknowledgments
We express our gratitude to the contributors
in this volume, whose participation has made
this collaborative study a success. We thank
Ronald Bishop of the Smithsonian Institution
who served as discussant for the session at
the SAA annual meetings in San Juan, Puerto
Rico and for his invaluable comments on the
papers in this volume. They provided a wide
array of research questions to the
archaeological ceramic sherds collected from
diverse Caribbean island locations and varied
temporal contexts. We also thank Jonathan
Dake, Mark Hammond, Nicole Little,
Rebecca Schmidt, and Tessa Schut for
carrying out the preparation and irradiation of
samples in all of the INAA studies. Operating
support for the MURR Archaeometry
Laboratory was provided by grants from the
National Science Foundation. The INAA
work was also partially funded by the US
Department of Energy Office of Nuclear
Energy, Science and Technology Award (No.
DE-FG07-03ID14531) to the Midwest
Nuclear Science and Engineering Consortium
under the Innovations in Nuclear
Infrastructure and Education program.
Finally, we would like to acknowledge the
important support and contributions of the
late Professors Irving Rouse and James B.
Petersen.

References Cited
Baxter, M. J.
1992 Archaeological uses of the Biplot-A
Neglected Technique? In Computer Applications
and Quantitative Methods in Archaeology, 1991,
edited by G. Lock and J. Moffett. BAR
International Series S577, 141-148. Tempvs
Reparatvm, Archaeological and Historical
Associates, Oxford.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Compositional Studies of Caribbean Ceramics

Bieber, A. M. Jr., D. W. Brooks, G. Harbottle, and E.
V. Sayre
1976 Application of Multivariate techniques to
Analytical Data on Aegean ceramics.
Archaeometry 18:59-74.
Bishop, R. L. and H. Neff
1989 Compositional Data Analysis in Archaeology.
In Archaeological Chemistry IV, edited by R. 0.
Allen, pp. 576-586. Advances in Chemistry Series
220, American Chemical Society, Washington,
D.C.
Bishop, R. L., R. L. Rands, and G. R. Holley
1982 Ceramic Compositional Analysis in
Archaeological Perspective. InAdvances in
Archaeological Method and Theory, vol. 5, pp.
275-330. Academic Press, New York.
Bohus L. Saho, Ma. M. Mackowiak de Antczak, E. D.
Greaves, A. Antczak, J. Bermudez, Zs Kasztovszky, T.
Poirier, A. Simonits
2005 Incipient Archaeometry in Venezuela:
Provenance Study of Pre-Hispanic Pottery
Figurines. Journal ofRadioanalytical and Nuclear
Chemistry 265(2):247-256.
Carini, S. P.
1991 Compositional Analysis of West Indian
Saladoid Ceramics and their Relevance to Puerto
Rican Prehistory. Ph.D. dissertation, Department
of Anthropology, University of Connecticut,
Storrs-Mansfield.
Cogswell, J., H. Neff and M. D. Glascock
1998 Analysis of Shell-Tempered Pottery
Replicates: Implications for Provenance Studies.
American Antiquity 63(1):63-72.
Crane, B. D.
1993 Colono Ware and Criollo Ware Pottery from
Charleston, South Carolina and San Juan, Puerto
Rico in Comparative Perspective. Ph.D.
dissertation, Department of American
Civilization, University of Pennsylvania,
Philadelphia.
Crock, J. G. and J. B. Petersen
2004 Inter-Island Exchange, Settlement Hierarchy,
and a Taino-Related Chiefdom on the Anguilla
Bank, Northern Lesser Antilles. In Late Ceramic
Age Societies in the Eastern Caribbean, edited by
A. Delpuech and C. L. Hofman, pp. 10-27. BAR
International Series 1273. Archaeopress, Oxford.
Cruxent, J. M. and I. Rouse
1969 Early Man in the West Indies. Scientific
American 221(5):42-52.


Descantes et al.


Curet, L. A.
1997 Technological Changes in Prehistoric
Ceramics from Eastern Puerto Rico: An
Exploratory Study. Journal of Archaeological
Science 24(6):497-504.
Descantes, C., R. J. Speakman, M. D. Glascock, and
D. V. Hill
2007 Chemical and Mineralogical Analyses of
Ceramics from the Indian Creek Site, Antigua:
Preliminary Results. In Proceedings of the 21st
Congress of the International Association for
Caribbean Archaeology (IACA). University of the
West Indies, St. Augustine, Trinidad and Tobago.
Fitzpatrick, S. M.
2000 Barbados Ceramics: Scientific Analysis-
Petrological and Chemical Analysis. In
Prehistoric Settlements in the Caribbean:
Fieldwork in Barbados, Tortola, and the Cayman
Islands, edited by P. L. Drewett, pp. 144-146.
Archetype Publications, London.
Glascock, M. D.
1992 Characterization of Archaeological Ceramics
at MURR by Neutron Activation Analysis and
Multivariate Statistics. In Chemical
characterization of ceramic pastes in
archaeology, edited by H. Neff, pp. 11-26.
Monographs in world archaeology no. 7.
Prehistory Press, Madison, Wis.
1998 Activation analysis. In Instrumental Multi-
Element Chemical Analysis, edited by Z. B.
Alfassi, pp. 93-150. Kluwer Academic Publishers,
Dordrecht and Boston.
Glascock, M. D. and H. Neff
2003 Neutron activation analysis and provenance
research in archaeology. Measurement Science
and Technology 14:1516-1526.
Glowacki, D. M. and H. Neff (eds.)
2002 Ceramic Production and Circulation in the
Greater Southwest: Source Determination by
INAA and Complementary Mineralogical
I,\,. ,,,. ,i, Monograph 44. Cotsen Institute of
Archaeology at University of California, Los
Angeles, Los Angeles.
Harbottle, G.
1976Activation Analysis in Archaeology.
Radiochemistry 3:33-72. The Chemical
Society, London.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Compositional Studies of Caribbean Ceramics

Hauser, M. W.
2001 Peddling Pots: Determining the Extent of
Market Exchange in E, h .... i i Century Jamaica
Through the Analysis of Local Coarse
Earthenware. Ph.D. dissertation, Department of
Anthropology, University of Syracuse, Syracuse.
Hofman, C. L., A. J. D. Isendoom and M. A. Booden
2005 Clays Collected. Towards an Identification of
Source Areas for Clays Used in the Production of
Pre-Columbian Pottery in the Northern Lesser
Antilles. Leiden Journal of Pottery Studies 21:9-
26.
Hofman, C. L. and A. J. Bright
2004 From Suazoid to Folk Pottery: Pottery
Manufacturing Traditions in a Changing Social
and Cultural Environment on St. Lucia. New West
Indian Guide 78(1&2):5-35.
Keegan, W. F.
1995 Modeling Dispersal in the Prehistoric West
Indies. World Archaeology 26(3):400-420.
2000 West Indian Archaeology 3: Ceramic Age.
Journal ofArchaeological Research 8(2): 135-167.
Knippenberg, S.
2006 Stone Artefact Production and Exchange
Among the Northern Lesser Antilles. Amsterdam
University Press, Amsterdam.
Leese, M. N. and P. L. Main
1994 The Efficient Computation of Unbiased
Mahalanobis Distances and their Interpretation in
Archaeometry. Archaeometry 36: 307-316.
Lundberg, E. R., J. H. Burton, and W. C. Lynn
2002 Investigation of Ceramic Variability at the
Tutu Site through Acid-Extraction Elemental
Analysis. In The Tutu Archaeological Village Site:
A Multidisciplinary Case Study in Human
Adaptation. Vol. 2, edited by M. K. Sandford, pp
166-198. Routledge, London and New York.
Neff, H.
1994 RQ-mode Principal Components Analysis of
Ceramic Compositional Data. Archaeometry
36:115-130.
2000 Neutron Activation Analysis for Provenance
Determination in Archaeology. In Modern
analytical methods in art and archaeology.
Chemical Analysis. Volume 155, edited by E.
Ciliberto and G. Spoto, pp. 81-134. John Wiley &
Sons, Inc., New York.
2001 Quantitative Techniques for Analyzing
Ceramic Compositional Data. In Ceramic


Descantes et al.


Production and Circulation in the Greater
Southwest: Source Determination by INAA and
Complementary Mineralogical I-,, 1 '1 ,
edited by D. Glowacki and H. Neff, pp. 15-36.
Monograph 44, Cotsen Institute of Archaeology,
UCLA.
Nichols, D. L., E. M. Brumfiel, H. Neff, M. Hodge,
T. H. Charlton, and M. D. Glascock
2002 Neutrons, Markets, Cities, and Empires: A
1000-Year Perspective on Ceramic Production
and Distribution in the Postclassic Basin of
Mexico. Journal ofA miil ,. ..... 1
Archaeology 21:25-82.
Padilla, R., P. J. M. Van Espen, R. R. Pla, E. M.
Rossi, R. A. Delgado, P. P. G. Torres and M. C.
Gonzalez
2003 Compositional Classification of
Archaeological Pottery Based on INAA and
SEM-EDX. Journal of Trace and Microprobe
Techniques 21(4):677-695.
Padilla, R., P. Van Espen, P. P. Godo Torres
2006 The Suitability of XRF Analysis for
Compositional Classification of Archaeological
Ceramic Fabric: A Comparison with a Previous
NAA Study. Analytica Chimica Acta 558:283-
289.
Rouse, I.
1992 The Tainos: Rise and Decline of the People
Who Greeted Columbus. Yale University Press,
New Haven.
Sayre, E. V.
1975 Brookhaven Procedures for Statistical
Analyses of Multivariate Archaeometric Data.
Brookhaven National Laboratory Report BNL-
23128. New York.
Speakman, R. J., and M. D. Glascock (editors)
2007 Special Issue: Fifty Years of Neutron
Activation Analysis in Archaeology.
Archaeometry 49(2), entire issue.
Steponaitis, V., M. J. Blackman, and H. Neff
1996 Large-Scale Patterns in the Compositional
Chemical Composition of Mississippian
Pottery. American Antiquity 61:555-572.
Veloz, M. M. and B. Vega
1982 The Antillean Perceramic: A New
Approximation. Journal oj ...- World
Archaeology 33-44.


Journal of Caribbean Archaeology, Special Publication ', 2008











Compositional Studies of Caribbean Ceramics

Weigand, P. C., G. Harbottle, and E. V. Sayre
1977 Turquoise Sources and Source Analysis:
Mesoamerica and the Southwestern U.S.A. In
Exchange Systems in Prehistory, edited by T. K.
Earle and J. E. Ericson, pp. 15-34. Academic
Press, New York.
Wilson, S. M., H. B. Iceland, and T. R. Hester
1998 Preceramic Connections between the
Caribbean and the Yucatan Peninsula. Latin
American Antiquity 9:342-352.
Wilson, S. M.
1999 Cultural Pluralism and the Emergence of
Complex Society in the Greater Antilles. Paper
presented at the XVIII International Congress for
Caribbean Archaeology, St. George's, Grenada,
July 1999.
Winter, J. & Gilstrap, M.
1991 Preliminary Results of Ceramic Analysis and
the Movements of Populations into the Bahamas.
In Proceedings of the 12th International Congress
of Caribbean Archaeology, edited by L. S.
Robinson. Martinique: International Association
for Caribbean Archaeology, 317-386.


Descantes et al.


Journal of Caribbean Archaeology, Special Publication 2, 2008








Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776


BACK TO THE SOURCE: PROVENANCE AREAS OF CLAYS
AND TEMPER MATERIALS OF PRE-COLUMBIAN CARIBBEAN CERAMICS

A.J. Daan Isendoorn
Faculty of Archaeology
University of Leiden
Reuvensplaats 3-4, Room 120
2311 BE Leiden
a.j. d. isendoorn @arch. leidenuniv. nl

Corinne L. Hofman
Faculty of Archaeology
University of Leiden
Reuvensplaats 3-4, Room 14
2311 BE Leiden
c. I. hofinan@arch. leidenuniv. nl

Mathijs Booden
Department of Geology
School of Geography
Geology and Environmental Science
University of Auckland
Auckland, New Zealand
m. booden(kauckland. ac. nz

Current knowledge regarding the provenance of pre-Columbian ceramic raw materials, i.e.
clays and temper materials, and the identification of exchange wares is largely insufficient.
Only in rare cases have technological aspects such as the temper constituents of the pottery
been analyzed, and even then rarely combining conventional and archaeometric techniques. In
this paper the methodology of a project that started at Leiden University in 2004 will be
discussed and illustrated by the preliminary results obtained from a micro-regional case study
carried out in mvi'thei I St. Lucia. Potsherds from three archaeological sites have been
analyzed. The results show that pre-Columbian potters on the island made use of both locally
available as well as non-local clays.


Leiden University initiated ceramic since. This research formed the impetus for a
technology research in the Lesser Antilles in PhD project on the provenance and exchange
the late 1980s. Results of these studies, of ceramics (raw materials and finished
mainly based on conventional methods products) within the Lesser Antillean
involving workability tests, fabric analysis, archipelago combining conventional and
and thin-sectioning have been published ever archaeometric techniques. The project is part


Journal of Caribbean Archaeology, Special Publication #2, 2008









Provenance Areas of Clays and Temper Materials

of the pluri-annual VIDI-program on
Mobility and Exchange in the pre-Columbian
Caribbean (started in 2004 and funded by the
Netherlands Foundation for Scientific
Research. It is directed by Professor Corinne
L. Hofman).
The main objective of the PhD project is to
identify clay sources on the islands and
reconstruct contact and/or distribution
networks in the Lesser Antilles between 400
BC and AD 1492.
A combined approach is chosen to tackle
these objectives. These are the collection of
clay and pottery samples from the various
islands, testing the workability properties of
clays, determining the mineral constituents of
both clays and pottery samples, and analyzing
the chemical composition of the clay and
pottery by means of conventional as well as
archaeometric techniques such as X-ray
fluorescence (XRF), and in the future,
thermal ionization mass spectrometry
(TIMS). In combination, these methods
complement each other rather than cancel
each other out (see Arnold et al. 1991; Bishop
et al. 1982).
Contemporary clay sources have been
sampled and clays and potsherds from several
islands of the Lesser Antilles have been
subjected to this combined approach (see
Stark 1966; Hofman et al. in press for similar
sampling methods). To date, 210 clays and
600 potsherds have been collected from
twelve islands in the Lesser Antilles. These
are currently being submitted for analysis at
the laboratories of the Faculty of
Archaeology at Leiden (Laboratory of
Ceramic Studies) and at the Faculty of Life
and Earth Sciences at the Vrije Universiteit
Amsterdam.
So far, the collection of clay samples and
potsherds, needed for the project, has been


Isendoom et al.


completed. The analysis of all clay samples
and of the potsherds, from the northern
Lesser Antilles, has also been finalized. All
analyses on the clay samples and the
potsherds that are described in this paper are
finalized1. Analyses of the remaining samples
will be conducted during the coming year and
are expected to provide an overall picture of
the provenance and distribution of ceramic
raw and/or finished products throughout the
Lesser Antilles.

Clay Sampling Strategies
Between 2004 and 2006 three field trips
were made in order to collect clay and pottery
samples from various islands in the chain. In
addition, pottery samples were obtained from
fellow institutions such as Yale University,
the University of the West Indies, Trinidad,
the Florida Museum of Natural History,
Calgary University, the Barbados Museum
and Historical Society, the St. Lucia
Archaeological and Historical Society, the
University of Vermont and the Antigua
Museum. This was accomplished with the
cooperation of many colleagues: Reg and
Nicky Murphy, Mary Hill Harris, Peter
Harris, Bill Keegan, Joe Moravetz, John
Crock, Birgit Faber Morse and Eric Branford.
Sampling of clay and pottery has been
carried out in the main geological regions that
characterize the Lesser Antillean archipelago;
the southern continental islands (Trinidad and
Tobago) (Boomert 2000, van Soest 2000), the
southern to central volcanic islands (Grenada,
St. Vincent and St. Lucia) and the central to
northern split-arc region with volcanic islands
(Guadeloupe's Basse Terre, St. Eustatius and
Saba) and limestone/extinct-volcanic islands
(Guadeloupe's Grande Terre, Antigua, St.
Martin and Anguilla). Barbados, the sub-
aerial expression of the sedimentary wedge
that overlies the southern to central part of


Journal of Caribbean Archaeology, Special Publication ', 2008









Provenance Areas of Clays and Temper Materials

the Lesser Antilles subduction zone, has also
been included. The geological survey was
conducted in collaboration with the Vrije
Universiteit Amsterdam (see also Hooijkaas
and Booden 2004)and the aim of the surveys
was to collect a representative sample of the
clay deposits on each island, by taking into
account the great variability of geological
formations. Sampling procedures were
proposed for each island and were designed
beforehand on the basis of archaeological
inventories, soil maps, and geological maps.
Furthermore, the sampling was aimed at
establishing an overall geochemical picture of
the variety in clay deposits present. A final
aim was to search for specific clay sources in
the vicinity of known archaeological sites.
Potsherds from three archaeological sites
on St. Lucia were analyzed. These include
Black Bay, a late Cedrosan Saladoid site (AD
400-600), Giraudy and Saltibus Point, multi-
component sites containing late Cedrosan
Saladoid and Suazan Troumassoid
components (AD 400-1200).
Of the 104 clay samples that were collected
on St. Lucia (Figure 1) in 2004, 12 samples
are geographically associated with the three
archaeological sites. Sampling of the clays
was carried out according to the sampling
strategy and clay sources were found in a
variety of locations such as inland hillsides
and valleys. Clays were often collected on
roadside cuts. Furthermore agricultural fields
and meadows were seen as possible sampling
areas, but to avoid contamination, the
samples were preferably collected on ridges,
because contamination of a sample can come
from fertilizers, the presence of industrial
waste dumps, waste from villages such as
laundry detergent, sea spray, and nearby
construction sites.


Isendoom et al.


Workability Tests and Experiments
To evaluate the suitability of the clay
mixtures for Amerindian manufacturing
techniques, the workability and plasticity
properties of the clays were tested by making
testpots and testbars (Hofman et al. 1993;
Hofman and Jacobs 2000/2001). Of the
twelve clays, five clays from Black Bay, two
from Giraudy and two from Saltibus point
were well suited for coil building. In the case
of the Black Bay clays, the coils tended to
adhere to each other easily while the paste
can be smeared to fix the coils firmly
together. They are also well suited for
pinching and modeling. On some of the
samples scraping and tapping could easily be
executed. These clays contained "enough
bones" and therefore probably would be
suitable for coiling and the production of
large vessels. The amount of natural grains
prevented the clays from shrinking too much
upon drying and firing of the testbars.
Although, these clays already have enough
plasticity, some of them might even improve
in workability with aging. Two of the clay
samples (STL-1 and STL-3) improved with
the addition of some fine sand. Furthermore
the presence of fine divided organic material
in the clay is recognizable on the testbars.
Testbars in this case show a black core after
firing at 6500C under oxidizing conditions.
Three of the clay samples (STL-2, STL-31
and STL32) contain clear quartz crystals (i.e.
approx. 6%), which resembled the fabric of
the analyzed potsherds from Black Bay.
The clays collected near Saltibus Point
(STL-19 and STL-20) also proved to be well
suited for coiling, pinching and modeling.
They have good workability properties and
the fabric of one of them (STL-20) resembles
one of the potsherds from the Saltibus Point
site (SP-175). Both the clay and the potsherds


Journal of Caribbean Archaeology, Special Publication ', 2008









Provenance Areas of Clays and Temper Materials


Alluvium, river sediment, beaches
Rhyolite
Andesite, altered andesite
Andesite agglomerate
Andesite porphyry
Belfond pumice flows
Caldera andesite agglomerate
Agglomerate tuffs
Andesite
Andesite breccia
Basalt, minor andesite
Agglomerate tuffs
Porphyritic/aphyric basalt


Um Dome lavas
Z_ Piton agglomerat
| Belfond pumice f:

Caldera rim fault


all


-K7-1

STL-20
0 15 km Q K- Saltibus Point
Black Bay a x-2
| STL-1,K3-2 sTL-2,STL3,STL-?1,5TL-32 ..-. Giraudy


Figure 1. Geological map of St. Lucia, after Tomblin (1964). Sample sites are indicated and
samples and settlements discussed are identified.


contain iron-oxide siltstone. Experiments
showed that the clay's workability could be
improved by the addition of a shorter clay,
that is, a clay of low plasticity, or some fine
sand.
The clay mixtures were experimentally
fired to test their firing and post-firing
behaviors and to facilitate comparison with


the pre-colonial potsherds. It is remarkable
that most of the St. Lucia clays could be used
to produce a well-fired pottery at relatively
low firing temperatures of 650C. This
implies that the pre-colonial pottery that was
manufactured with clays from St. Lucia clays
also could have been fired at such low
temperatures.


Journal of Caribbean Archaeology, Special Publication ', 2008


H

U


H



U
U
H--
Um
U--
U
U

m
m
U


Isendoorn et al.


e









Provenance Areas of Clays and Temper Materials

Fabric Analysis
Mineralogical fabric analysis2 was carried
out to identify the mineral and non-mineral
constituents in the clays and potsherds as has
been done in other regional ceramic studies
(e.g., Bullen and Bullen 1968, 1972;
Donahue et al. 1990).
Fabric analysis on 19 potsherds from the
sites of Black Bay, Giraudy and Saltibus
Point demonstrates that the majority of the
potsherds have a mineral assemblage
characteristic of volcanic islands in the region
in general.
The fabrics of the testbars that were made
with these clays resemble the fabrics of the
potsherds from the three sites. Typically the
same grain types occur in the various fabrics.
Angular crystals and broken grains of
transparent quartz are abundant in all the
sherds (i.e., approx. 6%). Additionally,
crystalline or partly weathered grains of
feldspar, kaolinite and iron-oxide siltstone are
present in all the samples albeit in lower
numbers (i.e., approx. 2%-4%).Dark minerals
like pyroxene and amphibole were found in
most of the sherds, but predominantly in
relatively low numbers (i.e., approx. 4%-
6%). They showed somewhat higher
percentages in only a few sherds from
Saltibus Point. The homogeneity of grain
types suggests a local origin of the majority
of the pottery from Black Bay, Saltibus Point
and Giraudy.
Exceptions are several sherds from Giraudy
that show a non-local origin on the basis of
their fabric. For example one sherd (GIR-
192, a Cedrosan Saladoid sherd) shows
differences in the color of the fabric and the
size and sorting of the grains. This sherd is
clearly different compared to other sherds
from Giraudy suggesting the use of non-local
clay.


Isendoom et al.


Archaeometric Techniques
Similar to instrumental neutron activation
analysis (INAA), which was used for the
analyses that are presented in other papers in
this issue, X-ray fluorescence (XRF) can be
used to determine the provenance of clay and
pottery samples. In this project, XRF was
preferred because of its ability to process
large numbers of samples (Fitton 1997). The
results will eventually be combined with
those from TIMS, which determines the
abundance ratio between certain isotopes
within a sample. This technique will be used
in the future to determine, for example,
87Sr/8Sr, 207Pb/26Pb and 143Nd/144Nd ratios.
XRF is a relatively inexpensive method and
sample preparation is a relatively fast
process. Four grams of each sample were
needed. The technique is destructive, because
the material to be sampled must be in the
form of a fine and homogeneous powder.
However, the samples do not require any
special treatment after analysis so powdering
and pressing the powder to hard pellets for
use in the spectrometer are the only
preparation steps. There is no need to
dissolve the sample chemically, and the XRF
technique is insensitive to the bonding state
of elements. This ensures that absorption of
secondary X-rays within the sample is the
only main source of error. Because most
absorption is by major elements, the
measured major element abundances can be
used to correct for this effect. The resulting
data generally have analytical precision and
detection limits in the order of 2 ppm for
trace elements. Analytical accuracy is within
a few percent for most of those same
elements, the error being mainly due to
absorption of secondary x-rays by the pressed
powder matrix.


Journal of Caribbean Archaeology, Special Publication ', 2008










Provenance Areas of Clays and Temper Materials

A Philips Panalytical Magix'Pro XRF
spectrometer was used in this study to
determine the concentrations of over 30
selected major and trace elements in each
sample. The resulting data are analyzed
graphically and through multivariate
statistical analysis3.
Both clays and potsherds from Black Bay,
Saltibus Point, and Giraudy that were
submitted for fabric analyses were also
submitted for analysis by XRF. Major
elements were examined to determine the
degree of geochemical weathering of the
clays. It is important to determine the degree
to which the compositions of bedrock, clays
and sherds were altered in order to be able to
compare potsherds to clays on the basis of
their geochemical composition. In addition,


6


5


4
Nb/Th
3


2


1


0


0,0 0,2 0,4 0,6 0,8
TIO2JTh


Isendoorn et al.


the trace element ratios of elements not
susceptible to chemical weathering were used
to distinguish between sherds produced with
different source clays.
The graphs contain the element ratios of
TiO2/Th plotted against those of Nb/Th of the
whole sample4 (Figures 2 and 3). These
elements have been chosen on the basis of
their immobile behavior during chemical
weathering, and because they provided the
clearest results. Recently the study of element
ratios based on Ba, Y, and Zr abundances
have also provided promising results
(Hofman et al. 2005; Hofman et al. in press).
The graph shows two different
"composition lines" that are defined by the
majority of the samples (Figure 2). One line
is defined by a majority of potsherds, and its


1,0 1,2 1,4 1,6


Journal of Caribbean Archaeology, Special Publication 2, 2008


0
D *Sherds St. Lucia
0 Clays St. Lucia








0
0
0
D


0 0 E


ArW]


Figure 2. Graph showing the results of the XRF measurements for TiO2/Th and Nb/Th of all clay
samples and potsherds from St. Lucia.










Provenance Areas of Clays and Temper Materials


5.0



4.0



3,0
Nb/Th


2,0



1,0



0,0


TI02/Th


Figure 3. Graph showing the results of the XRF measurements for TiO2/Th and Nb/Th
of clay samples and potsherds selected from the sites Black Bay, Saltibus Point and
Giraudy.


lower end, by a minority of clays ("sherd
composition line"). The other consists of few
sherds and mostly clays ("clay composition
line").
Four sherds from Black Bay fall on the clay
composition line. Their chemical composition
is similar. The clays plot at different positions
in the Southern End Member range. The
associated sherds are scattered within the
same range but mostly plot at positions
intermediate between the clay samples. This
indicates that these sherds were all produced
from mixtures of locally available clays and
temper materials.
Seven of the sherds from Giraudy plot on
the sherd composition line in the graph
shown in Figure 2. These sherds are
apparently mixtures between clays from the


Southern End Member range and an as yet
unidentified clay source. The two clays from
Giraudy plot in the Southern End Member
range, but all except one sherd plot outside
this range on the high-Nb/Th array (see
Hooijkaas and Booden 2004). These sherds
are characterized by very low Th contents. It
is possible that the sherds are mixtures of
Southern End Member clay and a low-Th
component, which would suggest import of
material. Low-Th material is found on St.
Lucia in the Sulphur Springs hydrothermal
crater. Sulphur Springs-type material is a
potential explanation. Alternatively, the
results could suggest that Th bearing phases
may have been preferentially removed from
the clay or temper components during
production. The majority of sherds from
Giraudy form a distinct group. Two sherds


Journal of Caribbean Archaeology, Special Publication 2, 2008


*Sherds St. Lucia (Black Bay)
0Sherds St. Lucia (Saltibus Point)
OSherds St. Lucia (Giraudy)
* Clays St. Lucia (Black Bay)
O Clays St. Lucia (Saltibus Point)
n Clays St. Lucia (Giraudy)



GIR-192




R20
s.l *-.'. a
nD '-'"'s


Isendoorn et al.









Provenance Areas of Clays and Temper Materials

(GIR-192 and GIR-195) do not plot in this
group. These are also the only Giraudy sherds
that do not have a low Th concentration.
Seven sherds from Saltibus Point fall on the
sherd composition line indicating a mixture
of southern clay and either one of the
anomalous clays that fall on that line or an
unknown clay, presumably from the Sulphur
Springs area. Figure 3 provides a view on the
composition line of potsherds from the
archaeological sites that are discussed in this
paper and the clay samples that were
geologically associated with these sites. As
opposed to Figure 2, Figure 3 only shows the
results of part the sample and not the whole
sample.
In summary, the majority of the pottery
from the three St. Lucian sites appears to be
of local origin, as is to be expected, given that
sources providing excellent clays for the
production of pottery are abundant on this
island.
Three main provenance areas can be
distinguished on the island based on ratios of
immobile trace elements in clays: the
northern, central and southern parts with
increasing Nb/TiO2 ratios. The majority of
the sherds, showing a relatively uniform
geochemistry, are made of clays with a
southern St. Lucia provenance. Several
sherds, however, have compositions
suggesting that they were manufactured with
non-local clays or that their constituents have
been imported from as yet unidentified
sources in other parts of the island.

Concluding Remarks
The islands in the Lesser Antilles are very
diverse geologically. This diversity is
responsible for the variation in availability of
clay sources on the various islands. Volcanic
islands such as St. Lucia are extremely rich in


Isendoom et al.


suitable clays whereas in general limestone
islands such as Anguilla offer a more limited
number of clay sources. This differential
availability of clays must have entailed the
establishment of a network for the
procurement and distribution of pottery raw
materials and/or finished products, which
probably paralleled similar networks in which
lithic raw materials, exotics, perishable
materials and ideas were carried throughout
the archipelago.

Acknowledgments
The authors would like to acknowledge
Arie Boomert and Alistair Bright for their
useful comments and corrections of the
original English text. The authors also would
like to acknowledge fellow institutions: Yale
University, UWI, Florida Museum of Natural
History, University of Calgary, the Barbados
Museum, the St. Lucia Archaeological and
Historical Society, The University of
Vermont, and the Antigua Museum for the
pottery samples that they provided for the
analyses. Many colleagues (Reg and Nicky
Murphy, Mary Hill Harris, Peter Harris, Bill
Keegan, Joe Moravetz, John Crock, Birgit
Faber Morse and Eric Branford) facilitated
the cooperation with these institutions. The
authors would like to extend their
appreciation to them here. The XRF analyses
and fieldwork are funded by the NWO VIDI-
project "Mobility and exchange: dynamics of
social, material and ideological relations in
the pre-Columbian insular Caribbean"
directed by Professor Corinne L. Hofman and
by the Byvanck Fonds, to which the authors
are much indebted.


Journal of Caribbean Archaeology, Special Publication ', 2008











Provenance Areas of Clays and Temper Materials

References Cited
Arnold, D.E., H. Neff, and R.L. Bishop
1991 Compositional Analysis and "Sources" of
Pottery: An Ethnoarchaeological Approach.
American Aih. ,, ', New Series 93:170-90.
Bishop, R.L., R.L. Rands, and G.R. Holley
1982 Ceramic Compositional Analysis in
Archaeological Perspective. InAdvances in
Archaeological Method and Theory, edited by M.
B. Schiffer, pp. 275-330. vol. 5. Academic Press,
New York.
Boomert, A.
2000 Trinidad, Tobago and the Lower Orinoco
Interaction Sphere: An Archaeological/
Ethnohistorical Study. Cairi Publications,
Alkmaar, The Netherlands.
Bullen, A.K. and R.P. Bullen
1968 Two strategraphic tests at the Grand Anse Site,
St. Lucia. In The Amerindians in St. Lucia
(louanalao), edited by Revd. C. Jesse, pp. 24-41,
The Voice Publishing Company, Castries.
Bullen, R.P. and A.K. Bullen
1972 Archaeological Investigations on St. Vincent
and the Grenadines, West Indies. In William L.
Bryant Foundation American Studies Report No.
8, Orlando Fl.
Donahue, J., D.R. Watters, and S. Millspaugh
1990 Thin Section Petrography of Northern Lesser
Antilles Ceramics. Geoarchaeology 5(3):229-254.
Fitton, G.
1997 X-ray Fluorescence Spectrometry. In Modern
Analytical Geochemistry. An Introduction to
Quantitive Chemical Analysis for Earth,
Environmental and Materials Scientists, edited by
R. Gill, pp. 87-115. Addison Wesley Longman
Limited, Edinburgh Gate, Harlow, England.
Hofman, C.L.
1993 In Search of the Native Population ofPre-
colonial Saba (400 -1450 A.D.), Part One:
Pottery Styles and their Interpretations.
Unpublished PhD Thesis, Leiden University,
Leiden.
Hofman, C.L., A.J.D. Isendoorn and M.A. Booden
2005 Clays collected. Towards an Identification of
Source Areas for Clays used in the Production of
pre-Colonial Pottery in the Northern Lesser
Antilles. Leiden Journal of Pottery Studies 21:9-
26.
Hofman, C.L., A.J.D. Isendoorn, M.A. Booden and


Isendoorn et al.


L.F.H.C. Jacobs
2008 In Tuneful Threefold. Combining
Conventional Archaeological Methods,
Archaeometric Techniques and
Ethnoarchaeological Research in the Study of pre-
Colonial Pottery of the Caribbean. In: New
Methods and Techniques in the Study of
Archaeological Materials from the Caribbean,
edited by C.L. Hofman, M.L.P. Hoogland and
A.L. van Gijn, pp. 21-33. University of Alabama
Press, Tuscaloosa.
Hofman, C.L. and L.F.H.C. Jacobs,
2000/2001 The Dynamics of Technology, Function
and Style. A study of Early Ceramic Age Pottery
from the Caribbean. Newsletter of the Department
of Pottery Technology. Leiden University. 18/19:
7-43.
Hooijkaas, G.R. and M.A. Booden
2004 Provenance ofPre-colonial Ceramics on St.
Lucia, West Indies. Unpublished Master's Thesis,
Vrije Universiteit Amsterdam, Amsterdam.
Tomblin, J.F.
1964 The Volcanic History and Petrology of the
Soufriere region, St. Lucia. Unpublished PhD
Thesis, University of Oxford, Oxford.
Van Soest, M.C.
2000 Sediment subduction and crustal
contamination in the Lesser Antilles island arc:
the geochemical and isotopic imprints on recent
lavas and geothermal fluids. Amsterdam:
Netherlands Research School of Sedimentary
Geology, publication number 20000101.

Notes
1Fieldwork in 2006 was generously funded by the
Byvanck Fonds.
2Mineralogical fabric analysis was performed with a
binocular microscope with the use of direct light.
Samples were first cut with a diamond saw and treated
with sandpaper to obtain a flat surface. Additionally
they were re-fired at 700 C under oxidizing
atmosphere before analyzing them under the
microscope.
3Geochemical analyses were performed at the
Faculty of Earth and Life Sciences of the Free
University Amsterdam under the responsibility of
Prof. Gareth Davies.
4Zr/Th and Hf/Th have also been plotted against Nb/
Th, but TiO2/Th-Nb/Th provided the clearest results.


Journal of Caribbean Archaeology, Special Publication 2, 2008












Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776










































































This Page Intentionally Left Blank.



Journal of Caribbean Archaeology, Special Publication #2, 2008








Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776


PRE-COLUMBIAN POTTERY IN THE WEST INDIES:
COMPOSITIONAL CHANGE IN CONTEXT

Peter E. Siegel
Department of Anthropology
Montclair State University
Montclair, NJ 07043
siegelp@mail. montclair. edu

Christophe Descantes
Archaeological Research Facility
University of California Berkeley
2251 College Building
Berkeley, CA 94720-1076
cdescantes@berkeley. edu


Jeffrey R. Fergurson
Archaeometry Laboratory
University of Missouri Research Reactor
1513 Research Park Drive
Columbia, MO 65211
fergusonj@,missouri. edu


Michael D. Glascock
Archaeometry Laboratory
University of Missouri Research Reactor
1513 Research Park Drive
Columbia, MO 65211
glascockm@,missouri. edu


Raw material selection in the production of pottery was examined in the context of site location,
developing settlement hierarchies, and evolving institutional inequality in pre-Columbian
Puerto Rico. Distinctive sherds ranging in age from 200 BC to AD 1200 were selected from
various contexts in the Maisabel site. This time frame spans occupations of the earliest single-
village egalitarian communities (Hacienda Grande and Cuevas complexes) through the devel-
opment of multi-village territorial polities (Monserrate and Santa Elena complexes). lh'e, J
were also selected from Site HU-7, occupied during the transition between two late prehistoric
complexes, ca. AD 1000. All sherds were subjected to instrumental neutron activation analysis
to characterize compositional variability regionally and through time.


Pottery was of fundamental importance to
pre-Columbian cultures in the Caribbean.
Broad similarities in surface decorations,
vessel morphologies, and technology within
and across many islands of the archipelago
hint at social networks, through which ideas
and perhaps pots were exchanged. Dramatic
social and political changes occurred from


the early to late ceramic age in some parts of
the Caribbean, documented archaeologically
in site distributions; settlement organization;
and such domains of material culture as
ceramic styles, rock art, and iconography.
The degree to which raw materials or finished
products flowed through exchange networks
and within and across settlement systems or


Journal of Caribbean Archaeology, Special Publication #2, 2008









Pre-Columbian Pottery in the West Indies


polities is poorly understood. In the present
study, raw-material selection in the
production of pottery is examined in relation
to site location, developing settlement
hierarchies, and emergent social inequality in
Puerto Rico. Selected sherds from two pre-
Columbian sites were subjected to neutron
activation analysis to assess compositional
variability geographically and through time.
The paper is divided into six sections. First,
background is provided into Saladoid and
post-Saladoid (Ostionoid) social and political
organization in the West Indies; second, we
present the problem and hypotheses
addressed in this study; third, a brief review
of the methods used in the study are
discussed; results are detailed in the fourth
section, followed by a discussion section; the
paper closes with avenues for additional
research.

Social and Political Context of
Pre-Columbian Cultures in the Caribbean
The earliest ceramic-age colonists to the
West Indies departed from northeastern South
America approximately 2,500 years ago.
These people were horticulturalists, who
relied extensively on fishing and the
collecting of marine and terrestrial faunal
resources (deFrance 1989, 1990; deFrance et
al. 1996; deFrance and Newsom 2005;
Newsom and Wing 2004; Siegel 199 l1a,
1991b). They produced thin-walled
elaborately painted, incised, and modeled
ceramic vessels and figurines; fine
groundstone celts, adzes, beads, and amulets;
carved and ground shell, bone, and coral
objects; in addition to many everyday items
fabricated from stone, bone, shell, clay, coral,
wood, cloth, and feathers (Rouse 1992).
Similarities in material culture across sites
and through time provide the basis for
assigning the groups to a single series of


Saladoid cultures, named after the Saladero
type site excavated by Irving Rouse and Jose
Cruxent (1963). It is generally agreed that
Saladoid peoples displaced pre-existing
Archaic groups who were already occupying
the Caribbean archipelago. However, the
extent and nature of interactions between the
ceramic and lithic-age groups in the
Caribbean are poorly understood, and
recently have become the topic of
considerable interest (Rodriguez Ramos
2005; Siegel 1989; Siegel et al. 2005).
The earliest Saladoid colonists arrived to
Puerto Rico by approximately 200 to 300 BC.
For the next six to seven centuries (ca.
300/200 BC-AD 400), early Saladoid
(Hacienda Grande complex) groups occupied
sizable villages located in coastal to near-
coastal settings. It is likely that Saladoid
newcomers came to landscapes already
modified by the previous Archaic residents.
Well-developed Archaic occupations have
been documented in the southern portions of
the Caribbean (Allaire and Mattioni 1983;
Boomert 2000; Harris 1973; Williams 2003).
Emily Lundberg (1980:135) observed long
ago that "the first pottery-making people to
migrate into the West Indies did not move
into a vacuum. Archaic groups (or at least
people who made no use of pottery) were
living all along their pathway." In addition,
ceramic-age colonists brought with them
established ideas for how to make a living,
how to organize their villages, and how the
universe was structured. Ceramic
iconography and village organization are
vivid expressions of the Saladoid connection
to the South American tropical rainforest (de
Hostos 1919; Moravetz 2005; Roe 1989;
Siegel 1995, 1996, 1999).
Settlement patterns and burials indicate that
early-Saladoid social structure was based on


Journal of Caribbean Archaeology, Special Publication ', 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


Table 1. Ceramic-age chronology for Puerto Rico.

Period Date Range Cultural Complex Cultural Series
IIa ca. 200 BC-AD 400 Hacienda Grande Saladoid
IIb AD 400-600/700 Cuevas Saladoid
IIIa AD 600/700-900 Monserrate Ostionoid
IIIb AD 900-1200 Santa Elena Ostionoid
IV AD 1200-1500 Esperanza/Boca Chica Ostionoid/protohistoric


an egalitarian ethic. Institutional social
inequality was not a feature of early-Saladoid
society (Rodriguez L6pez 1990; Siegel 1993,
1995, 1996; Versteeg 1989). By about AD
400, we see an increase in the number of sites
and habitats occupied compared to previous
occupations. The late-Saladoid period (AD
400-600/700) was associated with continued
habitation of coastal areas, in addition to
substantial occupations in interior valley
settings (Curet et al. 2004; Rodriguez L6pez
1990; Siegel 2004).
The post-Saladoid occupations of the island
are associated with an explosion in the
frequency of sites and site types. At this time,
formal civic-ceremonial plazas were
constructed in a number of settlements.
Combining lines of evidence from site
locations, relative site sizes, architectural and
structural organization, and mortuary patterns
there appear to have been fundamental
transformations in social relations beginning
around AD 700 (Curet and Oliver 1998;
Oliver 1998; Siegel 1996, 1999, 2004). Given
the rather coarse chronology that we
currently work with, where our finest degree
of control is no better than two to three
hundred years, the underlying shift in social
organization was probably more gradual than
it appears to us archaeologically (Table 1).
Over the span of about seven centuries, from
ca. AD 700 to AD 1400, the cultural


landscape of Puerto Rico progressed through
a series of gradual but dramatic shifts.
Tracking the locations of civic-ceremonial
centers, as a proxy for mapping the political
geography of the island, we see power
initially broadly dispersed in the south and,
through time, increasingly concentrated in the
high interior mountains (Curet et al. 2004;
Siegel 1999; Torres 2005). This trend was
"associated with [the establishment] of well-
defined group territories, increased solidarity
among group members, and notions of
exclusive rights over resources, land, and
people" (Siegel 2004:93). Ethnohistoric
documents reveal tensions between groups,
ranging from low-level rivalries to casual
feuding to out-and-out warfare and military
campaigns of conquest (Siegel 2004:89-90).
Numbers of sites in general, and ball courts/
ceremonial plazas in particular, increased to
their greatest levels by the Esperanza
(protohistoric) period. From demographic
trends apparent in the regional archaeological
database, we might infer that population
increases within the geographically
circumscribed border of the Puerto Rico
coastline combined with emergent and
aspiring leaders in post-Saladoid times were
responsible for changes in sociopolitical
organization.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


The Problem
In this context of institutional inequality
fundamental questions are raised regarding
the circulation of materials and loci of
production. Unambiguous ceramic-style
changes have been documented with the shift
from single-village communities to multi-
village polities (e.g., Roe 1989; Rouse 1992).
Assemblages from early-Saladoid sites
suggest a strong local focus on food supplies,
combined with systematic long-distance
exchange for semi-precious stones used in
craft production (Cody 1991; deFrance and
Newsom 2005). In the post-Saladoid world,
settlement patterns evince a linked and
hierarchical system of villages and camps
that eventually formed into territorial chiefly
polities (cacicazgos), well described in the
sixteenth-century Spanish accounts.
The trajectory of ceramic-style shifts are
well documented over approximately 2,000
years in the Caribbean, from about 500 BC to
AD 1500 (Petersen et al. 2004; Rainey 1940;
Roe 1989; Rouse 1992). In short, the early-
Saladoid series of styles include elaborately
decorated and technically sophisticated
vessels in a variety of forms, ranging from
bottles with multiple carinations, open bowls,
and restricted flying-saucer shaped bowls. In
contrast, the later Ostionoid series of styles
are characterized by less complexly decorated
and simpler vessel forms. The best
explanation that I've heard for this
revolutionaryy" shift relates to concomitant
changes in social organization: egalitarian to
institutionalized inequality (Roe 1989, 2005).
That is, with social and political changes
there were distinctive material
transformations in how power and prestige
were expressed and displayed. In Saladoid
communities, where power was of the
achieved variety, the locus of prestige was in


the small personal-presentation realm of
material culture (Roe 1989, 2005). We find
exquisitely carved and polished stone and
shell artifacts and fine pottery. At the scale of
small and portable, these diminutive objects
were designed to be admired up close and
personal. During the following Ostionoid
periods, in the context of developing chiefly
polities, the locus of power shifted from the
achievements of individuals to corporate
groups, where people were born into
positions of power and high status.
Materializations of this group power are seen
in the large easily visible and not-easily-
movable petroglyphs and ceremonial plazas.
It is in this context of shifting power
relations, developing settlement hierarchies,
and evolving institutional inequality that we
will address the production of pottery.
Maisabel is a large ceramic-age site located
on the north-central coast of Puerto Rico. It
was intensively occupied from about 200 BC
to AD 1200, spanning the full range of the
Saladoid period and much of the Ostionoid
period. Importantly, this occupational history
spans the transition from tribal egalitarian
communities to chiefdoms in Puerto Rico.
Maisabel was a highly structured village,
with a cemetery/plaza, series of mounded
middens, and residential area (Siegel 1992,
1995, 1996, 1999). Sherds were selected
distinctive of the Hacienda Grande, Cuevas,
Monserrate, Santa Elena, and Esperanza
styles from well-controlled contexts in the
site.
Site HU-7 is a small Ostionoid village or
camp located on the east coast of Puerto Rico
(Figure 1). The site contains a buried sealed
deposit of pottery that is stylistically
transitional between Monserrate and Santa
Elena (Siegel 2002). Sherds were selected
from several vessel types in the site to assess


Journal of Caribbean Archaeology, Special Publication ', 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


Figure 1. Map of Puerto Rico showing the locations of the Maisabel and HU-7 sites.


variability in pottery composition during this
period of polity formation and integrated
settlement hierarchies.
With the development of multi-village
polities and regional settlement hierarchies,
there is an expectation for the systematic
movement of goods within and across
settlement systems. The flow of tribute and
trade and exchange were important aspects of
Taino chiefly polities, especially in
negotiating alliances (Wilson 1990). With the
increasing importance of trade and tribute in
emerging and competitive polities we might
expect to find artifacts laden with symbols
and iconography to be moving through
exchange networks. Saladoid and Ostionoid
decorated pottery constitutes one class of
symbolically charged easily movable artifacts
and thus are uniquely appropriate to address
the flow of materials through and across
emergent polities (Roe 1989, 2005). Goals of


the research were to identify distinctive
compositional signatures in the fabrics of
sherds from the various time periods. Two
hypotheses guided this study:
Hi: Pottery vessels from early periods
exhibit a narrow range of compositional
variability compared to those from later
periods. Shifts in compositional
variability through time relate to changes
in settlement patterns, from single-
village communities to large multi-
village polities and the attendant
development of regional social
networks, through which pottery
circulated.
Ho: There is no discernible
compositional variability in sherds from
the various periods, suggesting that
pottery did not circulate through regional
social networks.
H2: The production, use, and ultimate


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


disposal of ceramic vessels were
tethered to individual villages or
communities, regardless of time period
and degree of social complexity.
Ho. Pottery was manufactured in places
different from where it was used and
eventually discarded.
Other relevant studies have been conducted
in the Lesser Antilles. In their petrographic
analysis of Saladoid and post-Saladoid
pottery from sites on Barbuda, Montserrat,
Anguilla, and St. Martin (northern Lesser
Antilles), Donahue et al. (1990) found
generally distinct temper associations on the
different islands. Saladoid and post-Saladoid
sherds from the Sufferers site on Barbuda
were compositionally/mineralogically
similar, leading them to conclude "that
virtually identical temper agents were being
used in the two periods," unless some sherds
were misidentified as to temporal placement
(Donahue et al. 1990:251). Viewing all
sherds as a group, they observed that post-
Saladoid pottery was more diverse in temper
associations than Saladoid pottery (Donahue
et al. 1990:252).
Based on her compositional analysis of La
Hueca and Hacienda Grande-style sherds
from the Hope Estate site on St. Martin,
Corinne Hofman (1999:184) concluded that
"although both styles are distinct in
decorative motifs, the composition of the
paste is often identical and that on this basis
the La Hueca pottery should indeed be
classified as a member of the Saladoid series,
rather than as a separate series." This
conclusion is consistent with Carini's (1991)
compositional analysis of early Saladoid
pottery. Although we are in agreement with
the conclusion that La Hueca pottery is of the
Saladoid series, we would only suggest that it
is entirely conceivable that potters of


different cultural series could have made pots
using similar, if not identical clay recipes,
especially if source materials were derived
from the same locales.

Methods
Fifty-three sherds from the Maisabel and
HU-7 sites were submitted to the University
of Missouri Research Reactor Center
(MURR) for instrumental neutron activation
analysis (INAA). Of these, 40 were collected
from Maisabel and 13 from Site HU-7.
Discussions of INAA, analytical procedures,
and sample preparations are presented by
Descantes and Glascock (2005) and
Descantes et al. (this volume).
The INAA produced concentration values
for 33 elements in most of the specimens
(Descantes and Glascock 2005:3). These data
were standardized to base-10 logarithms to
compensate for differences of magnitude
between major and trace elements and to
approximate normal distributions, especially
for the trace elements. Calcium levels for
most of the sherds were found to be quite
high, ranging from concentrations of 1
percent to values exceeding 25 percent
(approximate mean of 3%). This is not
surprising given the calcareous sediments and
rocks and karst topography that characterize
much of the coastal plains in Puerto Rico.
Calcium values were corrected using
MURR's UNSHELL program (Cogswell et
al. 1998). In their studies of the Maisabel and
HU-7 sherds, MURR staff employed
elemental plots to identify subgroups in the
compositional data compositionallyy
homogeneous groups within the analytical
database). Three compositional groups
ranged in size from 5 to 21 members, with 8
sherds not assigned to any group.


Journal of Caribbean Archaeology, Special Publication ', 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


Geological and Pedological Contexts
of the Maisabel and HU- 7 Sites
Puerto Rico is the smallest and most
easterly of the Greater Antilles. It is 3,421
square miles in area and is included in the
Greater Antilles Geologic Province. Puerto
Rico is surrounded by the Atlantic Ocean to
the north, Mona Passage to the west, the
Caribbean Sea to the south, and Vieques
Sound to the east. Approximately 75 percent
of the island is characterized by rugged
steeply sloped mountains. These mountains
form the Cordillera Central, the Sierra de
Luquillo, and the Sierra de Cayey. Much of
the island's periphery consists of a coastal
plain that averages approximately 5 km in
width. Although the climate of Puerto Rico in
general may be classified as humid sub-
tropic, considerable variation exists within
the island, from the mountains to the coast
and from the island's northern shore to its
more arid southwestern coastal plain.
Puerto Rico is divided into seven
physiographic regions: Mountain Uplands,
St. John Peneplain, Caguana Peneplain,
Foothill Zone, Interior Lowlands, Belted
North Coastal Zone, and Playas and Alluvial
Plains (Beinroth 1969; Mitchell 1954). The
Maisabel and HU-7 sites are located within
portions of the Playas and Alluvial Plains
physiographic region. Although most sections
of this region are situated along the north and
south coasts, sizable pockets of playas and
alluvial plains are found on the east and west
coasts. In general, the playas and alluvial
plains are flat and mostly sandy and located
along the mouths of larger streams. The
playas and alluvial plains date to the
Holocene and Pleistocene ages (Beinroth
1969:11).
The Maisabel site landscape encompasses
both beachfront and inland positions, each


with attendant geology, soil types, and
formation processes. An expansive mangrove
swamp at the mouth of the Rio Cibuco is
located about a half kilometer southeast of
the site. An assortment of calcareous
sediments and rocks occur in the north-
central coast region. These range from
Tertiary limestone in the higher uplands well
south of the coast to various Pleistocene and
Holocene terrace and beach deposits along
the coast itself (Beinroth 1969; Guillou and
Glass 1957). In the immediate vicinity of the
site the main geologic materials are
recognized simply as marine-terrace deposits
as well as eolian sands, typical for nearly any
low-lying coastline location.
Soil variability in the vicinity of Maisabel
is related to soil parent material type and
landscape position. Three main parent
material or deposit types occur within and
around the site. These consist of eolian sand
concentrated mainly between the shoreline
and the crest of an interfluve ridge, sandy
residual limestone comprising nearly all of
the upland landscapes south of the ridge, and
alluvial and peat deposits associated with the
Rio Cibuco and associated karst features.
Sizable deposits of clay are associated with
the Rio Cibuco and adjacent mangrove
swamp within 1 km of Maisabel.
Site HU-7 is located on the east-central
coast of the island, just south of the Rio
Ant6n Ruiz (Figure 1). The site is positioned
on sandy near-shore Holocene sediments,
approximately 300 meters from the beach. A
large area of tidal-marsh soils occurs west of
the site along with smaller pockets of salt
marsh. Portions of the tidal-marsh soils
developed on alluvium. The upland area
backing the site contains soils formed on
volcanic rocks. Deposits of clay and gravelly


Journal of Caribbean Archaeology, Special Publication ', 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


clay loam are located in close proximity to
the site (Boccheciamp 1977).

Results
Thirty-three elements were measured in 53
sherds selected from the Maisabel and HU-7
sites located on the north and east coasts of
Puerto Rico, respectively. Nickel was
subsequently deleted because of its absence
in most specimens (Descantes and Glascock
2005:5). The log-transformed dataset was
divided into two groups for an initial round of
analysis: Maisabel vs. HU-7 sherds.
Maisabel contains pre-Columbian
occupations spanning approximately 14
centuries, from ca. 200 BC to AD 1200.
These occupations are represented by the
Hacienda Grande, Cuevas, Monserrate, Santa
Elena, and Esperanza cultural complexes,
each associated with distinctive ceramic
styles. The test hypothesis states that potters


from each of the five cultural complexes
selected unique recipes in their manufacture
of ceramic vessels. If true, then the
compositional dataset should sort into five
groups of sherds, each representing a distinct
cultural complex or ceramic style. The null
hypothesis states that potters through time did
not create clay recipes distinctive of each
ceramic style.
Elemental plots and Mahalanobis distance
calculations using a subset of the principal
components were performed to define
compositionally homogeneous groups of
sherds (Descantes and Glascock 2005;
Ferguson 2007). Three compositional groups
were identified in the Maisabel sample of
sherds (Table 2, Figure 2). Group B, with 21
members, consisted of 9 Hacienda Grande-
style sherds, 3 or 4 Cuevas, 3 or 4
Monserrate, 2 Santa Elena, 2 Esperanza, and
1 undifferentiated Ostionoid (Table 2;


1.0 1.2 1.4 1.6 1.8 2.0 2.2
Chromium (log base-10 ppm)


2.4 2.6 2.8


Journal of Caribbean Archaeology, Special Publication 2, 2008


Group C


Figure 2. Compositional groups of Maisabel sherds. Unassigned samples shown (+).


Siegel et al.


I I











Pre-Columbian Pottery in the West Indies



Table 2. Compositional groups of the Maisabel sherds.


Group
B C D Unassigned
Cultural Cultural Cultural Cultural
I ii,,,.'-i Affiliation Affiliation Affiliation


SE
HG
HG
HG
HG
Mon
HG (ZIC)
HG (ZIC)
HG
Ost
Cu
Esp
Mon
Cu/Mon
Esp
Cu
Mon
SE
HG
Cu
HG


HG
Ost
Ost
SE
Mon


Cu
Cu
HG (ZIC)
HG (ZIC)
Mon
Mon
Mon


HG: Hacienda Grande complex
HG (ZIC): Hacienda Grande complex, zoned-incised cross-hatched
Cu: Cuevas complex
Mon: Monserrate complex
SE: Santa Elena complex
Esp: Esperanza complex
Ost: Ostionoid series


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.







Pre-Columbian Pottery in the West Indies


y "ll.

PUR 30


PUR 26 b

PUR 40


PUR 5


PUR 11


p
r~r~


PUR 23


PUR 21


PUR 2


PUR 29


Figure 3. Sample of sherds from Maisabel Compositional Group B.
Journal of Caribbean Archaeology, Special Publication 2, 2008


PUR 12


Siegel et al.










Pre-Columbian Pottery in the West Indies


Figure 4. Sample of sherds from Maisabel Compositional Group C.


PUR 33

Figure 5. Sample of sherds from Maisabel Compositional Group D.

Journal of Caribbean Archaeology, Special Publication 2, 2008


PUR 27


PUR 16


PUR 18


PUR 14


PUR 7


PUR 22


Siegel et al.


MINNOW









Pre-Columbian Pottery in the West Indies


Figure 3). Group C contained 1 sherd each of
the Hacienda Grande, Monserrate, and Santa
Elena styles and 2 of undifferentiated
Ostionoid (Figure 4). Group D consisted of 2
sherds each of Hacienda Grande, Cuevas, and
Monserrate and 1 of undifferentiated
Ostionoid (Figure 5). Eight sherds were
unassigned to any group and consisted of 1
Hacienda Grande, 1 Monserrate, 3 Santa
Elena, 2 Esperanza, and 1 undifferentiated
Ostionoid (Table 2; Figure 6). Based on
elemental compositions, there do not appear
to be unique recipes, by cultural complex, in
the production of ceramic vessels at
Maisabel. The null hypothesis is accepted in
this test.
A second hypothesis was posed to test for
geographic distinctions in ceramic
production: there are unique recipes in the
production of ceramic vessels based on
settlement locations. The null hypothesis
states that settlement location does not bear


on the compositional makeup of vessels. The
compositional databases of the Maisabel and
HU-7 sherds were combined to test the
hypothesis that pottery production is linked to
settlement location.
Four compositionally discrete groups and
one subgroup were identified (Figure 7).
Again, all ceramic styles in the assemblage
are represented in Groups B, C, and D.
Groups A and A2 consist exclusively of Site
HU-7 sherds. Two HU-7 sherds were placed
in Group D, along with 7 Maisabel
specimens. One HU-7 sherd was not assigned
to any group. Based on these data, the null
hypothesis is rejected. Regardless of time
period, geographic location of a settlement is
important in determining the compositional
makeup of the associated pottery.

Discussion
This pilot study resulted in unexpected
findings. Sherds were carefully selected from


Figure 6. Sample of Maisabel sherds unassigned to any compositional group.

Journal of Caribbean Archaeology, Special Publication 2, 2008


PUR 25


PUR 34


PUR 24 PUR 28PU8


Siegel et al.


PUR 8










Pre-Columbian Pottery in the West Indies


0
SC


+
Group C


T +
Group A2




V







Group A **


1.4


1.2 1.6 2.0 2.4
Chromium (log base-10 ppm)


Group D


Figure 7. Compositional groups of Maisabel and HU-7 sherds combined. Unassigned samples
shown (+).


Group


well-documented contexts in the Maisabel
site to investigate ceramic compositional
variability through approximately 14
centuries of pre-Columbian occupations,
spanning the Hacienda Grande, Cuevas,
Monserrate, Santa Elena, and Esperanza
cultural complexes. Current understanding of
ceramic-age settlement patterns and political
organization for Puerto Rico suggests that the
earliest (Saladoid) occupations (Hacienda
Grande [ca. 200 BC-AD 400] and to some
extent Cuevas [AD 400-600/700]) consisted
of egalitarian single-village communities.
Villages tended to be large and widely spaced
around the island in coastal to near-coastal
settings. The earliest formally constructed
ball courts on the island date to
approximately AD 700 (Monserrate


complex). During the post-Saladoid
occupations (Monserrate [AD 600/700-900],
Santa Elena [AD 900-1200], Esperanza/Boca
Chica [AD 1200-1500]), we see an explosive
increase in the number of sites and site types
and the development of settlement
hierarchies. Post-Saladoid settlement patterns
relate to the formation of and, through time,
increasingly competitive chiefly polities
(Siegel 2004).
Viewing for the moment the long trajectory
of the ceramic age from the perspective of the
terminal endpoints (Hacienda Grande vs.
Esperanza), we expected that ceramic
compositional variability would be lower at
the older single-village community terminus
(Hacienda Grande) than at the later
integrated-settlement hierarchy end


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


(Esperanza/protohistoric). This expectation
relates to the presumed systematic movement
of materials through the post-Saladoid
polities, especially in the context of tributary
relations (Moscoso 1986).
Results of the current study suggest that the
production and circulation of ceramic vessels
remained at the local domestic level, regardless
of the larger social and political context.
Individual groups of sherds from Maisabel,
produced by similarities in elemental
compositions, contained specimens from all
cultural complexes. Based on this analysis of
40 sherds from Maisabel, we conclude that the
occupants of the settlement procured clays and
temper materials from the local area
throughout the ceramic age. Importantly,
ceramic vessels found in the site apparently
were not transported from other places. When
elemental compositions of the Maisabel and
HU-7 sherds were combined into a single
analysis, there was nearly perfect segregation
of the sherds, by site, further support for the
emphasis on the local area in the production,
use, and final disposition of pottery throughout
the pre-Columbian occupations of Puerto Rico.
When multi-village polities formed during
post-Saladoid times some villages may have
maintained a certain degree of autonomy. For
example, Maisabel, and perhaps other large
Saladoid villages, continued to thrive as
relatively autonomous villages even as
networks of integrated settlements developed
across the post-Saladoid landscape of Puerto
Rico. Continuity of village autonomy would
explain the lack of differentiation in paste
content of ceramic vessels produced at
Maisabel from early to late ceramic times.
John Hoopes (2005:10) suggested a similar
phenomenon in the early formation of
chiefdoms for the Chibchan world of southern
Central America and northern Colombia:


"Although there are some indications of
settlement hierarchies, it remains to be
demonstrated that some centers exercised
political control over others and that some
villages lost autonomy as they were
incorporated into multivillage polities."

Future Research
This pilot study leads to three lines of
additional research that will help to clarify
the use of ceramic vessels during the ceramic
age: (1) clay sourcing, (2) functional
variability within ceramic assemblages, and
(3) regional variability in ceramic
assemblages.
Clay Sourcing
Results of the current study suggest that
settlement occupants relied on local sources of
clay in the production of pottery. In Maisabel,
ceramic vessels produced by artisans of all the
major ceramic-age complexes shared common
elemental compositional distributions. Except
for two sherds, the Maisabel and HU-7
samples sorted into two distinct ceramic
clusters.
It will be important to clarify raw material
sources in ceramic production by conducting
systematic soil surveys in relation to each
archaeological site. In particular, clay samples
need to be collected, analyzed for elemental
distributions, and compared to the database
created for the current study. Published soil
surveys may be used to guide this fieldwork
(Table 3; Acevido [1982]; Boccheciamp
[1977]). The soils map for the Maisabel setting
shows sizable patches of clay within 1 km and
especially within 3 km of the site (Table 4).
Functional Variability within Ceramic
Assemblages.
Sherds selected for the current study
generally targeted good examples of
distinctive styles recognized for Puerto Rico


Journal of Caribbean Archaeology, Special Publication ', 2008


Siegel et al.










Pre-Columbian Pottery in the West Indies


Table 3. Potential sources for clay for pre-Columbian pottery production in the vicinity of the Maisabel and HU-7
sites.

Maisabel HU-7


Bajura Deep, nearly level, and poorly
clay drained. Located on floodplains.
Permeability is slow and available
water capacity is high. Organic mat-
ter content is high

Jareales Deep, nearly level, and poorly
clay drained. Located in coastal low-
lands. Permeability is very slow and
available water capacity is high.
Organic matter content is high

Santa Clara Moderately deep, gently sloping to
clay sloping, and well drained. Located
on foot slopes and small hills. Per-
meability and available water ca-
pacity are moderate


Caguabo Shallow, well drained, and moderately
clay loam permeable. Formed from partly weath-
ered volcanic rocks. Moderate avail-
able water capacity


Coloso silty Deep, nearly level, somewhat poorly
clay loam drained. Located on floodplains and
occasionally flooded. Permeability is
slow. High available water capacity.


Fortuna
clay




Junquitos
gravelly clay
loam


Nearly level on river floodplains. Per-
meability is slow. High available water
capacity.



Moderately deep and well drained and
moderately slowly permeable. Located
on foot slopes in humid volcanic up-
lands. Formed in alluvial and colluvial
sediment derived from extrusive vol-
canic rocks. High available water ca-
pacity.


Pinones Deep, poorly drained, and very slowly
silty clay permeable. Located on coastal low-
lands. High available water capacity.


and neighboring islands. This selection
necessarily was weighted to decorated vessels
and thus undoubtedly limited the analysis to a
narrow range of functionally specific wares.
An important follow-up study will be to
systematically select a cross section of vessel
types, defined on the basis of morphology
and technology, from each of the cultural
complexes. Functionally specific wares may
be associated with distinct clusters of
elemental concentrations, which also may be
linked to specific clay sources identified in
the soil surveys.


Regional Variability in Ceramic Assemblages
A large follow-up study will entail the
systematic collection of sherds from other
sites in the vicinity of Maisabel and HU-7.
Maisabel is located in the Cibuco valley,
where there are numerous documented late
Saladoid and post-Saladoid sites. We have
good evidence for the gradual development
of settlement hierarchies and increasingly
integrated polities after the Hacienda Grande
period (Siegel 2004). In particular, it will be
important to analyze samples of pottery from


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.









Pre-Columbian Pottery in the West Indies


Table 4. Distribution of potential clay sources in the vicinity of Maisabel.

Catchment Bajura clay (m2) Jareales clay (i2) Santa Clara clay (m2)
Distance (km)

0.5 65,800

1.0 290,700

2.0 7700 96,700

3.0 973,400 403,300 93,000

Total 981,100 403,300 546,200


the large Paso del Indio site located a short
paddle ride up the Rio Cibuco from Maisabel
(Walker 2005) and samples from some of the
many smaller hinterland sites documented for
the valley (Siegel and Joseph 1993).
Site HU-7 represents a small household
settlement that was occupied for
approximately 10-20 years during the
transition between Periods Ilia and IIIb
(Monserrate to Santa Elena periods). The site
is located within the largest remaining
mangrove swamp on Puerto Rico. Based on
location, we can reasonably assume that the
inhabitants of Site HU-7 took advantage of
the bountiful resources surrounding them.
Further, the small settlement was
undoubtedly linked to a larger network of
settlements across the post-Saladoid
landscape. By Period IIIb, a site with a small
ceremonial plaza was located in Sabana
Arriba, approximately 22 km to the northwest
of Site HU-7 (Rodriguez L6pez and Rivera
1983). Assuming that ball court/plaza sites
served as political or administrative centers
then Sabana Arriba may have included the
east-central coast within its jurisdiction
during Period IIIb (Siegel 1999: Figure 6).
Social and economic relations revolving


around tribute may have been established by
Period IIIa, associated with the earliest
evidence for institutionalized social
inequality on Puerto Rico. As such, small
household-based camps dating to the post-
Saladoid occupations on the island must be
viewed as potential resource-extraction
places in the primary production of tribute. In
this regard, it will be important to compare
elemental distributions of pottery from sites
within and across polities.

Acknowledgments
The Maisabel site excavation, analysis, and
related research reported herein was supported
by the Centro de Investigaciones Indigenas de
Puerto Rico; National Science Foundation
(Doctoral Dissertation Research Grant BNS88-
22317); Wenner-Gren Foundation for
Anthropological Research (Grant No. 6167); H.
John Heinz III Fund Grant Program for Latin
American Archaeology; and the U.S. Army
Corps of Engineers, Jacksonville District
(Contract No. DACW17-93-C-0063). Site HU-
7 was excavated and analyzed with funding
from the U.S. Army Corps of Engineers,
Jacksonville District (Contract No. DACW 17-
98-D-0021-0002). John Milner Associates, Inc.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.










Pre-Columbian Pottery in the West Indies


and New South Associates, Inc. provided travel
support for Siegel to collect sherds to use in this
study. Funding to conduct the instrument
neutron activation analysis was provided by the
University of Missouri Research Reactor Center
(NSF grant BCS-05-04015). Addressing the
comments of two anonymous reviewers and Jeff
Speakman improved the paper from its original
version.

References Cited
Acevido, Gilberto
1982 Soil Survey of Arecibo Area oJ .... h., i Puerto
Rico. Soil Conservation Service, U.S. Department of
Agriculture, University of Puerto Rico, Mayaguez.
Allaire, Louis, and Mario Mattioni
1983 Boutbois et la Godinot: Deux gisements
aceramiques de la Martinique. Proceedings of the
International Congress for the Study of the Pre-
Columbian Cultures of the Lesser Antilles 9:27-38.
Montreal.
Beinroth, FriedrichH.
1960 An Outline of the Geology of Puerto Rico.
Bulletin 213. Agricultural Experiment Station,
University of Puerto Rico, Mayagtiez.
Boccheciamp, Rafael, A.
1977 Soil Survey ofHumacao Area of Eastern Puerto
Rico. Soil Conservation Service, U.S. Department of
Agriculture, University of Puerto Rico, Mayaguez.
Boomert, Arie
2000 Trinidad, Tobago and the Lower Orinoco
Interaction Sphere: An Archaeological/
Ethnohistorical Study. Cairi, Alkamaar, The
Netherlands.
Carini, Stephen P.
1991 Compositional Analysis of West Indian Saladoid
Ceramics and their Relevance to Puerto Rican
Prehistory. Ph.D. Dissertation, University of
Connecticut, Storrs. University Microfilms, Ann
Arbor.
Cody, Annie
1991 Distribution of Exotic Stone Artifacts through
the Lesser Antilles: Their Implications for Prehistoric
Interaction and Exchange. In Proceedings of the
Congress of the International Association for
Caribbean Archaeology 14:204-226. Barbados.


Cogswell, James W., Hector Neff, and Michael D.
Glascock
1998 Analysis of Shell-Tempered Pottery
Replicates: Implications for Provenance Studies.
American Antiquity 63:63-72.
Curet, L. Antonio, and Jos6 R. Oliver
1998 Mortuary Practices, Social Development, and
Ideology in Precolumbian Puerto Rico. Latin
American Antiquity 9:217-239.
Curet, L. Antonio, Joshua Torres, and Miguel
Rodriguez
2004 Political and Social History of Eastern Puerto
Rico: The Ceramic Age. In Late Ceramic Age
Societies in the Eastern Caribbean, edited by
Andre Delpuech and Corinne Hofman, pp. 59-85.
Paris Monographs in American Archaeology No.
14. BAR International Series 1273. British
Archaeological Reports, Archaeopress, Oxford.
deFrance, Susan D.
1989 Saladoid and Ostionoid Subsistence
Adaptations: Zooarchaeological Data from a
Coastal Occupation on Puerto Rico. In Early
Ceramic Population Lifeways and Adaptive
Strategies in the Caribbean, pp. 57-78, edited by
Peter E. Siegel. BAR International Series 506.
British Archaeological Reports, Oxford.
1990 Zooarchaeological Investigations of an Early
Ceramic Age Frontier Community in the
Caribbean, The Maisabel Site, Puerto Rico.
Antropol6gica 73-74:3-180.
deFrance, Susan D., William F. Keegan, and Lee A.
Newsom
1996 The Archaeobotanical, Bone Isotope, and
Zooarchaeological Records from Caribbean Sites
in Comparative Perspective. In Case Studies in
Environmental Archaeology, edited by Elizabeth
J. Reitz, Lee A. Newsom, and Sylvia J. Scudder,
pp. 289-304. Plenum Press, New York.
deFrance, Susan D., and Lee A. Newsom
2005 The Status of Paleoethnobiological Research
on Puerto Rico and Adjacent Islands. InAncient
Borinquen: The Archaeology and Ethnohistory of
Native Puerto Rico, edited by Peter E. Siegel, pp.
122-184. University of Alabama Press,
Tuscaloosa.
de Hostos, Adolfo
1919 Prehistoric Porto Rican Ceramics. American
J il ,.o ..i.. .,, ,t 21:376-399.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.











Pre-Columbian Pottery in the West Indies


Descantes, Christophe, and Michael D. Glascock
2005 Instrumental Neutron Activation Analysis of
Pre-Columbian Pottery from Puerto Rico, West
Indies. Unpublished manuscript, on file, Research
Reactor Center, University of Missouri, Columbia.
Donahue, Jack, David R. Watters, and Sarah
Millspaugh
1990 Thin Section Petrography of Northern Lesser
Antilles Ceramics. Geoarchaeology: An
International Journal 5:229-254.
Ferguson, Jeffrey R.
2007 Re-Evaluation of INAA Data for 53 Ceramic
Samples from Puerto Rico Submitted to Peter
Siegel. Unpublished manuscript, on file, Research
Reactor Center, University of Missouri, Columbia.
Guillou, Robert B., and Jewell J. Glass
1957 A Reconnaissance Study of the Beach Sands of
Puerto Rico. Geological Survey Bulletin 1042-1.
U.S. Geological Survey, Washington, D.C.
Harris, Peter O'B.
1973 Preliminary Report on Banwari Trace, a
Preceramic Site in Trinidad. In Proceedings of the
International Congress for the Study of the Pre-
Columbian Cultures of the Lesser Antilles 4:115-
125. St. Lucia.
Hofman, Corinne L.
1999 Pottery. InArchaeologicallf i. ,'it,. i,.i,a on
St. Martin (Lesser Antilles), edited by Corinne L.
Hofman and Menno L.P. Hoogland, pp. 149-187.
Archaeological Studies Leiden University. Faculty
of Archaeology, Leiden University, Leiden, The
Netherlands.
Hoopes, John W.
2005 The Emergence of Social Complexity in the
Chibchan World of Southern Central America and
Northern Colombia, AD 300-600. Journal of
Archaeological Research 13:1-47.
Lundberg, Emily R.
1980 Old and New Problems in the Study of
Antillean Aceramic Traditions. Proceedings of the
International Congress for the Study of the Pre-
Columbian Cultures of the Lesser Antilles 8:131-
138. Tempe, Arizona.
Mitchell, R.C.
1954 A Survey of the Geology of Puerto Rico.
Technical Paper 13. Agricultural Experiment
Station, University of Puerto Rico, Mayagtiez.


Moravetz, losif
2005 Imaging Adornos: Classification and
Iconography of Saladoid Adornos from St.
Vincent, West Indies. BAR International Series
S 1445. British Archaeological Reports,
Archaeopress, Oxford.
Moscoso, Francisco
1986 Tribuy classes en el Caribe tiiiii'.,
Universidad Central del Este, San Pedro de
Macoris, Repuiblica Dominicana.
Newsom, Lee A., and Elizabeth S. Wing.
2004 On Land and Sea: Native American Uses of
Biological Resources in the West Indies.
University of Alabama Press, Tuscaloosa.
Oliver, Jos6 R.
1998 El centro ceremonial de Caguana, Puerto
Rico: Simbolismo iconografico, cosmovisi6n y el
poderio caciquil Taino de Boriquen. BAR
International Series 727. British Archaeological
Reports, Archaeopress, Oxford.
Petersen, James B., Corinne L. Hofman, and L.
Antonio Curet
2004 Time and Culture: Chronology and
Taxonomy in the Eastern Caribbean and the
Guianas. In Late Ceramic Age Societies in the
Eastern Caribbean, edited by Andrd Delpuech
and Corinne Hofman, pp. 17-32. Paris
Monographs in American Archaeology No. 14.
BAR International Series 1273. British
Archaeological Reports, Archaeopress, Oxford.
Rainey, Froelich G.
1940 Porto Rican Archaeology. Scientific Survey
of Porto Rico and the Virgin Islands, vol. XVIII-
part 1. The New York Academy of Sciences, New
York.
Rodriguez L6pez, Miguel
1990 Arqueologia del Rio Loiza. Proceedings of
the Congress of the International Association for
Caribbean Archaeology 11:287-294. San Juan,
Puerto Rico.
Rodriguez L6pez, Miguel, and Virginia Rivera
1983 Sitio "El Destino," Vieques, Puerto Rico informed
preliminary. In Proceedings of the International
Congress for the Study of the Pre-Columbian Cultures
of the Lesser Antilles 9:163-172. Montrial.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.











Pre-Columbian Pottery in the West Indies


Rodriguez Ramos, Reniel
2005 The Crab-Shell Dichotomy Revisited: The
Lithics Speak Out. InAncientBorinquen: The
Archaeology and Ethnohistory of Native Puerto
Rico, edited by Peter E. Siegel, pp. 1-54.
University of Alabama Press, Tuscaloosa.
Roe, Peter G.
1989 A Grammatical Analysis of Cedrosan
Saladoid Vessel Form Categories and Surface
Decoration: Aesthetic and Technical Styles in
Early Antillean Ceramics. In Early Ceramic
Population Lifeways and Adaptive Strategies in
the Caribbean, edited by Peter E. Siegel, pp. 267-
382. BAR International Series No. 506. British
Archaeological Reports, Oxford.
2005 Rivers of Stone, Rivers within Stone: Rock
Art in Ancient Puerto Rico. In Ancient Borinquen:
Archaeology and Ethnohistory of Native Puerto
Rico, edited by Peter E. Siegel, pp. 288-336.
University of Alabama Press, Tuscaloosa.
Rouse, Irving
1992 The Tainos: Rise and Decline of the People
Who Greeted Colombus. Yale University Press,
New Haven.
Rouse, Irving, and Jos6 M. Cruxent
1963 Venezuelan Archaeology. Yale University
Press, New Haven.
Siegel, Peter E.
1989 Site Structure, Demography, and Social
Complexity in the Early Ceramic Age of the
Caribbean. In Early Ceramic Population Lifeways
and Adaptive Strategies in the Caribbean, edited
by Peter E. Siegel, pp. 193-245. BAR
International Series No. 506. British
Archaeological Reports, Oxford.
1991a Migration Research in Saladoid Archaeology:
A Review. The Florida, i-h,, ."y.. ./.' 't 44(1):79-
91.
1991b On the Antilles as a Potential Corridor for
Cultigens into Eastern North America. Current
Si ,, / ...J, -. 32:332-334.
1992 Ideology, Power, and Social Complexity in
Prehistoric Puerto Rico. Ph.D. dissertation, State
University of New York at Binghamton.
University Microfilms, Ann Arbor.
1993 Saladoid Survival Strategies: Evidence from
Site Locations. In Proceedings of the Congress of
the International Association for Caribbean
Archaeology 14:315-337. Barbados.


1995 The Archaeology of Community
Organization in the Tropical Lowlands: A Case
Study from Puerto Rico. In Archaeology in the
Lowland American Tropics: Current Analytical
Methods and Recent Applications, edited by Peter
W. Stahl, pp. 42-65. Cambridge University Press,
Cambridge.
1996 Ideology and Culture Change in Prehistoric
Puerto Rico: A View from the Community.
Journal of Field Archaeology 23:313-333.
1999 Contested Places and Places of Contest: The
Evolution of Social Power and Ceremonial Space
in Prehistoric Puerto Rico. Latin American
Antiquity 10:209-238.
2002 Archaeological Data Recoveries at Sites HU-
6 and HU- 7. Rio Ant6n Ruiz Flood Control
Project, Municipio de Humacao, Puerto Rico.
Technical Report No. 862. New South Associates,
Stone Mountain, Georgia. Submitted to U.S.
Army Corps of Engineers, Jacksonville District.
2004 What Happened After AD 600 on Puerto
Rico? Corporate Groups, Population
Restructuring, and Post-Saladoid Social Changes.
In Late Ceramic Age Societies in the Eastern
Caribbean, edited by Andr6 Delpuech and
Corinne Hofman, pp. 87-100. Paris Monographs
in American Archaeology No. 14. BAR
International Series 1273. British Archaeological
Reports, Archaeopress, Oxford.
Siegel, Peter E., and J. W. Joseph
1993 Archeological Data Recovery at El Palmar de
las Animas (Site VB-27) and the Concrete Well
Site (Site VB-32), Rio Cibuco Flood Control
Project, Municipio de Vega Baja, Puerto Rico.
New South Associates Technical Report 206. New
South Associates, Stone Mountain, Georgia.
Submitted to the U.S. Army Corps of Engineers,
Jacksonville District. Copies available from the
Office of the State Historic Preservation Officer,
San Juan, Puerto Rico.
Siegel, Peter E., John G. Jones, Deborah M. Pearsall,
and Daniel P. Wagner
2005 Environmental and Cultural Correlates in the
West Indies: A View from Puerto Rico. In Ancient
Borinquen: The Archaeology and Ethnohistory of
Native Puerto Rico, edited by Peter E. Siegel, pp.
88-121. University of Alabama Press, Tuscaloosa.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.











Pre-Columbian Pottery in the West Indies


Torres, Joshua M.
2005 Deconstructing the Polity: Communities and
Social Landscapes of the Ceramic-Age Peoples of
South Central Puerto Rico. In Ancient Borinquen:
The Archaeology and Ethnohistory of Native
Puerto Rico, edited by Peter E. Siegel, pp. 202-
229. University of Alabama Press, Tuscaloosa.
Versteeg, Aad H.
1989 The Internal Organization of a Pioneer
Settlement in the Lesser Antilles: The Saladoid
Golden Rock Site on St. Eustatius, Netherlands
Antilles. In Early Ceramic Population Lifeways
and Adaptive Strategies in the Caribbean, edited
by Peter E. Siegel, pp. 171-192. BAR
International Series No. 506. British
Archaeological Reports, Oxford.
Walker, Jeff
2005 The Paso del Indio Site, Vega Baja, Puerto
Rico: A Progress Report. In Ancient Borinquen:
The Archaeology and Ethnohistory of Native
Puerto Rico, edited by Peter E. Siegel, pp. 55-87.
University of Alabama Press, Tuscaloosa.
Williams, Denis
2003 Prehistoric Guyana. Ian Randle, Kingston.
Wilson, Samuel M.
1990 Hispaniola: Caribbean Chiefdoms in the Age
of Columbus. University of Alabama Press,
Tuscaloosa.

Notes
1 Twenty-one sherds contained nickel; all except for
one came from Maisabel. The Maisabel sherds with
nickel spanned all of the cultural complexes.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Siegel et al.








Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776


PRELIMINARY INTERPRETATIONS OF CERAMIC COMPOSITIONAL ANALYSIS
FROM LATE CERAMIC AGE SITES
IN ANGUILLA AND THE SALT RIVER SITE IN ST. CROIX

John G. Crock
Department of Anthropology
University of Vermont
111 Delehanty Hall
Burlington, Vermont 05405
john.crock@uvm.edu


Birgit F. Morse
Curatorial Affiliate
Yale Peabody Museum
145 Rocky Hill Road
Nelson, NH 03457
birgitFM@aol. com

James B. Petersen (deceased)
Department of Anthropology
University of Vermont
111 Delehanty Hall
Burlington, Vermont 05405


Christophe Descantes
Archaeological Research Facility
University of California Berkeley
2251 College Building
Berkeley, CA 94720-1076
cdescantes@berkeley. edu

Michael D. Glascock
Archaeometry Laboratory
University of Missouri Research Reactor
1513 Research Park Drive
Columbia, MO 65211
glascockm@missouri. edu


Late Ceramic Age ceramic samples from five sites in Anguilla, British West Indies, and from
the two most recent occupations at the Salt River Site, St. Croix, U.S. Virgin Islands, were
analyzed by instrumental neutron activation analysis (INAA) at the University of Missouri
Reactor Center (MURR). Compositional analysis of ceramic sherds from various
archaeological contexts on these two islands help address issues related to ceramic production,
interisland trade and exchange, and cultural affiliation.


Historically, ceramic research in the
Caribbean has focused on establishing and
refining typological and chronological
sequences based on decorative style and
vessel shape (cf. Rouse 1992). This necessary
and essential work has formed the basis for
reconstructions of pre-Columbian migration
patterns and general spheres of cultural
interaction. Unfortunately, due to the regional
homogenization of ceramic form and


decoration, style-based analyses cannot
typically be used to discern specific patterns
of intraisland and interisland ceramic trade
and exchange. Compositional analysis of
ceramics offers a complementary line of
inquiry that helps address these issues with
greater precision, allowing the identification
and/or exclusion of possible production
source areas, information that is vitally
important for evaluating the role of ceramic


Journal of Caribbean Archaeology, Special Publication #2, 2008









Preliminary Interpretations from Aguilla and the Salt River Site


Figure 1. Map showing the location of Anguilla and St. Croix within the Caribbean Region.


vessels within a system of socioeconomic
interaction.
In this paper, we report the preliminary
results of INAA of ceramic sherds recovered
from archaeological sites on the small
Caribbean islands of Anguilla and St. Croix.
Compositional analysis using INAA was
undertaken to characterize and enumerate
potential sources for the ceramics recovered
from sites on these two islands. The results
suggest that, during the Late Ceramic Age,
ca. AD 600-1500, there were multiple
manufacturing locales for the ceramics
utilized at sites in Anguilla and St. Croix, and
that, in many cases, production did not likely
occur on the same island where the vessels
ultimately were used and discarded. Though
only preliminary, these data contribute to a
better understanding of the dynamics of
interisland interaction and the potential
significance of ceramic trade and exchange in
the development and maintenance of social


organizations in the northern Lesser Antilles
and Virgin Islands during the Late Ceramic
Age.

Environmental Background
Anguilla is a relatively flat island with an
area of only 102 square km and a maximum
elevation of only 65 m. It is the northernmost
of the outer, lower, islands in the Lesser
Antillean chain (Figure 1) that are largely
comprised of Oligocene limestones overlying
a basal volcanic series and an intermediate
series of feldspathic tuffs, conglomerates, and
cherts (Earle 1923). Anguilla is made up of
bedded white or cream-colored limestones
with fossils that were laid down directly on
top of an igneous basement. The underlying
volcanic material is exposed only in a small
area of the uplifted north coast (Earle 1923),
below Crocus Hill, the island's highest point.
St. Croix is the southernmost of the Virgin
Islands and is situated at the southeastern end


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.









Preliminary Interpretations from Aguilla and the Salt River Site


of the Greater Antilles, about 100 km
southeast of Puerto Rico (see Figure 1). The
closest of the Lesser Antilles are Anguilla, St.
Martin and Saba, all of which are located on
a north-south axis about 150 km to the east.
Throughout the history of Caribbean
archaeology, St. Croix always has been of
great interest because of its somewhat
isolated geographic location between the
Greater and the Lesser Antilles, and its
ethnohistoric position between the Classic
Taino in the west and the Eastern Taino to the
east (Rouse 1992; Morse 2004).
St. Croix is divisible into several
geographical zones, the character of which
undoubtedly had a considerable effect on
prehistoric settlements. The two major
physiographical areas include the late
Cretaceous Oldland, underlain by rock of
volcanic origin, and the Central Kingshill
Plain, based on Tertiary marls and
limestones, which divide the Oldland into a
northwestern and an eastern part. The latter is
partly covered by more recent alluvium
(Cederstrom 1941, Nagle and Hubbard 1989).
St. Croix has an area of about 220 square km
and more than 50 known Amerindian sites
that can be found in almost every sector of
the island. Most sites are situated on or within
1 km of the coast, near sheltered bays or
reefs, and the fewer inland sites are near
waterways (Morse 2004).

Methods
The samples analyzed from Anguilla were
obtained from 0.5 m by 0.5 m test pits
systematically excavated at 25 m intervals
across five different sites occupied during the
Late Ceramic Age (ca. AD 600-1500). These
include the Sandy Ground, Shoal Bay East,
Barnes Bay, Sandy Hill and Forest North
sites (Figure 2). Sandy Ground and Shoal
Bay East have longer occupational histories


with deposits that date between ca. AD
400/600-1500, whereas the other three date
solely to more recent periods between ca. AD
900-1500. The 50 sherds studied were drawn
from a larger grab sample of close to 300
sherds that were selected previously for basic
temper analysis conducted by Crock and
Petersen (Crock 2000). Two sherds were
selected from each test pit excavated at each
of the five sites, one randomly from an
"upper" level and one randomly from a
"lower" level in an effort to produce samples
representative of earlier and later occupations
at each site.
For the St. Croix sample, a total of 50
ceramic sherds from the two most recent
Ceramic Age occupations at the Salt River
site were analyzed. Previously, a basic
temper analysis of these same sherds was
conducted by Morse. The site is located on
the island's northern shore at the base of Salt
River point, which is a small, prominent,
brush-covered peninsula stretching east
towards the Salt River inlet. It was here that
Columbus supposedly had an encounter with
Amerindians on his second voyage in 1493.
St.Croix's only ballcourt (batey) was also
discovered at the site (Hatt 1924, Figure 3).
The ballcourt is the easternmost known in the
Caribbean and its construction is attributed to
the last ceramic horizon (Taino). The Salt
River site seems to have been the largest
coastal settlement on the island at the
European Contact, and the only one with
continuous habitation throughout the Pre-
Columbian Ceramic Age, which in the Virgin
Islands, began shortly before AD 1 and lasted
into the historic period. The site's chronology
is known from two well-documented
excavations, the results of which have been
analyzed and compared with current research
findings in St. Croix and nearby islands
(Morse 1997). One of the excavations was


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.












Preliminary Interpretations from Aguilla and the Salt River Site


Angunilla


SHOAL BAY EAST


LUmentone


HILL


LWttle Harbour
Lockrum Bay

Point


Boy
NORTH

Key:
Archaic. Small Post-Soladold,
SCave and Spot Finds
0 Substantial
Late and Post-Saladoid
Substaotalo
Post-BaIadold
0. Sal Pond
0 1 2 km
scale


Figure 2. Map of Amerindian archaeological sites in Anguilla, B.W.I., with the five sites studied highlighted.


Figure 3. View of the Salt River site, St. Croix. Note petroglyph-lined ball court in foreground.
Photo by Gudmund Hatt, 1923.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.


SANDY


Mead's


BARNES









Preliminary Interpretations from Aguilla and the Salt River Site


led by Gary Vescelius, whose comprehensive
survey of the island in 1951 was carried out
under the auspices of the St. Croix Museum
Commission and the Yale Peabody Museum
where his collections are presently housed.
Vescelius excavated 12 different units (pits)
at the site, and the excavated ceramics proved
essential to developing a ceramic seriation for
St. Croix and the broader region (Vescelius
1951). For the present study, 10 sherds were
selected from each of Vescelius' pits 1, 2, 4,
7 and 10. Five sherds from each of these
excavation units were randomly selected
from levels containing Elenan Ostionoid
ceramics dating to between ca. AD 900-1200
(Figure 4), and five from shallower levels
containing Chican Ostionoid ceramics dating
to between ca. AD 1200-1500 (Figure 5).
These two ceramic horizons represent the
most recent Late Ceramic Age periods for the
site, island, and region and are basically
comparable to the temporal periods
represented by the Anguilla materials.

Results
A previously conducted temper study found
that 80% of Anguillian ceramics contain one
or more volcanic minerals, the majority
including rounded black sand grains (Crock
2000). Although Anguilla does have local
clays that are suitable for the manufacture of
ceramics, volcanic minerals, such as those
visible with a low power microscope in
sherds from Anguillian sites, do not occur
naturally on the limestone island. Exposures
of volcanic material do exist in deeper
basement sections of the uplifted north coast.
However, none has been readily identified as
sources for the temper observed in any
samples. These data have been used to
suggest that more than half of the pottery
found in Anguilla was manufactured on
volcanic islands, including those with black


sand beaches, and/or manufactured using
local clays tempered with sand from these
sources (Crock 2000). The closest islands to
Anguilla with black sand beaches include St.
Kitts, Nevis, Montserrat, and Antigua located
approximately 80-160 km to the southeast.
Finer grained, thin section petrography
conducted by Donahue and others on samples
from two Anguillian sites suggests the
percentage of "exotic" ceramics is closer to
100%, with no sherds exhibiting exclusively
carbonate temper (Donahue et al. 1990).
Principal components analysis conducted
by Descantes and Glascock (2005a) on a
subset of the same samples studied earlier by
more conventional means indicates the five
site samples are representative of three
chemically distinct groups (Figure 6). A total
of 10 of the 50 sherds (20%) fall within the
range of Group 1, 27 (54%) fall within the
range of Group 2, 6 (12%) fall within the
range of Groups 3 and 7 (14%) could not be
assigned to any of the three groups. Based on
the earlier analysis of the temper used in the
production of the same samples, all three
groups identified in the compositional
analysis likely represent off-island sources, or
at least recipes containing off-island or
"exotic" constituents. The identification of
three distinct groups in the principal
components analysis is surprising in that it
excludes the possibility that all ceramics
found in Anguilla derive from a single source
of manufacture as well as the possibility that
each village or household had its own unique
ceramic recipe. Multiple sources for pottery
vessels, like other commodities, is consistent
with evidence that indicates Anguillian
communities were major participants in an
extensive interisland trade and exchange
network during the Late Ceramic Age (Crock
and Petersen 2004; Knippenberg 2006).


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.










Preliminary Interpretations from Aguilla and the Salt River Site


Figure 4. Elenan Ostionoid decorated ceramic sherds recovered by Gary Vescelius
from the Salt River site in St. Croix. Yale Peabody Museum collection.


Figure 5. Chican Ostionoid decorated ceramic sherds recovered by Gary Vescelius from
the Salt River site in St. Croix. Yale Peabody Museum collection.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.










Preliminary Interpretations from Aguilla and the Salt River Site


In terms of possible site-based correlations,
Group 1 is represented by samples from three
of the five sites, Sandy Hill, Barnes Bay, and
Forest North. Group 1 samples are enriched
in the rare earth elements (REEs) and several
transition metals (Co, Fe, Mn, V, etc.)
relative to Groups 2 and 3. The latter groups
are enriched in the alkali metals (K, Rb, Cs,
and Ba) and in elements associated with
quartz or sand temper such as Hf and Zr. The
common trait shared by all of these sites is
that they were occupied solely during post-
Saladoid times, ca., after AD 900. The two
sites without representation in Group 1-
Sandy Ground and Shoal Bay East-have
occupations that date to earlier periods. It is
therefore possible that Group 1 may represent


0
N-


temporal patterns of ceramic production. That
is, Group 1 may be associated with pottery
production post-dating AD 900. However,
given that the analyzed samples from all sites
were derived from both earlier and later
contexts (deeper and shallower), the absence
of Group 1 ceramics from Sandy Ground and
Shoal Bay East might be more related to
social factors or simply a function of the
small sample studied.
Group 2 comprises the majority of samples
and includes pottery from all of the five sites
studied. Relative to Groups 1 and 3, Group 2
is a core group and intermediate on most
elemental concentrations. The lack of any
statistically significant patterning within or
between sites for Group 2 indicates that it


-0.20 -0.15 -0.10 -0.05 -0.00 0.05 0.10 0.15


0.20


Principal Component 1


Figure 6. Biplot of principal components 1 and 3 displaying three compositional groups
identified in the sample of ceramic sherds from five sites in Anguilla. Ellipses represent
90% confidence level for membership in groups. Vectors denote elemental influences
on the ceramic data. Unassigned specimens are not shown (from Descantes and
Glascock 2005a, Figure 1)


Journal of Caribbean Archaeology, Special Publication 2, 2008


Group 2


Group 3


Group 1


Crock et al.









Preliminary Interpretations from Aguilla and the Salt River Site


Figure 7. Broad-line incised rim sherd from the Barnes Bay Site in Anguilla. It was the
only decorated sherd within the samples studied from Anguilla and a member of
compositional Group 3


represents a consistent and enduring source
for Anguillian ceramics, throughout the
history of Amerindian occupation on the
island. Based on simple falloff curve models
of trade and exchange, the high proportion of
ceramics exhibiting a variation of this
particular "recipe" may represent a closer
source than others, perhaps on the nearby
island of St. Martin. St. Martin is the origin
of a large proportion of the imported lithic
materials, including greenstone used in the
production of stone axes and calcirudite used
to manufacture three-pointed zemis.
Group 3, the least represented of the three
groups, is highly enriched in the alkali
metals. It is worth noting that the three sites
with samples assigned to Group 3 also are
sites (Sandy Ground, Shoal Bay East and
Barnes Bay) that exhibit a higher relative
percentage of decorated sherds and other


objects argued to be related to social status
(Crock 2000). Of the two sites that did not
contain pottery assigned to Group 3, one of
them, Forest North, also happen to produce
the lowest recorded estimated proportion of
decorated sherds. Additionally, Group 3
included the only decorated sherd analyzed in
the study of Anguilla ceramics, a slipped,
broad-line incised sherd, possibly from a
"pelican bowl" (Figure 7). Though very
tentative due to sample size, this may suggest
that there is a relationship between the
provenance of Group 3 ceramics and the
production of rarer decorated ceramics during
the Late Ceramic Age, ca. AD 600-1500.
Three distinct groups were also identified in
the ceramic samples from Salt River, St.
Croix, based on principal components
analysis of INAA results by Descantes and
Glascock (2005b) (Figure 8). A total of 24


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.










Preliminary Interpretations from Aguilla and the Salt River Site


samples fall into Group 1 (48%), 20 in Group
2 (40%), and three in Group 3 (6%). Three
specimens (6%) could not be assigned to any
of the three groups identified. Relative to
Group 1, Group 2 clearly has higher
concentrations of the alkali metals and REEs,
and is also enriched in zinc and antimony,
when compared with compositional Group 1.
Group 3 is enriched in thorium. The small
member size in this group makes it difficult
to test the probability of membership for
samples in Group 3 relative to Groups 1 and
2 (Descantes and Glascock 2005b).
The identification of three separate groups
of ceramics in the principal components
analysis points to the use of at least three
separate recipes in the production of the


ceramics recovered from the two temporal
components studied at the Salt River site. It is
possible that all three groups represent
ceramics that were manufactured on St. Croix
from materials local to St. Croix.
Alternatively, one or more of the
compositional groups could represent off-
island production locales.
Given the geologic characteristics of St.
Croix with its two major physiographic areas,
the sherds from compositional Groups 1 and
2, which are represented in both ceramic
periods and date between AD 900-1500, may
well represent locally manufactured ceramics
utilizing clays from these two environments.
Alternatively, these groups may also
represent non-local sources, potentially


-0.20 -0.15 -0.10


-0.05 -0.00 0.05


0.10 0.15


Principal Component 1

Figure 8. Biplot of principal components 1 and 2 displaying the three compositional
groups identified in the sample of ceramic sherds from the Salt River site, St. Croix.
Ellipses represent 90% confidence level for membership in the groups. Vectors denote
elemental influences on the ceramic data. Unassigned specimens are not shown (from
Descantes and Glascock 2005b, Figure 1).


Journal of Caribbean Archaeology, Special Publication 2, 2008


Co


Group 2


Group 1 o
S I Group 3
L - -


Crock et al.









Preliminary Interpretations from Aguilla and the Salt River Site


including production locales on nearby
islands such as Puerto Rico. Ceramics from
sites in St. Croix and Puerto Rico are closely
linked based on stylistic attributes and the
geographic proximity of the two islands
certainly would have favored cultural
relatedness as well as trade and exchange.
The Salt River Bay with its well-protected
harbor, and the island's midway position
between the Greater and the northern Lesser
Antilles, were undoubtedly important factors
in the site's settlement and position within a
broader inter-island exchange network. When
compositional data for St. Croix clay samples
and for ceramics recovered from sites in
Puerto Rico becomes available we will be
able to better assess the relationship between
the two islands in terms of ceramic
production and consumption.
The three samples from the least well-
represented group, Group 3, all derive from
the latest ceramic period dating from AD
1200-1500. One of these samples exhibits a
compositional affinity to samples of ceramics
from the Dominican Republic (Descantes and
Glascock 2005b). The one sherd from the
Salt River site chemically resembles the
Dominican Republic samples as a whole,
which derived predominantly from the
southeast side of Hispaniola in and around
the ceremonial center of La Aleta. A possible
compositional relationship between ceramic
samples from Salt River and sites in
southeastern Dominican Republic is not
surprising based on the similarity between the
two in terms of ceramic style and
archaeological context (Hatt 1932). Though
tentative, the potential relationship between
ceramics found at a ceremonial site like Salt
River and ceremonial sites in Hispaniola such
as La Aleta, is intriguing. Based mainly on
geographic distance, Salt River has long been
more closely identified with Puerto Rico in


terms of potential sociopolitical ties. Possible
direct links with other ballcourt sites, like
those known in Hispaniola, may help better
identify and broaden the understanding of
Salt River's position relative to broader
issues of regional Taino sociopolitics.

Discussion
Compositional analysis presents an
extremely valuable "new" perspective on
regional interaction in the eastern Caribbean
as indicated by the provocative results
obtained in the pioneering analysis for two
island contexts presented here. Clearly, we
need to analyze samples of locally available
clays, both in the context of Anguilla and St.
Croix to identify/rule out potential clay
source areas. As is always the case, analysis
of additional ceramic samples from other
sites will also further advance our
understanding of ceramic production over
time and help identify the potential number
and variability of contemporaneous
manufacturing locales.
Even at this early stage, however, we are
able to discount the extremes when
reconstructing the production and distribution
of ceramics. As stylistic analysis already has
long suggested, the results from Anguilla
indicate that there were fewer ceramic
"recipes" than households and sites and
therefore, Late Ceramic Age vessels were
more likely the products of some degree of
specialized production at locations likely
determined by the differential availability of
suitable clays and tempering agents. The
opposite extreme can also be discounted
because of the compositional analysis. That
is, that there was not a single-source
manufacturing center for the Late Ceramic
Age pottery used by the occupants of villages
in Anguilla or for the pottery used by the
people living at Salt River, St. Croix. Rather,


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.










Preliminary Interpretations from Aguilla and the Salt River Site


ceramic production was more likely
dispersed. Based on the sites studied in
Anguilla and its geology, the majority of the
vessels were likely produced using a very
limited number of recipes at a limited number
of off-island locations. For the vessels used at
Salt River, St. Croix, ceramic production was
also seemingly limited to a handful of recipes
from locations that were more likely to have
been local to the site/island based on the local
availability of constituent materials.

Acknowledgments
The authors thank Scott M. Fitzpatrick and
Chris Clement for their helpful comments on
earlier drafts of this paper.

References Cited
Cederstrom, D. J.
1941 Notes on Physiography of St. Croix, Virgin
Islands, American Journal of Science, 239:553-
576.
Crock, J. G.
2000 Interisland Interaction and the Development
of Chiefdoms in the Eastern Caribbean.
Unpublished Ph.D dissertation, Department of
Anthropology, University of Pittsburgh,
Pittsburgh.
Crock, J. G., and J. B. Petersen
2004 Inter-Island Exchange, Settlement Hierarchy
and a Taino-Related Chiefdom on the Anguilla
Bank, Northern Lesser Antilles. In Late Ceramic
Age Societies in the Eastern Caribbean, pp. 139-
156, edited by A. Delpuech and C. L. Hofman.
BAR International Series, Paris Monographs in
American Archaeology. Archaeopress, Oxford,
14.
Descantes, C. and M. Glascock
2005a Compositional Analysis of Ceramics from
Anguilla, British West Indies. Report prepared for
J. B. Petersen and J. G. Crock, University of
Vermont, Burlington.
2005b Compositional Analysis of Ceramics from the
Salt River Site, St. Croix, US Virgin Islands.
Report prepared for B.F. Morse, Yale Peabody
Museum, Yale University, New Haven.


Donahue, J., D. R. Watters, and S. Millspaugh
1991 Thin Petrography of Northern Lesser Antilles
Ceramics. Geoarchaeology, 32:229-254.
Earle, K. W.
1923 The Geology ofAnguilla, B.W.I. Crown
Agents for the Colonies, London.
Hatt, G.
1924 Archaeology of the Virgin Islands,
Proceedings of the International Congress of
Americanists, 21(1): 29-42.
1932 Notes on the Archaeology of Santa Domingo,
Geografisk Tidskrif 35 (1-2), Copenhagen.
Knippenberg, S.
2006 Stone Artefact Production and Exchange
Among the Northern Lesser Antilles. University of
Leiden, Netherlands.
Morse, B. F.
1997 The Salt River Site, St. Croix, at the time of
the Encounter. In The Indigenous People of the
Caribbean, ed. S. M. Wilson, pp 36-45,
University Press of Florida.
2004 At the Onset of Complexity: Late Ceramic
Developments in St. Croix. In Late Ceramic Age
Societies in the Eastern Caribbean, eds. A.
Delbuech and C. L. Hofman, pp. 183-193, BAR
International Series 1273, Oxford.
Nagle, F. and D. K. Hubbard
1989 St. Croix Geology since Whetten: An
Introduction. In Terrestrial and Marine Geology
of St. Croix, U.S. Virgin Islands, ed. D. K.
Hubbard, pp. 1-7, Special Publication No 8, West
Indies Laboratory, Teague Bay, St. Croix,
U.S.V.I.
Rouse, B. I.
1992 The Tainos: The Rise and Decline of the
People Who Greeted Columbus. Yale University
Press, New Haven and London.
Vescelius, G. S.
1952 The Cultural Chronology of St. Croix, Ms. on
file. Yale University New Haven.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Crock et al.












Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776










































































This Page Intentionally Left Blank.



Journal of Caribbean Archaeology, Special Publication #2, 2008








Journal of CaribbeanArchaeology
Copyright 2008
ISSN 1524-4776


COMPOSITIONAL ANALYSIS OF CERAMICS FROM LA ALETA, DOMINICAN
REPUBLIC: IMPLICATIONS FOR SITE FUNCTION AND ORGANIZATION

Geoffrey W Conrad
William Hammond Mathers Museum
Indiana University
601 East 8th Street
Bloomington, IN 47408-3812
conrad( aindiana. edu


Charles D. Beeker
Office of Underwater Science
Indiana University
HPER 058
1025 East 7th Street
Bloomington, IN 4 7405-7109
cbeeker @indiana. edu

John W Foster
California State Parks
PO Box 942896
Sacremento, CA 94296-0001
jfni a irks.ca.gov


Christophe Descantes
Archaeological Research Facility
University of California Berkeley
2251 College Building
Berkeley, CA 94720-1076
cdescantes@berkeley. edu

Michael D. Glascock
Archaeometry Laboratory
University of Missouri Research Reactor
1513 Research Park Drive
Columbia, MO 65211
glascockm@missouri. edu


La Aleta is a major Taino site that we believe served as a regional ceremonial center. The site's
features include a flooded sinkhole and four stone-lined plazas similar to contact-period plazas
used for ball games and dances. This paper uses data derived from instrumental neutron acti-
vation analysis (INAA) of ceramics from the sinkhole and plazas to evaluate 1)po,,ithew about
La Aleta's importance and internal organization. The INAA data are not ini ,nii'nii n /ith the
interpretation that La Aleta served a regional, rather than local, population, but the results to
date are probably best described as equivocal. The data do not support the liypo,theii that each
of the four plazas was used by people from a different locality. Instead, the data are more con-
Si1,te./t 11 ith alternative 1i pithe we about the use of the plazas and or the manufacture of the
ceramics found therein.


This paper is an experiment in using data
derived from instrumental neutron activation
analysis (INAA) of ceramics to evaluate
hypotheses about inter-site functional
variability and intra-site spatial organization
in Taino settlements. A research team from
Indiana University, in cooperation with


Dominican colleagues, has been working in
and around the Parque Nacional del Este in
the southeastern Dominican Republic (Figure
1) since 1996. This zone formed part of the
core of the Taino chiefdom of Higtey, one of
the principal cacicazgos of Hispaniola at the


Journal of Caribbean Archaeology, Special Publication #2, 2008









Compositional Analysis of Ceramics from La Aleta

time of European contact (Las Casas
1967:1:22-26).
Among the sites we have been investigating
is La Aleta (Figure 1), which contains a
flooded sinkhole known as the Manantial de
la Aleta (Atiles and Ortega 2001; Guerrero
1981; Ortega and Atiles 2003). Elsewhere we
have argued that the Manantial de la Aleta
was an important center for ritual offerings
connected with the life-giving properties of
water, female fertility, the female qualities of
the universe, female supernaturals, and
ancestor worship (Beeker et al. 2002; Conrad
et al. 2001; Conrad, Beeker, and Foster
2005). Seven AMS radiocarbon dates on
organic artifacts from the Manantial range
from cal A.D. 1035 to 1420 (Conrad et al.
2001:14-15, Table 1).
In addition to the Manantial, La Aleta also
contains four bateyes (ceremonial plazas) that
are typical features of major Taino
ceremonial centers in the Greater Antilles,
especially in Puerto Rico and Hispaniola


Conrad et al.


(Atiles and Ortega 2001; Guerrero 1981;
Ortega and Atiles 2003). During the contact
period such plazas were used for ritual
activities, including dances and ballgames.
The presence of the Manantial and the four
plazas suggests La Aleta was a regionally
important ceremonial center that served a
relatively wide area. The four plazas also
suggest that the site was used by multiple
Taino sociopolitical groups, each of which
may have been associated with particular
zones within the settlement-for example,
with a particular plaza.
Our purpose here is to ascertain whether
compositional analyses of ceramics can
contribute to the evaluation of these specific
hypotheses about La Aleta, and by extension
to the investigation of broader questions
about Taino history and culture, and
Caribbean prehistory in general. Our data
derive from INAA studies of 175 sherds: 125
from La Aleta, consisting of 25 samples each
from the Manantial and the four plazas; 25


Figure 1. Schematic diagram of instrumental neutron activation analysis.

Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Analysis of Ceramics from La Aleta

sherds from a nearby shoreline village site
called La Cangrejera; and a total of 25 sherds
from three sites in the Punta Macao region,
farther to the north (Figure 1).
The samples from the Manantial de la Aleta
were selected from collections recovered by
the Indiana University team during dives in
1996 and 1997 (Beeker et al. 2002; Conrad,
Beeker, and Foster 2005). The sherds from
the plazas at La Aleta were excavated by
Elpidio Ortega and Gabriel Atiles in 1997
(Atiles and Ortega 2001; Ortega and Atiles
2003), whereas those from La Cangrejera
were excavated by Atiles and Harold Olsen
Bogaert in 1998 (Andujar and Atiles 2003).
The sherds from the Punta Macao sites,
which were investigated by Atiles in 2003
and 2004 (Andujar 2004), were initially
collected for a different purpose, but proved
to have some relevance to the problems under
consideration. In May of 2004, the Museo del
Hombre Dominicano, at that time directed by
Carlos Andujar, graciously made all of these
materials available to us for analysis.

INAA Results
The INAA was carried out at the
Archaeometry Laboratory of the University
of Missouri Research Reactor Center
(MURR), using the standard MURR
procedures described by Christophe
Descantes et al. in the introductory paper of
this volume (see also Baxter 1992; Bieber et
al. 1976; Bishop and Neff 1989; Descantes
and Glascock 2004; Glascock 1992; Neff
1994, 2001; Sayre 1975). We identified five
compositionally homogeneous groups within
the ceramics (Figure 2), designated La Aleta
1 (n=6), La Aleta 2 (n=l 11), Punta Macao
(n=24), Group 1 (n=101), and Group 2 (n=4);
29 samples were unassigned (Table 1). Each
of these groups is distinguished from the
others by being relatively enriched (or


Conrad et al.


diluted) in a different set of elements.
Compared to the other groups, La Aleta 1
ceramics are enriched in chromium and
cobalt, whereas La Aleta 2 ceramics are
enriched in antimony, arsenic, terbium,
samarium, lutetium, and uranium actinidee).
The Punta Macao group is enriched in
sodium, potassium, hafnium, rubidium, and
barium. Group 1 ceramics are enriched in
manganese, while those in Group 2 are
enriched in hafnium, zirconium, uranium, and
actinide. Following the "provenance
postulate" of Weigand et al. (1977), these
compositional groups are assumed to
represent geographically restricted sources or
source zones.
We should note that there may be only a
three-group structure, because there is a
possibility that compositional groups La
Aleta 2 and Group 2 are transformations of
Punta Macao and Group 1. All of the sherds
in those first two groups were recovered from
the flooded sinkhole known as the Manantial
de la Aleta, and their chemical composition
may have been altered by their prolonged
immersion in deep water. In other words,
there is the possibility that the underwater
postdepositional environment of the La Aleta
sherds may have affected the sherds by
enriching them in uranium.
The five (or three) groups have different
distributions. The groups designated La Aleta
1, La Aleta 2, and Group 2 occur only at the
site of La Aleta. As indicated above, all of the
Group 2 and La Aleta 2 specimens were
recovered from deep-water contexts in the
Manantial. In contrast, all of the La Aleta 1
sherds came from the four plazas at the site.
Group 1 ceramics were found at three sites:
La Aleta, La Cangrej era, and Punta Macao
proper. The Punta Macao group had the
widest distribution, being represented at all


Journal of Caribbean Archaeology, Special Publication ', 2008











Compositional Analysis of Ceramics from La Aleta


Table 1. Sherds by compositional group

Site La Aleta 1 La Aleta 2 Group 1 Group 2 Punta Macao

La Aleta

Manantial 0 11 2 4 0

Plaza A 2 0 21 0 0

Plaza B 2 0 22 0 1

Plaza C 1 0 20 0 1

Plaza D 1 0 18 0 0

La Cangrejera 0 0 16 0 4

Punta Macao Sites 0 0 2 0 18

Total 6 11 101 4 24


0
CO
0




o


0 0 o
0 d

a- Punta Macai


0 +OOR120



0
1 -0.10 -0.06


-0.02 0.02 0.06 0.10 0.14 0.18

Principal Component 1


Journal of Caribbean Archaeology, Special Publication 2, 2008


I Manantial de la Aleta 1
+DORI + ? 2


S "Wo(12 00 0+qoeM94 Manantial de
04io6 I et a s l+oR 0 la Aleta 2
S o o a v _) .0
i b, ( 2 tV V +IXHOISs


Figure 2. Plot of the five compositional groups (ellipses represent 90% confidence level).


Conrad et al.










Compositional Analysis of Ceramics from La Aleta

five sites: La Aleta, La Cangrejera, and the
three sites in the Punta Macao area (see Table
1).

La Aleta as a Regional Center:
INAA Evidence
In terms of our first hypothesis, if La Aleta
was indeed a regionally important center,
pottery found at the site would have been
drawn from-and manufactured across-a
wider area than the pottery from La
Cangrej era, a village site with only local
importance.2 If so, we would expect
compositional analyses of ceramics from the
two sites to produce different results. For
example, if the pottery from La Cangrejera
was indeed manufactured within a restricted
geographical area, we would expect the
chemical range of variation to be minimal in


Conrad et al.


terms of clay sources and paste "recipes," and
thus have few recognizable compositional
groups. In contrast, if La Aleta had a broader
regional importance, we would expect its
ceramics to be more heterogeneous in terms
of sources, recipes, and composition, and
therefore have more identifiable groups than
comparable sites in the area.
Compositional analysis does indeed reveal
that samples from La Aleta have a wider
range of variation than samples from La
Cangrejera (Figure 3). The first 50 samples
submitted for analysis were the 25 sherds
from the Manantial de la Aleta and the 25
from La Cangrej era. Clear chemical
compositional differences exist between the
ceramics from these two sites, most visible in
bivariate projections that include uranium and
arsenic concentrations, however, it should be


i.,, ,.,1i.i de la Aleta
-.. 'PNE-01-A
E o o
L o 0
C:- % o o0
0 0
U) o
0 0




0 0 08 12 16 20 2








Arsenic (log base-10 ppm)
from La AletaLa Cangrejera
0

I
1 0.0 0.4 0.8 1.2 1.6 2.0 2A4

Arsenic (log base-10 ppm)

Figure 3. Plot of arsenic and uranium concentrations showing the contrast between ceramics
from La Aleta and La Cangrejera.


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Analysis of Ceramics from La Aleta

noted again that the underwater
postdepositional environment may have
affected the sherds at the Manantial de la
Aleta by enriching them in uranium. Overall,
the La Cangrej era ceramics are more
homogeneous than the specimens from the
Manantial de la Aleta. When the ceramics
from the four plazas are included in the
comparison, the heterogeneity of the
ceramics from La Aleta increases
considerably, magnifying the contrast
between the two sites. All five (or three) of
the compositional groups are represented at
La Aleta, whereas only two groups are
represented at La Cangrej era. Admittedly,
when the samples from the plazas are
included, the sample size for La Aleta is five
times as large as that from La Cangrejera.
In brief, compositional analyses offer some,
albeit limited, support for the hypothesis that
La Aleta was a ceremonial center that served
a regional Taino population. The
compositional evidence is consistent with
some stylistic evidence we have published
elsewhere, although that evidence is also
limited (Beeker et al. 2002). At this point
much work remains to be done before we can
delimit the region served by La Aleta, which
may well have fluctuated through time. To
judge from the results of the preliminary
work presented here, however, it seems very
likely that additional compositional analyses
could be very helpful in resolving this
question.

Use of the Plazas:
INAA Evidence
Our second hypothesis is concerned with
the internal organization of La Aleta, and
specifically with the four plazas. If the
artifacts recovered from La Aleta were drawn
from-and manufactured across-a wide
area, ceramics fom the site as a whole should


Conrad et al.


be relatively heterogeneous in terms of
sources and recipes. A closer examination of
intra-site patterning, however, might provide
additional information about La Aleta. That
is, if each plaza was used by Taino groups
from a different locality or set of localities,
with every group providing its own ceramics,
we would expect compositional analyses of
sherds from each of the plazas to produce
different results.3 On the other hand, if all
plazas were used by all groups but for
different ceremonies, the results from all of
the plazas should look similar.
INAA data indicate that the compositions
of the plaza ceramics cannot be separated in
elemental or principal component space
(Figure 4). No statistical difference exists
among the chemical compositions of the
ceramics found on the four plazas; i.e., none
of the plazas appear to have chemically
distinctive or unique ceramics. At first glance
these data suggest that the use of any given
plaza was not restricted to a particular group
or groups. Instead, the data are more
consistent with the hypothesis that each plaza
was used by all groups for a different set of
activities distinguished by content, timing, or
other factors. In other words, it seems likely
that what differentiated the four plazas from
one another was the ceremonies that took
place in them, not the people who
participated.
In reality, however, the similarity in the
chemical compositions of the ceramics does
not preclude either possibility: that each plaza
was used by all groups, as above, or that the
four plazas were used by different groups.
The hypothesis as stated assumes that each
group was bringing its own pottery, made
with its own local clays, for use in
ceremonies at La Aleta, and that there was no
centralized ceramic production for the site. In


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Analysis of Ceramics from La Aleta


-0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04


0.06 0.08 0.10


Principal Component 1


Figure 4. Plot showing no compositional difference among the ceramics from the four plazas.


fact, the data for the ceramic groups from the
plazas may raise questions about this
assumption. The vast majority of the sherds
from the plazas-81 percent-are assigned to
Group 1, with only rare occurrences of
ceramics from other compositional groups
(Figure 5). The proportion of Group 1 sherds
in the individual plazas ranges from 72 to 88
percent. If we consider only samples that can
be assigned to one of the five compositional
groups, the proportion of Group 1 increases
to 91 percent, with a range of 88 to 95
percent for the individual plazas. The
overwhelming preponderance of Group 1 in
these samples might reflect centralized, and
possibly specialized, production of ceramics
for use in the ceremonies held in the plazas,


regardless of the group or groups
participating. If so, we do not know at present
where this production might have occurred.
A third possibility is that the ceramics have
nothing to do with the use of the plazas but
are simply rubbish deposited as construction
fill. In that case, the high percentage of
Group 1 ceramics might indicate that the fill
in the four plazas originated from a single
source, a large midden in which Group 1
pottery predominated. We cannot completely
discount this possibility, but we consider it
unlikely, as no one has yet identified any
substantial midden deposits at La Aleta
outside the plazas themselves. For the
moment, we consider it more probable that
the results of the INAA analyses reflect use


Journal of Caribbean Archaeology, Special Publication ', 2008


Conrad et al.










Compositional Analysis of Ceramics from La Aleta


CM
o
,o ,r:,. : de la Aleta 1
0o



0+


E + +



DL co
C o Punta Macao

o Group 2 0 PlazaA
0 E o Plaza B
A Plaza C
o + Plaza D
0
I -010 -0.06 -0.02 0.02 0.06 0.10 0.14

Principal Component 1

Figure 5. Plot showing the sherds from the four plazas affiliated primarily with Group 1.


of each plaza in different ceremonies,
localized and perhaps specialized production
of ceramics for use in the ritual activities
carried out in the plazas, or both.

Vessel Forms and Compositional Groups
In response to an earlier version of this
paper (Conrad, Descantes, VanderVeen, and
Glascock 2005), Emily Lundberg (personal
communication 2005) asked whether we had
attempted to control for vessel form, and
whether it might be an important variable.
The answer to the first question is "no,"
except unintentionally in the case of the
sherds from the Manantial de la Aleta. For
pragmatic reasons we chose nondiagnostic
body sherds from most contexts.4
Twenty-two of the 25 specimens from the
Manantial can be categorized by vessel form,


however. Sixty-four percent (51 of 80) of the
recognizable vessels or fragments recovered
from the Manantial were bowls (Beeker et al.
2002), and 68 percent of the identifiable
sherds from the Manantial that were sent to
MURR (15 of 22) were bowl fragments. Of
the identifiable sherds that could be assigned
to specific compositional groups, the
percentage of bowls in each group ranged
from 60 to 100 (Table 2). This limited
evidence suggests that the percentage of
bowls in each compositional group mirrors
the percentages in the sample submitted to
MURR and the total assemblage from the
Manantial; we suspect the same is true of
other vessel forms as well. Accordingly, there
does not appear to be a strong correlation
between vessel forms and compositional


Journal of Caribbean Archaeology, Special Publication ', 2008


Conrad et al.









Compositional Analysis of Ceramics from La Aleta

groups, but at this point the sample is really
too small to permit definite conclusions.

Conclusion
Clearly, our sample sizes are small, and
Ronald Bishop (personal communication
2006) has rightly characterized our results as
"equivocal." Accordingly, our interpretations
can only be viewed as preliminary.
Nonetheless, the results do indicate that
compositional analyses have the potential to
make major contributions to Caribbean
archaeology-even if all they do at this point
is call our attention to new questions. What
do the INAA results tell us about La Aleta?
Did the Manantial serve a regional
population, whereas the plazas were used
only by local groups? Was pottery "mass-
produced" for use in activities connected with
the plazas, but not the Manantial? If so, does
that mean the Manantial was reserved for use
by regional elites, while the plazas were
accessible to local commoners? Was the
Manantial in some way a more sacred space
than the plazas?
It is true that our results to date can provide
at best equivocal answers to these questions.
Nonetheless, without INAA analysis we


Conrad et al.


could not address these problems, because the
ceramics from the different contexts look the
same stylistically. In fact, without the INAA
analysis, would we even know that these
questions existed? The analysis of additional
samples from the sites under consideration
here, as well as samples from clay sources in
the region, would not simply support or cast
doubt on our current interpretations. As
above, they would lead us to new research
questions about the Taino occupation of the
Parque Nacional del Este region. In terms of
even broader, longer-term implications, the
work presented here indicates that
compositional analysis has the potential to
reveal significant spatial groupings, and
ultimately temporal groupings as well, in
Caribbean artifacts. If so, it can make
powerful contributions to a host of questions
in Caribbean archaeology that are now
debated inconclusively on the basis of
stylistic and documentary evidence alone.

Acknowledgments
We are grateful to the Secretaria de Estado
de Cultura of the Dominican Republic and
two of its agencies, the Museo del Hombre
Dominicano and the Oficina Nacional del


Table 2. Manantial de la Aleta: vessel form by compositional group.

Group 1 Group 2 LA2 Unass. Total

Bowl 2 3 6 4 15
Olla 0 0 1 0 1

Jar 0 1 0 2 3
Bottle 0 0 1 0 1
Platter 0 0 1 0 1
Buren 0 0 1 0 1
?? 0 0 1 2 3
Total 2 4 11 8 25


Journal of Caribbean Archaeology, Special Publication 2, 2008










Compositional Analysis of Ceramics from La Aleta

Patrimonio Cultural Subacuatico, for all they
have done to make our research possible. The
samples from the Manantial de la Aleta and
La Cangrej era were processed through the
operating support for the MURR
Archaeometry Laboratory by the National
Science Foundation (Grant BCS-0504015).
Analysis of the samples from the plazas at La
Aleta was made possible by a grant-in-aid
from the Office of the Vice President for
Research (now the Office of the Vice Provost
for Research), Indiana University. Tessa
Schut and Nicole Little of MURR carried out
the preparation and irradiation of the samples.
Figure 1 was prepared by Ellen Sieber of the
William Hammond Mathers Museum,
Indiana University. We are grateful to all of
those institutions and individuals for their
help. We also appreciate Ronald Bishop's
insightful comments on the version of this
paper we presented at the original SAA
symposium. Any errors of interpretation are
our responsibility.
Finally, and with a profound sense of loss,
we want to express our thanks to James B.
Petersen, who saw value in an early and quite
rough version of this paper and went out of
his way to tell us so. That was typical of Jim.
He was a fine scholar and a warm and
engaging human being. We miss him on both
counts.

References Cited
Anduijar Persinal, C.
2004 Presentaci6n. Boletin del Museo del Hombre
Dominicano 37:1-3.
Anduijar Persinal, C., and J. G. Atiles
2003 Investigaciones Arqueol6gicas y
Reconocimiento a los Sitios Cueva de Pefi6n
Gordo, Cueva de Jos6 Maria, y Sitio La
Cangrejera. Provinicia La Altagracia y Parque
Nacional del Este. Boletin del Museo del Hombre
Dominicano 33:161-175.


Conrad et al.


Atiles, J. G., and E. Ortega
2001 Un Sitio Llamado el Manantial de la Aleta.
Boletin del Museo del Hombre Dominicano
30:33-54.
Baxter, M. J.
1992 Archaeological Uses of the Biplot-A
Neglected Technique? In Computer Applications
and Quantitative Methods in Archaeology, 1991,
edited by G. Lock and J. Moffett, pp. 141-148.
BAR International Series S577. British
Archaeological Reports, Oxford.
Beeker, C. D., G. W. Conrad, and J. W. Foster
2002 Taino Use of Flooded Caverns in the East
National Park Region, Dominican Republic.
Journal of Caribbean Archaeology 3:1-26. http://
www.flmnh.ufl.edu/jca.Beekeretal.pdf, accessed
April 2007.
Bieber, A. M. Jr., D. W. Brooks, G. Harbottle, and E.
V. Sayre
1976 Application of Multivariate Techniques to
Analytical Data on Aegean Ceramics.
Archaeometry 18:59-74.
Bishop, R. L., and H. Neff
1989 Compositional Data Analysis in Archaeology.
In Archaeological Chemistry IV, edited by R. 0.
Allen, pp. 576-586. Advances in Chemistry Series
200. American Chemical Society, Washington,
D.C.
Conrad, G. W., C. D. Beeker, and J. W. Foster
2005 Underwater Archaeology at the Manantial de
la Aleta, Dominican Republic. InActas delXX
Congress Internacional de Arqueologia del
Caribe, vol. 2, edited by C. Tavhrez Maria and M.
A. Garcia Ardvalo, pp. 703-710. Museo del
Hombre Dominicano and Fundaci6n Garcia
Ar6valo, Santo Domingo.
2006 Compositional Analysis of Ceramics from La
Aleta, Dominican Republic. Paper presented at the
71st Annual Meeting of the Society for American
Archaeology, San Juan.
Conrad, G. W., C. Descantes, J. VanderVeen, and M.
Glascock
2005 Compositional Analysis of Ceramics from La
Aleta, Dominican Republic. Paper presented at the
21st Congress of the International Association for
Caribbean Archaeology, St. Augustine, Trinidad
& Tobago.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Compositional Analysis of Ceramics from La Aleta

Conrad, G. W., J. W. Foster, and C. D. Beeker
2001 Organic Artifacts from the Manantial de la
Aleta, Dominican Republic: Preliminary
Observations and Interpretations. Journal of
Caribbean Archaeology 2:1-20. http://
www.flmnh.ufl.edu/jca/Conrad et al.pdf,
accessed April 2007.
Descantes, C., and M. D. Glascock
2004 Compositional Analysis of Taino Ceramics
from the Dominican Republic. Report prepared
for Geoffrey W. Conrad, Department of
Anthropology, Indiana University, Bloomington,
Indiana. Copies available from Research Reactor
Center, University of Missouri, Columbia, or from
Conrad.
Descantes, C., M. D. Glascock, G. W. Conrad, C. D.
Beeker, and J. W. Foster
2006 Compositional Analysis of Ceramics from
Archaeological Sites in the Southeastern
Dominican Republic. Ms. on file, William
Hammond Mathers Museum, Indiana University,
Bloomington.
Glascock, M. D.
1992 Characterization of Archaeological Ceramics
at MURR by Neutron Activation Analysis and
Multivariate Statistics. In Chemical
Characterization of Ceramic Pastes in
Archaeology, edited by H. Neff, pp. 11-26.
Prehistory Press, Madison, Wisconsin.
Guerrero, J. G.
1981 Dos Plazas Indigenas y el Poblado de
Cotubanamd, Parque Nacional del Este. Boletin
del Museo del Hombre Dominicano 16:13-30.
Las Casas, B. de
1967 At ,/. ,...... Historia Sumaria. 2 vols.
Institute de Investigaciones Hist6ricas,
Universidad Nacional Aut6noma de M6xico,
Mexico, D.F.
Neff, H.
1994 RQ-Mode Principal Components Analysis of
Ceramic Compositional Data. Archaeometry
36:115-130.
2001 Quantitative Techniques for Analyzing
Ceramic Compositional Data. In Ceramic
Production and Circulation in the Greater
Southwest: Source Determination by INAA and
Complementary Mineralogical l.. I ,,,, 1 i
edited by D. Glowacki and H. Neff, pp. 15-36.


Conrad et al.


Cotsen Institute of Archaeology Monograph 44.
University of California at Los Angeles, Los
Angeles.
Ortega, E., and G. Atiles
2003 Manantial de la Aleta y la Arqueologia en el
Parque Nacional del Este. Fundaci6n Ortega
Alvarez Vol. IX. Academia de Ciencias de la
Repuiblica Dominicana, Santo Domingo.
Sayre, E. V.
1975 Brookhaven Procedures for Statistical
Analyses ofMultivariate Archaeometric Data.
Brookhaven National Laboratory Report BNL-
23128. Brookhaven, New York.
Weigand, P. C., G. Harbottle, and E. V. Sayre
1977 Turquoise Sources and Source Analysis:
Mesoamerica and the Southwestern U.S.A. In
Exchange Systems in Prehistory, edited by T. K.
Earle and J. E. Ericson, pp. 15-34. Academic
Press, New York.

Notes
1 For detailed descriptions of the INAA results see
Descantes and Glascock (21*" 4) and Descantes et al.
(2006).
2 The hypothesis as stated here assumes that the
pottery found at La Aleta was made elsewhere, or at
least with clays obtained elsewhere. Earlier reports on
our INAA work presented at conferences (Conrad,
Descantes, VanderVeen, and Glascock 2005; Conrad
et al. 2006) stated incorrectly that there are no clay
sources within 15 km of La Aleta. This claim was
based on the initial stage of a survey by Robert Green
of the Department of Anthropology, Indiana
University. More recently Green (personal
communication 2006) has found clay at La
Granchorra, about 5 km east of La Aleta.
3 This hypothesis assumes that that all of the plazas
are contemporaneous. The available evidence does not
indicate any major chronological differences among
the plazas (Atiles and Ortega 2001; Ortega and Atiles
2003). For our purposes here, however, the INAA
results render the point moot.
4 In his discussion of the papers presented at the
symposium that was the basis for this volume, Ronald
Bishop pointed out that for INAA analyses, diagnostic
ceramics are preferable to nondiagnostic ones. We
agree, but for the specific analyses discussed in this
paper, we deliberately asked for nondiagnostic sherds
because we thought it might facilitate approval of our


Journal of Caribbean Archaeology, Special Publication 2, 2008











Compositional Analysis of Ceramics from La Aleta

request. In 2004 Gabriel Atiles (personal
communication) told us that he did not know of any
earlier requests to remove artifacts from the Museo del
Hombre Dominicano for destructive analysis and that,
in his opinion, previous administrations of the museum
would not have granted one. As it turned out, the
administrations of the museum with which we have
worked have been extremely helpful in making
specimens available for INAA, and we have selected
diagnostic sherds for subsequent analyses.
For the initial analysis of specimens from the
Manantial de la Aleta, small, undecorated fragments
were removed from larger, diagnostic pieces, which is
why vessel form is known in most cases.


Journal of Caribbean Archaeology, Special Publication 2, 2008


Conrad et al.








Journal of Caribbean Archaeology
Copynght 2008
ISSN 1524-4776

PRELIMINARY PETROGRAPHIC AND CHEMICAL ANALYSES OF PREHISTORIC
CERAMICS FROM CARRIACOU, WEST INDIES

Scott M. Fitzpatrick
Department of Sociology & Anthropology
North Carolina State University
scottfitzpatrick@ncsu. edu

Jennifer A. Carstensen
Department of Geological Sciences
California State University Northridge
eitsyrc@aol. corn

Kathleen M. Marsaglia
Department of Geological Sciences
California State University Northridge
kathie. marsaglia@csun. edu

Christophe Descantes Quetta Kaye
Archaeological Research Facility Institute ofArchaeology
University of California Berkeley University College London
cdescantes@berkeley. edu quettakaye @hotmail. co. uk

Michael D. Glascock Michiel Kappers
Archaeometry Laboratory In-Terris Site Technics
University of Missouri Research Reactor Netherlands
glascockm@missouri. edu michiel@in-terris. corn

Petrographic and chemical analysis of prehistoric ceramic sherds from the island of Carriacou
in the vinhew uii Grenadines, West Indies offers preliminary insight into resource exploitation,
manufacturing techniques, and the distribution of pottery from ca. AD 400-1200. Thin-section
petrography of two different suites of sherds (Suite 1: n 24from six sites; Suite 2: n 54from
five of the same sites plus three othei %) indicates that there are at least five temper groups, all
or most of which appear to be exotic to Carriacou. Instrumental neutron activation analysis
(INAA) of 56 sherds from Suite 2 reveals the existence of two main geochemical compositional
groups. This may be a reflection of prehistoric potters selecting clays derived from two geo-
chemically different substrates in the locations) of manufacture. A comparison of these find-
ings %t'eit that: 1) exotic materials are predominant in manufacture; 2) both local and re-
gional transport of ceramics occurred prehistorically; 3) the correspondence between the tem-
per and INAA compositional groups is unclear, it ,eting' that the paste geochemistry might
mask minor temper differences; and 4) clay and temper preferences may have changed through
time, although this will require further testing.


Journal of Caribbean Archaeology, Special Publication #2, 2008









Analyses ofPrehistoric Ceramics from Carriacou

The analysis of ceramics from
archaeological sites, both on the surface and
excavated from stratified deposits, can help
us explore issues related to the production,
distribution, and movement of artifacts across
time and space. In the Caribbean, most
ceramic studies have focused on examining
stylistic motifs and morphological attributes
of pottery (e.g., Righter 1997; Roe 1989; see
also Keegan 2000), whereas technological
aspects of production such as firing
temperature, porosity, and density have only
been cursorily studied (for one exception see
Curet 1997). Unfortunately, there has been a
paucity of research dedicated to investigating
the mineralogical and chemical composition
of pottery in general (but see Carini 1991;
Donahue et al. 1990; Fuess et al. 1991; Fuess
and Donahue 1992;Lambert et al. 1990;
Mann 1986; Winter and Gilstrap 1991; van
As and Jacobs 1992), despite the usefulness
of such approaches for answering a number
of important provenance and manufacturing/
technology-related questions and their
widespread use in other regions worldwide.
To remedy this situation and provide the
first compositional analysis of ceramics from
the island of Carriacou in the southern
Grenadines, we conducted thin-section
petrography and INAA on nearly 80 sherds
(Suite 1: n=24 sherds, detailed petrography
only; Suite 2 = 56 sherds, all INAA and 54
cursory petrography). This study is one
component of the Carriacou Archaeological
Survey Project and part of a larger regional
effort by researchers at the University of
Missouri Research Reactor (MURR) to
develop a database of Caribbean ceramics
that will be beneficial to the greater
Caribbean archaeological community. In this
paper, we first provide a brief archaeological
and environmental background to
contextualize research thus far conducted on


Fitzpatrick et al.


Carriacou. We then discuss the methods we
used for sampling and analyzing prehistoric
ceramics. Our results suggest that there are at
least five major geological sources of temper
components, with two major chemical groups
(paste?) identified using INAA. Both types of
analyses indicate that prehistoric Carriacouan
pottery was produced using primarily exotic
materials however, there does not appear to
be any direct correlation between the temper
and chemical groups. Nonetheless, the data
still have implications for understanding local
manufacturing techniques, the movement of
pottery and other raw materials within the
Lesser Antilles, how cultural interactions and
resource exploitation may have changed over
time, and the utility of combining these two
methods for examining prehistoric pottery
production.

Archaeological Research on Carriacou
Jesse Fewkes (1907:189-190) was one of
the first scholars to investigate Carriacou and
adjacent islands and described the ceramics
found there as "among the finest West Indian
ware that has yet come to the Smithsonian
Institution." Bullen (1964) investigated
Grenada in the 1960s and made several short
trips to St. Vincent and the Grenadines to
collect artifacts and excavate exploratory
trenches, including the Sabazan site on
Carriacou (Bullen and Bullen 1972). Sutty
(1990) surveyed portions of Carriacou and
recorded surface finds at Grand Bay,
Sabazan, and a number of other prehistoric
sites, but did not conduct any excavation.
In July 1999, Kappers visited the Grand
Bay site on Carriacou and noted a substantial
amount of cultural material visible on the
surface. The project's field directors (Kaye,
Kappers, and Fitzpatrick) later surveyed
nearly the entire coastline of Carriacou in
March-April 2003, as well as interior areas


Journal of Caribbean Archaeology, Special Publication ', 2008









Analyses ofPrehistoric Ceramics from Carriacou

that were relatively flat or easily accessible.
The team recorded 11 locations with
evidence for prehistoric occupation, six of
which had significant finds (primarily
ceramics and faunal refuse) that were
indicative of long-term settlement activities
(Kaye et al. 2004). Of these six sites, Sabazan
and Grand Bay had the most extensive
stratified coastal profiles and an abundance of
faunal remains, artifacts, and archaeological
features, although other sites were also
thought to have good potential for further
study. As part of a long-term plan to
investigate Carriacou's prehistoric
occupation, the research team has focused on
conducting limited testing at Sabazan and
large area excavations at Grand Bay since
2004 (Fitzpatrick et al. 2004; Kappers et al.
2005; Kaye et al. 2004, 2005).
A total of 20 radiocarbon dates (charcoal,
marine shell, and human bone) from Grand
Bay, Sabazan, and Harvey Vale suggest that
the island was first settled by ceramic making
peoples during the terminal Saladoid period
around A.D. 400, with later periods of
cultural development characteristic of the
Troumassan Troumassoid (ca. A.D. 600-
1000) and Suazan Troumassoid (ca. A.D.
1000-1400) subseries of ceramics
(Fitzpatrick et al. 2004, in press; Harris
2005). Material recovered from excavations
at Grand Bay includes a vast array of mostly
undecorated pottery sherds, ceramic adornos
(modeled applique of animals or zoomorphs
attached to the rims of vessels), two rare
stone cemis, bone needles and tools, carved
turtle plastron, shell beads and adzes, charred
seeds, an enormous amount of fishbone
(LeFebvre 2005), turtle bone, and mollusk
shells, at least ten human burials, and
numerous posthole, pit, and hearth features.


Fitzpatrick et al.


Environmental Background
Carriacou is located in the southern Lesser
Antilles approximately 250 km north of
Venezuela and 30 km north of Grenada
(Figure 1). Politically, Carriacou is part of the
tri-island nation of Grenada along with Petite
Martinique, but also includes the smaller
islands of Petite Dominica, Petite St. Vincent,
Saline, and Frigate. Carriacou is the largest
island in the Grenadines chain measuring
10.4 km from north to south, 8.7 km across at
its widest point, and roughly 32 km2 in
area-it has a maximum elevation of 290 m.
Geologically, Carriacou lies on the southern
Lesser Antilles platform between the two
most active volcanoes of the Lesser Antilles
magmatic arc, the subaerial St. Soufriere
volcano on St. Vincent (Heath et al. 1998b)
and the submarine volcano Kick'em Jenny
near the volcanic island of Grenada (Heath et
al. 1998a). Outcrops of Neogene (2.7 to 11.2
Ma) magmatic rocks dominate the western
half of Carriacou, whereas outcrops of older
Miocene to Eocene sedimentary units
dominate the eastern half (as summarized by
Speed et al. [1993]). These outcrops include
basaltic to andesitic intrusive, extrusive, and
epiclastic volcanic rocks and fossiliferous
limestone (Caldwell 1983; Donovan et al.
2003; Jackson 1980; Pickerall et al. 2001,
2002; Robinson and Jung 1972; Speed et al.
1993).

Methods
One of the research goals of archaeological
investigation on Carriacou is to determine
how pottery, the most ubiquitous artifact
class found on the island (and the Caribbean
in general), was manufactured by native
groups, used for domestic or other activities,
and perhaps transported either across or
between islands through time. A fundamental


Journal of Caribbean Archaeology, Special Publication ', 2008










Analyses ofPrehistoric Ceramics from Carriacou


Figure 1. Map of Carriacou with locations of sites mentioned in this study.

Journal of Caribbean Archaeology, Special Publication ', 2008


Fitzpatrick et al.









Analyses ofPrehistoric Ceramics from Carriacou

part in answering these questions is
conducting mineralogical and chemical
analyses on sherds found at archaeological
sites to examine whether there are any
differences in sherd composition that are
spatially or temporally distinct. For this study
we used two techniques-thin section
petrography and INAA. Both of these have
had limited application, or in the case of the
latter, have never been used in the Caribbean
prior to MURR's initiation of the current
regional study. Both of these analytical
techniques have been successfully used
elsewhere for deciphering patterns of clay
and temper provenience and resource use
prehistorically (e.g., Bishop et al. 1982;
Dickinson 2001; Fitzpatrick et al. 2003;
Glascock 1992; Harbottle 1976; Neff 1992,
1994, 2000, 2001; Stoltman 2001) and
provide a strong foundation for examining
changes to ceramic assemblages over time
and space. Below we briefly discuss the
analytical procedures, followed by the results
of our analyses.
Petrography
A total of 78 sherds were examined
petrographically. The first suite of samples
(n=24) were undiagnostic surface finds from
the sites of Anse la Roche, Jew Bay, Grand
Bay, Sabazan, Great Bretache Bay, and
Tyrell Bay (Figure 1). The second suite
consisted of 54 thin sections of the same 56
sherds analyzed by INAA (described in the
section below). Many of the sections
prepared from the sherds were stained to
enhance the recognition of calcium and
potassium feldspars. All thin-sections were
petrographically examined and sherds were
classified into groups based on their
composition (Table 1). A representative
subset of 11 stained thin sections from suite 1
were point counted using a petrographic


Fitzpatrick et al.


microscope equipped with a Swift automated
stage and counter to better define group
compositions; up to -1000 points were
counted on each thin section using the Gazzi-
Dickinson method (Ingersoll et al. 1984).
Sand-sized temper components were
classified as individual mineral grains or rock
fragments according to criteria discussed in
Marsaglia (1992), with the addition of
carbonate and matrix silt/clay paste
categories (see Tables 2 and 3).
INAA
A total of 56 ceramic sherds from eight
sites on Carriacou (Lauriston, Sabazan, Jew
Bay, Tyrell Bay, Great Bretache Bay, Dover,
Grand Bay, and St. Louis), as well as the
nearby island of Petite Martinique, were sent
to MURR for INAA analyses (see Table 1).
Samples were primarily undiagnostic surface
finds with the exception of Grand Bay sherds
which were recovered from stratified deposits
during excavation of the site in 2004. A
selected portion of the sherd suite included
specimens that were identified by Harris
(2005) as Saladoid, Troumassan, or Suazan
Trousmassan. The ceramics were prepared
for INAA using standard MURR procedures
that are explained in detail in Descantes et al.
(this issue).

Results
Petrographic Analyses
Temper to matrix ratios in the 11 point-
counted sherds range from 1:2 to 1:6. The
samples contain mainly rock tempers with
carbonate debris and grog, rare to absent
(Figure 2). The main monomineralic temper
components in the sherds are plagioclase
feldspar and dense minerals such as pyroxene
and amphibole. Quartz is predominant in one
unusual sample and potassium feldspar is
predominant in another. Volcanic rock


Journal of Caribbean Archaeology, Special Publication ', 2008











Analyses ofPrehistoric Ceramics from Carriacou


Table 1. Petrographically analyzed sample sets. Group 1 = Volcanic/Pyroclastic; Group 2 = Igneous
Basement; Group 3 = Placer; Group 4 = Potassic Volcanic; Group 5 = Quartzose. FG = Fine Grained,
CG = Coarse Grained, VL = Volcaniclastic Lathwork, VM = Volcaniclastic Microlitic.

SUITE 1


Sample
Site 1-GB-i
Site I-GB-2
Site I-GB-3
Site 2-SZ-1
Site 2-SZ-2
Site 2-SZ-3
Site 2-SZ-4
Site 2-SZ-5
Site 4-AR-1
Site 4-AR-2
Site 4-AR-3
Site 4-AR-4
Site 5-GTBB-I
Site 5-GTBB-2
Site 5-GTBB-3
Site 5-GTBB-4
Site 5-GTBB-5
Site 6-TB-1
Site 6-TB-2
i; t TB I


Grand Bay
Grand Bay
Grand Bay
Sabazan
Sabazan
Sabazan
Sabazan
Sabazan
Anse La Roche
Anse La Roche
Anse La Roche
Anse La Roche
Great Bretache Bay
Great Bretache Bay
Great Bretache Bay
Great Bretache Bay
Great Bretache Bay
Terrell Bay
Terrell Bay
T1ll B


Cultural Petro
period group
--- 1
--- 1
--- 1
--- 1
--- 1
--- 1
--- 3
--- 1
--- 1
--- 1
--- 1
--- 1
--- 4
--- 1
--- 1
--- 1
--- 5
--- 1
--- 1
1


Detailed petro group
CGVL
CGVM
CGVL
CGVL
CGVL
CGVL
Placer
CGVL
Fine-grained Volcan. lathwork
FGVL
CGVM
FGVL
K-rich volcaniclastic
CGVL
CGVL
CGVL
Quartzose
FGVL
CGVL
CrG n


I%1e %-L I- lI L ady --- ---J V --
Site 10-JB-1 Jew Bay --- 1 FGVM ---
Site 10-JB-2 Jew Bay --- 2 Basement ---
Site 10-JB-3 Jew Bay --- 2 Basement ---
Site 10-JB-4 Jew Bay --- 1 FGVL ---

SUITE 2
Cultural Petro INAA
Sample Site period group Detailed petro group group


SMF001
SMF002
SMF003
SMF004
SMF005
SMF006
SMF007
SMF008
SMF009
SMF010
SMF011
SMF012


Dover
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay


Saladoid
Troumassan



Suazan
Suazan


1
1
1
1

2
1
1
1
1


CGVL
FGVL
CGVL
CGVM

Basement?
CGVL
CGVL
CGVL
tuff or tephra sample


1 CGVM


Journal of Caribbean Archaeology, Special Publication 2, 2008


INAA
group


1
2
2
2
1
1
1
1
1
U
2
2


Fitzpatrick et al.











Analyses ofPrehistoric Ceramics from Carriacou


Table 1 continued.


SMF013
SMF014
SMF015
SMF016
SMF017
SMF018
SMF019
SMF020
SMF021
SMF022
SMF023
SMF024
SMF025
SMF026
SMF027
SMF028
SMF029
SMF030
SMF031
SMF032
SMF033
SMF034
SMF035
SMF036
SMF037
SMF038
SMF039
SMF040
SMF041
SMF042
SMF043
SMF044
SMF045
SMF046
SMF047
SMF048
SMF049
SMF050
SMF051
SMF052
SMF053
SMF054
SMF055
SMF056


Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Grand Bay
Jew Bay
Jew Bay
Lauriston
Lauriston
Lauriston
Lauriston
Lauriston
Petite Martinique
Petite Martinique
Petite Martinique
Petite Martinique
Petite Martinique
Sabazan
Sabazan
Sabazan
Sabazan
Sabazan
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
Sparrow Bay
St. Louis
Tyrell Bay
Tyrell Bay
Tyrell Bay
Tvrell Bay


Saladoid


Suazan
Suazan
Suazan





Suazan











Saladoid











Suazan



Saladoid


Suazan
Troumassan


CGVL
CGVL
CGVL
CGVM
CGVM
FGVL
Placer
CGVL
CGVL
CGVL
CGVL
CGVM
Basement?
CGVL
FGVL
CGVL mixed CGVM
CGVL
CGVL
FGVL
CGVM
FGVL
CGVL
CGVL
Basement?
CGVL
CGVL
FGVL
CGVL w/ common carbonate
CGVL
CGVL
Basement
FG mixed volcaniclastics
CGVL
CGVL
K-rich volcaniclastic
FGVL
Placer
CGVL
FGVM
CGVL
FGVL
CGVL
K-rich volcaniclastic
FGVL


U: group indeterminable
---: undiagnostic or analysis not conducted


Journal of Caribbean Archaeology, Special Publication ', 2008


2
2


2
1


Fitzpatrick et al.










Analyses ofPrehistoric Ceramics from Carriacou


Table 2. Definition of counted categories and recalculated petrographic parameters.
Sand-Sized Temper Categories

Qm Monocrystalline quartz
Qp Polycrystalline quartz
P Plagioclase feldspar
K Potassium feldspar
Op Opaque dense mineral
Cpx Dense mineral clinopyroxene
Opx Dense mineral orthopyroxene
Hbl Dense mineral amphibole
Lvv Volcanic lithic with vitric texture
L vml Volcanic lithic with microlitic texture
Lvl Volcanic lithic with lathwork texture
Lvo Holocrystalline volcanic lithic composed of
plagioclase and dense minerals
Lm Metamorphic lithic
Biocl Calcareous bioclast
Biot Biotite
Musc Muscovite
Unk unknown Grain
Grog Grog (recycled ceramic fragment)

Recalculated parameters

QFL%Q=100*((Qm+Qp)/(Qm+Qp+P+K+Lvo+Lvv+Lvml+Lvl+Lm))
QFL%F=100*((P+K)/(Qm+Qp+P+K+Lvo+Lvv+Lvml+LvIl+Lm))
QFL%L=100*((Lvo+Lvv+Lvml+LvIl+Lm)/(Qm+Qp+P+K+Lvo+Lvv+Lw+Lvml+Lvl+Lm))

QmKP%Qm=100*(Qm/(Qm+P+K))
QmKP%K=100*(K/(Qm+P+K))
QmKP%P=100*(P/(Qm+P+K))

LvvLvmlLvl%Lvv=100*(Lvv/(Lvv+Lvml+Lvl))
LvvLvmlLvl%Lvml=100*(Lvml/(Lvv+Lvml+Lvl))
LvvLvmlLvl%Lvl=100*(Lvl/(Lvv+Lvml+Lvl))


fragments are present in most samples; these
exhibit microlitic to lathwork textures with
variable proportions of glass, plagioclase, and
dense minerals. Volcanic debris ranges from
fresh to altered. The compositional variability
of the tempers is represented in Figure 3
where modal results are displayed on a series
of ternary plots used in sandstone provenance
studies.


Overall the samples fall into five general
groups as illustrated by the photomicrographs
in Figure 2 and the descriptions outlined
below. Groups are roughly listed in their
order of importance:
Group 1 (n= 65): Volcanic Pyroclastic
Temper. This is the most varied of the groups
and could be further divided into several
subgroups depending on volcanic lithic types
and proportions present. Temper variability is


Journal of Caribbean Archaeology, Special Publication ', 2008


Fitzpatrick et al.











Analyses ofPrehistoric Ceramics from Carriacou


Table 3. Petrographic data and recalculated parameters.


Sample Qm Qp P K Op Cpx Opx Hbl Lvv Lvml Lvl Lvo Lm Biocl Biot Musc Unk Grog


1-GB-1 0 1 229 4
1-GB-2 2 1 45
2-SZ-4 31
6-TB-1 144 3
10-JB-2 1 1 149 5
10-JB-1 147
5-GTBB-5 116 10 11 2
6-TB-3 126 5
10-JB-4 117 1
4-AR-1 270
5-GTBB-1 2 104


13 14
2
2 161
2 2
6 11
2 11


21
4 11
1 1


Sample Total Matrix Matrix Total Total
Grains Clay Silt Matrix Count
1-GB-1 288 518 186 704 992
1-GB-2 107 331 100 431 538
2-SZ-4 303 409 21 430 733
6-TB-1 205 285 318 603 808
10-JB-2 200 362 199 561 761
10-JB-1 170 115 104 219 389
5-GTBB-5 165 621 224 845 1010
6-TB-3 161 287 448
10-JB-4 191 311 112 423 614
4-AR-1 315 0 315
5-GTBB-1 182 750 69 819 1001


0 0 1
15 12
4 81 3

1
2

1
31
3 1 1
29

QFL%
Q F


9 1
1
3 1 3


1? 4


3 4 15


2 18


QmKP% LvvLvmlLvl%
L Qm K P Lvv Lvml Lvl


0.4
3.7
0
0
1.3
0
88.1
0
0
0
0


93.8
25
0
100
0
80

96.7
40.4
71.4
59.7


a function of the proportions of volcanic
glass, plagioclase, and dense minerals within
individual volcanic fragments, as well as
degree of alteration and grain size of the
temper. The more holocrystalline varieties
are consistent with epiclastic sand derived
from basaltic to andesitic lava flows, whereas
the glassy varieties are consistent with
derivation from andesitic to basaltic
pyroclastic ash. Where altered, they could
have been mined from weathered outcrops or
diagenetically modified older units.
Group 2 (n= 6): Igneous Basement
Temper. These tempers are composed of
angular plagioclase and dense minerals


without volcanic glass. Some are limited to
plagioclase crystals only, whereas others
consist of plagioclase with minor dense
minerals. These suggest anorthosite to diorite
igneous source rocks.
Group 3 (n- 3): Placer Temper. These are
thought to be concentrated dense mineral
deposits derived from intermediate igneous
rocks based on their mineralogy
(clinopyroxene, amphibole, plagioclase) and
small amounts of intermediate volcanic rock
fragments. A high-energy beach environment
is indicated by the moderately-well sorting,
the high degree of grain rounding, and the
presence of calcareous bioclastic debris.


Journal of Caribbean Archaeology, Special Publication ', 2008


Fitzpatrick et al.


I










Analyses ofPrehistoric Ceramics from Carriacou


Figure 2. Photomicrographs of sherds illustrating temper provenance groups defined in this study [Temper Groups:
A = placer; B = Coarse-grained microlitic; C = Coarse-grained lathwork; D = Quartzose; E = K-spar volcanic; F =
Basement]. Top left (A) is a placer temper with grains of brightly birefringent carbonate bioclasts and strongly
colored dense minerals highlighted under crossed nicols. Top middle (B) is a volcanic temper of volcanic frag-
ments with microlitic texture under crossed nicols. Top right (C) is a temper of volcanic fragments exhibiting lath-
work texture under plane light. Bottom left (D) is a quartzose temper under crossed nicols to highlight the low bire-
fringence (grey to white) and shapes of the quartz temper grains surrounded by nonbirefringent paste. Bottom cen-
ter (E) shows the yellow stained, postassium feldspar crystals set in black paste under plane light. Bottom right (F)
shows a temper dominated by coarse euhedral (rectangular) to subheadral plagioclase crystals set in a dark paste.


Group 4 (n= 3): Potassic Volcanic Temper.
This temper is limited to potassium feldspar
and traces of volcanic lithic fragments.
Volcanics/pyroclastics with high potassium
feldspar are very unusual, but may be related
to hydrothermal alteration of tuffs.
Group 5 (n= 1): Quartzose Temper. This
sand temper is mineralogically mature (-90%
monocrystalline and polycrystalline quartz),
moderately sorted, with angular to
subrounded grains. The grains suggest some
mechanical abrasion and transport. Such
sediment is associated with a continental
provenance.


Discussion of Temper Provenance
The dominance of plagioclase feldspar
(QmKP plot, Figure 3), pyroxene, amphibole,
and microlitic to lathwork volcanic lithic
textures in temper Groups 1, 2 and 3 is
consistent with intermediate to basaltic
igneous sources. Overall, these tempers are
more feldspar enriched than natural basaltic/
andesitic sands from beach, fluvial, and deep
marine environments (e.g., Marsaglia 1993;
Marsaglia and Ingersoll 1992), suggesting
that many may represent rock ground up by
prehistoric potters, rather than sand gathered
from streams or beaches. Sherds with high


Journal of Caribbean Archaeology, Special Publication 2, 2008


A B C


Fitzpatrick et al.









Analyses ofPrehistoric Ceramics from Carriacou

concentrations of dense minerals (Group 3),
however, can be explained as beach/stream
players derived from volcanic and/or plutonic
igneous rocks. There is no evidence of
volcanic lithics being recycled from older
sedimentary units. Hypothetically, some of
the Group 1, 2, and 3 tempers could have
been derived from crushing shallow intrusive
to extrusive magmatic rocks on Carriacou,
but work in progress (Pavia, in prep.)
indicates little textural similarity between
Carriacou volcanic outcrops and Carriacou
tempers implying an extra-Carriacou source.
The "local" regional geology can be
subdivided into three main potential sources
of volcaniclastic temper: 1) the active
volcanic center to the north (e.g., St.
Vincent); 2) the active volcanic center to the
south (Kick'em Jenny, Isle de Caille, and
Grenada); and 3) the Grenadine islands on the
intervening southern Lesser Antilles arc
platform (SLAAP). The SLAAP includes
larger (e.g., Carriacou, Union, Canouan,
Mustique, and Bequia), as well as numerous
smaller islands (e.g., Petite Martinique,
Mayreaus, Jamesby). The series of islands on
the SLAAP between Grenada and St. Vincent
(the Grenadines), are distinctly different in
their geology, being mainly composed of
older Tertiary intrusive to extrusive igneous
rocks, epiclastic arc-derived volcaniclastic
units, chert, marl, and limestone. There is no
source for fresh pyroclastic debris on the
SLAAP islands, so the volcaniclastic temper
or pottery must have been imported from
active volcanic centers, the St. Vincent or
Grenada centers being the most proximal.
The surface of St. Vincent is mainly
covered by late Pleistocene pyroclastic
deposits, red oxidized scoria, and flows and
scoria falls from historic (1718, 1812, 1902,
1971, 1979) eruptions associated with


Fitzpatrick et al.


explosive eruptions of the Soufriere
stratovolcano (Heath et al. 1998a). The latter
consists of basaltic and basaltic andesite lava
flows and a significant rapidly deposited
scoriaceous to pumiceous yellow tuff unit
(Rowley 1978) that erupted 3600 to 4500
years B.P. (Heath et al. 1998a). Younger
overlying pyroclastic deposits were produced
by as many as 20 subsequent vulcanian
explosive eruptions. Thus, there are ample
sources of fresh pyroclastic temper on St.
Vincent. The St. Soufriere volcano also
ejected blocks of plagioclase-rich anorthosite
(Lewis 1973), a source rock that when
crushed could perhaps best explain the pure
plagioclase end member tempers.
In contrast, fresh pyroclastic sources
associated with the Grenada volcanic center
are not as extensive. Grenada is composed of
Pliocene to Pleistocene basaltic to andesitic
volcanic centers, the youngest of which is Mt.
St. Catherine. Young (1000 yrs. B.P.) fresh
pyroclastic outcrops available to prehistoric
potters would have been limited to a glassy
basaltic scoria cone near Radix village and
small volcanic centers on nearby Isle de
Caille (Arculus 1976; Devine 1995; Devine
and Sigurdsson 1995).
Future work on the geochemistry of glassy
pyroclastic tempers and plagioclase
phenocrysts may help to fingerprint the
volcanic sources) of this material. The St.
Soufriere basaltic-andesitic magmas of St.
Vincent are calc-alkaline to tholeiitic (Heath
et al. 1998a, 1998b), whereas to the south in
the southern Grenadines and Grenada the
lavas are alkaline (Brown et al. 1977).
Volcanic islands north of St. Vincent are
calcalkaline, but show no tholeiitic
tendencies (Heath et al. 1998b). There are
also distinct changes in the Sr isotopes of
plagioclase phenocrysts in volcanic centers


Journal of Caribbean Archaeology, Special Publication ', 2008











Analyses of Prehistoric Ceramics from Carriacou


Group It
Group 2 0
Group X
Groups 5


Figure 3. This series of ternary plots illustrates some of the compositional variability of the temper groups pictured
in Figure 2 and discussed in the text. Top is a QFL ternary plot where Q= total quartz grains, F = total feldspar
grains, L = total lithic or rock fragments. Note that in the case of these temper samples the lithics are all of volcanic
origin. This plot best discriminates the quartzose Group 5 from the other groups. Middle is a QmKP ternary plot
where Qm = monocrystalline quartz, K = potassium feldspar, and P = plagioclase feldspar with fields after LeMai-
tre (1989). This ternary discriminates Groups 4 and 5 from the other groups. Bottom is a ternary plot of different
volcanic lithic groups, those with vitric (glassy), microlitic and lathwork textures. This illustrates lithic texture
variability within Group 1 samples. These are the standard plots used in sand provenance studies, and so do not
completely discriminate the five groups because they do not include key components such as dense minerals, car-
bonate, or grog. For example, the placer group (3) which is dominated by dense minerals shows significant overlap
with the other groups.
Journal of Caribbean Archaeology, Special Publication 2, 2008 70


Fitzpatrick et al.









Analyses ofPrehistoric Ceramics from Carriacou

along the Lesser Antilles chain (Van Soest et
al. 2002)
The Group 4 temper is likely volcanic, but
with anomalously high potassium feldspar
content. In the Caribbean, such potassium-
rich rocks have only been described in the
literature in hydrothermally altered igneous
rocks from Puerto Rico (e.g., Pease 1976).
This does not preclude a local, here-to-fore
undocumented source on one of the Lesser
Antilles volcanic islands.
The high quartz content of the Group 5
temper (single sample) implies a continental
rather than a magmatic arc source.
Interestingly the island of Barbados is partly
constructed of quartzose sediments shed off
the South American continent into the
Atlantic Ocean basin and then scraped off
during the subduction process that has
produced the Lesser Antilles magmatic arc
(Kasper and Larue 1986). This accretionary
prism has grown to the surface and outcrops
of such quartzose sediments are present on
Barbados, which have and still serve as
temper sources for potters (Drewett and
Fitzpatrick 2000). Alternatively, the temper
could be from the parent South American
source. Unfortunately, the two Great
Bretache Bay samples were part of Suite 1
and thus not analyzed with INAA, which
might give us a better indication of whether it
was part of one of the two main
compositional groups identified chemically.
Furthermore, given that only one sample with
this temper type was found and it was a
surface find, there is also the possibility that
it represents an historic/modern fragment. In
sum, the temper compositions suggest that all
of the sherds from Carriacou appear to have
been produced outside of Carriacou and
imported to the island.


Fitzpatrick et al.


INAA
Exploratory data analyses of the 56 sherds
from Suite 2 were conducted on 33 elemental
abundance measurements before identifying
compositional groups (elemental
concentrations of nickel were removed from
subsequent analyses due to low detection
limits). Of the 33 elements, thorium,
potassium, rubidium, cesium, and antimony
created the most variance between groupings.
A two-group structure was identified in the
ceramic specimens: Group 1 (n=30) and
Group 2 (n=16). The compositional groups
can be graphically represented in principal
component space (Figures 4 and 5) and in
elemental space (Figure 6). Statistical tests
based on Mahalanobis distance-derived
probabilities using eight principal
components were conducted subsuming
90.7% of the total variance (see Tables 5-8)
to support the graphical representation of the
group structure (Figures 4-6). A cut-off of
1% was generally used to refine the
membership of Groups 1 and 2. However,
exceptions were made based on the graphical
representation of the data. Ten specimens
(18%) could not be assigned to any of the
identified compositional groups (Figures 5
and 6; Table 8). It is highly probable that
analyzing more samples could identify
additional compositional groups more clearly
in the Carriacou ceramic assemblage.
Chemical characteristics for the
compositional groups are represented in
Figure 4. Relative to Group 2, Group 1
clearly has elevated concentrations of
antimony, dysprosium, and ytterbium.
Antimony is a semi-metal whereas the other
several are rare earth elements. Antimony can
be enriched in soils developed on volcanic
materials (e.g., Terashima et al. 2002). Group
2, on the other hand, is enriched in several


Journal of Caribbean Archaeology, Special Publication ', 2008









Analyses ofPrehistoric Ceramics from Carriacou

rare earth elements, an actinide, and a
transition metal (that is Cs, La, Th, and Ta),
when compared to Group 1. Within the St.
Vincent rocks, feldspar phenocrysts contain
much less Th (-10%) as compared to the
volcanic groundmass (Heath et al. 1998a).
Thus, some of the temper chemical variation
could be explained by different proportions
of glassy groundmass versus plagioclase
crystals in sherd tempers. Furthermore, this
suggests that soils developed on more glass-
rich units might contain higher Th
concentrations than those developed on
plagioclase-rich units. Additional chemical
analyses of sherd paste and temper
components are needed to clarify these
relationships.
Given the relatively small sample size, only
tentative statements can be made about the


I-0.3


Fitzpatrick et al.


relationship between the provenance of the
sherds and their group assignments (see
Table 8). Interestingly, the Grand Bay site
has almost equal amounts of both
compositional types. Another possibly
significant observation is that only Group 1
ceramics were collected from the sites of
Dover, Lauriston, and Sabazan, while Jew
Bay only has compositional Group 2
ceramics. However, these distributions do not
appear to be geologically constrained (i.e.,
based on site location within a particular
geological formation). Finally, the
Troumassan sherds with potassic tempers
(SMF047, SMF055) do show unusual
chemistries and fall outside the compositional
range of the INAA Groups 1 and 2.
In sum, INAA of the 56 samples has
identified two distinct ceramic compositional


-0.2 -0.1 -0.0 0.1 0.2 0.3 0.4 0.5

Principal Component 1


Journal of Caribbean Archaeology, Special Publication ', 2008


Figure 4. Bivariate plot of principal components 1 and 3 displaying two compositional
groups. Ellipses represent 90% confidence level for membership in the groups. Vectors
denote elemental influences on the ceramic data. Unassigned specimens are not shown.










Table 4. Principal components analysis of 56 specimens.
Simultaneous R-Q factor analysis based on variance-covariance

PC Eigenvalue % Variance Cum. % Var.
1 0.4868 32.4368 32.4368
2 0.3097 20.6338 53.0705
3 0.2395 15.9603 69.0308
4 0.0938 6.2468 75.2776
5 0.0754 5.0249 80.3025
6 0.0605 4.0321 84.3346
7 0.0500 3.3312 87.6658
8 0.0457 3.0464 90.7122
9 0.0252 1.6769 92.3891
10 0.0233 1.5510 93.9401
11 0.0181 1.2067 95.1468
12 0.0159 1.0570 96.2038
13 0.0114 0.7624 96.9661
14 0.0105 0.6986 97.6647
15 0.0081 0.5397 98.2044
16 0.0059 0.3947 98.5992
17 0.0046 0.3039 98.9031
18 0.0035 0.2359 99.1390
19 0.0034 0.2275 99.3665
20 0.0021 0.1366 99.5032
21 0.0019 0.1241 99.6272
22 0.0012 0.0796 99.7068
23 0.0011 0.0712 99.7780
24 0.0009 0.0633 99.8413
25 0.0006 0.0409 99.8822
26 0.0005 0.0337 99.9159
27 0.0004 0.0277 99.9436
28 0.0003 0.0183 99.9619
29 0.0002 0.0140 99.9759
30 0.0002 0.0115 99.9873
31 0.0001 0.0088 99.9961
32 0.0001 0.0039 100.0000


Table 5. Mahalanobis distance calculated probabilities and
posterior classification for compositional Group 1 members.
Eight principal components were used. Probabilities are
jackknifed for specimens included in each group.


ID. NO.
SMF001
SMF005
SMF006
SMF007
SMF008
SMF009
SMF017
SMF020
SMF022
SMF023
SMF027
SMF028
SMF029
SMF030
SMF031
SMF033
SMF034
SMF035
SMF037
SMF038
SMF039
SMF040
SMF041
SMF042
SMF043
SMF044
SMF046
SMF049
SMF053
SMF054


Group 1
84.322
88.743
94.633
17.885
35.883
28.550
53.663
44.278
99.172
14.475
52.882
74.355
96.047
19.257
77.422
28.140
4.634
35.558
48.573
7.878
25.094
86.945
53.455
2.478
90.214
10.815
51.148
26.230
47.917
82.972


Group 2
2.143
1.224
2.335
0.104
0.523
0.694
3.327
3.244
1.422
0.303
2.843
0.244
0.777
0.117
0.950
3.970
0.914
2.375
1.747
0.328
0.466
1.037
0.349
0.533
0.463
0.020
1.051
1.004
0.091
2.220










Analyses ofPrehistoric Ceramics from Carriacou

Table 6. Mahalanobis distance calculated probabilities
and posterior classification for compositional Group 2
members. Eight principal components were used.
Probabilities are jackknifed for specimens included in
each group.

ID. NO. Group 1 Group 2
SMF002 0.000 78.209
SMF003 0.000 55.485
SMF004 0.000 87.512
SMF011 0.000 26.309
SMF012 0.000 97.785
SMF013 0.000 16.903
SMF014 0.000 47.163
SMF015 0.000 70.300
SMF018 0.000 59.265
SMF019 0.000 4.147
SMF024 0.000 81.479
SMF025 0.000 84.574
SMF026 0.000 31.292
SMF036 0.000 37.114
SMF050 0.000 1.010
SMF056 0.000 31.923

groups. Possible tendencies or associations
between the chemical compositions of the
sherds and their site provenance were found.
The submission of more samples from these
contexts could further test these identified
patterns as well as delineate more groups (7
of the 10 unassigned samples clearly have no
affinity with the two compositional groups)
and subgroups. Determining whether the
identified compositional groups refer to local
or exotic sources will require either the
chemical analysis of raw clay samples or the
mineralogical analysis of raw clay samples
and ceramics.
Discussion and Conclusions
Thin-section petrography (n=78) and INAA
(n=56) of two suites of ceramic sherds from
archaeological sites on Carriacou suggest that
pottery was made from primarily or even


Fitzpatrick et al.


Table 7. Mahalanobis distance calculated probabilities
and posterior classification for unassigned members
into compositional Groups 1 and 2. Eight principal
components are used.

ID. NO. Group 1 Group 2
SMF010 0.000 0.000
SMF016 0.000 0.001
SMF021 40.016 8.913
SMF032 0.000 0.168
CA 4-U'fXA A C 1 AO IQ AIA
IlSMFr V45 0 .148 0.540V
SMF047 0.000 0.000
SMF048 0.002 0.166
SMF051 0.000 0.001
SMF052 0.000 0.000
SMF055 0.000 0.839


exclusively exotic materials using volcanic
sand as temper with minor amounts of
carbonate and/or grog. The Quartzose and
Potassic groups suggests that some samples
may have tempers possibly derived from
quartzose outcrops on Barbados and potassic
outcrops on Puerto Rico. The identification of
two major chemical groups using INAA is
not suggestive of Carriacouan pottery having
either a local or exotic source; at least 10 of
the 56 (17.8%) samples in Suite 2 are outliers
that could not be assigned to either
compositional group, perhaps indicating that
they were produced and transported from
another source. Four of the five sherds from
Petite Martinique analyzed with INAA fell
into compositionally defined groups (n=3
[Group 1]; n=l1 [Group 2]), indicating that
inter-island transfer of ceramics was
occurring at least on a local scale.
Only 14 of the sherds from both suites were
stylistically identified; of these, none appear
to fall into any compositional pattern based
on temporality or cultural design, with both
early (Saladoid), middle (Troumassan
Troumassoid), and late (Suazan


Journal of Caribbean Archaeology, Special Publication 2, 2008










Analyses ofPrehistoric Ceramics from Carriacou


o 4..
4-
oGroup 12 +IM











-0.3 -0.2 -0.1 -0.0 0.1 0.2 0.3

Principal Component 1
Figure 5. Bivariate plot of principal components 1 and 3 displaying the two composi-

level for membership in the groups.







Site Group 1 Group 2 Unassigned
Lauriston 5 0 0










Sabazan 5 0 0

St. Louis 0 0 1
0- +SM OI
STyrrel -0.2 -0. -0.0Bay 2 .1




level for membership in the groups.





Table 8. Compositional group assignments and site provenience.

Petite Group Martinique Group 2 Unassigned
Total 3 0 10
Grand lBay 9 11 3

Lauriston 5 0 0
Sabazan 5 0 0
Sparrow Bay 5 1 4
St. Louis 0 0 1
Tyrrel Bay 2 1 1
Petite Martinique 3 1 1
Total 30 16 10


Journal of Caribbean Archaeology, Special Publication 2, 2008


Fitzpatrick et al.









Analyses ofPrehistoric Ceramics from Carriacou


CL Group 2
a.
V0
^o





0.8 1.0 1.2 1.4 1.6 1.8 2.0









confidence level for membership in the groups. Unclassified samples (+) are labeled.
0 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Cerium (log base-10 ppm)
Figure 6. Bivariate plot of base-10 logged cerium and thorium concentrations showing
the chemical distinctiveness of the two compositional groups. Ellipses represent 90%
confidence level for membership in the groups. Unclassified samples (+) are labeled.


Troumassoid) periods falling into Group 1
using either petrography or INAA, for
example. This could indicate that there was
little preference by prehistoric potters in
seeking out specific clay or temper resources.
It is interesting to note, however, that all
sherds from Sabazan and Lauriston in Suite
2, as well as the four Saladoid samples, fell
into the same petrographic and INAA groups
(as did four out of the five Sabazan sherds
analyzed petrographically in Suite 1;
Saladoid sample SMF051 could not be
grouped using INAA). It is also notable that
three of the six Jew Bay samples present in
both Suites 1 (n=4) and 2 (n=2) fell into
petrographic Group 2, of which only six are
known from the total number of sherds
analyzed. Whether these represent real
cultural preferences for potting is unknown.


Although the results are preliminary, thin-
section petrography, used in conjunction with
INAA, has provided a means for explaining
how ceramics were manufactured
prehistorically, the compositional diversity of
geological materials used in production, and
the differences between sherds found at
various archaeological sites. QFL and QmKP
ternary plots suggest that there are at least
five different possible sources of material,
most of which appear to be exotic. INAA
suggests two main chemical groups.
However, these observations will have to be
confirmed by additional testing of these and
other sherds. Without more information, if
the ceramics were made using non-local
resources, it would appear at this stage to
signify two distinct clay sources.
Alternatively, the geochemical groupings


Journal of Caribbean Archaeology, Special Publication ', 2008


Fitzpatrick et al.









Analyses ofPrehistoric Ceramics from Carriacou

may be related to the chemistry of the
volcanic temper components.
Research now in progress is focused on
petrographically analyzing the INAA suite of
sherds in more detail and comparing the
geochemisty from sherds and source material
from lavas and volcanic centers in the Lesser
Antilles. This will allow us to infer whether
the exotic sherds are anomalous or are a
ubiquitous component to sherd suites from
archaeological sites on Carriacou. The
analysis should also help us test models of
interisland interaction between Carriacouan
groups and those on other islands. The
preliminary data now suggests that all of the
sherds within the total Carriacou sherd suite
could be exotic, suggesting that movement of
these artifacts was widespread. This
phenomenon is not unheard of in the
Caribbean. Cordell (1998), for example,
estimated that all of the pottery found at the
Coralie site (GT-3) on Grand Turk, in the
Ostionan style dating from AD 700-1100,
was imported from Hispaniola nearly 200 km
away. In addition, three ceramic inhaling
bowls (one recovered from deposits at Grand
Bay dating between ca. AD 1000-1300 and
two unprovenienced from the Carriacou
Historical Society Museum) were dated with
luminescence to 430 + 192 BC (weighted
average). This is several hundred years older
than the earliest 14C dates of ca. AD 400 from
the island, suggesting that they may have
been heirlooms transported from elsewhere.
This hypothesis is supported by petographic
analysis which shows that these samples were
likely produced with non-local materials, and
luminescence dating of two diagnostic sherds
recovered from stratified deposits at Grand
Bay that overlap with the existing
radiocarbon chronology (Fitzpatrick et al.
n.d.).


Fitzpatrick et al.


Further examination of sherds that are
stylistically identifiable, versus undiagnostic
(which provided the bulk of samples used in
this study) will be critical for testing whether
Carriacou, or some communities within, had
greater access to imported goods versus
others and if these varied through time.

Acknowledgments
We gratefully acknowledge the assistance
given by the Carriacou Historical Society in
obtaining permission from local landowners
to survey the island, conduct excavation at
Grand Bay, and collect samples for testing.
We also thank the Ministry of Tourism in
Grenada, the Minister for Carriacou & Petite
Martinique Affairs, and the many local
individuals and businesses who have
supported the Carriacou Archaeological
Survey project over the past several years.
Petrographic analyses were made possible
through a grant to Marsaglia from California
State University Northridge and
professional development funds from NC
State University. Dozens of students
participated in the 2003-2005 field projects
and this research could not have been done
without their help. We also acknowledge the
National Science Foundation for its grant to
MURR (NSF BCS-05-04015) which made
this work possible. Thanks also to an
anonymous reviewer who provided useful
comments on an earlier draft of the
manuscript.

References
Arculus, Richard J.
1976 Geology and geochemistry of the alkali basalt-
andesite association of Grenada, Lesser Antilles
island arc. Geological Society ofAmerica Bulletin
87:612-624.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Analyses ofPrehistoric Ceramics from Carriacou

Baxter, Michael J.
1992 Archaeological uses of the biplot a
neglected technique? In Computer Applications
and Quantitative Methods in Archaeology, 1991,
edited by G. Lock and J. Moffett, pp. 141-148.
BAR International Series S577, Tempvs
Reparatvm, Archaeological and Historical
Associates, Oxford.
Baxter, Michael J.
1994 Stepwise discriminant analysis in
archaeometry: a critique. Journal of
Archaeological Science 21:659-666.
Bieber, Alan M. Jr., Dorothea W. Brooks, Garman
Harbottle, and Edward V. Sayre
1976 Application of multivariate techniques to
analytical data on Aegean ceramics.
Archaeometry 18:59-74.
Bishop, Ronald L. and Hector Neff
1989 Compositional data analysis in archaeology. In
Archaeological Chemistry IV, edited by R. 0.
Allen, pp. 576-586. Advances in Chemistry Series
220, American Chemical Society, Washington,
D.C.
Bishop, Ronald L., Robert L. Rands, and George R.
Holley
1982 Ceramic compositional analysis in
archaeological perspective. InAdvances in
Archaeological Method and Theory, vol. 5, pp.
275-330. Academic Press, New York.
Briden, J. C., D. C., Rex, A. M. Faller, and J. F.
Tomblin
1979, K-Ar geochronology and paleomagnetism of
volcanic rocks in the Lesser Antilles island arc:
Philosophical Transactions of the Royal Society of
London 291:485-528.
Brown, G. M., J. G. Holland, H. Sigurdsson, J. F.
Tomblin, and R. J. Arculus
1977 Geochemistry of the Lesser Antilles volcanic
island arc. Geochimica et Cosmochimica Acta 41
(6):785-801.
Caldwell, George
1983 A study of the volcanic rocks on the Southern
half of Carriacou, Grenadines, West Indies
Unpublished Master's Thesis: University of
Windsor, Windsor, London.
Carini, S. P.
1991 Compositional Analysis of West Indian
Saladoid Ceramics and Their Relevance to Puerto
Rican Prehistory. Ann Arbor: University
Microfilms International.


Fitzpatrick et al.


Carter, Sidney W., Bettina Wiegand, Gail A. Mahood,
Francis 0. Dudas, Joseph L. Wooden, Alan P.
Sullivan, and Samuel A. Bowring
2003 Strontium isotopic analysis as a tool for
ceramic provenance research. Abstract:
Geological Society of America, Annual Meeting,
Seattle, Nov. 2-5, 2003, p 399-405.
Cordell, Ann
1998 Possible manufacturing origins of Ostionoid
pottery from the Bahamas. Paper presented at the
55th annual meeting of the Southeastern
Archaeological Conference, Greenville, SC.
Curet, L. Antonio
1997 Technological changes in prehistoric ceramics
from Eastern Puerto Rico: An exploratory study.
Journal ofArchaeological Science 24:497-504.
Davidson, Jon
1985 Mechanisms of contamination in Lesser
Antilles island arc magmas from radiogenic and
oxygen isotope relationships. Earth and Planetary
Science Letters 72(2-3):163-174.
Davidson, J.
1987 Crustal contamination versus subduction zone
enrichment: Examples from the Lesser Antilles
and implications for mantle source compositions
of island arc volcanic rocks. Geochimica et
CosmochimicaActa 51(8):2185-2198.
Davidson, Jon, Frank Tepley, III, Zenon Palacz, and
Simon Meffan-Main
2001 Magma recharge, contamination and residence
times revealed by in situ laser ablation isotopic
analysis of feldspar in volcanic rocks. Earth and
Planetary Science Letters 184(2):427-442.
Devine, Joseph D.
1995 Petrogenesis of the basalt-andesite-dacite
association of Grenada, Lesser Antilles island arc,
revisited. Journal of Volcanology and Geothermal
Research 69:1-33.
Devine, Joseph D., and Sigurdsson, H.
1995 Petrology and eruption styles of Kick'em-
Jenny submarine volcano, Lesser Antilles island
arc. Journal of Volcanology and Geothermal
Research 69:35-58.
Dickinson, William R.
2001 Petrography and geologic provenance of sand
tempers in prehistoric potsherds from Fiji and
Vanuatu, South Pacific. Geoarchaeology 16
(3):275-323.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Analyses ofPrehistoric Ceramics from Carriacou

Donahue, John, David R. Watters, and Sarah
Millspaugh
1990 Thin section petrography of Northern Lesser
Antilles ceramics. Geoarchaeology 5:229-254.
Donovan, Stephen K., Ron K., Pickerill, Roger Portell,
Trevor A., Jackson, and David A. T. Harper
2003 The Miocene palaeobathymetry and
palaeoenvironments of Carriacou, the Grenadines,
Lesser Antilles. Lethaia 36(3):255-272.
Drewett, Lys and Scott M. Fitzpatrick
2000 Barbados Ceramics: Scientific Analysis. In
Prehistoric Settlement in the Caribbean:
Fieldwork in Barbados, Tortola and the Cayman
Islands edited by P. L. Drewett, pp. 141-146.
London: Archetype Publications.
Fitzpatrick, Scott M., William R. Dickinson, and
Geoffrey Clark
2003 Ceramic Petrography and Cultural Interaction
in Palau, Micronesia. Journal ofArchaeological
Science 30(9):1175-1184.
Fitzpatrick, Scott M., Quetta Kaye, and Michiel
Kappers
2004 A radiocarbon Sequence for the Sabazan Site,
Carriacou, West Indies. Journal of Caribbean
Archaeology 5:1-11.
Fitzpatrick, Scott, Michiel Kappers, and Christina
Giovas
In press The southern route hypothesis: Carriacou's
chronological position in Antillean prehistory. In
Island Shores, Distant Pasts: Archaeological and
Biological Perspectives on the Pre-Columbian
Settlement of the Caribbean edited by Scott M.
Fitzpatrick and Ann H. Ross. Gainesville,
University Press of Florida.
Fitzpatrick, Scott M., Quetta Kaye, James Feathers,
Jennifer A. Carstensen, and Kathleen M. Marsaglia
n.d. Evidence for inter-island transport of
heirlooms?: luminescence dating and petrographic
analysis of ceramic inhaling bowls from
Carriacou, West Indies (in preparation).
Fuess, Martin T. and John Donahue
1992 Post-Saldoid Age pottery in the Northern
Lesser Antilles: lessons learned from thin section
petrography. Paper presented at the 57th Annual
Meeting of the Society for American Archaeology
in the Sponsored Symposium: The Late Ceramic
Age in the Northeast Caribbean. April 8-12,
Pittsburgh.


Fitzpatrick et al.


Fuess, Martin T., John Donahue, David R. Watters,
and Desmond Nicholson
1993 Report on thin section petrography of the
ceramics of Antigua, Northern Lesser Antilles:
method and theory. In Proceedings of the
Fourteenth International Congress for Caribbean
Archaeology edited by A. Cummins and P. King,
pp. 25-39. Barbados: International Association for
Caribbean Archaeology.
Glascock, Michael D.
1992 Characterization of archaeological ceramics at
MURR by neutron activation analysis and
multivariate statistics. In Chemical
Characterization of Ceramic Pastes in
Archaeology, edited by H. Neff, pp. 11-26.
Madison, WI: Prehistory Press.
Harbottle, Garman
1976 Activation Analysis in Archaeology.
Radiochemistry 3:33-72.
Harris, Mary Hill
2005 Introduction to the pottery of Carriacou. Paper
presented at the 20h International Association of
Caribbean Archaeology Congress, Port-of-Spain,
Trinidad, July 2005.
Heath, Emily, Simon P. Turner, Ray Macdonald, Chris
J. Hawkesworth, and Peter van Calsteren
1998a Long magma residence times at an island arc
volcano (Soufriere, St. Vincent) in the Lesser
Antilles: evidence from 238U-230Th isochron dating.
Earth and Planetary Science Letters 160(1-2):49-
63.
Heath, Emily, Ray MacDonald, Harvey Belkin, Chris
Hawkesworth, and Haraldur Sigurdsson
1998b Magmagenesis at Soufriere Volcano, St.
Vincent, Lesser Antilles Arc. Journal of Petrology
10:1721-1764.
Herz, Norman
2001 Sourcing lithic artifacts by instrumental
analysis: In Earth Sciences andArchaeology edited
by P. Goldberg, V. T. Holliday, and C. R. Ferring,
p. 449-469. London: Kluwer Academic/Plenum
Publishers,
Ingersoll, R. V., Bullard, T. F., Ford, R. L., Grimm, J.
P., Pickle, J. D., and Sares, S. W.
1984, The effect of grain size on detrital modes: A
test of the Gazzi-Dickinson point-counting method.
Journal of Sedimentary Petrology 54:103-116.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Analyses ofPrehistoric Ceramics from Carriacou

Jackson, T. A.
1980 The composition and differentiation of the
volcanic rocks of Carriacou, Grenadines, West
Indies. Bulletin of Volcanology 43(2):311-324.
Kappers, Michiel, Scott M. Fitzpatrick, and Quetta
Kaye
2005 Automated Data Collection and 3D Modelling
of Archaeological Sites: Examining the Prehistory
and Destruction of Grand Bay in Carriacou. Paper
presented at the XX International Association of
Caribbean Archaeology Congress, Port-of-Spain,
Trinidad.
Kasper, D.C., and Lame, D.K.,
1986 Paleoceanographic and tectonic implications of
quartzose sandstones of Barbados. Tectonics
5:837-854.
Kaye, Quetta, Michiel Kappers, and Scott M.
Fitzpatrick
2004 A Preliminary Report on the Excavation at
Grand Bay, Carriacou, West Indies, June 28th -
July 31st, 2004. Papers from the Institute of
Archaeology 15:82-89.
Quetta Kaye, Scott M. Fitzpatrick, and Michiel
Kappers
2005 Continued Archaeological Investigations at
Grand Bay, Carriacou, West Indies (May 23rd-July
22nd, 2005) and the Impact of Hurricanes and other
Erosive Processes. Papers from the Institute of
Archaeology 16:108-114.
Keegan, William F.
2000 West Indian archaeology. 3. ceramic age.
Journal of Archaeological Research 8(2): 135-
167.
Lambert, J. B., L. Xue, J. M. Weydert, and J. H.
Winter
1990 Oxidation states of iron in Bahamanian pottery
by X-ray photoelectron spectroscopy.
Archaeometry 32:47-54.
Leese, Morven N. and Peter L. Main
1994 The efficient computation of unbiased
Mahalanobis distances and their interpretation in
archaeometry. Archaeometry 36:307-316.
LeFebvre, Michelle
2005 Prehistoric Exploitation of Vertebrates at the
Grand Bay Site, Carriacou, West Indies: A
Historical Ecological Perspective. Unpublished
M.A. thesis, University of Florida, Gainesville.
Lewis, Jared F.


Fitzpatrick et al.


1973 Petrology of the ejected plutonic blocks of the
Soufriere Volcano, St. Vincent, West Indies.
Journal of Petrology 14:81-112.
Mann, C. J.
1986 Composition and origin of material in pre-
Columbian pottery, San Salvador Island,
Bahamas. Geoarchaeology 1:183-194.
Marsaglia, Kathleen M.
1993 Basaltic Island Sand Provenance, in Johnsson,
M.J., and Basu, A., eds., Processes controlling the
composition of plastic sediments. Geological
Society of America Special Paper 284:41-65.
Marsaglia, Kathleen. M. and Raymond. V. Ingersoll
1992 Compositional trends in arc-related, deep-
marine sand and sandstone: a reassessment of
magmatic-arc provenance. Geological Society of
America Bulletin 104:1637-1649.
Marsaglia, Kathleen M.
1992 Petrography and provenance of volcaniclastic
sands recovered from the Izu-Bonin Arc, Leg 126.
Proceedings of the Ocean Drilling Program,
Scientific Results 126:139-154.
Neff, Hector
1992 Introduction. In Chemical Characterization of
Ceramic Pastes in Archaeology, edited by H.
Neff, pp. 1-10. Madison, WI., Prehistory Press,
Neff, Hector
1994 RQ-mode principal components analysis of
ceramic compositional data. Archaeometry
36:115-130.
2000 Neutron activation analysis for provenance
determination in archaeology. Chapter in Modern
Methods in Art and Archaeology, edited by E.
Ciliberto and G. Spoto, pp. 81-134. Chemical
Analysis Series, Vol. 155, John Wiley and Sons,
Inc.
2001 Quantitative techniques for analyzing ceramic
compositional data. In Ceramic Production and
Circulation in the Greater Southwest: Source
Determination by INAA and Complementary
Mineralogicali ,. hi,-ido.-'0i edited by D.
Glowacki and H. Neff, pp. 15-36. Monograph 44,
Costen Institute of Archaeology, UCLA.
Nordt, Lee C.
2002 Stable Carbon and Oxygen Isotopes in Soils:
In Earth Sciences and Archaeology, edited by P.
Goldberg, V. T. Holliday, and C. R. Ferring, pp
419-445. London: Kluwer Academic/Plenum


Journal of Caribbean Archaeology, Special Publication 2, 2008











Analyses ofPrehistoric Ceramics from Carriacou

Publishers.
Pease, M. H. Jr.
1976 The source of hydrothermal solutions at Puerto
Maunabo and its bearing on the base-metal-
potassium feldspar association in Puerto Rico.
Journal ofResearch of the U.S. Geological Survey
4:61-66.
Perini, Giulia, Frank J. Tepley III, Jon P. Davidson,
and Sandro Conticelli
2003 The origin of K-feldspar megacrysts hosted in
alkaline potassic rocks from central Italy: a track
for low-pressure processes in mafic magmas.
Lithos 66(3-4):223-240.
Pickerill, Ron K., Stephen K. Donovan, and Roger W.
Portell
2001 Bioerosional Trace Fossils from the Miocene
of Carriacou, Lesser Antilles. Caribbean Journal
of Science 38(10):106-117.
Pickerill, Ron K., Stephen K. Donovan, and Roger W.
Portell
2002 The bioerosional ichnofossil Petroxestes pera
Wilson and Palmer from the Middle Miocene of
Carriacou. Caribbean Journal of Science 37
(1):130-131.
Righter, Elizabeth
1997 The ceramics, art, and material culture of the
early ceramic period in the Caribbean Islands. In
The Indigenous People of the Caribbean, edited
by S. M. Wilson, pp. 70-79. University Press of
Florida, Gainesville.
Robinson, E. and Peter Jung
1972 Stratigraphy and age of marine rocks,
Carriacou, West Indies. The American Association
of Petroleum Geologists Bulletin 56(1): 114-127.
Roe, Peter G.
1989 A grammatical analysis of Cedrosan Saladoid
vessel form categories and surface decoration:
Aesthetic and technical styles in early Antillean
ceramics. In Early Ceramic Population Lifeways
and Adaptive Strategies in the Caribbean, edited
by P. E. Siegel, pp. 267-382. BAR International
Series 506, Oxford.
Rowley, Keith
1978 Late Pleistocene pyroclastic deposits of Soufriere
volcano, St. Vincent, West Indies. Geological
Society ofAmerica Bulletin 89:825-835.
Sayre, Edward V.
1975 Brookhaven Procedures for Statistical Analyses
ofMultivariate Archaeometric Data. Brookhaven


Fitzpatrick et al.


National Laboratory Report BNL-23128. New York.
Speed, R.C., Smith-Horowitz, P.L., Perch-Nielsen,
K.v.S., Saunders, J.B., and Sanfilippo, A.B.
1993 Southern Lesser Antilles Arc Platform: Pre-Late
Miocene Stratigraphy, Structure, and Tectonic
Evolution, Geological Society of America Special
Paper 277.
Steponaitis, Vincas, M. James Blackman, and Hector
Neff.
1996 Large-scale compositional patterns in the
chemical composition of Mississippian pottery.
American Antiquity 61:555-572
Stoltman, James B.
2001 The role of petrography in the study of
Archaeological Ceramics: In Earth Sciences and
Archaeology edited by P. Goldberg, V. T. Holliday,
and C. R. Ferring, C.R., pp. 297-323. London:
Kluwer Academic/Plenum Publishers.
Sutty, Leslie
1989 A Listing of Amerindian Settlements on the
Island of Carriacou in the Southern Grenadines and a
Report on the Most Important of These, Grand Bay.
In Proceedings of the 11th Congress of the
International Association for Caribbean
Archaeology, edited by P. Tekakis, A. Gus, I. Vargas
Arenas, and M. Sanoja Obediente, pp. 242-259. La
Fundacion Arqueologica, Antropologica e Historica
de Puerto Rico, San Juan (1985), Puerto Rico.
Terashima, S., A. Ohta, N. Imai. T. Okai, M. Mikoshiba,
and M. Taniguchi
2002 Geochemistry of toxic trace elements (As, Sb, Pb,
Cr, Mo, Bi, Cd, TI) in the soils from the Kanto
district, Japan; preliminary study for the soil
geochemical mapping; Part 3. Bulletin of the
Geological Society ofJapan 53:749-774.
Thirlwall, M.F., A. M. Graham, R. J. Arculus, R. S.
Harmon, and C. G. Macpherson
1996 Resolution of the effects of crustal assimilation,
sediment subduction, and fluid transport in island arc
magmas: Pb-Sr-Nd-O isotope geochemistry of
Grenada, Lesser Antilles. Geochemica et
CosmochimicaActa 60(23):4785-4810.
van As, B. and L. Jacobs
1992 A technological study of Golden Rock pottery.
In (The Archaeology ofSt Eustatius, the Golden
Rock Site. St Eustatius, Netherlands Antilles,
edited by A. H. Versteeg and K. Schinkel, pp
230-236. Publication of the Historical Foundation
St Eustatius, no. 2, and of the Foundation for
Scientific Research in the Caribbean Region, no.


Journal of Caribbean Archaeology, Special Publication 2, 2008











Analyses ofPrehistoric Ceramics from Carriacou

130.
Van Soest, Mathijs C, David R. Hilton, Colin G.
Macpherson, and David P. Mattey
2002 Resolving sediment subduction and crustal
contamination in the Lesser Antilles Island Arc: a
combined He-O-Sr isotope approach. Journal of
Petrology 43(1): 143-170.
Weigand, Phil C., Garman Harbottle, and Edward V.
Sayre
1977 Turquoise sources and source analysis:
Mesoamerica and the southwestern U.S.A. In
Exchange Systems in Prehistory, edited by T. K.
Earle and J. E. Ericson, pp. 15-34. Academic
Press, New York.
Winter, John and Mark Gilstrap
1991 Preliminary results of ceramic analysis and the
movements of populations into the Bahamas. In
Proceedings of the T i. ../uh International Congress
for Caribbean Archaeology, edited by L. S.
Robinson, pp. 371-386. Martinique: International
Association for Caribbean Archaeology.


Fitzpatrick et al.


Journal of Caribbean Archaeology, Special Publication 2, 2008












Journal of Caribbean Archaeology
Copyright 2008
ISSN 1524-4776










































































This Page Intentionally Left Blank.



Journal of Caribbean Archaeology, Special Publication #2, 2008








Journal of CaribbeanArchaeology
Copyright 2008
ISSN 1524-4776


COMPOSITIONAL ANALYSIS OF FRENCH COLONIAL CERAMICS:
IMPLICATIONS FOR UNDERSTANDING TRADE AND EXCHANGE

Kenneth G. Kelly
Department of Anthropology
University of .,in h Carolina
Columbia, SC 29208
kenneth. kelly @sc. edu

Mark W. Hauser
Department ofAfricana Studies
University ofNotre Dame
Notre Dame, IN 46556
mhauserl @nd.edu

Christophe Descantes Michael D. Glascock
Archaeological Research Facility Archaeometry Laboratory
University of California Berkeley University of Missouri Research Reactor
2251 College Building 1513 Research Park Drive
Berkeley, CA 94720-1076 Columbia, MO 65211
cdescantes@berkeley. edu glascockm@missouri. edu

There is a considerable literature that explores the significance of low-fired earthenware
production as a component of African Diaspora identity creation and maintenance both in the
West Indies and in the American S.,tinhe, ti Yet very little analysis has gone into understanding
the role of industrially-produced, low-fired, earthenware ceramics in the Caribbean. We
believe that these ceramics may be an overlooked archaeological resource. Although they do
not typically reflect the fairly rapid changes in style that make European ceramics useful for
chronology building, and whereas they do not usually exhibit stylistic and morphological
variations that enable clear identification of their origin, they were produced in great quantity
and transported around the West Indies to serve a wide variety of uses. We %t,-,-e't that
industrially-produced, low-fired, earthenwares may provide us ni ith more information than
simply their functional purpose. They may also prove useful as a key aspect of material culture
to aid in the reconstruction of trade and interaction patterns, dependant, of course on being
able to identify the place of origin of these ceramics. This article discusses compositional
analysis of archaeological ceramics and wasters (poorly fired ceramics) recovered from
historic kiln sites on the islands of Martinique and the Guadeloupe Archipelago. Compositional
data from kiln sites are then compared to ceramic sherds from excavated domestic contexts
elsewhere on these islands to begin to reconstruct trade and exchange patterns during the
French Colonial period. The results from these analyses not only point to expected routes of
trade, but also routes which contravened colonial boundaries.


Journal of Caribbean Archaeology, Special Publication #2, 2008









Compositional Analysis of French Colonial Ceramics

Historical archaeologists, ethnoarchaeologists,
and ethnographers of material culture have
documented the presence and persistence of
low-fired earthenware traditions on many
islands of the West Indies. Present day
earthenware production is known from
islands with colonial histories as diverse as
St. Lucia, Martinique, Nevis, St. Kitts, and
Jamaica (Beuze 1990; Ebanks 1984, 2000;
England 1994; Heath 1988; Hoffman and
Bright 2004). Low-fired earthenware also is
known archaeologically from these islands,
as well as Antigua, St. Croix, Montserrat, and
St. Martin (Gartley 1979; Handler 1963,
1964; Heath 1990; Petersen et al. 1999).
These low-fired earthenwares have been
studied by a number of scholars who have
viewed their production and use as aspects of
identity creation and resistance throughout
African Diaspora related sites in the West
Indies and the Southeast United States (e.g.,
Crane 1993; Ebanks 1984; England 1994;
Ferguson 1992; Hauser and Armstrong 1999;
Hauser and DeCorse 2003; Heath 1988;
Mathewson 1972; Mouer et al. 1999;
Petersen et al. 1999; ). However, an extensive
(indeed probably outnumbering hand-built
ceramics) and potentially very important
subset of low-fired earthenware production
has generally been overlooked in
archaeological investigations of the
plantation-era West Indies. These wares,
produced industrially in the Caribbean,
include sugar cones, drip pots, and other
"industrial" ceramics used in the plantation
economies, as well as other forms, and are
typically seen as evidence of the production
of sugar. This stands in contrast to hand-built
wares that are usually seen as a product of
African inspired cultural resilience and
creativity. In this article we explore how
these industrially produced low-fired
ceramics may be able to yield evidence of


Kelly et al.


their origin through compositional analysis
that will allow archaeologists to use them to
address anthropological questions of trade
and interaction rather than mere function.
This project originated as one of us (Kelly)
was recovering large quantities of low-fired,
but apparently industrially produced wares
(wheel-thrown, controlled firing, etc.) from
excavations at two plantation sites on
Guadeloupe (Kelly and Gibson 2005). The
sherds are similar to those observed on the
surface at two known pottery production sites
on the island. However, it was impossible to
ascertain their origin because they lacked any
distinguishing characteristics. We later
discussed this problem, and developed a
research plan to attempt sourcing sherds
through petrographic analysis of thin
sections. We were subsequently approached
by Christophe Descantes and invited to
participate in a region-wide survey of
ceramics using instrumental neutron
activation analysis (INAA) which was
supported by the University of Missouri
Research Reactor Center (MURR). This
allowed us to expand our samples and begin
addressing broader questions regarding
provenience and manufacture of ceramics.
In 2004, we began this study of ceramic
production in the French West Indies by
collecting ceramic sherds from as many
historic pottery production sites as we could
find on the islands of Guadeloupe and
Martinique. Although some sites were little
more than large waster heaps associated with
historically documented pottery production
sites, others were very impressive with
substantial standing remains. Of these, the
Fidelin kiln on Terre de Bas, Les Saintes (a
small island in the southern portion of the
Guadeloupe archipelago, between
Guadeloupe and Dominica) where there is a


Journal of Caribbean Archaeology, Special Publication ', 2008









Compositional Analysis of French Colonial Ceramics


Figure 1. Sampled sites in Martinique, Guadeloupe, and St Martin.


double-chamber kiln, each chamber of which
measures about 4 m square, and the still-
active pottery at Trois Ilets, Martinique, are
most notable. The sherds collected from eight
production sites (three on Guadeloupe, five
on Martinique) were augmented by sherds
excavated from two plantation sites on
Guadeloupe, an urban site in Basse Terre,
Guadeloupe, a surface collection from the
island of St. Martin, and sherds of modern
production in St. Anne, Martinique (Figure
1). In most cases these sherds were remains
of industrial pottery vessels such as sugar
drip jars and sugar cones, although some
domestic wares such as cooking and serving
vessels and pitchers were included as well.
These samples were analyzed by INAA and
optical thin-section petrography to identify
chemical, mineralogical, and physical
composition properties.


Context
Although locally- and regionally-produced
ceramics in the Caribbean have been studied
by a number of scholars (see Crane 1993;
Ebanks 1984, 2002; Handler 1963; 1964;
Hauser and Armstrong 1999; Hauser and
DeCorse 2003; Heath 1988; Mathewson
1972; McCusick 1960; Petersen et al. 1999;
Verin 1963 ), few have systematically
catalogued and completed compositional
studies of these wares. During the eighteenth
and nineteenth centuries, local earthenwares
were manufactured at scales of both craft and
industrial production as utilitarian household
wares for use by enslaved laborers and
others, and industrially as vessels used in the
processing and storing of sugar in the
plantation industries. Among the few
compositional studies that have been
undertaken of either hand-built or wheel-


Journal of Caribbean Archaeology, Special Publication ', 2008


Kelly et al.




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs