Publisher’s Note: A correction article relating to this paper has been published and can be found at

(1) Overview


The PAAF Project

This database has been created in the framework of the PAAF Project (Parures Amérindiennes en matériaux lithiques dans les Antilles Françaises), which were funded from 2016 until 2019 by the French Ministry of Culture and the Guadeloupe Regional Council. This project consisted in three workpackages, including chaînes opératoires studies, gemology, and the creation of a regional database and GIS. More information on the other results obtained during this project can be found in [1, 2, 3].

Archaeological context

Early ceramic sites in the Antilles, comprising ancient cedrosan Saladoid and huecan Saladoid sites, are well known to deliver remarkable collections of lapidary artwork (e.g. [4, 5, 6, 7, 8]). These first formative occupations of the Antillean archipelago are dated back to the second half of the first millennium BC until the end of the fourth century AD. Linked to a pioneering agro-ceramist dynamic, these groups are characterized by a predetermined economic system based on horticulture, fishing, hunting, foraging, and associated with the introduction of animal and vegetal species from the continent [9]. They are also distinguished by a ceramic and lapidary production testifying of an exceptional social, technological and symbolic investment, and by the settling of important long distance networks. After this specific period of time when the lapidary craftsmanship seems to be at the center of the symbolic production of the inhabitants of the Antilles, the middle, recent and late Ceramic periods decrease their investment in this type of personal ornaments (e.g. [2, 9, 10, 11, 12]).

Quite surprisingly, little work has been specifically dedicated to the study of the lapidary personal ornaments in the Antilles, despite the potential information one could extract from it. The most comprehensive study of these artifacts is clearly the one made by Cody [5, 13] based on a survey she sent by post to every archaeologist working in the Caribbean area to build a database. This major work enabled her to compare the results she obtained for the site of Pearls, on Grenada [14], to the rest of the Antilles by building a first framework of inter-islands relationship.

Since the work done at the beginning of the 1990’s, no comprehensive analysis of this part of the material production of the Amerindians has been conducted on a regional scale. Putting together the older data with the recent one will hopefully allow the research community to better understand the changes in the society of the first phases of the Ceramic age.

Spatial coverage

The geographic distribution of the dataset encompass the complete archaeological record of the Caribbean islands. However, some islands did not yield any lapidary artifacts, or at least none that we could find in the literature. We registered the lapidary artifacts for all the regions of the Caribbean arc which can be divided in the Lesser Antilles (Leeward Islands and Windward Islands), the Greater Antilles (Puerto Rico, Hispaniola, Jamaica, Cuba) and the Lucayan archipelago (Bahamas, Turks and Caicos Islands). The northernmost site in our dataset is Minnis-Ward, the southernmost is Erin Bay, the easternmost is Lovers Retreat (TOB-69) and the westernmost is E2 Fort Charles (Figure 1, Table 1).

Figure 1 

Map of the distribution of the sites registered in the database. Bahamian and Jamaican sites are represented in inserts.

Table 1

Spatial coverage of the dataset.


Trinidad TR-03 Erin Bay –61.72190 10.08838 Fewkes 1914

Bahamas BH-01 Minnis-Ward –74.51969 24.09849 Blick et al. 2010

Jamaica JA-04 E2 Fort Charles –77.80000 17.91667 Roobol & Lee 1976

Tobago TO-01 Lovers Retreat (TOB-69) –60.77424 11.22533 Harris 1980

Temporal coverage

The database compiled in this study aims at consolidating and disseminating data about lapidary artifacts in the Caribbean islands recovered from archaeological excavations or surveys for the period before the arrival of the Europeans in the archipelago. Thus, while the end limit is well known around the end of the 15th century (depending on the islands), the start limit may be different for each island based on the current knowledge of the first human occupations in the Antilles [15]. However, the start of Early Ceramic period, supposed to represent the beginning of lapidary production in this region, is generally set to ca. 400 cal BC. Most of the archaeological sites registered in the database relates to the Early and Middle Ceramic period (mainly Saladoid culture) and some to the Late/Final Ceramic period (mainly Troumassoid and Barrancoid cultures). Only one site from the Contact Period (Cayo culture) is registered. The periodization used in this work is the one proposed by Bérard [16].

(2) Methods


The database compiled in this work has been created based on two different methodologies: piece by piece first-hand analysis, and literature screening. Both methods were used in parallel for the duration of the project. Presently, the dataset of lapidary artifacts contains 4991 entries, originating from 87 sites. Data about lapidary artifacts in the French islands of the Antillean (from south to north: Martinique, Guadeloupe, St. Barthelemy and St. Martin) have been registered thanks to missions in the museums, the storage of the Ministry of Culture, and by extracting some collections to study them in continental France. The detailed methodology is described in the case studies articles [1, 2] and includes mainly photography, classic measurements with digital caliper, technological study, and mineralogical determination by eye and systematically confirmed by Raman spectroscopy. The literature review, which accounts for the most part of the dataset, has been conducted as a long-term job. As for every literature review, it includes the reading of the major works on the subject and the literature cited by these major works. In the case of our specific study, one of the major sources of information has been the proceedings of the twenty-eight International Association for Caribbean Archaeology (IACA) congresses, in which the words bead, pendant, perle, pendentif, cuenta, pendiente, perla, have been systematically searched for. Some unpublished information has also been recovered by directly contacting the archaeologists currently excavating sites, as well as diving into the reports from French commercial archaeology. Numeric literature was the main source of documentation and the search for specific words was thus done thanks to the pdf reading software. For less recent literature or scanned documents, Optical Character Recognition (OCR) was first applied to the documents. In physical books, we read the text, looked at figures and tables, and we used the index to find the information we were looking for.

Sampling strategy

The artifacts integrated in this database relates to the lapidary production chaîne opératoire from the raw material until the finished object. From sites located on French islands, they were exhaustively studied, measured, and analyzed. There was no sampling either as for the findable literature data.

Quality Control

The data that entered in the database is of heterogeneous quality. It goes from high resolution macro photos to no image at all, from Raman spectroscopy and X-Ray diffraction mineralogical studies to nakedeye greenstone determination, from recent excavations with complete sieving to surface collection. The authors made their best to find the best data about each artifact, including dissecting fieldwork reports, but the quality of the literature is very diverse. Data cleaning and consistency have been realized thanks to the use of standardized thesaurus with dropdown menus to avoid typos for most of the fields. For other fields, we created lists of values for each variable to spot the discrepancies. For numeric variable, graphics based on the measurements were explored in order to spot any outlier and check on its values. The mapping of the sites helped in checking the geographical coordinates values, since any typo would have probably set the archaeological site in the sea.


Most of the constraints relate to the literature-based part of the database, since the French artifacts have all been photographed, measured and analyzed during the project. The quality of the information in the literature is very heterogeneous, because of the seniority and/or the lack of exhaustiveness of the publications. The quality of information has been problematic for several topics of the database, including:

  • the mineralogical determination of the raw material used by the Amerindians, mostly done by naked eye by untrained archaeologists
  • the quality of the reproduction of ancient photographs in scanned or photocopied documents
  • the lack of complete description of lapidary assemblages in most of the sites. The table or text of the articles may list tens of artifacts, while the figures only depict 5 of them.
  • the difficulty of assessing the origin of the artifacts in multicomponents archaeological sites
  • the difficulty of cultural attribution for ancient excavations

Beyond these constraints related to the existing artifacts’ collections, the major issue related to the completeness of the archaeological record is of course very significant. An important part of the artifacts registered in this database come from ancient excavations or surface collections by amateur archaeologists or collectors. Therefore, even if the quality of the archaeological literature would be excellent, and we could have a perfect recording in the database of the artifacts recovered since the beginning of the 20th century, it would still lack much information for technological studies of the chaînes opératoires due to the lack of sieving, the picking of nice and complete artifacts by collectors etc. It is also necessary to underline the differences in comprehension of the archaeological stratigraphy between ancient and modern excavations, with or without radiocarbon dating, etc.

(3) Dataset description

The database created in this project is made of four related tables (Figure 2), in which the Source table is still a work in progress. Each table exists in French and English. For the sake of simplicity we will describe here only the English version.

Figure 2 

Relational map of the database.

Islands table (ISLANDS and ILES)

Island is the name of the island.

Country is the country from which the island is part of.

Region is the large area in which the island is located (Lesser Antilles, Greater Antilles, Central America, South America, North America). It includes the continent surroundings the Caribbean because this table is also used for the database of potential sources of raw materials, which is a work in progress.

Index_Island is a combination of two letters used as short notation.

ID_Island is an unique integer for each Island of the Caribbean, taken from the Global Administrative database (GADM) which gives a unique integer for each administrative subdivision in the world.

Sites table (SITES_EN and SITES_FR)

This table is related to the Islands table by the ID_Island field. It therefore automatically integrates the Island parameter from this table.

Index_Site is the unique chain of characters identifying the site. It is composed of the Index_Island, a dash, and two digits for the number of the site on this island. For example, GD-01 is the first recorded site for Guadeloupe.

Site is the name of the archaeological site.

ID_Island is an unique integer for each Island of the Caribbean, taken from the Global Administrative database (GADM) which gives a unique integer for each administrative subdivision in the world.

City is the name of the city in which the archaeological site is situated.

Longitude and Latitude are the geographic coordinates of the site. They are expressed in WGS84 decimal degrees.

Precision explains the origin of the geographic coordinates, whether from a GPS tracker, the reported data from a map or from an address given in a publication, the centroid of the city or of the island.

Dist_coast is the shortest calculated distance from the geographic coordinates to the coast.

Altitude is the altitude of the geographical coordinates taken from the SRTM Digital Elevation Model.

Type_site is the type of archaeological site, whether a cave, a village, a funerary site etc.

Nb_artifacts is the calculated number of artifacts related to this site in the BEADS table.

Period and Culture are the chronological and cultural attributions of the main occupation of the site that yielded the lapidary artifacts. They are based on the work by Bérard [16].

Date1_BP is the calendar age associated with its error Date1_BP_error and the material dated when it is known as Date1_BP_material. This is repeated three times with Date2_BP and Date3_BP and their associated errors and materials. Date1_IntCal20_S and Date1_IntCal20_E are respectively the starting and ending calibrated age using the OxCal online tool for calibration based on the IntCal and Marine20 calibration curves [17, 18]. For marine shell dates, the mean regional correction to the reservoir effect has been used (–146 +/– 114 year) based on DiNapoli et al. [19].

ref_biblio_1, 2, 3, 4 are the short citations of the references related to the archaeological site. ref_date refers to the document from which the date has been retrieved.

Beads table (BEADS and PERLES)

This table is related to the Sites table using the Site field. Some fields are thus used directly in the BEADS table thanks to the relation between both tables, so that the final user does not have to relate both tables himself. This is the case for Island, Longitude, Latitude, Period, and Culture.

Index_B is the unique character chain composed of the Index_Site, a dash, and the number of the artifact in the site.

Site is the name of the archaeological site.

Object is the kind of object that the artifact is, for example bead, pendant, raw material.

Type and Subtype specifies the shape of the object, for example a pendant can be from type zoomorphic and from subtype frog.

Colour relates to the visible main color of the artifact.

Gem_material is the gem material used to produce the artifact, it is based on the Mineral_1, 2, 3, 4, 5 and uses a list of values coming from the gemological vocabulary.

Progress states the advancement in the production of the artifact, it can be a finished object or a blank for example.

State specifies if the object is complete or broken.

Weight, Bead_length, Bead_dia_min, Bead_dia_max, Pend_Height, Pend_width, Pend_thickness, Blank_length, Blank_width, Blank_thickness are the measurement, expressed in millimeters, of beads, pendant and blanks respectively.

Nb_perforation, Pos_Perforation and Shape_Perforation specify the number, position and shape of the perforation(s).

Perfo_Dia is the smallest diameter of the perforation thus usable to hang the artifact.

Structure is the type of archaeological structure in which the artifact was unearthed, for example, a midden, a pot-hole, a burial etc.

US is the stratigraphical unit which the artifact comes from.

Square, Level and Z are the coordinates of origin of the object in the excavators’ system.

Year_excavation is the year of excavation of the site when this artifact was discovered.

Excavator contains the name of the archaeologist responsible for the excavation at the time of the discovery of this artifact.

Ref_storage and Inv_site are the references of the artifact in the curating location and the excavators’ system respectively.

Storage_Island, Storage_City, and Storage_Location resume the actual curating location of the artifact.

Year_study is the year of study for the artifacts that have been investigated by the PAAF project.

Method_carac is the analytical method used to determine the composition of the artifact.

RawMat_estim is the raw material estimated in the literature, or before the use of analytical techniques.

Notes contains remarks about the artifact that was noted during the literature screening.

Ref_biblio_1, 2, 3, 4, 5, 6 specifies the publications where the artifact has been described.

Some general information are summarized in the Table 2.

Table 2

Summary of the Site dataset.


Antigua Elliot’s (PH-03) 64 (Middle?) Ceramic Middle Cedrosan Saladoid Murphy et al. 2000

Antigua Mill Reef (PH-01) 2 Late Ceramic Mamoran Troumassoid (Mill Reef) Hoffman 1970

Antigua Royall’s (JO-11) 199 (Middle?) Ceramic Middle Cedrosan Saladoid Murphy et al. 2000

Antigua Doig’s (PA-15) 43 Early/Middle Ceramic Early/Middle Cedrosan Saladoid Gent & deMille 2003

Antigua Winthorpe Bay 1 (Middle?) Ceramic Middle/Late Cedrosan Saladoid deMille et al. 1999

Aruba Tanki Flip 2 Late Ceramic Dabajuroid Rostain 1995

Barbuda Seaview 18 Early/Middle Ceramic Early/Middle Cedrosan Saladoid Kendall et al. 2011

Bahamas Minnis-Ward 1 Late/Final Ceramic Lucayan Blick et al. 2010

Bonaire Wanapa 3 Ceramic Haviser 1990

Carriacou Grand Bay 17 (Middle?) Ceramic Late Cedrosan Saladoid Sutty 1990

Curacao De Savaan 4 Ceramic Haviser 1990

Dominique Soufrière 1 Early Ceramic Early Cedrosan Saladoid Bérard 2009

Guadeloupe Gare maritime 59 Early Ceramic Huecan Saladoid Romon et al. 2013

Guadeloupe Morel 61 Early/Middle Ceramic Cedrosan Saladoid/huecan Delpuech et al. 1996

Guadeloupe Anse à la Gourde 28 Late Ceramic Troumassoid Delpuech et al. 1997

Guadeloupe 24 rue Schoelcher 1 Early Ceramic Early Cedrosan Saladoid Etrich 2003a

Guadeloupe Allée Dumanoir 2 Early/Middle Ceramic Early/Middle Cedrosan Saladoid Etrich 2003b

Guadeloupe Anse à la Barque 1 Undertermined Undertermined Turpin 2015

Guadeloupe Anse Bertrand 2 Undertermined Undertermined Turpin 2015

Guadeloupe Anse Ste Marguerite 4 Delpuech 2007

Guadeloupe Cathédrale 6 Early/Middle Ceramic Saladoid Bonnissent & Romon 2004

Guadeloupe Butel 1 Undertermined Undertermined

Guadeloupe Grand Carbet 1 Early Ceramic Early Cedrosan Saladoid ToledoIMur 2003

Guadeloupe Ilet Gosier 1 Late/Final Ceramic Troumassoid Romon et al. 2003

Guadeloupe Les Mineurs 3 Undertermined Undertermined

Guadeloupe Plage de Roseau 1 Contact Cayo LeLay 2013

Guadeloupe Anse Vinaigri 3 Undertermined Undertermined

Guadeloupe La Ramée 3 Early/Middle Ceramic Saladoid Casagrande 2013

Grenada Pearls 1412 Early/Middle Ceramic Saladoid Murphy et al. 2000

Grenada Grand Anse Beach 3 Early/Middle Ceramic Saladoid Cody 1993

Grenada Caliviny Island 3 Early/Middle Ceramic Undertermined Bullen & Bullen 1968

Grand Turk Governor’s Beach (GT2) 5 Final Ceramic Ostionoid meillacan Carlson 1995

Jamaica C12 Logie Green 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica C7 Harmony Hall 3 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica C8 Wallman Town 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica E2 Fort Charles 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica E5 Alligator Pond 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica K13 Bellevue 6 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica Y19 Pepper 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica Runaway Bay 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica S12 Naggo Head 2 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica S8 Marlie Mount 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica T1 New Forest 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica Y19 Coleraine 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

Jamaica Y21 Fort Haldane 1 Early/Middle Ceramic Saladoid Roobol & Lee 1976

La Désirade Morne Cybèle 1 1 Final Ceramic Suazan Troumassoid Hofman 1995

La Désirade Petite Rivière 4 Late/Final Ceramic Troumassoid deWaal 2006

Martinique Anse Trabaud 2 Late/Final Ceramic Troumassoid Mattioni 1983

Martinique Vivé 40 Early Ceramic Early Cedrosan Saladoid Mattioni 1979

Martinique Moulin l’Etang 1 Early Ceramic Early Cedrosan Saladoid Bérard 2004

Martinique Macabou 1 Final Ceramic Suazan Troumassoid Allaire 1977

Martinique Diamant 1 (Middle?) Ceramic Middle Cedrosan Saladoid Vidal 1995

Martinique Pory-Papy 3 (Middle?) Ceramic/final Cedrosan Saladoid – Troumassoid

Martinique Perrinon-Doume 1 (Middle?) Ceramic/final Cedrosan Saladoid – Troumassoid

Marie Galante Cocoyer St Charles 1 Early Ceramic Early Cedrosan Saladoid Stouvenot 1999

Marie Galante Grotte Cadet 2 1 Late/Final Ceramic Troumassoid Courtaud et al. 2005

Marie Galante Stade J. Bade 4 Late Ceramic Troumassoid Serrand et al. 2016

Marie Galante Taliseronde 1 Early/Middle Ceramic Early/Middle Cedrosan Saladoid Durand & Petitjean-Roget 1991

Montserrat Trants 602 Early/Middle Ceramic Saladoid Crock & Bartone 1998

Nevis Hichmans 1 Ceramic Saladoid - Post-Saladoid Wilson 1989

Puerto Rico Hacienda Grande 16 Early Ceramic Huecan Saladoid Crock & Bartone 1998

Puerto Rico Tecla 72 Early Ceramic Huecan Saladoid NarganesStorde 1995

Puerto Rico Punta Candelero 592 Early Ceramic Huecan Saladoid Rodriguez 1991

Puerto Rico Punta Mameyes 4 Early Ceramic/récent Cedrosan Saladoid – Ostionoid elenan Ortiz-Montanez et al. 2019

Sainte Croix Prosperity 26 Early/Middle Ceramic Cedrosan Saladoid Hardy 2009

Sainte Croix Cane Bay 2 Late Ceramic Ostionoid Hardy 2008

Sainte Croix Jolly Hill 2 Late Ceramic Early Ostionoid Hardy 2008

Sainte Croix O30. Krause 9 Early Ceramic Cedrosan Saladoid/huécan Toftgaard 2019

Sainte Croix O18. Spratt Hall 2 Early Ceramic Cedrosan Saladoid/huécan Toftgaard 2019

Sint Eustatius Golden Rock 81 Early/Middle Ceramic Saladoid Versteeg 1999

Saint Lucia Lavoutte 1 Final Ceramic Suazan Troumassoid Hofman 2012

Saint Martin Baie Orientale 2 17 Late/Final Ceramic Troumassoid marmoran (Mill Reef) Bonnissent 2008

Saint Martin Hope Estate 115 Early Ceramic Cedrosan Saladoid/huecan Bonnissent 2008

Data type

The database contains mainly text and numerical information, being the description of the artifacts, their context, and their measurements. It is encoded with UTF-16. The database available via the Filemaker application also contains pictures and drawings of the artifacts.

Format names and versions

It contains the ISLANDS, SITES_EN, BEADS, ILES, SITES_FR and PERLES tables as csv files, and a bibtex file containing the references cited in the dataset.

Creation dates

The database have been created during the PAAF Project, funded from 2016 until 2019. Minor additions have been made until the publication of this article and will continue. Updated versions of this database will be uploaded, thanks to the DOI versioning support in online archives.

Dataset Creators

The database has been created in the framework of the PAAF project, led by Alain Queffelec and Pierrick Fouéré, with the technical assistance of Jean-Baptiste Caverne. The information about lapidary artifacts from French islands are the result of analysis done by Alain Queffelec, Pierrick Fouéré and Ludovic Bellot-Gurlet. The literature based records are the result of the work done by Alain Queffelec with the help of Pierrick Fouéré and Jean-Baptiste Caverne.


The database is proposed both in French and English, in the repository and in the web application. As for the GIS online application, it is proposed in French, English, Spanish and German.


License CC-BY 4.0

Repository location


The full dataset is available in the Data folder at It contains the ISLANDS, SITES_EN, BEADS, ILES, SITES_FR and PERLES tables as csv files, and a bibtex file containing the references cited in the dataset.

Filemaker server

The database, including photos and drawings, is accessible as a Filemaker server application managed by the Huma-Num service, an institutional repository of the Centre National de la Recherche Scientifique (CNRS) (Figure 3). Users can use the database in reading mode only or, if interested in participating in improving the quality of the database, can have more advanced rights by directly asking the authors.

Figure 3 

Screenshot of the online application.

The database is located here: (click on “se connecter en tant qu’invité” if you just want to be in reading mode). Database is available in French and English, by changing the model (arrow in the top left corner).

GIS availability (ArkeoGIS)

An online, free and multilingual GIS application allows to visualize the database with a cartographic projection (Figure 4). A simplified version of the dataset is indeed accessible via the ArkeoGIS platform ( Users must register first to access this geographical application, since it is controlled in order to prevent archaeological looting.

Figure 4 

A. Screenshot of the ArkeoGIS application, a simplified GIS online system. B. Zoom on Guadeloupe, showing the potential of the ArkeoGIS visualization tool.

Publication date

The last version of the database has been uploaded on the 28/02/2021.

(4) Reuse potential

This database will be very helpful for spatial and temporal analysis research in the Caribbean, including GIS and social networks studies. It provides information on the evolution and distribution of the raw materials, types of personal ornaments, stylistic evolution and distribution, for one of the major kind of personal ornaments for this region of the world.