Invited presentation at the 74th Annual Meeting of the Society for American Archaeology, Atlanta, GA (2009). Archaeological provenance research has increased exponentially since nuclear reactors at Berkeley, Brookhaven, and Michigan established archaeometry laboratories in the 1960s. Yet, few first-generation laboratories exist today. Data from these labs represent substantial investments of resources and energy; however, these data are slowly being lost through laboratory closure and personnel retirement. We discuss efforts to salvage records from laboratories at Manchester and Berkeley to ensure availability to future researchers. We recommend current laboratories adopt protocols that ensure data are preserved in perpetuity. Data preservation is relevant to both current and future research, and to an historical understanding of our discipline.

1. Rescue, Preservation, and
Dissemination of Geochemical Data
Matthew T. Boulanger
Michael D. Glascock
Archaeometry Laboratory
University of Missouri–Columbia
http://archaeometry.missouri.edu/
74th Annual Meeting
April 22–26, 2009, Atlanta, GA

3.  2 case studies demonstrating the centrality of
data-management policies
 What happens to our data once our
laboratories close?
 How can we minimize potential data loss?
 Past efforts to address these issues have largely
gone unheeded or have been forgotten
 SARCAR (Beck 1984, Bishop et al. 1984)
 Archaeometric Clearinghouse (J. of Field Arch.)

4.  Est. 1970s
 Operated Regularly
between 1970s and
1990s
 6000+ samples
analyzed
 Primarily European
and Mediterranean
Pottery
Newton, G.W.A., J. Bourriau, E.B. French, and A.J.N.W. Prag
2007 INAA of Archaeological Samples at the University of Manchester.
Archaeometry 49(2): 289-299.

5.  Digital Records
Transferred to MURR
lab in 2006
 Data represent most
archaeological ceramic
samples analyzed at the
lab
 Elemental abundances
 Minimal contextual and
descriptive information
 Small number of samples
with no clear
archaeological association

6.  Goals
 Standardize data formats and concentration data
 Preserve additional products of the lab
 Analytical software
 Standard conversion factors
 Bibliographic citations
 Make data available through the WWW
 Identify and obtain relevant publications
 Reconstruct contextual and descriptive information

7.  Est. 1950s
 Operated Regularly
between 1960s and
1980s
 10,000-12,000 samples
analyzed
 Samples from Europe,
Asia, Africa, and the
Americas
Asaro, F., D. Adan-Bayewitz
2007 The History of the Lawrence Berkeley National Laboratory Instrumental
Neutron Activation Analysis Programme for Archaeological and Geological
Materials. Archaeometry 49(2): 201-214.

9.  Goals
 Transcribe elemental abundances into digital format
 Scan/transcribe contextual/descriptive data
 Produce archival quality digital images of all
correspondence, paperwork, and photographs
 Assure long-term storage of all LBNL materials
 Identify and obtain relevant publications
 Combine descriptive data with compositional data
 Collect data from COM Microfiche

10. 2
1
KIT-4 Raw
KIT-4 Adjusted
0
As La Lu Nd Sm U Yb Ce Co Cr Cs Eu Fe Hf Ni Rb Sb Sc Sr Ta Tb Th Zn
2
1
LBNL MURR
KIT-11 Raw
KIT-11 Adjusted
0
As La Lu Nd Sm U Yb Ce Co Cr Cs Eu Fe Hf Ni Rb Sb Sc Sr Ta Tb Th Zn
2
1
KIT-26 Raw
KIT-26 Adjusted
0
As La Lu Nd Sm U Yb Ce Co Cr Cs Eu Fe Hf Ni Rb Sb Sc Sr Ta Tb Th Zn
2
1
KIT-33 Raw
KIT-33 Adjusted
0
As La Lu Nd Sm U Yb Ce Co Cr Cs Eu Fe Hf Ni Rb Sb Sc Sr Ta Tb Th Zn

11.  Digital file storage facilitates, but does not
ensure, data preservation
 Legacy software and hardware
 Inconsistent file structures
 Complimented with archival-quality paper records
 Contextual and descriptive data are at best
unstandardized, and at worst absent
 Typically provided in paper documentation
accompanying samples
 Analytical protocol are rarely specified

12.  Develop, implement, and maintain storage of:
 Elemental abundances and error determinations
 Contextual and descriptive data
 Analytical Protocol
 File-naming conventions
 “AC245.per” → “Rehman_1989_NAA_Ceramic.per”
 Metadata structures
 Metadata are used to describe the context, content, and
structure of data file
 Who, what, when, where, why, and how of associated
data

13.  Models from the geological community
 National Geochemical Database
 http://minerals.cr.usgs.gov/projects/geochem_database/index.html
 Canadian Geochemical Surveys
 http://gdr.nrcan.gc.ca/geochem/
 EarthChem
 http://www.earthchem.org/
 GeoRem
 http://georem.mpch-mainz.gwdg.de/

15.  Web pages
 HTML or XHTML
 Static
 Minimal graphics, scripts, or other
interactive features
 Data distribution
 Microsoft Excel format (*.XLS)
 Widely supported, but proprietary
 Cross platform, but dependent upon software
 Comma-separated (*.CSV)
 As close to “open” as possible
 Platform and software independent

16. http://creativecommons.org/

17.  Manchester Database [Online]
 http://archaeometry.missouri.edu/datasets/uman/index.html
 Berkeley Database [In Progress]
 http://archaeometry.missouri.edu/datasets/lbnl/index.html
 MURR Database
 http://archaeometry.missouri.edu/datasets/datasets.html
Questions About These Data?
boulangerm@missouri.edu

18.  Drs. F. Asaro, G. Newton, J. Tomlinson, J. Prag, and J.
Speakman provided significant assistance with this project
 K. Walton, J. Masters, B. Alex, A. Crawford, and A. Watkins
helped transcribe and check LBNL data
 Dr. D. Schwarz and B. Chancellor are assisting in the
development of the new Archaeometry Laboratory database
The MURR Archaeometry Laboratory
is supported by a grant from the National Science Foundation (#0504015)