North American Archaeomagnetism
AM Sample Collecting
Because the archaeomagnetic data recorded by archaeological features are spatial, rather than temporal, they must be calibrated by
independent dating techniques before they can be used in reconstructions of secular variation. A number of researchers have identified
this process as a potential source of error that could impact the precision of secular variation records (Lengyel 1999; Tarling and
Dobson 1995; Wolfman 1990a). In order to minimize this potential, great care must be taken to ensure that the event dated through
independent means matches the event recorded by the archaeomagnetic feature, i.e., the time the feature was last substantially fired
(Dean 1978). Chronometric dating sources such as historic documents (Lengyel and Eighmy 2001; Lengyel and Sternberg 2004; Tanguy et
al. 1999), dendrochronology (Doyel and Eighmy 1994; LaBelle and Eighmy 1997), radiocarbon (Kean et al. 1997), thermoluminescence (Becker
et al. 1994), or artifact seriation (Sternberg 1982) are typically combined with archaeological inferences to generate the independent
estimate for the target event, which is then assigned to the archaeomagnetic data.
Once a suitable feature has been identified for archaeomagnetic research, the collector must obtain a sample of this feature in such
a way that the original orientation of the sampled material is maintained. To do this, a set of between eight and twelve cubic pedestals,
or specimens, are carved into the material. Then, clay collars are placed around each specimen and a nonmagnetic mold, usually made
of aluminum, is seated on the clay. The purpose of the clay collar is to allow the mold to be leveled to control for the inclination
of the specimen. Once the mold is leveled, a nonmagnetic plaster, usually a gypsum cement, is poured into the mold, encasing the top
and sides of the specimen. The orientation of one side of the mold is then taken, and the direction of measurement is scribed on the
top of the specimen. The specimens are then broken free from the feature and the bottom of each is capped with plaster. The fully
encased specimens are then removed from the aluminum mold and packaged for transport to the lab.
When these specimens are processed in the lab, the data collected from each are averaged to obtain the mean values for the feature.
By averaging the data from at least six specimens, the archaeomagnetist is largely able to correct for errors derived from differences
in mineralogy, weathering, and firing times and temperatures over the whole of the feature, as well as for errors in specimen orientation
made by the collector (Tarling and Dobson 1995). The archaeomagnetic data recovered from an archaeological feature are typically described
in terms of the averaged declination (D) and inclination (I) for the feature, the precision (k) and angle of confidence (alpha95)
associated with the measured direction (these are measures of the error associated with the direction), and the number of specimens
(N) included in the feature average. In the U.S., these data are typically converted to the paleomagnetic pole location for the geomagnetic
field that would have produced the measured declination and inclination.
Paleomagnetic pole (red dot) for the field that would have produced the declination and inclination measured in the sampled feature. The green oval surrounding the pole indicates the region of 95% confidence.