Samples of ancient earthenware were prepared in 1:10 glass beads for the X-ray fluorescence quantitation of 10 major elements (Na, Mg, Al, Si, P, K, Ca, Ti, Mn, and Fe). Calibration standards of the glass beads were used with a mixture of reagents in an arbitrary ratio instead of rock standards. The calibration curves were constructed using the reported values of Japanese wares and clay materials. The adequacy of conventional sampling methods was statistically evaluated by examining the homogeneities of two shards (112g of Jomon ware and 92g of Yayoi ware) with respect to their crystalline phases and 10 major constituents. Both shards were found to be heterogeneous. An important finding is that in the sampling from a shard, unless at least half of the shard is homogenized, small sampling cannot represent the whole shard. The sampling was found to be only marginally reliable because the K/Ca scatter diagram for determination of provenance showed regional characteristics, and the measured analytical values showed normal (or Gaussian) distributions. Furthermore, taking eight samples or above from a shard enhances the reliability of small sampling.
ASJC Scopus subject areas