Supplementary materials for thesis "Provenance studies of British prehistoric greenstone implements using non-destructive analytical methods" - part 4
Geochemical data obtained using field-portable non-destructive X-ray fluorescence instrumentation (PXRF), supported by magnetic susceptibility (MS) measurements, is able to narrow the provenance of IPC Group I prehistoric greenstone axes to two locations in West Cornwall: St Ives and Mullion Island. It also supports the provenance of IPC Group III axes to a greenstone quarry near Perranuthnoe, Cornwall and supports the Cornish origin of seven Irish gabbroic and doleritic axes. Provenancing methodology involved the examination of 149 axe and 73 greenstone thin sections, followed by comparison of 272 axe and over 100 exposure geochemical and magnetic analyses using Simple Component Analysis (SCA), element discrimination diagrams, Discriminant Analysis (DA) and comparative statistics (MANOVA, West). Accuracy, precision and detection limits of the PXRF are determined using international reference samples, greenstone samples and greenstone axes. PXRF instrumentation is generally linearly calibrated with zero offset, and with a compensatable bias for K, Ca, Ti, Mn, Fe, Rb, Sr, Y, Zr, Nb, Ba, Ce and Pb. Precision in measuring greenstone rock composition is 5.3% relative standard deviation of the mean based on two PXRF measurements and from elements with concentrations tentimes the PXRF detection limit (generally less than 900ppm. for major elements and less than 30ppm for trace elements). Empirical studies using newly manufactured greenstone axes show that no adjustments to PXRF measurements are required when measuring polished curved axe surfaces, and that Ti, Y, Zr and Nb PXRF measurements made on weathered surfaces are statistically indistinguishable from those measured on fresh, flat surfaces of the same composition. KT5 magnetic susceptibility meter precision is 1.5% relative when measuring typical greenstone rock with MS of 0.6 x 10-3 SI. A robust correction procedure based on estimates of axe curvature, surface relief and dimensions is shown to correct the measured MS to that of the true rock value.
This is number 4 of 5 parts. The PCD images have been converted to JPG to enable accessibility. For details of how to open the original PCD files please refer to the README file.
For more information, please refer to the thesis.