- Two sets of phantoms have been used to calibrate a 109Cd y-ray induced K-XRF bone-lead measurement system. Both sets of phantoms are made of plaster of Paris, but the calibration lines are significantly different. This results in a significant difference for the derived concentrations of bone lead for the same person using these two sets of phantoms. This study shows that the different calibration lines are due to the different compositions of the phantoms, which can then be accounted for by adjusting the parameters related to the phantom composition in spectral analysis. Bone-lead concentrations for ten lead-exposed smelter workers were computed before and after analysis modification, and the results show that the bone-lead concentrations for the same person calculated from two sets of phantoms are not significantly different, only after the modifications are incorporated. Through these investigations, it was discovered that a common practice of setting the ratio of the calcium edge amplitude to the coherent scatter amplitude as a constant is only valid when all spectra are acquired at the same system resolution. When there is a change in the resolution between spectra, it has been determined that the ratio of the calcium edge amplitude to the coherent area should instead be used as the constant factor in the analysis program.