This paper describes X-ray fluorescence spectography analysis of small drilled samples of museum artifacts. Dissolve the sample, absorb on the paper, dry, mill, finally press into a pellet. Since the major sample element is now diluted in cellulose, matrix effects are avoided. Standards are simpler to make with this method, or one can use commercial standards for atomic absorption. The detection limit is in the microrange. Calibration curves are generally linear. Gold is dissolved by potassium cyanide, which is non-volatile, so the calibration curves are no longer linear. Detailed theory and experimental proof of a means of compensation follow. Measure emergent radiation from 1) the sample pellet, 2) the pellet backed by a thick disc of the element sought, and 3) from the disc of the element alone. Perform these measurements on two standards as well as the sample. Calculate the sample concentration by the equations given. The author states that this method is suitable for routine analysis of very small samples from antiquities.