Microanalysis data depe nd strongly on the electron beam energy used in the S E M and the effects are more pronou nced with bulk samples than with thin sections. The situation can be used to advantage but there are also significant problems. The data and any quantitative corrections can be quite different at different beam energies. Th e spectra look different and this can be both helpful and confusing, with low ene rgy x-ra y lines generally more prominent at lower voltages. S p ec tra at several different voltages are particularly informative, both for quantitative analysis and to deco nvo lute three dimensional geometries. There may be no single perfect voltage for an analysis.