The physical meaning of resolving power when the shape of the involved diffraction pattern is unknown is experimentally studied with the double crystal x‐ray spectrometer. Our concepts of resolving power have not been extended in effect beyond the formal numerical definition and now there is need for knowing more completely the perturbations introduced by thed&lgr; interval. Sixteen different pairs of crystals, quartz and calcite, cleavage and etched surfaces, were used in measuring the widths and shapes of theK&agr; doublet lines of Mo, Cu and Ti and theL&agr;1line of Ag with the spectrometer in various parallel and antiparallel positions. The crystal diffraction patterns are evidently asymmetrical, in qualitative agreement with theory; and the asymmetry evidently varies with different pairs of crystals as well as with wave‐length. The (1, −1) width is invalidated as an adequate criterion of resolving power. New criteria for ``spectrometric'' perfection of calcite crystals are proposed and an equation is given for correcting observed x‐ray line widths for the effects of ``spectrometrically'' perfect crystals.