Since the resolution-limiting factor in electron microscopy of biological macromolecules is not instrumental, but is rather the preservation of structure, operational definitions of resolution have to be based on the mutual consistency of a set of like images. The traditional measure of resolution for crystalline specimens in terms of the extent of periodic reflections in their diffraction patterns is such a criterion. With the advent of correlation averaging techniques for lattice rectification and the analysis of non-crystalline specimens, a more general - and desirably, closely compatible - resolution criterion is needed. Two measures of resolution for correlation-averaged images have been described, namely the differential phase residual (DPR) and the Fourier ring correlation (FRC). However, the values that they give for resolution often differ substantially. Furthermore, neither method relates in a straightforward way to the long-standing resolution criterion for crystalline specimens.
