Experimental observations of diffusional broadening of the Mo¨ssbauer resonance in polycrystalline solids have generally been analyzed on the basis of a liquidlike model which neglects the restriction of atomic jumps to crystallographically determined directions. A calculation based on the theoretical results of Chudley and Elliot for single crystals, averaged over all directions, shows that a substantial correction to the liquidlike result is required. Application of this correction, along with a simple assumption about the correlation factor largely eliminates the existing discrepancies between theory and experiment.