Although microstress is well-recognized as playing a very important, and sometimes dominant, role in ''pinning'' domain walls in minerals of interest to palaeomagnetists, it has often been treated in rock magnetism in a grossly oversimplified and incorrect manner. This paper attempts to review and extend the know of how microstress in general and dislocations in particular affect the motion of domain walls. It is found that only such unrealistic cases as a large grain containing a single dislocation is the microcoercivity, hc, independent of domain wall thickness, w. In a first order approximation, hc is linearly proportional to (λwm)/MS, where MS is the saturation magnetization, λ is some appropriate magnetostriction constant, and m is a constant determined by the distribution of microstress. A reasonable upper limit for m for magnetite based on experimental studies is found to be 0.5. Both positive and negative values for m are allowed. Microstress acts like a narrow-band-pass filter in the ''pinning'' of domain walls. Domain walls that have thicknesses roughly 1/5 of the microstress wavelength are far more effectively impeded by the stress than are larger or smaller walls. This implies that scientists can reduce their focus to a narrow range of stress wavelengths that effectively ''pin'' domain walls and that it is inappropriate to use sinusoidal ranging stress fields to model microstress. Moreover, because domain wall thicknesses usually vary with temperature, the most effective pinning sites are also temperature dependent, even when defects are completely immobile. ¿ American Geophysical Union 1989