Magnetic stability is directly related to the fidelity of the magnetic record and hence, it is of particular interest to scientists using paleomagnetic results. Surprisingly, magnetic stability is not well-understood. Hodych [1982, 1986] used a wide assortment of data to demonstrate that coercivity (one measure of stability) was linearly proportion to magnetostriction divided by the saturation magnetization in magnetite and titanomagnetite. This implies that internal stress affects the magnetic stability, as has been recognized for several decades. However, detailed studies of how domain walls are affected by defects, such as dislocations, seemed to indicate that the coercivity should not exhibit the linear relationship found by Hodych [ Xu and Merrill, 1990a]. A possible solution to this puzzle was given by Xu and Merrill [1990b], who suggested that this result could be explained if there were many dislocations involved in pinning domain walls. Moskowitz [1993] extended this work to show that dislocation dipoles are likely to play a more significant role in domain wall pinning than individual dislocations. Nevertheless, one does not expect there to be too many dislocations in small MD magnetite grains where the most stable remanent magnetization resides and hence the puzzle has not been completely solved.
Micromagnetic calculations of the screening of domain moments by the movements of unpinned walls were first done by Moon and Merrill [1986], extended by Xu and Merrill [1990], and further advanced by Xu and Dunlop [1993]. In particular, Xu and Dunlop have shown how screened regions in moderately large MD grains could be very stable with respect to alternating-field demagnetization. Their results provide one possible explanation for PSD ``behavior'' in moderately large MD grains.