Anisotropy of low field magnetic susceptibility (AMS) is a versatile petrofabric tool. For magnetite, AMS primarily defines grain-shape anisotropy; for other minerals, AMS expresses crystallographic control on magnetic properties. Thus, we may infer the orientation-distribution of a dominant mineral from the AMS of a rock. AMS principal directions can record current directions from sediment, flow-directions from magma, finite-strain directions from tectonized rocks and stress-directions from low-strain, low-temperature, neotectonic environments. AMS measurements may reveal some aspects of the strain-path, where carefully selected.For example, we may compare different parts of a heterogeneously strained domain, different minerals in a homogeneously strained site, AMS with schistosity/mineral lineation, and AMS with remanence-anisotropy. Such measurements isolate the orientation-distributions of different minerals, adding a temporal scale to the kinematic sequence. Normally, we can interpret the principal directions of AMS distributions as a physically significant direction, such as a current direction, magmatic flow or finite-strain axis. However, calibrating the AMS ellipsoid shape against the magnitude of the controlling physical process is very difficult. Primarily, this is because the shape of the AMS ellipsoid combines contributions from several minerals whose individual AMS ellipsoids are of different shape. Thus, small variations in the proportions of minerals change the shape of the rock's AMS ellipsoid, even if the alignment process were of constant intensity. In deformed rocks, AMS is more strain-sensitive than calcite twinning or the alignment of calcite or quartz c-axes. Not all AMS fabrics relate to crystallographic or grain alignment. First, displacement fabrics generate AMS where an isotropic matrix of high susceptibility displaces unevenly spaced objects of low susceptibility and suitable scales. Second, AMS location fabrics occur where sub-isometric magnetite grains are close enough, in certain directions, for their demagnetizing fields to interact. This accounts for the AMS of many magnetite-dominated signals where there is no aligned magnetite. Third, the AMS of single-domain magnetite is inverse to shape so that such grains may oppose the AMS contribution of parallel minerals. Finally, transitional sedimentary-tectonic or magmatic-tectonic fabrics yield smeared, temporal sequences of AMS principal directions that cannot be immediately attributed to a single alignment process. These transitional AMS ellipsoids mix primary and secondary AMS components, making it difficult to characterize either component. However, such fabric combinations may permit us to recognize the sense of shearing in flow processes.