We investigated the conditions under which inhomogeneity in electrical conductivity may significantly modify the magnetic evoked field (MEF) due to primary currents (i.e., neuronal currents) in the brain. In the case of an isolated turtle cerebellum immersed in a large bath of physiological saline, our theoretical analysis showed the cerebellar surface to significantly enhance the MEF due to a primary current, by a factor of as much as two, for experimentally determined values of the conductivities of the cerebellar tissue and saline. A further parametric investigation of the conductivity effect revealed that conductivity boundaries may significantly modify the MEF due to neuronal currents located within 1 mm of a conductivity boundary, as would be the case for active neurons near an edema, an anoxic fringe such as might occur during stroke, or a ventricle in the human head. For a stationary neural source, conductivity boundaries may modify the magnitude of its MEF without affecting its temporal waveform. However, this boundary effect was found to be small for a model geometry locally approximating cortical sources in a sulcus or a fissure, where the boundary effects from adjacent sulcal walls tend to cancel each other.