The depth of threshold magnetic nerve stimulation can be estimated by using thresholds from two different-sized stimulus coils and plotting their induced electric field vs. depth profiles. Stimulation is presumed to take place where the two field profiles are equal. If the two coils have unequal inductances, however, there is a relative shift in threshold between coils that alters the intersection point and the apparent stimulus depth. This systematic error arises from two sources: (1) there is a difference in the fraction of stimulator energy reaching each coil, and (2) pulse durations are different, causing threshold shifts governed by the nerve strength-duration curve. Both sources of error are additive. If the larger coil has the lesser inductance, stimulus depth is underestimated; if it has the greater inductance, it is overestimated. This can lead to large disparities in the measured depth, depending on the sets of coils used. In this paper, we show how to correct for errors introduced by unequal inductance and how this resolves discrepancies in depth measurement. Our own depth measurements in the motor area for threshold finger movements, and recalculated depths from Epstein et al., indicate that stimulation is slightly deeper (18-21 mm, average 19+ mm) than previously thought. This suggests that threshold magnetic stimulation in the motor area may arise from large, tangentially oriented fibers in the superficial white matter, or in the gray matter at the upper sulcus or lip of the gyrus.