No mention has been made in the preceding article by Hoffman and Horton concerning two limiting aspects of the magnetotelluric method. So that the reader is not misled about the significance of the analyses presented, the following remarks are added. First, as Price [1962] and Wait [1962] have noted, the Cagniard method of interpretation of magnetotelluric data assumes a uniform plane wave source. The magnetotelluric soundings made at Tikhaya Bay, U.S.S.R., unquestionably involve magnetic and electric field perturbations which arise in modulation of the auroral electrojet; these are sources of finite dimensions and demand a modified magnetotelluric interpretation. The idea that pi 2 micropulsations (as well as bay disturbances) originate in modulations of the auroral electrojet is due to Campbell [1964] and has been developed by O'Brien [1965], Other geomagnetic disturbances observed near the auroral zone may have originated deep in the exosphere but were propagated in the Alfven mode over only a limited region of space. Thus it is probable, in our opinion, that a substantial part of the geomagnetic spectrum observed at Tikhaya Bay and chosen for analysis by Hoffman and Horton has arisen in sources of finite dimensions. The question therefore arises, ‘Is the plane wave approximation valid?” To answer this question we follow an approach used by Wait [1962] and by Madden and Nelson [1964] to define a critical horizontal scale length Lc below which the fields change by an amount sufficient to cause the apparent resistivity to depart by more than 10% from the plane wave value. This critical length must be computed for each earth model studied; for the data of Horton and Hoffman (a layer 30 km thick with a resistivity of 10 ohm m underlain by an infinite half-space of resistivity 0.1 ohm m) the critical length is shown in Figure 1. For this particular model the relationship is very nearly linear over the period range considered. It is evident that at the longer periods, √T over 30, the fields must be ‘uniform’ over distances greater than 40 km, which may be questionable for the assumed sources in this region. It is also quite probable that the very low values of apparent resistivity at the longer periods are caused by the nonuniformity of the fields and are not necessarily related to the true resistivity distribution.