We report results of a ''blind test'' of the algorithm developed by Kamide et al. (1981) (KRM algorithm) for deriving ionospheric currents and electric fields using data from ground magnetometers. The ''ground magnetometer data'' used for the test came from the results of a substorm simulation performed with the Rice University Convection Model (Harel et al., 1981a). Pedersen, Hall, field-aligned, and ring currents are included in the Rice Convection Model, but earth currents and polar cap currents are not. A Biot-Savart routine was used to integrate over the model current system and compute vector magnetic fields ΔB at 96 grid points on the ground. At the Natĩonal Oceanic and Atmospheric Administration (NOAA), current j and electric-field E distributions were derived by using the conduc˜tivity model and ground ΔB values provided by Rice. Plots of j and E were then exchanged, one set giving the Rice-computed j and E on which the ''theoretical magnetometer data'' were based and one set showing j and E computed at NOAA by inversion of the ''magnetometer data.'' At auroral latitudes, where currents and electric fields are strongest, the j and E plots are in encouragingly goodd agreement, with total polar cap potential drops agreeing to better than 10%. Total electrojet strengths estimated from the KRM inversion algorithm average about 0.84 of the electrojet strengths in the source-current distribution. Overall, the test confirms the ability of the algorithm to deduce patterns of auroral ionospheric currents, Birkeland currents, and electric fields reasonably well from ground-based magnetic data when the ionospheric conductivity is known.