• Radio images at a single frequency give us two-dimensional information, but no height information • Different frequencies refer to different layers of constant magnetic field: multiple frequencies therefore reveal different heights (peel away layers), but the scale is not measured • Can in principle use div.B = 0 but we do not measure field direction (although information from comparing x and o from same layer?) and we do not measure weak fields, so we cannot solve this problem with radio data alone (?) • But in conjunction with extrapolations, which also have their own limitations, progress may be possible