Some criticisms made by Gokhale (1974) on the present author's (1972) work on the cooling of a sunspot are met. It is agreed that axisymmetric models of magnetic and velocity fields are inadequate with respect to the zero divergence conditions, and that nonaxisymmetric models must be considered even if only in qualitative form. It is pointed out that the failure of Wilson's (1972) axisymmetric models to satisfy the zero divergence condition at all points where r is below a critical value does not invalidate some of the other conclusions. Calculations are carried showing the existence of a region where the zero divergence condition is satisfied to first order. It is agreed also that the power converted to magnetic form during field amplification is not a likely explanation of the residual energy deficit associated with sunspots.