During the last two decades several methods have been reported in the literature (Corbato, 1965; Gerard and Debeglia, 1975; Bhattacharyya, 1978) for the automatic construction of models that fit observed magnetic and gravity measurements by using either space or frequency domains. Whichever approach is employed, the methods proceed as follows. An initial model is proposed and a sequence of corrections is generated in order to improve the fit between the observed and calculated anomalies. These corrections are successively applied to the initial model until a certain convergence criterion is satisfied. The above corresponds to the well-known trial-and-error procedure; the relative advantages of one method over another arise from the way in which the corrections are generated.