Most of the methods allow the interpretation of gravitational anomalies by the selection of masses without flexible connection with a priori data on their shape and density. Such an approach cannot be recognized as reasonable because of the uncertainty of the reference level and other reasons of ambiguity of the inverse problem of potential theory. In the framework of testing geologically based hypotheses about the shape and density of structures that are linked in advance to the observed anomalies, we propose, in our opinion, a more rational formulation of the problem. ...