[en] A brief outline is given of the concepts of microdosimetry and the development of the theory of dual radiation action, which was proposed as the link between the physical properties of radiation (microdosimetry) and their biological effects. A good deal of experimental evidence was consistent with the theory in its original site model form, but there were also areas of significant disagreement. More recent experiments with ultrasoft x rays and beams of correlated ions have provided critical tests which have shown that the site model does not give an adequate description of the mechanism of radiation action. A large proportion, at least, of the radiation-induced lesions are produced by highly localized energy concentrations without the need for long-range interaction of sublesions. Attempts have been made to allow for such lesions within the original postulates of the theory of dual radiation action by development of the distance model or generalized formalism using a distance-dependent interaction probability. Some of the present limitations of this approach are discussed. An alternative hypothesis, involving dose-dependent repair acting directly upon the lesions, is described. A comparison is made between consequences of these two alternative approaches. The validity of the application of microdosimetric measurements to a variety of problems is considered