Resonance phenomena, which offer a powerful tool of studying intricate properties of condensed matter, have for a long time been divided into electric and magnetic resonances. Electron cyclotron resonance (CR), for instance, belongs to the class of electric resonances, whereas electron paramagnetic resonance (EPR) belongs to the class of magnetic resonance phenomena. Electric resonances are excited by the electric vector of an electromagnetic wave, and electrons experience a change in their orbital state but not in their spin state.Naturally, the classification of the motion in terms of the coordinate and spin degrees of freedom, which in fact provides the grounds for such a simplified description, is possible only in the absence of spin-orbit (SO) interaction. The subject of this paper is an electron resonance of a more general type, excited due to SO interaction. The characteristic features of this resonance, called combined resonance (COR), are: (i) the electric mechanism of its excitation, and (ii) change of the spin quantum state, when the quantum numbers corresponding to the orbital motion either remain unchanged or are changing. In the former case the resonance occurs at the spin frequency of an electron and is called electric-dipole spin resonance (EDSR), or electric-dipole-excited electron-spin resonance (EDE-ESR). In the latter case it occurs at combinational frequencies, that is, linear combinations of orbital and spin frequencies. We shall call this the electric-dipole combinational frequency resonance. If the mechanism of its excitation is not specified, or if the emphasis is on the frequency of the resonance rather than on the mechanism of its excitation, we shall use the terms spin resonance (SR) or combinational frequency resonance (CFR).At present COR, first predicted by Rashba (1960Rashba ( , 1961, is being experimentally discovered and studied for various crystals with different types of symmetry. It has been observed in 3D systems, i.e. in bulk, in 2 D systems (on heterojunctions and inversion layers), in ID systems (on dislocations) and in 0 D systems (on impurity centres). Now COR is regularly used to investigate the band structure of semiconductors. The extensive use of the method is accounted for by: (i) the relatively high intensity of COR (which may exceed the EPR intensity by several orders of magnitude), (ii) the presence, as a rule, of several COR bands in the spectrum, and (Hi) the fairly specific angular dependence of their intensity. Now it is necessary to clarify, first, what the source is of the high COR intensity, and, second, why for more than 15 years after the discovery of EPR by Zavoisky (1945) and its extensive experimental investigation, COR was not observed. It is convenient to start this discussion by taking band electrons as an example.In the absence of SO interaction, an electron put in a constant homogeneous magnetic field H, performs two independent motions associated with orbital and E.L Rashba and V.I. Sheka *Very often only the electric-dipole CFR bands are ascrib...
