The problem of increasing of the efficiency η of light emission by cathodoluminophores attracts attention due to the necessity to develop highly efficient field emission sources of radiation. Nowadays the energy yield of cathodoluminophores does not exceed 20-25% [1-3].     The host matrix of a cathodoluminophore absorbs the energy of fast electrons and transforms it into the energy of multiple electron-hole pairs (e-h pars) that, captured by the radiation centers, excite them. The main losses (more than 60% of the initial energy) take place at the thermalization stage. Each thermalized e-h pair creates one photon in ideal conditions. The total efficiency can be approximately described by the following equation [2]: (1 ) g hv r E,   (1) where r is the reflection coefficient for electrons, <hν> is the average energy of the emitted photon, Eg is the forbidden gap of the host matrix, and β is a coefficient, which characterizes the average energy value of e-h pair creation. Starting from the experimental data, this coefficient is usually taken as β ~ 3  [4]. As Yu.M. Popov has shown [1], this is associated with the semiconductor band structure, and β ~ 3 for direct-gap semiconductors where the effective masses of electrons and holes are approximately equal. This means that η cannot exceed 33% or if to take into account nonradiative and Stokes losses 25-30%.     It was pointed out [5] that in wide-band dielectrics (Eg > 6-8 eV) the energy value for a pair does not amount to 3-4 Eg, but only 2Eg. This increases the theoretical limit of the cathodoluminescence efficiency from 33% to 50%. Since in classical cathodoluminophores each e-h pair creates no more than one photon, the Stokes losses of the luminophore emitting in the visible range (<hν> ~ 2-3 eV) are unacceptably high (more than 50%), and at most optimistic assumptions η can not exceed 1020%. However, a so called cascade luminescence exists, when one high-energy photon creates two photons with less energy [6,7]. Cascade emission of two photons occurs due to two subsequent radiation transitions in one luminescence center and other ways can be suggested too. For example, the processes reverse to those in antistokes photoluminophores can occur (emission of a photon by the luminescence center when the “residual” energy is transferred to the other center or transmission of energy to two other centers) or emission of several photons due to successive intercenter recombinations, etc. Therefore, in principle emission of two or greater number of photons in visible range per one e-h pair is possible in a luminophore with a wide-gap matrix. Lowering of the Stokes losses in wide-gap matrices due to multiphoton radiation down to 10-20% would allow one to reach the cathodoluminescence efficiency of 30-40%.     The author acknowledges valuable discussions and comments of N. P. Soshchin and M. I. Danilkin.