[en] One of the most important aims of analysis in the resonance region is the evaluation of neutron resonance parameters on the basis of a given formalism of the theory of nuclear reactions. However, the task of finding resonance parameters from experimental data on the energy dependence of cross-sections is subject to a number of difficulties. These difficulties are not only of a theoretical character associated with the selection of one version or another of the theory taking into account the effects necessary (interference between resonances, Doppler effect etc.), but also involve problems of principle. Whether the set of parameters found is the only possible one within the context of a single formalism used remains open. The specific features of processing the experimental data are such that even with good resolution a number of overlapping resonances (occurring as a result of the fluctuation in inter-level distances or the Doppler effect) may be classified as an isolated resonance. Moreover, even given a very weak inter-level interference and Doppler effect, unambiguous parametrization of the cross-sections is not always possible. In the present paper these questions (the choice of the approximation needed for describing experimentally observed cross-sections, allowance for inter-level interference and the Doppler effect and the possibility of ambiguous reproduction of the resonance structure of cross-sections) are examined with reference to the parametrization of the total cross-sections for non-fissionable nuclei in the low-neutron-energy region