In order to disclose the so far little-known nature and practical implications of possible dead-time phenomena in neutron spectra measured with sequentially operating reverse Fourier time-of-flight spectrometers, an analysis of non-homogeneous, especially periodic, Poisson processes deformed by constant dead time is carried out. Calculations based on this treatment show that, unless the neutron beam is modulated with frequencies restricted to integer multiples of the inverse dead time, notable spectrum distortions may occur even at counting rates not higher than a few hundred neutrons/s. Such deformations depend strongly on the shape of the true time-of-flight pattern: e.g. monochromatic peaks become asymmetric and give rise to “ghost” peaks or jumps shifted towards longer flight times by a distance equal to the dead time, whereas continuous spectra, besides producing similar but broader “ghost” effects, will be distorted throughtout with the result that their tails appear more or less depressed. Of the possible methods of preventing spectrum distortions, a simple but effective technique is discussed in which the dead time is deliberately extended so as to always make it equal to the nearest integer multiple of the modulation period. The usefulness of this procedure, as well as the dead-time phenomena theoretically predicted for monochromatic peaks, are well confirmed by two examples of real reverse Fourier time-of-flight experiments on pyrolytic graphite crystals.