Gaumond, Kim, and Molnar [J. Acoust. Soc. Am. 74, 1392–1398 (1983)] developed a Markov chain “wheel model” as a tool for estimating the stimulus-dependent and recovery-dependent factors of the spike discharge probability for auditory nerve spike trains. Using this model, we have calculated the expected PST histograms for various recovery functions (RF) and excitation functions (EF) in order to examine the distortion in the shape of the PST histogram resulting from post-spike recovery effects. For constant EF the Markov chain state occupancy distribution varies with value of the EF. Using an RF similar to that found by Gaumond, it is found that step changes in the EF lead to overshoot/undershoot in the PST histogram which decays as the state occupancy distribution approaches its new equilibrium. A plot of equilibrium firing rate versus EF value yields a compressive nonlinear function. For a slowly varying EF, the state occupancy distribution is never far from equilibrium, and the PST histogram is well-approximated by applying this nonlinear function to the EF waveform. [Work supported by NIH grants GM07564 and NS07498.]