In this paper the author proposes a hypothesis to explain a wavelength-fixing mechanism for rail corrugation. Corrugation growth was predicted using a theoretical model that took into account multiple-wheel interaction and the Doppler effect of rail vibration caused by wheel movement. The wheel-rail contact forces due to virtual sinusoidal irregularities on the rail head were calculated, and the corrugation growth at an arbitrary wavelength was predicted. To verify the hypothesis, the author conducted a precise investigation into the corrugation wavelengths on various track sections using vertical acceleration of the axle box. Even on a specific section of track, large-amplitude corrugation was not generally found at a specific wavelength but rather in a few wide-band ranges. Moreover, in each range where significant corrugation occurred, several sharp peaks were observed at constant frequency intervals. These peaks were found to appear at regular intervals approximately equal to fractions of a wheelbase length, and the theoretical prediction showed good agreement with measured corrugation characteristics. It was concluded that various kinds of corrugation are the result of rail vibration interference excited by multiple wheels.