Optical flow fields composed of spatiotemporal luminance modulations (first-order) versus spatiotemporal contrast modulations (second-order) differ in their ability to induce the illusions of motion aftereffect (MAE) and structure from motion. The inability of second-order (SO) motions to support structure from motion except in response to simple motion stimuli has led to the suggestion that SO motions cannot support computations relating to motion-in-depth. Three experiments were used to determine the ability of first-order (FO) and SO motion stimuli to induce vection, the illusion of 3-D motion induced by image flow. Subjects reported the duration and latency to onset of the MAE and vection illusions in response to stimuli composed of multiplicative combinations of a basic vection signal (radially expanding concentric rings) with three different carriers. The stimulus patterns contained different amounts of FO and SO motion energy (ME). The results indicated that a SO motion signal can induce vection but the vection responses will have slower onsets and shorter durations than those induced by a FO motion signal. As well, the vection response to the SO signal is not the result of contamination by FO artifacts, suggesting that the visual system contains a mechanism specialized for the detection of SO motions and that the activation of these mechanisms is sufficient to induce a sense of self-motion.