Vection has typically been induced in stationary observers (ie conditions providing visual-only information about self-motion). Two recent studies have examined vection during active treadmill walking-one reported that treadmill walking in the same direction as the visually simulated self-motion impaired vection (Onimaru et al, 2010 Journal of Vision 10(7):860), the other reported that it enhanced vection (Seno et al, 2011 Perception 40 747-750; Seno et al, 2011 Attention, Perception, & Psychophysics 73 1467-1476). Our study expands on these earlier investigations of vection during observer active movement. In experiment 1 we presented radially expanding optic flow and compared the vection produced in stationary observers with that produced during walking forward on a treadmill at a 'matched' speed. Experiment 2 compared the vection induced by forward treadmill walking while viewing expanding or contracting optic flow with that induced by viewing playbacks of these same displays while stationary. In both experiments subjects' tracked head movements were either incorporated into the self-motion displays (as simulated viewpoint jitter) or simply ignored. We found that treadmill walking always reduced vection (compared with stationary viewing conditions) and that simulated viewpoint jitter always increased vection (compared with constant velocity displays). These findings suggest that while consistent visual-vestibular information about self-acceleration increases vection, biomechanical self-motion information reduces this experience (irrespective of whether it is consistent or not with the visual input).