The differential contributions of static versus dynamic visual cues to postural control were studied in human subjects. Lateral body oscillations were measured with accelerometers located at head, hips and ankle levels, while subjects righted their balance under various mechanical conditions: on either a soft (foam rubber) support or a hard one, and in either the classical or the sharpened Romberg stance. The visual pattern (horizontal or vertical rectangular grating) was illuminated with either a stroboscopic bulb or a normal one, and control measurements were also taken in darkness for each mechanical condition. Acceleration signals were processed into their frequency power spectra, the mean area and shape of which were taken to characterize the postural skills involved and the effects of either the visual suppressions or the mechanical destabilizations. Although dynamic visual cues have already been found to play a major role in the control of lateral body sway (Amblard and Crémieux 1976), we demonstrate here that static visual cues, the only ones available under stroboscopic illumination, also make a clear though minor contribution. Hence we suggest the existence of two modes of visual control of lateral balance in man, which are well separated in terms of the frequency range of body sway: the first mechanism, which operates below 2 Hz and is strobe-resistant, seems to control the orientation of the upper part of the body; the second mechanism, which operates above 4 Hz, centers on about 7 Hz and is strobe-vulnerable, seems to immobilize the body working upwards from the feet. Thus static visual cues may slowly control re-orientation or displacement, whereas dynamic visual cues may contribute to fast stabilization of the body. In between the frequency ranges at which these two visuomotor mechanisms come into play, at about 3 Hz, there is what we call a "blind frequency", a visually neutral sway frequency which may arise from the incompatibility of visual reorientation with visual stabilization, and where vision appears unable to reduce postural sway to any marked extent. Transmission of the destabilization produced by suppression of visual cues or by mechanical methods from one anatomical level to another is also briefly discussed in terms of bio-mechanical constraints, and the correlations between various pairs of levels are considered.