1. Psychophysical experiments by scaling perceived velocities with recording of eye movements were made in 50 human subjects. 2. Velocity estimation showed linear functions for scaling with and without eye following. The comparison of motion detection with eyes stationary and following revealed different slopes for scaling but the same exponent of 1 for Stevens's power function. 3. Visual motion detection occurs under two different conditions: a) moving retinal images without eye movement result in afferent-retinal velocity information (inflow processing of visual motion signals from immobile eyes). b) Following movements of the eyes without moving retinal images are used as efferent-oculomotor information (outflow processing of angular velocity signals from moving eyes). Both conditions may be combined but were investigated separately in the experiments described. 4. Afferent-retinal motion detection causes a velocity sensation 1.6-1.66 times greater than that of efferent oculomotor motion detection. 5. When ocular following movements lag behind higher stimulus velocities, combined afferent and efferent motion detection occurs. Higher velocity sensation by moving retinal images inflow is added to the evaluated motor outflow according to the angular velocity of ocular following movements. 6. The linear scaling functions have similar slopes in velocity estimation from central and paracentral (25°) and peripheral (45° excentrie) parts of the visual field and similar standard deviations for estimations of central and peripheral projections. Stable border contrasts cause additional overestimations. 7. The results are discussed with regard to their significance for motion perception and sensori-motor regulation of ocular following movements and slow phases of optokinetic nystagmus. © 1969 Springer-Verlag.
