tiate smooth eye movements in the absence of a moving target, previous experiments have established that such movements can be evoked if the subject is required to pursue a regularly repeated, transient target motion stimulus. We sought to determine whether active pursuit was necessary to evoke such an anticipatory response or whether it could be induced after merely viewing the target motion. Subjects were presented with a succession of ramp target motion stimuli of identical velocity and alternating direction in the horizontal axis. In initial experiments, the target was exposed for only 120 ms as it passed through centre, with a constant interval between presentations. Ramp velocity was varied from ±9 to 45°/s in one set of trials; the interval between ramp presentations was varied from 640 to 1920 ms in another. Subjects were instructed either to pursue the moving target from the first presentation or to hold fixation on another, stationary target during the first one, two or three presentations of the moving display. Without fixation, the first smooth movement was initiated with a mean latency of 95 ms after target onset, but with repeated presentations anticipatory smooth movements started to build up before target onset. In contrast, when the subjects fixated the stationary target for three presentations of the moving target, the first movement they made was already anticipatory and had a peak velocity that was significantly greater than that of the first response without prior fixation. The conditions of experiment 1 were repeated in experiment 3 with a longer duration of target exposure (480 ms), to allow higher eye velocities to build up. Again, after three prior fixations, the anticipatory velocity measured at 100 ms after target onset (when visual feedback would be expected to start) was not significantly different to that evoked after the subjects had made three active pursuit responses to the same target motion, reaching a mean of 20°/s for a 50°/s target movement. In a further experiment, we determined whether subjects could use stored information from prior active pursuit to generate anticipatory pursuit in darkness if there was a high expectancy that the target would reappear with identical velocity. Subjects made one predictive response immediately after target disappearance, but very little response thereafter until the time at which they expected the target to reappear, when they were again able to re-vitalise the anticipatory response before target appearance. The findings of these experiments provide evidence that information related to target velocity can be stored and used to generate future anticipatory responses even in the absence of eye movement. This suggests that information for storage is probably derived from a common pre-motor drive signal that is inhibited during fixation, rather than an efference copy of eye movement itself. Furthermore, a high level of expectancy of target appearance can facilitate the release of this stored information in darkness.