It has been shown that, even for very fast and short duration movements, seeing one's hand in peripheral vision, or a cursor representing it on a video screen, resulted in a better direction accuracy of a manual aiming movement than when the task was performed while only the target was visible. However, it is still unclear whether this was caused by on-line or off-line processes. Through a novel series of analyses, the goal of the present study was to shed some light on this issue. We replicated previous results showing that the visual information concerning one's movement, which is available between 40degrees and 25degrees of visual angle, is not useful to ensure direction accuracy of video-aiming movements, whereas visual afferent information available between 40degrees and 15degrees of visual angle improved direction accuracy over a target-only condition. In addition, endpoint variability on the direction component of the task was scaled to direction variability observed at peak movement velocity. Similar observations were made in a second experiment when the position of the cursor was translated to the left or to the right as soon as it left the starting base. Further, the data showed no evidence of on-line correction to the direction dimension of the task for the translated trials. Taken together, the results of the two experiments strongly suggest that, for fast video-aiming movements, the information concerning one's movement that is available in peripheral vision is used off-line. (C) 2003 Elsevier Science B.V. All rights reserved.