Abstract
Flashed targets' positions are misjudged systematically under many circumstances. Often, misjudging the moment of the flash could be responsible for the error. For instance, flashes presented near the time of a saccade or during smooth pursuit may be mislocalised because they are considered to have taken place when the eyes had reached a later orientation. Similarly, flashes presented near a moving target may appear to lag behind the target because they are considered to have occurred when the target was further along its path. But why should flashes always appear to have occurred later? We propose that this is because it is impossible to distinguish between a short intense flash and a longer less intense flash. If people cannot reliably estimate the duration of the flash they cannot know that the experimenter used an extremely short one, so they are likely to overestimate flash duration and thereby the ‘moment’ of the flash. To show that this really occurs we conducted an experiment consisting of two parts. In the first part subjects synchronised and matched the luminance of two dark flashes of different durations that were presented on a white background (56 cd/m2). In the second part each pair was presented once, with the previously set parameters, and subjects indicated whether the two flashes were identical. The results indicate that flashes lasting up to 30ms look the same as ones lasting only one frame. Even considerably longer flashes were not always judged to be different. To appear simultaneous the flashes were synchronised at about a quarter of their duration. Thus unless subjects in experiments in which targets or the eyes move near the time of the flash (correctly) assume that the flashes are of extremely short duration, they will systematically misjudge the moment of the flash, resulting in systematic localisation errors.