Abstract

Several possible operations on visual curve tracing were investigated using elementary stimuli without interweaving curves. Experiment 1 required subjects to determine whether a line segment was connected or disconnected by one or two gap(s). The effects of line complexity, gap redundancy, two-point distance, and gap distance were assessed. Mean response time for both gap and no-gap trials increased linearly with two-point distance (i.e., linear distance between the ends of the line segment). Furthermore, two-point distance interacted with line complexity. Therefore, response time increased with the distance along the curve, consistent with the notion of "curve tracing"(Ullman, 1984). In the gap trials, however, mean response time increased linearly with gap distance (i,e.. distance between a central point and a gap). Furthermore, mean response time for double-gap trials was faster than that for single-gap trials. Using a half circle with a gap and two Xs. Experiment 2 tested whether the curve tracing operation is carried out by an exhaustive or by a self-terminating process. Again, mean response time in same trials increased linearly with the distance between two Xs. However, mean response time in different trials remained unaffected by the distance between two Xs. Taken together, effects of gap distance and gap redundancy are best explained by a self-terminating curve tracing process.