It was confirmed by the reproduction method with 20 adults that (a) the tau effect, i.e., the greater the temporal separation between successively presented stimuli with a constant spatial interval, the greater the apparent spatial interval, occurred, especially in the situation in which the participants were assumed to form a set easily using temporal interval as the cue for spatial interval estimation, and (b) the anti-tau effect occurred in the special situation in which the participants were assumed to form a set using speed of an imaginary movement as the cue for it. These findings supported the hypothesis of "cue-selection sets."