The purpose of the present experiment was to examine developmentally the effect of the degree of unity in a verbal auditory stimulus given during the standard time of 40sec on the length of the estimated time by the method of reproduction, that is, on the length of the reproduced time. The Ss were 16 kindergarten children, 48 elementary school children in 1st, 3rd and 5th grades, 16 junior high school students in 7th grade and 16 undergraduate students (see Table 1). The stimuli used consisted of 7 categories each of which included 3 stimuli of the same kind. They were (1) vain (two signal sounds of the beginning and the end mediated by no stimulus), (2) letters (59 letters), (3) nonsense syllables (20 syllables 59 letters), (4) shorter series of words (20 words, 59 letters), (5) longer series of words (44 words, 200 letters), (6) a tale for children (200 letters), and (7) a tale for adults (200 letters). A total of twenty-one stimuli were used. Each S was given all of 21 trials, but the order of stimulus presentation was at random.<br>The results were as follows:<br>1. The reproduced time generally decreased in length in the following order, that is, (1) vain, (2) letters and longer series of words, (3) nonsense syllables and shorter series of words, and (4) tales for children and for adults (see Tables 4 and 5). It seems that this order corresponds to the degree of unity in the stimulus, in other words, to the number of perceived changes stated by P. Fraisse.<br>2. The younger the subjects except kindergarten children were, the larger was the effect of the degree of unity mentioned in 1. This effect was in adults insignificant (see Fig. 1 and Table 6). Those findings show the general trend in children, implying the positive meaning of the developmental approach to this problem.<br>3. The younger the subjects were, the shorter was their reproduced time and that in adults was near the standard time (see Table 2).<br>It seems that the reproduction of the vain time requires patience on the subjects, so in younger children the frequency of perceived changes during the reproduced time may by larger than during the standard time because of their impatience. With development the frequencies of both standard and reproduced times may gradually coincide and become stabilized so that the reproduced time gets nearer the standard time.<br>4. The reproduced time of young girls was shorter than that of young boys, particularly at the first state of the trials (see Table 8). The causes for this difference cannot be detected from the present data.