It has been known that the mesocotyl of the japonica type of rice does not elongate more than 1 cm even in darkness. In the previous paper, however, it was shown that the mesocotyl of some seedlings of a japonica rice elongated more than 1 cm, whereas the mean mesocotyl length of total seedlings was less than 3 mm. Assuming that this variation of mesocotyl length may be caused by some physiological differences in the seeds, the following experiments were conducted. Rice seeds, Oryza sativa L. cv. Sasanishiki, belonging to the japonica type were used as materials. After harvesting, the panicles were dried 45 days under the room conditions in 1972. Seventy-two panicles with each 8 primary branches (Fig. 1) were selected for the experimental use. Terminal (I), middle (V) and lowest (VIII) branches were cut off at their bases. They were sterilized, washed and placed on a sheet of wet cotton wool, being incubated in darkness at 30°C for 11 days. The almost all seedlings of the terminal and the middle branches had short mesocotyls, whereas the some seedlings of the lowest (VIII) branches, especially located at 2' and 3' positions on secondary branches, produced long mesocotyls (Fig. 1, 2 and Table 1, 2). When the seeds located at 1, 2, 3, and 4 positions of the upper primary branches (x), and those located at 2', 3' and 4' positions of the lower secondary branches (●) were sampled (Fig. 1), and incubated separately in darkness at 30°C for 10 days, 4 % of the former and 50% of the latter plants produced long mesocotyls (more than 6 mm), respectively. The growth pattern of these plants with long mesocotyls was very similar to that of the plants treated with abscisic acid (ABA) as shown in the previous study (Fig. 2). The seeds harvested in the falls of 1970, 71 and 72, i.e. 2.1, 1.1 year and 1.5 month after harvest, respectively, were sterilized, and placed on a sheet of wet cotton wool in Petri dishes, being applied with or without 10 ppm of gibberellic acid (GA3), and were incubated in darkness at 30°C for 10 days. As shown in Fig. 3, a half of the 2.1 year seeds did not germinate, whereas almost all seeds germinated in both 1.1 year and 1.5 month seeds. In 2.1 year old seeds, the frequency of longer mesocotyls increased (Fig. 3). This mesocotyl, elongation was promoted by GA3 treatment, i.e. about one thirds of mesocotyls elongated to 15-36 mm in 2.1 year old seeds, while no mesocotyl reached over 15 mm in both current and 1.1 year old seeds. In this case too, the growth pattern of seedings with long mesocotyls was vas very similar to the plants treated with ABA (Fig. 4). From above results, it may be suggested that the variation in mesocotyl growth reflects the difference of some internal factors which are affected by the location of seed in a panicle or the time after the seed harvest.