Underyearling masu salmon were reared under a 8L:16D photoperiod for approximately 3 months, and subsequently transferred to 8L, 10L, 11L, 12L, 13L, 14L or 16L photoperiod in October, in order to determine the suitable photoperiod for induction of smolting (Experiment 1). Smolt was judged by the external appearance of fish, i.e., silvery body color, black margin on the dorsal and caudal fins, and slim body shape. As early as 1 month following this transfer, smolts appeared in the 16L group. After 2 months, the 13L group had the highest frequency of smolts. After 3 months, most smolts were found in the 11L group and presmolts appeared even under constant short daylength (8L:16D). Overall, the largest number of smolts appeared in the group transferred from 8L:16D to 13L:11D. Underyearling fish reared under natural photoperiod were also transferred to 11L, 12L, 13L, 14L, 15L or 16L photoperiod in October (Experiment 2). Only a few smolts appeared during this experiment. In order to determine a suitable short daylength for pretreatment prior to the transfer to long daylength, underyearling masu salmon were reared under 6L, 7L, 8L, 9L, 10L, 11L, 12L and 13L photoperiods for approximately 5 months, and subsequently transferred to a 13L:11D photoperiod in October (Experiment 3). One month after the transfer, a small number of smolts appeared in only the group transferred from 6L. After 2 months, a large number of smolts appeared in the groups transferred from 10L and downward. After 3 months, a small number of smolts newly appeared in the groups transferred from 12L and 11L. A few smolts appeared in the group under constant 13L after 6 months. Finally, effects of increasing daylength on the secretion of growth hormone (GH), thyroxine (T 4) and triidothyronine (T 3) were investigated (Experiment 4). Seven days after transfer from short (8L:16D) to long (13L:11D) daylength, plasma GH levels were significantly higher than those in the control group under constant 8L. Plasma GH levels tended to increase gradually in both groups throughout the experiment. Plasma T 4 and T 3 levels were not affected by increasing daylength. These results indicate that it is necessary to rear fish under short daylength for a sufficient period of time before daylength can be changed to long-day for purposes of artificial induction of smolting, and the larger the difference in daylength after changing photoperiod to long daylength, the earlier smolts appear. Furthermore, an increase in daylength may stimulate the transient secretion of GH, which may participate in the process of smolting.