Abstract.— Indoor and outdoor experiments were conducted to improve the rearing techniques for artificially hatched northern bluefin tuna Thunnus thynnus during growout culture. Collisions with the walls of tanks or nets caused mass mortality that occurs during growout. The period when collisions frequently occur and the types of injury caused by collision were examined in this study. Juveniles were reared in indoor tanks from 30 to 120 d after hatching, and in an open sea net cage from 42 to 150 d after hatching. Dead fish were collected and counted daily in both of the experiments. In the indoor experiment, the sampled fish were preserved in 10% formalin solution, and each of 10 specimens of about 30, 50, 70, 85, 100, 130, 160 and 225 mm in body length (BL) were examined using x‐rays to detect injury of the bones. Juvenile and young adult bluefin tuna showed a reduction in numbers caused by collision with the tank or the net wall during the experiments. In the indoor tank, there were 1,200 fish on day 30 but only eight on day 120. The daily mortality increased from day 30 after hatching, when juveniles reached 50‐mm BL and remained over 4%/d until day 60 when juveniles grew to 300‐mm BL. The proportion of dead fish with injuries of bone, especially of the vertebral column and the parasphenoid, increased after fish reached 50‐mm BL, and exceeded 60% in fish with BL 85 mm or greater. In the open sea net cage, there were 3,841 fish at the start of the experiment on day 42 and only 65 on day 150. In this experiment, the reduction was greatest from the start of the experiment until day 80, when fish grew to approximately 25 cm in total length. Significant bacterial, viral or parasitic diseases were not observed in these fish; the only findings were dislocations of the vertebral column and injuries to the upper and lower jaws. These results show that the loss of juvenile and young adult bluefin tuna was caused by collision with the tank or net wall that fatally damaged the bones of the vertebral columns and the parasphenoid.

No abstract

Aggressive behaviour, assessed by 24-h observations, and survival rates, determined in a 2-week rearing experiment, were examined in greater amberjack, Seriola dumerili, subjected to single and multiple factors (high density, restricted feeding and diverse size). Aggressive behaviour frequency in the group subjected to both restricted feeding and size diversity (FS) was signi¢cantly higher than that in the other groups (Tukey, n 5 4, Po0.05). In the FS group, aggressive behaviour increased with the starvation time from 19 days post hatching (dph) to 23 dph (twoway analysis of variance, Tukey, n 5 4, Po0.05), but decreased gradually thereafter to 46 dph. In the rearing experiment, the survival rates for the diverse size and FS groups were lower than those of the control or restricted feeding groups (Kaplan^Meier, log-rank test, n 5100, Po0.01). Daily mortality in all groups was the highest at 24 dph; although dead ¢sh were usually small, disappearance due to cannibalism was not observed during the rearing period. The results indicate that aggressive behaviour in greater amberjack is induced by hunger and size diversity in the early juvenile stages (23 dph). Mass mortality during seedling production was mainly caused by injury and the death of small ¢sh due to aggressive behaviour/cannibalism.

Sex reversal from female to male in the six-year-old longtooth grouper Epinephelus bruneus (mean body weight, 3.1 kg) was induced by the implantation of 17-alpha-methyltestosterone (MT) cholesterol pellets. We compared the e‹cacy of the cholesterol pellets at MT doses of 1 mg/kg BW (group A) and 2 mg/kg BW (group B) implanted in February, as well as 2 mg/kg BW implanted in April (group C). At 4 months after the treatment, 67