Modification of the plasma cortisol response to stress in rainbow trout by selective breeding.

Male and female rainbow trout were segregated into high- and low-responding individuals (HR, LR) on the basis of their plasma cortisol response to a 3-h period of confinement imposed at monthly intervals for 5 months. Consistent divergence was obtained in the responsiveness of the two groups, although the difference between LR and HR groups was greater in female fish (56 c.f. 116 ng ml(-1)) than in males (45 c.f. 69 ng ml(-1)). Progeny groups (full-sib families) were obtained from the pairing of HR males and females and LR males and females. A third progeny group (US) was obtained by random pairing of parents which were not selected as HR or LR. Poststress plasma cortisol levels in the progeny were first tested at 6 months after hatch and were significantly correlated with the response of the corresponding parental groups, HR > US > LR (178, 126, 81 ng ml(-1), respectively). The difference in responsiveness between LR and HR groups was demonstrated in all four subsequent tests over a 12-month period. There were no significant differences in baseline plasma cortisol levels in LR and HR groups prior to confinement. During a 4-h period of confinement, the differences in plasma cortisol levels between LR and HR fish were sustained throughout, indicating that the trait upon which the fish were selected was related to absolute maximum levels of circulating cortisol and not to the rate of change of cortisol levels during exposure to a stressor. A moderately high heritability (h(2)) for confinement-induced plasma cortisol of 0.41 was obtained by a parent-progeny regression. Manipulation of stress responsiveness in fish by selective breeding offers scope for optimizing performance under intensive rearing conditions but also provides a useful research tool for investigating the operation of the endocrine stress response.