Abstract
No systematic breeding work has been done so far with salmonids except for some attempts with rainbow trout (Salmo gairdneri) and Pacific salmon.The possibilities for obtaining genetic change are particularly dependent on selection intensity, heritability, phenotypic variance, and length of generation interval. The very high fertility in salmonids makes it possible to practice an intensive selection for fish in captivity and a more moderate selection in wildlife cultivation. Very few estimates of heritability have been calculated for production traits in salmonids and most of them are based on limited data. The few estimates for growth rate are moderately large (h2 = 0.1–0.2). For most production traits studied, a remarkably high phenotypic variance has been found, and the coefficient of variation is 2–4 times as high as for production traits in farm animals. The length of generation interval varies from one species to another and with the production system applied; 2–5 yr will be the usual.To make a selection program possible, one must control the whole life cycle of the fish. Therefore, hatching, fingerling production, and cheap and efficient tagging methods are necessary requirements.Because of the high fertility, the large variation and a moderately long generation interval, it should be possible to obtain much higher genetic gain per time unit in salmonids than in farm animals.