Carnivorous fish are two to three times as efficient as pigs and broilers in converting energy and protein to edible food for humans. As the domestication of fish continues, they will become more and more efficient and competitive with these industries. In the future, this will be very important, as more efficient utilization of available food resources is required to supply the growing human population with enough food. Today, about 1% of aquaculture production is based on genetically improved fish and shellfish. For salmonid fishes, we have the necessary knowledge to establish efficient breeding programmes. Large genetic variation has been associated with important economic traits. For growth rate, heritability ranges from 0.2 to 0.3, with a coefficient of variation of 20–30%. This implies that it is possible to obtain large responses from selection for growth rate. In several large-scale experiments and in breeding programmes, 10–15% genetic change has been obtained per generation, which is much higher than that reported for other farm animals. In Norway, extensive breeding experiments with Atlantic salmon and rainbow trout were started in 1971. For the first two generations of selection, the breeding goal was growth rate. Age at sexual maturation (measured as frequency of grilse) was then included. From the fifth generation, disease resistance (measured by challenge test for furunculosis and the virus ISA) and meat quality (measured as fat percentage, fat distribution and flesh colour) were included. Today, Norsk Lakseavl AS (Norwegian Salmon Breeding Company Ltd) or NLA runs the National Breeding Programme and has two breeding stations where 400 full-sib and half-sib families of Atlantic salmon are tested in each of four year classes. For rainbow trout, the number of families totals 120 in each of three year classes. In 1997, the Norwegian production was 310 000 tons of Atlantic salmon and 33 000 tons of rainbow trout. At present, about 65% of the salmon and trout produced in Norway use genetically improved fish from NLA and multipliers. The cost–benefit ratio for the National Breeding Programme in Norway is estimated to be 1:15.