This work is a part of a program that aims to analyze nutritional and genetic determinism of fat storage allocation in rainbow trout body compartments. Two lines of rainbow trout were selected for high or low muscle fat content using a non-destructive estimate of this feature in live fish (Distell Fish Fatmeter). Triplicate fish groups of the lean muscle line (L) and fat muscle line (F) obtained after 3 generations of selection were fed diets containing 100 or 230 g lipid/kg DM, for six months, right from first feeding. They were then all fed a commercial diet with 280 g lipid/kg dry matter for the following 7 months. Some males from the line F were maturing; they were discarded from the study. At the end of the growth study, relative weights of carcass, digestive tract including adipose tissue and fillet were recorded. Colour of fish fillet was measured immediately after sampling with a chromameter (CR200, Minolta) in 3 points per fillet (anterior, middle, and tail part). Muscle samples from six fish per treatment were frozen until analysis for lipoprotein lipase activity. Fat content was measured by RMN in dressed fillets. Thickness of subcutaneous adipose tissue was estimated using an image analyzer on slices sampled 1 cm before the dorsal fin. Fillets from other fish were cooked using a microwave oven and subjected to sensory analysis by a trained panel. Viscerosomatic index was higher in line L than in line F and was increased in trout of both lines previously fed the diet with 23% fat. However, carcass yield was decreased only in line F previously fed the fat-rich diet. As expected, muscle fat content was significantly different between lines but it was not affected by early nutritional history. Muscle lipoprotein lipase activity was diminished in line L in agreement with the lower muscle lipid content of this group. Thickness of subcutaneous adipose tissue tended to be greater in line F (p=0.07) but was not affected by early feeding (p=0.9). Dressed fillet yield was higher in line F compared to line L and lower in fish previously fed the diet with 23% fat. Lightness values of fillet were decreased in these latter fish. Chroma values were higher in line F than in line fed the L. None of the sensory features was significantly affected by early feeding composition. On the contrary, selected lines significantly differed for several sensory traits: odour intensity, orange colour, fat texture, oily fish flavour, and quantity of fat droplets in juice were increased in line F compared to line L. In conclusion, energy content of diets fed during the first six months of life did not affect muscle fat content and sensory parameters of 600 g trout but induced long term changes in viscero-somatic index, dressed fillet yield and flesh colour lightness.