A simplex design was used to test the effects of feeding Sea Grain(R) (fish meal process) processed from herring by-products (Clupea harengus), capelin (Mallotus villosus), small sandeel (Ammodytes sp.), and blue whiting (Micromesistius poutassou), to Atlantic salmon smolt (1+) for two 6-week periods. Seven diets similar in lipid (34%), protein (43%) and gross energy (24 MJ/kg - 1 ) were prepared from the Sea Grain(R) products in different combinations. Further, there was a shift in diet between the two experimental periods (applied to half the experimental groups) to examine the effects of changing feeds on appetite.During the experiment, average salmon weight increased from initially 119 to 407-447 g. Fish fed a mixture of 27% herring, 27% small sandeel and 45% blue whiting showed a higher specific growth rate (SGR) compared to fish fed a mixture of 55% small sandeel and 45% blue whiting only. Increased protein retention occurred concomitant with improved feed utilisation (feed conversion ratio, FCR) throughout the experiment. Some marine feed ingredients and mixtures thereof resulted in better growth and better feed utilisation than others, especially diet consisting of 27% herring, 27% small sandeel and with mixtures of more than 27% capelin and herring included the diet. Growth differences were apparently due to minor variations in protein or lipid qualities. Additionally, the dietary groups showing the best protein production value (PPV) and FCR also showed the highest concentrations of plasma triacylglycerol and protein. In these groups, aspartate amino transferase and alanine amino transferase showed increased levels, but were within ranges described not to influence health in any negative way. Good fish health was indicated in that all groups showed low mortality and haematological values within normal ranges. A special effect of feeding 55% capelin was found; a high condition factor simultaneously with a low hepatosomatic index. SGR was not significantly influenced by the shifts in diet. However, the shifts resulted in slightly reduced SGR in the groups fed a new feed just after introduction of the new ''taste''. This indicates that Sea Grain(R) can be processed from various species for direct use in salmon feeds without prior extensive mixing of different batches as common in the fish meal industry.