The proportion of amino acids which get deaminated varies with the food and the fish's circumstances. Fish which are not growing because of low temperature or have their ration at maintenance level or below, would deaminate most or all of their amino acids. Fish kept at high rearing temperatures or at high activity levels and therefore having very high metabolic rates would do likewise. On the other hand, fish having rapid growth and high protein intake would deaminate a relatively small proportion of their digested protein, although the absolute quantity of amino acids deaminated could still be large enough to produce a relatively large SDA. The energy for deamination need not necessarily come from amino acids, but will be preferentially taken from carbohydrate or lipid, if available. Thus, salmonid aquaculturists long ago discovered this "protein-sparing" action of limited amounts of inexpensive carbohydrate in the diet as a way of reducing the cost of feed and still achieving a desired level of growth. The protein-sparing action of lipids appears to have been minimally investigated. One can thus minimize SDA costs, but not avoid them completely.