except the identical pair of diets, '-'icts 12 and 13. This ration en the basis 

 of higher than average mortalities and the relatively small stimulus resulting 

 from the addition of salmon eggs in this combination may be considered an un~ 

 satisfactory cold water diet. The oth^r diets were all adequate during the 

 cold-water period on the basis of gro;rt.h except Diet iS, where the aided e>:^ense 

 of the yeast was nob ^justified by the resultant growth rate. 



Of the satisfactory diets, the most rapid growth during the cold- 

 water period was made by Diets 6 and 7, each consisting of 22.2 percent beef 

 liver, 22.2 percent hog liver, 22.2 percent hog spleen, and 33. U percent salmon 

 eggs. 



The next most rapid growth rate W;as achieved, not by one diet, but 

 by several. The diets in this group were: the ineat- viscera mixture of 22.2 

 percent each of beef liver, hog liver, and hog spleen, and 33^1; percent of 

 salmon viscera (Jiets 2 and 3)3 the meat-viscera piixture supplemented with 

 crab meal (Diets h and 5)^ the diet of equal parts beef liver, hog li-ver, and 

 salmon viscera (Diets 8 and 9)j and a similar diet supplemented \-n.th 10 per- 

 cent salmon milt^ (Diets 10 and 11), Of soecial interest is the meat-viscera 

 mixture supplemented with 5 percent crab meal since this diet had a growth rate 

 comparable to its control, ^iets 2 and 3, and would b./ reason of the cost of 

 crab meal be a less expensive diet. 



The lowest group of growth rates in the experiment, and again the 

 rates were quite comparable, wbs oroduced by the high-level salmon '/iscera diets 

 supplemented variously with either beef liver, or hog liver, or a combination 

 of beef liver and hog liver. The diet supplemented iirith hog liver in this 

 group has an advantage over the diet supplemented with beef liver sincd hog 

 liver is usually cheaper and more readily obtainable than beef liver, ■'•'he 

 beneficial effect on grow'th,of increasing levels of beef liver and hog liver 

 and decrecAsing levels of salmon viscera, should be noted in Diets 17, I8, and 8. 



The diets suitable for warm water may be limited at once by discarding 

 all diets xd-thout vacuum-dried salmon viscera meal since tliis meal or a similar 

 sJ-.aon waste m.eal reduces the cost and increases the growth rate so markedly. 

 Unfortunately f^iis slim' nates a factual discussion of the high-level salmon 

 viscera diets because space was not available to test the effect of meal on 

 them. A supposition as to the effect of salmon waste meals on these diets was 

 advanced, however, in the high-level salmon viscera section. The supposition, 

 based on existing evidence, was that the addition of a salmon waste meal prob- 

 ably would have increased the groirth rates but not enough to compare favor- 

 ably T-dth the diets with lower levels of salmon mixture supplemented with crab 

 meal. Diet 5, since the control with no crab meal. Diet 3, made much better 

 gains. 



This elimination leaves the following diets, Diets 3, 7, 9, II3 and 

 13, all of which had satisfactory growth rates and low mortalities during the 

 warm-water period. Of these diets. Diets 7 and 13 produced the highest rate 

 of growth. Diet 7 consisted of 20 percent each of beef liver, bog liver, and 



10 



