fish food, but has some shortcomings as shown in the 19U7 ■L'eavenworth feeding 

 trials in which this product demonstrated a lack of anti-dnenic properties. 

 Koweverj because of the f;rowth rate demonstrated that .'•earj further trials 

 were conducted in I9I48 ^Ath hake as a substitute for salnon viscera' in the meat- 

 viscera diet. Hake appeared quite promising at the end of the cold-water period, 

 but at the end of the warm-water period the salmon viscera variant had exceeded 

 the hake variant in mean lot Neif;ht<, Further tests were conducted in 19U9 with 

 hake as a substitute for hog spleen in the meat-viscera diet. Again the hake 

 variant looked most promising at the end of the cold-water period but failed 

 to hold the advantage during the warm-water period. Repetition of the warm- 

 water trials of 19h9 in the present experiment substantiated the former conclu- 

 sions that hake, although a satisfactory substitute for hog spleen in the raeat- 

 viscera-meal diet, did not produce growth comparable to that of hog spleen in 

 warm water. 



In the present experiment, hake was compared ^^'ith halibut sawdust and 

 whole cod as possible substitutes for hog spleen. In addition, hake was tested 

 in various modifications of the m.eat-viscera mixture utilizing salmon milt, 

 salmon eggs, and vacuum-dried salm.on viscera meal. The hake used in all of these 

 trials was disintegrated whole and then frozen. 



The diets and their components used in the hake tests vjere as follows: 

 Diet 3U consisted of 22.2 percent each of beef liver, hog liver, hake, and 33.4 

 percent salmon viscera; Diet 35 consisted of 20 percent each of beef liver, 

 hog liver, and hake, 30 percent of salmon viscera and 10 percent of vacuum-dried 

 salmon viscera meal_; Diet 36 consisted of 20 percent each of beef liver, hog 

 liver, and hake, 30 percent salmon viscera and 10 percent salmon milt; Diet 37 

 consisted of I8 percent ppcV^ of beef liver, ^og liver, and hake, 2? percent of 

 salmon viscera; Diet 38 consisted of 20 percent each of beef liver, hog 

 liver, and hake, 30 percent of salmon eggs, and 10 liercent of salmon milt; and 

 last. Diet 39 consisted of I8 percent each of beef liver, hog liver, and hake, 

 27 percent of salmon eggs, 9 percent of salmon milt, and 10 percent of vacuum- 

 dried salmon viscera meal. 



The mean lot weights indicated that the hake variant was the equal 

 of the whole cod variant but significantly inferior in growth potential to 

 the halibut sawdust and hog spleen variants (Table 2, Diets 3$, U2, UU, and U6) . 



In the comparison of diets with and without vacuum-dried salmon viscera 

 meal, the diets with salmon viscera meal. Diets 35, 37, and 39, exceeded the 

 controls without mral., Diets 3h, 36, and 38, at the five percent confidence 

 level. Moreover^ the growth increments by percentages were U7.5 percent for 

 Diet 35. It will be noticed that these percentages are in agreement with similar 

 tests of the effect of meals on growth in the first experiment of the 1950 trials. 



The evaluations of salmon eggs and salmon viscera and the com.oarisons 

 of diets with and without salmon milt confirmed the results secnred \-dth tliese 

 products during the warm-water oeriod of the first experiment in that salmon 

 eggs produced more weight than salmon viscera, and salmon milt had no deleter- 

 ious effect on growth or mortality (Table 2, Diets 35, 36, 37j 38, and 39). 



16 



