contribution of fish meal to the meat and viscera mixture., These 

 diets were exact duplicates of those fed during 19U7- In that year 

 (Table h) the Cortland diet and the meat-viscera-meal mixture pro- 

 duced comparable gains. Both indicated significantly higher growth 

 rates in the fish than was made by the meat and viscera mixture which, 

 in turn was superior to the beef liver diet. In 19k7 , the mortalities 

 incurred by these diets were not significantly different. 



In 19U8, the total weight of the fish fed each diet, as shown 

 in Figure 2, indicated that the addition of 10 per cent meal to the 

 meal and viscera mixture (Diet 2) made no measurable contribution to 

 the total weigbt of the fish during the first 12 weeks of the ex- 

 periment. The Cortland diet (Diet h) appeared to be definitely detri- 

 mental during the cold-water period. At the end of the first 12 weeks, 

 the meat-viscera-meal mixture and the meat and viscera mixture were 

 comparable in their total gains in weight. At the same time the 

 Cortland diet and beef liver control also showed no significant dif- 

 ferences from each other. The difference in weight that existed 

 between the two groups, however, was significant. These results 

 were in marked variance with the findings of the 19U7 feeding trials. 



At the conclusion of the 2li-week experiment, the addition of 

 meals to the diet had made a measurable contribution to the total 

 weight of the experimental fish. In the two groupings, the ad- 

 dition of 10 per cent meal resulted in a significant increase in the 

 total weight of Diet 2 over the fish in Diet 8, and the f?0 per cent 

 meal contained in the Cortland diet made a measurable contribution 

 to the total weight when compared to the fish fed the beef liver 

 control. Contrary to the 19U7 results, the Cortland diet was no 

 longer comparable to the meat-viscera-meal mixture but resulted in 

 a total weight of fish produced which was inferior to the meat and 

 viscera mixture. All four diets differed significantly from each 

 other. 



It will be noted that, with the advent of warmer water, the slope 

 of the weight curves altered sharply. It was during this period that 

 the acceleration in growth due to the addition of meal became ap- 

 parent. 



It has been the policy of this laboratory to make no compensation 

 for mortality when evaluating practical diets on the theory that 

 mortality is one measure of the efficiency of a diet. In this case, 

 however, where significant differences in mortality existed and 

 where there was a distinct possibility that the contribution of meals 

 to the growth rate was being obscured by the mortality, it appeared 

 desirable to plot the rate of growth for the diets in question to 



eliminate the effect of mortality. The rate of growth, as determined 

 by the average weight of the fish, for the four diets in question is 

 - tted in Figure 3. From this praoh it is , parent that the water 



28 



