CONDITIONS OF EXPERIMENT 



In the first experiment, 500 grams of fingerling blueback were placed in 

 each of the 36 troughs necessars'- for the l8 diets » By an actual count of U 

 troughs, the average niomber of fish per trough was determined to be 1,308.5 

 with a coefficient of variation of 2.5 per cento This number of fish represented 

 an initial average weight per fish of o3821 grams or an initial number per pound 

 of 1,188. 



In the second experiment 1000 grams of fingerling blueback were placed in 

 each of the 52 troughs.. The average number of fish per trough as determined 

 by an actual count of 6 troughs was l4ii8o5 fish, with a coefficient of variation 

 of 3o6 percent. The number of fish per pound was calculated at 203«6 and 

 the average weight per fish at l.llItS grams. 



In both experiments the diets were treated uniformly when prepared and 

 fed, and the bind resulting from the addition of 2 percent salt, although 

 not entirely uniform from diet to diet, was satisfactory for purposes of 

 experimentation (^urrows et al, 1951) » The fish were fed three times a day 

 with a modified potato ricer. 



The description of the food products common to most of the diets in both 

 experiments was as follows s all raw products were ground, and mixed while in 

 a frozen condition; the beef liver and hog liver were dyed and flukey; the hog 

 spleen was fit for human consumption and therefore not dyed; the salmon viscera 

 was obtained from Puget Sound pink salmon (0. gorbuscha ) and contained only the 

 gonads, livers^ air bladders, and digestive tracts; the vacuum-dried salmon 

 viscera meal was prepared from the same salmon viscera used in the experiments. 

 A small Stokes rotary vacuum-dryer operating at 100° F. was used to dry-render 

 the meal. 



Blood counts and hemoglobin determinations were made at the end of both 

 the first and second experiments » Two blood counts were taken from a single 

 sample of blood f rom e ach of four fish chosen at random from each lot. The 

 hemoglobin determinations were made with the Tallqvist-Mams readings » As 

 the hemoglobin determinations and erythrocyte counts were derived from different 

 samples, an anemic tendency was not indicated unless the means of both determin- 

 ations were in agreement (with the hemoglobin determination below 9«Ii grams 

 per 100 milliliters and the erythrocyte count below 900^000 red cells per cubic 

 millimeter) » 



The water temperatures for the first experimentaveraged k^" F. ^^^.th little 

 variation during the cold-water period of 12 weeks from April 5 to June 28^1950. 

 At th^ end of this period the water temperatiore rose rapidly. The water tem- 

 peratures averaged 53o9° ^» for the warm-water period of 12 weeks from June 

 29 to September 20,1950. In the second experiment (July 5 to September 27) the 

 water temperature averaged 51i° F. 



