CONDITIONS OF EXPERIMENT 



Before comparing the results brought about by feeding different 

 diets, a description of the experimental techniques and of the water 

 temperatures is given to qualify the conclusions based on the out- 

 come of the various trials. 



The techniques used in these trials are, in general^ much the same 

 as those used in previous -L-eavenworth feeding trials. A more complete 

 account of the experimental design, stocking, food preparation and 

 presentation, and of the. method of statistical analysis will be fouiid 

 in the paper by Burrows et al. (19i>l) . 



The experimental design, statistically speaking, consisted of 

 one replication and 20 treatments. The one replication represented 

 two. adjacent troughs, and the 20 treatments each represented a different 

 diet or lot. The kO troughs necessary for such a design were stocked 

 with equal weights and comparable nianbers of fish chosen at random from 

 a homogeneous population. Each trough received exactly 500 grams of 

 fish. By a hand count of 10 per cent of the troughs, the number of fish 

 per trough was found to be about 2,277 with a coefficient of variation 

 of l.oii per cent. This number of fish weighing 500 grams represents an 

 initial average weight per fish of .2196 grains or an initial number per 

 pound of 2,067. After the troughs were stocked, no changes in the num- 

 ber of fish occurred except those unavoidable losses brought about by 

 mortalities. 



In the preparation and feeding of the diets for the 20 lots, the 

 techniques with the exception of binding were comparable for each diet. 

 The unbound diets were the beef liver control, and the tuna liver diet. 

 The diets in the salmon segment group were not bound but received sal- 

 mon waste meal as an absorbing agent to prevent leaching. The rest of 

 the diets contained either hog liver or hog spleen, either of which in 

 the presence of salt produced a very satisfactory bind„ Salt was add- 

 ed at the rate of two graias per 100 grams of ration in the diets con- 

 taining hog liver or hog spleen. The fish were fed three times a day 

 with a modified potato ricer. 



Water temperatures had an important bearing on the outcome of 

 these feeding trials. For instance, several diets in this experiment, 

 although suitable for use when the water temperatures were above 50'-' F,, 

 were definitely harmful to the fish when fed at temperatures of h5° F, 

 Because of the importance of water temperatures, a brief description 

 will be given of the temperature patterns and their effect on this 

 experiment. 



The water temperatures for the first 12-week period averaged ii^^F, 

 with less than a degree variation. At the end of this first period the 



