48 American Fisheries Society 



better method was keenly appreciated, and attempts in this 

 direction were started. The main difficulty to be remembered, 

 was in the fact that the relatively small quantity of fat had to be 

 recovered quantitatively from a large bulk of water. 



All the determinations were in duplicate. The nitrogen was 

 determined by the Kjeldahl method, and one-tenth portions of 

 the entire water were used for the analysis. The sugar was deter- 

 mined by Bertrand's method in one-quarter portions. The sugar 

 analysis was made on the perfectly clear nitrate, which was 

 obtained by treating the water with aluminum cream to free it of 

 the protein. On extracting this clear filtrate with ether, I inva- 

 riably found that there was no fat present. This suggested a 

 method of determining the fat, which is evidently carried down 

 with the aluminum precipitate. By drying this voluminous 

 gelatinous precipitate, powdering it and extracting it with ether, 

 it was hoped that accurate results on the fat suspended in the 

 water would be ootained. While this method, or some similar 

 modification of it, has good possibilities, a number of difficulties 

 were encountered in its practical application, which were not 

 entirely removed before this work had come to an unexpected 

 halt, and the matter of utilization of the fat from liver remained 

 untouched in this investigation. 



The feeding of liver has not met with success in my hands. A 

 number of circumstances have probably conspired to make these 

 experiments less definite than those of the preceding series. 

 Many experiments terminated in failure, owing to the death or ill 

 condition of the trout. Three experiments are recorded in 

 Table VII, and these will perhaps throw light on the question of 

 liver as a trout food. It will be seen that the trout did not grow as 

 well as those fed on beef heart, indeed one specimen lost 8.5 grams. 

 The utilization of protein (nitrogen) was very low (90.5 to 83.6 

 per cent). The utilization of glucose shows likewise wide fluctua- 

 tion from 96.9 to 86.9 per cent. The utilization of the fat ranged 

 about 90 per cent, but as the analytical data were not entirely 

 reliable, these results have not been included in the table. 



Before concluding, one other point should be mentioned. In 

 the course of the half hour during which the liver pulp remained 

 in the aquaria, as much as 60 to 75 per cent of its nitrogenous 



