138 U. S. BUREAU OF FISHERIES 



During the summer of 1925 a lot of lake-trout fingerlings was 

 kept on a diet of sheep liver and clabbered sour milk for a period of 

 four months. In this case the milk was not mixed directly with the 

 liver, but the fish were given alternate feedings of liver and clabbered 

 milk. The growth was practically identical with that of the controls 

 fed straight sheep liver, but near the close of the experiment the 

 mortality of the fish on liver and milk showed a considerable increase 

 over that of the control. 



For some unknown reason dry buttermilk has not made as favorable 

 a showing with yearlings as with fingerlings. (Fig. 11'.) In com- 

 bination with pig liver dry buttermilk produced only a moderate 

 growth, although when shrimp meal was added to the ration the 

 growth was considerably better. With beef melts the buttermilk 

 made an even poorer showing than with pig liver. We are planning 

 more extensive exijeriments with dry buttermilk in the rations of 

 yearlings in the hope that some explanation will be found of the 

 apparent discrepancy in the results with fingerlings and yearlings. 

 No attempt has been made to include dry skim milk or " consolidated 

 buttermilk " in the diets of yearlings. 



CEBEALS AND SIMILAR PRODUCTS 

 WHEAT MIDDLINGS 



Probably no cereal products have been used more consistently than 

 Avheat middlings or low-grade flours. Although our results from 

 feeding middlings have been somewhat conflicting, we are of the 

 opinion that this product has comparatively little nutritive value in 

 trout rations. It is exceedingly doubtful if trout can make use of 

 raw starches, and it is probable that any apparent advantage derived 

 from the presence of middlings or flour in the ration is due to the 

 fact that they absorb the meat juices which otherwise would be 

 largely lost. Thoroughly cooked middlings are no doubt superior 

 to the raw product but, nevertheless, have made a poorer showing in 

 our experiments than the fish meals or dried milks. (Fig, 11.) 



SOYBEAN MEAL 



As the protein constituent of this meal more closely resembles 

 animal protein than is the case with cereals it was one of the first 

 vegetable products selected in our investigation of possible meat 

 substitutes. Soybean meal was incorporated into meat rations of 

 brook and rainbow fingerlings (figs. 7 and 8) at a 25 per cent level. 

 In both cases lower growth resulted, being more marked in the case 

 of the rainbows. The growth made by the brook trout when soy- 

 bean meal was included in their ration was not greatly inferior to 

 that produced by straight meat. In general, however, we found 

 that the addition of this meal to the meat rations of brook trout, 

 fingerlings, and yearlings alike, resulted in definitely retarding the 

 rate of growth. This was directly proporticmal to the amount of 

 soybean meal used. For some reason rainbow trout on a ration 

 containing 25 per cent soybean meal made u much poorer showing 

 than brook trout, 



