538 U, S. BUREAU OF FISHERIES 
first occurred only in lower Labrador, while the second was found 
along the entire coast. Both are sea run. 
INVESTIGATIONS PERTAINING TO FISH-CULTURAL OPERATIONS 
TROUT CULTURE 
Feeding experiments with various diets and with several species 
of trout were continued during the summer by Dr. H. S. Davis and 
M. C. James at the Holden (Vt.) experimental hatchery. A more 
detailed account of the work at the station may be found on p. 675. 
A comparison of three common foods—beef heart, beef liver, and 
sheep liver—indicated clearly the superiority of beef liver when 
trout are to be reared to a larger size than the small fingerlings 
commonly used for stocking waters. Apparently any one of the 
three foods is satisfactory with steelhead, rainbow, and brook trout 
when the fish are to be planted early in the season. However, when 
brook trout were held to the age of 6 months, beef liver was found 
to be markedly superior as a food and reduced the mortality approxi- 
mately one-half and produced twice the growth of its nearest com- 
petitor—sheep liver. 
A comparison of cooked sheep liver with the raw liver as ordi- 
narily fed yielded conflicting results, probably due to a number of 
complicating factors. With brook trout the superiority of raw liver 
became more marked as the experiment progressed, and at the end 
of 103 days the total mortality among the fish fed cooked liver was 
four times as great as among those on a raw diet. 
With young rainbow trout cooked liver showed a superiority over 
the raw product over a period of 94 days, both as regards mortality 
and growth. Inasmuch as many commercial growers have reported 
favorable results from the use of cooked food, it is planned to con- 
tinue experiments along these lines. 
Nothing has been found that offers any prospect of entirely re- 
placing the fresh-meat diets. However, three substances were tried, 
which yielded very encouraging results as substitutes for part of 
the meat, and in mixture may even be superior to the pure-meat diets. 
These are soy-bean oil meal, a dried shrimp product, and fresh-water 
mussel meal. 
Soy-bean oil meal is manufactured from the residue after the oil 
has been expressed from the bean. An important property is the 
high percentage of protein, which more closely resembles animal 
protein in structure than do proteins from vegetable sources. Ex- 
periments with this meal were unsuccessful at first, but after adjust- 
ing the quantity to a basis of 50 per cent meal and 50 per cent meat, 
both brook and lake trout fingerlings were maintained for several 
months with nominal loss, although the growth did not quite equal 
that of the controls. Further experiments will be undertaken on 
a larger scale, but it is believed that this material will be most useful 
as a food for yearlings and adult fish. 
The waste from shrimp factories, under the name of “shrimp 
bran” and “shrimp meal,” frequently have been used in trout diets 
with more or less beneficial results. These products usually contain 
a large percentage of chitinous shell, which is of little value except 
