676 U. S. BUREAU OF FISHERIES 
and the supplying of these deficiencies by the construction of 9 pools and 285 
feet of raceways during the last two years has accounted for much of the 
time spent there and for the somewhat limited output of a strictly scientific 
nature. 
A few words as to the conduct of such an enterprise may be permissible. It 
is probable that the keeping of exact, easily analyzed data and records on every 
condition occurring in the ordinary State or Government hatchery would 
bring to light information that would surprise the keepers. Such information 
as is generally recorded is fragmentary, unsystematic, and frequently an ap- 
proximate expression of personal judgment. The fish records cover the lots 
as groups, and analysis of diverse conditions is impossible. At the Holden 
station it is a cardinal principle that every factor having any bearing on the 
fish must be recorded. A ecard filing system has been begun to give ready in- 
formation on the history and status of virtually every fish on hand. These are 
cross indexed so that the fish in any trough or pond can be traced back to their 
origin with all deviations from normalcy exposed; or any particular lot of fry 
or eggs can be followed through to the final disposal of the last survivor. 
Routine records are kept of losses, temperature, where changes are occurring, 
growth (if desired), diet, etc. Any unusual manifestation in any lot is noted 
as to time and place; this has proved to be of value in checking up in the case 
of a subsequent outbreak of disease. In fact, such a plan will give, in time, a 
mass of information, from the ordinary routine of the hatchery, that will be as 
valuable as any to be gained from formal experiments. It may be stated that 
with 6 species on hand from a dozen or more different sources, distributed 
among 10 ponds and 15 raceway compartments, as well as 40 troughs, a sea- 
son’s haul of notes can well be described as voluminous. 
The feeding experiments cited at the beg nning are part cf a series initi- 
ated four years ago and carried on at other hatcheries. Based, at the start, 
on a belief that some of the many difficulties besetting the rearing of trout 
might reasonably be expected to have their origin, to some degree, in the diets 
of artificial culture, the assumption of vitamin deficiency was entertained. 
First attempts to improve such a condition (if it existed) by the enrichment 
of various foods with substances of high vitamin value, like cod-liver oil and 
yeast, were not highly successful. The most encouraging results were obtained 
in the augmentation of the diet of rainbow trout with vitamin A from cod- 
liver oil and vitamin B from yeast and natural sources. A short time later, 
investigations by other workers brought out the fact that liver, heart, etce., 
were much richer in vitamins than had been suspected. Less emphasis has 
been placed on this branch of the work of late; it has been developed that 
under some circumstances a vitamin enrichment of the diet of rainbow trout, 
by means of cod-liver oil, has been of some benefit, but our experiments have 
consistently shown the reverse to be true in the case of brook trout. 
The more recent work (that of the past season) was deyoted to determining 
the effects of various diets on a serious disease that prevails at Holden; to 
determining the relative values of the common diets of present practice (beef 
heart and liver and sheep liver) ; and to the development of three new prod- 
ucts to supplement the meat diets. Beef liver was found to be apparently 
better, as regards lower mortality and increased growth, when compared with 
the other meats mentioned. The superiority is more marked, of course, where 
fish are to be reared to large size rather than planted as early fingerlings. 
The new products utilized were soy-bean oil meal, a dried shrimp, and a clam 
meal. It is widespread practice among fish-cultur sts to feed their larger fish 
a mixture of fresh meat with a large proportion of a cheaper meal cereal, such 
as middlings or a low-grade flour. There is a serious doubt that the carbohy- 
drate of these materials enters into the nutrition of trout very extensively, and 
the purpose of the products mentioned above was to obtain the roughage values 
obtained from cereals, and at the same time feed a high protein ration to fur- 
ther growth and condition, and at the same time supplant part of the expensive 
meat. It can be said that experiments on a small scale have shown no objec- 
tionable features in these substances; with some promise of real benefit in the 
case of the clam meal, and their use will be continued in a more practical way. 
The procedure and technique of these feeding experiments do not appear 
to have the complexity that would warrant the application of the term “ scien- 
tific”’ The simple process of feeding one lot of fish a certain diet and an- 
other lot a different diet, and observing their reactions, growth, and mortality. 
would not seem to require a highly specialized training or a broad erudition. 
This is true; but at the same time it must be recalled that our present knowl- 

