THE EARLY FEEDING OF SALMONOID FRY. 
BY CHARLES E. ATKINS, EAST ORLAND, ME. 
‘The initial feeding of salmonoid fry has always been regarded 
by fish culturists as of critical importance and it has come to be 
generally considered of urgent necessity that the first manifes- 
tation of desire for food should be met promptly by its gratifi- 
eation. Either artificial food must be administered at once or 
the fry must be Liberated in water affording an immediate and 
constant supply of natural food. Some authorities have even 
urged feedings in advance of the absorption of the yolk sack. 
The consequences of even a brief delay in this matter have been 
supposed to be very serious, extending to the death of all fry 
subjected to a few days of hunger. 
At the Craig Brook station it has been one of the rules most 
rigidly enforced, to watch the fry approaching the completion 
of the sack-period very closely, anticipating their appetites by 
tempting bits thrown in tentatively, and to lose not a day in 
satisfying the first demand for food. As early however as 1897 
a single experiment in fasting had indicated that the question 
of the soundness of the theories accepted might well be taken 
up, and with the hope of accumulating data from which safe 
rules of procedure could be formulated, several experiments in 
the enforced fasting of fry were undertaken in 1904, and a 
more extended series in 1905. It is the purpose of this paper 
to present the most important results obtained from the experi- 
ments of 1905. 
The data which I will consider concern the treatment of 4 
lots of brook trout fry, 4 of lake trout, 8 of Atlantic salmon and 
4 of silver salmon, 20 lots in all, that were subjected to enforced 
abstinence from food; and of 3 control lots which were closely 
related to the fasting subjects, but were amply fed. All of these 
fry were hatched at Craig Brook. The feeding and fasting were 
conducted in troughs a little more than a foot wide, in which a 
depth of water of about + or 5 inches was maintained. The wa- 
ter was mainly derived from Craig pond, a lakelet of great 
n 
d 
