->? 
~2 
American Fisheries Society. 
: A 
the control lot seems to have reaped a slight advantage from the 
food eaten, having lost but four and nine-tenths per cent, while 
the fasters were losing six per cent. When we next compare the 
eaters with the 14-day fasters, we find the advantage very decid- 
edly in favor of the food; the eaters have lost but five and four- 
tenths per cent, while the fasters were losing 52 per cent; finally, 
the victims of the 19-day fast lost 78 per cent, while the eaters 
were losing 6 per cent. 
Whether trout fry of prime condition, from vigorous wild 
parents, would have suffered as severely as those treated in these 
experiments is a question we have no means of answering posi- 
tively; but the presumption favors a negative answer. 
The next series of experiments to be considered deals with 
lake trout fry. Of these there were four lots, of 100 fry each, 
and their fasts, as in the case of the trout, were, respectively, 
5, 9, 14 and 19 days, but in each case the fry had been fed 6 days 
before the fast began. Accounting in each case for the losses 
from the beginning of the fast down to 15 days after its close, 
it appeared that the mortality was a little heavier than with 
brook trout in the cases of the fry fasting 5 and 9 days, and a 
httle lighter in cases of the longer fasts. There was no control 
lot of lake trout. 
Of Atlantic salmon fry there were two series. The first series 
embraced 4 lots, of 1,000 each (marked 1847 A, B, C and D), 
and their fasts were, respectively, for 5, 10, 15 and 20 days. 
None of them had received any food before the experiments 
began. The total losses for periods corresponding with the com- 
putations for the brook and lake trout, that is, from the begin- 
ning of the fast down to 15 days after its close, were respectively, 
25, 43, 64 and 217 fry out of each thousand—the percentages 
being thus two and five-tenths, four and three-tenths, six and 
four-tenths and twenty-one and seven-tenths. 
The second series of Atlantic salmon fasters consisted of four 
lots of 500 each, from the same control lot as the first series, 
namely, No. 1847, and the members of this series were distin- 
guished by the letters EK, F, G and H. Their losses for similar 
periods as the other series were, respectively, in percentages, 
four-tenths, eight-tenths, six and six-tenths and fourteen and 
six-tenths. As compared with the first series, these were lighter 
