American Fisheries Society. Fees) 
the rate of 4 feeds per day of chopped hogs’ liver. The fasts 
were for 5 days, 10 days, 14 days and 19 days. The losses of 
the first three lots for the usual period, that is, from the begin- 
ning of the fast down to 15 days after its close, were exactly 
alike, 8 fish dying out of each lot, a percentage, for the entire 
period, of one and six-tenths. The lot that fasted 19 days lost— 
in 34 days just 6 fry, a percentage of one and two-tenths. Com- 
paring now the 4 lots with each other and stating the losses in 
the ratio of the daily losses per 10,000, the fasters for 5 days 
lost 8; the fasters for 9 days lost six and four-tenths; the fast- 
ers for 14 days lost five and one-half; the fasters for 19 days 
lost three and one-half; eight, six and four-tenths, five and one- 
half and three and one-half. That is, the longer the frv fasted 
the hghter the mortality. Comparing now the fasters with the 
control lot it is found that the latter, the feeders, lost at the rate, 
stated in daily loss per 10,000, for the different periods, 13; 
ten and four-tenths; nine; and seven and-seven-tenths. A mean 
of these losses would be 10 daily out of 10,000 or one-tenth of 
one per cent, while the mean loss of the fasters was five and 
eight-tenths daily out of 10,000, or one-seventeenth of one per 
eent. That is, taken all together, the feeders lost almost twice 
as heavily as the fasters. 
To sum up for all the species except lake trout, the 5-day 
fasts were in all cases accompanied by lighter mortality than 
that suffered by the feeding fish; the 9 and 10-day fasts by light- 
er mortality in some cases, by heavier in others; the 14 to 20-day 
fasts by heavier mortality except in the case of silver salmon. 
The subsequent behavior of the fry deserves a moment’s 
mention. The fry appeared, during the extended fast, to grow 
thinner in body, but when feeding began they were in every case 
ready, at once took to eating and in a few days showed that they 
were building up. 
What practical lessons are to be drawn? Far be it from me 
to insist that this series of experiments be taken as concluding 
the matter. It is only one series, and needs support from others. 
Yet the results agree in general with those obtained in 1904 and 
earlier, at the Craig Brook Station, and surely indicate that an 
early and abundant supply of food is by no means so essential 
to trout and salmon fry as we have supposed. They even go 
