78 Thirty-fourth Annual Meeting. 
losses except in case of the 15-day fasts, where it was almost 
exactly the same. The mean losses of all the lots of the first 
series were eight and seven-tenths per cent, and of the second 
series five and six-tenths per cent, a difference of three and one- 
tenth per cent in favor of the second series. As these fry, both 
series, were taken out of the same original lot, the difference 
suggests the query whether there was any difference in the treat- 
ment of the two series. There was a difference, which I will now 
state, without, however, claiming that it explains the difference 
in results. 
It had occurred to the experimenter that fry that were de- 
nied artificial food might still be able to pick up a trifle of food 
in the form of minute animal hfe brought in with the water- 
supply, to tide them over the waiting period. In order to elim- 
inate this source of uncertainty, one series of the Atlantic 
salmon fasters was kept in water filtered through gravel and 
sand which must have intercepted most, perhaps not all, natural 
food. Now did the fry so treated show any effect of being de- 
prived of an equal opportunity to snatch a possible bite of live 
food now and then? On the contrary, the fry so rigorously 
treated had lighter losses than those in unfiltered water. ‘The 
general summary for the series in unfiltered water show a loss 
of eight and seven-tenths per cent, for those in the filtered 
water five and six-tenths per cent, a difference of three and one- 
tenth per cent in favor of the filtered water. 
Now how do the losses of the fasters compare with those of 
the control lot from which they were taken and which had been 
fed 4 times a day? As in the case of the brook trout the five 
day fasts were accompanied in each series of fasters by a light- 
er loss than in the fish that were fed. The 10-day fast of the 
first series was accompanied by a little heavier, and that of the 
second series by a very much lighter loss than in the control. 
The facts already stated are certainly surprising, but the 
most astonishing part of this experience is still to be laid before 
you. .It pertains to the experiment with silver salmon. [Oncor- 
hynchus Kisutch! This Pacific species the station has handled 
this year for the first time. For the fasting experiment 4 lots of 
them were counted out, 500 fry in each. They were ready to feed 
May 18, and on that day the feeding of the control lot began, at 
