188 THE SALMON AND ITS MIGRATIONS 



hind leg bone and a small portion of the verte- 

 brae was present; all the other bones missing. In 

 a kelt, caught in a trout net, about the beginning 

 of May, was a piece of fat, with a shred of skin. 

 This appeared to come from some species of duck, 

 but as there were no feathers left on it I could 

 not determine what kind it had been. Four sal- 

 mon, out of fifteen or twenty thousand, had 

 enough food found in them to keep a small trout 

 alive one day. I think that we may well conclude 

 that they do not feed in fresh water. But the 

 question often posed by anglers naturally pre- 

 sent itself : " Why, if they do not feed, do they 

 rise to the fly ? A fly going through the water 

 looks very much like a live insect; why do they 

 chase it and snap at it sometimes so viciously, if 

 not for food ? ' ' This is certainly a good argu- 

 ment, but the fact remains that no flies are found 

 in them. There are myriads of flies and grubs of 

 all descriptions in a river, and surely, if the sal- 

 mon were taking them as food, the remains would 

 be found in the stomach. If we open a trout or 

 small parr, we will generally find it stuffed 

 with insects of various kinds. I have often sat 

 and watched these small parrs for hours at a time 

 in a sheltered corner or eddy of a river, throwing 

 into the water all sorts of things, bits of bread, 

 meat, paper, and all kinds of crawling or flying 

 insects that I could get hold of. The fish would 

 rise at everything that dropped, but would not 



