1 8 The Atlantic Salmon 



some rivers is an unnecessary confirmation of 

 the fact of a divided migration of salmon at this 

 stage of their existence, and it is now well estab- 

 lished that this divided migration occurs in all 

 stages of growth from the smolt to the adult fish, 

 the great surplus of grilse over salmon in some of 

 our rivers is hard to explain. One river may 

 yield one grilse to ten salmon, and another, near 

 by, one salmon to four or five grilse. As the 

 fewest grilse are found in the rivers of this coun- 

 try, so far as I know, which produce the largest 

 fish, it may be possible, though I only hint at it 

 as a theory, that the large average growth at- 

 tained by the salmon of the Romaine, the Cas- 

 capedia, and some other rivers is due to their 

 continued residence for years in the sea, where 

 their increase in size is not checked by the exer- 

 cise of the reproductive function and the severe 

 strain it involves. This would appear to show 

 that a very much larger proportion of the fish of 

 some rivers pass the grilse stage of existence at 

 sea than is the habit with natives of other 

 rivers, or, that the grilse of some rivers ascend 

 other streams than those in which they were 

 born. 



It is probable that the smolts of this country 



