194 THE SALMON AND ITS MIGRATIONS 



" spawning season, gone down, recovered, and 

 " returned, for numerous experiments show 

 " that the shortest period of return is about 

 " three months, and it is only about three months 

 " since the earliest fish had begun to spawn in 

 " the river which these are now ascending. They 

 " must have passed the autumn or earlier winter 

 " in the sea. Then they must have passed the 

 * ' winter without breeding, and there we have the 

 " discouraging fact or hypothesis that the salmon 

 ' is a fish which does not breed every year, a 

 " hypothesis which will have the less chance of 

 " acceptance just at present, when it appears, 

 * ' or is supposed to have been discovered, that the 

 " herring a fish resembling the salmon at least 

 " in the important respect of being migratory 

 " breeds twice in each year, or, at all events, 

 " breeds at two widely different seasons of the 

 5* year." 



Mr. Eussell, in the above, made the mistake of 

 supposing that these fish were ascending the riv- 

 ers at that early date, while as a matter of fact 

 they were descending after having passed the 

 winter in the river and undergone the change I 

 have mentioned. It is also quite possible that 

 the change from the parr stage into that of the 

 smolt, which has been such a puzzle to natural- 

 ists, may be effected in the same way, this change 

 coming on only when the parr reaches the salt or 

 brackish water in the estuaries. 



