Mr D. Ellis on the Natural History of the Scdiiwn. 271 



During the summer months, salmon from the sea proceed 

 sometimes high up those rivers, which are furnished with a due 

 supply of water, either permanently, or during occasional floods ; 

 and in subsequent floods they try again to return to the sea : at 

 this time of the year, however, their migrations into rivers are 

 often limited to the point to which the tide flows^ and they re- 

 turn again to the estuary and sea with the ebb-tide. 



In autumn, again, the male and female salmon ascend to the 

 shallow fords and sources of rivers to breed, remaining there 

 during the winter months, the male, howevel*, returning early 

 in the following year, and the female not till March or April. 



Beside the breeding-fish, which descend in the winter and 

 early spring months, clean salmon from the sea are constantly 

 ascending and descending those rivers, which, by the quantity^ 

 quality, and temperature of their water, are fitted to receive 

 them. 



With respect to the causes which influence these alternate 

 migrations of salmon, it would appear that they move towards 

 the sea chiefly in search of the food found on its coasts, and 

 on those of estuaries ; whilst the chief impulse that urges their 

 movement up rivers, is the propagation of their kind, whe|;e 

 alone the spawning process can be duly exercised. As to the 

 cause of their seeking fresh water, when not urged by that im- 

 pulse, we can ofler no other reason than that of a sort of instinct^ 

 which incites them to remove occasionally into fresh water, 

 in which alone they were at first able to exist; whilst the 

 appetite for food calls them again back to the sea. Perhaps, if 

 the water of rivers were always in sufficient quantity, and per- 

 fectly suited to the taste and economy of salmon, they would 

 be moving backwards and forwards from the sea to the river, 

 and from the river to the sea, at intervals more or less great ; 

 and, therefore, the different jxjriods in which they appear in 

 different rivers, is owing to the different circumstances and con- 

 ditions in which those rivers may be, rather than to any na- 

 tural difference in the economy or habits of the fish. How far 

 they move into the deep sea is not known, but that they roam 

 over the coasts, at great distances from the mouths of rivers, is 

 certain from the fact of their being captured in such situations. 



3 



