478 Memoir Sears Foundation for Marine Research 



tively) and in New England rivers, grilse have been reported as being seldom seen 

 (JOJ; I04\ 2g) or rare (7). 



Rivers having a high proportion of fish that have been at sea for three years are 

 known as "large salmon" rivers, two of the most notable being the Moisie and Grand 

 Cascapedia rivers, for which the following percentages have been given for two- and 

 three-year sea-life fish, respectively: i8.3and 58.8 for the Moisie (J2); 6 and 58.8 forthe 

 Grand Cascapedia {2^^. The tabulated data to the left, giving the percentage of Atlantic 

 Salmon of 30 pounds and over in the Moisie run, suggest that 

 there has been a decline in the proportion of heavier fish in 

 that river (j^). 



The average weights (in pounds) of Atlantic Salmon re- 

 ported for various rivers are: for those two years at sea— 10.8 

 for the Restigouche in New Brunswick, 8.9 for the Godbout 

 in Quebec, and 11.2 for the Wye in England, averaged from 

 records of more than 20 years (9); 11.7 for the Grand Casca- 

 pedia, 10 for the Miramichi,!^ and 11.4 for the east coast 

 of Newfoundland, including Labrador {20)\ for those three or more years at sea — 

 Grand Cascapedia, 23.6 and 37 for 3 and 3 + fish, respectively (29); Miramichi, 19.5 

 for 3 and 3 + fish combined;^* and the coast of eastern Newfoundland, including Lab- 

 rador, 19.5 for 3 and 3 + combined (20). 



It has been suggested for Newfoundland Atlantic Salmon that those having a short 

 river life tend to stay longer at sea before they return to spawn {q^. On the other 

 hand, the opposite suggestion is implied in Dahl's statement that findings in Norway 

 do not appear to support the idea that the longer parr remain in the river the less 

 time they will spend in the sea (J7). For Scotland, Calderwood found that smolts 

 that matured rapidly and well in the river were likely to stay away from fresh water 

 for some years after leaving it, while smolts that remained in the river for more than 

 two years and developed slowly were likely to return to it soon. However, he found 

 that this relationship did not apply to the Grand Cascapedia, Quebec. A tabulation 

 of the characteristics of Atlantic Salmon from various sections of the Gulf of St. 

 Lawrence does not show a consistent relation between those factors {12). 



There is no agreement from either observations or experiments concerning the 

 relative extent to which heredity or environment affect the early or late running char- 

 acteristics of fish of the different rivers {i44\ 26: 22), or the marked changes that 

 have sometimes occurred in the proportion of grilse running to a number of streams, 

 or other differences. The nature and extent of these changes are indicated in the fol- 

 lowing examples. 



A serious and almost catastrophic decline in the stock of grilse along the coast 

 and in the rivers of Scotland during the present century (lOO: J^), more especially after 

 about 1908, has prompted the question as to whether selective fishing or some other 

 factors may be responsible for these changes. 



13. Blair, M. A. thesis, University of Toronto (1932)- 



