115 



no 



60 



90 



'57 / 



61 

 / o56 



->55 



/ 



/o'59 



/o'53 



75 80 85 90 



MEAN LENGTH OF AGE I SMOLTS (MM.) 



95 



Figure 16. — Relation between mean lengths of age II and 

 age I sockeye salmon smolts in the same year of sea- 

 ward migration, 1953-61, Wood system. The number 

 beside each point indicates the year of smolt migra- 

 tion. 



The individual points in the relation between 

 the length of smolts and numbers of adults in the 

 escapement do not fit the regression closely, and, 

 in addition, exact data on natural mortality in 

 the ocean are lacking for the Wood system. 

 Hence, the illustration from the Wood River 

 smolt migration of 1955 given above can only 

 suggest the effect of the size of smolts on ocean 

 survival. It is probable, however, that ocean mor- 

 tality related to size attained by age I sockeye 

 salmon in fresh water is an important factor in 

 the determination of the numbers of adults in 

 the returns to the Wood system. 



The average length of age II smolts was better 

 correlated with the mean length of age I smolts 

 of the following year class than with the number 

 of fish in the parent escapement (fig. 16). This 

 relation suggests the presence of competition for 

 food between age classes of juvenile sockeye 

 salmon. 



A lack of significant competition for food 

 among individuals of an age class after their 

 first year in the lake is due to their comparatively 



small numbers — most sockeye salmon smolts of 

 the Wood River are age I (table 26). 



Kvichak System 



The data on mean lengths of age I and age II 

 smolts produced by each escapement to the Kvi- 

 chak system in 1952-60 (table 27) were used to 

 calculate a straight-line regression of smolt 

 length on numbers of adults in the parent escape- 

 ment for these years. The relation for age I 

 smolts is 



L, = 90.64 — 0.7X 



and for age II smolts 



L„ = 115.51 -1.4X 



where L is the length of smolts in millimeters 

 and X is the number of fish (millions) in the 

 parent escapement (fig. 17). The correlation 

 coefficient for age I smolts was r = —0.73 and 

 for age II smolts, r = —0.93. 



The slope of the regression line of smolt size 

 on escapement for age II smolts of the Kvichak 

 system is almost twice that of the age I ; the 

 growth retardation caused by competition for 

 food among the progeny of each escapement 

 seems to continue during their second year in 

 fresh water. The contrast to the Wood system in 

 this regard may be the result of the proportion- 

 ately greater holdover of young to age II smolts 

 in the Kvichak system. 



The regression of length of age I smolts on 

 numbers of adults in the parent escapement for 

 the Wood system has a slope six times that de- 

 rived for age I smolts in the Kvichak system. 



,5; 130 



• AGE I SMOLTS 

 o AGE H SMOLTS 



Lq = II5.5I-L4X 



^ ^ 2 4 6 8 10 12 14 16 



FISH IN PARENT ESCAPEMENT (MILLIONS) 



Figure 17. — Relation between mean lengths of age I and 

 II sockeye salmon smolts by age class and numbers of 

 adults in the parent escapements, Kvichak system. 



444 



U.S. FISH AND WILDLIFE SERVICE 



