FISHERY BULLETIN: VOL. 7 1, NO. 3 



5- 



3- 



2- 



^>e 



s 





o "* " 





10 



Figure 3. — Ratio of returns to number of spawners for the spring, summer, and fall groups of spawners 

 (see text) on the ordinate scale against hundreds of thousands of spawners (abscissa). 



Table 1. — Showing the number of odd- and even-numbered years on which each group of 

 spawners produced returns falling above or below the curve of Figure 3. 



(1972)3 and say (data not given) that Gard and 

 Drucker demonstrated the existence of these 

 races by comparing early spawning sockeye 

 salmon in the lateral tributaries of the lake, in 

 the upper Thumb, O'Malley River, and Canyon 

 Creek with later spawners in these terminal 

 streams, and on the Thumb Beach in Karluk 

 Lake. The late spawners were said to show 

 greater mideye to fork length, and greater fe- 

 cundity at comparable lengths. 



According to this statement Van Cleve and 

 Bevan show, not a comparison between geo- 



3Gard, R., and B. Drucker. 1972. Differentiation and 

 cause of decline of sockeye salmon of the Karluk River 

 system, Alaska. Unpubl. manuscr. Auke Bay Fish. Lab., 

 Natl. Mar. Fish. Serv., NOAA, Auke Bay, AK 99821. 



graphical localities (the same localities in each 

 season were not used) but merely between early 

 and late spawners. The validity of the quoted 

 comparison is open to serious question. In my 

 published report on the fecundity of North 

 American Salmonidae (Rounsefell, 1957) I 

 note that for Karluk sockeye salmon, over a 

 4-yr period, that 60-cm sockeye with 2 yr at sea 

 showed a consistently and statistically signifi- 

 cant higher fecundity than Karluk sockeye with 

 3 yr at sea. These data were for fish running 

 throughout the season. 



Davidson and Shostrom (1936:9) showed that 

 late-running salmon have longer heads and that 

 the distance from the eye socket (equivalent to 

 Gard and Drucker's center of the eye for general 



654 



