170' 



59° - 



58' 



57°- 



56' 



55' 



54' 



'P/. 



ST, PAUL I 



% 



© 



0©^^ 



NUNIVAKI 



• 1957 U.S 

 O 1958 U.S 

 n I960U.S 

 A I9 60 JAPAN 

 A 1961 JAPAN 

 ® 1964 U.S. 

 O 1965 U.S. 



©9^ 



ST GEORGE I. 



A^ 



Jd. 



9?' 



CAPE MORDVlNOF^.v'-^'^W>^-^o 





^^/ UNALASKA I 



170' 



165' 



160* 



Figure 12. 



-Areas of tagging by United States and Japan in the 6 yr selected for study to determine tlie distribution of sockeye salmon 

 stocks in offshore area of Bristol Bay. (See Appendix Table 1 for source of data.) 



greater proportion of the .2 fish, which were better 

 able to pass through the gill nets, might be captured; 

 .3 fish were usually held in the net by the gills. 



The size composition of sockeye salmon runs to 

 each fishing district in Bristol Bay varies from year to 

 year and between districts in a single year. In some 

 years the run to a given district may be composed 

 largely of the smaller .2 fish. In other years the larger 

 .3 fish may predominate. Large differences between 

 districts in the proportion of .2 and .3 fish could lead 

 to error in interpreting tagging results because of the 

 differences in the vulnerability of tagged fish in these 

 two size groups. No adjustments have been made for 

 the selective action of gill nets in calculating the ex- 

 pected number of tags from each district, and the 

 assumption is made in the analysis that tagged fish in 

 both size groups were as vulnerable to fishing as un- 

 tagged fish. 



Results of Tagging. — The tag recoveries, though 

 few in number and producing only small differences 



between the actual and expected number of tags 

 recovered in the four Bristol Bay fishing districts 

 (Table 1), are the results of the only significant tag- 

 ging to be carried out in the Bering Sea and outer 

 Bristol Bay. Therefore, I have used consistencies in 

 the recovery distribution for all 6 yr of tagging as an 

 indication of segregation of individual sockeye stocks 

 in the offshore area. The results of the chi-square 

 analysis to test the hypothesis of like distribution in 

 the offshore area for all sockeye salmon stocks of 

 Bristol Bay origin are given in Table 2. 



Except for 1958, the hypothesis of like distribution, 

 i.e., no segregation, for all sockeye salmon stocks of 

 Bristol Bay origin occurring in the eastern Bering Sea 

 is not rejected. A consistent feature in the recovery 

 data for those sockeye salmon tagged in the Bering 

 Sea in 1957, 1958, 1960, 1961, and 1964 (tagging sites 

 8, 9, and 10, Fig. 12) is the lower-than-expected 

 number of recoveries in the Nushagak fishing district 

 (Table 1). However, except for the 1958 tagging near 

 the Pribilof Islands, differences between the actual 



