and expected numbers of tag recoveries were not large 

 enough to result in significant chi-squares in tests 

 between sockeye salmon stocks from individual 

 fishing districts or between members of stocks cap- 

 tured in fishing districts on the east and west sides of 

 Bristol Bay (Table 2). The significant chi-square ob- 

 tained in the analysis of the 1958 tag recovery data 

 was largely due to the few Nushagak recoveries and to 

 a lesser extent the more-than-expected Egegik 

 recoveries (Table 1). The few tags recovered from the 

 taggings at sites near the Pribilof Islands in 1964 (sites 

 8 and 9, Fig. 12) did not result in significant chi- 

 squares and is in contrast to the results of the 1958 

 tagging in the same area. 



Tagging in outer Bristol Bay in 1964 at stations 12, 

 13, 15, and 16 (Fig. 12) was closer to the mouths of the 

 major river systems, where one might expect more 

 segregation of individual stocks. Tagging stations 12 

 and 15 were near shore, and stations 13 and 16 were 

 well offshore. For these reasons tag recoveries from 

 each of the four stations were treated individually and 

 then combined as nearshore and offshore tagging 

 (Table 1). In all tests the differences between the ac- 

 tual and the expected number of tag recoveries were 

 not sufficient to result in significant values of chi- 

 square (Table 2), and the hypothesis of like distribu- 

 tion for all sockeye salmon stocks is not rejected. 



Although the data are weak because of the few tags 

 recovered from taggings at stations 12, 13, 15, and 16, 

 the distribution of the tag recoveries suggests that 

 segregation of individual stocks may be beginning in 

 this region of Bristol Bay. For example, recoveries in 

 the Egegik and Ugashik districts (Table 1) were about 

 equal to or were greater than the expected number 

 from the tagging at station 12 (inshore), but no 

 recoveries were made from the offshore tagging at sta- 

 tion 13 (offshore). Also recoveries from the Naknek- 

 Kvichak district were less than expected from fish 

 tagged at station 12 but about as expected from fish 

 tagged at station 13. Recoveries from Nushagak Bay 

 were about as expected from tagging at station 12 and 

 greater than expected from tagging at station 13. 



My interpretation of the data from the outer two 

 stations is as follows: Ugashik and Egegik fish are pres- 

 ent at station 12 nearshore but not at station 13 

 offshore, an indication that these stocks are beginning 

 to leave the offshore waters and move toward their 

 home-river systems on the east side of Bristol Bay. 

 This would decrease the proportion of Naknek- 

 Kvichak stocks at station 12 near shore. The number 

 of Naknek-Kvichak recoveries from tagging at station 

 12 was, in fact, less than expected (Table 1). The ac- 

 tual number of Naknek-Kvichak recoveries from sta- 

 tion 13 is close to the expected number and indicates 



Table 2. --Summary of chi-square analysis of recovery data for tags 

 the United States and Japan between 1957 and 1965. 



released in the eastern Bering 

 (See Fig. 12 for location of 



Sea and outer Bristol Bay by 

 release sites. ) 



10 



