spring and a return migration toward the east 

 in the Subarctic Current (fig. 1). The fish appar- 

 ently repeat the winter-spring-sununer round of 

 migrations of their first year at sea (fig. llA). 

 Further support for this hypothesis is provided by 

 the recovery in the Slmmagin Ishxnds area in the 

 eastern Aleutians of .2-age sockeye salmon that 

 had been tagged as .1-age immature fish a year 

 earlier near Adak Island (fig. 12). The 1964 data 

 are typical of other years" results. At Adak they 

 had been part of the \yestward flow of immature 

 fish, and at Shumagin Islands they were among 

 maturing fish that characteristically migrate west- 

 ward, many of them en route to Bristol Bay 

 (Thorsteinson and Merrell, 1964). 



THIRD YEAR AT SEA 



The migrations during the third year at sea 

 must be considered separately for one group that 

 matures and returns to spawn at the end of its 

 second year at sea (early in its third summer at 

 sea), and another group that remains another full 

 year at sea. From the numbers of .2- and .3-age 

 fish returning in successive years, about 60 to 80 

 percent of the .2-age group mature and spawn. 

 The fish of the maturing group average about 51 

 to 53 cm. in fork length and are thoroughly mixed 

 with the older .3-age mature salmon. They ap- 

 l)roach the eastern Aleutian Islands area from the 

 south and southeast in late May, migrate westward 

 with the Alaskan Stream, northward through the 

 Aleutian passes, and then northeastward toward 

 their respective Bristol Bay estuaries (Hartt, 

 1966). The recovery distribution of sockeye salm- 

 on (mostly .2-age mature fish) tagged in the 

 Aleutian Island area in 1960 is illustrated in figure 

 14. Those approaching from far to the west ap- 

 parently migrate rather directly toward the Bay. 

 Kecaptures by the high-seas fleet illustrate the 

 course through the central Aleutian area. The lack 

 of high-seas returns from releases in the eastern 

 Aleutians (fig. 14D) indicates that this group must 

 have turned northward and entered the Bering 

 Sea before reaching lat. 175° W. Migration 5peed 

 is rapid chiring the final 30 to 60 days at sea, .iver- 

 aging 25 to 30 miles (46.3 to 55.6 km.) per day. 



Even as late as May 1, sockeye salmon from 

 Bristol Bay are distributed over vast areas at sea 

 extending from 1,200 miles (2,222 km.) to the west 

 off the Kamchatka Peninsula to 1,200 miles to the 



east in the central Gulf of Alaska. Figure 15 il- 

 lustrates the tagging areas during the period May 

 1 to June 15 from which Bristol Bay recoveries 

 were subsequently received in the year of release. 

 The bulk of the run passes through the Aleutian 

 Islands area between June 1 and 20, and through 

 the estuaries between June 23 and July 18, with 

 the peak between July 2 and 9. Tagging has indi- 

 cated a sequential correspondence between date of 



O TAGGING ARE* ft NO TAGGED 

 . ^BECOVERT LOCATION 

 6B MATURES TAGGED 

 20 RETURNS 



Figure 14.— Distribution of recoveries of sockeye salmon 

 tagged and recovered in 1060 (U.S. tagging). Panels A, 

 ('. and D each show the results from two areas of 

 oi>eration : panel B shows the results from one. 



Figure 1.5. — Areas of release of sockeye salmon tagged 

 between May 1 and .Time l.">, and subsequently recovered 

 in Bristol Bay in the same year lietween .Tune 26 and 

 .July 24 (U.S. data 19.">6-6.">. Canadian data 1061-G-'), 

 .Japanese data 1958-61 ) . 



MODELS OF OCEANIC MIGRATIONS OF PACIFIC SALMON 



453 



