In contrast to the low movement rates observed 

 for coho salmon in coastal waters off Oregon, 

 Washington, and Vancouver Island, net migra- 

 tion rates reported for salmon tagged in offshore 

 waters of the North Pacific were generally much 

 higher. Godfrey et al. (1975) calculated an aver- 

 age rate of 24 km/day for tagged coho salmon 

 recovered in the Japanese high seas fishery and 

 30 km/day for coho recovered in coastal waters. 

 Hartt (1962, 1966) estimated that the migration 

 rates of sockeye salmon into Bristol Bay averaged 

 44-50 km/day, whereas those of maturing sockeye 

 salmon caught on the high seas averaged about 

 32 km/day. Chum salmon had migration rates at 

 sea of 23-50 km/day; pink salmon had average 

 rates of 43 km/day for coastal returns and 50 km/ 

 day for high seas returns. 



Rapid migrations in coastal waters of British 

 Columbia and Washington were also found for 

 pink salmon (Vernon et al. 1964; Stasko et al. 

 1973) and for sockeye salmon (Madison et al. 

 1972) and off the Kurile Islands for chum salmon 

 (Ichihara et al. 1975). However, migration rates 

 slowed greatly as fish neared their home river 

 systems (Vernon et al. 1964; Groot et al. 1975). 



Because net migration rates of coho salmon in 

 coastal waters off Oregon, Washington, and Van- 

 couver Island are so much lower than movement 

 rates found for other salmon stocks, these coho 

 are probably spending less time migrating in a 

 single direction compared to meandering and 

 feeding. Similarly, Milne (1957) concluded that 

 coho salmon in coastal waters of British Colum- 

 bia probably meander during both feeding and 

 spawning migrations. Slow, feeding movements 

 off Oregon and Washington are also suggested by 

 the long time period (3-4 months) during which 

 individual stocks are available in the ocean fish- 

 eries (Hunter 1985). The relatively fast net mi- 

 gration rates observed for some coho salmon re- 

 captured within 33 days of release (Fig. 2B) 

 suggest that actual movement rates over short 

 time periods may be quite high but that meander- 

 ing courses over time produce low net migration 

 rates. Higher migration rates for fish tagged in 

 late summer, to be expected if movements were 

 changing from predominantly feeding to homing, 

 were not apparent (Fig. 2C). 



Roughly equal numbers of coho salmon were 

 recovered to the north (27) and to the south (35) of 

 release sites, although most (8 of 11) coho re- 

 leased from lat. 45°N and south were recovered to 

 the north (Fig. 2D). Van Hyning (1951) also found 

 that most coho tagged south of 45°N travelled to 



the north after release; however, he found that 

 most coho released off northern Oregon and the 

 Columbia River (46°15'N) were recovered to the 

 north as well. Fry et al. (1951) and L.B. Boydston 

 (California Department of Fish and Game, un- 

 publ. data) reported that most recoveries of ma- 

 turing coho salmon tagged off northern California 

 were to the north, off Oregon or Washington. 



Northward migration by most of the maturing 

 coho salmon tagged at sea south of 45°N during 

 their final summer in the ocean is consistent with 

 the distributional patterns of coastal Oregon and 

 early run Columbia River stocks in the ocean 

 fisheries. Peak catches of coastal Oregon coho 

 salmon stocks are off northern California in May 

 and June and shift to waters off Coos Bay in July 

 and August (compiled from Hunter 1985). Rela- 

 tively high percentages (24-37%) of the ocean 

 catch of coastal Oregon coho salmon stocks (all 

 combined) are off northern California (Garrison 

 1985; Hunter 1985; Oregon Department of Fish 

 and Wildlife 1982). Similarly, between 62 and 

 65% of early run Columbia River stocks are 

 caught south of the Columbia River (Hunter 

 1985; Howell et al. 1985; Oregon Department of 

 Fish and Wildlife 1982). Therefore, many fish 

 from these two stock groups, which make up a 

 substantial fraction of the coho catch off Califor- 

 nia and Oregon, migrate south and then migrate 

 north sometime later during the summer to re- 

 turn to their natal systems. Southward migration 

 into waters off northern California and southern 

 Oregon may be advantageous to these coho 

 salmon stocks because of the potentially high food 

 production fueled by strong coastal upwelling 

 during the summer in this area. 



Other stocks of coho salmon are caught to the 

 north of where they entered the ocean during 

 their final summer in the ocean. About 47% of the 

 late run Columbia River coho are caught north of 

 the Columbia River in Washington and British 

 Columbia (Howell et al. 1985). Smaller, but sig- 

 nificant percentages of other stock groups from 

 the south (early Columbia River, private hatch- 

 ery, and other coastal Oregon groups) are also 

 caught as maturing adults north of their natal 

 streams. Thus, these fish would eventually have 

 to migrate to the south to return to their natal 

 streams. Therefore, the subsequent southward 

 movement of many of the maturing coho salmon 

 we tagged north of 45°N (Fig. 2D) is not surpris- 

 ing. 



The slow net migration rates, prolonged resi- 

 dence in coastal waters, and mixed north and 



824 



