fewer returns undoubtedly are related to a di- 

 minished effort to recover tags from these experi- 

 ments. Fish counting and tag observing were dis- 

 continued at Zosel Dam from August 20 to August 

 26, at a time when tagged fish were still passing 

 and when the number of tags from the August 

 10-13 experiments should have peaked there. 

 Therefore, with comparable recovery effort, tag 

 returns from these experiments probably would 

 have been greater than indicated. It is noted 

 that the 1955 returns from this same time period, 

 August 9-11, and for comparable water flows 

 (fig. 10), were consistent with returns from earlier 

 experiments occurring at high flows. It is evi- 

 dent, as shown by upstream tag returns, that 

 changes in flows at Rock Island Dam had little 

 effect on the ability of sockeye to pass the dam. 



Chinook 



In order to compare recovery ratios of chinook 

 salmon tagged above and below the dam, it was 



necessary to separate the chinooks by race. 

 Races at Rock Island Dam were reported by 

 Fish and Hanavan (1948) as spring and summer 

 chinook. Spring chinook, passing Rock Island 

 Dam earlier in the season than summer chinook, 

 migrated to smaller, more remote spawning 

 streams. In these streams, tagged fish were 

 much easier to observe and recover than in larger 

 streams, such as the Wenatchee and lower Methow 

 Rivers (the summer chinook spawning areas). 

 Thus, recoveries from spring chinook tagging 

 experiments were much greater than from experi- 

 ments on summer chinook. 



Upstream tag recoveries for all chinook experi- 

 ments are given in table 9. Many samples were 

 too small to compare returns statistically. Only 

 in 1956 were sufficient fish obtained to release at 

 the two banks below the dam and above the dam. 

 Returns of spring chinooks tagged below the dam 

 were virtually the same (27 percent) for the 3 



LOSS AND DELAY OF SALMON PASSING ROCK ISLAND DAM 



351 



