364 



Fishery Bulletin 103(2) 



correlations was low, but the consistent pattern among 

 stocks, age groups, and time periods indicates that the 

 negative correlations were not spurious. 



Discussion 



Previous research documented reduced annual scale 

 growth of Nushagak Bay (Bristol Bay) sockeye salmon 

 during odd-numbered years of their second and third 

 years at sea (Ruggerone et al., 2003). The primary find- 

 ing of our investigation was that salmon scale growth 

 reduction during odd-numbered years did not occur 

 throughout the second and third years at sea. During the 

 second year at sea, scale growth reduction began three 

 to five circuli after peak scale growth. During the third 

 year at sea, scale growth reduction began immediately 

 after peak growth. This finding was consistent among all 



age groups of both Kvichak and Egegik sockeye salmon 

 prior to and after the mid-1970s regime shift that led to 

 greater sockeye salmon abundance. Comparison of sea- 

 sonal scale growth patterns before and after the regime 

 shift indicated that the recent period of high sockeye 

 salmon abundance was associated with relatively high 

 growth 1) immediately after entry to Bristol Bay, 2) after 

 peak scale growth during the first growing season, 3) 

 during the middle of the second growing season, and 4) 

 during the third spring but followed by below average 

 growth during the remaining summer and fall. 



Timing of peak scale growth and differences in 



scale growth between odd- and even-numbered years 



The approximate time period of peak scale growth can 

 be estimated from previous studies of salmon circuli for- 

 mation at sea and timing of peak prey production. Bilton 



