GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON 



160° 150° 



64= 



140"^ 



130° 



64°N 



60= 



56= 



52= 



48= 



Figure 8. — Map of the North Pacific Ocean showing the sites sampled by an oceanographic cruise in 1959 

 (S.I.O.U.C. 1965) and Station P, where sea surface temperatures and salinities were also recorded. 



Table 3. — Temperature at depth, recorded by the RV 

 Brown Bear, 16-24 July 1959 (data from S.I.O.U.C. 

 1965). 



iData not included In calculations of gradients. 



mer when the sockeye salmon migrate through 

 these areas, the vertical and horizontal gradients 

 are presumably smaller. 



Available information indicates that salmon in 

 general and sockeye in particular do not restrict 

 themselves to one depth but rather have a diel 

 vertical movement pattern while at sea. Manzer 

 (1964) reported that most sockeye were caught in 

 gill nets at or near the surface during the night, 

 but in the daytime they were caught in substan- 

 tial numbers as deep as 48 to 60 m. Mishima and 



Shimazaki (1969) reported a more complex pat- 

 tern: sockeye were most abundant on the surface 

 at 13:00-15:00 h but a second peak of abundance 

 occurred at 03:00-05:00 h. Whereas variations in 

 diel movements and depth distribution may 

 occur, it seems likely that sockeye experience 

 temperature changes of 1°C during their daily 

 movements, and may experience changes of 4° to 

 5°C if they dive below the mixed layer. 



The slight changes in temperature associated 

 with horizontal movement relative to vertical 

 movement make it unlikely that the long- 

 distance migration of homing sockeye is deter- 

 mined by physiological responses to temperature 

 (Laevastu 1983). Moreover, the temperatures ex- 

 perienced by sockeye salmon at sea do not seem to 

 reflect physiological optima (Brett 1974, 1983). 

 Nevertheless, there is a west-east gradient of in- 

 creasing temperature over much of the homeward 

 path of Fraser River sockeye salmon. Therefore, 

 "predictive behavioural thermoregulation" (Neill 

 1979) may play a role in homing, though 

 "reactive behavioural thermoregulation" (e.g., 

 011a et al. 1975) probably does not. However, gra- 

 dients are an inefficient aid to migration unless 



465 



