The routes followed during the migrations seem 

 unrelated to land or continental slielves. Some- 

 times a considerable concentration of salmon ap- 

 pears in passes and near points, but the normal 

 migratory routes seem to be across open water, 

 even where, as in the approaches to inner Bristol 

 Bay, it would be convenient to follow close to the 

 shore. 



The remarkable directness of the final migration 

 in the open sea is illustrated by the Karaginski 

 district pink salmon that migrate through tlie area 

 of the Japanese high-seas fishery. "^^Hien the fish 

 have been tagged and released from different 

 points along a north-south line, the recaptures 

 show the tendency to proceed rather directly to- 

 ward their destination (fig. 8). ObA'iously, their 

 migration is not random as suggested by Saila and 

 Shappy (1963). 



Many of the migration routes traveree different 

 ocean domains. The Karaginski pink salmon, for 

 example, within 2 months apparently travel from 

 the eastwardly flowing central Subarctic Current 

 across the Alaskan Gyre, across the westerly flow- 

 ing Alaskan Stream, and through the Bering Sea 

 Gyre in tlie western Subarctic Domain (see Dodi- 

 mead et al., 1963 :167). Bristol Bay sockeye salmon 

 regularly occupy the Alaskan Gyre and the west- 

 ern Subarctic Gyre in late spring and migrate 

 through the Bering Sea Gyre on the homeward 

 migration. Such routes clearly take the salmon 

 tlu'ough parts of the ocean far removed from any 

 recent mixing ^^•ith home-stream waters, and we 

 conclude, as does Neave (1064), that the olfactory 

 sense cannot provide a significant source of guid- 

 ance information except at the end of the route. 



Much of this migration takes place through and 

 during some of the most prolonged and violent 

 ocean storms in the world. The weather of the 

 Aleutians area (U.S. Department of Commerce, 

 1955 : 343) is characterized by persistently overcast 

 skies, high winds, and violent storms. No other 

 oceanic area in the wox-ld is recognized as having 

 worse weather in general than that of the Aleutian 

 Islands — clear weather over large areas is rare. 

 Even in the milder summer periods, the sky is 

 obscured by fog, mist, haze, or clouds most of the 

 time. In the outer parts of Bristol Bay (U.S. De- 

 partment of Commerce, 1955:631) the average 

 weather in June ( the month when most salmon are 

 homewai'd bound) is 44 percent fog, mist, or haze, 



and mean cloud cover is 8/10. In the central Gulf 

 of Alaska (U.S. Department of Commerce, 1955: 

 619) the average June weather is 25 percent fog, 

 mist, or haze; the mean cloud cover is 9/10. Our 

 own experience with several years of vessel opera- 

 tion in this area indicates that it is impossible to 

 navigate by celestial observations alone, and we 

 suggest that salmon have far too little opportunity 

 to observe tlie direction of either the sun or the 

 polarization of light to keep the kind of sched- 

 ule that they manage. 



POSSIBLE GUIDANCE MECHANISMS 



Little can be said about the physiology of salm- 

 on which specifically explains how they navi- 

 gate during transoceanic migrations, because the 

 necessary experiments have not been performed. 

 If certain generalizations are made, however, some 

 interesting possibilities emerge from the migra- 

 tion data just presented and from the literature of 

 fish physiology' as a whole. 



Two of the preceding generalizations from the 

 tagging and seining data seem particularly sig- 

 nificant. First, the most common direction of 

 travel follows the various North Pacific currents. 

 Second, salmon in the open sea do not drift with 

 the current, but actively swim with it. We, there- 

 fore, conclude that the migrations of salmon on the 

 high seas are actively directional in a way which 

 somehow relates to the ocean currents. Directional 

 cues for animal orientation have included celestial 

 bodies, water movement, olfactory stimuli, and 

 electrical or magnetic fields. Let us examine 

 whether any of these are compatible Avith our data 

 on salmon migrations. 



SUN ORIENTATION 



The sun, a prominent object by which some ter- 

 restrial animals navigate, could similarly serve 

 salmon on the high seas. Indeed, certain lake fish 

 have such an orienting mechanism (Hasler and 

 Schwassmann, 1960). Also, some arthropods ori- 

 ent to the plane of polarization of sunlight (Ivan- 

 off and Waterman, 1958). Arthropods and fish 

 both require good visibility of the sun, however, 

 because both become disoriented when clouds ob- 

 scure the sun (Waterman, 1959: Ilasler et al., 

 19.i8). Salmon, on the otlier hand, migrate at night 

 as well as during the day and through regioiis 

 wliere clouds obscure the sun almost continuously. 



JIODBLS OF OCEANIC MIGRATIONS OF PACIFIC SALMON 



457 



