Furthermore, small lakes have only minimal 

 amounts of wave action to complicate a fish's view 

 of the sun, but marine Hsh must view the sun 

 through a water surface which is never completely 

 calm. Also, lake fish can migrate in straight lines; 

 salmon migrate along circular as well as straight 

 roiites. This circumstantial evidence casts doubt on 

 sun orientation as a primary navigational mech- 

 anism for Pacific salmon. 



Although the cloudy weather of the North 

 Pacific severely limits the possible use of sun orien- 

 tation by salmon, use of the sun cannot be entirely 

 eliminated from consideration. An oceanic bird, 

 the slendcrbilled sliearwater {Puffinus ten.uiros- 

 fns) , which makes an annual circum-Pacific, trans- 

 equatorial migration, travels near the Aleutian 

 Islands at about the same time that some salmon 

 there begin their final liomeward migration. The 

 slenderbilled shearwater appears to use sun ori- 

 entation as its primary navigational cue (Ser- 

 venty, 1963), even though a related shearwater 

 becomes disoriented during overcast conditions 

 (Matthews, 1964). Despite such evidence on pos- 

 sible use of the sun, a simpler hypothesis to ex- 

 plain navigation by salmon on the high seas seems 

 preferable. 



ORIENTATION TO WATER MOVEMENT 



Fish in ri\'ei's and streams are very sensitive to 

 current direction and usually orient upstream to 

 maintain position. Salmon smolts often show an 

 active downstream orientation during their sea- 

 ward migration. Optical, tactile, and latei-al-line 

 senses all seem to be involved in these rheotropisms, 

 but the need for some kind of stationary reference 

 point, such as a shore or the stream bottom, makes 

 this an unlikely mechanism for use on tlie higli 

 seas where reference points are very distant. Direct 

 detection of water movement also seems unlikely 

 because of the very large size of the water bodies 

 and, therefore, the correspondingly slight velocity 

 gradients. A gradient does exist, however, and tlic 

 sensitivity of fish to velocity gradients in the 

 absence of other cues appears untested; it cannot 

 be entirely excluded as a ]ws>sible navigational 

 mechanism. 



Salmon are perliaps capable of detecting tlie 

 interfaces between moving bodies of water either 

 through sensing of chemical differences between 

 two bodies of water or by detecting tlie water tur- 



bulence at tlie interface. This navigational cue 

 would lead to a great deal of random swinnning, I 

 however, wliiie tlie fish searches for tiiese margins 

 and would tend to concentrate fish near tlie mar- 

 gins of ocean currents: neither behavior is char- 

 acteristic of salmon migration. Salmon are rela- 

 tively evenly distributed across the ocean currents 

 and the migration routes appear well defined and 

 "purposive," often converging on the spawning 

 streams from se^•eral directions. 



ELECTRICAL POTENTIALS AVAILABLE FOR 

 ORIENTATION 



Because sea water is an electrical conductor mov- 

 ing tlirough the eai-tlvs magnetic field, the ])roduc- 

 tion of an electrical voltage can be expected. 

 Oceanographers have recognized the electrical po- 

 tential of sea water for some time. Stommel (1954) 

 found potential dift'erences of 0.2 to 2.6 v. across 

 long distances in the Atlantic Ocean among sev- 

 eral submarine cables. Similar voltages were ob- 

 served between Florida and Cuba. Snyder (1966), 

 in describing the underwater search for the atomic 

 submarine Thresher, reported a voltage of about 

 140 my. (millivolts) between towed electrodes, one 

 on the surface and one in about 2,440 m. of water. 

 Hughes (1962) attempted to use these voltages to 

 measure ocean currents by towing electrodes, 

 spaced 46 m. apart, behind a ship. He found volt- 

 ages of 3 to 5 mv. when the ship crossed the cnn-ent 

 and a reversal of polarity when the ship traveled 

 in the reverse direction. Voltage was nil when the 

 ship moved either with or against the current. The 

 voltages per knot of current \aried considerably 

 in sliallow water but consistency was greater in 

 deep water. 



The electrical gradient that miglit lie available 

 for na'\igational use, therefore, is about 0.1 to 0.;") 

 fiv. (microvolts) /cm. Because these voltages are 

 directly related to the current and arc jiolarized 

 with respect to its direction, electrical cues seem 

 to be a possible navigational device for salmon on 

 the high seas. Magnetic sensitivity has been shown 

 for several animals, but the receptor organ is un- 

 known. Presumably, detection occurs as a i-e.sult 

 of a voltage induced within the receptor (Brown, 

 Barnwell, and Webb, 1964). The basic question, 

 therefore, is whether fish can detect minute 

 voltages. 



458 



U.S. FISH AND WILDLIFE SERVICE 



