Odor Perception and Taste 223 



ment. Gates at the openings of tlie tubes were opened and the fi)' were 

 allowed to enter the tubes of their choice. Their distribution after the test 

 was recorded and compared with the distribution obtained when no odor 

 \\as introduced. 



Many organic odors were tested in this manner. None was found which 

 \A'ould attract salmon, but morpholine seemed to fit the necessary require- 

 ments—soluble in water permitting accurate dilutions, detectable in very 

 low concentrations (easily detected at 1 X 10~^ ppm) making the treat- 

 ment of large volumes of water feasible, and chemically stable under 

 stream conditions. At low concentrations morpholine was neither an 

 attractant or repellent, so that salmon could be conditioned to it in either 

 direction. Thus salmon fry which hatched in a stream treated with a 

 low concentration of morpholine should associate that odor with the 

 spawning site when they returned several years later. This would give the 

 stream a characteristic odor [and satisfy condition (1)]. Salmon have 

 already shown evidence of detecting this odor at concentrations as low 

 as 1 X 10~^ ppm [condition (2)] and if they are able to retain odor 

 impressions from youth to maturity they should have no difficulty in 

 arriving at the original spawning site. A further and a very crucial test 

 would be an attempt to decoy the salmon conditioned to morpholine up 

 the wrong stream branch by treating that branch with morpholine. If 

 the salmon entered the wrong branch, it would be certain proof that they 

 were migrating entirely by odor impressions. However, if they did not 

 enter the wrong branch, the odor theory would not necessarily be dis- 

 proved because the odor impression of the 'liome" stream might consist 

 of a complex of odors including morpholine, and the salmon might not 

 have accepted the other components making up the odor complex (ex- 

 cluding morpholine) of the wrong stream branch. Salmon could be 

 expected to enter a "wrong" stream branch when die odor impression for 

 morpholine completely subordinated other stream odors. Field tests to 

 measure the possibilities given above have not been completed. 



In further experiments on odor recognition in salmon, sexually ripe 

 coho salmon were captured in each of two branches of the Issaquah River 

 in Washington, marked and returned downstream below the junction of 

 the two branches to remake the upstream run and reselect the correct 

 branch.103 j^ one half of the fish the nasal sac was plugged with cotton. 

 Most of the normal fish repeated their former stream choice upon reaching 

 the fork while the plugged-nose fish selected one or the other of the 

 branches in nearly random fashion. While the pressure of the cotton plugs 

 in the nasal sacs of the plugged-nose salmon may have influenced their 

 behavior (and no similar pressure was applied to the normal fish), the 

 experiment is indicative of the importance of the olfactory system in the 

 migration of these fish. 



