CHEMORECEPTION: LOCOMOTION AND ORIENTATION 301 



distance from the source increases. Given the proximity of the nostrils in all 

 but a few fishes, a bilateral differential in stimulus strength, dependent on dif- 

 ferences in molecular density of the odor substance, is an unlikely orienta- 

 tion mechanism at best, in all but the steepest gradients. In most of the 

 natural environment, therefore, whenever olfaction is involved, a different 

 orientation principle must prevail. The strongest biological evidence in favor 

 of this postulate is the highly efficient olfactory orientation in Petromyzon 

 marinus (Kleerekoper and Mogensen 1963), although, as a cyclostome, it has 

 a single median nostril (monorhinic). 



The results of earlier experiments (Kleerekoper 1967) suggested an orien- 

 tation mechanism in which the role of olfaction would be unrelated to the 

 gradient of the odor substance. It was observed that the teleost Diplodus 

 sargus, when stimulated with an attractant odor, modified its turning be- 

 havior as described above and oriented toward the odor-releasing compart- 

 ment of the tank, but was unable to localize that compartment. The fish 

 generally entered the neighboring compartments as if it were turning "too 

 early." In these experiments, the slow rate of water flow, from the periphery 

 toward the center, was the same in all compartments (1 1/min). When, how- 

 ever, a differential flow was created increasing the rate in the odor-releasing 

 compartment by 10%, localization by the fish became absolutely accurate, 

 although no preference for that compartment occurred in response to the 

 differential flow rate alone, that is, in the absence of the olfactory stimulus. 



These results strongly suggested a similarity with the behavior of various 

 unrelated organisms, described in the literature, in which odor acted as a 

 releaser of rheotaxis. This phenomenon was first observed in the aquatic 

 snails Alectrion and Busy con by Copeland (1918) and confirmed by 

 Henschel (1932) for Nassa reticulator. Odor stimulation in stagnant water 

 evoked random "alarm" movements but when a current was produced, loco- 

 motion became at once positively rheotactic. Similar behavior has been ob- 

 served in terrestrial organisms that display anemotaxis when stimulated with 

 an attractant odor, such as Triton (Czeloth 1931), Drosophila (Fliigge 1934), 

 and Planaria (Doflein 1926). 



The significance of the findings with Diplodus, later confirmed in a series 

 of experiments, is in the fact that a diffusion field of a chemical source 

 contains no direction vector that might be used in orientation toward the 

 source. Localization could only occur by random sampling and a topo- 

 graphic comparison of relative densities of odor molecules, the results of 

 which would have to bias locomotor behavior in the general direction of 

 increasing odor concentration. One such chemotropotactic orientation mech- 

 anism, applicable in a steep molecular gradient such as prevails at short 

 distance from the odor source, was already discussed. However, it could not 

 function, as was pointed out, beyond relatively short distances from the 

 source. Water flow, on the other hand, contains a strong direction vector to 

 which rheotactic movements respond in a highly efficient fashion. Hence, it 

 was postulated that, in fish, the perception of an attractant odor, rather than 

 allowing efficient oriented locomotion, merely releases positive rheotaxis 

 (Kleerekoper 1967, 1969). In an odor-dependent displacement, be it in 



