Behavioral and physiological reactions were examined in juvenile coho salmon (Oncorhynchus kisutch) before and after the chemodetection of alarm substance and chemical stimuli released from predators and nonpredators. Chemical stimulus from northern squawfish (Ptychocheilus oregonensis) was avoided in a two-choice Y trough, whereas stimulus from the largescale sucker (Catostomus macrocheilus) was not. Paradoxically, both stimuli induced a stress response as indicated by elevations in plasma Cortisol and glucose. Plasma thyroxine was not a sensitive indicator of stress. Extracts from the broken skin of squawfish or suckers induced a physiological stress response, thereby raising the possibility of a nonostariophysan fish recognizing an ostariophysan alarm substance. Rinses from human skin or L-serine were behaviorally avoided, but neither stimulus induced physiological stress responses. A conclusion from these results is that behavioral and physiological reactions to chemical stimuli from predators do not necessarily co-occur. These data also suggest that fright is not necessarily a sufficient condition for inducing a stress response of the general adaptation syndrome type in fish.

A B S T R A C T Beidler's work in the 1950s showed that anions can strongly influence gustatory responses to sodium salts. We have demonstrated "anion inhibition" in the hamster by showing that the chorda tympani nerve responds more strongly to NaCI than to Na acetate over a wide range of concentrations. Iontophoretic presentation of CI-and acetate to the anterior tongue elicited no response in the chorda tympani, suggesting that these anions are not directly stimulatory. Drugs (0.01, 1.0, and 100 IxM anthracene-O-carboxylate, diphenylamine-2-carboxylate, 4-acetamido-4'-isothiocyanatostilbene-2,2'-disulfonate, and furosemide) that interfere with movements of Cl-across epithelial cells were ineffective in altering chorda tympani responses to 0.03 M of either NaCI or Na acetate. Anion inhibition related to movements of anions across epithelial membranes therefore seems unlikely. The chorda tympani contains a population of nerve fibers highly selective for Na ÷ (N fibers) and another population sensitive to Na + as well as other salts and acids (H fibers). We found that N fibers respond similarly to NaCI and Na acetate, with spiking activity increasing with increasing stimulus concentration (0.01-I.0 M). H fibers, however, respond more strongly to NaC1 than to Na acetate. Furthermore, H fibers increase spiking with increases in NaCI concentration, but generally decrease their responses to increasing concentrations of Na acetate. It appears that anion inhibition applies to taste cells innervated by H fibers but not by N fibers. Taste cells innervated by N fibers use an apical Na + channel, whereas those innervated by H fibers may use a paracellularly mediated, basolateral site of excitation.

Timidity, vigilance and response to alarm substance were examined in shoals consisting of one to 12 zebra danios, Brachydnnio rerio. Measures of timidity were not clearly related to shoal size.Fish in larger shoals spent more time in central areas of the tank away from cover than fish in smaller shoals. However, fish in small shoals appeared to be as aggressive as those in larger shoals. Vigilance and foraging rates did not appear to be related to shoal size. Per capita foraging rates and shoal size were not correlated. After being frightened by alarm substance, danios in larger shoals did not return to foraging sooner than those in smaller groups. Zebra danios in all shoal sizes responded behaviourally to alarm substance. It appears that the presence of conspecifics is unnecessary for alarm behaviours to occur and that the nature of the behaviours are independent of shoal size.

Pearl dace, Semotilus margarita (Pisces, Cyprinidae), respond behaviorally and physiologically to conspecific alarm substance. The behavioral alarm reaction was biphasic. A brief initial phase of rapid and unpredictable swimming was followed by a period of inactivity that was observable even after 5 h. In nature, biphasic behavioral reactions may function to remove alarmed fish from the area of greatest danger and then render them inconspicuous. The physiological alarm reaction included elements of a stress response. Concentrations of plasma cortisol and glucose were elevated at 15 min after the detection of alarm substance, but had returned to control levels by 5 h. There were no odor-induced changes observed in brain concentrations of dopamine, norepinephrine, 5-hydroxytryptamine, or tryptophan. The biphasic behavioral response and the physiological stress response were interpreted as adaptations that permit pearl dace to successfully react to threats of predation.