There are many anecdotal references in pet fish and industry literature as well as scientific experiments pointing up the presence of chemical messengers enhancing/stimulating fight/flight reactions of fishes. The sight of another individual can be proved to effect a fish's metabolism/behavior, but even a very small amount of material from another member of the same species or just water from their system can elicit an extraordinary overt reaction on another individual. As keepers of aquatic life these elements of chemical communication are extremely important to our success, especially in the transport of same-species schools in small volumes of water.
Likened unto the "alarm calls" of birds, zebras and monkeys, alarm substances of fishes are ecto- or social hormones (pheromones) communicating the presence of danger to conspecifics (members of the same <or closely related> species). Pandey retells a history of as Von Frisch (1938) describing his accidental discovery of a fright reaction in a European minnow Phoxinus phoxinus . He noted that on introduction of an injured fish the school dispersed, retreated and hid. Von Frisch further discerned that this fright substance (schreckstoff) was deriveable only from the fish's skin extract. Subsequently, other investigators have been busy studying the distribution of alarm substances and fright reactions in fishes. These inquiries have mainly centered on the important pet fish SuperOrder Ostariophyi (minnows <barbs, danios, rasboras, goldfish, koi, "sharks">, tetras, catfishes, many knifefish families (the suborder Gymnotoidei)), and two of the important food-fish orders Gonorhynchiformes (milkfish), and Salmoniformes (salmons and trouts, of course). These pheromonal behavior effects have now been demonstrated in virtually all living orders of fishes.
Alarm substances have been isolated from excretions, body slime and most-celebratedly, in large, specialized goblet or club cells (Kolbenzellen) located in the skin of fishes in the principal groups listed above. These cells do not open to the surface, but rather release their contents after mechanical injury.
An elaboration of the adaptive significance of alarm pheromone systems has been elucidated by Pfeiffer (1977). Some possible advantages are: 1) alerting members of the same species (conspecifics) to the presence of a predator, or away from a hunting spot, 2) reduction of cannibalism of young; that is, a strong fright reaction elicited by eating young of it's own species.
These chemicals are secreted or otherwise released into the environment where they alter the behavior and/or physiology of other's by ingestion, absorption, taste (not just through the mouth), or smell.
Other Related Minutiae:
There is the opposite of fright contagions released by the same species as well, id est chemicals that incite gathering, such as group feeding behavior, and as they say on T.V., but wait, there's more...
Chemicals from other species may also elicit a fright reaction. The classic bear sticking a paw in upstream water and scaring away salmon is an okay example. My favorite, though I lifted from a Wall Street Journal article, concerns the Prez, George Bush and his lack of fishing success. Seems like George is of a type of human that secretes a large amount of the amino acid L-serine which is very repugnant to fish. The head of state would do well to bring along a professional bass fisherperson who secretes very little L-serine, as baiting his own hook assuredly sends fish scattering.
Alarm Substances probably represent the most important aspects of chemical communication in fishes. Their presence and effects in crowded closed systems with inadequate filtration almost certainly leads to tremendous negative stress and "anomalous" loss of growth, reproduction, vitality, even life. These compounds are found in and have been demonstrated to have significant behavioral and physiological consequences in the whole "evolutionary" range of fishes, lampreys, sharks to flatfishes (pleuronectids). The home hobbyist,industry professional, and large-scale aquaculturist would do well to engineer their facilities and/or arrange maintenance procedures to minimize negative effects of secretory, excretory and damage-source alarm substances.
Where I Got This Stuff; & You Can Too:
Backhaus, Dieter. 1960. More findings about fish behavior, Alarm Substances among Fishes. Trop. Fish Hobbyist 11/60
Bryant, P.B. 1987. A study of the alarm system in selected fishes of northern Mississippi. 86 pp. Diss. Abst. Int. Pt. B -Sci. & Eng.; vol. 47, no. 7.
Fenner, Robert & Candace D.V.M. 1986. The Function of Body Slimes in Fishes, FAMA 6/86.
Jacobsen , J. & Johnsen, G.H. The influence of alarm substance on feeding in Zebrafish (B. rerio). J. Lepid. Soc. 82(4), 325-327, 1989.
Jaismal, S.K. & Waghray, S. Quantification of defense reactions of a cichlid fish, Oreochromis mossambicus (Peters) Trewavas, in response to warning chemicals. Indian J. Anim. Sci. 60(9) 1990.
Jonsson, L. 1979. Chemical Stimuli: Role in the Behavior of Fishes. Environmental Physiology of Fish.; Plenum Press: New York, N.Y. pp. 353-363
Pandey, A.K. Chemical signals in fishes: Theory and application. Acta Hydrochim. Hydrobiol.; vol. 12, no. 5. pp. 463-478; 1984.
Pfeiffer, W. The distribution of fright reaction and alarm sub stance cells in fishes. Copeia 1977, pp. 653-665.
Sterba, James P. 1990. Something About George Bush Just Sends Fish Reeling; Lure Maker Finds That Fish Like Schools but Loathe The Education President. Wall Street Journal, Sept. 17, 1990, p. 1.