Visual Communication

The importance of vision for moving, feeding, and communicating in the aqueous environment has resulted in well-developed eyes in many fish species (Walls, 1942; Yager, 1968; Ingle, 1971; Munz, 1971; Tomita, 1971). Color patterns and shapes may have evolved for environmental needs such as camouflage or locomotion rather than as signaling devices. For instance, Barlow (1974) generalized that cichlids living in clearer waters tended to be blue or green, while those in more turbid waters showed more yellow, orange, or red. Teleosts typically reflect the vertically oriented light regime of natural waters by being dark above and light below (countershading). Even brightly colored reef species are countershaded. At close range conspicuous markings render these fishes highly visible; but some colors, especially yellows, appear muted at greater distances. Thus, countershading may still be functional in forms such as butterfly fishes (Hamilton and Peterman, 1971). The cryptism of countershading can be forsaken for communicative needs; some cichlids display reverse countercoloring during the mating season, making them more conspicuous to potential partners (Albrecht, 1962; Barlow, 1974).

FUNCTION

Sexual Reproduction

Spawning is a major event in the life of a fish. Many species undergo extensive migrations to areas favorable to their young (Harden Jones, 1968). Unique visual signals have evolved in many cases to insure successful mating. Schooling species (e.g., many freshwater cyprinids) often swim to the bottom and stake out territories for reproduction. The details of spawning behavior in the 20,000 species of fish (Cohen, 1970) are largely unknown (Breder and Rosen, 1966); and the variety of social and sexual behaviors can be impressive even within single families, i.e., the pomacentrids (Reese, 1964; Albrecht, 1969; Fishelson, 1970; Swerdloff, 1970; Russell, 1971; Keenleyside, 1972a; Brown et al., 1973; Robertson, 1973).

Barlow (1970) has divided the functions of courtship into arousal and appeasement. According to the arousal hypothesis, courtship behavior stimulates the reproductive physiology of the animal to whom the behavior is directed. The appeasement role of courtship reduces the probability of attack by the mate. Both of these functions are minimized in fishes that spawn in the water column away from the bottom. Such fishes are nonterritorial and are often in groups that deny individuals the time and isolation necessary to develop a complex spawning ritual. Pelagic species are not typically known to form pairs of aggregations for extensive periods. In fact, many pelagic species have mass spawnings in which eggs and sperm are shed together. The cue for release of gametes may often be rapid swimming of the group. Here physical factors must control oogenesis and the female does not receive prolonged stimulation by a male. By nature, pelagic fishes are difficult to observe and many do not adapt readily to captivity. Our knowledge of their behavior is minimal, and generalizations are suspect because in at least certain instances transitory pairing and courtship behavior are known. Perhaps extended courtship is replaced with the self-stimulation of rapid swimming, and in some cases the continued attendance of several males stimulates the females (tactile). Magnuson and Prescott (1966) have described courtship behavior of captive Pacific bonito (Sarda chiliensis). The male appeared unable to identify the sex of a conspecific except by behavioral characteristics. A female would wobble, and one or more males would follow. If only one male was following, there would be a gradual transition from wobbling to circle swimming. The wobbles would become more and more pronounced, until the female continued one phase of the wobble into a circular path instead of turning back to the next phase of the wobble. Gametes were released in circular swimming. If two or more males followed a female, the males became involved in lateral threat displays inhibiting courtship. Brawn (1961a) also demonstrated pairing with courtship and agonistic behavior in the cod (Gadus morhua), another schooling species.

Because of internal fertilization with claspers, it is necessary for sharks and rays to form pairs. Information on reproduction in sharks has been garnered from fisheries studies and indirect evidence, rather than from extensive observation (Springer, 1967). Sharks often assemble in unisexual groups of about the same size. Spawning migrations bring these groups together for reproduction. During courtship male sharks may not feed, but females do. Courtship may thus become a hazardous activity for the males of some of the larger species, for since they are inhibited from making strong attack, the females they are courting may sometimes respond by killing them. Among the large carcharhinid sharks, the female is aroused to mate by rough courtship of the male, who uses his teeth to slash the skin on the female's back.

Ethologists have devoted much more time to the study of freshwater and near-shore marine fish associated with the bottom than to typical pelagic species or deepwater forms. Spawning on the bottom typically involves claiming a territory, preparing or maintaining a nest, courting, laying and fertilizing eggs, and some guarding of the eggs. Cichlids and scattered species in other families form pair bonds for reproduction and care of eggs, but in many families the female either leaves the nest or is driven away from it by the male, who assumes parental duties. The amount of the male's courtship is highly variable. Three examples, a minnow, a nandid, and a stickleback, have been chosen to demonstrate the range in behavior.

Males of the fathead minnow (Pimephales promelas) are content merely to defend their territories, and it is the female who initiates spawning (McMillan and Smith, 1974). The male performs no displays outside of his territorial object. The female herself is weakly territorial and probably attracted by territorial objects. The bright banding of a male, along with his vigorous movements within the territory, are also stimuli attracting the female. The male is aggressive toward the female, butting, charging, and chasing. A female ready to spawn would try to remain close to the undersurface of the male's territorial object, despite his attacks. When the female succeeded in positioning herself laterally close to the male, butting would typically stop and spawning vibrations begin.

Badis badis provides an intermediate example (Barlow, 1962). Here, too, the female normally seeks out the male, but an unpaired male may look for a female and stimulate her to follow him back to his burrow. The male usually attacks the female when she first enters the burrow. The attack is diminished by two complementary actions of the female (appeasement). She remains nearly motionless and leans against the male with tautly spread median fins. The fishes start carouseling and butting before the complicated act of enfolding, in which the fishes wrap around each other and the female pulsates. Enfolding appears to determine the moment of ovulation in the female (arousal). The early enfoldings serve to marshal the eggs for subsequent enfoldings. Eggs fall from the pair as they disengage. While this complicated act is necessary to stimulate the female, it is not certain how important it is to the male. In general males are assumed to be able to spawn with less coordination than is required for females. For example, males of the colonially nesting sunfish (Lepomis megalotis) leave their nests for no more than two or three seconds to intrude on a neighbor's nest and try to fertilize freshly deposited eggs (Keenleyside, 1972b). It is interesting to contrast the complicated spawning behavior of Badis with poeceleiids. Liley (1966) felt that signals at the start of courtship activities show greater species specificity and are more divergent than the signals and responses which occur later in the sequence.

In the final example the male stickleback, Gasterosteus aculeatus, actively courts the female (Tinbergen, 1951). The female appears in the territory of a nesting male and approaches in reaction to the male's zigzag dance. The male leads to the nest, and the female follows. The male then shows the nest entrance, the female enters, the male quivers on the female's tail region, and the ritual terminates with spawning and fertilization. The male stickleback is exceptionally aggressive and undergoes an approach-avoidance conflict as the female approaches his nest. In a variant of the normal courtship the male uses a pattern known as dorsal pricking, a somewhat jerky pushing of the female with the dorsal side (Wilz, 1970). This behavior pushes the female away when she would normally be following the male to the nest. By inducing the female to cease following, the pricking display functions to facilitate the male's switch from high aggression to greater sexual motivation. According to Barlow's appeasement hypothesis, the male in this case actually affects a behavior that serves to appease him.

The primary function of courtship behavior is to bring males and females of the same species together and to avoid wasting gametes. Sunfish are often known to hybridize, and Keenleyside (1967) and Steele and Keenleyside (1971) have experimentally studied species recognition in Lepomis megalotis and L. gibbosus, the longear and the pumpkinseed. Longears nest in crowded colonies where they are constantly contending with intruders, leaving them little or no time for active discrimination between the two species. In an experimental chamber, male longears did not distinguish between the two species, though females did. Male pumpkinseeds nest individually, and consequently they could discriminate between approaching females before they reached the nest. Nesting males were shown to court conspecific females preferentially. Since both female longears and male pumpkinseeds chose conspecifics under experimental conditions, Steele and Keenleyside doubted that they hybridize. Rather, they assumed that hybridization occurs between female pumpkinseeds and male longears.

Territories and Dominance

Two primary functions of territory are to insure reproduction and survival of the young. Feeding and safety may provide proximal benefits, particularly for species in which the male assumes parental tasks. For instance, the male garibaldi (Hypsypops rubicunda) cares for and defends its nest site all year in order to assure it for himself during the breeding season (Clark, 1970). The male maintains his territory for many years, and the red algal nest he cultures requires two or more years to develop. To sustain his time and effort, he locates his nests in areas with abundant food and shelter. Clark assumed that the male's requirement for reproductive success must be important enough to offset the resulting disadvantages to females, which are forced to live in poorer areas and which probably have higher mortality rates.

Fishes hold territories for varying amounts of time (Reese, 1973), with a range in the pomacentrids of twenty to thirty minutes in Chromis multilineata (Myrberg et al., 1967) to years in Hypsypops rubicunda (Clark, 1970). Along with extremes in duration, there can be tremendous variation in social structure and use of territories even within a species. Typically, male Chromis multilineata leave their school or aggregation to defend spawning territories for short periods (Myrberg et al., 1967). Other males were chased from the territories, but females were led to the center if they resisted the males' initial chase by "holding ground." However, occasionally two members of the aggregation would actually pair in the water column and swim rapidly together onto a small undefended portion of the substrate to spawn. Another damselfish, Dascyllus aruanus, forms either pairs or harems of one male and two to five females (Fricke and Holzberg, 1974). The females form a linear size-dependent dominance hierarchy, and spawning takes place in order according to rank. The wrasses (Labridae) often undergo sex changes (Reinboth, 1972, 1973) that lead to interesting behavioral situations. Robertson (1972) studied harems of Labroides dimidiatus consisting of one territorial male, three to six mature females, and several immature individuals. Females are territorial and also have a size-related linear dominance hierarchy. When the male dies the α-female reverses her sex and starts showing the male aggressive display toward the other females within one and a half to two hours. The behavioral changeover can be completed within a few days, and sperm can be released fourteen to eighteen days after the start of the reversal. Robertson believes that dominance suppresses sex reversal and that release from aggressive signals is the hormonal trigger.

A fish is dominant in its territory, but it may find occasions for leaving it. Males of Hypsypops rubicunda court and pair in aggregations above their territories where aggressiveness is reduced. Spawning occurs on the male's territory (Clark, 1971). Various species of cichlids leave their territories without being vigorously attacked. They rise to the surface, turn pale, and depress their fins when swimming through neighboring territories (Baerends and Baerends-van Roon, 1950).

Not all aspects of dominance interactions are innate; experience may play a part. Habituation is involved in the termination of hostilities between territorial neighbors of the stickleback Gasterosteus aculeatus and of the convict cichlid Cichlasoma nigrofasciatum (van den Assem and van der Molen, 1969; Peeke et al., 1971) and most likely plays a similar role in all territorial species. The status of a green sunfish (Lepomis cyanellus) is determined, at least in part, by its experiences (McDonald et al., 1968). Similarly sized dominant-subordinate pairs of sunfish were broken up and manipulated so that dominant fishes were placed with larger conspecifics and subordinates with smaller ones. After five days of treatment the original pairs were reunited, and in fifteen of the twenty pairs the original dominance status, as indicated by coloration, was reversed.

Whether fish can recognize individual conspecifics is a question basic to the study of hierarchies. Although this question has not been answered thoroughly, there are apparently fish species whose members can recognize individual conspecifics and other species whose members lack this ability. Jenkins (1969) found evidence of nip-right relationships in salmonids, whereas Myrberg (1972c) found nip-dominance in pomacentrids. Although Jenkins's trout did form stabilized relationships among confined fish, the persistent occurrence of revolts, even in relatively peaceful groups with large size ranges, and the exceptional cases of rank change and social mobility suggested to him the presence of either a definite limit to the effectiveness of learning or an irreducible social instability in groups of stream resident trout. Another possibility, consistent with his suggestions, might be that individuals recognize each other imperfectly, if at all, so that the outcome of an interaction would depend on the prior successes (wins and defeats) of the fishes, coupled with their internal motivation at the moment. Also, working with damselfish in the wild, Myrberg (1972c) found that smaller fish challenged and even occasionally chased larger members of the colony. This again argues against the individual recognition one would assume in a nip-right hierarchy. However, Nelson (1964) seems to have established the occurrence of individual recognition in dominance hierarchies and pairs for some glandulocaudine fishes. Recently Fricke (1973) claims to have demonstrated individual recognition between members of breeding pairs of the anemone fish (Amphiprion bicinctus) and within groups of Dascyllus aruanus (Fricke and Holzberg, 1974). Catfish recognize individuals by chemical means (Bardach and Todd, 1970).

Fish territories, at least in some pomacentrids, appear to vary, depending on the species of intruding fish (Clark, 1970). Hypsypops rubicunda attacked bottom-grazing fishes frequently or almost every time. Predatory fishes were generally tolerated and plankton-eating forms were nearly ignored altogether. The number of interspecific attacks was higher when males were guarding eggs than at any other time. Pomacentrus flavicauda directed only one-fifteenth of its agonistic responses toward conspecifics (Low 1971); responses were typically directed toward competitors for algae rather than toward carnivores. Myrberg and Thresher (1974) attacked this problem experimentally by presenting various species held in jars to a territorial Eupomacentrus planifrons to see how close they could place the captives before eliciting attack. The maximum distance of attack was different for each species, even though individuals of various sizes were involved. Wrasses (male and female Halichoeres garnoti) were reacted to in a similar manner, even though their color patterns are highly divergent. The authors suggested that forms, rather than color pattern, allowed species discrimination, assuming of course that vision is important. They implied that the damselfish reacted to male and female wrasses as members of the same species. An alternate explanation is that the fish discriminated the two sexes as different entities but of equal threat potential, in which case color might still be important.

Cleaning Symbiosis

Cleaning has evolved in tropical and temperate freshwater and marine fishes from many families. In view of the polyphyletic origins and diverse situations involved, it is not surprising that early thoughts about cleaning (Limbaugh, 1961; Feder, 1966) have been questioned (Hobson, 1969). Cleaners are generally supposed to eat parasites, fungi, and necrotic tissue from their hosts; thus, cleaning is a case of mutualism with obvious benefits to both species. In an experimental demonstration Limbaugh (1961) removed all the known cleaning organisms from two small Bahamian reefs and found that within days the number of fishes on the reefs was drastically reduced. Many of the remaining fishes developed fuzzy white blotches, swellings, ulcerated sores, and frayed fins. Since Hypsypops rubicunda does not allow cleaners to enter its territory while it is guarding eggs, it provides a natural experiment. In southern California Hobson (1971) found an average of sixty-seven parasitic copepods (Caligus hubsoni) on egg-guarding males as opposed to four to eight on males outside the reproductive season. However, Youngbluth (1968) and Losey (1972) removed Labroides phthirophagus, an obligate cleaning wrasse, from patch reefs in Hawaii and found no dramatic effect on either the numbers of remaining fishes or on their health. Losey (1972) noted that Labroides eat not only ectoparasites but scales and associated dermal and epidermal tissues or mucus as well; thus, he suggested, the relationship could be considered parasitic or commensal. Perhaps these examples should not be thought of as conflicting but merely as reflecting various systems in which cleaning plays a role.

The process of bringing two species together for cleaning involves a highly developed form of communication. The fact that many different species allow themselves to be cleaned makes the process even more interesting. Losey (1971) listed the interactions involved in cleaning as follows:

Cleaner fish

(1) Inspect, swimming close to the host and apparently exploring its body surface;

(2) Clean, feeding on matter on the body surface of the host;

(3) Dance, a dorso-ventral oscillation of the body.

Host fish

(1) Pose, the position or orientation of the host's body and fins and the swimming movements frequently assumed during cleaning interactions;

(2) Body jerk, any short quick movement of the body or head;

(3) Attack, darting quickly toward the cleaner.

We will briefly highlight some of the factors involved in communication in Labroides phthirophagus, an obligate cleaner (Youngbluth, 1968; Losey, 1971), and Oxyjulis californica, a facultative cleaner (Hobson, 1971). Labroides maintains a cleaning station, the position of which is learned by the host fish. Either to pose or to inspect may be the initial action in a cleaning sequence. The wrasse recognizes individual species and has definite preferences. Large jacks and parrotfish may be pursued by the cleaner as they appear near his station, while smaller wrasses, butterfly fish, or damselfish may pose continually and be ignored by the cleaner. The senorita (Oxyjulis californica) maintains no station and only some individuals seem to clean. In addition, individuals appear to specialize, in that some clean the damselfish Chromis punctipinnis and others the topsmelt (Atherinops affinis). Since Oxyjulis obtains much of its food without cleaning, the average encounter with a host fish will not result in cleaning, and the hosts do not bother to pose until a cleaner initiates the activity.

The cichlid Etroplus suratensis has a more ritualized pose display than most other species (Wyman and Ward, 1972). While performing the head-up display, it rapidly flickers its dark pelvic fins while simultaneously quivering its entire body. This display may be necessary because it approaches the cleaner (Etroplus maculatus) in its territory; failure to emit the submissive display results in attack.

One of the most important behavioral questions about cleaning concerns mutual recognition. Fricke (1966) and Losey (1971) elicited posing by a presentation of cleaner models. Losey claimed that hosts show a graded response to increasingly real models of Labroides phthirohe responses of approach and attraction based solely on visual cues have also been reported for a sphagus, but his results were not complete enough to explain the relative effectiveness of different elements of coloration and body shape. The conspicuous coloration of many cleaners prompted Eibl-Eibesfeldt (1955) to hypothesize a guild mark theory (occurrence of some similarities in color pattern). Recently, Ayling and Grace (1971) and Potts (1973) have described guild marks in cleaners. Hobson (1969) felt that evidence supporting a direct relationship between bright coloration and cleaning is still weak. Further, distinctive coloration will render a cleaner more conspicuous to predators, and Hobson (1971) doubted that cleaners are immune from prédation during noncleaning situations. Darcy et al. (1974) experimentally demonstrated that gobies, which clean piscivores, will not usually be eaten by them, while a cleaning wrasse that does not normally service these fishes will be eaten.

Guild coloration or not, experience is probably a factor in the recognition of a cleaner by the many hosts involved. By presenting a moving model of a cleaner fish for positive reinforcement, Losey and Margules (1974) operantly conditioned a butterfly fish Chaetodon auriga to occlude a light beam. They believed that the reward is probably tactile stimulation of the model. In nature, hosts might learn to recognize a new cleaner as a source of tactile stimulation.

Schooling

The configurations or qualitative characteristics of schools of different species of fish are manifested in highly varying degrees of cohesion and polarization. Presumably for the initiation and maintenance of the school, each individual would be able to provide the appropriate stimuli (visual, olfactory, acoustic, etc.) as well as to perceive and respond to stimuli arising from conspecifics. Other factors determining the formation and the quality of a school are the life stages of the organism as well as its relationship with other environmental input, such as solar and lunar cycles. No definition of a school will be entirely satisfactory because of the varied forms that mutual attraction takes. However, for the purposes of this chapter a working definition of a school will be "an aggregation formed when one fish reacts to one or more other fish by staying near them" (Keenleyside, 1955), with further emphasis on the biosocial mutual attraction between individual fish (Shaw, 1970).

In spite of their movement and size schools often act as whole units, frequently exhibiting what appears to be instantaneous and synchronous reactions to a variety of stimuli. Within a school, specific stimuli or "messages" might convey information about attraction, changes in speed, cohesion, direction, presence of food, or approach of predators. To understand the mode of transmission and rapid integration of the information throughout an entire grouping of fish, it is necessary to identify these stimuli and determine how they are interpreted via the sensory components.

Vision appears to be one of the more important senses utilized by schooling species (for review and discussion see Morrow, 1948; Atz, 1953; Breder, 1959; Shaw, 1970; and Radakov, 1973). The dominant role played by vision was established in a study by Keenleyside (1955) in which purely visual cues were effective in stimulating the initial phases of intraspecific approach and attraction among stickleback, Gasterosteus aculeatus, roache, Leuciscus rutilis, rudds Scardinius erythrophthalmus, and the characid Pristella riddlei. For the characid, a conspicuous black patch on the dorsal fin was found to be an important visual signal used in the recognition and attraction of species mates. The responses of approach and attraction based solely on visual cues have also been reported for a sea catfish (Plotosus anguillaris: Sato, 1937), the atherinids (Menidia menidia and M. beryllina: Shaw, 1960), the mullets Mugil cephalus (Olla and Samet, 1974) and M. chelo, a roach (Rutilus rutilns), a pomacentrid (Chromis chromis: Hemmings, 1966), and a jack (Caranx hippos: Shaw, 1969). When the tuna Euthynnus affinis were separated by transparent barriers, and hence could rely only on visual communication, the long-term maintenance of attraction and the schooling tendency persisted for as long as nine days (Cahn, 1972). Topp (1970) has implicated a dramatic color change and an associated behavior in maintaining the integrity of schools of rudderfish (Kyphosus elegans).

Besides its function in the attraction response, vision appears to be important in the communication of changes in movements of the school. In studies on tuna (E. affinis: Cahn, 1972) and jacks (C. hippos: Shaw,1969), fish separated by transparent barriers tended to turn simultaneously and react mutually to other positional changes. Hunter (1969) studied the communication of velocity changes among schooling jack mackerel (Trachurus symmetricus) by electrically stimulating one fish and observing the responses of others in the school. He suggested that the latency and the velocity of responses among the group of fish reacting to the stimulus fish depended upon their visual perception of movement although the roles of other sensory systems were not analyzed.

Despite the inaccessibility of many species that exhibit schooling, experimental work on selected groups has provided a basis for biologists to assess the value of schooling. Theoretically it would appear that the most beneficial and adaptive characteristic is the increased or more efficient performance of certain activities once the fish have formed into a schooling group. A recent study on social facilitation in feeding behavior (Olla and Samet, 1974) established that the initiation of feeding behavior of single mullet (Mugil cephalus) was greatly facilitated solely by visual cues arising from a feeding group. Additional studies (Welty, 1934; Uematsu, 1971) have also established the role of the group with respect to feeding facilitation. Other adaptive features of the school may be related to the facilitation of reproductive behaviors, lowered predation (Seghers, 1974), improvement in orientation during a migration, and energy conservation resulting from the hydrodynamic advantage of swimming behind other moving fish.

The existence of heterotypic aggregations or schools, such as Breder (1959), Collette and Talbot (1972), Ehrlich and Ehrlich (1973), and Hobson (1974) observed, leads to further questions about the degree to which communicative processes are used in social groupings of fish. If one is to rely on Shaw's (1970) analysis of schooling in which the critical factor is the biosocial mutual attraction among individuals, then these observed heterotypic groups may represent only the formation of temporary associations that may be highly adaptive in the natural environment, but in which mutual communication between species may also be less well developed or absent. For example, Collette and Talbot (1972) observed the transient appearance of several different species within the body of a resident school of bonnetmouths (Inermia vittata) and suggested "that a resident school can itself become a habitat for other schooling species." Breder (1959) suggested that heterotypic groupings of goatfishes and various gerrids may be viewed as feeding associations in which gerrids catch food particles missed by goatfishes as the latter feed along the bottom.

Other adaptive features of these groups may be that resident schools provide camouflages for certain species and hence serve as a predator defense mechanism. Since the formation of these groups may be based possibly on a mutual toleration between species, rather than on a mutual attraction, further investigations of the biological causal factors are needed to fully understand the formation and maintenance of these “schools.”

Parent-Young Interactions

Except for the cichlids, information on parent-young interactions comes from a diffuse and incidental literature (Breder and Rosen, 1966), and even the cichlid literature focuses largely on behavior of adults (Noakes and Barlow, 1973). Many of the species that care for their young have evolved mouth breeding habits (Oppenheimer, 1970). Further work is needed on a wide variety of species from different families.

Parent cichlids appear to recognize their offspring by a combination of visual and chemical cues, although some species may accept young of other broods or even other species (Kühme, 1963; Myrberg, 1966; Noakes and Barlow, 1973; Barlow, 1974).

Parent cichlids of many species shepherd their young for a while and can alert them to danger (Noble and Curtis, 1939; Baerends and Baerends-van Roon, 1950; Künzer,1962). Baerends and Baerends-van Roon's work on alerting in Tilapia natatensis will be described. The response of the young is released by disturbance of the water and not by visual stimuli; but once aroused, the young direct themselves toward the female by means of visual clues. Slow movement of the mother elicited following by the young, but violent movement accompanied by a color change to black caused them to scatter and head for the bottom. As danger became less imminent, the mother would retreat toward the young and assume a diagonal position, her longitudinal axis making an angle of about 10° or 20° below the horizontal. In this position she moved slowly backward and the young swarmed toward her head and into her mouth. The young were attracted to the trailing edge of the underside of disc models as they were moved away. Young would also be attracted to dark spots on a model, but would then wander along the surface of the disc looking for the mouth.

SIGNAL VARIATION AND CODING

Color Patterns

Coloration in fishes essentially represents a balance between those factors which maximize signal value for communication and those which function to make an animal less conspicuous, i.e., cryptic coloration for both predators and prey. The wide array of colors found in fishes suggests a use of visual signaling. Some fishes (notably cyprinids) are known to have good color vision (Beauchamp and Lovasik, 1973; Daw, 1973), and perhaps most teleosts have it to varying degrees; but most species have not been checked for the presence of cones. Although much work has been done on the possible uses of color in communication, few studies make the important distinction, either in the experimental design or in the interpretation of data, between color and brightness. It is always possible in some circumstances that varied color merely forms a pattern of contrasting brightness.

In their classic work on cichlids, Baerends and Baerends-van Roon (1950) designed experiments that included the consideration of discrimination based on color versus brightness. As an attempt to identify the visual components that were involved in the attraction of young Aequidens latifrons and Cichlasoma bimaculatum to adults, experiments were performed using discs of several different colors as well as grays of varying brightness. Both species demonstrated a preference for colors of shorter wave length with little discrimination of brightness. The authors concluded that color was an important releasing stimulus for following in these two species. However, in similar experiments on Tilapia natalensis attraction of young fish to the female was based perhaps on discrimination of brightness rather than of color.

Responses to brightness seem also to be involved in the experiments of Picciolo (1964), who found that the blue color pattern (contrasting brightness) displayed on the throat and breast of the male gourami Cotisa lalia functions as a visual stimulus for sexual discrimination for both sexes. Model experiments demonstrated that the color pattern need not be blue but merely dark in order to attract and release aggressive responses in males.

Whether fish are responding to brightness or to color per se, the presence of the color enhances the communication of signals by vision. The examples cited below, although generally not considering the difference between brightness and color, show the communicative function of color even if it only enhances contrast.

The shallow inshore marine environment of the subtropical and tropical zones contains many colorful types. Lorenz (1966) hypothesized that bright, poster-colored fishes were more aggressive toward conspecifics than the more modestly colored species. More recent studies, however, demonstrate that drably colored fishes may defend territories and that territories are defended against other species (Rasa, 1969; Clark, 1970; Low, 1971; Myrberg and Thresher, 1974; Tavolga, 1974).

The degree of conspicuousness of coloration in certain species has been shown to be important in communication. For example, Haskins et al. (in Liley, 1966) found that male guppies (Poecilia reticulata) that were conspicuously marked tended to have greater mating success when in competition with other males not so brightly marked. In another study Keenleyside (1971) found that the conspicuous opercula patch, black eye with light red iris patches, and black pelvic fins of Lepomis megalotis were the features that appeared to be most important in eliciting aggression from nest-guarding males.

An important aspect of the role that coloration plays in communication is its location on a fish. For example, the blue color pattern of the male Colisa lalia must be located on the anterior ventro-lateral surface in order to function as a cue for sex discrimination (Picciolo, 1964). When the color pattern was placed on the center of the lateral surface of a model, it did not particularly attract males or females. A model Lepomis megalotis with the eye and opercular patch placed near the tail was relatively ineffective in eliciting aggression although any aggression that was elicited was directed at the abnormally located eye and patch (Keenleyside, 1971). In a number of African mouth breeders of the genus Haplochromis, the males bear yellow or orange spots that look much like eggs (Fig. 1) on the anal fin (Wickler, 1962). The female lays her eggs in the male's pit, but takes them in her mouth before he can fertilize them. The male then spreads his anal fin and fertilizes the empty pit. The female grasps the egg dummies on his anal fin with her lips, engulfing the sperm ejected by the male. Another example of position of color is found in the blue-throated darters, a subgenus within the genus Etheostoma. These fish raise their heads when fighting, exposing their blue throats. In contrast, other members of the genus lack the blue throat and fight with their heads in the normal position (Winn, unpublished).

Fish coloration may be controlled by both the central nervous and the endocrine systems (Bagnara and Hadley, 1969; Fujii and Novales, 1969; Gentle, 1970a, 1970b). Many fishes take on breeding coloration during the spawning period. The classic example is the male stickleback Gasterosteus aculeatus, whose eye turns blue and whose belly becomes red (Tinbergen, 1951). Fishes may also exhibit transient, fairly rapid changes in color to communicate information about other behavior states. In Pomacentrus jenkinsi aggressive motivation is indicated by the darkening of the normally yellow eye to gray (Rasa, 1969). The amount of red color in the iris of Lepomis megalotis is related to dominance rank order (Hadley, quoted in Keenleyside, 1971). Nakamura and Magnuson (1965) observed the transient appearance of black spots ventral to the pectoral fins, faint vertical bars on the flanks, and a yellowish middorsal stripe in the tuna Euthynnus affinis during feeding. They interpreted this display as a possible social releaser signaling the presence of food to other members of the school. In further work with the related Pacific bonito (Sarda chiliensis), Magnuson and Prescott (1966) decided that this feeding display was agonistic. In the leaf fish (Polycentrus schomburgkii) a severely dominated male loses its dark coloration, becoming yellow-white, with the top of the head brown, a color pattern corresponding closely to that of a female ready to spawn (Barlow, 1967).

Leong (1969) performed a series of experiments on the cichlid Haplochromis burtoni, which exhibits rapid transient changes in coloration. In this study Leong was able to show the extreme specificity of visual signaling as well as indications of how the signal is being processed by the receiving fish. Only two components of the male's territorial coloration were found to affect attack readiness: the black vertical component of the head pattern and the orange patch above the pectoral fin (Fig. 1). The vertical component alone increased the attack rate by 2.79 bites/ min. The orange patch alone decreased the attack rate by 1.77 bites/min., while a dummy with both colorations present increased the attack rate by 1.08 bites/min., i.e., the sum of the effect of both components if presented separately (2.79 minus 1.77). Fig. 2 shows some of the models and their effects. Leong explained this cumulative effect by the rule of heterogeneous summation, which is represented by a unimodal peak. Had there been two peaks, each coinciding with one of the two stimuli, the fish might have been responding to either the vertical bar or the orange patch rather than a composite.

Fig. 1. Different color patterns of the male cichlid Haplochromis burtoni: a. Juvenile; b. adult ready to establish territory; c. territorial male; d. spawning male; e. and f. fleeing male. Crosshatch is orange and black is black. (From Leong, 1969.)

These two patterns were observed by Leong to appear and disappear within seconds; thus, they most likely reflect the internal state of the animal. He hypothesized that a combination of these patterns promotes a balanced level of aggression in established colonies, while an intruding male displaying only the vertical bar would rapidly be attacked.

Information about the role of bioluminescence in fish communication is meager and in most cases purely speculative (McAllister, 1967; Nicol, 1967; Tett and Kelly, 1973). Photophore patterns that are often species-specific and in some cases gender-specific suggest a recognition function. Further, it appears that the light emitted may be flashed on or off, increasing the possibilities of signal complexity and specificity as well as enhancing a signal.

The same balance that exists in coloration and shading in fish exists with bioluminescence, with a compromise necessary between those aspects that may be selected for on the basis of signal function and those that act to lessen the probability of attack from enemies. In this vein Clarke (1963) suggested that the ventrally located photophores may help to prevent detection inasmuch as the light emitted ventrally will match incident light, rendering the animal less conspicuous than if the light were emitted from the dorsal part of the body.

We know of only two cases in which bioluminescence has been observed in actual communication. Crane (1965) injected a gravid female Porichthys notatus with adrenaline, causing her to luminesce and turn pale. He placed the female in a tank with a nesting male, who courted her for about an hour, producing light intermittently in five-to ten-second displays, grunting, nudging, and grabbing her in his jaws. The nocturnal reef fish Photoblepharon palpabratus has a large bacterial organ under each eye (Morin, et al., 1975), which produces intense light. The fish is capable of covering the organ and flashing its light on and off. These fish spend the day in caves and come out on dark nights when they may use their lights for attraction, leading to group formation of three to twenty-five species mates. In addition, the defense of territory by male-female pairs from conspecific intruders is correlated with light emissions. When intruding Photoblepharon approached, the female swam back and forth rapidly. She would then turn off her light, swim directly toward the intruder, and turn on the light when she was just next to the other fish. This was invariably effective in driving intruders away (Morin et al., 1975).

Fig. 2. Dummies and the effect of color patterns on the average increment in attack rate after their presentation. Vertical lines represent the variances of the mean. (From Leong, 1969.)

Movements

This section will be devoted to signal movements that are typically considered fixed action patterns. Barlow (1968) was dissatisfied with this term, feeling that it was applied uncritically to almost any behavior that has a degree of regularity sufficient to permit one to recognize it. He offered the term "modal action pattern" (MAP), feeling that it conveyed the essential features of the phenomenon without implying a degree of fixity that has seldom been tested.

Baerends and Baerends-van Roon (1950) divided the signal movements of various cichlid species into thirteen patterns: lateral displays, tail beating, tail fluttering, frontal display, mouth fighting, butting, swimming on the spot and tail wagging, attitude of inferiority, jerking and quivering, inviting, nipping of a substrate, skimming (pseudolaying and pseudofertilizing), calling the young, and jolting. Barlow (1974) recognized these categories and pointed out that the similarity between the different species is striking, although there are some statistical differences in the frequency of occurrence and sequencing of various MAPS and some patent but small differences in the MAPS used by the different species.

Frontal and lateral displays occur in agonistic encounters of many species. These displays are often bluffs, ritualized fighting, or the result of conflicting tendencies. In juvenile Atlantic salmon (Salmo salar) high attack tendency results in charging, nipping, and chasing, while fleeing is the result of high escape tendency (Keenleyside and Yamamoto, 1962). Frontal and lateral displays occur as a result of conflict between attack and escape with frontal display indicative of a relatively high level of attack tendency and lateral display indicative of escape tendency. The gray reef shark (Carcharhinus menisoirah) displayed toward divers with laterally exaggerated swimming and rolling along with spiral looping (Johnson and Nelson, 1973). The display occurred under approach-withdrawal conflict situations. It was elicited by rapid diver approach and was most intense when there was maximum escape-route restriction. More data are needed to determine objectively the underlying causes of the types of postures described above. Myrberg (1965) has interpreted some fin-spreading actions in cichlids as intention movements rather than as either fright- or attack-motivated behaviors.

Many fishes make themselves appear larger during agonistic displays by erecting fins and branchiostegals and by spreading the opercula. In most cichlids the intensity of a lateral display can be gauged by the extent of fin erection (Baerends and Baerends-van Roon, 1950), but even within the family there are exceptions. In the oscar (Astronotus ocellatus) full erection of the fins indicates a beaten or frightened fish. Rasa (1969) has separated fright and aggression in Pomacentrus jenkinsi on the statistical occurrence of external changes. Gray eyes and lowered dorsal fin are present in aggressive fish, while yellow eyes and raised dorsal fin are typical of frightened fish. Both of these conditions are components of agonistic interactions.

As with other aspects of visual communication, movements can be profoundly influenced by the environment. McKenzie and Keenleyside (1970) compared the reproductive behavior of the stickleback Pungitius pungitius from South Bay, Canada, with the behavior of the European form (Morris, 1958). The European form breeds in quiet, weedy streams, while the South Bay form breeds along rocky, barren, often turbulent lakeshores. South Bay males do not dance-jump when leading the female to the nest but, more like Gasteros tens males in open areas, swim directly and quickly to the nest after courtship jumping. Morris argues that jumping during the leading phase of courtship in the European form slows down the male's return to his nest, allowing the female to maintain visual contact as she follows him through the dense weeds.

Size

Generally, larger members of a species are dominant over smaller ones. For example, male Tilapia mossambica as little as 2 mm longer than conspecifics won aggressive encounters between them (Neil, 1964). However, size alone may not always be the only cue, but rather may act in concert with other signs to bring about a particular response pattern. Barlow (1970) demonstrated that visual perception of size was important in pair formation of the cichlid Etroptus maculatus. Males normally attack fish as large as or larger than themselves, making it difficult for a male to mate with a large female. Females generally attack fish that are smaller than themselves, making it difficult for a female to mate with a small male. Various chaetodontids (butterfly fish) show a great deal of intraspecific aggression toward equal-sized individuals, but little toward smaller or larger conspecifics (Zumpe, 1965). Hurley and Hartline (1974) showed that schools of Chromis cyanea escaped from larger geometric models at a greater distance than from smaller ones.

There have been cases in which factors other than size appear to be involved in dominance-hierarchical situations. DeBoer and Heuts (1973) contend that in Hemichromis bimaculatus stable dominance relationships need not be dependent on physical strength relationships between individuals but are probably determined by continuous stimulation from the environment and by mutual perception of each other's overt aggression and/or flight behavior. Keenleyside (1971) found the same level of agonistic response to small, medium, and large plywood models in the male longear sunfish (Lepomis megalotis). Myrberg and Thresher (1974) found that Eupomacentrus planifrons has a variable territory, the size of which depends on the particular species of intruder. But within the range tested, the size of the intruder does not affect the maximum distance of attack. Pomacentrus jenkinsi will attack any moving object, even as large as a human swimmer, that invades its territory (Rasa, 1969). As in other species, the possession of a territory by a small individual allows it to drive away larger nonresidents.

Shape

There is great diversity in body shapes of fishes, with the selection pressure for a particularconfiguration related to the environment. Fastmoving pelagic species have streamlined bodies selected mainly for hydrodynamic considerations, while the shapes of the more sedentary demersal species are selected for other adaptive reasons. Although it is clear that body shape evolved through selective pressures attuned to basic habitat and niche adaptation, we assume that shape is used in species recognition in many forms. Within a family of fish such as the surgeonfish (Fig. 3), there may be obvious differences in shapes between the different species (Barlow, 1974). However, it has not been demonstrated that fishes generally recognize members of their own and different species on the basis of form. For example, nest-guarding male longear sunfish (Lepomis megalotis) responded more aggressively to circular models than to models shaped like fishes (Keenleyside, 1971). Oval and triangular models were less effective, and the author suggested that "increasing the vertical dimension in relation to the horizontal enhances the stimulus value of the model." The results indicated, however, that there is a limit beyond which the vertical (or depth) feature of a model no longer serves as an effective stimulus in aggression.

Fig. 3. Surgeonfishes, illustrating the diversity in body shapes. (From Barlow, 1974.)

The shape and size of many fishes are sexually dimorphic (Breder and Rosen, 1966). Experimental investigation of those secondary sex characters has been limited. Extreme attention to shape has been demonstrated in gouramis through the use of models (Picciolo, 1964). In Trichogaster trichopterus and T. leeri the male's dorsal fin is longer than the female's, and the fin appears to function as a visual cue for sex discrimination. A female model of T. trichopterus that displayed a swollen abdomen had a strong attraction value for males and was capable of evoking following behavior from them. The male sword-tail (Xiphophorus hellerii) develops a blackedged, tapering spike as the lower third of the caudal fin. Hemens (1966) investigated the behavioral significance of this "sword-tail" and found that it is an important visual stimulus in releasing aggression in other males but is not significant in male-female interactions.

Chemical Communication

Since living organisms produce biochemical products, it is not surprising that primitive plants and animals developed a chemosensory capability (Kittredge et al., 1974). The chemical senses of fishes have recently been reviewed by Kleerekoper (1969), Bardach and Todd (1970), Hara (1971), and Bardach and Villars (1974). This literature demonstrates the fundamental importance of the chemical senses for feeding, orientation, and communication of fishes. In some species the chemical sense is a primary modality mediating the way in which the organism perceives and reacts to its environment, while in others it may be auxiliary, acting in conjunction with other senses or perhaps as a "priming" mechanism before vision or other modalities become operative.

In the area of communication, studies equivalent to the playback of sounds or the presentation of visual models have succeeded in experimentally establishing that purely chemical cues, derived from skin washings or extracts, gonadal fluids, or urine, can elicit a variety of responses, including species recognition, intraspecific attraction (serving as a factor in schooling and aggregating behaviors), fright reactions, agonism, and various reproductive activities. Although the specific composition of these chemicals is not understood or documented for fish, these substances are, nevertheless, quite prevalent throughout the animal kingdom (Marler and Hamilton, 1966). These chemical stimuli were first designated as pheromones by Karlson and Luscher (1959) and were defined as "substances which are secreted to the outside by an individual and received by a second individual of the same species in which they release a specific reaction, for example, a definite behavior or a developmental process."

For fish whose behavior is not governed predominantly by visual cues, such as catfish of the genus Ictalurus, chemoreception is an important sensory modality for social behavior, particularly in the recognition of other individuals. Todd, Atema, and Bardach (1967) successfully conditioned blinded yellow bullheads (Ictalurus natalis) to discriminate between the odors of individuals. Test fish with nares cauterized were unable to discriminate between odors. When water containing only dermal mucus washings of donors was used as a test substance, normal test fish were able to distinguish the odors, but to a lesser degree than when tank water in which donor fish were maintained was used, suggesting that some factor in the excretory products is the major source of the chemical stimuli. Richards (1974) found that a chemical substance in the urine of blinded brown bullheads (I. nebulosus) is important in individual conspecific recognition. In this study another important chemical factor involved in the recognition process was found in extracts of the urophysis, a component in the caudal neurosecretory system. Göz (1941) demonstrated that Phoxinus phoxinus could chemically discriminate conspecifics.

In addition to their role in individual recognition, chemical secretions may also be utilized in the discrimination of another fish's hierarchal status (Bardach and Todd, 1970). Pairs of yellow bullheads were held in tanks and allowed to fight until one of each pair became dominant. Then each pair was separated. When water from a tank in which the dominant fish resided was added to the tank of the fish originally paired with it, the formerly subordinate individual responded by avoiding the area where water from the dominant's tank was introduced. In the converse experiment, a dominant fish swam toward the area of inflow of water from a subordinate's tank and sometimes exhibited aggressive behaviors at that point. In another series of experiments, when a previously dominant fish was returned to its tank after having unsuccessful encounters with more dominant fish, its formerly subordinate tank mate attacked it as if it were of inferior status. These results suggested that an alteration (decrease) in status may have been chemically communicated. In a final series of tests, when low-ranking bullheads with their nares cauterized were returned to their former tanks, they were unable to recognize their tank mates and immediately attacked them. Whereas normally low-ranking fish occupied particular areas of a community tank, the cauterized fish swam throughout the tank and attacked dominant fish in their shelters. These results indicated that, among bullheads, chemical stimuli appeared to mediate individual and status discrimination among species mates as well as to maintain and enhance normal social structuring of a community.

For fish in a reproductive state pheromones may serve as signals for recognizing appropriate mates as well as for identifying intruders in an established nest site. In the blind goby (Typhlogobius californiensis) male-female pairs, living as commensals in burrows of the shrimp Callianassa affinis, appear to retain their status for life, probably ten or fifteen years (MacGinitie, 1939). The gobies were found to be sensitive to intrusions of other individuals of the same sex through chemical cues. After an invasion, a fight between the stranger and the resident of the same sex ensued until one of the antagonists was killed or driven from the burrow. The opposite sexed resident fish became passive and accepted the victor (either the former resident or a stranger) as a mate. This suggested that pair formation may be dependent more on gender-specific rather than on mate-specific pheromones in this species. MacGinitie was also able to determine experimentally that odors of a strange fish introduced into the burrow of an established pair would elicit aggressive behavior by the inhabitant corresponding to the sex of the donor.

Pheromones may also stimulate nuptial behaviors. In the gobiid fish Bathygobius soporator gravid females produce a chemical secretion that stimulates courtship behavior in males even in the absence of visual cues (Tavolga, 1956). This substance was found to elicit courting movements by males within five to ten seconds after its introduction, and it continued to stimulate the male's response for approximately one hour. Tavolga tested extracts from various tissues and determined that the pheromone was produced in the ovaries and that it was sensed by olfaction.

Male glandulocaudine characids have a caudal gland that may produce a substance which, when directed toward the female during "dusting," increases the probability that she will pair (Nelson, 1964a, 1964b). Interestingly, the members of the genus Glandulocauda that do not appear to have an intact gland produce croaking sounds, which may have taken over much of the gland's function. Courting males of the genus Hypsoblennius produce a pheromone that attracts other ripe, nonparental males (Losey, 1969). Although not fully explained, these results suggest that possibly on population level the male pheromone may act to facilitate and enhance the sexual receptivity of other males during the breeding season. In a similar study by Leiner (1930) it was found that mucus secreted by a male stickleback Gasterosteus aculeatus elicited courtship behavior among other conspecific males. Visual cues are primarily responsible for pairing in the blue gourami ( Trichogaster trichopterus: Cheal and Davis, 1974). Although chemical cues from a female of this species to some extent increases nest building in isolated males, the full expression of this response is mediated by a combination of chemical and visual stimuli. Similarly, in the angelfish (Pterophyllum scalare) chemical stimuli from males will increase spawning rates in females, but this response is increased further by the effects of visual and chemical cues (Chien, 1973). Additional examples of the role of chemical secretions in reproductive behavior are reviewed by Bardach and Todd (1970).

Many species of fish shed their gametes into the water, where fertilization takes place. But meeting of eggs and sperm is not totally haphazard. Spermatozoa generally move and may be attracted to eggs by a chemical message. Unlike that of most species, the sperm of the Pacific herring(Clupea pallasii) are almost motionless in the water(Yanagimachi, 1957). Once in the vicinity of the micropyle of a mature egg, however, the sperm move actively and instantly enter the micropyle. Yanagimachi found that the sperm attractant quality emanated from the egg membrane around the micropyle area—the interior of the egg is ineffective—and suggested that the essential groups of the sperm-activating factor either are proteins or are intimately associated with protein. Suzuki (1961) also implicated a messenger from the micropyle area for sperm attraction in the Japanese bitterling (Acheilognathus lanceolata).

Parental care and recognition of the young among several species of fish also appear to be mediated by chemical cues. Klihme (1963) studied adult jewel fish (Hemichromis bimaculatus) whose young had recently hatched. When water from the tank containing their young was introduced into a parent's tank, the adult oriented to that specific inflow locus and displayed fanning and other parental behavior. This response persisted for three weeks, which is the normal period of parental care. By exchanging these fry with younger or older offspring, Kühme was able to extend or shorten, respectively, the time period of parental care. He also established that parents could distinguish their young from those of other parents solely on the basis of chemical cues. Similar results of parental discrimination of and attraction to their own broods have been established in the dwarf cichlid (Nannacara anomala: Kunzer, 1964) and the Central American cichlid (Cichlosoma nigrofasciatum: Myrberg, 1966, 1975).

The utilization of chemical cues for the initiation and maintenance of schooling behavior in fish has been previously studied with varying interpretations of its degree of importance. Keenleyside (1955) studied blinded rudd (Scardinius erythrophthalmus), which perceived and were attracted to odors of their species mates. Following destruction of the olfactory epithelium, the fish ceased to respond to these stimuli. Keenleyside suggested that olfactory cues may keep these schools from scattering at night. The roach Rutilus rutilus exhibited comparably high levels of attraction to both visual and chemical stimuli of species mates (Hemmings, 1966). The author also hypothesized that mutual attraction and schooling maintenance may be mediated to a large degree by chemoreception during nocturnal conditions and by vision during the day, each being further integrated with stimuli received through the lateral line system. Kühme (1964) found that young jewel fish (Hemichromis bimaculatus) could orient positively to the odors of other comparably aged conspecifics. Although chemical cues appear to serve a function in the attraction phases of schooling behavior, as Shaw ( 1970) pointed out, other stimuli, particularly visual cues, must necessarily be integrated for the manifestation of all the finer spatial and positional adjustments of a school.

There is another group of chemical compounds among fish that have been found to stimulate alarm reactions or fright responses. The alarm substance, or Schreckstoff (von Frisch, 1938, 1941), is generally released from club cells in the epidermis of an injured fish (Pfeiffer, 1960; see Kleerekoper, 1969; Bardach and Todd, 1970; and Hara, 1971 for recent reviews).

Pfeiffer (1963) reported the presence of alarm substances and reactions in several species of North American Cyprinidae and Catastomidae and believed that only fish in the order Ostariophysi possessed these substances. Verheijen (1963) identified nine species of cyprinid fishes that also exhibited the flight reponse to intraspecific skin extracts. An alarm substance has been described for the top smelt (Atherinops affinis: Skinner et al., 1962), but Rosenblatt and Losey (1967) discounted it.

Reed (1969), following von Frisch's methods for measuring the alarm reaction, found that Gambusia affinis and Fundulus olivaceus (nonostariophysans) and Notropis venus tus, N. texanus,and Hybopsis aestivalis (Cyprinidae) exhibited fright responses to skin extracts from their own species. The fright reaction among Fundulus and Gambusia consisted of the fish becoming motionless. In some cases Gambusia also darted downward and began digging in the gravel as if attempting to hide. The Cyprinidae of this study, in general, became excited and formed tight schools that moved to the tank bottom. In the same study Reed also found that the same five prey species exhibited a fright response when exposed to the odors of three North American and two South American predatory species. The significance of this finding was that in several cases the prey and predator species are ecologically isolated and hence would normally not encounter each other in their natural environments.

The alarm substance has two beneficial effects for the species involved. The fright reaction should move fishes away from feeding predators as rapidly as possible. It may also be an anticannibalism device so that a fish preying on a young fish of its own species or a related species will be inhibited from further feeding upon release of the alarm substance.

Acoustic Communication

Unlike higher vertebrates, which typically share homologous sources of sound production, fishes early in evolution apparently did not possess specialized mechanisms to produce sound. Mechanisms of this sort appeared as later developments, evolving sporadically and independently in various fish taxa.

Fishes produce sounds in two basic ways: by stridulation of bony elements or by movement of the swimbladder (Barber and Mowbray, 1956; Burkenroad, 1931; Skoglund, 1961; Tavolga, 1962; Gainer and Klancher, 1965; Markl, 1971). Stridulation can be caused by grinding of the teeth, moving of the skull, pectoral girdle, or fins. The swimbladder can be set into motion by the rapid contraction of specialized intrinsic or extrinsic muscles. In addition the swimbladder can pick up and amplify vibrations produced by stridulation of other parts of the body.

Stridulatory sounds are usually of short duration and spread over a wider frequency range than swimbladder sounds. The croaking gourami (Trichopsis vittatus), for example, produces sound energy up to 12 kHz (Marshall, 1966), although it is doubtful that the fish can hear frequencies this high.

Sounds produced by swimbladder mechanisms generally contain energy in a frequency range from less than 100 Hz to several kHz with greatest amplitude in the lower frequencies. In many species the sound fundamental corresponds directly to the rate of muscle contraction (Packard, 1960; Skoglund, 1961; Winn and Marshall, 1963; Cohen and Winn, 1967; Markl, 1971).

Fishes also produce sounds incidental to swimming, known as hydrodynamic sounds (Moulton, 1960). The significance of these sounds in communication has not been well established but they may attract animals to sources of food in some instances. More extensive reviews on sound-producing mechanisms are available in Marshall (1962, 1967), Schneider (1967), Tavolga (1964, 1971a), Demski et al. (1973), and Fine (1975).

Sounds are usually named by either their function or their behavioral context (e.g., courtship sound or threat sound) or by how the observer may describe the sound (e.g., onomatopoeic description such as knock, thump, purr, staccato). That these representations of sound are, of course, subjective and will vary from observer to observer (Fish and Mowbray, 1970) makes it quite difficult to compare sounds from the literature where different investigators have named the same sound differently. Contributing to the problem is the failure of many investigators to publish the sonagrams or oscillograms of the fish signals or even to describe the signals completely.

With these limitations in mind, we have attempted to summarize in Table 1 the sounds and coincident behavior of various fishes. The fishes are listed taxonomically according to Greenwood et al. (1966). The examples used were of fishes that modulate their acoustic signals in some way, or conversely, those that fail to vary them in diverse situations. It is quite possible in many instances that signals are more variable or are graded in a finer manner than described in the original publications. Acoustic variations are presented in this table in an attempt to discern how fishes may code their sounds. Tavolga (1974) warned that parameters that are blatantly obvious to humans may be irrelevant to a fish. Meaningful parameters may be established experimentally by playing back whole sounds and their components while observing responses. (Playbacks will be reviewed later in this chapter.)

Hearing in fishes has been extensively reviewed (Moulton, 1963; Cahn, 1967; Enger, 1968; Flock, 1971; Lowenstein, 1971; Tavolga, 1971a; Erulkar, 1972; Hawkins, 1973; Popper and Fay, 1973). The problem is complicated by the dual nature of underwater sound; it is made of a vector velocity or displacement component and a scalar pressure component. Fishes receive sound vibrations through the lateral line and the labyrinth. Receptors respond to shearing forces that move the kinocilium on hair cells (Hilliman and Lewes, 1971). Lateral line neuromasts are deformed by near-field sounds and mechanical movements in the water, and are therefore tuned to have higher thresholds than units in the ear. Lateral line organs can act as directional sensors close to the emitting source (Harris and van Bergeijk, 1962). Recent evidence indicates that the labyrinth may also function in localization of sound (Schuijf and Siemelink, 1974).

FUNCTION OF SOUND

Fish and Mowbray's (1970) work on fishes of the western North Atlantic is indicative of the large number of fish species known to produce sounds. Impressive as this number is, it is important to realize that the behavioral significance of only a small fraction of these sounds is known. The sounds that observers have been able to identify as playing a role in specific behavioral acts are primarily related to either aggression or reproduction. In many cases the close relation between these two categories makes a complete dichotomy difficult.

Many species, such as the tiger loach (Botia hymenophysa: Klausewitz, 1958), use sound in defense of territory. In the toadfish Opsanus tau the boatwhistle, which can attract a female to a male's nest, may also indicate that the territory is occupied by a male, and hence other males are warned that they are unwelcome in the area (Gray and Winn, 1961; Winn 1964, 1967, 1972). Should another male approach the nest, the resident male will grunt at the intruder. The northern midshipman (Porichthys notatus) has three agonistic calls: a grunt spaced both regularly and irregularly and a buzz (Cohen and Winn, 1967). All three signals are elicited under identical conditions, making the individual message content for each sound unclear.

Perhaps uniquely, sea anemone fishes (Amphiprion spp.) make a sound that is specific for fighting and another that is used as a territorial threat sound (Schneider, 1964a). In addition these fishes have a submissive sound emitted by vanquished fish after a fight.

The female in a courting pair of the cichlid Hemichromis bimaculatus emits a "br-r-r" sound while aggressively holding ground after the male has bitten or rammed her or shows intentions of doing so (Myrberg et al., 1965). This behavior deters further attacks by the male, suggesting to Barlow (1970) that the action of the female induces fear in the male.

The squirrelfish Holocentrus rufus produces two calls of markedly different types (Winn et al.,1964). The grunt, although used when a territory is invaded by another species, is used chiefly against neighboring conspecifics. When the staccato sound is made, it acts primarily as a warning sound elicited when the territory is invaded by large fish, including predators, or by almost any fish that appears suddenly. Grunts appear to be associated primarily with aggression and experimentally were shown not to habituate readily, while the staccato is associated more with escape tendencies and did habituate readily.

A schooling, nonterritorial squirrelfish, Myripristis berndti produces both a grunt and a staccato sound (Salmon, 1967), just as the territorial species Holocentrus rufus does (Winn et al., 1964); but in Myripristis both sounds act as warning calls. In addition, M. berndti also produces two calls that could be termed agonistic against species members, but they serve the purpose of increasing distance between fish. This call might function in determining optimal spacing within the school rather than as territorial behavior. Salmon speculates from these observations that in various species of Myripristis schooling, rather than occupying a territory, encourages the evolution of a complicated vocabulary. Among different species of squirrelfish, at least, this conjecture is supported by the fact that the sounds emitted by M berndti and M. violaceus (Salmon, 1967; Horch and Salmon, 1973) are more complicated than that of the territorial Holocentrus rufus (Winn et al., 1964). The problem may be more complex because H. rufus forms schools in certain habitats (Winn, pers. obs.). However, observations made of other schooling species appear not to support the idea as a generalization.

Cod (Gadus morhua) and haddock (Melanogrammus aeglefinus), which are schooling species, appear to have simple vocabularies (Brawn, 1961b; Hawkins and Chapman, 1966). The cod apparently uses the same grunt for agonism and courtship, as well as for the breakup of courtship (Brawn, 1961b), but possibly further study will show a more differentiated acoustic system. It is possible that this sound may function as an alerting or display-enhancing device and be devoid of a specific message content. In the sea catfish (Galeichthys felis: Tavolga, 1971b) and the croaking gouraitti ( Trichopsis vittatus: Marshall, 1966) the agonistic displays and vocalizations may have a hierarchical function. For example, once beaten, a gourami may remain near the dominant conspecific without eliciting further aggression.

Most soniferous fishes can be induced to vocalize when they are held or chased (Burkenroad, 1931; Fish, 1954; Fish and Mowbray, 1970; Horch and Salmon, 1973). We assume that the sounds may be produced in the wild when a fish is attacked by a predator and may act as warning calls. The squirrelfish Myripristis violaceus produces both a growl and a grunt (the latter were produced only when the fish were hand-held) when confronted by threatening situations. When these sounds were played back, the fish responded by showing attention to the sound source (a speaker) and producing growl sounds. On the basis of their observations, Horch and Salmon (1973) surmised that these sounds may be used in nature to alert species mates to danger.

Since Smith's work in 1905, sciaenid calls have been generally assumed to play a courtship role because the sounds occur only during the mating season. Dijkgraaf (1947) observed a group of four Corvina nigra, of which two appeared to be a pair. The smaller of the pair, probably the male, would swim closely behind and below the sluggish and larger female, with his head below her abdomen. At odd intervals and without apparent external motivation, the male would swim to one of the other fishes and chase it around the tank while emitting its knocking sound. This behavior occurred around twilight, when the apparently dominant male searched for the female and began his chasing.

Males are often more vocal than females during close courtship exchanges. In toadfish, as well as in many other species, there may be an endocrine as well as a physical basis for sexual differences in sound production. For example, courtship sounds are produced only by male toadfish Opsanus tau. However, both Demski and Gerald (1974) in O. beta, and Fine (unpublished) in O. tau, have elicited boatwhistle-like sounds in female toadfish by brain stimulation; though the female is not known to produce this sound in nature (Gray and Winn, 1961). Although the swim bladder intrinsic-muscle complex grows faster in the male than in the female, the bladder complex is equally developed in the two sexes, implying a hormonal rather than a morphological basis for differential sound production in the male (Fine, 1975).

Although not common, there are cases in which females have been observed to produce courtship calls. Delco (1960) described sounds by females during the courtship behavior of Notropis lutrensis and N. venustus. Stout (1963) was somewhat skeptical of this work since he found that sounds were largely, if not entirely, produced by males in the closely related species N. analostanus. Fish (1954), describing mating of a pair of sea horses Hippocampus hudsonius stated that "preliminary activity consisted of slow swimming, either together or apart, accompanied by occasional noisy snapping of the head. Clicks were often produced alternately by the two fishes, and during their actual embrace, these sounds were loud and almost continuous." Although no sounds were heard during nonaggressive courtship, spawning, or caring for the offspring, the "br-r-r" sound of the female Hemichromis bimaculatus seems to inhibit the male's aggression (Myrberg et al., 1965). In the croaking gourami (Trichopsis vittatus) both the male and female croak during lateral displays, aggressive interactions, and the early stages of courtship (Marshall, 1966). During courtship and spawning the female purrs in a headup posture. According to Marshall, this display operates as a distance-decreasing mechanism in contradistinction to the hypothesized function of the croaking sound. It is probable that in both the cichlid and the gourami courtship sounds have evolved from a system designed primarily for aggression.

There are few known examples of males producing courtship sounds during the reproductive season when no female is within visual distance. Toadfishes (Opsanus tau: Fish, 1954; Fish and Mowbray, 1959; Tavolga, 1958c, 1960; Gray and Winn, 1961; Winn, 1967, 1972; 0. beta: Breder, 1941, 1968; Tavolga, 1958c, 1960; 0. phobetron: Tavolga, 1968b) and Porichthys notatus (R. Ibara, pers. comm.) produce sounds of this type. Such a system of spontaneous calling is analogous to bird song and shares similar properties (Winn, 1964, 1972). Schleidt's (1973) concept of tonic communication should also be applicable to the toadfish. A background chorus of boatwhistles could enable females to enter final spawning readiness, at which time the call would become attractive and thereby serve as an orienting stimulus. Similarly, a priming function was shown by Marshall (1972), who played back male sounds to females of the cichlid Tilapia mossambica, causing them to lay eggs several days earlier than control females. There may be other examples where sounds act in communicating to distant fish. Although Gerald (1971) observed sound production associated only with active courtship behavior in various species of sunfishes, playbacks of the male's courtship grunts attracted both males and females. The sunfish call could be evolving for communication over longer distances or could simply have an incidental (to the emitter) attractive quality. Delco (1960) found that male and female Notropis lutrensis were attracted to a chamber from which female N. lutrensis sounds were produced. Only males of N. venustus responded to conspecific sound in a comparable experiment.

Schwarz (1974a) used information analysis, a fundamental tool for determining if communication occurs, for the first time on fish sounds. She showed that there were three lines of evidence that indicated a function for the sounds of Cichlasoma centrarchus: sound was associated with aggressive behavior; responses to silent behavior differed significantly from responses to behavior accompanied by sound; and finally, the amount of information transmitted when animals could hear as well as see one another differed from that transmitted when they could only see one another. A recipient conspecific responds to sound by ceasing aggressive acts and usually moving quickly away from the emitter. Aggressive displays not accompanied by sound emission have a much lower threat value. Breeding pairs do not use the sound for courtship, but rather as a threat. From introduction up to the time of spawning, the male produces most of the sounds, directing them toward the female. From this point on, the female makes most of the sounds, but directs them at fishes in adjacent tanks (Schwarz, 1974a, 1974b, in preparation).

SIGNAL VARIATION AND CODING

Most fish sounds are basically percussive, and the various messages sent between fishes are elaborations of pulses (Fig. 4). In an attempt to summarize how fishes code their signals, Winn (1964) categorized them in five basic ways (Fig. 5), which suggested to him that the temporal patterning of sounds was an important carrier of information. The first and second were variabletime-interval and fixed-time-interval signals, in which the time between units viewed as one sound on a spectrogram (grunt, knock, etc.) were variable or fixed. This scheme referred to whole sounds rather than to components of sounds related to individual muscle contractions. In the third and fourth types of signal, the duration of any signal is lengthened (unit-duration signal) or the amount of time during which units are produced is varied (time-length signal). The final category was harmonic-frequency signals of longer duration.

There are five basic ways in which a signal produced by a fish can be varied: amplitude, duration, repetition rate, number of pulses within a signal, and frequency. While in many cases, categories such as duration, repetition rate, number of pulses, and intervals are obviously related, at other times they can refer to quite disparate quantities.

Amplitude

By analogy with humans, we might expect fishes to reserve their louder sounds for higher emotional states or levels of arousal, and in fact such modulation does occur. The cod (Gadus morhua,) grunts more loudly when using sound as a determined threat and more faintly at the end of an aggressive encounter (Brawn, 1961b). Fish (1954) observed a sea robin, Prionotus Carolinas, that would lie quietly on the bottom clucking softly while being gently stroked by the experimenters. However, if this stimulation was applied too long or too heavily, the fish would break away in apparent annoyance and emit a louder burst. The knocking signal of the sciaenid Corvina nigra becomes more energetic under higher stimulation (Dijkgraaf, 1947).

Amplitude is also used to differentiate sounds that have different meanings. The threatening sound of the sea anemone fish (Amphiprion xanthurus) is loud enough to be audible in air 10 m from the aquarium. The fighting sound of the fish is emitted much closer to a conspecific than the threatening sound and is of lower intensity (Schneider, 1964a). The courtship purr of the satinfin shiner (Notropis analostanus) is less intense than its agonistic vocalizations (Winn and Stout, 1960; Stout, 1963). The threatening sound in the tiger fish Therapon jarbua is louder and is elicited by a higher level of agonism than the drumming sound (Schneider, 1964b). The short grunt is quieter than the long, loud grunt in the sea catfish (Galeichthysfelis: Tavolga, 1960). The squirrelfish Myripristis berndti produces grunts and staccatos as warning calls. The staccato, which is elicited under a more stressful situation than the grunt, is also louder (Salmon, 1967). Winn (1972) found that a toadfish boat whistle with an amplitude equivalent to that emitted by a fish less than one foot away would cause another toadfish to stop calling, suggesting that the call aids in territorial spacing. This effect is not obtained with lower amplitudes.

Fig. 4. Examples of fish sounds demonstrating squirrelfish Holocentrus rufus; c. chorus of croakers Mi- their pulsed nature, a. rapid series of knocks of the cropogon undulatus (note sounds occur in pairs); satinfin shiner Notropis analostanus; b. staccato of the d. grunts of a toadfish Opsanus tau. Narrow-band filter.

Fig. 5. Diagrammatic representation of how fish sounds are coded. (From Winn, 1964.)

Duration

Fishes vary the duration of their acoustic signals, producing different calls and modulating the signal within a call. The short grunt naturally emitted by Galeichthys felis is 20 to 40 msec long, while the grunt emitted by a prodded fish is over 100 msec; in the catfish Bagre marinus the sobbing sound may be over half a second long compared with the yelp, which varies from 100 to 200 msec (Tavolga, 1960). The squirrelfish Myripristis berndti has one of the most varied vocabularies known for a fish, with the signals largely separated on the basis of duration (Salmon, 1967). Both the knock and the growl are agonistic calls expressed to conspecifics. Given during chasing, the knock is short, while the growl, produced only after physical contact, is many seconds long. This fish also produces grunts and staccatos in warning situations. While the staccato is considerably longer than the grunt, the individual components of the staccato, made of a series of sounds, appear to be of shorter duration than the grunts.

The tiger fish Therapon jarbua produces two agonistic calls under different levels of motivation. The drumming sound consists of 10 msec pulses, produced at irregular rates; and the threatening sound is made of a burst of these pulses, up to 200 per second. The more intense a fish's threat, the longer it produces the sound (Schneider, 1964b).

The threatening, fighting, and shaking (submissive) sounds of the sea anemone fish (Amphiprion xanthurus) separate easily on the basis of duration, being respectively 25 to 30, 45 to 60, and 250 to 400 msec long. These signals are longer than comparable ones produced by A. polymus (Schneider, 1964a).

Both male and female toadfish Opsanus tau produce a single coarse grunt (Fish, 1954) that can be continuously graded into a rapid series of grunts called a growl. The grunt, which is produced in aggressive encounters, may become a growl when the intensity of the encounter increases (Winn, 1972). The two signals are differentiated by their duration and rate (Winn, 1972).

Agonistic sounds of several sympatric triggerfishes (Balistidae) vary in duration and pulsing (Salmon et al., 1968).

The northern midshipman (Porichthys notatus) makes a continuous mating call that can vary from less than ten minutes to an hour (Ibara, pers. comm.). The call is made by a nesting male to attract a female and is produced without a direct external stimulus, but it may be elicited indirectly by the calls from other nearby males. The duration of the call is probably related to some aspect of the male's internal state.

Delco (1960) described a female courtship call for the minnows Notropis leutrensis and N.venustus having durations of 0.84 and 0.047 to 0.07 seconds, respectively. The duration of the components of the male satinfin shiner's (Notropis analostanus) courtship purr varies between 11 and 24 msec, with the purring sound that is directed at the female during courtship consisting of a more rapid series of lower intensity knocks (Winn and Stout, 1960; Stout, 1963).

The courtship grunt produced by the male damselfish Eupomacentrus partitus is variable in number of pulses and duration (Myrberg, 1972a, 1972b), with the variation apparently correlated with the amount of time the male spends swimming close to the female. Moulton (1958) observed a meeting of a pair of angelfish Pomacanthus arcuatus in which vocalizations changed from short grunts to longer moanlike sounds.

The freshwater drum (Aplodinotus grunniens) produces a drumming sound during a long session during the day (Schneider and Hasler, 1960). Early and late in the session it produces a main sound of three to five seconds with one to several short sounds preceding and following. During the peak hours of the session the sound series reaches a minute or longer before a break, and the short sounds decrease or disappear. Dijkgraaf (1947) described a similar phenomenon in another sciaenid, Corvina nigra. Under periods of increasing stimulation, the knocking signal became more frequent and more energetic, and the duration increased. However, the number of knocks in a signal varied from five to seven and did not increase as it did in Aplodinotus.

Sounds may function as an ethological isolating mechanism in fish, with duration being an important component of the system. Gerald (1971) has studied the male courtship grunts in six sympatric species of sunfish (Centrarchidae). He was able to separate the vocalizations on the basis of call duration and percentage of call pulsed (Fig. 6). Although he has some evidence that the sounds are discriminated by conspecific females, the system is not perfect and hybrids occasionally occur. Another interesting example comes from three closely related Atlantic toad-fishes, Opsanus tau, 0. beta, and 0. phobetron (Walters and Robins, 1961; Fish and Mowbray, 1959). The toadfish species share similar meristics and morphometries and are separated primarily on size and color pattern. The boatwhistle calls of the three are distinct (Fig. 7). 0. beta produces a double hoot, with the second sound shorter than the first, while O. tau and 0. phobetron produce a single boatwhistle. While the call of an O. tau rarely exceeds half a second, 0. phobetron's call may last up to a second (Tavolga, 1968b). Although the three species are currently allopatric, they may have been sympatric in the recent past and perhaps the divergence of their calls can be ascribed to character displacement. Duration is an important parameter in boatwhistle call recognition in Opsanus tau (see playback experiments).

Fig. 6. Average grunt duration and pulses per second of six species of sunfish (Lepomis). The crosses represent the intersection of the means for each species, and the closed rectangles represent two standard errors on either side of the means. L. macrochirus, L. microlophus, and L. cyanellus had essentially no pulsation. (From Gerald, 1971.)

Sounds of a nonsexual nature may also diverge among related forms. Such sounds could function for intraspecific species recognition, or they might be a reflection of a change in a fish's whole acoustic repertoire during speciation. For instance, the hand-held sounds of Priacanthus meeki range between 76 and 150 msec, while those of P. cruentatus go from 300 to 600 msec (Salmon and Winn, 1966).

Fig. 7. Boatwhistle call of three species of toadfish. A. Opsanus tau; B. 0. beta; C. 0. phobetron. Narrow band filter. (B. and C. from tape accompanying Tavolga, 1968b.)

Repetition Rate

The temporal patterning of fish sounds, a combination of duration and repetition rate (on and off times), is of paramount importance in their coding. Winn (1964, 1967, 1972) has focused on the idea embodied in repetition rate by dividing sounds into fixed and variable interval types. It is largely variable rates that make the drumming sounds of various sciaenids species-specific. Corvina nigra produces eight pulses a second (Dijkgraaf, 1947), Cynoscion regalis twenty-four (Tower, 1908), and Aplodinotus grunniens eighteen to twenty-seven (Schneider and Hasler, 1960). This sort of variation is shown by Fish and Mowbray (1970) in the sonagrams of a host of sciaenid species.

Individual species often produce different calls of varying repetition rates. The rapid series of knocks of Notropis analostanus is variably paced, while the purr has a constant but faster repetition rate (Winn and Stout,1960; Stout, 1963), i.e., the fixed interval of Winn (1964,1972). Likewise, squirrelfish of the genera Holocentrusand Myripristis produce variable-interval grunts and fast-paced equal-interval staccatos (Winn et al., 1964; Salmon, 1967). Toadfish grunts vary from individual pulses to rapidly pulsed growls (Winn, 1964). In fact, the boatwhistle of the toadfish probably represents an extreme within this spectrum, with the sonic muscles undergoing a sustained contraction at their maximal rate, which is considerably higher than that used in the grunt and growl. The thump produced during a fight between conspecific males in Hemichromis bimaculatus has a pulse repetition rate of 12 Hz, while the "br-r-r" produced before attack is pulsed at thirty-five times a second (Myrberg et al., 1965).

Motivational states and their meaning can be graded by varying the rate of production of a sound. The courtship grunts of Bathygobius soporator increase in repetition rate during active courtship (Tavolga, 1958a). Chasmodes bosquianus produces its courtship grunt every second or two, but occasionally the male will emit a burst of three to four sounds in quick succession if a female approaches his shelter (Tavolga, 1958b). Likewise, male toadfish have been observed to increase their rate of boatwhistling when a gravid female approached (Gray and Winn, 1961; Fish, 1972). Experimental studies on rate are given in the playback section.

Recent work on the swimbladder's role in hearing supports our contention that the temporal nature of sound, rather than variations in its frequency content, is important for communication in fishes. The swimbladder presents an acoustic discontinuity to underwater sound, which causes it to resonate. This motion is rera-diated and translated to the ears. Unlike an underwater bubble, which has a highly tuned resonant frequency determined by its size and depth, the swimbladder is highly damped within the body of the fish (Alexander, 1966; McCartney and Stubbs, 1970; Demski et al., 1973; Sand and Hawkins, 1973; Popper, 1974). Popper (1974) found that the response of the swimbladder of an insonified living goldfish was flat from 50 to 2,000 Hz, indicating that the bladder did not selectively favor any particular frequencies. A strongly tuned bladder would change a fish's hearing sensitivity with depth and size (age) and therefore be deleterious (Popper, 1971; Sand and Hawkins, 1973). The damped nature of the swimbladder makes it responsive to the time-based nature of a signal. A strongly resonant structure takes time to start responding and continues to respond after stimulation has terminated (Popper, 1974). Such a reverberating system would be poorly adapted to discriminate broad-band pulsed sounds.

Popper (1972) studied the auditory threshold in the goldfish as a function of signal duration and indicated that there were no differences between short pulses and continuous tones, and that thresholds were the same whether there is a long or short off-time between pulses. Popper concluded that temporal summation does not occur in goldfish and that the enhanced sensitivity for long sounds found in mammals is unnecessary for the short duration of fish sounds.

Number of Pulses or Units

Looking at the number of pulses or units in an acoustic display could be identical to focusing on the duration orrepetition rate (or the on and off time). But, whereas the lattertwo categories are on an obvious continuum, the former may belooked at as a binary or incremental function. For instance, manyfishes make sounds that contain one or multiple pulses (Table 1). If the sounds merely grade from unito multipulsed with some sortof proportional behavioral motivation, they do not belong in this category. This question cannot be answered in most cases. In N. analostanus single knocks are used in the initial stages of courtship and in chasing between males, while two different types of series of knocks are used for more intense fighting and courtship (Winn and Stout, 1960; Stout, 1963). The courtship purr of the female croaking gourami (Trichopsis vittatus) is separated from the male and female agonistic croaking sound by the number of pulses, averaging 2.3 pulses per croak and 4.5 pulses per croak respectively (Marshall, 1966). This type of coding reaches a peak in the damselfish Eupomacentrus partitus, where five different calls can be separated in this manner (Myrberg, 1972a, 1972b). The pop, chirp, long chirp, and burr have respectively 1, 3, 4-6, and 8-12 pulses per sequence, while the grunt has a variable number of pulses. Each drumbeat sound of the black grouper (Mycteroperca bonaci), caused by opercular movements, consists of five sound pulses (Tavolga, 1960).

Table 1

Message and modulation of fish vocalization.

Repetition of entire sounds seems to be patterned in the clown fish (Amphiprion xanthurus: Schneider, 1964a). The submissive, shaking sound is given in groups of three to six, and the fish does not usually emit more than four attack sounds in one fight. The threatening sound varies with the situation: long series of ten or more sounds are emitted before a fight and shorter series of three to five are emitted after a lull in fighting, initiating a new phase of battle.

Frequency

The role of frequency in coding fish sounds is not well understood. Although some sounds (particularly stridulatory ones) have wide-frequency components, other sounds clearly show characteristic species differences in frequency (particularly fundamentals). The frequency spectrum is typically a property of the sound-producing mechanism and is sometimes a function of the sonic muscle contraction rate. Sustained muscle contractions often produce harmonics, but in some cases, as Tavolga (1962) has shown, harmonics may also be related to the acoustic properties of the environment. Playbacks to toadfish demonstrated that they respond positively to tones of 200-400 Hz without the presence of harmonics (Winn, 1972), and any harmonics above the second are beyond the toadfish's hearing range (Winn, 1972; Fish and Offutt, 1972). At this time there is no proof of relevance of harmonics to meaning in a fish signal.

In a system where cavity size and internal pressure remain relatively constant, frequency is not widely varied. Generally, certain frequencies or groups of frequencies are favored because of the physics of the system, but it is not typically known what portion of the acoustic energy fish focus on within their hearing range. In many instances the higher frequencies of a sound may be above the fish's hearing range.

Species-specific frequency differences occur because of morphometric and physiological changes in the sound-producing mechanism during evolution. There are differences in the fundamental frequency of the boatwhistle calls of the three species of toadfish that were discussed at length in the section on duration. In addition, there are indications of a possible clinal variation in the fundamental frequency of Opsanus tau feoatwhistles up and down the Atlantic coast of North America, with frequency increases in the lower latitudes (Tavolga, 1968b, 1971a; Fish and Mowbray, 1970). However, the effects of temperature and fish size on the fundamental are unknown for this species. Schneider (1964b) has shown that increasing temperature increases the contraction rate of the sonic muscles in Therapon jarbua. In frogs, rising temperature generally causes an increase in the repetition rate and fundamental of a call, with a coincident drop in duration (Blair, 1958; Schneider, 1968, 1974; Lorcher, 1969). The rate of boatwhistling does increase with temperature in Opsanus beta (Breder, 1968), although the mechanism behind this correlation is unexplained.

The effect of animal size on frequency is an open question and may depend on the species. An inverse correlation between size and frequency was established in the croaking gourami ( Trichopsis vittatus: Marshall, 1966), was assumed in the cichlids Hemichromis bimaculatus and Cichlasoma nigrofasciatum (Myrberg et al., 1965), and is probably true of many fishes. Myrberg et al. (1965) also found that the "br-r-r" sound of the female is higher pitched than the same sound of the larger male.

In species like the toadfish, where sonic muscle contraction rate governs the fundamental frequency (Skoglund,1961), bladder size may not be an important consideration. In the Japanese gurnard (Chelidonichthys kumu) the dominant frequency of the "gu" call lies around 300 Hz in fishes of three different size groups, even though the sound energy ranges somewhat higher in smaller specimens (Bayoumi, 1970). The fundamental of low-pitched grunts from Bagre and Galeichthys is about 150 Hz, regardless of the size of the fish (Tavolga, 1962). The principal frequencies of the toadfish grunt do not appear to be related to size (Fish, 1954).

Frequency modulation within a call is common in birds and mammals (Armstrong, 1963; Poulter, 1968; Tembrock, 1968) but occurs rarely, if at all, in fishes. Caldwell and Caldwell (1967) found a double click in the pinfish (Lagodon rhomboïdes), where the second click has a slightly higher or lower principal frequency than the first. Several types of courtship sounds of the bicolor damselfish (Eupomacentrus partitus) have the same frequency range (Myrberg, 1972a, 1972b), except for the burr, which has a restricted and different peak frequency range. How the fish accomplishes this modulation is unknown. Tavolga (1960) shows several sonagrams of catfish and toadfish vocalizations in which the fundamental frequency decreases slightly through the course of a long sustained call. This decrease undoubtedly results from a diminution of muscle-contraction rate, perhaps due to fatigue, and is probably unrelated to communication. This change in contraction rate does provide a theoretical basis of frequency modulation.

Enger (1963) showed that some auditory neurons of Cottus scorpius responded to sounds up to 200 Hz and others up to 300-500 Hz. Within certain limits, frequency could be detected by the volley principle. Jacobs and Tavolga (1968) demonstrated good frequency discrimination in the goldfish. Given this and the fact that in some cases frequencies differ between species, frequency must be important in some cases (see playback section), although the fishes' system is primitive when compared with that of higher vertebrates. The ability to use frequency will vary from one group of fishes to the next, depending on the complexity of the auditory system.

THE PLAYBACK EXPERIMENT

Playing back their sounds to fish is an experimental method of unraveling messages, equivalent to presenting painted models in the visual modality. The experiment can be used for at least four distinct purposes: (1) to establish species recognition of a sound; (2) to separate different calls; (3) to observe the effect of the sound on behavior and discover or confirm how it is used; and (4) to establish how the call is coded or what parameters of the call are relevant to communication.

Moulton (1956) attempted one of the first playback experiments when he transmitted a staccato call and an electronic imitation to the sea robins Prionotus Carolinas and P. evolans. Occasionally a fish would respond with a staccato call after the playback was turned off. Moulton was also able to suppress the staccato call by playing back 200-600 Hz signals for the approximate duration of the call. Recently Fish and Mowbray (1970) questioned Moulton's species determination, suggesting that on the basis of frequency range and pulse form his sounds came most likely from a sciaenid, probably Cynoscion regalis.

Another early work using the playback technique involved the playing of a tape loop of male courtship sounds of Bathygobius soporator to males and females (Tavolga,1958a). Females responded within a minute by increasing their activity, respiration, and contacts between individuals (nipping, butting, and approaching), though they did not orient to the sound. When a male was confined in a flask during playback, the females oriented toward him, bumping the flask repeatedly. Male responses to playback were similar, except that they approached and remained near the sound if in a courtship set, i.e., isolated or recently exposed to females. Males in a combat set (exposed to other males) exhibited no response to the playback. Tavolga played back electronic pulses to the fish to explore the significant parameters of the call. Animals responded maximally to frequencies between 100 and 300 Hz, to pulse durations of 75-150 msec, and to variable repetition rates depending on duration. Pulses shorter than 75 msec were effective only at high repetition rates. Intensities higher than normal were also effective. Within these limits the sound was not specific and positive responses were elicited to the courtship sound of the blenny Chasmodes and to Tavolga saying "ugh-ugh."

Various aspects of the vocalization of the toadfish Opsanus tau have been investigated. Winn (1967) found that he could increase a fish's rate of boatwhistling by playing back boatwhistles at a rate of eighteen or more per minute. Grunts and slower paced boatwhistle playbacks (thirteen per minute and below) did not increase calling. In further work, Fish (1972) and Winn (1972) found that the maximum stimulatory threshold calling rate was one sound every 4.0 to 4.5 seconds. When sounds were played at this rate, the toadfish and the playback alternated. Faster playbacks and an experiment with a delayed playback of each sound the toadfish produced did not establish antiphony. Continuous tones suppressed calling, but when intermittent tones were played, toadfish placed their boat-whistles in the silent period. Winn (1972) played tone signals to toadfish to determine what part of the signals were relevant in communication. A frequency of 180 Hz was stimulating, but a sound with more energy at 100 Hz than at higher frequencies was not. Sounds too loud, equivalent to the call of a fish less than one foot away, caused a reduction in calling. Durations of 75 and 150 msec resulted in a loss of stimulatory value compared to a duration of 300 msec, the approximate average of natural calls from Solomons, Maryland. It was clear that amplitude, duration, repetition rate, and frequency are all important parameters in social facilitation of boatwhistling by toadfish.

Winn also showed that females were attracted to boatwhistles coming from cans in front of a speaker. Ripe females of Porichthys notatus, a member of the same family, were attracted to continuous tones of 105-120 Hz. The females became excited and swam around the tank and to the speaker (Ibara, pers. comm.). Nonripe females, a juvenile, and nest-guarding males were not attracted to the sound.

After a latency of thirty seconds to one minute, courtship sounds (chirps and grunts) of the bicolor damselfish (Eupomacentrus partitus) caused the male to take on color patterns of "white body and black mask" associated with courtship (Myrberg, 1972b). They also caused an increase in courtship behavior (tilt, dip, nudge, and lead) and vocalization. However, the fishes did not orient to the hydrophone. The agonistic pop of the bicolor and a squirrelfish staccato decreased courtship behavior compared to controls. Playbacks to bicolors of chirps from congeners E. planifrons and E. leucostictus were stimulatory to a lesser degree (Myrberg and Spires, 1972). In unpublished experiments (E. Spanier, pers. comm.) it was determined that the important parameter for species recognition of Eupomacentrus calls was the off-time or interval between sounds and not pulse duration and frequency.

Marshall (1966) performed a series ofexperiments with the croaking gourami (Trichopsisvittatus). Playback of croaking sounds to males produced no significant change in the rate of air-gulping or locomotion, and inconsistent increases in aggressive behavior. In some experiments behavior did change consistently between playbacks of croaks and background noise, but not between the croak and no sound control. Similarly, Winn (1967) inhibited toadfish boat-whistling through playback of background noise. It appears that such playbacks are suppressing behavior with an undefined stimulus. Marshall's (1966) reasons for the lack of positive results are instructive. The croaking sounds may have been only a minor part of the total complex of aggressive behavior and by themselves were not sufficient to elicit a consistent response. Playbacks are usually made from sound recorded on a loop and therefore are played without variation. This alone may make the sounds less realistic. The lack of correct temporal answering by the "playback" fish may be important. In several instances the dominant fish of a pair seemed aware that the sound was coming from the speaker and not his opponent. Habituation can quickly become a problem since there is no association of the sound with either a conspecific or specific postures of the conspecific. Finally, there is a problem with fishes maintaining variable levels of motivation. Some fish in Marshall's experiments increased and others decreased their aggressive activity during the playbacks. Preplayback and playback periods should be contiguous to minimize this problem.

Studies have been performed on several species of squirrelfish. Winn et al. (1964) caused Holocentrus rufus to enter their shelters by playing their warning staccato call, toadfish boatwhistles, and scad sounds. The fish stayed in their shelters for the first minute of sound emission, but after a while started investigating the speaker. When the sound was turned off, the fish became more active again. Lobster sounds and background noise had no effect. Playback of their grunts and staccatos to Myripristis berndti caused immediate orientation to the sound source (Salmon, 1967). The fish swam toward the speaker within five to ten seconds and ceased sound production. There was no response to playbacks of background noises or knocks. Myripristis argyromus will follow these same warning calls to their source. Thus, there does not appear to be discrimination between these calls by the two species (Popper et al., 1973). Playback of conspecific growls and hand-held grunts also caused M. violaceus to approach the speaker (Horch and Salmon, 1973). During playback of these sounds, knocks and thumps normally produced at irregular but frequent intervals ceased, and growl sounds not normally produced were elicited. Many fishes make sounds when held (Burken-road, 1930; Fish and Mowbray, 1970), but Horch and Salmon have been the only ones to demonstrate that such a sound can have a communicatory function.

Delco (1960) demonstrated a positive approach to the call of a conspecific female in Notropis lutrensis and N. venustus. Stout (1963, 1966) played back the fighting and courtship calls of Notropis analostanus. With two males of unequal dominance together, the fighting sound caused an increase in number and duration of aggressive encounters. The purring sound also caused an increase in aggressive encounters and often caused males to exhibit solo spawning. With a male and two females together, the courtship purr increased the average duration and number of the male's courtships; the fighting knocks produced a decrease in courtship that was not statistically significant. Stout concluded that the differential response to the two sounds demonstrated that the fish discriminated them.

Gerald (1971) played courtship grunts from species pairs of sunfish out of underwater speakers adjacent to lift nets. A significant difference in the capture ofLepomisn megalotis and L. humilis was found, with the calls attracting more conspecifics than heterospecifics. Note that the calls of these two species appear to be very much alike (Fig. 6). A somewhat similar test with L. humilisand L. macrochirus calls was not significant. However, samples were small, and further work might well establish the distinctiveness of the various sunfish calls.

Schwarz (1974b) played back the conspecific low growling sound to one individual of male-male and male-female pairs of the cichlid Cichlasoma centrarchus, which were acoustically but not visually isolated from each other. Playback of the growling sound to a male markedly lowered the number of highly aggressive encounters he directed at either his male or female partner, while playback of control noise or silence had no significant effect. She concluded that the sound functions to inhibit aggressive behavior in the recipient. In another cichlid, Tilapia mossambica, sounds appear to function as a priming stimulus; recordings of the male's low-pitched drum sounds caused females to lay eggs several days earlier than controls did (Marshall, 1972).

Electrical Communication

Various aspects of electric organs, receptors, and electrical communication have been recently reviewed (Black-Cleworth, 1970; Bennett, 1971a, 1971b; Bullock, 1973; Hopkins, 1974a, 1974b). Six groups of fishes have independently evolved a system of electric organs and specialized receptors (Rajidae, Torpedinidae, Mormyriformes, Gymnotoidei, Malapteruridae, and Uranoscopidae). Other fishes might be expected to respond to low electrical currents. Since Hopkins treats the subject in chapter 13 of this volume, only a short discussion follows.

The discharges are species-specific (Bullock, 1969) and are used in courtship and in hetero-specific (Bauer and Kramer, 1974) and conspecific agonistic communication. According to Hopkins (1974a), the coding of electrical communication in various roles lies in the diversity of the signals, which can be classified according to certain parameters such as the shape of the electric field, the wave form of the electrical discharge, the discharge frequency, the timing patterns between signals from sender and receiver, the frequency modulations, and the cessations of the discharge.

Although electrical signals related to attack and threat behaviors appear to be species-specific, a common appeasement display has apparently evolved in several species. The cessation of discharge signaling by subordinates of G. carapo (Black-Cleworth, 1970), Gymnarchus niloticus (Hopkins, 1974a), and Gnathonemuspetersii (Bell et al., 1974) reduced the rate of attack by the dominant. Such a display is effective in reducing aggression because the subordinate becomes electrically inconspicuous (Hopkins, 1974a).

Electrical discharges of Sternopygus macrurus have been shown to play a significant role in courtship of this species (Hopkins, 1972). During the breeding season the discharge frequencies of males and females are distinctly different. The male shows no response to the discharge of other males but responds to the female discharge with a distinct courtship signal.

Kastoun (1971) strung wire leads between two otherwise separated tanks, each containing an electric catfish (Malapterurus electricus). He stimulated one of the fish (stimulus fish) and monitored its signal output as well as the behavior of the fish in the adjacent tank (response fish). Knocking on the aquarium glass of one tank resulted in the emission of two to five impulses (defense or flight volleys) and the escape of the response fish to its living tube. By variously manipulating the stimulus fish, the researcher elicited from the response fish a search reaction and attacks on closely situated objects. The electrical volleys were discriminated through the number of impulses and the type of signal, i.e., impulse intensity, form, and grouping. The response fish was not observed to emit answering signals. In a similar experiment with electric eels (Electrophorus electricus), prodding or feeding the stimulus fish resulted in emission of impulses that attracted eels to electrodes in an adjacent tank (Bullock, 1969). Eels would cluster around and nuzzle a captive specimen lowered into their tank in a net, and they were attracted to electrodes emitting artificial pulses of variable parameters, including potentials up to 150 V.

Conclusion

Even using Burghardt's general concept of communication, as opposed to Tavolga's (1968a, 1970) more restrictive definition, it is obvious that fishes typically communicate for only short periods of their lives and about restricted subjects. Social behavior is much less complex than in higher vertebrates. There are certainly exceptions, such as schooling behavior, cleaning symbiosis, and defense of a feeding territory, all of which may last for long periods. Still, communication is a requisite for the continued existence of fishes, and it is used mostly during the mating season.

Fishes have undergone an extensive adaptive radiation unparalleled in other vertebrate classes. They are an old group with many phylogenetic lines. These two factors are reflected in the importance different modalities assume in the life styles and communicatory systems of various species. Even within a single medium, like sound, various groups appear to have evolved different strategies of detection.

The Umwelt of most fishes seems limited to a small surrounding pocket. Vision, though often limited by turbidity, is probably the primary sense of most species dwelling in surface-lit waters. Sound travels about five times as fast in water than in air and may be propagated for long distances. For these reasons it is often assumed that acoustic sensitivity is particularly important for aquatic vertebrates. This assumption is equivocal for most fishes, which cannot localize sounds more distant than a few meters. Banner (1972) felt lemon sharks could generally detect prey by sound at distances greater than their estimated visual range (still only about 5 m), but no one has demonstrated communication over long distances. The maximum distance a toadfish can hear the boatwhistle of another toadfish is 3 to 4 m. It is also noteworthy that sound signals have evolved in only a limited number of species. A superiority of light over sound was demonstrated in the pinfish Lagodon rhomboïdes by Jacobs and Popper (1968), who trained the fish in an avoidance task using either light or sound as the conditioned stimulus. Fish learned to avoid light in a median of four and a half training days, while it took twenty one days to avoid the sound. The authors felt these results were consistent with a primarily visual orientation of the fish. Squirrelfish and grunts learn a sound avoidance much more easily than Lagodon, though all three species are sound producers.

Olfaction is a highly developed sense for many species, but examples of its use in communication are not widespread. The number of examples should be readily extended by further studies. An electrical sense is developed in a few species. Fishes have been shown to react to an odor in water or to an electric current as if an individual emitting these stimuli were actually present.

Although we have discussed the subject on a topic by topic basis, communication typically involves several sensory channels. A courtship ritual might utilize visual, acoustic, and olfactory signals, but responses to signals of one modality indicate that redundancy, a safety factor, is built into the system.

References

Albrecht, H., 1962. Die Mitschattierung. Experientia, 18:284-86.

Albrecht, H., 1969. Behaviour of four species of Atlantic damselfishes from Colombia, South America {Abudefduf saxatiles, A. taurus, Chromis multilineata, C. cynaea, Pisces, Pomacentridae). Z. Tierpsychol., 26:662-76.

Alexander, R. McN., 1966. Physical aspects of swimbladder function. Biol. Rev., 41:141-76.

Armstrong, E. A., 1963. A study of bird song. London: Oxford University Press.

Atz, J. W., 1953. Orientation in schooling fishes. In: Proceedings of a Conference on Orientation in Animals, T. S. Schneirla, ed. Washington, D.C.: Office of Naval Research, pp. 115-30.

Ayling, A. M., and Grace, R. V., 1971. Cleaning symbiosis among New Zealand fishes. New Zeal. J. Mar. Freshw. Res., 5:205-18.

Baerends, G. P., and Baerends-van Roon, J. M., 1950. An introduction to the study of cichlid fishes. Behaviour, Suppl. 1.

Bagnara, J. T., and Hadley, M. E., 1969. The control of bright colored pigment cells of fishes and amphibians. Amer. Zool., 9:465-78.

Banner, A., 1972. Use of sound in prédation by young lemon sharks, Negaprion brevirostris (Poey). Bull. Mar. Sei, 22:251-83.

Barber, S. B., and Mowbray, W. H., 1956. Mechanism of sound production in the sculpin. Science, 124:219-20.

Bardach, J. E., and Todd, J. H., 1970. Chemical communication in fish. In: Advances in Chemoreception, vol. 1,J. W. Johnston, Jr., D. G. Moulton, and A. Turk, eds. New York: Appleton-Century-Crofts, pp.205-40.

Bardach, J. E., and Villars, T., 1974. The chemical senses of fishes. In: Chemoreception in Marine Organisms, P. T. Grant and A. M. Mackie, eds. New York: Academic Press, pp.49-104.

Barlow, G. W., 1962. Ethology of the Asian teleost, Badis badis. IV. Sexual behavior. Copeia, 1962:346-60.

Barlow, G. W., 1967. Social behavior of a South American leaf fish, Polycentrus schomburgkii, with an account of recurring pseudofemale behavior. Amer. Midi. Nat., 78:215-34.

Barlow, G. W., 1968. Ethological units of behavior. In: The Central Nervous System and Fish Behavior. D. Ingle, ed. Chicago: University of Chicago Press, pp.217-32.

Barlow, G. W., 1970. A test of appeasement and arousal hypotheses of courtship behavior in a cichlid fish, E tropins maculatus. Z. Tierpsychol., 27:779-806.

Barlow, G. W., 1974. Contrasts in social behavior between Central American cichlid fishes and coral reef surgeon fishes. Amer. Zool., 14:19-34.

Bauer, R., and Kramer, B., 1974. Agonistic behaviour in mormyrid fish: latency-relationship between the electric discharges of Gnathonemus petersii and Mormyrus rume. Experientia, 30:51-52.

Bayoumi, A. R., 1970. Underwater sounds of the Japanese gurnard, Chelidonichthys kumu. Mar. Biol., 5:77-82.

Beauchamp, R. D., and Lovasik, J. V., 1973. Blue mechanism response of single goldfish optic fibers. J; Neurophysiol., 36:925-39.

Bell, C. C.; Myers, J. P.; and Russel, J. C.; 1974. Electric organ discharge patterns during dominance related behavioral displays in Gnathonemus petersii (Mormyridae). J. Comp. Physiol., 92:201-28.

Bennet, M. V. L., 1971a. Electric organs. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp.347-491.

Bennet, M. V. L., 1971b. Electroreception. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp.493-574.

Black-Cleworth, P., 1970. The role of electrical discharges in the nonreproductive social behaviour of Gymnotus carapo (Gymnotidae, Pisces). Anim. Behav. Monogr., 3(1): 1-77.

Blair, W. F., 1958. Mating call in the speciation of anuran amphibians. Amer. Nat., 92:27-51.

Blair, W. F., 1974. Character displacement in frogs. Amer. Zool., 14:1119-25.

Brawn, V. M., 1961a. Reproductive behaviour of the cod (Gadus callarias L.). Behaviour, 18:177-98.

Brawn, V. M., 1961b. Sound production by the cod (Gadus callarias L.). Behaviour, 18:239-55.

Breder, G. M., Jr., 1941. On the reproduction of Opsanus beta Goode and Bean. Zoologica, 29:229-32.

Breder, G. M., Jr., 1959. Studies on social groupings in fishes. Bull. Amer. Mus. Nat. Hist., 117:397-481.

Breder, G. M., Jr., 1968. Seasonal and diurnal occurrences of fish sounds in a small Florida bay. Bull. Amer. Mus. Nat. Hist., 138:327-78.

Breder, G. M., Jr., and Rosen, D. E., 1966. Modes of reproduction in fishes. New York: Natural History Press.

Brown, J. H.; Cantrell, M. A.; and Evans, S. M.; 1973. Observations on the behavior and coloration of some coral reef fish (family: Pomacentridae). Mar. Behav. Physiol., 2:63-71.

Bullock, T. H., 1969. Species differences in effect of electroreceptor input on electrical organ pacemakers and other aspects of behavior in electric fish. Brain Behav. Evol., 2:85-118.

Bullock, T. H., 1973. Seeing the world through a new sense: electroreception in fish. Amer. Sei., 61:316-25.

Burghardt, G. M., 1970. Defining "communication." In: Advances in Chemoreception, vol. 1: Communication by Chemical Signals, J. W. Johnston, Jr., D. G. Moulton, and A. Turk, eds. New York: Appleton-Century-Crofts, pp.5-18.

Burkenroad, M. D., 1930. Sound production in the Haemulidae. Copeia, 1930:17-18.

Burkenroad, M. D., 1931. Notes on the sound-producing marine fishes of Louisiana. Copeia, 1931:20-28.

Cahn, P. H., 1967. Lateral Line Detectors. Bloomington: Indiana University Press.

Cahn, P. H., 1972. Sensory factors in the side-to-side spacing and positional orientation of the tuna, Euthynnas affinis, during schooling. Fish. Bull., 70:197-203.

Caldwell, D. K., and Caldwell, M. C. 1967. Underwater sounds associated with aggressive behavior in defense of territory by the pinfish, Lagodon rhomboides. Bull. S. Calif. Acad. Sei., 66:69-75.

Cheal, M., and Davis, R. E., 1974. Sexual behavior: social and ecological influences in the anabantoid fish, Trichogaster trichopterus. Behav. Biol., 10:435-45.

Cherry, C., 1957. On Human Communication: A Review, a Survey, and a Criticism. New York: Wiley.

Chien, A. K., 1973. Reproductive behaviour of the angelfish Pterophyllum scalara (Pisces: Cichlidae). II. Influence of male stimuli upon the spawning rate of females. Anim. Behav., 21:457-63.

Clark, T. A., 1970. Territorial behavior and population dynamics of a pomacentrid fish, the garibaldi, Hypsypops rubicunda. Eco I. Monogr., 40:189-212.

Clark, T. A., 1971. Territory boundaries, courtship, and social behavior in the garibaldi, Hypsypops rubicunda (Pomacentridae). Copeia, 1971:295-99.

Clarke, W. D., 1963. Function of bioluminescence in mesopelagic organisms. Nature, 198:1244-46.

Cohen, D. M., 1970. How many recent fishes are there? Proc. Calif. Acad. Sei., 38:341-46.

Cohen, M. J., and Winn, H. E., 1967. Electrophysiological observations on hearing and sound production in the fish, Porichthys notatus. J. Exp. Zool., 165:355-70.

Collette, B. B., and Talbot, F. H., 1972. Activity patterns of coral reef fishes with emphasis on nocturnal-diurnal changeover. Nat. Hist. Mus. Los Angeles Sei. Bull., 14:98-124.

Crane, J. M., Jr., 1965. Bioluminescent courtship display in the teleost Porichthys notatus. Copeia, 1965:239-41.

Darcy, G. H.; Maisel, E.; and Ogden, J. C.; 1974. Cleaning preferences of the gobie Gobisoma evelynae and G. prochilos and the juvenile wrasse Thalassoma bifasciatum. Copeia, 1974:375-79.

Daw, N. W., 1973. Neurophysiology of color vision. Physiol. Rev., 53:571-611.

DeBoer, J. N., and Heuts, B. A., 1973. Prior exposure to visual cues affecting dominance in the jewel fish, Hemichromis bimaculatus Gill 1862 (Pisces, Cichlidae). Behaviour, 44:299-321.

Delco, E. A., Jr., 1960. Sound discrimination by males of two cyprinid fishes. Tex. J. Sei., 12:48-54.

Demski, L. S., and Gerald, J. W. 1974. Sound production and other behavioral effects of midbrain stimulation in the free-swimming toadfish, Opsanus beta. Brain Behav. Evol., 9:41-59.

Demski, L. S.; Gerald, J. W.; and Popper, A. N.; 1973. Central and peripheral mechanisms of teleost sound production. Amer. Zool., 13:1141-67.

Dijkgraaf, S., 1947. Ein Tone erzeugender Fisch in Neapler Aquarium. Experientia, 3:493-94.

Eibl-Eibesfeldt, I., 1955. Über Symbiosen, Parasitismus, und andere besondere zwischenartliche Beziehungen tropischer Meeresfische. Z. Tierpsychol., 12:203-19.

Enger, P. S., 1963. Single unit activity in the peripheral auditory system of a teleost fish. Acta Physiol. Scand., 210 (Suppl.).

Enger, P. S., 1968. Hearing in fish. In: Hearing Mechanisms in Vertebrates, A. V. S. and J. Knight, eds. London: J. Churchill Ltd., pp.4-17.

Ehrlich, P. R., and Ehrlich, A. H., 1973. Coevolution: heterotypic schooling in Caribbean reef fishes. Amer. Nat., 107:157-60.

Erulkar, S. D., 1972. Comparative aspects of spatial localization of sound. Physiol. Rev., 52:237-360.

Feder, H. M., 1966. Cleaning symbiosis in the marine environment. In: Symbiosis, vol. 1, S. M. Henry, ed. New York: Academic Press, pp.327-80.

Fine, M. L., 1975. Sexual dimorphism of the growth rate of the swimbladder of the toadfish Opsanus tau. Copeia, 1975:483-90.

Fish, J. F., 1972. The effect of sound playback on the toadfish. In: Behavior of Marine Animals: Current Perspectives in Research, vol. 2: Vertebrates, H. E. Winn and B. L. Olla, eds. New York: Plenum Press, pp.386-432.

Fish, J. F., and Offutt, G. C., 1972. Hearing thresholds from toadfish, Opsanus tau, measured in the laboratory and field. J Acoust. Soc. Amer., 51:1318-21.

Fish, M. P., 1953. The production of underwater sound by the northern seahorse, Hippocampus hud-sonius. Copeia, 1953:98-99.

Fish, M. P., 1954. The character and significance of sound production among fishes of the Western North Atlantic. Bull. Bingham Oceanogr. Coll., 14:1-109.

Fish, M. P., and Mowbray, W. H., 1959. The production of underwater sounds by Opsanus sp., a new toadfish from Bimini, Bahamas. Zoologica, 44:71-76.

Fish, M. P., and Mowbray, W. H., 1970. Sounds of Western North Atlantic Fishes. Baltimore: Johns Hopkins Press.

Fishelson, L., 1970. Behaviour and ecology of a population of Abudefduf saxatilis (Pomacentridae, Teleostei) at Eilat (Red Sea). Anim. Behav., 18:225-37.

Flock, A., 1971. The lateral line organ mechanoreceptors. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp.241-63.

Fricke, H., 1966. Zum Verhalten des Putzerfisches, Labroides dimidiatus. Z. Tierpsychol., 23:1-3.

Fricke, H. W., 1973. Behaviour as part of ecological adaptation. In situ studies in the coral reef. Helgolander wiss. Merresunters., 24:120-44.

Fricke, H. W., and Holzberg, S., 1974. Social units and hermaphroditism in a pomacentrid fish. Naturwissenschaften, 61:367-68.

Frisch, K. von, 1938. Zur Physiologie des Fischschwarmes. Naturwissenschaften, 26:601-606.

Frisch, K. von, 1941. Über einem Schreckstoff der Fischaut und seine biologische Bedeutung. Z. vergl. Physiol., 29:46-145.

Fujii, R., and Novales, R. R., 1969. Cellular aspects of the control of physiological color changes in fishes. Amer. Zool., 9:453-63.

Gainer, H., and Klancher, J. E., 1965. Neuromuscular junctions in a fast-contracting fish muscle. Comp. Biochem. Physiol., 15:156-65.

Gentle, M. J., 1970a. The central nervous control of color change in the minnow (Phoxinus phoxinus L.). 1. Blinding and the effects of tectal removal on normal and blind fish. J. Exp. Biol., 54:83-91.

Gentle, M. J., 1970b. The central nervous control of color change in the minnow (Phoxinus phoxinus L.). 2. Tectal ablations in normal fish. J. Exp. Biol., 54:93-102.

Gerald, J. W., 1971. Sound production during courtship in six species of sunfish (Centrarchidae). Evolution, 25:75-87.

Göz, H., 1941. Über den Art- und Individualgeruch bei Fischen. Z. vergl. Physiol., 29:1-45.

Gray, G. A., and Winn, H. E., 1961. Reproductive ecology and sound production of the toadfish, Opsanus tau. Ecology, 42:274-82.

Greenwood, P. H.; Rosen, D. E.; Weitzman, S. H.; and Myers, G. S.; 1966. Phyletic studies of teleostean fishes, with a provisional classification of living forms. Bull. Amer. Mus. Nat. Hist., 131:341-455.

Hamilton, W. J. Ill, and Peterman, R. M., 1971. Countershading in the colourful reef fish Chaetodon lunula: concealment, communication or both. Anim. Behav., 19:357-64.

Hara, T. J., 1971. Chemoreception. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp.79-120.

Harden Jones, F. R., 1968. Fish Migration. London: Edward Arnold Ltd. 325pp.

Harris, G. G., and van Bergeijk, W. A., 1962. Evidence that the lateral line organ responds to near-field displacements of sound sources in water. J. Acoust. Soc. Amer., 12:1831-41.

Hawkins, A. D., 1973. The sensitivity offish to sounds. Oceanogr. Mar. Biol. Annu. Rev., 11:291-340.

Hawkins, A. D., and Chapman, C. J., 1966. Underwater sounds of the haddock, Melanogrammus aeglefinus. J. Mar. Biol. Assoc. U.K., 46:241-47.

Hemens, J., 1966. The ethological significance of the sword-tail in Xiphophorus hellerii (Haekel). Behaviour, 27:290-315.

Hemmings, C. C., 1966. Olfaction and vision in fish schooling. J. Exp. Biol., 45:449-64.

Hilliman, D. E., and Lewes, E. R., 1971. Morphological basis for a mechanical linkage in otolithic receptor transduction in the frog. Science, 174:416-19.

Hobson, E. S., 1968. Predatory behavior of some shore fishes in the Gulf of California. U.S. Bur. Sport Fish. Wildl. Res. Rep., 73:1-92.

Hobson, E. S., 1969. Comments on certain recent generalizations regarding cleaning symbiosis in fishes. Pac. Sei., 23:35-39.

Hobson, E. S., 1971. Cleaning symbiosis among California inshore fishes. Fish. Bull., 69:491-523.

Hobson, E. S., 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. Bull., 72:915-1031.

Hopkins, C. D., 1972. Sex differences in electric signaling in an electric fish. Science, 176:1035-37.

Hopkins, C. D., 1974a. Electric communication in fish. Amer. Sei., 62:426-37.

Hopkins, C. D., 1974b. Electric communication: functions in the social behavior of Eigenmannia virescens. Behaviour, 50:270-305.

Horch, K., and Salmon, M., 1973. Adaptations to the acoustic environment by the squirrelfishes Myripristis violaceus and M. pralinius. Mar. Behav. Physiol 2:121-39.

Hunter, J. R., 1969. Communication of velocity changes in jack mackeral (Trachurus symmetricus) schools. Anim. Behav., 17:507-14.

Hurley, A. C., and Hartline, P. H., 1974. Escape response in the damselfish Chromis cyanea (Pisces: Pomacentridae). Anim. Behav., 22:430-37.

Ingle, D., 1971. Vision: the experimental analysis of visual behavior. In: Fish Physiology, vol. 5, W. S. Hoar and D. J. Randall, eds. New York: Academic Press, pp.59-77.

Jacobs, D. W., and Popper, A. N., 1968. Stimulus effectiveness in avoidance behavior in fish. Psychon. Sei., 12:109-10.

Jacobs, D. W., and Tavolga, W. N., 1968. Acoustic frequency discrimination in the goldfish. Anim. Behav., 16:67-71.

Jenkins, T. M., Jr., 1969. Social structure, position choice and microdistribution of two trout species (Salmo trutta and Salmo gairdneri) resident in mountain streams. Anim. Behav. Monogr2:57-123.

Johnson, R. H., and Nelson, D. R., 1973. Agonistic display in the gray reef shark, Carcharhinus menisoirah, and its relationship to attacks on man. Copeia, 1973:76-84.

Karlson, P., and Luscher, M., 1959. "Pheromones": a new term for a class of biologically active substances. Nature, 183:55-56.

Kastoun, E., 1971. Elektrische Felder als Kommunikationsmittel beim Zitterwels. Naturwissenschaften, 58:459.

Keenleyside, M. H. A., 1955. Some aspects of the schooling behaviour of fish. Behaviour, 8:133-247.

Keenleyside, M. H. A., 1967. Behavior of male sunfishes (genus Lepomis) towards females of three species. Evolution, 21:688-95.

Keenleyside, M. H. A., 1971. Aggressive behavior of male longear sunfish (Lepomis megalotis). Z. Tierpsychol., 28:227-40.

Keenleyside, M. H. A., 1972a. Intraspecific intrusions into nests of spawning longear sunfish (Pisces: Cen-trarchidae). Copeia, 1972:272-78.

Keenleyside, M. H. A., 1972b. The behaviour of Abudefduf zonatus (Pisces, Pomacentridae) at Heron Island, Great Barrier Reef. Anim. Behav., 20:763-74.

Keenleyside, M. H. A., and Yamamoto, F. T., 1962. Territorial behaviour of juvenile Atlantic salmon (Salmo salar L.). Behaviour, 19:139-69.

Kinzer, J., 1961. Über die Lautäusserungen der Schwarzgrundel Gobius jozo. Aquar. Terras-Kunde, 7:7-10.

Kittredge, J. S.; Takahashi, F. T.; Lindsey, J.; and Lasker, R.; 1974. Chemical signals in the sea: marine allelochemics and evolution. Fish. Bull., 72:1-11.

Klausewitz, W., 1958. Lauterzeugung als abwehrwaffe bei der hinterindischen Tiger-schmerle (Botia hymenophysa). Natur. Volk, 88:343-49.

Kleerekoper, H., 1969. Olfaction in Fishes. Bloomington: Indiana University Press.

Kühme, W., 1963. Chemisch ausgelöste Brutpflege- und Schwarmreakionen bei Hemichromis bimaculatus (Pisces). Z. Tierpsychol., 20:688-704.

Kühme, W., 1964. Eine chemisch ausgelöste Schwarmreaktion beim jungen Cichliden (Pisces). Naturwissenschaften, 51:120-21.

Künzer, P., 1962. Die Auslösung der Nachfolgreaktion durch Bewegungsreize bei Jungfishen von Nannacara anomala Regan (Cichlidae). Naturwissenschaften, 22:525-26.

Künzer, P., 1964. Weitere Versuche zur Auslösung der Nachfolgreaktion bei Jungfischen von Nannacara anomala (Cichlidae). Naturwissenschaften, 51:419-20.

Leiner, M., 1930. Fortsetzung der ökologischen Studien an Gasterosteus aculeatus. Z. Morphol. Ökol. Tiere, 16:499-540.

Leong, C. Y., 1969. Quantitative effects of releasers on the attack readiness of the fish Haplochromis burtoni (Cichlidae, Pisces). Z vergl. Physiol., 65:29-50.

Liley, N. R., 1966. Ethological isolating mechanisms in four sympatric species of poeciliid fishes. Behaviour (Suppl.), 13:1-197.

Limbaugh, C., 1961. Cleaning symbiosis. Sei. Amer., 205:42-49.

Lorcher, K., 1969. Comparative bio-acoustic investigations in the fire-bellied toad, Bombina bombina (L.). Oecologia, 3:84-124.

Lorenz, K., 1966. On Aggression. New York: Harcourt, Brace, & World.

Losey, G. S., Jr., 1969. Sexual pheromone in some fishes of the genus Hypsoblennius Gill. Science, 163:181-83.

Losey, G. S., Jr., 1971. Communication between fishes in cleaning symbiosis. In: Aspects of the Biology of Symbiosis, R. C. Chens, ed. Baltimore: University Park Press, pp.45-76.

Losey, G. S., Jr., 1972. The ecological importance of cleaning symbiosis. Copeia, 1972:820-33.

Losey, G. S., Jr., and Margules, L., 1974. Cleaning symbiosis provides a positive reinforcer for fish. Science, 184:179-80.

Low, R. M., 1971. Interspecific territoriality in a pomacentrid reef fish Pomacentrus flavicauda Whitley. Ecology, 52:648-54.

Lowenstein, O., 1971. The labyrinth. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp.207-40.

McAllister, D. E., 1967. The significance of ventral bioluminescence in fishes. J. Fish. Res. Bd. Can., 24:537-54.

McCartney, B. S., and Stubbs, A. R., 1970. Measurements of the target strength of fish in dorsal aspect, including swimbladder resonance. In: Proceedings of an International Symposium on Biological Sound Scattering in the Ocean, G. B. Farquhar, ed. Maury Center for Ocean Science MC Report 005. Washington, D.C.: Government Printing Office, pp. 180—211.

McDonald, A..L.; Heimstra, N. W.; and Damkot, D. K.; 1968. Social modification of agonistic behavior in fish. Anim. Behav., 16:437-41.

MacGinitie, 1939. The natural history of the blind goby, Typhlogobius californiensis Steindachner. Amer. Midi. Nat., 21:489-505.

McKenzie, J. A., and Keenleyside, M. H. A., 1970. Reproductive behavior of ninespine sticklebacks (Pungitius pungitius (L.)) in South Bay, Manitoulin Island, Ontario. Can. J. Zool., 48:55-61.

McMillan, V. E., and Smith, R.J. F., 1974. Agonistic and reproductive behaviour of the fathead minnow (Pimephales promelas Rafinesque). Z. Tierpsychol., 34:25-58.

Magnuson, J. J., and Prescott, J. H., 1966. Courtship, locomotion, feeding and miscellaneous behaviour of Pacific bonito (Sarda chiliensis). Anim. Behav., 14:54-67.

Markle, H., 1971. Schallerzeugung bei Piranhas (Serrasalminae, Characidae). Z. vergl. Physiol., 74:39-56.

Marler, P., 1961. The logical analysis of animal communication.J Theoret. Biol., 1:295-317.

Marler, P., and Hamilton, W.J. Ill, 1966. Mechanisms of Animal Behavior. New York: John Wiley and Sons.

Marshall, J. A., 1966. The social behavior and role of sound production in Trichopsis vittatus, with a comparison to Trichopsis pumilis. Ph.D. diss., University of Maryland.

Marshall, J. A., 1972. Influence of male sound production on oviposition in female Tilapia mossambica (Pisces, Cichlidae). Amer. Zool., 12:663-64.

Marshall, N. B., 1962. The biology of sound-producing fishes. Symp. Zool. Soc. Lond., 7:45-60.

Marshall, N. B., 1967. Sound-producing mechanisms and the biology of deep-sea fishes. In: Marine Bioacoustics, vol. 2, W. M. Tavolga, ed. Oxford: Pergamon Press, pp. 123-33.

Morin, J. G.; Harrington, A.; Nealson, K.; Krieger, N.; Baldwin, T. O.; and Hastings, J. W.; 1975. Light for all reasons: versatility in the behavioral repertoire of the flashlight fish. Science, 190:74-76.

Morris, D., 1958. The reproductive behaviour of the ten-spined stickleback (Pygosteuspungitius L.). Behaviour (Suppl. 6).

Morrow, J. E., Jr., 1948. Schooling behaviour in fishes. Quart. Rev. Biol., 23:27-38.

Moulton, J. M., 1956. Influencing the calling of sea robins (Prionotus spp.) with sound. Biol. Bull., 111:393-98.

Moulton, J. M., 1958. The acoustical behavior of some fishes in the Bimini area. Biol. Bull., 114:357-74.

Moulton, J. M., 1960. Swimming sounds and the schooling of fishes. Biol. Bull., 119:210-23.

Moulton, J. M., 1963. Acoustic behavior of fishes. In: Acoustic Behavior of Animals, R. G. Busnel, ed. New York: Elsevier, pp.655-93.

Münz, F. W., 1971. Vision: visual pigments. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp. 1-32.

Myrberg, A. A., Jr., 1965. A descriptive analysis of the behavior of the African cichlid fish, Pelmatochromis guentheri (Sauvage). Anim. Behav., 13:312-29.

Myrberg, A. A., Jr., 1966. Parental recognition of young in cichlid fishes. Anim. Behav., 14:565-71.

Myrberg, A. A., Jr., 1972a. Ethology of the bicolor damselfish, Eupomacentrus partitus (Pisces: Pomacentridae): a comparative analysis of laboratory and field behaviour. Anim. Behav. Monogr., 5:199-283.

Myrberg, A. A., Jr., 1972b. Using sound to influence the behaviour of free-ranging marine animals. In: Behavior of Marine Animals, vol. 2, H. E. Winn and B. L. Olla, eds. New York: Plenum Press, pp.435-68.

Myrberg, A. A., Jr., 1972c. Social dominance and territoriality in the bicolor damselfish, Eupomacentrus partitus (Poey) (Pisces: Pomacentridae). Behaviour, 41:207-31.

Myrberg, A. A., Jr., 1975. The role of chemical and visual stimuli in the preferential discrimination of young in the cichlid fish, Cichlasoma nigrofasciatum (Gunther). Z Tierpsych., 37:274-97.

Myrberg, A. A., Jr.; Brahy, B. D.; and Emery, A. R.; 1967. Field observations on reproduction in the damselfish, Chromis multilineata (Pomacentridae), with additional notes on general behavior. Copeia, 1967:819-27.

Myrberg, A. A., Jr.; Kramer, E.; and Heinecke, P.; 1965. Sound production by cichlid fishes. Science, 149:555-58.

Myrberg, A. A., Jr., and Spires, J. Y., 1972. Sound discrimination by the bicolor damselfish, Eupomacentrus partitus. J. Exp. Biol., 57:727-35.

Myrberg, A. A., Jr., and Thresher, R. E., 1974. Interspecific aggression and its relevance to the concept of territoriality in reef fishes. Amer. Zool., 14:81-96.

Nakamura, E. L., and Magnuson,J.J. 1965. Coloration of the scombrid fish Euthynnus affinis (Cantor). Copeia, 1965:234-35.

Neil, E. H., 1964. An analysis of color changes and social behavior of Tilapia mossambica. Univ. Calif. Publ. Zool., 75:1-58.

Nelson, K., 1964a. Behavior and morphology in the glandulocaudine fishes (Ostariophysi, Characidae). Univ. Calif. Publ. Zool., 75:59-152.

Nelson, K., 1964b. The evolution of a pattern of sound production associated with courtship in the characid fish, Glandulocauda inequalis. Evolution, 18:526-40.

Nicol,J. A. C., 1967. The luminescence of fishes. Symp. Zool. Soc. Lond., 19:25-55.

Noakes, D. L. G., and Barlow, G. W., 1973. Ontogeny of parent-contacting in young Cichlasoma citrinellum (Pisces, Cichlidae). Behaviour, 46:221-55.

Noble, G. K., and Curtis, B., 1939. The social behavior of the jewel fish, Hemichromis bimaculatus Gill. Bull. Amer. Mus. Nat. Hist., 76:1-46.

Olla, B. L., and Samet, C., 1974. Fish-to-fish attraction and the facilitation of feeding behavior as mediated by visual stimuli in striped mullet, Mugil cephalus. J. Fish. Res. Bd. Can., 31:1621-30.

Oppenheimer, J. R., 1970. Mouthbreeding in fishes. Anim. Behav., 18:493-503.

Packard, A., 1960. Electrophysiological observations on a sound-producing fish. Nature, 187:63-64.

Peeke, H. V. S.; Herz, M.J.; and Gallagher, J. E.; 1971. Changes in aggressive interaction in adjacently territorial convict cichlids (Cichlasoma nigrofasciatum): a study of habituation. Behaviour, 40:43-54.

Pfeiffer, W., 1960. Über die Schreckreaktion bei Fischen und die Herkunft des Schreckstoffes. Z vergl. Physiol., 43:578-614.

Pfeiffer, W., 1963. Alarm substances. Experientia, 19:113-23.

Picciolo, A. R., 1964. Sexual and nest discrimination in anabantid fishes of the genera Colisa and Trichogaster. Ecol. Monogr., 34:53-77.

Popper, A. N., 1971. The effects of size on auditory capacities of the goldfish. J Aud. Res., 11:239-47.

Popper, A. N., 1972. Auditory threshold in the goldfish (Carassius auratus) as a function of signal duration. J. Acoust. Soc. Amer., 52:596-602.

Popper, A. N., 1974. The response of the swim bladder of the goldfish (Carassius auratus) to acoustic stimuli, J Exp. Biol., 60:295-304.

Popper, A. N., and Fay, R. R., 1973. Sound detection and processing by teleost fishes: a critical review.J Acoust. Soc. Amer., 53:1515-29.

Popper, A. N.; Salmon, M.; and Parvulescu, A.; 1973. Sound localization by the Hawaiian squirrelfishes, Myripristis berndti and M. argyromus. Anim. Behav., 21:86-97.

Potts, G. W., 1973. Cleaning symbiosis among British fish with special reference to Crenilabrus melops (Labridae).J. Mar. Biol. Assoc. U.K., 53:1-10.

Poulter, T. C., 1968. Marine mammals. In: Animal Communication, T. A. Sebeok, ed. Bloomington: Indiana University Press, pp.405-65.

Radakov, D. V., 1973. Schooling in the Ecology of Fish. New York: John Wiley and Sons.

Rasa, O. A. E., 1969. Territoriality and the establishment of dominance by means of visual cues in Pomacentrus jenkinsi (Pisces: Pomacentridae). Z Tierpsychol., 26:825-45.

Reed, J. R., 1969. Alarm substances and fright reaction in some fishes from the Southeastern United States. Trans. Amer Fish. Soc., 98:664-68.

Reinboth, R., 1972. Aormonal control of the teleost ovary. Amer. Zool., 12:307-24.

Reinboth, R., 1973. Dualistic reproductive behavior in the protogynous wrasse Thalassoma bifasciatum and some observations on its day-night changeover. Helgoländer wiss. Meersunters, 24:174-91.

Reese, E. S., 1964. Ethology and marine zoology. Oceanogr. Mar. Biol. Annu. Rev., 2:455-88.

Reese, E. S., 1973. Duration of residence by coral reef fishes on "home" reefs. Copeia, 1973:145-49.

Richard, I. S., 1974. Caudal neurosecretory system: possible role in pheromone production.J. Exp. Zool., 187:405-8.

Rigley, L., and Marshall, J. A., 1973. Sound production by the elephant-nose fish, Gnathonemus petersi (Pisces, Mormyridae). Copeia, 1973:134-35.

Robertson, D. R., 1972. Social control of sex reversal in a coral-reef fish. Science, 177:1007-9.

Robertson, D. R., 1973. Field observations on the reproductive behavior of a pomacentrid fish Acanthochromis poly acanthus. Z. Tierpsychol., 32:319-24.

Rosenblatt, R. H., and Losey, G. S., Jr., 1967. Alarm reaction of the top smelt, Atherinops affinis: reexamination. Science, 158:671-72.

Russell, B. C., 1971. Underwater observations on the reproductive activity of the damoiselle, Chromis dispilus (Pisces: Pomacentridae). Mar. Biol., 10:22-29.

Salmon, M., 1967. Acoustical behavior of the menpachi, Myripristis berndti, in Hawaii. Pac. Sei., 21:364-81.

Salmon, M., and Winn, H. E., 1966. Sound production by priacanthid fishes. Copeia, 1966:869-72.

Salmon, M.; Winn, H. E.; and Sorgente, N.; 1968. Sound production and associated behavior in triggerfish. Pac. Sei., 22:11-20.

Sand, O., and Hawkins, A. D., 1973. Acoustic properties of the cod swimbladder. J. Exp. Biol., 58:797-820.

Sato, M., 1937. On the barbels of a Japanese sea catfish, Plotosus anguillaris (Lacepede). Sei. Rep. Tohuku Imp. Univ. Ser. 4, 11:323-32.

Schleidt, W. M., 1973. Tonic communication: continual effects of discrete signs in animal communication systems. J. Theor. Biol., 42:359-86.

Schneider, H., 1964a. Bioakustiche Untersuchungen an Anemonenfishen der Gattung Amphiprion (Pisces). Z. Morphol. Ökol. Tiere, 53:453-74.

Schneider, H., 1964b. Physiologische und Morphologische Untersuchungen zur Bioakustik der Tigerfische (Pisces, Theraponidae). Z. vergl. Physiol., 47:493-558.

Schneider, H., 1967. Morphology and physiology of sound-producing mechanisms in teleost fishes. In: Marine Bio-acoustics, vol. 2, W. N. Tavolga, ed. New York: Pergamon Press, pp. 135-58.

Schneider, H., 1968. Bio-akustiche Untersuchungen am Mittelmeerlaubfrosch. Z. vergl. Physiol., 61:369-85.

Schneider, H., 1974. Structure of the mating calls and relationships of the European tree frogs (Hylidae, Anura). Oecologia, 14:99-110.

Schneider, H., and Hasler, A. D., 1960. Laute und Lauterzeugung beim Süsswassertrommler Aplodinotus grunniens Rafinesque. Z. vergl. Physiol., 43:499-517.

Schuijf, A., and Siemelink, M. E., 1974. The ability of cod (Gadus morhua) to orient towards a sound source. Experientia, 30:773-74.

Schwarz, A., 1974a. Sound production and associated behavior in a cichlid fish, Cichlasoma centrarchus. I. Male-male interactions. Z. Tierpsychol., 35:147-56.

Schwarz, A., 1974b. The inhibition of aggressive behavior by sound in the cichlid fish, Cichlasoma centrarchus. Z. Tierpsychol., 35:508-17.

Schwarz, A. Sound production and associated behavior in Cichlasoma centrarchus. II. Breeding pairs. In preparation.

Seghers, B. H., 1974. Schooling behavior in the guppy (Poecilia reticulata): an evolutionary response to prédation. Evolution, 28:486-89.

Shaw, E., 1960. The development of schooling behavior in fishes. Physiol. Zool., 33:79-86.

Shaw, E., 1969. The duration of schooling among fish separated and those not separated by barriers. Amer. Mus. Novitates, 2373. 13pp.

Shaw, E., 1970. Schooling in fishes: critique and review. In: Development and Evolution of Behavior, L. R. E. Tobach, D. S. Lehrman, and j. S. Rosenblatt, eds. San Francisco: W. H. Freeman, pp.452-80.

Skinner, W. A.; Mathews, R. D.; and Parkhurst, R. M.; 1962. Alarm reaction of the top smelt, Atherinops affinis (Ayres). Science, 138:681-82.

Skoglund, C. R., 1961. Functional analysis of swimbladder muscles engaged in sound production of the toadfish. J. Biophys. Biochem. Cytol. (Suppl.) 10:187-200.

Smith, H. M., 1905. The drumming of the drum-fishes (Sciaenidae). Science, 22:376-78.

Springer, S., 1967. Social organization of shark populations. In: Sharks, Skates, and Rays, P. W. Gilbert, R. F. Mathewson, and D. P. Dall, eds. Baltimore: The Johns Hopkins University Press, pp. 149-74.

Steele, R. G., and Keenleyside, M. H. A., 1971. Mate selection in two species of sunfish (Lepomis gibbosus and L. Megalotis peltastes). Can. J. Zool., 49:1541-48.

Stout, J. F., 1963. The significance of sound production during the reproductive behavior of Notropis analos tanus (family Cyprinidae). Anim. Behav., 11:83-92.

Stout, J. F., 1966. Sound communication in fishes with special reference to Notropis analostanus. Proc. 3rd Annu. Conf. Biol. Sonar and Div. Mam., 3:159-77.

Suzuki, R., 1961. Sperm activation and aggregation during fertilization in some fishes. VI. The origin of the sperm-stimulating factor. Annot. Zool. fapon., 34:24-29.

Swerdloff, S. N., 1970. Behavioral observations on Eniwetok damselfishes (Pomancentridae: Chromis) with special reference to the spawning of Chromis caeruleus. Copeia, 1970:371-74.

Tavolga, W. N., 1956. Visual, chemical and sound stimuli as cues in the sex discriminatory behavior of the gobiid fish Bathygobins soporator. Zoologica, 41:49-64.

Tavolga, W. N., 1958a. The significance of underwater sounds produced by males of the gobiid fish. Bathygobius soporator. Physiol. Zool., 31:259-71.

Tavolga, W. N., 1958b. Underwater sounds produced by males of the blenniid fish, Chasmodes bosquianus. Ecology, 39:759-960.

Tavolga, W. N., 1958c. Underwater sounds produced by two species of toadfish, Opsanus tau and Opsanus beta. Bull. Mar. Sei., 8:278-84.

Tavolga, W. N., 1960. Sound production and underwater communication in fishes. In: Animal Sounds and Communication, W. E. Lanyon and W. N. Tavolga, eds. Amer. Inst. Biol. Sei.: Publ. No. 7 (Washington, D.C.), pp.93-136.

Tavolga, W. N., 1962. Mechanisms of sound production in the ariid catfishes Galeichthys and Bagre. Bull. Amer. Mus. Nat. Hist., 124:1-30.

Tavolga, W. N., 1964. Sonic characteristics and mechanisms in marine fishes. In: Marine Bio-acoustics, vol. 1, W. N. Tavolga, ed. New York: Pergamon Press, pp. 195-211.

Tavolga, W. N., 1968a. Fishes. In: Animal Communication, T. A. Sebeok, ed. Bloomington: Indiana University Press, pp.271-88.

Tavolga, W. N., 1968b. Marine animal data atlas. Tech. Rep., Naval Training Device Center (Orlando, Fla.), 1212-2. 239pp.

Tavolga, W. N., 1970. Levels of interaction in animal communication. In: Evolution and Development of Animal Behavior, L. R. Aronson, E. Tobach, D. S. Lehrman, and J. S. Rosenblatt, eds. San Francisco: Freeman, pp.281-302.

Tavolga, W. N., 1971a. Sound production and detection. In: Fish Physiology, vol. 5, W. S. Hoar and D.J. Randall, eds. New York: Academic Press, pp. 135-205.

Tavolga, W. N., 1971b. Acoustic orientation in the sea catfish. Galeichthys felis. Ann. N.Y. Acad. Sei., 188:80-97.

Tavolga, W. N., 1974. Sensory parameters in communication among coral reef fishes. Mount Sinai J. Med., 41:324-40.

Tembrock, G., 1968. Land mammals. In: Animal Communication, T. A. Sebeok, ed. Bloomington: Indiana University Press, pp.338-404.

Tett, P. B., and Kelly, M. G., 1973. Marine bioluminescence. Oceanogr. Mar. Biol. Annu. Rev., 11:89-173.

Tinbergen, N., 1951. The study of instinct. Oxford: Clarendon Press.

Todd, J. H.; Atema, J.; and Bardach, J. E.; 1967. Chemical communication in social behavior of a fish, the yellow bullhead (Ictalurus natalis). Science, 158:672-73.

Tomita, T., 1971. Vision: Electrophysiology of the retina. In: Fish Physiology, vol. 5 W. S. Hoar and D. J. Randall, eds. New York: Academic Press, pp.33-57.

Topp, R. W., 1970. Behavior and color change of the rudderfish, Kyphosus elegans, in the Gulf of Panama. Copeia, 1970:763-65.

Tower, R. W., 1908. The production of sound in the drumfishes, the sea robin and the toadfish. Ann. N. Y. Acad. Sei, 18:149-80.

Uematsu, T., 1971. Social facilitation and feeding behavior of the guppy. II. Experimental analysis of mechanisms, fap. J. Ecol., 21(l-2):54-67.

van den Assem, J. and van der Molen, J., 1969. Waning of the aggressive response in the three-spined stickleback. I. A preliminary analysis of the phenomenon. Behaviour, 34:268-324.

Verheijen, F. J., 1963. Alarm substance in some North American cyprinid fishes. Copeia, 1963:174-76.

Walls, G. L., 1942. The vertebrate eye and its adaptive radiation. Bloomfield Hills, Michigan: Cranbrook Inst. Sei.

Walters, V., and Robins, C. R., 1961. A new toadfish (Batrachoididae) considered to be a glacial relict in the West Indies. Amer. Mus. Novit., 2047.

Welty, J. C., 1934. Experiments in group behavior of fishes. Physiol. Zool., 7:85-128.

Wickler, W., 1962. "Egg-dummies" as natural releasers in mouth breeding cichlids. Nature, 194:1092-93.

Wilz, K. J., 1970. Causal and functional analysis of dorsal pricking and nest activity in the courtship of the three-spined stickleback Gasterosteus aculeatus. Anim. Behav., 18:115-24.

Winn, H. E., 1964. The biological significance of fish sounds. In: Marine Bio-acoustics. vol. 1, W. N. Tavolga, ed. New York: Pergamon Press, pp.213-31.

Winn, H. E., 1967. Vocal facilitation and thebiological significance of toadfish sounds. In: Marine Bio-acoustics, vol. 2, W. N. Tavolga, ed. New York: Pergamon Press, pp.283-304.

Winn, H. E., 1972. Acoustic discrimination by the toadfish with comments on signal systems. In: Behavior of Marine Animals: Current Perspectives in Research, vol. 2: Vertebrates, H. E. Winn and B. L. Olla, eds. New York: Plenum Press, pp.361-85.

Winn, H. E., and Marshall, J. A., 1963. Sound producing organ of the squirrelfish, Holocentrus rufus. Physiol. Zool., 36:34-44.

Winn, H. E.; Marshall, J. A.; and Hazlett, B. A.; 1964. Behavior, diel activities, and stimuli that elicit sound production and reaction to sounds in the longspine squirrelfish. Copeia, 1964:413-25.

Winn, H. E., and Stout, J. F., 1960. Sound production by the satinfin shiner, Notropis analostanus, and related fishes. Science, 132:222-23.

Wyman, R. L., and Ward, J. A., 1972. A cleaning symbiosis between the cichlid fishes Etroplus maculatus and Etroplus suratensis. I. Description and possible evolution. Copeia, 1972:834-38.

Yager, D., 1968. Behavioral analysis of color sensitivities in the goldfish. In: The Central Nervous System and Fish Behavior, D. Ingle, ed. Chicago: University of Chicago Press, pp.25-33.

Yanagimachi, R., 1957. Some properties of the spermactivating factor in the micropyle area of the herring egg. Annot. Zool. Japon., 30:114-19.

Youngbluth, M. J., 1968. Aspects of the ecology and ethology of the cleaning fish Labroides phthirophagus Randall. Z Tierpsychol., 25:915-32.

Zumpe, D., 1965. Laboratory observations on the aggressive behaviour of some butterfly fishes (Chaetodontidae). Z Tierpsychol., 22:226-36.