CHAPTER 7

VISUAL SYSTEMS

PETER MARLER

Central to any definition of communication is the reception by an organism of information conveyed by a stimulus from the external world. However, this encompasses not only stimulus exchange between members of the same species but also exchanges with the inanimate environment. The whole subject of exteroception is thus included. How can we restrict the definition to phenomena which we intuitively accept as communicative? The basic components of a communication system are a sender, a receiver, and a medium or channel for signal transmission. In stimulus exchanges with the environment, or exchanges between an animal and its prey, the relationship between sender and receiver is one-sided; while one participant tries to maximize the efficiency of the stimulus exchange, the other is at best neutral and often seeks to minimize it. In true communication, however, both participants seek to maximize the efficiency of information transfer. Usually two-way communication is possible.

Defined in this way, communication systems have value for both sender and receiver. We usually find this synergistic interplay between sender and receiver only when both are members of the same species. However, it can also occur in exchanges between different species. In some cases this is obvious, as in the relationship between honey bees and the flowers of plants that they pollinate. In these and other ways, the communication systems of animals may be in some degree affected by the relationship that they bear to other species, a subject that we shall return to later. For the moment we shall restrict ourselves to intraspecific communication.

Visual communication signals are used in varying degrees by most vertebrates. In some, such as certain fish, lizards, many birds, and primates, vision is probably the most important sensory modality for communication. Among invertebrates only certain arthropods use visual communication to an equivalent degree, particularly some insects and Crustacea. What distinctive property of visual systems can we discern that leads to this extensive exploitation? Most important is surely the directionality of light and the consequent ease with which an animal with suitable light receptors can locate the source of a visual signal. The same considerations lead predators to rely ex- tensively on vision for locating prey, so creating something of a dilemma for prey animals that seek to exploit visual systems of communication. The conflict between a demand for conspicuousness and the advantage of crypticity has a strong influence in shaping the evolution of visual signals.

PROPERTIES OF VISUAL SIGNALS

The efficiency of a communication system is partly a function of the number of different signals generated and how the signal types are related to one another, whether they form discrete categories or grade into one another. Optical systems are uniquely suited to provide a very rich variety of signals. The stimulus variables that are available and perceptible to animals include intensity or brightness, the wavelength or color, the degree and plane of polarization, the spatial pattern in which different stimuli are presented, and the temporal pattern of stimulus delivery. With the exception of the plane of polarization, all these variables are employed in the communication systems of some species. Their exploitation depends, in the first place, on the species possessing receptors that can perceive them. Obviously, a species that lacks color vision cannot exploit wavelength as a variable in intraspecific communication except as a way of varying brightness. It also depends on the particular selective influences to which a signal is subject.

The structure of visual communication signals in a given context will be very much a function of such considerations as the needs for successful transmission to the receiver, accurate identification of signal type, selection among the potential audience of a certain class of respondents, and particular types of responses elicited by signals in the respondent. Consider first the problems of transmission.

TRANSMISSION OF VISUAL SIGNALS

If a species is to achieve a successful system of communication, it is necessary that respondents be able to distinguish the signals from other perceptible events in the environment at the appropriate distance. The signals must be perceived against a certain background, and thus require a degree of conspicuousness and improbability (Lorenz, 1935). Unlike acoustic signals, the intensity of visual signals cannot normally be varied above a certain maximum except in those few organisms that generate their own light source. Intensity will usually be limited by the level of ambient lighting, though brightness can be maximized by use of reflecting surfaces, as the Anna’s hummingbird does in its display dive by placing the respondent between itself and the sun (Hamilton, 1965) (Fig. 1). Most animals are driven to other means of achieving conspicuousness, particularly when signals must be transmitted over a considerable range. The patterns of light, dark, and color in the displays of many animals are obviously designed to achieve such conspicuousness.

It is possible that some patterns have an inherently high attentioncatching quality to a variety of species. Fantz (1965) has shown that when monkeys or children are presented with pairs of figures, they are most likely to fixate on those with a complex outline. A star is fixated more than a circle, a checkerboard more than a simple square, a bull’s-eye more than a pattern of stripes, and so on. Similarly, Blest (1957a, b) has evidence that bull’s-eye patterns are inherently conspicuous to birds. The wing eyespots exposed in the alarm displays of a variety of butterflies and moths seem designed to exploit this conspicuousness and serve to alarm predatory birds. Honey bees find such bull’s-eye patterns attractive, for such designs induce visits by feeding workers. Many intraspecific displays generate patterns that share some of these properties of complexity of outline, concentric markings, and a central focus. Perhaps a widespread property of the visual systems of animals is being exploited here.

If signals are required to operate over distances, the need for conspicuousness is likely to dominate their physical structure. At close range the problems of transmission are reduced and there is correspondingly greater latitude in the physical properties that are consistent with successful transmission. It thus becomes important to know the natural circumstances in which a signal is used, for example, within a close-knit, coherent group or in a highly dispersed population.

SIGNAL IDENTIFICATION

Closely related to signal transmission is the problem of distinguishing between the Various signals that members of a species employ. Signals used over long distances, susceptible to interference from the environment, must be distinctively different from one another if they are to be accurately and quickly identified. At long range, a respondent will be unable to perceive subtle variations in signal structure; there is likely to be a trend toward highly stereotyped signals. In other situations, small variations in signal structure can have great value for communicating certain types of information and may be exploited when circumstances permit.

The relevance of signal stereotypy to the efficiency of communication will also be affected by the number of signals that must be discriminated. This number includes not only those generated by members of a species but also those of other species living in the same community and active at the same time and place. When function requires accurate discrimination of a signal from the signals of another species, the existence of a large number of competing stimulus patterns is also likely to favor stereotypy, at least in some signal properties. However, the fact that most visual signals comprise several separate components which can be independently varied to some extent makes it possible to combine stereotyped and more variable elements in the same signal type.

The advantages of highly graded signal patterns are most likely to be exploited at close range. Not only are transmission problems minimal, but the question of specific identification is likely to have been already solved during the process which initially brought the animals into proximity. The elaborate and highly graded nature of many of the visual signals used for intraspecific communication in various group-living primates may be explicable in this way.

We find a spectrum of degrees of variability within a given signal category, some being highly variable so that the category itself may be even hard to define. Others are so stereotyped that it becomes a matter of interest to define the physiological basis of their regularity. Dane and his colleagues have called attention to the extraordinary stereotypy of some displays of ducks (Dane et al., 1959; Dane and Van der Kloot, 1964).

CONSEQUENCES OF SIGNAL RECEPTION

Without having access to physiological evidence, the investigator’s only clue that the signal has in fact been received will be a change in the behavior of the respondent. Although the accumulated observations of the field observer can be a good guide to the occurrence of communication, it is difficult to get quantitive supporting evidence. A thorough statistical picture of the likelihood that the behavior of the respondent might change without reception of a signal is necessary before a confident statement about the effect of the signal can be made. Hazlett and Bossert (1965) have shown how the occurrence of visual communication in hermit crabs can be so established. Long sequences of observations are made of the actions of crabs which can see each other. Analysis shows that certain actions effect predictable changes in the behavior of other crabs. For example, in one species the action “major cheliped extension,” is very likely to be followed by retreat of a respondent (Fig. 2).

By similar methods, Stokes (1962a,b) and Nelson (1964) have shown how aggressive postures in birds and courtship actions of fish, respectively, change the probability that respondents will behave in a certain way. Demonstrations of communication occurring under natural conditions are thus possible in spite of the confusing effects of the many uncontrolled variables that also affect the behavior. By experimentation the effect of these other variables can be controlled further, giving a more precise picture of the effects of signal perception. There are many analyses of small portions of the visual communication system of certain species done in this way, as, for example, the studies of social responses of roosters to models of hens (Carbaugh et al., 1962) and the aggressive and sexual responses of anabantid fishes to artificial models of varying shape and color (Picciolo, 1964).

SELECTION OF A RESPONDENT

If we observe the effect of emission of a visual signal in a group of animals under natural conditions, as often as not only certain individuals within the group will respond. Sometimes we can see the appropriateness of this selection of a certain class of respondents. The point is obvious with signals that accompany readiness to mate, for example, or cries of a hungry infant. The specification may be more refined than simply members of a species, sex, or age class. It may extend to certain individuals, or it may specify dominance status or some other social role. In addition to such long-term conditions, the specification may also refer to transient states. Only individuals in a certain physiological state may respond, such as those that are hungry or in reproductive condition.

The more elaborate the social structure and the more refined the social relationships between its individual members, the larger will be the number of respondents to be specified within the system of communication that the species possesses. The more potential classes of respondents signals must specify, the larger the repertoire of signal types needed. By the same token, it may have communicative advantage for animals to transmit signals that characterize the class to which they belong, whether defined by sex, age, dominance status, physiological state, and so on.

As a further complication, the specification of a respondent may also include the environmental circumstances in which the individual is placed. For example, the threat display of a male bird may elicit a different response from other reproductive adult males depending on whether the latter are within their own territory or outside it. Similarly, an alarm signal may elicit precipitant flight from respondents out in the open and little visible response from animals of the same class which are in cover. The same alternatives occur in monkeys, as Struhsaker (1967) found in the vervet Cercopithecus aethiops. Whenever social behavior is organized around external referents such as a nesting place, a feeding site, or space to live, there will be repercussions on the structure of the communication system used.

SELECTION OF A RESPONSE

Given that the reception of a visual signal elicits some change in the behavior of a respondent, what can we say of the different types of response that may occur? It is common to classify signals as sexual, aggressive, escape, and so on, according to the general type of response that they elicit. However, we should note that the first change in the behavior of a respondent on receipt of a signal is often less specific than this type of classification would imply. After becoming alert, the first action of a respondent is often some change in spatial relationship to the signaling animal, approach or withdrawal. We may be hard put to detect any sign of a more specific response until later in the behavioral sequence.

Both aggressive and sexual signals may elicit approach of a specified class of animal as a first response. Sometimes, it is true, one may also see signs of a more specific response, such as preparatory movements for attack or for copulation. Often, however, such signs are initially lacking and appear only when the animals get into closer proximity. The maintenance of a certain pattern of spatial distribution is, after all, an important function of the social behavior of many species. It is important not to lose sight of this by prematurely assuming that the first response elicited by signals in nature is necessarily a highly specific one. This is important not only in considering the function of communication signals but also in trying to work out their motivational basis as Schneirla (1959, 1965) has pointed out. Much of the social behavior of primates can be interpreted as functioning to maintain patterns of spatial distribution of individuals and troops (Marler, in press).

Bearing in mind the generalized nature of the first response to many communication signals, we can now return to the more widely used functional categories. Can we see any correlations between the physical structure of visual communication signals and the type of function that they serve—aggressive, sexual, parental, and so on?

ALARM SIGNALS

When danger threatens, animals commonly become alert and then take flight. The same actions may be visible to companions and often serve to communicate alarm without any further specialization for a signal function. Some species do have special visual signals for alarm. Flicking movements of the extremities in birds and fish may serve to alarm companions. Jerking of the tail seems to serve as an alarm signal in ground and tree squirrels (Eibl-Eibesfeldt, 1951; King, 1955; Ewer, 1966). Similarly, in many ungulates the actions of fleeing are accompanied by tail movements that seem to signal alarm. The movements are often accentuated by special markings, sometimes by a distinctive gait (Fig. 3) (e.g., Walther, 1964a, b). However, in general there seems to be little elaboration of special visual signals for communicating alarm. There may be a good reason for this. Vision is the very modality by which a predator is likely to be hunting, and to generate conspicuous visual signals might place the animal in serious danger. This is perhaps the explanation for the greater emphasis upon sound and chemical signals for disseminating alarm.

AGGRESSIVE SIGNALS

Visual stimuli from another member of a species will often elicit withdrawal. The most obvious are those generated by aggressive displays or combat, but there are many others. Indeed, it is not easy to draw a line between aggressive displays and any stimulus that elicits withdrawal. In some species, more or less durable characteristics of the external morphology will suffice to elicit withdrawal in a respondent, without any special posture or other behavior.

One example occurs in the European chaffinch (Marler, 1955a, b, 1956, 1957). In winter, flocks of these birds maintain a certain spacing. By providing movable food hoppers, the minimal distance at which individuals will tolerate each other without fighting can be measured, and proves to be smaller between a male and a female chaffinch than between two males. Males and females differ in appearance, the male’s breast being reddish brown and that of the female gray. The significance of this difference in coloration in eliciting withdrawal can be shown by coloring the breasts of females red. These disguised females will then be tolerated by males only at the same distance as other males would be. Evidently the red breast alone suffices to elicit withdrawal under these conditions. Variations in the size of the comb in hens may have a similar effect in a flock of chickens, a larger comb eliciting withdrawal (Collias, 1944; Marks et al., I960).

In some species such static sustained signals generated by the external appearance may be sufficient to elicit withdrawal, as Eisenberg (1964) has found in shrews. However, in chaffinches arid chickens they are often accompanied by special postures and movements that reinforce the withdrawal response in others. The actions employed are many and varied, and it is not easy to generalize about their characteristics. Many of them may have a largely arbitrary structure, as Lorenz (1935, 1951) has pointed out. However, we can sometimes discern some general principles. Many aggressive signals involve an increase in the apparent size of the signaling animal. The head is often held high, with the legs extended. Feathers and hair may bristle. In fish the fins are often expanded and the gill covers raised. Expansion of the body may occur, as in the lateral aggressive display of many lizards (Evans, 1961).

However, this is only a part of the story. Another trend is toward presentation of the main fighting weapons, often in postures which are preparatory to their use in combat. Darwin (1872) gave many examples within his category of so-called serviceable associated habits. In different birds we can see emphasis on the beak, wings, or feet, depending on the special structure of the species. In many mammals the teeth are used, but in some the feet or horns provide the major weapons, and again the postures adopted differ accordingly. There are often special markings or coloration which throw further emphasis on the structure displayed.

Although these principles are widespread, they leave many details of the structure of aggressive displays unexplained. A striking variety of animals have at least two different aggressive displays, often more, and we still know little of the functions that they serve. In many birds, for example, there is one aggressive display in which the head is held high, a so-called “head-up posture,” and another in which the head is held low and the beak open and pointing toward the opponent, a socalled “head-forward posture.” These two types are widespread in passerine birds (Andrew, 1961), and they also occur in gulls (Andrew, 1961; Tinbergen, 1959). The head-up posture seems designed to increase apparent size, while the head-forward posture more obviously displays the weapons and prepares for attack. Perhaps the intensity of opposition determines which of these two alternatives is more appropriate in a given situation. The same alternatives can be discerned in some mammals such as moose (Geist, 1963), rats (Barnett, 1958) and other rodents (Eisenberg, 1963).

In one of the few direct attempts to demonstrate the role of postures in eliciting withdrawal of an opponent, Stokes (1962a, b) studied the consequences of some postures of titmice of the genus Parus. Certain actions clearly shift the probability that an opponent will withdraw or attack, the efficacy apparently varying according to circumstances, which under field conditions are out of control. This is always a difficulty with field observations, and one may hope eventually to be able to subject the problem to experimental analysis, perhaps by the use of movable models or motion pictures. This method might also serve to reveal the significance of another characteristic of many aggressive displays, namely, their variability. In contrast with many sexual displays which are highly stereotyped, aggressive displays of many animals seem to have a graded quality, often varying along a continuous spectrum from attack to flight. The manifestation in behavior of subtle changes in the motivation of the signaling bird should add to the potential information content of such signals, providing that the receiver can perceive them accurately.

SUBMISSIVE SIGNALS

Just as aggressive signals tend to encourage the withdrawal of the respondent, so there is another set of signals which are used to maintain proximity with others or to elicit approach. Sometimes this approach will be succeeded by another set of actions with some independent significance for survival such as suckling or nursing; food sharing, exchange, or stealing; genital contacts of various kinds; and so on. At other times, signals seem to occur primarily to facilitate proximity. One prerequisite of such a signal is, of course, that it should discourage aggressive or avoidance behavior. Thus we commonly see in submissive or appeasement behavior a tendency to minimize the generation of stimuli which might elicit aggression. The postures used are often the reverse of those employed by the same species in aggressive display. Darwin (1872) drew attention to this in his principle of antithesis. Some examples are widespread, such as the direct frontal gaze at the opponent in aggressive display compared with aversion of the gaze or a shifty way of looking around in submissive behavior. Sudden vigorous movement is often avoided in submissive displays, and the weapons are often concealed to some degree as well. While such actions may reduce the probability of aggression, they do not necessarily facilitate the approach of a companion.

Apart from sequels to approach that have some obvious independent significance, such as parental or reproductive behavior, one of the most common activities following an approach sequence seems to be allogrooming (as contrasted with autogrooming, when an animal grooms itself). Members of many social species spend too much time in this activity, to be explicable solely in terms of care of the body surface. Several investigators have independently pointed out the reduction of aggressive behavior during this social grooming. Allogrooming occurs commonly in social estrildine finches (Sparks, 1964, 1965). Harrison (1965) reviews the occurrence of this behavior in different bird families, and finds some common characteristics in the species that engage in allogrooming. Most characteristically, close proximity of individuals is enforced by some property of the environment or social system, as when species such as penguins, shearwaters, gannets, and cliff-nesting marine birds breed in restricted nesting sites. Other correlated factors seem to be the long absence of one mate from the nest, a long pair-bond, and sexual monomorphism, all perhaps to be construed as requiring special development of bond-forming activities and signals for the reduction of potential aggressiveness.

Similarly, in social mammals Walther (1964) finds that allogrooming is common in social ungulates of the genus Tragelaphus. He notes that it seems to serve a placatory function (Fig. 4). In social rodents, allogrooming often seems to elicit and maintain proximity between individuals (King, 1955; Barnett, 1958; Eisenberg, 1963). Among primates at all phyletic levels, from macaques, baboons, and chimpanzees (Hall and DeVore, 1965; Sade, 1965; Goodall, 1965, in press) to lemurs (Jolly, 1966), allogrooming along with several other types of contact behavior plays a prominent role in the maintenance of peace and cohesion within the troop (Andrew, 1964). It seems clear that bodily contact has rewarding consequences in primates quite apart from any sexual connotations (Falk, 1958). The gestures used in greeting behavior by higher primates such as the chimpanzee (Goodall, in press) often involve actual contact or movements which are normally preludes to contact. Both in parent-young interactions and in relationships between adult animals, visual signals that introduce various kinds of tactile stimulation seem to have a very important social function.

Visual signals which are preludes to other types of social activity such as food exchange or mounting behavior have also been incorporated into the patterns of submissive activity for facilitating proximity and averting aggression. In gulls, the hunched posture associated with food begging seems to have a submissive function (Tinbergen, 1959). Courtship feeding plays a role in bringing mates together in many birds (Lack, 1940). In several primates, sexual presentation seems to function in part to avert potentially aggressive interaction (Altmann, 1962; Hall and DeVore, 1965). The submissive function is similar in principle to that of the grooming function; namely it elicits some rewarding activity which requires proximity and employs motor patterns which are incompatible with aggressive behavior.

SEXUAL SIGNALS

The series of actions leading to the fertilization of an egg are peculiarly significant for the survival of a species, and visual signals figure prominently in bringing together potential mates at the appropriate time. The stage in the life history at which pair bonds first become established varies widely. In some species, the sequences begin early in youth. Many social primates, for example, are born into the group from which they themselves will select a mate. In the hamadryas baboon, males bring some females into their harem while they are still little more than infants (Kummer, in press).

Often bonds are established in adulthood, but several weeks before the act of fertilization will occur. There are still other animals in which the bond between a mated pair is very brief, only as long as is necessary for them to identify each other, achieve copulation, and separate. The function required of sexual signals varies somewhat according to which type of pair-bonding prevails. In all cases the function of reproductive isolation must be served, calling for some degree of specific distinctiveness in the signals used. However, if the interval between pairing and mating is long, there is ample opportunity for correction of errors that might arise. With a short bond, the danger of errors in mating is increased, and the pressure on specific distinctiveness is correspondingly greater (Mayr, 1942; Sibley, 1957). In such species the sexes are often strikingly different in external appearance and in the behavior patterns they employ in courtship. Such birds as manakins, hummingbirds, birds of paradise, and many ducks come into this category.

Often the visual stimuli generated by distinctive differences in the external appearance of males and females seem to suffice to elicit sexual behavior in the appropriate circumstances. For example, Howell and Bartholomew (1952) found that a model of a female brewer blackbird would elicit copulation from adult males, provided it met certain minimum requirements. Wings were not necessary. Either a head or a tail had to be present but one or the other could be removed without eliminating the response. Further removal of parts inhibited the mating reaction. If the tail was present, it had to be at an angle above the horizontal. Eye color was not important, even though the males and females differ strikingly in this trait. The plumage color had to be predominantly that of a female, but a female head and neck on an otherwise male-colored dummy was effective when the male was in a state of high sexual excitement.

In four species of gull which have similar displays and differ only slightly in external appearance, Smith (1966) has shown that the color or contrast of a fleshy ring around the eye is significant. In two sympatric species the ring is yellow in one, the glaucous gull, and purple in the other, Thayer’s gull. By capturing paired females in the field and releasing them after painting the eye ring to resemble the other species, he could disrupt the pairs. Although males would establish a pair-bond with a disguised female, they refused to copulate and eventually the pairs broke up. Controls painted with their own color were undisturbed. At an earlier stage of the breeding cycle, the females also exerted a choice. At pair formation, females would refuse to pair with males painted with the eye color of the other species. By an ingenious combination of these experiments Smith could establish hybrid pairs. Males of species A were colored with the eye ring of species B. They were then accepted as mates by females of species B. However, the males would not copulate with them, and normally the pairs would then have broken up. Before this happened Smith gathered the paired females, recolored them like species A and released them. They then became acceptable to the males of species A and stable hybrid pairs were established. Unfortunately, he had to leave the area before the outcome was determined of the fifty-five mixed matings so created.

In such a case, more or less stable properties of external appearance of adults are involved. Often there is some transient change in appearance or behavior which is necessary to elicit sexual responses. Many fish undergo a change in color of the integument as they come into the reproductive season (e.g., Neil, 1964; Barlow, 1963). The effectiveness of a model in eliciting sexual behavior may vary with its external coloration, as a variety of investigators have shown (Picciolo, 1964). Seasonal changes in the coloration of the plumage or soft parts in many birds may have a similar significance.

Some signals serve to establish a pair-bond some time before copulation will take place. Other signals will bring the mates into full reproductive condition and synchronize their physiological cycles, as Lehrman (1959, 1961, 1964) has shown in doves. In addition, signals may be presented only for a brief time as an immediate accompaniment to readiness to mate. In many animals there is a distinctive posture or movement by which the female invites a male to copulate. Species distinctiveness is often only slight or lacking in such displays, presumably because problems of specific identification have been resolved earlier. A curious situation exists in ducks and geese, where postcopulatory displays are often more specifically distinct than those preceding copulation (Johnsgard, 1963). Whether postcopulatory displays have a role in reproductive isolation remains to be explored. Perhaps they help to determine whether a female will ovulate after mating has occured.

Although we have spoken of sexual signals as though they all elicit the same sort of response, separate elements can, in fact, often be discerned even for elicitation of a single act such as mating. In cockerels, Carbaugh et al. (1962) find that different elements in the posture of a crouching hen serve to elicit and to orient mounting behavior. In cichlid fishes, Wickler (1965) has found a subtle example of a sexual orientation signal. In one of the mouth-breeding species, in which the female takes the eggs into her mouth after spawning, the male has egg-like spots close to the genital opening on the anal fin. The female nibbles at these after collecting eggs, while the male is still ejecting milt, so maximizing the chances of fertilization of the eggs (Fig. 5). Other signals serve to bring the pair together and to initiate the spawning sequence. The anal spots apparently just serve to orient this one phase.

INTERSPECIFIC RELATIONS AND VISUAL SIGNALING

Visual signals generated by the external morphology of animals often play a vital role in social behavior. The pattern of distribution of animals in space is often influenced by signals from the external morphology, and parent-young relationships may also be profoundly affected by morphology. In monkeys, the distinctive appearance of infants, contrasting with the adult pelage in many species, seems to elicit parental behavior (Booth, 1962; Jay, 1962). However, the external morphology is also subject to many other selective influences. The dictates of locomotion and food capture have obvious repercussions. The relationships to other species will also have a strong influence, both as manifest in competitive or cooperative relationships and also as they arise in the interaction between predator and prey. Systems of intraspecific visual communication can only be understood if these other factors impinging upon the external morphology and behavior are taken into account.

Predators may exert a selective influence upon the appearance of potential prey animals in a number of ways. Concealment may be at a premium, calling for cryptic appearance and behavior. Another strategy for reducing predation is to alarm the predator, either by presentation of inherently startling patterns of movements, or, in species which have protective devices such as noxious or unpalatable secretions or weapons, by conspicuousness that aids predators in learning to avoid the species. Still other animals employ some form of deception, by mimicry of another. There are many variations on these themes (Cott, 1957; Tinbergen, 1965), often having some impact on the visual communication system that the species employs for intraspecific purposes.

The visual signaling of the two European freshwater sticklebacks, Gasterosteus aculeatus and Pygosteus pungitius, differs in a number of respects. They also differ in their vulnerability to predation. Both have protective spines, but those of Gasterosteus are longer and stronger, and experiments demonstrate that they are more effective in discouraging attack by predatory pike and perch (Hoogland et al., 1957). The shyness of Pygosteus, its occupation of less open habitats, and its less conspicuous nuptial male coloration all seem attributable to the difference in vulnerability to predation (Morris, 1958).

The patterns of social organization in some gull species seem to be related to the vulnerability of ground-nesting birds to such predators as foxes (Kruuk, 1964; Tinbergen, 1965). The relationship is especially apparent in the kittiwake, a gull which has avoided predation by nesting on narrow ledges on sheer cliffs. This unusual habit seems to be responsible for many of the behavioral differences distinguishing kittiwakes from typical gulls, such as the use of submissive behavior by the young. Most gull chicks simply flee from adult attack, but a young kittiwake on the nesting ledge cannot flee from an aggressive adult (Cullen, 1958).

Among ducks, the signaling behavior of the Steller and the common eiders differ in a number of respects, the displays of the Steller eiders tending to be curtailed and inconspicuous by comparison. McKinney (1965) suggests that many of the differences can be attributed to their feeding habits. Steller eiders tend to feed more often in shallow water and seem to be more vulnerable to attack by birds of prey, especially the bald eagle. The apparent curtailment of display activities is one facet of a more general tendency for the Steller eider to be more wary in its behavior. Thus differences in the visual signals of these two species can only be understood by taking their differences in feeding behavior into account.

Predation also seems to be a determinant of social behavior in weaverbirds. Crook (1964, 1965) has shown that the elaborate woven nests, built in inaccessible places, often with a neck to discourage entry by climbing predators, have repercussions upon the types of action used in display. The general habitat type and the predominant food also seem to influence the type of social organization. The insectivorous weaverbirds living in forests normally nest in a dispersed fashion. By contrast, grain-eating weaverbirds living in the savannah nest in colonies. The mating system also differs in these two groups, the forest birds being monogamous, the savannah species predominantly polygynous, again with repercussions on their signaling systems.

Food is important; it is a prerequisite of polygynous systems that the female should be able to raise the young with minimal aid from the male. Thus monogamy will be favored where food is relatively sparse and both parents are needed to feed their young. The correlation between polygyny, and food available in abundance close to the nest, recurs in other bird groups such as icterids, wrens, and manakins (Orians, 1961; Orians and Collier, 1963; Verner and Willson, 1966; Verner, 1964, Snow, 1962, 1963; Crook, 1965). Even within a species, it is sometimes possible to correlate variation in the extent of polygyny in different habitats with the food supply that is available locally (Verner, 1964; Willson, 1966).

The evolution of a polygynous type of mating system seems to have a profound impact upon the signal repertoire of the species, particularly on the visual signals generated by the external morphology. As Darwin (1871) pointed out, sexual selection will be acute in such circumstances (Selander, 1965). There is intensification of intrasexual competition for mates among males and an increased requirement that male signaling behavior be stimulating to females. Both trends call for elaboration in the signal repertoire of males, which is especially evident in their external morphology. Sexual dimorphism is striking, both in body size and in the extent of external ornamentation, in those members of many animal groups that show some tendency toward polygyny, whether they be birds, carnivores, ungulates, or primates; perhaps sexual dimorphism even makes some contribution to the evolutionary heritage of man (Bartholomew and Birdsell, 1953).

The correlation between polygyny and large size and excessive ornamentation of the male as compared with the female is especially clear in birds (Amadon, 1959). The excesses are often sufficiënt to suggest reduced prospects of survival in the individual male. Among the icterids, for example, there is a trend toward increased dimorphism in species having polygynous or promiscuous mating relationships (Selander, 1958). Study of natural populations of the great-tailed and boat-tailed grackles shows that the large and ornamented male, with its long ungainly tail, actually has less chance for individual survival than the female. As a result, the sex ratio among adults is strongly unbalanced in favor of females (Selander, 1965). Presumably this disadvantage is balanced by a greater reproductive success for those that do survive. Once again there must be something distinctive in the ecology of grackles that permits the evolution of this polygynous condition and the extensive involvement of the external morphology in signaling behavior.

In addition to the relationship with predators, other kinds of interspecific interaction can have repercussions upon visual signaling systems. Ecological competition between two sympatric species, for example, may favor the use of signals involved in spacing which are effective in interspecific encounters as well as intraspecific ones. We know that interspecific territoriality is common and ecologically significant in certain groups of birds (Orians and Willson, 1964). This may tend to discourage evolutionary divergence in the signal characters employed in such encounters, including the external morphology.

Sympatric species sometimes engage in noncompetitive social activities, as when birds form mixed flocks to forage for food or exchange signals evoked by a predator that preys on several of them. In such situations uncontrolled emission of signals eliciting aggressive behavior could be a disadvantage because of the disruption of flock integration. Hamilton (1961; Hamilton and Barth, 1962) has argued that the assumption of a neutral or dull plumage in the nonreproductive season by males of some migratory birds in North America favors mixed flocking by the minimization of visual signals that might generate hostility. Moynihan (1960, 1962) has a similar interpretation for the neutral plumage of some neotropical birds, where participation in mixed feeding flocks seems to be particularly important for some species. The need for distinctive external morphology may also be reduced if a species lives in an impoverished fauna, and in many island races of birds the males tend to revert to a female-like plumage (Lack, 1947; Grant, 1965).

Thus the system of visual communication that a species employs cannot be understood in isolation from the rest of its biology. Contingencies as apparently remote as the associated fauna, the staple diet or the habitat in which it is sought may have profound repercussions upon the type of signals used. It would be hard to find better justification for the emphasis upon study of a species in the field, within its natural environment voiced by Lorenz and Tinbergen. No more than a fragmentary understanding of a communication system is possible until we know the type of social organization within which it is used. The understanding of social organization requires a knowledge of ecology. For this reason a merger of behavioral and ecological approaches promises to throw new light on how systems of visual communication evolve.

REFERENCES

Altmann, S., 1962. A field study of the sociobiology of rhesus monkeys, Macaca mulatta. Ann. N.Y. Acad. Sci., 102:338-435.

Amadon, D., 1959. The significance of sexual differences in size among birds. Proc. Am. Phil. Soc., 103:531-536.

Andrew, R. A., 1964. The displays of the primates. In: J. Buettner-Janusch (Ed.), Evolutionary and Genetic Biology of Primates, Vol. 2. New York, Academic Press, pp. 227-309.

Andrew, R. J., 1961. The displays given by passerines in courtship and reproductive fighting: A review. Ibis, 103a:315-348.

Barlow, G. W., 1963. Ethology of the Asian teleost Badis badis. II. Motivation and signal value of the colour patterns. Animal Behaviour, 11:97-105.

Barnett, S. A., 1958. An analysis of social behavior in wild rats. Proc. Zool. Soc. London, 130:107-152.

Bartholomew, G. A., and J. B. Birdsell, 1953. Ecology and the protohominids. Am. Anthropologist, 55:481-498.

Blest, D., 1957a. The function of eyespot patterns in the Lepidoptera. Behaviour, 11:209-256.

Blest, D., 1957b. The evolution of protective displays in the Saturnoidea and Sphingidae (Lepidoptera). Behaviour, 11:257-309.

Booth, C., 1962. Some observations on the behavior of Cercopithecus monkeys. Ann. N.Y. Acad. Sci., 102(II):447-487.

Brown, J. L., 1964. The evolution of diversity in avian territorial systems. Wilson Bull., 76:160-169.

Carbaugh, B. T., M. W. Schein, and E. B. Hale, 1962. Effects of morphological variations of chicken models on sexual responses of cocks. Animal Behaviour, 10:235-238.

Collias, N. E., 1944. Aggressive behavior among vertebrate animals. Physiol. Zool., 17:83-123.

Cott, H. B., 1957. Adaptive Coloration in Animals. London, Methuen.

Crook, J. H., 1964. The evolution of social organization and visual communication in the weaverbirds (Ploceinae). Behaviour Suppl., 10:1-178.

Crook, J. H., 1965. The adaptive significance of avian social organizations. Symp. Zool. Soc. London, 14:181-218.

Cullen, E., 1957. Adaptations in the kittiwake to cliff-nesting. Ibis, 99:275-302.

Dane, B., C. Walcott, and W. H. Drury, 1959. The form and duration of the display actions of the goldeneye (Bucephala clangula). Behaviour, 14:265-281.

Dane, B. C., and W. G. Van der Kloot, 1964. An analysis of the display of the goldeneye duck (Bucephala clangula (L.)). Behaviour, 22:282-328.

Darwin, C., 1871. The Descent of Man and Selection in Relation to Sex. London.

Darwin, C., 1872. The Expression of the Emotions in Man and the Animals. London, John Murray.

Eibl-Eibesfeldt, I., 1951. Beobachtungen zur Fortpflanzungsbiologie und Jugendentwicklung des Eichhörnchens (Sciurus vulgaris L.) Z. TierpsychoL, 8:370-400.

Eisenberg, J., 1963. The behavior of heteromyid rodents. Univ. Calif. Puhl. Zool., 69:1-114.

Eisenberg, J., 1964. The behavior of the wandering shrew, Sorex vagrans. Am. Midland Naturalist, 72:417-425.

Evans, L. T., 1961. Structure as related to behavior in the organization of populations in reptiles. In: W. F. Blair (Ed.), Vertebrate Speciation. Austin, Univ. Texas Press.

Ewer, R. F., 1966. Juvenile behavior of the African ground squirrel, Xerus erythropus (E. Geoff.). Z. Tierpsychol., 23:190-216.

Falk, J. L., 1958. The grooming behavior of the chimpanzee as a reinforcer. J. Expt. Anal. Beh., 1:83-85.

Fantz, R. L., 1965. Ontogeny of perception. In: A. M. Schrier, H. F. Harlow, and F. Stollnitz (Eds.), Behavior of Nonhuman Primates, Vol. 2. New York, Academic Press, pp. 365-403.

Geist, V., 1963. On the behavior of the North American moose (Alces alces andersoni Peterson 1950) in British Columbia. Behaviour, 20:377-416.

Goodall, J., 1965. Chimpanzees of the Gombe Stream Reserve. In: I. DeVore (Ed.), Primate Behavior. New York, Holt.

Goodall, J., Expressive movements and communication in the Gombe Stream chimpanzees (in press).

Grant, P. R., 1965. Plumage and the evolution of birds on islands. Syst. Zool., 14:47-52.

Hall, K. R. L., and I. DeVore, 1965. Baboon social behavior. In: I. DeVore (Ed.), Primate Behavior. New York, Holt, pp. 53-110.

Hamilton, T. H., 1961. On the functions and causes of sexual dimorphism in the breeding plumage characteristics of North American species of warblers and orioles. Am. Naturalist, 95:121-123.

Hamilton, T. H., and R. H. Barth, 1962. The biological significance of reason change in male plumage appearance in some new world migratory bird species. Am. Naturalist, 96:129-144.

Hamilton, W. J., Ill, 1965. Sun-oriented display of the Anna’s hummingbird. Wilson Bull., 77:38-44.

Harrison, C. J. O., 1965. Allopreening as agonistic behavior. Behaviour, 24:161-209.

Hazlett, B. A., and W. H. Bossert, 1965. A statistical analysis of the aggressive communications system of some hermit crabs. Animal Behaviour, 13:357-373.

Hoogland, R., D. Morris, and N. Tinbergen, 1957. The spines of sticklebacks (Gasterosteus and Pygosteus) as means of defense against predators (Perca and Esox). Behaviour, 10:205-236.

Howell, T. R., and G. A. Bartholomew, 1952. Experiments on the mating behavior of the Brewer blackbird. Condor, 54:140-151.

Jay, P., 1962. Aspects of maternal behavior among langurs. Ann. N.Y. Acad. Sei:, 102(II):468-476.

Johnsgard, P. A., 1963. Behavioral isolating mechanisms in the family Anatidae. Proc. 13th Intern. Ornithol. Congr., 531-543.

Jolly, A., 1966. Lemur social behavior and primate intelligence. Science, 153: 501-506.

King, J. A., 1955. Social behavior, social organization and population dynamics in a blacktailed prairiedog town in the Black Hills of South Dakota. Contrib. Lab. Vert. Biol Univ. Mich., 67:1-123.

Kruuk, H., 1964. Predators and anti-predator behavior of the black-headed gull (Larus ridibundus). Behaviour Suppl, 11:1-129.

Kummer, H., Two basic variations in the social organization of the genus Papio (in press).

Lack, D., 1940. Courtship feeding in birds. Auk, 57:169-178.

Lack, D., 1947. Darwin’s Finches. Cambridge, Cambridge Univ. Press.

Lehrman, D. S., 1959. Hormonal responses to external stimuli in birds. Ibis, 101:478-496.

Lehrman, D. S., 1961. Hormonal regulation of parental behavior in birds and infrahuman mammals. In: W. C. Young (Ed.), Sex and Internal Secretions, Vol. 2, pp. 1268-1382. Williams and Wilkins Company, Baltimore.

Lehrman, D. S., 1964. Control of behavior cycles in reproduction. In: W. Etkin (Ed.), Social Behavior and Organization among Vertebrates, Chicago, Univ. of Chicago Press, pp. 143-166.

Lorenz, K. Z., 1935. Der Kumpan in der Umwelt des Vogels. J. Ornithol., 83: 137-213, 289-413.

Lorenz, K. Z., 1950. The comparative method in studying innate behavior patterns. Symp. Soc. Exptl. Biol., 4:221-268.

Marler, P., 1955a. Studies of fighting in chaffinches. (1) Behaviour in relation to the social hierarchy. Brit. J. Animal Behaviour, 3:111-117.

Marler, P., 1955b. Studies of fighting in chaffinches. (2) The effect on dominance relations of disguising females as males. Brit. J. Animal Behaviour, 3:137-146.

Marler, P., 1956. Studies of fighting in chaffinches. (3) Proximity as a cause of aggression. Brit. J. Animal Behaviour, 4:23-30.

Marler, P., 1957. Studies of fighting in chaffinches. (4) Appetitive and consummatory behaviour. Brit. J. Animal Behaviour, 5:29-37.

Marler, P., Aggregation and dispersal: Two functions in primate communication (in press).

Mayr, E., 1942. Systematics and the Origin of Species. New York, Columbia Univ. Press.

McKinney, F., 1965. The spring behavior of wild Steller eiders. Condor, 67:273-290.

Marks, H. L., P. B. Siegel, and C. Y. Kramer, 1960. Effect of comb and wattle removal on the social organization of mixed flocks of chickens. Animal Behaviour, 8:192-196.

Morris, D., 1958. The reproductive behavior of the ten-spined stickleback (Pygosteus pungitius L.). Behaviour Suppl., 6:1-154.

Moynihan, M., 1960. Some adaptations which help to promote gregariousness. Proc. 12th Intern. Ornithol. Congr. 1958, Vol. 2, pp. 523-541.

Moynihan, M., 1962. The organization and probable evolution of some mixed species flocks of neotropical birds. Smithsonian Inst. Misc. Coll., 143(7): 1-140.

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

Nelson, K., 1964. The temporal patterning of courtship behavior in the glandulocaudine fishes (Ostariophysi, Characidae). Behaviour, 24:90-146.

Orians, G., 1961. The ecology of blackbird (Agelaius) social systems. Ecol. Monographs, 31:285-312.

Orians, G., and G. Collier. 1963. Competition and blackbird social systems. Evolution, 17:449-459.

Orians, G., and M. F. Willson, 1964. Interspecific territories of birds. Ecology, 45:736-745.

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

Sade, D. S., 1965. Some aspects of parent offspring and sibling relations in a group of rhesus monkeys, with a discussion of grooming. Am. J. Phys. AnthropoL, 23:1-18.

Schneirla, T. C., 1959. An evolutionary and developmental theory of biphasic processes underlying approach and withdrawal. Nebraska Symp. Motiv., 1-42.

Schneirla, T. C., 1965. Aspects of stimulation and organization in approach-withdrawal processes underlying vertebrate behavioral development. Advan. Study Behavior, 1:1-74.

Seiander, R. K., 1958. Age determination and molt in the boat-tailed grackle. Condor, 60:355-376.

Seiander, R. K., 1965. On mating systems and natural selection. Am. Naturalist, 99:129-141.

Sibley, C. G., 1957. The evolutionary and taxonomie significance of sexual dimorphism and hybridization in birds. Condor, 59:166-191.

Smith, N. G., 1966. Evolution of some Arctic gulls (Larus): An experimental study of isolating mechanisms. Ornithol. Monographs, 4:1-99.

Snow, D. W., 1962. A field study of the black and white manakin, Manacus manacus, in Trinidad. Zoologica, 47:65-104.

Snow, D. W., 1963. The evolution of manakin displays. Proc. 13th Ornithol. Congr., pp. 553-561.

Sparks, J. H., 1964. Flock structure of the red avadavat with particular reference to clumping and allopreening. Animal Behaviour, 12:125-136.

Sparks, J. H., 1965. On the role of allopreening invitation behavior in reducing aggression among red avadavats, with comments on its evolution in the Spermestidae. Proc. Zool. Soc. London, 145:387-403.

Stokes, A. W., 1962a. Agonistic behaviour among blue tits at a winter feeding station. Behaviour, 19:118-138.

Stokes, A. W., 1962b. Comparative ethology of great, blue, marsh, and coal tits at a winter feeding station. Behaviour, 19:208-218.

Struhsaker, T. T., 1967. Auditory communication among vervet monkeys. In: 5. A. Altmann (Ed.), Social Communication among Primates. Chicago, Univ. Chicago Press, pp. 281-324.

Tinbergen, N., 1959. Comparative studies of the behaviour of gulls (Laridae): a progress report. Behaviour, 15:1-70.

Tinbergen, N., 1965. Behavior and natural selection. In: J. A. Moore (Ed.), Ideas in Modern Biology, Proc. 16th Intern. Zool. Congr., Washington, 1963, Vol 6. New York.

Verner, J., 1964. Evolution of polygamy in the long-billed marsh wren. Evolution, 18:252-261.

Verner, J., and M. F. Willson. 1966. The influence of habitats on mating systems of North American passerine birds. Ecology, 47:143-147.

Walther, F., 1958. Zum Kampf und Paarungsverhalten einiger Antilopen. Z. Tierpsychol, 15:340-380.

Walther, F., 1964a. Verhaltensstudien an der Gattung Tragelaphus De Blainville, 1816 in Gefangenschaft unter besonderer Berücksichtigung des Sozialverhaltens. Z. Tierpsychol., 21:393-467.

Walther, F., 1964b. Einige Verhaltensbeobachtungen an Thomsongazellen. (Ga-Zella thomsoni Gunther, 1884) im Ngorongoro-Krater. Z. Tierpsychol., 21: 871-890.

Wickler, W., 1965. Die Evolution von Mustern der Zeichnung und des Verhaltens. Naturwiss., 52:335-341.

Willson, M. F., 1966. Breeding ecology of the yellow-headed blackbird. Ecol. Monographs, 36:41-77.