The Tools of Communicating

As birds interact they provide each other with information in many ways. Every feature of each individual's appearance, its every action, could in principle be informative for another individual. This is not to say that every such source of information is equally relevant at any given time, or even that all the incidental aspects of appearance and demeanor provide enough information to keep interactions from becoming chaotic. It merely means that birds bring abundant sources of information to their interactions. Among these sources are some that have evolved to be informative. That is, there are tools specialized for communicating.

Of these tools, the behavioral unit that has been central to ethological studies of communication is the "display"; displays and simultaneous or sequential compounds of displays are specialized acts that are performed by individuals. There are also behavioral units with repeatable, formal patterns that can be performed only through the cooperation of two or more participants; these "formalized interactions" have been studied less thoroughly than have displays. Other communicative specializations are objects produced by behavior, and yet others are not behavioral at all: plumage patterns and colors, for instance.

DISPLAYS

The concept of displays, or at least of behavior adapted to provide recipient individuals with information that is not otherwise readily apparent was used by Darwin (1872), who saw such acts as means of making available information about an individual's internal emotional state. The term "display" began to be used more or less consistently in reference to these units when Huxley's (1914) studies of great crested grebes (Podiceps cristatus) appeared. It was not until relatively recently that it was defined as an act specially adapted in "physical form or frequency to subserve social signal functions" (Moynihan, 1956, 1960, in the course of work with gulls and flocking passerines). Although the term connotes visible behavior, Moynihan recognized that the concept to which it applies is much more general, and that the behavior can be suited to reception by any sensory modality.

As a class of animals, birds are known to have a remarkable diversity of display behavior. This enormous diversity can be categorized, when this is useful, in terms of different sensory modalities. Visible displays, for instance, abound, and the conspicuous and bizarre posturings and movements of ducks, geese, grebes, and gulls figured importantly in the work of such pioneer ethologists as Heinroth, Lorenz, Huxley, and Tinbergen. Displaying birds may wave their wings, waggle their tails, stretch their necks, point with or deflect their bills, crouch, cower, leap, flutter, raise or lower crests or other tracts of feathers, manipulate token nest material or food, or do these and other things in complex combinations. The number of distinctive possibilities is limited, but large. The number of possible audible displays is extremely large, encompassing clear musical whistles, trills, harsh rasps, hoots, bell-like vocalizations, and many other products of the respiratory tract; as well as mechanical snaps, winnowings, and drumming sounds produced with wing or tail feathers; and sounds made by rapping the bill on a resonant object. Opportunities for tactile displays are less readily available to birds than to, say, mammals or social insects, and yet a diversity of special touching patterns has evolved, ranging from allopreening and various forms of bill-to-bill touching and even nipping, to bodily pushing and treading on the partner's back during mounting for copulation. Not surprisingly, the main sensory modality for which evolution has been much less productive of displays for birds than for other animals has been olfaction. Birds are not known to have displays involving the release of chemicals as scents.

The countless displays, like any other products of evolutionary processes, did not arise de novo. For visible displays, at least, the close obversation of birds has provided very important clues about the kinds of acts that probably serve as precursors. Thus Daanje (1950) recognized in many of the stereotyped postural displays of birds positions that were parts of other acts; in particular, of acts with which a bird prepared to take flight, strike, turn away, etc. For instance, before taking flight a bird will first draw back its head and neck, lower its breast toward the ground, raise its tail, flex its legs, and begin to extend its carpals (wrists) outward. It may pause there, ready to spring, or begin to spring by extending its legs, head, and neck, depressing its tail, and spreading its wings. Flight does not always follow such preparations, even the second set, and so Daanje called them "intention movements" and showed how positions very similar to them occurred very commonly in postural displays. That is, the movements performed in preparing to fly are informative and are highly suited for evolutionary elaboration that enhances their effectiveness as signals. Intention movements appear to be a major evolutionary source of visible display patterns, not just in birds but in most other animals as well.

Other evolutionary sources of display behavior that have been postulated for birds also appear to be significant for other kinds of animals. Morris (1956), for instance, saw in the thermoregulatory lifting or depressing of body feathers, which are autonomic responses in tense situations, the sources of displays in which birds lift their crests, ruffle their flanks, or take on a sleeked appearance. Grooming and some other acts that occur during aggressive encounters with surprising regularity are seemingly irrelevant; the internal causes of these "displacement" activities, as they have been termed, remain controversial but the class is recognized as yet another evolutionary source of displays (Tinbergen, 1952, 1959). Bastock, Morris, and Moynihan (1953) saw in the Grass Pulling of disputing gulls a display that probably arose from yet another kind of source, the regular redirection onto substitute targets of striking and tearing acts in circumstances in which a bird was highly motivated to attack and yet afraid or for some other reason inhibited from directing the attack to its rival.

Ethologists have been less successful in proposing the evolutionary origins of vocal and other kinds of displays (except for most tactile acts, which appear to have obvious origins), but have made much use of research on birds in describing the ways in which displays come to diverge from their predecessor acts as evolution proceeds: increased conspicuousness, uncoupling and differential modification of component movements and their motivational thresholds, increasing the amount of use of an act beyond that which is directly functional, and so forth (see Hinde and Tinbergen, 1958; Tinbergen, 1959; Morris, 1966; and W.J. Smith, in press). And, finally, as comparisons of the displays of diverse species of birds have also contributed to our understanding of the kinds of selection pressures that mold the physical forms of displays, it can fairly be said that the study of bird communication has had a very pervasive influence on ethological concepts of display evolution.

While the concept of displays as formal units has been extremely productive in the study of communication, it is not without its problems (these are discussed relatively fully by W. J. Smith, in press). It can be very difficult, for instance, to demonstrate that an act has been specialized to facilitate the sharing of information. In practice, we tend to accept relatively stereotyped acts that appear to have no direct function as being displays if it seems likely that they will be of use in interactions. For vocal displays this criterion is probably sufficient, but for visible displays it is sometimes hard to be sure. For example, many birds flash their wings open while foraging, an act that looks very much like display behavior but that may be specialized to startle prey. Other acts—slightly rigid postures, small amplitude movements of the wings, or slight sleeking or fluffing of feathers, for instance—are so little specialized that it is very difficult to know whether they should be classified as displays. Such problems are not unexpected and for many analytic purposes may not be very important. Greater difficulties are posed by the fact that no display is fully constant; many are quite variable in form and may even intergrade with other displays. This can make it very difficult, at least by criteria of form, to define display units. Comparable definitional problems arise when specialized acts are used only in combinations: are the display units the combinations or the recombinable components? I shall suggest below an amplification of the traditional display concept that permits a practical resolution of such difficulties (see section on Interpretation of Bird Communication).

FORMALIZED INTERACTIONS

There remains the problem that acts specialized to be informative do not all fall within the capacities of single individuals. There are behavioral units that are essentially cooperative, that can only be performed by the combined actions of two or more participants. If we reserve the term "display" for formalized acts of individuals, we can refer to these cooperative performances as "formalized interactions." For instance, when two gulls (Laridae) greet each other on a male's pairing territory, they tend to go through a predictable behavioral sequence (reviewed by Tinbergen, 1959; Moynihan, 1962a). As the female arrives, the male emits Long Calls in the Oblique posture, facing toward her. They then align in parallel with each other, and both utter a vocal display known as the Mew Call. Remaining in parallel, they go through a somewhat variable succession of progressively less agonistic postural displays (displays less likely to be followed by attack or escape). The interaction is often terminated with mutual Head Flagging, in which they stand side by side, stiffly turning their heads away from each other. The exact sequence of displays is subject to omissions and some changes in order, but the greeting participants cooperate and each accommodates by tending to perform the same display as its partner most of the time, shifting displays when the other does. Similar greeting ceremonies, involving fixed mutual orientations and simultaneous performance of displays, are known in many other kinds of bird, although few are as complex as the gulls'.

Formalized interactions are not limited to greetings but may occur in appeasing and reassuring, courting, and other encounters, usually those that are rich in uncertainty. Their main characteristic is the recurrence of a pattern of cooperative behavior. Each participant has a definable, preestablished part to play in an interactional pattern of regular, classifiable moves and responses. Participants may play their parts somewhat differently from event to event, but even as each affects the others it is also constrained by the formalized pattern to accommodate to their acts.

OTHER FORMAL SOURCES

Behavior is a transient source of information, but birds can act to make information more persistently available. They may do so by making objects that will subsequently have direct functions, for example, the nests that males of savannah-dwelling weaverbirds (e.g., Quelea spp.: Crook, 1964) and diverse other species build and show to prospective mates as they attempt to form pair bonds; or their constructions may serve a similar communicative function but lack direct utility, as do the bowers and stages of bowerbirds (Ptilonorhynchidae: Gilliard, 1956, 1963) and the cleared lek areas of manakins (Pipridae: Chapman, 1938; Sick, 1959, 1967). Birds are not known to scent-mark sites, a common practice used by mammals and invertebrates for leaving a source of information to act in a communicator's absence. But at least one species, the Japanese quail (Coturnix coturnix japonica)_y does have a special chemical product with which it marks its droppings. It produces a foam that may be a visible rather than an olfactory marker (Schleidt and Shaker, 1972).

The principal specializations of birds that serve as persisting sources of information are neither behavior nor behavioral products, however. They are the colors and patterns of plumage, bare skin, epidermal outgrowths, or other "soft parts" that birds wear as badges or uniforms. The plumage of a male bird may be more distinctive and hence perhaps more important than his displays in identifying him to unmated females (Hinde, 1956, 1959). For instance, N. G. Smith (1966) found that four species of arctic gulls (Larus species) that are very similar in plumage and displays are differentiated primarily by the colors of their irides, the fleshy rings around their eyes, and the extent of the contrast these make with their white heads. By altering eye-ring colors Smith demonstrated that both males and females are prepared to respond to these characteristics of their partners in the period before egg laying; the badges are "prezygotic isolating mechanisms" preserving the integrity of each species' gene pool (see Mayr, 1963). That the identifying information provided by some badges may also be important for interactions other than pair bonding has been shown by experiments such as those done by Peek (1972) with red-winged blackbirds (Agelaius phoeniceus). He removed the bright red epaulets from seventeen otherwise normal males and found that although the birds displayed and appeared otherwise normal, they could not maintain their territories against other males.

Badges are not important solely because they persist. Indeed, some badges are kept hidden or inconspicuous much of the time, then revealed in conspicuous acts. Various species of tyrannid flycatchers, such as the kingbirds (genus Tyrannus: W.J. Smith, 1966) have a brilliant reddish or yellowish concealed crown patch that they rapidly uncover and may even erect when displaying while preparing to attack. Many birds have crests or other elongate feathers that are conspicuous primarily when erected. Gregarious species often have patches of white or bright colors on their wings, back, or tails that become visible as "flash patterns" when the birds take flight (Moynihan, 1960, 1962b). Canada geese (Branta canadensis), which have a flash patch at the base of the tail, have a white chin patch that is not concealed but that becomes very noticeable when the geese perform a Head-Tossing display before flying (Raveling, 1969).

Displays, their compounds, formalized interactions, constructions, and badges are all tools evolved by birds to facilitate information sharing. They provide formal means of contributing information to events. Communication, however, depends on further sources of information, sources that are not specialized to this end but are simply inherent in the structure of all events and entities. Some of these sources are public, accessible to all participants. Some are private, part of the genetic and experiential stores of information carried by each participant. Such sources can be thought of as acting contextually to the formal sources, and their role is considered below in discussing the evolution of messages.

The Interpretation of Bird Communication

Ethologists have had a number of goals in interpreting bird communication. A central one has always been to understand the motivational causes that underlie the performance of displays: the physiological states and their stimulus relationships that lead a bird to respond to a particular circumstance by displaying. The objective of such work is to understand the internal mechanisms that control the behavior of an individual bird, a goal that is not central to the study of communication per se. Communication, in the sense that the term is used in this book, is an interactional process, a procedure by which information is shared between or among individuals. The goal of defining what kinds of information are made available by what sources (i.e., by displays, formalized interactions, badges, etc., and by incidental nonformalized sources) is more pertinent, as is the study of responses to this information and the kinds of function that are generated.

MOTIVATIONAL CAUSATION

The earliest trends in the interpretation of bird communication were "causal" and functional, and they remain the predominate trends. The focus, not inappropriate for a branch of evolutionary biology arising from comparative physiology and comparative anatomy, has been on the mechanisms and adaptiveness of individual behavior. Tinbergen described this perspective succinctly when he stated that the "central" problem in analyzing the behavior of a communicator in events involving social cooperation is "what urges the actor to signal?" (1953:73).

The mechanisms of internal causation are not directly accessible through the study of behavior and have usually been represented by hypotheses about emotional states or motivation. The various behavioral criteria used to determine motivational causation include primarily the following (abstracted from accounts by Moynihan, 1955, and Tinbergen, 1959): (1) If a display posture or movement resembles other motor patterns it may share causal states with them. Thus if it resembles acts used in striking, it is assumed that it may be aggressively motivated. (2) If a display is performed simultaneously or in quick succession with the behavior of, say, approaching or turning away, it may share motivational causes with these acts. (3) If a display is performed when the displaying individual is somehow balanced by circumstances between two courses of action that are more direct, the display may be caused by a conflict of two motivational states. For example, if a bird that will attack intruders in its territory but will flee if attacked when it is an intruder in theirs does neither, but instead displays when facing a rival at their common territorial border, then those displays are probably caused by the conflicting motivational states of aggression and fear.

The concept of motivational "conflict," seen in the third situation above, has been one of the most useful contributions of this research. Most of the display behavior of birds appears to be performed when the communicators are in conflict states. Observational techniques for revealing the different absolute and relative strengths of conflicting motivations were developed by Moynihan (e.g., 1955, 1962a) in the course of work with gull displays. The existence of such differences in conflict states underlying different displays was subsequently verified experimentally by Blurton Jones (1968), working with threat displays of the great tit (Parus major).

Behavioral studies have provided better evidence for the existence of motivational conflict and for intensity differences among the conflicting motivations that cause the performance of different displays than they have for describing the motivational systems themselves. In practice, although not always in theory, there has been a predominating trend to lump the hypothetical causes of bird displays into three "unitary" motivational systems: aggression, fear, and sexual motivation (e.g., Tinbergen, 1959). Yet these gross, functionally oriented categories are each far from unitary. As Hinde has argued, variables such as the persistence, directiveness, and temporal clustering of the behavior patterns each category is said to control do not always change in concert (summarized in Hinde, 1970). Further, there are diverse behavior patterns assigned to each causal category, and at least in the case of "sex" these are remarkably heterogeneous (Moynihan, 1962a): at one extreme is copulation, at the other a tentative associating without contact.

Three motivational systems are in fact inadequate to explain the internal causation of bird displays. How many systems should be recognized is by no means clear, however, as the observation of behavior alone does not permit discrete physiological mechanisms to be identified (Hinde,1970). Hinde has suggested that for some purposes it is appropriate to proceed by recognizing a class of "intervening variables," which he calls "tendencies." For instance, observations suggest that there is a tendency to drink a certain amount of water (measurable as a dependent variable) after an individual has been deprived for a certain period (measurable as an independent variable). The "tendency" to drink implies an internal mechanism that somehow relates the other two variables, but it does not postulate properties of this mechanism—e.g., it does not require it to be unitary. A tendency postulated as a cause of a display would be described in terms of the behavior correlated with the use of that display, be it attack, escape, or mounting, or some other activity not adequately described by the traditional motivations. For example, tendencies such as "staying," "flying," "approaching," and "being gregarious" have all been described for bird displays (see review by Hinde, 1970, and such studies as Moynihan, 1960; Andrew, 1961a; Stokes, 1962; Delius, 1963; Crook, 1963, 1964; Tinbergen, 1964; Fischer, 1965).

KIND OF INFORMATION MADE AVAILABLE

Correlations such as those used in research on causal tendencies are also useful as indications of the kinds of information that a display can make available about the communicator's behavior. The analytic perspective shifts very markedly, however. The superordinate behavioral unit is no longer an act of an individual bird, but an interaction based on the behavioral contributions of participating individuals. The important mechanisms at this level of integration are found less within individuals than among individuals in their moves, countermoves, and accommodations.

In most of their encounters, at least in natural circumstances, birds do not interact chaotically. The relevance of a display, a badge, or any comparable specialization is in the information it can contribute to these interactions, information that makes it easier for each participant to anticipate the actions of the others, to be more prepared to respond in ways that can further develop or stabilize their interaction.

Analysis of the behavioral information content of a display (its "messages" about behavior) requires that the range of circumstances in which the display naturally occurs be found and thorough observations be made of the behavior with which its use correlates throughout this range. Many displays correlate with different acts in different events, but unless a display varies consistently as these activities change it cannot provide detailed information about them. For example, the vocal display called song in many species of birds is uttered by an individual that is prepared to attack, greet, attempt to form a pair bond or copulate, or join another individual in close association without making contact. The most that can be learned from such a display is that the communicator will probably seek interaction if an opportunity arises, but just what kind of interaction a recipient should expect in any particular event cannot be known from the utterance of song alone—although it may be readily predicted from other sources of information that are available in the event.

Analyses of this sort are clearly at variance with traditional causal analyses, especially if the latter are phrased in terms of a small set of a priori motivational variables. Inasmuch as causal interpretations continue to adhere largely to this tradition, it may be useful to compare the results of the two approaches in some detail by attempting a provisional reinterpretation of some gull displays. Thanks primarily to the very careful work of Tinbergen and his students, gull displays are among the most intensively studied of any group of birds, and they have provided the main model for most current research. Further, Tinbergen's (1959) review of this work provides a detailed example of the usefulness and shortcomings of interpretation based on the motivational trio of aggression, fear, and sex.

Of the nine display patterns analyzed in that review, the most convincing interpretations involve the simpler conflicts of motivational states. Consider, for instance, the Aggressive Upright, a posture in which the neck is extended upward, the bill is angled downward, and the carpals are lifted away from the body. A communicator in this posture may approach and then attack an opponent; or it may hesitate, cease approaching, and then withdraw. Tinbergen indicates its motivational state as a conflict of aggression and fear. An interpretation of the behavioral messages encoded would be closely comparable: one can expect attack and escape activities or indecisive acts while the gull chooses. However, in two closely related postures the carpals are not lifted and the bill angle differs, being horizontal in the Intimidated Upright and upward in the Anxiety Upright. Tinbergen's interpretation that these postures are a result of less aggression and more fear yields only a partial motivational description, which, moreover, does not tell us what limits to expect of the posturing gull's behavior. The existing descriptions imply that gulls adopt these postures when in a state of conflict between fear and motivations for virtually any behavior from foraging (e.g., when intruding on a shoreline territory) to joining and associating (e.g., when forming a pair bond) to attack (see Tinbergen, 1965, and the evidence he presents from his student Manley, 1960).

This large range of alternatives to escape is not adequately characterized by the traditional motivational terms, and what is needed is a concept that implies simply that the displays are performed when some unspecifiable motivation conflicts with fear. The motivation is unspecifiable because it can be very different at different times, and what is consistent is simply the conflict it engenders. A behavioral message interpretation of these two displays would recognize that the communicator will either escape or select some activity from a broad set of alternatives incompatible with escaping. That its motivational state varies widely simply means that the behavioral alternatives to escape are not specified precisely by the display. A lack of precision is by no means uncommon in behavioral messages, and in gulls this very broad message (termed the "general set" of incompatible alternatives by W.J. Smith, 1969a) (see Table 1) appears to recur in the songlike Oblique-cum-Long Call display, there indicating acts that are incompatible not with escape but with attack.

There are several display patterns of gulls for which the traditional trio of motivational causes encounters more obvious difficulties. Prominent among these are Choking and Mew Calling. Tinbergen interpreted the former as caused by a very strong conflict between defensive aggression and some motivation to remain with the nest site or territorial boundary, but he recognized that Choking can be performed in both "friendly" and "hostile" circumstances. Neither the behavioral messages nor the motivational causes of Choking are fully evident from existing descriptions, although the display apparently does correlate with the tendency of the communicator to remain at a special site. In the case of Mew Calling, however, it is evident that aggression, fear, and sexual motivation are not consistent causes of the display. The first two can be part of the motivational state at times; for instance, when a male herring gull turns from a hostile encounter at a territorial boundary and proceeds to walk in parallel with his mate, both of them Mew Calling. The parallel walking suggests a balanced conflict between approach and withdrawal, and, at least for the male, part of the motivation to approach could be aggressive (this is less obvious for the female). Yet neither attack nor escape is seen in many situations in which one individual approaches its mate or its offspring on the nest, e.g., a male bringing nest material to his mate, or an individual of either sex coming to relieve its partner from incubation or to feed the small chicks.

There are interesting behavioral correlations for Mew Calling, however, that involve interacting and also flying, walking, or, rarely, swimming. For instance, in the circumstances just described, Mew Calls usually just precede alighting in the territory and are then used while walking toward the mate or chicks. A parent herring gull returning to a territory when its chicks have left the nest and hidden in the vegetation usually stands and Mew Calls; its chicks then run to it and are fed. It seems characteristic of all cases in which herring gulls perform this display that the communicator is prepared to interact, either agonistically or not, and that its locomotory behavior is likely to be indecisive—slowing as it shifts from one form to another in alighting, approaching partway to another individual, stopping short, or deviating from an approach to walk in parallel. With hidden chicks the parent stops and Mew Calls when it cannot approach them further because it does not know where they are within the territory. Note that the traditional motives of aggression, fear, and sex are inadequate causal explanations for this display not because they never occur, but because they do not occur consistently and because many other motives may be causal to it in different cases. The information that the display makes available about behavior, however, includes at least that the communicator is prepared to interact and to locomote, although it will most likely do so in some indecisive fashion. Such information is probably useful to recipients of the display, especially since they will usually have a reasonable expectation of the most likely kind(s) of interaction. Such an expectation, however, is based on sources of information other than the display.

Table 1

Behavioral selection messages widely encoded in the displays of diverse species of birds.

Bird displays very commonly encode the information that the communicator will behave indecisively, alternately starting different incompatible activities or pausing, and this undoubtedly relates to the finding that a conflict of motivational causes is often characteristic of the states of displaying individuals. Other displays, such as much of the singing referred to above, appear to carry information less about indecisive acts than about the behavior of seeking the opportunity to be decisive. In many species, singing customarily occurs when no other individuals are present (see review by Andrew, 1961b). Singing individuals of species as phylogenetically diverse as the chaffinch (Fringilla coelebs: Marler, 1956), European blackbird ( Turdis merula: Snow, 1958), green-backed sparrow (Arremonops conirostris: Moynihan, 1963), Carolina chickadee (Parus carolinensis: S. T. Smith, 1972), and eastern phoebe (Sayornis phoebe: W. J. Smith, 1969b, 1970) behave as if seeking other birds: they take high, conspicuous perches, or in some species sing in special display flights, or may actively patrol. Song ceases immediately and some form of interaction is usually attempted if a suitable recipient becomes available; for instance, when a field sparrow (Spizella pusilla) hears a recording of its species' song played from a territorial border or beyond, it increases its singing in reply; but if the recording is played within the territory the owner does not sing—it flies directly toward the source of the sound (Goldman, 1973). (Birds do countersing across territorial boundaries that neither will cross to attack the other, and are then inaccessible to each other because of the boundary convention. The countersinging is apparently a less-intense interaction than they are seeking, but the boundary thwarts closer approach.) Still other displays provide information less about what a communicator will attempt to do or seek opportunity for than about the kinds of behavior it is prepared to accept. A female bird who is receptive to copulation, for instance, will solicit a male by adopting a posture that facilitates his mounting.

The information made available by a Mew Call, indicating that a bird will evidence indecision about locomotory behavior, is widely encoded by the displays of diverse other species. Many New World flycatchers (family Tyrannidae) have vocalizations that are employed as they alight, make flight-intention movements, or slow down or veer in flight in circumstances in which indecisive flying and perching may alternate: e.g., approaching an agitated mate or potential mate, including mutual greeting performances; begging by an offspring following parents whose tempers have worn thin; veering off from pursuit or attack of a predator; choosing between taking a station or continuing to patrol a territory, or between flying to a singing perch and foraging, or between remaining in a territory and following a foraging mate or a fleeing intruder out of it; staying with the mate or flying to a border in response to the calls of a neighbor in that region; or leaving the nest in the absence of its mate (e.g., W. J. Smith, 1966, 1969a). The flightless Adelie penguin (Pygoscelis adeliae) has vocal displays with remarkably comparable employment. One is used by a penguin as it approaches its territory after a long absence, on leaving its nest after incubating, when responding to fights in nearby territories, during stalemated aggressive encounters, after unsuccessfully threatening an intruding crowd (e.g., a creche of chicks), and in response to slowly approaching humans or predatory skuas (Stercorarius skua: Ainley, 1974). The other display is uttered while a penguin is walking between its nest and the beach, standing or walking by the water's edge (particularly as individuals begin to dive from a flock into the water), swimming at sea, and apparently when being chased.

The Carolina chickadee utters a fixed sequence of three distinct vocal displays as it alights in various circumstances in which it appears somewhat indecisive: a High Tee as it approaches the perch, a Chick on alighting, and a Dee immediately afterward. It also has a Lisping Tee, which is used primarily when in flight or when flight is very imminent (S. T. Smith, 1972). Many other flocking species, from red-legged partridges (Goodwin, 1953) to species of Chlorospingus tanagers (Moynihan, 1962c), also have displays used as flight becomes likely. In Canada geese, the Head-Tossing display mentioned above in discussing badges is performed by members of a family group as they get ready to fly. If the gander performs the display the whole family flies shortly afterward, but if other members begin it they may have to continue it for tens of minutes before the gander is ready to lead them in flight—if he gets ready at all (Raveling, 1969).

Such displays provide more information than just the likelihood of flight, of course. The Head Tossing of geese, for instance, is performed by individuals who would fly if that did not require them to leave their families, and the Lisping Tee of chickadees is often emitted in flights that sever association with their mates or flocks. Both displays provide information about the readiness of the communicators to associate with their companions, although the relative probabilities of flying away differ for the two species. Displays providing information about the behavior of associating are customarily used only if something makes that association difficult, anything from an incompatible behavioral alternative or the absence of companions to an obscuring environment. Green-backed sparrows have a Plaintive Note that mates utter when they become separated; they usually reestablish association immediately afterward (Moynihan, 1963). Ornithologists have described what they often call "contact calls" in the repertoires of many species that customarily forage actively and socially, the calls apparently conveying information about association behavior and helping the birds to maintain the coherence of their social groups.

While other messages about behavior are made available by the displays of some birds, most species appear to encode primarily the dozen messages listed in Table 1. The table may not contain all the messages that are widespread among the displays of diverse species, but it contains all those that appear at present to be widespread, and the eventual list may not be a great deal longer. (There are indications that the same list is also widespread among other classes of vertebrate animals; W.J. Smith, in press). Note that many of the messages are only broadly predictive of a communicator's behavior. The message of interaction, for instance, does not specify what kind of interaction to expect, that of locomotion does not specify the functional class of the locomotion, and the so-called general set message merely indicates that some of the acts that may be selected when a communicator displays are incompatible with other acts whose possibilities are also encoded by the display. In contrast, messages such as attack, escape, and copulation are much more narrowly predictive. It would appear that the evolution of the displays of most birds has responded to pressures favoring either messages that are serviceable in many different situations and very dependent on contextual sources of information to elicit appropriate responses, or messages that are less widely useful but that are more capable of eliciting precise responses immediately, with minimal dependence on other sources of information. It would also appear that only a few messages are either suitable to be very broadly useful or are needed in relatively narrowly defined circumstances.

A basic feature of bird communication that was probably crucial to the evolution of these limitations on the kinds of message that are widely used is the limited size of each species' display repertoire (W.J. Smith, 1969a). By traditional criteria for the recognition of display units, no bird appears to have more than about forty to forty-five displays, and many have fewer (Moynihan, 1970). Further, these tallies include both audible and visible displays, and the two classes usually overlap very considerably in the kinds of information they make available. With such small display repertoires, it is evident that a species that ties down too many of its displays with narrowly predictive messages will be unable to display at all in a great many circumstances. Evolution should favor precise messages in special circumstances, but it appears that in most species this sacrifice of displays for other than attack, escape, association, and copulation messages is not adaptive.

Because this argument turns on the fact that birds (and other animals) have only limited repertoires of display behavior, the problems discussed earlier that are inherent in the traditional display concept become a matter of concern. That some acts may be only slightly or partially specialized to be informative may not be crucial, as these are necessarily poorly distinguished from nondisplay acts and thus probably not as efficient in providing information as are more striking displays; nonetheless, it would be useful to have an estimate of the magnitude of this problem. The chief difficulty may lie in the fact that displays vary in form and even intergrade. Yet the concept of displays as vehicles for information suggests a means of resolving this issue. When a communicator is found to use markedly variable or intergrading displays, their different forms are usually found to vary in correlation either with differing probabilities of intensities of communicator acts (see below) or with different kinds of acts (different behavioral selections). If the specification of a particular set of behavioral messages is the fundamental task of a display, then we might accept a shift in the kinds of message encoded as marking the boundary between related displays and amplify the traditional definition of display behavior accordingly. Using this criterion we can determine the number of behavioral packages for different sets of information that are available to an individual of any species. Present indications are that the number of displays estimated by this criterion is usually very similar to the numbers estimated by somewhat more arbitrary criteria of form.

Displays provide more information about communicator activities than simply an indication of the selections that an individual may make from its behavioral repertoire. Supplemental information (see Table 2) is made available for each selection that gives at least the probability of its being performed. Although a display may occur simultaneously with a behavioral selection it is also likely to precede it or to occur when there is some possibility that the selection will follow. And since many, perhaps most, displays provide information about two or more incompatible selections it is obvious that only one can follow, or at least be the first to follow. For instance, the green-backed sparrow has at least three vocal and three visible displays that are performed in agonistic circumstances: some when attack is more likely than escape, some when attack and escape are about equally probable, and some when escape is more likely than attack; in most of these circumstances, indecisive behavior is more likely than either attack or escape (Moynihan, 1963).

It is easier to compare the probabilities of different behavioral selections from among these displays than it is to measure the probability for any one display, because the use of each display can change the circumstances of the interaction. After a severe threat, for instance, an opponent may flee and in fleeing obviate the need for attack by the communicator. Severe threats may indicate a high probability of attack if the situation remains unchanged, but an observer may rarely see that eventuality in a close interaction. Other kinds of circumstances may change much less abruptly, on the average, than do agonistic encounters; but in principle the problem remains, and even a continuously static situation may alter a communicator's readiness to act. The probability that each behavioral selection encoded by a display will be performed is in effect only at the instant of displaying and cannot, therefore, be measured accurately; the probability information made available by displays is to some degree indeterminate for an observer. Nonetheless, it is of crucial importance to the displaying birds.

Table 2

Widespread behavioral supplemental and nonbehavioral messages.

In addition to indicating the probability that a behavioral selection will be performed, many displays appear to provide some information about the expected intensity of the activity. Thus some displays providing information about escape behavior may correlate primarily with slight avoidance movements or local withdrawal, others with headlong fleeing. Different displays with association messages may correlate with different distances among the participants or perhaps different degrees of readiness to let association be interrupted. Very little study of such correlations has been undertaken, however, and this message, although it appears to exist independently of the probability message and need not vary in parallel with it, is still largely unknown.

In being repeated in a predictable pattern, displays can also provide information about the stability of a communicator's behavior or behavioral predispositions. As songbirds utter bouts of singing each morning and evening twilight, for instance, they reaffirm their continuing readiness to defend their territories. Schleidt (1973) has termed this procedure "tonic" communicating.

Finally, some displays also provide information about the direction to be taken by the communicator's attending or other behavior. That is, without specifically naming its intended recipients, the whereabouts of a predator, or other external entities such as a nest site, a bird can provide effectively comparable information by indicating the direction in which it will attack, from which it will flee, etc. Much of this is done by employing formally restricted angles of orientation, as in the mutual adoption of a parallel relationship between two greeting gulls. A male green-winged teal (Anas carolinensis) indicates that a female is the object of his attentions by orienting broadside to her when performing seven different displays and swimming directly away from her when performing an eighth (McKinney, 1965). When he is distracted by other males and is trying to avoid them to align with the female he uses yet another display in a more variable orientation that reflects his divided attention. A female mallard (Anasplatyrhynchos) will perform Nod swimming to a group of males, but with the Inciting display takes a fixed bodily orientation with respect to one individual she has chosen as her mate (Weidmann and Darley, 1971). Both displays appear to indicate that she is ready to interact, but the latter indicates that she will take a greatly narrowed direction. Among many other possible examples of the importance of formally restricted orientations and the directional information they make available are two cases in the Triumph Ceremony of the gray-lag goose (Anseranser: Fischer, 1965). In the first phases of this display a gander displays and orients directly toward the gander of another family and approaches to attack him. In the second phase he returns to his own family using the same display postures and movements, but carefully maintaining an oblique orientation to each bird; they respond by similar displaying and adoption of mutually oblique orientations.

If the information provided by displays were entirely concerned with behavior then most birds would be awash in sources of information, relevant and irrelevant, and not know to which to attend. In fact, however, ethological studies of bird communication have demonstrated that a great deal of the information provided identifies communicators.

Most vocal displays identify the species of the bird that uses them, particularly the louder displays, which are effective over large distances. Visible displays may not need to do this if they reveal or are seen along with badges that accomplish the same task, and there are also some vocal exceptions. For instance, among species that flock together, the "alarm calls" that are uttered on sighting a predator may converge onto a common form, usually one with physical characteristics that make the source of the sound difficult to locate by a binaural recipient (Marler, 1955a). Identifying information is not wholly lost in such cases, but it specifies membership in an assemblage of interdependent species rather than in a single species.

Within a species some aspects of the form of displays may vary regionally and so identify local populations. Detailed mapping of the precise geographic distributions of such dialectal differences has very rarely been done with samples sufficiently large to show whether variation is continuous or stepped, however, and much remains to be learned about the limits of dialectal regions (Thielcke, 1969). The exact amount of variation that is truly geographical is also not known for most existing samples, as little attention has been paid in most cases to the behavior of the birds that were being recorded (see critique by S. T. Smith, 1972).

Many displays have pecularities of form that are specific not to populations of individuals but to single individuals. (For a detailed review of the numerous observations and experiments demonstrating this phenomenon in the vocal displays of birds see Beer, 1970.) Individual identifying information permits the formation of bonds between mates, parents and offspring, and even territorial neighbors. In some species, infants still in the egg learn to identify the pecularities in the voices of their incubating parents (Tschanz, 1968; Norton-Griffiths, 1969; Beer, 1970).

Displays may make available information that identifies not only individuals and populations but also the sex and, sometimes, the age classes to which communicators belong. For instance, in some birds song is used by only one sex, usually the male. Care has to be taken before concluding this for any species, however, as members of the opposite sex may sing on the relatively rare occasions that they take on the role of advertising. And there are species for which the use of song identifies the communicator as a female, e.g., the red phalarope (Phalaropus fulicarius: Tinbergen, 1935).

Finally, in some species displays take forms indicating that the communicator is a member of a particular group of bonded individuals. In the American goldfinch (Spinus tristis), for instance, one vocalization is adjusted in the course of the year to identify the individual's bonds in a pair, a family, and a small flock, as each of these successive groups becomes important (Mundinger, 1970). In a Trinidadian hummingbird known as the little hermit (Phaethornis longuemareus), the males who gather near each other as a distinct group on a larger lek all share a single song form, different from that of other groups singing on the same lek (Wiley, 1971). In many species, members of mated pairs come to share combinations of vocalizations, which they use in duetting performances that are distinct from the performances of neighboring pairs (discussed below).

One other category of nonbehavioral information is provided by vocal displays: information that enables a binaural recipient to locate the source of the vocalization by comparing temporal changes in the phase and amplitude of the sound reaching its two ears (Marler, 1955a). Except in the relatively few cases in which these clues are reduced by natural selection, the complexity of bird vocalizations usually appears sufficient to provide this information in abundance. The complexity of form of a vocalization determines both its locatability and how many kinds of identifying information it can carry, how much of the hierarchy of classes can be represented. Within the limits of form, each display provides sufficient identifying information to permit it to be useful in many kinds of circumstances (Marler, 1959).

Another category of information that has often been said to be encoded in some bird displays is information about the "environment," for example, about hawks or habitats. However, there is still no convincing evidence that referents external to the communicator are encoded by formal behavior in any species except humans (see W.J. Smith, in press). Environmental information is often available in events when birds are communicating, but its sources would not appear to include their formal acts.

RESPONSES AND FUNCTIONS

Communication occurs only when information is shared between or among individuals. While it is essential to determine the informative potential of displays, badges, and other sources, it is also necessary to determine the kind of response to which this information contributes if the process of communicating is to be understood. Further, it is the responses that lead to the adaptive advantages of sharing information, the functions of the process, for both the suppliers and the recipients of the information. As a branch of evolutionary biology, ethology has been very interested in adaptive significance, and hence concerned with the functions of communicating. Unfortunately, research on functions and the responses that generate them is not easy. When a bird that is apparently a recipient of, say, a display does something, it is often very difficult to know to what the bird is responding, since many sources of information are available to it. Further, its response may be largely inaccessible to an observer, for instance, a slight alteration of the bird's state of responsiveness to subsequent stimuli but not an immediate and overt change in its behavior.

There has been some tendency to oversimplify the role of recipient individuals in events in which communication occurs. In part this is due to the ethological theory that displays act as "releasers" of responses (see Lorenz, 1950;Tinbergen, 1959). The theory is soundly based on the observation that animals do treat some stimuli as being especially significant for particular classes of responses, and on the knowledge that perceptual processes organize stimuli according to various preestablished criteria of salience. Carried to extremes, however, this perspective underestimates the dependence of recipients on the relationships among stimuli. Thus, while a particular badge—for instance, the red spot on the mandible of an adult herring gull (Larus argentatus)— may be very important in eliciting and directing pecks by its hungry chicks (Tinbergen and Perdeck, 1950), as the chicks mature and gain experience they learn additional characteristics of their parents' heads and come to demand much more detail of the visual stimuli at which they will peck (at least in laughing gulls Larus atricilla: Hailman, 1967, 1969). The releaser is not a key to unlock responses, but a stimulus specialized to be particularly noticeable to recipients, and to be accepted as that stimulus to which other stimuli have a contextual relationship in some particular frame of reference (W.J. Smith, in press).

Much experimental work has been done with birds in testing the responses elicited by badges and displays. Birds' appearances have been modified by removing or altering badges by clipping feathers or painting fleshy areas (see references to studies by Peek and by N. G. Smith under the initial discussion of badges, above) and by dying feathers. For instance, Marler (1955b) gave female chaffinches reddish breast feathers matching those of the males, and found that the altered birds usually became dominant over other females and under some conditions were even successful in agonistic encounters with males. Stuffed birds or portions of mounts have been used to elicit aggressive responses (e.g., by Lack, 1940, with the European robin, Erithacus rubecula) or sexual display and mounting (e.g., MacDonald, 1968, with the spruce grouse, Canachites canadensis). Models of birds have been used to test responses to particular postures; for example, Stout and Brass (1969) sought differential responses by placing two models in different postures within territories of breeding glaucous-winged gulls (Larus glaucescens). However, all these methods of providing specific visible stimuli share in some degree the problem of appearing unnatural. A female colored to look like a male will not always act like one, and a bird altered to look like another species may persist in trying to mate with members of its own. More serious, a stuffed mount or model is unresponsive. Even if it is built so that its poses can be changed, it cannot be manipulated to respond as naturally as its recipients would expect of a real bird. Static models elicit responses, but beyond the initial moves of their recipients the circumstances are difficult to interpret.

Yet another procedure is to record and play back vocal displays. This technique has been used frequently to test whether birds can discriminate their own song (or sometimes just an unstudied and variable sample of their vocal repertoire) from that of related species (e.g., Dilger, 1956;Thielcke, 1962; Lanyon, 1963; Gill andLanyon, 1964; Thompson,1969), or can recognize the song of their own subspecies (e.g., Thönen, 1962) or local dialect (e.g., Lemon, 1967). Tests of this sort have even shown individual recognition, as territorial males distinguish between the songs of their neighbors and of more distant males (e.g., Weeden and Falls, 1959; Falls, 1969; S. T. Emlen, 1971; Goldman, 1973). Although the responsiveness of test individuals to playback lessens with habituation and varies with such factors as time of day and phase of the nesting cycle (Verner and Milligan, 1971), and the strength of a response can be difficult to measure (see S. T. Emlen, 1971, 1972), the technique has revealed a good deal about the identifying information made available by vocalizations. Further, because sound recordings can readily be modified by filtering frequencies, rearranging sequences, and altering intervals it has been possible to ask what physical components of bird song carry the identifying information (e.g., Busnel and Brémond, 1961, 1962; Brémond, 1968a, 1968b; Falls, 1963, 1969; Schubert, 1971; S. T. Emlen, 1972; Helb, 1973). The pure tones of species such as the whitethroated sparrow (Zonotrichia albicollis) can even be reproduced with audio oscillators and then modified in predetermined ways to study which characteristics provide species- and which individual-identifying information (Falls, 1969).

Playback has also been used to test the responses of females to song (e.g., Falls, 1969; Milligan and Verner, 1971; Payne, 1973b), of parents and their offspring to each other's vocalizations (see review by Beer, 1970), and of various species to their own and each other's alarm calls (e.g., Brown, 1962; Curio, 1971; Thielcke, 1971). Where responses to more than just the identifying information are sought, however, the technique encounters the same problem as does the use of models—the sound source is unresponsive to the actions of its recipients. Further, it is necessary to accompany the sound with a model or a mirror (e.g., Stout et al., 1969) to provide a visible participant, and this involves adding the information from a static posture or a careful mimic. This is not to argue that playback and the use of models and mounts are not necessary procedures, but that their employment introduces information that makes the experimental circumstances unnatural and hence difficult to interpret. The further development of techniques for presenting paired stimuli, done with careful attention to maximizing the naturalness of at Jeast the initial impingement of these stimuli on the test animal, may contribute a good deal to our understanding of responses.

If we do not experiment we are left with the problem of deciding what responses accrue to displays and other formal information sources in natural events in which we lack control over many relevant variables, and we may find it difficult to quantify observations. Nonetheless, natural events do provide very useful clues, and under some circumstances particular encounters may recur very frequently. Tinbergen (1959) has described a very fruitful procedure, which he refers to as seeking "natural experiments." In one of the examples he gives, an act is repeated several times without eliciting overt responses from nearby individuals, and then a largely similar act is performed with a display and they do respond: A male black-headed gull (Lams ridibundus) gathering nest material on his territory repeatedly approached five individuals who were resting just outside. While he was preoccupied they continued to rest. Then he stopped, preened near his nest and again approached them, now in an Aggressive Upright posture. When he came within three meters they all adopted Anxiety Uprights and walked away. Since an observer sees the Aggressive Upright primarily or only in circumstances in which the communicator appears to be threatening, such an instance tends to confirm the suspected functions of intimidation and territorial defense and the expected response of withdrawal. Other responses are also seen when the recipients are different; e.g., counterthreat would be one response of a territorial neighbor.

Slight manipulations of natural situations can help an observer obtain repeated samples of a given kind of event. For instance, in order to increase the incidence of agonistic encounters that he could observe in a winter flock of blue tits (Parus caeruleus), Stokes (1962) provided a rich food source, and by stringing peanuts and chunks of coconut on a wire, he forced individuals to remain at it while eating. He counted the number of times communicators used particular displays in the presence of one other individual and tallied the subsequent behavior of each participant as attack, escape, or staying. The recipients behaved appropriately following displays that correlated with different communicator activities; e.g., the probability of the recipient's fleeing after three of the displays was about twice what it was when none of them was used, even when Stokes analyzed only events in which the communicator did not actually attack. This suggests that the displays and/or such other sources of similar information as intention movements and known status relationships were influencing recipient behavior.

Natural experiments and slight manipulations of natural circumstances will probably continue to provide the least-distorted insights about the responses made to displays and other formal sources of information. The advantages of more-controlled experiments will be necessary, however, and techniques need to be developed that do not grossly violate the expectations of the recipient birds. For instance, if models are placed in a territory they should mainly be in postures that a territorial intruder might adopt, and these are typically not those correlated with the highest probabilities of attack. By using more than one experimental technique in conjunction with natural observations the contributions of each procedure can be pooled and their disadvantages minimized.

The study of functions is also based on observation of natural events and involves consideration of the ways in which the responses made by the recipients of displays are adaptive for them and for the individuals who make the information available. Some functions are obvious. A bird that flees into cover when a companion utters an "alarm call" may avoid being caught by a hawk. Many other functions are less obvious. In the same example, for instance, the call may help the bird that utters it to escape if the sudden scattering of its companions confuses the hawk, but it may not help otherwise. Still, the call may function for that communicator in various ways; for example, by protecting the investment it has made in its offspring, if they are in the group, and by helping to keep alive members of the group who in subsequent events may give it warning when hawks approach.

The functions of a display in any given event may be quite different for a communicator and for a recipient, as they are in the above example. They may also differ among recipients, and differ from the same display in different kinds of situations—particularly when the displays encode broader behavioral selection messages. Further, the functions of a single communication event may have to be stated for more than one time span. Thus the immediate functions of, say, a male bird's song may be to attract an unmated female and to repel a neighboring male. In the longer view, the attraction of a female is a necessary step if the male is to pass on his genes to a future generation, and the repelling of the neighbor may help train that individual to remain away, thus reducing subsequent competition for the female. Thus, the functional assessment of displays is a complex task.

Many functions have been proposed for the displays of birds, although few have been thoroughly studied or experimentally tested. Cautioning that much more study is needed, Thielcke (1970) has reviewed the literature on bird vocalizations and offered the following descriptive list of functions: territorial defense, attraction of a mate, maintenance of a pair bond, mutual stimulation of mates, synchronization of pair activities, facilitation of the simultaneous hatching of the eggs in a clutch, familial recognition, group coherence, assembly of roosting groups, attraction to feeding sites, control of agonistic encounters, maintenance of pair relations and facilitation of the activities in which mates cooperate, raising of young, and alarming of young and other conspecific individuals.

Bird Song

Although much research has been done on display postures and movements, and on the vocalizations usually referred to as "calls" by ornithologists, a great deal of attention has centered on a kind of display that is complexly developed and frequently used by many species of birds: song. Other animals sing, of course, from cicadas and frogs to humpback whales (Megaptera novaeangliae: Payne and McVay, 1971) and titi monkeys (Callicebus moloch: Moynihan, 1966), but bird songs have attracted human attention for a very long time, even if initially primarily through their aesthetic appeal. They are interesting for reasons beyond aesthetics, however. Studies have now revealed that in their organizational complexity and in the ways in which they are learned, the songs of some birds show very interesting parallels with human speech behavior, parallels further extended by the use of song in formally patterned performances that require more than one participant.

Song is not a precisely defined behavioral category (see Thorpe, 1961; Armstrong, 1963). Although complexity of form is an important attribute of most of the songs that are intensively studied, many species have comparable performances in which the songs are very simple (e.g., the chipping sparrow, Spizella passerina: Borror, 1959; Marler and Isaac, 1960; and the chiffchaff, Phylloscopus collybita: Schubert, 1971). Further, some species have complex vocalizations that would not usually be called songs or that comprise a series spanning the distinction between songs and "call notes" (e.g., in the genera Chlorospingus: Moynihan, 1962c; Tyrannus: W. J. Smith, 1966; and Parus: S. T. Smith, 1972), and in many species the song vocalizations are partly or wholly specialized amalgams of call notes (Howard, 1920; Thorpe, 1961; Immelmann, 1969). The singing performance, in fact, is a better indicator than are the forms of the vocalizations. The most useful working definition of song seems to be a vocalization or a set of vocalizations that is repeated in more or less continuous, regular patterns, often in sustained bouts.

Recent research on bird song has concentrated on a number of issues. One is the identifying information songs carry, as is discussed above. Another is the temporal organization of the complexity, which makes many of them so distinctive among vocal signals. Because portions of this complexity may be typical only of local populations, the question has arisen as to how it is passed from generation to generation, leading to very detailed studies of song learning. Finally, the specialized behavior of duet singing has brought considerable attention to an interesting set of formalized interactions.

Before reviewing this research, a caution is in order. While the repetitive use of patterned vocalizations is an extremely interesting phenomenon, it should not be assumed (as it very often is) that the remainder of the vocal repertoire is necessarily of less interest. In fact, in most species the non-song vocalizations are used much more frequently than the songs, and are much more likely to be employed during active interactions. Birds communicate with whole repertoires of displays, not just with their most complex vocalizations.

COMPLEXITY AND TEMPORAL PATTERNING

The song units of many bird species are complex in that they are made up of more than one sort of component, arranged in an orderly fashion (see, e.g., the descriptions of chaffinch song and its three subsections in Thorpe, 1961). In some species each male has only one such complex song unit, but in others he characteristically has two or more, sometimes a great many. The different song forms or song types in a male's repertoire typically consist of rearrangements of a limited set of the basic components, and in several species at least some of these rearrangements have been shown to be rule-bound. For example, in the mistle thrush (Turdus viscivorus: Isaac and Marler, 1963) most of the components are not repeated while other components always are; in the olive-backed thrush (Hylocichla ustulata: Nelson, 1973) each component tends to be a variant or an elaboration of the preceding one; in the rose-breasted grosbeak (Pheucticus ludovicianus: Lemon and Chatfield, 1973) the components of songs occur in regular couplet and triplet combinations; in the yellow-throated vireo (Vireo flavifrons) approximately nineteen kinds of components can be divided into three categories on the basis of gross form, and the various positions members of these categories take within different song forms are limited in several ways (Smith, Pawlukiewicz, and Smith, ms.). Because an extremely large number of recorded samples must be obtained, the songs of none of these species have yet been described in sufficient detail to exhaust the limits of the variability and permit us to write rules that are characteristic of the entire species population. Nonetheless, the phenomenon is well established. It is interesting that it is comparable to one of the grammatical levels of human speech: the basic, distinctive, vocal units (songs and words or morphemes) are in both cases constructed of a limited number of components that are sequentially arranged according to systems of rules.

In human speech the grammatical level of which we are most aware is the organization of words into phrases, so the question arises whether the sequences of bird songs are also patterned and, if so, how complexly. Evidence for sequential patterning of songs is again available from diverse species. Ohio song sparrows, for instance, go through their entire song repertoires before they repeat runs of any one kind of song (Nice, 1943), and members of the species from central California, who have larger song repertoires, usually go through about ten different song forms before repeating a run of any one (Mulligan, 1966). Cardinals (Lemon and Chatfield, 1971) show a first-order Markovian relationship between successive song forms; i.e., each is largely determined by the one that went before, if they are not separated by a relatively long pause. In the eastern wood pewee (Contopus virens: Craig, 1943) the long daily bout of predawn singing comprises three song forms arranged in a number of orderly sequences, and singing during the day employs two of the songs according to yet other rules (W.J. Smith, 1968, and in prep.). Members of another tyrannid genus, the phoebes (Sayornis, three species), employ either two or three different units—one a very variable vocalization—in patterned bouts governed by rules similar to those apparent in the daytime singing of the pewee: one song form occurs primarily singly and terminates runs of the other unit(s). The runs vary in length in a nonrandom fashion (W.J. Smith, 1969b, 1970). The yellow-throated vireo, however, has much greater freedom: individual males use up to eight song forms in a number of couplet combinations, most of which have members that overlap with those of one or more other couplets and so can be used to produce recurrent runs of triplets; a few longer sequences are also found (Smith, Pawlukiewicz, and Smith, in press.). In populations of other vireo species, such as V. solitarius in the Chiricahua Mountains of Arizona, individual males may have as many as seventy-five different song forms (Smith and Smith, in prep.).

Since at least simple combinations and recombinations of song form are rule-bound in these species, birds can be said to have singing grammars that are in part analogous to the phrase-structure grammars of human speech. In one very important respect, however, they appear to fall short. There is no reason to suspect, on the basis of existing evidence, that the various song forms of any species are differentiated into lexical categories such as the nouns and verbs with which we are familiar in English. (This lack of categorical distinction greatly limits the possible informative diversity of singing.) Rather, it appears that each song, like each call, encodes information about the same referent classes as in every other avian vocalization: behavioral selections and supplemental behavioral classes, classes of membership that identify the singer, and the class of "location" information.

The considerable experimental evidence for the nonbehavioral classes of information that are made available by bird song is reviewed briefly in the section on responses and functions, above. Experimental testing of the behavioral message classes is much more difficult, however. Further, since birds very often sing when they are not interacting, it is difficult to formulate precise and appropriate descriptions of the behavioral messages from observations of naturally occurring events. Nonetheless, a number of studies have reported correlations between the use of different song forms and different activities of the singer (see Armstrong, 1963, for a review), and in some cases in sufficient detail to suggest how messages differ among the song forms or their formal combinations.

Carolina chickadees, for instance, have two different song forms, either of which can be used loudly or faintly (S. T. Smith, 1972). The more commonly used form is sung by territorial individuals seeking any of a considerable range of kinds of interactions with their mates or neighbors; with use of its fainter version the singer is less likely to attack. The other song form correlates almost entirely with the most active kinds of seeking behavior, e.g., patrolling and intruding into neighboring territories.

In Africa, males of the parasitic indigobirds (genus Vidua) have both nonmimetic songs, which are peculiar to the genus, and mimetic songs, with which each male mimics a host species that is used by indigobirds to raise their young. The nonmimetic songs correlate with agonistic behavior in which the singer chases other males from his singing station, and the mimetic songs correlate with visits by females, becoming especially frequent if a female flies from a male's station without having completed a mating sequence (Payne, 1973a).

Various studies have shown that the two or more songs of many species of New World warblers are used in different circumstances (e.g., Ficken and Ficken, 1965, 1966; Morse, 1966, 1967, 1970; Lein, 1972). The most detailed study is that of Lein (1973) on the chestnut-sided warbler (Dendroica pensylvanica), which he has shown to have at least five song forms, each of which can also be shortened or muted. One is used primarily in territorial encounters; two are used primarily at territorial borders, or at least away from the center of the territory; and the remaining two are used almost exclusively in the interior. The distance from a singing neighbor also influences the choice of song. As the probability of a territorial encounter increases or decreases while a male is singing (e.g., by the onset or cessation of a neighbor's singing, movement of either male toward a border), he switches song forms in a very predictable fashion.

The two song forms of the eastern phoebe (Sayornis phoebe: W.J. Smith, 1969b) may be alternated, or one may occur in runs terminated by single occurrences of the other, so that their proportional representation in singing bouts varies; the varying proportions correlate with different activities of the communicator. At one extreme the bird is very likely to interact if it gets the opportunity. At the other it is more likely to forage or preen, and if it can interact it will usually just associate, not attempting to make contact (e.g., to fight or to copulate). In the daytime singing of the eastern wood pewee, which also comprises differing proportions of two song forms, runs of different lengths of one of the forms correlate with different probabilities of the communicator's flying, apparently providing a measure of its activity level.

The story for the yellow-throated vireo is much more complicated. Each male has a number of patterns of singing and uses different selections of patterns when patrolling his territorial borders, attempting to confront a neighbor, foraging away from the nest but not in a boundary region, approaching the nest, or remaining on or in the vicinity of the nest. Patterns correlated with the last activity are the simplest, those used in patrolling the most complex. Unmated males patrol and tend to sing complex patterns that resemble those used by mated males when the latter are patrolling. Of the eight males studied, no two sang identical selections of song forms or identical patterns (Smith, Pawlukiewicz, and Smith, in press.). However, they had many similarities and did not contradict the uses of one another's patterns, and all differed in important details of their circumstances, such as the presence or absence of neighboring males, having mates, the attentiveness of their mates, and their successes in nesting. Although the analyses in that study were based on 9,419 songs recorded during over forty-one hours of observations, the variation among the males and the complexity of the correlations between behavior and singing patterns suggest that a much larger sample will be required to reveal more precisely the messages of the song forms and combinations and to clarify the species' grammatical rules.

SONG LEARNING

The basic forms of most bird displays develop even in individuals experimentally reared in isolation. Thus they do not involve learning based on hearing and seeing other members of the species perform. Remarkably, this is true even for quite complex songs, for example, those of the song sparrow (Melospiza melodia: Mulligan, 1966). However, there are species in which development of the full form of complex songs does require such learning and also subsequent learning through practice.

In these species young birds must usually hear appropriate songs, which will act as models, during a critical period early in their lives, well before they start to sing. By experimentally restricting what potential models they can hear, it has been shown that what they will accept is limited. In chaffinches and white-crowned sparrows (Zonotrichia leucophrys), for instance, it has to be the song of their own species or of a species with a similar song pattern. In nature, of course, they will almost always have the opportunity to hear their own species sing during the period in which they are fledglings and, in fact, will obtain as models songs that are typical of their own dialect group. In the bullfinch (Pyrrhula pyrrhula: Nicolai, 1959) and in various grass finches (Estrildidae: Immelmann, 1969) it is the song of the male that helped rear them: the father or, in experimental conditions, a foster father of another species. (In parasitic indigobirds what is learned appears to be the song of the host species raising the young; whether their species-specific songs require hearing a model has not yet been tested, but the appropriate model might not be available to a young bird; see Payne, 1973b, and above.) There tends to be a second learning period in which singing is practiced, and a variable, rambling "subsong" is perfected into a full adult song. This may come months after the auditory model is stored. The experience of countersinging during this practice enables members of some species (e.g., the chaffinch) to elaborate on their learned model. Inventive elaboration has also been found even in isolated cardinal males (Richmondena cardinalis: Dittus and Lemon, 1969), but it is not characteristic of other species such as the estrildids. The latter have an early end to the critical period in which they can learn by hearing other individuals, and this is correlated with a very early onset of sexual maturity.

There have been many studies of song learning in which young birds of various ages have been acoustically isolated. By controlling their exposure to sounds this research has determined many of the characteristics of the sounds that the birds will accept as models. By deafening and other procedures the roles of auditory and proprioceptive feedback have been studied, and by adopting procedures from psychology the reinforcing properties of song have been investigated. The following reviews provide a detailed coverage of these experiments: for the chaffinch, Thorpe (1958, 1961), and Hinde (1969); for the white-crowned sparrow, Marier (1970); for several species of sparrows, Marler (1967); for estrildid finches, Immelmann (1969); and for general coverage Hooker (1968), Konishi and Nottebohm (1969).

Among the most interesting things to emerge are a number of similarities between the learning of song by birds and of speech by humans. Marler (1970) reviews these, pointing out that both involve critical periods before sexual maturity in which young individuals select a set of sounds that they will learn from the many available to them in natural circumstances. They are predisposed to recognize the appropriate set and will work at reproducing it without extrinsic reinforcement, practicing through a stage known as "subsong" or "babbling," respectively, until they achieve an adult form that they recognize through a process that requires auditory feedback. (Female birds may produce little song, but it has been shown that they do learn it, for testosterone injections will elicit malelike singing behavior; presumably their learning serves primarily to help them identify males of their own population with whom to form pair bonds.) Marler has suggested that these parallels are not unexpected in any system of vocal learning in which what is learned cannot be left to chance and which requires the development and refinement of skill. Why some species learn their songs and others do not is a more difficult question, although part of the answer would appear to lie in the use of songs by some species as means of identifying local populations.

The rhythm and phrase structure of bird song are also part of what is copied by many species (see, e.g., Güttinger, 1973), just as human neonates copy with their body movements the rhythmic organization of the human speech they hear (Condon and Sander, 1974; these patterns of body movement continue to be a part of listening behavior and become part of speaking behavior as the infants mature). In humans the learning of patterns of organization in this fashion may facilitate the learning of hierarchical grammatical patterns, but it is too early even to speculate whether this is necessary in the ontogenetic development of hierarchically patterned singing by birds.

DUETTING

It is not wholly appropriate to consider the topic of duetting under bird song as in many species the performances are based on non-song vocalizations, or even on such nonvocal sounds as drumming (e.g., in the hairy woodpecker, Dendrocopus villosus: see Kilham, 1960). Nonetheless, the recent increase of research activity on the topic has centered largely on studies of song duets, especially those of African shrikes of the genus Laniarius (beginning with Thorpe, 1963, and Thorpe and North, 1965). Birds in this genus also provide a case of the learning of pairspecific song patterns, as pairs develop repertoires of duets in which some of the songs are peculiar to individuals and each mate knows and responds to its partner's special forms (most recently reviewed by Thorpe, 1972). Learning, like song, is not essential to the duetting performances of many other kinds of birds, however. The basis of duetting is simply a formally patterned mutual performance that, by Thorpe's definition, is characteristically used by paired individuals.

Because duets cannot be performed by single individuals they are formalized interactions in the sense described at the beginning of this chapter, and are distinctive within that category primarily by being audible (usually vocal) performances by pairs. In principle, visible duets would be largely comparable, but the term has not been applied to postures and movements. When mates duet one leads and the other comes in either with the same vocalization, creating a polyphonic performance, or a different vocalization, yielding antiphony. Other individuals, offspring for instance, may sometimes join in. The duet then becomes a trio, quartet, or "communal" performance (Thorpe, 1972). In many species a mate of either sex will initiate a duet, although the male does so much more often than the female. In Laniarius and some other genera, if one mate is missing the other sometimes sings both contributions (the vocalizations do not overlap and the performance is no longer a duet, but it may serve to recall the missing partner).

Duetting performances have not been distinguished by precise criteria from formalized vocal interactions used as greetings or as appeasingreassuring ceremonies, and probably should not be, although Thorpe (following Armstrong, 1963) has ruled that patterned countersinging between neighbors of the same sex is not duetting. Because of the overlap between duetting and related performances, attempts to catalogue the taxonomic families of birds in which duetting occurs remain provisional. The present evidence, however, indicates that many families are involved. Some species that duet can be found in virtually every bird fauna throughout the world, although they may be proportionately more common in tropical than in temperate regions. Duetting species are typically birds that pair for life and are territorial for prolonged periods (although males and females may hold adjacent territories; see Kilham, 1960); often the sexes are very similar in appearance, and in many, but by no means all, cases the habitats are so densely vegetated that continuous visual monitoring of one mate by the other is hindered (Thorpe, 1972).

Thorpe's (1963) initial interest in duetting was spurred in considerable part by the precise timing with which some shrikes (genus Laniarius) and sylviid warblers (genus Cisticola) answer their mates, and by the brevity of the interval between the onset of the first bird's call and the onset of the answering call. He proposed that these intervals can be used to measure the auditory reaction time of birds, a physiological parameter that appears to be briefer than the corresponding reaction time of humans. It has subsequently been found, however, that duetting birds are frequently within sight of each other and use visible cues (Hooker and Hooker, 1969). Thus until a reasonably large sample has been recorded in which it is known that the answering bird is unable to see the initiator the value of this procedure for measuring auditory reaction time cannot be fully assessed. In addition, Payne (1970, 1971, 1973c) has argued that throughout a duetting sequence of vocalizations it is possible that each individual is following its own, autochthonous calling rhythm, a rhythm to which it can and does adhere even if its mate fails to call. He notes, for instance, that when duet sequences were recorded from a pair of L. barbaries over a period of more than an hour, the interval between a call by the first bird and the answer differed from sequence to sequence, which would not be expected if the interval were set largely by the auditory reaction time.

Birds do answer their mates' duetting calls much of the time, however, and there has been much speculation over the functions of a vocal ceremony that recurs as frequently as this one does. Thorpe has pointed out that for those species in which mates do duet when out of each other's sight in dense vegetation the vocalizations should help them remain in contact and aware of each other's whereabouts. The same applies to nocturnal duets of owls (e.g., J. T. Emlen, 1973) and perhaps to birds living in the dense mists and clouds of tropical mountaintops (Thorpe and Hall-Craggs, in Thorpe, 1972). Duetting must have other functions, however, as duetting vocalizations are much more complex than most "contact" calls, and because duetting birds are often in each other's sight: in some species they are usually perched beside each other, or within the same bush, or out in the open.

For species that form large flocks, one function of duetting with group-specific vocalizations might be to enable bonded individuals to remain together in the crowd. This has not been tested, although this possibility is suggested by the fact that pairs of such flocking species as orange-chinned parakeets (Brotogeris jugularis: Power, 1966) duet when agonistically aroused and that common crows (Corvus brachyrhynchos) duet when in large roosting or foraging flocks (Chamberlain and Cornwell, 1971).

Many duetting species that do not inhabit dense vegetation or join large flocks live in regions in which the seasonal changes that precede breeding are rapid and unpredictable in onset (Diamond and Terborgh, 1968), or where seasons are not sharply differentiated (Kunkel, 1966). Changes in their patterns of duetting might help mates to synchronize their physiological states and to change together rapidly enough to meet the demands of such circumstances. Yet the duetting continues all year, and so it must do more than that.

Duetting is just one of several kinds of mutual, formalized exchanges of display behavior that are characteristic of animals with persistent bonds, patterns that range from brief greetings to prolonged allopreening or allogrooming. Whatever else these exchanges accomplish, they probably serve the continuously necessary function of reaffirming the adherence of both individuals to their bonded relationships. Subsumed by this function is the mutual checking of each other's states or conditions—of the physiological "status quo," as Estes (1969) has put it. This may well be what has too often been described as the "identifying" function of duetting, as individual identification in any simple sense undoubtedly need not be repeated many times a day by formal behavior patterns between birds who know each other intimately. Not identification, then, but reaffirmation may be the most widespread function of duetting.

In addition to reaffirming their relationships, duets can serve more-immediate functions for a pair. They may enable the birds to coordinate their activities in such events as nest relief, in which the birds take turns incubating the eggs. Duets may often serve as greetings and as appeasing-reassuring performances in other circumstances, as seems to be the case in diverse species of tyrannid flycatchers (W. J. Smith, 1971, and research in progress). Further, in many cases duetting vocalizations must function in the wider social sphere beyond the pair. They are, for instance, audible to pairs on neighboring territories. Duetting by one pair can be answered by the duets of a neighboring pair, with the subsequent development of territorial countersinging. At least in the bell shrike (Laniarius aethiopicus) the timing of the duets of each pair degenerates within such encounters, and all the birds begin to countersing "more or less at random" (Thorpe, 1972). Yet in some species countersinging itself may be patterned in ways that are comparable to duetting patterns. For instance, neighboring pairs of the slate-colored bou-bou shrike (Laniarius funebris) tend to imitate each other's duets in countersinging (Wickler, 1972). Conspecific neighboring males of cardinals (Lemon, 1968a), pyrrhuloxias (Pyrrhuloxia sinuata: Lemon and Herzog, 1969), and black-crested titmice (Parus atricristatus: Lemon, 1968b) will match each other's selections of song forms when countersinging. Pairs of these species do not duet in such territorial disputes, but when mated cardinals do duet the female matches the male's song forms in replying to him.

No published studies of duetting include sufficient observations of the behavior of the duetters to permit analyses of the behavioral messages of the vocalizations, although some information about readiness to interact is being made available by the duetting vocalizations. Species in which individuals have large repertoires of duetting vocalizations may be able to provide different behavioral messages by selecting different vocalizations. Individuals of Laniarius aethiopicus, for instance, each have several flutelike notes as well as a "snarling" vocalization (Thorpe, 1972). The last seems to be used in duets by individuals who are "aggressive," although details about the likelihood of attack behavior have not been published. On the other hand, it is amply evident that duetting vocalizations provide nonbehavioral information about each communicator's location and about several features of its identity: its species and sometimes its individual identity, in some cases its sex (see Hooker and Hooker, 1969; Todt, 1970; Thorpe, 1972), and (by being used in duets with particular other individuals) its bonded relationships.

References

Ainley, D. G., 1974. Displays of Adelie penguins: a re-interpretation. In: The Biology of Penguins, B. Stonehouse, ed. London: MacMillan.

Andrew, R. J., 1961a. The motivational organisation controlling the mobbing calls of the Blackbird (Turdusmerula): I, II, III, and IV. Behaviour,17:224-46, 288-321; 18:25-43, 161-76.

Andrew, R. J., 1961b. The displays given by passerines in courtship and reproductive fighting: a review. Ibis, 103a:549-79.

Armstrong, E. A., 1963. A Study of Bird Song. NewYork:Oxford.

Bastock, M.;Morris, D.;Moynihan, M.; 1953 Some comments on conflict and thwarting in animals.Behaviour, 6:66-84.

Beer, C. G., 1970. Individual recognition of voice in the social behavior of birds. Advances Study Behavior, 3:27-74.

Blurton-Jones, N. G., 1968. Observations and experiments on causation of threat displays of the Great Tit (Parus major). Animal Behavior Monographs, 1:75-158.

Borror, D. J., 1959. Songs of the Chipping Sparrow. Ohio J. Science,59:347-56.

Brown, R. G. B., 1962. The reactions of gulls (Laridae) to distress calls. Ann. Épiphyties, 13:153-55.

Brémond, J.-C., 1968a. Valeur spécifique de la syntaxe dans le signal de défense territoriale du Troglodyte (Troglodytes troglodytes). Behaviour,30:66-75.

Brémond, J.-C., 1968b. Recherches sur la sémantique et les éléments vecteurs d'information dans le signaux acoustiques du Rouge-gorge (Erithacus rubecula L.). La Terre et la Vie, 2:109-220.

Busnel, R.-G., and Brémond, J.-C., 1961. Étude préliminaire du décodage des informations contenues dans le signal acoustique territorial du Rougegorge (Erithacus rubecula L.). Comptes rendus des Séances, Acad. Sei., 252:608-10.

Busnel, R.-G., and Brémond, J.-C., 1962. Recherche du support de l'information dans le signal acoustique défense territoriale du Rouge-gorge (Erithacus rubecula L.). Comptes rendus des Séances, Acad. Sei., 254:2236-38.

Chamberlain, D. R., and Cornwell, G. W., 1971. Selected vocalizations of the Common Crow. Auk, 88:613-34.

Chapman, F. M., 1938. Life in an Air Castle.Nature Studies in the Tropics. New York: Appleton-Century.

Condon, W. S., and Sander, L. W., 1974. Neonate movement is synchronized with adult speech: interactional participation and language acquisition. Science, 183:99-101.

Craig, W., 1943. The song of the wood pewee (Myochanes virens). N.Y. State Mus. Bull., 334:6-186.

Crook, J. H., 1963. Comparative studies on the reproductive behaviour of two closely related weaver bird species (Ploceus cucullatus and Ploceus nigerrimus) and their races. Behaviour, 21:177-232.

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

Curio, E., 1971. Die akustiche Wirkung von Feindalarmen auf einige Singvögel. J. Ornithol., 112:365-72.

Daanje, A., 1950. On the locomotory movements in birds and the intention movements derived from them. Behaviour, 3:49-98.

Darwin, Charles, 1872. The Expression of the Emotions in Man and Animals. London: Appleton

Delius, J. D., 1963. Das Verhalten das Feldlerche. Zeits. Tierpsychol., 20:297-348.

Diamond, J. M., and Terborgh, J. W., 1968. Dual singing by New Guinea birds. Auk, 85:62-85.

Dilger, W. C., 1956. Hostile behavior and reproductive isolating mechanisms in the avian genera Catharus and Hylocichla. Auk, 73:313-53.

Dittus, W. P. J., and Lemon, R. E., 1969. Effects of song tutoring and acoustic isolation on the song repertoires of cardinals. Anim. Behav., 17:523-33.

Emlen, J. T., 1973. Vocal stimulation in the Great Horned Owl. Condor, 75:126-27.

Emlen, S. T., 1971. The role of song in individual recognition in the Indigo Bunting. Zeits. Tierpsychol., 28:241-46.

Emlen, S. T., 1972. An experimental analysis of the parameters of bird song eliciting species recognition. Behaviour, 41:130-71.

Estes, R. D., 1969. Territorial behavior of the wildebeest (Connochaetes taurinus Burchell, 1823). Zeits. Tierpsychol., 26:284-370.

Falls,}. B., 1963. Properties of bird song eliciting responses from territorial males. Proc. 13th Int. Orn. Congr., 259-71.

Falls, J. Bruce, 1969. Functions of territorial song in the White-throated Sparrow, In: Bird Vocalizations, R. A. Hinde, ed. Cambridge: Cambridge University Press, pp.207-32.

Ficken, M. S., and Ficken, R. W., 1965. Comparative ethology of the Chestnut-sided Warbler, Yellow Warbler, and American Redstart. Wilson Bull., 77:363-75.

Ficken, M. S., and Ficken, R. W., 1966. Singing behavior of Blue-winged and Golden-winged Warblers and their hybrids. Behaviour, 28:149-81.

Fischer, H., 1965. Das Triumphgeschrei der Graugans (Ariser anser). Zeits. Tierpsychol., 22:247-304.

Gill, F. B., and Lanyon, W. E., 1964. Experiments on species discrimination in Blue-winged Warblers. Auk, 81:53-64.

Gilliard, E. T., 1956. Bower ornamentation versus plumage characters in bowerbirds. Auk, 73:450-51.

Gilliard, E. T., 1963. The evolution of bowerbirds. Scientific American, 209(2):38-46.

Goldman, P., 1973. Song recognition by Field Sparrows. Auk, 90:106-13.

Goodwin, D., 1953. Observations on voice and behaviour of the Red-legged Partridge Alectoris ruf a. Ibis, 95:581-614.

Güttinger, H. R., 1973. Kopiervermögen von Rhythmus und Strophenaufbau in der Gesangsentwicklung einiger Lonchura-Arten (Estrildidae). Zeits. Tierpsychol., 32:374-85.

Hailman, J. P., 1967. The ontogeny of an instinct. Behaviour, Suppl. 15.

Hailman, J. P., 1969. How an instinct is learned. Scientific American, 221(6):98-106.

Helb, H.-W., 1973. Analyse der artisolierenden Parameter im Gesang des Fitis (Phylloscopus t. trochilus)mit Untersuchungen zur Objektivierung der analytischen Methode. J. Ornithol., 114:145-206.

Hinde, R. A., 1956. A comparative study of the courtship of certain finches (Fringillidae). Ibis, 98:1-23.

Hinde, R. A., 1959. Behaviour and speciation in birds and lower vertebrates. Biological Reviews, 34:85-128.

Hinde, R. A., 1969. Bird Vocalizations: Their Relation to Current Problems in Biology and Psychology. New York: Cambridge University Press.

Hinde, R. A., 1970. Animal Behaviour: A Synthesis of Ethology and Comparative Psychology, 2d ed. New York: McGraw-Hill.

Hinde, R. A., and Tinbergen, N., 1958. The comparative study of species-specific behavior. In: Behavior and Evolution, A. Roe and G. G. Simpson, eds. New Haven: Yale University Press, pp.251-68.

Hooker, B. I., 1968. Birds. In: Animal Communication, T. A. Sebeok, ed. Bloomington: Indiana University Press.

Hooker, T., and Hooker, B. I. (Lade), 1969. Duetting. In: Bird Vocalizations,R. A. Hinde, ed. New York: Cambridge University Press.

Howard, E., 1920. Territory in Bird Life. London: John Murray.

Huxley, J., 1914. The courtship habits of the Great Crested Grebe (Podiceps cristatus); with an addition to the theory of sexual selection Proceedings Zoological Soc., 35.

Immelmann, K, 1969. Song development in the Zebra Finch and other estrildid finches. In: Bird Vocalizations, R. A. Hinde, ed. New York: Cambridge University Press.

Isaac, D., and Marler, P., 1963. Ordering of sequences of singing behaviour of Mistle Thrush in relationship to timing. Animal Behav.11:179-88.

Kilham, L., 1960. Courtship and territorial behavior of Hairy Woodpeckers. Auk, 77:259-70.

Konishi, M., and Nottebohm, F., 1969. Experimental studies in the ontogeny of avian vocalizations. In: Bird Vocalizations, R. A. Hinde, ed. New York: Cambridge University Press.

Kunkel, P., 1966. Quelques tendances adaptives du comportement de certains oiseaux tropicaux. Chronique de TInst. Rech. Sei. Afrique Central, 1(3):29-37.

Lack, D., 1940. The releaser concept of bird behaviour. Nature,145:107-108.

Lanyon, Wesley, E., 1963. Experiments on species discrimination in Myiarchus flycatchers. Amer. Mus. Novitates, 2126:1-16.

Lein, M. R., 1972. Territorial and courtship songs of birds. Nature, 237:48-49.

Lein, M. R., 1973. The biological significance of some communication patterns of wood warblers (Parulidae). Ph.D. diss., Harvard University.

Lemon, R. E., 1968a. The relation between organization and function of song in Cardinals. Behaviour, 32:158-78.

Lemon, R. E., 1968b. Coordinated singing by blackcreased titmice. Canadian J. Zool, 46:1163-67.

Lemon, R. E., 1967. The response of cardinals to songs of different dialects. Anim. Behav., 15:538-45.

Lemon, R. E., and Herzog, A., 1969. The vocal behavior of Cardinals and Pyrrhuloxias in Texas. Condor, 71:1-15.

Lemon, R. E., and Chatfield, C., 1971. Organization of song in Cardinals. Anim. Behav., 19:1-17.

Lemon, R. E., and Chatfield, C., 1973. Organization of song of rose-breasted grosbeaks. Anim. Behav., 21:28-44.

Lorenz, K., 1950. The comparative method in studying innate behaviour patterns. Symp. Soc. Experimental Biol., 4:221-68.

MacDonald, S. D., 1968. The courtship and territorial behavior of Franklin's race of the Spruce Grouse. Living Bird, 7:5-25.

McKinney, F., 1965. The displays of the American Green-winged Teal. Wilson Bull., 77:112-21.

Manley, G. H., 1960. Agonistic and pair formation behaviour of the black-headed gull. D. Phil, diss., Oxford University.

Marler, P., 1955a. Characteristics of some animal calls. Nature, 176:6-8.

Marler, P., 1955b. Studies of fighting in chaffinches. (2) The effect on dominance relations of disguising females as males. Brit.J. Animal Behaviour, 3:137-46.

Marler, P., 1956. The voice of the chaffinch and its function as a language. Ibis, 98:231-61.

Marler, P., 1959. Developments in the study of animal communication. In: Darwin's Biological Work, P. R. Bell, ed. New York: Wiley.

Marler, P., 1967. Comparative study of song development in sparrows. Proc. 14th Internat. Ornithol. Congr. Oxford: Blackwell's.

Marler, P., 1970. A comparative approach to vocal learning: song development in White-crowned Sparrows. Journal Comp. Physiol. Psychol. 71, Monograph 2(2): 1-25.

Marler, P., and Isaac, D., 1960. Physical analysis of a simple bird song as exemplified by the Chipping Sparrow. Condor, 62:124-35.

Mayr, Ernst, 1963. Animal Species and Evolution. Cambridge: Harvard University Press.

Milligan, M. M., and Verner, J., 1971. Inter-populational song dialect discrimination in the Whitecrowned Sparrow. Condor, 73:208-13.

Morris, D. J., 1956. The feather postures of birds, and the problem of the origin of social signals. Behaviour, 9:75-113.

Morris, D., 1966. Abnormal rituals in stress situations. Phil. Trans. Roy. Soc. B, 251:327-30.

Morse, D. H., 1966. The context of songs of the Yellow Warbler. Wilson Bull., 78:444-55.

Morse, D. H., 1967. The contexts of songs in Black-throated Green and Blackburnian Warblers. Wilson Bull., 79:64-74.

Morse, D. H., 1970. Territorial and courtship songs of birds. Nature, 226:659-61.

Moynihan, M., 1955. Some aspects of reproductive behavior in the Black-headed Gull (Larus ridibundus L.) and related species. Behaviour, Suppl. IV.

Moynihan, M. H., 1956. Notes on the behavior of some North American gulls. I. Aerial hostile behavior. Behaviour, 10:126-78.

Moynihan, M. H., 1960. Some adaptations which help to promote gregariousness. Proc. 12th Internat. Ornithol. Congr.: 523-41.

Moynihan, M., 1962a. Hostile and sexual behavior patterns of South American and Pacific Laridae. Behaviour, Suppl. VIII: 1-365.

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

Moynihan, M., 1962c. Display patterns of tropical American "nine-primaried" songbirds. I. Chlorospingus. Auk, 79:310-44.

Moynihan, M. H., 1963. Display patterns of tropical American "nine-primaried" songbirds. III. The Green-backed Sparrow. Auk, 80:116-44.

Moynihan, M., 1966. Communication in the Titi monkey, Callicebus. J. Zool„ Lond., 150:77-127.

Moynihan, M., 1970. The control, suppression, decay, disappearance and replacement of displays./ Theoretical Biology, 29:85-112.

Mulligan, J. A., 1966. Singing behavior and its development in the song sparrow, Melospiza melodia. Univ. Calif. Publ. Zool., 81:1-76.

Mundinger, Paul C., 1970. Vocal imitation and individual recognition of finch calls. Science, 168:480-82.

Nelson, K., 1973. Does the holistic study of behavior have a future? In: Perspectives in Ethology, P. P. G. Bateson and P. H. Klopfer, eds. New York: Plenum Press.

Nice, M. M., 1943. Studies in the life history of the song sparrow. Trans. Linn. Soc. N.Y., 6:1-328.

Nicolai, J., 1959. Familientradition in der Gesangsentwicklung des Gimpels (Pyrrhula pyrrhula L.)./. Ornithol., 100:39-46.

Norton-Griffiths, M., 1969. The organisation, control and development of parental feeding in the oystercatcher (Haematopus ostralegus). Behaviour, 34:55-114.

Payne, R. B., 1970. Temporal pattern of duetting in the Barbary Shrike Laniarius barbarus. Ibis, 112:106-108.

Payne, R. B., 1971. Duetting and chorus singing in. African birds. Ostrich, Suppl., 9:125-46.

Payne, R. B., 1973a. Behavior, mimetic songs and song dialects, and relationships of the parasitic indigobirds (Vidua) of Africa. A.O.U. Ornithol. Monogr. 11.

Payne, R. B., 1973b. Vocal mimicry of the paradise whydahs (Vidua) and response of female whydahs to the songs of their hosts (Pytilia) and their mimics. Anim. Behav., 21:762-71.

Payne, R. B., 1973c. (Review of) Duetting and antiphonal singing in birds, its extent and significance. Auk, 90:451-53.

Payne, R. S., and McVay, S., 1971. Songs of humpback whales. Science, 173:585-97.

Peek, F. W., 1972. An experimental study of the territorial function of vocal and visual display in the male red-winged blackbird (Agelaius phoeniceus). Animal Behav., 20:112-18.

Power, D. M., 1966. Antiphonal duetting and evidence for auditory reaction time in the Orange-chinned Parakeet. Auk, 83:314-19.

Raveling, D. G., 1969. Preflight and flight behavior of Canada Geese. Auk, 86:671-81.

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

Schleidt, W. M., and Shaker, M. D., 1972. Cloacal foam gland in the quail Coturnix coturnix. Ibis, 114:558.

Schubert, M., 1967. Probleme der Motivwahl und der Gesangaktivität bei Phylloscopus trochilus (L.)./. Ornithol., 108:265-94.

Schubert, G., 1971. Experimentelle Untersuchungen über die Artkennzeichnenden Parameter in Gesang des Zilzalps, Phylloscopus c. collybita (Vieillot), Behaviour, 38:289-314.

Sick, H., 1959. Die Balz der Schmuckvögel (Pipridae). / Ornithologie, 100:269-302.

Sick, H., 1967. Courtship behavior in the manakins (Pipridae): a review. Living Bird, 6:5-22.

Smith, N. G., 1966. Evolution of some arctic gulls (Larus): an experimental study of isolating mechanisms. A.O.U. Ornithol. Monogr., 4:1-99.

Smith, S. T., 1972. Communication and other social behavior in Parus carolinensis. Nuttall Ornithol. Club Publ. (M.C.Z., Cambridge, Mass.), 11:1-125.

Smith, W. John, 1966. Communication and relationships in the genus Tyrannus. Nuttall Ornithol. Club Publ. (M.C.Z., Cambridge, Mass.), 61:1-250.

Smith, W. John, 1968. Message-meaning analyses. In: Animal Communication, T. A. Sebeok, ed. Bloomington: Indiana University Press.

Smith, W. John, 1969a. Messages of vertebrate communication. Science, 165:145-50.

Smith, W. John, 1969b. Displays of Sayornis phoebe (Aves, Tyrannidae). Behaviour, 33:283-322.

Smith, W. John, 1970. Song-like displays in the genus Sayornis. Behaviour, 37:64-84.

Smith, W. John, 1971. Behavioral characteristics of serpophaginine tyrannids. Condor, 73:259-86.

Smith, W. John, in press. The Behavior of Communicating: An Ethological Approach. Cambridge: Harvard University Press.

Smith, W. J.; Pawlukiewicz, J.; and Smith, S. T.; in press. Kinds of activities correlated with singing patterns of the yellow-throated vireo. Anim. Behav.

Snow, D. W., 1958. A Study of Blackbirds. London: Allen and Unwin.

Stokes, Allen W., 1962. Agonistic behaviour among Blue Tits at a winter feeding station. Behaviour, 19:118-38.

Stout, J. F., and Brass, M. E., 1969. Aggressive communication by Larus glaucescens. Part II. Visual communication. Behaviour, 34:42-54.

Stout, John F.; Wilcox, C. R.; and Creitz, L. E.; 1969. Aggressive communication in Larus glaucescens. Part I. Sound communication. Behaviour, 34:29-41.

Thielcke, G., 1962. Versuche mit Klanattrapen zur Klärung der Verwandtschaft der Baumläufer Certhia familiaris L., C. brachydactyla Brehm und C. americana Bonaparte./. Ornithol., 103:266-71.

Thielcke, G., 1969. Geographie variation in bird vocalizations. In: Bird Vocalizations, R. A. Hinde, ed. Cambridge: Cambridge University Press.

Thielcke, G., 1970. Die sozialen Funktionen der Vogelstimme. Vogelwarte, 25:204-29.

Thielcke, G., 1971. Versuche zur Kommunikation und Evolution der Angst-, Alarm- und Rivalenlaute des Waldbaumläufers (Certhia familiaris). Zeits. Tierpsychol., 28:505-16.

Thompson, W. L., 1969. Song recognition by territorial male buntings (Passerina). Anim. Behav., 17:658-63.

Thönen, W., 1962. Stimmgeographische, ökologische und verbreitungsgeschlichtliche Studien über die Mönchmeise. Ornithologische Beob., 5/5:101-72.

Thorpe, W. H., 1958. The learning of song patterns by birds, with especial reference to the song of the chaffinch, Fringilla coelebs. Ibis, 100:535-70.

Thorpe, W. H., 1961. Bird Song: The Biology of Vocal Communication and Expression in Birds. Cambridge: Cambridge University Press.

Thorpe, W. H., 1963. Antiphonal singing in birds as evidence for avian auditory reaction time. Nature, 197:774-76.

Thorpe, W. H., 1972. Duetting and antiphonal singing in birds. Its extent and significance. Behaviour, Suppl. 18:1-197.

Thorpe, W. H., and North, M. E. W., 1965. Origin and significance of the power of vocal imitation with special reference to the antiphonal singing of birds. Nature, 1208:219-22.

Tinbergen, N., 1935. Field observations of east Greenland Birds. I. The behaviour of the Rednecked Phalarope (Phalaropus lobatus L.) in spring. Ardea, 24:1-42.

Tinbergen, N., 1952. 'Derived' activities; their causation, biological significance, origin and emancipation during evolution. Quarterly Rev. Biol., 27:1-32.

Tinbergen, N., 1953. Social Behaviour in Animals. London: Methuen.

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

Tinbergen, N., 1964. Aggression and fear in the normal sexual behaviour of some animals. In: The Pathology and Treatment of Sexual Deviation, I. Rosen, ed. New York: Oxford University Press.

Tinbergen, N., 1965. Some recent studies of the evolution of sexual behavior. In: Sex and Behavior, F. A. Beach, ed. New York: Wiley.

Tinbergen, N., and Perdeck, A. C., 1950. On the stimulus situation releasing the begging response in the newly hatched Herring Gull (Larus argentatus argentatus Pont). Behaviour, 3:1-39.

Todt, D., 1970. Die antiphonen Paargesänge des osafrikanischen Grassängers Cisticola hunteri prinoides Neumann. J. Ornithol., 111:332-56.

Tschanz, B., 1968. Trottellummen. Zeits. Tierpsychol., Suppl. 4.

Verner, J., and Milligan, M. M., 1971. Responses of male White-crowned Sparrows to playback of recorded songs. Condor, 73:56-64.

Weeden, J. S., and Falls, J. B., 1959. Differential responses of male Ovenbirds to recorded songs of neighboring and more distant individuals. Auk, 76:343-51.

Weidmann, U., and Darley, J., 1971. The role of the female in the social display of mallards. Animal Behaviour, 19:287-98.

Wickler, W., 1972. Aufbau und Paarspezifität des Gesangduettes von Laniarius funebris (Aves, Passeriformes, Laniidae). Zeits. Tierpsychol., 30:464-76.

Wiley, R. H., 1971. Song groups in a singing assembly of Little Hermits. Condor, 73:28-35.