CHAPTER 15

BIRDS

BARBARA I. HOOKER

In the adaptation of birds to an aerial environment, the evolution of feathers and a remarkably efficient respiratory system have incidentally enabled birds to develop complex systems of communication. Communication with one another is achieved by means of displays of characteristic feather patterns (Armstrong, 1947) and by highly specialized sound signals, most of which are vocalized. The vocal apparatus, in many cases, is capable of producing phenomenally varied sound signals, as, for example, in the case of the Indian hill mynah (Gracula religiosa intermedia) (Thorpe, 1959). This chapter will, of necessity, be restricted to a consideration of the vocal communication of birds.

Cherry in 1957 defined communication as “the establishment of a social unit from individuals, by the use of language or signs” (p. 303). This was refined by Smith (1965), who emphasized that a communication event must be characterized by three essential features: a communicator, a signal, and a recipient. The communicator's role is to encode a message which describes the condition of its central nervous system and transmit it as a signal to the recipient, to whom it comes as one of many sensory inputs to the C.N.S. The code must be held in common by the two individuals, and its critical features should be in genetic storage; nongenetic storage—”memory”—probably being important in most higher vertebrates and certainly in the birds. Smith also draws attention to the significance of the “immediate and historical context” of the signal to the recipient.

By utilizing vocal signals, birds have evolved a sophisticated communication system similar, in many ways, to that of man. Sounds are very economical to produce and decay quickly, while at the same time traveling far and fast. They are comparatively independent of physical obstacles, and a great spectrum of frequency and intensity is available. Simple signals are transmitted by simple sounds which may be shared by several species. The alarm calls given by passerine birds when a hawk flies over are almost identical for the species shown in Fig. 1 (Marler, 1959). Similarly, the mobbing calls given by small birds in response to a predator such as an owl are all very alike (Fig. 2). Kear has shown that the distress calls of day-old ducks and geese are strikingly similar in all cases (Fig. 3). However, since the call notes of passerines and nonpasserines are comparatively simple and almost certainly transmitted genetically only, most present-day research into bird song refers to the full song, or primary song. This can include extremely complicated sound signals, e.g., the song of the Gouldian finch illustrated in Fig. 4 (Thorpe, 1961). The full song is usually emitted only under very special circumstances—effecting reproductive isolation and breeding success.

The full song of a passerine bird is given, typically, by the male in full reproductive condition for the purpose of transmitting sexual and motivational information to a prospective mate. It is therefore a speciesisolating mechanism and is frequently valuable for the field identification of new species. The chiffchaff (Phylloscopus collybita) and willow warbler (P. trocholus), for example, were first distinguished as separate species by means of their song (Fig. 5) by Gilbert White in 1879. Recently Kear (personal communication) has analyzed the calls of ninety species of the Anatidae at twenty-four hours age, in looking for taxonomie relationships, and she has found that even in these very juvenile birds, certain sounds concerned with social contact vary between species.

As well as acting as a specific recognition mark, the full song also proclaims the male bird to be in charge of a territory, thus transmitting environmental information. The full song is loud and readily located, being delivered in a temporal sequence, patterning, frequency, rhythm, tonality, accent, and grouping of notes characteristic of the species. Once a female has established herself as the singers sexual partner, it is usually important for her to be capable of distinguishing him from other males of the same species. Each male’s song must therefore contain individual variants which play an important role in maintaining the pair-bond.

Since tape-recording and sound spectrographic analysis became available for the study of animal communication systems, a significant amount of research into the acquisition of vocal behavior has been carried out using many species of birds as experimental subjects. Much of this work has been concentrated on the development of song within a species and the learning processes involved. The laboratory procedure adopted by most workers involves sound isolation of baby birds so that they do not hear the normal parental song pattern until exposed to it artificially. If an isolated bird develops a normal vocal pattern of calls and songs, then its vocal behavior is considered to be wholly inherited. Call notes are almost without exception unaffected by limiting early auditory experience in this way, but in many cases the full song is affected, being dependent on some sort of learning process in early life if normal development is to take place. Unless the young birds are hatched in a soundproof incubator, it is not possible to know how much of their subsequent behavior has been affected by very early exposure to the specific song pattern or indeed to the whole spring chorus. It is extremely difficult to raise young nidicolous birds by hand from the day of hatching, though studies of two examples are available: the whitethroat (Sylvia communis) (Sauer, 1954) and the blackbird (Turdus merula) (Messmer and Messmer, 1956). In each case the birds sang within the normal range for the species on reaching maturity.

Very little is known about the auditory experience of unhatched birds. Nidifugous birds are easily incubated and raised in isolation and at maturity vocalize normally, e.g., Gallus domesticus (Schjelderup-Ebbe, 1923), which suggests that a learning process is not involved in the acquisition of the normal species repertoire—though his results show that there was some delay in the first appearance of some calls. However, it has been shown recently (Vince, 1966a) that in several species of game bird the embryos emit simple “clicks” which serve to stimulate the whole clutch in such a way that hatching is synchronized. Vince also showed (1966b) that isolated eggs which are artificially stimulated by a mechanical clicking or by vibration hatch significantly earlier than do unstimulated eggs. Another study (Gottlieb, 1966) showed that Peking ducklings are more responsive to sounds after hatching if they have previously been exposed to auditory stimulation. Clearly, very early auditory experience plays a most important role in the over-all development in these birds.

Most passerine birds sing seasonally. With increasing day length in the spring and with the development of the testes, the males commence to subsing. Subsong is defined as differing from full song (Thorpe and Pilcher, 1958) in being quieter, of different pattern, with longer song bursts and notes of lower fundamental frequency but with a greater frequency range (Fig. 6). Subsong has perhaps been most fully studied in the chaffinch (Fringilla coelebs) (Thorpe, 1958) in which species the full song “crystallises out from the amorphous subsong” (p. 553) once a male has established himself in a territory. The long, rambling, chirping and rattling subsong gives way to the short and stereotyped full song bursts through a rapid process of tightening and integration, eventually producing a song (Fig. 12a) two to two and one-half seconds in duration and divided into three phrases corresponding to: phrase la, chip-chip-chip-chip; phrase lb, tell-tell-telltell; phrase 2, cherry-erry-erry-erry; phrase 3a-b, tissy-che-wee-000 (Garstang, 1922). The frequency envelope falls from phrase 1 to phrase 3b in a stepwise fashion, and the last phrase is usually a single, but complicated, flourish.

During the first breeding season, a male chaffinch will produce up to five or six themes conforming to the general pattern and similar to the songs of its neighbors in the same area, but with some individual characteristics (Thorpe, 1958a). In subsequent seasons it does not elaborate new themes, tending to breed in the same area and presumably being recognized by female birds, which have also returned to their first breeding grounds after winter flocking.

The development of song in the chaffinch was reported ten years ago (Thorpe, 1958a,b) using sonographic analysis, and although a great deal of investigation into the songs of many species of bird has subsequently been achieved, the chaffinch is still perhaps the most satisfactory species for experimental studies on song learning. The early work of Thorpe will therefore be summarized here.

If a male chaffinch is taken from the nest in the wild four to five days after hatching and is hand-reared in sound and visual isolation from all other birds until the following spring, it will sing a normal subsong, but the full song which develops from this will be very simple (Fig. 7a) and will differ from normal in having only a slight trace of division into three phrases and in lacking the terminal flourish. The song is of about the right length and number of notes, with the normal tonal quality. If this bird is now tutored with a model of a normal chaffinch song, using a continuous-loop tape recorder, it will attempt to match the model and its song will become more normal (Fig. 7b). If hand-reared isolates are placed together in a room in spring, the individuals will counter-sing and stimulate one another to produce a highly complex song divided into phrases and even, in some cases, with a terminal flourish, but comprised of notes of more or less abnormal form. All members of the group will sing a similar song.

If young male chaffinches trapped in their first autumn, having heard their parents sing during spring and summer and perhaps having tried to sing with them, are then isolated and allowed to sing alone in their first spring as adults, the song they produce (Fig. 8a) is much more normal than those of hand-reared isolates. If a group of similar autumn-trapped birds is isolated but allowed to counter-sing in the following spring, then the songs produced are more complex and quite normal (Fig. 8b). From these results it is assumed that three-monthold chaffinches have acquired, or memorized, a normal chaffinch song model which they later attempt to match and reproduce. Hand-reared and isolated birds, with no normal song model, sing a very simple song unless allowed the stimulus of countersinging in groups in which they produce a more complex and phrased song, though still composed of abnormal notes.

The next questions asked in this study were: Why do wild chaffinches, exposed to many kinds of bird song, produce only chaffinch songs, and why does a male chaffinch, which can sing five to six themes at the end of its first breeding season, not add to its repertoire in subsequent seasons?

The first of these problems was tackled by attempting to train male chaffinches to sing alien songs by tutoring them in their first spring with tapes of sounds ranging from a tune on a bird flageolet to a tonally similar bird song. Tutoring with a normal song in spring has some effect on a first-year wild-caught bird; but much more obvious effect on a hand-reared bird (Fig. 7b). Similar sets of birds were then exposed to abnormal chaffinch songs, with the terminal flourish placed in the middle (Fig. 9). Wild-caught birds did not attempt to match this model, but hand-reared birds achieved success. Hand-reared birds would also match “a model of an alien bird such as tree pipit (Anthus t. trivialis). If the model had less resemblance than this to a chaffinch song, they either failed completely to mimic it or scraps from the model were incorporated into the subsong only and later discarded. From these experiments it was concluded that early exposure to a normal chaffinch song renders a male chaffinch capable of selecting only chaffinch song from its auditory environment during its first spring; a chaffinch deprived of this early exposure will select alien sounds from its environment and incorporate them into its song if these sounds are similar to chaffinch sounds in frequency, duration, and rhythm.

Male chaffinches often have more than one song theme, the themes following one another in the same outburst of singing, but not in a random way, so that they cannot be alternative song, or motor, patterns. Hinde (1958) suggested that each song is associated with both inhibitory and facilitative effects of the repetition of that type. When songs are played back to the chaffinch, those song types which it utters most frequently itself are most effective in evoking singing, irrespective of whether they resemble completely normal song or not. Further, the song type played tends to occur more frequently in the bird’s own singing. This supports the view that song learning consists mainly in the acquisition of a selective responsiveness to a particular song type. Hinde suggested that the significant aspects of performance of song by a chaffinch would be the act of uttering, a proprioceptive feedback, and the auditory perception of the consequent sound pattern.

That this is not the whole story applicable to all avian species has recently been shown by Busnel and Brémond (1962), who found that if the component parts of the song of a robin (Erithacus rubecula L.) were scrambled and played back to the robin, the bird would reply as to normal territorial song. The authors then analyzed the song to find which features were effective stimuli in evoking a response. The phrases of the normal song alternate between higher and lower frequencies. By rearranging the scrambled phrases at random, they could evoke a normal response; a single phrase when repeated at the same frequency produced no response; two low-frequency phrases in alternation were “not entirely ignored,” whereas with high- and low-frequency phrases in alternation, such as are found in the normal song, the response was sometimes as good as to a normal song! It must be stressed that this result is as yet unique and applies only to the territorial defense situation in the robin. It is certain that since many species can imitate fine detail, they can pay attention to it, and presumably respond to it, in particular situations. It would be interesting to know if the robin is capable of imitation (Thorpe, personal communication).

Another aspect of the problem is currently being studied by Stevenson (1966, personal communication), who suggests that development of full song from subsong in a chaffinch depends on notes which resemble the adult song having reinforcing properties. If so, then hearing these notes, either from himself or from another source, should be rewarding to a young chaffinch whose song is developing. This has been tested in an operant situation, where a young bird can play a recording of an adult song by hopping onto an “active” perch, but not when on an “inactive” one. A reinforcing effect of hearing the song is said to be shown if the bird perches relatively more frequently on the active than on the inactive perch; that is, when the perch does play the song rather than when it does not. The results of experiments so far completed indicate that adult song does act as reinforcement for autumn-caught, testosterone-injected male birds; whereas a burst of white noise similar in duration to the adult song does not—the response, therefore, being specific to the adult song. However, adult song was not reinforcing for hand-reared birds with no early song experience. Since it has been shown (Thorpe, 1958) that hearing adult song in the early months of life does improve subsequent song development, it is suggested that this difference between autumn-caught and hand-reared birds is due to previous song experience rather than to experience of other kinds.

Recent work on female chaffinches has attempted to throw light on the second question raised above, the apparently limited “sensitive period” in which chaffinches can acquire new songs. Male birds do not elaborate new song themes after their first twelve to thirteen months of life (Thorpe, 1958a). This could be due to a limited sensitive period in the imprinting sense (Thorpe, 1961), dependent on the bird’s first exposure to high levels of testosterone. Chaffinches do not pair for life as far as is known. Females are therefore presumably capable of responding to a different set of male song themes in the next season. A series of experiments was carried out on female birds to test their learning ability with regard to male song (Lade and Thorpe, in press). This involved subjecting female chaffinches to large doses of male sex hormone which induces singings—thus exposing the song pattern which the female had learned.

Old and experienced female birds injected with large doses of testosterone phenylacetate will sing normal songs (Fig. 10). Similarly, autumn-trapped young females will sing normal songs in their first spring, but in both cases there was a range within the group from fairly simple to very complex songs. Nearly all fell within the normal range of male chaffinch songs. Hand-reared and isolated female chaffinches will, on exposure to testosterone, sing a simple isolate song similar to the isolated hand-reared male bird’s song (Fig. 11a). Exposure to a tutor tape of a normal male song (Fig. 12a) before exposure to testosterone, so that the female does not produce song herself and “practice” the appropriate motor pattern while hearing the model, shows that the model has some effect on the song eventually produced after injection, in that the song is somewhat phrased (Fig. lib). If hand-reared, isolated female birds are exposed to a normal model song and testosterone simultaneously, two effects are shown. The female may attempt to match the model (Fig. 12b), or her song may show normal structure but be abnormally elongated to four seconds (Fig. 12c). This elongation of song was also seen when tutoring autumn-caught female birds in their first spring. From these experiments it may be concluded that, for a significant amount of learning to take place, simultaneous exposure to the normal model and to testosterone, with the production of the motor pattern involved in singing, is necessary. During the present experimental season (1966) both naive and previously treated females were brought into song and allowed the opportunity to elaborate and change their song patterns by countersinging with a group. The results of this experiment have yet to be analyzed.

Different species of bird vary very greatly in their song-learning ability (Thorpe, 1961), ranging from the nonlearning situation of birds such as doves, whose species-specific cooing rhythms are unaffected by any interference with early auditory experience (Whitman, 1919; Lade and Thorpe, 1964), to the inventive blackbird, whose incredible song development has been worked out in detail by Hall-Craggs (1961). This worker broke up the song into its smallest component parts and compared these with the bird’s song in the first week of the season. The twenty-six basic phrases used throughout the entire season were included in the song of the first week; but the number of élaborations of new themes and phrases from these basic components during the season was phenomenal. An example is shown in Fig. 13.

Investigations into the role of auditory experience in the development of bird vocalizations was continued more directly by experiments involving the deafening of birds at various stages, effectively producing a more “complete” auditory isolation. This deafening prevents birds from “referring” to inherited or acquired material.

If a bird reared in auditory isolation without deafening develops normal vocalizations for the species, this behavior pattern is presumably genetically controlled, unless the bird listens to its own voice through a sensory feedback mechanism and can therefore be described as “self-taught” (Konishi, 1963). However, to produce the right pattern it must match a genetically acquired model coded somewhere in the CNS—which the chaffinch only has in part; its genetic model, as shown above, is only a simple and unphrased cascade of notes. When a blackbird is isolated, at a few days old, it can eventually produce a normal song (Messmer and Messmer, 1956). It must therefore have a “complete genetic model.” The question next asked was whether or not this species relied on being “self-taught.” Using a technique developed by Schwartzkopff (1949) for deafening passerine birds by extirpation of the cochlea, Messmer and Messmer deafened a baby blackbird and found that it could still produce the normal species song, indicating that sensory feedback from the cochlear nerve was not necessary for normal song development.

A number of species has now been investigated in this way. Konishi (1963) deafened a group of four domestic fowl chicks (Gallus domesticus) within a few hours of hatching and found their vocalizations at maturity to be identical with those of a control group. Furthermore, one cock, which had been deafened when mature, maintained his vocal repertoire without auditory feedback. Konishi pointed out that the involvement of a nonauditory feedback mechanism of the actual motor output has not yet been excluded.

Konishi (1964) next turned to a passerine species, the Oregon junco (Junco oreganus), eight individuals of which had already been raised in acoustic isolation (Marier, et al., 1962). These birds sang several song types, including normal “wild type” songs, but differing from the norm in being “somewhat longer, with fewer longer syllables” (p. 29).

Wild Oregon juncos have simple trill-type songs (Fig. 14), and Konishi found that when they were deafened before the onset of subsong, they produced songs within the normal range of wild birds. However, deaf birds tend to produce songs with fewer syllables, but with mean intervals between syllables significantly longer and with variation of intervals and syllables less distinct (Fig. 15). Thus deafening before the onset of singing in this species caused an instability within the songs and from song to song, presumably due to the lack of complete crystallization of song in the absence of auditory feedback. Once again he found that the deaf birds could maintain their song pattern from year to year.

Konishi (1965) next investigated a species known to have a short early critical period for song learning, long before the bird itself begins to sing—the white-crowned sparrow (Zonotrichia leucophrys nuttalli) (Marler and Tamura, 1964). This species is in many respects similar, with regard to vocal development, to the chaffinch. If the bird has heard the normal song very early in its development, long before it has sung itself, it can sing normally in its first spring. If the bird fails to hear a wild song of the species during its first ten to one hundred days, it fails to produce a normal song (Fig. 16) even if it is exposed to normal birds’ singing during its first spring. Song learning in this species must therefore involve acoustic learning, or the bird must be furnished with an “acquired template.” Regardless of whether the birds have or have not been exposed to normal song during the critical period, if they are deafened before the onset of their own singing, they fail to produce a normal song—indeed they fail to sing as normally as undeafened hand-reared isolates (Fig. 16). Once the birds have sung, they can maintain their song pattern after deafening.

From these results Konishi proposed a neurological model for the acquisition of song in the white-crowned sparrow, as illustrated in Fig. 17, in which the thick arrows symbolize the motor output from the central nervous system to the syrinx and other organs involved in the motor control. Diagram 1 of Fig. 17 shows the condition in the intact bird which develops an abnormal song if it has not been exposed to a normal song model during the critical period. In diagram 2 the model is matched by the vocal output to the template it has so acquired early in life. Experiment 1: Deafening, or removal of auditory feedback, causes development of an extremely abnormal song, regardless of having an acquired template (IB) or not (1A). The acquired template cannot be used without auditory feedback, and the resultant vocal motor output is patterned centrally and/or by nonauditory feedback. Experiment 2: In the case of a bird which has already performed the motor pattern producing normal song, nonauditory feedback alone can maintain the established vocal pattern after the bird has been deafened. Information for auditory feedback must therefore be matched and evaluated in some way by a mechanism responsible for motor output deriving its information from a nonauditory feedback source.

Until the mechanism involving nonauditory feedback loops to the CNS is further investigated, this line of approach to the problems of avian vocal communication can go no further. The aspect of the subject on which it fails to throw light is that of song development in those species in which the function of song is other than territorial. As shown by the work of Hinde (1958), song learning in the chaffinch seems to indicate a selective responsiveness to the species-specific stimulus song which will result in a chaffinch producing a song very similar to that of its neighbor. The chaffinch is restricted very rigidly in what it will copy to chaffinch-like noises and can also, therefore, be described as matching an “acquired or inherited template/’ However, many of the other European species show a much greater power of vocal mimicry than does the chaffinch (Thorpe, 1961). Various members of the Sturnidae, for example the Indian hill mynahs (Thorpe, 1959), can mimic almost any noise presented to them, and thus it is difficult to apply the theories of Hinde and Konishi to these species. Other workers have therefore concentrated on a study of mimetic birds.

As Thorpe and North (1965) have pointed out, some mimics such as the mocking bird (Mimus poly glottis) incorporate an enormous variety of other bird sounds into their repertoire, but all the sounds are of bird-like timbre, timing, and pitch. These mimics are usually territorial species. However, birds such as the parrots and mynahs, which are capable of reproducing almost any noise when trained in captivity, are believed to be comparatively nonterritorial, and in the wild have not been observed to incorporate alien songs into their repertoire. Thorpe and North suggest that the extreme development of imitative ability occurs where the main function of song is to provide for social recognition and cohesion rather than for territorial advertisement. This seems to apply especially in the tropics.

Among European birds, the bullfinch (P.pyrrhula) can be trained in captivity to sing highly complicated songs, but extensive studies of this species (Nicolai, 1959) failed to indicate a function for this ability. Thorpe (1961) points out that the bullfinch does have a family tradition in song, the young birds singing like their parents and grandparents. This may be extremely important in maintaining social units. The dense cover in which bullfinches move would make social recognition by sound signals more important than by visual ones; the white rump, flashing in flight, is very conspicuous, but it does not distinguish between family groups (T. Hooker, personal communication). Recently, Tretzel (1965) has illustrated a prédisposition for specific vocal mimicry in the case of the crested lark (Galerida c.cristata). He recorded an individual which incorporated a totally alien sound—a shepherd’s command whistle (Fig. 18)— into its song. The mimicked whistles were usually sung in series, or separated only by short bursts of other song (Fig. 19). Two other larks in the same area, very likely the progeny of the first lark, also sang the shepherd’s whistle phrases, and, if so, a new song tradition was developing in a small population in a manner which favored the imitation of certain song types and motifs originally alien to the species.

Another territorial species, the shama thrush (Copsichus malabarieus) is a good mimic, and it has been suggested (Gwinner and Kneutgen, 1962) that in this species the ability has been adapted for the maintenance of the pair-bond. The male of a pair sang a certain song motif which was never sung by the female until the male bird was removed. She would then sing the whole song of the male, which would have the immediate, effect of recalling the male bird “as if by name,” since there is nothing more stimulating to a mated male than to hear its own repertoire repeated in its own territory (cf. Hinde, 1958).

An extremely interesting example of vocal pair-bonding has been described for the antiphonally singing shrikes of East Africa (Thorpe, 1963; Thorpe and North, 1965, 1966). These birds can sing in duet with such a rapid reaction time (Fig. 20) that unless an observer is actually standing between the two birds it is impossible to recognize that more than one bird is singing. In a species such as Laniarius aethiopicus, the bou-bou shrike, a pair of birds can elaborate a whole repertoire of duet patterns (Fig. 21) by which they can recognize one another in dense undergrowth and be distinguished frorri other pairs in the neighborhood. In this species either sex can start or finish and either bird can sing the whole pattern alone in the absence of the partner. When the partner returns, the pair can either sing in perfect unison or sing antiphonally again. Trio singing has also been observed and at present is not understood at all, but it has been suggested that the third bird may be the progeny of the pair.

If it can be shown that the extreme imitative ability of certain species such as parrots and mynahs has a function in establishing and maintaining pair-bonds, then this imitative ability, apparently unused in the wild, may receive a plausible explanation (Thorpe and North, 1966). It would also obviate the necessity for assuming that such birds inherit a tendency to match an acquired template in order to maintain the species-specific vocal pattern.

SUMMARY

The nature of vocal communication in birds, as revealed by current research, has been reviewed with special reference to the development and inheritance of song patterns. The functions of song have been considered, and possible explanations for some special cases such as duet singing and mimicry have been offered.

Investigations into the development of vocal communication in various species of bird have shown that, in common with other behavior patterns, vocal behavior varies between species with different ecological requirements. The strongly territorial and competitive birds of the more temperate climates tend to be dependent on very early auditory experience for fixing their subsequent vocal behavior, while the less territorial and tropical birds tend to retain their ability to learn new social vocal patterns for unlimited periods. However, research into bird song has been limited so far to only a small number of species exhibiting such a wide variety of patterns of vocal development that any attempt to summarize for the whole class Aves would be very premature.

REFERENCES

Armstrong, E. A., 1947. Bird Display and Behaviour. London.

Busnel, R.-G., and J.-C. Brémond, 1962. Récherche du support de l’information dans le signal acoustique de défence territoriale au Rouge-gorge (Erithracus rubecula L.). C. R. Acad. Sei., 254:2236-2238.

Cherry, C., 1957. On Human Communication. New York, Wiley.

Garstang, W., 1922. Songs of the Birds. London.

Gottlieb, G., 1966. Species identification by avian neonates; contributory effect of perinatal auditory stimulation. Anim. Behav., in press.

Gwinner, E. von, and J. Kneutgen, 1962. Uber die biologische Bedeutung der “zweckdienlichen” Anwendung erlernter Laute bei Vögeln. Z. Tierpsychol., 19:692-696.

Hall-Craggs, J., 1961. The development of song in the Blackbird (Turdus merula). Ibis, 104:227-300.

Hinde, R. A., 1958. Alternative motor patterns in chaffinch song. Anim. Behav., 6:211-218.

Kear, J. 1966. Personal communication.

Konishi, M., 1963. The role of auditory feedback in the vocal behaviour of the Domestic Fowl. Z. Tierpsychol., 20:349-367.

Konishi, M., 1964. Effects of deafening on song development in two species of Juncos. Condor, 66, No. 2:85-102.

Konishi, M., 1965. The role of auditory feedback in the control of vocalization in the Whitecrowned Sparrow. Z. Tierpsychol., 22:770-784.

Lade, B. I., and W. H. Thorpe, 1964. Dove songs as innately coded patterns of specific behaviour. Nature, 202:366-368.

Lade, B. I., and W. H. Thorpe, The song-learning ability of female chaffinches on exposure to testosterone (in press).

Marier, P., 1959. Developments in the study of animal communication. Ch. 4 in Darwins Biological Work: Some Aspects Reconsidered. P. R. Bell (Ed.), Cambridge.

Marler, P., 1960. Bird Songs and Mate Selection. Animal Sounds and Communication. American Inst, of Biological Sciences, Publ. No. 7.

Marler, P., and M. Tamura, 1964. Culturally transmitted patterns of vocal behaviour in a sparrow. Science, 146:1483-1486.

Marler, P., M. Kreith, and M. Tamura, 1962. Song development in hand-raised Oregon Juncos. Auk, 79:12-30.

Messmer, E., and I. Messmer, 1956. Die Entwicklung der Lautäusserungen und einiger Verhaltensweisen der Amsel. Z. Tierpsychol., 13:341-441.

Nicolai, J., 1959. Familien Tradition in der Gesangsentwicklung des Gimpels. J. Orn., 100:39-46.

Sauer, F., 1954. Die Entwicklung der Lautäusserungen vom Ei ab Schalldicht gehalteren Dorngrasmücken (Sylvia c. communis Latham) in Vergleich mit später isolierten und mit wildleben Artgenossen. Z. Tierpsychol., 11:10-93.

Schelderup-Ebbe, 1923. Weitere Beiträge zur Sozialeund individualpsychologie des Haushuhns. Z. Psychol., 92:60-87.

Schwartzkopff, J., 1949. Uber Sitz und Leistung von Gehör und Vibrationssinn bei Vögeln. Z. vgl. Physiol., 31:527-608.

Smith, J. W., 1965. Message meaning and context in ethology. Amer. Naturalist XCIX, No. 908.

Thorpe, W. H., 1958a. The learning of song patterns by birds, with especial reference to the song of the Chaffinch, Fringilla coelebs. Ibis, 100:535-570.

Thorpe, W. H., 1958b. Further studies in the process of song learning in the Chaffinch (Fringilla coelebs gengleri). Nature, 182:554-557.

Thorpe, W. H., 1959. Talking birds and the mode of action of the vocal apparatus of birds. Proc. Zool. Soc. Lond., 132:441-455.

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

Thorpe, W. H., 1963. Antiphonal singing in birds as evidence for avian auditory reaction time. Nature, 197:774-776.

Thorpe, W. H., and M. E. W. North, 1965. Origin and significance of the power of vocal imitation: with special reference to the antiphonal singing of birds. Nature, 208, 219-222.

Thorpe, W. H., and M. E. W. North, 1966. Vocal imitation in the tropical Boubou Shrike Laniarius aethiopicus major as a means of establishing and maintaining social bonds. Ibis, 108:432-435.

Thorpe, W. H., and P. M. Pilcher, 1958. The nature and characteristics of subsong. Brit. Birds. 51:509-514.

Tretzel, E., 1965. Imitation und Variation von Schäfenpfiffen durch Haubenlerchen (Galerida c. cristata L.). Ein Beispiel für spezielle Spottmotiv-Prädisposition. Z. Tierpsychol., 22:785-810.

Vince, M. A., 1966a. Potential stimulation produced by avian embryos. Anim. Behav., 14:34-40.

Vince, M. A., 1966b. Artificial acceleration of hatching in quail embryos. Anim. Behav., 14 (in press).

White, G., 1879. The Natural History of Selbourne. London, Fredk. Warne & Co.

Whitman, C. O., 1919. The Behaviour of Pigeons. Posthumous Works of C. O. Whitman, 3:1-161. In: A. H. Carr (Ed.), Publ. Carnegie Inst.