Song learning in brood-parasitic indigobirds Vidua chalybeata: song mimicry of the host species

Transcription

1 Anim. Behav., 1998, 55, Song learning in brood-parasitic indigobirds Vidua chalybeata: song mimicry of the host species ROBERT B. PAYNE, LAURA L. PAYNE & JEAN L. WOODS Museum of Zoology and Department of Biology, University of Michigan (Received 18 March 1997; initial acceptance 22 June 1997; final acceptance 5 September 1997; MS. number: A7694R) Abstract. Brood-parasitic village indigobirds, Vidua chalybeata, were bred in captivity and foster-reared by their normal host, red-billed firefinch, Lagonosticta senegala, or by an experimental foster species, Bengalese finch, Lonchura striata. Male indigobirds reared by Bengalese finches developed the songs of Bengalese finches, and males reared by firefinches developed songs of firefinches. Males copied their foster father only when they had lived with him long after independence (45 days post-fledging), while males separated normally at independence (22 24 days post-fledging) copied songs of other individuals and not songs of their foster father. Males reared by Bengalese finches showed no preference to learn firefinch song over songs of the experimental foster species or other control finch species even when they had lived with firefinches as companions from the time of fledging to independence. Males copied several song themes, acquired the same number of mimicry songs, and acquired their songs at the same age, whether reared by Bengalese finches or by firefinches. When they lived with other indigobirds, the male indigobirds copied mimicry songs of male indigobirds that mimicked the same foster species. We predicted mimicry-song specificity and repertoire size in experimental indigobirds from a hypothesis of an early developmental period when young indigobirds focus their attention on their foster parents, and a later period when they direct their attention to other birds with similar songs. The predictions, based on field observations of wild birds, were that (1) males reared by a novel foster species other than the normal host would learn the song of that foster species, and (2) males that left their foster parents at the normal time of independence would copy the songs of other individuals, including other adult indigobirds that mimicked the same foster species. Begging calls of young indigobirds did not mimic the calls of young firefinches. Indigobirds reared alone, or with young of the normal host or of the experimental foster species, all developed begging calls in adult song that resembled their own begging as nestlings and fledglings, and only males that heard other adult indigobirds with firefinch-mimicry begging developed firefinch begging in their song. The incorporation of the innate begging calls as well as the learned begging calls into adult song, and the modification of the song themes of their individual song models, suggest that song development involves processes in addition to copying the songs of their own foster species and of older adult male indigobirds with songs like their own foster parents The Association for the Study of Animal Behaviour The African village indigobird, Vidua chalybeata, is a brood-parasitic songbird. Adult male indigobirds mimic the songs and begging calls of their host species, the red-billed firefinch, Lagonosticta senegala, and female indigobirds visit and mate with males that mimic the song of their host species (Payne 1973, 1990). Although the breeding plumage of male Vidua indigobirds differs only slightly in colour, the species differ conspicuously Correspondence: R. B. Payne, Museum of Zoology, University of Michigan, Ann Arbor, MI , U.S.A. ( rbpayne@umich.edu). in the songs they mimic. Each indigobird species parasitizes a different host species, and each male s mimicry songs indicate his own species identity as well as his foster species (Payne 1973, 1983, 1996; Payne & Payne 1994, 1995). The brood-parasitic nestlings also mimic the host nestlings in mouth pattern and colour (i.e. each brood parasite host species pair displays a different mouth pattern), which may allow the nestlings to be accepted in the host brood and receive the care of a discriminating foster parent (Nicolai 1964; Payne 1996). In contrast, while the begging calls /98/ $25.00/0/ar The Association for the Study of Animal Behaviour 1537

2 1538 Animal Behaviour, 55, 6 of adult indigobirds appear to resemble those of their host species (i.e. begging calls in adult mimicry song; Nicolai 1964; Payne 1973), the begging calls of dependent indigobirds recently recorded in the field differ from those of their host species (Payne 1996). Insofar as a male indigobird mimics the species that reared him and this song mimicry is learned, his songs inform the female about his own early experience and function as a sexually selected signal to the female of the genetic quality of the male. If she mates with a male whose mimicry songs are like those of her own foster parents, then the pair will produce offspring having a mouth pattern that matches their common host species. Learning the songs of the host species may also be important in colonization of a new host and in speciation. If males learn the song of their own foster species, then song learning may be part of the process through which indigobirds successfully colonize a new species of host, as follows. Colonization may begin when an errant female lays in the nest of a novel foster species and her young survive, the males reared by the new foster copy the songs of the novel host, and females are attracted to males with songs like their own foster parents and lay in the nests of their own foster species. The process of host switching and mimicry of the new host s songs may lead to indigobird speciation when genetic differences accumulate through selection between the descendants of parasites on old and new hosts (Payne 1973; Payne & Payne 1994). Molecular and morphological evidence suggest that indigobirds colonized their host species well after the host species had diverged, rather than having co-speciated along with them (Payne et al. 1993; Klein & Payne 1998; Payne 1998). Although indigobirds may learn mimicry songs and calls from their foster parents, the evidence is indirect, based on the widespread occurrence of song learning in songbirds (Kroodsma & Baylis 1982) including the estrildid finches, the group to which the Vidua finches are perhaps most closely related (e.g. Goodwin 1960; Güttinger 1972, 1973; Güttinger & Achermann 1972; Dietrich 1980; Baptista 1996). In one of the best-studied estrildids, the zebra finch, Taeniopygia guttata, males learn from their father or from other males depending on social conditions (e.g. Immelmann 1969; Clayton 1987a; Eales 1989; Williams 1990; Slater & Mann 1991; ten Cate et al. 1993; Zann 1996). Straw-tailed whydah, V. fischeri, another brood-parasitic finch that mimics the songs of its host species in the field (purple grenadier, Granatina ianthinogaster), copied the songs of Bengalese finch, Lonchura striata, when reared by it in experimental conditions; whydahs copied both the foster father and other male Bengalese finches (Nicolai 1973). Field observations suggest that male indigobirds learn their mimicry songs, but they do not learn only from their foster parents. Each adult male indigobird has a repertoire of three or four mimicry songs, rather than one or two as expected if he simply had the songs of his foster parents (a male firefinch usually has one song theme, a female has one song theme but a female rarely sings, Payne 1990). Neighbouring male indigobirds match the details of all three or four mimicry songs with each other. Because firefinches are individualistic in song and the indigobirds are not (Payne 1979, 1985, 1990), it is likely that male indigobirds copy their mimicry songs from another male indigobird rather than from their own firefinch foster parents. Their population ecology also makes it unlikely that all male indigobirds in a neighbourhood would have the same foster parents: indigobirds parasitize about 30% of the firefinch nests (Morel 1973; Payne 1977) and indigobirds that match each other s mimicry songs are only about 10% as numerous as firefinches within the same area (Payne & Payne 1977). In addition, a few male village indigobirds (1%) do not mimic the songs of red-billed firefinch but mimic the songs of another species of finch; these may be males that were reared by the other species (Payne 1973; Payne et al. 1992, 1993). Also, juvenile indigobirds reared and captured in the field in Zambia and then kept with firefinches from another part of Africa copied the individualistic songs of these novel firefinches. When the indigobirds sang as adults, their novel mimicry songs were copied by other yearling indigobirds from Zambia (Payne 1985). A hypothesis to account for the learning of mimicry songs and the matching of mimicry songs among individual indigobirds is that song learning is a two-stage process. Each bird in effect is guided through its rearing by the foster parents to the general features of their song, then learns the details of songs later from other firefinches or from other adult indigobirds that mimic the same species of firefinch (Payne 1985, 1990). The strong

3 Payne et al.: Song mimicry in a brood parasite 1539 predictions of this hypothesis are that (1) males reared by another foster species will learn the song of that novel foster species and (2) males that disperse from the foster brood at the normal time of independence will copy the songs of other birds that sing their foster species songs, including adult indigobirds that mimic these songs. An alternative hypothesis is that indigobirds have a restrictive predisposition to learn the songs of their normal host species, the firefinch. The strong predictions of this hypothesis are that (1) males learn firefinch song or conspecific indigobird song even after they are reared by another foster species, (2) males learn firefinch songs at an earlier age than they learn other species songs and (3) males learn more songs of their normal host species than of the alternate foster species. Finally, some features of mimicry song such as the begging calls may be innate. This hypothesis predicts that song develops without the experience of hearing other birds. We tested these predictions in experiments with indigobirds that were reared by a species of finch other than their normal firefinch host. METHODS Wild-caught village indigobirds were kept in an aviary with their host species, the red-billed firefinch. The indigobirds were from an unknown source in Africa, probably Senegal, because the males were identified when compared with museum specimens as the West African subspecies V. c. chalybeata. This form occurs in Senegal, and Senegal was known to export small birds in avicultural trade. Males mimicked the songs of red-billed firefinches, and the captive male indigobirds had a song repertoire typical of indigobirds in the field. Females were identified by size (wing <65 mm) and orange legs (Payne 1973, 1979, 1985). The breeding aviary was located outdoors on a walled rooftop and was built of 6-mm wire hardware mesh. We divided the aviary into six units, each with a closed, roofed indoor area and a mesh outdoor area. We provided openings between areas and units to allow birds to use several units as a common breeding chamber (Fig. 1). The arrangement allowed several pairs of hosts to nest without interfering with each other, so that several active nests were available at any time to each 1 m Outer Inner Figure 1. Roof aviary, units 1 6, each with an outer and an inner area. In this configuration, the units are connected by openings so that 1, 2 and 3 form one breeding chamber, and 4, 5 and 6 form a second breeding chamber. Only one pair of indigobirds was housed in each breeding chamber; several pairs each of firefinches, Bengalese finches and other finches shared a breeding chamber. pair of indigobirds. Inner areas had heat lamps available 24 h a day, food and water, and a heater to maintain a minimum temperature of 14 C. Outer areas were planted with grass, herbs, beans and courgettes, brush and white cedar trees to provide cover and live insects to supplement the standard diet of millet seeds, fresh egg mix (chopped boiled egg, calcium powder, Rowdybush Breeder Crumble, Life s Finest 90% protein powder), fresh fruit (sliced apples and oranges) and vegetables (frozen peas and corn, fresh broccoli, courgettes), sprouted seeds, romaine lettuce sprinkled with Nekton-E or -S vitamins, small mealworms, ant pupae, and crushed eggshell, grit, oyster shell and charcoal for minerals. Wicker nestboxes were provided for nests in the inner and outer areas; birds also built nests in the brush and trees. In each breeding chamber we kept a pair of indigobirds, two to six pairs of red-billed firefinches, and two to six pairs of Bengalese finches. Other estrildid finches were also present and nested in some years (purple grenadier; blue-capped cordon blue, Uraeginthus cyanocephalus; melba

4 1540 Animal Behaviour, 55, 6 finch, Pytilia melba; red-winged pytilia, P. phoenicoptera; yellow-winged pytilia, P. hypogrammica; red-headed finch, Amadina erythrocephala; blackfaced waxbill, Estrilda erythronotos; black-rumped waxbill, E. troglodytes; Dybowski s twinspot, Euschistospiza dybowskii; Peters twinspot, Hypargos niveoguttatus; goldbreast, Amandava subflava; and quail-finch, Ortygospiza atricollis). Several of these finches are host species of other kinds of Vidua including other indigobirds (Nicolai 1964, 1969; Payne et al. 1992; Payne & Payne 1994, 1995). The diversity of finch species reflected the field conditions where indigobirds feed in mixedspecies flocks after the young fledge and are independent of their foster parents. Female indigobirds laid eggs in firefinch nests in the roof aviary. We either left the eggs in these nests to be reared by the firefinches, or we moved the eggs to nests of Bengalese finches in the aviary or in cages in a bird room inside a building. The cages were cm or twice this size, and the room had several Bengalese finches and sometimes other finch species. Bengalese finches are a domesticated form of Asian sharp-tailed munia, Lonchura striata, and have been artificially selected by aviculturists to breed in small cages and rear the young of other finch species that do not regularly rear their own young in captivity (Immelmann et al. 1977; Baptista 1978; Goodwin 1982). We varied the early experience among birds to test the development of song at different ages and social conditions. Nestlings were reared either alone or together with young of the foster or other estrildid species. Both members of a foster pair of firefinches or Bengalese finches care for the nestlings. Three nestlings were reared by a female female pair of Bengalese finches to test song development when the foster parents do not sing, and other nestlings were reared by groups with two or more male fosterers: Bengalese finches sometimes nested in trios in a cage or in an aviary where as many as four adults fed a fledged young indigobird. We either kept an indigobird with its foster parents or separated it shortly after independence, as early as 18 days after fledging. Because the age when a young bird normally leaves its foster parents may affect whether it copies the songs of the foster parents, we predicted that a bird would retain the foster parents songs only if it remained with them longer than the normal period after independence. In the field, the young firefinches and indigobirds remain with the firefinch parents for less than a month after they fledge (Morel 1973; Payne & Payne 1977), and in our aviaries the fledglings beg from their foster parents up to 24 days after fledging and often remain close to them for another week. When we separated the juveniles from their foster parents, we caged them with other Bengalese finches, with both Bengalese finches and firefinches, or with these and other species of estrildid finches. Because the indigobirds mimic only the songs of their host species even when they live as juveniles in mixed-species flocks in the field, we predicted that they would not copy the songs of these birds other than their foster species. We used the variation in song among the individual firefinches and Bengalese finches to test whether a bird copied the song of the foster bird that reared him, other birds of the foster species in the bird room, other estrildid finches, or other singing indigobirds. For each indigobird we noted which birds were together in the same cage ( companions ), were visible to the male in an adjacent cage ( neighbours ), or were present in the bird room and could be heard but not seen by the indigobirds ( others ) (Table I). A young male indigobird heard as many as 18 male Bengalese finches and 17 firefinches by its second year, regardless of whether the male was caged with these singing birds. All male indigobirds heard at least four male Bengalese finches in the outdoor aviary or indoor bird room in addition to their own foster parents. In addition to Bengalese finches and firefinches, each young male indigobird also heard other finch species in the aviary or the bird room, as fledglings or juveniles in their first year or in later years. These other species provided a control for the specificity of song learning. If indigobirds have a preference for learning the songs of firefinches, then the acoustic or social availability of firefinches might overcome the early experience of an indigobird reared by an abnormal foster species. We reasoned that if there is a predisposition to learn the normal host species songs, then indigobirds should acquire songs of the firefinch, either instead of or in addition to the songs of their experimental Bengalese finch foster species, but should not learn other control species songs. We also tested whether males copied the songs of other male indigobirds that were reared by the

5 Payne et al.: Song mimicry in a brood parasite 1541 same foster species and that mimicked the foster species songs. We kept the Bengalese finch-reared young indigobirds where they could not see or hear an adult indigobird, except that birds reared on the roof could hear their own father. In addition, three birds reared in a cage in 1993 were kept in a room where they could hear a male indigobird that had been reared by a Bengalese finch in We separated from each other the birds reared by Bengalese finches and the birds reared by firefinches late in the year before they sang, and we kept these groups in different buildings. Except for their time as nestlings and fledglings for three males (o/y, y/b, pw/o) in the breeding aviary, we made certain that a young bird could not hear any indigobirds that had been reared by the other foster species. Songs of the two foster species are distinct. In Bengalese finches, males sing and have a simple contact call, but females rarely sing and have a complex contact call (Immelmann et al. 1977; Yoneda & Okanoya 1991). Song phrases consist of a repeated sequence of rattling elements or notes, and a song consists of a complex phrase of several types of elements, usually with three such phrases repeated in series in a song. The elements are short (<0.05 s), with a wide-band frequency and are not whistled. Songs are individualistic and each male has a distinct song (Dietrich 1980; Goodwin 1982; Clayton 1987b). Female contact calls are not as individualistic as male songs, and we were unable to trace transmission from individual females to the fostered young, although the indigobirds often included these contact calls in their song. In firefinches, males are individualistic in song. Firefinch song has an abrupt introductory chick note that rises rapidly in pitch, and a series of three to six whistled notes that extend typically s, often rise in pitch, and sound like pea ; the whistled notes repeat or intergrade in a series and the song sequence is chick, pea, pea... Females sing infrequently; males and females have similar, non-individualistic contact calls (Immelmann et al. 1965; Immelmann 1968; Payne 1985, 1990). We recorded songs on a Sony TC-D5M cassette-recorder and Sennheiser ME40 microphone and examined the recorded songs on a VGA screen with a Kay Elemetrics DSP-5500 Digital Sound Processor with settings of 0 16 khz and 256 points/s (234 Hz) with a filter to remove noise below 1.5 khz. We printed the songs with a Mitsubishi P61U printer. We recorded each bird when it was a year old; we also recorded most birds in later years. We compared the spectrograms of their songs with those of our other captive birds. The song reference catalogue included songs of 53 male Bengalese finches, 48 firefinches, and the other estrildid finches and the wild-caught indigobirds. We also compared each song with all songs of the other captive-reared indigobirds. We traced the origin of mimicry songs in each subject bird by their resemblance to the songs of the foster parents, to songs of companion and neighbouring birds, to songs of other birds in the catalogue, and to songs of the bird s own father and other captive adult indigobirds. Two persons independently scored the closest match of each indigobird song to the songs in this catalogue, using the spectrogram form and sequence of song elements. The match-to-sample method assumes that a song is learned intact from an individual song model (Payne & Payne 1993, 1997), and this is indicated in natural populations in this species by the precise matching of the details of songs among neighbours, rather than matching only parts of the songs of different neighbours (Payne 1973, 1979, 1985, 1990). The method allows for a bird to modify its song when it does not copy a song precisely. In cases where a bird had more than one mimicry song, he could match more than one source. Where a bird had a song that equivocally matched more than one model, because the songs of the problematic models were similar, we compared time and place to determine whether he could have heard the song before his own songs were recorded, and we allowed a match only when the model source was present before the young indigobird was recorded. A few songs could not be attributed to one particular model source by these criteria. RESULTS Indigobirds hatched in days and remained in the nest another days before they fledged. They began to feed themselves by pecking seeds from the ground within 6 days after they fledged, they hulled their own seed by 10 days after fledging, and they were independent by days after fledging. Some birds continued to beg

6 1542 Animal Behaviour, 55, 6 Table I. Song learning in captive-reared village indigobirds, Vidua chalybeata* Male Vidua Year born Age (days after fledging) when separated from fosterer Vidua father Fosterers (male), BorF Companions and neighbours N Bengalese songs copied N Firefinch songs copied Song sources copied, known [and uncertain or modified] r/w V1 B1 None; P1, G1; W1 2 0 B1, [B6, B14, B15 or B39], G1, W1 w/g V1 B2, B3 None; G B2, B3, B6, [1 3 other B songs] B/y V2 B (none)** [B21, B22, B23, F15, F1] 3 0 B23, B35, [B36, B37 or B38] y/v V1 B2, B3 None 3 0 B2, B3, [B6 or B15] B/o V1 B1, B7 F3, T1; B4, B5 2 0 B1, B7, T1 g/g V1 B4, B5, B13 U1, G2, Vidua y/v 3 0 B4, 2BsongsofVidua y/v o/y V1 B8 (B9, F2, F1, F3, F4, Q1, V1); B6, B18, B19 y/b V1 B10 (F11, F12, F13, F14, Q1, V1); B1, B11, [B4, B13, B20, B5, B6] v/v V3 B7 [B30, B27]; B26, Vidua r/b-b, [F15, F16] 4 3 B6, B12, B33, B34, F2, [F20, F21] B1, B11, B4, [1 3 B songs of B13, Vidua B/o or g/g], F3, [F22 or F23] 2 0 B30, 1BsongofVidua r/b-b r/b-b V2 B7 [B24, B25, B27, F15, F16]; B26; B B24, [2 other B songs] r/bw V6 B40 [B6, B30], Vidua o/so, wb/v, g/pp 3 0 [3 B songs] o/so V6 B40 [B6, B30], Vidua r/bw, wb/v, g/pp B songs of Vidua r/bw g/pp V6 B (none)** Vidua r/bw, o/so, wb/v B songs of Vidua r/bw wb/v V6 B (none)** Vidua r/bw, o/so, g/pp B songs of Vidua r/bw

7 Payne et al.: Song mimicry in a brood parasite 1543 Table I. Continued. Male Vidua Year born Age (days after fledging) when separated from fosterer Vidua father Fosterers (male), BorF Companions and neighbours N Bengalese songs copied N Firefinch songs copied Song sources copied, known [and uncertain or modified] r/b V1 F2 (B15, B16, B17, B32, F5, F6, F7, F8, I1, P2, Q1, U2); F18, F19, [V4, V1, V2, V5] v/b V1 F1 (B14, B15, B31, F9, I1, P2); F18, F19, [B3, V1, V2, V5] pw/o V6 F14 (F3, F24, F25, F26, F27, B3, B4, B18, B40, B41, B42, O1, O2, U3, V7, W1), [V6, 12 additional F, 4O, 12 additional B, T1, W1]; F F2, F songs of V5, [V4,V1orV2] 0 4 [F1], F18, F songs of V1, [V5] 1 5 F27, F song of V6, 3Fsongs of F28, B song of F28 *Colour-band combinations (e.g. r/w) identify the subject indigobirds. Each letter-number combination is a possible or actual tutor male (a foster parent, a companion, a neighbour, or an other bird, or more than one of these; e.g. B1 was a foster parent of r/w and a companion of y/b). B: Bengalese finch; F: red-billed firefinch; G: violet-eared waxbill; I: purple grenadier; O: goldbreast; P: red-winged pytilia; Q: quail-finch; T: Dybowski s twinspot, U: blue-capped cordon bleu (U1 was a female that sang); V: village indigobird (wild-caught adult); W: black-rumped waxbill. B1 B10, B40, foster father of one or more indigobirds; B11 B32, B41 B43, companion or neighbour; B33 B39, other bird whose song was copied. F1, F2, F14, foster father; F3 F19, F24 F28, companion or neighbour; F20 F23, other bird whose song was copied. Boldface indicates fosterers, companions or neighbours whose songs were copied. Songs copied from these three sources are also indicated in the song sources column. Five birds were reared in the roof aviary (o/y, y/b, v/b, r/b, pw/o) with companions in breeding chamber; other birds were reared by Bengalese finches in the bird room in small cages. Companions and neighbours. Symbols: (parentheses), male birds in addition to parent and foster parents in the natal aviary; [brackets], neighbours visible in an adjacent cage or aviary; no parentheses or brackets, companions in same postnatal cage; semicolon, separates an earlier social group from a later social group (companions or neighbours) (e.g. r/w had no companions other than his foster parents; W1 was added 2 years after P1 and G1 had been housed with r/w). Songs of birds other than those in the fosterers and companions and neighbours columns (non-boldface) were other birds that could be heard but not seen. Songs of birds in [brackets] were uncertain in source (songs were similar to each bird indicated but were somewhat modified, so it was uncertain which birds were copied), or were much modified from any source (for w/g, r/bw, r/b-b and v/b). **Fosterers were female Bengalese finches, there was no male fosterer. r/bw was 5 days older than wb/v and 37 days older than g/pp, had more early experience with male Bengalese finches than these two, and was heard singing subsong 60 days before o/so, and thus was probably the tutor for the other three; the four Vidua were companions after fledging and developed a shared set of identical songs.

8 1544 Animal Behaviour, 55, 6 Indigobird r/w (a) Bengalese finch B1 (c) (b) (d) B39 0 khz s Figure 2. Song learning of indigobird r/w, which was reared by Bengalese finches. (a, b) Mimicry songs of r/w, (c) Bengalese finch fosterer B1, (d) Bengalese finch later companion B39. for 24 days after fledging. They remained with their fosterers until they fed independently and no longer begged from them, and they associated with their foster parents for another 10 days in the aviary. The time of development of captive-reared birds was consistent with that of wild birds in the field (Payne & Payne 1977; Payne 1985), where the first young indigobirds appeared at the call-sites of the singing adults about 8 weeks after the first copulations of the season. At this time, the young from the first matings would have separated from the foster parents and become independent about 4 weeks after they fledged. Specificity of Song Mimicry Fourteen male indigobirds were reared by foster Bengalese finches and their songs were recorded in a later year. All copied the songs of the experimental foster species that reared them, while two also copied songs of their normal host species, the firefinch (Table I). The three indigobirds reared by their normal host firefinches copied firefinch songs and one also copied a Bengalese finch song. Time of Independence and Early Song Learning The five indigobirds that remained with their Bengalese finch foster father for 46 days or longer after fledging sang a song that matched the song of the foster father (Fig. 2, Table I). The other bird (B/y) that remained with his fosterers (both fosterers were females and did not sing) developed the song of a Bengalese finch that was visible in the next cage both during B/y s period of dependence and for 2 months after he fledged. The six birds that we separated early (days after fledging) from their foster father did not copy him, but copied other Bengalese finches. Four birds were fostered by two or more male Bengalese finches in addition to one or more females. The three birds (y/v, w/g, B/o) that remained with their Bengalese finch fosterers for at least 50 days after fledging copied songs of both foster fathers (Table I). The three birds reared by firefinches copied firefinch songs both from the foster father and from the adult indigobirds they heard in later months when caged in the same room. The foster firefinches and the adult indigobirds had very similar songs, and the details of song (shape of the whistled element, sequence of phrases within a song) suggest that the young indigobirds copied the older indigobirds (Fig. 3). The third bird reared by a firefinch did not copy the foster father but copied the variable songs of its later social companion firefinch. Number of Song Models and Song Learning Outside the Foster Group All Bengalese finch-fostered indigobirds sang the songs of at least two Bengalese finches and most sang songs of three or four. Only eight of the

9 Payne et al.: Song mimicry in a brood parasite 1545 Indigobird v/b (a) Firefinch F1 (d) (b) Indigobird adult V1 (e) (c) Indigobird adult V2 (f) khz s Figure 3. Song learning of indigobird v/b, which was reared by firefinches and later heard the adult wild-caught indigobirds V1 and V2, which mimicked the same species of firefinch. (a, b, c) Songs of v/b, (d) song of foster firefinch F1, (e) song of indigobird V1, (f) song of indigobird V2. Songs of foster F1 and indigobird V2 were similar. ca 34 Bengalese finch song themes of the indigobirds were traced to their foster fathers. The remaining songs were learned from sources outside the foster group (Table I). About half (five of seven) of these birds copied the songs of their companion or neighbour Bengalese (the others had no companion or neighbour Bengalese finches), and some (eight of 12) copied other male Bengalese finches that they heard but could not have seen. An average of 16 male Bengalese finches (companions, neighbours and other males out of the indigobird s sight) were in the room and all were potential sources of songs for each male. Counting all models separately for each indigobird, no more than 16 (7%) of the 224 potential model Bengalese finches that were not companions or neighbours were copied, and at least seven (27%) of the 26 companion and neighbour Bengalese finches were copied (in addition to foster Bengalese finches that remained with their fostered indigobird), so the birds tended to copy their companions and neighbours. A few mimicry songs did not closely match the catalogue sample and may have been modified from an unidentified Bengalese finch source. The three firefinch-fostered indigobirds had several mimicry song themes, although the foster fathers each had only a single song theme. Two birds also copied the song of another firefinch with which they lived after independence (Table I). Two birds heard their own father when they were fledglings and later when they were neighbours, and both developed songs like their father (Fig. 3). Improvisation and Modification The differences between some of the indigobirds songs and songs of birds they heard may be due to modification or improvisation on these song themes. For example, only one Bengalese finch theme of male r/w was nearly identical to the song of a bird that he could have heard (Fig. 2, Table I). The other Bengalese finch-like song was about equally similar to that of each of four other birds, and it was not clear which bird was copied, although the song was clearly like a Bengalese finch song. Perhaps r/w improvised upon or modified a song theme that he heard earlier (Fig. 2d, B39), but we

10 1546 Animal Behaviour, 55, 6 Indigobird g/g (a) Bengalese finch B4 (d) (b) Indigobird adult y/v (e) (c) Indigobird adult y/v (f) khz s Figure 4. Song learning of indigobird g/g, which was reared by a trio of Bengalese finches, and songs of his two song models, a foster Bengalese finch and an adult indigobird. (a, b, c) Songs of g/g, (d) song of fosterer B4, (e, f) Bengalese finch songs as mimicked by indigobird companion y/v. could find no single Bengalese finch with a most-similar song. Modification and improvisation may also have been involved in the development of song themes by four male indigobirds (r/bw, o/so, wb/v, g/pp) that were caged together from independence and whose song themes sounded like Bengalese finches, but matched each other and did not match any Bengalese finch model. Other Indigobirds as Song Models Indigobirds copied the mimicry songs of older indigobirds in certain conditions. Male g/g was reared by Bengalese finches, then moved to a cage with an older indigobird (y/v) and other finches. Male g/g copied two Bengalese finch songs from male y/v, which had been reared by Bengalese finches (Table I, Fig. 4). Another male (v/v) was caged with male r/b-b, which fledged 4 months earlier, and v/v copied a Bengalese finch song from r/b-b. Two other males that heard an older indigobird did not see him; y/b developed songs that were more like those of the indigobird mimicry songs than of the Bengalese finch he heard but did not see. Also, four males (r/bw, o/so, wb/v, g/pp) that were companions from independence and had no other companions or neighbours, developed identical song themes, none of which closely matched a male Bengalese. One male (probably r/bw, which was older and more experienced with male Bengalese finches in early life than two of the others, and began to sing earlier than the other three) may have modified a song of one or more models, then the four together improvised the song themes to match each other. In addition, the three indigobirds reared by firefinches copied the mimicry songs of

11 Payne et al.: Song mimicry in a brood parasite 1547 Bengalese finch fledglings (a) (b) Indigobird fledglings reared by Bengalese finches (c) (d) Firefinch fledgling (e) Indigobird adults reared by Bengalese finches (f) w/g (g) y/b Reared by firefinch (h) v/b Indigobird adult V2 (i) (j) Indigobird adult V1 (k) 0 khz s Figure 5. Begging calls of indigobirds and their foster species. (a, b) Bengalese finch fledgling, (c, d) indigobird fledglings, reared by Bengalese finches, 3 5 days out of nest, (e) firefinch fledgling, (f, g) innate begging call of indigobird adults, reared by Bengalese finches, (h) indigobird adult, reared by firefinch (acquired from V1, cf. k), (i, j) indigobird adult, wild-caught, reared by firefinch, (k) indigobird adult, wild-caught song model for v/b. Note the different begging calls of young indigobirds and Bengalese finches, the similarity in begging calls of young and adult indigobirds reared by Bengalese finches, and the similarity in firefinches, firefinch-reared wild-caught adults, and the adult firefinch-reared indigobirds that heard the older adult indigobirds. an adult indigobird that mimicked the firefinch species (Fig. 3). One of them (pw/o) developed a fair match to a firefinch mimicry song of his father in the neighbouring breeding chamber of his aviary; he did not copy the Bengalese finch mimicry songs of the male indigobird in his own aviary chamber. Development of Begging Calls If their foster siblings influence the development of song mimicry in adult indigobirds, then the young or adult indigobirds may copy the fosters begging calls. Although the song of adult indigobirds includes calls that have been identified as mimicry begging of young firefinches (Nicolai 1964; Payne 1973, 1985), the begging calls of young indigobirds have not previously been described in captives, and the juvenile begging calls of indigobirds differ from those of young firefinches in the field (Payne 1996). Indigobird nestlings and fledglings, when begging from a foster parent, repeat a single element which they vary between bouts with the intensity of begging (Fig. 5c, d). The call elements are short (less than 0.05 s), have a co-dominant overtone rather than a harmonic band, and rise rapidly in pitch to more than 2 or 4 khz, then often dip in pitch at the loudest point. In contrast, begging firefinches give two kinds of elements, a

12 1548 Animal Behaviour, 55, 6 short call and a longer whistle which develops into the distance contact call of the juvenile, and both have a complex structure with two rising sounds produced at the same time; within a begging bout they give the two elements in irregular alternation (Fig. 5e). Adult indigobirds give these firefinch-like begging calls in song. Indigobird adult song both in the field (Payne 1973) and in captivity (Nicolai 1964) includes not only mimicry of the firefinch song but also the irregularly alternating begging calls of the firefinch young. Because the vocal begging themes of the young indigobirds differ from begging of the young firefinches, and the adult indigobirds have both kinds of begging themes in their songs, the young may learn the calls from their foster siblings, then incorporate the calls into their own song when adult. Or, they may copy the calls from the songs of other adult indigobirds. All experimental males had the two singleelement indigobird begging themes in their adult song. In indigobirds reared by Bengalese finches, three were the only nestling in the brood, four (r/w, y/v, w/g, r/b-b) were reared with other young indigobirds, two (g/g, o/y) were reared with a young Bengalese finch, and two (r/b-b, v/v) were reared with a young firefinch (r/b-b had both). No birds reared together with a young Bengalese finch had elements in their adult song like the begging calls of Bengalese finches, which are longer ( s) and either fluctuate in pitch or rise in pitch, and have a conspicuous first-harmonic band (Fig. 5a, b). Because of the regular phrasing, and because the details of time and frequency in call elements appear to be identical in the young indigobird as it begs and in the adult as it sings, even when it has heard no other young or adult indigobird in its brood, the song incorporates the indigobird s own begging calls (Fig. 5c, d, f, g), and the indigobird s nestmates appear to have no effect on whether these begging calls are integrated in adult indigobird song. The three indigobirds (v/b, r/b, pw/o) reared by firefinches developed songs that included the two classes of begging call themes, those like young firefinches (Fig. 5h) and those like young indigobirds. Only one of the three indigobirds had a firefinch nestmate. Each heard a wild-caught adult male that mimicked firefinch songs and begging calls during the first year, so the indigobirds apparently learned the firefinch-like begging calls from the adult male indigobird (Fig. 5i k). Predispositions to Learn Species-typical Songs As the indigobirds are species-specific brood parasites, they might have a predisposition to learn their normal host species song and to select it as the only model to copy in their natural many-species environment. We tested the specificity of song learning by providing social and acoustic access of young Bengalese finch-reared indigobirds to firefinches and other species of finches. Eleven of 14 indigobirds reared by Bengalese finches also heard the normal host, the firefinch, during the first year of life. Their experience ranged from hearing but not seeing a firefinch, to living in the same aviary with a firefinch after they fledged. The three indigobirds (B/y, wb/v, g/pp) reared by two females copied the songs of a Bengalese finch neighbour or of an indigobird companion with Bengalese finch songs, and not the songs of firefinches, although all three heard firefinch songs. The two aviary-reared males (o/y, y/b) developed songs of both Bengalese finches and firefinch (Table I, Fig. 6). On the day he fledged, o/y begged from his foster parents, but not from firefinches and firefinches did not feed him. Both o/y and y/b lived in the roof aviary with an adult indigobird (V1) that mimicked firefinch songs, although the mimicry songs of o/y and y/b did not closely match V1 s mimicry songs. The early experience of o/y and y/b with an adult indigobird may have influenced their own songs. Both o/y and y/b lived with Bengalese finches and firefinches during their first winter, but the other Bengalese finch-reared birds that did so did not develop firefinch songs. One indigobird (pw/o) reared by firefinches gave both firefinch songs and Bengalese finch songs. His companion from independence was a firefinch that had been reared by Bengalese finches and gave both variable firefinch-typical songs and songs of Bengalese finches. Male pw/o apparently acquired both firefinch and Bengalese finch songs from his firefinch companion. Occasionally indigobirds copied the songs and calls of other finches (Table I). Male B/o was separated from his Bengalese finch fosterers and caged with a pair of firefinches, and a month later a Dybowski s twinspot was caged with them. B/o copied songs of the twinspot and not the firefinch (Fig. 7). Male r/w, caged when 6 months old with a violet-eared waxbill, Granatina granatina, copied

13 Payne et al.: Song mimicry in a brood parasite 1549 Indigobird o/y (a) Bengalese finch B34 (d) (b) Firefinch F2 (e) (c) Firefinch F20 (f) khz s Figure 6. Mimicry songs of indigobird o/y, which was reared by Bengalese finches, and his Bengalese finch and firefinch song models. (a, b, c) Indigobird o/y, (d) Bengalese finch B34, (e) firefinch F2, (f), firefinch F20. the alarms and contact trills of the waxbill (Fig. 8). Male r/w also copied the calls of a black-rumped waxbill that we acquired only after r/w was 3 years old. Although indigobirds reared by Bengalese finches occasionally copied song of firefinches, they as often copied the song of other species (one of two males housed with the violet-eared waxbill copied the waxbill, the male housed with the Dybowski s twinspot copied the twinspot, and the male housed with the black-rumped waxbill copied this waxbill). Because all male indigobirds reared by Bengalese finches copied songs of Bengalese finches, but only occasionally copied songs of firefinches, and the indigobirds reared by male Bengalese finches copied songs of other estrildid finches as often as they did songs of firefinches (two of six birds given the opportunity with companions or neighbours in each case), we conclude that indigobirds show no specific predisposition to learn the song of their normal host species, the firefinch. DISCUSSION Young brood-parasitic village indigobirds learn the songs of the species that rear them, either their normal firefinch host or an experimental foster, the Bengalese finch. The experimental results suggest that song learning involves two stages, a first stage where they become attuned to the songs of their foster parents, and a later stage where they copy other birds with songs like those of their foster parents. As predicted from this hypothesis, (1) males reared by another foster species will learn the song of that species and (2) males that leave the foster parents at the normal time of independence copy the songs of other birds, either adults of the foster species or the other adult

14 1550 Animal Behaviour, 55, 6 Indigobird B/o (a) 10 8 Dybowski's twinspot T1 (b) khz s Figure 7. Song learning of indigobird B/o, which was reared by Bengalese finches and copied songs of both Bengalese finch and Dybowski s twinspot. (a) Indigobird B/o, (b) Dybowski s twinspot T1. Indigobird r/w (a) (b) (c) 8 6 Violet-eared waxbill G1 (d) (e) (f) khz s Figure 8. Song learning of an indigobird that was reared by Bengalese finches and copied both Bengalese finches and violet-eared waxbill. (a, b, c) Indigobird r/w, (d, e, f), violet-eared waxbill G1. indigobirds that mimic the same foster species. Males reared by Bengalese finches do not change their own songs to mimic firefinch songs when they leave the foster parents and associate with firefinches at independence. That is, the indigobirds are not constrained to copy only the song of their normal host species, but can learn and copy the song of other estrildid finches that rear them and of other indigobirds that mimic the same kinds of songs. Whether an indigobird copies the song of the individual that rears him depends on the length of time he remains with his foster parents. Birds that remain with their foster father long after fledging copy the song of the foster father, whereas birds that leave within 6 weeks do not copy his song. All

15 Payne et al.: Song mimicry in a brood parasite 1551 indigobirds later learn the songs of other birds having songs similar to those of their own foster parent. The time course of song learning after independence from the foster parents is the same in birds reared by Bengalese finches and in birds reared by their normal host the firefinch. The results indicate that indigobirds learn the general features of the birds that rear them, but the details of song are acquired later from other birds with similar songs, after the age when the indigobirds normally disperse from the foster parents (Morel 1973; Payne & Payne 1977). Because an indigobird learns the songs of other individuals of his foster species and other birds with similar songs, and not only the songs of the individual that reared him, he acquires a repertoire of mimicry songs. The male that was reared by a firefinch and then had a firefinch companion with Bengalese finch song as well as firefinch songs copied both kinds of songs from his companion. In our experimental males that were reared by Bengalese finches and later had both Bengalese finch and firefinch companions and/or neighbours, five of six copied these Bengalese finches and not their normal host species the firefinch. The five males that were reared by Bengalese finches and then lived with an older indigobird companion that mimicked Bengalese finch song all copied Bengalese finch song from the older indigobird, as did another male that heard an older indigobird in the room and had no visual contact with him. The number of mimicry songs in the repertoire of birds reared by Bengalese finches was the same as in the birds reared by firefinches (three or four songs). In field conditions, the indigobirds acquire their mimicry song repertoire when they copy an adult indigobird with songs that mimic the same host species. In the field, males share their mimicry songs with each other, rather than singing the individualistic songs of their own foster parents. Also, a male indigobird has several firefinch song themes, whereas if it recalled songs only from its foster parents it would have only one or two (Payne 1985, 1990). Indigobirds can also learn the mimetic begging calls of their host species directly from other adult indigobirds. Song development of indigobirds reared by Bengalese finches is similar to that of straw-tailed whydahs reared by Bengalese finches (Nicolai 1973). The whydahs had no opportunity to hear their normal host, the purple grenadier, or adults of their own species (in contrast to the design in our indigobirds). The whydahs copied songs from their foster parents and from other Bengalese finches. The whydahs also copied the songs of other Vidua (including the firefinch song of indigobirds as song models; personal observation with Nicolai at Seewiesen in 1970), when no experienced males of their own species were present in the aviary. Finally, the whydahs had a begging call like that of young grenadiers (Nicolai 1964, 1973); this was thought to be an innate call of the young whydah that is recalled in the song of the adults, since the whydahs had not heard begging young of the host or songs of other adult whydahs. In addition, the development of begging calls in the indigobirds independently from the begging calls of their host young is similar in another species of whydah. Young nestling paradise whydahs, V. paradisaea, have a begging call like that of their host melba finch. When the whydahs were reared together with melba finches, as they became older the calls of the nestling parasite diverged from those of the nestling host (Nicolai 1969). The begging calls of paradise whydahs, like those of straw-tailed whydahs and indigobirds, appear to develop without imitation of their foster species. In the indigobirds, vocal development was not affected by the presence of foster siblings, as suggested for some other finches (ten Cate 1984). Development of song mimicry in captive-reared indigobirds was similar in several ways to the song behaviour observed in the field and in experiments where wild-reared juveniles were captured and caged with other birds: males mimic the songs of their foster species, they copy songs heard after independence, and they copy the mimicry songs of other indigobirds (Payne 1985). A male indigobird had three or four mimicry song themes both in the experiments and in the field. The diversity of mimicry song themes is thought to increase the chance that a male can attract a female indigobird that was reared by a foster parent with a similar song theme, and is driven by intersexual selection (Payne 1983, 1990). The differences in song learning observed in the field and in captivity can be explained by social conditions, where birds in the field are free to associate by choice rather than forced by circumstance in a cage. In the field, the juveniles have more social interactions and opportunities to learn songs from older male indigobirds. In the field, these songs are copied precisely between male indigobirds, which continue to