At Aristophanes, Birds 172ff., Peisetaerus persuades the Hoopoe that the birds would be better off building a city in the clouds. The Hoopoe announces that he will go off to summon the other birds to an assembly, so that the proposal may be approved. ‘How will you summon them?’, asks Peisetaerus. ‘That's easy’, replies the Hoopoe:

Why does the Hoopoe leave the stage at all? It is dramatically unnecessary for him to go inside the skênê building to sing his wake-up song (209ff.).Footnote ² Indeed, there may even be a distinct disadvantage in this set-up, in so far as singing from backstage would almost certainly impede the projection of the Hoopoe's voice, thereby spoiling a fine bit of monodic lyric.Footnote ³ What is, essentially, the right solution was suggested more than four decades ago by C.F. Russo: ‘Hoopoe's demanding virtuoso monody … would not have been delivered by the third actor, but by a professional singer.’Footnote ⁴ In Russo's wake, Kakridis (n. 2) speculated that the actor playing Tereus, being unmusical, withdrew into the skênê while a chorus member sang the monody from backstage (the chorus have not entered the orchestra yet). However, both Russo's and Kakridis's solutions seem to disregard the serious difficulty mentioned above: backstage singing would probably impede audience appreciation of the Hoopoe's elaborately constructed song.Footnote ⁵

This last point needs to be further elaborated. Wearing a mask seems to impede the projection of the performer's voice anyhow,Footnote ⁶ and so it would be reasonable to assume that whoever performed the Hoopoe's backstage song would have to do so without wearing a mask. However, the Hoopoe actor would probably not have enough time between 262 (end of monody) and 270 (reappearance on stage) to put his mask back on. Significantly, in Euripides' Medea, the Medea actor is given forty-seven lines (167–214) to put on his mask and come out from the skênê building after finishing his backstage monody (96–167). The Medea parallel is useful from another point of view too: Medea's backstage singing involves only eighteen straightforward anapaests (96–7, 111–14, 144–7, 160–7), whose main points (curses against her children, her husband and his bride) are variously repeated, no doubt for the audience's sake, in the Nurse's more controlled and, one may assume, more audible chanted anapaestic responses (115–30, 168–72), as well as in Medea's first rhêsis (214–66). Medea's backstage singing is highly unusual, and only partially paralleled by such backstage utterances – mainly appeals or requests – as Ajax's two iambic trimeters in Sophocles' Ajax 342–3 or Dionysus' seven glyconic/dactylic lines in Euripides' Bacchae (576–7, 580–1, 585, 594–5).Footnote ⁷ By contrast, the Hoopoe's monody consists of forty-eight sung lines: after a prelude of fourteen sung anapaests (209–22), there follow thirty-four lines (227–62) in a variety of highly unusual metres, the likes of which are only rarely to be found in comedy.Footnote ⁸ If, as seems likely, playwrights had good reason to limit backstage singing to relatively brief and uncomplicated utterances, it would be a shameful waste of musical expertise to have the Hoopoe actor deliver this charming song behind the skênê wall.

One may profitably compare this with modern opera practice, where offstage singing is usually limited to the first bars of an aria. For instance, in Rossini's Il barbiere di Siviglia, Figaro's short prelude (‘La ran la lera, la ran la là’) to his famous aria ‘Largo al factotum della città’ is habitually performed offstage. In Verdi's La traviata Alfredo's offstage ‘A quell'amor ch’è palpito' at the end of Act 1 provides a contrapuntal contrast to Violetta's onstage ‘Sempre libera’. In the same composer's Rigoletto, the Duke of Mantua sings offstage the first few bars of ‘La donna è mobile’ in the final act, by way of a brief allusion to the memorable aria, which had been sung in full at an earlier point. A master of the offstage song, Verdi used the same device again in Il trovatore (Act 4), where Leonora hears Manrico sing his farewell lament from inside his prison cell. Another composer known to have made extensive use of the device is Puccini: in La Bohème, Mimi sings the first bars of ‘Si, mi chiamano Mimi’ from backstage for special effect, thereby directing ‘our attention to that spot on the stage where the singer will make her entrance’, and inspiring ‘curiosity about the author of the voice.’Footnote ⁹ Puccini was to exploit the anticipatory potential of the offstage voice also in Tosca, in Madama Butterfly and in La rondine. In Donizetti's Don Pasquale, Ernesto's serenade (‘Com’è gentil la notte a mezzo april') to Norina in Act 3 functions as a somewhat extended prelude to this scene's centrepiece, which is Norina's and Ernesto's (onstage) duet ‘Tornami a dir che m'ami’. As far as I know, the only genuine exception to the practice described here is Turridu's ‘O Lola ch'hai di latti la cammisa’, sung even before the first curtain is drawn just after the instrumental overture in Mascagni's Cavalleria rusticana. This novelty represented a startling departure from established practice, and Mascagni was fully aware of its shocking potential, so much so that he did not dare to submit the song to the jury for a competition, for fear they might ‘find it too risky’.Footnote ¹⁰

A further parameter one needs to take into account is that the Hoopoe must be assumed to have sung from the stage-house roof, as if from a high perch. This is suggested both by the meaning ‘step (up) on to’ that ἐμβάς seems to have in 202Footnote ¹¹ and by the observation that such an arrangement ‘puts Tereus in the best possible position to summon the other birds, as he does, from far and wide’ in 227–63.Footnote ¹² Even the Hoopoe's wake-up call to his bedfellow (209–22) could not inappropriately be delivered from the rooftop, with only a minor compromise of naturalistic verisimilitude (assuming that this was a requirement of the comic stage). Indeed, that the Hoopoe is visible as he delivers his wake-up song is strongly suggested by 225–6 where Peisetaerus is able to deduce, presumably from the Hoopoe's movements, attitude or posture, that the bird is ‘getting ready to sing again’ (226 οὕποψ μɛλῳδɛῖν αὖ παρασκɛυάζɛται).Footnote ¹³

It is sometimes suggested that the actor playing the Hoopoe is hidden inside the ‘thicket’ (λόχμη) because Peisetaerus at 265–6 compares the Hoopoe's cries to those of a χαραδριός (perhaps a plover or a stone curlew), a bird supposedly notorious for hiding during the day or when in danger. This explanation goes back to ancient scholarship, purportedly to one Euphronius (perhaps the poet of the Pleiad),Footnote ¹⁴ but is based on the unlikely and unprofitable notion that Aristophanes attributes to the bird χαραδριός the behaviour of some of its sellers, who supposedly hid it from public view for fear that passers-by might avail themselves of the bird's jaundice-curing properties for free (Hipponax fr. 52 West explicitly, and logically, associates such behaviour not with the bird but with its sellers.) It is more likely that Peisetaerus, in comparing the Hoopoe's cries (266 ἐπόπωζɛ Schroeder: ἐπῶ(ι)ζɛ codd.) to those of a stone curlew, is belittling, tongue-in-cheek, the Hoopoe's splendid song: the χαραδριός was possessed of an ‘unremarkable’ (ϕαῦλος) voice according to Aristotle.Footnote ¹⁵

Thus, we are presented with what appears to be a conundrum. If the actor playing the Hoopoe was a singer, then there was no reason for him to be removed offstage. If on the contrary he could not sing, and had to be substituted by an unseen virtuoso, having him appear on the stage-house roof or some such visible place would defeat the purpose of removing him offstage in the first place. The only solution that lies to hand, it seems, is to presume that as soon as the actor playing the Hoopoe disappeared inside the skênê building, an identically costumed singer wearing an identical Hoopoe's mask climbed up the stairs and perched on the rooftop. This would, additionally, remove the difficulty of having the Hoopoe actor get up to roof level between 208 and 209.Footnote ¹⁶

The arrangement suggested here is consistent with Sommerstein's likely hypothesis that, since Aristophanes ‘would hardly have envisaged this clumsy arrangement when he originally composed the play’, he presumably resorted to it as a solution of despair because in the event he was ‘unable to find a single person with the requisite abilities both as an actor and as a singer.’Footnote ¹⁷ Since Birds was produced/directed by Callistratus,Footnote ¹⁸ it may even be that, when writing the script, Aristophanes simply failed to foresee or take into account the unavailability of a suitably qualified actor-singer, hence the seeming mismatch between original composition and eventual performance. As Parker (n. 8), 302 reminds us, ‘exigencies of this kind have left their mark on masterpieces of music and drama’: as a notable case in point she adduces Twelfth Night 2.4, where the disguised Viola, having previously announced her plan to ‘sing | and speak to [Orsino] in many sorts of music’ (1.2.57–8), fails to fulfil her promise, and Orsino asks instead for Feste to be brought onstage to sing ‘that old and antique song we heard last night’ (2.4.3). Thus, on principle it does not seem entirely implausible that, in the case of Birds too, a last-minute arrangement should have been permanently incorporated in the text.

In addition, it appears that the chorêgos for the Birds production was anything but niggardly: apart from the Hoopoe and his Servant Bird, he funded what must have been a lavishly costumed chorus and was even determined to bring on four extra birds (268–93) purely to make a show and provide a feed for a few jokes.Footnote ¹⁹ We may thus reasonably speculate that he will not have baulked at the relatively small additional cost of a duplicate costume.Footnote ²⁰