1 The original version of Madama Butterfly was premièred in 1904; it was revised several times until 1906, when the final version was premièred. Turandot was incomplete when Puccini died in 1924; it was completed after his death by Franco Alfano and premièred in 1926.

2 See, for example, Helen M. Greenwald, ‘Picturing Cio-Cio-San: House, Screen, and Ceremony in Puccini's Madama Butterfly’, Cambridge Opera Journal, 12 (2000), 237–59 (pp. 239–40), and Kunio Hara, ‘Rudolf Dittrich's Nippon Gakufu and Giacomo Puccini's Madama Butterfly’, Music Research Forum, 19 (2004), 1–25 (pp. 21–3). On the ‘complex and contradictory’ nature of Madama Butterfly in terms of orientalist representation, see Ralph P. Locke, Musical Exoticism: Images and Reflections (Cambridge, 2009), 206–13. For an overview of scholarship on operatic orientalism, see my ‘Exoticism’, The Oxford Handbook of Opera, ed. Helen M. Greenwald (New York and Oxford, 2014), 795–816.

3 I have traced these themes in the 1904 Italian edition of the vocal score (Ricordi 110000), the 1904 piano reduction score (Ricordi 110001), the 1905/6 English/Italian editions of the vocal score (Ricordi 111200), the 1907 French edition of the vocal score (Ricordi 111360) and the 1907 full score (reprinted by Dover in 1990). (In this article, I will refer to rehearsal figures and bar numbers in the 1907 full score, unless otherwise noted.) On Puccini's revisions to the Madama Butterfly score, see Julian Smith, ‘A Metamorphic Tragedy’, Proceedings of the Royal Musical Association, 106 (1979–80), 105–14; idem, ‘Tribulations of a Score’, Madam Butterfly/Madama Butterfly, ed. Nicholas John (London, 1984), 15–23; William Ashbrook, ‘Reflections on the Revisions of Madama Butterfly’, Giacomo Puccini: L'uomo, il musicista, il panorama europeo, ed. Gabriella Biagi Ravenni and Carolyn Gianturco (Lucca, 1997), 159–68; Dieter Schickling, ‘Puccini's “Work in Progress”: The So-called Versions of “Madama Butterfly”’, Music and Letters, 79 (1998), 527–37; and Philip Gossett, ‘Some Thoughts on the Use of Autograph Manuscripts in Editing the Works of Verdi and Puccini’, Journal of the American Musicological Society, 66 (2013), 103–28.

4 On Puccini's multiple sources of information on Japanese music, see especially Arthur Groos, ‘Cio-Cio-San and Sadayakko: Japanese Music-Theater in Madama Butterfly’, Monumenta nipponica, 54 (1999), 41–73, and Kunio Hara, ‘Puccini's Use of Japanese Melodies in Madama Butterfly’ (M.M. thesis, University of Cincinnati, 2003).

5 On Puccini's use of Japanese themes with apparently meaningful associations in the opera, see for example Michele Girardi, Puccini: His International Art, trans. Laura Basini (Chicago, IL, 2000), 215–16.

6 I will discuss the Fassini music box and its impact on Puccini scholarship below. I should note that although I have decided to follow Puccini scholars and American musicologists more generally in employing the term ‘music box’, the term ‘musical box’ is preferred by many in the community of mechanical music collection and scholarship. See Arthur W. J. G. Ord-Hume, Musical Box: A History and Collector's Guide (London, 1980). The Swiss manufacturers referred to these mechanical musical instruments as boîtes à musique.

7 Ord-Hume reports that the term ‘harmoniphone’ was used by Louis Ducommun for his vox celeste cylinder boxes, to which a reed organ had been added, and that this manufacturing practice began around 1850. See Ord-Hume, Musical Box, 134–5. The dimensions of the Guinness box with lid closed are: 12⅞ inches high by 20¼ inches wide by inches deep. The serial number that appears in multiple locations on the Guinness box is 2098. This number, as well as the construction of the cylinder mechanism, helps to identify this instrument as having been manufactured by C. Paillard et C^ie in Ste Croix, Switzerland, just prior to 1880. (I am grateful to Jere Ryder, conservator of the Guinness Collection, for assistance in dating this music box.) Markings on the box made by a restorer based in Rhode Island indicate that Guinness acquired the box by 1964. For recordings of all six tunes on the Guinness box, see <http://www.morrismuseum.org/mechanical-musical-instruments-automata> (accessed 18 September 2014). On the Guinness Collection, see Ellen M. Snyder-Grenier, Musical Machines and Living Dolls: The Murtogh D. Guinness Collection of Mechanical Musical Instruments and Automata (Morristown, NJ, 2011). On my discovery of this music box, also see my ‘Music Box as Muse to Puccini's “Butterfly”’, New York Times (17 June 2012), Arts and Leisure section, pp. 1 and 9, and ‘The Sound of Scholarly Serendipity’, The Institute Letter, Institute for Advanced Study (Fall 2012), 10. A colour version of Figure 1 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

8 Scholars continue to refer to the Fassini music box as a ‘Chinese music box’ and an ‘authentic Chinese’ musical source. See, for example, Andrew Davis, Il Trittico, Turandot, and Puccini's Late Style (Bloomington, IN, 2010), 178. (Online comments by collectors and lot descriptions by auction houses reveal that music-box collectors also assume that the music played by these boxes is fully Chinese.) On the marketing of Swiss boxes in China, see for example John E. T. Clark, Musical Boxes: A History and an Appreciation (3rd edn, London, 1961). Clark states that large numbers of reed-organ boxes were shipped to Persia, India and China (p. 61). On the Swiss music-box industry more generally, see Vincent Arlettaz, ‘Histoire d'un amour: Sainte-Croix et la boîte à musique’, Revue musicale de Suisse Romande, 60 (2007), 6–23.

9 Georg Capellen, ‘Madame Butterfly und die Exotik’, Neue Musik-Zeitung, 30 (1909), 465–8. Capellen himself published arrangements of Japanese songs for piano. See Shogaku Shoka: Japanische Volksmelodien des Isawa Shuji (Leipzig, 1904).

10 Mosco Carner, ‘The Exotic Element in Puccini’, Musical Quarterly, 22 (1936), 45–67 (p. 47; see also p. 45).

11 Ibid., 50–4. Carner cited two collections of Japanese melodies published by Breitkopf & Härtel: Rudolf Dittrich's Nippon Gakufu (1894) and the Shuji/Capellen 1904 collection. Carner's discussion of Puccini's operatic exoticism includes La fanciulla del West. In a later publication, Carner mistakenly referred to Butterfly's theme as ‘the first of the identifiable Japanese tunes that Puccini uses in the opera’. See Mosco Carner, Madam Butterfly: A Guide to the Opera (London, 1979), 50.

12 Mosco Carner, Puccini: A Critical Biography (3rd edn, London, 1992), 415.

13 Such publications include Karl Gustav Fellerer, Giacomo Puccini (Potsdam, 1937), 65–75, 98–108; Duiti Miyasawa, ‘Madama Butterfly's Original Melodies’, Opera News, 16 (28 January 1952), 7–9, republished in a slightly revised version as Juichi [sic] Miyazawa, ‘Some Original Japanese Melodies in Madama Butterfly’, Giacomo Puccini: Nel centenario della nascita (Lucca, 1958), 157–61; Norbert Christen, Giacomo Puccini: Analytische Untersuchungen der Melodik, Harmonik und Instrumentation (Hamburg, 1978), Chapter 5, pp. 207–92; and Mary Renner Heath, ‘Exoticism in Puccini: The Japanese Melodies in Madama Butterfly’, Opera Journal, 13 (1980), 21–8. In reference to the composition of Turandot, William Ashbrook and Harold Powers stated that Puccini ‘felt he needed to find authentic melodies, as he had for Madama Butterfly twenty years earlier’, and they refer to Puccini's use of ‘authentic Chinese tunes’ in that opera, although they place ‘authentic’ in scare quotes in some cases. See Ashbrook and Powers, Puccini's Turandot: The End of the Great Tradition (Princeton, NJ, 1991), 61 and 94. On Puccini's use of Chinese melodies in Turandot, also see Peter Korfmacher, Exotismus in Giacomo Puccinis ‘Turandot’ (Cologne, 1993), 80–95, and Kii-Ming Lo, ‘Turandot’ auf der Opernbühne (Frankfurt am Main, 1996), 318–36.

14 Hara, ‘Puccini's Use of Japanese Melodies in Madama Butterfly’, Chapter 4, and idem, ‘Rudolf Dittrich's Nippon Gakufu’.

15 Hara, ‘Rudolf Dittrich's Nippon Gakufu’, 2–3.

16 The flowing quavers and other details somewhat resemble other specific bars in ‘Rakubai’. I should note that ‘Rakubai’ is also included as no. 6 in Collection of Japanese Koto Music (Tokyo, 1888) at the same pitch but not with the same harmonization as in the Dittrich arrangement.

17 Carner, ‘The Exotic Element in Puccini’, 65. On Puccini's special orchestration for Butterfly's entrance, also see Carner, ‘L'elemento esotico in Puccini’, La musica occidentale e le civiltà musicali extraeuropee, ed. Stelio Felici (Florence, 1971), 140–6 (p. 144). Betraying a lack of familiarity with Japanese music, Carner claimed that the final bars of the opera are ‘an almost exact reproduction of the timbre of Japanese temple-music’. See Carner, Puccini, 417.

18 Carner, Madam Butterfly, 53 and 56.

19 Greenwald, ‘Picturing Cio-Cio-San’, 237–8.

20 Ibid., 239–40.

21 Peter Ross, ‘Elaborazione leitmotivica e colore esotico in “Madama Butterfly”’, Esotismo e colore locale nell'opera di Puccini, ed. Jürgen Maehder (Pisa, 1985), 99–110 (pp. 105–7); and Girardi, Puccini, trans. Basini, 237. David Rosen has pushed back against the notion that the exotic features of Butterfly's theme are stripped away in ‘Io seguo il mio destino’, and instead argues that it remains strongly ‘oriental’ in its marking and function, indicating that she cannot escape her Japanese identity no matter how hard she may try. See Rosen, ‘“Pigri ed obesi dei”: Religion in the Operas of Puccini’, Madama Butterfly: L'orientalismo di fine secolo, l'approccio pucciniano, la ricezione, ed. Arthur Groos and Virgilio Bernardoni (Florence, 2008), 257–98 (p. 266).

22 Locke, Musical Exoticism, 187.

23 Ibid., 204.

24 Kimiyo Powils-Okano, Puccinis ‘Madama Butterfly’ (Bonn, 1986), 49 and 61–2. Powils-Okano's supposition that the theme had no East Asian source is incorrect, however, as we shall see.

25 Julian Budden, ‘Forte e nuova, ma non facile’, Madama Butterfly 1904–2004, ed. Ilaria Narici (Milan, 2004), 11–29 (p. 26). See also Budden, Puccini: His Life and Works (Oxford, 2002), 248–9.

26 Arthur Groos, ‘The Sketches at the Accademia Filarmonica in Bologna’, Madama Butterfly: Fonti e documenti della genesi, ed. Groos (Lucca, 2005), 143–94 (p. 143). Also see Luigi Verdi, ‘I manoscritti della Madama Butterfly nell'Archivio dell'Accademia Filarmonica di Bologna’, Madama Butterfly, ed. S. Camerini, catalogue of the exhibition ‘Butterfly: I manoscritti ritrovati nell'Archivio dell'Accademia Filarmonica di Bologna (19–31 March 1996) (Bologna, 1996), 57–90.

27 Groos, ‘Cio-Cio-San and Sadayakko’, 45. Describing Butterfly's entrance, Groos writes: ‘The heroine directs the others to genuflect to the accompaniment of a more individualized melody, apparently derived from the Vivace of the 1901 Bologna sketch, and orientalized through the use of octave grace notes and pizzicato’ (ibid., 64). See also Arthur Groos, ‘Madama Butterfly between Comedy and Tragedy’, Madama Butterfly, ed. Groos and Bernardoni, 159–81 (p. 166).

28 Girardi emphasized the importance of Puccini's use of ‘authentic themes’ and offered ‘Hana saku haru’ as a source for Butterfly's theme (despite Powils-Okano), before following Groos by pointing to the Bologna manuscript as a more likely model. See Girardi, Puccini, trans. Basini, 211–13.

29 Capellen, ‘Madame Butterfly und die Exotik’, 468.

30 Powils-Okano, Puccinis ‘Madama Butterfly’, 59–60. Several commentators have followed her with this attribution. For example, see Budden, Puccini, 249.

31 Groos, ‘Madama Butterfly between Comedy and Tragedy’, 166, and idem, ‘Cio-Cio-San and Sadayakko’, 47, 64 and 66–8. Also see Hara, ‘Puccini's Use of Japanese Melodies in Madama Butterfly’, 11–14.

32 Groos, ‘Cio-Cio-San and Sadayakko’, 65. Also see idem, ‘Madama Butterfly between Comedy and Tragedy’, 166, 169.

33 Groos, ‘Cio-Cio-San and Sadayakko’, 66–8. On this theme and its transformations, also see Ross, ‘Elaborazione leitmotivica e colore esotico in “Madama Butterfly”’, 100–4.

34 Rosen, ‘“Pigri ed obesi dei”’, 264.

35 Capellen, ‘Madame Butterfly und die Exotik’, 468. Although Capellen was correct in suspecting that some of the themes in the opera might be of Chinese origin, he was mistaken in emphasizing differences between a ‘Chinese’ and a ‘Japanese’ pentatonic scale.

36 Quoted in Groos, ‘Cio-Cio-San and Sadayakko’, 49. This quotation comes from a letter Puccini wrote to Luigi Illica dated 30 April 1902.

37 The multiple differences between the Swiss music-box realizations of ‘Shiba mo’ and folk performance versions of the melody in China are striking. Puccini's setting, of course, further obscures identification of the Chinese melody by a listener familiar only with Chinese folk performance styles. This is one reason why hearing the music-box version of the tune is crucial in attempting to notice and identify the tune as employed in Puccini's opera. The transcriptions from the Guinness music box are my own, and are based primarily on the reed-organ realizations of each melody. A sound clip for Example 2 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

38 Julian Budden, ‘La dissociazione del Leitmotiv nelle opere di Puccini’, Giacomo Puccini, ed. Ravenni and Gianturco, 453–66 (p. 460). Groos has stated that ‘in spite of Puccini's extensive use of Japanese materials to represent his vision of a “true Japan”, there seems to be no Japanese melody associated particularly with the heroine’ (‘Cio-Cio-San and Sadayakko’, 62). I argue that this Chinese melody is quite closely associated with Butterfly in the opera, with only a few exceptions.

39 Kii-Ming Lo has made a similar claim regarding Puccini's use of three melodies from a music box in Turandot, arguing that they serve a more integral function in the score than do the Chinese melodies that he derived from printed sources. See Lo, ‘Giacomo Puccini's Turandot in Two Acts: The Draft of the First Version of the Libretto’, Giacomo Puccini, ed. Ravenni and Gianturco, 239–58 (pp. 252–3).

40 Here I am in agreement with Locke (Musical Exoticism, 207–8) but not with James Parakilas, who labelled Butterfly's entrance ‘non-exotic’. See Parakilas, ‘The Soldier and the Exotic: Operatic Variations on a Theme of Racial Encounter’, Opera Quarterly, 10/2 (winter 1993), 33–56 (p. 47), and 10/3 (spring 1994), 43–69 (pp. 50–1).

41 A sound clip for Example 3 may be accessed at <http://music.williams.edu/profile/wsheppar/>.

42 For a related example of the influence of music boxes on musical style in the nineteenth century, see Jeffrey Kallberg, ‘Chopin's Music Box’, Chopin's Musical Worlds: The 1840's, ed. Artur Szklener (Warsaw, 2008), 189–202. Kallberg points to a connection between exoticism and music boxes, and notes a ‘clear association in the nineteenth century between automata and various notions of the Other’, postulating that a music-box style itself ‘would likely evoke a link with various forms of difference’ (pp. 201–2).

43 Hara has stated that ‘Che tua madre’ is the only aria of Butterfly's to be based ‘entirely on Japanese or Japanese-influenced music’. This remains correct, given that ‘Shiba mo’ is a Chinese folksong. See Hara, ‘Rudolf Dittrich's Nippon Gakufu’, 19.

44 A sound clip for Example 4 may be accessed at <http://music.williams.edu/profile/wsheppar/>.

45 Budden, ‘La dissociazione del Leitmotiv nelle opere di Puccini’, 459–60. This is actually a full statement of ‘Shiba mo’.

46 Vera Micznik, ‘Cio-Cio-San the Geisha’, A Vision of the Orient: Texts, Intertexts, and Contexts of Madame Butterfly, ed. Jonathan Wisenthal et al. (Toronto, 2006), 47–55. Micznik refers to this theme as a ‘seldom mentioned motive that appears in the most crucial moments’ (p. 47). I fully agree that the theme is crucial, but it has hardly been ignored by Puccini scholars.

47 Ibid., 49.

48 Ibid., 52.

49 Ibid., 53.

50 Antoinet Schimmelpenninck, Chinese Folk Songs and Folk Singers: Shan'ge Traditions in Southern Jiangsu (Leiden, 1997), 87. Schimmelpenninck states that in Jiangsu ‘many different versions circulate’ of this song (p. 167). Also see Frank Kouwenhoven and Antoinet Schimmelpenninck, ‘Zhao Yongming: Portrait of a Mountain Song Cicada’, Lives in Chinese Music, ed. Helen Rees (Urbana, IL, 2009), 23–44 (p. 30). On this genre and its appearance in nineteenth-century Chinese songbooks, see Stefan Kuzay, ‘Life in the Green Lofts of the Lower Yangzi Region’, Lifestyle and Entertainment in Yangzhou, ed. Lucie B. Olivová and Vibeke Børdahl (Copenhagen, 2009), 286–314. A 1920s recording of ‘Shiba mo’ as performed in a Cantonese opera is available at <http://www.civilization.ca/cantoneseopera/opera132-1-e.shtml> (accessed 18 September 2014). A field recording made by Schimmelpenninck in the 1980s is available on a CD accompanying Chinese Folk Songs and Folk Singers and at <http://www.iias.nl/oideion/journal/issue01/zanten/5-audio4.html> (accessed 18 September 2014).

51 In an unpublished paper delivered at the 2002 CHIME conference in Sheffield and entitled ‘On “Eighteen Touches”’, Liu Yong referred to ‘Shiba mo’ as ‘the most popular erotic folksong in China’, and stated that ‘when a writer needs an example of an erotic song, “Eighteen Touches” is always the first one mentioned’. (I am grateful to the author for providing me with a transcript of that paper.) I have identified fictional works that employ reference to ‘Shiba mo’ in this way. See, for example, Han Z. Li, The Water Lily Pond: A Village Girl's Journey in Maoist China (Waterloo, ON, 2004), in which a sexually charged village woman is said to hum an ‘underground song, “Eighteen Touches”, by which a man engaged a woman prior to the act’ (p. 6).

52 See Schimmelpenninck, Chinese Folk Songs and Folk Singers, 118, 168. Also see Shi-Zheng Chen, ‘The Tradition, Reformation, and Innovation of Huaguxi: Hunan Flower Drum Opera’, Drama Review, 39 (1995), 129–49 (p. 134).

53 In discussing the huaguxi play Shi ba mo, Chen reports that in ‘the music and movement, through solos and duets, the feel of the woman's body from head to toe is graphically explored’. According to Chen's informant, this traditional play – ‘about a sophisticated playboy who goes to a prostitute’ – apparently drove young audience members wild, and they ‘would rush home to make love as soon as the play finished’ (ibid., 132). I should note here that the Japanese term ‘shiba-shiba mo’ (meaning ‘often’ or ‘many times’) occurs in the classic Manyoshu poetry collection in a telling of the famous Tanabata love story in which two celestial lovers do not often have the opportunity to meet, given their separate positions in the night sky. I believe that this linguistic echo is entirely coincidental to Puccini's use of the ‘Shiba mo’ melody.

54 Schimmelpenninck, Chinese Folk Songs and Folk Singers, 167.

55 Jin Jiang, ‘Dubious Prosperity: Women and Entertainment in Wartime Shanghai’, Frontiers of History in China, 4 (2009), 124–48 (p. 138).

56 Liu Yong reported on the use of ‘Shiba mo’ in foreplay in ‘On “Eighteen Touches”’.

57 Jonathan P. J. Stock, Huju: Traditional Opera in Modern Shanghai (Oxford, 2003), 65. In this instance, the male role was taken by a female performer.

58 W. H. Blake, The Adventures of a Naval Chief Gunner (Brisbane, 1906), 245. Here Blake refers particularly to the Yu-ling-tong dining house in Ching-kiang where there were ‘probably two hundred good-looking singing girls’. The significance of the other song mentioned by Blake here, ‘Sinfa’, for Puccini's Turandot will be discussed below.

59 Groos, ‘Madama Butterfly between Comedy and Tragedy’, 159–81. As Groos notes, the revisions removed many (but not all) of the comic elements in the opera (p. 180).

60 Smith, ‘A Metamorphic Tragedy’, 105.

61 On interpreting Pinkerton, see, for example, Susan McClary, ‘Mounting Butterflies’, A Vision of the Orient, ed. Wisenthal et al., 21–35. On the transformations of the Pinkerton character in film versions of the Butterfly story, see my ‘Cinematic Realism, Reflexivity and the American “Madame Butterfly” Narratives’, Cambridge Opera Journal, 17 (2005), 59–93, and for discussion of the Pinkerton character in Tin Pan Alley songs and in the alternative rock band Weezer's album Pinkerton, see my forthcoming Extreme Exoticism: Japan in the American Musical Imagination.

62 A sound clip for Example 5 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

63 After the Bonze and the rest of Butterfly's family exit we hear a gentle, extended version of the Patrimony theme (including the rocking quaver gesture) as Pinkerton comforts the weeping Butterfly and she declares that she will not allow her family's repudiation to bother her. In the 1904 version of the opera, at a moment leading up to the love duet when Butterfly suddenly breaks free of Pinkerton's embrace and considers her past (immediately prior to rehearsal figure 131; also see figure 126 in the 1907 full score), the opening D–C–E gesture of the Patrimony theme is asserted in the basses. A sound clip for Example 6 may be accessed at <http://music.williams.edu/profile/wsheppar/>.

64 The theme that closes the opera (Act 2, rehearsal figure 58) and that is associated with Butterfly's own death is clearly associated with – but does not supersede – the Patrimony theme. This theme was based on the Japanese song ‘Suiryo-bushi’. See Powils-Okano, Puccinis ‘Madama Butterfly’, 57–8.

65 ‘Bangziqiang’ refers to the distinctive melodies in Bangzi opera and ‘the Bangzi melody’ is referred to as one of the four great characteristic melodies or forms in Chinese opera. ‘Pang-tzu’ (‘wooden-block tunes’) served as an older transliteration of the term.

66 James Bell, A System of Geography, Popular and Scientific (London, 1849), 112. Bell also refers to this form of punishment as the ‘bastonade’. Coincidentally, this form of punishment appears in another opera from the early twentieth century. The Courtiers at the end of Act 1 in Stravinsky's Chinese musical fairy tale Le rossignol (1908–14) state that if they had not successfully located the nightingale for their Emperor they would have received a hundred blows of the bamboo rod: ‘Nous aurions certes échoué! / Et pour souper nous aurions eu les reins / de cent coups de bambou frottés!’

67 St Jean Gabriel Perboyre (1802–40), for example, is believed to have suffered pantse (either 110 or 100 strokes from the bamboo rod) for not trampling on the crucifix when required to do so in China.

68 I have not been able to identify a Chinese model melody that would explain the choice of ‘Bangzi’ for the title of this music-box tune. It is possible that this melody derives from bangzi opera and that Puccini or an associate mistakenly read the transliterated term ‘poutzi’/‘pantse’ in its meaning as a form of punishment rather than in its alternative definition referring to a musical instrument and an operatic genre.

69 A sound clip for Example 7 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

70 Scholars have noted, instead, the similarity between the opening gesture in Turandot and the opening of Iago's ‘Credo’ in Verdi's Otello (see, for example, Carner, Puccini, 529). Helen Greenwald has previously noted the rhythmic similarity between the end of Butterfly and the opening of Turandot, suggesting that the rhythmic style served as a general sign of Asian local colour for Puccini, but she does not develop this observation further. See Greenwald, ‘Character Distinction and Rhythmic Differentiation in Puccini's Operas’, Giacomo Puccini, ed. Ravenni and Gianturco, 495–515 (pp. 504–5, note 17). On the unresolved ending of Madama Butterfly, see Deborah Burton, Recondite Harmony: Essays on Puccini's Operas (Hillsdale, NY, 2012), 30–1.

71 Maria Ng has offered a brief general comparison of the two operas, claiming unconvincingly that Butterfly and Turandot reflect the differences between Japanese and Chinese cultures and the differences in Western perceptions of these two nations c.1900. See her ‘The Taming of the Oriental Shrew: The Two Asias in Puccini's Madama Butterfly and Turandot’, A Vision of the Orient, ed. Wisenthal et al., 170–80.

72 The journalist Luigi Gualtiero Paolini published the news of Puccini at work on Turandot in Bagni di Lucca in the Giornale d'Italia on 19 August 1920. Quoted in Gabriella Biagi Ravenni and Daniela Buonomini, ‘“Caro Ferruccio …”: Trenta lettere di Giacomo Puccini a Ferruccio Giorgi (1906–1924)’, Giacomo Puccini, ed. Ravenni and Gianturco, 169–209 (p. 190, note 3). Paolini reported that the music box had been manufactured in Germany and imported to China. Although it was actually probably made in Switzerland, it is striking that the earliest report was correct on the European origins of the music box, a fact missed by most other commentators on the role of the Fassini music box in the composition of Turandot. For example, Carner repeatedly referred to Puccini's use of an ‘ancient musical-box’ or an ‘ancient Chinese musical-box’. See, for example, Carner, Puccini, 247, 522.

73 Giuseppe Adami, Il romanzo della vita di Giacomo Puccini (Milan, 1942), 241.

74 Ibid.

75 Giuseppe Adami, Puccini (2nd edn, Milan, 1935), 200.

76 On Fassini's career, see, for example, Gaetano Bazzani, Soldati italiani nella Russia in fiamme 1915–1920 (Trent, 1933), 228–9, and Dionigi Roggero, ‘Il barone Fassini Camossi’, Il Monferrato (December 2004), available at <http://www.comune.penango.at.it/files/Monferrato/monf2.pdf> (accessed 18 September 2014). Fassini's career was accurately referenced in Ravenni and Buonomini, ‘“Caro Ferruccio …”’, 198. These authors also cite a letter from Puccini of 19 January 1921 in the form of a poem that suggests the composer had borrowed money from Fassini (ibid., 196). I am grateful to Bibiana Gattozzi for contacting the Italian Embassy in Beijing and reporting that there was no record of any diplomatic post for Edoardo Fassini in the first decades of the twentieth century.

77 Fassini published at least seven short songs and piano pieces between 1911 and 1915 with lyrics in English, French and Italian. The following are held at the Biblioteca Nazionale Centrale, Florence: ‘Midshipmen's Two-step’, ‘Little Spring’, ‘La valse aérienne’, ‘Parvenza’, ‘Avant l'oubli’, ‘Quand je te dis …’ and ‘Invan tu chiedi’.

78 The British, for example, apparently held loot auctions every day except for Sundays, and men of many nationalities took advantage of the countless bargains to be had. See ‘Loot Auctions in Peking’, New York Times (20 January 1901), 11. Also see James L. Hevia, ‘Looting Beijing: 1860, 1900’, Tokens of Exchange: The Problem of Translation in Global Circulations, ed. Lydia H. Liu (Durham, NC, 1999), 192–213, and idem, ‘Looting and its Discontents: Moral Discourse and the Plunder of Beijing, 1900–1901’, The Boxers, China, and the World, ed. Robert Bickers and R. G. Tiedemann (Lanham, MD, 2007), 93–113. On Italian participation in the international allied force in the Boxer War, see Ludovica de Courten and Giovanni Sargeri, Le regie truppe in estremo oriente 1900–1901 (Rome, 2005), and Fabio Fattore, Gli italiani che invasero la Cina: Cronache di guerra 1900–1901 (Milan, 2008).

79 I am grateful to Dr Marie-Nolle Snider-Giovannone, who shared her research on this Italian military expedition with me. On this unique expedition, see Antonio Mautone, Trentini e italiani contro l'armata rossa: La storia del Corpo di Spedizione in Estremo Oriente e dei ‘Battaglioni Neri’ 1918–1920 (Trent, 2003), 161–7, and ‘La legione redenta’, <http://digilander.libero.it/fiammecremisi/eramoderna/apocrifi35.htm> (accessed 16 September 2014).

80 See Daniele Varè, Laughing Diplomat (New York, 1938), 143–4. Varè relates that ‘many of the irredenti were musicians. We bought musical instruments for them in Shanghai and started a band. There was also a chorus of fifty voices’ (ibid., 144).

81 Ibid., 146–7. Varè reports that when he asked Baron Fassini about this story, Fassini ‘laughed and would not confirm or deny’.

82 Hammelmann and Rose were able to locate the Baroness Fassini in Rome thanks to a tip from a gas-station attendant in Lucca. I am very grateful to Michael Rose for reconstructing the history of this episode for me and for providing copies of the original tapes that were made in Rome. My account here is based on our extensive correspondence, particularly on Rose's detailed account sent to me by email on 23 February 2012. Rose has recently published a book based on the ‘Birth of an Opera’ series: Michael Rose, The Birth of an Opera: Fifteen Masterpieces from Poppea to Wozzeck (New York, 2013).

83 The 1971 RAI television broadcast featured the musicologist Giorgio Pestelli and conductor Giancarlo Chiaramello discussing the role of the music box in the composition of Turandot. Both Pestelli and Chiaramello clearly felt awkward in ‘accusing’ Puccini of not having composed these melodies, and they carefully defended the composer's ‘originality’. In both the 1971 RAI and 1983 Cronache italiane televised programmes, the music box is accurately described by the collector Gigi Capra as having been manufactured in Switzerland. Capra also felt compelled to defend Puccini against charges of plagiarism. I thank Bibiana Gattozzi for viewing these television programmes with me and for translating from the Italian.

84 The account of the rediscovery of the Fassini music box offered by Mary Jane Phillips-Matz in her biography of the composer is inaccurate and was based on a 2001 interview with Weaver, who was ailing at the time and was being asked to recall details of an event that occurred nearly four decades earlier. For example, Weaver is quoted as saying that ‘there were five tunes [on the music box], and three of those were quite obviously in Turandot. It was an eerie experience with this tinkly music and the idea that Puccini had had this box. He kept it forever.’ The music box contains four melodies and Puccini did not have the box in his possession. See Phillips-Matz, Puccini: A Biography (Boston, MA, 2002), 266.

85 See Ashbrook and Powers, Puccini's Turandot, Chapter 4. Powers's transcriptions of the Weaver broadcast – made from a tape loaned to him by Weaver – and his brief notes on the music-box melodies are held in the Powers collection at the Mendel Music Library, Princeton University. I am grateful to Gene Lavergne for providing me with a recording of the Weaver broadcast.

86 The Fassini music box is currently held by Lionello Ghiotti, son-in-law of the late Gigi Capra. I am very grateful to Dr Ghiotti for providing me with a video recording and photographs of the music box. The existing tune sheet on this box appears to have been cut from a larger original, as none of the typical printed decoration remains and part of the border is clearly missing. Dr Ghiotti has explained how he acquired the music box in ‘Gigi Capra e il carillon di Turandot’, Graf Boccassi, 2 (2009), 11–13. On the basis of the available evidence, including the Italian television broadcasts noted above, I am able to reconstruct the migrations of the Fassini music box. The box appears to have been manufactured in Switzerland and exported to China in the late nineteenth century. Baron Fassini returned with the box to Italy either after the Boxer War or in 1920 and brought the instrument to Bagni di Lucca, where Puccini encountered it in August 1920. The music box then travelled with Fassini's widow to Rome, but appears to have been given to one of her children and was lent back to her for the BBC recording in 1965. The box was then owned by Pietro Fassini, who gave it to Gigi Capra in Alessandria. Ghiotti's wife, the daughter of Gigi Capra, inherited the music box from her father. (Note that on p. 12 of Ghiotti's article Baron Pietro Fassini-Camossi is said to have given the box to Gigi Capra in 1950. Dr Ghiotti has clarified to me that this date was a mistake on his part and that the correct year was probably 1970, in time for Capra to present the box in the 1971 Italian television broadcast. Dr Ghiotti's box is quite clearly the same one that was featured in the Italian television broadcasts mentioned above.) A colour version of Figure 3 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

87 Powers notes that a transcription of ‘Mo Li Hua’ had appeared in John Barrow's 1805 Travels in China in a slightly different version, but he assumes that Puccini was unaware of that source and thus would not have known the Chinese title of this melody. However, Puccini would have seen the tune card on the Fassini box and might have been told that ‘Sinfa’ can be translated as ‘Fresh Flowers’. See Ashbrook and Powers, Puccini's Turandot, 95; 176, note 5. Powers notes that a ‘different form of this tune, having nothing whatever to do with Turandot’ was published in J. A. van Aalst's Chinese Music (Shanghai, 1884) under the title ‘Hsien Hua’ (p. 176, note 9). However, the title ‘Sinfa’ appears on both the Guinness and Fassini music boxes as well as on numerous other music boxes from the period, and van Aalst's publication represents a somewhat divergent transcription of a variant of this melody. Following Barrow, Puccini scholars have consistently referred to the melody as ‘Mo Li Hua’, or ‘Moo-Lee-Wha’ (see, for example, Carner, Puccini, 523).

88 The fourth melody on this box appears also as the fifth melody on the Guinness box, but with a different title on the tune sheet.

89 Swiss music boxes were also exported to Turkey, Persia and Japan, and occasionally featured melodies from those exotic lands. See Alfred Chapuis, History of the Musical Box and of Mechanical Music, trans. Joseph E. Roesch (Summit, NJ, 1980), 261–70.

90 A colour version of Figure 4 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

91 For the fifth tune on the Guinness box (‘Shanghai Pier’) the characters are transliterated on the tune sheet as ‘Shanghai Mody’ (they read ‘Shanghai Matou’ in modern transliteration). As we saw, a similar mistake appears on the Fassini box tune sheet and, in general, the surviving tune sheets for Swiss boxes featuring Chinese tunes are riddled with errors and a wide range of variant spellings of the same transliterated titles.

92 The fact that ‘She pah moh’ on the tune sheet does not line up correctly with the ‘Shiba mo’ tune played on the box and used by Puccini might at first seem to make it less likely that Puccini encountered the Guinness box. (The Guinness box tune sheet's ‘Poutzi’ rather than ‘Pantsé’ spelling of the ‘Bangzi’ tune presents a somewhat similar problem.) However, under any scenario, Puccini's understanding of this title or of the song's basic subject matter would have required either the intervention of an informant who knew how to read the Chinese characters or, more likely, an informant who simply remembered being told that the final tune on the box was an erotic folksong about a woman's body. The tune sheet's decoration style might have suggested to Puccini that the box was of European manufacture or, conversely (given the long history of the Chinese export market), that this was a Chinese imitation of European design.

93 I am grateful to Renato Meucci and Hendrik Strengers, who both suggested that I consult music establishment address books covering this period to determine when this particular shop existed. According to Paul de Wit's Weltadressbuch der gesammten Musikinstrumenten-Industrie (Leipzig, 1890; 1903; 1906; 1926), Augusto Bianchetti's establishment existed at this address as early as 1903. By 1926 the shop's name was ‘A. Bianchetti e figli’, which suggests that the Guinness box was serviced at this location at least before 1926 but possibly as early as 1903. I am also indebted to Patrizio Barbieri, Fabrizio Della Seta, Federica Riva and Emanuele Senici for crucial assistance as I sought information about the Bianchetti shop.

94 Arthur Cunliffe, president and registrar of the Musical Box Society of Great Britain, reported to me that in the current registry of over 10,000 musical boxes compiled over the past 30 years, only 13 boxes are listed as playing Chinese melodies. This gives a rough indication of how few musical boxes have returned to the West from China. Philippe Rouillé of the French Association des Amis des Instruments de Musique Mécanique informed me that he saw many of these musical boxes rusting away in warehouses in China in the 1970s.

95 A colour version of Figure 5 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

96 For one famous example of Puccini's drawing style, see his self-caricature found on a sketch for La bohème held in the Morgan Library and Museum and available at <https://www.themorgan.org/collections/collections.asp?id=186> (accessed 18 September 2014).

97 This copy of the 1904 vocal score (Ricordi 110000) has been digitized and is available at <http://hdl.handle.net/1802/21913> (accessed 18 September 2014). Given that this score at Eastman is readily available online, I will provide in my notes below page numbers from it that offer many additional examples of similar markings to supplement the Figures. I am grateful to David Peter Coppen at the Sibley Library for providing me with timely access to this score.

98 At the Morgan Library and Museum, Puccini's manuscripts are held in the Robert Owen Lehman, Mary Flagler Cary and Heineman collections as well as in the general music manuscripts collection. At the Beinecke Library, Yale University, Puccini's manuscripts are found in the Frederick R. Koch Collection, GEN MSS 601. The Puccini sources discussed by Allan Atlas in his ‘Newly Discovered Sketches for Puccini's Turandot at the Pierpont Morgan Library’, Cambridge Opera Journal, 3 (1991), 173–93, are now held at the Beinecke. The Bologna manuscripts are reproduced in Groos, ‘The Sketches at the Accademia Filarmonica in Bologna’.

99 A colour version of Figure 6 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

100 Colour versions of Figures 7 and 8 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

101 Also see the image reproduced in Groos, ‘The Sketches at the Accademia Filarmonica in Bologna’, 157.

102 See the Eastman score, pp. 7, 110, 112, 115, 156, 164, 200, 204, 211, 216, 221, 337, 342, 378–9, 380, 382 and 391–2. Puccini's numeral 4 is also found in the image reproduced in Groos, ‘The Sketches at the Accademia Filarmonica in Bologna’, 149.

103 See, for example, the Eastman score, pp. 266, 318, 342 and 377.

104 See, for example, the Eastman score, pp. 75, 112, 194, 262, 277 and 392.

105 See ibid., pp. 95, 110, 246 and 373.

106 See, for example, Groos, ‘The Sketches at the Accademia Filarmonica in Bologna’, 153.

107 I am indebted to Giuseppe Tasca, nephew of the late opera director Alberto Fassini, who was a grandson of Puccini's friend Baron Alberto Fassini, for providing me with these letters and for offering crucial information on his family's history in Palermo. This correspondence appears to be all from 1916 and concerns Puccini having been filmed by Alberto Fassini's company and, perhaps, the sale of a car by Puccini to Alberto. I am also grateful to Joseph Franconi Lee for putting me in contact with Giuseppe Tasca, and to Dieter Schickling for alerting me to the existence of another letter, dated 21 February 1916, from Puccini to Alberto Fassini.

108 On Alberto Fassini, see, for example, Francesco M. Biscione's article ‘Fassini, Alberto’ in the Dizionario biografico degli italiani, <http://www.treccani.it/enciclopedia/alberto-fassini_%28Dizionario_Biografico%29> (accessed 22 November 2014). In 1938, Alberto Fassini authored a 15-page pamphlet (held at the Library of Congress) on tourism in the African colonies entitled ‘Ospitalità e turismo nell’ Africa orientale italiana’. On Alberto Fassini's work as vice president of the Corporazione dell'Ospitaliti in the Mussolini government, see Brian L. McLaren, ‘Mediterraneità and Modernità: Architecture and Culture during the Period of Italian Colonization of North Africa’ (Ph.D., Massachusetts Institute of Technology, 2001), 320–1.

109 I am grateful to Dieter Schickling for providing me with a copy of this 8 November 1906 autograph portrait. I have determined that ‘Baronessa Stena Fassini’ – the dedicatee – was Alberto Fassini's wife. (Stefania Spadafora was her maiden name, and ‘Stena’ was her nickname. According to Giuseppe Tasca, she met Alberto in Palermo in 1904.)

110 Puccini could have met up with Alberto Fassini at any one of a number of the baron's homes during the period in which he composed Madama Butterfly. Puccini was in Palermo in 1901 and Alberto had a famous villa designed by Ernesto Basile built there in 1903, but it is not clear whether Alberto was in Palermo at the time of Puccini's visit. Puccini was in Rome in 1902 and the Fassinis owned property there. We know that Puccini finished a draft of Act 1 of the opera on 7 September 1902, and so his encounter with a music box must have taken place well before that date, and most likely before May of that year. For detailed chronologies of Puccini's travels, see Dieter Schickling, Giacomo Puccini: Catalogue of the Works, trans. Michael Kaye (Kassel, 2003); and Schickling, Giacomo Puccini: Biografie (Stuttgart, 2007).

111 On this general topic, see in particular Ysia Tchen, La musique chinoise en France au XVIIIe siècle (Paris, 1974). On Amiot, see Lam Ching Wah, ‘A Highlight of French Jesuit Scholarship in China: Jean-Joseph-Marie Amiot's Writings on Chinese Music’, CHIME Journal, 16–17 (2005), 127–47; and François Picard, ‘Joseph-Marie Amiot, jésuite française à Pékin, et le cabinet de curiosités de Bertin’, Musique, images, instruments: Revue française d'organologie et d'iconographie musicale, 8 (2006), 68–85. Also see Kii-Ming Lo, ‘In Search for a Chinese Melody: Tracing the Sources of Weber's Musik zu Turandot, op. 37’, Tradition and its Future in Music: Report of SIMS 1990 Osaka (Tokyo, 1991), 511–21, and eadem, ‘New Documents on the Encounter between European and Chinese Music’, Revista de musicología, 16 (1993), 16–31. For extracts from and discussion of European writings on Chinese music from this period, see Frank Harrison, Time, Place and Music: An Anthology of Ethnomusicological Observation c.1550 to c.1800 (Amsterdam, 1973), 161–94 and 207–18, and idem, ‘Observation, Elucidation, Utilization: Western Attitudes to Eastern Musics, ca.1600-ca.1830’, Slavonic and Western Music: Essays for Gerald Abraham, ed. Malcolm Hamrick Brown and Roland John Wiley (Ann Arbor, MI, 1985), 5–30. On English reception of Chinese music in this period, see David Clarke, ‘An Encounter with Chinese Music in Mid-18th-Century London’, Early Music, 38 (2010), 543–57.

112 John Barrow, Travels in China (1st American edn, Philadelphia, PA, 1805), 209–10, 216. Barrow's transcription and translation of ‘Moo-lee-wha’ appears on pp. 211–12. Barrow referred to the melody as ‘one of the most popular songs in the whole country’ (p. 210).

113 J. A. van Aalst, Chinese Music (Shanghai, 1884). Also see the contemporaneous publication in English, China: Imperial Maritime Customs: Illustrated Catalogue of the Chinese Collection of Exhibits for the International Health Exhibition, London, 1884 (London, 1884), which provides transcriptions of multiple songs and has a chapter (Chapter 26) devoted to Chinese music. On van Aalst, see Han Kuo-huang, ‘J. A. van Aalst and his Chinese Music’, Asian Music, 19 (1988), 127–30.

114 On the importance of van Aalst's publication for Turandot, see Ashbrook and Powers, Puccini's Turandot, 96–7, and for a table detailing the appearances in the opera of all the melodies Puccini borrowed from van Aalst and from the music box, see Kii-Ming Lo, ‘Turandot’ auf der Opernbühne, 334.

115 For French sources, see particularly Louis Laloy, La musique chinoise (Paris, 1910), 122–6, and Georges Soulié, La musique en Chine (Paris, 1911). The notated music in Judith Gautier's Les musiques bizarres à l'Exposition de 1900: La musique chinoise (Paris, 1900), offers a rather bizarre representation of, or musical response to, Chinese music.

116 I should note that a genealogical website created by a distant descendant gives Bovet's dates as 1825–1913 and states that he was born and died in London. See <http://hibbertfamily.org/html/connections/frederick%20bovet%201.htm> (accessed 22 November 2014). I am grateful to Rosemary Haden for offering information on the Bovet family and to Barbara Galimberti, Martin Wyss and Karla Vanraepenbusch, who offered assistance in tracking down information on Fritz Bovet. Although Puccini was in London in 1900, which is where he saw David Belasco's production of Madame Butterfly in June, I have found no evidence suggesting that he encountered Fritz Bovet.

117 The Bovets were from Fleurier in the Neuchâtel canton – a centre of Swiss watch-making and music-box manufacture. On the marketing of Bovet watches in China, see Alfred Chapuis, La montre ‘chinoise’ (Neuchâtel, 1919), Chapter 6, and Sylvia Robert, ‘Edouard Bovet dit-de-Chine: Négociant en horlogerie (1797–1849)’, Biographies neuchâteloises, ed. Michel Schlup, 5 vols. (Hauterive, 1996–2008), ii: Des lumières à la révolution (1998), 49–55.

118 Music boxes played a major role in disseminating music in the nineteenth century, with a particular emphasis on opera melodies. On this topic, see, for example, Denise Gallo, ‘Verdi's Music on Mechanical Boxes’, Verdi Forum, 28/29 (2001–2), 4–7.

119 On Fritz Bovet's role in the creation of Swiss music boxes playing Chinese melodies, see Chapuis, La montre ‘chinoise’, 135–6. Chapuis based his discussion on Bovet family correspondence (now apparently lost) and on his interviews with Swiss watch-making families. Chapuis reports that a large number of music boxes were exported to China (pp. 221–2) and that E. Paillard et C^ie of Sainte-Croix was particularly active in this trade (p. 223). Chapuis lists several of the tune titles that appear on these music boxes, mistakenly splitting the title ‘Loc-tee Kuntzin’ into two separate titles (p. 224). Also see Chapuis, History of the Musical Box and of Mechanical Music, trans. Roesch, 269, or the original French edition, Histoire de la boîte à musique et de la musique mécanique (Lausanne, 1955), 282. In this landmark volume, Chapuis reproduces the notation employed for pinning ‘Sinfa’ which was given to him by MM. Gueissaz (p. 265; p. 277 in the French edition). For brief references to Fritz Bovet, also see Lindsay and May Ride, An East India Company Cemetery: Protestant Burials in Macao (Hong Kong, 1996), 13, 187. For information on Bovet's time in China during the Opium Wars, see Auguste Jeanneret-Oehl, Souvenirs du séjour d'un horloger neuchâtelois en Chine (Neuchâtel, 1866). For references to Bovet's service as vice-consul for France in Canton in 1856, also see Henri Cordier, L'expédition de Chine de 1857–58 (Paris, 1905), 55–6, and Papers Relating to the Proceedings of Her Majesty's Naval Forces at Canton (London, 1857), 75–8. I would like to thank Suzanne Aspden for providing me with a copy of Bovet's published string quartet held at the Bodleian Library, Oxford.

120 J. Dyer Ball, Things Chinese, or Notes Connected with China (4th edn, New York, 1904), 725.

121 I thank Nicholas Simons of the Music Box Society of Great Britain and Rosanna Harris of the Musical Box Society International for their assistance. Steve Ryder offered his wealth of knowledge on music-box history and technology as well as crucial guidance. Timothy Reed has brought to my attention the existence of several music boxes and sources and I am indebted to him. I am particularly grateful to the late Philippe Rouillé, who sent me copies of tune sheets and recordings of several music boxes that he owned. I am also grateful to Donald Day and Beatrice Farmer for sending recordings of their music boxes and to the staff at the Musical Wonder House in Wiscasset, Maine. For reports on specific surviving boxes that play Chinese melodies, see John Turner, ‘L'Epée: The Chinese Musical Box’, The Music Box, 20/3 (autumn 2001), 87, and Alec Reid, ‘A Pre-Telly Vision Box’, Mechanical Music (May/June 2012), 56. I am grateful to Mr Reid for providing me with a recording of this ten-tune cylinder box, which includes ‘Sinfa’, ‘Bangzi’, ‘Shiba mo’ and ‘Loc Tee Kun Tzin’, as well as dancing dolls. The Mira & Stella Company continued to manufacture boxes featuring these tunes, as is evident in lists available at <http://antiquemusicbox.us/mirastella.pdf> (accessed 18 September 2014) and in The Music Box, 9/5 (spring 1980), 235. The appearance of Bovet's set of Chinese melodies on later Thorens disc music boxes is evident in several surviving examples and in published lists, such as that provided at <http://www.aerialphotolab.com/Thorens_Discs.htm> (accessed 18 September 2014).

122 Associating the mechanical with the exotic was a common trope in the early twentieth century. On representations of mechanical music in Stravinsky's Le rossignol (1908–14), for example, see Daniel Albright, Stravinsky: The Music Box and the Nightingale (New York, 1989), Chapter 4. In this tale, exotic Japanese ambassadors bring a mechanical nightingale, which then displaces the Emperor's natural, living bird.

123 Alexandra Wilson, ‘Modernism and the Machine Woman in Puccini's Turandot’, Music and Letters, 86 (2005), 432–51. Arman Schwartz, ‘Mechanism and Tradition in Puccini's Turandot’, Opera Quarterly, 25 (2010), 28–50.

124 Schwartz, ‘Mechanism and Tradition’, 29. Schwartz is less convincing in suggesting that Turandot herself resembles a music box in her first appearance (pp. 41–3).

125 Schwartz notes that Ping, Pang and Pong are associated with a mechanical style, but incorrectly states that Puccini employed only four bars from the Fassini box for their initial music (ibid., 37). Schwartz makes a similar mistake – while offering perceptive comments on the orchestration – in referring to Puccini's use of a ‘four-bar pentatonic fragment’ of ‘Mo Li Hua’ (p. 45). Puccini scholars have long known that the composer used both tunes from the Fassini box in their entirety, and both tunes are quite a bit longer than four bars in duration.

126 A sound clip for Example 9 may be accessed at <http://dx.doi.org/10.1080/02690403.2015.1008863>.

127 A waltz occupies the lengthy middle section of the work and ‘Loc Tee Kun Tzin’ abruptly returns for a final appearance in the concluding three bars. This piece has been published in an edition by Paul Wehage (Lagny sur Marne, 2008). Ali Ben Sou Alle also composed Shanghai Redowa Waltz for soprano saxophone and piano.

128 For an overview of the position of ‘Mo Li Hua’ in China, its transmission to Europe and its appearance in Tan Dun's Symphony 1997, see Christian Utz, ‘Ein Feld “entorteter” Identitäten: Essentialismus und Differenz in der neuen Musik Chinas und Japans’, Reflexionen der kulturellen Globalisierung, ed. Ute Hoffmann (Berlin, 2003), 105–25. On Tan Dun's relationship to Puccini, also see my ‘Blurring the Boundaries: Tan Dun's Tinte and The First Emperor’, Journal of Musicology, 26 (2009), 285–326. For a critical assessment of Tan Dun's use of ‘Mo Li Hua’ in his work, see Yu Siu Wah, ‘Two Practices Confused in One Composition: Tan Dun's Symphony 1997: Heaven, Earth, Man’, Locating East Asia in Western Art Music, ed. Yayoi Uno Everett and Frederick Lau (Middletown, CT, 2004), 57–71.

129 Quoted in Cui Xiaohuo, ‘Classical Piece Will Ring in Ears of Winners’, China Daily (6 August 2008), available at <http://www.chinadaily.com.cn/olympics/2008-08/06/content_6907444.htm> (accessed 27 March 2014).

130 See Oliver Chou, ‘From Nation's Favourite to Populist Poison’, South China Morning Post (6 March 2011), available at <http://www.scmp.com/article/740037/nations-favourite-populist-poison> (accessed 1 April 2014). Chou reports that the government's negative view of the song started in December 2010 when Lynn Chang performed it at the Nobel Peace Prize ceremony in tribute to the jailed dissident Liu Xiabo.