¹ Homer, Sidney, My Wife and I (New York: Macmillan, 1939), 191 Google Scholar.

² Freitas, Roger, ‘Towards a Verdian Ideal of Singing: Emancipation from Modern Orthodoxy’, Journal of the Royal Musical Association, 127 (2002), 226–57 (p. 235)CrossRefGoogle Scholar.

³ Will Crutchfield, ‘Vocal Ornamentation in Verdi: The Phonographic Evidence’, 19th-Century Music, 7 (1983–4), 3–54. Crutchfield paid particular attention to the topic of vocal ornamentation and demonstrated its survival into the performance practice of the early twentieth century. In his study he relied on early recordings as an essential medium for the reconstruction of past performance practices.

⁴ The idea that each register of the vocal compass yielded a different colour is confidently asserted in more than 70 volumes of vocal pedagogy written between the 1840s and 1930s that I surveyed, including vocal treatises, singing methods and exercises of daily practice. Vocal pedagogues’ traditional advocacy of a ‘consistent’ transition between the registers should not be confused with the idea of ‘timbral consistency’; these are two separate notions which coexisted in the pedagogical discourse well into the twentieth century.

⁵ Rebecca Mara Plack complains about the endemic lack of agreement on even the most essential terms of such a specialized lexicon, which compelled her to build a tailor-made vocabulary in order to discuss stylistic gestures in the vocal recordings that she examined. See Plack, ‘The Substance of Style: How Singing Creates Sound in Lieder Recordings, 1902–1939’ (Ph.D. dissertation, Cornell University, 2008), 16.

⁶ ‘Voce omegenea, timbratura squillante, piena, spontanea’. La perseveranza, November 1897, quoted in Pietro Gargano and Gianni Cesarini (with Michael Aspinall), Caruso (Milan: Longanesi, 1990), 42.

⁷ ‘Meravigliosa voce, spontanea, che ha tutte le pastosità, tutta la gamma in suo potere’. Il mondo artistico, 34/48 (1900), 11. On this occasion, Caruso was singing Cavaradossi at the Teatro Sociale in Treviso.

⁸ ‘Del personaggio di Mario Cavaradossi egli fa una creazione del tutto nuova; di ciò si sono convinti coloro i quali assistettero as esecuzioni di Tosca, con altri tenori, come il Borgatti e il Giraud.’ Ibid., reporting a review first published in Il gazzettino.

⁹ It was owing to Borgatti’s indisposition that Caruso made his début at La Scala earlier than anticipated, on 26 December 1900. On the opening night of that season, Borgatti was scheduled to sing Tristan in Wagner’s Tristan und Isolde. With Borgatti indisposed, the general director of La Scala, Giulio Gatti-Casazza, decided to replace Wagner’s opera with Puccini’s La bohème, in which Caruso was cast as Rodolfo.

¹⁰ ‘La declamazione, troppo sdolcinata per essere soave, di De Lucia; esagerata in leziosaggini, abusante dei falsi, erompente con stacco notevole negli effetti dei medii; ha ricordato il secondo Osaka, il Borgatti, al quale si adattava così poco la parte, per la natura del suo organo vocale, tetragono agli effetti falsi, della voce mista.’ Italia, November 1899, quoted by Aspinall in Gargano and Cesarini, Caruso, 241.

¹¹ Both Guardabassi and Galli trained first as baritones before transitioning into dramatic tenor roles. They made only a few recordings, among which Guardabassi’s Siciliana was recorded for a minor label, the Lyrophon Record (Matrix J 403), while Galli’s Miserere was a Gramophone & Typewriter Company (G&T) disc, Cat. 52408 (1902). Martinez-Patti’s substantial recording legacy deserves much more attention than it has actually received; his recording of the Siciliana mentioned above was made for Pathé, Matrix 84500, Cat. 10156 (1908).

¹² Gargano and Cesarini, Caruso, 44.

¹³ See Frederick William Gaisberg, The Music Goes Round (New York: New York Times Company, 1977), 50.

¹⁴ The first ten tracks made by Caruso, which were recorded by Frederick William Gaisberg in April 1902, have been transferred by Ward Marston for Naxos Historical CD 8.110703. Later in this article some of the relevant passages from ‘Una furtiva lagrima’ will be considered together with the associated audio excerpts.

¹⁵ Melba, Nellie, Melodies and Memories (London: Thornton Butterworth, 1925), 130 Google Scholar. To be noticed again are the adjectives ‘spontaneous’ and ‘natural’, used by Melba in describing Caruso’s singing.

¹⁶ Because of our different historical perspective, when we listen to the early recordings of Caruso the characteristics of ‘spontaneous’, ‘natural’ and ‘modern’ singing which were immediately and consistently recognized by his contemporaries remain rather elusive to our ears.

¹⁷ ‘In the remainder of the opera he has been delightful. All the fioriture, mezze voci, falsettos […] progressively became full, ringing, powerful sounds’ (‘Nel resto dell’opera [Borgatti] ha deliziato. Tutte le fioriture, le mezze voci, i falsetti […] diventavano grado a grado note piene, squillanti, potenti’). Nino Creso, in Il mondo artistico, 34/48 (1900), 7. About mezza voce, James Stark underlines that this technique ‘is [generally] taken for granted’ in both pedagogical literature and voice science, and is ‘not explained’. Stark suggests, on the basis of his own interpretation of Manuel García II’s definition, that mezza voce can be obtained by ‘disengaging the vocal tract to make it as passive as possible, thereby reducing the resonances of the vocal tract, while keeping airflow rates low’. Stark, Bel canto: A History of Vocal Pedagogy (Toronto: University of Toronto Press, 1999), 247.

¹⁸ In 1905, Borgatti recorded a number of titles for Fonotipia, including ‘E lucevan le stelle’ from Puccini’s Tosca, Matrix XPh 1516, Cat. 39406, and (in Italian) ‘Deh non t’incantan’ from Wagner’s Lohengrin, Matrix XPh 1517, Cat. 39407. In 1900, Cesarani (c.1860–?) recorded Puccini’s solo, together with 14 other tracks, for Berliner Records in Milan, Matrix 2883, Cat. 52585, while the recording by Giraud (1868–1928) of the aria from Carmen, sung in Italian as ‘Il fior’, was made for G&T, Matrix 272i, Cat. 052069.

¹⁹ For contemporary observations regarding the naturalness of Adelina Patti’s singing, see Roger Freitas, ‘The Art of Artlessness, or, Adelina Patti Teaches Us How to Be Natural’, Word, Image, and Song, ed. Rebecca Cypess, Beth L. Glixon and Nathan Link, 2 vols. (Rochester, NY: Boydell & Brewer, 2013), 213–42.

²⁰ New York Times, 24 November 1903.

²¹ The Sun (New York), 24 November 1903. Note the reference to Caruso’s ‘natural and free delivery’.

²² Pre-performance essay by Richard Aldrich on the Met’s production of Bellini’s La sonnambula, New York Times, 10 December 1905.

²³ This date range can be estimated by considering the beginning of his active stage career together with the chronological analysis of his recorded legacy. Post-1906 recordings demonstrate that Caruso was able consistently to control heavier systems of voice production and display a departure from his previous practice of producing some of his top notes in a sort of reinforced falsetto. This observation also aligns with Michael Aspinall’s theory of a change in breathing techniques in the period 1906–7, namely, from intercostal respiration to abdominal breathing – the latter favouring more deeply set laryngeal postures. See Aspinall in Gargano and Cesarini, Caruso, 247.

²⁴ A distinction between male and female voices, with a two-register theory for male and a three-register theory for female voices, was generally accepted. Brent J. Monahan states that this latter division had ‘greater acceptance during the nineteenth century’ as well as surviving well into the twentieth century; see Monahan, The Art of Singing (Metuchen, NJ, and London: Scarecrow Press, 1978), 146. The terminology adopted for register classification was admittedly confusing, with voice teachers making their personal choices of lexicon. For instance, what Luigi Lablache (1842) defined as the ‘head’ register is described as the ‘mixed’ register in Lamperti’s treatise (1877), while both William Shakespeare (1898) and Mathilde Marchesi (1896; she, however, dealt exclusively with female voices) used the term ‘medium’. Beniamino Carelli (1875) would refer to the secondo registro (‘second register’) and voce mista (‘mixed voice’) in a confusing way, and García (1841) preferred the term ‘falsetto’ – for which he was criticized by his pupil Marchesi.

²⁵ Francesco Lamperti, A Treatise on the Art of Singing, trans. J. C. Griffith (Milan, Naples, Rome, Florence and London: Ricordi, 1877), 17. This belief is unanimously shared among the voice teachers of the period, including Virginia Boccabadati in Osservazioni pratiche per lo studio del canto, 3rd edn (Pesaro: Federici, 1893), 10–11; Beniamino Carelli in Cronaca d’un respiro (Naples: Tipografia dell’Ariosto, 1875), 22, and L’arte del canto: Metodo teorico-pratico, 7 vols. (Milan: Ricordi, 1905–9), i, 52; Alessandro Guagni-Benvenuti in L’odierna scuola di canto in Italia (Rome: Tipografia Metastasio, 1886), 43–4; Mathilde Marchesi in The Marchesi School: Theoretical and Practical Vocal Method (London: Enoch & Sons, 1896), p. iv; and William Shakespeare in The Art of Singing (London: Metzler & Co., 1898), 36–46.

²⁶ Manuel García, ‘Extract from the Mémoire on the Human Voice’, in A Complete Treatise on the Art of Singing: First Part, complete and unabridged edns 1841 and 1872, ed. and trans. Donald V. Paschke (New York: Da Capo Press, 1984), p. xli (emphasis added).

²⁷ David C. Taylor, The Psychology of Singing: A Rational Method of Voice Culture Based on a Scientific Analysis of All Systems, Ancient and Modern (New York: Macmillan, 1908), 16–17.

²⁸ ‘Osò tentare di far dell’arte del canto una scienza’. Carelli, Cronaca d’un respiro, 33, where he also attributes to the physiologist Segon the statement that ‘on the anatomical and physiological knowledge of the vocal organs, García founded the mechanics of singing’. In modern times, the impact of García’s anatomy-based studies on the history of vocal pedagogy is revealed in the different ways in which contemporary writers have interpreted his insights. Nevertheless, a connection with Taylor’s idea of a ‘mechanical turn’ is a persistent view, even if the writer does not state it openly. On this topic, see David Mason, ‘The Teaching (and Learning) of Singing’, The Cambridge Companion to Singing, ed. John Potter (Cambridge: Cambridge University Press, 2000), 204–20.

²⁹ Contemporary research in the field of voice science supports these ideas. Johan Sundberg quotes Harry Hollien, ‘On Vocal Registers’, Journal of Phonetics, 2 (1974), 125–43, who defines register as ‘a totally laryngeal event; it consists of a series or a range of consecutive voice frequencies which can be produced with nearly identical phonatory quality’. Sundberg, The Science of the Singing Voice (Dekalb: Northern University Press, 1987), 49–50.

³⁰ See Brian White, Singing Techniques and Vocal Pedagogy (New York and London: Garland Publishing, 1989), 59. Meanwhile, Sundberg points out, ‘Voice science has not developed to the point of any complete understanding of the glottal mechanism and its voice quality aspects.’ Sundberg, The Science of the Singing Voice, 50.

³¹ Sundberg, The Science of the Singing Voice, 51. The lengthening of the folds is managed by the cricothyroid muscles, which in contracting provoke the backward and upward tilting of the cricoid cartilage. This movement, given the position of the vocal folds within the larynx, increases the distance between the the posterior and anterior attachments of the folds and provokes their stretching and lengthening. For a detailed description of the lengthening and shortening of the vocal folds, see ibid., 15–18.

³² See Minoru Hirano, William Vennard and John Ohala, ‘Regulation of Register, Pitch and Intensity of Voice’, Folia phoniatrica et logopaedica, 1 (1970), 1–20.

³³ Sundberg, The Science of the Singing Voice, 53.

³⁴ See Natalie Henrich Bernardoni, ‘Mirroring the Voice from García to the Present Day: Some Insights into Singing Voice Registers’, Logopedics Phoniatrics Vocology, 31 (2006), 3–14 (p. 9). Following current linguistic usage, Henrich Bernardoni labels these registers as M1 and M2, terms which I avoid for reasons of consistency as I put late nineteenth-century vocal pedagogy in dialogue with early recordings.

³⁵ García, Manuel, Hints on Singing, 2nd edn (London: Ascherberg, Hogwood & Crew, 1911), 16–17 Google Scholar.

³⁶ García, A Complete Treatise on the Art of Singing: First Part, ed. and trans. Paschke, 28.

³⁷ I am not asserting that this idea is universally accepted by singing teachers, who actually follow a number of different approaches, but it is nevertheless widely shared among those who belong to the English-speaking world.

³⁸ As Sarah Potter points out, ‘García’s published works as a body do not suggest continual larynx-lowering.’ See Potter, ‘Changing Vocal Style and Technique in Britain during the Long Nineteenth Century’ (Ph.D. dissertation, University of Leeds, 2014), 35. The use of the dark timbre is not intended to be continuous.

³⁹ See Carelli, Cronaca d’un respiro, 46–8. Carelli also provides the reader with drawings of the vocal tract’s different shapes.

⁴⁰ Lamperti, A Treatise on the Art of Singing, 17; Shakespeare, The Art of Singing, 36–46; Carelli, Cronaca d’un respiro, 22; Guagni-Benvenuti, L’odierna scuola di canto, 43–4.

⁴¹ Lena Doria Devine, ‘Francesco Lamperti and his Methods’, The Etude, 26 (1908), 259.

⁴² The ability of Caruso to handle elements of the older singing style is debatable, as his early recordings are already littered with his attempts towards shaping a new type of vocalism, as we have seen. Reviews of his 1890s performances, however, praise the suppleness of his singing and the light vocal weight of his upper range, supporting the idea that he could have carried on with the old school of singing had he wished to do so.

⁴³ This recording was made for the Gramophone & Typewriter Company (G&T), with Salvatore Cottone at the piano, Matrix C-18822-3, Cat. 88580 (1902).

⁴⁴ A sound clip of the relevant passage from this recording may be accessed in the Supplemental Material online at <sound clip 1>.

⁴⁵ A sound clip of the relevant passage from this recording may be accessed in the Supplemental Material online at <sound clip 2>.

⁴⁶ This is a Victor recording, Matrix B 998, Cat. 81028 (1904).

⁴⁷ A sound clip of the relevant passage from this recording may be accessed in the Supplemental Material online at <sound clip 3>.

⁴⁸ Sound clips of the two relevant passages from this recording may be accessed in the Supplemental Material online at <sound clip 4> and <sound clip 5> respectively.

⁴⁹ As the discussion above regarding the various possible shapes of the vocal tract demonstrates, different laryngeal heights and configurations of the soft palate were known and formed an essential part of late nineteenth-century singers’ training.

⁵⁰ This recording was made for Victor, with orchestra, Matrix A 8364, Cat. 7-52002 (1909).

⁵¹ A sound clip of the relevant passage from this recording may be accessed in the Supplemental Material online at <sound clip 6>.

⁵² This phenomenon is known as ‘the singer’s formant’, which, as Sundberg explains, ‘is produced by a clustering of the third, fourth and fifth formant frequencies. An important articulatory means for achieving such a clustering is the shape of the larynx tube and a wide pharynx, often produced by a lowering of the larynx.’ See Johan Sundberg, ‘Where Does the Sound Come From?’, The Cambridge Companion to Singing, ed. Potter, 231–47 (p. 242). Also to be noted: in the second ascent to A4, Caruso applies vowel modification on the second vowel of the word ‘amato’, which is sung on the G4 preceding the top note. This procedure is also audible in the 1904 recording, though it comes across in a much milder manner.

⁵³ Sound clips of the two relevant passages from this recording may be accessed in the Supplemental Material online at <sound clip 7> and <sound clip 8> respectively.

⁵⁴ If the issue of the passaggio to the upper notes was resolved in the period 1895–1906, the recording evidence seems to indicate that experimentations with the various breathing systems continued until much later. As mentioned above (n. 23), Aspinall discusses the evolution of Caruso’s breathing technique, and identifies the years 1906–7 as the turning point towards abdominal respiration. My own chronological listening to Caruso’s recordings led me to think that in the years after 1906, Caruso continuously swapped between different combinations of abdominal and intercostal breathing, thus continually varying the dynamic interaction of the different parts of the vocal mechanism. Although an overall shift towards prevalent abdominal breathing and heavier registration choices is confirmed by the majority of the 1911–21 recordings, many sections of solos and ensembles are still managed within the intercostal system.

⁵⁵ New York Times, 10 December 1905.

⁵⁶ The Sun (New York), 21 November 1905.

⁵⁷ It should, according to Sansone, be applied only to the subcategory of the ‘plebeian tragedy’ first exemplified by Cavalleria rusticana. Matteo Sansone, ‘Verga and Mascagni’, Music and Letters, 71 (1990), 198–214 (p. 198).

⁵⁸ Nicolaisen, Jay, Italian Opera in Transition 1871–1893 (Ann Arbor, MI: UMI Research Press, 1980), 243–4Google Scholar.

⁵⁹ Carl Dahlhaus, Nineteenth-Century Music, trans. Bradford Robinson, California Studies in Nineteenth-Century Music, 5 (Berkeley, CA: University of California Press, 1989), 351. For analogous views, see David Kimbell, Italian Opera (Cambridge: Cambridge University Press, 1991), 621; Adriana Guarneri Corazzol, ‘Opera and Verismo: Regressive Points of View and the Artifice of Alienation’, trans. Roger Parker, Cambridge Opera Journal, 5 (1993), 39–53; Kimbell, ‘Opera since 1800’, The Cambridge History of Italian Literature, ed. Peter Brand and Lino Pertile (Cambridge: Cambridge University Press, 1996), 450–6; and Wagner, Hans-Joachim, Fremde Welten: Die Oper des italienischen Verismo (Stuttgart: Metzler, 1999), 7 CrossRefGoogle Scholar.

⁶⁰ Giger, Andreas, ‘ Verismo: Origin, Corruption, and Redemption of an Operatic Term’, Journal of the American Musicological Society, 60 (2007), 271–315 CrossRefGoogle Scholar.

⁶¹ On the concept of the ‘concrete’ observation of reality, see Jules Champfleury, Le réalisme (Paris: Michel Levy frères, 1857). Another criterion listed by Champfleury was the inclusion of areas previously excluded from art on the ground that the representation of their subjects was considered artistically unsuitable and broke classicist rules of stylization by introducing stylistic mixture through the depiction of such subjects. On the influence of French naturalism on verismo, see Spinazzola, Vittorio, Verismo e positivismo (Milan: Arcipelago, 1993)Google Scholar, and Gorizio Viti, Verga verista (Florence: Le Monnier, 1994).

⁶² An idea of the number of French operas performed on the stages of Italian opera houses in the 1870s is offered by the table in Daniele Pistone, Nineteenth-Century Italian Opera from Rossini to Puccini, trans. E. Thomas Glasow (Portland, OR: Amadeus Press, 1995), 92–3. Alan Mallach offers a statistic relating to the presence of French operatic repertoire in the economically active but culturally conservative city of Livorno over the last four decades of the nineteenth century. Mallach, , The Autumn of Italian Opera (Boston, MA: Northeastern University Press, 2007), 15 Google Scholar.

⁶³ Carl Dahlhaus, Realism in Nineteenth-Century Music, trans. Mary Whittall (Cambridge: Cambridge University Press, 1985), 70. This was an element underlined by the critic Amintore Galli following the Milanese première of the opera, in relation to which he stated, ‘Mascagni has completely renounced the forms of the classic recitative: he keeps with a melodic and dramatic declamatory style, which has a strong and profound effect on the audience’ (‘Il Mascagni ha interamente rinunciato alle forme del recitativo classico: egli si attiene ad una declamazione melodica e drammatica che ha forte e profonda presa sull’animo del pubblico’). Galli, in Il teatro illustrato e la musica popolare, 121 (1891), 7–9 (p. 7).

⁶⁴ In the writing of the third-act duet between Radames and Amneris in Aida, Verdi at first freed Ghislanzoni from any preoccupations regarding musical form (in order to develop the dramatic situation as well as possible), but later ended up listing the number of different metres he needed in order to create a beautiful melody for Radames: three lines of settenari, a quinario and two endecasillabi. See the excerpts of the correspondence sent by Verdi to Ghislanzoni and quoted in Kimbell, Italian Opera, 548–9.

⁶⁵ ‘L’uggia della cantilena, della simmetria […] della prosodia italiana, che genera quasi inevitabilmente nella frase musicale povertà e grettezza di ritmo’. Il Figaro, 11 February 1864, quoted in Tutti gli scritti di Arrigo Boito, ed. Piero Nardi (Milan: Mondadori, 1942), 1119. The innovative influence of Boito on Italian opera is not limited to the language but extends also to musical dramaturgy. For him, ‘I. the complete obliteration of formula; II. the creation of form; III. the realization of the most immense tonal and rhythmic development possible today; [and] IV. the supreme incarnation of the drama’ (‘I. la completa obliterazione della formula; II. la creazione della forma; III. l’attuazione del più vasto sviluppo tonale e ritmico possibile oggi; IV. La suprema incarnazione del dramma’) were all crucial elements to the renovation of opera. Il Figaro, 21 January 1864, quoted in Tutti gli scritti, 1107–8.

⁶⁶ Julian Budden, Operas of Verdi, 3 vols. (Oxford and New York: Oxford University Press, 1992), iii, 280.

⁶⁷ Lamperti, A Treatise on the Art of Singing, 14.

⁶⁸ ‘Scrivano i maestri come una volta si scriveva, profittando dei progressi della scienza armonica, non per sopraffare l’espressione melodica, ma solo per corredarla di valevoli artifizi.’ Giraldoni, Leone, Compendium: Metodo analitico analitico, filosofico e fisiologico per la educazione della voce (Milan: Ricordi, 1889), 57 Google Scholar.

⁶⁹ ‘La sensualità s’esprime soprattutto nel ‘medium’ della voce, tende ai colori scuri. Di qui, un accorciamento della gamma d’estensione in zona acuta e una certa centralizzazione delle tessiture. Questo fu un fatto generale.’ Rodolfo Celletti, ‘La vocalità mascagnana’, Atti del Primo Convegno Internazionale di Studi su Pietro Mascagni, ed. Fedele D’Amico (Milan: Casa Musicale Sonzogno, 1987), 39–48 (p. 41).

⁷⁰ Ibid., 42. Celletti specifically refers here to the tenor voice in the operas of Mascagni.

⁷¹ ‘Personalmente ho sempre creduto, non soltanto che questo procedimento [il passaggio dal registro centrale a quello acuto] sia stato caro ai veristi […] ma che costituisca il lato propriamente “veristico” del lore stile, insomma che il verismo musicale consiste principalmente in questo procedimento […] Infatti quel poggiare su un settore della voce come quello, che è il meno naturale e il più problematico, richiede uno sforzo, una tensione che inclinano a volgere il canto verso un’espressione “direttamente” passionale, assimilabile all’eccitazione d’un linguaggio parlato.’ Fedele D’Amico from the conference proceedings which report his question/observation following the paper presented by Celletti (see above, n. 69).

⁷² ‘D’una tempra nobile e capace ad un tempo d’un amor tenero, ma degno’. See Panofka, Heinrich, Voci e cantanti (Florence: Cellini, 1871), 89 Google Scholar. More references to the masculine character of spinto tenors occur at pp. 37–8, 83, 91, 97.

⁷³ Zenatello recorded this solo for Fonotipia, Matrix XPH16111, Cat. 39243 (1905), while Caruso’s recording is on Victor, Matrix C-4317, Cat. 6001-A (1907). The age gap between Zenatello (born in 1876) and Caruso (born in 1873) was just three years. The ageing factor, which bears great relevance to the vocal performance, is not therefore considered when discussing these recordings, both made when the singers were in their prime.

⁷⁴ Marconi’s freedom and easiness in reaching the upper notes struck J. H. Duval, while for Max De Schauensee, the high placement of pre-verismo tenors’ upper range was the most distinctive characteristic of Marconi’s singing. See Michael E. Henstock’s liner notes accompanying the CD Symposium 1069, Francesco Marconi, Antonio Cotogni, The Harold Wayne Collection, 2 (1989). By contrast, when Caruso strikes his top B♭4, the listener has the impression that such an explosive mass of sound could not easily be carried higher. A look at the short discography of Tamagno will reveal that the titles more often recorded were from Verdi’s Il trovatore or Otello and Meyerbeer’s Gli Ugonotti or L’Africana, all rigorously sung in Italian.

⁷⁵ Unlike Zenatello, Caruso also subtly recurs to vowel modification on the exclamation ‘Ah’ and in the ‘a’ vowels in ‘Pagliaccio’, on the G4, and ‘amore’, on the A4.

⁷⁶ Sound clips of the relevant passage from these two recordings by Zenatello and Caruso may be accessed in the Supplemental Material online at <sound clip 9> and <sound clip 10> respectively.

⁷⁷ Zenatello possessed a dramatic voice and began his career as a baritone. In 1899, when he was only 23, he performed a tenor role (Canio) at the Teatro del Fondo in Naples. This was not a planned début, as Zenatello was covering for the indisposed tenor Sarcoli in the production in which he had played, until then, the baritone role of Silvio. Celletti, ‘Giovanni Zenatello’, Le grandi voci: Dizionario critico-biografico dei cantanti, ed. Rodolfo Celletti, Luisa Padovani, John B. Richards and Raffaele Vegeto (Rome: Istituto per la Collaborazione Culturale, 1964), 906. In the same season, which according to Tom Hutchinson and Clifford W. Williams was hosted at the Teatro Mercadante and not the Teatro del Fondo, Zenatello sang the baritone roles of Lothario in Thomas’s Mignon, Tonio in Pagliacci and Alfio in Cavalleria rusticana (all of them at the Teatro Mercadante). Two months after his début in a tenor’s role, he sang Manrico in Verdi’s Il trovatore (June, Teatro Mercadante), as well as the title role in Gounod’s Faust and Edgardo in Donizetti’s Lucia di Lammermoor (both in August, at the Teatro Bellini, Naples). After this intense season he decided to take a break of a few months in order to retrain as a tenor under the guidance of Maestro Moratti in Milan. See Hutchinson and Williams, ‘Giovanni Zenatello’, The Record Collector, 14 (1961–4), 100–43.

⁷⁸ Giraud’s ‘Quando le sere al placido’ was a G&T recording, Matrix 273i, Cat. 052070 (1904), which has been transferred by Marston on Nineteenth Century Italian Tenors, CD 53018-2 (2016). Borgatti’s ‘E lucevan le stelle’ was one of the Fonotipias that he recorded in 1905, Matrix XPh 1516, Cat. 39406, which can be heard in the CD transfer by Symposium in The Harold Wayne Collection, 29 (1997). In this aria, Borgatti also displays a strikingly incisive pronunciation which, if not ideal in a cantabile by Puccini, made him an extremely suitable interpreter of Wagnerian roles – witness his Pathé recording of ‘Cede il verno’ from Wagner’s Die Walküre, Matrix 86684 (1916–19).

⁷⁹ While both Michael Scott (in The Great Caruso (London: Hamish Hamilton, 1988), 98–9) and John Potter (in Tenor: History of a Voice (New Haven, CT, and London: Yale University Press, 2009), 94) state that the antagonism was more fictional than real, a 1903 letter from Caruso to the Italian American banker and impresario Pasquale Spinelli seems to suggest that he actually feared and disliked Bonci. See <https://www.schubertiademusic.com/items/details/5205> (accessed 20 May 2019).

⁸⁰ Bonci’s recording is on Columbia, Matrix 36460-1, Cat. A 5449 (1912).

⁸¹ Sound clips of the relevant passages from this recording may be accessed in the Supplemental Material online at <sound clip 11> and <sound clip 12> respectively.

⁸² A sound clip of the relevant passage from this recording may be accessed in the Supplemental Material online at <sound clip 13>.

⁸³ Robert Philip, Early Recordings and Musical Style: Changing Tastes in Instrumental Performance, 1900–1950 (Cambridge: Cambridge University Press, 1992), broke new ground in its re-evaluation of early twentieth-century recordings as an invaluable resource for the history of performance, dealing with essential stylistic gestures such as rubato, portamento, tempo and vibrato. The AHRC-funded project CHARM (Centre for the History and Analysis of Recorded Music, 2004–9) was preoccupied with, among other issues, the analysis of stylistic parameters as audible in early recordings. In terms of the voice specifically, the studies of both Freitas and Crutchfield cited in nn. 2 and 3 above focus on the performance practice of the nineteenth century, relying on the support of the recording evidence among other sources.

⁸⁴ Bel canto was a term retrospectively created in the last decades of the nineteenth century by the vocal pedagogues who trained, among others, singers of Caruso’s generation. With reference to the Italian pedagogical writing that I have studied, the existence of a ‘traditional Italian way’ of teaching singing does clearly emerge. This Italian ‘school’, which proudly claimed its links with a system of teaching developed over the preceding two centuries, was apparently based on widely shared tenets of ‘good’ singing.