62. Edgar

Shortly after the first performance of Le Villi (60), in addition to purchasing the rights to that opera, the publishing house of Ricordi instructed Puccini to write a new opera, to another libretto by Ferdinando Fontana. Ricordi announced this information on 8 June 1884 in their magazine, Gazzetta musicale. Even prior to that, although still deeply involved in the reworking of Le Villi, Fontana had proposed a new subject that summer (see Marchetti 1973, No. 88), perhaps already concerning Edgar.
The subject of Edgar is taken from Alfred de Musset’s “armchair drama [spectacle dans un fauteuil]” entitled La coupe et les lèvres (1832). However, Fontana only utilized several of the dramatically spectacular elements for his libretto and left little of Musset’s romantic and extravagant pre-symbolist writing about a Faustian hero with a longing for redemption.Note: *A comparison of Musset’s model and the libretto is included in Cesari 1994, pp. 325-340; a detailed description of the origins of the opera is also published in Cesari 1993, pp. 459-467, and Cesari 1994, pp. 204-247.Puccini began the composition in March 1885 (see Quaderni 1992 II, No. 6), when the libretto only existed in vague outlines. Not even the name for the eponymous hero had been fixed at that time and, for a long time, other principal characters still had different names.
Right from the start, there were great problems with Fontana’s libretto (see Quaderni 1992 II, No. 6 ff.). Nevertheless, Puccini obviously had already begun with the continuous composition draft of Act 1 (see Carteggi 1958, No. 19).Note: *This letter is certainly from the end of April 1885 and not, as the editor asserts, from November, as it is definitely the answer to Fontana’s letter of 27 April 1885 (= Quaderni 1992 II, No. 10). At a meeting in Lucca at the end of May or the beginning of June 1885, the composer and librettist temporarily eliminated the open questions (see Marchetti 1973, Nos. 98-99), and while Fontana sent further variants of text and plot to Puccini (see Quaderni 1992 II, No. 15 ff.), he probably finished the composition of Act 1 in November, for by the end of that month he was already in the midst of Act 2 (see Quaderni 1992 II, No. 24, and 62.A.II.10.b), the finale of which he had reached by the end of December (see Quaderni 1992 II, No. 27, and Carteggi 1958, No. 22 — certainly of 29 December 1885). In spite of pressure from Fontana as well as Ricordi, Puccini worked remarkably slowly (see Fontana’s further letters in Quaderni 1992 II). That was obviously also due to his private situation, for he had to arrange his life with his pregnant mistress Elvira (their son Antonio was born on 22 December 1886).
Nevertheless, Puccini probably finished the composition at the end of October 1886 (see Carteggi 1958 No. 24). A short time before that he had begun the instrumentation at his Alpine holiday resort of Sant’Antonio d’Adda; the beginning of the autograph full score was dated there on 12 September (see 62.B.1). Obviously however, he interrupted this work in the first months of 1887. Perhaps he was inspired to revise Edgar considerably once again after having attended Verdi’s Otello (first performed on 5 February 1887). On 9 March 1887, Puccini documented the fact that he was just then finishing Act 4 and correcting the entire work (Marchetti 1973, No. 114). But in June there was still some talk about Act 4, while the first two acts were already orchestrated (ibid. No. 117), which — at least for Act 1 — did not mean the definitive version. On 19 July, his brother Michele, who stayed with Giacomo at the summer retreat in Caprino Bergamasco (where Fontana also resided), deemed the composition to be ready and enthusiastically referred to Act 4 as a successful piece of work (ibid. No. 120). In September, Puccini communicated that he was intensively occupied with the instrumentation of Edgar (ibid. No. 122). This concerns Act 3, the autograph full score of which is dated at the beginning “settembre 87” (see 62.B.1), but also not yet in its definitive version.
Contrary to Puccini’s assertion on 27 October, that Edgar was now finished (Carteggi 1958 No. 26), he still worked on the opera for months thereafter. Ricordi’s official statement concerning the completion, published in the Gazzetta musicale of 11 December 1887 (p. 395), included that misinformation. Perhaps this was a consequence of the disastrous performance of Le Villi in Naples on 15 January 1888, an event that deeply affected Puccini and possibly caused him to occupy himself with Edgar once again. In any event, he rewrote the end of Act 1. The last 25 pages of the autograph full score (from fig. 46-3 in the current Ricordi edition) have a separate heading and a new pagination (see 62.B.1), and are dated “4 maggio 88”. At the end of the autograph full score of Act 3 there is the date of 13 July. The last part of it (from the trio, before what is now fig. 33 in the current Ricordi edition) also has a separate pagination and, therefore, probably was rewritten in the weeks before (see 62.B.1), in Puccini’s new summer vacation residence, the Swiss Vacallo above Chiasso (a part of which is also the little village of S. Simone mentioned in the autograph). But on 11 August, Giulio Ricordi complained that the opera was not yet completed (MRC  1888/89 - 5/173). In any case, Act 4 still was missing. It was only on 9 November 1888 that Puccini delivered the full score to the publisher (MRC 1888/89 - 10/47).
Furthermore, immediately before the first performance, Puccini once again made other changes, some particularly due to the replacement of the originally designated singer of Tigrana by Romilda Pantaleoni (Verdi’s first Desdemona), who was a soprano and not a mezzo-soprano (see Quaderni 1992, p. 105). For her, Puccini is also said to have written an additional solo in Act 3, “Pensai la mia bellezza” (see the Gazzetta musicale of 31 March 1889, p. 206). Indeed, Tigrana’s aria in 62.B.1 represents a separate complex (leaves 69–89), possibly written later (compare Quaderni 1992, p. 107, footnote 5). Even during the performances there was some experimentation: the prelude to Act 3 was omitted on the first and second nights and only played in the third and last performance.Note: *Contrary to statements in Companion 1994, p. 106, it is true that the Preludio already dated from autumn 1887, as is clearly shown in 62.B.1.
The first performance on 21 April 1889 was not successful — at least with the critics,Note: *See the Gazzetta musicale of 12 May 1889, p. 301, and of 9 June 1889, p. 372.  — and the version played on that occasion has never been printed. Since the performance material (full score and orchestral parts) used for that production was destroyed in Ricordi’s Milan storage facility during World War II at the latest, this first version also cannot be reliably reconstructed. The very detailed reports on the premièreNote: *Gazzetta musicale of 28 April 1889, pp. 267-272; Corriere della sera (reprinted in Carteggi 1958 No. 30, footnote 1); La Perseveranza and La Lombardia (both reprinted in Cesari 1994, pp. 135-152).  do not provide any certainty about it, nor does a description of the changes in the score announced on the occasion of the publication of 62.E.1.Note: * Gazzetta musicale of 9 February 1890, p. 91; see also Quaderni 1992, p. 105.  However, there are hints to the very first version to be gleaned from a careful comparison of the autograph full score, the first piano-vocal score, and the libretto printed for the first performance (as Cesari has done).Note: *See Cesari 1994, pp. 174-182, 254-260, 516-521.
In any case, it is certain that the first performance differed at several points from the autograph full score. This is indicated by a leaf (now in I-Lgp) containing Puccini’s handwritten notes on a back cover of the libretto printed for the first performance, which he most certainly made during the Milan rehearsals. It follows from this that there certainly were cuts in the first part of Act 4 (Gualtiero’s “Preghiera” and the duet of Edgar and Fidelia), but perhaps also in the duet of Edgar and Tigrana and in the finale of Act 2. However, it is also possible that Puccini made those notes only as corrections for an intended further revision.
Far more substantial changes were agreed between Puccini, Fontana, and Giulio Ricordi immediately after the La Scala performances (see Ricordi’s letter of 28 April 1889),Note: *The original is in I-Lmp; English translation in Marek 1951, pp. 67-69. and Puccini began his work at once (see Carteggi 1958 No. 31 of 7 May 1889). According to MRL, already at the beginning of June Ricordi also began the production of the piano-vocal score, although Puccini was nowhere near finished with his work. Fontana’s textual drafts for an expansion of the second finale are found in his letters from the second half of August.Note: *Quaderni 1992 II, Nos. 133 and 134.  On 8 September, Ricordi received the revised first act,Note: *See Pintorno 1974, No. 1. and Puccini announced the delivery of Acts 2 and 3 for the end of the month (Carteggi 1958 No. 32), so that the complete revision probably was concluded in October 1889. On this basis, the first printed piano-vocal score of Edgar (62.E.1) was published in January 1890. On that occasion, Ricordi pointed to the fact that the edition considerably differed from the version of the first performance.Note: *See the Gazzetta musicale of 9 February 1890, p. 91. Consequentially, the Act 1 concertato was “totally altered in form ... rendering it much more compact and also more natural” (“cambiata assolutamente la fisionomia al ‘concertato’... rendendola molto più spiccia ed anche più naturale”); in Act 2, the initial chorus (“Splendida notte”) was cut, the finale extended by 250 bars, and Tigrana’s concluding solo (“O della morte o mio tu sarai”) added; in Act 4, the introductory scene and the love duet were changed and the “battaglia de’fiori” was cut.Note: *See also the summary in Cesari 1994, pp. 522-523.  To a large extent, these changes originated from Puccini’s notes made during rehearsals for the first performance, discussed above.
This version with changesNote: *See 62.B.2 and Cesari 1994, pp. 183-186 and 273-275. was the basis for the second performance of the opera on 5 September 1891 in Puccini’s home town Lucca, only two and a half years after the first performance. In contrast to the unsuccessful La Scala première, Puccini had an overwhelming success in Lucca.Note: *See Arrighi 1957, with reports from the local newspapers, as well as the Gazzetta musicale of 13 September 1891, pp. 591- 
594, and the following four issues of that weekly publication.  The newspapers were already mentioning a forthcoming further revision and reported that Puccini was going to eliminate Act 4. Evidently that news, originally disseminated by Puccini’s friend Carlo Paladini, was based on insider information. So the intention to delete the fourth act probably stemmed from the composer and Ricordi during the Luccan rehearsals and performances.
The additional revisions Puccini made soon afterwardsNote: *See MRC 1891/92 - 7/467 of 28 October, Quaderni 1992 II, No. 141, of 10 November, and Carteggi 1958 No. 56 of 
27 November 1891.  were not restricted to the elimination of Act 4 and the integration of the end of its plot into the finale of Act 3 (musically nearly unchanged; Puccini later adapted a portion of the Fidelia-Edgar duet from Act 4 for use in Act 3 of Tosca (69). They also led to other considerable retouches and cuts. Act 1 was the least effected by the revisions; the shortened prelude of the former Act 4 was played at its beginning and, in all, only about 50 bars of the act were cut. On the other hand, Act 2 was shortened by a third, and Act 3 was reduced by about one quarter of its original size.Note: *For further references to these changes found in MRC, see Cesari in Quaderni 1992, p. 99, footnote 7; see also the
overview in Martinotti 1977, pp. 504-09.
These changes were originally destined for the production in Madrid, to be conducted by Luigi Mancinelli.
It already had been scheduled for a long time,Note: *See Carteggi 1958 Nos. 56–58, and earlier mentions in Nos. 41, 43-45, 48-51; see also Cesari 1994, pp. 267-272. postponed for various reasons until 19 March 1892, but already in rehearsal for several months.Note: *See the Gazzetta musicale of 13 December 1891, p. 812.  The first performance of the new version, therefore, took place somewhat unexpectedly on 28 January 1892 in Ferrara. It was based on the new piano-vocal score, just published at the beginning of 1892 (62.E.2). Even earlier than the Madrid production, there was also a performance in Turin on 5 March 1892.Note: *See the reports in Cesari 1994, pp. 283-86. There are reasons to believe that the Madrid performance differed, at least in modifications to the orchestration, from the Ferrara/Turin version, for Puccini was in direct contact with Mancinelli and wanted to make changes in Madrid at the last moment. In particular, in Madrid Puccini once again wrote out the full score of the slightly shortened prelude to Act 1 (formerly to Act 4, see 62.B.3).Note: *The whole complicated process is described in detail by Cesari in Quaderni 1992. It is uncertain whether the Madrid changes became established in Ricordi’s hire material for later performances, but no new piano-vocal score was printed because the voice parts were obviously not affected by the changes. There was only one performance (in August 1892 in Brescia) before the next revision. There were apparently also plans in the following years for performances in Naples and Bologna,Note: *See MRC of 28 August 1896, and the Gazzetta musicale of 22 September 1898, p. 550, respectively. connected with new reworkings (compare also 62.B.5); but they did not take place.
As indicated in the note inserted in 62.B.1, it was only on 15 January 1901 that Puccini received a copy of the full score from Ricordi in the 1892 version (62.E.2a), along with the autograph Act 4, which has been missing ever since. So it is clear that at that time he intended to work on the opera again.Note: *This is confirmed by Carteggi 1958 No. 247 (of 11 March 1901), in which Puccini wrote: “riduco ‘Edgar’!!! in galvanoplastica! [I reduce ‘Edgar’!!! in electrotyping!]”. Perhaps Puccini was not joking here at all, as is often thought, but actually was referring to the printing process for the new piano-vocal score.  If letter No. 26 in Epistolario indeed dates from 6 March 1901, as it probably does,Note: *Cesari also concurs with that opinion; see Quaderni 1992, p. 103, footnote 32 Puccini wanted to reinsert Act 4 and cut Act 2 instead. Such a version was not made, because a short time later Puccini finally obtained the operatic rights to David Belasco’s play, Madame Butterfly (74), and began working on that piece. Therefore, he obviously put Edgar aside again.
Only during his next hiatus, after the first performances of Madama Butterfly when Puccini was looking for a new subject for an opera, did he return once more to Edgar. On 24 January 1905, Puccini retrieved the autograph full score of Act 2 from Ricordi. That means he not only wanted to shorten Act 2 again, but also to revert to its original version, or to the status of 62.E.1 not preserved in the copy. On 23 February 1905, Puccini informed Ricordi that the revision of Act 1 was finished (Epistolario 1982 No. 88). Three weeks later, the rest was settled, including a new version of the piano-vocal score, prepared by Carignani to parallel Puccini’s changes (Epistolario 1982 No. 27 of 14 March 1905). The earliest known printed copies of the new piano-vocal score (62.E.4) already are from April 1905, and according to MRL, Ricordi only began the production of the edition on 4 March. The reason for this great haste was an imminent performance, the first one after more than a decade, in the context of a nearly complete cycle of Puccini’s works (only Le Villi was missing) in the summer of 1905 in Buenos Aires, for which the composer travelled to Argentina. This was the last performance of Edgar during Puccini’s lifetime, the last of only seven productions in all. So Edgar is Puccini’s least performed opera.
For the definitive version, Puccini again made substantial cuts in several places. He also reinstated other music, for example, the original introductory chorus, “Splendida notte” (probably in an altered version) in Act 2. Ultimately, the definitive version of Edgar was only half as long as the original version. No other Puccini opera shows such serious interventions, also drastically affecting the dramaturgy.
It is obvious that in 1944 Ricordi produced their first duplicatable full score of Edgar — although still not made available for sale, and not engraved. That full score is a revision of 62.E.4 in accordance with 62.B.6ì. It also corrects some inaccuracies, particularly in the indication of dynamics, of 62.E.4. Understandably, the textual variants of 62.E.4 compared with 62.B.6 were, however, retained, since they certainly originated with Puccini’s approval — if not directly due to his own initiative.