This is a series of studies of individual operas, written for the serious opera-goer or record-collector as well as the student or scholar. Each volume has three main concerns. The first is historical: to describe the genesis of the work, its sources or its relation to literary prototypes, the collaboration between librettist and composer, and the first performance and subsequent stage history. This history is itself a record of changing attitudes towards the work, and an index of general changes of taste. The second is analytical and it is grounded in a very full synopsis which considers the opera as a structure of musical and dramatic effects. In most volumes there is also a musical analysis of a section of the score, showing how the music serves or makes the drama. The analysis, like the history, naturally raises questions of interpretation, and the third concern of each volume is to show how critical writing about an opera, like production and performance, can direct or distort appreciation of its structural elements. Some conflict of interpretation is an inevitable part of this account; editors of the handbooks reflect this – by citing classic statements, by commissioning new essays, by taking up their own critical position. A final section gives a select bibliography, a discography and guides to other sources.

Victorien Sardou (1831–1908) is today scarcely more than a footnote to the history of late nineteenth-century French drama. Were it not for Puccini's opera, based on his play, Sardou's name would probably be known today only to students of French literature. Yet in his time he was a most successful playwright, generally recognized as a born man of the theatre who had all the tricks of his trade at his finger-tips. Modelling his technique on that of his more famous predecessor, Eugene Scribe, who may be said to have initiated what is nowadays called (with some contempt) the ‘well-made’ play, Sardou wrote a number of stage-works in this genre. An extremely well-thought-out plot, basically naturalistic and which unfolds with almost mathematical logic, stunning coups de théâtre, and a dialogue flexible, well-turned and often razor-sharp – these were the positive aspects of Sardou's dramas. Yet there was in them no profundity of thought and feeling, no spiritual, moral or social ‘message’, no poetry. Sardou's were boulevard dramas in which his supreme aim was to entertain – to entertain as a high-class thriller entertains, creating an atmosphere of mounting suspense and riveting the spectator's attention by means of sensational stagehappenings that were to their author of far greater importance than the exploration of a character's psychology. In a Sardou play, action dwarfs character which is, as it were, only one storey high. ‘Sardoodledom’ was the unflattering term that G. B. Shaw coined for this kind of play.

I remember distinctly the first time Floria Tosca entered my life. It was years before I, as Baron Scarpia, was to take a hand in directing her tragic fate, but I can see her now in my mind's eye. I still hear the sweet impatience of her ‘Mario! Mario!’ sung off-stage. I still feel the frisson of excited expectancy which passed through the audience. And then she was there – entering the Church of Sant'Andrea della Valle with royal bearing, her left arm encircling her offering of flowers, extended so that the light caught the sheen on her sleeve and shoulder. She hardly touched the floor in the lightness of her tread – a marvellous vision, complete in its expression of art: La Divina Claudia Muzio.

Even then I think I sensed something of the elements which combined to make that perfect entry. Infinite research and study had prepared the groundwork, imagination had been brought to bear on the knowledge thus absorbed, and a tremendous flair for the projection of a unique personality had completed the miracle. Fascinated by the way in which singing and acting were blended in a harmonious whole, I dreamed that one day perhaps I would be able to do the same.

Tosca is, strictly speaking, not a music-drama in the accepted (Wagnerian) sense, but rather the early herald of the modern musictheatre, that is, a drama with a powerful, action-packed plot round which the music coils and recoils with snake-like suppleness and pliancy. Were it not for the lyrical episodes in which Puccini the musician asserts himself against Puccini the dramatist, the music might be denned as a function, or extension in sound, of the action. Nowadays when Salome and Elektra form part of the ordinary repertory, and when Wozzeck and, latterly, the three-act Lulu have come into their own – all these operas partake of the nature of musictheatre – it is difficult to realize the daring novelty that Tosca represented in the early 1900s. On the face of it, the subject seemed to defy operatic treatment, and it is a measure of Puccini's stature as a musical dramatist that he not only overcame its apparent unsuitability for musicalization but found valid musical equivalents for such scenes as Cavaradossi's torture and Scarpia's sexual frenzy. Sardou's play has vanished into limbo while Puccini's opera has been proclaiming its extraordinary vitality in countless productions over more than eight decades. Moreover, Tosca provides an outstanding example, like the large majority of successful operas, of the composer's acute instinctive awareness of the inner and, largely, imponderable laws that govern action and music as expressed in the right balance between the dramatic (action-music) and the lyrical (aria, duet).

For this analytical chapter I have chosen from all three acts scenes that seem important from either the musical or the dramatic point of view, or both. Before this, something must be said about the overall musico-dramatic structure of the work. We have noted the striking acceleration of the action in Act III, as compared with the leisurely exposition of the drama in Act I, yet any imbalance that might be felt is deceptive. For Act III makes up for its comparative shortness by the weight of its dramatic happenings – the shooting of Cavaradossi, the discovery of Scarpia's murder and Tosca's suicide. Moreover, there is a certain symmetry and correspondence between the musical structures of the two acts which, though not of the strict Bergian order, is marked enough to reinforce my view that Tosca is one of the best constructed operas in the repertory. Consider: both acts open, after preliminaries, with an aria for the tenor which is followed by a love duet for soprano and tenor, and both close with a coup de théâtre – the Te Deum scene of Act I and the execution and suicide of Act III. In the context of the drama the closely corresponding scenes stand worlds apart. Yet Cavaradossi's ‘Recondita armonia’ and ‘E lucevan le stelle’ express in their different way the same thing – his all-consuming love for Tosca.

The history of Tosca in the United States following its successful introduction at the Metropolitan on 4 February 1901 has been one of apparently unflagging popularity. A staple in the repertory of the resident American companies, it has been nothing less in the offerings of the various touring companies, such as the San Carlo, which used to criss-cross the continent in fat times and lean. It seems safe to assert that there must be scarcely a city or a town in the United States where opera has been given in the last eighty years that has not been exposed to at least one Tosca. Because of the nature of its plot and its historic setting in well-known Roman monuments there has been little temptation to produce Tosca in alien epochs or styles; consequently the American history of Tosca tends naturally to focus upon the many exceptional performers who have appeared in it.

The first American cast of Tosca was headed by Milka Ternina as Tosca, Giuseppe Cremonini (who had been the first des Grieux in Puccini's Manon Lescaut) as Cavaradossi and Antonio Scotti as Scarpia; the conductor was Luigi Mancinelli. With the exception of the tenor, the principals and conductor were those of the Covent Garden premiere the previous July, a sequence typical of the regime of Maurice Grau that then dominated the principal opera houses of both London and New York.

It was Ferdinando Fontana, the librettist of Puccini's two early operas Le Villi (1884) and Edgar (1889), who first suggested Sardou's play to Puccini as an operatic subject. This was in early spring 1889, a few months after the first (La Scala) production of Edgar. The composer, who acquainted himself with the play through Fontana, was greatly impressed by it, and wrote to his publisher, Giulio Ricordi, on 7 May 1889:

(Fontana was much piqued when seven years later, after Puccini had decided on Tosca, Ricordi called in Giacosa for the versification of the libretto, a task which, as the first to have thought of the French play, he had been hoping for himself and about which he had corresponded with Sardou.)

It seems likely that, because Puccini was at the time a composer without name or mark outside Italy, Sardou at first refused permission for the musicalization of his play. Also, in spite of his evident enthusiasm for the subject Puccini, then sailing in the deep waters of romanticism (Manon Lescaut) and later attempting a mixture of romantic and realist elements (La Bohème), began to have qualms about a subject of such realistic violence and brutality.

To recast a play or a novel into an effective opera text requires a special technique. Yet, as Ernest Newman wrote some 30 years ago, while there are textbooks for composers, conductors, instrumentalists and singers, there is no manual for the art of the librettist. More recently Gary Schmidgall has singled out certain features in literary sources on which opera thrives, such as passionate force, lyrical expression, individuality and vigour of characters, direction of impact and simplicity. Schmidgall, however, only refers to what might be called the ‘opera-genic’ in literary sources – he does not touch on the specific method a librettist employs in translating a literary text into a viable opera text. His work appears to be entirely empirical. The good librettist is born, not made. His instinct, supported by ingenuity and technical skill, will guide him in the deployment and balancing of those features listed by Schmidgall. At the same time he must be flexible and adaptable enough to accommodate himself to the special musico-dramatic requirements of a given composer. Thus, a librettist employed by one composer will approach his task, so far as the details of his adaptation are concerned, from an angle different from that of a librettist working for another composer. Hofmannsthal could never have been the ‘right’ poet for Puccini, while Illica and Giacosa could never have been the right collaborators for Richard Strauss.

With the action of the opera set in Rome, Ricordi's choice (first suggested by Illica) of the Italian capital for the first production was as appropriate as it was diplomatic, since it was calculated to flatter the Roman public's amour propre as well as its local patriotism. But evidently the publisher had not reckoned with two things. The first was the implied anti-clericalism of the opera manifest not only in Cavaradossi but also in the fact that its villain was himself portrayed as a devout believer; this threw an odd light on the Catholic Church and would probably go against the grain of many spectators, to say nothing of the Vatican. The second thing likely to have an adverse influence on the opera's reception was the traditional antagonism that existed (and seems still to exist) between Rome and the cities of northern Italy. But what Ricordi could not foresee was that the disturbed political atmosphere of the time would greatly contribute to the nervous tension that reigned at the Teatro Costanzi on the evening of the first night – 14 January 1900. There was, incidentally, a notable coincidence: ten years earlier Mascagni had launched at the same theatre his Cavalleria rusticana, which marked the birth of verismo.

The three principal roles were taken by Haricleé Darclée (Tosca), Emilio De Marchi (Cavaradossi) and Eugenio Giraldoni (Scarpia). Darclée seems to have been chosen for the part less for her vocal accomplishments than for her striking beauty and her talent as an actress.

What was it that caused Puccini to revert in the mid-1890s to the Tosca subject, a subject a tinte forti, highly coloured which he had rejected some six years earlier? The answer lies in an important change in the operatic climate during the last few decades of the nineteenth-century, a change that followed the turn in European literature from romanticism to realism and its more extreme form, naturalism. This new realist movement, originating in France, was heralded by such writers as Balzac, Flaubert and Dumas fils, and culminated in Zola, the foremost practitioner and theoretician of naturalism. In opera the first stirrings were sensed in Verdi's Luisa Miller (1849) and notably La Traviata (1853), in which for the first time a sexually tainted heroine was brought on to the operatic stage. Naturalism was firmly placed on the operatic map with Bizet's Carmen (1875) which became a work of seminal importance for naturalist opera of the following decades. Thus, Alfred Bruneau wrote operas after texts drawn from the novels of Zola, Massenet made a single excursion into naturalism with his two-act La Navarraise (1898) and was followed by Charpentier with his Louise (1900). In Italy the chief proponents of operatic naturalism or verismo (from vero = true) was the group of composers known under the collective name, giovanescuola italiana – Mascagni, Leoncavallo, Giordano, and Puccini (Tosca, Madama Butterfly, La fanciulla del west and Il tabarro).

It is perhaps because of our over-familiarity with Puccini's musical language – and of course its exploitation in the hands of numerous imitators, both in Hollywood and elsewhere – that critical writings on the composer have tended to polarize. On the one hand, there are those who seem to resent profoundly his popular success, who find his music coarse, vulgar and cynical, and who present his image as a blunt stone on which to sharpen their wit. This started almost from the beginning. La Bohème and Tosca, for example, were initially greeted coolly by the press: both represented a change of direction which the critics found hard to follow. And the eventual canonization of these operas, far from redeeming Puccini in the eyes of the journalists, simply added more fuel to their fire: as his later operas appeared, they were constantly arraigned for falling below the standard of the early works, now proclaimed ‘masterpieces’. A little later, the Italian critic Fausto Torrefranca devoted a full-length book to the composer. It turns out to be a polemic of rare severity: a study whose approach to its subject is almost entirely negative. As Torrefranca says in his Preface, he chose Puccini because: ‘he seems to me the composer who personifies with greatest completeness the decadence of today's Italian music, and who represents its cynical commercialism, its lamentable impotence, its celebration of the international vogue’.

Il Pigmalione. This one-act ‘scena lirica’, as Donizetti called it, is his only attempt at a mythological subject; it tells the familiar story of the sculptor whose masterpiece is brought to life. The final page of the autograph states that it was ‘begun on 15 September and finished 1 October at almost two in the morning, Tuesday, the day of the arrival of the new legate’; the year was 1816, when Donizetti was studying with Padre Mattei at Bologna. Most probably this score was composed for the experience, and since Mayr visited Donizetti in Bologna in September 1816, it would seem the project was either suggested by him or undertaken by Donizetti to demonstrate to his former teacher his current prowess. Since Donizetti's studies at Bologna were made possible only by contributions from Bergamo, he was scarcely in a position to commission a new libretto. Readily available to him was Sografi's text for Cimadoro's Pimmaglione, which was widely performed, especially in northern Italy.

Il Pigmalione, in every sense a modest work, contains only two characters (Pigmalione, tenor, and Galatea, soprano); the vocal writing is limited in range and discreetly embellished, and the work is scored for flute, oboe, two clarinets, two horns, two bassoons and strings. The autograph is unique among Donizetti's operas in not being divided into separate numbers – a strong indication that the work was written as an exercise, without expectation of performance or publication.

From the point of view of half a century ago no recent musical phenomenon would have seemed less likely than the extensive and so far unabating revival of Donizetti's operas. In spite of the unevenness of his output, works such as Anna Bolena, Maria Stuarda and Roberto Devereux have become established in the repertory and seem unlikely ever to slip back into obscurity. Donizetti was not a great originator; he found his mature voice more slowly than most of his contemporaries; he eludes the handy labels that help us to pigeonhole composers from the past. He is a central figure, much more consistent and true to himself than one might expect of a composer whose arena was the highly eclectic and adaptive world of Italian opera in the first half of the nineteenth century. At that time a composer enjoyed little status in the musical hierarchy, especially in the jungle world of the Italian opera house, where, for instance, Méric-Lalande was paid more than nine times as much for singing the Carnival season at La Scala as the composer was paid for writing Lucrezia Borgia.

A fair estimate of Donizetti's qualities can only be achieved by considering his music in the context of the theater within which he worked, its musical conventions, and its personnel of singers and other composers, all of whom were struggling to capture and maintain the loyalty of the public.