Avoiding silent opera: the ‘grand’ performing right at work in nineteenth century Paris

The music industry has been made possible through performing rights based on a law introduced by the post-revolutionary French national assembly in 1791. However, it took until the mid-nineteenth century until a system of royalty collection was established in France (and another half a century or more in other countries). In France, this new system for non-dramatic performing rights was preceded by royalty regulations in theatres. This study describes how nineteenth century composers were compensated for their work in the Paris Opera through this performing right for drama, known as the ‘grand right’. The tariff-based compensation method had been put in place by a royal réglement in 1713. It created a classic winner-take-all phenomenon in which composers such as Auber, Meyerbeer, and Halévy thrived in the nineteenth century. A contributing fact was the opera house programming which, contrary to the programming of today, favoured new pieces. ‘Grand opéras’ were à la mode and they contributed to the financial success of their composers. However, these operas eventually lost their attraction. In 1884, the Paris Opera adopted a compensation system based on a percentage of box office revenues. The study is based on primary data for 1810–1866.


Introduction
One of the many costs that makes opera a very expensive art form is the royalty attributed to composers (in addition to librettists, translators, and choreographers). Without music there would be no aria. Nevertheless, composers' compensation is, in most cases, only a small part of the budget for an opera production. However, the composer, in contrast with performers, has an advantage in the ability to reap such tiny income from simultaneous performances in several theatres. Moreover, there will be revenues for all future performances during the composers' lifetime and, presently, for 70 years in the future for their heirs. The sum of miniscule performing right revenues may be considerable.
Much has been written about the 'small right' form of performing right, especially from around a decade ago when the debate on copyright piracy peaked (for instance : Liebowitz 2005: Liebowitz , 2008: Liebowitz , 2010Zentner 2006;Oberholzer-Gee and Strumpf 2007;Andersen and Frenz 2010;Albinsson 2013a). However, opera royalties are excluded from the 'small right' as the latter only applies to 'nondramatic performances' (which is the term used, for instance, in U.S. Code Title 17. Copyrights). For dramatic performing rights, the term 'grand right' (GR) is used. It is this kind of per performance royalty that is the focus of this study.
The author has found only one previous-and very recent-article concerned with the GR specifically: Cuntz's 'Grand rights and opera reuse today' (2020). Cuntz finds 'that changes in copyright status increase the total number of performances that individual works receive on stage once copyright expires'. Furthermore, he provides 'preliminary evidence on chilling, long-term effects of status around premiering operas and revivals at the beginning of the copyright term'. This is, however, a comment regarding the present situation in opera house programming. As we shall see, the Paris Opera programming was the exact opposite during the nineteenth century.
The author finds it slightly surprising that GR is more or less unaccounted for in Agid and Tarondeau's (2006) book on the Paris Opera and in their multi-opera comparative study (2011). The former mentions, in passing, the 1713 règlements with a clause on the remuneration of authors (including composers), but nothing else regarding this topic. The details of the 1713 règlements will be discussed below. Annunziata and Colombo's (2018) Law and Opera compilation contains 25 chapters contributed by many different authors. None even slightly deals with GR. The recent article by Giorcelli and Moser (2019) about the effects of French copyright on operatic creativity in the parts of Italy under Napoleonic rule does not include GR, although it is the intellectual property rights method peculiar to the opera stage. They find 'first, … that the adoption of copyrights led to a significant increase in the number of newly created operas. Second, copyrights raised the quality of new operas, measured both by their immediate success and by their longevity. Third, there were no benefits from copyright extensions beyond the life of the original creator.' It could be argued that these conclusions regarding some Italian provinces are relevant to Paris during the studied period as well. Presented below are details regarding performing rights royalties-the kind of information that Giorcelli and Moser's paper lacks. A paper by Cancellieri and Turrini (2016) presents nine hypotheses regarding the way in which economics and policies affect the programming strategies of opera houses-none of which deals with GR. Sudhanva Deshpande (2009) claims that 'experience suggests that copyright is designed to protect corporate profits rather than artists' creativity. It is imperative for artists and theatre people to work out alternatives to copyright ….' This seems to be a futile wish as copyright, the performing rights in music in particular, provides an ever-increasing number of jobs and massive tax revenues from commercial mass culture companies and their employees. Hence, this study takes GR as a given legal condition. This was certainly the case during the studied period. As this study provides information on a little-researched field, it seems that it is 'a first'. The information that it provides is detailed but limited to one particular domain, that of the GR system in the Paris Opera in the nineteenth century.
Before GR, composers could receive an income on their operas from: α. employment by a royal or aristocratic household or a theatrical entrepreneur β. lump sum: public and private ex-ante grant, commission fee, or 'Gegenverehrung', that is, a reciprocal veneration -a token of gratitude for a dedication in a score (Pohlmann 1962, 137). γ. the net box office takings for one or more, in rare cases, 'benefit concerts' δ. ex-post awards or 'blue-sky-prizes' for operas found to be of high enough quality to be performed (Scotchmer 2004. ch. 2.3 and 2.5). Sometimes composers were given bonuses after a sufficiently large number of performances ε. publication fees from publishers of printed scores and parts The lump sums received by composers may have depended on the number of planned performances. Yet, the compensation was still not based on an actual performing right system. The Paris Opera seems to have been the first to remunerate composers for each performance on a tariff basis. It is the presentation of the unique tariff system applied by the Paris Opera that is the new contribution of this study. The presentation is based on the collection of a multitude of primary, historical data.
After the introduction of performing rights, composers could reap income from concerts and opera productions. Furthermore, copyright regulations have been expanded by every new technological innovation (Albinsson 2013b): (A) performing right-grand right royalty. This can be included in the commission fee, if any, but can, alternatively, be negotiated separately. The GR is negotiated for every single opera production by the composer or a publisher. (B) performing right-small right royalty. Applies when the whole or parts of the opera is performed non-dramatically, i.e. in concerts, or broadcasted or streamed. The small performing right is licensed through the 'blanket' use of all music covered by a royalty collecting association. In most countries, there is a single such collector-SACEM in France.
1 3 (C) Mechanical rights-apply to the sale of recorded tangible items, mainly vinyl, CD or DVD.
The study of GR is hampered by the fact that most music theatre companies are private entities and so little has been archived; however, the affairs of opera companies run within the realm of state budgets have fared better. In some cases, their archives are extensive. The aim of this paper is to investigate the financial effects of the tariffbased GR used by the Paris Opera, now the Opéra national de Paris, in the nineteenth century. Although the GR is the paper's main focus, other kinds of income are mentioned as historical examples. The absolute worth over time (rather than the relative worth) from Rodney Edvinsson's on-line historical currency converter is used for present-day comparisons in this study. 1 More precisely, Edvinsson has estimated the quantity of consumer goods and services that x French francs in year y could buy in Sweden, his home country, in 2015.
How much could the prescribed royalties buy when they were paid? A baguette and a bottle of wine are essential and historical ingredients on French household menus. In 1855, a kilogram of white bread cost 0.40 francs and a litre of wine cost 0.55 francs (Sabot 2012, 34). Costs for specific items directly related to composers in this study have not been found with the exception of Meyerbeer's rent for the flat on Rue neuve St. Augustin, 49, which in 1842 was 270 francs per month (Becker and Becker 1975, 410).

The regulation of grand rights in Paris
The modern global system of small right licensing organisations was first introduced in Paris in the mid-nineteenth century (Albinsson 2014). However, there was a GR regulation for the Royal Opera in Paris long before. The first GR regulation to be used by the Paris Opera was introduced in the réglements issued by King Louis XIV after the establishment of l'Académie Royale de Musique in 1669 (the predecessor of the current Opéra National de Paris). Through the eighteen-paragraph Réglement concernant l'Opéra donné à Versailles le 11 Janvier 1713, King Louis XIV tried to regulate how the opera should be run, including the first GR clause ( §15). The composer remuneration clause was repeated (as §16) in the even more It may be that the réglement ratified and clarified a system that was already in place. The clause stipulated 100 livres (equal to approximately €1200 in 2015) for each of the first ten performances and 50 livres for twenty more performances thereafter. After these thirty performances, the piece belonged to the opera company, which could stage it again without further royalties; see Table 1. Thus the composer could be given a maximum of 2000 livres if the opera was sufficiently popular with the audience. The lead singers had a yearly salary of 1500 livres. With a successful opera the composer could thus earn more or less what the main actors received (Durey de Noinville 1757). The use of this Parisian proto-GR elsewhere in France is unlikely. For instance, the entrepreneur who produced opera performances at the Théâtre du Capitole in Toulouse in the 1750s, Monsieur Prin, recorded no honoraires d'auteur costs in his bookkeeping (Archives Municipale de Toulouse, code GG 942).section Thus, the compensation from the royal opera to composers was based on fixed sums per performance. For the few playwrights whose oevres were performed at the Comédie-Française, a réglement was issued by King Louis XIV in 1667 stating they would receive compensation of one-ninth of the net box-office revenue (approximately 11%) (Petri 2000, 68;Brown 2004, ch. intermission, 1). The struggle by dramatists, led by Beaumarchais, to gain legal possession of their plays has been documented in detail by Boncompain (2001) and Brown (2006). When it came to the actual ownership of their works, the circumstances were principally similar for composers. None could expect to maintain long-term property rights. This issue was not addressed until the Décret rendu sur la Pétition des Auteurs dramatiques (Décret 1791) was passed by the post-revolutionary National Assembly on January 13, 1791. The decree was based on a report by the Jacobin lawyer and orator Jean le Chapelier; Article No. 3 reads: The works of living authors may not be represented in any public theatre within the extent of France without formal consent in writing by the author; the penalty for violation is to be the confiscation of the total proceeds of all performances for the benefit of the authors. 2 This was the first legal safeguarding of the performing right in the hands of the originators. In the 1 September 1793 revision it is specified that playwrights are guaranteed the means to dispose of the property of their works with equal freedom regardless of whether they are printed or performed (Le droit d'auteur 1893;Brown 2006, 147).
The 1791 copyright law introduced the new concept of post mortem autoris, i.e. the right for the heirs of a deceased copyright owner to collect the royalties for a prescribed number of years. In this introductory case the duration of the post mortem autoris clause was ten years. It was left unchanged until a new law, Loi sur les droits des héritiers et des ayants cause des auteurs, was passed on 14 July 1866. A giant leap was then taken as the post mortem duration was extended to 50 years (Law 1866).
The organisation established by Beaumarchais, La Société des Auteurs Dramatiques, to work for the recognition of the right for playwrights to be sufficiently remunerated (primarily) by La Comédie-Française succeeded in its pursuit through the new law. A new, more operative organisation, the Bureau Central de Perception des Droits d'Auteur, was created by composer and librettist Nicolas-Étienne Framéry to implement the new legislation. It soon reached agreements with La Comédie-Italienne and four other theatres. These agreements stipulated that one-seventh of the net box-office proceeds of all performances should be granted to the author (Brown 2006, 144). A remuneration size in this range, 12-14% of box office revenues, is, again, the common business praxis today (Albinsson forthcoming). A similar agreement must have been signed with L'Opéra-Comique. At least, that opera company, presenting somewhat lighter, more operetta-like performances, used a percentage system in 1831 (Archives Nationales-henceforth AN, code AJ13/1034). The Bureau was succeeded by the Société des Auteurs et Compositeurs Dramatiques/SACD in 1829.
The division between grand and small performing rights was established after nondramatic performing rights were accepted by courts of law in Paris and the first collective licensing society, Société des Auteurs, Compositeurs et Éditeurs de Musique (SACEM) was established (Tournier 2006, 26-27). In SACEM's first charter, works already protected and administered by the much older SACD, 'operas, vaudevilles and comic scenes,' were exempt from handling by SACEM. The new society explicitly wanted to 'in no way affect the powers or rights of the Société des Auteurs Dramatiques, as they remain today' (SACEM's Act of Constitution 1851. Article 18). This exemption has been included in most, if not all, other national charters for collective licensing agencies. Thus, the separation of grand and small rights remains globally today.
In the revised réglement from 1776 a pension scheme was introduced: 'After three successful pieces large enough to make those pieces longlasting in the repertory the composer will receive a pension of 1000 livres per year. It will increase with 500 livres for each of the following two pieces. For the sixth piece the pension will increase by 1000 livres'.
The tariff system of GD introduced by Louis XIV in 1713 held sway until the late nineteenth century. In an Arrêt du Conceil d'Etat de Roi of 10 April 1778, it is made clear that the increased royalty tariffs that had been introduced in a similar arrêt two years earlier should again be valid under the new opera entrepreneur, Jacques livres for the following ten and 100 livres for each of the others, up to and including the fortieth; see Table 2. Furthermore, 'His Majesty wishes that in the event that the number of the representations exceeds, without interruption, that of forty, it is paid to each of the authors a gratuity of five hundred livres' (Arrêt 1778). Gerhard (1992, 40) claims that the tariff was raised to 250 francs (1 franc = 1.0125 livre) for each of the first 40 performances and 100 francs for each of the following after the July Revolution of 1830. However, this tariff had already been decided upon by Louis XVIII on 18 January 1816 shortly after the re-establishment of the monarchy (AN, code AJ13/109; see Table 3). The 1816 tariff was augmented from 1 January 1861 through a décret signed by Napolén III (AN, code AJ13/221; see Table 4). It seems, however, that the opera managers had a rather slack attitude towards the tariffs. Sometimes royalty payments were not precisely according to the monarch's directives.  The currency used in Tables 1 and 2 is the livre. This was replaced in 1795, after the French Revolution, by the franc, which was worth 1.0125 livres when introduced.
Although not explicitly stated in the Réglements, the actual payments reveal that only composers residing in France-preferably in Paris-received royalties. International treaties appeared at a later stage. 3 Place of residence was clearly what mattered. Gioachino Rossini received compensation after his move to Paris in 1824.
The tariff system did not distinguish between operas composed for the Paris Opera or those performed there after they had been premiered elsewhere. No sign of the separate present-day 'commission fee' has been found.

Performing right in other European countries
In an international perspective France was the pioneering country. The debate in Britain in the beginning of the nineteenth century dealt with the question of whether it was the published play or something non-copyright that was staged when the text was enhanced by acting, set design, lighting, costumes, and sometimes music. Only alterations of published plays when staged were at first considered as copyright infringements. In 1830 playwright James Robinson Planché suggested a bill to alter and extend the provisions of the Copyright Act of 1814 'with respect to Dramatic Writings.' His initiative eventually resulted in the Dramatic Literary Property Act (1833). Its introductory clause ruled that: … the Author of any Tragedy, Comedy, Play, Opera, Farce, or any other Dramatic Piece or Entertainment, composed, and not printed or published by the Author thereof or his Assignee… shall have as his own Property the sole Liberty of representing, or causing to be represented, at any Place or Places of Dramatic Entertainment whatsoever, in any part of the United Kingdom of Great Britain… or in any Part of the British Dominions… (Dramatic Literary Property Act 1833, first paragraph).
How legal matters should be conducted was specified as well as a fine for infringements of 'not less than Forty Shillings' (Dramatic Literary Property Act 1833, second and third paragraphs). The playwrights immediately organized the Dramatic Authors' Society for the administration of performing rights. The new, revised Law of Copyright of July 1, 1842, clarified 'that the Provisions of the said Act [i.e., the Dramatic Literary Property Act of 1833]… and this Act, shall apply to Musical Compositions…' (Copyright Act 1842, 412).
A copyright law, Gesetz zum Schutz des Eigentums an Werken der Wissenschaft und Kunst (Law for the Protection of Property in Scientific and Artistic Works), was 1 3 European Journal of Law and Economics (2021) 51:183-200 issued for Prussia by its Nationalversammlung (Parliament) in 1837. 4 It included a section ( § §32-34) on performing rights: The public performance of a dramatic or musical work, be it wholly or with insignificant abridgements, may only take place with the permission of the author, his heirs, or his legal successors, as long as the work has not been published by means of printing. The exclusive right to grant this permission is vested in the author for life and in his heirs or legal successors for a period of 10 years after his death. (Prussian Copyright Act, 1837) Similar performing rights protection was introduced in other Germanic states. Saxony included a performing rights directive in its Code of Law in 1841 (Kawohl 2008). Austria followed in 1846 (Austrian Copyright Act, §7). When the German Empire was officially constituted on 1 January 1871, a copyright act, which the North German Confederation had decided on 6 months earlier, was implemented for the entire empire (Copyright Act for the North German Confederation 1870, 350-351).
Spain also issued regulations for performing rights. The Royal Order on Music Copyright (1839) declares that the performing rights for literary works that were established 2 years earlier should be extended to musical compositions. Portugal's first copyright law of 1851 includes a performing rights clause (Copyright Law of Portugal 1851, titulo II).
In 1855, King Oscar I of Sweden issued a law regarding the prohibition of public performances without the consent of the 'owner' (Kongl. Maj:t, 1855). The new law includes both a kind of proto-post-mortem and a provision concerning fines. The 1863 'Royal Majesty's Gracious Regulation for the board and administration of the Royal Theatres' (Kongl. Maj:t 1863, §72) states that 'to a piece originated in Sweden and accepted for a first performance an acceptance fee shall be immediately paid; for a spectacle for the full evening it is settled at 225 Riksdaler' (approximately €1485 in 2015). §73 provides royalties as shares of box office revenues for performances following the first. Hence, a mix of the percentage and the tariff systems was to be used.

The primary, historical data
Both primary and secondary data has been collected for this study. The collection of a large quantity of primary data is the unique contribution of this study. These data have been collected from the Archives Nationales de France/AN (The National Archives of France). They have been extended by data from the chronopera.free.fr website, which presents collected performance data for the Paris Opera from 1749. Information on titles, composers, and librettists, where such information were lacking in the AN or Chronopéra data, have been collected from La Bibliothèque nationale de France (The National Library of France). The primary data cover the period 1810-1866 plus 1871. Chronopéra lacks data for 1810 and 1817-1821. Data for these years were found in Archives Nationales (codes AJ13; 94, and 164-168). By contrast, for some periods for which the Archives Nationales lack data Chronopéra has been used. There is a general lack of data for 1867-1870. However, for some important composers Chronopéra information for these years are used.
The primary data have been transferred to an SQL relational database. It includes 8417 posts from the Archives Nationales and 1268 posts from Chronopéra. For the latter posts, likely royalties were added based on what was found in the Archives Nationales data immediately before or after the individual Chronopéra additions. Moreover, 136 operas and 84 ballets were performed during the studied period, which were created by 95 composers and 117 individual librettists or librettist in collaborations.

Supplementary income
When an opera was successful, the right to print and distribute was bought by publishers from composers through contracted lump sum payments (Income A). Until the latter part of the nineteenth century, publishers covered their costs by selling copies to opera houses. Even more important as a source of income were the adaptions for piano, two or four hands, of the score and the single arias. They were widely sold to the international bourgeoisie. After the mid-nineteenth century publishers began instead to rent out the music. The rental included the use of the printed score and parts plus what could be estimated as a fair share of the ticket revenues. Hence, the rent included the performing right. Everything was included in the publisher calculus: the venue, the season, the impresario and the mood of the city (Toelle 2007, 54). As neither Toelle nor Rosselli (1884, ch. 5) describe a per performance royalty system in Italian opera houses during the nineteenth century, it seems that the only fee that composers received in Italy was lump sum remunerations from, first, impresarios and, later, publishers when new operas were sold to them. 5 According to Neal Zaslaw (1990, 540) Jean-Baptiste Lully (1632-1687) was the only man in history to run the Paris Opera at a profit. In the process, he died a very rich man leaving an estate worth 800,000 livres (equivalent to approximately €15,810,000 in 2015) (Blanning 2008, 77). During the 1750-1751 season, the Paris Opera paid 7500 livres (equivalent to approximately €70,000 in 2015) as honoraires des auteurs, i.e. royalties (Albinsson 2018). The accounts for 1766-1767 lists such costs at the same level as the costs for the luminarie/illumination: 25,000 livres (AN/Archives Nationales, code AJ 13/2).
Before Rossini came to Paris in 1824, he worked as the principal music director at the Teatro San Carlo in Naples with an annual salary of 12,000 francs. The opera impresario Domenico Barbaja also controlled gambling operations in the royal theatres in Naples. Rossini was given a share of their revenues of at least 1000 ducats. It seems that Barbaja managed much of Rossini's earnings even after he had left Naples. Stendahl informed a friend that Rossini 'has invested 100,000 francs with Barbaja at seven and a half per cent' (Kendall 1992, 139;Eisenbach 2013, 52).
It cannot only be the Paris Opera GR royalties that enticed Rossini to move to Paris. His earlier operas were performed there only four times before 1826. Already in December 1823, the French royal household had negotiated with Rossini to compose an opera buffa for the Théâtre-Italien and a bigger opera for the main Paris Opera. He asked for 40,000 francs and a benefit performance. The contract was signed on 27 February 1824. The work on the opera for the Paris Opera was delayed as Rossini took on the role of director of the Théâtre-Italien. His Il viaggio a Reims was premiered there on 19 June 1825 (Kendall 1992, 125-128).
After the fourth performance of Il viaggio a Reims, Rossini staged Giacomo Meyerbeer's Il crociato in Egitto at the Théâtre-Italien-a year after its first performance in Florence-and he invited the composer to Paris. In this way, Rossini facilitated Meyerbeer's remarkable career there. Rossini's commissioned piece for the main Paris Opera, Le siège de Corinthe, finally received its premiere on 9 October 1826. Its 106th and last performance was in June 1844. By then, it should, according to the tariff in Table 3, have earned Rossini a GR income of 16,600 francs. Although these GR royalties were significant, they were, in total, substantially less than the 25,000 francs that Rossini received in 1826 as annual salary from the royal household as the appointed Premier Compositeur du Roi et Inspecteur Général du chant en France (Kendall 1992, 131-133). Hence, it is not likely that it was the GR tariffs that drew Rossini to Paris.
Rossini had substantial income from the publishers of his operas (Income ε). In 1826, he signed the first contract with publisher Eugène Troupenas at 6000 francs for Le Siège de Corinthe (Kendall 1992, 134). Only 2 years later, he received 16,000 francs from Troupenas for the publication rights for Le Comte Ory. In 1829, Troupenas paid Rossini 24,000 francs for Guillaume Tell.
Giacomo Meyerbeer (1791-1864) received 24,000 francs for Les Huguenots from Schlesinger and later 44,000 francs for Le Prophète from Brandus & Cie. According to the Breitkopf & Härtel business historian von Hase (1918Hase ( /1968, the contract with Brandus & Cie included a paragraph concerning the relation between the main publisher, Brandus, and the German partner, Breitkopf & Härtel. The latter was to pay the former 8000 FFR for the right to publish various arrangements of the opera and parts of it-many translated into German.

General findings
Operabase.com lists the globally most played titles during the 2018-2019 season. For all of the top 50 operas the post mortem periods had expired and, thus, no composer royalty payments were demanded. In the nineteenth century Parisian data in this study, only 6.8% of the 9685 performances were by deceased composers. Only 0.3% were of operas by composers who, by the current post mortem standard, would not receive royalties. This probably depicts a change of audience taste which has been financially beneficial for the opera companies of today.
Mozart's Don Giovanni was performed 68 times from 1834 by the Paris Opera with no royalty to the composer. Firstly, he was foreign and, secondly, the post mortem period was long past. However, 7120 francs were paid to Castil-Blaze (1784-1857) for his arrangement of Mozart's music.
The mean royalty for opera performances was 132.16 francs, ranging from a minimum of 33 francs to a maximum of 345 francs. A basic OLS regression shows that the royalty for a particular piece decreased by 2.31 francs per annum. However, as shown in Tables 1, 2, 3 and 4, the decrease was, in fact, step-wise rather than continuous.
When Carl Maria von Weber's Der Freischütz was put on in 1841 as Le Freyschütz, the composer had been dead for 15 years. Neither he nor his heirs were entitled to royalties. The Paris Opera management demanded that no opera should have spoken dialogues so Hector Berlioz set them to music. For this, he received a royalty of 233.34 francs for each of the first 15 performances, 170 francs for each of the next 45 performances and a final royalty level set at 100 francs, comprising a total of 12,940 francs for 74 performances.
The median opera performance royalty for the whole period was 100 francs. Through the 1810s and slightly into the following decade, this was the same amount as the monthly salary for rank-and-file bass players (Albinsson 2016). Their salary was later decreased to 75 francs per month. The leader of the first violinist section, the premier violon, was paid roughly three to four times as much. In the opera budget for 1819, solo singers, les premiers sujets, were listed with monthly salaries of 833 francs with ample possibilities for traitements supplémentaires. The premières sujets, i.e. principal female solo singers, were paid equally much making it an early example of a non-discriminating pay-roll.
In society at large, the average hourly wage for a labourer was increased from 0.18 francs in 1810 to 0.205 francs in 1860; that is, a 14% raise. By contrast, the Paris Opera composer tariff did not increase at all.
For conseillers d'état (state councillors), the annual salaries were halved between 1810 and 1830, from 24,000 to 12,000 francs, perhaps as a consequence of the July Revolution in 1830. By 1860, they were again at the same level as in 1810. In 1820, professors at the Collège de France had a yearly income of 6000 francs. This was raised to 7500 francs in 1870 (Fourastié 1962, 28-29). The most successful opera composer during the studied period, Giacomo Meyerbeer (1791-1864), earned an average yearly income of approximately 6000 francs from the Paris Opera and publishers.
Royalties were paid in cash at the opera. Already in 1810, the first year investigated, a large share had been signed by intermediary agents on behalf of the composers. There is only a small number of names of such agents in the archived documents. One possibility is that they represented La Société des Auteurs Dramatiques.
Composing is, generally, a 'winner-takes-all' trade. Consumers of opera, the audiences, all demand the highest possible quality. After the break-through of the 'grand opéra' style in the late 1820s, the number of successful composers diminished to only a handful. The grand opéra was a spectacular dramatic genre that demanded lavish stage effects, large-scale casts, big orchestras and a ballet. In the primary data, it is obvious that a grand opéra almost always took up an entire evening as opposed to the prior double bill (sometimes even triple-bill) programming with one (short) opera and a ballet. The first work commonly considered to be a grand opéra is La Muette de Portici composed by Daniel-François-Esprit Auber (1782-1871, first performed on 29 February 1828. Gioachino Rossini followed suit. His last opera, Guillaume Tell, premiered 3 August 1829. Giacomo Meyerbeer (1791-1864) settled in Paris after the success of his Il crociato in Eggito at the Théâtre Italien. On 21 November 1831 at the Paris Opera, his Robert le diable was considered as one of the greatest operatic triumphs of all time. It marked the beginning of Meyerbeer's domination. This was made possible by only a few more operas: Les Huguenots (1836) Rossini viewed Meyerbeer's success story with malice and, perhaps, envy. In his opinion, they owed a great deal to him and especially to Guillaume Tell. According to Ethan Mordden (1985, 37), Rossini raged: 'Je ne suis qu'un compositeur, moi! Eux, ils sont des hommes d'affaires!' (I am but a composer! They are businessmen!).
Two years after Meyerbeer's death, the post mortem autoris was expanded to 50 years. Before the prolonged period expired there were, according to Chronopéra, 1695 more performances of Meyerbeer oeuvres at the Paris Opera. Whether bilateral treaties with other countries and, later, the Berne Convention of 1886 covered GR revenues and, if so, how much was paid to Meyerbeer's wife and the three surviving children-none of whom resided in France-is not known and a possible topic for further research.
Fromental Halévy (1799-1862) contributed to the grand opéra repertoire with La juive which was first performed on 23 February 1835. He later composed two more grand opéras: La Reine de Chypre (1841) and Charles VI (1843). Gaetano Donizetti (1797-1848) was internationally well-established when he settled in Paris for the premiere of Lucia de Lammermoor at the Théâtre de la Renaissance. Although that piece is not generally considered a 'grand opéra', his later Dom Sébastien, roi de Portugal (1843) is.
From a total of 614 performances, including all other works, Halévy was, during his lifetime, compensated with a total royalty payment of 180,040 francs. During his lifetime, Halévy received 50,800 francs for 328 performances of La juive. Auber received total royalty compensation for La Muette de Portici of 15,300 francs for 154 performances during his lifetime. He composed four other operas of four acts or more. Unlike Meyerbeer, Auber also composed shorter pieces-not considered as 'grand' operas. Consequently, his total royalty income was 86,800 francs for 785 performances of grand and not-so-grand opéras.
The real losers were, of course, those composers whose pieces were rejected by the opera's scrutinising committee. Ferdinand Hérold (1791-1833) may have been among them. He had several operas performed at the Opéra-Comique but at the royal opera only his music to three short ballets were accepted. Some operas by a few other composers were accepted but performed only a small number of times. Hector Berlioz's (1803-1869) opera Benvenuto Cellini was performed six times after its first performance in 1838 with total royalties of 1420 francs. 6 In 1851, Jakob Rosenhain (1813-1894) had his opera Le Démon de la Nuit performed four times, with 680 francs compensation. Emanuele Biletta's (1825-1890) opera La Rose de Florence was performed five times in the autumn of 1856 for a compensation of 850 francs.

The shift from the tariff to the percentage system
When the first GD tariffs were declared by the king in 1713 the opera venue, Salle du Palais-Royale, had a capacity of 1270 spectators. The Opéra Garnier had only just opened at a time in which there were signs of a transfer to the percentage system. It is still in use and has 1900 seats, succeeding the 1800-seat Salle Le Peletier. Both of these venues had audience capacities of approximately 50% more than the Salle du Palais-Royale.
The tariffs had more than doubled from 1713 to 1860. Despite this, it became increasingly obvious that the percentage system used in other theatres would be beneficial for composers. In the data from the Paris Opera found for this study, two pieces without music were compensated using the percentage system. The playwright Sophie de Bawr's one act comedy La suite d'un bal masqué was compensated with 303 francs, 6% of the gross box office revenue, as one of four pieces performed on 16 May 1857. On 15 December 1859, Eugène Scribe was awarded 994 francs, 8% of the gross box office revenue, for a performance of La Dame blanche. He had written this libretto for a production, with music by Boieldieu (whose right for compensation had ceased as he died in 1834), at the Opéra-Comique-hence the exception from the tariff system. From the Opéra-Comique, Scribe was already used to an 8.5% royalty in 1831 (AN code AJ13, 1074). 7 When Ambroise Thomas's opera Hamlet was performed for the 66th time on 7 January 1870, with Swedish soprano Christina Nilsson in her most celebrated part as Ofelia, the gross box office income was 11,906 francs. The 1860 tariff, Table 4, stipulated a maximum royalty of 500 francs, equal to only 4.2% of the gross box office revenue that night. A contemporary British theatrical manager has data for Meyerbeer's box office revenues: 'The maximum receipt of the Grand Opera House is about 12,000 francs … which sum was taken nightly for the first 100 representations of Meyerbeer's "Africaine"' (Maynard 1867, 28). This, too, means that a performance for a full-capacity audience would bring only 4.2% of the box office revenues to the originators-in this case Meyerbeer and Scribe.
Perhaps it was such examples that in 1879 enticed 15 composers and librettists, as representatives of the SACD, to present a demand for a transfer to the percentage system for the Paris Opera. The standard percentage was set at 6.5% (to be split among composers and librettists) 'of the gross recette, including subsciptions, box rental per annum or per day, and all tickets sold for or in any way attributed to that night' (AN code AJ13, 221, my translation). Among the SACD representatives were Alexandre Dumas (fils), Ludovice Halévy, Charles Gounod and Léo Delibes. However, the archived manuscript copy is not signed so we do not know for certain that it was actually agreed upon by Auguste Vaucorbeil, the opera manager, at this stage. As a struggling composer himself, he was most likely in favour of the demand. Moreover, as he was elected president of SACEM at least once he must have had the trust of his fellow composers. His obituary in Le Matin (1884) described him as 'very eager to be always faithful to the laws of the most strict justice.' A contract with the SACD was certainly signed by Vaucorbeil on the 12 May 1884 (AN code AJ13/1184).
It is likely that Vaucorbeil and the SACD were influenced by the Paris Convention for the Protection of Industrial Property, which had been signed by 11 countries the year before. It was followed in 1886 by the first intellectual property treaty to cover music-the Berne Convention for the Protection of Literary and Artistic Works.

Conclusion
Naturally, only leading composers, as perceived by contemporary audiences, had their operas staged at the Paris Opera in the nineteenth century. The GR remuneration they received from the management was substantial in some cases. After the introduction of the 'grand opéra' in 1828, the most prosperous composers were Meyerbeer, Halévy and Auber. They were the winners by far. However, their operas are rarely performed today while, for instance, Donizetti's Lucia di Lammermoor is staged more frequently (20th most played opera globally in 2018-2019 according to operabase.com). During Rossini's lifetime, Guillaume Tell was performed at the Paris Opera 311 times between 1833 and 1866. It was shown 18 times by 6 opera companies during the 2018-2019 season according to operabase.com. Rossini's most popular opera at present, Il Barbiere di Siviglia (5th on the operabase.com list for 2018-2019 with 539 performances worldwide), was performed only 5 times in Paris during his lifetime, while the operas that he composed after his move to Paris in 1824 were performed 682 times.
Of course, the losers were, primarily, those who composed operas that were rejected. Hector Berlioz, Jakob Rosenhain and Emanuele Biletta, had operas staged but with only a few performances each. For the 487 performances of Robert le diable, Meyerbeer earned 40 times what Berlioz received for seven performances of Benvenuto Cellini.
In contrast to current operatic programming, which favours pre-WW2 pieces, 93.2% of the performances in this study were by contemporary composers.
Although not explicitly mentioned in the Réglements, royalties were only paid to composers residing in France. Bilateral or multilateral copyright treaties with other countries were not signed until after the period studied here.
The tariff-based royalty system used by the Paris Opera seems to have been an exception to the percentage-based alternative that was more commonly used in other venues. By 1884, this had also become the standard for the Paris Opera. As little has been archived in Paris for the percentage system another state-run opera company will be the focus of a presentation of its effects in a forthcoming article.