One of the first occasions I saw an opera lit (or partially lit) with candles was a staging of Rameau's Castor et Pollux at the Estates Theatre in Prague. Despite many imperfections, the effect – pardon the upcoming pun – was electric: the upward throw of light from the footlights, the sideways throw of light from the torchères and the variety of shadows created in the upper area of the proscenium were a revelation. Until recently, an often-given solution to historical theatre lighting was to invoke the Drottningholm candle (and related technologies), a device with a synthetic flame that hovers according to the heat generated by the light source, used at the baroque theatre outside Stockholm. One visitor has commented:

Magic? Well, not really. And the introduction of fibre optics has altered things again. At Český Krumlov, the haunting and more spectacular theatre built by the Schwarzenberg family at their seat in the Czech Republic, similar candles are in use (though are not always on!) throughout the auditorium, to similar hypnotic effect. But at Český Krumlov the pit is still lit by real candles, and there is absolutely no mistaking the difference in the quality of light the two different sources produce. And to be clear: my lust after candlelight is not an antiquarian search for ‘authenticity’, but a search for answers to questions about the staging of opera in now non-existent buildings (especially those in London), and an examination of the assumptions of commentators – particularly modern ones – about just what can be seen of and therefore ‘read’ into the creators’ intentions, whatever we may believe these to have been.

As always, modern performers pick and choose the aspects of historically informed performance they wish to observe, and rightly so: no one, for example, would attempt to recreate the extreme discomfort of the overcrowded pit in an eighteenth-century London theatre, with its noise, smells and outside privies, usually round the corner and in the dark. But concerning the stage, where modern efforts toward historically informed performance have been concentrated, we seem happy to put an early-music band in the pit, employ gesture with much preparation and convincing effect, and use dramatic costumes with enough feathers to make one fear for all avian species, yet show almost no interest in presenting those elements of production with anything like credible eighteenth-century lighting. There is simply no general appetite for the darker stage with a lighter frame, with the scenery in shadow and the grand costumes catching light at the front of the stage. No wonder ‘authentic’ gesture often looks ridiculous, frequently resembling an over-active clergyman semaphoring from a pulpit.

Not for the first time, though, we are confronted with a system that the eighteenth century itself may well have thought was imperfect. The ballet master Jean-Georges Noverre was less than complimentary about the available mechanisms used to light the theatre:

Footlights, then, do not show the players at their best, or indeed at all. ‘But how to light the proscenium from on high and by the sides?’ (‘Mais comment éclairer le Proscænium par le haut & par les côtés?’; Observations, 29). Noverre suggests that the angle of the proscenium columns be adjusted towards the auditorium, that light in the pit be increased by a polished material curved to reflect the maximum amount of illumination toward the stage, and that the painter be given control over how the wings are lit.

These remarks, modest enough but highly revealing, represent some of the few incisive comments on stage lighting that have come down to us. Among the others are those of Pierre Patte, who in July 1781 published his ideas on lighting in the Mercure de France, and who would later go on to write the Essai sur l'architecture theatrical. But as Petr Peřina has pointed out, there is little specific mention of theatre illumination either in commentaries on the theatre or in documents of reform (Petr Peřina, ‘Candle Lighting in Theatres in the 17th and 18th Centuries’, in The World of Baroque Theatre, ed. Pavel Slavko and Hana Srbová (Český Krumlov: Society of Friends of Český Krumlov, 2007), 131). And what there is tends to relate either to expenses (‘I shall receive all such monies as shall be above the actual expenses of Rent, Lighting, Printing, Carpenters, &c’; W. M. Balfe, English Opera House (London: W. S. Johnson, 1841), 5) or to the auditorium (‘nor is there any Reason the Managers should be obliged to pay an exorbitant Sum nightly for lighting the House’; Anonymous, A Fair Enquiry into the State of Operas in England (London: printed for M. Cooper, 1759), 6).

Noverre's candid views on lighting were expressed in a book advocating the building of a new opera house on the Place du Carrousel, and they are among other specifications for the proposed theatre that give us an idea of the shortcomings he found in the buildings in which he worked. His views touched most areas of the theatre, both inside and out: the new structure needed to be wide enough to accommodate the creativity of ‘the painter, the ballet master, the composer and often the poet’ (‘le Peintre, le Maître des Ballets, le Musicien, & souvent le Poëte’; Observations, 2); it should have ‘a great foyer, adequate to practise ballets, when the theatre was crowded’ (‘un grand Foyer, propre à faire répéter des Ballets, dans les instans pressés’; Observations, 9); and the backstage areas had to be adequate in terms of both size and access for the requirements of the ballet master, with areas on either side of the stage having enough space for the crew, actors, dancers, chorus and the walk-on parts. Being able to manoeuvre in the space would enable the theatre personnel to perform as an ensemble, with ‘a precision, a variety in the effects and a silence that has never existed before at the Opera’ (‘une précision, une variété dans les effets & un silence qui n'ont jamais existé à l'Opéra’; Observations, 10). The audience had to be seated comfortably, and not with a view behind the sets that would spoil the illusion. In essence Noverre, like every theatre promoter since the dawn of time, wanted space.

Noverre's Observations had, however, been inspired by a particular event not unrelated to theatre lighting: this was the burning on 8 June 1781 of the theatre in which he had worked from 1774 to 1779. Opened on 20 January 1770, the theatre – the first purpose-built opera house in Paris – formed a wing of the Palais Royal and was designed by the architect Pierre-Louis Desproux. The 1781 fire gutted the structure: a painting entitled ‘Incendie de l'Opéra’ by Hubert Robert (nicknamed ‘Robert des Ruines’ after his penchant for representing remains in a very romantic manner) shows nothing left but the outer walls of the stage, proscenium and auditorium. Noverre did nothing that was not self-interested, and there seems little doubt that the Observations was an attempt to ensure that he was consulted on or in some way part of the rebuilding of the Paris Opéra. The dangers of fire were (and are) well known, and indeed my evening at Prague's Estates Theatre had emphasized the dangers of the eighteenth-century theatrical experience, with my architect father (really a fan of Ponchielli rather than Purcell) spending much of the performance looking out for fire exits. In eighteenth-century accounts of theatres there are many reports of ‘smoulderings’, small fires and even medium-sized blazes, and there were dancers whose dresses caught the footlights, singers whose wigs were set on fire by brushing past the candles mounted on the flats and stage hands who had accidents of all types behind the scenes. Few eighteenth-century theatres lasted longer than twenty years; that the London theatre buildings of Drury Lane, Covent Garden and the King's had had a combined life of some 280 years when they burned down between 1789 and 1808 was a remarkable feat of endless vigilance on the parts of the lighting crews and theatre staff.

Noverre was well aware of these dangers, and in the disposition of his model theatre's spaces, safety came first, for ‘the fortunes and the lives of citizens in danger, the masterpieces of the fine arts devoured by flames and the lives of the men that cultivate them, are all exposed to a danger that is constantly present’ (‘la fortune & les jours des Citoyens en danger, les chef-d’œuvres des beaux Arts dévorés par le feu, & la vie des hommes qui les cultivent, exposée à un danger sans cesse renaissant’; Observations, 2). He favoured an island site that made the building accessible on all sides, while his model of two pavilions – one for the auditorium and one for the personnel and backstage areas linked by bridges – was crucial to his approach to fire fighting. It placed the flammables relating to set building, scene painting and costume making in one space, and the theatre lighting in another. Thus a fire started by stage lighting in the scenery would not be given a boost by the paint, turpentine, sawdust and material off-cuts in the scene and costumes shops. In the courtyard created by the space between these two buildings Noverre specified there would be pumps and water tanks for use if a blaze broke out. And the two rooms on each side of the theatre, with access to the stage but used for storage, would also contain ‘water tanks and mobile pumps’ (‘reservoirs & des pompes de mobiles’; Observations, 11). If firefighters – Noverre recommended a team of eight – were always kept on duty and at the ready throughout performances, they would be able to ensure that all tanks were kept full and react to an outbreak of fire at an early stage. And Noverre was not content simply to provide these advantages, but proposed that they be tested on a regular basis. The firemen must ‘make an inspection round every evening after the rehearsals and performances’ (‘faire une ronde d'inspection tous les soirs après les Répétitions & les Représentations’; Observations, 17), sleep in the building in rotation, inspect the pumps every fifteen days and renew all the water in the tanks once a month. These measures, together with the instruction to the architect that all the doors should be accessible from outside, and should open outwards, shows Noverre to be ahead of his time; many lives would be lost before such things became the norm in the management of theatre buildings.

Unlike nearly every other eighteenth-century commentator on theatre buildings, Noverre had the opportunity to see at least some of his ideas put into practice. When he arrived in London in November 1781, just after his Observations had been published, it was to work at the King's Theatre, the auditorium of which had remained essentially the same since 1709. The theatre's 1778 alterations had given the stage more space, for in removing ‘the heavy columns which gave the house so gloomy an aspect that it rather resembled a large mausoleum or a place for funeral dirges, than a theatre’ (The Morning Chronicle, 25 November 1778), the new owners had increased both the stage space and the size of the proscenium. In this new form, the stage now fitted the brand of dance that Noverre promoted, the ballet d'action. This genre was grander and more elaborate than most theatrical dance seen in previous London seasons, with stories that involved more dancers, more gestures and ‘a wider palette of emotions’ (Michael Burden, ‘Visions of Dance at the King's Theatre: Reconsidering London's “Opera House”’, Music in Art: International Journal for Music Iconography 36/1–2 (2011), 122). And while the ballet d'action did not entirely replace plotless divertissements, it ‘abruptly and radically altered the use of dance. . .in the opera house’ (Judith Milhous, ‘The Economics of Theatrical Dance in Eighteenth-Century London’, Theatre Journal 55/3 (2003), 482).

But even these alterations did not provide enough space for Noverre's ambitions, nor did they answer the ideas about space laid out in the Observations. When he was booked to return for the 1782–1783 season, the summer months were spent altering the stage area again. This time, the alterations were in the hands of the architect Michael Novosielski, who was also the theatre's set designer. He lengthened the forestage and increased the width of the auditorium by ‘building corridors and box access outside the walls of the original building’ (Curtis Price, Judith Milhous and Robert D. Hume, Italian Opera in Late Eighteenth-Century London: The King's Theatre, Haymarket, 1778–1791 (Oxford: Clarendon, 1995), 70). All in all, these alterations were reputed to have cost between £8,000 and £10,000, and there can be little doubt that they were done to Noverre's specification. But all this effort was in vain. Noverre did reappear for the next season, but not to dance: he arrived ‘for the purpose of recovering some debts he left due to him on his last departure from this Country’ (Public Advertiser, 14 December 1782). If, as seems likely, the money was due from the opera house manager William Taylor, there was clearly no incentive for him to stay and take up his position! This was doubtless frustrating for the theatre, but it is also frustrating for us, since it denies us the possibility of finding out what Noverre made of these alterations in practice.

Noverre's version of the King's Theatre went the way of the Paris Opéra: it burned to the ground in 1789. All of the establishment was destroyed with the exception of the room over the arcade to the pit door. (Facing the Haymarket, the room was used successively by Jean Marie Léger as the opera dance academy, by one M. Ridaut to teach swordsmanship and by Henry Angelo for a fencing school.) The fire broke out ‘five minutes before ten o'clock’, when the performers were ‘practising a repetition of the dances’ on a non-performance night; it began at the top of the building in the flies, or perhaps in the scene room, and the ‘first notice they had of the mischief was sparks falling on their heads’ (The Times, 18 June 1789). Angelo's own account of the blaze starts while he was walking through Berkeley Square. From there he saw black smoke ascending, and hearing the fire was in the Haymarket, he hastened in that direction. On reaching that street, he recounted that

Angelo's dramatic account, with its exciting narrative of rescue, looting and loss, ends with what was for him a final reprieve: the fire engines, placed at the theatre's frontage in Market Lane (the narrow street running from Pall Mall to Charles Street, and parallel to the Haymarket), prevented the fire from spreading across the street from the opera house. Had it done so, it would probably have spread further westward to the next street over, St Alban's Street, where Angelo had his lodgings, rooms now containing the rescued portrait of Joseph Bologne, Chevalier de Saint-Georges, by the American artist Mather Brown.

So ended the theatre that had served London so well since 1705. The house in which many of Handel's operas had been premiered, in which star singers such as Senesino, Mingotti and Pachierotti had performed, and in which Dauberval, Vestris and the Simonets had danced, was now a gutted shell. Had Novosielski acted upon Noverre's other guidance in the Observations and installed the recommended water tanks, constructed proper divisions between the production departments and rostered the all-important fire-watchers, the theatre could have survived, and it is possible that there might today be a building through which we could understand the performance of Italian opera and elite dance in London.

The question is: just how much notice in a practical sense should we take today of Noverre's Observations? Certainly the author had an agenda and was known for his own self-promotional abilities. But as the Journal de Paris noted, his propositions were taken seriously in the eighteenth century precisely because he was a practitioner, not simply an architect with interesting ideas or a philosopher wanting to centre the opera at the heart of the city:

In a survey of writers on theatre buildings published in the following decade, George Saunders made similar remarks, noting that in his Observations Noverre had ‘exposed the faults of stages in general, and pointed out their remedies; and these faults are inseparable from every edifice of that kind’ (George Saunders, A Treatise on Theatres (London: author, 1790), ix).

Noverre did not, of course, provide us with all the pointers we need in understanding the past or in applying his ideas to performances of eighteenth-century opera in the twenty-first century. Indeed, his disclosure that ‘as for the cost, it would be considerable, no doubt; but it is not up to me to calculate the resources and to determine the use of the public purse’ (‘Quant à la dépense, elle seroit considérable, sans doute; mais ce n'est pas à moi à calculer les ressources & à fixer l'emploi de la richesse publique’; Observations, 36) would now attract ridicule from any funding organization considering a budget proposal based on such an airy and grandiose approach. But the minutiae of his remarks on lighting chime with those of one modern commentator, who has suggested that ‘research into historically accurate lighting. . .needs to be absorbed into a more robust approach to historically informed lighting on today's stages’ (Francis Reid, ‘Historically Informed Lighting in Baroque Opera Staging’, in The World of Baroque Theatre, 156). We could manage this by following Noverre's own precepts:

And, indeed, many scholars and practitioners, including Petr Peřina, have followed Noverre's advice to the letter, conducting experiments that have yielded revealing results concerning burn rate, smoke production and different qualities of light, depending on the composition of the candle. But Noverre's remarks also lead to Reid's conclusion (‘Historically Informed Lighting’, 162) that modern lighting should be based on ‘best practice’ rather than ‘average practice’, and with the key aim that it ‘should embrace all the visual results from baroque technology, both established and experimental’.