The two parts of Davenant’s The Siege of Rhodes launched the new theatre at LIF in June 1661. The first part had been published and performed as one of Davenant’s recitative ‘operas’ in the 1650s, but part two may have been receiving its premiere. Part two was first published in 1663 along with a revised version of part one; however, it appeared on the Stationers’ Register on May 30, 1659, and the published text may have no LIF connection. Nevertheless, it is instructive to compare the scenic demands of this text not only with those in the first part of Rhodes, but also with those in at least two of Thomas Killigrew’s pre-Restoration plays, Bellamira and Thomaso. Both future patent holders are visualising potential production on a scenic stage. Davenant’s vision in Rhodes 2 is conservative, there are no major advances in staging over Rhodes 1, but there are suggestions in the text that Davenant is thinking of a scenic stage with more flexible arrangements than those available for the Commonwealth productions of Rhodes 1. In general, though, the scenic resources called for in Rhodes 2 are of the same type, if not dimensions, as those supplied by Inigo Jones and his assistant John Webb for Davenant’s masque Salmacida Spolia in 1640. In Bellamira 2, however, we encounter something new. This play was written in Venice in the early 1650s during Killigrew’s exile, perhaps with little thought of realisation, but its stage directions suggest sophisticated, Italian scenic arrangements, rather than anything that would be found on simpler English stages in the 1660s. For example, the discovery scene in 3.1 implies an area upstage of the backshutters deeper than that available in the limited relieve spaces of any stage design by Jones or Webb. The associated stage direction reads:
The Scene opens and discovers a Prison, where Pollidor and Phillora appear next the Stage chained to a Ring fastned to the ground [;] upon the other side of the Prison, and in a darker part of the Scene lies Palantus chained behinde them in the dark, Bellamira chained, and afar off in prospective other Prisoners and dead Carcases.
This direction with its assumed depth, layered arrangement of actors, and perspective scenery would be difficult to accommodate in either the 7 feet 7½ inches behind the Hall shutters, or in the 8 feet 2½ inches of Graham Barlow’s LIF model, but it would present fewer problems on the highly flexible stage of the Teatro SS. Giovanni e Paolo, Venice (1639). A plan of this theatre, with its several shutters dispersed along the length of the stage, may be found in Allardyce Nicoll’s The Development of the Theatre. Looking at the plan with Bellamira in mind it is surprising that Killigrew, who presumably had ample opportunity to visit this theatre, did not become the greater scenic innovator of the two patent holders. Bellamira may have been written as a closet drama, but its scenic implications anticipate Restoration staging of the 1670s, rather than anything Killigrew achieved in the 1660s. By contrast, Davenant’s The Siege of Rhodes 2 anticipates exactly the kind of scenic staging he was to produce in 1661. While the text cannot be used to infer resources at LIF, it does not call for anything that could not be realised on a basic scenic stage. However, two stage directions in particular suggest an advance on the Rutland House production: “Enter Ianthe and her two women at the other Door”, and:
The Scene is Chang’d./ Being wholly fill’d with ROXOLANA’S Rich Pavilion, wherein is discern’d at distance, IANTHE sleeping on a Couch; ROXOLANA at one End of it, and HALY at the other; Guards of Eunuchs are Discover’d at the wings of the Pavilion; ROXOLANA having a Turkish Embroidered Handkerchief in her left hand, and a naked Ponyard in her right.
Taken in relation to the earlier play, which makes no reference to doors in the stage directions, and which was staged throughout with fixed wings, the first of these directions may imply forestage doors, and “wholly fill’d” in the second hints at the use of changeable wings. As Ann-Mari Hedbäck suggests, the two references to doors in part two may indicate that the printer’s copy was a manuscript connected with a pre-Restoration performance (albeit one for which no record exists). However, they may also suggest that by 1659, or thereabouts, Davenant had already decided on the form of his future scenic stage (if he had not done so already for his aborted Fleet Street theatre of 1639). If we accept that the second suggestion is plausible, these two directions may be seen as indicating the nature of the future LIF stage: a forestage with doors of entrance, fully changeable scenery, and a separate discovery or relieve space (Ianthe “discern’d at a distance”).
The contrast between the depths of Killigrew’s imagined discovery space and Davenant’s carefully delimited area – one actor on a couch and one at either end – suggests another index we may use to check the universality of the proposed LIF model. An exact measurement of the LIF relieve/discovery space is of course impossible, but figures for other stages may be used as a guide. To add to the figures noted above for the Hall stage and Barlow’s model, the relieve area on the larger Salmacida Spolia stage, for example, had a total depth of 8 feet 7 inches (in each case the relieve area is approximately one third the depth of the whole scenic stage). Sightlines obviously determine the extent to which an audience member would be able to see into such a relieve area. At best, the view is a rectangle bounded by the rearmost wing edges and the background scenic element; at worst, in a side seat, one of the corners would be lopped off. For the purposes of this study I assume the optimal viewpoint. It is helpful at this point to recall the two other indices of universality for my LIF model: no more than two forestage doors in any one scene may be indicated as practical, and discovery/relieve scenes cannot occur successively; also to restate the main scenic specification, the model allows the loading of three wing settings and four backscenes (three backshutter pairs and one relieve), but additional settings may be accommodated by replacement (removal) during act breaks. Should more than three wing settings or four different backscenes be required mid-act, settings may be replaced at scene changes, but there are obvious practical limitations involved should multiple changes be demanded. The evidence from Restoration promptbooks is that while scene keepers may have simplified an author’s demands depending on their scene stock, it was standard practice to match as far as possible appropriate scenery to locations stated or implied in the play text.
 Comedies and Tragedies, London: Henry Herringman, 1663, p.542. Act 2.1 of Killigrew’s Thomaso (p.326) calls for the backshutters to open and discover a piazza and several practical balconies.
 The Hall figure is stated on Webb’s plan (see Orrell, The Theatres of Inigo Jones and John Webb, p.174). Barlow places his backcloth approximately 10 feet 3 inches from the rear wall of the theatre, the relieve depth stated above is therefore the gap between the shutter frame and the backcloth. Barlow does not provide a specific reason for placing the backcloth where he does, but the relieve depth is in proportion both to this stage and to the other stages discussed here (see ‘From tennis court to opera house’, unpublished PhD thesis, University of Glasgow, 1983, vol. 3, fig. 16).
 London: Harrap, 1966 (5th ed.), p.169.
 Killigrew had a semi-official position in Venice as Charles Stuart’s Resident, but as Alfred Harbage suggests he probably “found ample time for his own diversions”, including writing Bellamira (Thomas Killigrew, New York: Benjamin Blom, 1967 , p.94).
 Hedbäck, ‘The Siege of Rhodes: A critical edition’, Studia Anglistica Upsaliensia, 14, Uppsala, 1973, p.56, 80.
 See ibid. p.xxiii.
 On the Salmacida plan Webb has marked two dimensions in the relieve area: 3ft 10in from the backshutters to the front of the vertical support for the cloud machine, and 3ft 9in from the rear of this support to the backcloth or board. He has also marked 1ft as the width of the support giving a total of 8ft 7in. This last figure may also be obtained by scaling from Webb’s stated measurements; hence Richard Southern errs when he states this space to be “nearly 7 ft. deep” (Changeable Scenery, London: Faber & Faber, 1952, p.69). The best reproductions of the Salmacida plan and section are to be found in Stephen Orgel & Roy Strong, Inigo Jones: The Theatre of the Stuart Court, 2 vols., London and Berkeley: Sotheby Parke Bernet and University of California Press, 1973, pp.738-41.
 See also my analysis of Guzman on this blog and Peter Holland’s discussion in The Ornament of action (Cambridge University Press, 1979, pp.45-6).