1 Though one should bear in mind the Choral Dances from Gloriana and the dance element of Death in Venice.

2 For example we find him writing to Erwin Stein (his publisher, at Boosey & Hawkes) on 13 November 1957: ‘I was delighted with your sweet letter, but please don't think I was cross with anyone particular about the Ballet proofs. I was only just cross in the abstract to have to go back to that beastly work, of which at the moment I am heartily sick. The maddening thing is that after we have all spent hours reading it there will quite clearly remain dozens of mistakes. I frankly don't know what we can do, but I am clear at the moment I don't want any more to do with it! But I must not be silly, and your nice remarks about it make me feel that the work was not just a waste of a year's work’.

3 Britten was also reluctant to approve the idea of a suite, to be drawn from the ballet by another and sympathetic hand, on the grounds that this was something he eventually wanted to do himself. But, though he may have contemplated it, he never got round to doing it. This meant that it was, in the main, only the Pas de six—the final set of dances from Act III scene 2—that was heard in the concert hall, and that infrequently. It was given a separate opus number, Op. 57a, but was available only on hire, presumably because everyone was waiting on the appearance of the composer's own, or other authorized, suite, which was never to materialize. Not long before his death in 1976 Britten at last authorized the publication of a suite, taken from the ballet by Norman Del Mar. While this is more extensive than opus 57a, it is, in a sense, more ‘Prince’ than ‘Pagodas’ since, for practical reasons—the extra percussion required—it excludes the Pagodas’ music. The ‘Del Mar’ Suite was given a first broadcast performance on 7 December 1963 by the BBC Scottish Orchestra conducted by Norman Del Mar. The first concert performance took place at the Edinburgh Festival on 29 August 1964. A study score of the ‘Del Mar’ Suite, , Prelude and Dances from the Prince of the Pagodas (Op. 57b: Boosey & Hawkes, 1980 Google Scholar [HPS 919] constituted the first publication of any of the music from the ballet. (The first recording of this suite was issued in 1982 by EMI [ASD 4073]). A much more comprehensive suite, which does include the Pagodas' music, has been devised by Michael Lankester, who conducted its first performance, at a BBC Promenade Concert, on 21 July 1979. The materials for this suite can be hired from the publishers, Boosey & Hawkes. There is also at least one other concert suite, compiled by André Previn, and other conductors may have made individual compilations that have reached performance but have not otherwise been documented.

4 Originally issued by Decca on two discs, LXT 5336–7; now available as a boxed set, GOS 558–9.

5 Edited from my contribution to The Decca Book of Ballet, ed. Drew, David (Decca, 1958)Google Scholar.

6 ‘Variation’ in its choreographic, not strictly musical, sense. But while it is indeed diversity that the Divertissement dazzlingly unfolds, there is room too for inter-dance sharing of some powerful melodic gestures and above all for some thematic transformation which directly relates to the dénouement that the Divertissement actually celebrates. For example, there can be no doubt, in my view, that in the Entrée we are (pace Evans, Peter, The Music of Benjamin Britten, [Dent, 1979] p. 234 Google Scholar) expected to hear the motif which the oboe gaily introduces at Figure 27—

–as a new diatonic version of the Salamander motif (Ex. 1) After all, the Salamander has just since shed his green skin; and what could be more appropriate—or symmetrical—than the appearance, at the very beginning of the concluding Divertissement that signifies the Prince's final release, of his old motif, first revealed in the Prelude to Act I, in a new—hisnew—guise? No more clusters, no more heterophony: all that has been jettisoned along with the Salamander and the Pagodas. ‘Transformation’ is precisely what we have immediately experienced; and this familiar motif, given a fresh format, is a delicious and effective way of making the point.

7 The coda heard at the first Covent Garden performances was later deleted by the composer and a new number substituted.

8 I refer throughout to the rehearsal letters and figures that appear in the piano reduction of the full score and in the full score itself.

9 An ostinato seldom encountered. But there is a precedent in the last song of Mahler's Kindertotenlieder, well known to Britten.

10 The only writer who, to my knowledge, has noted the implications of this passage is Evans, Peter (The Music of Benjamin Britten, p. 225)Google Scholar. But then he is virtually alone in having written anything serious at all about the music of the Pagodas.

11 Cf. Act II Figure 74f., where the clusters reappear and have a very important role to play and where the Salamander music itself is incorporated into a fully fledged gamelan texture.

12 It is hardly surprising, given Britten's own obsession with scales, that it was this number that so powerfully appealed to him.

13 A rather amusing though no doubt unconscious quotation is to be found in Act III at Figure 16f. (cf. also its repetitions 16 bars later and most extensively at Figure 18f.), where Britten shows that for all his well-known looking down his nose at The Rake's Progress, he had stored away a memory of Stravinsky's brilliant prelude to Act III of the opera.

14 The only significant stretch of canonic writing I have found in the Pagodas.

15 Britten would have assimilated this particular Balinese practice in the first instance from his acquaintance with Colin McPhee's two-piano transcriptions of Balinese Ceremonial Music, even before hearing it for himself on the island.

16 Cf. Pagodas Act II, I bar after Figure 74f., and Death in Venice Scene 16 Figure 3Olf.

17 I have already pointed out (above, fn. 6) that the chord-clusters initially emerge from the Salamander music as it first appears in the Prelude. They are as it were the vertical manifestation of the impact made on Britten by Balinese music, of which just such clusters are a prominent feature. They are also the direct predecessor of the similar cluster which identifies Tadzio in Death in Venice.

18 (New Haven: Yale University Press), Ex. 337, pp. 348–9. See also Douglas Young's sleeve-note for East-West Encounters, Cameo Classics GOCLP 9018(D), a most valuable source of information about McPhee. Britten must have met McPhee not long after his arrival in New York in August 1939. At the back of his pocket diary for that year we find scribbled there: ‘Colin McFee [sic] 129 East 10. Algonquin 4–4980’. The registering of that address and telephone number undoubtedly marked the beginning of their friendship. McPhee died in Los Angeles, where he was teaching at the University of California, on 7 January 1964. It is possible that the original gramophone recording of the ballet came to his attention. Further evidence has come to light of the close association of the two men during Britten's years in the United States. It was McPhee who prepared an ingenious and skilful transcription for two pianos of Britten's Variations on a Theme of Frank Bridge, which was used for performances in New York in 1942 Google Scholar of a ballet, Jinx, presented by the Dance Players at the National Theatre and first performed on 24 April. (We must remember in this context McPhee and Britten as performers at two pianos of the Balinese Ceremonial Music). The manuscript of this transcription, now in the Britten-Pears Library at Aldeburgh, is dated ‘Feb-March 1942’. A dyeline of a copyist's copy (also in the Library) was clearly used for performance: on it appear cues related to the dancers. ( Balanchine, George and Mason, Francis, Festival of Ballet [W. H. Alien, 1978], pp. 326–7Google Scholar, give a detailed account of the story of the ballet. It was revived by the New York City Ballet in 1949.) There is also an intriguing undated letter of Britten's from these yeats, drafted for him by Elizabeth Mayer and addressed to David Ewen, an American popular encyclopaedist. Ewen had evidently asked for information about Britten and his music, to which the composer replied: ‘Of course I shall be delighted to co-operate with you in your new book. Unfortunately I have so far been unable to obtain copies of the best articles written about me. They were published in periodicals in England some time ago & I am afraid I have not got them with me. But Mr. Colin McPhee is engaged in writing a comprehensive survey of my work at the moment, which Arden Music is considering using sometime—but at any rate not before the fall—& anyhow there would be no objection to you using it I know'. Presumably, and regrettably, McPhee's study was never completed and never published; perhaps Britten's departure from the States in 1942 killed off the project.

19 Britten's manuscript notes of scales and fragments of typical Balinese gamelan figuration and textures, including indications of instrumentation, are preserved in the Britten-Pears Library. At the top of one page is written ‘Kapi Radji (Overture)’; then follows this notation of the scale on which the particular piece Britten had heard was based:

(See also The Music of Benjamin Britten, p. 234, where Peter Evans has correctly deduced the scale.) Britten also wrote out. as part of the same sketch scraps of motifs and rhythmic figuration and indications of the basic pulse or beat. At a later date, clearly, he looked through these notes in order to locate something appropriate for the ballet: under the notation given above appear the underlined words ‘This for beginning of Pagoda scene’; and it was indeed by the scale and the subsequent outline of motif, rhythm, instrumentation and texture that the music for the Pagodas was generated. Cf. Figure 71 f. with this transcription of Britten's on-the-spot, seminal sketch.

There is little probability that Britten would even have known of McPhee's Music in Bali. It is exceptionally interesting, however, that McPhee's 1966 excerpt from and commentary on the Balinese gambang style relate back to his two-piano transcription of the same Gambangan that he had published in 1940—one of the very transcriptions he played at that time with Britten. While there can be no doubt that Britten consulted not the almost forgotten transcriptions from 1940 when composing the Pagodas but his own manuscript notes from 1956, there can be little doubt that it was his unconscious memory of playing the 1940 Gambangan that influenced him to choose the very same music again from his own 1956 notations. I am much obliged to Douglas Young who has shared his thoughts with me about the history of this fascinating passage.

20 One recalls Erwin Stein's famous remark about Britten's discovery of ‘the sonority of the second’ in his discussion of the Sinfonietta ( Benjamin Britten: A Commentary on His Works by a Group of Specialists [Rockcliff, 1952, and Greenwood, USA, 1972] p. 249 Google Scholar). The piled-up clusters one finds in oriental music must have made a special appeal to a composer who had long been devoted to the smallest type of cluster. Thus in the Pagodas two favourite sonorities—one fresh, one of long standing—are integrated. Chains of seconds abound, most of them matching up to Stein's description of Britten's exploitation of them as ‘beautiful and tender’. Furthermore, the very first initiating chord of the work, the added sixth, might be thought of in this context as particularly appropriate—a chord as it were with a built-in cluster. See also Peter Evans's comment in his review of the full score of Death in Venice in Music & Letters 62 (1981), pp. 112–14CrossRefGoogle Scholar:

Britten's conversion, for long stretches of this opera, of harmony from an agency of movement into one primarily of motivic amplification ensure that one's ears soon become acutely sensitive to the harrowing ubiquity of certain intervallic complexes; and when the same nuclei are operating melodically, often in tenuously related two-part writing or a single part heterophonically tensed against itself (the logical final stage in Britten's lifelong addiction to the ‘sonorous second’), then ‘analysis’, whether or not verbalized, appears an unusually immediate, and a required, part of the listening process. [My italics.]

21 It seems beyond belief—certainly beyond my belief—that Kennedy, Michael (Britten: Master Musicians, [Dent, 1981] p. 214)Google Scholar can commit himself to the opinion that Pagodas ‘of all Britten's large-scale works’ is ‘the least characteristic in sound’. Almost any bar of the score demonstrably shows this to be the reverse of the truth. And this from a student of Britten's music! Scarcely less disconcerting is Walsh's, Stephen suggestion (Observer, 20 06 1982)Google Scholar in a review of the first recording of the ‘Del Mar’ Suite (see above, n. 3) that in comparison with Britten's work from the 1930's the ballet ‘is merely a work of effortless talent … more like a well-rehearsed high-wire act; the creative muscles are so attuned to it that it barely stretches them any more’. It seems curious to come to such a sweeping conclusion on the basis of the suite, which represents only a fraction of the total music and from which the Pagodas' innovative music is altogether excluded. But perhaps this is what Mr. Walsh means by writing, as he puts it, ‘tongue-in-cheek’.

22 I am thinking not just of the gamelan music, extraordinary though that is, both in its own right and as a marvellous example of Britten's ‘photographic’ ear: it seems hardly possible that so authentic a gamelan-like sound could by conjured out of the modern symphony orchestra. I also have in mind—but how does one choose amid such riches?—Variations I and II from the Pas de six, both of them built around very particular instrumental timbres and agilities (horn and piano in Variation I, solo violin in Variation II), both of them representing opposed extremes of colour and density (the one dark and heavy, the other light and weightless), both products of a common uncommon imagination.