Pioneering Haydn On Record
by Christopher Howell

An occasional series dedicated to commercial and live recordings of Haydn symphonies up to and including those of the first two complete cycles by Ernst Märzendorfer and Antal Dorati

Symphony no. 53 in D major “L’Impériale” Hob. I/53 (1777-79)

A pdf version of this article can be downloaded for offline reading and printing here.

The origin of the nickname “L’Impériale”, attached to Symphony 53, seems unknown. Neither Hodgson[1] nor Robbins Landon[2] offer any explanation. Even the dating is fluid – the Wikipedia articles vary according to which language you look at, the Italian article even predating it to 1774[3]. There are, moreover, four alternative finales. B and D are lifted from opera overtures by Haydn – under the labels Overture in D Hob.1a:1 and Overture in D 1a:4 respectively[4], they have enjoyed some independent life on record. Finale C is not thought to be by Haydn at all. Robbins Landon notes an autograph dated 1777 with finale B, an entry in the Breitkopf catalogue without the slow introduction, dated 1779-80, a version with the slow introduction and finale B in Seitenstetten Monastery, a version without introduction and with finale B published in London by Blundell “as performed at Messrs. Bach and Abel’s Concerts” [which ended in 1781] and a version copied by Elssler between 1779 and 1782 with finale A. Finale C was substituted by the Parisian publisher Sieber. The symphony was one of the first to gain wide popularity, not least in England, and according to Hodgson the version known there lacked the slow introduction and used finale D. Despite which, the piano (or organ or harpsichord) arrangement by one J. Bland published in London around 1785 “as performed at the Nobility’s Concerts” has the slow introduction and finale B[5].  

Whatever its purpose, this symphony, with its bright D major tonality and timpani joining the horns – except in finale C – certainly has a festive, celebratory air. Timpani without trumpets seems a little odd, but the trumpets could easily have doubled the horns, especially if there was an Empress listening. The slow introduction, even if an afterthought, is imposing and the first movement proper, marked Vivace, impresses with its rocking, arpeggio-based principal theme, appearing first on the cellos. It was the second movement, though, which earned the symphony its celebrity. This is one of the first, and one of the best, of Haydn’s double variation movements, in which a friendly, elegant theme in the major alternates with a song-like melody in the minor key, allowing a touch of gentle pathos. Double variations inevitably have a lot of repeats, eight in this case, and there is a sting in the tail here for a conductor who decides to omit them. The first variation in each key has the repeats written out, since the repeats themselves are varied, so the conductor who did not play the repeats in the first statements will end up with some odd proportions. On the principle that variations do sometimes shed repeats in their later stages, less damage will be done by omitting those in the last two major key variations (from bar 97).

The first recording was made in Paris on 24 April 1939, with Edvard Fendler conducting the Paris Conservatoire Orchestra on French HMV (DB5 131-2). Fendler (1902-1987) was born in Leipzig. After fleeing Nazi Germany and making a few records in Paris, his post-war career took him to out-of-the way centres such as Guatemala and Mobile, Alabama. He reconstructed this symphony and other rare 18^th century scores from orchestral parts found in European libraries during the 1930s. His edition of “L’Impériale” was published in 1949, leading to the American première of the symphony under Leopold Stokowski[6]. After an imposing introduction, his first movement Vivace can only be called manic – the sole object apparently being to get to the end, without repeats, before the wax runs out. The orchestra is pretty brilliant, I must say. The second movement likewise has no repeats, but here, at a lilting, two-in-a-bar Andante, there is elegance, charm and even a touch of pathos. The minuet goes at a breezy tempo, but not unreasonably so. There is substantial rephrasing of the melody, or did the parts discovered by Fendler contain this phrasing? The trio is somewhat bland. The finale – finale C – is manic once more, but this raises the question whether Fendler’s fast movements were manic and repeatless by choice and not just necessity. At 02:32 there was plenty of time to take a tempo at which the music could speak and maybe even give the repeat.

Set down on 24 April 1949 with “His Symphony Orchestra”, Symphony 53 is Stokowski’s only commercial recording of a Haydn symphony (RCA LM-1073), though live recordings have been found of 31, 45, 60 and 101. There is also a live performance of 53 with the New York PO, given a couple of months before the recording. His introduction is grand yet spacious, without the bullishness of Fendler. He takes the Vivace considerably slower than Fendler, proving it is quite swift enough for full-throated Beethovenian drive. No repeats and a four-bar cut near the end. Was this tiny snip needed to squeeze the music onto a longish 78 side?[7] The second movement is slowish and in this case there might be something to be said for Fendler’s more flowing tempo. Stokowski shades it beautifully, though, without exaggeration and much gentle expression. No repeats and two eight-bar cuts – one in the first major key variation, the other in the last. These were presumably needed to fit the music on a single side, but would Stokowski not have had sufficient whack to tell RCA the first two movements required two sides each? The minuet is deliberate, but with a jaunty lilt and a gracefully characterized trio. All repeats are played. I gasped when I heard how slowly Stokowski took finale C, but this allows the triplets towards the end of each section to be articulated with fierce clarity – evidently Stokowski felt that these triplets “fixed” the tempo. There is a splendid, almost Beechamesque, bonhomie to it. No repeat. There is more cause than not to regret that Stokowski did not come this way more often.

Stokowski’s recording was subsequently issued on LP with the incongruous coupling of Liszt’s “Les Préludes”. The first recording of the LP epoch was set down by the Vienna Symphony Orchestra under Paul Sacher on 14-16 April 1953 (Philips 642-108 DXL). Sacher (1906-1999) contributed notably to the history of music in the 20^th century by channelling a substantial part of his immense wealth into the commissioning of new works from such leading figures as Stravinsky and Bartók as well as many lesser-remembered figures from his native Switzerland. He had studied conducting with Weingartner and formed the Basle Chamber Orchestra in 1926, the Basle Chamber Choir following in 1928. He conducted the premières of most of his commissioned works as well as developing a healthy interest in music from the baroque and early classical periods. For better or worse, the recording companies seemed to have looked on his conducting activities as a rich man’s hobby. Recordings under his baton of works he commissioned and premièred sound as if they should have high documentary value, but they have to be sought mainly in European radio archives. His work in the recording studio, mostly for Philips, was largely confined to baroque and early classical works which, the recording companies perhaps thought, would go well enough without a conductor at all. The numerous recordings in the Swiss radio archives show, at the very least, that he was able to obtain sympathetic and nuanced performances of contemporary music in a middle-of-the-road idiom. Likewise in Haydn, while he may not have the absolute control over every detail of nuance and articulation that Stokowski achieved – this is a gift shared by few mortals – he gets spruce, vivacious playing from the VSO.

The Boston Haydn Society Urtext edition, edited by Helmut Schultz, is dated 1951, so Sacher might have earned a few extra kudos by using it. So far as I can judge without actually having seen the Fendler edition, this is what he is using. As with Fendler himself, there is notable rephrasing of the minuet and finale C – Schultz published finales A and B as well. Also, there are added wind parts, for example from bar 56 of the first movement and if any of the recordings heard so far, including Sacher’s, include the timpani part, I can only say the engineers kept it well out of earshot[8]. That said, Sacher manages an imposing but not inflated introduction and has the Vivace spin along nicely with plenty of dynamic shading. The first repeat is played. In the second movement, his tempo sounds ideally poised between those of Fendler and Stokowski, lending the music a sort of pawky grace.  However, without a Stokowski-like control over nuance, it is the pawkiness that predominates as the music proceeds. He plays the first two repeats and ignores the others except, oddly, for the penultimate one. The rather ceremonial minuet slogs at times but the finale, with the repeat, has plenty of drive. The tempo seems at first a little closer than Stokowski’s to what might logically be thought a Presto, but when the triplets arrive the playing is rather blurred, suggesting that Stokowski really had judged the maximum tempo at which this passage could be articulated clearly.

Thanks to the Cameo Classics/Itter Collection set, we can hear Sacher conducting this symphony live with the Royal Philharmonic Orchestra later the same year. Sacher enjoyed a brief but intense relationship with the RPO around 1953 beginning, it would seem, with Stravinsky’s “The Rake’s Progress” at Glyndebourne, for which the RPO was at that time the house orchestra. A performance was broadcast on 1st January 1953. Presumably it was this that led to a series of engagements towards the end of the year, broadcast on 11 November (Pergolesi, Hartmann, Veress and Haydn 39), 13 November (Bach, Honegger, Mozart) and 18 November. This last was given in the Royal Festival Hall for the Royal Phiharmonic Society. It began with the Haydn 53 issued by Cameo, followed by the Ritual Dances from Tippett’s “A Midsummer Marriage”, Mozart’s 4^th Horn Concerto (with Dennis Brain) and Martinu’s Concerto Grosso for two pianos and orchestra.

Presumably, Sacher is still using the Fendler edition, since we have the added wind parts at bar 56 of the first movement and the variant phrasing in the minuet and finale C. However, this time Sacher has timpani and trumpets. The latter are not allowed to dominate, they simply add a little richness to the forte passages. The recording, which sounds very acceptable for the date, defines the timpani sufficiently well to show that they are not playing the part in the Schultz edition. Moreover, timpani and trumpets play in the finale in which, according to the Schultz edition, they are silent. If you are going to have timpani and trumpets in movements I and II and conclude with finale C, it seems logical to construct parts for them, but who did this work?[9]

As for the performance itself, Sacher once again drives the first movement along with plenty of vitality and a fair amount of dynamic shading. Here, too, he plays the first repeat. In the second movement, he is a tad slower and the extra quality of the RPO enables him to produce a more expressive account. This time, he gives the first two repeats and no others. The minuet is more heavy-handed in London – one can imagine the players grimacing at the thought of what Beecham might have done with this. The finale, again with the repeat, bustles along merrily. In spite of the minuet, this seems to be the Sacher recording to hear. It does have the curiosity of presenting, trumpets and drums included, an edition nobody would be likely to play today, but which apparently draws upon early parts that, whether or not with Haydn’s approval, were heard in his own time.  

A 1966 recording by the Köln Kammerorchester under Helmut Müller-Bruhl, issued on Schwann, has eluded me, as has a version by the Vienna Baroque Ensemble under Theodor Gulschbauer listed with one star in the 1968 EMG Art of Record Buying (Erato STU70333). Likewise, two out of three made or issued in 1970. The version by Ludwig Wieland and the RBM Kammerorchester, issued on Sastrophon, is listed in Grey’s Classsical Discography. For information on the recording by the Heidelberg Chamber Orchestra under Jean-Louis Joubert (Oryx EXP38), I can only refer to reviews by Trevor Harvey in Gramophone[10] and the anonymous review in the EMG Monthly Letter[11].  Oryx had a brief life – the present-day company of that name has no connection with the one existing around the 1970s. Their selling point was that, at 14 shillings and sixpence (72.5 pence), their Explorer series offered the cheapest LPs on the market and the recordings were new ones[12].  Saving money is a laudable activity, but Harvey felt the purchaser “will be ill-advised to buy something that will commend [these symphonies] to him as little as these performances and recording are likely to do”. His advice was to wait, since neither symphony (the coupling was no. 50) was “at present otherwise available”. The EMG critic also felt they were “none too distinguished performances” but oddly enough noted that other recordings – in the plural – were available of Symphony 53, though they were “not particularly exciting” either. Harvey’s review does tell us that timpani are not used, which he thought “odd … even though there is one source that does not include them” and that finale B is played.

The 33 Haydn symphonies recorded over the years by Sir Neville Marriner and the Academy of St. Martin-in-the-Fields are largely subsequent to my cut-off date, but no. 53 was set down on 25-26 March 1970[13]. He is clearly using the Urtext edition, with the timpani well in evidence, but no trumpets, which Schultz and Robbins Landon evidently rejected as inauthentic – it would be interesting to know more about the parts that Fendler found. . Never mind, we have no added wind parts and a properly phrased minuet. Marriner’s introduction has the notes of the opening gesture slightly separated, as was becoming the norm in baroque and classical music in the 70s, rather than sustained as earlier conductors preferred. His Vivace is well articulated at a brisk, rather than headlong, pace. The first repeat is played. His second movement has a good middle-of-the-road tempo and is nicely turned with all the repeats. The minuet has a fair lilt and a suave trio. Best, perhaps, is the finale. He plays finale B and there is some sizzling articulation of the semiquavers, while finding space for some elegant phrasing in the contrasting material. Oddly enough, Marriner had issued a recording of finale B, under its “Overture in D” guise, the previous year. I have not checked whether the performances are actually different. If anyone reading between the lines is suspecting a somewhat muted enthusiasm on my part, perhaps this is because I sense a somewhat muted enthusiasm in the performance. For all its expertise and – up to a point – vitality, not to speak of a clean text and a finale Haydn actually wrote, there is more sense of discovery and sheer enjoyment in Sacher’s RPO performance.

There is immediately greater weight and stature in Märzendorfer’s introduction. The timpani are clearly present – they seem to be using hard sticks. The Vivace is marginally slower than Marriner’s, but with trenchant delivery it seems to mean more. There is no denying that Marriner has the more refined orchestra, however. First repeat only. There is all of three seconds difference between Märzendorfer and Marriner in the second movement – he, too gives all the repeats. Märzendorfer seems a little more chunky, but this may be down to the orchestra and recording. I am not sure that Marriner’s relative smoothness is necessarily an advantage. Märzendorfer’s minuet also goes at about the same tempo as Marriner’s, with a rumbustious lilt, but when the trio arrives a Viennese upbeat leads to a slightly slower tempo. Purists may frown, but at least he goes beyond mere suavity. Märzendorfer offers both finale A and finale B. The former, in which he gives all repeats, is a quirky affair which he plays with evident enjoyment. In finale B, in which Haydn indicated no repeats, the semiquavers that Marriner’s violins executed with such brilliant aplomb are rather obscured by the wind, perhaps for good reason, but there is edge-of-the-seat excitement here which I missed in Marriner. 

Dorati’s introduction has sharper staccatos than Märzendorfer’s, making it peremptory rather than majestic – a fair alternative. His hushed beginning of the Vivace is promising and, at a similar tempo to Märzendorfer’s, there is both breadth and vitality, though with slight sagging in secondary material and during the development. First repeat only. Dorati’s second movement is very nice indeed, elegant and neatly phrased, with all repeats. So far, so good, but his minuet is insufferably slow. Maybe in his mind’s eye he sees bewigged dancers in 18^th century costume, bowing and gracing with elegant formality and, if we could see them too, this may even work. But this is a symphony, not a ballet, and for sheer listening it is plain dull. As first issued, Dorati gave finale B. Like Marriner, he has a string-dominated recording so when the violin semiquavers arrive, it is clear that his orchestra is not the equal of Marriner’s. While the ASMF had us wondering at the virtuosic articulation, the Philharmonia Hungarica sound rather frantic. Better, then, to cover them with a more forward wind section, as with Märzendorfer did. That said, there is plenty of vitality and a suave presentation of the secondary material.

Towards the end of the series, Dorati set down a supplementary album with various alternatives movements, including finales A, C and D of the “Impériale”. I commented on the quirkiness of finale A in Märzendorfer’s performance, but it would hardly have struck me that way it I had heard Dorati’s humdrum performance first. He gives the repeats. Finale C, with the repeat, fares no better, but finale D seems to engage the conductor more. Those who seek scores at IMSLP will find the Schultz edition, which has finales A, B and C. Finale D can be found there under Overture in D Hob.1a:4. There is not quite the eagerness that Jochum can find in Haydn’s presto finales, but Haydn wrote the piece as a curtain raiser not a finale and from that point of view Dorati has it spin along nicely.

Up to this date, then, Märzendorfer leads the field among the commercial recordings, with Marriner preferable to Dorati. Sacher, especially with the RPO, is a special case for the edition used, and presents finale C to good effect. If you particularly want to hear finale D, you will not need to buy Dorati’s cycle just for that since this piece had fared quite well as Overture in D Hob.1a:4. Archipel have issued a performance by the Vienna Symphony Orchestra under Fritz Busch, undated but apparently contemporary with Busch’s Remington recordings of Symphonies 100 and 101. A little later, in June 1956, Hans Swarowsky set it down for Urania with something called the Vienna Philomusica Orchestra (UX129/USD-1029). The BNF issue has optimistically upgraded this to the “Vienna Philharmonic” and indeed, the UK issue on Saga had the name of the orchestra written in such squiggly italics that most people would have read it as the VPO – as they were no doubt intended to. Later still, in May 1967, Robert Rudolf recorded it with the Vienna Radio Orchestra for Westminster (WST 17141). The sleeve of this LP claims a major career for Cleveland-born Robert Rudolf, including an acclaimed appearance with the London Philharmonic Orchestra[14]. In 1968 he was auditioning for the post of assistant conductor to Stokowski with the American Symphony Orchestra and told the “Tucson Citizen” “It could be embarrassing if I don’t get the post … and I may well not get it”[15]. In the absence of further references, I presume he did not. As well as these three, Gray’s “Classical Discography” lists performances of an Overture in D by the Vienna Philharmonic under Robert Heger (9 January 1929), the Paris Conservatoire orchestra under Edvard Fendler (8 November 1937), the Master Virtuosi of New York under Gene Farrell (1967) and the Vienna Chamber Orchestra under Helmut Froschauer (1968). I have not heard these and since Gray does not specify which overture it is, presumably because the recording data available did not do so, I cannot say whether they are finale B or finale D.

Returning to the three I have heard, they seem in remarkable agreement over tempo and, having sensed a certain lack of eagerness in Dorati’s performance, I can only say he sounds positively dashing in comparison. That said, all three convince on their own terms, without quite resolving my perplexity as to why Haydn should have written Presto if he meant Allegretto. Best, probably, is Busch. His phrasing has considerable detail and, as behoves one of the great interpreters of Mozart operas, it is the sort of performance that, without knocking you out of your seat, provides assurance of an enjoyable evening to follow. Swarowsky is a little more sternly symphonic. Rudolf makes a few points here and there – notably, after the pause at bar 141, he has the next phrases played at a slower tempo, rather hesitantly. Interesting, but perhaps it would become fussy on repeated hearing.

So much for the commercially issued recordings up to 1972. Some live performances that have been preserved are of uncommon interest. Sacher’s RPO recording has been discussed. Hard on its heels came Charles Munch and the Boston Symphony Orchestra on 26 November 1954. This has quite good sound and includes the announcer, who tells us this was the first time the symphony was played in Boston. Rather ominously, he says the symphony has three movements but, thankfully, Munch gives all four. Given that the Schultz Urtext had been issued by the Boston Haydn Society, Munch might have honoured local musicology by using it. Instead, we have the Fendler version with its revised phrasing of the minuet and finale C. We know from other recordings that Munch was no fan of repeats so it will be no surprise that there are none in movements I, II and IV. In addition, he cuts eight bars from II and omits the second repeats in both the minuet and the trio. One for the bin, then? Well, not exactly, since what is left is one of the most adorable Haydn performances I have heard. The introduction impresses with its fine dynamic contrasts – it is interesting that these live captures can give a better impression of this orchestra’s quality in the pianos than the rather coarse RCA recordings from the same period. In the Vivace, Munch finds a tempo that allows vitality and drive, but also spaciousness and, in the development, mystery. His second movement is rather slow, but the theme is phrased with much tenderness, as well as delicacy in the later violin figuration. There is, too, a wistful poetry in the minor key episodes and Munch makes much of the bassoon towards the end of the second of these – indeed, I got the impression that the cellos and basses are not playing here. The minuet trips along delightfully and, in the pianissimo passage in the second part, wonderful poetry. The trio is affectionately phrased. In the repeat of the first part, the strings are either cut right back or eliminated to have the flute stand alone. The finale is a tad faster than Stokowski’s, enough to add jokey humour to bonhomie. The pauses are beautifully timed. Played like this, I cannot help thinking that, whoever wrote finale C, it is the best of the four. Which makes it difficult for me to believe it is not real Haydn.

The most intriguing performance must be that by “a small section” of the BBC Northern Orchestra under Stanford Robinson. This opened a concert broadcast on 9 August 1967 and was followed by Boris Blacher’s Fantasy for orchestra and Tchaikovsky’s Second Symphony. Robinson (1904-1984) was born in Leeds and owed his unusual first name to his father’s fervent admiration for Sir Charles Villiers Stanford, who had just been appointed conductor of the Leeds Festival[16]. His career was mainly with the BBC, which increasingly typecast him as a conductor of the lighter repertoire – Baxians will know that there was a lot more to him than that. Whether in more serious or lighter repertoire, he had firm ideas about doing things properly. Back in 1930, he conducted a recording of Mendelssohn’s “Elijah” that, by the side of the 1947 version under Sargent, sounds like a blueprint for a more recent “authentic” performance. In 1966, he set down all the Gilbert and Sullivan operettas for the BBC, performances that circulate avidly among aficionados. For the occasion, he studied the manuscripts and corrected numerous errors in the D’Oyly Carte orchestral parts. You would expect the D’Oyly Carte company to have been eternally grateful. Instead, they paid someone to reinstate all the mistakes and sent him the bill[17].  The Haydn performance has been preserved in passable sound with an announcer who tells us that the realization of the harpsichord part by Robinson himself is played by Peter Smith.  

The harpsichord announces its presence early on, filling in two of the rests (bars 8 and 9) with little flourishes. After an imposing introduction, Robinson takes a very relaxed view of the Vivace, definitely four-in-a-bar. Once I had adjusted, I found plenty to enjoy in this spacious, rather grand reading. “Small section” or not, the manner of playing, with generous bows from the strings, is full-textured. In second subject territory, Robinson adds trills to the grace notes. Just once (bar 152), there is actually a trill in the score, which I suppose is some justification for adding them systematically[18]. No repeat. In the second movement, Robinson takes a fairly slow tempo, nicely phrased and duly expressive in the minor key episodes, though it has to be said that the BBC NO, decent band though it was, is rudimentary beside the refinement of Munch’s Boston SO. Robinson adds a number of trills and other variants to the melodic line, as well as a few minor interpolations by the harpsichord. In the first variation, the more florid parts are played as a violin solo, nicely handled by the leader, Reginald Stead. There are rallentandos towards the end of some of the variations, though not all of them, and the last two are interpreted as a general winding-down. Robinson plays the first four repeats but drops two later. The minuet goes at a fresh, bouncy tempo. The pause in the second half, before the pianissimo passage, is filled with an upward flourish by the harpsichord. In the trio, the more melodic part, which is scored by Haydn for flute doubling strings, Robinson has strings only the first time, flute only in the repeat. Finale A is used. Robinson takes a broad, almost hymn-like view of the Moderato tempo. The contrasting sections, which are interpreted as quirky under Märzendorfer and not really interpreted at all under Dorati, carry more expressive weight here. An interesting view that lingers in the mind. There are a number of variants to the violin parts but the most interesting moment comes in the eight bars from 129. The strings here have bare arpeggios – too bare for Robinson apparently, since he adds some melodic phrases played by the harpsichord.

This performance is truly sui generis. In certain respects – the harpsichord interpolations, the added trills, the variants to the string parts[19],  the allotting of certain passages to solo violin, the alternation of strings and solo flute in the trio – it anticipates the most extreme “historically informed” renderings of more recent decades. At the same time, these elements are grafted onto a performance manner that, with its full, legato bowing and unhurried, flexible tempi, harks back to pre-war days. Somewhat disturbingly, I find the symphony resonates in my mind more like this than in brisker, sprucer modern interpretations.

Stanford Robinson was no stranger to Haydn by 1967. In the three years from 1964 to 1966, the BBC presented an “extended series” entitled “The Haydn Symphony”. Not all the entries in the BBC Programme Index show which symphonies were played, and the relevant Radio Times issues are not yet available as scans. It is not clear, then whether all the symphonies were covered, or how many were conducted by Robinson, but he contributed, at the very least, nos. 1, 8, 20, 38, 56, 57, 67, 86 and 93. The other conductors, all with BBC regional orchestras, were Rae Jenkins, George Hurst, Bernard Keeffe (including no. 53 in 1964), Arwel Hughes, Graham Treacher, Maurice Handford, Leo Wurmser, John Carewe, James Loughran, Michael Krein, Harry Newstone, Alberto Bolet, David Lloyd-Jones, Mansel Thomas, Norman Del Mar and Michael Hall. It would be interesting to reconstruct more fully the programmes of this enterprise. It would be interesting, too, to hear from anyone who has any of these on old tapes[20].  

Lastly, a 1969 performance by the Bavarian Radio Symphony Orchestra under Dean Dixon. Something of a firebrand in his youth, Dixon seems, in his later phase, to have aimed at the purest, most unadorned presentation of the music[21]. To this end, he gives a very precise, well-phrased rendering of the Urtext score with all repeats except the second in the first movement. His tempi in the first two movements are similar to Robinson’s unhurried traversals, with the emphasis on majesty in the first and gentle poise in the second. The minuet is also taken slowly and the finale – he chooses finale B – is brisk without being the full-tilt Presto some other conductors have given us. The orchestral sound is full though not heavy, with long bowings in the old-fashioned manner. In spite of the slow tempi, the performance is energized, and its forthright truthfulness takes us into Klemperer territory. Given that Dixon was one of the very few conductors who troubled to conduct a Klemperer symphony[22], it rather looks as if the older conductor had become a role model for Dixon. What, if anything, Klemperer himself could bring to a Haydn symphony that Dixon, adopting a similar manner, could not, is something that will have to be discussed when the possibility arises to hear both of them in the same symphony – there are at least two instances. I found Dixon mightily impressive in Symphony 103. In this case, I found a certain marmoreal quality. The effect in the second movement is hypnotic as it ticks patiently through its more than nine minutes. My preference for faster minuets must be obvious by now, though the playing under Dixon is alert. On the whole, I think I would not forgo the excitement generated by Märzendorfer, among the Urtext recordings, while Munch has to be my guilty pleasure and I must say Robinson tugs at the memory.

One tragi-comic affair remains. On 16 June 1933, in the BBC Radio Theatre, the BBC inaugurated its first theatre organ, designed by John Compton and consisting of 2,826 pipes in 35 ranks. It is still there, though little used. Three organists played for the inauguration – George Thalben-Ball, G. D. Cunningham and Walter Alcock. The first of these used it to set down, presumably around the same time, his arrangement of an “Air and Variations from Symphony in D” by Haydn (HMV B.8142). This turns out to be the second movement of Symphony 53, shorn of the minor-key episodes and with a flamboyant coda added. I have the idea that this could be a pleasant voluntary on an organ less removed from Haydn’s own time. Thalben-Ball exploits the characteristic woozy vibrato of a typical theatre organ and, as the variations gather notes, seems to wish to put about the idea that Haydn was some sort of precursor of Widor. ’Nuff said.

A few notes to the table below: repeats are played in the minuet and trio unless otherwise stated. Finale A has some short repeats at the beginning. All the performances play them. Finales B and D have no repeats. Finale C has one.

	I	II	III	IVA	IVB	IVC	IVD
Fendler	04:22 neither repeat	04:31 no repeats	03:31			02:32 no repeat
Stokowski	05:14 neither repeats 4-bar cut	04:56 no repeats 2 8-bar cuts	04:15			03:44 no repeat
Sacher VSO	06:37 first repeat only	05:21 5 repeats omitted	04:16			04:33 with repeat
Sacher RPO	06:43 first repeat only	05:26 6 repeats omitted	04:25			04:32 with repeat
Marriner	07:11 first repeat only	06:32 all repeats	03:50		04:28
Märzendorfer	07:30 first repeat only	06:29 all repeats	04:24	04:34	04:06
Dorati	07:40 first repeat only	06:32 all repeats	04:53	05:18	04:13	05:44 with repeat	03:34
F. Busch							03:56
Swarowsky							03:58
R. Rudolf							04:02
Munch	05:26 neither repeat	05:04 no repeats 8-bar cut	03:00 No 2nd repeats in minuet or trio			03:56 no repeat
Robinson	07:14 neither repeat	08:05 2 repeats omitted	04:18	06:11
Dixon	08:44 first repeat only	09:18 all repeats	04:45		04:45

Christopher Howell © 2026

---

[1] Antony Hodgson: The Music of Joseph Haydn: The Symphonies, Fairleigh Dickinson University Press 1976, p. 81.

[2] H.C. Robbins Landon: Haydn. Chronicle and Works. Vol. II. Haydn at Esterhaza 1766-1790. Thames and Hudson 1978, pp.560-1.

[3] Somewhat puzzlingly, another article in Italian, by Ennio Melchiorre, while dating it 1778-80, states that it was intended as a homage to the Empress Maria Theresa after a visit to Esterhazy and that “according to certain [unspecified] biographers, Haydn dedicated it to her on the occasion of the proclamation of her daughter, Marie Antoinette, as Queen of France”. https://www.flaminioonline.it/Guide/Haydn/Haydn-Sinfonia53.html, viewed 26.4.2026. An attractive theory that would have to date the symphony back to 1774 since Maria Theresa visited Esterhazy in 1773 and Marie Antoinette became Queen of France upon her husband’s succession as Louis XVI in 1774. It would be interesting to know who Melchiorre’s “certain biographers” are. Scholars have argued whether the symphony played in Maria Theresa’s honour in 1773 was no. 48, which has been named after her, or no. 50. I have seen no suggestion that it may have been no. 53.  

[4] In some places, “Italian Overture”.

[5] Downloadable from IMSLP.

[6] Information from notes to the Cala reissue of Stokowski’s recording https://dn720300.ca.archive.org/0/items/cd_schumann-symphony-no-2-haydn-symphony-no_leopold-stokowski-and-his-symphony-orchest/cd_schumann-symphony-no-2-haydn-symphony-no_leopold-stokowski-and-his-symphony-orchest.pdf#:~:text=2%2D5%20During%20the%201930s%2C%20the,orchestral%20parts%20he%20found%20in viewed 26.4.2026

[7] The cut is not inherent in the Fendler edition, since Fendler’s recording does not make it.

[8] Finale C has no timpani parts anyway. The parts Fendler discovered in European libraries perhaps derived from someone who saw no point in having the timpani play in movements I and III and then stay silent for the finale. Arguably, if you are going to have timpani, you should have trumpets too, if only to double the horn parts. In truth, there are moments where Stokowski does seem to have timpani and even trumpets, but with a very backward balance, and there are also moments where timpani and trumpets would logically be playing, if present, and do not seem to be. 

[9] The sleeve note by William Flanagan for the Epic release of the Sacher performance perhaps sheds some light: “We are, as a matter of fact, fortunate to have Symphony no. 53 … fifteen years ago it could not have been heard at all; its score, in fact, had been lost. Fendler, the German musicologist and conductor found orchestra parts to the symphony in the course of a research project undertaken by the French State Radio. And finding a complete set of parts was a wild and woolly chase: the horn and trumpet parts, for example, were found in Munich, the violin parts in Marseilles, still others in Bordeaux. … The symphony is scored for flute, 2 oboes, 2 bassoons, 2 horns, 2 trumpets, timpani and strings”. According to this, then, trumpets and timpani should be playing on the Sacher VSO recording. I can only insist that, if they are, they are so backwardly balanced, here as well as in the Stokowski recording and Fendler’s own, that they may as well have stayed at home. If you want to hear them, go to the Sacher RPO recording. Does Fendler’s score perhaps indicate these elements as optional? Flanagan’s note, in any case, had been overtaken by events – the Boston Haydn Society edition had come out in the meantime.

[10] Gramophone June 1970 p.38.

[11] June 1970 p.2

[12] Music for Pleasure and Fontana SFL LPs cost only fractionally more but were reissues from the back catalogue.

[13] So too, I now realize, was no. 52. At some stage, many of these articles will have to be updated to take into account recordings of which I have since become aware or of which I have found a copy.

[14] 23 April 1960 in the Royal Festival Hall, but I have found no review with which to measure the acclaim.

[15] 8 April 1968

[16] Information from Donald Brook: Conductors’ Gallery, Rockliff 1946 p.119.

[17] Caution is always advisable when you read something on the internet, but this account rings true: https://www.good-music-guide.com/community/index.php/topic,30294.0.html#:~:text=The%20BBC%20recruited%20excellent%20casts%2C,had%20top%2Dflight%20light%20opera%20credentials. Viewed 5.5.2026.

[18] I have no definite knowledge that this was Robinson’s idea or whether it appears an the edition he was using. I suspect that he used the Urtext as a base and the added ornaments were his own. The phrasing of the minuet shows that he was not using Fendler’s edition.

[19] Though the question remains – how far are these variants Robinson’s own work? The announcer says nothing of this, but he does say that the symphony was very popular in England in the 19th century. Did Robinson draw on old scores and parts that, if not exactly authentic, were representative of how the symphony was played in Haydn’s day or not long after?

[20] A performance of no. 67 under Robinson, not from this series, will be discussed in a later article.

[21] My article on this conductor is in need of updating, but it touches on this aspect. 

[22] No. 2, with the Hessian Radio SO, 1 March 1973