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. 104 in D major Hob. I/104 (1795)
Part 2: 1951-1958
A pdf version of this article can be downloaded for offline reading and printing here.
As explained in the first part of this article, I am dividing my discussion of this symphony into four parts. The second part begins from the dawn of LP and analyses recordings by Annovazzi, Krauss, Ginzburg, Szell, H. Blech, Rowicki, Rosbaud (2), Horenstein, Kempe (2), Wöldike, Münchinger and Bernstein.
Of unknown date, but issued in 1951, is a performance by the “Austrian Symphony Orchestra” (probably the Tonkünstler Orchestra) conducted by Napoleone Annovazzi (Remington RLP-199-56). Annovazzi (1907-1984) had a career mainly outside Italy, embracing Riga, Barcelona, Valencia, Vienna, New York City Opera and a 23-year association with the Dublin Grand Opera Society, though at the time of his death he was in charge of the opera house in Lucca. As was inclined to happen with Italian conductors, he was mostly asked to conduct Italian opera, though his repertoire in Dublin included Lohengrin, Eugene Onegin and Der Rosenkavalier. He recorded for several smaller labels, including the sole “official” recording by Virginia Zeani of La Traviata, made in Hamburg. To what extent he also got to conduct the concert repertoire I have not been able to discover. For the curious, quite a lot of his Dublin opera performances can be heard at Internet Archive. This Haydn recording seems to be his only non-operatic work in the studio. The LP was completed by Mozart’s Symphony K200 under Paul Walter.
There is something rather theatrical about this performance. The introduction is very slow indeed, really eight-in-a-bar – and promises solemn things to come. Instead, the Allegro is quite speedy and has considerable joi de vivre. No repeat. The Andante is almighty slow, unashamedly four-in-a-bar. The playing is warm and affectionate but, as tends to happen when this movement goes too slowly, there are times when too little is happening for the music to have much sense. The first repeat is taken. The minuet lilts – just about – in a stately, courtly manner. The trio is nicely played. Annovazzi does not hold up the rising third motifs, but he consistently treats the second as an echo of the first. The finale, without repeat, has the rising jubilation of an opera buffa finale. This can hardly be my go-to London Symphony but it suggests that Annovazzimay be a conductor worth tracking down.
Still in Vienna, on 22 March 1951, the Viennese conductor Clemens Krauss set the London down with the Vienna Symphony Orchestra (Supraphon 14113/16-V). I was seriously disappointed by Krauss’s recording of Symphony 93. Happily, this is much better. The introduction is broad without exaggeration. The unauthentic crescendi and diminuendi in the Breitkopf edition are observed very sparingly indeed – did Krauss instinctively realize they should not be there? Other details show that he is using the Breitkopf edition. The Allegro bowls along at a good but not overdriven tempo. Interestingly, Krauss makes a lot of small adjustments to the dynamics here. Repeated phrases in forte passages are often made so that the second becomes an echo of the first. At other times, he cuts back the strings to give prominence to woodwind lines that are hardly evident in previous versions. No repeat. The Andante, with the first repeat, is given in a leisurely two-in-a-bar tempo. Krauss’s careful balancing brings to the fore the cello and viola lines after the first double bar. Likewise, he relishes the bassoon’s doubling of the violins from bar 17. There is another case of “edited” dynamics at bar 50, where he makes a steep diminuendo so as to treat the following five bars as a crescendo, rather than maintain a steady forte as written. Highly effective, though I suppose there was nothing to stop Haydn from indicating it is he wished. Rather oddly, the sudden forte in the second half of bar 83 – during the dotted rhythm passage – is ignored[1]. Liberties apart, this is a lovely performance of the movement. The minuet goes at a nice Viennese lilt. A slowing of the rising third phrases introduces a slightly slower trio, still with a good lilt, but the real schmaltz comes at the end of the second part of the trio, with some portamenti that sound very odd to modern ears. The repeat of the second section of the minuet is omitted. The finale, with repeat, is disconcertingly slow. I tried to come to terms with it, but it seems earthbound. I got the idea the orchestra was not convinced either. An early attempt to move it on is countered by the conductor, but in the development a faster tempo is allowed to encroach for quite a few bars. Krauss’s tempo does allow certain details to emerge and in the scurrying passages you can really hear that the lower strings are in contrary motion to the upper ones. The chorale-like passages assume a Brahmsian warmth without any noticeable slackening of pace. A curious, but curiously attractive performance overall.
The first recording with all repeats was issued by Melodiya in 1951. The USSR State Symphony Orchestra was conducted by Leo Ginzburg. Ginzburg (1901-1979), a Soviet citizen of Polish-Jewish origin, was conductor of the Moscow State Symphony Orchestra from 1950 to 1954. He also taught conducting at the Moscow Conservatoire, where a distinguished roll-call of students includes Vladimir Fedoseyev. His recordings, insofar as I can trace them, are few – a pity, on the strength of this one. The introduction is expansive, enunciating the dotted rhythms clearly and with a sense of underlying tension. The main theme of the Allegro is warmly expressed and the tempo is kept steady, but with great inner energy in the articulation of the quavers, which are often played with separate bows even when marked as slurred. The effect is one of magisterial strength. The Andante is slow but manages, where Annovazzi failed, to maintain a feeling of two-in-a-bar. There is delicacy and elegance yet with an overall elegiac feeling. The minuet is fairly fast with a fine lilt. A slightly halting interpretation of the rising third motif introduces a deliciously sprung trio. The finale is swift but, unlike Dobrowen’s, never becomes a gopak, all is exuberant joy. Ginzburg finds a wider range of feeling and expression in this symphony than any of his predecessors, and furthermore manages to do this without straying beyond the confines of Haydn’s own world. This performance does not evoke earlier or later composers and seems to speak with Haydn’s own voice. As well as being the only really complete recording so far, I have an idea it might be the best performance so far as well. The recording has some distortion in heavy passages and the treble needs taming, but it is considerably better than I expected.
The 1955 edition of The Record Guide[2] listed just two recordings of Symphony 104, both starless – the Krips and the Munch. In its lofty way, it limited itself to remarking that “A good recording of the last symphony of all still eludes us”. Critics in those days were evidently not required to explain their summary judgements. It would have been good to know what was supposed to be so wrong with the Krips, in particular. Other recordings discussed here were presumably no longer in the catalogue, or had never reached the UK at all. Two recordings set down in 1954, but perhaps not issued by the time The Record Guide went to press, might have won their approval.
Or maybe not. The first of these was set down on 11 April 1954 by the Cleveland Orchestra under Georg Szell (Epic LC.3196). The conductor clearly means business with a forceful and relatively swift introduction. The music sounds a little restless at this tempo but I suppose this is a valid interpretation. The Allegro is brisk and impresses with its fiery discipline – the gentler moments are no more than decorously shaped. No repeat. The Andante is definitely two-in-a-bar and is shaped neatly but without apparent feeling. As found in other too-dogmatically two-in-a-bar interpretations, the music sounds excessively busy as Haydn branches into triplets and the dotted rhythm section goes for little. First repeat only. The tempo for the minuet is about the same as Ginzburg’s – Szell is five seconds longer – and the comparison shows what is wrong here. The rigid beat and smartly punched-out staccatos deprive it of a real lilt. The brutal playing of the final chord sums it all up. Szell holds back the tempo for the rising third motif that introduces the trio. It sounds merely mannered in such a loveless performance. The finale, repeatless, goes pretty fast, but some conductors have made such a tempo, or even a faster one, sound effervescent. Szell does not let the music breathe so it degenerates into a display of orchestral virtuosity, no more. I always go to a Szell performance hoping to discover at last evidence of his greatness. Oddly enough, I found it in his Rome performance of Symphony 93, but I am afraid the supreme disciplinarian holds sway here.
If Ginzburg deserves credit for making the first recording with all repeats, Harry Blech and the London Mozart Players deserve no less for making the first to be readily available outside the Soviet bloc. This was set down on 29-30 November 1954 (HMV CLP 1055). This was also the first time since Barbirolli and Fischer that anyone had used an orchestra scaled down to a classical size. Compared with the Cleveland and most others, the string sound is thin with a corresponding presence of the wind band. This apart, the differences are not great since Blech, as noted in other articles, had his strings play in the traditional manner with long bows and avoidance of stabbing staccatos. Indeed, from this point of view, Szell’s performance might sound more “modern”, if you like that sort of thing. Blech’s introduction, fairly slow, is portentous rather than dramatic. His Allegro then scuds along merrily. The curious thing is how similar Blech’s tempi are to Szell’s from this point on, and how different they sound. The players are not regimented but allowed space to breathe the music. The woodwind interventions at the start of the development are delightfully piquant, for example. Elsewhere, Blech could take his time over slow movements, but he obviously saw that this one should not be indulged too much. The question is, why, when the tempo is about the same as Szell’s, does the music not sound over-busy in the same way when it branches into semiquaver triplets, dotted rhythms and so on? “Take time in the tempo”, I once heard a singing teacher say. Blech, who had until not long before been the leader of a string quartet and had played many of the Haydn quartets, evidently understood this point. The tempo does not audibly slacken, yet all the details find their proper space. There is also much affectionate phrasing here. Another feature common to all the Blech Haydn performances I know is his merry minuets. This is no exception. Since the stopwatch tells us that the tempo is the same as Szell’s (two seconds longer!), how can this be? The six repeated crotchet Ds from bar 20 are symptomatic. Under Szell they are punched out in a sharp staccato, all exactly the same. With Blech, they are detached – Haydn did not actually ask for staccatos – but allowed to sing, and are played in crescendo, the last one being held a little longer so as to lead on to the next phrase. The sort of innately musical solution that a violinist would likely do automatically. The suavely lilting trio can stand as testimony to the gentlemanly yet characterful music-making that endeared Blech to London audiences in the 1950s. The fact that he became increasingly laissez-faire over the next decades does not detract from his achievement here. The finale spins along like a joyful rustic dance, with the chorale passages “taking time in the tempo”. Anybody in 1955 who could get both Ginzburg and Blech had a fine perspective of this symphony.
Clevelanders, maybe Americans generally, will be shaking their heads in disbelief. Am I claiming Blech was a greater conductor then Szell? In one sense, clearly not. The ensemble of the opening phrase would have demanded an immediate remake from Szell, maybe several. I think Blech, as an ex-quartet leader, saw himself as leading his band rather than dictatorially conducting them. Sometimes, in performance, he would put his baton down and let them play. His luck was to assemble a group of fine London musicians and give them the space to enjoy their music. In the early years of his orchestra, then five years old, the formula seems to have paid off.
A recording by the Polish Radio Symphony Orchestra under Witold Rowicki was issued by Muza in 1955 (L 0072). This is the only recording, unless there are others I am not even aware of, which I have been unable to hear. I do, however, have a live recording of Rowicki conducting the symphony with the Warsaw National Philharmonic Orchestra on a visit to Leningrad on 31 October 1959, so I will discuss it here on the assumption that Rowicki’s overall view did not change much over four years. This may not be the case, of course, and in one sense it clearly is not. Frustratingly, a Rowicki performance, presumably the 1955 Muza, is listed at Spotify, but after undergoing the rigmarole of registering, fortunately free, it turned out that “this content is not available”. They do, however, show the timings – 07:12, 08:40, 05:21 and 04:46. Comparison with the table below will show that they are very different. In the live performance, Rowicki plays all the repeats. If the Spotify timings are correct, it looks as if there are no repeats in the outer movement and, at most, only the first in the Andante. Another query is whether anyone would conduct the symphony in the studio with the almost desperate intensity Rowicki displays before an audience. The introduction is powerfully brooding, an impression exacerbated by the compressed dynamics of the radio recording, which rarely allow the music to drop below mezzo forte. Both the Allegro of this movement and the finale are ripped through at an incredible lick. Maybe I would have been riveted to my seat if I had been there in the hall. Sitting at home, while I could admire the fiery intensity, it quickly became unremitting. In the context, the Andante seemed quite slow. In reality it is a fairly middle-of-the-road tempo, but with a lack of dynamic shading and a bleakness of phrasing that would better suit Shostakovich, I found little to admire. The minuet is on the slow side and heavy-booted into the bargain. Not much elegance even in the trio. I had expected better.
Gray’s Classical Discography[3] lists a recording on Concert Hall by the Pasdeloup Orchestra under Louis Martin, but the date and number are those of Martin’s recording of Symphony 101. This was coupled with Symphony 104 under Boult, of which more later, hence the confusion.
An impressive version was set down on 5-6 March 1956 by the Berlin Philharmonic Orchestra under Hans Rosbaud (DGG LPM 1863). Impressive not least for the orchestral sound. Rosbaud draws lean and clear textures from a string section that could sound thick in other hands, and balances this against a well forward wind choir and strongly present trumpets and drums. He was not a conductor who went in for charm and his introduction is imposing with a granitic initial attack. The Allegro is brisk but not headlong and has a stern inner vitality. The repeat is played. The Andante is taken about as slow as it can go, but not slower – it retains a two-in-the-bar feeling and the dotted rhythms and triplet semiquavers fall perfectly into place. It could be described as intense rather than affectionate. First repeat only. The minuet is a ceremonial affair, but he manages to keep it alert. The finale is not exactly Spiritoso (witty) but it has strength and great vitality. No repeat. Without particular point-making, this performance has that indefinable thing called stature.
Rosbaud was not a prolific recorder, so his posthumous reputation has been enormously enhanced by the issue of live and broadcast performances, including a notable series with the South-West German Radio Symphony Orchestra of Baden-Baden, of which he was chief conductor from 1948 to his death in 1962. These include a considerable number of Haydn symphonies, including many that were little known at the time. His recording of the London with this orchestra, set down in the studio (without an audience) on 27 June 1952, testifies in many ways to his utter consistency. The excellent strings/wind balance is the same, with the trumpets perhaps further forward but without blaring. In terms of repeats, back then he also omitted that of the first movement. However, the Berlin performance may also testify to the difficulty of getting the Berlin PO to play differently from their usual manner. With his own orchestra, the first movement has an extra urgency, the second has greater fluidity without any sense of haste and the finale is a knockout. If the tempi in these movements are just slightly faster, the minuet breezes along here, much to its advantage. Fine as the Berlin performance is, the Baden-Baden one is better still.
Another conductor whose live performances are avidly sought by collectors is Jascha Horenstein. In truth, his official legacy, mainly on Vox, is quite extensive, but compromised by second-rate orchestras and sub-standard sound. To boot, he was not always asked to do the repertoire in which he could be most revelatory. He set down Symphony 104 during sessions for Vox from 27 to 29 April 1956 (PL 9330). The orchestra is described on the sleeve as the Vienna Pro Musica Orchestra, a convenience name regularly used by Vox. Later issues have called it the Vienna Symphony Orchestra but, if so, were there contractual obstacles to naming it on the original issue?
Horenstein’s massive opening bars hold Haydn’s pauses for what seemed an incredibly long time, until I reflected that most conductors barely hold them beyond the marked minim. The way in which Horenstein conjures the following piano section as if from the depths of the fortissimo – most conductors pause slightly before going on – shows a master conductor at work. Though not, apparently, one who could get the fortissimo bars exactly together or precisely in tune. Nevertheless, this dramatically slow introduction is impressive. Impressive, too, in its way, is the Allegro, taken so slowly as to be almost four-in-a-bar. The sheer conviction with which Horenstein weighs into the quavers that often scud along as a generalized rush just about carries the day, and he manages not to be heavy, but it is a joyless ride, all the same. Towards the end – bars 238 to 247 – he introduces a rallentando which seemed excessive considering how slowly he is going anyway. No repeat. The Andante is quite slow, though not entirely out of sight of two-in-the bar, and gets slower towards the end. Nevertheless, I found this the most convincing movement. Horenstein finds an attractive tenderness in the music, while investing the forte passages with a tragic power. First repeat only. Slower-than-usual minuets are unlikely to produce revelations, since these movements do not usually aspire to metaphysical dimensions. Horenstein’s Brahmsian warmth in the gentler passages does not save this from being a trudge, but I did appreciate the wistful grace of his trio. The finale, without repeat, is once again slow. The generosity with which the chorale-like passages emerge without any need to slow down is part compensation, but the lack of joy is troubling. On the other hand, I did notice that in several passages, notably in the later stages of the development, the orchestra was playing with such intense commitment as to momentarily convince me in spite of myself. It is a pity Horenstein, peripatetic to the end, never had an orchestra with which he might have honed his interpretations rather than having to persuade an orchestra afresh every time he came back to a particular work.
It is as well that Vox did not shy away from long sides. Horenstein’s London, which shared an LP with the Clock, stretched to 29:28 even with most of the repeats omitted. A little calculation tells me that the performance would have lasted 35:30 with all the repeats – longer than some romantic symphonies.
Back in London, Rudolf Kempe’s recording with the Philharmonia Orchestra was made on 8 June 1956 (HMV ALP 1471 in the UK, Capitol G-7150 in the USA). After Horenstein, this is a pleasing return to normality. After a grave introduction, the Allegro bowls along pleasingly. No repeat. The Andante is perhaps the best movement here. It has much of the tender wistfulness of Horenstein’s though with a little more fluidity. The minuet, at a midway tempo, has a pleasing lilt and I enjoyed the mellifluousness of the trio. Here we can find pointers to what this performance has and has not. As the bassoon joins the violins in the theme, Kempe has the staccatos played gently, while Horenstein and many others prefer to have it punch out the staccato notes to more earthy, bucolic effect. If this performance is not really memorable, it is because Kempe does not seize – maybe did not want to seize – opportunities to go beyond mere niceness. The finale, without repeat, is sufficiently effervescent. Its only oddity is a slowing of the chorale-like theme the last time round, dwelling on the upward scales by the flute. I feel a little guilty over my lack of great enthusiasm. This performance brings no new insights but it pleasantly rehearses old ones and was doubtless thoroughly enjoyed when choice was more limited – though not so limited as to exclude, even in HMV’s own stable, the more vital (and completer) performance by Harry Blech.
1975 is normally beyond my cut-off date, but given Kempe’s reputation I was curious to see what, if anything, had changed in his live performance on 8 October of that year with the BBC SO in the Royal Festival Hall (BBCL 4188-2). The outer movements have gained their repeats, which is all to the good, and there is more fire to the playing, though it is often slipshod, culminating in some shockingly crude trumpet playing at the end of the finale. The slowing noted above towards the end of this movement is not present. The Andante, still without the second repeat, sounded exactly the same at the beginning, but seemed to drag towards the end. Since the stopwatch tells me this movement is slightly faster now, I can only assume that the only moderately engaged playing, resulting in generic phrasing, gave the impression that it was slower and heavier. The minuet seems an exact replica of the previous version, though less well played. I am not sure where this leaves us. Kempe’s interpretative insights seem to have been reserved for other music, maybe even for other Haydn symphonies – his late interest in Symphony 6 produced good results.
In 1956, Kempe’s use of the old Breitkopf edition might have seemed pardonable. By 1975, he could have taken on board the Urtext edition edited by Robbins Landon, which had appeared in 1967. He was not alone in failing to do so.
The name of the Danish conductor Mogens Wöldike (1897-1988) may not resonate with the many, but in Haydn circles his name is iconic. His set of the last six symphonies was recorded in June 1956 with the Vienna State Opera Orchestra for Vanguard. In the words of Antony Hodgson, “One early group of recordings using the latest researches in authentic text was the magnificent set of Symphonies 99-104 by the VSOO under Mogens Wöldike. Here in stereo sound was a conductor who, like Scherchen, was not afraid to let the instruments strike through in suitably forceful style. Tempos were well judged and there was always a strong sense of impulse”[4]. He does add the proviso that, for the LP issues, many repeats – Wöldike played them all – were “brutally hacked out”, but as far as Symphony 104 is concerned, on my LP copy (AVRS 19027) they survive intact.
With the passing of time, a few other provisos may be added. The introduction is gravely impressive and the Allegro portrays an upfront, no-holds-barred Haydn that may have shocked in 1956. With pounding timpani and ringing trumpets, the surging vitality is still thrilling. The same may be said of the finale, though several others have approached Wöldike’s excitement here. The Andante has a wistful charm similar to that found by Horenstein and Kempe, while erupting forcefully in the fortissimo passages. The minuet lilts at a good speed without trying to make it become a scherzo. Thus far, so magnificent. The VSOO, however, was not a chamber orchestra and, while the trumpets and drums sound through splendidly, the balance between strings and wind was perhaps better managed on Harry Blech’s record. Moreover, while the VSOO is nominally the Vienna Philharmonic minus the front desks, in practice they played, at least back then, so consistently less well under the “State Opera” title as to raise the suspicion they did so deliberately. Wöldike certainly managed to fire them up, but there are moments of ensemble where they sound rather provincial. Still, this is not too serious so long as the spirit is there.
But what of the “latest researches in authentic text”? Following the performance with the Robbins Landon score, I was left nonplussed. The hairpin crescendo-diminuendo in bar 6, present in the Breitkopf edition but not in the Robbins Landon edition or in the holograph score that can be downloaded from IMSLP, is observed here even to a greater degree than in several other performances. The sforzati in bars 54 and 56 have been removed, but what of the bowing slurs in bars 32 and 34? Every other performance so far has punched the notes out with separate bows, and so does Wöldike. These slurs are not present in the holograph, so how do they come to be in the Robbins Landon edition? More of this in a moment. I will give a few other examples. In bar 16 of the Andante, the last two quavers, slurred in the Robbins Landon edition and in the holograph, are played as delicate staccatos here as in every other performance so far. In bars 98-99 of this movement, the semiquaver triplets are played, by Wöldike and everybody else so far, as perky staccatos, whereas they are marked with a legato slur in Robbins Landon. But there is no trace of a legato slur in the holograph! Last but not least, at bar 123, the descending scale for the second violins, absent in Robbins Landon and in the holograph, is present here as in every other performance. The impression is that Wöldike is using the Breitkopf parts with a few corrections here and there. None of this research had arrived in time for the 1956 recordings. So far as the text is concerned, then, Wöldike has been overtaken by events and I am not sure that his performance, fine as it is, comprehensively supersedes those of Blech or Ginzburg.
But what else had he to go by, and how to account for the discrepancies between the Robbins Landon score and the holograph? Robbins Landon relates: “In 1958 and 1959, we were able to announce … our discovery of awhole series of authentic parts of the Salomon Symphonies, corrected by Haydn and partly on British paper with the watermark date ‘1794’, at Donaueschningen Castle. Still later, we persuaded the British Museum’s reading room to purchase a series of printed editions by Robert Birchall which proved to be based on Salomon’s long-disappeared orchestral material”[5]. I take it, then, that discrepancies between the holograph and the Robbins Landon edition are variants found in these parts which Robbins Landon took to be improvements and fresh thoughts made by Haydn himself when preparing the symphony for performance.
It may be difficult for younger listeners to realize the extent to which the early LP catalogue was dominated, at least as far as the baroque and early classical period was concerned, by Karl Münchinger (1915-1990) and the Stuttgart Chamber Orchestra which he founded in 1945. His recording of Haydn’s The Creation was made with the Vienna Philharmonic Orchestra, and it was with this orchestra that he set down the London Symphony in sessions from 2 to 7 May 1957 ). The opening bars show that the presence of the VPO under its own glorious name is no guarantee of pure intonation, but with the piano bars my ears pricked up. Münchinger does not invest the music with tragic weight or dramatic tension, but I had not heard it played before with such natural warmth. Natural warmth tinged with that golden string sound the VPO can produce when it feels like it also informs the statement of the opening theme of the Allegro. This is fairly brisk but also steady and springy, so the impression is of openhearted enjoyment. No repeat. Warmth rather than wistfulness also characterizes the Andante which, though fairly slow, is beautifully poised and fluent. First repeat only. I have taken issue with other minuets as slow as this one, but here we have the VPO providing those upbeats of which it alone seems to know the secret and which have you swaying gently as you listen. I enjoyed every minute. The finale, repeatless, is not a riot, let alone a gopak, but at its steady pace it becomes a joyful country dance. A lot of the success of this most enjoyable performance is likely due to the fact that Münchinger had the good sense to let the VPO give the performance it wanted to give, but let us not underestimate him. Others have tried the same thing with this orchestra and the results are not often this good.
It is perhaps a pity that Leonard Bernstein’s recording, set down on 27 January 1958 with the New York Philharmonic Orchestra (Columbia MS 6050), came so early in his love-affair with Haydn. Later on, he used the Robbins Landon editions as they became available, and was inclined to be more generous with repeats. On the other hand, his tendency to introduce certain mannerisms became more pronounced as time went on. The only startling deviation from the norm here is his separation on the opening gestures with unwritten pauses – not just between the fortissimo and piano phrases, which several conductors have done, but even between the two fortissimo phrases. That said, his introduction is grave but not portentous. The main theme of the Allegro is introduced with an innocent flow and the movement then develops with forward-moving energy without being over-driven. No repeat. Flowing innocence again characterizes the Andante. Most of this is touching and delicious but, like others who have taken Haydn’s two-in-a-bar too literally, he cannot entirely avoid the music sounding excessively busy when the semiquaver triplets arrive. First repeat only. The minuet bounces along merrily. Bernstein holds up slightly the rising third motifs that introduce the trio, but what follows if delightful. The finale, without repeat, is swift but, thanks to the virtuosity of the NYPO, it does not sound furious or frenetic, just joyful and exuberant. A lovely performance.
Christopher Howell © 2026
| I | II | III | IV | |
| Annovazzi | 07:33 repeat omitted | 09:57 1st repeat played | 05:19 | 04:58 repeat omitted |
| Krauss | 07:15 repeat omitted | 07:12 1st repeat played | 4:23 2nd section repeat of minuet omitted | 07:57 with repeat |
| Ginzburg | 10:29 with repeat | 11:10 with both repeats | 04:49 | 06:32 with repeat |
| Szell | 06:26 repeat omitted | 06:44 1st repeat played | 04:54 | 04:43 repeat omitted |
| Blech H | 09:00 with repeat | 09:34 with both repeats | 04:56 | 06:26 with repeat |
| Rowicki 1959 | 08:02 with repeat | 09:30 with both repeats | 05:30 | 06:05 with repeat |
| Rosbaud 1956 | 08:51 with repeat | 08:16 1st repeat played | 05:31 | 05:17 repeat omitted |
| Rosbaud 1952 | 06:29 repeat omitted | 07:37 1st repeat played | 04:36 | 05:05 repeat omitted |
| Horenstein | 09:24 repeat omitted | 08:48 1st repeat played | 05:54 | 05:22 repeat omitted |
| Kempe 1956 | 07:18 repeat omitted | 08:35 1st repeat played | 05:12 | 05:02 repeat omitted |
| Kempe 1975 | 08:43 with repeat | 08:22 1st repeat played | 05:12 | 06:30 with repeat |
| Wöldike | 08:54 with repeat | 09:06 with both repeats | 05:01 | 07:07 with repeat |
| Münchinger | 07:21 repeat omitted | 07:59 1st repeat played | 05:37 | 05:21 repeat omitted |
| Bernstein | 07:21 repeat omitted | 06:44 1st repeat played | 04:54 | 04:51 repeat omitted |
[1] In the Eulenberg edition, edited by Ernst Praetorius, this is marked down to mezzo forte followed by an immediate diminuendo, but Krauss does not make any dynamic variation at all.
[2] Edward Sackville-West and Desmond Shawe-Taylor with Andrew Porter and William Mann: The Record Guide, Revised Edition, Collins 1955, p.355.
[3] https://classical-discography.org/
[4] Antony Hodgson: The Haydn Seekers, CRC Magazine, Winter 2001, p.23.
[5] H. C. Robbins Landon, Haydn. Chronicle and Works, Volume III, Haydn in England 1791-1795, Thames and Hudson 1976, p. 498.
