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. 93 in D major Hob. I/93 (1791)
A pdf version of this article can be downloaded for offline reading and printing here.
Haydn’s 93rd Symphony is numerically the first of the “London” symphonies, though in reality it was preceded by nos. 96 and 95. It needs no particular introduction. Those who follow older performances with the Robbins Landon, or other Urtext, edition, will notice many small variations in phrasing and dynamics but the difference that will strike lay listeners comes at the beginning of the second movement, the first eight bars of which Haydn marked to be played by solo strings. This indication was suppressed in older editions so only the later performances in this survey, with its cut-off date around 1972, will observe this direction. This same movement is famous, of course, for the “fart”. After the music seems to be tapering away into silence, the bassoon (two of them according to the score, but in most recordings it sounds like just one) plays its bottom note fortissimo. Some conductors revel in this, others seem embarrassed and try to give it a purely musical sense, as a prelude to the fortissimo entry of the whole orchestra. The other controversial issue may be the tempo for the third movement, which is entitled Menuetto but is also marked Allegro.
I have drawn on the combined resources of WERM[1], CHARM[2] and Mike Gray’s Classical Discography[3] to produce the following list of commercially issued recordings up to and including the two cycles issued in the early 1970s.
Beecham LPO 18.12.1936
Cantelli NBCSO 2.3.1949
Beecham RPO 24.4.1950
Scherchen Vienna State Opera O 6-7.1952
Singer Austrian SO 7.1952
Krauss Bavarian RSO 4.4.1953
Beecham RPO 4.10.1957
Jochum Dresden Staatskapelle 6.11.1967
Leinsdorf Boston SO issued 1968
Szell Cleveland O 19.4.1968
Jones L Little Orchestra of London 4.1968
Kempe Munich PO issued 1971
Bernstein NYPO 7.12.1971
Märzendorfer Vienna Radio O (cycle completed in 1971)
Dorati Philharmonia Hungarica 5-6.1972
Of the various live/broadcast performances around, some more easily found than others, the following caught my eye: Beecham Philadelphia O 28.11.1941 (1st movement without conductor, see explanation below)
Krauss Bavarian RSO 7.1952
Szell RAI Roma 29.5.1954
Szell Vienna SO 17.6.1954
Cantelli NYPO 6.3.1955
Newstone Haydn O 22.5.1955 (included in Cameo/Itter box)
Ančerl 24.4.1957 (issued on Tahra)
Sargent BBCSO 4.8.1961
Swarowsky Cologne Radio SO 2.1962
Kempe MPO Liverpool 2.3.1971
Leinsdorf Chicago SO 1.11.1979 (after cut-off date, mentioned for comparison with 1968 recording)
Leinsdorf Philadelphia O 1 or 2.4.1983 (as above)
Some conductors find pre-echoes in this work of Beethoven’s Second Symphony in the same key, but Sir Thomas Beecham, in his 1936 LPO version, is not one of them. After a stern opening rap, the answering string phases are delicate curlicues. The Allegro assai is taken steadily, almost suggesting a minuet with its suave airs and graces. Beecham exacts very short staccatos from his string players, with plenty of air between the notes and very detailed phrasing, though with an emphasis on daintiness and none of the spikiness of some modern performances. The violins’ repeated triplets in the transition to the second subject are not an exuberant tremolo but are clearly articulated. Occasional touches of sternness are allowed in the development (the repeat is omitted) but the prevailing mood is of a baroque ceiling with wafting cherubims. In the second movement, he makes each separate phrase a bob and a curtsey, usually with a smile, sometimes with a sigh. At the end, Beecham, who surely liked a good fart as much as anyone, has the bassoon play as delicate a piano as it can down there (the marking, at least in the Robbins Landon score, is fortissimo). Solo strings are not used at the beginning. Rococo grace is again to the fore in the slowish – but not heavy – minuet. The second repeat is omitted, but back in the days of 78s it is impossible to know if a missing repeat really reflects the conductor’s wishes. Dynamics are altered here and there – the minuet finishes with a diminuendo the first time round, a crescendo in the da capo. Haydn’s marking is simply forte. The double-stopped string chords in the second part of the trio are bowed as marked the first time but replaced with pizzicatos in the repeat, with what authority I do not know. The Presto ma non troppo finale has the emphasis very much on the non troppo. No trace of pre-Beethovenian ebullience here. All is gentle and graceful until … Dare I suggest the unmarked accelerando at the end is rather vulgar? The brief second repeat is omitted. In a way, this is wonderful. Several generations found it wholly so and I respect those for whom it still is. Personally, I find too much of the “Papa Haydn” here and my attention is continually drawn to the clever things the conductor is doing with the orchestra. Haydn sounds a much greater composer in several later performances.
Maybe already in Beecham’s own 1941 Philadelphia performance. This is a piece of history for the conductor’s fans. Beecham arrived late for the concert, blaming the hotel clock. Since the Philadelphia orchestra, like time and tide, waits for no man, the orchestra manager had them start without him. By the end of the first movement the great man was waiting in the wings and, as the announcer explains, came on stage, amid much applause, to conduct the rest of the concert. Given the unsettled circumstances and the scrunchy, crumbly recording, it might seem that I have said enough already, but you pass over Beecham at your peril. Determined to show the public what they had just missed, he straightway took complete command and drew the most exquisitely nuanced performance of the second movement imaginable. Even the fart is given with the best of galante manners. The minuet, if on the slow side, has a robust lilt. There are some odd dynamic manipulations but the double-stopped string chords in the second part of the trio are not played pizzicato as in 1936. The finale is dashing. This should have avoided the need for an accelerando at the end, as in 1936, but he makes one all the same. Repeats are in place in the minuet and finale, but that in the first movement is omitted. The conductorless first movement is straightforward and robust. Perhaps even this tells us something. Probably Beecham rehearsed it that way, but with the idea that this would give him the raw material on which to build during the concert itself. Most significantly, this performance has nothing of the “Papa Haydn” aura of the 1936 one.
The NBC SO provides Guido Cantelli with fiercely clear, often very staccato, articulation in the outer movements. Tempi are not excessively fast, but the music is firmly driven. You would not think of Beethoven here, but you might think of Mozart in resilient mood, such as in the “Prague” Symphony. The first movement repeat is given, the brief second repeat in the finale is omitted. The second movement is similarly conceived to Beecham’s but while Beecham, by some particular magic all his own, has you waiting during the pauses and wondering what he will do next, under Cantelli the music, though beautifully shaped, simply stops and starts. He appreciates the bassoon joke, but earlier in the movement, when the bassoon doubles the melody and was well to the fore under Beecham, its colour is barely perceptible. This may be the fault of the engineers. No solo strings in the opening bars. I thought the minuet and trio very effective – brisk and tough. Maybe I would have enjoyed this performance more in a less dry acoustic, one that emphasizes less the staccato severity of the performance, so does anything change with his 1955 live recording?
The sound is pretty good for an off-the-air recording of that date and I have the idea that careful equalization could make it sound better still. Six years is a long time in the artistic development of a greatly gifted young conductor and, while the strings’ articulation is excellent, there is none of the fierceness of the NBC version. The first movement has fire but also breadth and the phrases hang together in the second movement, where he adopts a slightly more flowing tempo. The minuet seemed a little deliberate at the beginning but it soon develops a fine swagger, while the finale has energy without sounding driven. All repeats are now present. Above all, this performance does not make me think of Beethoven or even Mozart, just of Haydn. This is clearly the performance to remember Cantelli by.
In the meantime, Beecham had re-entered the fray. As in Philadelphia, he has freed his interpretation of the “Papa Haydn” image. Indeed, the first movement – still without repeat – is so strongly spelt out as to verge on the pedantic at times. The second movement lacks that magic that can be dimly perceived in the Philadelphia recording – perhaps he really needed an audience to do this. The minuet is steady but punchy – and has both repeats now. There are still some odd dynamic manipulations and the strange substitution of pizzicato for the double-stopped string chords in the second part of the trio on its repeat. The finale has plenty of drive but lacks the second repeat and still retains that unseemly dash to the finishing post at the end. I prefer Cantelli 1955 to this, but in 1950 it was the best version so far.
Hermann Scherchen immediately announces that he means business (when does he not?) with a real Adagio of an introduction that suggests both drama and spiritual depth. His Allegro assai has a dignified gait. With more sustained playing and a reverberant acoustic, this is pre-Beethovenian Haydn, but not the ebullient Beethoven of the Second Symphony. At a tempo that many (though maybe not Scherchen) might think well suited to the opening movement of the “Eroica”, this is grand, fiery and dramatic. The repeat is omitted. The Largo cantabile is so slow that I initially gasped with astonishment. Yet it is a part of Scherchen’s genius that he can take a tempo that seems perversely wrong, and play it so convincingly that you end up by thinking, at least for the duration of the performance, that perhaps everybody else is wrong. In spite of the slowness, and the consequent length of the rests, his intensity and moulded phrasing lead the ear on. The music never falls apart. It seems a visionary movement, divorced from its time and place. In this context, the bassoon incident seems no longer a joke but a mysterious intrusion from another world. No solo strings at the beginning. The minuet is rather like Cantelli’s 1949 recording, brisk and tough. The finale is urgent and fiery, though the timing shows that it is not so very fast. The second brief repeat is missing. At the end, Scherchen brings out the trumpet’s triadic theme followed by hammered out repeated notes to suggest a parallel with the end of Beethoven’s Fifth Symphony. Say what you will, this is a remarkable performance.
George Singer’s Remington recording has eluded me, though a few of this American conductor’s performances have appeared on various blogs.
I list above both a studio recording[4] and a live one by Clemens Krauss. It is the latter, from 1952, which is most commonly found – on Orfeo and Archipel – and I have not heard the other (if there are really two). Krauss points up the parallels with Beethoven in ebullient mood but has difficulty in finding tempi where everything fits in perfectly. The first movement goes at a real lick, at the expense of some scruffy playing from the Bavarian Radio Symphony orchestra, which evidently achieved a state of polished excellence some years later. The tempo slackens for the second subject, but not in the lead-in, which would be plausible, if hardly the best “classical” style. After one or two bars, it is as though Krauss realizes this is too fast for comfort and puts on the brakes. The repeat is not played. The second movement is the most successful – mobile and songful. Haydn actually marked this movement 2/2 not 4/4. I do not really see how it could swing along two-in-a-bar, and would be curious to hear if anyone has tried this, but Krauss makes a nod in that direction. No solo strings at the start and a very polite bassoon at the end. The minuet trips along nicely, though with some odd departures from the written dynamics. The second part is made to end with a diminuendo each time. The trio has a slower tempo, particularly in the answering string phrases. Something strange happens towards the end of the first part of the Trio. The tempo slackens further, portentously, and the music almost grinds to a halt. The return to the minuet is another odd moment. Krauss begins at the same tempo as the trio, then accelerates to his first tempo after a couple of bars. The finale is excitingly driven if a bit messy, with some loss of momentum in the second subject. There is a weird dynamic alteration in the last eight bars, of which bars three to six are suddenly played piano. The two small repeats are played. All in all, this sounds like a gifted conductor busking it on limited rehearsal time.
After this initial small burst of interest, Symphony 93 did not find much favour with the recording companies in the 1950s and 60s. The next recording appeared in 1957 – Beecham’s third. Given the date, it might have been in stereo, but it was mono only. Beecham again meant still no first movement repeat, still no second repeat in the finale, further unscripted dynamic manipulations in the minuet and the weird pizzicato variant in the trio repeat. That said, it was his best. The introduction begins sternly – the soft string phrases are beautifully tailed off but with no suggestion of rococo curlicues. Beecham seems unable to have rid himself of the idea that the Allegro assai was a minuet writ large, but he manages a greater surge this time, erasing the suggestion of pedantry I found in the 1950 version. The second movement is simply lovely this time. This is all to the good except for the fart, which is simply lovely too and presumably should not be. In spite of the oddities, the minuet has an infectious swagger and the finale has abundant dash. There is still the grandstand accelerando at the end, but given the more dashing context, the adjustment is smaller and less disturbing. Purchasers in 1957 had a quandary. This was certainly the best recorded so far, even if it was in mono. Those who went back to Scherchen got a fuller idea of the greatness of the symphony, but the sheer uniqueness of the experience might have been uncomfortable for repeated listening.
And so the situation remained for a decade[5]. Moreover, the next recording to appear, and the first in stereo, was not widely known at the time since it hailed from the other side of the Iron Curtain. This was Eugen Jochum’s 1967 recording with the famed Dresden Staatskapelle. Jochum’s complete set of the “London” Symphonies with the LPO falls just outside my cut-off date, but there is general consensus that the conductor’s occasional Haydn recordings from the 1960s were better. Certainly, it is difficult to imagine Jochum, or anyone else really, improving on this Dresden recording. The very slow introduction, full of dramatic contrasts, is deceptive – this is not going to be a Brucknerian reading at all. The Allegro assai[6] is exactly that. Without reducing the music to a gabble, it has a coursing vitality that carries all before it. As this part of the movement starts we note at once from the phrasing that Jochum is using the Robbins Landon edition, or at any rate an Urtext score. Small differences, but they all show that Haydn had it right. The repeat is played. The second movement differs from previous recordings on two vital points. The first eight bars are played – at last – by solo strings. And Jochum has seen that this movement, while marked Largo cantabile, is not in 4/4 but in 2/2 (cut time). The old Breitkopf edition had altered this to 4/4. Jochum is not so slow that he can actually be conducting in two, but he does manage to give the feeling of a very slow two-in-the-bar. The result does not yield in sheer loveliness to Beecham – and the bassoon is allowed to blow off properly. The minuet is just that little bit faster than Beecham – without trying to create a Beethovenian scherzo – to set up a swinging lilt. The Presto ma non troppo finale goes at a lick – with both repeats – that might have been troppo with a lesser orchestra, but here everything is under control and the symphony spins to a whirling conclusion. What is there not to like? Well, in the light of what was to come, you might miss the clarity of a chamber orchestra, combined with ringing trumpets and hard timpani sticks. This is plainly a full-scale symphony orchestra in a warm acoustic. Given that premise, it is a wonderful performance.
Recorded at about the same time[7] and issued in 1968, the next recording also circulated little outside its country of origin, though for different reasons. This was the version by the Boston Symphony Orchestra under Erich Leinsdorf (RCA LSC 3030). Even in the USA, Leinsdorf’s Boston tenure aroused no great enthusiasm. In the UK, his records did not sell and RCA practically stopped marketing them. Ironically, when his Haydn coupling of Symphonies 93 and 96 came out at mid-price in 1971, it was one of the select few that enjoyed a generally good press. Like Jochum, Leinsdorf uses an authentic edition with correct phrasing and solo strings at the beginning of the second movement. It sounds as though he may have scaled down the orchestra somewhat. We have the word of the Philadelphia Enquirer[8] that for a much later performance with the Philadelphia Orchestra he had just 44 players on stage. On the other hand, the Chicago Tribune[9] tells us that in 1979 with the Chicago SO he used “a larger body of strings and wind than the 40 originally employed” without specifying the number used. I doubt if the Boston recording had all that few and Leinsdorf presumably relied on refined playing and careful balancing. A search through the Boston press in those years shows that Haydn was programmed frequently and for the earlier symphonies Leinsdorf led the performance seated at the harpsichord – not the case with this late symphony, clearly. Compared with Jochum, Leinsdorf is a shade more formal, interpreting Haydn perhaps from the Mozartian end of the telescope. It is beautifully phrased, especially the delicately drawn second movement that goes at a tempo not far from Jochum’s – though the fart is kept firmly in check. The first movement is bright and perky, there is a fine strutting rhythm to the minuet and a vivacious finale. And yet, of the two, I would have to choose the Jochum. The Leinsdorf is not one of his didactic performances and is fully enjoyable, but Jochum provides that extra surge of electricity that has you not just satisfied but dancing in your seat.
Though my cut-off date is 1972, the two later performances mentioned above circulate among collectors and I will briefly mention a few of their salient points. In 1979, Leinsdorf reverted to the old edition, with differences of phrasing and no solo strings at the beginning of the second movement. Neither performance has the first movement repeat. The 1983 performance has the authentic edition back and sounds big and full in spite of the scaled-down band. It has an almost fierce clarity to the articulation. The second movement now sounds more droll than graceful and the bassoon is allowed a more substantial fart. The minuet is deliberate but punchy and finale is slower and more sedate than the other two though there is compensating energy and the timings show that the difference is not all that great. The studio recording is the one to remember Leinsdorf by, though it would be interesting to hear the 1966 Boston performance if a recording exists.
Hard on the heels of Leinsdorf/Boston came George Szell’s Cleveland recording. It is a truism that Szell had that touch of greatness which eluded Leinsdorf and this version has always been held in high esteem. On the face of it, the two do not seem all that different. While Szell made it an article of faith to conduct Bruckner from corrupt scores, he has obviously got an Urtext edition of the Haydn so we get solo strings at the start of the second movement. Moreover, as the Allegro assai starts, he not only respects the differences of phrasing but points them up as if he wants anyone with ears to hear them. At a tempo similar to Leinsdorf’s – so slower than Jochum’s – he combines grandeur with elegance. The second movement is also similar in manner to Leinsdorf, though I noted a few points in which Leinsdorf showed greater warmth where Szell maintained a serene poise. In the third bar of the minuet, Szell has no truck with the Urtext edition and punches out the quavers staccato – the Urtext edition has them bowed legato –as though saying “This is how I’ve always done it and I’m not changing for anyone”. The movement never quite recovers from this didactic start. Though the tempo is practically identical to Leinsdorf’s, Leinsdorf is better at setting up a buoyant swing. Towards the end of the da capo minuet, Szell makes a rallentando and concludes the final statement at a slower tempo, something that might have worked better if the playing had exuded a greater sense of enjoyment. In the finale, it is Haydn himself with whom Szell has an issue, eliminating his second repeat. The movement is elegantly turned but Szell holds things up here and there for a little point making, something Leinsdorf avoids to good effect. So, if you want an elegantly formal performance as a foil to Jochum’s élan, it seems as if Leinsdorf is your man. Except for the small matter of the celebrated Cleveland fart. This really is the biggest, blackest bacon-and-beans blow-off you are ever likely to hear. Szell allows the player to hold it for longer than the written crotchet and he then curdles below pitch at the end. As the full orchestra comes in you can imagine the whole party rounding on the perpetrator and demanding to know where he has left his manners. After which the final piano phrases are elegantly turned as if butter would not melt in the players’ mouths. This is as hilarious as anything imagined by Hoffnung, but would you make this your go-to version for that alone? I realize there are people who can listen to the same dirty joke over and again and laugh like a drain every time as if it were new. For those who, like me, fancy they have a slight artistic-intellectual streak, once you know how the story ends, you know. Funny to tell your friends, maybe, otherwise best kept as a memory. So I am glad I have heard it but the rest of the performance does not especially tempt me back.
Even if Szell had to wait till 1968 to record this symphony, he was no stranger to it and two recordings survive of him conducting it in Europe in 1954. The earliest, by a small margin, was made with the Orchestra Sinfonica di Roma della RAI. Szell was not an assiduous visitor to the RAI orchestras. In 1954 he conducted a concert with the Rome orchestra consisting of Mozart, Schubert and Bloch. It was common practice for RAI to grab visiting conductors and have them set down something extra in the studio if time permitted, thus Szell got to do his first Haydn 93. Useless to expect an Urtext back then so no solo strings in the second movement, and he omits the second repeat in the finale (but plays that in the first movement). It sounds as if Szell enjoyed challenging the orchestra with tempi that would have had the Clevelanders on the hop, and it sounds as if the Rome musicians enjoyed meeting the challenge, for all is clean and confident. There is tremendous verve in the outer movements and in the second, just at the point where, in the Cleveland version, I noted that Leinsdorf produced more warmth (the excursion into B flat major), here Szell produces more still. The minuet has splendid swagger. The fart is well-behaved but in every other respect this performance leaps to life where the Cleveland one sits on the page. It sounds like the work of a great conductor.
The following month with the Vienna Symphony Orchestra, Szell is a tad more relaxed but this is still a splendid performance. I did wonder if the challenge of an untried relationship in Rome spurred both orchestra and conductor to produce that extra frisson of a great performance. Here he is among old friends. But the only difference of real note is that at the end of the minuet, he makes a schmaltzy upbeat and concludes the finale statement at a slower tempo, as he did later in Cleveland, but hardly at all in Rome. Live in Vienna, it works better though I could personally do without it. The Rome performance remains the one that is clearly the work of a great conductor. It has been issued by Archipel, but I cannot comment on the quality since I have it off-air from a broadcast some years ago.
At about the same time as Szell was recording Symphony 93 in Cleveland, Leslie Jones and the Little orchestra of London were setting down the first complete cycle of the “London” Symphonies with a scaled-down band and Urtext scores. For this particular symphony, indeed, it was the first recording at all with something less than a full-size symphony orchestra – a 1955 recording by Harry Newstone and the Haydn Orchestra has turned up more recently and will be discussed below. Not that it sounds so very small. I recall from my student days, when the now-disbanded Reid Orchestra performed regularly in Edinburgh University’s Reid Concert Hall, which seated an audience of slightly more than 200 and had a fairly reverberant acoustic, that Haydn and Mozart symphonies played by a chamber-sized orchestra in such a venue could be impressively noisy affairs. The Conway Hall, London, where the Jones recording of Symphony 93 was made, seats about double that, but is still classed as a small hall. The Hanover Square Rooms, where Haydn premièred the work with an orchestra of forty-odd, could seat, at a pinch, double that again but, with a wooden floor and vaulted ceiling, the effect is likely to have been quite fulsome. From that point of view, the Jones recording sounds about right. Wind, brass and timpani blaze through the texture in the fortes and there is a harpsichord continuo played by Harold Lester. How much he plays, I cannot say – so far as we know, Haydn himself, while presiding over the performances at a harpsichord (and later a piano), did not play all the time. Once or twice I had a vague idea that a harpsichord might be there, not necessarily in the same room. It was rather a surprise, then, when it filled in an empty bar just before the end of the first movement with a little flourish. Either you do this all through, I would have thought, or you stay at home. It is not as if these late symphonies have harmonic holes to be filled in. The other question is, how to play? There is plenty of vitality, rather less grace – the second movement is rather bumpily phrased. Then, in the forte passages, there is no attempt to grade the dynamics and pace the climaxes – the instruments simply blast away at each other until the next piano passage comes along. I found it rather tiring to listen to. But, you will, say, before the invention of the conductor, that was the best they could do. Maybe, but are we so sure that an age in which such beautiful music could be written, did not find a way to play it beautifully too? Are we so sure that Haydn, who surely had a strong personality, could not influence proceedings while seated at his harpsichord just as much as Erich Leinsdorf could while seated at his? This induces the reflection that, just as this is the first recording with a scaled-down orchestra, it is also the first under a conductor without some claims to “great” status[10]. Refreshing as this enterprise sounded in its day, I have to say that the finest of the recordings mentioned so far offer an experience of a different order.
Two more recordings were set down in 1971, just before my cut-off date. That by Rudolf Kempe and the Munich PO seems a strange affair, coupled with Genzmer’s Ballet-Suite Der Zauberspiegel under Reinhard Peters (Da Camera Magna SM 91509). I have not been able to hear it, but I have heard a live Kempe recording made in the same year, which I will discuss below.
CBS’s New York PO recordings had the reputation, at least in the UK, of being thick and unsubtle, but Leonard Bernstein’s recording of this symphony falls much more pleasantly on the ear than Jones’s, both for the transparent orchestral textures and the warm, unforced recording quality. Bernstein is using an Urtext edition and plays all the repeats. In the introduction, he differs from other conductors in his treatment of the turn in bar four and on its subsequent appearances. He plays it as four equal 32nd notes, while all other conductors heard so far have held the first note longer and squeezed the turn in at the end. I am not aware of any musicological reason for saying one or the other is right. His Allegro assai goes at a relaxed tempo – he seems to share Beecham’s idea that this is a minuet writ large. It is good-humoured and elegant but I feel there is a grandeur to the music that is missing. For his second movement, all is forgiven. His string quartet seems to have come straight down from heaven and I don’t think anyone has shaped this movement so exquisitely since Beecham in Philadelphia. His bassoon is made to mind his manners, but you cannot have everything. Bernstein’s minuet is so slow that not even he can keep it buoyant but his finale is a delight. Though taken at a good presto, he keeps it all light and dancing, without attempting Beethovenian drive. A mixed bag, but I cannot throw out a version with a wonderful second movement and one of the best finales.
Time for a look at the two cycles.
From the outset, with a three-in-a-bar tempo for the introduction and the accompanying quavers staccato as marked, it is clear that Märzendorfer is not in the business of creating proto-romantic atmospheres. His Allegro assai is as fast as Krauss’s, but with no slackening for the second subject. With exciting staccato articulation, ringing trumpets and hard timpani sticks, there is an exhilaratingly festive, outdoors feeling to the music-making. With a small orchestra sounding big, it is hard to say why this works where Jones’s did not, but somehow the pacing and balancing puts it in another world. The repeat is played. The Largo cantabile, with solo strings at the beginning, does not evoke the Elysian fields. Instead, the separate, bobbing phrases have a pawky humour. There is at least an attempt at a two-in-a-bar feeling. The bassoon joke is adequately fruity. The minuet has a vigorous stride while the finale, not taken too fast for clarity, is again festive, even bucolic in the second subject. As so often, Märzendorfer seems to stand as a gateway to the HIP performances that were to come.
Dorati is less radical – trumpets, drums and wind generally are more absorbed into the texture and the string articulation, while clear and crisp, is less gutsy. For those who find Märzendorfer too abrasive, this could be the answer. The first movement (with repeat) bowls along very nicely, the second (with solo strings at the beginning) is above all elegant though with a properly relished bassoon joke, the minuet is buoyant and the finale is vital and neatly turned. If you expect Haydn to sound delightful, devoid of disturbing features and not looking beyond his own time, this should be about ideal. I have to confess I began to lose interest towards the end.
I have already discussed several of my listed live performances in conjunction with their conductors’ commercial recordings. A few remain.
Of particular interest is Harry Newstone’s 1955 performance with the Haydn Orchestra[11], since this symphony had a long wait for a recording with a Haydn-sized orchestra – Leslie Jones’s – and I did not find the result very convincing. This earlier outing is significantly different and it has to be said that, for all the conductor’s dedication to the cause, an Urtext edition was evidently not available in 1955, so there is some “wrong” phrasing and no solo strings in the second movement. Like Jones’s Little Orchestra of London, the Haydn Orchestra makes a pretty full sound, but in a different sort of way. I began this article by remarking that several conductors find anticipations in this work of Beethoven’s Second Symphony in the same key, but this matter has not been particularly present in the subsequent discussion. Newstone was one of the conductors I was thinking of. After an introduction that is not excessively slow but alert to the unexpected key changes and dynamic contrasts, his Allegro assai bowls along at a tempo in which everything falls into place. The repeat is missing, but this does not prevent the music from coming across with real stature. In the Largo cantabile second movement, the conductor’s well-judged tempo allows the music to sing in long lines in spite of the many rests and, once again, the various incidental details all fall into place, including a properly fruity bassoon. The minuet has a broadly swinging gait while the Presto ma non troppo finale is swift and vivacious with an unbuttoned fullness that may strike us today as Beethovenian rather than Haydnesque. Both repeats are observed. No one hearing this performance would doubt that they had heard a masterpiece of real stature. As with other Newstone performances in the Cameo/Itter set, I think this is a studio recording made for broadcasting – I hear no sign that an audience was present.
Karel Ančerl’s Berlin Radio SO performance was likewise a studio recording made for broadcasting. I reviewed this Tahra box more than twenty years ago[12], noting that “Tempi are well-chosen, the phrasing is well-prepared and always musical, but it has left me with no particular memory.” If proof of the latter were needed, I have to confess that, had I not come across a reference to this recording on the internet while looking for something else, it would not have found a place here since I was – I thought – unaware of its existence. I did, on the other hand remember being mightily impressed by the Schubert 9 and Dvořák Slavonic dances in the same box. My colleague Jonathan Woolf reacted similarly[13] in saying that “It’s not a revelatory reading but it’s hardly a negligible one either.” Looked at in the historical perspective of this article, perhaps there is more to be said. The situation re this symphony was not too good in 1957 and did not improve for another decade, so an Ančerl commercial recording, if Supraphon had had the wit to make one, could have held an honourable position. Ančerl is tough but lively. After a fairly literal introduction, the Allegro assai is brisk and vigorous. Not much charm but plenty of conviction. No repeat. This, at the time, might have seemed a refreshing foil to Beecham’s minuet-style reading or Scherchen’s breadth and grandeur, imposing though it is. The second movement, at a slowish tempo, hardly has Beecham’s magic or Scherchen’s amazing concentration, but it does not hang fire. The bassoon plays as per the score. The minuet is swifter than usual in those days and again tough, while the finale, with both repeats, is driven very briskly indeed. And yet, for all its excellence, the music does not leap off the page as it sometimes could with Ančerl, or as it did under Jochum in 1967, or had under Szell and Newstone in their broadcasts a few years earlier.
Or as it did under Sir Malcolm Sargent. Haydn does not seem to have occupied a prominent place in Sargent’s repertoire, though he recorded Symphonies 94 and 98 in the 1940s and there is a tape of him conducting no. 88 in Montevideo in 1950. I was uncertain what to expect of his Prom performance of no. 93 on 4 August 1961. What I did not really expect was that it would be very similar to Newstone’s in style and conception – a taut pre-Beethoven-Second view with tempi in the outer movements ideally chosen so that everything falls into place. The second movement retains a long singing line through all the rests and relishes the bassoon joke adequately, while the minuet is buoyant with no slackening for the Trio. Furthermore, Sargent seems fully engaged, conducting with the sort of disciplined exuberance one imagines he might have applied to Beethoven’s Second. We need not only imagine this, by the way, since a performance from the 1965 Proms has been preserved, but that will be for another time. I should be hard put to choose between this and Newstone, but Sargent did have that extra bit of “flash”. The first movement repeat is omitted, those in the finale are played and, inevitably, solo strings are not used at the beginning of the second movement.
In my article on Hans Swarowsky[14], I described his performance, set down with the Cologne Radio Symphony Orchestra in February 1962, as “magnificently fiery, grand but delicate where required”. In the light of the traversals I have just heard, how does this stand up? Swarowsky is less radical than Märzendorfer, but by clear articulation and clean textures he manages to be energetic without in the least suggesting pre-Beethoven. In some ways, this is very similar to Dorati, even in his tempi apart from a notably swift introduction to the first movement. He has, though, an extra strength and engagement that goes beyond the merely delightful. It is also interesting that, though the Robbins Landon edition is dated 1965, Swarowsky in 1965 was using solo strings at the beginning of the second movement and in several respects his phrasing corresponds to that in the Urtext edition, which removes various fussy slurs in older editions. I could wish his bassoon joke had not been so delicately tasteful, but in other respects my previous description can stand. All repeats are played.
Though Rudolf Kempe’s rare commercial recording has eluded me, he included this symphony in the repertoire for a tour in the same year which included a visit – which was broadcast – to the Liverpool Philharmonic Hall. Neither Haydn nor Mozart figured largely in Kempe’s programmes so I wondered what to expect, especially since I remembered him conducting a Mozart symphony with the RPO in Edinburgh that I found so over-preened as to be rather repulsive. Nothing of that here. I wonder what edition he was using? At first, the phrasing seemed to correspond to the Urtext score, but there are no solo strings at the beginning of the second movement and staccato quavers in bar three of the minuet. But bars 165-8 of the finale are correctly played by a solo cello. Very oddly, in bars 72 and 211 of the first movement the appoggiaturas are not only played as equal quavers – about 60% of the conductors here follow this option – but has them bowed separately and brilliantly so they do not sound like appoggiaturas at all. A pic’n mix edition of his own? He does at least give all the repeats. A dignified introduction is followed by an Allegro assai that has a certain gracious quality, though it is sufficiently fast to achieve majestic energy – more Mozart than Beethoven – as it proceeds. The second movement is quite mobile. There is some beautiful shading here and there, but the forte passages run ahead slightly disconcertingly. The fart is as well relished as by anyone outside Szell/Cleveland. The minuet seemed deliberate at first but actually develops a good swing. Best of all is the finale, with real verve and some excellent phrasing. A lot of good things, then, but a little uneven.
Conclusions? From Märzendorfer you get a radical view that looks ahead to later HIP performances. If this is too much for you, but you want a clean text and all repeats, Jochum is simply super, with Swarowsky not far behind and Leinsdorf 1968 for a more formal, but not didactic, alternative. If repeats and editions do not worry you, Szell’s Rome performance is splendid, as is the live Cantelli. For a big-limbed, exuberant pre-Beethovenian slant, Sargent marginally outdoes Newstone, though the latter is readily available and Sargent is not – it would make a nice addition to his discography. If the fart is the most important thing for you about the symphony, it will have to be Szell/Cleveland. Scherchen’s iconoclastically individual interpretation has the stamp of greatness. As has, in its way, Beecham’s 1936 version – Beecham’s genius if not Haydn’s. Beecham moved away from his “Papa Haydn” interpretation, but so did others, more effectively. Nor should I forget Bernstein’s best two movements, and most of the others, except perhaps Krauss, have their points of interest.
One final remark. According to the Robbins Landon edition, this symphony has a duration of c. 26 minutes. I wonder what tempi he considered to be ideal? It is true that, if you made yourself a compendium version using the slowest performance, with repeats, of each movement – so Bernstein-Scherchen- Beecham/Philadelphia-Dorati – this would stretch out to 29:12, but in reality I doubt if a single conductor has ever given such a performance. The longest all-repeats performance below, Bernstein’s, takes 24:56. Märzendorfer whittles this down to 20:37, Jochum to 22:00.
| I | II | III | IV | |
| Beecham 1936 | 06:42 no repeat | 06:17 | 04:08 2nd repeat of minuet omitted | 04:39 no 2nd repeat |
| Beecham 1941 | 06:48 no repeat | 08:06 | 05:03 | 05:02 both repeats |
| Cantelli 1949 | 07:46 with repeat | 06:22 | 04:15 | 04:47 no 2nd repeat |
| Cantelli 1955 | 08:13 with repeat | 05:40 | 04:24 | 05:14 both repeats |
| Beecham 1950 | 06:39 no repeat | 06:09 | 04:48 | 04:36 no 2nd repeat |
| Scherchen | 06:54 no repeat | 09:39 | 04:10 | 05:17 no 2nd repeat |
| Krauss 1952 | 05:32 no repeat | 05:28 | 04:30 | 05:00 both repeats |
| Beecham 1957 | 07:17 no repeat | 06:16 | 04:49 | 04:46 no 2nd repeat |
| Jochum 1967 | 07:26 with repeat | 05:26 | 04:20 | 04:48 both repeats |
| Leinsdorf 1968 | 08:05 with repeat | 05:15 | 04:45 | 05:06 both repeats |
| Leinsdorf 1979 | 06:20 no repeat | 05:26 | 04:41 | 05:16 both repeats |
| Leinsdorf 1983 | 06:16 no repeat | 05:03 | 04:39 | 05:27 both repeats |
| Szell 1968 | 08:12 with repeat | 05:35 | 04:46 | 04:45 no 2nd repeat |
| Szell 1954 Rome | 07:36 with repeat | 05:19 | 04:39 | 04:29 no 2nd repeat |
| Szell 1954 VSO | 07:47 with repeat | 05:25 | 04:46 | 04:39 no 2nd repeat |
| Jones L | 08:44 with repeat | 06:02 | 04:36 | 05:06 both repeats |
| Bernstein | 08:52 with repeat | 05:51 | 04:55 | 05:18 both repeats |
| Märzendorfer | 06:56 with repeat | 04:57 | 04:00 | 04:44 both repeats |
| Dorati | 07:55 with repeat | 05:24 | 04:45 | 05:38 both repeats |
| Newstone | 06:03 no repeat | 06:29 | 04:07 | 04:42 both repeats |
| Ančerl | 05:59 no repeat | 06:05 | 04:21 | 04:52 both repeats |
| Sargent | 06:32 no repeat | 06:59 | 04:22 | 04:58 both repeats |
| Swarowsky | 07:43 with repeat | 05:23 | 04:22 | 05:11 both repeats |
| Kempe | 07:54 with repeat | 05:07 | 04:36 | 04:40 both repeats |
[1] Clough, Cuming: The World’s Encyclopedia of Recorded Music, Sidgwick & Jackson Limited, 1952; ditto, Second Supplement, 1953; ditto, Third Supplement, 1957
[2] https://charm.rhul.ac.uk/index.html
[3] https://classical-discography.org/
[4] Listed in Gray
[5] The 1963 EMG Art of Record Buying, a select catalogue of recommended available recordings, listed the 1957 Beecham with their maximum two stars, the 1950 version with one, presumably for the poorer sound. The 1965 edition had just the 1957 Beecham, but this was a time when mono records began to disappear – the 1966 and 1967 editions listed no recording of this symphony at all.
[6] Maybe I should point out to English-speaking readers who know more French than Italian that the Italian word assai, though it may be etymologically related to the French assez, has the opposite meaning. So, while Assez vif is only “quite fast”, allegro assai is “very fast indeed”.
[7] I have no exact date since it is not listed in Gray’s Classical Discography. This recording might actually precede the Jochum since Leinsdorf gave Symphony 93 in Boston on 11 November 1966 so quite likely recorded it not long afterwards.
[8] 1 April 1983.
[9] 2 November 1979
[10] Apart from the George Singer recording which I have not been able to hear.
[11] I gave a brief summary of Newstone’s career in my article on Symphony 34.
[12] https://www.musicweb-international.com/classrev/2004/jun04/Ancerl_vol1.htm
[13] https://www.musicweb-international.com/classrev/2002/Dec02/KarelAncerlVol1.htm
[14] https://www.musicweb-international.com/classrev/2014/jul14/swarowsky_forgotten.htm
