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. 67 in F major Hob. I/67 (1779 or earlier)

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

Symphony 67 has some unusual features even by Haydn’s standards. Contrary to the practice at the time, the lightest movement is the first – a tripping Presto in 6/8 which begins out of nothing. Without any introduction, the first violins take up the theme, pianissimo, the other strings offering the lightest of pizzicato accompaniments. There follows a quite extended Adagio in which the strings are muted. At the end, moreover, they are asked to play col legno – that is to say, turning the bow over to play with the wood. Given that the notes are also staccato, should the players really play the notes with the wood or just tap the string lightly? Several conductors go for the latter, which does not sound very different from pizzicato. Having them actually play their notes, however briefly, is a far more eerie effect. The minuet is more “normal” but the trio is played by just two solo violins, muted. The first has the melody, and is asked to play it on a single string – all on the E string, therefore. The second is asked to tune the G string down to an F, which then serves as a drone throughout. The effect suggests bagpipe-like folk instrument¹. The finale sets off with a vigorous Allegro di molto but having reached the end of the first part – which should be repeated – Haydn suddenly inserts an expansive Adagio movement, which begins with just three solo strings² and gradually brings in the entire orchestra, including a section for wind only. The Allegro di molto returns to round things off. It seems that it is preparing to wind down and end pianissimo, but two sharp fortissimo chords close the symphony.

The first recording of this symphony was set down in New York by the Orchestra of the New Friends of Music under Fritz Stiedry on 9^th January 1939 (Victor M-536). As reported in the New York Times of 4^th February 1939, it was one of five Haydn symphonies they presented in concert, specially edited by Alfred Einstein from early sources to replace the “highly inaccurate” mid-19^th century editions which were hitherto the only ones available. The others were 71, 77, 80 and 87. If you follow the performance with the subsequent Robbins Landon edition, Einstein’s conclusions seem virtually identical. Stiedry’s col legno passage is the only one in all these performances where I can really appreciate the effect since, as I discussed above, he has the wood of the bow actually play each note, however briefly, rather than just bounce on the string, so we can hear the strange timbre. The solo strings in the Adagio section of the finale are observed. Indeed, the latter continue even after Haydn has directed that the whole section should resume – maybe the sources are ambiguous over this.

With its concentration on sharp articulation, brittle staccatos in the second movement, driven brilliance in the opening Presto and the Allegro di molto sections of the finale, a minuet that does not hang around and what seems a very small orchestra, Stiedry’s manner may seem a blueprint for later HIP performances. On the other hand, there is no harpsichord – which some HIP conductors do not use either – and no embellishment except a slight elaboration of the cadenza-like moment in the second movement, taken by a solo violin. Repeats are ignored in the first two movements and one is missing from the minuet. One should beware of criticizing 78 recordings on these grounds, given the space available. However, the Adagio section of the finale has all repeats and is taken very slowly. This seems to be the part of the symphony that most engaged Stiedry. Somewhat heartless elsewhere, he shows deep feeling here. It is for this, and the col legno passage, that I would return to the performance.

The second recording, and the first on LP, was by Paul Sacher and the Vienna Symphony Orchestra – it came on the other side of the his performance of no. 53. It was set down on 22-23 April 1953³ and issued on Philips. The full strings play in the Adagio section of the finale though the cadenza-like moment in the second movement is taken by a solo violin and elaborated just as in Stiedry – the embellishments are so similar, and are also heard identically in other performances, as to suggest that they derive from a single source.

Under Sacher, the Presto first movement is hardly more than an Allegretto, though it is played with much delicacy and elegance. The first repeat is played. Unfortunately, it sags in second subject territory, proving that the tempo is too slow after all. Something similar happens in the Allegro di molto sections of the finale, where Sacher seems unable to shape the contrasting material affectionately without losing steam. The minuet goes at a chirpy pace and the trio is very beautifully handled by the solo violin, though his occasional portamento will raise some eyebrows today. It is for the second movement, though, that I will return to this recording. Sacher takes it relatively swiftly – a two-on-a bar feeling rather than four – and invests it with grace and even a little gentle humour. Considering that the emotional weight of the symphony turns out to be in the middle section of the finale, this seems a plausible way to play this movement. It would have carried more point, though, if Sacher had shaped the Adagio in the finale with more than generic niceness. No repeats in the second and fourth movements.

In 1971, roughly contemporary with the two complete cycles, Vilmós Tatrai set down this symphony leading – not conducting – the Hungarian Chamber Orchestra (Hungaroton SLPX 11571). Several instalments of Tatrai’s quite considerable series of Haydn symphonies have emerged in these articles, showing the advantages, and sometimes the disadvantages, of an approach closer to that of an enlarged string quartet than a scaled-down symphony orchestra. The delicate opening of no. 67 sounds an obvious candidate for this approach, and so it proves – it scampers along delightfully. The first repeat is played. The second movement is taken slowly, with a mixture of grace and wistful intimacy. No repeat. The cadenza-like passage is given literally, without any embellishment. The minuet is slowish but alert. Unless the failing is of my own ears, the drone note is not present in the trio, not even an octave higher. The finale, with all repeats, has an unhurried but friendly Allegro di molto and an expansive Adagio played with deep feeling.

And so to Märzendorfer and Dorati. Märzendorfer’s first movement is properly swift, a little more than Tatrai’s, but dances along delightfully, without the fierce drive of Stiedry. The first repeat only is taken. His second movement is sufficiently slower than Sacher’s – and definitely four-in-a-bar – to wear a graver face, yet it is no less light and graceful. No repeat. Märzendorfer is very similar to Tatrai here and he, too, has the cadenza-like passage played literally by all the violins, without embellishment. The minuet is somewhat formal, only just avoiding pompousness, and his violin soloist in the trio is bland – I much prefer Sacher’s, regardless of the portamenti. Märzendorfer is back on finest form in the finale, which has all repeats. His Allegro di molto is merry yet unhurried – this section of the symphony convinces me here as none of the others till now did. The Adagio section is played slowly and lovingly, using solo strings – and returning to them in places after Haydn has indicated tutti. With slight reservations over the minuet, this is a very fine performance.

Dorati’s first movement is so swift as to create little impression beyond that of speed – there is no time for the music to breathe. Märzendorfer, just a tad slower, seems to me to get this movement absolutely right. Dorati gives both repeats, but omits that in the following movement, where he, too, admits no embellishment to the cadenza-like passage for violins. At a tempo pitched midway between Sacher and Märzendorfer, Dorati finds much delicate grace in the music. The minuet is not one of his flabby ones, though there is an archness to the phrasing that sends me elsewhere. The solo violin is bland in the trio. Dorati’s finale, with all repeats, is very good. He does not hurry the Allegro di molto, adopting a friendly gait similar to Märzendorfer’s. His Adagio is likewise slow and reverent, using solo strings as marked, if slightly more formal in tone than Märzendorfer. Despite two excellent movements, this yields to the latter and also, I would say, to Tatrai on account of the over-swift first movement.

A couple of live/broadcast recordings have emerged. In my article on Symphony 53, I discussed an unusual performance by Stanford Robinson. I also gave a few details of the extensive series of Haydn symphonies given by the BBC in the years 1964-1966, involving numerous conductors and all their regional orchestras. Stanford Robinson contributed no. 67 to this series, conducting the BBC Northern Orchestra, a performance broadcast on 6 September 1964. However, the performance I have is from a concert with the BBC Symphony orchestra in the Maida Vale Studio on 7 December 1963. Unfortunately, it is not possible to see from the BBC Programme Index what else was played – though whatever it was involved a fair number of soloists. Later still, in a broadcast of 23 April 1970 with the BBC Scottish SO, Robinson conducted it again. This time the programme describes it as “arranged by Stanford Robinson”. Without hearing that performance, I cannot say whether this was a snide comment by some BBC producer on Robinson’s credentials as an editor, or whether in the meantime the conductor had upgraded his interventions from mere editing to full-blown arrangement. The 1963 performance is somewhat less extreme than that of Symphony 53. There is a harpsichord which is not often evident but suddenly turns up to fill the otherwise empty bar before the reprise of the main theme in the second movement with some flourishes and imitative material. There are some added trills and mordents – the latter sounding agreeably chirpy at the end of the second movement. The cadenza-like passage in this movement uses the elaboration heard in several other versions, but played by all the violins. Evidently he was drawing upon an older edition, since he uses tutti rather than solo strings in the central part of the finale. None of this would justify calling it an “arrangement”. If less remarkable than the performance of no. 53, it is nonetheless a likeable affair. The opening Presto is taken relatively steadily but, after a smudged start, it is beautifully sprung and attentively phrased. Neither repeat is played. The Adagio is played broadly and gravely. No repeat. As with most performances except Stiedry’s, the col legno passage at the end sounds more like pizzicato. The minuet goes at a cheery pace and the solo violin in the trio is allowed a few extra trills. The outer parts of finale are vivacious at an un-hectic speed. Only one of the repeats in the Adagio e cantabile section, tenderly played, is observed.

An unexpected conductor in this repertoire was István Kertész, though he did record Symphonies 45 and 104 and the Sinfonia Concertante with the Bamberg Symphony Orchestra in 1960-61. He gave this symphony as the opener for his concert with the Philadelphia Orchestra on 13 January 1967. He takes Haydn’s Presto marking for the first movement at its word – there are times when even the Philadelphia orchestra seems to be only just managing it. He demonstrates a rare capacity to get feathery pianissimo playing at high speed but, with no repeats, this is Haydn-lite. The second movement is taken as a two-in-the-bar Adagio – this is what the marking says, after all – and emerges as a nocturnal serenade, gossamer light without a trace of heaviness even in the occasional forte passages. No repeat. The minuet is rather dogged, but the trio is interesting. Most conductors have the second violin play its drone accompaniment in a fairly bouncy manner. Kertész has it very sustained and treats the whole trio as a sort of dream episode, almost like one of Chopin’s more arcane mazurkas. Memories of his native Hungary, perhaps. I am glad to have heard the performance if only for this. The finale is fleet in the Allegro non troppo sections while in the central section Kertész evidently interprets the Adagio marking as applying to the bar not to the quaver – arguably correct since the time signature is 3/8 not 3/4. No repeats in this movement. The result is that, whereas in most performances the emotional weight of the symphony comes in this Adagio, under Kertész it is simply an agreeable episode inserted in a lively finale. Haydn-lite again. The conclusion is that Kertész, at this early stage in a career that sadly knew no later stages, had not engaged very deeply with this composer.

Christopher Howell © 2026

1 It would be interesting to see how this was notated in the older editions, since some earlier performances seem to have the drone note played an octave higher. In the case of the Tatrai performance, it does not seem to be present at all. I can see the inconvenience of stopping a live performance in order for the violin to tune the string down, then up again afterwards, but surely this is not a problem for a recording. This might be resolved in a concert by having the viola play rather than a violin. You would not get the resonant effect of the open string, but at least it would provide the correct note. The drone is heard in the Märzendorfer and Dorati recordings. Interestingly, it is present, by whatever means, in the live performances under Robinson and Kertesz.

2 Some of the performances discussed do not use solo strings. Evidently the “highly inaccurate” editions mentioned omit this indication.

3 Information from the Mike Grey Classical Discography.