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. 28 in A major Hob. I/28 (1765)

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

Symphony 28 has several unusual features. Much has been said about Haydn’s “monothematic” symphonic movements – movements, that is to say, where the second subject, if you insist on seeking one, is the same as the first. This one could be called “mono-motivic” since it is entirely dominated by a single motif – three repeated notes on the upbeat of nearly every bar. It was presumably this dogged insistence that inspired Robert Simpson, in 1972, to devise a programme entitled “The Sincerest Form of Flattery” in which the London SO under Harry Newstone played this symphony, Beethoven’s Fifth and Vaughan Williams’s Fourth. Compared with Haydn, Beethoven, who permitted a brief but lyrical second subject to contrast with the relentless hammering that is the main business of his first movement, appears almost digressive[1]. Yet, aside from the improbability that Beethoven knew the early works that Haydn himself had put behind him, the effect is quite different. The Haydn, lightly scored, has a conversational elegance[2] and, moreover, whereas Beethoven invariably follows his three repeated notes with a long note a third lower to create what is vulgarly known as the “fate” motif, Haydn keeps us guessing as to where he is going after his three notes – sometimes the same note, sometimes up, sometimes down. The second movement has muted strings while the oboes and horns stay silent. As with Symphony 8, Haydn contrasts songful phrases with quizzical staccato ones. The minuet is marked Allegro molto – alone in Haydn’s symphonies apart from no. 94. It, too, plays with repeated notes, and in a very odd way. The violins’ opening phrase has six repeated notes, alternating between open strings and stopped notes played with the fourth finger on the string below. The game is to contrast the ringing timbre of the open string with the more contained sound of the stopped one. The performances discussed below show that, if you interpret the Allegro molto marking too literally, there is not time to appreciate the changing timbres. The trio, for strings only, is in the minor key. The first violins play a repetitive phrase while the seconds have double notes. The effect is of a distant, melancholy hurdy-gurdy. The finale resembles a gigue, though without the typical upbeat.

The first recording was set down for the Haydn Society in June 1949 with the American conductor Jonathan Sternberg conducting the Vienna Symphony Orchestra (HSLP 1001, Parlophone PMA 1001 in the UK). Coupled with Symphony 1, it was the Haydn Society’s first LP. Sternberg (1919-2018) was then assisting Robbins Landon in the search for Haydn manuscripts in Europe and conducted several of the Society’s records. He also accompanied, this time for Nixa, Alfred Brendel’s first record – an uncharacteristic choice of Prokofiev’s Fifth Piano Concerto. The Record Guide found the Haydn recording “dry, scratchy and top-heavy”[3], in which case René Gagnaux, from whose currently offline blog Mon Musée Musicale I obtained this, must have done some good work since it sounds reasonable for the date. What he could not do, unfortunately, was improve the performance, which is so heavy-handed and unphrased that I see no reason to discuss it in detail. I will mention one point only. In the Minuet, the players get the hang of the rhythm about half-way through, after which the performance achieves a degree of swing. This sort of thing typically happens in rehearsal at the first reading. Evidently, this being a shoe-string operation, the first reading had to go on the disc willy-nilly. One appreciates that the cash-strapped young society wanted to extract as much music as possible from a single session, but they did Haydn no favours, and unfortunately this recording remained the only one until the two cycles appeared in the early 1970s. The omission of repeats in all but the Minuet and Trio is, under the circumstances, a relief.

The two cycles offered a major improvement on the situation. Both Märzendorfer and Dorati brought a combination of elegance and vivacity to the first movement. Märzendorfer played both repeats, Dorati only the first, which seems a pity in such a short movement. The other difference is that Märzendorfer has his oboes play out in fortes within an inch of overblowing, producing a strong sound similar to a period instrument. Under the circumstances, the engineers need not have gilded the lily further by close-miking them. The result might seem overbearing for some tastes, in which case the more rounded tone of Dorati’s oboists and his more traditionally string-based recording could be preferred. Märzendorfer treats the second movement as a delicate serenade, with charming results. Dorati is marginally slower and perhaps burdens the music with more expressivity than it can bear. In order to keep things interesting, he makes some rather fidgety dynamic shading. Both conductors agree to play just the first repeat. The minuet is an unusual case of a Dorati minuet that seems too fast. Whatever Haydn meant by Allegro molto, Märzendorfer’s slightly slower tempo allows us to hear better the alternation between open and stopped strings. More seriously, Dorati’s trio sounds perfunctory at this tempo and he introduces more fidgety dynamics to compensate. Under Märzendorfer, its melancholy, folksy poetry emerges naturally. Märzendorfer’s finale has plenty of energy, though I would call his tempo a lively Allegro rather than a Presto assai.[4] Dorati is only a little faster but has more dash. The question also returns of how you wish your oboes to sound. Both conductors make both repeats. I could prefer Dorati in this movement, and those who think the second repeat should not be played might prefer him in the first movement too. Märzendorfer’s superior treatment of the middle movements would be the deciding factor for me.

Meanwhile, this symphony had not gone unnoticed by the BBC. Harry Blech and the London Mozart Players broadcast it on 1^st February 1952, Stanley Pope, with a section of the London SO led by Hugh Maguire, programmed it with nos. 45 and 90 on 9 December 1956 and Rae Jenkins gave it with the BBC Welsh Orchestra, as part of an extensive Haydn series involving several conductors, on 29 November 1962. The Pope was included in the Cameo Classics set of performances from the Itter Collection. It would be interesting to hear from anybody with a tape of one of the others.

Stanley Pope (1916-1995) apparently disliked making recordings and left only two – the original version of Schumann’s Fourth Symphony and four of Elgar’s Pomp and Circumstances (omitting no. 3), both on Philips. As well as this Haydn, several of his Havergal Brian premières have been issued from broadcast performances. He was noted for his Bruckner and broadcast Symphonies 5 and 7 as part of a BBC cycle divided between different conductors in the 1960s.

I do not know how many strings constituted a “section” of the LSO but they sound quite numerous. The oboes are not a notable presence except on the rare occasions when they have something on their own, but the horns shine through boldly. As was the manner in the 1950s, the strings play with full bows and do not peck at staccatos, so the effect is quite full-blooded. Pope has evidently prepared his score carefully, with a number of added crescendos and diminuendos. But it cannot only be for this reason that the music comes across as a major symphonic statement in a way it does not in the two very good performances just discussed. The first movement simply leaps into life. Every phrase knows where it is going and leads inexorably to the climax of each section. First repeat only. The second movement is expressed warmly and songfully, the curious dotted phrases offering a droll contrast. Neither repeat, though. The minuet is upfront. I have stated my preference for a slower tempo here, and this is the one movement where I might prefer Märzendorfer. However, one reason for alternating open and stopped strings is that you can make more noise that way, and Pope’s violins certainly do that. He manages to avoid the trio sounding humdrum at the faster tempo, as it did under Dorati. The finale, with both repeats, surges vigorously to its conclusion. The sound is fairly good for the date. 

A year later, in 1957, the symphony was broadcast by Hans von Benda (1888-1972), conducting the Berliner Kammerorchester which he had founded in 1932. His first movement is energetic and purposeful and has the first repeat. As regards dynamics, he plays by the book. That is to say, minimal shading where Haydn has not indicated anything. In truth, the hammering forte passages get a little overbearing without the crescendos and diminuendos that Märzendorfer, Dorati and Pope, in their different ways, all saw fit to add. Still, that is how Haydn wrote it. I am not sure the same can be said of the second movement. The strings do not sound muted to my ears and they sound anything but pianissimo. Indeed, they dig in as if it were a Mahler symphony. I checked the old Bärenreiter to see if it had different markings from the Robbins Landon Urtext but no, as far as this is concerned they are identical. This became oppressive and I was glad Benda did not play the repeats. The minuet is a sturdy affair at a steady tempo that I appreciated. Appreciation turned to dismay when he took a much slower tempo for the trio. The finale, with both repeats, is adequately lively if hardly dashing. This performance used to be available at René Gagnaux’s site. While it is an improvement on the Sternberg, also on that site, it is interesting only as a piece of German radio history.

Of the commercial recordings, the Märzendorfer seems the best up to this time but, good as it is, I think I shall return more often to Pope.

Christopher Howell © 2026

[1] I have not been able to hear this broadcast so I am only guessing at the parallels Simpson might have drawn. The Beethoven-RVW connection is fairly obvious since Vaughan Williams himself described the crescendo linking his scherzo and finale as a “crib” from Beethoven.

[2] Unless, of course, Newstone managed a more Beethovenian performance than any I have heard.

[3] Edward Sackville-West, Desmond Shawe-Taylor, with Andrew Porter and William Mann: The Record Guide, Revised Edition, Collins 1955, p.349.

[4] Readers are reminded that the Italian assai is not equivalent to the French assez – Presto assai means very fast indeed.