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. 3 in G major Hob. I/3 (1759/60)

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

Symphony no. 3 has fascinated me since my student days, when I discovered its finale which seems almost a blueprint for that of Mozart’s Jupiter Symphony – a combination of fugue with elements of sonata form, though much shorter and with no repeats, and a four-note subject of which the first two notes are identical to those of the Mozart symphony.

It would be easy to underestimate the forward-looking aspects of the first movement, precisely because it seems so “conventional”. We have seen so many examples of Haydn first movements that are “monothematic”, that have “development” sections that turn out to be something else, and “false recapitulations”. Here we have text-book sonata form, with a contrasting second subject, a development based on motives from the first subject and a by-the-book recapitulation. A movement that follows the rules, then – until we remember that, in 1759-60, these were rules of Haydn’s own evolving. The first part of the first subject is a four-note figure which, though it does not have the same pattern as that in the finale, may represent an early attempt to present a symphony as a unified statement.

The second movement, as in so many of the earlier symphonies, is scored for strings only. It is a tender, poignant Andante moderato in G minor. The minuet – this is a fairly early case of a four-movement symphony – is a canon, generally with the lower strings following the violins and oboes at a bar’s distance, though towards the end of the second part this is reversed. The trio shows Haydn’s burgeoning interest in instrumental colour, giving an independent role to the oboes and horns rather than use them to enrich the sonority.

Given the extraordinary finale, it is slightly surprising that this symphony seems to have gone unrecorded till Max Goberman set out to do the entire cycle in 1960-62. The orchestra, exceptionally for this series, was the New York Sinfonietta, though according to Gray’s Classical Discography[1] it was recorded in Vienna (HS 8M/8S). The playing is precise and nimble. Furthermore, the symphony entered the lists with a clean text, all repeats and a harpsichord continuo. This latter, however, is so distantly recorded as to suggest, conversely, that it was not really needed at all. The outer movements are taut and urgent but not overdriven while the minuet has a vivacious, if sometimes dogged, stride. It shows how times and attitudes have changed that such a well respected critic as Roger Fiske could write in Gramophone that “the conductor insists on all the repeats and thereby kills our interest in the very fragile slow movement”[2]. Here I wondered if the movement need be so fragile and whether it would not take a little more depth of expression than the gentle serenade Goberman offers. My misplaced romantic notions, no doubt, but I wonder what Harry Blech, for example, might have made of this.

Goberman was soon followed, in 1964, by Leslie Jones and the Little Orchestra of London (Pye GGC 4021/GSGC 14021). Jones adopts a lighter, more dancing manner in the first movement and his harpsichord is slightly more audible though still, I would say, not enough if you think it is really needed. This is sheer delight, leaving Goberman to sound a little over-earnest in comparison. First repeat only. His Andante moderato is notably slower than Goberman’s and confirms my feeling that there is more emotional depth to be extracted from this movement. I only wish he had allowed at least one repeat. Jones has the minuet dancing lightly, removing the dogged air of Goberman. The trio, while nicely despatched, sounds rather too brisk – could the movement take a more stately tempo without lapsing into heaviness? Honours are fairly even in the finale but, whether for the recording or for the playing itself, the counterpoint sounds clearer under Jones. In spite of the shortage of repeats, this version is more likely to win friends for the symphony than Goberman’s and it is the finest performance I have heard from Jones so far in this survey.

Before coming to the two complete cycles, I mention for completion’s sake a version in which the Bamberg SO is allegedly conducted by Alexander von Pitamic (PDU AC 60012), conjecturally dated 1971 by Mike Gray[3]. It is well known that von Pitamic, like Alberto Lizzio, was an invention of Alfred Scholz, applied to recordings made by Scholz himself or obtained, officially or not, from German radio stations. Given the paucity of recordings of this work, I should have been interested to hear it, no matter who the conductor was, but this was not possible[4].

Märzendorfer, like Jones, gives only the first repeat of the first movement and neither in the second. His first movement has a festive, exultant air, investing the quavers in the lower strings at the beginning with great energy. I like this very much, perhaps more even than Jones. About the second movement, I am less sure. Märzendorfer is a tad slower than Jones and certainly recognizes the poignant character of the music. But by making each pair of quavers an individual sigh he rather loses sight of the overall line. So perhaps I prefer Jones here. Märzendorfer’s minuet is bright and perky, which is all to the good. Theoretically, it is equally to the good that he realizes the trio needs more space, but maybe he relaxes the tempo too much – a tempo which would, I imagine, be untenable for the minuet itself. Then, as I have remarked on other occasions, the problems with playing trios at a slower tempo comes with the return to the minuet. Märzendorfer makes a pause before the reprise of the minuet but this is not enough to cushion us from the shock as the original tempo breezes in. His finale is splendid, festive and jubilant. He is particularly good at creating a cumulative effect, each new stage capping the previous one. Even so, and despite some fine things, perhaps Jones remains preferable of those heard so far.

However … Anyone who suspects I undertook these articles with the secondary agenda of Dorati-bashing is in for a surprise. Dorati plays the first movement with splendid vitality. He has smoother contours than the others and there is a graciousness that I have often missed in his Haydn. Both repeats are played, and so they are in the second movement. The stopwatch says he is slightly faster than Goberman but, with finely shaded phrasing, he realizes its poignant nature much better. The stopwatch also tells us his minuet has the same tempo as Goberman’s, but in place of doggedness there is a joyful lilt – not one of his flabby minuets at all. When the trio arrives, the tempo seems absolutely right. His finale is the fastest of all, but with transparent textures, every contrapuntal strand is clear while there is some fizzing articulation from the strings. At last I have found a symphony where Dorati comes out top.

No live or broadcast performance has come my way. In the BBC survey of all the Haydn symphonies in the 1960s, no. 3 fell to David Lloyd-Jones in 1965, which should be interesting if a tape were to emerge.   

Christopher Howell © 2026

[1] https://classical-discography.org/search.php

[2] Gramophone May 1969 p.1560.

[3] https://classical-discography.org/search2.php

[4] It is theoretically available at Spotify, but the links lead to Rosbaud’s performance of Symphony 12.