The world has not been short of Mendelssohn symphony cycles in recent years. The 2010s was a particularly felicitous decade for Felix. Within a brief period complete recorded cycles of the five numbered symphonies emerged from Jan Willem de Vriend and the Netherlands Symphony Orchestra (Challenge Classics), Andrés Orozco-Estrada and the Tonkünstler-Orchestra Niederösterreich (Preiser/Oehms Classics), Edward Gardner and the CBSO (Chandos), John Eliot Gardiner and the LSO (LSO Live), Antonello Manacorda and Kammerakademie Potsdam (Sony), Andrew Manze and the NDR Orchestra (Pentatone), Yannick Nézet-Séguin and the Chamber Orchestra of Europe (DG), and Pablo Heras Casado with the Freiburger Barockorchester or Bavarian RSO (Harmonia Mundi).Footnote ¹ These came hot on the heels of an even more ‘complete’ collection (as explained below) of Mendelssohn symphonies from Thomas Fey and the Heidelberger Sinfoniker (Hänssler Classic), recorded in 2006–9 and released together as a set in 2017, plus a cycle from Andrew Litton and the Bergen Symphony Orchestra made around the 2009 Mendelssohn anniversary (BIS). The 2022 release of Symphonies Nos. 1 and 3 from the Swedish Chamber Orchestra and Thomas Dausgaard (again on BIS) points to a likely ongoing cycle. And only in the last year the Orchestra of the Age of Enlightenment put on a complete concert series of Mendelssohn’s ‘mature’ symphonies and concertos under the hands of András Schiff (QEH London, April 2024).

It wasn’t always like this. For long, only a handful of sets offered listeners the opportunity to hear Mendelssohn’s ‘mature’ symphonic corpus: most significant were perhaps Wolfgang Sawallisch’s pioneering mid-1960s cycle with the New Philharmonia (Philips), and Claudio Abbado’s early ‘80s recording with the LSO (DG), which rapidly became seen as a benchmark. On the continent, Christoph von Dohnányi’s late ‘70s Decca cycle with the Vienna Philharmonic never quite received the critical acclaim to rival Abbado’s, but it is a fine set nevertheless and contains an especially strong Third Symphony. Then there is the inevitable Herbert von Karajan recording with the Berlin Philharmonic (DG, recorded 1971–72), while Kurt Masur committed the five symphonies to disc with Mendelssohn’s own Leipzig Gewandhaus Orchestra in the late ‘80s. The following decade saw a moderate increase in volume, Klaus Peter Flor setting down these works with the Bamberg Symphony Orchestra (RCA), Vladimir Ashkenazy with the Deutsches Symphonie-Orchester Berlin (Decca), and accounts from both Francesco d’Avalos and Walter Weller with the Philharmonia (ASV and Chandos respectively). Some presence, to be sure, but nothing like the flurry of recent activity.

Against this backdrop, the release in 2024 on Alpha of a set from Paavo Järvi and the Zurich Tonhalle Orchestra marks the first ‘Covid-Cycle’ – all but one of the works being recorded in partial lockdown during those uncertain spring days of 2021 (the exception being the choral Lobgesang or Song of Praise, captured in 2023 after the lifting of restrictions). It would, no doubt, be too hasty to claim a brave new world of Mendelssohn performance to have been inaugurated through such circumstances, but nonetheless this occasion does give pause for reflection on the direction of travel in recent decades, as well as offering a more immediate opportunity for assessing the qualities of this latest set. For as I will return to at the end of this article, there is much to admire about these performances. Bright, clean, carefully considered with plenty of attention to detail, yet lively and with its own wiry elegance, the current offering has numerous qualities that can be best understood against the backdrop of evolving approaches to Mendelssohn performance.

*

Yet isn’t the very idea of a Mendelssohn symphony cycle still a little peculiar – at least when taken as analogous to the symphonic outputs of most other comparable nineteenth-century composers? For a start, how many works are we talking of? I used the term ‘mature’ above in scare quotes, as there exist 12 earlier symphonies or sinfonias, written between the ages of 12 and 14, that are scored for strings, plus another opening movement written shortly after the ‘First Symphony’ in late 1824 (Fey actually includes these 13 string symphonies as part of his larger survey, hence its greater ‘completeness’ as it were).Footnote ² If completeness were therefore the primary criterion, there would need to be 18 symphonies in order to create a Mendelssohn cycle. Or 18 and a little bit – as the opening of a final, incomplete symphony in C major for full orchestra, dating from the mid-1840s, exists in manuscript.Footnote ³ Even narrowing down the criteria by designating the ones of interest as being for ‘full orchestra’ – thus apparently distinguishing between the earlier works for strings alone – runs into difficulties, as the eighth string symphony also exists in a form with added wind and timpani parts, thereby making it a full symphony.Footnote ⁴ (As far as I know, though, no one has ever pulled this one out from its neighbours and included it with the later five symphonies.) ‘Published’ would be equally unhelpful, as all 13 string symphonies have long been available in print. ‘Published by the composer’ proves even more problematic, as this would disqualify not only the string symphonies, but also the ‘Reformation’ and ever-popular ‘Italian’ Symphonies, both of which were shunned by Mendelssohn as not fit for publication and released by his enterprising publishers Breitkopf und Härtel only after his death.Footnote ⁵ This fact explains the confusing numbering, which reflects the order of publication but manifestly not the order of these works’ composition. And coming full circle by returning to our first suggestion, ‘mature’ is also questionable, even without the distrusted organicist connotations, as the First Symphony in C minor, op. 11, dates from a year or so before the time when Mendelssohn is typically assumed to have attained his ‘first maturity’ – in 1825 at the age of 16 with works like the Octet (although the two pieces once shared the same scherzo!).Footnote ⁶ Probably the easiest, and most neutral, way of getting at the five works conventionally included in a Mendelssohn symphonic cycle is to describe them as the ‘numbered’ symphonies – with the proviso that ‘for full orchestra’ can be added if needed.

The musings above may seem to be little more than scholarly quibbling about definitions and membership – ‘academic’ questions, amusing if one has a taste for paradox but of little real import. Herein, though, lies one of the main stumbling points in Mendelssohn’s position as one of the pre-eminent symphonists. However we divvy up the works termed ‘symphony’ – even just the five numbered works – they do not easily form a unified set. While the so-called ‘Italian’ and ‘Scottish’ Symphonies remain among the most popular and successful of nineteenth-century orchestral works, a Mendelssohn symphonic ‘cycle’ is undeniably a slightly odder affair than a Brahms or even a Schumann cycle. Think of it: five works, one of which, while supremely well written, is still showing a teenager forming his own voice; two of which were considered by their own composer as not good enough to publish (one of these he even claimed he wished to burn!); two of which have more or less questionable associations with landscape painting and ‘exotic’ locales that have long distracted from their remarkable musical qualities;Footnote ⁷ and one of which is a curious fusion of symphony with cantata written for a civic function, beloved by Victorian choral societies, reviled by New German critics, whose short-lived similarities with Beethoven’s Ninth have never worked in its favour. Even if one loves the music, it doesn’t quite add up to a consistent ‘cycle’ in the way many comparable symphonic oeuvres.

Yes, Beethoven wrote a choral symphony, but he was Beethoven, and in the eyes of critics can do what he wants. The Lobgesang still seems an outlier. An otherwise complete cycle by Bernard Haitink and the LPO delegates the second symphony to Riccardo Chailly – owing more to scheduling than aesthetic grounds I’m led to believe, but still revealing of the distance that separates this work from the others.Footnote ⁸ And while Heras Casado has given us a complete set of the five, his usual partners, the Freiburger Barockorchester, are replaced by the Bavarian RSO and Chorus for the second. Yes, a Schubert symphony cycle is also a little noteworthy: but at least with the earlier Viennese we have a split (clear, at least as far as reception goes) into six charming but still ‘early’ symphonies and two mature masterpieces. Mendelssohn has this split (the first symphony belongs to the early but charming camp); but he also has a choral work (like Beethoven, but more obviously a cantata and an occasional piece); a ‘problematic’ programmatic work (the ‘Reformation’); and then two popular but apparently poetically inclined symphonies.

If Mendelssohn’s five numbered symphonies are a ‘cycle’, the centrifugal force of generic disintegration is always in danger of overwhelming the centripetal tendency towards a single conceptual centre. None of this is a negative reflection on the qualities of the music in question; indeed, one could view Mendelssohn’s imaginatively diverse range of approaches to the genre as a distinct virtue. The A major and A minor works (the ‘Italian’ and ‘Scottish’ in popular parlance) are among the very finest of all nineteenth-century symphonies. Despite what the composer viewed as its possible liabilities, the ‘Reformation’ is a central document of musical programmaticism and a locus classicus of nineteenth-century musical historicism. Indeed, in musicological terms, the two historic ‘problem pieces’ (the ‘Reformation’ and Lobgesang) are perhaps the most intriguing and rewarding for scholarly and critical discussion. Even the early C minor symphony is astonishing for a 15-year-old and can pack quite a punch. The very need for a ‘cycle’ – for completeness, for a monumentality that seeks to ensure canonic stability in the ephemeral medium of music – is moot and might be productively questioned more often. (Interestingly, Mendelssohn himself, while one of the first to programme near-complete cycles of Beethoven symphonies over a concert season, never seems to have put on the First. Correspondence with friends suggests that he didn’t generally find Beethoven’s early works that noteworthy, much preferring the later music. Musical quality trumped any abstract desire for completeness in his mind.)

Might it be refreshing to move on from an ingrained but perhaps outdated aspiration to canonic monumentality? While record companies might still see the neatly packaged totality as an ineliminable selling point, do we really need complete sets of a composer’s symphonic output in this day and age? For instance, rather than placing the choral Lobgesang alongside the other four instrumental symphonies (where it inevitably sticks out), would it not be both more fitting and more provocative to place it beside the other symphonic cantata Mendelssohn completed around the same time, Die erste Walpurgisnacht, its wilder, heathen counterpart?Footnote ⁹ After all, the Lobgesang was never directly termed purely a symphony or even as ‘Symphony No. 2’ by Mendelssohn (the latter was merely the inference from the A minor Symphony being published subsequently as ‘No. 3’). Or could the early C minor Symphony be coupled alongside one of the later string symphonies (No. 11 in F minor, for instance, perhaps the most impressive of the set, or the full orchestral version of No. 8), which directly preceded it and chronologically as well as aesthetically are much closer to it than the later works?

*

But, moving on from such foundational reflections, how have these works fared in performance? For the sake of comparison, sharpening distinctions through the introduction of somewhat artificial heuristic categories, one might identify two broad approaches to performing Mendelssohn’s orchestral music. First, what we might call a mainstream later twentieth-century performance practice for Romantic music, using the full resources of the modern symphony orchestra, introducing light and shade and varying the tempo without ever pushing too hard, essentially treating Mendelssohn like any other nineteenth-century composer. (Abbado’s set, for instance, is a good example of this approach working successfully.) Second, one that adopts almost unfailingly brisk tempi, keeping the music running and emphasizing the dynamic, almost visceral quality of the writing (Fey is perhaps the most pointed exemplar of this). This bifurcation may have been sharpened by the historically informed performance movement and the incorporation of aspects of the associated performing styles into a new mainstream in recent decades, but it can be traced much further back. Obviously, Mendelssohn’s music is not the only, or even the first, repertory for which the second style is applicable: HIPsters have been doing something comparable with Beethoven’s symphonies for around half a century. But this manner does have a pedigree of being popularly supposed to date back to Mendelssohn himself – although the evidence for this may have been mildly overstated or warped by rival accounts long after his death, most signally Richard Wagner in his 1869 tract On Conducting, who set up a ‘Mendelssohn school’ as the foil to his own, evidently superior interpretive manner.Footnote ¹⁰

Mendelssohn, one of the leading conductors of his day, was known even during his lifetime for fast tempi: his landmark 1836 performance of Beethoven’s Ninth Symphony at the Lower Rhine Festival in Dusseldorf, for instance, while a great success with the public, led to prolonged dispute between admiring and hostile critics owing to the brisk speeds he had adopted, and Schumann (generally a staunch admirer, especially in later years) later remarked upon what he considered Mendelssohn’s unduly fast tempo for the first movement.Footnote ¹¹ Certainly by the twentieth century a division between a supposedly ‘Mendelssohnian’ conducting tradition (lithe and incisive, marked by quick tempi for fast movements and a reluctance to depart significantly from the music as noted) and a later, rival ‘Wagnerian’ one (weightier, permitting much greater tempo fluctuation and freedom with the score) had been set up.Footnote ¹² Such is the simplicity of the picture that most subsequent conductors could be placed on one side of this hypothetical divide (thus, for the critic John Ardoin, the Mendelssohnian tradition was continued in the twentieth century by Arturo Toscanini, the Wagnerian by Wilhelm Furtwängler).Footnote ¹³ One might well be sceptical about how useful such schematic divisions, with their proliferating pairs of binary oppositions (‘classical’ versus ‘romantic’, ‘objective’ versus ‘subjective’ …) actually are either as an accurate representation of Mendelssohn’s practice or for categorizing performances, but it nevertheless is clear that lively tempi were important for Mendelssohn.

Although the two approaches outlined earlier may sound antithetical, and are perhaps rarely encountered together, they are not necessarily entirely mutually exclusive. To my mind, one of the keys to a really persuasive performance of Mendelssohn’s instrumental music, at least in his faster outer movements, is managing to maintain a singing, expressive melodic line at a brisk tempo, allowing sufficient expressive nuance (such as subtle fluctuation where needed) without letting the tempo sag to the extent of losing the underlying momentum. Scherzos, perhaps to an even greater degree, often rely on speed, on the ability to articulate sounds at almost impossible rapidity, but again need to be kept clear-textured, and when tonal and timbral sensitivity is lost can become garbled. Still, to place the dynamism in greatest relief, moments of stillness and calm are necessary, as well as to do full justice to the beauty and poetic expressivity of the slow movements.

There are places where the balance is hard to get: the closing minutes of the Third Symphony, for instance, contains a haunting A minor coda with clarinet winding its way across a frozen horizon of sustained strings, where the music appears to disappear into the mists and off the edge of the world. (Passages like this reveal how hard it is to forget the unauthorized ‘Scottish’ nickname the symphony acquired decades after Mendelssohn’s death.)Footnote ¹⁴ As a conductor it almost impossible not to dwell on this hypnotic passage, slowing down the tempo to a virtual standstill; yet too great a relaxation plays havoc with the ensuing A major coda, whose Allegro maestoso, if all maestoso and too little allegro, can sound portentous rather than exhilarating, and abrupt rather than inevitable in succession. The preceding music has to sound slow, almost static, and yet keep that hidden underlying pulse, with as little let up as possible. Normally some compromise is taken: best, perhaps, is to avoid milking the A minor mists for every drop of expressive melancholy, and allowing the final triumphant moments to emerge convincingly as part of an overriding sweep, but almost invariably a choice has to be made between apparent perfection of detail and overall coherence.

The challenge is how to combine the beauty, charm, timbral sensitivity and textural refinement of Mendelssohn’s orchestral writing (which may emerge more easily with slightly more relaxed speeds and a willingness to pull the tempo about) with the vitality and dynamic elan that may appear more readily when adopting brisk tempi. How to square the circle? An instructive case is given by the great Greek conductor Dimitri Mitropoulos, who left studio recordings of the Third and Fifth Symphonies with the Philharmonic Symphony Orchestra of New York (i.e. the NYPO) in the early 1950s, and (amongst several others) a live account of the Third with the Berlin Philharmonic given at the 1960 Salzburg Festival, a few months before his death.Footnote ¹⁵ In these performances Mitropoulos manages almost miraculously to combine lyricism, sensitivity to phrasing and repose in the more inward moments with a dynamic drive that is all-but overwhelming in places, by turns ferocious or exultant. Indeed, such is the lick that he takes that in concert the Berlin Philharmonic (who despite their reputation were probably not yet back among the world’s best at the time) are barely able to keep up in the scherzo. A review of the Salzburg concert could scarcely contain its surprise at the revelatory result: ‘there was an ovation after the end of the first part of the concert, such as is rarely encountered at the very end of one, and achieved not by Brahms 1 or the “Eroica” but – listen and marvel – by Mendelssohn’.Footnote ¹⁶ (A Mendelssohn symphony? Who would have thought of it!) For comparison, one could turn to the Swiss conductor Peter Maag’s celebrated studio interpretation of the Third from the same period with the LSO (Decca, 1957). Maag’s account is loving, generously phrased, singing, eminently musical and arguably equally fine; but lacking just that bit of bite, that sweeping energy in comparison with Mitropoulos.

Whether one prefers one way or the other is obviously largely a matter of personal taste. There is so much to admire in each; but I would propose that there is something quite distinctive to Mendelssohn, his own exhilarating brand of musical dynamism, that is easily and too often left out when the propulsive aspects of his music are downplayed (and when the performance is otherwise not as sensitive as Maag’s). This is largely due to his remarkable and unique feeling for movement and momentum, which in technical terms can be put down to his command of phrase rhythm, his control of cadences and large-scale cadential deferment, his transparent yet ever colourful textures and refined deployment of the orchestra. Unlike numerous other nineteenth-century symphonies, his rarely get bogged down by the insistent over-repetition of set rhythmic figures, confinement within four-square phrases, or fussy displays of contrapuntal and motivic working-out. The thematic process is so smooth and seemingly natural, spontaneous, his melodies so captivating, as to hide the skill with which they are handled. His music flies, fleet-footed, buoyantly, surging towards its (usually joyful) conclusion.

*

The examples from Mitropoulos and Maag are among the finest of earlier renditions of an individual symphony (neither left a complete Mendelssohn cycle, though Maag would go on to re-record No. 3 along with No. 4 in Bern in the 1980s). But many subsequent accounts tend towards one or other side in the hypothetical division sketched above. Increasingly common in the last two decades is the fast-paced, dynamic Mendelssohn, which (depending on one’s perspective) might occasionally topple into hard-nosed over-assertion. Unsurprisingly, this is normally most apparent those renditions with a clear link with the historically informed performance movement, including the use of smaller forces such as chamber orchestras. Thomas Fey with his Heidelberg orchestra (a ‘hybrid-HIP’ band, mostly on modern instruments but with period trumpets and timpani, in the manner of Nikolaus Harnoncourt’s Chamber Orchestra of Europe performances), is probably the most extreme in this direction, and while his readings have found plenty of admirers, some reviewers have complained of an excessively driven quality. Pablo Heras Casado’s set, primarily with the Freiburg Baroque Orchestra, is likewise fast-paced and tending to the austere in orchestral sonority, and while Jan Willem de Vriend’s Netherlands Symphony Orchestra (HET Symfonieorkest) is not ostensibly a ‘period’ group, it is very much influenced by this style with lean textures and rapid tempi. A similar reaction – divided between thrills and shivers – has also accompanied the first release in Thomas Dausgaard’s nascent cycle with the Swedish Chamber Orchestra.

Then we have accounts that seem to merge elements of each style, often in intriguingly diverse ways. Antonello Manacorda’s Sony set with the Potsdam Chamber Academy, for instance, maintains the punchiness of a smaller chamber ensemble (violins antiphonally divided, vibrato rationed when needed, with rasping brass and timpani) alongside elements redolent of more traditional ‘Romantic’ practice. Notwithstanding the brisk tempi adopted for faster movements, the slower music is never hurried, and despite occasional roughness around the edges, what is persuasive here is the dynamic sense of line and phrasing within a purposefully moving tempo – that all-important ‘singing allegro’ identified above. (Manacorda himself was a violinist and former concertmaster of the Gustav Mahler Youth Orchestra in the days of Claudio Abbado, so he has a link with a notable Mendelssohn forebear.) This set contains a particularly impressive account of the ‘Reformation’, which in Manacorda’s intense reading takes on an imposing symphonic quality.

Indeed, one of the interesting features of both Manacorda’s and Yannick Nézet-Séguin’s cycles is their preference for Christopher Hogwood’s 2009 Bärenreiter edition of the ‘Reformation’ Symphony, which offers additional passages that Mendelssohn deleted at various stages of composition, rather than the familiar 1832 version of the work, published in 1868 some two decades after the composer’s death.Footnote ¹⁷ (All cycles prefer the original 1833 form of the ‘Italian’ Symphony to the 1834 revision – not only, one assumes, owing to the incompleteness of the latter version (planned changes to the opening movement were never made), but because no one has ever preferred the revision – apart from the composer.)Footnote ¹⁸ Hogwood’s edition, nominally an Urtext, is in fact a highly problematic amalgam of variants that doesn’t represent any single form of the symphony as conceived by the composer: it transmits draft material that Mendelssohn deleted during the initial composition in 1830, but also prefers passages from a later 1832 revision over the finished 1830 version (which, apparently unknown to Hogwood, still exists in a separate full-score copy).Footnote ¹⁹ Apart from a few moments in the opening movement, the main difference listeners will note in these performances is in the considerably extended transition between slow movement and finale, whose instrumental recitative gestures and cyclic recall of earlier movements connect back to Mendelssohn’s instrumental works of the late 1820s, as well as pointing up the links to a recent D minor symphony by Beethoven with a choral – as opposed to chorale – finale.Footnote ²⁰ Judging by these two excellent performances, I might even prefer the symphony with this linking passage, even if the philological grounds for this particular version are dubious. What would be even more intriguing is a recording of the completed 1830 version, which has significant differences in the finale and a quite distinct ending.Footnote ²¹

As for Nézet-Séguin, he, in his very different way, achieves another seemingly unlikely fusion between opposed stylistic tendencies and means. Given the chamber orchestra forces, one might expect a leaner, cooler, brisker affair; but there is nothing hard-driven, austere or cold about these accounts. One of the glories of the set is the playing of the Chamber Orchestra of Europe, whose woodwinds, in particular, create such glowing, vibrant colours. Tempi are by no means sluggish but are never forced, details emerge without being underlined (occasionally indeed I miss a touch more shaping of phrases), and to that extent this interpretation seems much more in the ‘traditionalist’ camp – with the proviso that the use of chamber-orchestra forces ensures there is inevitably a lightness and transparency to the texture that often eludes older, more stately readings. With the exception of the cycle’s opening First Symphony (the first movement disappointingly literal), there are few if any recent accounts conveying as much charm, colour, and warmth in Mendelssohn’s writing.

Then we have Edward Gardner on Chandos with a full symphony orchestra (the CBSO) and Chandos’s typically rich sound, but adopting brisk, incisive tempi – a kind of chiasmic reversal of Nézet-Séguin, but again very successful in its own way. While using the familiar revised version of the ‘Reformation’, Gardner’s unique selling point is in offering a Song of Praise – an English-language Second Symphony in place of the German Lobgesang. Both forms are perfectly ‘authentic’ (in the sense of authorized by the composer): not only was the piece published simultaneously in both languages, but in fact the second ever performance was given in English under Mendelssohn’s baton in Birmingham Town Hall (the location used for the recording), three months after the Leipzig premiere. This piece has almost invariably been recorded in German, even by British forces, but as with Haydn’s oratorios and Mendelssohn’s own later Elijah (written indeed for Birmingham) there’s no good reason for singing the work in German in the Anglophone world. As so often Mendelssohn was not quite satisfied with the first version and made changes to the score after the 1840 Birmingham performance.Footnote ²² His most important addition – the dramatic ‘Watchman, will the night soon pass?’ episode, one of the real highlights of the piece – is especially effective in this recording.

John Eliot Gardiner, with a slightly slimmed down LSO, is more obviously aiming at a stylistic fusion of a modern symphony orchestra and the historically informed style of which he has so long been a leading practitioner. And once again, the results are generally very persuasive, with the full tonal resources of a top orchestra moulded by the judicious use of passages of reduced string vibrato and Gardiner’s characteristic shaping of phrases. (Here it is an unconvincing Third that lets down the otherwise fine performances in the cycle; like Järvi’s, Gardiner’s set also includes a selection from the incidental music to A Midsummer Night’s Dream.) Andrew Manze, highly respected in his own right as a solo violinist and again closely associated with historically informed performance practice, actually is rather less overtly HIP-inclined as a conductor than many others in this repertory. His accounts with the Hamburg NDR Orchestra and chorus are more ‘mainstream’ than many recent interpretations, but very satisfying: his Lobgesang, for example, sometimes a tricky work to pull off, sounds entirely ‘right’, the result of unfussy and sensitive music-making.

*

How does Järvi’s new account stand up in relation to the recent competition, then? On first listening, I found these performances a little hard to place, even a touch undistinctive; but returning to them after several months I’ve warmed considerably to their qualities. It may be easier initially to say what they’re not – as they avoid extremes in any of the hypothetical performance categories I set up above. The Zurich Tonhalle is obviously a modern symphony orchestra, and while vibrato is rationed in places (such as the start of the ‘Reformation’, understandably given its neo-renaissance counterpoint) and a few timpani thwacks suggest a nod to earlier performance styles (achieved perhaps through the use of harder sticks) the overall effect is of a fairly traditional ‘modern’ orchestra. Nevertheless, there is a control of ensemble and transparency to the sound that avoids any suggestion of out-dated stodginess. Tempi are by no means sluggish (the finale of the ‘Italian’ is certainly brisk, for instance), but one easily finds more driven, unremitting, hard-hitting Mendelssohn performances (Fey, Heras Casado, Daumsgaard, Manacorda). Still, this is not in the grand old style either – no heavy, portentous Mendelssohn here. Witness the opening of the ‘Song of Praise’, where the trombones are jovial and animated rather than solemn and stately, without quite becoming a bouncy tripping hither, tripping thither in the manner of Gardiner. (The one possible exception is the end of the Third Symphony, the point discussed earlier – too ponderous, too little allegro, and timbrally undistinguished – a rare misjudgement.)

My first impression was of a relative indifference to phrasing on the part of Järvi: not that it is shunned at the smaller level of the theme, but that there was a less pronounced sense of line, of dynamic ebb and flow, over the longer expanse of a section or movement. Contrast this with, first, the consistent low-level intervention of Gardiner in moulding melodic phrases and springing rhythms, which are rarely under sprung in his hands, or the strong sense of melodic line, of rising and falling intensity Manacorda achieves despite the sometimes-astringent sonority of his Potsdam Chamber Academy. On returning to Järvi’s recordings, though, I’m not sure this criticism is entirely fair. Without exaggerating this aspect, there is a consistent shaping of phrases and bringing out of salient details in accompanimental voices suggestive of a great deal of care. Being hypercritical, one might say the performances are always good, well considered and thought through, but rarely take wing on their own or risk abandoning themselves to the flow of the music, with the attendant possibility of loss of control. Listen, for instance, to the final climax of the opening movement of the ‘Reformation’: the notes are all there, but there is no sense that the performance has gone further to reach that (illusion of?) musical flight. This could, of course, be a result of the studio recording (and a studio recording during the pandemic to boot), although my experience of Järvi in concert doesn’t entirely contradict this sense of well-intentioned control.

What seems to be most characteristic, idiosyncratic even, is the concern with texture and articulating blocks or passages of sound. Especially in fugal or contrapuntal passages, Järvi has a neat way of clarifying textures and bringing out the almost tactile ‘feeling’ and patterning of the music at these points. This can be heard particularly clearly in the fugatos in the development section of the First Symphony’s finale and first movement of the ‘Reformation’, but there are numerous contrapuntal passages throughout these works that benefit from such sensitivity. An additional factor that aids this textural clarity is the antiphonally divided first and second violins: hearing the wonderful interplay across the podium in passages like the coda to the first movement or internal retransition of the minuet in the First Symphony makes one wonder why so many conductors, even those offering an apparent rapprochement with aspects of historically informed style, still do not reinstate what was common practice until the latter half of the twentieth century.

For their part, the Tonhalle Orchestra plays precisely and with poise. While there may be more voluptuous depths of sound on offer elsewhere, there is a pleasing sweetness to the violins, and an airiness and textural transparency that comes from the separation between firsts and seconds. Lower strings may seem throaty or sinewy rather than rich, though the cello countermelody in the recapitulation of the Third Symphony is very atmospherically done, softly and slowly, with distinctive melancholy. Woodwind and brass are light and flexible (the clarinets are brighter than the sound from a German orchestra, for instance); without conveying enormous individual character as individual voices, they contribute to the well-controlled ensemble. Neither overly romantic nor astringently modern, neither soft nor brutal, the performances are clean, focussed, well-considered.

As for highlights among the cycle, there isn’t a single symphony that is either a standout or a dud, but instead a pretty consistent level across all five. The First Symphony is amongst the finest accounts I’ve heard and holds its own well in comparison with any other recorded cycle. I can find no fault with the ‘Italian’, which is lithe, supple, a little on the lighter side, but very successful. The ‘Reformation’ is fine but perhaps just a little too safe, too disciplined, and to my mind doesn’t supplant the recent accounts from Nézet-Séguin and Manacorda, dodgy edition or not. The Third is generally good, apart from the slightly ponderous ending: expressive changes of mood and atmosphere are taken smoothly and there’s plenty of zip in the finale. The Lobgesang, finally, while not monumental or the most dramatic, is lively and expressive. Of the singers, the tenor possesses neither the most powerful nor mellifluous voice, but the sopranos and the chorus acquit themselves admirably, the choral ‘Nun danket alle Gott’ being very beautifully taken.

Or almost finally, as there’s the additional fourth disc with excerpts from the Midsummer Night’s Dream music. Given that this is not the complete incidental music (omitting the melodramas), and the disc runs to only 42 minutes, it might have been worthwhile including more material, at least the other three major concert overtures (Calm Sea and Prosperous Voyage, The Hebrides, The Fair Melusina, which would also have fitted as fillers on each of the previous symphony discs). Once again, the performances are fine, but here miss that touch of magic. The opening Scherzo shows the poise and control of the woodwind at its best, though the Nocturne is a little prosaic.

Those wanting a modern recording with a traditional symphony orchestra that is nevertheless firm and incisive might be very satisfied with Edward Gardner. Those preferring a slightly smaller and more agile body need look no further than Nézet-Séguin. For period/modern fusion Gardiner and Manacorda offer much of interest. For a fairly HIP-tending take, Fey is stimulating, if not perhaps for repeated listening. Looking further back, those happy with older recordings will still find enormous value in classic sets by Abbado, Dohnányi, and Sawallisch. Then of course there are accounts of individual symphonies. I have mentioned Mitropoulos and Maag with the Third; simply remaining with interpretations of that even earlier vintage one might continue to Cantelli or Klemperer with the ‘Italian’ (both Philharmonia), and Maazel’s ‘Reformation’ (Berlin) is worth hearing alongside Mitropoulos (New York). And yet there is room for Järvi amid these varied offerings. While it may not be an automatic first choice for all within the increasingly crowded Mendelssohn marketplace, it is a worthy addition the recent spate of recordings, sensitive and well thought through, striking a fine balance between precision and expression.