Vivaldi: The Four Seasons. Remy Principe (violin), Ferruccio Vignanelli (harpsichord), Gino Nucci (organ), Orchestra dell’Accademia di S. Cecilia, Bernardino Molinari 1942. Ermitage ERM 116 S (1991), no barcode then (sticker) 8014394101160

Ermitage ERC CD 12006-2 (1991) 8014394250066:

Reissue Ermitage France CD 12006-2 (1999): 3760020070069:

Aura Music AUR 129-2 (1999) 8014394521296 (also 0697833001290):

Second to none: the first of all Seasons
23 April 2025
I can’t insist enough on the historical significance of this recording: 1942, an orchestra from Rome called back then “Orchestra dell’Augusteo” (Augusteo is the Mausoleum of Emperor Augustus, and was used in the early 20th Century as a concert venue), but soon rechristened “Orchestra Stabile dell’Academia Nazionale di Santa Cecilia” (I’ve also seen a Cetra ad from 1943 where it was designated as “Orchestra Stabile della R.[eale] Accademia di S.[anta] Cecilia”, and I haven’t seen any other indication that the Academy was ever called “Royal”). The conductor: Bernardino Molinari (1880-1952), a friend and champion of the composers of his time, Debussy (he made an orchestration of L’Isle Joyeuse that you can hear on YouTube, link will open a new tab), Strauss, Malipiero, Alfano, Casella, Pizzetti, and Stravinsky. The recording was first issued on six 78s from the label Cetra, label numbers CC 2308/13 (12 sides) in Set 6, then CB 20228/33 (according to my research this is a reissue from late 1942) and BB 25067/72 (published at the latest in 1947; it is listed in the Cetra catalog of that year – much gratitude to Istituto Centrale per i Beni Sonori ed Audiovisivi for making it, and many other early record catalogs, available online, link will open new tab to their website -, but I haven’t been able to find the catalogs from 1943 to 1946, to check if it came out earlier). It was reissued in the US in 1948 on 78s Cetra Soria set 107 (individual discs BB 2043-8), and transferred to 33rpm LP the year later, Cetra-Soria 50.004…




…AND THE FIRST EVER RECORDING of The Four Seasons. Talk about being a standard bearer! (or is it a trail-blazer?) The solo violinist wasn’t credited in any of the Cetra releases. French Vivaldi specialist and discographer Roger-Claude Travers suggests Remy Principe, concertmaster of the Accademia Santa Cecilia until 1943, who is known to have played the concertos with conductor Molinari. Not only can I confirm with certainty that it was indeed Principe, but I can also provide the names of harpsichordist and organist. Thanks to the marvelous website worldradiohistory.com, I chanced on the June-July 1942 issue of the Italian radio magazine Radiocorriere, with an ad for the recently published Cetra set:

Welcome back from oblivion, harpsichordist Ferruccio Vignanelli (1903-1988) and organist Gino Nucci (the spelling “Nuci” appears to be a typo, Gino Nucci was organist then chorus master of the Accademia di Santa Cecilia until the 1960s).
Arcane but important fact for the discographers and Vivaldi specialists: Bernardino Molinari performs here his own edition, published in 1927 by Ricordi, and effusively dedicated to none other than… Benito Mussolini, “powerful and impressive builder of a new social harmony” (to quote the ironic review by Roland Gelatt in The Saturday Review of 28 August 1948, who hilariously concluded: “We can thank Maestro Molinari for bringing this music to our attention and for conducting it in the present Cetra album so agreeably. As for his inscription, we adjure him to pick future dedicatees with greater prescience and better judgment”). The Wikipedia entries (even the Italian versions) on The Four Seasons or Bernardino Molinari are curiously silent about this – rather embarrassing – dedication, although it is duly mentioned in the liner notes to the original CD reissue from 1991 on Ermitage ERM 116 S. And I’ve been able to verify that the dedication is indeed printed in the 1927 scores (one per concerto):

This is how it translates (my translation):
TO
BENITO MUSSOLINI
POWERFUL AND MARVELOUS ARTISAN
OF NEW AND FECUND SOCIAL HARMONIES,
ARE DEDICATED THESE TRANSCRIPTIONS
OF OUR ANCIENT MUSIC
PERPETUATING IN THESE TIMES
THE MELODIES OF ITALIANATE BEAUTY
There is a certain importance, in documenting the early recorded history of The Four Seasons, to these arcane matters of editions. While the now widely accepted and performed edition is the one edited by composer Gian Francesco Maliperio for the Istituto Italiano Antonio Vivaldi and publisher Ricordi, it wasn’t published until 1951. As late as 1955 with the New York Philharmonic (and already in 1951 with the NBC Symphony Orchestra), Guido Cantelli conducted Molinari’s edition. It is in fact fascinating to retrace the circulation of the Molinari edition in the US: in January of 1928, the conductor was invited to perform the concertos with the Saint-Louis Symphony Orchestra, giving them their American premiere (see this story here, link will open a new tab); he
then travelled to New York where he handed out score and parts to Toscanini, who conducted the Spring Concerto in February 1928. It was then Molinari himself who conducted Autumn and Winter with the Philharmonic in August of the same year; and then almost a decade later, in 1936, Toscanini reprised Winter (gratitude to the New York Philharmonic Archive for giving public access to that information). That goes a long way to explain why it was Molinari’s edition that Cantelli used for his 1951 and 1955 performances: probably because it was available in the orchestra’s library.
On the other hand, when Bernstein tackled the work with the same orchestra in the 1963 and 1964 seasons (reissued in 1993 on The Royal Edition vol. 98, Sony SMK 47 642), he used neither Molinari nor Malipiero, but yet another early edition, by Alceo Toni (who in 1920 had published the first modern edition of The Four Seasons – arranged for piano 4-hands…). Toni’s orchestral score had originally been published in 1924 by the Milan publisher Sonzogno (say the scholarly sources, but I haven’t seen trace of it in the online library catalogs), then reissued in 1942 by Carisch (another Milanese publisher) and reprinted in 1947. I haven’t actually seen any of these two editions, Toni nor Molinari, but playing by ear Toni is much closer to Malipiero’s edition – and hence, presumably, to the original parts published in 1725 in Amsterdam by publisher Le Cène. One finds online past articles by Roger-Claude Travers, in which he indicates that Cantelli performs Toni’s edition; but in a private correspondence he kindly recognized his mistake, having confused with Bernstein’s recording, as both share the same violinist, New York’s concertmaster John Corigliano Sr.
It is sometimes said and read (including on the Naxos double-disc reissuing it) that the first recording was Louis Kaufman’s with the Concert Hall Society String Orchestra under Henry Swoboda, made in December 1947, released at the end of 1948 on Concert Hall Society’s set AR (78s) and in the early months of 1949 on LP, CHC 1, Concert Hall’s first release on the new medium (reissued by Naxos, 8.110297-98, with the complete opus 8, link will open new tab to my review). But not so, and the precedence of Molinari’s recording is undisputable, even if it was Kaufman’s recording rather than Molinari’s that put The Four Seasons on the international music map (witness the French Grand Prix du Disque it was awarded in 1950). In his autobiography “A Fiddler’s Tale”, Kaufman is not specific about which edition he played (“from a Milanese publisher” – but both the Molinari and Toni editions were issued by Milanese publishers…), but it is audibly Toni’s. And it does make sense, as in 1947, Toni’s version had just been reprinted by Carisch. Some of the reviews from back then that I’ve read (in The Saturday Review) were rather more positive about Molinari’s recording than Kaufman’s, complaining about the imbalance of the latter’s forward placement, and his too romantic adagios (the reviewer in The Gramophone Shop Supplement was far less positive though).
But a recording made in wartime Italy, of a transcription that was not entirely “Urtext” – and dedicated to the wrong person to boot – probably didn’t stand much of a chance, especially in the face of all the versions that first trickled (Münchinger’s first out of three recordings for Decca in 1951), then flooded the market, and Molinari was soon forgotten.
One also frequently reads online that the premiere recording was made by British fiddler Alfredo Campoli with the Boyd Neel String Orchestra, a recording issued in 1995 on Pearl Gemm CD 9151. Not so. First, it is a broadcast, said to be from the French radio and to date from circa 1939, of a performance that was recorded on acetates and kept in Campoli’s private collection until its publication by Pearl in 1995. So even if the dating were true, it would make it the oldest documented performance of The Four Seasons, but pride of being the premiere recording, e.g. the first to have been made for publication and published, would definitely remain to Molinari. Moreover, the 1939 dating (and location in France) is dubious, if not impossible, for reasons that I develop in my review of the Campoli version (link will open a new tab). I’ve found documentation showing that Campoli and Boyd Neel performed the Seasons in England at the end of 1949 and again in the early months of 1951, and if confirmed, this would be more likely. By the way, whenever and wherever it was, Campoli performs Molinari’s edition.
Back to Molinari then. Even before listening, a number of questions arise. What will be an interpretation of The Four Seasons with no recording precedent to rely upon or refer to, no model, no previous recording, one in which the interpreters are alone before the score and its suggestions, one in which their interpretive choices are not, consciously or unconsciously, shaped by- and reproducing what they heard before, but determined only by their own autonomous reaction to the score? Will, as so many did after and until Harnoncourt’s recording (1977), Molinari play just “the music”, or will he play the theatre of Vivaldi’s music, graphically depicting in music the underlying program announced by Vivaldi’s explanatory sonnets, the birds’ happy greeting of spring, the flowing waters, the soft breezes and the storms erupting, the dog barking, the plaintive villager, the drunkard, the hunting party and its galloping horses, the gunshots, the cautious walking on ice, the cozy sitting by the fire with rain pouring outside, the ice breaking?
AND the sircocco in winter?
Or, as the first fiddler would ask, after having patiently waited for the conductor to finish his long-winded explanation of the whys and whats and context and meaning: would the “Allegros” be played as real Allegros?
Well, what immediately strikes as soon as the first measure of Spring, especially in view of the dominant interpretive tradition that came in force soon after, is the liveliness of the opening Allegro, beautifully illustrating Vivaldi’s program: the bird’s joyous greeting of spring. It’s not just a matter of tempo, but also of (staccato) articulation, where the score might have invited for more legato and less bounce; and the two accompanying violins also chirp with great character. There’s also the presence of the harpsichord – 1942, that’s three decades after Wanda Landowska had the Pleyel company build her monster of a harpsichord – and the leanness of the string ensemble – contrary to some comments that I’ve read, there’s not much that sounds like a “large” symphony orchestra here. The author of the Ermitage CD liner notes, esteemed Italian musicologist Piero Rattalino, talks about “a piano fitted with a device to sound like a harpsichord”, but I wonder where he gets that from. I don’t hear a modified piano, I hear a harpsichord. One also notices the use of the organ to underpin by pedal notes the figurations of the solo violin (very conspicuous at 1:38 in the first movement of Spring). The singing birds trio is taken in strict tempo, where performers since Harnoncourt play it as a freely-pulsed cadenza, but it’s played with fine character. And at that tempo, the gently flowing waters (1:04) have the agitation of a cascade, and the ensuing storm is suitably stormy. Likewise in the Finale, animated and robustly phrased, a real dance of nymphs and shepherds, not a stately and dignified courtly minuet as so many after turned it into.
Another great surprise: the opening of Summer is, truly, the “Allegro non molto” written by Vivaldi, not turned into a slow introduction, which even period-instrument ensemble, influenced by false interpretive traditions and exaggerating the “languor” described by Vivaldi’s explanatory sonnet, are still accustomed to doing. To be entirely honest, although the tempo in the ensuing “cuckoo” Allegro (0:45) isn’t exceptionable in itself, in relation to that opening tempo it is probably too slow: in fact, it is taken practically at the same tempo (and the sense of tempo unity that ensues was probably intended). It also sounds as if the score is doctored here: I think I hear not one but TWO violins playing, one playing the rapidly repeating ostinato 32nd notes, and the other the cuckoo melody itself. The organ also continuously plays a pedal-note. The summer storms erupt with great power, and if you find that the solo melody at 3:01 sounds whining – it should: it depicts the weeping of the shepherd, fearing the cruel storm and his fate. Again this passage is underpinned by organ continuo, and starting at 3:28 (measure 126), Molinari adds an upward counter-melody played by (I think) viola that’s not in the Malipiero (nor Toni) edition, but sounds just fine.
Another great effect, not explicitly written in the score but very welcome, is the way Molinari has the hunting party in the Finale of Autumn approach from afar, from pianissimo to forte, as in the “Bydlo” movement from Ravel’s orchestration of Mussorgsky’s Pictures at an Exhibition. With that, this is a cozy, unhurried hunting party, with horses trotting more than galloping, and the sonics don’t let the gunshots register with sufficient impact.
Compared to the invitations of Vivaldi’s theatrical program, the opening Allegro of Autumn, depicting the dance to celebrate the new harvest, is robustly phrased, but somewhat held-back in tempo – but nothing shockingly plodding either. The drunkard (1:13) doesn’t particularly sound like one, although, at the return of the solo violin’s figurations at 2:37, solo violin and ensemble play with a welcome flexibility of tempo. In truth, I’ve seldom heard the Autumn’s drunkard passage played as I think it could and should, everybody, even Harnoncourt, plays it much too straight to evoke a brawling, tottering and hiccupping drunkard. The only exceptions I’ve encountered are in arrangements, by Red Priest and The Palladian Ensemble (the latter playing Nicolas Chédeville’s 18th century arrangement with hurdy-gurdy). In the Largo of Spring the viola interjections are played so softly and gently, with such disregard for Vivaldi’s explicit instruction to play them “always molto forte and a strappato”, e.g. with a strong tearing motion from the bow, that they evoke nothing of the barking dog they should – maybe a gentle wind, or maybe the softly tolling drum of destiny. But at least, even at a slower tempo, the solo violin manages to avoid the Kreisler-like sentimentality of Louis Kaufman in 1947. In the Largo of Summer, the accompanying figure played by violins sounds like just that, an accompanying figure, not hovering horseflies (but the distant thunder is evocative).
And one thing that was “invented” later is playing the opening of Winter “sul ponticello” to enhance the freezing impression. But I appreciate Molinari’s rise from pianissimo to forte there, and the urgency of his tempo, which makes sense of Vivaldi’s reference to “the biting blast of the terrible wind” in the solo violin runs, and his rise from pianissimo to forte.
On the other hand the Largos sound more their age, slow (no pouring rain in Winter, but big and heavy drops dripping from the roof after the rain), and with a sentimental solo violin (except in the Largo of Winter, a slumber scene inspired by that of “Il Sonno/The Dream”, the Largo from Vivaldi’s Flute Concerto “La Notte”, which doesn’t feature the violin in a solo role). And it gets even worse in Winter, because at 1:27 Molinari assigns to the full violin section the repeat of the solo violin melody. It wouldn’t have been out of place in Love Story. But one may be easily caught by the suave and kitsch charm of it all. There’s no added ornamentation either – that’s a liberty, based on later understanding of baroque performance practice, that interpreters will assume only in the 1970s. In the Autumn’s Finale as already in its first movement, Molinari lacks the forward-moving vigor that Kaufman brought to them five years later. Another feature of this recording’s age are, evidently, the 1942, 78rpm-transferred sonics, but, for the vintage, I find that they are very acceptable, somewhat opaque but still clear enough to my ears, and with hardly any surface noise that would betray its 78rmp origin.
Oh, and let me not forget the sirocco… There’s a passage in the Finale of Winter, which repeats the opening melody of Summer, and because its initial statement was associated with the image of man and animal’s languidness under a crushing heat (which, as mentioned above, many have translated into a “Lento” tempo rather than Vivaldi’s “Allegro non molto”), many have likewise transferred that languidness and suddenly slow tempo to that passage in Winter. But Vivaldi prescribes there no slowing down of tempo, it is still an Allegro, all the more so as it is here supposed to depict the sirocco wind. And check on Wikipedia: when it hits Italy coming from North Africa’s Sahara, the sirocco is not a sweet zephyr but “can reach hurricane speeds”. It is also, specifically in the Venetian lagoon (although Vivaldi’s Seasons may have been inspired rather by the surroundings of Mantua), the cause of floods. That passage played Largo certainly has a musical validity, but it remains that everybody who plays it that way misguidedly contradicts Vivaldi’s explicit program.
With Molinari? The tempo is kept Allegro (1:53) and the passage sounds like a sirocco – “at last”, am I tempted to write. But no, not “at last”: at first. Subsequent performers would have been better inspired to derive more inspiration from that recording, the first of all Seasons.
Molinari: hmmm…. That he was a staunch admirer of Mussolini, dedicated his 1927 edition of The Four Seasons to him and was a member of the Fascist party and cultural organizations… and a Catholic, didn’t prevent him from immigrating to Israel (then still Palestine) after the war (fleeing the jeers and boos that now greeted his concert appearances in Italy), becoming a conductor and musical advisor to the Israel Philharmonic Orchestra and arranging the Israeli national anthem (see his entry on Wikipedia). The most complete bio I’ve found online is in Italian – and it too is silent about both the Mussolini dedication and the Israeli episode.
As already mentioned, playing by ear, Molinari’s edition takes some liberties with the ‘Urtext’ score as featured in the now-standard Malipiero edition and subsequent ones: organ underpinning of the violin’s solos, violas’ countermelody in the “weeping shepherd” episode of the slow movement of Summer… There is also the generous use of the harpsichord, sometimes even doubling violins rather than taking the part of continuo (as in the outer movements of Spring), and at times (as in the hunting scene that makes the Finale of Autumn, or at one point in the Finale of Summer, underpinning the figurations of solo violin) the replacement of Vivaldi’s “arco” bowing of accompaniment figures by pizzicato, or again (as at 1:27 in the slow movement of Winter) the attribution to full first violin deck of a melody ascribed by Vivaldi to the soloist. There is one other small difference with the Malipiero edition, but that is not proper to Molinari as it is found also in Toni’s 1924 edition: in the Finale of Summer, the echo-melody of second violins isn’t played at the beginning (0:13), although it is present in the coda (2:19). But those liberties are usually quite inconspicuous and, I find, don’t betray the score.
The one exception would be the slow movement of Winter, that I already mentioned above, with its attribution to the full complement of first violins of the melody written only for the soloist. On top of that, the recording makes an unrequested repeat of the whole passage (at 2:20 / end of measure 13, it “jumps” back to the point where the full first violins came in at 1:27 / bar 9, then, having reached again end of bar 13 at 3:14, segues into bar 14 as it should have in the first place, only for solo violin to reassume its role at 3:25); and since it doesn’t happen in the other recordings using the Molinari edition (Cantelli, Campoli), I have to conjecture that it is an editing error in splicing together two sides of the original 78s (it was often the case that, as the end of sides, in order to avoid a brutal cut, 78rpms would end with a cadential pedal-note)… As for the the occasional presence of a continuo organ, while it is not an option that I remember encountering often in subsequent and more recent versions, I don’t see how that breaks any “historically-informed” rule. Only in the first movement of Winter would I object to Molinari’s use of an pedal-note by the organ to underpin the solo violin passages, as it clogs the texture.
A note of warning. If you fish online, you may chance on a 2020 article by art historian Myriam Zerbi, “A Flow of Music – Antonio Vivaldi At The Origins of a Rediscovery”, published by the Fondazione Giorgio Cini and Istituto Italiano Antonio Vivaldi (available here). Therein, referring to a 1979 article by Michelangelo Abbado (father of conductor Claudio Abbado), “Antonio Vivaldi nel nostro secolo con particolare riferimento alle sue opera strumentali”, published in Nuova Rivista Musicale Italiana, XIII,1, 79-112, she writes:
“The most egregious transcription was surely that made by Bernardino Molinari of The Four Seasons, published by Ricordi in 1927. Molinari’s rewrite bristles with liberties and interpolations: the original concertos, conceived for violin and strings, found their ensembles swollen by oboes, clarinets and bassoons resulting in an orchestration as lush as it was distant from Vivaldi’s conception. And it was in this guise that The Four Seasons were first recorded in 1942, for the Cetra label, in what would remain the benchmark edition for years to come.”
To make a long story short, obviously this is not true, and in a private correspondence Mrs Zerbi kindly admitted to her mistake. My guess is that she confused with a passage in Abbado’s article (kindly sent to me by Fondazione Cini) in which he describes a doctored and overblown transcription, made as late as 1951, by the Russian Alexander Siloti – Rachmaninov’s teacher – of Vivaldi’s op. 3-11, “abolishing” the three solo violins (quite daring indeed in a Concerto for three violins!) and, “completely taken by a mania of grandeur”, adding, together with harpsichord, pairs of flutes, oboes, clarinets, bassoons, and, “as if that were not enough”, counter-bassoon (page 96).
TT 44 minutes, the duration of a Seasons with no fillers. I first acquired the 1999 reissue by Ermitage France, 12006-2, whose liner notes, at least in my copy, are… in Italian, and anyway so succinct (just short blurbs about conductor and orchestra, nothing about work and recording history) as to be useless (but the booklet nicely reproduces – much-reduced – facsimiles of the original score’s title page and a page of the solo violin part). But in the course of my research I was made aware of the liner notes of the first 1991 edition on Ermitage ERM 116S, by musicologist Piero Rattalino, that I’ve already quoted above. They are wordy and convoluted (as you, somehow, expect of Italian scholars) and highly disputable, in their effort to demonstrate (or rather; in their claim rather than effort to demonstrate) that Molinari plays not in a “historically-informed” style (which, according to him, would later inform the performances of I Virtusosi di Roma, I Musici and all those chamber ensembles that would specialize in 18th Century music), but guided by an “aesthetic sensitivity”. But when you compare Molinari’s performance to those that came out later, in the 1950s and 1960s, other than the details of doctored orchestration, Molinari is absolutely not outlandish.
Still, Rattalino’s essay is filled with priceless information on the early editions – Molinari vs Toni – and early performances, and they alone warrant the acquisition of the 1991 edition.
Of course, this is not a version for the casual listener of The Four Seasons. But for serious fans, fascinated as myself with the history of interpretive traditions, this is an indispensable addition to their collection of Four Seasons. Many other versions from the 1950s to 1970s also belong there, from I Musici’s to, say, Marriner’s, but more for documentary reasons than for musical ones: because they were important and much-feted in their time, although their interpretive approach is now passé. Not so with Molinari: there are still many things that are entirely valid and inspirational today. The long-uncredited fiddler Remy Principe does a very acceptable job at it, sometimes a little coarse tonally in the fast runs. For the “historically-inclined” music-lover like myself, I cannot recommend a choice between Molinari and Kaufman: both versions have their flaws and both their unique insights – I’m happy to keep both in my collection.
Keep up the good work.
Interesting: Vivaldi’s explicit instructions regarding the barking dog. That is why I love Harnoncourt.
Yes, exactly: Harnoncourt was the first one to play the theatre of Vivaldi. I still wonder why all the previous performers didn’t. It’s all there, very explicitly, in the score!