‘Great Conductors of the 20th Century’ is the boast of this enterprising but uneven series, though as the late Professor Joad would have warned: ‘It all depends on what you mean by a great conductor’.

To deal with the worst first. It takes a perverse kind of genius to transform Bruckner’s Seventh Symphony, the loveliest and most perfectly wrought of his mature works, into the Frankenstein’s monster Mravinsky conjures up in this live 1967 performance. It is a reading which kills the things it loves. Every note in the first two movements is played con amore yet even if one sets aside the fact that the orchestra, with its brazen trumpets and wobbly Wagner tubas, is ‘wrong’ for Bruckner, you still have to contend with the fact that this is an over-emphatic, over-intense, and (by Leningrad standards) really rather poorly played performance. The finale beggars belief: a brassy uproar, humourless and, in the final analysis, inhumane. Things are marginally better on CD2 though I would not willingly trade this account of Haydn 88 for Furtwängler’s or this Francesca da Rimini for Stokowski’s. The Leningrad orchestra is at its persuasive, idiosyncratic best in the Glazunov but Mravinsky bullies the tuttis and, as in the Haydn, gives an unduly rushed and regimented account of the finale.

It was Klemperer who famously countered the accusation ‘Szell is a machine’ with the retort ‘Yes, but a very good machine’. At its best, Szell’s conducting married sense and sensibility to an astonishing degree. Sadly, this latest anthology does him few favours. The 1970 Cleveland Dvoák Eighth is a studio-bound affair, finely shaped and played but mannered and over slow in places. This is not a symphony that takes kindly to being made to put on its Sunday best. The most persuasive Dvoák Eighths are sensitive to Dvoák’s skill in adapting symphonic procedures in a manner which allows subsidiary ideas to dance attendance in a quasi-improvisational way on what is, in effect, a single commanding theme. Szell’s performance is more an inventory of orchestral effects, every drum-roll itemised, than a living performance.

Perversely, the live 1966 Tchaikovsky Fifth is also sombre and heavy-handed, a huge disappointment after Szell’s (largely forgotten) 1959 Cleveland studio recording where the rhythms live, move and have an altogether more dance-like being. Would he have sanctioned the release of a recording in which the Cologne players can hardly get a pizzicato properly together? I doubt it. The WDR orchestra gives a better account of itself in Debussy’s La mer but what is the point of hearing the notes if the music is missing? The overtures do nothing to redeem the anthology. Auber’s Rossini-influenced overture to Fra Diavolo, is well done – the chamber music-like dimensions of the opening, the orchestral barrage at the end – but real Rossini it is not. In attempting to out-Toscanini Toscanini, Szell gives us as brutal and unstylish an account of the overture to L’italiana in Algeri as you are likely to hear.

After five hours of what Stravinsky used to call ‘musical execution in the firing-squad sense of the term’, the Kempe anthology comes as musical balm. Here in the Wolf, the Bruckner, the Johann Strauss and the Ravel, the sound is the man: civilised, warm, discriminating without fuss. As an anthology, it is compromised, not unusually for this series, by the inclusion of a performance of a big, standard repertory symphony that simply does not merit inclusion. True, Kempe’s live 1974 RPO Eroica has greater musical concentration than the Eroica which sits at the centre of the (prize-winning) Fricsay volume in this same series but it is nonetheless a hopelessly dated reading. The first movement is far too slow, uses a bad text, and has a nasty horn fluff at a crucial moment of transition; after which the slow movement sounds more like a Song of Autumn than a Funeral March.

By contrast, CD1 is a tour de force. It begins with a fiercely argued 1960 studio recording of Brahms’s Tragic Overture of which Trevor Harvey wrote in these columns: ‘This is the most uninhibited Kempe that I have come across. A general impression has been made that he is an intellectual conductor, thinking more about the design of a movement than its emotional content, working to his climaxes with deliberation.’ Interestingly, the live 1971 Munich recording of Bruckner’s Fourth has all these qualities: warmth and drive and a fine sense of musical architecture. Other partnerships could be said to match Kempe and the Munich Philharmonic in the opening movements but Kempe scores a particular success with the far from foolproof finale. The Brahms shows EMI’s uncertain touch with the acoustic of Berlin’s Grünewaldekirche but the Herkulessaal Bruckner is predictably fine.

And so to Mitropoulos, the least regarded of these four conductors by the record industry – and, by devolution, the world at large – yet the purveyor of four performances which stand head and shoulders above anything else I have heard in this series. Mitropoulos delivers the five scenes from Berlioz’s Romeo and Juliet new-minted, as though hot off the press. The La mer is elemental and exalted, the Dance of the Seven Veils no tawdry showpiece but an 8'49" distillation of what Salome is terrifyingly about. All these are studio recordings made in clear, well defined sound in New York in the early 1950s. The Berlioz has been on CD, but as late as 1997 when John Hunt published his discography, Back from the Shadows, the Debussy and Strauss were still awaiting their first UK release. Toscanini was largely to blame. As DH Lawrence said of Michelangelo, it became damnably difficult for people to see past him.

If that isn’t a sufficient feast, there is also CD1 which is given over to a towering, emotionally riven account of Mahler’s Sixth recorded by WDR, Cologne a year or so before Mitropoulos’s death from heart failure in Milan in 1960. This was a musical dark age where Mahler’s Sixth was concerned, yet Mitropoulos had conducted the work way back to 1947 when he gave its American première, an event condescendingly ‘approved’ by Mahler pupil Bruno Walter who had persistently ducked the challenge. By 1959, Mitropoulos probably knew the symphony, musically and within his own mind and imagination, better than anyone on the planet, yet he was still re-thinking it. (The present performance places the Scherzo second, anticipating Ratz’s Critical Edition by four years.)

The Sixth is the only properly ‘tragic’ work Mahler ever wrote. This accorded with Mitropoulos’s mood at the time. Writing to his long-standing friend and admirer Katy Katsoyanis only days after the present performance, he remarked: ‘I have drawn into myself more than ever before and I long for moments of isolation. I look at the people around me with the same kindness but with more indifference, as though I weren’t living among them any more, and was watching from another world… It’s as though I am expecting the end.’

The performance has been on CD before in ill-sounding pirated form but what we have here is WDR’s own transcription. And very fine it sounds. The reading is astonishingly personal yet totally coherent. For a man who was as ill as Mitropoulos already was, it has remarkable tenacity and power. The Cologne players were clearly on their knees by the end but that detracts nothing from the sustained purposefulness and resilience of their playing. As Michael Tanner, author of the booklet essay, says of Mitropoulos’s La mer: ‘It is exhausting and exhilarating to listen to, but one is hardly sure which faculty has taken the strain.’