When Leopold Nowak’s edition of the original 1873 version of Bruckner’s Third Symphony was published in 1977, there was hope among Bruckner scholars that it would be performed sufficiently often until – as Robert Simpson put it – ‘the flaws of the two later versions become obvious to all’.

More than 40 years have passed since then, and it’s a revolution that’s refused to happen. Eliahu Inbal took up the cause in 1983 (Teldec, 6/84) and there have been two distinguished recordings from Roger Norrington: the first with the period-instrument London Classical Players (EMI, 1/97), the second with the Stuttgart RSO (Hänssler, 9/08). That Stuttgart version, recently reissued as part of a five-CD budget-price set of all Norrington’s Stuttgart Bruckner recordings (A/21), is probably the pick of the bunch.

François-Xavier Roth’s evolving Cologne cycle appears to be dealing only in first versions, hence this 1873 Third. As a performance it’s not unlike Norrington’s, the only difference being an even greater edge-of-the-seat tension in the Cologne playing, something to which Bruckner regulars may or may not warm.

That’s certainly the case at the very end of the symphony, that famously disconcerting 50-bar D major dash, which Roth plays as directed – ‘very quick’ and very loud (triple forte). It makes a devil of a racket. Norrington takes a slightly more measured approach, quietly acknowledging what Bruckner himself decided in his final, 1889 version of the symphony. There the tempo is adjusted to ‘quick’, the dynamic level to double forte, with the trumpets asked to play broadly and with restraint.

As for one of the symphony’s more unforgettable moments – the finale’s counterpointing of polka and chorale, said to have been inspired by Bruckner hearing dance music from a nearby house while a revered colleague lay in his coffin in the church across the way – that’s disappointingly garbled in all recordings of the 1873 version, even Norrington’s. Here Bruckner’s worryingly vague instruction, ‘somewhat slower’, is the principal culprit, heedless of the fact that an Allegro movement in 2/2 will encourage conductors to take the opening at quite a lick. In the 1889 revision, the polka-chorale’s marking is changed to a less ambiguous ‘slower’.

No one does that better than Karl Böhm in his 1970 recording with the Vienna Philharmonic (Decca, 9/71). Like many old-school Brucker conductors – Günter Wand another prime example – Böhm uses the 1889 text. The scholarly argument is that playing an 1873 composition in an 1889 re-composition is an aberration. Yet, if that means missing out on some of this oddball symphony’s most characteristically Austrian features, I’m at one with those old-school Brucknerians in going for 1889.