On comparing different versions of the Fourth Symphony (9/01)‚ I suggested Dausgaard’s may well prove the best if it combined his first volume’s fine sound with Frandsen’s interpretative edge. Suffice to say that those expectations have been more than fulfilled: this is the best recording yet of this splendid‚ unconventional symphony. Its more unassuming successor has a hopelessly complex history (of which more below). This is the third recording of the Fifth’s second (1931) version: comparing them shows how strong a Langgaard interpreter Stupel remains. Danacord’s sound is still acceptable if not as warm or precise as the competition’s and the playing is full­blooded if without the DNRSO’s knowledge of the style – one of the newcomer’s strongest attributes. It must also be said that Järvi’s tempi are very fast: the opening‚ for instance‚ is marked Lento misterioso‚ after all. True‚ the quasi­Sibelian Andante conclusion sparkles more for him‚ but Dausgaard’s new account almost matches it and throughout more of the detail can be heard. Dausgaard’s account is the slowest of all‚ at times a touch ponderous‚ but I suspect that this is right; Stupel runs him very close‚ however. The real novelty here is the Fifth’s previous (1926) version‚ actually the third of five possible incarnations of a 1917­18 orchestral fantasia‚ Summer Legend Drama (‘Sommersagnsdrama’)‚ overhauled with new material in 1920 as Symphonic Festival Play‚ which was revised in 1940­41. The difference between the Symphony’s two versions is startling‚ explained by their being based respectively on Summer Legend Drama and Symphonic Festival Play rather than each other. Nowhere is the expressive divergence more fascinatingly audible than in Version 1’s conclusion: a haunting violin solo depicting the nix (a Nordic sprite)‚ which throws a wholly different light on the lowering‚ rushing string passage at the close.