The dark opening chord of the Scenes historiques already proclaims its author. The Lydian modality and muttering string textures of this overture and the revolving accompaniment to the ceremonial march in the following ''Scena'' suggest a date around the time of the Fourth or Fifth Symphonies. In fact, the first set of the Scenes, although revised in 1911-12, was composed in 1899, the year of the First Symphony and Finlandia (itself originally one of the set). The Op. 66 pieces were indeed composed in the aftermath of the Fourth Symphony, and with hindsight one could point to a greater concentration and more epigrammatic style. But there are fine things in both sets—that ceremonial march haunts me mercilessly, and the halting tenderness of ''At the drawbridge'' (last of the second set) is scarcely less memorable.
Recording and performance on the Chandos issue have been highly praised, and although I have slight preference for the new BIS version, both sound absolutely superb, especially of course on CD—choice may largely depend on the coupling. Neeme Jarvi and Sir Alexander Gibson are in close agreement on tempo and characterization, but the Gothenburg woodwinds are superior in tone and intonation (especially in the last two Scenes) and just occasionally the BIS recording is a franction more realistic (compare the castanets in the curious ''Festivo'' from Op. 25).
I would also recommend the BIS En saga in preference to the others listed, though again by no great margin, since all have considerable merit. Jarvi seems to judge the weight of his reading perfectly at each stage—the unexpectedly measured tread at fig. H pays great dividends later on. Interestingly enough, his is the only performance which sounds in the right key at the end—with Ashkenazy (Decca) and Berglund (HMV) the fuller sonority of the pedal E flat sets up an unwanted overtone tierce de picardie; with Davis (Philips) an overtone from the cymbal roll creates a peculiar kind of added sixth. Ashkenazy and the Philharmonia come up with both the most exciting and the most questionable features—undeniable magnificence in the Berliozian surge after fig. Q, but a piu lento instead of a tempo after fig. O which makes for a fatal sag in the structure. In this music the view of the whole can redeem miscalculations in detail, but not the other way round. The Chandos record is completed by fine performances of Rakastava and the Valse lyrique, the former, as RL noted, with a rather slow last movement, and the latter not Sibelius at his finest. A minor annoyance on the Jarvi LP is the narrowness of bands between individual pieces.'