The numbering of the Leonore overtures is misleading. No.1, thus called because Beethoven's biographer Anton Schindler asserted it was the first to be written, was instead almost certainly composed last (in 1807). No.2 was the first, for the original version (1805) of the opera that Beethoven wished to call Leonore but which actually appeared with the title Fidelio. When he revised this opera in 1806 the overture was what we now call No.3. The final version of the opera (1814) has an entirely different overture, in E major.
This “No.2” was long considered no more than an Urfassung for No.3, hence not published at all until 1842, but it is very much a complete and valid piece in its own right, and perhaps especially the passage between bars 214–35 presents an attractive alternative turn of phrase that is worth preserving and cherishing.
sources
(A Autograph score, lost)
(PX First performance parts, lost; some manuscript parts in the Beethoven- Haus, Bonn were copied later from B and are of no significance)
B Copyist's score (1805) with some corrections by Beethoven, others (not all necessarily correct) by Mendelssohn, housed in the Beethoven-Haus (shelf mark H.C. Bodmer Bk 3). Some pages, containing bars 38–52, 433–42, 484– 519, are missing entirely. B was very possibly copied from A
C Copyist's score (1808) written by Joseph Klumpar; copied from PX (but Pos, and off-stage Tr, missing); no LvB corrections. Discovered in 1852 by Otto Jahn, now in the Staatsbibliothek zu Berlin (shelf mark Mus. ms. autogr. Beethoven 66)
B and C are mutually independent and, as authentic sources, are of equal value; many of the differences between them are listed in Hess 1972 (but he calls our C, “A”)
E,P First edition score and parts, published by Breitkopf & Härtel in 1842; text inevitably taken from B, hence bars 38–52 and 433–42 are simply missing. But in order to bridge the gap between 483 and 520 the equivalent passage was lifted from Leonore No.3 and inserted here
F New edition by Otto Jahn, published by Breitkopf & Härtel in 1854 following the discovery of C, and taking the text of C as its model (instead of E's B – apart from Pos, which of course stem from B)