The bassoon solo in the “Berceuse” is the Firebird’s lullaby to Kastchei and his monsters that have danced themselves to exhaustion during the “Infernal Dance.” We might assume that Stravinsky’s own recordings could be used as reference for an appropriate tempo, but the tempos of each are quite varied. For example, Stravinsky takes a tempo close to  ♪ = 92 in his 1946 recording with the New York Philharmonic; fifteen years later, however, he takes a tempo closer to  ♪ = 120 with the Columbia Symphony Orchestra. I prefer a tempo in the former range (about  ♪ = 90-100), which can be heard in the recordings of the Royal Concertgebouw Orchestra and the San Francisco Symphony.

The Case of the Mysterious D♮/D♭
The main issue I want to discuss with this excerpt is the now infamous question of whether to play a D♭ or D♮ in the fourth bar of Rehearsal 186. Luckily, the long and complex circumstances leading to this confusion have been expertly documented in Jeffery Lyman’s 2008 Double Reed article “D or D flat?: Stravinsky’s Berceuse and the Long Story of a Short Note."1 The controversy began with Stravinsky’s lesser-known 1912 revision of the “Berceuse,” when the natural sign in front of the D was omitted from the published score. I believe that we can dismiss this missing natural sign as an error on the publisher’s part, since this version of the “Berceuse” was the only one to appear before 1928 in either print or recorded form without the original D♮. We know that Stravinsky’s scores were often published in poor condition, and that Stravinsky himself often had little control over the publishing rights to his own music. Robert Craft, Stravinsky’s close friend and colleague, gave us an idea of just how many mistakes could be possible in one edition of The Firebird when he pointed out that the score for the 1919 suite contained over three hundred errors in total.2

Some bassoonists attempt to explain the necessity of a D♮ by looking at the accompanying harmonies, and argue that the note must be a D♮ so that it does not clash with the D♮ in the accompaniment. However, Stravinsky’s second reduction for violin and piano (published in 1932) appears to silence any such assertions. Here, for the first time, Stravinsky directly indicates that the note should be different than in his original ballet score. In this version (transposed to E Minor) Stravinsky marks what corresponds to a cautionary flat sign before the D, while leaving the dissonant accompaniment intact. By the time of the 1945 suite, the note had officially been changed to D♭ in the bassoon solo as well.

Did Stravinsky make this change in his 1945 suite because he started to believe the D♭ version sounded better? In 1959, he lamented:

I laboured again and again on that piece, but could do no better, and an awkward orchestral handicap remains, though I cannot say exactly what it is. I have already criticized The Firebird twice, in my revised versions of 1919 and 1945, and these direct musical criticisms are stronger than words.3

Interestingly, Stravinsky often expressed his disdain for Firebird’s popularity, and complained about the number of times he was requested to conduct it. Could it be that he simply began caring less about the piece, therefore allowing bassoonists to choose either note they wished? Accounts of Stravinsky yelling directions at the orchestra in rehearsals—including specifically for the bassoonist not to play D♮—do not portray a composer who no longer cared about the finer details of his work. Furthermore, he retroactively applied the D♭ to the original ballet version as well, which can be heard in the included 1961 recording with the Columbia Symphony Orchestra.4

There is little doubt that Stravinsky preferred the D♭—after all, he did make the conscious decision to change the note in his suites, and most of his own recordings feature the D♭ as well. But the original version—which went unchanged for eighteen years—was a D♮. Just because Stravinsky made this change to his own piece, does it automatically mean we must follow it? To claim that an artist can do absolutely no wrong when altering his own work is a bit too reverential, in my opinion. After all, George Lucas made a number of changes to his Star Wars movies twenty years after they were released, yet most fans agree that many of these revisions actually make the movies worse—not better—than the originals.

So what are we supposed to think? Should it be a D♮, a D♭, or should we just choose whichever note we prefer? Though this final option is both a tempting and easy choice, I believe that one of these notes is far more musically interesting and logical—the D♮. Instead of examining how the solo relates to the underlying accompaniment, we should actually take a closer look at the melodic structure of the solo itself.

Similarities of the Fundamental Line
To understand why the D♮ is the better note to play here, we first need to figure out where this section of new material comes from. In the first half of the solo, we have two consecutive statements of a four-bar melody. At Rehearsal 132, this melody appears for a third time, but it is interrupted by a new four-bar phrase in the third bar. At the end of this interruption, we are finally presented with the last two bars of the main melody that began at Rehearsal 132.

The original melody and the “interrupting” melody both span four bars, and this should be our first indication that there may be a connection between the two. Looking closer, we can also see that the pitch structure and melodic contours of each are also very similar. McGill explains exactly how these melodies are related:

When this theme is repeated later, it is expanded by simply ornamenting the main melody. First, there’s part A of the theme and its variation, then part B’s variation, and finally, part B essentially as it originally sounded… Whether to play a D♭ or a D♮ in the fourth bar, here, causes a lot of controversy among otherwise levelheaded bassoonists. To me, it’s simply an ornamentation of the original melody, so I play a D♮.5

In fact, if the note is played as a D♮, then the fundamental lines of both the original melody (Example 18.1) and the interrupting melody (Example 18.2) are identical.6 As both examples show, the fundamental structure of the melody is B♭ – D♭ – D♮ – E♭ – F – E♭ (scale degrees 5 – 7 – #7 – 1 – 2 – 1),7 but even more interesting is that, as shown in the fourth level of Example 18.2, the D♮ of the interrupting melody does not resolve to E♭ until the fourth beat of the bar. This means that, at a deeper level, the questionable D♮ of the interrupting melody actually lasts an entire beat longer than the D♮ of the original melody! From this perspective, we can see that by changing this note to a D♭, we are really changing two beats of the bar, not just one note.

Example 18.1. Stravinsky, The Firebird – bars 3 to 6 of Rehearsal 183, melodic reduction

Example 18.2. Stravinsky, The Firebird – bars 3 to 6 of Rehearsal 186, melodic reduction with D♮

Example 18.3 shows the fundamental line of the solo with the D♭ instead of the D♮. As you can see, it’s not nearly as elegant as the version with the D♮—it is almost the same as the original melody, but not quite the same. In fact, I would argue that it is just different enough to be distracting—once you are aware of the fundamental line that repeats in each of the first two statements of the melody (Example 18.1), you will start to expect the D♮ (sounding as the leading tone to the subsequent E♭) here as well. To me, hearing a bassoonist play a D♭ instead of a D♮ is tantamount to hearing a bassoonist play a completely wrong note.

Example 18.3. Stravinsky, The Firebird – bars 3 to 6 of Rehearsal 186, melodic reduction with D♭

One argument for playing the D♭ over the D♮ could be that the D♭ – C♭ M2 interval is easier to play than the D♮ – C♭ augmented interval. It is possible this played into Stravinsky’s decision to first change the note in the 1932 reduction for violin and piano, but most bassoonists would agree that the interval of D♭ – C♭ is actually more awkward because of the additional keys needed to play the D♭. Bearing that in mind, this argument of decreased difficulty seems to hold little water in regards to the bassoon solo of the orchestral version.

At the end of this short discussion, what have we discovered? The D♮ is more interesting melodically because of the exotic augmented D♮ – C♭ interval; listeners—at least subconsciously—expect to hear the D♮ because it recreates the same fundamental line as the first two statements of the melody; and the D♮ – C♭ interval is even physically easier to play than the D♭ – C♭ interval. In my opinion, the choice is clear—the note should be a D♮.

1 Jeffrey Lyman, “D or D flat?: Stravinsky’s Berceuse and the Long Story of a Short Note.” Double Reed, Vol. 31, No. 1 (2008), 75-85.

2 Robert Craft, Stravinsky: Glimpses of a Life (St. Martin’s Press: New York, 1992), 10.

3 Stravinsky and Craft, Expositions and Developments, 132.

4 Columbia Symphony Orchestra, Stravinsky: The Ballets, Vol. I, Igor Stravinsky, conductor,Sony Classical SM3K 46291 (CD), 1991.

5 McGill, Orchestral Excerpts for Bassoon. Quote transcribed and punctuated by the author.

6 To arrive at their identical fundamental lines, the B melody—a more florid line containing a few more notes than the A melody—must be reduced one level further than the A melody.

7 This is essentially a prolonged motion from B♭ to E♭ (scale degree 5 to 1).