The gestation of the Nocturnes is, of all of Debussy's works, the most difficult to reconstruct. The work's first iteration—that of the Scènes au crépuscule —was, as we have seen, hypothetical to say the least. We know a bit more about his second idea, conceived for violin and orchestra, since the composer wrote to Lerolle on 28 August 1894 that he had begun three Nocturnes in which he intended to use “separate orchestral groups, in order to try to find nuances within these groups alone, because composers don't really risk enough in their music.” Whistler's influence is obvious here and indeed would have been perceived as such at the time when the work became known. The painter's biographer, Théodore Duret, would even attempt, as early as 1904, to compare the two art forms. For various critics, Debussy would be the “Whistler of music.” We will recall that one month later, he outlined his plan to Ysaÿe, for whom the work was intended. On some undetermined date, perhaps near the beginning of 1894, Chausson wrote to Ysaÿe: “Debussy has already written the first part of your violin piece. I don't know it, and don't want to see it until my own piece is finished.” The composition of Pelléas subsequently delayed the continuation of Debussy's work, but he had scarcely completed his opera when he announced to Lerolle that he now wanted “to finish the three Nocturnes”— his typical equivocal expression. In November 1896, everything changed when Ysaÿe let him know that he could not premiere his work. During the following year, the version for violin and orchestra was transformed into a symphonic version.

From the end of 1897 onward, the completion of these new Nocturnes became a leitmotif in Debussy's letters to Hartmann. On 25 June 1898, the composer reported for the first time that the score was finished, but of course he was only talking about a rough draft, and the following months were spent orchestrating it.