In February 1934, a sick Jean Vigo had finished editing L’Atalante during a constant exchange with his ‘band’, a group of close friends, who, together with the editor Chavance, kept the director informed when he could not be present in person. Vigo then left Paris to recover, while his friend Maurice Jaubert was completing the music. Vigo would never again find the strength to continue working on the editing as he desired. His collaborator, Albert Riéra, proposed to tighten the cutting, but producer J.L. Nounez refused that anyone should substitute the director. Only following the disastrous reception at a trade screening did he accept the proposal of co-producer and distributor GFFA to substitute Jaubert’s music with a ‘realistic song’ adapted from the Italian, Le Chaland qui passe; the title with which the film would be released that September, shortly before Vigo’s death. In the meantime, a print of the original L’Atalante has been sent to London, entrusted as it seems by Maurice Jaubert to Alberto Cavalcanti.
This is the print that has been used by the present restoration as a reference, for both picture and sound. We reintroduced the final aerial shot, taken by Kaufman instructed by Vigo, whose length was foreseen in the recorded score. A number of cuts (caused by film wear, censorship, voyeur projectionists?) have been restored with the help of some prints of Le Chaland qui passe, mutilated only after its the release at the Colisée Theatre, when GFFA heeded – or took advantage of – the protests of some spectators. The approach taken was to return to the original 1933-1934 film, without trying to adapt it to the viewing habits of 21st century spectators.
Bernard Eisenschitz

At the end of the 1940s, the Cinémathèque française acquired an ensemble of outtakes of L’Atalante. It was first inventoried by Panfilo Colaprete and Vigo’s first biographer, Paulo Emilio Sales Gomes, who already back then declared that an “ideal edit” of the film was impossible. Langlois himself gave it up: “I have seen some superb things that Vigo had cut, because what he was striving for was total simplicity. I have made a version, into which I inserted some scenes out of pure curiosity, to see what effect they would have. Magnificent, declared those who saw them. But it wasn’t what Vigo wanted”.
In his 1950 restoration, Langlois thus confined himself to inserting two shots in the most plausible points possible: the newlyweds walking around haystacks, and that (inspired by L’Âge d’or, as Colaprete noted) of Jean licking a block of ice.
Many viewers have been struck, and with good reason, by the beauty of these images outtakes when they were edited into various restored versions, such as Soviet filmmakers Otar Iosseliani and Andrej Tarkovskij after seeing the Langlois print, or British mythographer Marina Warner, after seeing the 1990 restoration. Nonetheless, others commented on “the overenthusiastic tendency of the 1990 team to include… every piece of material available” (Michael Temple, Vigo, Manchester University Press, Manchester-New York, 2005).
Now, it is finally possible for us to discover, together with the restoration of the film in its pristine form, precisely what material was shot and then discarded – material that both illustrates the story of the working process, and the nature of the director’s inspiration.

Bernard Eisenschitz