Jorge Luis Borges once said that he had always imagined paradise as a library. This is the story of how, in one such library hidden in the Southern Hemisphere, a film reel waited eighty years for its moment—a reel capable of changing our understanding of one of the greatest films in cinema history.
In 1925, director Fritz Lang began shooting a film that UFA could not afford. A budget of 5 million Reichsmarks—an astronomical sum that nearly bankrupted Germany’s largest film company. Shooting stretched over 17 months. 36,000 extras passed through the studios at Babelsberg. Lang was building a city of the future with the obsession of a man who truly believed cinema could contain an entire civilization—with its underground chambers, cathedrals, machine halls, and plazas where crowds moved like the gears of a clockwork mechanism.
The premiere of Metropolis took place on January 10, 1927, at Berlin’s UFA-Palast am Zoo cinema. The full version ran for roughly two and a half hours. Critics were bewildered: some marveled at its visual force, while others condemned Thea von Harbou’s screenplay as naïve. But its commercial failure decided the film’s fate faster than any reviews could. By the summer of 1927, UFA and the American distributor Paramount had begun hacking the picture to pieces. Entire plotlines were cut, characters’ motivations simplified, episodes that distributors deemed “superfluous” vanished. The film was, in effect, rewritten with an editor’s scissors—and in this truncated form it spread around the world.
The complete Berlin version ceased to exist. For decades, film scholars and restorers worked with what remained: shortened release prints, each cut differently depending on the country and the tastes of the local distributor. The screenplay, cutting continuities, production stills—all of it made it possible to understand exactly what had been lost, but not to see it.
In 1928—just one year after the Berlin premiere—a 35-millimeter export print of Metropolis was sent to Buenos Aires. It had been commissioned by the Argentine distributor Terra. This was a version close to the original Berlin cut—one of those prints sent abroad before the distributors’ scissors completed their work on the domestic German market.
What follows is a story typical of Latin American film heritage: the reel did not end up in a state archive, but settled into private hands. It came into the possession of film critic Manuel Peña Rodríguez—a man who apparently understood that he was holding something valuable, but lacked the means to preserve it professionally. At some point, a 16-millimeter copy was made from the 35-millimeter original—a reduced duplicate, a kind of insurance policy in case the main reel perished. It was this copy that would outlive everything else.
In the 1960s, the materials passed to Argentina’s national arts fund. In 1992, the 16-millimeter copy arrived at the Museo del Cine Pablo Ducrós Hicken—Buenos Aires’s municipal film museum. There it lay among thousands of other reels, catalogued but never truly examined. No one carried out a frame-by-frame comparison with the known versions of Metropolis. No one checked whether this particular copy contained something absent from every other archive in the world. Put bluntly, the film was in plain sight—and invisible at the same time.
What fascinates me about this story is not the discovery itself, but its mechanics. This was not a case of someone opening a dusty suitcase in an attic and gasping. It was the work of expertise—slow, painstaking, demanding highly specific knowledge.
In 2008, Argentine film historian Fernando Martín Peña and museum curator Paula Félix-Didier began systematically studying the Museo del Cine collection. When they reached the 16-millimeter copy of Metropolis, something happened that can only be called professional shock. Peña knew Lang’s film inside out—he knew every frame of every known version. And there he was, seeing footage that existed in none of them.
The copy contained approximately 25 minutes of sequences lost from all known material worldwide. These were entire plotlines: the story of Georgy (the worker who exchanges clothes with the protagonist Freder and enters the world of the upper city), the thread of Josaphat (an employee fired for a moment of weakness), the actions of the Thin Man—a spy hired by the magnate Fredersen to keep watch. There were also scenes at the monument to Hel—Freder’s mother, the woman because of whom Rotwang and Fredersen became enemies. Extended episodes in the Yoshiwara entertainment district surfaced, along with details of the children’s rescue during the flood at the film’s finale.
Each of these fragments was known from the screenplay and editing documents. Film scholars had known for decades that these scenes existed. But knowing them from a description and seeing them on film are as different as a bridge blueprint and the bridge itself. The recovered episodes changed the very dramaturgy of Metropolis: the political conflict between the upper and lower cities became more layered, the characters’ motivations more comprehensible, Lang’s editing logic more transparent.
The 16-millimeter copy had serious problems. The format itself meant a loss of quality: the original had been shot on 35-millimeter film, and transferring it to 16 mm reduced the resolution by roughly four times. The image was grainy, high-contrast, damaged in places—scratches, dirt, sections where the emulsion had chemically degraded. Some shots were difficult to make out.
But it is important to understand the logic of film archiving here. Restorers do not need a perfect copy—they need information. Every frame, even a cloudy, scratched one, carries composition, staging, the duration of an episode, the order of the edits. It is like a palimpsest: the text is damaged, but legible. And if that text is the only surviving copy on the planet, the quality of the medium becomes secondary.
The work of comparing the recovered scenes with existing materials was, in essence, philological. The surviving screenplay by Thea von Harbou, the original production’s cutting continuities, censorship cards, production stills—all of it served as a coordinate system. Every recovered episode had to be identified, its place within the film’s structure determined, compared against its description in the documents, and verified as the very material in question rather than an alternate take or version of the scene. The process resembled an archaeological reconstruction: fragments of mosaic being fitted into an image known from a blueprint, but never before assembled.
The rights to Metropolis belong to the German Friedrich-Wilhelm-Murnau-Stiftung, based in Wiesbaden. The foundation’s director Helmut Possmann and lead restorer Martin Koerber became key figures on the German side of the project. The Deutsche Kinemathek—Berlin’s film archive, headed by Rainer Rother, which holds one of the world’s largest collections of material on German silent cinema—was brought into the work.
Collaboration between the Buenos Aires museum and the German institutions required resolving dozens of technical and legal issues. The reel had to be scanned at the highest possible resolution without damaging the already fragile medium. The recovered 16-millimeter footage was to be integrated into the existing restoration, based on 35-millimeter sources. The difference in quality was obvious: the new scenes looked markedly worse than the “old” restored fragments. The restorers decided not to conceal that difference, but to mark it—the viewer should be able to see where material from the Argentine copy begins and understand why it looks different.
It seems fundamentally important to me that no one tried to create a “seamless” version by disguising the footage’s origin. That would have been restoration fraud—like a museum filling in the cracks of an ancient vase until it looked new. Here, the honesty of restoration became an ethical choice: the viewer received a film in which the seams remained visible, but the structure had been restored.
On February 12, 2010, the restored Metropolis was screened simultaneously in two cities. In Berlin, the premiere took place as part of the Berlinale at the Friedrichstadt-Palast: the Berlin Radio Symphony Orchestra, conducted by Frank Strobel, performed the original score by Gottfried Huppertz, written for the 1927 premiere. At the same time, the film was projected on a screen by the Brandenburg Gate—for those who could not get into the theater. In Frankfurt, it was shown at the Alte Oper. Two cities, two screens, one film—at last, for the first time in decades, close to what Lang had intended.
The orchestral accompaniment was not a decorative gesture. Huppertz’s score is a full-scale symphonic work, designed for the complete version of the film. When Metropolis was cut down, the music lost fragments too. Restoring the missing scenes made it possible to return musical passages that had existed only on paper for decades. Strobel reconstructed the film’s sonic fabric in parallel with the visual restoration—the two labors moving toward one another.
The reaction was predictably intense. Critics and film scholars who had spent years studying Metropolis through truncated versions saw for the first time how Lang’s editing logic worked as a whole. The Thin Man’s storyline, for example, transformed the film’s political plot from a schematic parable into something more unsettling—a story of surveillance, control, and a ruling power’s paranoid distrust of its own subjects. The scenes in Yoshiwara gave the upper city a dimension of decadence without which its conflict with the lower city had seemed too abstract.
The discovery did not make Metropolis absolutely complete. After the 2010 restoration, small fragments were still missing from the film—several scenes absent even from the Argentine copy. At those points, the restorers inserted text intertitles describing the lost action, based on the screenplay and editing documents. The film came within roughly 95% of Lang’s original structure—a colossal leap compared with earlier versions, but still not a one-hundred-percent restoration.
Those remaining five percent are a reminder of irreversibility. The original 35-millimeter negative from the Berlin premiere appears to be lost forever. Copies dispatched around the world in 1927–1928 were cut, lost, destroyed by war, negligence, the chemical decomposition of nitrate film. The fact that the Argentine copy survived to the twenty-first century at all was the result of a chain of decisions, each of which might have gone differently. Peña Rodríguez might not have preserved the reel. The arts fund might have written it off. The museum might not have accepted the collection. Any link—and the chain breaks.
Since 2001, Metropolis has been included in the UNESCO “Memory of the World” Register—a program created to protect humanity’s documentary heritage. It was added to the register before the Argentine discovery, on the basis of what was known at the time. After the 2010 restoration, the film’s status in the UNESCO program received, in effect, fresh material confirmation: now what was being protected was a film that could finally be seen almost in its entirety.
The central drama of this story is not the reel itself. The reel lay on a shelf and had nowhere to go. The drama lies in expertise. In that moment when Fernando Martín Peña looked at footage that dozens of people before him had seen, and understood what no one else had understood. That required not luck, but knowledge—detailed, almost maniacal knowledge of one particular film, its history, its versions, its losses.
Archives are strange institutions. They preserve things whose value is often unknown even to their keepers. The Museo del Cine in Buenos Aires is a small municipal museum, not the Cinémathèque Française or the Library of Congress. Its collection took shape haphazardly, from private collections, donations, chance acquisitions. It is precisely in places like these—at the periphery of the global archival system—that things hide which the center has long considered lost. Not because the periphery preserves them better, but because the center does not know what lies there.
The 2010 restoration of Metropolis became possible because two people in Buenos Aires had enough expertise to recognize the discovery, and enough persistence to carry it through to international cooperation. It was also possible because German institutions—the Murnau-Stiftung and the Deutsche Kinemathek—were prepared to take the Argentine material seriously despite its technical imperfections. The result was a film in which Lang’s vision can finally be seen—albeit with scratches, grain, and several gaps filled with text. Somewhere in the world’s archives, there may still be reels capable of closing that remaining five percent—but for now, nothing is known about them.