The Washington PostDemocracy Dies in Darkness

National Gallery confirms one of its Vermeers is actually not a Vermeer

Experts had long wondered whether ‘Girl With a Flute’ was really one of the world’s few paintings by the Dutch master. The pandemic gave the museum a chance to investigate.

“Girl With a Flute,” one of four paintings in the National Gallery's collection attributed to Vermeer, has been determined to have been painted by someone else. (National Gallery of Art)
8 min

It has long been designated as dubious. Now it’s official: “Girl With a Flute,” one of the National Gallery of Art’s four paintings attributed to Johannes Vermeer, is not, in fact, by Vermeer. Four are now three, and thanks to new combinations of scientific analysis, art historical insight and informed looking, a vexing, long-standing problem has been resolved.

In a news conference Friday, the museum shared the finding that an interdisciplinary team of curators, conservators and scientists has determined that the painting was made “by an associate of Vermeer — not by the Dutch artist himself.”

So Vermeer didn’t paint ‘Girl With a Flute.’ Why think less of it?

Vermeer (1632-1675) is one of the world’s most beloved painters. In normal times, people come to the National Gallery expecting to see all its Vermeers on display. It’s hard to justify removing them to the conservation lab for more than a day or two. But the pandemic changed that.

According to curator Marjorie (Betsy) Wieseman, head of the National Gallery’s department of northern European paintings, the museum’s extended closure meant that she and her colleagues had “a unique opportunity to take all four paintings off the wall and have them in the conservation lab at the same time.”

“Other people did needlepoint and learned to bake bread,” she joked in an interview Thursday. “This was our pandemic project.”

The halo of exceptionality around Vermeer’s name is made more luminous by the fact that his output was paltry. There are only about 35 paintings by Vermeer in the world. That partly explains why — although he was esteemed during his lifetime — for two centuries Vermeer was largely forgotten until his rediscovery in the 19th century. (“Girl With a Flute” was rediscovered in 1906 and donated to the NGA by Joseph Widener in 1942.)

Today, Vermeer is not just admired but adored. His life, about which little is known, is the subject of best-selling novels and movies. But the paintings themselves float above the noise and the hype. Incredibly quiet, exquisitely colored, breathtakingly intimate, they stand as a rebuke to the noise and mayhem of modern life and a salve to harried, information-age sensibilities.

Ten Vermeers, but a whole world of Dutch art

With time and space in the lab, the NGA’s researchers, led on the science side by senior imaging scientist John Delaney, subjected the paintings to sophisticated imaging. They were building on a rich history of Vermeer research at the NGA, notably by Melanie Gifford, now-retired research conservator of painting technologies. It wasn’t clear at the outset that they would come up with anything new.

But what resulted, according to Wieseman, was an “exponential increase in our understanding of Vermeer’s working process.” That leap in knowledge, she said, “enabled us to determine that [‘Girl With a Flute’] is not by Vermeer.”

Gifford had analyzed minuscule samples taken from the NGA’s Vermeers, so there was already a lot of data about the paintings, according to Delaney. Now, a combination of microscopic analysis and advanced imaging allowed Delaney and his fellow imaging scientist, Kathryn Dooley, to map the materials Vermeer had used. The techniques included X-ray fluorescence imaging spectroscopy and reflectance hyperspectral imaging, which uses a light-dispersing spectrometer to collect and process information from across the electromagnetic spectrum.

Visitors to a new NGA exhibition, “Vermeer’s Secrets” (Oct. 8-Jan. 8), can see some of what the research team uncovered before the works are sent to the largest-ever Vermeer retrospective at the Rijksmuseum in Amsterdam (Feb. 10-June 4). The display includes the NGA’s four Vermeer paintings (now three) and two 20th-century forgeries that are still in the gallery’s collection. (How these grotesque parodies were ever taken seriously as Vermeers is difficult to say.)

The research team, which also included Alexandra Libby, Dina Anchin, Lisha Deming Glinsman and Gifford, began by looking at the two masterpieces whose attribution to Vermeer has never been in question. Studying “A Lady Writing” and “Woman Holding a Balance” first, said Delaney, was “a great way to establish a baseline for his practice.”

Among the discoveries was that Vermeer was more vigorous in parts of his process than previously thought. He brushed on his first layers with surprising speed and freedom, at one point even applying a layer of copper-containing material known to hasten the drying process, as if he were in a hurry to move on to the next stage.

“We have this impression of Vermeer being the master of these smooth, satiny surfaces, where you can’t identify individual brushstrokes,” Wieseman said. “But then you look at how he set up that glow on the background wall [depicted in “Woman Holding a Balance”] and it’s exciting, vigorous brushwork. You get a sense of the artist really going at it.”

Lucian Freud captured intimacy, even when his subjects were clothed

The research team then turned to the two smaller, more problematic works, “Girl With the Red Hat” and “Girl With a Flute.” The two paintings have long been considered a pair. Both are “tronies” — the Dutch term for paintings of heads that were not portraits of specific people, but studies of types, often idealized or particularly expressive. (Vermeer’s “Girl With a Pearl Earring” is the most famous example.)

There were two main takeaways: “Girl With a Flute” was made by an artist — perhaps a student, an apprentice in training or an amateur taking lessons from the master — who, in Delaney’s words, “understands the technique but has very limited skill in executing it.”

The research team also concluded that Vermeer probably painted “Girl With the Red Hat” a couple of years later than previously thought, in a period — 1669 rather than 1666-1667 — when he was experimenting with new colors and slightly bolder paint application.

The NGA’s tronies both show young women with similar faces and expressions. Both subjects wear unusual hats and large pearl earrings. The backgrounds of both are rather summarily sketched in. Both show a tapestry on the wall and a chair with lion’s-head finials. And both are painted on wooden panels, which was extremely unusual for Vermeer.

Despite all that, scholars have long doubted whether Vermeer painted “Girl With a Flute.” It just didn’t look good enough. The transitions from light to dark, especially around the face, looked awkward and abrupt. The green shadows were heavily applied, creating what the “Vermeer’s Secrets” wall label calls “a blotchy appearance under the nose and along the jawline.”

In the 1990s, NGA curator Arthur Wheelock, an acknowledged Vermeer expert and recently retired, had “Girl With a Flute” designated as “attributed to Vermeer.” That designation, said Wieseman, was Wheelock’s “way of explaining why it generically looks like Vermeer but qualitatively doesn’t come up to the standard.”

Most scholars agreed, although Wheelock’s colleague at the Metropolitan Museum of Art, the late Walter Liedtke, maintained that it probably was a Vermeer, and Wheelock himself later altered his position, saying, “I have concluded that removing “Girl With a Flute” from Vermeer’s oeuvre was too extreme given the complex conservation issues surrounding this image.” (Abrasions to the painting’s surface had made it especially difficult to study.)

The new analyses seem to have confirmed the doubters. “At pretty much every level in the buildup of the painting,” said Wieseman, “it’s ‘close, but no cigar.’ ”

The research team uncovered that, although some of the same materials are present in both paintings (as Gifford had previously established), the paint handling is very different. Where the technique in “Girl With the Red Hat” is subtle and adroit, the application of paint to “Girl With a Flute” is relatively clumsy and coarse.

Instead of using coarsely ground pigments in the underlayers and finely ground pigments for the final layers (as Vermeer did), whoever painted “Girl With a Flute” did the opposite, giving the surface a granular quality. There are even fragments of bristles in the painting’s surface layers, suggesting that the artist was using an old or poorly made brush.

“The artist has a conceptual understanding of how Vermeer built up his paintings but just can’t manage the finesse,” Wieseman said.

There are also defects in the underpaint. For instance, in some of the blue areas, there are “traction crackles” indicating that the surface paint dried before the underlayers. “An experienced artist would know how to mix their pigments so that didn’t happen,” said Wieseman.

Similarly, in areas where white pigment was applied, the artist used too much medium (oil) in the underlayers, causing it to dry in a wrinkled fashion. The artist had to scrape that wrinkling down to get a smoother surface to apply the final layer of paint.

“These are beginner’s mistakes,” Wieseman said. “Vermeer knows why he’s doing things. He knows what the end result will be, whereas with this artist, you just don’t have that sense of understanding.”

If all this is true, it alters our understanding of Vermeer, who has long been considered a lone wolf working without assistants or students. The question becomes: Who was this artist who had access to Vermeer’s studio and used many of the same materials? And what might one day be discovered about their relationship?

The new findings are revelatory, but there will always be an air of enigma around Vermeer.