One day in 17th century London -- probably -- a painter stepped into a studio with a palette full of pigments, a canvas 7 feet high and a naked woman. Maybe she was his mistress, maybe not.
Either way, the result was "Andromeda Chained to the Rock," credited for more than 150 years to the Flemish master Anthony Van Dyck. The Ahmanson Foundation acquired it in 1985 as a 20th birthday present to the Los Angeles County Museum of Art, which put it up right away.
"It sends chills up and down my spine," said Scott Schaefer, then LACMA's curator of European paintings, welcoming the work to the collection.
But these days, "Andromeda" is all but invisible. Although it probably cost about $1 million, it hasn't been hung in a public area for several years, and the museum has never announced a reason. The answer is there, however, for those who dig into LACMA's online collection database: In July 1998, the museum decided it wasn't a Van Dyck after all.
"It's a casualty of art history," says J. Patrice Marandel, who took over as LACMA's curator of European paintings in 1993. "We're not hiding it. We're not ashamed of it. Perhaps we're sorry, but I wouldn't even go that far."
In the art world, this is called a "re-attribution" -- sometimes a painful event, sometimes a happy discovery, often a test of institutional candor. The story of how this "Andromeda" fell from grace is a lesson in how curatorial wheels grind when names such as Rembrandt, Rubens and Van Dyck are at stake, and a reminder that for all the technological advances of the last century, many of the most important questions in art history are still a matter of argument among experts squinting at old brush strokes.
"I still, quite frankly, believe in this picture," says Schaefer, who is now curator of paintings at the J. Paul Getty Museum. Whether you embrace "Andromeda" or reject her, he says, "depends on experience, emotion, feelings, hunches, impressions -- all the very things that are impossible to quantify."
Marandel leads a visitor down into LACMA's dim storerooms and plays a flashlight beam across "Andromeda." Set in a massive and ornate golden frame, the painting seems larger than 4 feet by 7 feet, and Andromeda looks like an Amazon.
"This is not a fake. It's a painting of the period, a painting that has some quality," Marandel says. "But the painting, when you start looking at it, is full of these awkward moments.... That body is, to me, awkward. The head is hopelessly small, compared to the huge vastness of torso, which is rather unpleasant."
And the belly button, which is placed dead in the middle of the composition, "is the biggest belly button in the history of belly buttons," Marandel says.
Marandel and Schaefer, believe it or not, are friendly. It's mysteries like this, both say, that offer a big part of the joy in their jobs. No major museum ever stops studying its own collection, especially when a new curator arrives, and every new possibility is fuel for daily drama that museum visitors might never notice.
Still, it's more fun to find a Van Dyck than lose one. For this and other reasons, the way museums handle attribution questions varies widely -- sometimes outright refusals to acknowledge doubters, sometimes public explorations of open questions.
The Amsterdam-based Rembrandt Research Project, a group of scholars convened by the Dutch government in 1968, has suggested that half of the paintings attributed to Rembrandt, including one apparent self-portrait at the Norton Simon Museum in Pasadena, were actually by Rembrandt's students and others. The project's findings are disputed by many museums, including the Norton Simon, and sometimes complicated by internal disagreements as well. New York's Metropolitan Museum of Art in 1995 staged a "Rembrandt / Not Rembrandt" exhibition.
In local history, the best-known re-attribution may be the Getty kouros, an arresting marble sculpture of a young man bought by the Getty in 1985 for an undisclosed amount estimated at about $10 million. The Swiss seller of the piece said it had come from a private collection after first appearing on the market in 1930, but many experts voiced doubts, saying its condition was suspiciously good and that it seemed to incorporate stylistic elements from several regions, perhaps a hint that a forger had combined elements from other surviving works of the era.
With questions lingering, the museum in 1992 shipped the work to Athens and sponsored a colloquium on the subject and published the results -- still inconclusive -- the following year. Today the Getty Villa displays the work with text saying, "Greek, about 530 B.C. or modern forgery." But in many other disputed cases, there's no modern skulduggery suspected, just murky history.