In a few weeks, Andrew Robison faces a difficult choice that only he can make. A festival in the Piazzetta in Venice, captured in a stellar 18th-century drawing by Canaletto, hangs on a wall of the National Gallery. It also hangs in the balance, as Robison must consider the highest of curatorial stakes: to let this masterwork survive or risk it being destroyed.

Of course Robison wants it to survive. He was the one who recently guided the museum to buy it. As the curator of prints, watercolors, drawings and rare illustrated books, he cares deeply about protecting all 106,000 works in his domain.

But in an emergency, not everything can be saved, and so he carefully ranks which works should be spared. The Canaletto is an obvious candidate for his top-priority list of 74 works on paper, but if it is included, something else has got to go.

In 1979, with Washington worried about 52 hostages in Tehran and terrorist threats at home, Robison’s boss asked him to create a big container for works of the highest value. If catastrophe hit, the container could be spirited away to an undisclosed location. Today, Robison has seven boxes in two separate storerooms — four for European holdings, three for American. These do not include the museum’s 10,000 photographs, 3,800 paintings and 2,900 sculptures, outside of Robison’s purview and mostly too big for any mad dash out the building. And because his works are so fragile and light-sensitive, they live most of their lives in protective storage, going on the walls for viewing only in short spurts.

In the two storerooms that Robison asked not be photographed or their locations disclosed, the black, cloth-lined boxes, each the shape of very large books, bear the label “WW3,” drawn in calligraphy. These in-case-of-World-War-III containers lie ready for any possibility, and in Robison’s absence, security guards have a floor plan that shows their exact location, like an X on a pirate map.

The National Gallery’s director, Earl “Rusty” Powell III, said that Robison’s box system is not the primary way the nation’s treasures are protected, but offers little elaboration: “We’ve got a very, very fine security program.”

But Robison, who trained in philosophy and theology at Prince­ton and Oxford universities, is devoted to the task he’s been given. It helps him focus on what’s important in each work and in the entire collection.

The box system is thus less a model for how museums should think about war and more a model for how a curator thinks about value. “I think it’s pretty unique to Andrew,” Powell said of the system. “It’s mainly a philosophical one.”

Power rankings

Born 71 years ago in Memphis, Robison is neither a rogue nor a prig, the two usual poles of cultural debate. He enjoys rankings in an era when art historians sneer at anything canonical. However conservative that makes him, he’s entirely self-possessed, open to new forms and ideas. He can talk as movingly about the louche John Currin of today as he can about the satanic flourishes of a late-15th-century German engraver with the rapper-like handle of Master LCz. After the cheese course in the National Gallery refectory, Robison casually rolled some red grapes in his hand like dice and popped a few in his mouth.

Showing off a box’s contents, he growled lusty approval of Peter Paul Rubens’s huge reclining Pan, which Robison counts among the most powerful in the WW3 boxes. He admired beautiful women in the works of art as robustly as Rubens did when he sketched a model for a figure in “The Assumption of the Virgin” — “Clearly he was in love with this woman,” Robison said, holding up the black chalk sketch.

Though he must know prices and trends, the art market, he said, cannot be the arbiter of what gets included in the WW3 boxes. “Money was just not a good guide,” he said. Picassos and Leonardos fetch huge sums that don’t always match their importance. Then, more recently, hedge funders have bought canvases the size of garage doors, inflating a contemporary genre that will likely falter.

To merit inclusion in the box, each work gets a thorough going-over by Robison’s team. The first criterion is aesthetic: Is it pleasing to the eye, well-made in both concept and execution? Next, historic: does it say enough about when it was made and who made it? Of all the moments of human history to which art can transport us, is this one worth remembering?

And then he has a more nebulous but convincing factor that Robison merely calls “power.” Of all the things that could be demonstrated with lines on paper, does this — through imagery alone — have a pronounced psychological impact? Does it change minds, just by viewing it?

“Very great quality will be more intriguing, more telling, more meaningful for visitors. It won’t be as meaningful if we had some sort of spread — one work from every great artist. It’d be more meaningful if we had really great works regardless of whether the artist is normally known as great or not.”

He compared his mode of thinking to that of other curators, like at the neighboring history museums on the Mall. “They’d be much more into, say, personalities,” Robison said. “If you’ve got Abraham Lincoln’s spectacles or George Washington’s false teeth, that becomes very intriguing. But I don’t want to demean it. . . . The original copy of the Declaration of Independence? Of course! It deserves its vault! I’m dealing with works of art. It’s a different thing.”

Brent Glass, the outgoing director of the National Museum of American History, explained that his curators do maintain lists of top priorities and do rely on red “disaster carts” with blotter paper and plastic wrap, in case of minor mishap if not catastrophe. And Cristián Samper, director of the National Museum of Natural History, admitted that no curator of anything precious wants to think about obliteration: “Most of us don’t want to know more than we know,” he said. And yet from a scientist’s viewpoint, the natural-history specimens they possess are valued for the data they yield, and digital recordings of the specimens might be seen as adequate.

A peek in the boxes

During Robison’s 38-year tenure, the works-on-paper collection more than doubled, from 50,000 to 106,000, and routinely his team found additions that pushed out other masterworks. In fact, only 27 percent of what Robison first put in the boxes in 1979 is still inside them.

When the debate resumes this fall, focused on the Canaletto, Robison will lead a handful of curators as they spread the boxes’ contents, in acid-free matboard, under protective tissue, across any available counterspace. Will Robison dump “Winter Landscape With Skaters”? It’s a nice little Dutch scene by a so-so artist, Hans Bol, who achieved what Robison called a “chilly, humid, vibrating atmosphere” that is exceptional. But is that enough?

Will anyone groan that Edvard Munch is hogging space in the lifeboat with his six multi-block woodcuts, revisions of an original composition over 25 years as the Norwegian saw new spectral possibilities for his seaside characters? Maybe the bravado of “A Satyr” by Benvenuto Cellini — himself a “bad boy” of the Renaissance — won’t stand up so well next to two other expressions of obvious self-confidence: the red chalk Rembrandt self-portrait or Picasso’s salute to himself, with gentle rouging of his cheeks and cravat.

The American and European works on paper are kept separate, physically and mentally. Robison would never let a grand European rarity bump out, say, the Winslow Homer watercolor of the farm girl cradling a chick. “It’s calendarish,” Robison admitted, but Homer once declared it his own favorite.

In a larger box nearby, a John Marin watercolor of New Mexico’s stormy mountains is deemed extra-special because it was donated by that doyenne of Taos herself, Georgia O’Keeffe. O’Keeffe’s blessing alone does not earn Marin his place in the boxes, along with two of her own. “It just turns out she was right,” Robison explained.

But whether he’d acknowledge it or not, as he looks around the American wares, Robison expressed more reservations than in the European room. As he looked at “Beach Scene,” by James McNeill Whistler, he appeared skeptical. It’s American impressionism, and a good example at that. But aren’t there better impressionist landscapes in America? Yes. Isn’t there a better Whistler in the box already? In fact, there are two.

“Hey, I’m not knocking it,” Robison said of “Beach Scene,” but then, under protest, he had to admit his own power. “Okay, I may be.”