A Diplodocus is being assembled. Ed Yong

As I walk into the Smithsonian National Museum of Natural History’s partially reassembled dinosaur hall, the first thing I notice is a Diplodocus peeking out at me. Its long neck cranes out from behind a cutout of a Jurassic tree, and its 150-million-year-old skull is angled in an almost cheeky way. “Oh hi,” it seems to say. “It’s you. Welcome.”

The famous 107-year-old hall closed down in 2014 for a five-year renovation, and is set to reopen in June 2019. When I visited it a month ago, it was cavernous and mostly empty. Now it’s a hive of activity, and a labyrinth of boxes, ladders, and paint pots. Stacks of ancient bones and steel brackets dot the space. Three dozen crates sit in a corner with “T-REX” stamped on them in a bold, aggressive font.

This is the fun bit, when the hall starts coming to life—when long-imagined exhibits are finally constructed, and when the museum’s specimens come home to be reassembled. A mammoth, a mastodon, and a giant ground sloth are already complete. The enormous, long-necked sauropod dinosaurs, including the Diplodocus, are next in line.

A set of crates with T-REX stamped on their sides.
Ed Yong

When I walk under the dinosaur’s neck and around the tree, I find that the animal is very much a work in progress: The skeleton is missing its front legs, ribcage, and shoulders. This is fairly typical. When assembling a dinosaur, you almost always start by building up one hind leg and down the other, before extending forward to the head and backwards to the tail. “Everything flows from the pelvis,” says Siobhan Starrs, who works at the museum and has been managing the renovation. It provides stability, and is often the heaviest bone; a tyrannosaur pelvis can weigh upwards of 1,000 pounds.

The bones aren’t stuck to each other. Instead, they’re assembled around a steel armature that provides support for the whole specimen—a metal skeleton for the actual skeleton. For the Diplodocus, two vertical columns will run up between each pair of legs, while a sinuous horizontal bar will hold up the neck, spine, and whiplike tail. So far, six vertebrae from the tail are in place. Four men from the dinosaur-mounting company Research Casting International (RCI) are getting ready to attach the seventh. Their shirts say SKELETON CREW on the back, under pictures of skeletal cowboys riding rodeo on dinosaurs.

The fossilized tail bone looks like a stony leg of ham, and weighs around 70 pounds. It sits on its own custom-made steel bracket, which has been locked into a lifting harness. The bone is both heavy and irreplaceable, so the men spend a long time testing it to make sure that it’s secure. Once they’re confident, they winch it up and carefully maneuver it into place. They bolt the bracket to the armature, and remove the harness. Seven down, dozens to go. A Diplodocus tail has around 80 bones in it, and each will take around 40 minutes to assemble.

The feet go on last, so for most of their construction the skeletons look like they’re levitating just off the ground. “When you’re walking around, [the feet] get in your way,” says Matt Fair, a production manager at RCI.

Fair, like many of the people at RCI, possesses an unusual blend of fine-arts expertise, trade skills, and paleontological knowledge. He has an intuitive sense for fossils and how strong or fragile they are, given the species they come from, their anatomical location, and the conditions in which they were created. “I’ve dealt with some stuff where if you breathe on it, it cracks,” he says. “And ribs are very fragile. I don’t have my hard hat on now because I was just up inside that ribcage.”

The ribcage he’s referring to belongs to a Camarasaurus, another giant sauropod that’s standing opposite the Diplodocus on the other side of the tree. In the old hall, it was lying on the floor in a death pose, and visitors frequently overlooked it. They’ll have a hard time doing so now, as the animal rears up dramatically on its hind legs, with its neck stretching from 20 to 25 feet into the air.

A Camarasaurus fossil being assembled.
Ed Yong

Re-posing a dinosaur is difficult work. In the Camarasaurus’s previous position, most of the Camarasaurus’s bones were still encased in exceptionally hard rock, and had to be liberated—a common problem for many of the museum’s old fossils. Some of the vertebrae had become warped and contorted over geological time, which meant that the neck had to arch in a specific way. I suggest that it’s a bit like ordering IKEA furniture, finding that the pieces don’t quite fit, and having to jimmy them into place.

“It’s a little more complicated than IKEA furniture,” Fair says, bristling slightly.

“And there’s no instruction manual,” Starrs says.

“There’s an ongoing argument about whether these animals could get up on hind legs,” says Matt Carrano, the museum’s dinosaur curator. “But they would have had to, at least once, if they wanted to make another generation of sauropods.”

Carrano consulted on all the specimens’ new poses to infuse them with both biomechanical plausibility, and some much-needed drama. Previously, most of them were just standing in static poses that reflected the old view of dinosaurs as ponderous creatures. The new hall will portray them as actual animals. “They’ll be performing,” Carrano says. “It’s like they’re on a nature show, and they’re doing the thing that the cameraperson waited 15 years to catch.”

For the Camarasaurus, that means craning toward high leaves. The Tyrannosaurus will be munching on a Triceratops. And the Allosaurus—another large, two-legged meat eater—will be sitting and caring for a nest of eggs and hatchlings. That’s a rather revolutionary posture for an animal that is often seen as nothing more than a slavering predator. “It’s not the super-obvious thing to do,” Carrano says, “but it was important not to have every animal biting another animal.”

“No one’s pooping, but we do have a lot of fossil poop,” he adds. It’ll go in a special display, next to the restrooms.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.