The dinosaur floor at the American Museum of Natural History is one of my favorite places to spend an afternoon. It feels like walking around a fantasy made real—this idea that our Earth, which is now ruled by small hairless apes, used to be the domain of monsters—giant, clawed beasts with teeth the size of my hand.

And among all this wonder, there's also a mystery. Where did they all go? They evolved and evolved, and then at the end of the Cretaceous period, they just disappeared. And no one's really sure why.

One of the most popular theories has to do with comets hitting the Earth, causing giant impact craters and wiping most living things off the planet.

"That was the big news back in the ’80s," says Michael Rampino, a professor of earth sciences at New York University, who was studying such craters at the time. Back then a research paper came out suggesting there was about a 30-million-year cycle for mass extinctions, like the one that killed the dinosaurs.

"And that got a lot of attention," says Rampino. "And I was working at NASA at that time with a friend down the hall who was an astronomer. And he said, 'You know there's an astronomical cycle that's also about 30 years.' So we started talking about a model in which the astronomy, the motion of the sun and the planets through the galaxy could trigger these effects on life."

Three decades later, Rampino's published his fullest version yet of that model. It links extinctions on Earth, with dark matter in the galaxy. Basically, he's saying one mystery of science might explain another.

There's a number of things you have to buy into to follow Rampino's dark matter and extinctions theory.

First you have to get that our Milky Way galaxy is shaped like a giant pancake. And our solar system revolves around the outer edge of it.

"We go up and down like a horse in a carousel," explains Rampino. "Up and down through the plane of the galaxy as we go around, and we pass through the plane about every 30 million years."

We're passing through the plane of the galaxy every 30 million years, and it's been suggested mass extinctions happen every 30 million years, and, some scientists say these big collisions with comets happen about every 30 million years, too. Rampino started to think about how these things all happening on about a similar cycle might be connected.

"It seemed like too much of a coincidence to just be a coincidence," he said.

So there're these strangely similar cycles. And then there's the other thing you need to believe in: dark matter.

When astronomers look out into space, they see galaxies moving in a way that, based on the galaxies' apparent mass, would break the laws of physics. So scientists theorize that there must be some extra mass out there in the galaxies, matter that human eyes can't see, i.e. "dark matter."

And when our solar system passes through the dense plane of the galaxy every 30 million years, Rampino thinks it bumps into a lot of dark matter. This disrupts the orbit of comets in our solar system, his theory goes, making some of them come crashing into Earth.

"Like apples falling out of a tree," Rampino explains.

And he thinks dark matter might trigger more than just these comet collisions. He argues that when the solar system crosses the galaxy pancake, dark matter particles collide with the Earth.

"They interact a little bit," he says, "they lose energy, and they fall down toward the core of the Earth. And when the concentrations of these things get high enough, they actually annihilate each other. And that produces a lot of heat."

That extra heat in the core could cause volcanism—another possible trigger for mass extinctions.

"That's the interesting thing," Rampino says, "because people have been debating for some time whether it's impacts that cause extinctions or volcanism—these big flood basalt volcanoes. And this would kind of take both of those things and say they both may be right. Since they're happening at the same time, because they occur when we're passing through the disc of the galaxy."

And while Rampino's theory affirms both sides in the volcanism-versus-comet-impacts debates, it also provokes a lot of skepticism. He expects people to tell him he's wrong.

"We don't know," he admits, "the extinction may not be periodic, the craters might not be periodic, the period may not be 30 million years, the dark matter may not be there in the way I'd like it to be. So all these pieces of the puzzle are controversial. But that's good. That means you're doing things on the edge on the frontier of the science."

Rampino says what needs to happen next is more testing. Better dating of comet craters and volcanoes. And we need to have a better sense of how dark matter actually works.

But he says even if his idea's proved wrong, just having set up a framework for how we might connect geologic events with astronomy, that's worth it to him.