Devon Island, the bulk of which straddles 75 degrees North latitude, well above the Arctic Circle, is a geologic wonderland. Today it is mostly polar desert. But it wasn't always so.

Its tale begins about 2.5 billion years ago.

"The oldest rock in this area are rocks of what you call the Canadian Shield," Pascal Lee, the director of the Haughton-Mars Project (HMP), told me. We were sitting inside the greenhouse at the project's research station on Devon Island, sweating, in t-shirts.

The Canadian Shield, he went on to explain, "is the original chunk of continental material that was amassed to form the bulk of North America, and Canada in particular."

Back then, when Earth's continents were being born, there was no Devon Island. What at present is the island's most deeply buried layer, was then part of the continent Arctica, the ancient foundation of today's Arctic landmasses.

These 2.5-billion-year-old rocks are exposed in places on Devon Island: on the east of the island as a result of uplift and tilting; in other locations, deep within canyons cut by glaciers. But hold off on the glaciers. They come later.

After the 2.5-billion-year old rocks, there's a big gap. The island's story picks up again around 600 million years ago and extends to the time of impact that formed Haughton Crater, 39 million years ago.

"What we are standing on today," Lee said, referring to the terrain visible in the vicinity of the Haughton-Mars Project Research Center (HMPRS), "is the cumulative product of layer upon layer of ancient marine sediments that were deposited in seas of different depths."

You can tell that because the rocks are full of fossils. Ammonites. Other little shelly things. From back when shells were a recent invention. On an island in the Arctic.

The fossils cover a long span of time. "You're talking about a sequence that goes from the late pre-Cambrian to the Devonian," he said. That's a very long span of time: 250 million years; say, 600 million years ago to 350 million years ago.

In other words, the fossils on Devon Island, which have not yet been extensively studied, are a rough catalogue of the first half of the existence of animal life on Earth.

Of course there are pieces missing. That's the way geology works. Rocks form, rocks get worn down, sediments get deposited, deposits get washed away. Only under certain conditions do deposits of new materials build up. And only those deposits that don't subsequently get eroded away remain visible.

Narrowing his focus a bit, Lee informed me that at the site of HMPRS, "we're at the boundary between the Ordovician and the Silurian." So "we're standing on rocks" he delivers this casually - that are "roughly 450 million years old."

"In the Silurian, we start seeing fish, but they're at the beginning, where they're still barely showing bones and jaws and they're at first small, very small."

The beginning of fish. As in: before that Earth had no fish; after that it did.

Lee concedes that he has "yet to see a fossil of anything that's like a fish here." But he adds, "There are other fossil indicators that tell us we are near the beginning of the Silurian."

Meanwhile, back in the late Ordovician, just before the Silurian, I now learn, the land that today is Devon Island was under shallow water. "Right here, we had a coral reef."

Oh, okay. So we're not just hanging out - a stone's throw from the North Pole - on any old marine sediments. We're on a coral reef. "The rocks that you're standing on here are reef."

By way of illustration, Lee points to two small mounds that lie about 100 yards from the cluster of buildings that make up the "downtown" area of HMPRS. One of the mounds is littered with fossils of ancient coral. Once Lee points out what fossilized coral looks like, you can't miss it.

Of course, coral reefs don't as a rule form above the Arctic Circle. But over time, the landmass that today is Devon Island has moved. Land on planets with plate tectonics has a tendency to do that. Devon used to lie farther south.

Exactly how far south is hard to say because, "the climate of the Earth was different at the time." Tropical conditions have at times extended farther toward the poles than they do today.

Since its earlier stint as a coral reef, "Devon Island has been making its way farther north." So, a quick summary. The oldest rock on Devon Island formed 2.5 billion years ago. The land where the Haughton Mars Project Research Station is sited, the former coral reef, dates to about 450 million years ago.

Now fast forward to the Eocene: a mere 39 million years ago. "All of the Earth was substantially warmer than it is today," Lee said. "What is now Devon Island was still a bit farther south, but not that much farther south." And: "There were woods here."

No more coral reef. We're not underwater any more. Now we're in the forest.

But don't get too attached to the forest. Because then came the impact. A massive object - no-one is exactly sure what it was - hurtling towards Earth, aimed squarely at Devon Island. It was big. Not finish-off-the-dinosaurs ginormous. But big enough that within about 10 seconds of striking its target, it had carved out a hole that the entire city of San Francisco, the Golden Gate and Bay Bridges, and bits of Oakland, southern Marin County and northern San Mateo County could fit inside. (See map.) Big.

The impact shattered the land. Lying around on the floor of the crater are shock cones, rocks that display radial lines in a pattern formed only when large impactors slam into the Earth.

The collision also pumped such intense heat into the ground that, in some places, rock melted, while in others hydrothermal vents suddenly sprang into existence. A massive umbrella of boulders ripped from the ground - geologists call it an ejecta blanket - was thrown upward and outward, crashing back to Earth to form a glowing ring around the crater's rim.

Remnants of the impact melt, the vents and the ejecta blanket are on display today, as though they were part of a large outdoor geology museum. Devon Island, being both cold and dry, has preserved them well.

By the way, scientists initially thought the crater formed 23 million years ago, because lake sediments inside the crater have been dated to that period.

"There was a little rhinoceros running around" back then, Lee said, "whose bone remains were found here." A primitive-seal-like animal, Puijila darwini, a previously missing link between land mammals and marine mammals, discovered on Devon Island in 2007, also hails from around that time.

But subsequent analysis revealed that the crater had formed some 16 million years earlier and that the lake sediments, along with the small rhinoceros and the missing link to marine mammals, were more-recent additions to the island's landscape.

After that - yes, there's more - Earth's climate cooled and Devon was buried under ice. The most recent ice age "reached its maximum at about 12,000 years ago," Lee said. As recently as 10,000 years ago there was "probably no visible land" on Devon Island.

As the glaciers spread, they carved canyons. When they retreated, the canyons were exposed. They are still visible today. Lee believes these canyons explain the formation of valley networks on ancient Mars.

Some scientists who study Mars point to the presence of its valley networks as evidence that Mars was once warm and wet. Based on what he's seen on Devon Island, though, Lee argues that Mars has always been cold.

"Imagine Devon Island has an ice cap," he said. "Most of it is just sitting on the ground, but locally, the ice was streaming along ... gouging the land beneath it ... grinding and breaking the rock away as it went."

Canyons carved by this process, under ice, "are called glacial trough valleys," Lee said. They look different than canyons, like the Grand Canyon, cut by "a river slowly meandering and eating its way into rock over time." Glacial trough valleys don't simply branch, like a tree. They spread and rejoin themselves, spread again, and rejoin again.

Canyons on Devon Island take this form. So do the valley networks on Mars. "The profiles of the canyons that end up happening here are mirror images of what you see on Mars," he said.

If liquid water played a role in forming Mars' canyons, Lee suggested, it wasn't water flowing out in the open, on the surface, it was water flowing underneath ice. That can happen even when the atmosphere is freezing cold.

"We often compare Valles Marineris and the other canyons on Mars to the Grand Canyon." Valles Marineris is huge. "But the comparison really just stops at ... the fact that it's a big hole in the ground."

The canyons on Devon Island, he said, canyons carved by ice, their shapes preserved to the present day by the island's climate, show you what took place on ancient Mars.

That's past work, though. Lee no longer has time to do much geology on Devon Island.

These days, when he and his colleagues return each summer to the Haughton-Mars Project Research Station, their focus is largely on practicing for missions of exploration to places you need a rocket ship, instead of a Twin Otter, to get to.

Testing out robotic arms, for instance. Or drills. Or spacesuits. Or growing food in a greenhouse by remote control. Or driving across the island in a Humvee, simulating the experience of exploring an airless world in a pressurized rover.

But every once in a while, when he doesn't know you're watching, you can catch him wandering about, eyeing the landscape, or picking up a rock and turning it over in his hand, wondering what unread chapters of Devon Island's story are still waiting to be told.