Throughout Earth's history, one question rises, again and again: what were the factors that killed off a species or a population of animals?

Many of them are terrestrial in origin, but a mysterious event that killed off 36 percent of the ocean's genera 2.6 million years ago in the Pliocene, including the famous giant, bus-sized shark Megalodon, could have been from elsewhere: an exploding star, or supernova, just 150 light-years from Earth.

In fact, it may even have been more than one, bombarding the planet with cosmic radiation that could have triggered climate change, as well as other effects that contributed to a mass extinction of the oceanic megafauna.

"I've been doing research like this for about 15 years, and always in the past it's been based on what we know generally about the universe - that these supernovae should have affected Earth at some time or another," said physicist Adrian Melott of the University of Kansas.

"This time, it's different. We have evidence of nearby events at a specific time. We know about how far away they were, so we can actually compute how that would have affected Earth and compare it to what we know about what happened at that time - it's much more specific."

And the proof is in the pudding - if the pudding is deposits of isotopes of iron-60 found in the seabed.

Iron-60 is radioactive, with a half-life of about 2.6 million years. This means that any iron-60 that may have formed with Earth 4.54 billion years ago would have decayed long, long ago. Whatever iron-60 remains on the planet today has to have come from elsewhere - blasted through space by a supernova.

Radioactive isotopes are relatively easy to date, which means we know roughly how old they are. There's a big spike about 2.6 million years ago, Melott said, and a smattering of small events dating back to around 10 million years ago.

This, the researchers said, is supported by the Local Bubble, a relatively gas-free cavity of space about 300 light-years across through which the Solar System is currently moving.

"It's basically very hot, very low-density gas - nearly all the gas clouds have been swept out of it," Melott said. "The best way to manufacture a bubble like that is [if] a whole bunch of supernovae blows it bigger and bigger, and that seems to fit well with idea of a chain."

And that has an effect, too. Rather than just continuing out through space, at least some cosmic rays blasted out by a supernova would reflect off the bubble's boundary, resulting in an ongoing bombardment that could last up to 100,000 years. With a chain of supernovae, this could go on much longer.

The idea that supernovae could trigger mass extinctions is not a new one. It's long been thought that a gamma-ray burst originating in a supernova could have caused the Ordovician extinction 450 million years ago, its radiation stripping the ozone layer and exposing life on Earth to the Sun's deadly UV light.

But gamma-ray bursts within the Milky Way are rare. So rare, in fact, that we've never actually seen one (although we might if we're lucky - or unlucky, if the Ordovician hypothesis is correct).

According to the work of Melott and his team, it was an entirely different kind of mechanism that behind the Pliocene marine megafauna extinction event. Instead off gamma-ray bursts, it was a type of cosmic-ray elementary particle called muons - a bit like electrons, but with much more mass and more energy.

"They're very penetrating. Even normally, there are lots of them passing through us. Nearly all of them pass through harmlessly, yet about one-fifth of our radiation dose comes by muons," Melott explained.

"But when this wave of cosmic rays hits, multiply those muons by a few hundred. Only a small fraction of them will interact in any way, but when the number is so large and their energy so high, you get increased mutations and cancer - these would be the main biological effects.

"We estimated the cancer rate would go up about 50 percent for something the size of a human - and the bigger you are, the worse it is. For an elephant or a whale, the radiation dose goes way up."

And because muons are so penetrating - they're even detected deep underground - they'd be able to penetrate the ocean, reaching the marine life therein. Of course, greater depths would have lower muon penetration, and that's consistent with the extinction event. Shallower coastal waters experienced the event much more severely.

Of course, we can't know for sure, and there are holes in the idea. Melott mentioned elephants and whales as an example of a larger animal that should, in theory, experience higher rates of cancer - yet their cancer rate is bafflingly low. Whales, too, get cancer at a much lower rate than humans. This phenomenon is called Peto's Paradox, and hasn't been addressed in the team's paper.

But since we don't yet have an explanation for what caused that extinction event, supernova radiation could be as interesting a place to look as any - even if it turns out that it was climate change alone that killed the beasts.

The team's research has been published in the journal Astrobiology.