The shifting face of a 200-million-year-old mystery
Five times in the last half a billion years, tremendous, global-scale extinctions have wiped out a significant fraction of life on Earth - and each of them presents a grand puzzle.
The most recent and the most familiar is the extinction that wiped out the dinosaurs - between the Cretaceous and Tertiary periods, about 65 million years ago.
But before that, 205 million years ago, was the "End-Triassic Event" - it set the stage for the Jurassic Period, which saw the rise to prominence of the dinosaurs.
Just what happened that killed off half the species on the planet, though, remains a mystery.
On the coast of Somerset in the UK this week, one researcher from the US is trying to peel away - actually, to hammer away - some of that mystery, and in so doing providing a picture of science at work.
Once upon a time, Paul Olsen of Columbia University's Lamont-Doherty Earth Observatory was one of the few people who believed that a giant impact of an object from space started the extinction, much like the impact that many people believe caused the demise of the dinosaurs.
"I argued quite passionately for that on the basis of the pattern of extinction, which looks very much like that at the Cretaceous/Tertiary boundary, and because I discovered an 'iridium anomaly' at the level of the extinctions.
"Iridium is very rare in the earth's crust but relatively abundant in extraterrestrial material, so an abundance of that element suggests that there was an impact at that time," he explained.
"I thought it was pretty clear that that was a plausible cause."
But science has moved on. Other workers in the field have found in that same period, enormous lava flows from a flurry of volcanic activity in the then-developing Atlantic Ocean also occurred at exactly the same time.
And Prof Olsen himself found more iridium anomalies at different, nearby times.
"The way many scientists work is that while they're pushing one idea passionately, they always have in the back of their mind that they may be wrong, and they have alternative explanations for the same observations - and I did too.
"It wasn't so simple to explain the extinctions as a result of a giant impact, but simpler to explain as a result of the giant lava flows," he said.
So it is that Prof Olsen is at St Audrie's Bay in Somerset this week, where rocks of the right age are pristinely preserved.
What is at issue is the precise timings of the events. Prof Olsen and his colleagues believe that an impact did occur, probably at Rochechouart in south-western France, but that it occurred a few thousand years before the extinction.
While it may have had an effect on the life on Earth, the impact does not seem to line up in time with the volcanic activity and the extinction.
But, as is so often the case with the record of events laid out like the pages of a book on the Earth's surface, more data is needed, and Prof Olsen is hammering away chunks of it, carrying the samples back to the US for analysis in neatly labelled bags.
Gareth Collins, an impact expert at Imperial College London, says that anyone trying to unravel the details of extinction events, "if they're honest with themselves, would say that we're never going to know".
Just which species died off and disrupted ecosystems, and how stressed various species may have been before dying off, are simply too tricky to figure out from the fossil record.
In contrast to the guiding principle of science, any guesses do not lead to testable predictions.
"I think we're still some way off understanding exactly what was the cause of the mass extinction, and the more general question of whether mass extinctions have a common cause or whether each one is an event that needs to be considered in isolation," he told BBC News.
The extinction at the Cretaceous/Tertiary boundary, for example, again shows a great deal of volcanic activity, and irrefutable evidence for a giant impact.
Might it be that all of the "big five" extinctions involved both a disrupted climate due to volcanic emissions, as well as something smashing into the Earth? These mysteries will endure for some time.
But Dr Collins argues that the efforts to unravel them must press on.
"I think it's incredibly imperative that we understand mass extinction events; they punctuate evolution and they are absolutely as important if not more important than the sort of incremental changes that occur between these extinctions.
"They sort of wipe the slate clean and then allow the next era to occur and the next dominant species to evolve."
So as Prof Olsen hammers layers of 200-million-year-old shale out of the Somerset foreshore, he is keeping an open mind about what he and others will find.
"When you find evidence that directly contradicts your favourite idea and you have to switch modes, switch paradigms to a different concept, that's real progress," he said.