During my morning sweep of tech feeds, a story popped up that looked like pure archaeological sensation: an international team found cave art in southwestern Ireland at least 15,000 years old — roughly five millennia older than any previous traces of humans on the island. Nice bonus for our era: the caves weren't found by someone with a shovel, but by a computer algorithm trained on characteristics of Welsh sites. The model spit out 28 candidates, people checked them on foot — and eight had art.
But the BBC headline was bolder than the find itself: "first known people in Ireland walked from Wales." And here I got suspicious. The thing is, "walked" implies a hypothesis of a glacial land bridge across the Irish Sea — and that hypothesis has long had an opponent in academic circles with a telling article title: "The Island of Ireland: Drowning the Myth of an Irish Land-Bridge?" Geologists Robin Edwards and Andrew Brooks concluded back in 2008 that Ireland separated from Britain around 16,000 years ago, and no later bridge carried flora and fauna to the island.
It was a classic fork: fresh sensation versus stubborn consensus, and both sides armed with numbers. I love these stories most — here we either find out who's right, or discover the question is trickier than both headlines suggest.
Let's start with the find, because it deserves its own story. Project coordinator is George Nash, archaeologist from the University of Liverpool; together with Genevieve von Petzinger (renowned researcher of signs in cave art) and Barbara Oostervyk he assembled the Lost Land Bridge Project. The team included specialists from University College Dublin, University of Adelaide, and scientists from Abu Dhabi, Wales, and Portugal.
The method was nontrivial. Nash spent years working with cave sites in South Wales — a region where modern humans lived from 35,000–40,000 years ago, seasonally inhabiting caves from Chepstow to St David's. He, along with speleologist Shane Difely and digital cartography consultant Darren Lee, loaded everything known about Welsh cave-site characteristics into an algorithm and sent the model to look for similar places in Ireland. Result — 28 candidates in southwestern Ireland, which the team checked personally.
From eight caves with art, the scientists described three techniques: engravings on limestone, painting with ochre and charcoal, and so-called finger flutings — grooves traced by three to four fingers through the soft clay crust on walls. There are figurative images: the front part of a deer's body (head, neck, side, about a meter long) and, Nash is convinced, a mammoth with a pronounced curved tusk. The team is keeping exact locations secret — they fear vandals and looters, and after the Lascaux caves story, that's reasonable.
The tastiest part of this story is the project's motto. "We just haven't searched enough and not in the right places," Nash told the BBC. "Absence of evidence is not evidence of absence." Behind this phrase — decades of professional frustration: in Wales hundreds of inhabited caves (on the Gower Peninsula alone about a hundred), and across the strait in Ireland — like a scorched sheet.
Either people really didn't go west, which for hunters following herds is implausible, or we looked poorly. The algorithm is a way to look systematically. And this is, perhaps, the first case in history where the eternal question "where to search?" was answered not by a veteran's intuition, but by a model trained on his experience — and both were searching in the same stone wasteland.
Now about what geology says, because "on foot" is primarily a physics question. During the last glacial maximum, around 26,000 years ago, giant ice sheets locked up so much water that the world ocean stood roughly 120 meters below modern level. Shelves were exposed: Britain and Ireland merged with the continent, herds of reindeer ran where ferries now sail. The question is only where exactly the land boundary ran and when it disappeared.
And here begins the boring but decisive mathematics. The thing is, sea level isn't the only player. Under the weight of an ice sheet, the earth's crust sags, and at its edges — bulges: this effect is called glacio-isostasy. On Ireland's coast, standing at the sheet's edge, crustal uplift partially compensated for the ocean's fall: water went down, but land rose up to meet it. Result: the deep Irish Channel between Wales and Ireland may never have dried out.
The Dublin reconstruction (Edwards and Brooks; Ireland is their longtime specialization) modeled this carefully: glacio-isostatic model with an updated British-Irish Ice Sheet component, calibrated against coastal geological markers, plus seafloor bathymetry. Conclusion: the last land bridge disappeared around 16,000 years ago, when the climate was still cold, and the model found "no support whatsoever" for a hypothesis of a bridge that later carried temperate species. Early reviews also found no grounds for a southern crossing: the Wales–Rosslare route was covered not by an isthmus but by the ice sheet's edge.
There is, however, a second candidate — the North Channel between Antrim and Scotland, shallower there. But it closes earlier, and it's on the opposite side from the Kerry and Cork caves. So the route map Nash draws — Pembrokeshire, 76 kilometers of tundra, Rosslare — conflicts with the best available model. Not with "some" model, but with the very one considered the regional standard.
To understand the price of claims, it's useful to lay out the chronology in one line. Around 40,000 years ago — settlement of South Wales by modern humans. Around 33,000 years ago — the most provocative Irish claim: a reindeer femur from Castlepook Cave with alleged stone tool marks, which zooarchaeologist Ruth Carden reported in 2021. The age is plausible, but the human authorship of the marks is disputed, so academia doesn't count this date officially.
Next — 26,000 years ago, glacial peak: the whole island under the sheet, no life. Around 16,000 — break of the last bridge according to the Dublin model. And here in the schedule appears the new find: "at least 15,000 years," and possibly older. Notice the window: if dates confirm closer to the lower boundary, people will be on the island AFTER the bridge break — and then the caves will testify to sea voyages, not a walk from Wales.
The official human presence benchmark today rests on an elegant find: a bear patella from Alice and Gwendoline Cave in County Clare, with cuts left when butchering the carcass on fresh bone. Two independent radiocarbon measurements gave 12,600–12,800 years — a peer-reviewed 2016 work in Quaternary Science Reviews. After that — the famous Mount Sandel site near Coleraine, circular Mesolithic semi-dugouts, roughly 9,000 years, a classic Peter Woodman excavation. Hence the gap in headlines: the BBC compares the new find to a rounded "10,500 years," the academic baseline starts at 12,700. In both cases the new dating, if it holds, pushes the boundary back thousands of years.
Cave art, unfortunately, has an inherent problem: charcoal can be dated directly, but engravings and ochre cannot. Their age is determined by the "ceiling and floor" method: uranium-thorium dating of the calcite crust that grew over the image, and the layer the image cut through from below. You get a range. The team hasn't published a single number in a peer-reviewed journal yet — and ZME Science honestly calls things by their names: both key news claims, dating and route, currently hang on researchers' words.
And now my favorite part of the program — because the question "was there a bridge" has witnesses who are silent in geology: the animals themselves. Irish fauna is strange: on the island there's no mole, no common shrew, no weasel, and no snakes. But there are species not found on the neighboring island in this combination. If a bridge existed and was passable — fauna would have filled the island evenly, as it fills the mainland. Instead we see a sieve with very characteristic holes.
The answer was found in genes. Take the pygmy shrew — a tiny creature with spacecraft metabolism. Its Irish mitochondrial line matches the line from Northern Spain — seemingly a greeting from the south, a "Lusitanian trace." But the full genetic analysis (journal Heredity, 2011) showed: Irish population diversity is pathetically low, haplotypes diverge in a "star" from one central one — classic sign of recent colonization by a small group.
Authors' verdict: the pygmy arrived from Britain in the Holocene as a human stowaway. The mechanism is confirmed by distribution: all islands with Irish pygmy lie within 25 km of the mainland, and it's absent from Shetland (175 km) and Iceland — an animal with furious metabolism simply doesn't survive a long voyage without food.
The greater white-toothed shrew's story is fancier. It was first found in Ireland in 2007 — from bones in raptor pellets. Genetics and skull morphometrics showed the source: France. The animal crossed the sea in bales of imported hay and staged an invasion on the island, from which the pygmy now suffers. A living demonstration of how the "corridor from Wales" works — only not glacial, but hay-based.
The wood mouse carries Iberian mitochondrial lines — the same "Lusitanian" signature, the same conclusion from modern reviews: multiple-period human introductions, not a walk across a bridge.
But the stoat is a possible rarity: the Irish subspecies Mustela erminea hibernica, by estimates, has lived on the island for over 12,000 years, and this is the only small animal that could have managed to slip through at the end of glaciation, while some northern passage still existed. One for the whole island. If the southern bridge had lived longer — the rest should have come behind it. And the weasel, the stoat's closest relative, isn't on the island — though in Britain it's found everywhere.
About the absence of snakes on the island there's a beloved folk story: Saint Patrick drove them out with his staff. Science answers more boringly and interestingly: there were never snakes in Ireland — the paleontological record documents this without any lyricism. The logic is the same as with fauna: when the last glacier retreated and it got warm, Ireland was already an island. Reptiles — creatures whose geography is entirely written by temperature and terrestrial connectivity — simply couldn't reach the freed land.
Today's herpetofauna composition underscores this: one native terrestrial species — the common lizard, a cold-hardy champion capable of wintering, plus the introduced slow worm. That's it. Snakes, which need stable warm seasons, didn't swim to the island. The same pattern holds New Zealand, Iceland, and Greenland — zero snakes everywhere isolation overlapped with cold.
Let's tie this to our theme. Nash's bridge is a plateau with summer temperatures minus ten to twenty, where hunters should have followed herds of mammoths, bison, and deer. But if this corridor existed even after the break — the first to pass through wouldn't be humans but that very cold-hardy fauna: the same common shrew, neighboring Welsh lizards, voles. They're not there. Fauna votes against the bridge, and it, unlike cave art dating, has already been peer-reviewed.
Let's strike a balance. The most likely and most valuable development: if uranium-thorium and radiocarbon dates confirm the age of the art, Ireland gets the oldest artistic tradition in the British Isles, and the lower boundary of human presence moves back several millennia. This is a textbook-level sensation — on an island archaeology long considered "young."
But here's the irony that made me dig into this topic: even full triumphant confirmation of the dating will most likely hit the bridge hypothesis rather than support it. The "at least 15,000 years" window lies predominantly AFTER the modeled break (~16,000). People on the island at this time means people in boats or walking along the ice edge from the north, not tourists from Pembrokeshire. The only scenario where "on foot from Wales" survives is if dates turn out substantially older than 16,000, meaning if the art is the same age or older than the glacial peak. That would require settling Ireland ten thousand years before the bear bone and before everything we know about the end of the Ice Age in Europe.
Perhaps the mysterious deer trace from Castlepook hints that the early chapter of history isn't closed yet — but this will only be verified in the lab.
The find's on thin ice in another place: until numerical dates and a peer-reviewed article are published — the entire construction rests on a press release and coordinator enthusiasm. Nash himself is an experienced rock art specialist, and the collection of techniques (engravings, ochre, finger flutings) looks plausibly Paleolithic. But archaeology is full of cases where "looks like Paleolithic" crumbled at first analysis, and the secrecy of locations, necessary to protect caves, unfortunately also prevents independent verification. So the right position here is halfway enthusiasm: applaud the art, trust the dates conditionally, believe the route after uranium-thorium.
What I like about this story isn't the sensation, but its architecture. On one hand — a team that did everything right: didn't go "wandering blindly," but trained a model on known sites, produced 28 candidates, checked them on foot, and found what by consensus shouldn't be there. This is exemplary work with the hypothesis "absence of evidence is not evidence of absence." On the other hand — the same team in the press release skipped past their own destination point: if their caves are dated, geology says people more likely sailed there than walked. The headline "on foot from Wales" is beautiful, but their own finds are about to be reframed as "in boats from nowhere."
And a second lesson, already general for any science. An island is an ideal detector of historical hypotheses: its fauna can't lie. Shrew and mouse genetics, absence of mole and snakes, the single ancient stoat — all these are independent witnesses that accumulated over decades and don't depend on any press releases. Cave art is more spectacular, but fauna is more evidential. In the dispute about how islands were settled, listen to those who can't interview themselves to the BBC.
Where I'd dig further: first, for dates — the moment a peer-reviewed article with numerical ranges comes out will decide everything; second, for "ice edge" as an alternative route — the hypothesis of movement along the ice sheet boundary has been gaining weight in recent years and could reconcile early dates with a broken bridge; third, forgive me, for hay — the story of how Irish fauna assembled for twenty thousand years from pellets, bales, and holds deserves a separate longread.