A previously unknown Roman fort discovered in Pembrokeshire in Wales overturns assumptions that the area’s indigenous Celtic tribe was on peaceful terms with the Roman invaders.
The site, which has excited archaeologists, had been hidden until now beneath an enormous, overgrown field. It explains why the land had been unsuccessful for farming: the farmer kept hitting stone.
The discovery was made by Dr Mark Merrony, a leading Roman specialist and tutor at Oxford University, who said: “It is a humongous fort, an incredible find of national importance.”
He is all the more excited because it is right next to a Roman road that he has also identified for the first time.
The fort is thought to date from the first to the third centuries, when the Celtic Demetae tribe inhabited the south-west area of modern Wales.
They were thought to have been pro-Roman, meaning there was less need for a major military presence to quell local resistance.
Merrony said that this fort suggested this part of Wales was considerably more militarised than previously thought: “I now don’t think they were pro-Roman at all, but that the Romans were hitting the area with an iron fist.”
He noted that its form and scale was like the only other Roman fort known in Pembrokeshire, excavated at Wiston near Haverfordwest in 2013. Both forts were now linked to a Roman road network that had not previously been known, he said.
As “a native of Pembrokeshire”, Merrony had often travelled along a particularly straight road – wondering whether it was Roman – and, when he looked at satellite imagery recently, his eye was drawn to a field with dimensions likely to be a Roman fort.
He drove to that field. “Sticking out of the ground was a triangular piece that looked like a Roman roofing slate. I thought: ‘Surely not?’ I pulled it up and lo and behold, it’s an archetypal Roman roofing slate, an absolute peach. Flip it upside down and you can see underneath a diagonal line where it was grooved to fit into the one that was underneath it. It’s a real beauty.”
He added: “That was the diagnostic evidence I was looking for, which is a miracle, because it’s a huge site.”
Taking his computer, he knocked on the door of the local farmer: “I said: ‘I think you’re in the middle of a Roman fort here.’”
The farmer told him that the field was full of slate and stone. “That suggests there’s a lot of material under the ground,” Merrony said. “That’s because there’s obviously several collapsed buildings here. The slates are left, the timbers have rotted.”
He said that some of the roofing slates were stained with rust deposits from nails and showed chiselling marks and incisions that accommodated overlying or underlying slates: “These are diagnostically consistent with other slates from Romano-British buildings.”
The fort measures about 185 by 155 metres. Its footprint was “fossilised” by the parliamentary enclosure of large tracts of common land in the area during the 19th century, through banks, hedges and walls.
Merrony said that its features ticked “all the boxes”. It has a characteristic “playing-card” form with rounded corners and is surrounded by a huge ditch with an outer bank or counterscarp and inner bank (scarp), “which was almost certainly the base for the rampart”: “The fort is strategically placed on sloping ground with an excellent vista overlooking a river valley, as was often the case. Its water supply drew upon two nearby springs.”
Merrony is the founding editor-in-chief of Antiqvvs, a quarterly magazine dedicated to archaeology, ancient art and history, where he will publish his latest research in August.
His previous discoveries include two Roman villas, at Ford in Pembrokeshire and at Cutteslowe in Oxfordshire, which he found in 2002 and 2016 respectively. He also identified a Roman road along the southern shoulder of the Preselis in 2022, overturning assumptions that the Romans did not venture far into Wales.
To protect the new discovery, its location cannot be disclosed at present. A geophysical survey would need to be conducted, Merrony said: “This will require funding and, given the extensive nature of this extraordinary discovery, it is desirable to involve the regional archaeological trusts.”
He expects to find evidence of an adjacent “vicus”, a typical civilian settlement that developed alongside a fort. He suggests that this was a Roman auxiliary fort, intended perhaps for a single unit of about 500 troops.
He believes that the site will be declared a scheduled monument: “It is an absolute classic.”
