Dunnicaer

As early as 1819 one of the Pictish symbol stones had already been prised from the stack and carried off to serve as an ordinary hearthstone, its carved symbols exciting little interest at the time. The rock's isolation and the danger of the climb meant that whatever else lay on its grassy summit went largely unexamined for another generation.

In 1832 a group of local youths scaled the stack and found the remains of a wall crowning its summit plateau. They tipped several of the loose stones down into the sea below, and when these were later hauled back out some were found to bear incised Pictish symbols. The discovery marked the first clear evidence that the isolated sea stack had once carried a substantial, deliberately built structure rather than a bare outcrop of rock.

In 1857 the antiquarian Alexander Thomson formally documented the recovered stones for the Society of Antiquaries of Scotland, and they were subsequently illustrated in John Stuart's two volumes of The Sculptured Stones of Scotland, published in 1856 and 1867 by the Spalding Club. These publications fixed Dunnicaer's stones within the wider record of Pictish carving, even though the difficulty of reaching the stack meant no proper excavation followed for well over a century.

Archaeological investigations on the summit in 1977 and again in 1982 found nothing of significance, and for a time it seemed the stack held little more to discover. A thorough scholarly description of the symbols themselves was published in 1992 by Leslie and Elizabeth Alcock, adding detail to the record even as the fate of the structure that had once held them remained unresolved.

That changed in 2015, when a University of Aberdeen team led by Gordon Noble began new excavations that continued over the following two years. They uncovered a stone rampart reinforced with timber framing running southwest towards the mainland, together with a hearth and the footings of internal rooms. Radiocarbon dating of the timbers placed the fort's use between the second and fourth centuries AD, making it the earliest known Pictish fort yet identified in Scotland.

Among the debris the excavators recovered glass, samian ware and black-burnished ware pottery, and a lead weight — an unusual assemblage so far north of the Roman frontier, and evidence that stone and oak timber had been brought in from some distance to raise a high-status stronghold. The fort was eventually abandoned, perhaps when its inhabitants relocated to nearby Dunnottar, perhaps because the sea was already eating away at the stack that supported it.

The symbol stones themselves are now thought to have been set into that same third- to fourth-century rampart, considerably earlier than the fifth- to seventh-century date conventionally assigned to the simplest Class I Pictish symbol stones — a finding that has reopened debate over how early the Pictish symbol tradition actually began.