In the closing years of the seventeenth century, Edward Lhwyd and three assistants journeyed through Wales, inter alia conducting surveys of various antiquities. This pioneering project generated some of the earliest plans of Welsh hillforts, though these survive solely as anonymous copies made some years after the tour in Wales ended in 1699. One of the most informative plans depicts Pen y Gaer in Caernarvonshire, complete with chevaux-de-frise, the first known illustration of such a phenomenon anywhere in the British Isles, where fewer than twenty instances have yet been reliably demonstrated, despite the profusion of hillforts in certain regions. Their early recognition of the Pen y Gaer chevaux-de-frise testifies to perceptive archaeological fieldwork by Lhwyd and his team, though no version of their plan reached print in over 260 years. Regardless of repeated recent claims to the contrary, it remains the only chevaux-de-frise yet identified in north Wales, and one of only five on record throughout Wales. Reviewing of numerous published accounts of Pen y Gaer and certain other hillforts reveals how an over-reliance on secondary referencing can result in disconcerting misconceptions.
Graeme Guilbert (Thu,) studied this question.