Castell Caer Seion is an Iron Age hillfort situated at the top of Conwy Mountain, in Conwy County, North Wales. It is unusual for the fact that the main fort contains a smaller, more heavily defended fort, complete with its own distinct defences and entrance, with no obvious means of access between the two. The construction date of the original fort is still unknown, but recent excavations have revealed evidence of occupation as early as the 6th century BC, whilst the smaller fort can be dated with reasonable certainty to around the 4th century BC. Whilst the forts were constructed in different periods, archaeologists have uncovered evidence of concurrent occupation, seemingly up until around the 2nd century BC. The larger fort contained around 50 roundhouses during its lifetime, whereas examinations of the smaller fort have turned up no more than six. The site was traditionally associated with Maelgwyn Gwynedd (c. 480 – c. 547 AD), but there is no evidence pointing to a 6th-century occupation. The fort and wider area beyond its boundaries have been said to retain significant archaeological potential, and are protected by law as a scheduled ancient monument.
Description
The fort is on the summit of a ridge of rhyolite at an altitude of approximately and is enclosed by a single rampart, with more complex works protecting a smaller fortified area at the western end. The northern side of the ridge is steep enough to offer a natural defence meaning that no outer wall was needed for the first phase of construction. The smaller fort was built inside the western margins of the larger fort sometime around the 4th century BC; it was completely enclosed, and defended by thick stone walls on all sides, with defensive ditches running along the eastern and north-eastern boundary and a defensive outwork to the southeast. The site has a commanding position overlooking Conwy Bay, and is situated close to a major trackway that follows the coastal ridge. The total area of the fort is , it is roughly long, and at its widest measures around from the southern ramparts to the cliffs on the northern boundary.
Confusion surrounding name
Writing in Archaeologia Cambrensis, W.E. Griffiths and A. H. A. Hogg identify 'Caer Seion' as the older and more genuine name of the hillfort, for which they cite evidence dating as far back as the 9th century. The alternative name 'Castell Caer Lleion' seems to date to the late 17th century and probably came about due to a mistranslation. Indeed, 'Castell Caer' is somewhat of a tautology, translating into English as "Castle Fort". Because there is no standardised name for the site, it is referred to variously as 'Caer Seion', 'Castell Caer Seion', 'Castell Caer Lleion', 'Castell Caer Leion' and 'Conwy Mountain hillfort'.< An antiquated name, no longer in popular use but recorded by John Leyland in his 1530s Itinerary in Wales, is 'Sinnodune'.
Archaeological excavations
There have been three principal excavations at the site: first in 1906 and 1909, then in 1951 and 1952, and finally in 2008. The dearth of dateable finds meant that, until the early part of the 21st century, the dating of the site was largely inferential. Each excavation has revealed a little more about the site, with modern techniques enabling archaeologists to date the earliest known occupation of the fort to within 200 years.
First excavation: 1906, 1909
The first recorded excavations – carried out by Harold Picton and W. Bezant Lowe – were begun in 1906, before a delay caused them to briefly abandon the project and resume it in 1909. Their excavations revealed little, but they are the first to record finding sling stones at the site, and they also turned up a rubbing stone. Bezant Lowe noted the possible existence of a semi-circular defence facing the entrance to the smaller fort. In one of the huts at the south-western end of the small fort they uncovered a stone floor, haphazard for the most part, but carefully laid and fitted together at one side of the hut. Picton notes that there was no evidence of the floor being used as a hearth, but does not speculate further as to the nature of the feature.