Caerleon Net


Pre Roman


Post Roman


1840 Tithe Survey

Parish Records

Articles From Gwent Local History Journal

Caerleon Net

Video About the Hill Fort





Picture of the Month


Notice Board


Local Information

See Also

Other Hill Forts in the Area


The Roman Army In Caerleon King Arthur Parish Records Directories History Index Home

Lodge Hill Fort

Interim Report
Excavation of Lodge Wood Camp
Summer 2000

by Ray Howell and Josh Pollard, SCARAB, University of Wales Newport

CAERLEON, Lodge Wood Camp, (ST 323 914)

Lodge Wood Camp is an imposing hillfort situated on the north-west edge of present day Caerleon, overlooking the mouth of the Usk valley and parts of the Severn estuary.  The monument comprises a triple-banked enclosure with out-works on the south, west and north, enclosing an area of c. 2.2ha.  A small oval enclosure (100 x 50m across) lies within the western third of the interior.  Several phases of construction may be represented by the earthworks (Whittle 1992, 45).  Despite its size, and location adjacent to the Roman legionary fortress at Caerleon, until this year the monument had not been subject to archaeological investigation.

An opportunity to examine the site was provided by the generous offer of a grant from the Charles Williams Trust (a Caerleon-based educational charity).  Scheduled Monument Consent was obtained, and work was undertaken by a team from the University of Wales College, Newport during June and early July 2000.  The principal aims of the excavation were to establish the date and sequence of rampart construction, the character of activity within the interior (e.g. evidence of occupation, craft activity, etc), and whether evidence existed for pre- and post-hillfort activity (Neolithic, Bronze Age, Roman and/or post-Roman).

Following limited geophysical survey within the interior, three areas were selected for excavation: an area of 106m2 within the western part of the interior of the hillfort close to the southern rampart (Trench 1); a trench (2) through the first sequence of bank and ditch adjacent to Trench 1; and a third trench sited over the innermost section of the western entrance (Trench 3).  Additional small-scale interventions were made in and around areas of mountain bike damage within the western entrance.  The results are summarized according to trench:

Trench 1: the hillfort interior

Intensive biological activity, in particular through tree and scrub root action, had resulted in severe reworking of the soil matrices, effectively removing obvious boundaries between originally discrete layers and deposits.  Consequently, it was often impossible to determine context changes on the basis of variations in soil structure, colour and composition, and stratigraphic boundaries were generally identified by changes in stone density.  The acidic nature of the soils had also led to the destruction of unburnt bone and

The sequence begins with the cutting of a substantial quarry hollow, running the whole length of the southern side of the trench, and two irregular hollows up-slope from this.  Over 4.2m wide and up to 1.0m deep, the cut of the quarry hollow followed natural bedding planes within the rock, the inner side varying from vertical to shallow and the base being stepped in profile.  Lying within the lee of the rampart, the feature was almost certainly dug to provide material for bank construction.

The quarry hollow was subsequently filled in a seemingly deliberate act with deposits of soil and stone in order to provide a level surface.  Cut into this surface in the south-western part of the trench were a number of post-holes, one group of seven (F.14, 34-39) describing a small sub-rectangular building c.3x2m (Fig. 1).  The interior of the structure was marked by a spread of irregular stone paving.  This was bounded on its northern side by two phases of shallow ditch or gully (F.3/20), running down-slope from the north-west corner of the trench and terminating adjacent to the structure and a zone of cobbling.  The second phase ditch F.3 had been deliberately backfilled with a deposit of stone rubble, and contained a La Tène 1 iron brooch and sherds of middle Iron Age pottery.

To the east, a low earth and stone bank ran on from the ditches after a gap of c.0.8m, both elements apparently forming an insubstantial enclosure of unknown extent within the interior of the hillfort.  The 'entrance' gap was marked by a single large post-hole.  The rock-cut scoops were contained within this, and provided platforms upon which a series of timber structures had been constructed.  Situated against the edge of excavation, and marked by stone-packed post-holes, it was not possible to determine the precise extent, layout or function of these buildings.  Quantities of metalworking slag and furnace base were, however, recovered from these areas in association with middle Iron Age ceramics.

At a later date, a series of narrow terraces were created across the east half of the area, taking advantage of the already stepped profile created by the slope of the ground and the infilling of the quarry hollow.  Cut into these, from a level that produced a small amount of late Roman pottery, were a number of small post-holes.  Both pottery and post-holes may relate to a limited phase of re-occupation of the hillfort.

Trench 2: the bank and ditch

A 2m wide cutting was taken through the inner bank and ditch.  The ditch was fully excavated, but time constraints resulted in only partial excavation of the bank deposits.  V-profiled, 5.7m wide and 2.2m deep, the ditch is of a single phase.  Making use of geological boundaries, it was cut into rock of the north side and a stiff red clay on the south.  The basal fills of the ditch comprised clay marl.  Sealing this was a massive (0.5m thick) collapse deposit of stone rubble from the rampart, perhaps representing a deliberate slighting of the defences.  This in turn was followed by further deposits of clay loam and rubble.  Sherds of middle Iron Age pottery came from the primary fill and fragments of Roman tile and pottery

The bank was only excavated to its full depth on the northern side of the cutting.  At least 8m wide and over 1.2m high, it is made up of a series of complex deposits of soil and rubble.  A minimum of two phases are represented; the first a stone-revetted timber-laced rampart, and the second a timber revetted stone bank.  The latter only survived as a thin rubble spread overlying a soil that had formed subsequent to the collapse of the primary rampart.  This episode of re-modelling is most likely of latest Iron Age, late Roman or even post-Roman date.

Trench 3: the western entrance

A 10x3m trench was sited to take in the northern half of the entrance proper and the terminal of the northern inner bank.  It became evident that the entrance was an original feature, though it may have been infilled early in the life of the hillfort and then re-established at a later date.  The sequence begins with a simple entrance formed by a low revetted stone bank, 6.2m wide, incorporating a recessed 'guard-chamber' and cobbled entranceway.  The format is similar to that of the period III-V south-east entrance at Dinorben (Savory 1964, fig. 5).  An oval post-hole off-set from the centre of the entrance may be part of a gate structure.  Subsequent phases of modification and addition followed.  First, an elaborate hornwork was constructed on the exterior side.  At a later date the entrance was infilled with dumps of earth and stone, supported on the outer face by stone revetting.  This was subsequently cut through late in the life of the monument, re-establishing the original line of the entrance.  With the exception of one small fragment of tile and a piece of modern glass, both from immediately below the topsoil, no artefactual material was present.  This in itself might imply the recutting pre-dates the post-medieval period


The excavations provided clear evidence of an early-middle Iron Age date for the construction of the hillfort, and for contemporary activity (both occupation and metalworking on some scale) within the interior.  Pre-hillfort activity is attested by only two pieces of worked flint, and these may, anyway, be curated.  A low density scatter of late Roman ceramics, along with limited structural evidence (post-holes within the interior and a late phase of bank rebuild) indicates renewed activity focussed on the monument, following a period of abandonment roughly coeval with the establishment and occupation of the Roman fortress at Caerleon.  A substantial deposit of stone rubble within the ditch, obviously derived from the inner rampart and associated with Roman pottery and tile, could tentatively indicate deliberate slighting of the bank.
Given our relatively impoverished understanding of the Iron Age in south-east Wales beyond the margins of the Severn Levels, the excavation has the potential to contribute significantly to our understanding of settlement practices, activities, and to a limited degree economies and environments, in the region during the 1st millennium BC.  The excavation has also provided the potential to examine the later (post-Iron Age) history of one of the region's larger hillforts, and, whilst to some extent ambiguous, the results do hint at the possibility of later Roman and/or post-Roman reoccupation.  In this respect, the key aims and objectives of the excavation have been, or will be, met.


First and foremost, we would like to thank the Williams Trust for their generous sponsorship of the excavations, Mr T. Prichard and the Trustees of Lodge Farm Church, on whose land the hillfort lies, and Mike Yates and Felicity Taylor of Cadw.  The work was supervised by the authors and Adrian Chadwick, Ian Dennis and Mike Hamilton, with additional assistance from Lesley McFadyen and Steve O'Rourke.  Thanks also go to Neil Phillips and Kate Smith for undertaking a survey of the earthworks under very difficult conditions.

Ray Howell and Joshua Pollard, SCARAB, University of Wales Newport


Savory, H.N. 1964. Dinorben: A hill-fort occupied in Early Iron Age and Roman times. Cardiff: National Museum of Wales
Whittle, E. 1992. A Guide to Ancient and Historic Wales: Glamorgan and Gwent. London: HMSO

Many thanks to: Ray Howell; Joshua Pollard; SCARAB and University of Wales Newport for permission to publish this report.