The Glamorgan-Gwent Archaeological Trust
Historic Environment Record
 

The following information has been provided under the terms and conditions of access as detailed on GGAT’s website www.ggat.org.uk.  Copyright is reserved on all data supplied by the GGAT HER Charitable Trust. All output resulting from the use of the data must acknowledge the source as follows:-
Derived from information held by the GGAT HER Charitable Trust copyright.

The data below is intended to be used for information and research only and is not for use as part of a commercial project.  If you wish to use information derived from material held by the GGAT HER Charitable Trust for publication in printed or multimedia form or to compile resources for commercial use, prior permission must be obtained in writing. For further information or to arrange a visit to the Trust please send an enquiry form http://www.ggat.org.uk/her/english/enquiry_form.html to her@ggat.org.uk.

Cardiff Roman Fort

Primary Reference Number (PRN) : 00101s
Trust : Glamorgan Gwent
Community : Castle
Unitary authority : Cardiff
NGR : ST1806076590
Site Type (preferred type first) : Roman Fort
Status : Scheduled Monument

Summary :
A Roman fort of four separate phases, 1st-4th. The first fort (probably c AD 55 - c AD 80), constructed in timber, may have covered an area of 12 ha and was probably for a conquest-period battlegroup or for a legion. It was replaced at the end of 1st century by an auxiliary fort, also in timber with earth defences, which was quickly replace by the third fort whose defences were on a slightly different line. A fourth, masonry fort of 'Saxon shore' type, was constructed after AD270.

Description :
Although it is now level, the site of the Roman fort at Cardiff appears originally to have been higher at the east end, with the highest point in the north-east corner, sloping down towards the west. It is possible that in the 2nd or 3rd century that this slope may have been terraced along the line of the north-south road which ran through the middle of the first three forts (Webster 1981b, 210).

The first fort:
From the evidence of the finds, the first fort appears to have been constructed early in the third quarter of the 1st century AD, and occupation lasted until some time between AD 70 and 90; however, given the history of the Roman conquest of South Wales, it is probable that this fort was founded c AD 55 and decommissioned c AD 80. Besides the north-south axial road noted above, parts of a number of buildings were discovered, all of timber construction with earth fast sills. Only two of these buildings were excavated sufficiently for any understanding to be gained of their plans, though neither lay completely within the area of excavation. Both were large, that in the north-east of the Castle Green, of which only the southern extent could be established for certain, measuring more than 45m north-south and more than 28m east-west, and that beneath the south centre of the Green at least 12x20m. In their fragmentary state, neither building can be identified as matching one of the standardised building types to be found in a 1st-2nd century fort, a military origin is virtually certain because of the Neronian construction date, and the type of construction technique employed. The most likely identification of the very large building at the north-east is as a principia or headquarters building, but one which will have been much larger than required for the standard auxiliary units of 500 or 1000 men. This has implications for the size and type of fort involved, for which there is otherwise no evidence since none of the defences lay within the area of excavation. The principia was normally placed close to the centre of a fort unless there were compelling topographical reasons which made this inappropriate, a constraint which will not have existed at Cardiff. A possible line for the north-western limits of the first fort can be arrived at from an assumption that the second fort which replaced it is unlikely to have extended beyond the area of the first fort, and on this basis the excavators suggest that the fort may have covered an area of as much as 12 ha, designed for part of a legion or an ad hoc grouping of units, such as was not uncommon during the period of the conquest. The implication is therefore that the fort extended substantially beyond the area of the castle on all sides. The other building is interpreted as the praetorium, or residence of the commanding officer on the basis of the imperfectly understood though clearly complex plan; it had undergone major repairs during its period of use (Webster and Webster 1981a, 43; Webster 1981b, 203).

Second fort:
Very little of the second fort was available for excavation, since the greater part of it lies to the north of the fourth fort/castle. Part of the southern defences were examined, consisting of a clay and turf rampart fronted by a V-shaped ditch, together with a portion of the south gate, through which entered the north-south axial road noted above, which was retained from the first fort. The gate seems to have consisted of three massive post-holes at either side of the road, presumably supporting a guard-tower at or above the level of the rampart walk. The lengths of road ditch now inside the gate were filled and the remaining lengths of ditch linked by a culvert under the road, which was resurfaced, the new surface covering the filled-in ditches. The dating evidence for this work suggests that the second fort was short-lived and did not last much beyond the end of the 1st century (Webster and Webster 1980, 48; Webster 1981b, 205-6). The size of this fort is a matter for conjecture; it has been reconstructed as covering an area of c 140x140m.

Third fort:
Again, only the defences were excavated. The defences of the second fort were carefully dismantled, with the ditch being filled in and the postholes for the gates being sealed with grey clay. Either a new rampart was built or the old one refaced, its south side now extending partly over the filled-in ditch of the second fort. The ditch was redug, partly into the fill of the old ditch, and new postholes dug in undisturbed ground for the replacement gate which was larger and more elaborate with six posts either side of the roadway, indicating the provision of gate-towers. In all, the defences were moved some 1-2m further south, and as it seems unlikely that so much trouble would be taken unless the fort were being redesigned, the excavators suggest that this was part of a scheme to reduce the fort in size, perhaps for a 500 strong infantry unit, a pattern which has been observed elsewhere in Roman Britain. There is little evidence for dating from the fort itself, but the excavators believed that a series of fairly regularly sized and spaced pits in the north-east centre of the Castle Green, containing material dating up to the beginning of the 3rd century, are likely to represent rubbish disposal from the third fort (Webster and Webster 1980, 49; Webster 1981b, 206-7).
Area to the south of the second and third forts:
Besides the pits mentioned above, the area which lies within the fourth fort/castle but outside the second and third forts was used for a variety of functions. The fort axial road continued in use in this area, and a number of fragmentary buildings were observed to either side of it, occupying the area which had been used for the praetorium of the first fort. The excavators suggested that this was a civilian settlement which had sprung up outside the south gate of the fort, a favoured spot for such establishments, to take advantage of the trading opportunities afforded by the military presence. An alternative explanation is that this area formed an annexe to the fort, providing space for activities which could not be conveniently carried out within the defences. The most noticeable activity which was recorded as taking place in this area was metalworking; the timber buildings over the praetorium, which appeared to be contemporary with the second fort and may have continued for some time after its replacement by the third fort, gave way to smelting activity which generated quantities of tap slag and furnace lining, and secondary processing of iron took place further north in the area previously occupied by the principia. There is also some evidence for the manufacture of tiles, since overfired and warped roof-tiles were found in the area. Such industrial processing could have been carried out either as military or as civilian activity (Webster 1981b, 206-7). There is also a report of a Roman hypocaust having been found in the 18th century near the south-west corner of the Castle Green (Ward 1908, 30; noted as the south-east corner of the castle court by Clark 1862, 253). This may relate to the bath-house which will have been provided for the fort and which is not otherwise attested. Its position is close to what was, before the formation of the castle moat, an arm of the Taff (Ward 1908, 44), and thus suitable for a bath-house to an auxiliary fort since these tended to be built close to a water supply where possible.

Fourth fort:
The fourth Roman fort is the structure which later became the outer ward of the castle. Investigations carried out during the restoration of the castle under the third and fourth marquesses of Bute established that the medieval defences followed the Roman line throughout (Ward 1913b, 86), and consisted of an earthen bank retained at the front by a wall of coursed rubble, and with a stone kerb at the back (Webster 1981b, 207). The wall was of typical Roman construction; a foundation of river boulders was capped with mortar and an offset foundation course (`plinth'), upon which was built the main wall structure, 3.2m wide, faced on both sides by courses of roughly squared local stone (lias limestone) with a core of undressed blocks of the same limestone and boulders. The mortar was a high-quality lime mortar with no admixture of crushed tile. At a height of 2.28m the wall was reduced in width to 2.59m by four single-course internal offsets; in the south-east corner, where the inner facing survived to a higher level, a second series of offsets reduced the width to 1.90m. At this point too, the base of the wall is at a lower level than elsewhere, probably as the result of a drop in the Roman ground level (Ward 1908, 48-9; 1913b, 88-9). The upper limits of the surviving outer facing of the Roman work can clearly be seen on the north east and south walls of the standing monument, where it is marked by a line of tiles; internally however, the core rises very considerably higher, to 3-5m (`10 to 12 ft': Ward 1908, 45; `from 12 ft to 16 ft': Ward 1908, 48), and the inner facing, although not surviving to the same height as the core, was preserved behind the bank (Clark 1880, 289; Ward 1908, fig 1, fig 2, fig 7; Ward 1913b, 88). Polygonal bastions survived along part of the enceinte, and there was evidence that there had been more. Gates are known from the middle of the north and south sides, with polygonal gate-towers. Most of the bastions had solid bases, and stood on offset foundation courses which were at a different level from those of the walls to which they were not bonded, though the upper part of the bastions were (Ward 1908, 49-50). This suggests that the decision to incorporate bastions was made during the course of building work, when also the form of the fort corners was changed from rounded to square (Webster 1981b, 207). Nothing is known of the Roman ditch, which seems to have been cut away by the medieval ditch. Lord Bute's reconstruction of the defences on the whole draws from excavated evidence and gives a good impression of what the fort probably looked like. The main point at which it can be shown not to be accurate is in the provision of a mural gallery, a feature which is not known from any contemporary fort in Roman Britain, and whose width is based on the 3m standard of the lower part of the wall, with no allowances being made for the offsets which reduce its width above the height of 2.28m.
Internally the north-south road may have continued in use, even though it was no longer aligned with the gates, since no new road has been found between the new gates. Like other similar late Roman forts, it does not seem to have had any internal buildings in stone, but hard-standings of cobble and slag were discovered which seem to have been designed as rafts for timber buildings, at least some of which probably had daub infill (Webster 1981a, 43; Webster 1981b, 207-8). The fourth fort appears to be 3rd century in origin; if the kerb retaining the bank was an original feature, which is not certain, a coin of AD 270 stratified underneath it gives a terminus post quem for construction, and occupation continued into the second half of the 4th century (Webster 1981b, 207-8). (Evans 2001)

Excavations in 2005-06 produced evidence of a timber building belonging to the first fort, represented by earth-cut slots for the walls and the remains of some clay floors. From its ground-plan, it was possibly a barrack-block. There were also two small ovens, and evidence for the manufacture of glass and the working of lead, iron and bronze. A line of five post-pits behind the bank of the fourth forts defences may represent a revetment to its rear. (Webster and Marvell 2010)

During excavations in 2005-06, in advance of a new visitor centre at Cardiff Castle, the central part of the excavation area was found to be occupied by a road, originally one of the main thoroughfares of the 1st century fort, the via principalis, and later continuing as a route running south from the second and third forts. It has a metalled surface. The structure was originally constructed with a 15m wide base, eventually narrowing to 10m, with associated side ditches, although the eastern ditch was preserved to a higher degree than the western ditch. A central camber was also evident, giving the road a maximum depth of 1.25m. It is also joined by a road approaching from the east at the southern edge of the excavated area. Towards the south of the road, where the stratification was better preserved, it was apparent that in its final phase the road-side ditch had been partially filled in and the metalled surface slightly extended over it.

A number of earth-cut features that appeared to be associated with the 1st century Roman fort were discovered. These consisted of:
(a) 41 pits, mainly clustered in the western part of the excavation
(b) 29 post holes all distributed fairly evenly across the western area of the site, with the exception of thirteen closely grouped examples, aligned east-west, in the eastern half of the site.
(c) The beam slots of a rectangular building located to the west of the Roman road. The remains suggest that the building was aligned roughly east-west and had a minimum length of 13.6m, and a minimum width of 11.33m. At least one internal division was present, again aligned east-west, which sectioned off a 3m wide room in the north of the larger building. Structures of a similar size have previously been excavated in the southeast quarter of the castle, and have been tentatively identified as barracks.
(d & e) Clusters of stake holes.
(f) 11 linear features on either side of the road. They included beam slots associated with the corner of at least one other building.
There were a number of destruction deposits associated with this end of this period.

Beneath the Modern, Post-medieval and medieval deposits were the remains of the defensive bank that ran inside the stone wall of the fourth Roman fort. It measured a maximum of 21.9m east-west, and 12.6m north-south. It was preceded by two sections of bank that respected the alignment of the road, would appear to represent part of either a temporary encampment; an uncompleted defensive circuit around a previously unknown fort or the enclosure of part of the vicus of the second or third forts.
(Dunning 2011)

Although the presence of a stone Roman fort at Cardiff has been known since the beginning of the 20th century, it is only in comparatively recent years that the castle was investigated in any detail. Excavations between 1974 and 1981 led to the proposal of a series of four successive forts occupying the same site, the sequence for which can be summarised as follows (after Webster 1981; 1990):
I. A large fort, the date range for construction is c AD 55/60 - c AD 75 (probably c AD 55 on historical grounds, as it is likely to be contemporary with Usk). Closure 70-90 (perhaps AD 80?)
II. A smaller fort (probably Flavian). Did not survive long into the 2nd century
III. Fort, smaller than its predecessor, but on the same alignment. Occupied during the 2nd and 3rd centuries
IV. A fort of Saxon Shore type, built c AD 260 or later, occupied until the 370s

The suggested plan of the forts (somewhat subjective) is shown as Figure 7. Webster appears to favour the idea of a largely continuous inhabitation of the site from the foundation of Fort I until the abandonment of Fort IV at some point in the later 4th century. He is not explicit as to whether there was a break in occupation between Forts II and III, but certainly he argues for continuity between Forts III and IV.

The area of Fort I, which encompassed all of the later installations, is bounded to the south by modern Duke Street and Castle Street, and extends across Kingsway to the east and into Cathays Park to the west. A north-south road running centrally through Fort I has been identified, but although Webster plan shows it continuing south of the defences, this is conjecture. Forts II and III, although much smaller, appear to have continued to use this road as their central axis. Both lie partly within the area of Fort IV, and partly to the north, on ground much disturbed by the later digging of a moat, and by modern development. The largely reconstructed defences of Fort IV form the remains presently visible.
(Pearson 2002, 15)

Sources :
Cadw , Application for Grant Aid
Cadw , Application for Scheduled Monument Consent
Cadw , Application for Scheduled Monument Consent
Dunning, R , 2011 , New Interpretation Centre, Cardiff Castle, Cardiff: Post-excavation analysis project design.
Dunning, R. , 2008 , External Signs and Viewing Platform, Cardiff Castle, Cardiff: archaeological evaluation and watching brief ( © GGAT)
Evans, E M , 2001 , Romano-British southeast Wales settlement survey: Final report ( © GGAT)
Glamorgan Gwent Archaeological Trust , Site Visit Record
Glamorgan Gwent Archaeological Trust , Site Visit Record
Glamorgan Gwent Archaeological Trust , Site Visit Record
Lewis, R. , 2017 , Bradley Court, Cardiff Archaeological Desk-based Assessment
Pearson, A , 2002 , Roman roads and vici in Southeast Wales. GGAT report no. 2002/061
Randall H.J , 1954 , The Roman Period. (A Hundred years of Welsh Archeology). , Archaeologia Cambrensis : 103 : 80-104
Robinson G.E. , 1877 , Roman Cardiff
Simpson G. , 1963 , Caerleon and the Roman Forts in Wales in the Second Century AD. , Archaeologia Cambrensis : 112 : 13-76
Smith E.G. , 1934 , Summer Excursion, 1934 - Cardiff Castle and Museums
Storrie J. , 1893 , Discovery of Ancient Remains at Cardiff , Archaeologia Cambrensis : 48 : 277-281
Ward J. , 1908 , Roman Cardiff , Archaeology Cambrensis : 63 : 29-64
Ward J. , 1913 , Recent Discoveries of Roman work at Cardiff Castle , Transactions of the Cardiff Naturalist Society : XLVI : 85-90
Ward J. , 1913 , Roman Cardiff II , Archaeologia Cambrensis : 68 : 159-164
Ward J. , 1914 , Roman Cardiff III , Archaeologia Cambrensis : 69 : 407-410
Webster P. , 1981 , Archaeological Notes: Cardiff Castle Excavations, 1974-1981 , Morgannwg : XXV : 201-211
Webster, P V and Marvell, A , 2010 , Cardiff
Wheeler R.E.M. , 1923 , Excavations , The Bulletin of the Board of Celtic Studies : I pt.IV : 340
Wilson D.R. , 1975 , Roman Britain in 1974
Wilson D.R. & Taylor M.V. , 1961 , Roman Britain in 1960. I. Sites Explored , Journal of Roman Studies : 51 : 157-160
Clark, G T, 1862, Some account of Cardiff Castle, Archaeol Cambrensis 3 ser 8, 249-71
Clark, G T, 1890, Cardiff Castle, Archaeol Cambrensis 5 ser 7, 283-92
RCAHMW, 1991, An inventory of the ancient monuments in Glamorgan. Vol III Part 1a, Medieval secular Monuments: The early castles: From the Norman Conquest to 1217
Ward, J, 1901, Cardiff Castle: Its Roman origin, Archaeologia 57, 335-52
Ward, J, 1908, Roman Cardiff, Archaeol Cambrensis 6 ser 8, 29-63
Ward, J, 1913a, Roman Cardiff II, Archaeol Cambrensis 6 ser 13, 159-64
Ward, J, 1913b, Recent discoveries of Roman work at Cardiff Castle, Trans Cardiff Naturalists 46, 85-9
Ward, J, 1914, Roman Cardiff III, Archaeol Cambrensis 6 ser 14, 407-10
Webster, J and Webster, P, 1981a, Cardiff Castle, Archaeol Wales 21, 42-44
Webster, P, 1981b, Cardiff Castle, Morgannwg 25, 201-11
02/PM Desc Text/Grant JP/1923/Cardiff Castle/pp1-19
01/MM Record Card/OS/1957/ST 17 NE I
08/PM Desc Text/Savoury HN/1973/Illustrated Guide Ancient Monuments of Wales no4 p35
07/PM List/RCAHM/1976/Glam Invent/p735
04/PM Desc Text/Wheeler REM/1925/Prehist and Roman Wales/pp234-6
05/PM Desc Text/Nash-Williams VE/1954/Roman Frontier in Wales/pp94-5
03/PM Desc Text/1927-1929/Bull Board Celtic Stud/Vol4 p256
06/PM Desc Text/Ward J/1901/Arch Camb/Vol57 pt 2 pp335-52
09/PM Mention/Webster J & P/1977/Arch in Wales pp33-4 no64
020/Desc Text/Cadw/Application for Scheduled Monument Consent/2007 Copy in further information file.
021/Desc Text/Cadw/Application for Grant Aid/2008 Copy in further information file.
022/Desc Text/Cadw/Application for Scheduled Monument Consent/2008/Copy in further information file.

Events :
E000202 : Cardiff Castle and Roman Fort (year : 1974-1981)
E000781 : Disabled access at Cardiff Castle (year : 2009)
E003402 : Platform Lift, Western Apartments, Cardiff Castle (year : 2011)
E003401 : Platform Lift, Western Apartments, Cardiff Castle (year : 2011)
E003113 : Lift Installation Western Apartments (year : 2009)
E004117 : External Signs and Viewing Platform, Cardiff Castle (year : 2008)
E002251 : Visitor centre site, Cardiff Castle (year : 2005-06)
E001657 : Romano-British Southeast Wales Settlement Survey (year : 1998-2001)
E004936 : Welcome Facilities Site at Cardiff Castle (year : 2004)
E004937 : Cardiff Castle (year : 2002)
E004938 : Cardiff Castle Welcome Facility (year : )
E006000 : Bradley Court, Cardiff (year : )
E008117 : Westgate Hotel, Parkgate, Westgate Street, Cardiff (year : 2019)
E008488 : Clwb Ifor Bach, 9 and 11 Womanby Street, Cardiff (year : 2022)
E007411 : Cardiff Castle and Roman Fort, Cardiff (year : 2023)

Related records
GGAT Historic Environment Record (HER) 03509s
GGAT Historic Environment Record (HER) 00102s

Compiled date : 01-02-1986


Images :



April 2, 2025, 1:07 pm - HTML file produced from GGAT HER Charitable Trust Ltd. Heathfield House, Heathfield, Swansea SA1 Tel. 01792 655208; Fax 01792 474469 website: www.ggat.org.uk  email: her@ggat.org.uk Registered Charity no. 505609
Data supplied by in partnership with Local Authorities, Cadw and the partners of ENDEX © GGAT HER Charitable Trust, 2025 (and in part © Crown, 2025).