Roman Shore Fort & Minor Settlement

Cardiff, South Glamorgan

NGRef: ST180765
OSMap: LR171
Type: Minor Settlement, Fort.

The Latest Roman fort at Cardiff (4th C.)
(Adapted from Collingwood, fig.6)
Roads
N (8) to Caerphilly (Mid Glamorgan)
Coastal road: NW (36) to NIDVM (Neath, West Glamorgan)
NE (13) to ISCA (Caerleon, Gwent)

The Roman Forts at Caerdydd/Cardiff

"Cardiff (Fig. 6) is a rectangular, almost square, fort with angular corners and polygonal bastions. Its walls are 10 feet thick at the base, reduced by offsets to 8 feet 6 inches, and have an earthen bank behind them. The fort measures about 650 by 600 feet (7¾ acres internally) and has a single gate, with projecting towers, at each end, and perhaps a postern in the middle of each side. The Roman walls have been restored recently, and it is therefore possible at Cardiff to see what a Saxon Shore fort looked like in the fourth century (Wheeler, Prehist. and Rom. Wales, 234 seqq.)." (Collingwood, p.54)

There have in fact been a succession of at least four Roman forts at Cardiff, all built in the same area and with much the same orientation:

The Flavian Vexillation Fortress - Cardiff A

N.G.REFDIMENSIONSAREA
??c.1,030 x 1,030 ft
(c.315 x 315 m)
c.24¼ acres
(c.9.8 ha)

The earliest fortifications were also the largest, housing of all its successors within its defensive enclosure, which defined an almost perfect square measuring about 1,030 feet (c.315 m) on each side and covering an area of over 24 acres (c.9.8). This is clearly not an ordinary auxiliary fort, and belongs instead to the class of large, short-lived structures known as "vexillation fortresses". These camps held a large garrison comprised of a backbone of legionary troops backed by a substantial auxiliary contingent of mounted and part-mounted units. These were battle-hardened troops engaged on active campaigns, who held occupancy only for a short period, often for only a single campaign season, until all local resistence in the area had been contained, before the fortress was abandoned in favour of a garrison fort of conventional size.

The First Auxiliary Garrison Fort - Cardiff B

N.G.REFDIMENSIONSAREA
??c.460 x 460 ft
(c.140 x 140 m)
c.4¾ acres
(c.1.9 ha)

The majority of this encampment lies within the northern part of the Flavian vexillation fortress and, although the fort's northern defences lie on the same alignment as those of the fortress, the older bank and ditch was not incorporated within the defences of the new fort, which were built just to the north instead. This suggests that for a period of time the Cardiff site was abandoned, the original rampart perhaps being slighted and thrown into its fronting ditch, or if not, the time period between the abandonment of the fortress and the building of the fort was sufficient to have caused deterioration of the original defences beyond use in the new garrison fortification. This second fort was, like its larger predecessor, square in outline but only one-fifth of its size, measuring around 460 feet (c.140 m) on each side and covering an area of about 4¾ acres (c.1.9 m).

The Second Garrison Fort - Cardiff C

N.G.REFDIMENSIONSAREA
??c.475 x 360 ft
(c.145 x 110 m)
c.4 acres
(c.1.6 ha)

The third fortification on the site was built upon the same alignment as its two predecessors and used the entire line of Fort B's south defences and the southern four-fifths of its east and west sides as a cursor. The fort measured roughly 475 feet ENE-WSW by about 360 feet transversely (c.145 x 110 m) and covered an area of just over 4 acres (c.1.6 ha), the new defences again being built just forward of the existing earthworks. The fact that pre-existing defences were not utilised probably indicates another period of abandonment between the occupation of Forts B and C.

The Roman Shore Fort - Cardiff D

N.G.REFDIMENSIONSAREA
??c.650 x 600 ft
(c.198 x 183 m)
c.9 acres
(c.3.6 ha)

A large stone fort was built on the site at roughly the same time as the forts of the Saxon Shore were appearing on the south-east coast of England; the Shore Fort here at Cardiff being similar in form and size to that at Portchester in Hampshire. This stone fort lay in the southern part of the original fortress encampment, wholly within its defences but upon a slightly different alignment to all of its predecessors, being more closely aligned east-west. The stone fort was supplied with polygonal corner- and interval-towers and substantial double gateways were situated in the centre of its north and south sides, but unlike Portchester no postern gates were provided in its eastern or western defences.

Defences comprised of a substantial stone wall backed by an earthen rampart. The wall was set upon a foundation plinth measuring just over 11 feet wide by 2 feet high (c.3.4 x 0.6 m), laid upon a foundation of river boulders 15 feet wide set in the bottom of a foundation trench 2 feet deep (c.4.6 x 0.6 m). The wall itself thus started at ground level and was about 10 feet (c.3 m) wide at its base, tapering by stepped off-sets at the rear but maintaining a vertical front face to a width of about 8 feet (c.2.4 m) at a height of about 7½ feet (c.2.3 m) above the ground. The Roman wall still exists in places to a height of 17 feet (c.5.2 m) and is not thought to have risen much higher than this when new. The original earthen rampart, now buried beneath Norman and later Medieval additions, survives to a height of about 12 feet (c.3.7 m).

The Dating Evidence From Roman Cardiff

"Finds of coins and pottery indicate that the site was occupied ... before the end of the 1st century ... and occupation seems to have ceased before the middle of the 2nd century." (Glamorgan Inventory, p.92)

A tentative history of the site may be as follows:

  1. The "vexillation fortress" may be dated to the campaigns of governor Julius Frontinus sometime during the years 74-78A.D.. It was presumably abandoned after a short period of time, as is the case with such encampmants.
  2. The first auxiliary fort on the site was probably occupied from late Flavian times until it was abandoned, probably demolished, perhaps during the time of Trajan but more likely during the Hadrianic period, its garrison possibly being withdrawn to serve on the northern frontier of Hadrian's Wall in the early-2nd century. It is possible, however, that the garrison was merely reduced during Hadrianic times, which may account for the second fort.
  3. A few sherds of Antonine pottery recovered from the site perhaps points to a military re-occupation during these times. This may very-likely be associated with the second incarnation of the auxiliary fort, but may also be an indication of continuing civilian occupation of an otherwise abandoned military site throughout this period, the implication being that the second fort may be dated to an earlier, Trajanic?/Hadrianic? reduction of the original garrison (see above).
  4. It is thought that the stone-built Roman or "Saxon" shore fort was constructed sometime between AD260 to 300, along with similar forts at Burgh Castle, Bradwell-on-Sea and Portchester. A coin dated to the reign of emperor Marcus Aurelius (c.A.D.175) was found beneath the Roman rampart bank in the SE corner.

Other Roman Sites in the Area

The coastal road to Nidum (Neath) proceeded twelve miles due west from Cardiff within a few hundred yards of the villa nearby at Ely (ST1476), and probably served the villa's at Whitton (ST0871) and Cae'r-mead (SS9569), both lying a couple of miles to the south of the road nearer the coast. After this the road turns obliquely west-north-west and proceeds a further eleven miles, crossing the Aventio Fluvius (Afon Ewenni) before turning again, this time heading north-north-west a further thirteen miles into Neath.

Bibliography and References

See: The Archaeology of Roman Britain by R.G. Collingwood (1930);
An Inventory of the Ancient Monuments in Glamorgan by the RCAHMW (HMSO, Cardiff) vol.I, pt.ii, pp.90-4 & fig.51;
Historical Map and Guide - Roman Britain by the Ordnance Survey (3rd, 4th & 5th eds., 1956, 1994 & 2001);
The Roman Shore Forts - Coastal Defences of Southern Britain by Andrew Pearson (Tempus, Stroud, 2002) fig.33 p.63.
Many thanks to Richard Gregory, who has a better map than me.

GoTop

This page was last modified: