NGRef: SN414203
OSMap: LR159
Type: Possible British Town, Fort

Roads
Possible road: SE (19) to Levcarvm (Loughor, West Glamorgan)
E (13.5) to Llandeilo
ENE (27) to Llandovery (Llandovery, Dyfed)

"Below the peoples we have mentioned, but more toward the west are the Demetae,
whose towns are: Luentinum 15*45 55� Maridunum 15*30 54�."

The Geography of Ptolemy (extract above) in the early-2nd century AD assigns the Iron-Age Demetae tribe within the region of Dyfed and ascribes to them the two poleis of Luentinum and Maridunum. The first is the Roman name for the gold workings at Dolaucothi near Pumsaint, the second-named polis being the earliest mention of the Roman town at Carmarthen.

In addition to the passage from Ptolemy, the town also receives mention in the Antonine Itinerary, produced in the late-2nd century, as the south-western terminus of route number twelve of the British section. Iter XII is entitled "the route from Muridunum to Viroconium", and details the 186 mile journey from Carmarthen, the civitas capital of the Demetae to Viroconium Cornoviorum (Wroxeter, Shropshire) the capital of the Cornovii tribe. In this itinerary the first town along the route is named as Leucarum (Loughor, West Glamorgan), which is reported as being 15 miles distant from Muridunum.

The last appearance of Carmarthen in the classical sources is thought to occur in the Ravenna Cosmology of the seventh century, where the name Macatonion (R&C#61), is listed between the entries for Glevum colonia (Gloucester) and the unidentified stations Epocessa & Ypocessa. The R.C. is notably corrupt, and it is not inconceivable that this entry has been derived from the earlier name for Carmarthen.

The Carmarthen Forts

N.G.REF DIMENSIONS AREA
SN417204 c.600 x 400 ft
(c.183 x 122 m)
c.5½ acres
(c.2.23 ha)

There are two superimposed Roman forts lying in the area between St. Peter's Church and the site of the Medieval Castle, the smaller and later fort of stone being built in the eastern corner of the earlier (and larger) timber-built fort. The timber fort's dimensions are given above, the later stone fort was aligned to the north-west and measured some 130 x 100 metres (c.425 x 330 feet) - established by the discovery of a major roadway dating to this period at 33 Spilman Street in 1993 - and therefore covered an area of about 1.3 ha (c.3¼ acres). Excavations within the forts defences have uncovered a wealth of pottery; aside from some early Dr.29 (dated 40-75AD), the recovered samian includes examples of Flavian, Hadrianic, and Antonine wares, followed by a long gap in the pottery-record during the third century before a reappearance of fourth-century forms, and a coin of Probus and two coins of 343AD? cannot be taken as signs of occupation during the third. It would seem, then, that the fort was abandoned (or destroyed) around the Severan period and later reoccupied.

The original military defences on the site consisted of a V-profile ditch 18ft (c.5.5m) wide and 9ft (c.2.7m) deep backed by a rampart of turf and clay 19ft (c.5.8m) wide, surviving to a height of 4¾ feet (1.45m) during excavations in the north-west corner of the town in 1969. These defences were built sometime after the early-Antonine period, as attested by a samian fragment of this date which was found sealed beneath the rampart.

The original ditch (on the NW at least) was filled-in so that the settlement's defences could be constructed, which consisted of a 58ft (17.7m) wide rampart fronted by a stone wall 5¼ ft (c.1.6m) thick. The original height of these civilian defences is not known, but must have been considerable. Several buildings were uncovered in the interior, the earliest being timber buildings dated to the early-2nd century, many of which were demolished to make way for cobbled roads and stone-built town houses by the beginning of the 3rd century.

Industrial buildings, mainly of timber, and identified by furnaces and evidence of bronze-working, occupied the area closer to the town walls. Towards the interior were gathered the principle residential areas with houses constructed of stone and possessing hypocaust under-floor heating and tessallated pavements. Construction was still going on in the western corner of the town during the latter half of the 4th century.

Latin Inscription from Carmarthen

BONO R[ei] P[ublicae] NATO

"Born for the benefit of the republic."

(RIB 412; 4th century)

  1. Inscriptions bearing variations of this same text have been uncovered in Britain at: Voreda (Old Penrith, Cumbria; statue base; RIB 930); Vinovium (Binchester, Durham; RIB 1040a); Hadrian's Wall, somewhere in Cumbria (RIB 2314).

There are only two Latin inscriptions on stone recorded for Carmarthen in Volume I of the R.I.B., one dedicatory inscription (RIB 412 supra) which was found during construction of a house within the western corner angle of the fort in 1804, and another (RIB 413) for which no text has been recorded.

The 'Inventory of Carmarthenshire' published by the HMSO records two uninscribed altarstones in the front garden of a house on Priory Street, one of which was placed in the garden of a house named 'Ystrad', where it has suffered from the effects of the weather, the other now resides in the porch of St. Peter's Church. The handfull of Roman coins recovered from the site prior to 1895 ranged from a 1st brass of Vitellius (A.D.69) to a small brass of Crispus (Imp. 317-26).

There is a substantial Roman building nearby at Abercyfor (SN4217), and a Roman fortlet lies about fifteen miles to the west at Tavernspite (SN1813).

See: Inventory of the Ancient Monuments in Wales and Monmouthshire - V County of Carmarthen by the RCAHMCWM (HMSO, London) pp.248/9 & fig.191;
Historical Map and Guide - Roman Britain by the Ordnance Survey (3rd, 4th & 5th eds., 1956, 1994 & 2001);
The Towns of Roman Britain by John Wacher (2nd Ed., BCA, London, 1995) pp.391-394 & fig.175;
Britannia xxv (1994) p.248 & fig.3 p.249;
Britannia i (1970) p.125 & fig.2;
The Roman Inscriptions of Britain by R.G. Collingwood and R.P. Wright (Oxford 1965);
Air Reconnaissance in Britain, 1958-1960 by J.K. St. Joseph in J.R.S. li (1961) p.127;
All Latin translations, including any inherent mistakes, are my own.