Cantonal capital of the Demetae of SW Wales. Its name (sea fortress) is attested in the Antonine Itinerary
and aptly describes its position at the tidal limit of the Towy
estuary, much used by shipping until the late 19th c.
There is no evidence of a pre-Roman settlement. The
first occupation appears to have been an auxiliary fort
founded ca. A.D. 74 on the elevated area overlooking the
river crossing, now demarcated by Spilman and King
Streets. The probable S side of the defenses is close to
the Ivy Bush Hotel and clearly belong to the Flavian
period. The vicus of the fort would by implication have
grown up to the NE in the area of St. Peter's church;
timber buildings of early 2d c. date, laid out in random
fashion, were located here in 1969. Eventually this area
was formalized into the tribal capital of the Demetae.
This was a small town of ca. 6 ha, with a massive
stone-faced rampart and an internal street grid. Changes
in modern street levels and building subsidence in Priory
Street suggested the position of the SW and NE ramparts, and the line of the NW rampart, the best preserved
of the four, was suggested by the line of a visible bank
behind Richmond Terrace (confirmed by excavation in
1968-69). Only the fourth side has not been determined
precisely, owing to modern buildings, but there is the hint
of a rampart through the garden of the present vicarage.
The original rampart proved to have been ca. 6 m wide;
it still stands nearly 2 m high in places. The turf and clay
bank was fronted by a V-shaped ditch of roughly the
same width and 3 m deep. The ditch of the first period
was filled in and the front of the rampart extended, to
allow the construction of a stone face in the second period. Extensive dumping to the rear brought the width of
the defenses to 18 m. The town wall thus underwent the
normal development familiar in Romano-British civil defenses. A terminus post quem for the construction of the
original rampart was provided by Antonine Samian.
Excavation in 1969 showed that creation of the street
grid also belonged to the Antonine period, and timber
structures were shown to have lain on either side of the
decumanus farthest N. By the 3d c. stone structures had
become more common and the buildings more complex.
A large town house built at that time continued in use
till after A.D. 320, when the area was leveled again for
the construction of an even larger building ca. A.D. 353.
Allowing this structure a normal life (there was no sign
of violent destruction), this must extend urban life in the
cantonal capital farthest W in Roman Britain into the
last quarter of the 4th c., lending some credence to Welsh
mythological associations between Carmarthen and Maxin Wledig (Magnus Maximus, emperor A.D. 383-388).
At 150 m outside the presumed position of the E gate
lies a large oval depression partly cut into a hillside.
Excavation in 1968 disclosed the site of an amphitheater;
the arena floor was covered with up to 2 m of silt. It
was constructed by cutting into the hill and using the
excavated soil to create the S bank of the cavea. The
arena is 46 by 27 m, the circumference of the cavea ca.
92 x 67 m. The arena wall was nearly 2 m thick and built
in the 2d c. Seating, however, was in timber with elaborate sub-frames to maintain the units in position on the
hillside. The site of a bath house is also known on the
Parade on the SE edge of the town. Finds are housed in
the local museum.
G.D.B. Jones, “Excavations at Carmarthen,” Carmarthenshire Antiquary
5 (1969) 1ff; 6