Gwynedd Archaeological
Trust Regional Historic Environment Record
Segontium Roman Fort, Caernarfon
Primary Reference Number (PRN) : 3089 Trust : Gwynedd Community : Caernarfon Unitary authority : Gwynedd NGR : SH48536240 Site Type (preferred type first) : ROMAN FORT Status : Scheduled Monument , Cared for by the State
Summary : The Roman auxilliary fort of Segontium was founded by Agricola in AD77 or 78. It started life with defences of earth and wood and with timber buildings inside. These original defences were rebuilt in stone in the first half of the 2nd century and the internal buildings were gradually replaced by stone structures from the same period onwards.
Named after the nearby river Seiont, the fort was orginially designed to house about 1000 men and is clearly identified in late Roman military route descriptions. The fort is laid out in typical fashion with 4 gateways, central headquarters buildings and ranks of barracks, each housing 80 infantrymen or 64 cavalrymen, and stores. The defences consist of a rampart fronted double ditch, now filled in.
The barracks in the south-eastern quarter of the site were demolished in the 2nd century and replaced by a large courtyard house and bath building, possibly as the accomodation for an imperial procurator - a high ranking official in charge of mining in the area. Segontium was garrisoned until the end of the 4th century - longer than any other fort in the area - which may also suggest that it had a role in the organisation of mining operations and in the collection of taxes.
The site has been extensively excavated over the years, notably be Sir Mortimer Wheeler in the 1920's and most recently by P.J.Casey and J.L.Davies in 1975-6.
Description : The Roman fort of Segontium was built in AD78 and abandoned in AD383. It was excavated in 1921-3 by Wheeler. A Mithraic temple, 150yds E of the fort, was excavated by Boon in 1959. <1> <2>
Further excavations were carried out in the S part of the fort in 1975-6, by P. J. Casey and R. L. Davies.
Foundations of barrack blocks and administration building clearly visible 0.3m high. The wall of the fort is best preserved on the SW-SE side where it is 2.8m high and 1m thick. About 1500m square of the SE corner of the fort was excavated in 1975-6, revealing a complex history of structures dating from the 1st to 4th centuries.
Barracks of the earliest phase are visible but 3 main buildings have been excavated:
i) Building of 13 rooms, 36m square with verandah and central well. probably the commander's quarters in the late Hadrianic or Antonine period.
ii) Unfinished bath house with no stoking pit. Drainage system running to S pre-dates the baths.
iii) Small bath house discovered in C19th then lost, now relocated. Dated by coins in the stoking debris to 350-400AD. The long curving feature running N-S was shown to be a very late revetted drain post-dating all other features in this area. It seems this area of the fort was unoccupied for much of the C3rd.
Excavations in 1978-9 in the praetentura found the earliest levels to be Agrocolan, beneath timber barrack blocks of the Flavian and Trajanic periods.
Later, the barrack blocks were found to have replaced a multi-period native settlement.
Excavations outside the NE gate showed that the ditch and causeway of the original Flavian fort remained in use until some period shortly after AD330. The ditch was then recut and a new outer ditch dug. Access to the gate was now across wooden bridges, and the structural evidence for these was found in association with the inner ditch. New evidence for the nature of the Flavian timber gate was also found. <3> to <22> (Casey, 1971. Casey & Davies, 1975, 1976, 1977, 1978, 1979)
Segontium Roman Fort, Caernarfon SH 485 624. In March and April 1991, the Gwynedd Archaeological Trust conducted excavations and a watching brief during the installation of a new water-main within the Roman fort of Segontium. The project was financed by Welsh Water plc.
A watching brief was maintained on the digging of a trench for a new pipe along the N side of the A4085 Caernarfon-Beddgelert road, which runs through the fort. The pipe was connected to an existing main within the fort, from which the trench ran S-E for some 135m, crossing the line of the rampart approximately at the site of the SE gateway.
Archaeological deposits were best preserved along the W half of the trench, where a number of features were recorded, including wall foundations, pits and spreads of burnt material. Pot sherds, fragments of iron and lead, and broken tiles were recovered from these features.
There was much modern disturbance where the trench crossed the line of the S-E rampart, and no trace was found of any structures at the presumed site of the gate, but beyond this two ditches were recorded, 5m and 3m wide, and 3.5m apart. The trench appeared to cut across the terminal of at least one of these, thus defining the N edge of the road into the fort.
Three trenches were excavated at the back edge of the SE wall of the fort, between the reservoir at the E corner and the modern road.
Immediately S of the reservoir a trench (A) across the line of the pipe enabled a section to be recorded through part of the defences. The evidence here corresponds well with the results of the 1975 excavations towards the S corner of the fort (Davies and Casey in AW 15, 42-43). A bank of stacked turves, at least 2m thick at its base and surviving 1m high with clay and gravel dumped against its inner edge, appeared to represent the Flavian rampart, although it produced no finds. The front face of this bank has been cut away, presumably when the defences were faced with stone, but again this could not be established without question as a Victorian water-main runs out through the line of the fort wall at this point.
A second trench (B) was opened roughly midway between the E corner of the fort and the site of the SE gate. The new pipe has already been laid here, and excavation was limited to recording a longitudinal section of part of the rampart about 4m back from the outer face of the wall. Lenses of clay and gravel confirmed the dump construction of the bank recorded in the trench close to the reservoir, but more importantly the record section produced evidence of three posts, with centres about 2.3m apart, running down through the bank material. The presence of the new pipe made it possible to examine these posts fully, but they did not appear to have been inserted into the bank after its construction. Their position suggests an interval tower, mirroring that recorded in 1975 midway between the SE gate and the S corner.
In the third trench (C), immediately N of the road, most archaeological deposits had been eroded away, although at the N end traces of the clay bank survived overlying a soil similar to that recorded at the bottom of the trench S of the reservoir.
The Trust expects to carry out further work in 1992 during the replacement of a valve chamber adjacent to the reservoir, after which a fuller report will be prepared. S. D. Boyle, Gwynedd Archaeological trust. <23>
Segontium was a 2.27 hectare fort forming part of a network of forts in Snowdonia. Caset and Davies (1993. 10) conclude that the "balance of probability indicates that the site is Agricolan and dates to 77 AD or shortly after." The fort would have been built atfer Gnaeus Iulius Agricola suppressed a rebellion of the Ordovices and occupied Anglesey (Nash-Williams, 1969, 5). Segontium would probably have been garrisoned by a 1000-strong infantry unit commanded by a senior officer, and in the early period it was the largest fort in North Wales (Casey and Davies, 1993, 10). At the end of the Trajanic period (c. 117 AD) the garrison seems to have been reduced with barracks being demolished and a decline in pottery and coins (Casey and Davies, 1993, 11). Although the garrison was reduced several high status buildings were constructed, possibly for a procurator overseeing mining in north Wales (Casey and Davies, 1993, 13-14). In the late 3rd and 4th centuries the fort seems to have been more densely occupied. the supply of coins seems to have ended around 393 AD and Casey (1993, 132) suggests that the garrison was withdrawn by Eugenius, who briefly held the Western Empire, to help defend his position. (Parry, 2013).
“ This stone had been re-used as a paving slab in the late structure within the eastern corner of the sacellum. Under an adjacent slab of the same floor was a coin of Valentinian I (364-375 A.D.). The letters are very roughly cut, but the strokes are certain, and are IANR the punctuations which appear in and near some of the strokes are probably accidental. At first sight the letters seem to read M(arcvs) Avr(elivs), or the like the first letter, however, is not an M, and either IANR or IAA/R are probably intended. In any case, no interpretation of the fragment is possible. Below these letters is the stroke of another (incomplete).” (Wheeler, 1923) (Photograph on p. 1).
Large number of Samian pottery sherds found within the walls of Segontium. Wheeler includes illustrations and photographs of the sherds, along with a brief description of each, including makers' marks (Wheeler, 1923).
Coarse Roman pottery found in and around the fort. Wheeler includes diagrams and descriptions of each sherd (Wheeler, 1923).
Bronze Roman brooches, buckles, clasps and pins found at Segontium. Wheeler includes illustrations of each (Wheeler, 1923).
Iron Roman arrowheads, spearheads, knives and tools found at Segontium. Wheeler includes a photograph of the finds (Wheeler, 1923).
“Fragmentary bronze fibula, found outside the south-west wall of the fort. Hollow trumpet-shaped head with lug (rudi- mentary loop) at top and, on each side, a projecting socket to receive axial rivet for the spring, which is missing. At the summit of the bow are lateral expansions somewhat resembling an Amazon's shield and containing remains of blue enamel. The foot is incomplete, but may have terminated in a hollow ring. Type very similar to Curle, Newstead, Pl. LXXXVII, 26, probably second century." “ (Wheeler, 1923) (Illustration on p. 25).
Animal bones found in Roman strata around Segontium. In the well were several oxen, pigs, parts of a horse and several birds. A rubbish pit from the fourth century contains red deer. Various other bones were found scattered around the site (Wheeler, 1923).
“Vicarage site” placed in the guardianship of the Ministry of Works, Ancient Monuments Branch in 1953 with a 99 year lease (Inspector, 1957).
“The Ministry has purchased Llanbeblig Vicarage and its garden containing the southern part of the Roman fort. The Vicarage has been demolished, and a start made on preserving the south wall of the fort which forms the boundary of the property.” (The Inspector, 1957).
Finds at Segontium: Barrel staves, amphorae with inscriptions, fragments of cut and stitched leather, a boar-tusk ornament, coins, glass beads, and a golden brooch. Article includes illustrations of each, and a diagram of the Roman barracks (Boon, 1975). Analysis of coin hoards found at Segontium in 1920-23 (Boon 1976).
Description and diagram of the sword and chatelaine found at Segontium (Hughes, 1910).
Further analysis of the sword found at Segontium, mentioned on pp. 324-325 (Baynes, 1910).
Description and photograph of the vessel found at Segontium (Roberts, 1910).
Discovery of engraved stones and slates, domestic refuse, and a coin, all from the Roman period (Mealy, 1846).
Analysis of a corn depository at Segontium (Jones, 1846).
Inscribed stone and coins found at Segontium (Foster, 1846).
Report on excavations at Segontium. Including list of finds and diagrams of the site (Arch Cam, July 1846).
More coins found at Segontium (Foster, 1847).
Description and illustration of an inscribed stone found at Segontium (Foster, 1853).
An account of Cambrian Archaeological Association excursion to the fort and museum in August 1937 with detailed, illustrated description and interpretation. (Lloyd Jones 1937).
A detailed account of the excavations which includes a full list of coins found as a hoard during the excavations. (Wheeler 1922).
Excavation report (Hughes 1921).
Detailed description and drawings relating to various excavations and finds at Segontium (Simpson, 1962). Prolific find of Valentinianic coinage for period 25-27 at Caernarvon [sic] Fort (Waddelove, 1985).
Segontium (Anon, 1919).
Excavations at Caernarvon 1976-1977 (White, 1985).
As far as the periods of occupation are concerned, coh. I Sunucorum was at Segontium in the time of Severus (and doubtless for some time thereafter). ( The elder Pliny mentions the Sunuci as a people of Lower Germany.) There must have been considerable movements of alae and cohorts within the province, not merely in the reign of Vespasian, when the conquest of Wales and the North meant that virtually every unit of the army of Britain was transferred to a new fort, but in the early years of Hadrian (as a result of the building up of his new frontier in the north of England) and again in the early years of Pius (when the Antonine Wall in Scotland was built, with its network of supporting stations) and on the establishment of each new frontier in the north of the province, there must have been some reduction of the garrisons in Wales, and transfer of alae and cohorts thence. But in the first few years of Trajan's reign it may be permissible to suppose that there was a relatively stable situation, already indicated by the archaeological evidence for a widespread policy of converting legionary fortresses and auxiliary forts from earth and timber (the materials which sufficed in a period when the definitive distribution of garrisons had not yet been decided upon) to stone (the outward sign of what was intended to be a permanent system). (Birley, 1952).
The report discusses the findings of the excavations in the South East part of the Praetentura: The internal structures were originally timber structures but were later replaced with stone. Coin finds indicate that there was a military presence after the originally thought abandonment date (AD 383) (Arch Cam, 1980).
Segontium was identified by Professor Rhys as ‘Caer Seon,’ a place associated with some of stories about Gwydion, a figure in Welsh mythology. John Rhys, the author of this article, argues that Professor Rhys’ “excellent note on the philology of the word” (Rhys, 1888) makes his suggestion plausible. Though he does go on to say that “the ancient fortress crowning the hill above the town of Conwy is known to this day as Caer Seion.” (Rhys, 1888).
Description and drawings of pottery fragments found at Segontium. Finds made prior to first formal excavation of the site (Turner, 1894).
Description and pictures of a bronze Cupid found at Segontium. Find made prior to first formal excavation of the site (Peake, 1893).
Discoveries at Segontium, excavations in the garden of the Vicarage. Description of building and samll finds (Mealy, 1846).
Further discussion of findings at Segontium (DW, 1846).
Recording of the north-west wall of Seontium fort prior to consolidation was underaken by Tim Morgan on behalf of Cadw. Excavation of an area of the north-west gate was undertaken, largerly clearing out Mortimer Wheeler's trenches of the 1920's, in an attempt to understand the original Roman and 1920s reconstructed masonary, prior to consolidation. No facing survived for the two guardroom walls which sat upon the natural clay. The apparent apsidal end of th south-west guardroom (as interperted by Wheeler) could not be confirmed. (Halfpenny, 2001).
Events : 42684 : Segontium and the Roman Occupation of Wales (year : 1923) 42685 : Segontium Excavations at Caernarfon 1976: Interim Report and Visitor's Guide (year : 1976) 42686 : Segontium (year : 1991) 44421 : Llys and Maerdref: an Investigation into the Location of the Royal Courts of the Princes of Gwynedd (year : 1995) 44557 : Early Celtic Societies in North Wales (year : 2010) 43333 : Field Monument Warden Site Visit: Segontium Roman Site (year : 1985) 44824 : Former Ysgol Yr Hendre, Ffordd Eryri, Caernarfon: Archaeological Trial Trenching and Limited Desk-Based Assessment Report (year : 2016) 44824 : Former Ysgol Yr Hendre, Ffordd Eryri, Caernarfon: Archaeological Trial Trenching and Limited Desk-Based Assessment Report (year : 2016) 40541 : Roman Fort Environs (year : 2001) 40536 : Roman Fort Environs 2000/2001 (year : 2002) 45272 : Unmanned Sites Interpretation Installations - Watching Brief Report. CN006 Segontium Roman Site, AN011 Trefignath Burial Chamber, AN016 Holyhead Mountain Hut Circles (year : 2016) 45969 : Segontium Pods, Caernarfon, Heritage Impact Assessment (year : 2019)
