CIMRM 2374 - Mithraeum. Caernarvon, Britain.

From Boon.

A Mithraeum was discovered at Caernarvon in 1958, outside the Roman fort, and excavated in 1959. Built in the 3rd century and abandoned then, it was destroyed at the end of the 4th century AD. Finds were few, but four small altars were found, one inscribed, and some interesting ritual iron-work. A housing estate now covers the site.

1. CIMRM entry

2. Details

The Roman auxiliary fort of Segontium (modern Caernarvon) in North Wales has a vicus on three sides of the fort. A mithraeum was discovered in 1958 during clearance for development of an area on the eastern corner of the fort, and excavated in 1959. The site is located 150 m east of the fort, on the western side of a shallow marshy valley, oriented 30 degrees east of north. The ancient parish church of St Peblig stands on the opposite side of the same valley.

The Mithraeum is 48 ft x 21.5 ft, and displays three structural phases, all in the 3rd century A.D., followed by a period of disuse, and then by deliberate destruction in the late 4th century. But the dates are unclear because of the lack of finds.1 The layout is normal; an ante-chamber, a long sunken nave or naos, flanked by benches or podia, and an alcove for the tauroctony. No fragments of this or any other Mithraic statuary were found. Four small altars were found, one only inscribed, with the name of a centurion. Some ritual iron-work was found.

UK Grid reference: SH/4875, 6240.

2.1. Pre-Mithraic building

Cobbles and some slab-flooring were found from an earlier Roman building. A trench 100 feet away also contained cobbles, but also two pieces of samian ware pottery (types: Dr. 37 and Curle 11) dated to the Flavian period of the late 1st c. AD. Chips of other late 1st century pottery were found over and around the cobbles. All this indicates a building of some kind from the end of the 1st c. A.D. The low situation - 20 feet below the fort - suggests a bath-house but there is no actual indication of the nature of the building. The building was then abandoned and as much as 14 inches of brown soil had accumulated over the floor before the Mithraeum was constructed.

The stone work was of poor quality; much more so than the masonry at the Roman fort, indicating the private nature of the construction, for otherwise the same masons could have been used.2

2.2. Phase 1

The Mithraeum was built of stone, probably up to the eaves, using untrimmed beach boulders up to 2 ft long. The entrance was in the south wall, 5 ft wide, and with a sill of soft pink sandstone, of which fragments were found. The building was roofed (in all three phases) with local welsh purple slate. Each slate was about 16 x 14 inches, and 1/2-5/8 in. thick. The slates were hung edge-wise and held by two iron nails. Slate roofing appears in the fort first in the 3rd century AD, and not before, although it appears at Chester ca. 100 AD. The ridge was clad with imbrices, portions of which were found in the rubble.

The ante-chamber or narthex was 6 x 18 ft, and totally destroyed long before excavation. There may have been a wooden shelter outside the door, as possible post-holders were found. The main chamber was 35 x 18 ft. The cult niche on the axis was rectangular and 8 x 1.5 ft. The nave and benches were 30 ft. long; the nave was 8 ft wide, the benches probably 5 ft although subsequent phases leave this unclear. Steps of brick led down from the ante-chamber to the nave floor. The benches ran to the end of the room, so had to be ascended from the side, rather than the end. The benches were 15 in. above the nave floor, and of earth, covered perhaps by a thin layer of fine gravel.

There is a foundation deposit in the alcove, containing a worn denarius of Faustina I, puts construction at around 200 AD. The fort was being repaired extensively between 198-209. The lower parts of two colour-coated beakers were also found; presumably ritual beakers laid down as part of the foundation deposit.

Fragments of 12 small lamps were found in the phase 3 bench and floor construction, all in the northern part of the building where the light would have been needed.

2.3. Phase 2

This rebuilding involved the insertion of a series of timber columns on each side, probably to support the phase 1 roof which was probably of green timber and so may have become unsafe. Ten column bases were found, evidently reused.

A drain was also inserted to contain water from a spring disturbed in the building of phase 1. It terminates half-way along the nave, against the middle of the west bench. At the Housesteads Mithraeum a spring-fed cistern was found in a similar position; no doubt a similar tank existed here.3

The phase 1 floor was patched crudely, rather than replaced, in such a way as to suggest that phase 1 and phase 2 were not far apart in time. No other data on the date of phase 2 is available.

2.4. Phase 3

The roof did eventually collapse, and some of its fallen slates were incorporated in the floor of a completely restored Mithraeum. This raised the floor around 9 inches, and left the drain blocked. The new flooring was 4 inch flagstones of weathered limestone, for some feet from the entrance northward, and, after a straight joint, cobbles thereafter, up to a second joint, beyond which the flooring was whatever rubble was in position. The benches were raised to preserve or even increase their height above the floor, to around 12 inches. The surface was again of earth. A prominent sill or coping about 6 inches above the level of the benches was installed on either side.

At the north end of the east bench was a cobbled area, 3 ft x 5 ft., level with the sill, of unknown purpose unless to support a statue. The state of the evidence on the west bench leaves it unclear whether there was a corresponding area there, but none was found.

Stone was inserted at the north end of the nave, probably, by analogy with Dura-Europos, to form a flight of steps up to the alcove. The stone had gaps, which would permit lighting to be inserted to illuminate the relief and possibly the main altars.4 There may have been seven steps, each tread being 6-7 in. wide, and the whole 4 ft in total.

The benches contain pottery of late 3rd-early 4th c. patterns. The garrison was withdrawn between c. 290-350, and phase 3 is followed by a period of desertion, so it is likely that phase 3 ended with the departure of the soldiers in c. 290.

Four subsidiary altars were found. They are 15-20 in. high, and made of local grit. Three were found smashed and scattered throughout the nave under the rubble. The fourth lay broken where it had fallen. Altar 1 was inscribed with four letters, "S.C.O." (probably the initials of the dedicator) followed by a backwards C, the sign of a centurion. But the remains of altar 1 are particularly widely-scattered and the fragment with the name of the god was not found at all, presumably destroyed. Altar 2 had a carefully smoothed surface, which may have had a painted dedication on it.

The post bases from phase 2 were reused as altar bases. A base at the end of the eastern bench was found, probably for a statue of Cautes or Cautopates.

Some iron-work was also found. It consists of a cross-shaped iron object, two iron sockets, an object resembling a wall-sconce, a short length of squared iron bar with loops attached to each face, and a small bronze bell shrouded in rust. X-rays show that there was no decoration. They were all found together, and presumably formed a single object; probably a two-branched candelabrum, which could possibly be fixed to a pole and carried.

2.5. The end of the Mithraeum

The rubble in the nave was embedded first in brown soil and then in burned ash and charcoal, in which large fragments, indicative of squared oak beams of some size, were found. Below this lay a layer of brown earth, nowhere more than 1/4 in, and often half that, on top of the phase 3 pavement. It was not contaminated by the burnt material above it. This shows that the Mithraeum was abandoned for a period. Little slate was found in the phase 3 destruction layer, suggesting that the building was roofless after it was abandoned, and the slates reused for other purposes. The absence of any statuary suggests that this was removed c. 290 by the garrison.

The fort was reoccupied 60 years later, by soldiers serving a Christian emperor, Constantius II; and a worn, burnt coin of Constans in the burned layer shows that the burning happened in the later years of the 4th century. Boon suggests that the nearby church founded by the early British saint (no other church bears his name) "Peblig, son of Maxen Weblig", i.e. Publicius, son of Magnus Maximus, may indicate that the saint presided over the destruction of the Mithraeum.5


  • G.C. Boon, "A Temple of Mithras at Caernarvon-Segontium", Archaeologia Cambrensis 109 (1960) pp. 136-178. This contains a full and detailed list of finds.
  • R. E. M. Wheeler, "Segontium and the Roman Occupation of Wales", Y Cymmrodor, XXXIII (1924). This is the publication of the Segontium site. Online here.

1Boon, p.138: "The Mithraeum measures 48 ft. by 21 1/2 ft. overall and displays three structural phases preceded by a distinct, no doubt nonreligious, usage of the site, and succeeded by a period of desertion culminating in desecration. This clear sequence (Fig. 6) cannot be transformed with equal precision into terms of chronology, for finds were few and badly preserved. Internal evidence shows that the life of the temple fell into the third century, which at Segontium was opened by a period of great activity at the fort, and was closed by the apparent withdrawal of the garrison, the Cohors I Sunicorum. The destruction of the Mithraeum took place in the later fourth century."
2Boon, p.161.
3Boon, p.149, n.28: "Housesteads, pp. 259 (plan), 260. Tanks, or other provision for a supply of water, commonly occur in Mithraea: T. & M., I, 297, II, passim; Corpus, passim, for plans and notes of installations. Ptuj, Mithraeum II (CIMRM 1509), is particularly apposite." and n.29: "Despite the dryness of the season, the northern part of the nave was waterlogged at drain-level during excavation, and the drain, once cleared, filled over one day to the brim with water. The southern part of the nave remained dry."
4Boon, n.37: "The altar shown in Pl. XIV (c), found many years ago in the footings of Llanbeblig church (Segontium, p. 127) is of similar date and type to those found in the Mithraeum, and may well have been taken thence to the church-a very short distance. It will have been about 3 ft. high originally, and may have been one of the main altars. Its wreath or crown is possibly of Mithraic significance, but this is rather special pleading.

On the possibility of concealed lighting, jeux de lumière most certainly had their place in Mithraea: T. & M., I, p. 81, and cf. Carrawburgh, pp. 37-8, Pl. xiiB, for an altar with a pierced radiate bust in relief, the holes communicating with a cavity at the back for a lamp to stand in. Also to be cited are the Housesteads relief (Housesteads, p. 256) showing the Egg-birth of Mithras, and also a leaden plaque from the S. Prisca Mithraeum (Corpus, mono 494) with a pierced radiate bust of Mithras."
5Boon p.156, and n.47: "See S. Baring-Gould and J. Fisher, Lives of the British Saints, IV (London, 1913), pp. 86-7. The strong traditional linkage of Maximus with Segontium must surely have some foundation in fact. There seems no reason why Peblig should not be 'historical'."

comments powered by Disqus