Gamesley (Ardotalia or Melandra) Roman Fort

Bath House, Flavian Auxiliary Fort (AD 69–96), Mansio and Vicus

Melandra Roman Fort, also known by its Latin name of Ardotalia, was originally built in Flavian times (c.75AD) as one of a network of outposts which connected Manchester to the wider road network. It continued to be occupied until at least the end of the second century, as Trajanic and Hadrianic pottery has been recovered from the site, along with Antonine samian ware by the potter Advocisus.  The fort had its original clay rampart reinforced with a stone wall and gateways during late-Hadrianic/early-Antonine times, and occupation into the third and fourth centuries is suggested by the discovery of coins of Marcus Aurelius and the usurper Carausius, the last dateable coin being one of Magnus Maximus.

Melandra Castle,  is of the common ‘playing card’ shape, covers approximately three and a half acres. It has a close resemblance in plan and size to the fort at Hardknot pass.  The defences consisted of an outer line of ditches (some of which were filled in while the fort was still in use) and stone walls, backed by earth ramparts. It has a stone wall 5 feet thick, contemporary with a 15-foot clay bank ; of its four gates the decumana (south gate) is single, the rest double, and there are no guard-chambers. All these features follow the Hardknot pattern. The headquarters building is of stone, and post-holes show that the barracks were of wood. The evidence of date was taken by the excavators to suggest the late first century; but the structural features point rather to the early second. The excavations hitherto carried out do not enable us to assert positively that there was not an earlier fort on the site.

The Headquarter Building

The H.Q. or PRINCIPIA seems to have been the only building made wholly of stone: it contained the Assembly Hall (which will hold a full century of men); the C.O’s office; the shrine and the unit pay and records office.   The small buildings which line the courtyard were probably for ‘unit stores’ of the more valuable sort.

The SHRINE, the centre room at the rear of the building, has a floor of crushed tile: the Imperial Altar was kept at the southern end of the room and smaller altars, one for each century, stood along the walls with the various standards of the unit.   The top of the Imperial Altar has survived but the altars themselves, which were renewed annually, the previous one being securely hidden, have never been found: five of the small uninscribed altars are still in existence.   At the rear of the building there is a small cobbled pavement which marks the spot from which the ‘official witnesses’ could watch ceremonies in the shrine through the grille above the Imperial Altar.

The room at the western end, the C.O’ S office, has a door leading to the platform at the end of the Assembly Hall: the room at the other end, judging by the lines of nails found during excavation, once had a board floor and at some point in its life had a hearth in the centre of the room.   Some indications have been found of an earlier timber building on a different alignment to the present H.Q.

Fort Bath House

Protected by steep slopes to the north and the west, a small bath-house stood outside the north-west corner of the fort. A bath-building, of coursed sandstone slabs set in mortar, was discovered outside the north west corner of the fort in 1973. It proved to be of Reihentyp, with an apsidal hot room, a warm room and a cold room. The hypocaust pilae were of tile. Much of the masonry had slipped down the hillside. This first phase appears to be Flavian but additions were made early in the 2nd. century about 120 A.D. A secondary wing of finely dressed gritstone on a base of sandstone slabs, was located to the south of the cold room in phase 1 running N-S. It consisted of two rooms, one heated and one unheated. The pilae were of squared sandstone blocks and the flues of box tiles. Between AD 120-140 during a third phase of building a possibly heated room, maybe a dressing-room, was added to the east of the secondary wing. This room was 5.6m by 5.0m and was built of sandstone blocks.


Part of the civil settlement adjacent to the Roman fort was excavated in 1963. A ‘MANSIO’ or ‘posting station’ lay between the fort and the present Melandra Castle Road (160′ x 60′) which was partly excavated in 1966. It was constructed of timber and fronted by a road to the W, and backed by a rampart and ditch system. The excavations indicate demolition rather than destruction as the ultimate fate of the building. A date for the demolition of c.AD 140 was arrived at from pottery evidence.


Further excavations in  1969 concluded that there was a sizeable defended vicus which lies under the present Gamesley Estate.. Considerable evidence for iron-smelting and lead and glass working came from the vicus

The decline of the Fort

Melandra was used for centuries as a convenient source of stone, rubble and gravel for local builders: re-used stone has been found at Woolley Bridge; Melandra Farm and in various walls throughout the district. It is believed that considerable quantities of the stone were used in the building of Mottram Church and that the large amounts of gravel taken from the sides of the site for road levelling the 18th – 19th century may explain some of the casual finds of roman material last century.

Classical reference for Zerdotalia

The only classical reference for the Roman name of the Melandra Castle fort is the Ravenna Cosmology of the seventh century. In this work the name Zerdotalia (R&C#108) appears between the entries for Aquae Arnemetiae (Buxton, Derbyshire) and Mamucium (Manchester, Greater Manchester). This name has been associated with the Melandra Castle fort, but is thought to be somewhat corrupt, and the name now commonly accepted is Ardotalia.

It is thought that the Roman name for the fort was ARDOTALIA (the place of the high, dark hill – ‘talia’ is a celtic word for a steep hill which passed into Latin but it has also been suggested that the fort and the river ETHEROW both took their name from the winding, heather covered valley).

Military Units at Ardotalia Fort

Ardotalia was constructed by  Cohors Primae Frisiavonum – The First Cohort of Frisiavones.

RIB 279 - Centurial stone of Valerius Vitalis

From the First Cohort of Frisiavonians the century of Valerius Vitalis (built this).


No commentary.

RIB 280 - Fragmentary building dedication to the emperor Hadrian?

For the Emperor Caesar … Trajan ..

IMP C[...]
T[...  ...]

No commentary.

They were recruited north of the Rhine – as this area was in revolt in 96 AD, the raising of this cohort was probably done some time between 98 AD and 100 AD: there is no mention of their being in Britain until early in the second century and it is likely that they were brought in as a re-inforcement in the early Trajanic period. Evidence for the existence of this unit exists not only from the building stone found at the site but also from various diplomas and other Roman writings.

The Ist Cohort of Frisians being a ‘1st Cohort’ would have roughly a thousand men for the 1st Cohort included the specialist craftsmen such as carpenters and stone-masons who could do the skilled work of building – which explains why the ‘centurial stone’ from the walls of Melandra is of the Frisians and not the Bracara.

Cohors Tertiae Bracaraugustanorum – The Third [Infantry] Cohort from Bracara Augusta came from the colonies of BRACARA AUGUSTANOREM (BRAGA in Portugal) and were probably Iberian Celts. They were transferred from the Legionary Headquarters on the Rhine to Caerleon in 89 AD and seem to have been attached to the XX Legion Valeria Victrix at Chester.

Whilst it is unknown which of these Cohorts manned the fort, it seems more likely that the 3rd Cohort of Bracara Augustani performed this duty, as they were from a hilly region and so were more experienced in holding terrain such as that found around Glossop. The Frisiavones were from low-lying lands beyond the Rhine and so may have been divided between the lower terrain of Manchester and Northwich.


Investigations in the 1960s discovered a large rectangular building immediately to the east of the fort which has been interpreted as a mansion house. This was demolished circa-AD 140 which may also represent the date the fort was abandoned.

References for Ardotalia

  • Britannia i (1970) pp.283/4 & fig.7;
  • The Roman Inscriptions of Britain by R.G. Collingwood and R.P. Wright (Oxford 1965);
  • The Archaeology of Roman Britain by R.G. Collingwood (Methuen, London, 1930). 

Map References for Ardotalia

NGRef: SK009951 OSMap: LR110

Roman Roads near Ardotalia

SE (12) to Brovgh-on-noe (Brough-on-Noe, Derbyshire) W (13) to Mamvcivm (Manchester, Greater Manchester). Doctor’s Gate Roman road ran between Ardotalia fort and Navio fort at Brough-on-Noe.

Sites near Gamesley (Ardotalia or Melandra) Roman Fort