Carzield Roman Fort – Cavalry in the Nith Valley

At Carzield, near Kirkton north of Dumfries, a quiet landscape of houses, lanes and fields conceals a Roman cavalry base. Its soldiers belonged to the army that occupied southern Scotland during the reign of Antoninus Pius. Here, well behind the Antonine Wall, mounted troops could watch the routes through Nithsdale and help keep the army’s communications open.

Carzield at a glance

Feature	Evidence and interpretation
Occupation	Antonine period, in the mid-second century AD. Often placed around AD 142–158; these are working dates, not securely established opening and closing years.
Historical setting	The advance under Antoninus Pius; GUARD associates the fort’s foundation with the campaigns of c. AD 139–143.
Garrison	Probably an auxiliary cavalry regiment, an ala quingenaria. Its name is not securely identified in the sources used here.
Size	152 × 170 m, approximately 2.58 hectares (6.39 acres). The archaeological record measures the fort over the ramparts, so this is not all usable accommodation.
Probable troop strength	Around 500 cavalrymen at nominal establishment; a standard ala quingenaria is conventionally reckoned at 512. The actual strength at Carzield is unknown.
Roman name	Not securely known.

Why build a fort here?

An army holding the Antonine frontier needed more than the Wall itself. Men, messages and supplies had to move through southern Scotland. Nithsdale offered an important corridor, and Carzield occupied a position where north–south travel could connect with routes across the valley.

The South West Scotland Archaeological Research Framework highlights this road junction role. A cavalry garrison made sense here: mounted patrols could cover a wider area than men on foot. Protecting communications and monitoring movement are reasonable interpretations of its position, although no surviving orders tell us the garrison’s daily duties.

When was Carzield occupied?

Pottery recovered during the 1939 excavations placed the fort in the Antonine period. GUARD Archaeology’s account of later discoveries connects its establishment with the Roman advance of approximately AD 139–143. Carzield was built on a different site from the earlier Flavian forts at nearby Dalswinton.

A working date range of approximately AD 142–158 places Carzield within the Antonine occupation of southern Scotland. The evidence does not establish the exact year the first troops arrived or the last patrol departed.

A fort built for horses as well as men

Excavation revealed a turf rampart and several defensive ditches. The national site record also records a bath-house and evidence interpreted as an annexe outside the main enclosure. Timber buildings occupied the interior.

The most revealing clue to the garrison came from the accommodation. Birley and Richmond identified a barrack with eight soldiers’ rooms and unusually large officers’ quarters. Comparing its arrangement with cavalry barracks elsewhere, and considering the space within the fort, they argued for a mounted regiment of roughly 500 men.

Such a base needed room for horses as well as soldiers. Stabling and the practical demands of keeping a mounted unit supplied help explain why a cavalry fort cannot be assessed using troop density alone. For the usual buildings and their functions, see Roman fort layout.

How many troops were stationed here?

Around 500 mounted soldiers is the preferred working estimate. Paul Holder’s study of auxiliary organisation lists 512 as the nominal strength of an ala quingenaria. That is an organisational model, not proof that exactly 512 men were present at Carzield: vacancies, detachments and other absences could change the number on site.

The unit’s name remains uncertain. No named regiment is assigned here simply because it served elsewhere in Roman Britain. See Auxiliary units in Roman Scotland for the different types of troops, and Explore the Roman Army for the wider military organisation.

A soldier’s bowl: the inscriptions

One small find brings us closer to an individual. RIB 2501.10 records a samian bowl fragment recovered from an Antonine rubbish deposit in 1955. Scratched beneath it were part of a name and the abbreviated rank sesquiplicarius.

A sesquiplicarius received one-and-a-half times basic pay. RIB’s discussion of the rank, alongside a comparison from Chesters, notes its strong association with mounted units. The Carzield graffito therefore fits the cavalry interpretation. The owner’s full name is uncertain, and the text does not identify his regiment.

Two further records show other uses of writing: RIB 2501.773 is a fragmentary graffito on a samian plate, while Britannia 38.33 records a lead weight found approximately 150 m north-east of the fort in 2006. Its mark means “half a pound”, although its surviving weight is appreciably lower. Neither names a military unit.

Boots, weapons and life behind the frontier

Modern service trenches have offered narrow glimpses into the buried fort. GUARD’s investigations recovered a javelin head retaining traces of its wooden shaft, pottery and the hobnails of a Roman boot. The leather had disappeared, but the arrangement of the nails preserved evidence of the footwear.

Samian pottery connected the occupants with suppliers in Gaul. Heating-system tiles and cereal remains add to the picture of a garrison supplied with food, equipment and some domestic comforts. Carzield’s importance lies partly in these everyday traces: the frontier depended on feeding, equipping and housing troops far beyond the line of the Wall itself.

References and further reading

• Historic Environment Scotland: Carzield, NRHE 65890 — location, defences, excavations, bath-house and finds.
• Birley and Richmond, ‘The Roman Fort at Carzield’, Transactions, volume 22, pp. 156–163 (PDF) — original excavation report and cavalry interpretation.
• Warren Bailie, Wellington Bridge, Carzield Roman Fort, Kirkton, Dumfries (2011, PDF) — GUARD Archaeology report, Project 3410.
• GUARD Archaeology: ‘Boots on the Ground’ the Roman way — readable account of later finds, with photographs and a fort plan.
• ScARF: Roman campaigns in South West Scotland — introduction and settlement — regional setting and interpretation.
• RIB 2501.10 — the sesquiplicarius bowl; RIB 2501.773 — samian plate graffito; Britannia 38.33 — inscribed lead weight.
• Paul Holder (2003), ‘Auxiliary Deployment in the Reign of Hadrian’, pp. 101–145, especially p. 120 — nominal unit strengths; used from the supplied PDF. This is background on organisation, not evidence identifying Carzield’s regiment.
• Romans in Scotland fort-size spreadsheet — fort dimensions, checked against the supplied spreadsheet copy.

← Return to the complete guide to Roman forts in Scotland