Their uses, sizes, layouts and dates
Roman temporary camps survive across Scotland as earthworks, buried ditches and cropmarks. They record armies moving through the country, stopping near river crossings, undertaking construction work and returning to ground used before. Some enclosures cover only a few hectares; others extend across much of the surrounding farmland.
The familiar term ‘marching camp’ describes only part of this evidence. Archaeologists generally use ‘temporary camp’ more broadly, because a perimeter alone rarely tells us how long troops stayed or exactly what they were doing. The differences between camps are best explored through four separate questions: what was their purpose, how large were they, how were they planned, and when were they occupied?
Camp, fort or fortlet?
A camp was a defended enclosure providing accommodation for troops, generally under canvas. A fort normally housed a garrison in barracks, with headquarters, stores and other buildings. A fortlet was a smaller military post. These are distinctions in use and organisation, not simply three bands of size: a temporary camp could be much larger than a permanent fort.
‘Temporary’ should not be read as ‘occupied for one night’. Discussing repeated oven use at Kintore and Milltimber, Jones considers stays of roughly one to six weeks, depending on the interval between firings. This illustrates a possible duration, rather than supplying a fixed rule for camps. Reoccupation and oven rebuilding also need consideration. See Rebecca Jones, Roman Camps in Scotland, Callander’s Landscape (57:06 onwards).
What were camps used for?
Campaign camps offered defended accommodation while troops marched or operated in the field. Their distribution helps identify routes through the landscape, although a line of camps is not automatically the itinerary of a single army. Forces separated by decades could find the same valleys and crossing places useful.
Construction camps housed soldiers close to building work. The clearest Scottish group lies beside the Antonine Wall. Their position, relatively small size and, at some sites, relationship to other frontier works support an association with its construction. Temporary enclosures near other military bases have also been interpreted as working camps, but proximity alone cannot establish their purpose.
At Burnswark, camps flank a hillfort and the wider evidence includes Roman projectiles. This is a specialised military setting, with a strong assault interpretation. The precise sequence and date still require care. It should not be used as the model for an ordinary overnight camp.
Practice camps are another recognised category in Roman archaeology, but the Scottish evidence is much less secure. Jones’s 2011 assessment found no convincing practice camps north of the Hadrian’s Wall zone. Small size alone is insufficient to identify a training enclosure.
Why do the sizes differ?
A hectare is 10,000 square metres, approximately 2.47 acres. Jones recorded many camps between about 10 and 18 hectares, alongside proposed larger groups around 25, 45, 54 and 67 hectares. These figures describe broad archaeological patterns rather than official Roman categories.
The approximately 25-hectare group is particularly interesting because similarities extend beyond area to entrances, proportions and annexes. Other proposed series are less consistent, and some members have incomplete boundaries. Assigning a camp to a size group is therefore a judgement that can change when more of its perimeter is discovered.
The Wall-side construction camps illustrate both a pattern and its exceptions. Jones identifies a concentration around 2–2.5 hectares, but records Easter Cadder at about 1.5 hectares and Tamfourhill at about 2.7 hectares. There was no universal construction-camp size.
Nor is there a reliable conversion from hectares to soldiers. Roads, animals, baggage, equipment and space inside the defences all occupied ground. The useful camping area could also be smaller than the total enclosure. An estimated capacity is not evidence that a particular legion was present.
Layouts, gates and annexes
Many camps have straight sides and rounded corners, but their outlines vary considerably. Roman surveyors adapted their work to the terrain. Raedykes has a notably irregular perimeter, while other camps are almost square, elongated or closer to a parallelogram. An irregular outline need not be un-Roman.
Entrances could be protected by a titulus, a detached stretch of rampart and ditch in front of the opening, or by a clavicula, a curved extension of the defences. Stracathro-type gateways combine an external clavicula ditch with an oblique traverse. These features can help identify Roman military planning, but cannot independently establish a date.
Camps did not all have four gates. Dunning, Raedykes and Normandykes provide examples with six; at other sites some entrances are inferred from incomplete evidence. Our guide to six-gated camps separates recorded gateways from reconstructions. Jones also notes that the literary model attributed to Pseudo-Hyginus had six entrances, so four named principal gates should not be mistaken for a universal total.
Some camps had an annexe, an additional enclosure attached to the main perimeter. Annexes are particularly characteristic of the approximately 25-hectare group, but their contents and functions are not securely established. They should not be labelled as cavalry quarters or baggage parks without supporting evidence.
When were the camps occupied?
The campaigns of Agricola, the Antonine advance and the operations of Septimius Severus provide a historical framework. They do not supply dates for every enclosure. ‘Flavian’ refers to a wider late-first-century period and is not interchangeable with ‘built by Agricola’.
Coins and pottery can assist dating when their archaeological context is secure. A coin in a later ditch fill need not date the digging of that ditch. Radiocarbon dates also require interpretation: they relate to the sampled material and normally provide ranges, rather than a single year of occupation.
At Ayr and Kintore, excavated ovens and radiocarbon evidence support Flavian activity without fixing an exact campaign year. At Dunning, ditch recutting establishes reuse, while second-century pottery from the western titulus, found in 1992, adds evidence absent from the 2014 trenches. A Flavian foundation followed by Antonine reuse remains a proposal, not a securely dated two-stage history.
Why did armies return to the same places?
Multiple camps occur around Ardoch, Newstead, Castledykes, Glenlochar and Camelon/Lochlands. Routes, river crossings and access to military bases made these places repeatedly useful. Jones describes such concentrations as gathering grounds.
Several camps beside one another do not necessarily represent soldiers arriving together. They may belong to separate visits. Intersecting ditches can establish which enclosure came first; recutting a ditch can indicate renewal of an existing circuit. A plan combining all recorded features is not a picture of one moment.
Finding camps and reading the evidence
Aerial photography has revealed many camps through cropmarks: buried ditches alter soil conditions and the growth of crops above them. Earthwork survey, LiDAR, geophysical survey and excavation can add different parts of the picture. Failure to see an entrance or an internal feature does not prove that it never existed.
Excavation inside a camp can be particularly informative. Ovens, pits and traces of everyday activity may survive where tents and equipment have disappeared. Recording a perimeter is only the beginning of understanding the people who used it.
Why so many camps are known in Scotland
In her Callander’s Landscape lecture, Jones reports that about 82 per cent of the Scottish camps in her study were discovered from the air. This is a figure for the body of sites discussed in the talk, not a fresh count of today’s catalogue. She traces the work from Robert Melville and William Roy in the 1750s through O. G. S. Crawford, Kenneth St Joseph and Gordon Maxwell. See Rebecca Jones, Roman Camps in Scotland, Callander’s Landscape (10:35–14:51).
Jones also considers why Scotland has more recorded camps than England and Wales combined. One possible explanation is that substantial ditches were more routinely used during northern campaigning, leaving cropmarks after the banks had been ploughed away. Differences in preservation, land use and discovery also affect the map. This is an explanation to investigate, not proof that earlier armies in southern Britain dispensed with ditches. See Rebecca Jones, Roman Camps in Scotland, Callander’s Landscape (19:58–20:50).
The lecture also revisits the chronology of the approximately 25-hectare, or 63-acre, series. At Innerpeffray West the camp preceded a road, encouraging an earlier date than the traditional Severan attribution. The road’s dating and the transfer of that conclusion to other camps require care. Our dating guide sets out the evidence.
Explore Roman camps
Explore the evidence in more detail through these five companion articles:
- Camp sizes and army strength — What enclosure area can tell us, and why hectares cannot be converted directly into soldiers.
- Dating Roman camps — How finds, radiocarbon evidence and overlapping ditches relate to the Flavian, Antonine and Severan periods.
- Antonine Wall construction camps — The evidence for camps associated with building the Wall, and the limits of assigning them to particular workforces.
- Reuse and gathering grounds — Why armies returned to useful places, and how archaeologists distinguish separate visits.
- Camp gateways: tituli and claviculae — An illustrated guide to entrance defences, including cuspate and Stracathro-type gateways.
Inside a Roman marching camp explains accommodation and daily organisation. Roman marching camps with six gates examines plans and entrances. Dunning explores reuse, Ayr introduces dating evidence, and Burnswark considers an assault landscape.
Use Maps & GIS to explore the distribution, or watch Rebecca Jones’s lecture on Roman camps in Scotland. The Fairmilehead article also explains why some traditional identifications need reconsideration.
References and further reading
Rebecca H. Jones, Roman Camps in Scotland (Edinburgh: Society of Antiquaries of Scotland, 2011; open-access edition 2026), especially pp. 5–12, 29–36, 47–55, 76–89, 99–124, 195–196 and 307–308. Page numbers refer to the printed pagination. The book’s text is available under CC BY-NC 4.0; its illustrations have separate rights.
Campbell, E., Durnan, J. and Driscoll, S. (2014), Excavations at Dunning Roman Camp: Data Structure Report, especially the summary and discussion.
Arabaolaza, I., with contributions by D. Alldritt and B. Ballin Smith (2019), ‘A Roman Marching Camp in Ayr’, Britannia 50, 330–349.
Updated 13 October 2026: expanded the discussion of use, size, design and dating, and clarified the difference between archaeological evidence and campaign attribution.
