A Roman camp can be clearly recognisable yet difficult to date. Its rounded corners and protected entrances may identify a military enclosure, while the evidence needed to place it within a particular campaign is missing. This distinction matters whenever a map labels camps as Agricolan, Antonine or Severan.
Historical periods provide a framework
| Period | How it should be used |
|---|---|
| Flavian: late first century AD | A broader archaeological and historical period. A Flavian date does not by itself identify Agricola as the commander. |
| Antonine: second century AD | The Wall-building programme provides a strong context for some construction camps. Other camps require their own evidence. |
| Severan: campaigns of AD 208–211 | A possible context for some large camps. Size and proximity to a Severan base are insufficient on their own. |
| Roman, more precisely undated | A valid conclusion where the military identification is secure but chronological evidence is lacking. |
The accounts of Agricola’s campaigns and Severus’s invasion explain why armies were operating in Scotland. They cannot allocate every surviving enclosure to a particular summer. A route proposed from literary evidence must still be tested against archaeology.
Which features came first?
Where two ditches intersect, excavation may show that one was cut after the other. That gives a relative sequence: earlier and later. It does not necessarily give a calendar date. The sequence becomes more informative if one of the structures can be independently dated.
The same principle applies to a recut ditch. Renewal after silting demonstrates a change in the life of the enclosure. It cannot, without further evidence, be translated into a return by a particular governor or emperor.
Dunning: reuse without a secure campaign date
At Dunning, the 2014 report documents ditch recutting but does not securely date the two occupations. Separate excavation in 1992 recovered mid-second-century pottery from a lower fill of the western titulus. Comparison with late-first-century pottery at Carey makes a Flavian foundation and Antonine reuse plausible. The pottery dates a deposit, however, and does not independently date both ditch phases. See Rebecca Jones, Roman Camps in Scotland, Callander’s Landscape (43:57 onwards).
A later paper on the excavations was listed by the University of Glasgow as accepted for publication in 2026. Its text was restricted when this article was prepared, so it has not been used to infer new dating results. The discussion here follows the accessible report.
Objects and radiocarbon samples
A coin’s issue date is not the same as the date it entered the ground. Pottery in an occupation deposit may help date activity, while material in an upper ditch fill could have arrived after abandonment. Residual objects can also be older than the deposit containing them.
Radiocarbon dating measures organic material, not a camp directly. The interpretation depends on the sample, its context and its relationship to the military activity. Results normally provide probability ranges, rather than the exact year a ditch was dug.
At Ayr, dates from six fire-pits span a broad range but overlap in the Flavian period. This supports a late-first-century interpretation; it does not establish occupation specifically in AD 80.
Can shape or gateway type supply a date?
Jones found considerable variation in camp outlines. Some Antonine Wall construction camps are almost square, while others are more rectangular. A square outline therefore cannot be treated as an Agricolan signature. Gateway forms are useful comparisons, but their distribution and survival complicate chronological arguments.
Similar camps may belong to related operations. That possibility becomes stronger when area, entrances, proportions, location and independent dating evidence agree. It remains important to avoid circular reasoning: dating one camp from another, then using the first to confirm the date of the second.
Innerpeffray West and the 63-acre camps
The traditional allocation of the approximately 25-hectare, or 63-acre, series to Severus is no longer a safe default. A trench at Innerpeffray West in 2007 established that a road crossed the camp’s defences after the camp had been built. In the lecture, Jones uses this relationship to argue for an earlier, potentially Antonine context for the series. See Rebecca Jones, Roman Camps in Scotland, Callander’s Landscape (45:45–47:00).
The strongest conclusion is the relative order: camp before road. The calendar date depends on the road’s identification and chronology. Jones’s published discussion, ‘Chasing the army: the problems of dating camps’, explicitly raised uncertainty over whether the recorded road was Roman. The Perth and Kinross research framework favours a second-century engineered road reusing an earlier track. An Antonine camp is therefore a reasoned interpretation, not a direct absolute date.
Nor does this one trench directly date every comparable enclosure. At Ardoch, the smaller camp of this series precedes the larger, traditionally Severan camp. That is valuable supporting sequence evidence, but other proposed members need their own assessment. Kirkpatrick-Fleming is discussed in the lecture’s second-century dating argument; pottery must be tied to the correct enclosure and deposit before assigning a date to all the camps there. See Rebecca Jones, Roman Camps in Scotland, Callander’s Landscape (42:53 onwards).
Kintore provides a stronger basis for late-first-century activity through excavation and radiocarbon evidence. Its bearing on Normandykes, Ythan Wells I and Muiryfold is comparative. Beattock’s Barnhill camps illustrate a different route to dating, through deposits overlying an earlier fortlet. These should be recorded as different kinds of evidence, rather than given equally certain campaign labels.
What a responsible date label should say
A useful description gives both the proposed period and its basis: ‘Antonine, supported by the relationship to the adjacent frontier works’; ‘possibly Severan, on comparative grounds’; or simply ‘Roman, undated’. Keeping those distinctions visible makes it possible to revise an interpretation when new evidence appears.
Sources
Jones, R. H. (2011; open-access edition 2026), Roman Camps in Scotland. Edinburgh: Society of Antiquaries of Scotland, pp. 52–55 and 109–124. Page numbers refer to the printed pagination.
Campbell, E., Durnan, J. and Driscoll, S. (2014), Excavations at Dunning Roman Camp: Data Structure Report. For the later paper’s status, see the University of Glasgow repository record (checked 13 October 2026).
Arabaolaza, I. (2019), ‘A Roman Marching Camp in Ayr’, Britannia 50, 330–349.
Return to Roman Camps in Scotland.
