The Antonine Wall was built across the narrow waist of Scotland from Old Kilpatrick on the River Clyde to Bridgeness on the Firth of Forth, whose strategic potential had been recognised in the Flavian period (see Pre-Antonine military presence). For much of its length the Wall followed the Midland Valley, sitting on its southern slopes overlooking the marshy ground to each side of the Rivers Carron and Kelvin. To the east, it was situated on top of the raised beach overlooking the Forth. To the west, beyond the River Kelvin, the Wall line utilised the drumlins of the Clyde Valley before ending on the north side of the river.
With some exceptions, the Wall line follows the most elevated, north-facing ground, resulting in frequent changes of direction, but it is not always placed in the most advantageous defensive position there (Hanson and Maxwell 1986, 162–3; Poulter 2009, 116). It is generally agreed that the location of many of the installations was determined first, both forts and fortlets, though there is a dispute about whether this applies to all of the former (see below). Recent analysis has highlighted what appears to have been a short-distance approach to planning the Wall line, usually downhill from convenient high points (Poulter 2009, 115–18).
Until the 1970s the Wall was thought to have been designed as a unitary monument (eg Macdonald 1934, 162; Robertson 1960, 27). However, the differing structural relationships between the forts and the Wall (see Forts) led Gillam to suggest that its original plan had been modelled on Hadrian’s Wall in its developed form, with six forts some 12km apart and several fortlets between each (Gillam 1975). This hypothesis was tested and seemingly supported by a successful search for more fortlets (eg Keppie and Walker 1981; see Fortlets). Where examined, all fortlets were either contemporary with or preceded the construction of the rampart. More recently, however, inconsistencies in the structural relationships between the Wall and some of the forts, the strategic positioning of supposedly secondary forts and the apparent primacy of fort locations in relation to the planning of the Wall line have resulted in a re-assertion of the view that all of the forts were also part of the original design, though constructed at different stages in the building process (Poulter 2009, 117–24; Poulter 2018, 114–22; Graafstal et al 2015; Graafstal 2020; Symonds 2017, 144–6). The juxtaposition of a fortlet and fort at both Croy Hill, Duntocher and probably Castlehill (Hanson 2020b; Hanson and Jones 2020), however, remain difficult to reconcile with this hypothesis, though there are some issues with the excavator’s interpretation of the layout of Duntocher (Robertson 1957). Hodgson has further argued that an original intention to build the Wall in stone was also changed at an early stage in order to speed up the building process (2020).
There have been various attempts to calculate how the construction of the Wall was organised and managed utilising a range of archaeological and epigraphic evidence (eg Keppie 1974; Hanson and Maxwell 1986, ch 6). Attempts to correlate variations in the dimensions of the elements of the Wall and the changing character of the materials employed in the construction of the rampart with the building sectors indicated by the Distance Stones have not proved successful (Macdonald 1934; Hannon, Rohl and Wilson 2017).
While it has been widely accepted that the Wall was constructed from east to west (eg Macdonald 1934, 393–400; Keppie 1974, 151), there have been various suggestions about its starting point, including that the sector east of the Avon was a late addition because of a misalignment of the Wall line on the opposite banks of the river (Maxwell 1989, 163), or even that the original eastern terminus of the Wall lay at Watling Lodge (Bailey 1995, 595). The measurements on the Distance Stones (see Distance Stones) led Hassall (1983) to argue for an initial central sector of 20 Roman miles followed by the eastern sector and then the final section west of Castlehill. However, there is no inherent reason why the building process would have commenced at one end and progressed in a linear manner to the other once the line of the Wall had been defined. The subdivision into sections allocated to different legions, combined with considerations of efficient manpower distribution, could imply that work commenced in different sections simultaneously.
Some military posts can be shown to have been prioritised in the building process. Thus, the fortlets at Duntocher and Cleddans at the western end of the Wall were originally built as freestanding structures. While that at Seabegs Wood is located on a slight salient angle, suggesting that its location was chosen before the path of the straight line stretch of the Wall was decided. Similarly, the forts at Old Kilpatrick, Castlehill, Balmuildy, Castlecary and Auchendavy clearly predate the construction of the rampart (see Forts). Building inscriptions from forts indicate both legionary and auxiliary involvement in their construction, in some cases (Castlecary and Bar Hill) at the same site.
The soldiers building the Wall were accommodated in temporary camps; these form a body of evidence unique to the Antonine Wall (see Camps). Some 20 of these have been recorded along the length of the Wall, most of them relatively small at between 2–2.5ha (Hanson and Maxwell 1986, 117–21; Jones 2005; for the most up-to-date plans and descriptions see Jones 2011). In the eastern half of the Wall the camps are found in pairs at each end of a legionary building sector, suggesting a possible division of labour. The pattern is less clear to the west, however, where fewer camps are known. Two small temporary enclosures that underlie the forts at Bar Hill and Croy Hill presumably relate to the surveying, planning or possibly construction of the Wall (Jones 2005, 553–4; see Pre-Antonine military presence).
- Seek more structures, especially fortlets, by means of more detailed LiDAR survey, geophysical survey and excavation, and confirm, where possible, their structural relationship with the Wall.
- Clarify, where possible, the structural relationship with the mural barrier for those forts where this is not known or is disputed, especially where there is ambiguity between the evidence from stone bases and turf superstructures.
- Find more camps, check their relationship with the Wall, their phasing and dating, and seek to understand their interiors better, whether through remote sensing or excavation, to improve understanding of layout, potential numbers of men; and explore their vicinities for traces of camp followers during the construction phases.
- Critically review and re-evaluate the data and interpretation of the remains at Duntocher.
- Review the evidence for the variations in the stone base, the turf and earth building materials and techniques, and the location of the Distance Stones in the light of recent excavations to gain further understanding of the planning of the Wall.
- Undertake a detailed study of building materials, their sourcing, and any changes in construction method or building technologies, to improve understanding of the phasing of the building of the Wall, as well as the labour requirements and timescales of the processes involved.