Return to normalcy? Building materials and the making of the Antonine Wall

ABSTRACT

The past 15 years have seen a lively discussion of the building sequence of the Antonine Wall, focusing on the apparent changes of plan at an early stage of its construction. Key issues in this debate are the change from stone to turf as the main building material and the choice for an unprecedented 2-mile spacing norm for forts, making this the most densely fortified frontier line of the Roman world. This paper reviews the main arguments and proposes a return to the central thought of John Gillam’s seminal 1975 paper: the army that came to build the Antonine Wall took with it the experience of Hadrian’s Wall and naturally started replicating this legacy on the Forth–Clyde isthmus – in stone, initially. In this light, the switch to turf and timber, very early in the sequence, stands out as a major breakaway from Hadrianic legacy. The structural and, consequently, operational implications of this decision are considered, in an attempt to make sense of the unique disposition of the Antonine Wall.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Notes

1. The inspiration for this contribution came from two recent papers by renowned authorities on the Antonine Wall, Bill Hanson (2020) and Nick Hodgson (2020). In many respects, this article builds on their work; slight differences in interpretation do not relieve my debt to both. As so often, I thank David Breeze for his comments and encouragement. Any remaining errors are mine, of course.

2. The rectangular interruption in the south cheek of the Rampart documented at Callendar Park (Bailey 1995, 585–586 with fig. 3) may be little more than a patch of soil material of different composition, possibly explicable with reference to the proposed ‘coffer-dam’ mode of construction of the cheeks (Bailey 1995, 586). Moreover, its position midway two kerbstones looks structurally unsound. Finally, timbers of the implied size (30 by 30 cm) appear to have been in very short supply; witness the rather flimsy fort gates of the Antonine Wall (Graafstal 2020b, 179). See also Hanson and Breeze (2020, 23).

3. I leave to one side the enigmatic ‘expansions’ and ‘enclosures’ that have been found abutting to the rear face of the Rampart in a few places. ‘Their irregular spacing (…) and inconsistent relationship with the Wall would seem to preclude their interpretation as elements in a regular watch-tower system’ (Hanson and Maxwell 1986, 98). For an alternative explanation see Poulter (2018).

4. This is why the stone base for the Antonine Wall where it formed the north rampart of forts was often terraced in relation to the stone bases of the forts’ side walls, resulting in the impression of secondary construction of the latter at e.g. Mumrills and Westerwood (Graafstal 2020b, 145).

5. These estimates are based on Hodgson (2017, 194–195) (for Hadrian’s Wall), and Snyder et al. (2023, 154), respectively. The latter paper is to some extent superseded by new insights into the fabric of the Antonine Wall gained at Laurieston (Romankiewicz et al. 2022) which have significant implications for the work load.

6. Assuming six men per tower (cf. Baatz 1976, 43) and an average 7.2 Wall miles for every fort (Breeze and Graafstal 2022, table 4).

7. At Cadder, the causeway had been left undug and the Rampart at the northeast corner built ‘with the needs of the fort in full view’ (Macdonald 1934, 302). The causeway at Croy Hill, likewise, suggests ‘foreknowledge of the existence of the fort on the part of the ditch-diggers’ (Hanson and Maxwell 1986, 108). The existing causeway over the Ditch at Rough Castle is a modern recreation after Iain Macivor had the original one removed as part of his excavations in 1957-1963 (pers. comm. David Breeze). However, Buchanan’s report (1903, 455) clearly implies that the original causeway consisted of undug earth.

8. See Hanson and Maxwell (1986, 108). In fact, Macdonald’s description clearly implies that the waterworks and associated construction pits preceded, and continued under, the fort wall of ‘Croy 2’ (1932, 251; Graafstal 2020b, 167).

9. A Roman fortlet thought lost to time is rediscovered: see historicenvironment.scot. Accessed September 9, 2023. https://www.historicenvironment.scot/about-us/news/a-roman-fortlet-thought-lost-to-time-is-rediscovered/

10. Based on archaeological evidence for Germanic societies, Steuer arrives at a typical strength of retinues of c. 10 to several dozen at most (1992, 2009). Plinius mentions dugouts carrying raiding bands of up to 30 men (Nat. hist. XVI.203). The early third century Illerup A war-booty deposit in Denmark can be broken down to ‘cells’ of comparable size (Ilkjær and Iversen 2009, 140–1), but is itself evidence of the potential of such war bands to scale up, making use of ‘clientele pyramids’ (Roymans 1990, 40–41).