3.3 Flavian Scotland (c. AD 77-86/90)

1st century AD Flavian temporary camps (later first century AD) © Crown Copyright Historic Scotland, with the addition of Glenluce © ScARF

Sites along the Gask ridge © Crown Copyright Historic Scotland

With the exception of the dating of the first incursions (see section 3.2 Questions of pre-Agricolan activity), the broad outlines of the extent and chronology of Flavian Scotland are generally accepted. However, there remain a number of issues of debate and uncertainty. Knowledge of the distribution of temporary camps is far from complete and their confident attribution to particular campaigns remains a matter of speculation rather than hard evidence (Jones 2006a). Similarly, knowledge of the site of the battle of Mons Graupius remains elusive (section 4.1). Despite some assertions to the contrary (e.g. Gregory 2001), there is no evidence of fort building north of the Mounth. The postulated sites at Thomshill and Easter Galcantray lack the distinctive morphological characteristics of Roman military works and have not provided any artefactual support for a Roman date, but questions remain over the distribution of sites in the south-west of Scotland and the existence of Flavian precursors to Antonine Wall sites.

Considerable survey and excavation has been undertaken over the last decade or so on sites associated with the Gask Ridge (e.g. Woolliscroft 2002; Woolliscroft and Hoffmann 2006). In addition to challenging the dating of the conquest of the area, the work has led to a re-assessment of the function of the chain of towers that accompany the road. Opinion is currently divided between those who follow the view that this represents an artificially-defined frontier (e.g. Hanson 1991b), and those that see it as simply a controlled supply line to the legionary fortress at Inchtuthil (e.g. Dobat 2009), or part of a wider frontier zone (Woolliscroft and Hoffmann 2006).

Inchtuthil itself is the key site of Flavian Scotland, and one of international importance as an early Imperial fortress unencumbered by later developments. The excavations of Richmond and St Joseph reconstructed a seminal plan (Pitts and St Joseph 1985); recent geophysical survey (DES 2009, 145; Britannia 41 (2010), 347, fig 2), combined with aerial survey data, will serve to put these extrapolations from small trenches onto a firmer basis. Our horizons should not be limited to the plan of the fortress alone; from the early excavations comes a small but significant (and incompletely published) assemblage of material beyond the headline-grabbing massive nail hoard, while recent survey work has expanded the material range and looked at the setting of the fortress (Britannia 41 (2010), 347-8; 42 (2011), 328-330). Its significance comes both from its tight dating and its information on legionary supply and equipment at the time; further study of the existing material and renewed excavation would be of value far beyond Scotland.


Cropmarks of the Flavian legionary fortress at Inchtuthil, Perth & Kinross. SC972093 © RCAHMS

Evidence points to a staged withdrawal from Scotland, with the forts north of the Forth-Clyde line and those south of it as far as Newstead abandoned by AD 86-87, with most of the remaining southern forts shortly thereafter. No Scottish forts show any certain Trajanic occupation, although some of the southern forts such as Broomholm may not have been abandoned until the early Trajanic period. With the construction of Hadrian's Wall in the 120s, Birrens (Dumfriesshire) was established as an outpost fort (Robertson 1975) at this time.

1st century AD Flavian forts and fortlets © Crown Copyright Historic Scotland

Inchtuthil represents a key site that would reward further study both in the field and in the archive. Assessment and publication of the Broomholm excavations would also considerably advance knowledge of Flavian Scotland (publication is in progress). Geophysical survey and fieldwalking at the fort of Ladyward would also provide useful information, as the site lies at a key position for the SW, but its chronology is unknown.