4.3 Landscapes of settlement

Forts were not established in isolation. They both sat within a natural and settled landscape, and then created their own surrounding landscape infrastructure of camps, fields, vici, temples, roads, bridges and cemeteries. Thus the fort acted as a node with profound influence on the surrounding area. How extensive this area of influence was is difficult to define and no doubt varied, but known examples indicate it is likely to have extended for at least a kilometre beyond the ramparts in terms of visible infrastructure.

To begin with the indigenous settlement landscape, there is a fundamental problem of chronology when it comes to identifying sites contemporary with the Roman occupation. Archaeologists have been reliant in the main on the presence of Roman material on indigenous settlement sites, but generally there is little of it and most sites fail to produce material later in date than the second century AD. However, given that Roman artefacts seem to have been differentially available according to social, political or economic factors, no chronological implications need necessarily be drawn from their absence from sites of potential Roman date. Though certain sites, such as lowland brochs and some souterrains, have been fairly extensively studied and regularly show occupation in the Roman period, there are large numbers of settlement enclosures recorded from aerial survey which cannot be dated without excavation. Recent excavation and survey in East Lothian has indicated that at most settlement sites occupation continued uninterrupted into the Roman Iron Age, though with a marked reduction in the number occupied by the 3rd century and a greater focus on Traprain Law (Haselgrove 2009; Lelong and MacGregor 2007; Hunter 2009). Bayesian statistical approaches to radiocarbon dating, as applied in the Traprain Law Environs Project (Hamilton and Haselgrove 2009) offer the potential to refine the dating of Iron Age settlements so that their attribution to the Roman Iron Age is less reliant on the presence of Roman artefacts.

No towns ever developed north of Hadrian's Wall, but civilian settlements outside forts (vici) are attested at several sites in the Antonine period, mainly associated with the Antonine Wall, notably at Carriden, Croy Hill and Rough Castle, where extensive field systems are attested (Hanson and Maxwell 1986; Jones 2006b, 2007). Identification of the vicus at Carriden is further confirmed by an altar dedicated by vikani (RIB III, 3503). The most extensive civilian settlement known in Scotland is at Inveresk, where a combination of stray finds, aerial photography and excavation by different organisations over a number of years has revealed structures extending for approximately a kilometre from the fort (Bishop 2002; 2004), In general, however, current knowledge of vici is little more than rudimentary and as yet there has been only one postulated in relation to a Flavian fort, at Easter Happrew, where aerial survey has identified buildings outside the fort on the opposite side from the annexe, supported by an extensive spread of finds (RCAHMS 1967, 171, fig 181; Wilson 2010, 49).

One related debate concerns the character and function of fort annexes. Some consider them to be defended vici (Sommer 1984, followed by Bidwell and Hodgson 2009, 31-3), though this interpretation is hotly disputed (e.g. Hanson 2007). Some substantive excavations have taken place (at Camelon, Newstead, and Elginhaugh), but there is still considerable scope for further field work and excavation to define function. Attempts to identify vici by means of geophysical survey along the Antonine Wall, however, have not proved successful (e.g. Burnham et al. 2007, 256-9), but were much more successful at Newstead, where a contrast between magnetically 'busy' and 'quiet' areas was very clear. Possible mansio buildings have been identified within attached annexes at Glenlochar and Newstead (Black 1991; Frere and St Joseph 1983) and postulated elsewhere, but none have been subject to detailed modern investigation.

Geophysical and other survey, and excavation as appropriate, should be undertaken in the vicinity of fort sites of all periods and particularly within fort annexes and at the sites of potential vici. Basic questions remain about the nature of both annexes and vici, and the activities taking place within them.

An evaluation should be a standing requirement of any ground-breaking work within 1 km of a fort.

The publication of the backlog of excavation reports on a range of important sites, including Newstead and surrounding settlements, Camelon and Croy Hill (see Table 4) should be actively encouraged.

Every opportunity to examine potentially Roman Iron Age settlement sites should be taken. Multiple C14 samples from these are required so that their chronology can be statistically refined and the detailed development of indigenous society studied.


Excavations by GUARD Archaeology in 2014 in advance of development near Falkirk revealed significant archaeological remains and finds including Roman pottery, metalwork and metalworking waste associated with the southern annexe of Camelon Roman Fort. Analyses of these finds sheds new light upon life during the early stages of settlement outside a Roman Fort in central Scotland. The report, ARO22: Outside the walls: Excavations within the annexe at Camelon Roman Fort, is available as a PDF from http://www.archaeologyreportsonline.com/PDF/ARO22_Redbrae_Road.pdf