The extensive, highly articulated agrarian landscape of the region in the last few centuries BC revealed by pollen analyses (SESARF Chapter 7.5) can also be traced archaeologically (Alexander and Watkins 1998; Haselgrove and McCullagh 2000; Lelong 2007; Haselgrove 2009, 2015, 2016, Armit and McKenzie 2013). Before AD 80 this was a very busy agrarian landscape although Bidwell (2020) has argued that cereal production by native farmers was too low to have supported the Roman army on the Antonine Wall in the 2nd century AD.
Van der Veen (1992) differentiated two types of arable economy in the Late Iron Age in north-east England. Group A where crop-growing was intensive but small-scale, because of the high energy needed to maintain soil fertility, through manuring, and to discourage perennial weeds. Group B arable economies where crop-growing was extensive, but lower energy inputs were needed. Group B rural economy is sometimes associated with household subsistence practices while Group A attributed to the intervention of elites in defining productivity (Styring et al 2017). The recovery of plant remains only been related to the Later Iron Age economy along the Forth coastal plain in one of Van der Veen’s (1992) models (Huntley in Haselgrove and McCullagh 2000, 188). Here the assemblage suggested Group A arable economy, but we do not know whether the Fisher’s Road East assemblage was typical of the region and South east Scotland may have contained both arable economies.
New work has shifted van der Veen’s (1992) north-south divide northward (Hodgson et al. 2001; Proctor 2009, 2016, 208; Hodgson et al 2012, 176; Haselgrove 2016, 415) so that we might now say that communities south of the River Tweed, loosely, were engaged in extensive arable agriculture while those to the north practiced a more intensive but smaller-scale farming. The geography is important. Van der Veen (2023) has suggested that in north-east England, native grain supplies to the Roman army should not be discounted because the crop regime was capable of yielding large harvests and agricultural surpluses. Heslop (2023) also identifies the rural economy in lower Teesdale to have generated huge disposable surpluses in the century before Roman occupation, sufficient to construct the Stanwick oppidum (Haselgrove 2016). Bidwell (2016, 302) also suggested that the Roman army in northern England may, in the later 3rd century have been supported through native agricultural expansion. Recent work on crannogs and dryland sites in south west Scotland has identified bread/club wheat as a consistent if rare find. Often assumed to be a Roman import from southern England (Holden in Lowe 2006, 152; Cavers 2008, 121), Haggerty and Haggerty (1983, 43) suggested local cultivation on soils requiring high fertiliser and labour inputs, in fields much like those of Van der Veen’s (1992) Group A economy, grown for an emergent local elite (Cavers and Crone 2018, 170).
The pattern of different rural economies in south east Scotland is also a major unanswered question. Signer (2022) has hinted that this might have been the case in northern England when she suggested a shift from surplus to subsistence in the post-Roman rural economy there. Rizzetto and Albarella (2022) make the same point, native communities moving from surplus generation to self-sufficiency as the Roman army left and the rural economy was restructured. Armit and McKenzie (2013, 499) certainly saw Roman occupation of the Lothian Plain as good for ‘trade’, seen in the supply of prestige goods to elite native settlements (MacInnes 1984), their numbers increasing at times of direct Roman occupation (Harding 2004, 191–192). Traprain Law was a ‘bustling native hill-town with close Roman contacts’ in the Roman Iron Age (Armit et al. 2006, 606) and into the 5th and 6th centuries AD (Hunter 2013). Agricultural production is assumed to have driven a prestige goods economy by Harding (2004, 171) with cattle the principal driver of wealth. In the organisation of this economy, Cook et al (2019) saw the need for hierarchical social structures. Hunter (1997, 121) and Armit (2019) have favoured less stratified societies.
Cattle were the dominant livestock at Roman sites in the region, greatly exceeding numbers of sheep and pig. Beef was the primary meat consumed and livestock increased in size during the Roman Iron Age (Pigière 2023). Cattle died at different ages, implying a range of uses, not just for meat, including being kept for breeding and for traction, although not for dairying (Barnetson 1988). Sheep and pigs were usually slaughtered at prime meat-production ages (Stallibrass 2009). Isotopic evidence suggests people consumed high levels of animal protein but little for aquatic (marine) resources. There is little evidence for the consumption of wild animals (Jay and Richards 2007). Hay meadows appear to have emerged on valley floors, most likely for herds of cattle (Tipping 2010). However, further exploration of animal use is the region is needed. Elginhaugh produced 1604 fragments of bone in stratified contexts, of which 1528 could only be determined as indeterminate mammal (Smith 2007, 632). Recently developed techniques such as ZooMS (Buckley 2016) and environmental DNA have potential to provide information on the exploitation of animals where bones preservation is poor.
Mercer (2018) proposed that the Southern Uplands were oriented towards a surplus-generating cattle (or, perhaps, horse) economy with Roman occupation. He identified native forts modified by annexes to create corrals for large herds. Stallibrass (2009, 2018) has proposed similar models of long-distance droving to supply the military bases in northern England. Strontium analyses on cattle teeth at Owlesbury in southern England, Minitti et al (2014) suggested they were sourced from some distance away – possibly as far as Scotland. Groot et al (2020) report similarly long ‘trade’ routes for cattle in The Netherlands.
Notably in the Cheviot Hills, the RCAHMS (1956), Halliday et al. (1981), Halliday (1982), Topping (1983, 1989) and Burgess (1984, 151) all argued for later prehistoric origins for cultivation terraces because they were overlain by either unenclosed roundhouses or other field system elements such as cairn fields (see also Harding 2004, 228). Brown et al (2023) suggested an early Bronze Age date for the excavated system at Plantation Camp in the Cheviots, and proxy evidence from Holyrood Park in Edinburgh has suggested that the Crow Hill terraces are also of Bronze Age date.
There are Iron Age examples of coaxial field systems in northern England (Fleming 1998; Hoaen and Loney 2013; Rippon et al 2015, 292; Heslop 2023), though none as yet in southern Scotland. North of the River Tyne around Morpeth, there is archaeological evidence for landscape re-structuring in the late 1st century AD. This may have been undertaken to manage stock, perhaps for the Roman army (Proctor 2009). The fertilising with seaweed of fields long distances from the shore on the Lothian Plain (Huntley and O’Brien 2009, 159, 162) suggests that access to such resources was not simply local to a farm but involved negotiation and organisation. Roymans and Gerritsen (2002) described the working of fixed fields in the southern Netherlands and northern Belgium as ‘intensive’.
Roman field systems have also been recorded in the region, although sometimes only dated by association. Field systems are notablearound Inveresk (Thomas 1988; Cook 2004; Hunter 2014), at Rough Castle, Falkirk (Maté 1995), where there are two discrete periods, at Carriden (Dunwell in Keppie et al 1995) and at Auchendavy (Hastie in Dunwell et al 2002). At Carriden (Keppie et al 1995, 602 – 605) the system is defined by ditches forming an ‘ordered system of small fields or plots’ respecting a Roman road within metres of the fort, associated with a vicus, a ‘semi-regular arrangement of sub-rectangular ditched enclosures or plots … characterised by rounded angles and slightly offset junctions’.