5.3 History of Research

Study of the region’s Iron Age and Roman sites has a long and varied history. It begins, as result of wider antiquarian interests in the Roman Empire, with the Roman military sites of the area. Descriptions of sites such as the exceptionally well-preserved Roman auxiliary fort of Ardoch, Braco, are published as early as the 16th century (MPK665; Boece 1527) and continued in antiquarian accounts into the 18th century. These included Sibbald (1695, 1707), Gordon (1726), Horsley (1732), Maitland (1757), Pennant (1771), and Adamson (1774). Gough’s edition of Camden’s Britannia (1790) and both the Old and New Statistical Accounts of Scotland (Sinclair 1799; Gordon 1845) also had accounts of Roman sites. Roy’s pioneering mapping (1793) is noteworthy as his surveys capture the earthworks of prominent Roman sites that have since been lost or severely deteriorated, mainly through agriculture. From the 18th century, as a result of agricultural improvements and treasure hunting, the recovery of Roman material became common; frequent references appear in the New Statistical Accounts of Scotland and later in more bespoke works such as Stuart’s Caledonia Romana of 1852. The Perth Literary and Antiquarian Society, founded in 1784, ranked the study of Perthshire’s Roman remains very highly on its agenda of antiquarian studies, an approach it maintained through the 19th century. As was common at the time, Bronze Age swords were often interpreted as Roman ones (Cowie and Hall 2001, 151–3). 

Pennant’s accounts of his travels to the region in the 1770s also deserve particular attention as an early commentator on Iron Age monuments and in particular the ‘circular buildings’ of ‘Highland Perthshire’ (1776). These represent some of the earliest comprehensive field observations of Iron Age sites in the region using an approach that is recognisable today as an early expression of landscape archaeology (Strachan 2013, 4). While mistakenly attributing Danish or Norwegian origins to sites, as so many also did with Pictish stones, his work remains an early and important regional thematic study which saw little if any revision through the subsequent Statistical Accounts of Scotland (Sinclair 1799; Gordon 1845). 

The 18th century excavations, early even in Scottish terms, at Dunsinane fort, Collace, by Playfair (MPK4823; Robertson 1799), were no doubt inspired by association with Shakespeare’s Macbeth. The small, heavily defended summit-fort remains scarred by these, and the following investigations by Nairne in 1854 were reported in the second volume of PSAS (Wise 1856). Also around this time, many of the region’s crannogs were recorded in the first gazetteer of crannog sites in Scotland (Stuart 1865, 174–7). 

Antique plan drawing of an oval hillfort, drawn in pencil with hashed lines showing the gradient of the hill surrounding the fort. There is a dividing line one third of the way from the left of the fort, and thick ramparts are outlined the whole way around the structure.
Dunsinane Hill – Playfair plan from 1832 ©️ HES
Aerial oblique photograph of an oval hillfort, showing the surrounding landscape of open land and limited trees. The fort has three levels which are separated by step-like gradients, with the top level having bumps and features still visible.
Dunsinane hillfort ©️ HES

Pennant’s external origin theories were to prevail, however, for almost 100 years until MacLagan (1871) suggested that monumental Iron Age structures were built by indigenous communities. Despite gaining support from Anderson (1876), the national debate continued to associate brochs, duns and forts with Scandinavian influences (Fergusson 1877) and MacLagan’s suggestion was only fully accepted in the 20th century. 

From the end of the 19th century, more formal programmes of archaeological work commenced. The large-scale excavations at Ardoch Roman fort (Christison and Cunningham 1898; Anderson 1898) are particularly notable, along with excavations at Castle Law, Forgandenny (MPK1905; Bell 1893; Christison 1900b) and Castle Law, Abernethy (MPK3069; Christison and Anderson 1899). Christison was influential, contributing significantly through classification and synthesis of Iron Age and Roman sites in the wider area (Christison 1900b) and through excavation. Recognising the potential and value of Iron Age monuments, he was an advocate for their study. In the annual report of the Society of Antiquaries of Scotland in 1900, he wrote 

I could almost regret that the Society have undertaken the excavation of Roman ‘Camps’ in preference to our Native Forts. The secrets that lie beneath the ruins of the Caterthuns, Dunsinnan, and hundreds of other native fortresses, are no less worthy of being brought to light than the relics left behind by the Romans (Christison 1900a, 12). 

Aside from Watson’s (1915) work on monumental roundhouses, including the excavation at Borenich (MPK1248 and MPK1246), Loch Tummel, on the eve of World War I, relatively little excavation was carried out on Iron Age sites in the first half of the 20th century. Interest focused on major Roman sites such as Inchtuthil legionary fortress, Spittalfield (MPK3639; Abercromby 1902) and Fendoch auxiliary fort, Glen Almond (MPK1479; Richmond and McIntyre 1936; Richmond et al 1939). This focus may in part have resulted from a lack of typologically distinct Iron Age material culture, which presented a significant challenge to the culture-historical approaches of the time. The shortage of chronologically definable data also resulted in narrative construction based on historically documented events, such as the Roman military campaigns. Diffusionist models for aspects of the Iron Age, such as for broch building in the region, also persisted long into the 20th century (e.g. MacKie 1987, 16). It has only been with the chronological security of radiocarbon dating across different site types that historically-driven chronologies have become increasingly scrutinised and the region’s Iron Age narratives have become more independent. 

Between the 1950s and 1970s Iron Age and Roman investigations continued to be undertaken in relative isolation from each other and were generally smaller scale in nature. Following Thorneycroft’s pioneering work on the Late Bronze Age settlement at Dalrulzion (MPK4038; 1933; 1948), Stewart excavated roundhouses at Dalnaglar (MPK4338), Kirkmichael (Stewart 1962) and Tulloch Field (MPK2854), and Enochdhu (Thoms and Halliday 2014). She then reviewed Watson’s work on monumental roundhouses, through ‘The Ring Forts of Central Perthshire’ and excavations at Litigan (MPK413; Stewart 1969) and Queen’s View (MPK1212), later published by Taylor (1990). A range of Roman site types were also excavated in the latter part of the 20th century: from forts at Carpow (MPK4624; Birley 1963; Dore and Wilkes 2000) and Strageath (MPK714; Frere and Wilkes 1989), to fortlets, such as Glenbank (MPK675; Woolliscroft and Hoffmann 2009). Other types of excavated Roman sites include watch towers – Westerton (MPK786; Hanson and Friell 1996) and Woodhead (MPK3672; Woolliscroft and Hoffmann 2010; Chapman et al 2011), and  camps, including Dunning (MPK1981; Dunwell et al 1995; Hunter 2015a), Forteviot (MPK1929; Taylor 1953) and Longforgan (MPK4799; Hunter 2015b). However, in the early years of the 21st century, it was observed that Iron Age research in the area had fallen behind that in neighbouring regions such as Stirling and Angus, with notably fewer excavations across Perth and Kinross in the 20th century (Davies 2007, 270). 

Air photographic survey has contributed significantly to understanding of the region’s Iron Age and Roman presence since the 1950s, having been initiated by Crawford (1949) of the Ordnance Survey. St Joseph of the Cambridge University Collection of Aerial Photography (CUCAP) flew annual sorties over the region until the late 1980s. This work was built on through the commencement of aerial survey by the Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS) in 1975/6 (Maxwell 1983, 27–40).  

Two important RCAHMS primarily ground-based surveys, of north-east and south-east Perthshire, were carried out in the late 1980s and early 1990s (1990; 1994). These set the paradigm for understanding most aspects of the Iron Age remains of the region, and many of the key underlying assumptions in subsequent studies can be traced back to these influential publications.  While extensive, the coverage was limited to eastern Perthshire and amounted to less than half of the PKARF area. Consequently, there is a survey bias to the west of the River Tay where the lack of a county-wide RCAHMS inventory has never been alleviated (Strachan 2013, 115). 

The advent of development-led archaeology has helped to broaden fieldwork from what had become an established focus on Roman military and monumental Iron Age sites. The resulting discovery and excavation of both new sites and monument forms has enriched our understanding of the period. It began to have impact from the 1980s, for example through excavations at Aldclune, Blair Atholl (MPK3; Hingley et al 1998) in advance of construction of the A9. Monitoring of development in the area significantly improved from 2000, when a high-quality Sites and Monuments Record and consistent, properly resourced screening of planning applications was introduced by Perth and Kinross Heritage Trust on behalf of the local authority. Nearly 40 years on from Aldclune, important sites are again being discovered and explored through the ongoing A9 Dualling Programme and its associated works. These include post-built roundhouses at Bertha Park, Perth (MPK20178-88; Engl 2020), two roundhouses, a souterrain and four-post structures at Northleys, Luncarty, and west of Newmill, Bankfoot, where a third roundhouse with an enclosing palisade was excavated (EPK1387; Paton and Wilson 2019). A roundhouse, 6-post timber structure, and a possible kiln were uncovered at the Stanley Road Junction  (MPK20176; Airey 2020). Perhaps most significant though are results from Loak Farm (MPK20126), where seven roundhouses, four-post and six-post structures, ring-ditches, and an oval post-built structure indicate extensive settlement of Late Bronze Age to Iron Age date (Kirby 2019). A second phase of monitoring at the same site has discovered three souterrains (Demay 2021). Another important developer-led site, excavated in advance of a golf resort at Blackford in 2007–8, uncovered extensive enclosed and unenclosed settlements, principally of the Middle and Late Bronze Age, but continuing into the Early Iron Age (O’Connell and Anderson 2021). More recently in 2021, excavation of targeted areas following trial trenching at the James Hutton Institute, Invergowrie, has revealed a well-preserved roundhouse with several associated elongated/oval buildings and outlying postholes and linear features that may also be contemporary. 

Black and white oblique photograph of an open excavation trench in the centre of tree-covered land and hills. The trench is full of cobble stones and looks like organised paved flooring. A central drain can be seen running the whole length of the site.
Aldclune 1980s excavations ©️ HES

Independent academic and project-based research has also contributed in recent decades. The Roman Gask Project, established in 1995 by Woolliscroft and Hoffmann, has helped to shape understanding, and questioned assumptions on the nature of the Roman presence in the region. Combining aerial and geophysical survey with fieldwalking and excavation, a wide range of sites along the Gask Ridge and the wider area have been investigated (Woolliscroft and Hoffmann 2006) and visual signalling explored (Woolliscroft 2001). In addition, the influential historical text by Tacitus on Agricola, the provincial Roman Governor from AD 77–83 who is attributed with the conquest of Scotland, has been reassessed (Hoffmann 2004; 2012). Research specifically on Roman camps in Scotland has reviewed current knowledge in the region and placed it in a wider context (Jones 2011). In addition, further geophysical surveys have been conducted by Morris (Blairgowrie Geoscience) at both Roman and Iron Age sites. 

The value of GIS analysis of Iron Age and Roman military sites has been explored through viewshed analysis to examine intervisibility (Strachan 2013, 61–67; Lindsay 2006; Murphy 2018; Tibbs 2021). Advancements in remote data gathering, such as satellite imagery, photogrammetry and LiDAR, offer increased potential for future research. GIS has also been used in the examination of small finds, as at Inchtuthil legionary fortress, where it was employed to better understand trade and coin loss (Trezzi 2017). 

Postgraduate research has contributed to the understanding of both settlement patterns and specific architectural forms, including a comparative study of later prehistoric settlement north and south of the Forth (MacInnes 1983). There is also a valuable comparative synthesis of later prehistoric settlement in ‘Perthshire’ and ‘Stirlingshire’, which uses dating and palaeoenvironmental data to aid discussion of settlement types known from survey and excavation (Davies 2006). Hooper’s (2002) thesis on Highland cultural landscapes, while focused on the early medieval period, also includes much of value regarding Iron Age settlement. 

In recent decades several thematic and landscape studies have been carried out. The multi-disciplinary Ben Lawers Historic Landscape Project (1996–2005) was primarily focussed on medieval and later periods, but it did explore Iron Age roundhouses at Tombreck (MPK15945) and Croftvellick (MPK9467; Atkinson 2016). Excavations at the Black Spout, Pitlochry (2005–9) responded to unresolved questions regarding the chronology of the monumental roundhouses following earlier work (MPK1607; Stewart 1969; Taylor 1990; Hingley et al 1998). They illustrate how targeted work at a single-site can help to address regional narratives through the context of wider synthesis (Strachan 2013). It was also an early Scottish example of ‘community archaeology’, highlighting the valuable contribution that can be achieved through citizen science. The Strathearn Environs and Royal Forteviot (SERF) project (2007–19) included the excavation of 10 forts, predominantly along the Ochils hills and Strathearn. While full results are in preparation (Poller forthcoming), much has been published online through The Archaeologist’s Trail: Hillforts of Central Scotland, including interactive media and data structure reports. Similarly, The Tay Landscape Partnership Scheme (2014–18) ‘Hillforts of the Tay’ project excavated three hilltop forts at the head of the Tay estuary (Moncreiffe, MPK3203 and Moredun Top on Moncreiffe Hill, MPK5232 and Castle Law, Abernethy, MPK3069). Radiocarbon dating confirmed an Iron Age date for all, which is significant because Moredun Top had been for many years mooted as an early medieval ‘nuclear’ fort (Feachem 1955, 79–80; Alcock et al 1989, 206 and Alcock 2003, 189). The excavations suggest earlier activity at the smaller, multivallate Moncreiffe fort, while on Moredun Top, a monumental roundhouse was uncovered within a series of three Iron Age forts occupied from the late 6th century BC to the mid-1st century BC. It confirms that this was a particularly prominent power centre, which dominated both the important estuarine and terrestrial routes and was a key site in local Iron Age society (Strachan et al forthcoming). The project saw considerable volunteer and community engagement, with web-based Minecraft, Virtual Reality and artistic reconstructions, interpretive leaflets and on-site panels plus a popular publication (PKHT 2020) making the findings widely accessible to a broad audience. Finally, building on Dixon’s excavations at Oakbank crannog in the 1980s, the Living on Water project (2017–20) has investigated seven of the 11 crannogs on Loch Tay with known Early Iron Age phases (see Oakbank Crannog Case Study). The project has aimed to develop chronological precision for the Hallstatt Plateau period, which coincides with the emergence of crannogs in the loch, using a combination of dendrochronology, radiocarbon dating and Bayesian statistical analysis. The initial results seem to suggest that crannog building began no earlier than 600 BC with a notable increase in activity from 400–350 BC. Further detail on the local economy and environment has been gleaned from the exceptionally well-preserved deposits of organic material (Cook et al forthcoming), along with publication of Oakbank crannog, which details the excavations there since 1980 (Dixon and Andrian forthcoming).