7.1 Introduction

1. Traditionally, the start of the Iron Age (around 800 BC to AD 500) was marked by the introduction of iron, and by the technological advances needed to work a metal with a melting point much higher than that of copper-tin alloys. Evidence of early ironworking is becoming increasingly apparent in the late Bronze Age. It may therefore be more appropriate to view the demarcation between the Bronze Age and the Iron Age as the point when the circulation, exchange and deposition of bronze ceased to play such a visible and potentially significant role in the structuring of social relations.

2. The deposition of metalwork continues to play a role during the Iron Age. More broadly, a contrast has been made between Bronze Age and Iron Age depositional strategies. During the former, the quantity of metal appears to have been important, at least in the closing stage. In the latter, the quantity of items tends to be small, but the quality of their workmanship and manufacturing skill is extremely high (Garrow et al 2009, 111). Whether this model applies in South West Scotland can be questioned. While it is true that there are several particularly fine pieces of Iron Age metalwork from the region, there is a marked absence of large Bronze Age hoards. Large hoards of more utilitarian – and largely fragmented – metal items are also not unknown in Iron Age contexts (see Chapter 6 Chalcolithic and Bronze Age).

3. Much of the Iron Age metalwork recovered from across the region is invariably decorated in the ‘La Tene’ style associated with the ‘Celts’. Items decorated in this distinctive curvilinear style can be found all over the British Isles and further afield, into continental Europe, suggesting some kind of cultural affiliation at the supra-regional level. This supposedly pan-European similarity is supported by the accounts of Classical authors, in particular Greeks and Romans, who interacted with this ‘barbarian’ world. Although the indigenous Iron Age inhabitants of the country lived in a non-literate society which transmitted knowledge through an oral tradition, contact – whether peaceful or violent – with the ever-expanding Roman Empire meant that for the first time they featured in written accounts.

4. Roman writers like Julius Caesar encountered the ‘Celtic’ peoples at first hand. There was, however, another literary tradition available to those researching the Iron Age. This comprised works by indigenous writers who much later captured versions of earlier myths and reworked them into epic literature. In both cases, such texts should be viewed with caution, as they may not necessarily reflect the diverse range of people and practices which together made up this so-called ‘Celtic’ Iron Age world over the roughly 1,300 years of its existence.

5. In South West Scotland, the apparent adherence to a supra-regional ‘La Tene’ decorative style (suggested by metalwork finds like the Balmaclellan mirror and the Torrs pony cap) must be balanced against evidence for regional settlement patterns. These become increasingly marked during the Middle Iron Age. Early Iron Age settlements tend to comprise roundhouse structures, like those occurring in the Late Bronze Age. However, unlike most of their Bronze Age counterparts, these houses are often enclosed by a boundary feature, usually comprising a timber palisade. This could have had a defensive role, or it could have been intended for visual impact in the wider landscape. By the Middle Iron Age, this physical demarcation became transformed into something which was much more impressive visually.

6. All the recognised regional types associated with the Scottish Iron Age can be found within the region: hillforts, promontory forts, crannogs, brochs, souterrains, duns and homesteads. All can arguably be categorised as falling within a broad group of monumental roundhouse styles. With some types – particularly the forts – unravelling their time depth remains problematic. Occupation potentially begins in the Late Bronze Age and continues onwards into the Early Medieval period (see Chapter 9 Early Medieval) and sometimes even later, into the Medieval period (see Chapter 10 Medieval). The medieval sites at Dundonald Castle and Auldhill, Portencross have both produced residual traces of Iron Age occupation and use (Ewart and Pringle 2004; Caldwell et al 1998). Crannogs, too, appear to have been a settlement type which was subject to longevity (Crone 1993).

7. Sometimes these settlements survive in association with field systems, again potentially encouraging confusion with the preceding later Bronze Age (see Chapter 6 Chalcolithic and Bronze Age). This is particularly true of upland areas, now classed as marginal land, where arable land use was no longer viable following climatic deterioration in the Late Bronze Age.

8. Evidence for Iron Age funerary practices throughout the region remains very sparse, though a recent find at Adie’s Brae, near Moffat, has helped shed some light on this important aspect of Iron Age life and death. Balanced against this relative scarcity are increased quantities of faunal and palaeoenvironmental evidence. This is boosted by the rich assemblages recovered from waterlogged deposits associated with crannog sites.

9. Early antiquarian efforts to interpret the physical evidence that we now recognise as Iron Age in character often viewed both sites and material culture through a prism shaped and influenced by Roman literature. One site in particular dominated antiquarian interest. This was the hillfort at Burnswark. Not only was it an imposing site with a complex array of earthworks, it was also in close proximity to the Roman fort at Birrens (see Chapter 8 Roman Campaigns), and its location in Annandale – on one of the main routes north into Scotland – would have made it a target for Roman military activity (RCAHMS 1997, 3).

10. Burnswark was first mentioned in 1727 by Alexander Gordon, who – having concluded that Annandale was the most likely place for the Agricolan advance – included the site in his search for potential contenders for Roman camps (RCAHMS 1997). Towards the end of the same century, Major General William Roy surveyed the complex of earthworks, along with various Roman military sites and the indigenous Iron Age forts at Fairholm, Castle O’er and Woodycastle. All these sites were included amongst a collection of Roy’s survey plans, published posthumously as Military Antiquities of the Romans in Britain (Roy 1793). Although Roy appeared to favour a Roman origin for all the defensive elements he identified at Burnswark, others took a more nuanced approach. The anonymous author of the Old Statistical Account entry for Tundergarth parish, by contrast, favoured the suggestion that Burnswark represented a site where the Romans besieged a camp of the Ancient Britons (Anon 1797, 446).

11. Although it was the monuments of eastern Dumfries and Nithsdale that initially proved more attractive to the 18th-century antiquarians, objects that we now recognise as Iron Age in date were also being recovered and reported more widely across the region. Some of these early metalwork finds came from the vicinity of Burnswark, such as the Middlebie hoard of horse gear and chariot fittings, found in 1732. The owner of the hoard was Sir John Clerk, a Scottish antiquarian who was a friend of Sir William Stukeley (MacGregor 1976, 2). Clerk considered the hoard of sufficient importance to present a detailed report of the find to the Society of Antiquaries of London. Later that century, before 1785, a bridle bit of similar character was recovered from a moss near Burnswark.

12. Early in the 19th century, a particularly fine piece of Iron Age metalwork was recovered from Torrs near Kirkcudbright. This was the celebrated ‘pony cap’ or chamfrein, comprising an unusually shaped sheet bronze headpiece found in association with two curved bronze ‘horns.’ The configuration of these items has challenged researchers to this day: some suggest that the horns were originally attached to the cap, while others argue that they represent different objects, perhaps drinking horn terminals. Decorated in the ‘La Tene’ style, the Torrs ‘pony cap’ was – and still is – an unparalleled find, and it remains a popular focus of interest and research. Originally interpreted as a ‘mummer’s mask,’ it was presented to Sir Walter Scott in 1859 (Atkinson and Piggott 1955; MacGregor 1976, 2).

13. Throughout the 19th century, the metalwork finds continued. A neck ring or torc from Lochar Moss was found before 1851; the Balmaclellan hoard, comprising a sheet bronze mirror and an engraved sheet metal strip, was discovered in 1861. The year 1866 saw the recovery of a diverse collection of weapons and utensils placed in a large cauldron from Carlingwark Loch, Castle Douglas. All these finds would be significant in developing our understanding of Iron Age art and material culture. Many items are listed in Morna MacGregor’s 1976 corpus of Early Celtic Art in North Britain (MacGregor 1976), with a good proportion regularly featuring in museum exhibition catalogues devoted to Celtic Art (for example, Műller 2009).

14. During the 1850s, another distinctive form of Iron Age settlement site started to attract increasing levels of interest. It was around this time that the Irish crannogs were first examined by Sir William Wilde. News of ongoing work on the Swiss lake dwellings was also becoming more widely known within the antiquarian community. However, it was not until 1863 that a similar site in South West Scotland was first described, when an account of the ‘lake dwellings’ of Dowalton Loch was presented as a paper to the British Association by Lord Lovaine (Munro 1880, 19). The site was subsequently explored by Dr John Stuart. This was the first of several such sites to be investigated in the region: a site at Black Loch was examined by Charles Dalrymple (1873), and another at Barhapple Loch was the focus of work by the Reverend George Wilson (1882). Wigtownshire thus became the first county in Scotland to have the presence of crannogs proved by systematic investigations (Munro 1885, 77).

15. Robert Munro excavated a crannog in Lochlea, initially working under the direction of Robert Cochrane Patrick and reported, too, on the examination of another in Kilbirnie Loch (Munro 1885). Lochlea crannog revealed a series of hearths indicative of successive phases. Investigations at Lochspouts revealed copious quantities of animal bones and anthropic material: worked timbers, quern stones, spindle whorls, stone hammers, worked bone and boars’ tusks, iron tools including a saw, and a snaffle-bit for a horse (Munro 1887). Ornamental items included bronze brooches or dress fasteners of the fibula type, glass beads and jet armlets. Particularly unusual survivals were the wide range of wooden objects preserved in the waterlogged deposits. The faunal remains revealed sheep, cow, pig, red deer, horse, limpets and whelks.

16. Further excavations of crannogs took place at Buiston, Kilmaurs (Munro 1882a) and Barhapple Loch, Glenluce (Wilson 1882). The former revealed a wide range of items including a dug-out canoe, iron objects, a gold finger ring, bone combs, an early forged coin and a crucible for gold working. The crannog site at Dowalton Loch also proved rich in terms of its material culture. As well as producing an unusual find of a fragment of preserved textile, this site also revealed metalworking evidence (again, in the form of a crucible) as well as the bronze ring-handle from a cauldron or bucket, beads and an enamelled bronze fitting. A Roman patera and a sherd of Samian ware were also recovered. Munro (1885) also noted the potential for further crannog sites at ‘Loch Inch Cryndil’ (Black Loch), ‘Airrieoulland’ (Rough Loch) and White Loch (‘of Ravenstone’), an indication that the South West had the potential to be a hotspot for this particular type of settlement.

17. Placing these sites within a longer chronological sequence was soon accomplished. Joseph Anderson ascribed them to the very latest prehistoric age, though it was noted that their occupation sometimes continued through the Early Medieval and Middle Ages (Munro 1880, 21). Robert Munro ascribed both the Scottish and the Irish examples to the ‘Celtic’ civilisation, citing the recurrence of similar material culture and similar construction methods, as well as the common presence of dug-out canoes.

18. Crannogs dominated the literature during the last decades of the 19th century, particularly at the regional level, but other types of Iron Age settlement site were also subject to increasingly detailed levels of study. Burnswark had, for more than a century, dominated antiquarian awareness as far as Iron Age defended settlements were concerned. More widely, however, a broader interest in comparable sites had been slow to take hold amongst the antiquarian community. They had long been recognised at the local level, their original role often recorded through place names or traditional tales and folklore. It was only towards the end of the 19th century that several individuals began to actively identify, record and survey this potentially rich resource. Following on from this came efforts to create a classification system so these sites could be understood and discussed more widely amongst the archaeological community.

19. The 1880s saw a flurry of papers devoted to these forts and other prehistoric enclosed (and potentially defensive) structures. George Wilson presented an inventory of forts in Wigtownshire (Wilson 1885), a paper which is noteworthy for his mention of ‘hut circles’ in association. Wilson did not, however, seem to have reached a demonstrable understanding of the considerable time-depth which may be inherent in their occupation. That this floruit of interest in such visually prominent sites should happen at such a late date was commented upon by Frederick Coles: in 1890, he noted that ‘it is all the more a matter of surprise that objects so conspicuous as large earthworks (motes and forts) should, till quite recently, have escaped my special notice. My experience … differs little from that of others, who, whether natives of the district or not, seem never to have even seen the majority of these remarkable structures’ (Coles 1891, 352). Coles also observed (ibid.) that they had in many cases even escaped the notice of the Ordnance Survey surveyors.

20. Coles went on to publish two papers, which together provide an early inventory of 89 different defensive sites located within the Stewartry of Kirkcudbright. He classed these structures variously as forts, motes and doons depending upon morphology. His inventory includes sites which we would now recognise to be Early Medieval or even Medieval in date, but his work focused very much upon survey and recording rather than interpretation. Rare exceptions are apparent, however. He argued, for example, that the features at Castledykes, Kirkcudbright were representative of a ‘British fort’ (Coles 1891, 371) and suggested, too, that the site named ‘Roman Camp’ at Bombie was unlikely to be Roman on account of the poor standards of engineering evident (Coles 1891, 370).

21. Active in Ayrshire at around the same time was the antiquarian John Smith. Although he often sketched site layouts, he did not produce the detailed survey plans which were produced by figures such as David Christison. Smith was, however, mindful of the indigenous origins of such monuments, postulating a defensive role for them in the context of wider Roman military activities across South West Scotland. He noted, of these Ayrshire sites: ‘When the Romans came… they would find the people in possession of such places, and no doubt would have the considerable labour of dislodging them…’ (Smith 1895a, 2).

22. Final mention should be made of David Christison, a prominent antiquarian who served as Secretary to the Society of Antiquaries of Scotland. He carried out a series of regional surveys within lowland Scotland, focusing on prehistoric defensive structures and publishing regional inventories within the pages of the Proceedings of the Society of Antiquaries of Scotland. Classification and time-depth were essential elements of his work: he discussed the various classes of defensive earthwork – forts, duns, motes – and assigned them a likely chronological origin based on their morphology and character. Prehistoric indigenous British forts could be identified through the lack of engineering skill evident in their construction. These sites contrasted with the more formal arrangement of defences (invariably accompanied by a rectilinear, as opposed to curvilinear, plan) associated with Roman and later defensive structures. His regional gazetteers included Dumfriesshire (1891) and Ayrshire (1893). His work complemented that of Coles and slightly overlapped that of Smith, but he recognised the efforts of both men and acknowledged them in his own work.

23. In the late 1890s, Christison went on to author a report detailing the first recorded excavations at Burnswark. These had been undertaken during the 1880s by James Barbour for the Society of Antiquaries of Scotland. This work aimed to try and elucidate the complex arrangement of earthworks apparent at the site, with a view to establishing the nature of the Roman military activities there (Christison et al 1899). It marked the first of a long succession of archaeological investigations at the site, which continues to the present day.

24. On a much more modest scale was the small fort at Seamill excavated by Robert Munro (Munro 1882b). It revealed a range of occupation debris broadly similar to the material culture recovered from the crannog sites. Amongst the finds noted by Munro were a spindle whorl, the upper stone of a rotary quern, bronze, iron and shale objects, and the bones of ox, deer, pig, sheep and marine shells.

25. Amongst all the reports which record the investigation of various forts and crannog sites, there is one relating to a particularly unusual site at Borness Cave. Exploration of sediments within the cave revealed a range of material culture manufactured from bone and coarse stone, including whetstones, combs, pegs and awls. Some dateable material was also found, including a small Samian ware cup, a glass armlet and a bronze brooch. All inferred a date range spanning the 1st to 2nd century AD, which is contemporary with the Roman occupation of South West Scotland. The excavator, W Bruce Clarke, also reported finding human remains, including two skulls (Clarke 1876; Clarke 1878).

26. Meanwhile, finds of Iron Age decorative metalwork continued. Some small items of bronze were recovered from the productive dunes of Luce Bay which were still consistently revealing a broad range of finds which spanned an equally broad chronological span. These pieces – which were described as being ‘of a Celtic character-’ were included in an inventory of prehistoric metalwork from Wigtownshire (Maxwell 1885, 44). A later report by Maxwell focused on an enamelled bronze rein-guide or ‘terret’ found at Auchendolly (Maxwell 1886). The year 1894 saw the publication of an article by Robert Munro reporting the discovery of a bronze scabbard-sheath decorated in the La Tene style (Munro 1894), apparently found during the digging of field drains near Bargany.

27. The period 1900-1910 saw James Barbour carry out a number of investigations in the region. He joined forces with David Christison to aid landowner and archaeologist Richard Bell in the excavation of a prehistoric fort found on his property: Castle O’er (Bell 1905). Then, in September 1901, Barbour excavated a rectilinear enclosure defined by double banks and ditches named Rispain Camp (Barbour 1902). Here, the investigations produced little in the way of material culture, but a fragmentary human skull was recovered. Without the benefit of radiocarbon dating, the site – based on its rectilinear form – was interpreted as a medieval moated settlement, an interpretation which went unchallenged until the late 20th century. Barbour also carried out the systematic excavation of a crannog – Lochrutton – on behalf of the Dumfries and Galloway Natural History and Archaeology Society (DGNHAS), but the dateable occupation material in this case indicated a medieval date.

28. The sites that Barbour investigated were wide-ranging and varied, but amongst the most unusual was surely Castle Haven, near Kirkandrews, Borgue. Work began when local landowner James Brown, owner of the Knockbrex estate, sent workmen in to clear and consolidate the site. Barbour was invited to examine the monument, and he subsequently published a detailed survey and analysis of his findings. He noted the fort’s broch-like qualities, on account of its intra-mural chambers and passages, but acknowledged too that it lacked the height of a true broch, stating, ‘The fort may be said to stand to the broch in the relationship of a one-storey cottage to the four-storey house’ (Barbour 1907, 80), He compared it instead to Irish ‘cashel’ sites and suggesting it represented an influx of Celts from Ireland to Galloway (Barbour 1907). Amongst the items of material culture recovered were the upper stone of a rotary quern, various stone discs, glass and amber beads and several bronze finger rings and brooches, as well as animal bone, and, apparently, a human skeleton. Today, we recognise Castle Haven as a unique example of a galleried dun in Galloway, occupied during the pre-Roman Iron Age and the Early Medieval period.

29. Another unusual site investigated by Alexander Curle in the early 1900s was a broch at Teroy Fort, Craigcaffie (Curle 1912), which Curle claimed as the first broch excavated in Wigtownshire (Curle 1912). Again, a range of material culture was recovered with the upper stone of a rotary quern amongst the finds, along with metalworking debris, including a fragment derived from the nozzle of a bellows, or tuyere.

30. It was becoming increasingly evident that considerable time-depth was represented in the occupation material occurring on these late prehistoric settlement sites. This had first been noted during the crannog excavations undertaken decades earlier; now, two phases of occupation – separated through the identification of distinctive items of material culture – were often identified. These indicated use during the Roman Iron Age and the Early Medieval period (see Chapter 9 Early Medieval). A similar pattern was observed on several Ayrshire hillforts of varying plan explored by John Smith (Castlehill, Dalry and Aitnock Fort, Dalry – Smith 1919) and on an important coastal fort excavated by Alexander Curle: Mote of Mark, near Rockliffe (Curle 1914). What was also clear from some of these sites was the frequency with which the by-products of industrial production were recovered, with iron-working slags and the broken moulds from casting decorative bronze objects often found. Disentangling the information occurring on these complex, multi-period sites could, however, be difficult, particularly in this period, which came before absolute dating.

31. It was during the decade 1910-20 that the first detailed systematic field survey in the region was carried out by the recently created Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS). It focused on the county of Wigtown, including the Machars and the Rhins, and it was particularly productive in terms of the previously unknown prehistoric defensive structures it revealed. Large numbers of hillforts and promontory forts, as well as the much smaller duns, were added to the list of those already known.

32. It is perhaps surprising that in the decades following the RCAHMS survey, no notable episodes of fieldwork were recorded on potential Iron Age sites in the region. This apparent dearth of activity has led later researchers to suggest that the relative paucity of material culture finds encouraged figures like Alexander Curle to focus their fieldwork activities elsewhere (Toolis 2018, 11). Only Arran saw very limited work undertaken on the vitrified fort at An Cnap, Sannox by Victor Paton (Paton 1928). Two further examples of vitrified forts in Ayrshire – at Auldhill, West Kilbride and Kildoon Hill, Maybole – were included in a paper by Vere Gordon Childe and Angus Graham, which highlighted recent discoveries made by the RCAHMS surveyors (Childe and Graham 1943). This publication merely highlighted the sites with no further work undertaken in association. Instead, the attention of those working in the region focused more upon the Roman military sites, now becoming increasingly better understood through the onset of aerial photographic survey.

33. Published works of synthesis and analysis which focused on the region were similarly lacking, though there were two notable exceptions. Both were published in the Transactions of the Dumfries and Galloway Natural History and Archaeological Society (the Transactions), and both comprised broad regional syntheses of the evidence for prehistoric activity. The first was J Graham Callendar’s account of Dumfriesshire, published in the 1923-4 volume (Callendar 1924); the second, authored by John Corrie, focused on Kirkcudbrightshire and was published in the 1926-8 volume (Corrie 1928a). Both works included a section on the Early Iron Age, noting both the transition from prominent burial monuments in the Bronze Age to more modest affairs in the Iron Age, and the variation evident in the settlement types. The consistent presence of forts was noted in both areas, with crannogs more numerous in the Stewartry, but still occurring in Dumfriesshire, and souterrains (or ‘earth-houses’) still thought to be absent throughout the region.

34. Work undertaken during the 1940s was also sparse. Vere Gordon Childe and Angus Graham included the forts at Auldhill (named ‘Portencross’ by Childe), Seamill and Kildoon Hill, Maybole, within a wider survey of vitrified forts carried out in the wake of the various RCAHMS surveys. As far as excavations were concerned, however, the first reported investigations undertaken in the region appear to have been Cecily Piggott’s excavation of the Milton Loch 1 crannog, Kirkcudbright. This work revealed a wealth of palaeoenvironmental evidence, but the excavations took place before the onset of radiocarbon dating, so its 2nd century AD date was obtained through the typological analysis of a looped bronze fastener (Piggott, CM 1955).

35. A significant quantity of research and analysis into material culture was, however, being undertaken, shedding more light onto some of the region’s earlier discoveries. One important example was the Carlingwark Loch hoard, discussed in detail by Stuart Piggott in 1953. Piggott’s work largely focused on creating an inventory of the hoard’s contents and characterising them variously as either of Roman or ‘native’ British origin (Piggott, S 1955). Piggott subsequently presented a paper co-authored with Richard Atkinson on the Torrs chamfrein, further highlighting this important piece of Iron Age metalwork and exploring its early history as an antiquarian discovery and curio (Atkinson and Piggott 1955). Several years later, Robert Stevenson’s paper on indigenous British and Roman glass forms (including bangles) was published. It included several examples recovered from the region (Stevenson 1956).

36. Both Piggott and Stevenson had explored the interactions between the invading Romans and the indigenous community and how these may have been expressed through access to – and influence from – the exotic items of material culture the Romans brought with them. Eric Birley – also a prominent Romanist – similarly explored the relationships between local social groups and the Roman occupying forces, but his 1950-1 paper in the Transactions was concerned with the Roman written sources. He used these to obtain a clearer picture of the various indigenous groups living in the region at the time of the Roman incursions (such as those occupying the hillfort at Burnswark). Although his paper focused largely on the Brigantes (whose tribal epicentre was Yorkshire), he also considered the Novantae, who had traditionally been associated with South West Scotland and Burnswark. Birley suggested that they were probably allies of the prominent Brigantian chieftain Venutius, who rebelled against the Romans (then allied with his wife, Cartimandua). The siege of Burnswark could thus potentially be viewed as resulting from Roman efforts to quash this rebellion (Birley 1951). Just a few years later, the archaeological evidence from in and around Burnswark was placed in jeopardy through afforestation: in response to that threat, William Fleming Cormack carried out fieldwalking on the site, recovering further examples of lead slingshots, as well as evidence for domestic occupation in the form of a spindle whorl (Cormack 1958).

37. Working at around the same time was an Ayrshire-based amateur archaeologist Alexander McLeod of Darvel, who investigated the vitrified dun at Kemp Law, Dundonald. McLeod was active with both the Darvel Antiquarian and Natural History Society and the Ayrshire Archaeological and Natural History Society (AANHS), serving on the committee of the latter organisation between 1948 and 1954. McLeod’s activities on Kemp Law were reported in the local press, but his contribution to wider archaeological knowledge appears to have been overlooked by the mainstream archaeological community and his findings never found their way into the literature.

38. The onset of aerial photography was having a major impact in the study of Roman military sites, but its potential for shedding light on the indigenous Iron Age was appreciated less quickly. One unusual exception was the earthwork at Blacketlees, near Annan, identified first on aerial photographs and subsequently excavated in 1957 by Alfred Truckell, who had by this time been appointed Curator of Dumfries Museum. From the findings of the excavation, Truckell interpreted the site as a homestead dating to the 4th century AD, lying just within the Iron Age as it is defined here, as opposed to the Early Medieval period (Truckell 1957).

39. A larger and ultimately more high-profile series of excavations took place at Trusty’s Hill, Anwoth, during 1960, under the direction of Charles Thomas. Two weeks of trial trenching were undertaken at various locations on the site, revealing evidence for two phases of occupation. The later levels dated to the Early Medieval period; the earlier phase, which comprised a hilltop enclosure with a rock-cut ditch providing an additional defence across the promontory, was interpreted by the excavator as likely to be contemporary with the Roman Iron Age (Thomas 1961, 10).

40. Excavations were also carried out in 1960 on the hillfort site at Camp Hill, Trohoughton, by Major General James Scott-Elliot. His work revealed the remains of timber roundhouse structures within two stone-faced ramparts. Two years later, over several seasons in 1962 and 63, Scott-Elliot directed excavations by DGNHAS at a much more modest semi-circular ditched enclosure at McCulloch’s Castle, Arbigland. The site was located on a clifftop, with the line of its diameter following the cliff edge. Though no traces of internal structures could be identified, there was a hearth which contained several sherds of Roman pottery, including a sherd of Samian ware dating to the 2nd century AD.

41. The early 1960s also saw the excavation of another smaller site, the palisaded settlement at Castle Hill, Darvel. This was undertaken by members of the Kilmarnock and West Kilbride classes on Local History and Archaeology – organised by Glasgow University Department of Extra-Mural Studies and AANHS. Two phases of roundhouse structure were identified on the site, which also produced four sherds of Roman pottery, dated to the Antonine period. These included a sherd of Samian ware. The excavation report was written up and prepared for publication by Alastair Hendry (Hendry 2019).

42. The year 1961 marked a milestone with the organisation of the Conference on the North British Iron Age, held in Edinburgh (MacGregor 1976, 4). This event saw important contributions by Stuart Piggott and others. Five years later, a milestone volume was published, edited by Albert Rivet (1966): The Iron Age in Northern Britain. The book included an important paper by Stuart Piggott (1966), ‘A Scheme for the Scottish Iron Age’, presented as a counterpoint to Christopher Hawkes’ recently released ‘Scheme for the British Iron Age’, which had been weighted towards evidence from the south of mainland Britain (Hawkes 1959). Piggott subdivided the Iron Age of northern Britain into four phases: Iron Age 1 (550 to 350 BC); Iron Age 2 (350 to 150 BC); Iron Age 3 (150 BC to AD 80) and Iron Age 4 (AD 80 to 300).

43. Piggott considered that Strathclyde, Galloway and Dumfriesshire formed part of a wider Solway-Clyde province. In South West Scotland, evidence for the various phases was scarce. At the time, there was no evidence for Piggott’s Iron Age 1; Iron Age 2 was represented only by the Torrs pony cap, and Iron Age 3 was once again unrepresented. It was only with Iron Age 4 that artefactual evidence (in this case, represented by the Carlingwark Loch hoard) was supported by monumental evidence in the form of crannogs (Piggott 1966, 9). Posing a particular challenge for interpretation was, Piggott suggested, the absence of pottery and burials from the Late Bronze Age onwards, and a settlement type (the roundhouse) which remained consistent from the 2nd millennium onwards into the Iron Age.

44. In addition to Piggott’s ‘Scheme,’ the Rivet volume included two important papers which referenced the region, authored by Robert Stevenson and Richard Feacham, respectively. Stevenson’s contribution examined Iron Age metalwork, while Feacham discussed hillforts. Stevenson’s paper, which dealt with chronology and cultural affinities, noted that the lengthening of the Iron Age brought about in the wake of radiocarbon dating meant that objects once viewed as contemporaneous now required to be placed into chronologically distinct horizons (Stevenson 1966).

45. Feacham highlighted the substantial number of hillforts identified in northern Britain over the previous decade or so. He proposed that these demonstrated firstly, the existence of regional types, and secondly, an evolution from ‘small semi-nomadic pastoral societies’ into larger, ‘at least partly organised’ ones (Feacham 1966, 60). South West Scotland presented a similar pattern seen elsewhere in Scotland: small settlements such as brochs and duns dominated to the west, while large settlements, such as hillforts, were predominant in the east. Feacham also endeavoured to identify ‘tribal centres’ for each of the various tribes identified by Roman writers. Walls Hill, Renfrewshire, was proposed as the tribal centre for the Damnonii of Ayrshire and the Clyde Valley, while the territory of the Novantae, in modern Dumfries and Galloway, had four potential contenders. The most obvious candidate was Burnswark, but Castle O’er, The Moyle and Giant’s Dike were also possible contenders (Feacham 1966, 80).

46. Rivet’s publication was released at around the same time that Stuart Piggott was supervising the PhD research of Morna MacGregor. MacGregor’s work examined the various categories of Iron Age material in Northern Britain, which featured decoration that fell within the descriptive category of ‘Celtic Art.’ Objects embellished in this way mainly, though not exclusively, comprised non-ferrous metalwork.

47. MacGregor looked at these objects’ cultural affinities and defined four main ‘schools’ of decoration, each combining various elements of the ‘B’ and ‘C’ styles of Celtic Art, as defined previously by Sir Cyril Fox (Fox 1958). MacGregor’s approach was rooted in the diffusionist model, with subtle changes in artistic style linked to population movements. The work would subsequently be published in the two-volume set Early Celtic Art in North Britain (MacGregor 1976).

48. In 1970, just before the publication of MacGregor’s work, a groundbreaking exhibition focused on Early Celtic Art was held in the National Museum of Scotland. Stuart Piggott organised the exhibition and negotiated the loan of the various exhibits, which were sourced from a number of museums located throughout Europe. He also wrote the accompanying guide catalogue, which provides a useful definition of the established terminology used to define Celtic Art: the Waldalgesheim style, the Plastic Style and the Sword Style (Piggott 1970). Amongst this illustrious collection of material culture, four items from South West Scotland took their place: the Torrs Chamfrein/pony cap and its now-detached horns (displayed separately as drinking horn terminals), the Balmaclellan mirror and associated metal strip, and the Lochar Moss neck ornament.

49. Ongoing research into this aspect of material culture suggested that – in the wake of the Roman military invasion – the relationship between the indigenous local communities and the occupying Roman forces was not clear-cut. Roman objects were obtained and used at a local level, while some kind of ‘Celtic’ influence was apparent amongst material culture used by Roman forces. Perhaps this reflected the ethnically diverse nature of the Roman army. This meant that some instances where a mixture of Roman and indigenous material culture was apparent could not be easily categorised as resulting from either Roman or indigenous practice. The prime example where this fusion was at its most pronounced was the Carlingwark Loch hoard with its mixture of Roman and indigenous metalwork discarded in a waterlogged or wet place. Such a practice was entirely in keeping with local practices of metalwork deposition which predated Roman activities in the area.

50. The 1960s saw several small-scale excavations undertaken within the region. One was undertaken on the exposed surface of an artificial island in Loch Arthur, New Abbey, a body of water which had revealed a complete dug-out canoe with carved animal prow in 1874 (Henderson and Cavers 2011, 105). The excavator exposed a series of clay-packed drystone footings which were interpreted as the undercroft of a wooden building and ascribed a 15th or 16th century AD date (Williams 1971a). It was almost five decades later that the artificial island’s origins as an Iron Age crannog would be confirmed through excavation (Henderson and Cavers 2011).

51. Also active at this time was Alastair Hendry, who undertook a long-running series of excavations at Gourock Burn, West Kilbride. Working over an extensive period spanning 1961 to 1980, Hendry investigated two roundhouse structures, both contemporary, and dating to a period spanning the 1st century to the mid-2nd century AD (Hendry 2020).

52. At around the same time Hendry was also involved with the investigation of a homestead at Bankhead, near Darvel, in association with local students from the Glasgow University Department of Extra-Mural Studies and AANHS. The site had previously been explored by members of the local archaeological society during the 1920s. They had sought to find traces of a medieval castle (‘Castle Lowrie’) at the site, but the later excavations instead confirmed that the site was an enclosed settlement of Iron Age date. These earlier investigations had disturbed around a third of the site, but despite this damage to the archaeological deposits, it was possible to identify three phases of occupation, which included two successive roundhouse structures within, and partly utilising, enclosing timber palisades. Finds included a sherd of 2^nd century AD Samian ware, a scrap of copper alloy sheeting and a fragmentary quernstone (Hendry 2019).

53. Throughout the 1970s, little in the way of fieldwork graced the pages of the Proceedings of the Society of Antiquaries of Scotland (the Proceedings). There was, however, one exception. This was the report by John Hunter (1975), written in the wake of excavations on the site of a drystone passageway at Ardeer, Stevenston. The site had been re-used as an ornamental grotto during the 19th century, but the excavator argued that it had all the characteristics of a souterrain. This challenged the established view that to date, no souterrains had been identified in the region. Nonetheless, Hunter stood by his interpretation. He argued that the type of materials recovered from the site, including kitchen refuse and metalworking slag, but no pottery, as well as the methods of construction employed, were entirely consistent with a souterrain of the Angus group. The latter had been first defined by Frederick Wainwright following the publications of his excavations at Ardestie and Carlungie I (Wainwright 1963).

54. Some important work was, however, being reported at the regional level. This included the publication of excavations at Burnswark carried out by George Jobey in 1965. Equipped with a more nuanced understanding of local Iron Age settlement forms and patterns, informed at least in part by information obtained through aerial photography, Jobey re-visited the areas previously excavated by Barbour in 1898.

55. Jobey examined both the hillfort and the associated Roman camps. Amongst the new findings were the identification of four superimposed timber roundhouses within the interior of the hillfort, and the presence of a timber palisade predating the defensive ramparts. This produced sherds of a flat-rimmed ware consistent with a Late Bronze Age or Early Iron Age date, and a radiocarbon date of 800 – 430 BC, suggesting occupation of the hill in at least the earliest Iron Age, if not earlier.

56. In addition, Jobey demonstrated two phases of construction for the ramparts, with an earlier timber structure modified by the addition of a stone facing on the front edge. Jobey also identified traces of what he identified as earlier, local Iron Age or Romano-British features, including a possible settlement, at the site of Barbour’s ‘western fortlet’ (Jobey 1978).

57. An unenclosed roundhouse structure at Moss Raploch, Clatteringshaws was excavated at around the same time by J Condry and Michael Ansell. The structure was dated to the 1st or 2nd century AD based on two fragmentary glass rings. Also of interest at the site were potential traces of roasting hearths and iron-working slags, which James Williams suggested provided evidence that bog ore had been processed into iron at the site (Williams 1978). Williams was also active at the hillfort site of Tynron Doon, although the focus of his excavations was not on the prehistoric occupation of the site. He was concerned instead with later phases of occupation during the Early Medieval and Medieval periods (see Chapter 9 Early Medieval). Material culture recovered from the site nonetheless confirmed earlier occupation during the Iron Age.

58. In addition to his work at Burnswark, George Jobey investigated several types of enclosed settlement associated with timber roundhouse structures in the eastern part of Dumfries. These included unenclosed platform settlements, which – although identified in prolific numbers to the west – did not occur within his study area. Better represented were the types which he defined as palisaded settlements, hillforts, and the smaller defended settlements, as well as the non-defensive settlements with levelled, or ‘scooped’ interiors. He demonstrated how some of these sites – in particular, the hillforts and defended settlements – made clever use of natural defences coupled with either rock-cut ditches or ramparts. In addition to these various groups, he identified a more ambiguous form of enclosure. This was described as having ‘doubtful function and context’, potentially because examples tended to be in areas subjected to prolonged intensive cultivation. One such site, at McNaughton’s Fort, Kirkcudbright, yielded a pre-Roman Iron Age date of 280 ± 100 BC (Jobey 1971) following excavations undertaken by Major General James Scott-Elliott in 1965.

59. Jobey’s work on Iron Age settlements culminated in the excavation of one of these ‘scooped settlements’ at Boonies, Westerkirk. This settlement was similar in many respects to excavated examples in the southeastern part of the Tyne-Forth province, but it differed in one important aspect. The eastern examples consisted of an early, timber-built phase, followed by a later phase where both rampart and circular house-structures were rebuilt with a stone component. Western equivalents, like Boonies, lacked the later rebuilding in stone. Excavation demonstrated that the interior had once held thirteen certain or possible timber roundhouse structures of ring-groove construction, representing successive phases of occupation. The enclosure revealed evidence for a formal gatehouse structure with a paved roadway, and the excavations revealed items of Iron Age material culture. These included Roman and indigenous Iron Age pottery, coarse stone objects such as fragmentary rotary querns and pounders, a fragment of a glass bracelet and a copper alloy penannular brooch. A radiocarbon date of 108 ± 47 AD was also obtained from the site, with Jobey suggesting that its origins pre-dated the Agricolan invasion. Jobey argued for a minimum of seven successive phases of occupation, meaning the site was occupied, possibly on a continuous basis, right through into the Roman occupation of the area.

60. Several sites which were morphologically similar to Boonies had been identified throughout the eastern part of Dumfries. Termed ‘birrins’ or ‘burians,’ they had previously been interpreted as medieval cattle corrals, but while a medieval re-use could not be excluded, Iron Age origins for these sites were finally confirmed in the wake of the Boonies excavations. Jobey also identified another of these settlements close to the site of the hillfort at Burnswark. Known as the ‘West Fort’, it had been excavated by Barbour during his excavations at the hillfort and the associated Roman military fortifications (Christison et al 1899). Finds included a similar broken rotary quern and opaque glass bracelet to those recovered from Boonies, further confirming both sites’ similarity.

61. In 1976, Roger Mercer investigated the Iron Age enclosure at Long Knowe, Eskdale. He argued that this site differed from the lowland scooped settlements like Boonies, on account of its elevated location on a hill spur and its differing character, which featured a much larger internal area and a much less pronounced external boundary feature (Mercer 1981). Mercer’s excavation of what he described as a ‘farmstead enclosure of quasi-defensive type’ (1981, 69) revealed the stone-packed ring grooves of ten roundhouse structures. At least five of these showed two or even three phases of occupation. What was particularly remarkable about this site – which had been damaged by forestry ploughing and tree-planting just four years earlier – was the waterlogged remains from its flooded counterscarp ditch. The items recovered included wood (such as birch, alder and willow), and the bones of cattle, sheep and horse. No cereal grains were identified, which suggested to the excavator that occupation may have been seasonal and focused on pastoral production. The site produced early Iron Age dates spanning the 7th to 4th centuries BC, and occupation predated a period of peat expansion which later covered the site.

62. Until 1980, literature relating to Iron Age settlement in Scotland and Northern Britain tended to focus upon classification and the challenges of creating a definitive typology of settlement types. This approach went hand in glove with the cultural historical (or diffusionist) school of thought that predominated in archaeological discourse throughout much of the 20th century. Constructing detailed and accurate classificatory schemes and typologies was key to improving understanding. Various types could be differentiated by their distinctive attributes. Brochs, for example, could be distinguished by their drystone construction (often forming tall towers), which featured intra-mural stairs and chambers. Once defined, types could be mapped over space and time, and origin points sought for their various attributes. The absence or presence of types was then interpreted as reflecting movements of people. During the Iron Age, such population movement was invariably linked with the movement of indigenous tribes in relation to Roman military activity occurring either in the British Isles themselves or further afield in continental Europe.

63. The cultural historical approach was being challenged by followers of the processualist school, who might consider, for example, brochs or forts as central places situated to exploit a wide range of resources occurring throughout their immediate hinterland. This approach was advocated by figures like Colin Renfrew with reference to the Neolithic and Bronze Age, but its impact on the Iron Age of South West Scotland appears to have been minimal. Nevertheless, both the cultural historical and the processualist approach came under fire in the early 1980s when John Barrett challenged both orthodoxies in a paper on the Iron Age in Atlantic Scotland (Barrett 1982). Barrett rejected the view that artefacts were either type fossils/cultural badges, an approach exemplified by the approaches to Celtic Art evident in the work of Morna MacGregor (MacGregor 1976) or Robert Stevenson (Stevenson 1976). He also disputed that artefacts – and architecture– might represent adaptations to an environmental regime. Instead, material culture, in all its forms, played an active role in the formation, reproduction and transformation of social systems.

64. Barrett pointed out how the Iron Age of Atlantic Scotland was defined through several monument types. These included brochs, duns and crannogs, but there were others which were less well known. Settlements could, however, be seen as more than the sum of their architectural traits. Once they were established, they might be expanded or elaborated in episodes that required the mobilisation of labour and resources. They provided a framework of spaces and boundaries that distinguished between the occupants and their activities. Within these same spaces, cultural and social values might be learned or even challenged. As a means of illustrating how this approach might work, Barrett selected brochs as a specific category. Rather than seeking a point of origin for these distinctive structures, he suggested that there should be a focus instead on the patterns of activity and relationships structured through the medium of the house, that is, the ‘structured set of relationships between things, people and practices’ (Barrett 1981, 212).

65. How such an approach might be applied to the Iron Age of South West Scotland was demonstrated just a few years later in a paper by Lesley Macinnes, published in the Proceedings. This work (Macinnes 1984) challenged the orthodox views that saw brochs as defensive structures and examined instead how brochs might have functioned as an architectural manifestation of wealth in lowland Scotland. Macinnes suggested that the development of the lowland brochs resulted from a period of relatively peaceful interaction with Rome. The acquisition of Roman material culture and the development of the brochs could both be viewed as an expression of wealth acquired through this interaction, with the collapse of the brochs linked with the Roman withdrawal from the area.

66. The 1980s also saw the first application of a new dating technique to the region’s archaeological sites: thermoluminescence. The target for this work was the vitrified material occurring in various defended settlements across the region. Undertaken as a joint project by Paisley College of Technology and the National Museums of Scotland in November 1983, thermoluminescence dates were acquired from the fortified hilltop sites of Mote of Mark, Portencross (Auldhill), Dundonald Castle and the much smaller vitrified dun at Kemp Law, Dundonald. The method yielded dates that were earlier than had been anticipated, routinely falling within the 2nd millennium BC (Bronze Age) rather than the 1st millennium BC (Iron Age) as had been expected. The authors (Strickertsson et al 1988) saw no reason to doubt the veracity of these findings, but some caution should perhaps be exercised until the method is more widely used and better understood. Later archaeomagnetic dating on two of the sites covered by the project (Knock Farril in the Highlands and Tap O’Noth in Aberdeenshire) suggested much later, Iron Age, dates than those obtained through thermoluminescence (Caldwell et al 1998, 67).

67. A particularly informative fieldwork project was undertaken in this decade in Upper Eskdale, Eastern Dumfriesshire. Directed by Roger Mercer, the Eskdale Project focused on the hillfort at Castle O’er, and a nearby enclosure at Over Rig. The project included a programme of field survey in the vicinity of these two quite different sites, as well as excavation of the sites themselves. These investigations revealed that occupation of the hillfort at Castle O’er had spanned several centuries, with the earliest phase commencing perhaps in the mid-1st millennium BC and apparently declining in the first few centuries BC. There followed a period of hiatus when occupation ceased, and the function of the site changed with its incorporation into a wider land management system. Several centuries later, during the 3rd or 4th century AD, Castle O’er resumed its previous role as an important settlement and centre of political power (see Chapter 9 Early Medieval).

68. Unlike the long-lived Castle O’er, the enclosure at Over Rig appears to have been used over a limited period coeval with Roman military occupation in the area. It was around this time that Castle O’er was re-used as part of a complex arrangement of fields and enclosures potentially used to corral livestock. Over Rig appears to have formed part of the same system, but its role is enigmatic. Its location is prominent but liminal, inviting various interpretations regarding its purpose and function. These range from a ritual enclosure to a demarcated space where livestock such as cattle or ponies could be gathered for exchange or for tribute to the local Roman garrison (Mercer 2018, 234).

69. Developer-led work undertaken at this time included investigations by Halpin at the prominent Ayrshire hillforts, Harpercroft and Wardlaw Hill (Halpin 1992). Work at Harpercroft was instigated by modernisation work to the radio transmitter which forms a prominent intrusive feature in the interior. Halpin’s excavations revealed traces of roundhouse structures and some curvilinear ditches of ambiguous character within the fort’s inner enclosure. At Wardlaw Hill, both interior and defences had been greatly reduced through modern cultivation across the site. Nonetheless, Halpin was able to demonstrate two phases of rampart construction, with the potential for a timber element to have been present.

70. The late 1980s and 1990s saw further research into Iron Age site types that fitted in well with this post-processual interpretative framework. Sally Foster, who was undertaking postgraduate research at the University of Glasgow, explored the spatial arrangement and architecture of the Atlantic brochs, while Ian Armit reviewed evidence for the abandonment of souterrains in east central Scotland. Neither of these works included sites located within South West Scotland, but the arguments and observations proposed would be referenced by later fieldwork carried out in the region, such as Claire Williamson’s excavation of a complex souterrain at Auchrannie, Arran.

71. One important research project which had direct relevance in the region was a Historic Scotland-funded survey of crannogs, which included the recovery of radiocarbon dates from six crannogs. The results of this survey were published in a 1993 report by Anne Crone. The material used for dating comprised structural timbers, usually stakes, which were smaller and less likely to be derived from long-lived tree species like oak. Dates were obtained from crannogs at Barean Loch, Buiston, Loch Arthur and Lochrutton, confirming three phases of crannog construction. Most fell within a date range spanning 850 BC to AD 200, within the pre-Roman and Roman Iron Age. There was, however, some evidence of later occupation during the Early Medieval period (4th to 6th century AD), while two of the sites – at Barean Loch and Lochrutton – provided medieval dates. In most cases, the presence of artefacts consistent with occupation in the pre-Roman and Iron Ages suggests that most of these apparently much later sites had been established in the earliest phase (Crone 1993), with Lochrutton the only one likely to have a medieval origin. Crone also explored the chronological relationship between Scottish and Irish crannog sites, and concluded that crannog-building first appeared in Scotland early in the 1st millennium BC before being taken up in Ireland perhaps as late as the 6th century AD.

72. The crannog site at Dowalton Loch was also the focus of a paper by Fraser Hunter, published in the Transactions of 1994. Hunter examined the history of antiquarian and archaeological investigations at the crannog, which began with a visit by John Stuart, the Secretary of the Society of Antiquaries of Scotland, in 1886. Examining the site in the context of the recent crannog survey reported by Anne Crone, Hunter acknowledged that the crannog structures themselves might well date from the Early Medieval or even Medieval periods. Material culture from the site, however, indicated occupation stretching back into the Iron Age, and potentially even earlier into the Late Bronze Age. Hunter noted that the loch had produced fragments of a variety of copper vessels, with objects of Late Bronze Age, Roman Iron Age and Early Historic date all represented. The majority of these survived as complete, though sometimes well-used, objects. These vessels may all have been placed into the loch as offerings, as they were considered symbols of plenty. Similarly, the cauldron from Carlingwark Loch may have been a similar offering in a body of water which had also housed a crannog and produced high-status metalwork (a sword) dating to the Late Bronze Age. In both cases, Hunter suggested, the presence of crannogs appeared to coincide with ritually important locations.

73. The presence of objects of Roman origin and manufacture supported the possibility that both sites had been high-status dwellings, with ongoing research suggesting that Roman material culture was confined to only the wealthiest indigenous sites. Hunter’s interpretation marked changing attitudes to the origins of the three large Roman/indigenous Iron Age hoards found in southern Scotland at Cockburnspath and Blackburn Mill, both in the Scottish Borders, and of course, at Carlingwark Loch. Hunter ascribed them to indigenous ritual practices rather than the activities of Roman auxiliary troops operating in the area.

74. In addition to the Crannog Survey, the late 1980s saw an in-depth survey of linear earthworks. This class of monument was represented, in South West Scotland, by the Deil’s Dyke, a massive earth-and-bank feature that could be traced over 8km, but which had originally extended between three and five times that. John Barber, who had previously examined threatened prehistoric landscapes at Machrie Moor on Arran, led the work on this project. He established a classificatory scheme for these monuments, which often formalised Medieval or Early Historic political or economic boundaries. Crucially, he raised the possibility that they may have incorporated earlier monuments, originating as far back as the Iron Age or even earlier (Barber 1998). This raised the potential for the identification of Iron Age linear monuments within the landscape of the region, although the presence and purpose of such monuments must, for the time being, remain conjectural.

75. While the theoretical approaches to how Atlantic Iron Age sites and material culture were being debated and explored, new excavations were being undertaken on several sites across the region. Most could be classed as rescue excavations, such as work carried out in response to an unauthorised intervention at a broch at Stair Haven, Wigtown. The rescue excavation and survey carried out by Michael Yates focused on the structure’s northwest side and entrance (Yates 1988). Later in this decade, an excavation was undertaken at Tailburn, Moffat Water. Here, James Rideout examined an earthwork that appeared to comprise the much-denuded remains of a defended knoll or promontory: what Rideout described as a small promontory or terrace edge fort. Although the site exhibited the characteristics of an Iron Age defensive structure, no artefacts or environmental data were recovered to confirm this (Rideout 1988).

76. A more productive site was Rispain Camp, excavated by Alison and George Haggerty over a three-year period in the late 1970s and early 1980s (Haggerty and Haggerty 1983). The site was distinctive in being rectilinear on plan, as opposed to circular, sub-circular or oval. This had been suggestive, to early researchers, of either Roman or Medieval activity. Excavation of comparable sites in the north of England and South East Scotland had, however, established Iron Age origins. The excavations at Rispain Camp confirmed that this site, too, had similar early origins. The site produced several interesting artefacts, including a fragmentary bronze object with a glass or enamel inlay (Close Brooks 1983).

77. A decade later, excavation of a comparable square, double-ditched enclosure was undertaken at Carronbridge, Thornhill. The site revealed a long and complex history of construction and occupation, with three phases evident in the enclosure ditches. Traces of six intersecting circular roundhouse-type structures were identified in the interior, as well as three large sunken features with complex, charcoal-rich fills. The excavations also identified traces of a Roman temporary camp as well as a third enclosure, which, although overlain by the camp, had been abandoned before the Roman military activities on the site. A range of dates was obtained, but most fell securely within a range spanning from 70 ±50 BC to AD 210 ±50, that is, within the Late Iron Age and coeval with the period of Roman occupation.

78. A changing approach to interpretation and social function was also evident in this report, with the author arguing that the enclosure ditches may have had a purely social, rather than a defensive, function, serving as ‘boundaries of social exclusion’ (Johnston 1995). Moreover, it was suggested that the nature and perception of this boundary may have changed over time as the number of houses within the enclosure changed. Enclosures associated with just one building may, for example, have defined the property and status of the householder. Johnston also compared the morphological differences between the Carronbridge example and its Northumbrian counterparts. Johnston suggested that the different internal organisation might reflect the fact that Carronbridge was built on flat terrain, while its Northumbrian counterparts were built on sloping ground.

79. Works associated with the upgrading of the A74 in the early 1990s resulted in the excavation of two important Iron Age sites located in the eastern part of the region. In both instances, work was undertaken by the Glasgow University Archaeological Research Division (GUARD), directed variously by John Terry and Iain Banks.

80. The first to be investigated was Uppercleuch, Annandale, which comprised an oval enclosure first identified on aerial photographs in the late 1950s. Within the bounds of the enclosure was a single roundhouse structure, a smaller enclosure and a stone spread identified as a cobbled yard. The site was interpreted as a settlement functioning within a pastoral economy, a hypothesis supported by the large quantities of pollen derived from plant species characteristic of grasslands and pasture. Small quantities of barley, oats and in particular wheat were also present (Terry 1993).

81. The second was Woodend Farm at Johnstonebridge, Annandale. Excavated in 1994 and 1997 respectively, this site comprised a multi-vallate sub-circular enclosure with three enclosure banks and ditches. The site was dated to the Romano-British period, although it is likely that some of the enclosing features had earlier origins in the pre-Roman Iron Age. The interior revealed several superimposed roundhouse structures, forming eight separate blocks of features. Unlike comparable ditched settlement sites like Boonies, it showed no evidence for having had an earlier palisade feature pre-dating the more substantial enclosing banks and ditches (Banks 2002).

82. Developer-led work during the 1990s also led to the discovery of roundhouse structures in a part of the region where they had previously been unknown. An unenclosed settlement comprising two ring-groove type roundhouse structures was identified at Gallow Hill, Girvan, as part of extensive archaeological works carried out there. It was suggested by Michael Donnelly and Gavin MacGregor, who excavated the site for GUARD, that the settlement was likely to date to the Romano-British period (1st to 2nd century AD), on the basis of comparison with the ring-groove structures at Fox Plantation, Stranraer. This recently excavated site had returned dates spanning this range, although it was admitted that an earlier, pre-Roman Iron Age date could not be ruled out. The excavators noted, too, the proximity between this indigenous Iron Age site and the Roman temporary camps at Girvan, located just 1km away.

83. Archaeological investigations during the 1990s were not confined to developer-led excavations of individual sites. The late 1990s saw a major programme of work undertaken by Julian Thomas on several Neolithic sites in the vicinity of Glenluce and Dumfries (see Chapter 5 Neolithic). Amongst the sites excavated between 1994 and 1997 was a Neolithic henge at Pict’s Knowe, Dumfries, which had been re-used in the Iron Age (Thomas 2007a). The ditch of the henge had been re-cut, but the site itself was ambiguous, revealing evidence of a timber platform and hints of votive deposition, including one Roman shoe. The site also revealed evidence for metalworking, which included a crucible, but unfortunately, recent disturbance by rabbits had destroyed much of the archaeological information.

84. Important survey work was also carried out by RCAHMS, active in the east of the region at this time. Their ensuing volume devoted to Eastern Dumfriesshire was the culmination of an intensive survey and recording programme. It yielded detailed information relating to relict prehistoric landscapes, settlement and field systems, providing insights into Iron Age land use, occupation and agriculture (RCAHMS 1997).

85. Whether these land divisions were established in the actual Iron Age or made use of earlier field systems remains open to debate. Archaeological investigations of these features are infrequent, with the settlement itself usually being the target for excavations. Developer-led excavations at Warden’s Dykes, Gretna in 1990-1 did, however, investigate features in proximity to a banked enclosure that was inferred, on typological grounds, to be of Iron Age date. These included a small portion of a field system re-cut during the Roman Iron Age but set out earlier, possibly during the early Iron Age or even earlier. Traces of ephemeral structures and a roadway leading to the enclosure were also found. These could not be dated, but an Iron Age date was inferred from their proximity to the enclosure.

86. In a later counterpoint to this work, RCAHMS carried out a further survey in 2002-3 in the Kirkcudbright Training Area, part of the Ministry of Defence Estates. This encompassed a condition survey of upstanding monuments, including those deemed to be Iron Age in date. These comprised the circular footings of a roundhouse structure, an enclosed settlement (or ‘homestead’), three forts, and a modified littoral cave, which revealed Iron Age and Roman items amongst archaeological deposits (Cowley 2003).

87. The work of RCAHMS confirmed and highlighted the complex nature of these prehistoric field systems, but field survey could only go so far in informing the time-depth inherent in such sites. Archaeological investigations undertaken in association with the North Western Ethylene Pipeline in 1990-1 by Andrew Dunwell, Ian Armit and Ian Ralston provided an opportunity to examine a small portion of one example at Coats Hill, Moffat. An initial survey of features originally noted by David Christison almost a century earlier revealed a dense palimpsest of features. A total of 40 small cairns were recorded in association with two annular structures, a platform structure and a more substantial cairn to the north. Unfortunately, no dateable evidence was recovered which could place these various elements into a particular chronological phase. There were, however, two enclosed settlements of Iron Age type, which could be matched with similar excavated examples of 1st to 2nd century AD date (Dunwell et al 2000). Some of the small cairns were also found in proximity to Medieval rig-and-furrow cultivation, which suggested that this landscape had been actively farmed, though perhaps only sporadically, for at least two millennia.

88. Another excavation, which looked at the relationship between a settlement site and the field system in its immediate hinterland, was undertaken by Richard Gregory at Hayknowes Enclosures, Annan. The site comprised a double-ditched enclosure with a roundhouse structure in its interior, which produced a mid- to late pre-Roman Iron Age date. A series of linear boundaries and one rectilinear enclosure were also explored, suggesting to the excavator that the site would have been placed to control the local agricultural resource, and in particular, livestock.

89. The last few decades of the 20th century also saw further light shed on the Iron Age as unintended consequences of several excavations targeting much later medieval castles. The excavators investigating three sites – Cruggleton Castle, Auldhill and later, Dundonald Castle – discovered residual material culture indicating early phases of occupation during the Iron Age, or potentially even the Late Bronze Age. The excavations at Cruggleton Castle, which took place between 1978 and 1981, revealed traces of a roundhouse structure dating to the 1st century AD (Ewart 1985). Excavations at Auldhill took place slightly later, between 1987-9, and revealed evidence of a timber-framed rampart that had been burned and was, in places, vitrified. This, along with some possibly associated features, was ascribed, potentially, to the Bronze Age. It is, however, possible that these features – as well as the smaller dun which succeeded the timber-framed fort – had Iron Age origins (Caldwell et al 1998). A bronze enamelled trumpet brooch previously recovered close to the excavated area certainly confirmed occupation during the 1st to 2nd century AD. At Dundonald Castle, prehistoric pottery was recovered in association with possible hearth deposits (Ewart and Pringle 2004). At a fourth site – Rowallan Castle – structural evidence for two roundhouse structures was identified on the site of the medieval castle’s ruined northeast tower.

90. One last discovery made during this period which is certainly worthy of mention is a very modest artefact recovered from amongst sand dunes at Brighouse Bay (Maynard 1994). This was the mould for a Roman coin of a type known as the double denarii that was introduced after AD 288. What was interesting about the coin mould was that it was one of several moulds used to counterfeit Roman coins at around the same time. The location of the find lay a considerable distance from known centres of Roman occupation and activity, suggesting that the counterfeiting was by indigenous craftspeople working from a local site.

91. The year 2001 was a landmark year for archaeological investigations when it marked the first confirmed discovery of a souterrain in the region. Arguably, this statement should be qualified: it seems quite likely that the stone-lined passageway identified at Ardeer almost a century previously was another example of this monument type. Indeed, John Hunter, who excavated the site in the 1970s, had argued that this was indeed the case. However, the Ardeer example had been greatly modified for re-use as a modern ornamental landscape feature, which meant that for decades, this apparently unique example had been hiding in plain sight.

92. Souterrains were thought to be characteristic of the frontier zones which lay beyond the Antonine Wall. They are particularly numerous in east central Scotland (Angus and Perthshire) with examples also found in the Western Isles. Interpretation of these sites first favoured a defensive origin, but by the 1950s, a role in the storage of surplus foodstuffs was proposed. A potential contemporaneity with the Roman military occupation of Scotland led to the suggestion that the food storage was linked with Roman military logistics. Surplus grain may have been acquired by the Romans either through purchase or as tribute. The 2001 works at Auchrannie identified what the excavators described as a ‘complex souterrain’ (Mudie 2007, 1). It was associated with a roundhouse structure 14.5m in diameter, which was represented by a post circle. The roundhouse was part-excavated by George Mudie of CFA Archaeology during this phase of works, with the balance of the roundhouse and the souterrain remaining unexcavated, and the results of these initial excavations published in 2007 (Mudie 2007).

93. The year 2001 also marked the publication of an important paper by various senior academics which included Colin Haselgrove, Ian Armit and Timothy Champion amongst their number. They argued that South West Scotland was one of a number of ‘black holes’ across the country with gaps in our understanding resulting from a comparative lack of research in recent times and a corresponding absence of theoretical paradigm through which evidence could be understood and interpreted (Haselgrove et al 2001).

94. Just six years later, an article challenging their analysis was presented by Graeme Cavers (Cavers 2007), who argued that this perceived paucity was largely illusory. Instead, certain elements of Iron Age material culture could be used to explore questions of identity and social structure. These issues could be addressed by the variation apparent in domestic settlement types, and the fact that some of the settlement forms were monumental in character.

95. It was during the same decade that several important excavations reached publication, following works undertaken during the late 1990s. A good proportion of these were developer-led, carried out in association with major roadbuilding works or large-scale utilities projects.

96. Sometimes the Iron Age material was encountered during excavations targeted at sites associated with much later buildings that were being investigated, with the discovery of residual prehistoric material at Dundonald Castle being a particularly obvious example. In other instances, important in situ Iron Age deposits were revealed as part of larger multi-phase sites. One such example was The Leven, Loudon Hill (Atkinson 2000). Here, two phases of double palisaded enclosure were identified. The later enclosure, which was larger than its predecessor, was dated to the Iron Age. The earlier may have had an earlier Iron Age origin, but a late Bronze Age date could not be discounted.

97. Cavers was the first of several postgraduate researchers who actively explored aspects of Iron Age sites in the region during the first two decades of the 21st century. Focusing on crannogs, Cavers (2005) highlighted the significant role played by the South West crannogs in the recent survey of these sites carried out by Anne Crone and others (Crone 1993). Cavers’s thesis explored the crannog sites of South West Scotland, as well as those elsewhere in Scotland, considering all aspects of crannog construction and occupation from a post-processual and contextual perspective.

98. Crannogs continued to be a focus for investigations in South West Scotland during the first decade of the 21st century. In 2002-4, a crannog at Loch Arthur was investigated by Cavers, working with AOC Archaeology and in conjunction with John Henderson of the University of Nottingham’s Underwater Archaeology Research Centre. The crannog was revealed to comprise a massive ‘packwerk’ structure composed of timber stakes stabilising a brushwood mound. This had been constructed during the second half of the 1st millennium BC, but the site was later re-used in the Medieval period. Slightly later, in 2006, Dorman’s Island, Whitefield Loch was subject to test-pitting by Cavers, Crone and others as part of the Scottish Wetlands Archaeology Programme. This was aimed at addressing the under-representation of Scotland’s rich wetland archaeological resource in wider research programmes (Cavers et al 2011).

99. The excellent results obtained by applying modern scientific techniques to the crannog site at Dorman’s Island led to further work. A promontory crannog at Cults Loch, Castle Kennedy, was subject to detailed study, in association with various sites in its hinterland. These included a promontory fort and a ditched and palisaded enclosure (Cavers and Crone 2018). While the crannog itself revealing a relatively tight date range which spanned several decades during the 5th century BC, the investigations were able to demonstrate that it formed just one part of a larger, integrated landscape which was established prior to the crannog’s construction and extended after it fell into disuse.

100. Attention then turned to the Black Loch of Myrton, which was excavated in 2013 and 2015 by AOC Archaeology as part of the Cults Landscape Project. Like Cults Loch, the crannog site at Black Loch of Myrton revealed an exceptionally well-preserved roundhouse structure, which was able to inform on the ongoing modifications of the site and the way in which the roundhouse form played a crucial role in structuring the daily life and activities of an Iron Age household (Crone and Cavers 2026).

101. Another possible crannog site investigated during the first decade of the 21st century was Newbarns, which was subject to ten seasons of excavation. Work was directed by Alastair Penman, in association with the Stewartry Archaeological Trust (Penman and Penman 2002; Penman and Penman 2004). The site comprised two artificial islands in a marshy area which had formerly comprised a body of water. Unlike many of the crannogs identified and investigated in the region, the Newbarns comprised a mound built of stone, as opposed to mounded brushwood secured by stake piles (the ‘packwerk’ type). Further investigations revealed that Iron Age occupation of the site formed a late phase of activity, while the site itself did not fit the criteria expected of a crannog. Instead, the mounds appeared to be funerary monuments originally constructed during the Bronze Age, even potentially having origins in the Neolithic. The change of use to a crannog came about after the transition to a colder, wetter climate, which had resulted in the hollow which contained the cairns becoming much wetter and more waterlogged.

102. Tessa Poller’s research presented a contextual approach to the analysis of Iron Age settlement sites in Galloway (Poller 2005), moving beyond traditional typologies and applying instead more flexible processes of categorisation. Paul Murtagh took a similarly contextual approach to the Iron Age settlement of Ayrshire and the Clyde Valley (Murtagh 2014), challenging the established and largely uncritically accepted typologies first proposed by RCAHMS. Also undertaking postgraduate research in Galloway was Ronan Toolis, who focused on forts and promontory forts, undertaking targeted small-scale excavations on several sites during his research (Toolis 2018). These sites included the promontory fort at Carghidown and the hillfort at Trusty’s Hill, Anwoth.

103. Developer-led excavations at Whitecrook Quarry, Glenluce were undertaken by Douglas Gordon of Rathmell Archaeology in 2006, revealing an unenclosed settlement comprising two ring-groove structures located outwith a palisaded enclosure. No traces of any structures were found within the palisaded enclosure, with the excavator suggesting that the palisaded enclosure would have been used for stock control. Radiocarbon dates covered a period spanning the Late Bronze Age and Early Iron Age, with a plain ceramic vessel, found crushed but complete, also consistent with this date range and suggestive of an act of structured deposition (Johnson 2009).

104. Further discoveries of roundhouse structures in the first decades of the 21st century included two possible examples excavated by Ross Murray of Headland Archaeology at Corton Road, Ayr, and a single ring-groove roundhouse excavated by Christine Rennie of GUARD Archaeology at a multi-period site in Monkton (Rennie 2015).

105. The year 2007 saw the publication of the first phase of works at Auchrannie, Arran, but it was not until 2013 that the second phase of archaeological investigations was undertaken at the site by Claire Williamson of Rathmell Archaeology (Williamson 2017). This comprised the excavation of a souterrain along with the remaining unexcavated portion of the roundhouse structure, heavily truncated by later agriculture. Further pit features were also identified and examined.

106. The souterrain at Auchrannie proved particularly interesting as its various elements differed markedly in form and composition. On plan, the souterrain itself was complex, with side passages and branches. Some elements were stone-lined; others had been lined with timber (Williamson 2017). One of the passages was backfilled around or slightly after the 2nd century AD, a date consistent with souterrain abandonment across Scotland. The other appeared to have been left open, as a void, until the medieval period. The decommissioning of the souterrain did not, however, mean the abandonment of the site, with radiocarbon dates suggesting that some form of occupation had continued, though at a much lower intensity.

107. Later that year, Rathmell Archaeology excavated another site in the region which would prove to be both important and unusual. This was the complex multi-period site at Hunterston, excavated by Douglas Gordon over the winter of 2013-14. The site occupied a hollow, located near the coast and in the vicinity of the hilltop site of Auldhill, Portencross, which had evidence of occupation in the later prehistoric period (Caldwell et al 1998). The hollow revealed residual evidence for use in the Mesolithic, and it also produced evidence for a succession of substantial timber settings dating to the Neolithic and Bronze Age. Occupation of the site culminated with the construction of four circular structures. One of the structures measured 4m in diameter, making it slightly smaller than average for a roundhouse-type dwelling. Large quantities of ironworking debris, including primary metalworking residues such as hammerscale and slag spheres, were found inside its footprint and in its immediate vicinity. Metallurgical analysis indicated that both smithing and smelting had taken place at the site (McDonnell and Turner forthcoming). Traces of a stake-built internal structure, described as ‘snail-shaped’ on plan, were also identified within the structure; this was comparable to a similar structure identified at Bryn-y-Castell, Llan Festiniog, north Wales (Gordon and Turner forthcoming, 140).

108. Iron-working at Hunterston appeared to take place late in a complex sequence of occupation and use at the site, with evidence for an earlier Late Bronze Age/Early Iron Age palisaded settlement still discernible as vestigial traces. The evidence obtained from this site was consistent with a model for Early Iron Age metalworking put forward by Andrew Heald in his 2005 PhD thesis (Heald 2005). In the Early Iron Age, he suggests that the smelting and smithing of iron took place in liminal, socially ambiguous places located outwith the boundaries of settlements. Later in the Iron Age, ironworking ceased to be viewed as socially, culturally and ritually polluting. Instead, it became incorporated into settlement sites like hillforts. In essence, the process became ‘domesticated.’ This hypothesis fitted in with the evidence from Hunterston, which suggests that the iron-working site in the hollow was abandoned after the Early Iron Age, while occupation – and potentially craft production, including ironworking – continued at nearby Auldhill. It is possible that Hunterston’s re-use of a much earlier site for Iron Age metalworking was comparable with that occurring at Pict’s Knowe. Unfortunately, later disturbance by rabbits at the latter had potentially erased primary evidence for such activities, leaving only secondary traces (including a crucible) behind (Thomas 2007a).

109. The year 2015 saw the start of the Scottish Iron Age Vitrified Hill Fort Project (SIAVH), undertaken by the University of York under the direction of Carol Lang. The project included geophysical survey and excavation work at one of South West Scotland’s so-called ‘vitrified’ hillforts at The Knock, Largs. Trial trenching targeted the fort’s southern ramparts, revealing evidence for burning and traces of an external wall face. However, no traces of vitrification were identified, with the excavator suggesting instead that earlier explorations had misinterpreted weathered basalt bedrock as evidence for vitrification. Whether the situation was similar on the north side of the site remained unknown, as this area remained unexplored.

110. Work at a more modest hilltop enclosure was also taken at Little Wood Hill, Threave during 2019 as part of the Galloway Glens Landscape Partnership ‘Can You Dig it?’ Project. Little Wood Hill is the smaller of two univallate ditched enclosures occupying the summits of adjacent low hills near Castle Douglas. Excavations produced a 1st century BC to 1st century AD date for the site, confirming an Iron Age date for the enclosure ditch, but no evidence for internal structures was found.

111. During the early 2020s, there was further research exploring the indigenous settlement patterns of eastern Dumfriesshire, with a view to better understanding interactions with the Romans. This began with a pilot study named ‘Between the Walls’ before it progressed into the main project, ‘On the Edge of Empire.’ The project involved collating data relating to earlier excavations as well as carrying out new survey work using remote sensing and geophysical survey. The aim was to improve the understanding of settlement patterns in the wider environs of Burnswark fort. A geophysical survey was undertaken by the German Archaeological Institute, directed by C Rummel, in September 2021 (Fernández-Götz et al 2022). This examination of a limited area raised the number of known or probable Iron Age settlements from 570 to 704, an increase of 20%. This is remarkable considering that half the area had already been included in the Eastern Dumfriesshire study area covered by RCAHMS in the 1990s (Cowley et al 2022, 16).

112. Iron Age sites have continued to be revealed and excavated in the region. A double ring-groove roundhouse was identified at Cowdie Knowe, Urr during the construction of a stretch of gas pipeline linking Cluden and Brighouse Bay in 2017. It yielded radiocarbon dates which spanned a range extending from around 350 to 120 BC. A more unusual discovery was the region’s third souterrain, discovered at Kilmarnock Road, Mauchline by Fraser McFarlane and Stephen Cox of Headland Archaeology. Unlike the previously excavated examples at Ardeer and Auchrannie, the Kilmarnock Road example was located much further inland. The souterrain appears to have been timber-lined and was found in association with a roundhouse structure and a former fence line. However, the radiocarbon dates from these structures dated from the Middle Bronze Age, with the excavators arguing for an earlier than previously thought origin for them (Cox and McFarlane 2026).

113. Other recent archaeological investigations of Iron Age sites include excavations by GUARD Archaeology at the Curragh South, Girvan led by Jordon Barbour and Dave McNicol. Here, two phases of Iron Age settlement, each comprising a roundhouse with an associated four-post ancillary structure, were identified, with the first phase originating from a period spanning the Late Bronze Age and Early Iron Age (1550 to 600 BC). A brooch of Roman origin was also recovered from the site. The site revealed a pattern identified elsewhere in the region: the earliest phase of settlement was unenclosed, while the second phase comprised a roundhouse and ancillary structure within a palisaded enclosure. The latter featured a substantial post-defined gateway, potentially constructed to make the site appear imposing and impressive.

114. Many of the Iron Age sites investigated in the first few decades of the 21st century represented examples of the earth-banked or ditched settlements characteristic of the eastern part of the region. One unusual exception was High Milton Dun, Craigoch, which was a stone-walled variant of the monumental roundhouse form. These structures are classed either as small stone-built forts, homesteads or duns. Damage from stone-quarrying required emergency excavations. This work was carried out by Graeme Cavers and Jamie Humble of AOC Archaeology, revealing a stone-walled roundhouse with occupation deposits likely to date from the early 1st century AD.

115. New discoveries continue to be made. Archaeological investigations associated with the Dunragit Bypass revealed a multi-period site at South Boreland, which was excavated by Kathy MacIver and Katy Walker of AOC Archaeology in 2020. Two concentrations of Iron Age settlement were revealed, comprising an area of unenclosed settlement to the north, and two enclosed areas of settlement to the south (MacIver and Walker 2024). Evidence for ring-groove roundhouses was found in association with both areas, and while the relationship between the areas remained ambiguous, the northern, unenclosed, settlement appeared to postdate the earliest buildings to the south. Initial dating has suggested an Early Iron Age date for all these elements.

116. There are still, however, instances of new discoveries which do not quite fit within the current understanding of regional settlement forms. Excavations at Weirston Road, Kilwinning, by Alex Wood and Kevin Paton of AOC Archaeology in 2017 revealed a settlement comprising a number of overlapping ring-groove structures with two opposing entrances (Wood and Paton 2025). These roundhouse structures of unusual form yielded dates spanning the Early to Middle Iron Age. Later excavations led by Steven Watt of AOC Archaeology to the south extended the evidence for occupation and cultivation at this site (Watt 2026).

117. Another unusual site is Adie’s Brae, Moffat, which was excavated in October 2024 by volunteers associated with the Destination Tweed community archaeology programme, supported by Katie O’Connell and Graham Cavers of AOC Archaeology. The site comprised a scooped settlement which revealed evidence for two roundhouses, pits and postholes. What was particularly striking was the discovery of the cremated remains of a human male aged at least 17 years old, on a site which produced a date range spanning the second half of the 1st century BC and the end of the 1st century AD, which is the Middle Iron Age. It is thought the deceased had been placed either within or beneath the building (https://destinationtweed.org/iron-age-human-remains-identified-in-south-of-scotland-excavation: Accessed 18^th May 2026).

118. Recent re-examinations of well-known datasets have included Tiffany Treadway’s doctoral research into patterns of wetland deposition in Wales and Scotland. Her thesis, submitted to the University of Cardiff in 2021, included case studies from South West Scotland. In addition to well-known finds like the Carlingwark Loch and Balmaclellan hoards, Treadway included some more obscure items. These ranged from an iron axe from Loudon Hill to a wooden plough beam from Whitereed Moss, Lochmaben (Treadway 2023). In several cases, however, an Iron Age date had to be inferred rather than confirmed, as it was not possible to date the objects through either radiocarbon dating or through typological means.

119. Meanwhile, there have been new investigations at some of the region’s most well-known and intensively studied sites. In the early 2020s, staff from AOC Archaeology supported volunteers in site clearance, survey and consolidation works at the Doon Castle broch, Ardwell as part of the Rhins Revealed project. Work included the creation of a detailed computerised image of the site using 3d modelling.

120. Still ongoing are the continuing investigations of Burnswark. This centres upon the nature of the occupation on the hillfort itself and its environs, aspects of which are crucial to interpreting the Roman siegeworks located at the base, and to assessing the nature of interactions between the occupying Roman military and the indigenous population. One of the approaches used in its ongoing study applied methods of analysis employed by the US military to assess the relationship between the fort and the Roman works to better understand the nature of the Roman action at the site (Brown et al 2025).

---