If you look at an Ordnance Survey map from the 1950s or 1960s you will see plenty of archaeological features – but search for them on the ground today and you will be disappointed. Many of the barrows that were still visible when the maps were produced have since been ploughed flat. That leaves hillforts as the most numerous and most visible of upstanding prehistoric monuments, though, as Sir John Lubbock – one of the prime movers behind the Ancient Monuments Protection Act of 1882 – observed, they have been ‘utilised as manure, road metal, and building material’ in the past.
Numerous as they are, they are not well understood – to quote a recently drafted research framework, the basic questions about hillforts – built by whom, when, and why – remain to be answered. At least we now know where they were built. Between 2012 and 2016, a team from the universities of Oxford, Edinburgh, and Cork set out to draw up an atlas of hillforts (see ‘Further reading’ on p.42), which now exists as an open-access dataset complementing a 500-page monograph, packed with distribution maps with commentaries largely written by Gary Lock and Ian Ralston.
Defining hillforts
Inevitably, the first question anyone compiling such a resource will encounter is the definition of a hillfort – not so easy to say, it turns out, because ‘hillforts’ are very varied in form, scale, and size. As Barry Cunliffe observed in The Wessex Hillforts Project (2006), ‘there may be no such thing as a typical hillfort’. Date was rejected as an essential criterion for inclusion because there is no dating evidence for many hillforts, and although many are of Iron Age origin, there are similar monuments of middle and late Bronze Age date and, notably in Scotland and Wales, some from the early medieval period. Nor was much attention paid to function in selecting sites, if for no other reason than that interpreting hillfort-use is all but impossible in most cases – some served multiple purposes that changed over time, and many remain unexcavated so their function is unknown.
The team eventually decided on three criteria: topographical position, the scale of the enclosing works, and the size of the enclosed area. To qualify for inclusion, the site needed to be set in a dominant position within its landscape, though that does not mean that it had to be located on the highest point: a valley site set in a bend in a river would count, and in lowland landscapes such as East Anglia or the gently undulating plains of Lothian, topographical dominance can be relatively slight.
Since there are many types of enclosed site, including farmsteads with banks and ditches to control livestock, the site should also ‘show some pretension’ in the form and scale of its enclosing ramparts, ditches, entrances, and outworks. Upstanding multivallate sites were regarded as undoubtedly meeting this criterion, whereas ploughed-out sites only did so with a minimum ditch width of 4m.
These first two criteria are entirely subjective and, for this reason, more than one team-member had to agree on a site’s inclusion, and it had to satisfy at least two of the three requirements. The third – size – is the only one that can be quantified rigidly, and yet it proved the most contentious of the three: setting the minimum size at 1ha would have excluded sites that have traditionally been recognised as hillforts and, after field visits to small sites in Galloway, Cumbria, and Northumberland, the team adopted a threshold of 0.2ha, though sites smaller than 0.2ha are included if they satisfy the other two criteria.
Headline figures
The Atlas database contains 4,147 sites, of which 3,354 (81%) clearly met the criteria for inclusion, and the remainder are given the benefit of the doubt. An attempt was made to gauge how many of these had been subject to excavation, either of their enclosing works or their interiors. In England, 34.9% have had some degree of interior excavation, being more than half (53.2%) of all hillfort excavations, which explains the dominance of English sites in many discussions and publications: only 15.3% of sites in Wales have been excavated, 10% in Scotland, and 7.2% in Ireland – only the Isle of Man, at 31.8%, compares with England. Similar percentages apply to the excavation of hillfort entrances and enclosures, though more of these have been dug overall, confirming that banks and ditches have proved more attractive to archaeologists than interiors, which might account for the emphasis on fortification as distinct from other uses.
This touches on an important strand in the interpretation of hillforts: the long-standing belief that their purpose is military and that those who constructed them are assumed to have chosen sites with defensive characteristics. More recent commentators have suggested a much wider range of possibilities for the siting of hillforts in the landscape: the availability of fresh water for people and livestock; access to timber and stone; the need for shelter from storms; the agricultural uses of the surrounding land for pasture, arable, or both; the movement of people, animals, and resources in the surrounding landscape; and proximity to established land or riverine trade routes.
Others have argued that the visibility of hillforts is of considerable importance – what can be seen looking out from the hillfort, but also what can be seen of the hillfort by people moving around the surrounding landscape. This has led to the interpretation of hillforts as places where power is symbolically displayed, the architecture and the location making a statement about the prestige of those who constructed the monument, and perhaps about the ownership and control of the landscape and its resources.
In considering this issue, the authors of the Atlas cite the work of Toby Driver, of the Royal Commission on the Ancient and Historical Monuments of Wales, whose study of hillforts in Ceredigion (see CA 318 and ‘Further reading’) has identified a number of features that offer clues to the choice of the site, the form of the ramparts, and the intentions of the builders. Some hillforts have a large number of ramparts that are bigger around their entrances than elsewhere around their circuits, where the ramparts can be slight or even absent. This suggests the intention was display rather than defence, and a symbolic rather than functional architecture. This public front, with imposing ditches and ramparts, implies the choice of a site that can be seen from a variety of different directions, and perhaps a deliberate channelling along particular routes to heighten the experience of those approaching it.
This insight, along with the recent discovery of a late Iron Age chariot burial associated with a Pembrokeshire hillfort (CA 355), has encouraged archaeologists to begin to survey the entrances and routes taken by users and visitors to hillforts, with the expectation that there is much to be learned about such monuments from the study of their hinterland as well as the fort itself. Jessica Murray, in a PhD undertaken at Oxford as part of the Atlas project, has added substance to Toby’s ideas by using GIS (Geographical Information Systems) software to model pathways through the landscapes surrounding selected hillfort sites in Ceredigion, the Gower Peninsula, Dartmoor, Wiltshire, and Aberdeenshire. By looking at which features are visible from different distances, she has been able to demonstrate that the most-accessible approach routes often face the most-prominent and numerous ramparts in a site’s circuit.
The effect is especially impressive when approaching hillforts built on high peaks from below. The highest 100 hillforts are mainly found in Wales, Scotland, and Ireland. The highest of all is Faha, Co. Kerry, at 820m. In Britain as a whole, Ingleborough in North Yorkshire bears the crown at 720m. In Wales, the highest fort is at The Whimble, Radnorshire (590m); in Scotland, at Ben Griam Beg (580m), and in the Isle of Man, at South Barrule (586m). Sometimes the highest hill is chosen – in Berkshire, Walbury Camp (295m) occupies the highest point in the southern English chalklands; Oxfordshire’s highest point (260m) is crowned by Uffington Castle. But 40.8% of hillforts lie at 100m or less above sea level, including some (such as Cuddy’s Cove in Northumberland and various sites in East Anglia) that lie below 50m, once again confirming that such sites are not exclusive to high-altitude settings, but are chosen for the hinterland that they overlook and their visibility from the rest of the landscape.
Are there different types of hillfort? The Atlas includes maps based on topographic setting (‘contour fort’, ‘promontory fort’, ‘hill top’, ‘cliff top’, ‘plateau fort’, ‘hillslope’, ‘defended enclosure’, ‘level terrain’, ‘marsh fort’), shape (rectangular, circular, sub-circular), and the size and complexity of their ramparts (univallate, bivallate, multivallate with closely or widely spaced ramparts). No particular classification has gathered universal acceptance, however, and the final verdict lies with Ian Brown, who wrote in Beacons in the Landscape (2021) that ‘because of the inherent variability of both hillfort types and siting, any classification will inevitably be spurious’, and Dennis Harding, who argued in Iron Age Hillforts in Britain and Beyond (2012) that if we accept that hillfort siting is determined by inherently variable local topography, there cannot be any standard plan for a hillfort.
The title of Harding’s book raises the question of where else hillforts are found beyond Britain and Ireland, and whether there is any evidence for cross-cultural influence. Hillforts and cognate sites are noteworthy features of the later prehistoric settlement record in many parts of Continental Europe. Dating from the middle and late Bronze Ages and the subsequent Hallstatt Iron Age (1500- 800 BC), these sites have been linked to the centralisation of social and political power and related conflict, being associated with such evidence of warfare as quantities of weaponry and signs of destruction by fire.
Regional variation
In southern central Britain, the apogee of hillfort-building is the middle Iron Age, corresponding to a period on the Continent when hillforts are remarkable for their rarity. There is then a major resurgence of hillfort construction on the Continent in the last century BC, from which emerged the numerous great oppida of Gaul, examples of which are numerically rare in England. The Atlas authors therefore conclude that hillfort construction ‘evolved according to rather different chronological rhythms’ on the two sides of the Channel.
The best evidence for shared knowledge is the use of offset rows of upright stones to impede attackers, known as chevaux de frise in France (the term means ‘Friesland horses’ and dates from a much later era, when various structures were used to prevent a cavalry charge). These have an Atlantic fringe distribution, from the Iberian peninsula to northern Scotland, but in purely architectural and morphological terms, it is difficult to see any evidence for external influence on the character of British and Irish hillforts.
Instead, one might look for regional variation within Britain and Ireland, and in this respect the main difference seems to be in the density and distribution of hillforts in the landscape. In very broad terms, there is an east/west divide in England, along a ragged line that starts at the Hampshire/Sussex border on the south coast and takes in much of Wessex before running up the eastern edges of adjoining counties up to Worcestershire, Staffordshire, and Cheshire. West of the line, there are double the number of hillforts per 100 square miles than to the east. There is an east/west divide in Ireland too, but one that is not quite so marked. The counties with the greatest density of hillforts all lie in Scotland, Wales, or the western parts of England.
Examining this density pattern further, the Atlas authors look for patterns relating to the size of hillforts, with similar results, except that the larger the hillfort area the more counties there are that lack any examples. Very large hillforts, of 5ha or more, are absent from 43 counties in Britain and Ireland. The counties with the greatest density are all in Wales and the Marches and southern central England, led by Herefordshire and Wiltshire. Pretty much the same pattern emerges when the total area enclosed by hillforts in each county is calculated. Wales dominates, followed by Wessex and Cornwall in England, East Lothian and Flintshire, and Co. Wicklow and Co. Dublin.
Hierarchy and community
This pattern surely tells us something, but what? Perhaps the answer is related to the difficult question of the functions of hillforts, and while warning that it is impossible to generalise, the authors of the Atlas consider a range of possibilities. They make a distinction of those hillforts that are fully integrated into the landscape in the sense that they are surrounded by contemporary settlements and field systems. As one might expect, many of these are coastal and lowland hillforts. Barry Cunliffe’s Danebury Environs Project has influenced thinking here, by characterising Danebury hillfort as a defensible central place serving as a redistribution centre for the produce of surrounding farmsteads and for more exotic goods from further afield, occupied by a social elite including warriors.
Others have argued against such a hierarchical interpretation on the grounds that the effort to construct and maintain hillfort ramparts over a long period of time implied a community effort, and hence hillforts might have performed a social and religious role as a gathering place at certain times of the year, not unlike that envisaged for Neolithic causewayed camps. This model of community use emphasises the use of hillforts for feasting and gift exchange, by contrast with the Danebury trade and redistribution model. Highly visible, ramparts are a constant reminder in the landscape of the social relationships and common bonds between all the households dispersed across the hillfort hinterland, in a loosely flat society that lacks a dominant elite.
By contrast, there are hillforts occupying extreme coastal edge or high-altitude settings that appear, in the present state of knowledge, not to have adjacent settlements or farmsteads. Often exposed to harsh weather conditions, they seem to have a different social and economic relationship with their hinterland. Some have dwellings, but perhaps for sporadic occupation, and the possible uses of such sites range from seasonal trade to ritual activities and practices considered sacred (including industrial practices such as metalworking).
The difficulty of generalising is illustrated by an analysis of finds from three near-adjacent sites on the Oxfordshire Ridgeway. Alfred’s Castle, with its many pits and roundhouses, was occupied for several centuries. Uffington Castle and Segsbury Camp have only intermittent occupation over a long period: the first seems to have had a ceremonial function, and the latter to have been involved in sheep-trading. The three sites were thus used for different purposes, but possibly by the same people.
Though most of the dated hillforts included in the Atlas were constructed in the late Bronze Age and Iron Age, there are 69 cases where the dating evidence shows use in the early medieval period. Most are reoccupied or refortified after a period of disuse, the best-known examples being those at Cadbury Castle and Congresbury, both in Somerset; Tap o’ Noth in Aberdeenshire; Dinas Emrys in Caernarfonshire, and Dalkey Island in Co. Dublin – though 28 are entirely new creations (some using recycled Roman masonry), primarily in Scotland. In each case, the rampart timberwork is a continuation in the use of the same fundamental materials and architecture as was employed in the previous millennium.
Hillforts today
The Atlas authors end by considering the place of hillforts in the present day. Many survive as impressive monuments in the landscape, though 355 hillforts included in the Atlas – 10.6% of the total – are now so reduced that they are known only from cropmark evidence. Not all benefit from statutory protection: 961 sites (31.9%) lack scheduling and, whether scheduled or not, some are being eroded by development, farming, forestry, and quarrying. Their visibility on the landscape can be compromised – the people of Oswestry have been campaigning for a decade to prevent housing developments that would encroach on Old Oswestry Hillfort and obscure its setting and views, while the lower slopes of Pen Dinas hillfort, south of Aberystwyth, Ceredigion, are now encircled by housing estates and recently built blocks of flats. And encroachment by woodland, bracken, and scrub can make them less easy to see and understand, as well as threatening the buried archaeology.
Whether or not hillforts ever were the products of egalitarian society, there is no doubting that many of them are valued today as community resources – accessible by public footpath and heritage trails, enjoyed by visitors for the sweeping views and the natural heritage of bluebell woods and grassland wildflowers. Caerau Hillfort, Glamorgan, in the western suburbs of Cardiff, is the focus of a model community archaeology project involving community groups, local schools, and heritage professionals in the co-production of archaeological and historical research, and a similar project is now being set up to explore Pen Dinas.
The impetus for the latter project, a partnership between Dyfed Archaeological Trust and the Royal Commission on the Ancient and Historical Monuments of Wales, has come from the people who live in the shelter of the hillfort wanting to know more about the monument. A recent public meeting to initiate the project received overwhelming support, with several attendees commenting that they had never before been to a gathering without a single dissenting or uncertain voice. There is no doubt that hillforts have worked their way into our affections despite – or maybe because of – the many questions that remain to be answered.
Further reading
Gary Lock and Ian Ralston (2022) Atlas of the Hillforts of Britain and Ireland (Edinburgh University Press, ISBN 978-147447126, £150).
The online Atlas of Hillforts (https://hillforts.arch.ox.ac.uk) contains supplementary information and data on all the sites included in the project.
Toby Driver (2021) The Hillforts of Cardigan Bay: discovering the Iron Age communities of Ceredigion (Logaston Press, ISBN 978-1910839560, £12.95).
