Beacons in the Landscape: The Hillforts of England and Wales

By Ian Brown

Of all Britain's great archaeological monuments the Iron Age hillforts have arguably had the most profound impact on the landscape, if only because there are so many; yet we know very little about them. Were they recognised as being something special by those who created them or is the 'hillfort' purely an archaeologists' 'construct'? How were they constructed, who lived in them and to what uses were they put? This book, which is richly illustrated with photography of sites throughout England and Wales, addresses these and many other questions. After discussing the difficult issue of definition and the great excavations on which our knowledge is based, Ian Brown investigates in turn hillforts' origins, their architecture, and the role they played in Iron Age society. He also discusses the latest theories about their location, social significance and chronology. The book provides a valuable synthesis of the rich vein of research carried out in Britain on hillforts over the last thirty years. Hillforts' great variability poses many problems, and this book should help guide both the specialist and non-specialist alike though the complex literature. Furthermore, it has an important conservation objective. Land use in the modern era has not been kind to these monuments, with a significant number either disfigured or lost. Public consciousness of their importance needs raising if their management is to be improved and their future assured.

• Language: English
• Publisher: Oxbow Books
• Release date: Jul 20, 2009
• ISBN: 9781909686250

Ian Brown

Ian Brown’s police career started as a beat constable in Cumberland before transferring to the Metropolitan Police and completing his service as a Detective Superintendent in The British Virgin Islands where he headed a very successful drug squad.His involvement in high profile and regulation cases led to some thirty commendations from courts and law enforcement agencies around the world and earned him a Queen’s Commendation for Bravery.Now an ‘enrichment speaker’ for Cunard, P&O and other cruise lines, he regularly presents to packed houses. He is also an accomplished after dinner speaker.Ian has been engaged as a police advisor on several television crime programmes. He lives in Kent with his wife and his main interests are criminal research and sport as he is also a professional Lawn Tennis Association coach.

Book preview

Preface to second edition

Hillforts are one of the most common archaeological monuments but, despite more than a century of investigation, they are still one of the least understood. Were they recognised as something special by those who created them, or are they an archaeologists’ ‘construct’, defined to help us delve into the past? Their great variability poses many problems, but an inherent prominence in the landscape does suggest a deliberate act on the part of their creators, and something more than the many thousands of farmstead enclosures that probably housed the majority of the late Bronze Age and Iron Age populations.

This book will look at these and many other questions and develops and expands themes discussed in the first edition. It is intended, therefore, as an up-to-date textbook on hillforts in England and Wales, but now also includes those of the Isle of Man with their unique Viking heritage. Although often seen as creatures of the Iron Age, use and reuse of hillforts after the Roman period was significant and this is given prominence in Chapter 10.

The location, significance and function of hillforts are considered, and of special importance is placing them in the context of prehistoric and later societies and the environment in which those involved lived. The subject of hillforts is immensely wide, encompassing not only the structure of these gigantic monuments but also the social, political and economic make-up of the people who constructed them; their beliefs, customs and superstitions. Overshadowed as they have been in the national psyche by the more recognisable prehistoric stone henges and medieval castles, hillforts are of equal importance to the heritage of both England and Wales (Figure 1).

The book incorporates over one hundred photographs and line drawings to illustrate points, many in colour, with a comprehensive bibliography, references within the text for clarity and a substantial index for future reference. It is intended to be of interest to professionals, students and those just with an interest in the subject alike.

Land use in the modern era has not been kind to hillforts, with a significant number being either disfigured or lost, but abuse began in earlier times. One of the most important objectives of this book, therefore, is to ensure that, by explaining them to a wider audience, they will be better appreciated and their future conservation and management improved and assured.

FIGURE 1. Moel Arthur contour hillfort with the partial contour hillfort of Penycloddiau in the background, Clwydian Range, Flintshire.

Part 1

THE ‘ELUSIVE’ HILLFORT

CHAPTER ONE

Hillforts – an introduction

The context

Occurring, for the most part, in well-defined areas of England, Wales and the Isle of Man, hillforts remain one of the most common archaeological monuments. Despite being the subject of interest to antiquarian investigators since at least the seventeenth century, and major advances in archaeological thought ever since, they are still little understood. Perhaps this is because of their large size and difficulty in excavating as a result, or possibly because of their inherent diversity of form and probable function; there are indeed many questions to answer. It would be wrong to call them ‘settlements’ per se – they were much more than that and were no doubt considered as ‘something special’ by their creators. Their inherent variability and undoubted prominence in the landscape suggest a deliberate act, and something over and above a simple farm or enclosure that was, probably, the norm in the late Bronze Age, Iron Age and Romano-British landscape. They have been overshadowed in the public psyche by the great henges, the later medieval castles and historic houses, but hillforts are of equal importance to the national heritage. Regretfully, land use has not been kind to hillforts, and not many sites have escaped some form of disfigurement during the modern era, some major sites being destroyed completely.

The ‘elusive’ hillfort

Before we look at what a hillfort actually is, let us first look at a distribution pattern of what may be involved in England, Wales and the Isle of Man (Figure 2). They cover most of southern England, from Cornwall and Devon, then across Wessex to Sussex. Many are to be found in Somerset and along the River Severn estuary to Gloucestershire, extending up the Welsh Marches and western Midlands into Wales, where they occupy most of the areas outside of the high central Cambrian core. The Bath area and Oxfordshire and Warwickshire have notable examples. To the east, in Northamptonshire and Leicestershire, numbers decline, with a few on the edge of the Cambridgeshire Fens and in Norfolk and Suffolk. Numbers increase in Essex but drop again in Surrey and eastern Kent. To the north Midlands and north-west, Derbyshire and Cheshire have significant numbers, the latter along the sandstone ridge, becoming less frequent in Lancashire and the Lake District periphery. To the north-east, Yorkshire has few, if spectacular examples, but the many clusters of sites in Northumberland have, perhaps, more in common with those across the border in southern Scotland. In effect, apart from Northumberland, there is a distinct bias towards southern and western areas. Apart from a few notable exceptions the forts of the Isle of Man are mostly coastal. Certainly, the numbers of hillforts are daunting. Hogg, in his 1979 index of hillfort sites (pp. 1–5), listed 3,840 sites in Britain as being: ‘enclosures with substantial defences, usually on high ground and probably built between 1000 BC and AD 700, but showing no significant Roman influence’. His invaluable inventory included ‘related structures’ as well as ‘hillforts’, and it is implicit that an unspecified number ‘merely acted as homesteads’ or something similar.

FIGURE 2. Distribution of confirmed and possible hillforts in England, Wales and the Isle of Man.

PAULA LEVICK – ATLAS OF HILLFORTS IN BRITAIN AND IRELAND.

The Atlas of Hillforts in Britain and Ireland project, running from 2012 to 2017, is referred to in the Notes on page xv and is the latest attempt at determining the number and character of hillforts in the British Isles. The overall criteria used for hillfort definition in this study was of sites with potentially substantial defences in a prominent location with an area of not less than 0.2 ha, with two of these required as a base. Because of problems involved in definition at some sites, three degrees of certainty were used: ‘confirmed’, ‘unconfirmed’ and ‘irreconcilable issues’. That is not to say that inclusion in these two latter categories meant that such sites were not hillforts as such, but that certain problems of definition should be recognised. It is not intended to go into any detail on this (see Lock 2019 and Ralston 2019), but, for the purposes of this book, ‘confirmed’ and ‘possible’ categories will be used to cover these alternatives. Thus, with these varying degrees of certainty, in 2017 a total of 4,127 hillforts were recorded in England, Wales, Scotland, Northern Ireland, the Republic of Ireland and the Isle of Man, with 1,225 being located in England, 690 in Wales and 30 in the Isle of Man. Of this 1,945 total, 878 were ‘confirmed’ hillforts according to the Atlas criteria in England, 626 in Wales and 22 in the Isle of Man. Hillforts are still being found as our knowledge increases, and some others now proven to be just ‘figments of imagination’, but how a ‘hillfort’ is defined, as opposed to other forms of enclosure, has taxed archaeologists for years.

When looking at hillforts ‘in the round’, it appears just how immensely complex they are, and how difficult it is to define what a hillfort actually is. This may seem a basic question, but one that still remains to be answered. Any definition must encompass not only the physical nature of these vast monuments (their place in the landscape, topography, earthwork or stone wall, form etc), but also the social, political and economic make-up of the people involved, their beliefs, customs, superstitions and rituals, and the timescale and possible function of these sites.

In general terms, therefore, what were the Ages of the Hillfort? Is anything clear-cut? The answer is probably ‘no’, as the debate as to whether there was continuity or change at the end of the Bronze Age and into the Iron Age has continued for some time. The latter has tended to hold sway, no doubt to differing degrees about the country, but ‘Bronze Age’ and ‘Iron Age’ wrongly imply a rigid cut-off point in the use of bronze and start of the use of iron. This is not so, and iron was used for some items before the traditional end of the Ewart Park phase of the Bronze Age and the cessation of bronze hoarding at the end of the ninth century BC. Around 800 BC, however, there appear to have been changes in society manifested in different ceramic forms, metalwork, land use, settlement pattern and belief systems, including grain storage in pits (Haselgrove et al. 2001, 26–7; Haselgrove and Pope 2007, 6). Then, rather than technological advance, the gradual move towards the use of iron appears to have been spurred on by a lack of metal ore generally beginning about 100 BC and resulting in a reduced supply of bronze (Bradley 2007, 227).

Chronologies differ, but a subsequent transitional period (sometimes called the ‘earliest Iron Age’), which may have lasted for several hundred years, is thought to have merged into an ‘early Iron Age’ ending in about 400 BC. Around this time further changes in pottery, settlement pattern, social organisation and land use, again differing from area to area, heralded another transitionary period. The ‘middle Iron Age’ finally lasted to around 100 BC and a ‘late Iron Age’ followed on into the first century AD. By the time of the Roman Conquest of AD 43 various types of enclosure were already a feature of non-Mediterranean continental Europe and Ireland, with an estimated 60,000 now surviving, 20,000–30,000 of which might be considered as hillforts (Ralston 1995, 60; 2006, 16). Some achieved massive proportions, as at Manching in Bavaria, whilst others were very small indeed.

‘Enclosure’ is an ‘architectural’ form, but, because of the inherent diversity of enclosures generally, there are substantial problems where hillforts lie within this form and thereby their subsequent definition (Brown 2019; Lock 2019; Ralston 2019). What are we actually looking at? When an area is enclosed, whether it be by palisade, bank, ditch or wall, or indeed a thorn hedge (Ralston 2006), an area is defined away from the outside world, so providing a ‘defensible’ space in the interior, and giving to those in this interior, whether it be community or individual unit, a sense of security and ownership. It is being increasingly felt by archaeologists that ‘enclosure’ was a complex process having multiple connotations and serving different purposes at different times and places in later prehistory and early history (Harding 2004, 298). Such purposes could have been social, economic, political or ritual in nature, or a combination of all of these; all would have been very significant to the people concerned. The basis of hillfort design, therefore, was that, by enclosing a site, a special place was defined, distinct from the outside world; the surrounding banks possibly themselves imbued with, or imbued an interior with, symbolic powers.

It is inevitable, therefore, that there will be a ‘continuum’ of sites (Wigley 2007b); from a space defined by purely a small ditch, to one with large and multiple banks and ditches; but where along this continuum do we place ‘hillforts’ and which enclosures are actually hillforts and which are not? Cunliffe (2006, 154) puts it succinctly that ‘there may be no such thing as a typical hillfort’. Opinions differ between researchers.

There is also the small enclosures problem: as sites get smaller so the scope for confusion increases, Carver (1991, 4) pointing out for lowland Shropshire that the variety of enclosures ‘does not divide easily into hillfort and lowland farm’. In Pembrokeshire many univallate sites under 1 ha in area, commonly called ‘hillforts’, are manifestly not. Most are the ‘raths’ (small ‘defended’ settlements or farmsteads), common to the area and located in lowland situations on better soils. Similar enclosures are located in south-west England, and especially Cornwall where there are nearly 2,000 of these ‘rounds’. Although the best known of these small enclosures in Pembrokeshire is the excavated Iron Age Walesland Rath near Haverfordwest (Wainwright 1969; 1971), there are literally thousands of small defended enclosures in south-west Wales, some rectangular in outline, some round, many now just crop marks, and most showing few, if any, artefacts on excavation (Harold Mytum pers. comm. 2007). The small sites of north-west Wales, some less than 0.1 ha in extent, often perched on prominent craggy outcrops, and sites in the Lake District in similar positions, also pose problems of definition. Former surveys by the Royal Commission in the Llŷn peninsula of Caernarfonshire also suggest a variety of smaller sites. Yet many of the above are called ‘hillforts’ in common parlance.

Such is the situation in the Cheviot Hills, Oswald and Pearson (2005, 119–20) question whether Yeavering Bell may be the only ‘true’ hillfort in Northumberland. Many ‘hillforts’ there enclose less than 0.1–0.5 ha, but small defensive enclosures, such as those at Dod Law and others surrounding the Milfield Plain show hillfort characteristics in terms of both location and landscape prominence. Nevertheless, the complexity of these sites and the nature of the kinship or other groups that occupied them has been the subject of substantial project work.

‘Ringforts’ are another case in point and the Llawhaden group of enclosed settlements in Carmarthenshire includes examples (Williams and Mytum 1998). Typically, a weak defensive position has a strong univallate defensive bank and ditch, the interior containing up to half-a-dozen roundhouses in less than half a hectare, as at Dan y Coed and Woodside. Similar sites are found from Glamorgan to the upper Severn and into northern England. Again, ringforts have often been defined as hillforts in the past but are now better looked upon as ‘defended farmsteads’.

Inevitably small enclosures have very variable characteristics, and those of the central Marches, which Wigley (2007b) calls ‘non-hillfort’ or ‘small settlement’ enclosures, are well explained – some are on hilltops, some on hillslopes and others on the valley floor. High-status living is indicated at Collfryn in Montgomeryshire (Britnell 1989), Bryn Eryr on Anglesey (Longley 1998), and Castell Odo on the Llŷn peninsula in north-west Wales (Alcock 1960). On the other hand, the small stone-walled enclosure of Bryn y Castell, Ffestiniog, with cobbled yard and circular hut, was a much more frugal affair (Crew 1984; 1986). A parallel situation occurs in north-west England, where the difference between enclosed farmsteads of generally under 1 ha (the majority of sites) and small hillforts ‘becomes an issue of semantics’ (Harding 2004, 50). A wide variety of other examples are coming to light nationwide, as in the ‘rounds’ of Somerset.

Whatever the case, as suggested above, these small farmstead and settlement enclosures would have formed the most numerous features of the late Bronze and Iron Age landscapes, but most would have been very different from the dominating hillforts (we will look at hillfort size in Chapter 4), which are certainly a ‘step up’ in character, and have been interpreted by Cunliffe (1991, 312) as ‘representing a level of social organization above that of the farmstead or hamlet and may legitimately be considered as a separate phenomenon’. Undoubtedly, for some sites, things are more diffuse, one form merging into another, but size alone cannot be the overriding factor when defining hillforts. As a ‘rule of thumb’, hillforts can be recognised as being located in commanding, controlling and strategic positions in the landscape and exhibit substantial and often massive defences, but within this ‘definition’ are wide variations.

Considering the above, the question must be asked, when and out of what, did hillforts evolve. Dating evidence tends to be diffuse and can be unreliable, especially in earlier investigations. In Britain, hilltops appeared to exert a magnetic pull for enclosure during the Neolithic and this trend continued into the Bronze Age. In southern England a series of ‘hilltop enclosures’ appeared, and lengths of earthwork defined large areas of upland, possibly for communal gatherings or storage, with or without ritual intent, as exemplified by the earlier Neolithic causewayed enclosures. Harrow Hill in Sussex, for example (the site of extensive Neolithic flint mines (Curwen and Curwen 1926)), showed much animal bone when excavated, possibly serving as a central facility for the scattered farms at the base of the hill, perhaps for slaughtering. Rams Hill in Berkshire, Highdown Hill and Harting Beacon in Sussex, Norton Fitzwarren in Somerset and Bathampton Down in the Avon Valley near Bath could have performed similar functions, and these will all be returned to in more detail in Chapter 3 when we look at ‘hillfort beginnings’.

Other lengths of earthwork enclosed substantial areas of flat upland by cutting off spurs, these sometimes called ‘plateau enclosures’, as at Cold Kitchen Hill in Wiltshire and Butser Hill in Hampshire. In the Long Mynd of Shropshire, some sites are associated with the extensive cross-dyke systems of the central Marches, and were possibly similar pastoral enclosures used for gathering, shearing, culling and general husbandry, possibly using portable hurdles (Guilbert 1976).

These earlier hilltop and plateau enclosures dating to the late Bronze Age and early Iron Age, together with the earlier still Neolithic ‘tor enclosures’ of the south-west, are better seen as an earlier type of enclosure. Some developed into later hillforts, as at Rams Hill and Highdown Hill, for example, whereas others, as at Harting Beacon, did not. Therefore, to define an enclosure, just because it is located on a hilltop and of some size, as a ‘hillfort’ per se, is erroneous and ignores the complexities involved in hillfort definition.

The hillforts seen dominating the landscape today, as at the rocky Carn Ingli in Pembrokeshire (Figure 3), are characterised by enclosing ‘defences’ (‘ramparts’), of bank or stone wall, with or without a ditch, which can occur singly, the so-called ‘univallate’ sites, or in greater multiples, the ‘multivallate’ sites, double defences sometimes being called ‘bivallate’. The simpler univallate type does not necessarily signify an earlier design; multivallation was a feature of some early hillforts, such as Danebury in Hampshire and Rainsborough in Northamptonshire. However, to categorise hillforts by these rampart sequences is to ignore their complexity, as both single and multiple sections can be found at the same location depending on topography and need, and, of course, many sites went through periods of alteration as time progressed, some of which were very extensive indeed. But these terms have proved to be a useful and convenient, if not entirely accurate, means of description. Whatever is the case, boundaries define, surround and enclose an interior which is accessed by, usually, one, two or three entrances, but sometimes more. Often, but by no means exclusively, sites are located on a dominant hilltop or significant geomorphological feature in the landscape. Today, grassy humps and hollows, rocky outcrops and fallen stone tend to be the only visible remains, but some hillforts still form very impressive features indeed, with deep ditches and high banks – up to four in number at Maiden Castle near Dorchester in Dorset and at Castle-an-Dinas, St Columb Major in Cornwall, and five to seven at Old Oswestry in Shropshire (Figure 4).

FIGURE 3. The prominent and rocky Carn Ingli hillfort above Newport and the Pembrokeshire coastal plain. Walled enclosures are located downslope of, and outside, the main enclosure.

Looking at the above, it appears that hillfort location and place in the landscape is extremely diverse. Categories tend to be devised to take account of most general situations, ‘hilltop’, ‘valley’ etc, as well as coastal situations, and aspect determined as well as land use. Some sites are not located on ‘hills’ at all – Ebury Hill Camp in Shropshire, for example, sits on a low-lying outcrop of hard andesitic tuff at only 90 m OD (Stanford 1985a, 9) and Old Oswestry appears to rise only marginally from the surrounding land, perched as it is astride a glacial outcrop. Lower still, are the marsh forts of Shropshire, Cheshire, Cambridgeshire and Yorkshire, originally surrounded by marsh, fen and carr. The term ‘hillfort’, therefore, does not describe the true situation well and has been used indiscriminately for a wide variety of sites surrounded and enclosed by these banks, walls and ditches, some on hills, some not, some large, some small, with little regard to their function or place in prehistoric or later society.

FIGURE 4. The multiple banks and ditches of Old Oswestry in Shropshire. The impressive western entrance with its flanking hollows, in the foreground, contrasts with the lesser more sinuous entrance on the east.

© CLWYD-POWYS ARCHAEOLOGICAL TRUST 95-C-1041.

As we will see in the next chapter, as early as the sixteenth century these enigmatic enclosures had become mixed up with folklore and Arthurian legend, some being considered mystical ‘druid’s circles’ or the realm of fairies. Later, into the seventeenth century, they gradually became seen as something more ‘of this world’ and certainly of ‘great antiquity’, and as the mid nineteenth century progressed, more scientific explanations emerged. The banks were then viewed as pure defensive mechanisms, and the term ‘rampart’, although not ideal, was coined to imply an element of fortification, refuge and defence to the enclosures, which became known as ‘encampments’, ‘camps’ or ‘temporary-camps’. It is indeed tempting to regard an enclosing obstacle, whether it be a crude wall or palisade or a more elaborate bank and ditch, as a simplistic defensive device, either to keep those with malevolent intent out, whether they be people or wild animals, or valuable stock in, and it is this interpretation that has been in the past the unquestioned basis of hillfort study. The early most influential investigators happened to be military men, such as Lt General Pitt Rivers (see page 23) and later Sir Mortimer Wheeler, and it was only natural that they would use military terms. This interpretation has until fairly recently been the unquestioned basis of hillfort study, as reflected in the work of these and other notable researchers and reinforced by Christopher Hawkes’ classic paper of 1931. For example, A.H.A. Hogg (1975) proposed the following definition: ‘A hillfort may be described as an enclosure, apparently fortified, and so placed as to gain some defensive advantage from its position; as being a type generally built during the native Iron Age; and normally not less than 0.2 ha in enclosed area.’ Michael Avery (1976, 3) had a different interpretation, defining a ‘fort’ as ‘a site fortified so as to be defensible against human attack’ and above a quarter of an acre in area; below that size, enclosures were to be called ‘fortlets’. Not all ‘forts’ were also ‘hillforts’ and he defined the latter as ‘a fort which not only has artificial man-made defences raised to protect it, but also exploits the natural terrain so as to give defenders an advantage of height over those approaching the site’. He later proposed that ‘all hillforts are connected by the fact that each was in one way or another used for defence of the community which built it’ (1993, 1). ‘Hillfort’ therefore has become the general descriptive term in use today and is unlikely to change. The defensive argument will be returned to in later chapters.

Hillfort classifications

‘Hillfort’ definitions, therefore, tend to be regarded as ‘all things to all people’ and to perform any sort of classification tends to end up the creature of the project concerned. Clearly, Dennis Harding (2004) was correct when he wrote that ‘the association of sites, large and small, simple or complex, into a single category of hillforts has undoubtedly been an oversimplification of archaeological classification’.

There have been various attempts, predominantly based on topographical location, at classifying types of hillfort. Some of these have been successful and are still in use, some not. James Dyer (1981, 6), for example, suggested five main groups: contour forts; promontory forts; plateau forts; valley forts; and multiple-enclosure or hillslope forts.

The Atlas of Hillforts in Britain and Ireland has defined seven hillfort types as a basis for classification, as follows: contour forts; partial contour forts; promontory forts; hillslope forts; level terrain forts; marsh forts and multiple-enclosure forts. These categories are subdivided into their topographical positions: hilltop; coastal promontory; inland promontory; valley bottom; knoll/hillock; outcrop; ridge; plateau/cliff edge; hillslope; lowland; and finally, other location. Virtually all combinations are therefore accounted for.

Contour and partial contour forts account for the most categories, at approaching three-quarters of total hillforts in the Atlas confirmed category in England, Wales and the Isle of Man combined. England and Wales have around 60 per cent of their hillforts so defined and the Isle of Man about 14 per cent, reflecting the coastal promontory status of most forts on the island. A contour fort will follow the contours almost exactly, whilst the latter takes into account the common occurrence of sites lying across the contours. Fine examples of contour hillforts are Caesar’s Camp at Easthampstead in Berkshire, Borough Hill in Northamptonshire, British Camp (Herefordshire Beacon) on the high ridge of the Malvern Hills in Herefordshire and Hambledon Hill in Dorset. However, the smaller Moel Arthur fort in the Clwydian Range of Denbighshire and Flintshire shows this particularly well (see Figure 1), and the partial contour sites of Moel Fenlli and Penycloddiau are nearby. At partial contour hillforts the apparent relationship between the contours masks a considerable variation in altitudinal range from one end of the fort to the other. This phenomenon appears to ‘tip’ the site towards a particular aspect, in the case of Earl’s Hill in Shropshire (Figure 5) to the south-west and at Ravensburgh Castle in Hertfordshire to the south.

Another major category, the promontory forts, as the name implies are located where the land forms a defined promontory, either inland or on the coast. Notable well-known examples inland, of which there are many, are Crickley Hill (Figure 6) and Bredon Hill Camp (also called Kemerton Camp) on the Gloucestershire/Worcestershire border, Sharpenhoe Clappers in Bedfordshire and Bryn Alun near Wrexham in Denbighshire located on a sharp bend of the River Alun. Coastal promontory forts (called ‘cliff castles’ in Cornwall) extend from around the Thames Estuary, south and west to Pembrokeshire (Figure 7) and the Gower and north to Anglesey, with fine small examples on the Isle of Man and the Isles of Scilly. Of special interest are the small inland promontory forts of Pembrokeshire, which complement their many neighbours on the coast.

Hillslope forts take a wide variety of forms and their character will be outlined in succeeding chapters, but an excellent example is Caer Drewyn located above the River Dee at Corwen in Merioneth in north Wales. Here the outstanding stone and earthen/rubble ramparts are clearly visible from below and dominate the valley (Figure 8). Level terrain forts on the other hand are situated on flat ground with no natural physiographic advantage, as at Rainsborough in Northamptonshire, and can cover a wide variety of different topographical situations, as can lower-lying valley forts, such as Risbury in Herefordshire. Marsh forts, as their name implies, are located on low-lying flat land formerly protected by marsh, fen, carr or mere, some probably accessed by causeways; Stonea Camp, in Cambridgeshire and Warham Camp, east of the Wash in Norfolk on the 15 m contour, for example. Wall Camp (Kynnersley) in Shropshire is located on a low sandstone and boulder-clay outcrop surrounded by peat, and The Berth in Shropshire and Sutton Common in Yorkshire are also notable; at The Berth the banks possibly acting as a dam to create a lake (Morris and Gelling 1991, 58).

Widely spaced ramparts are a feature of some sites, as at Warbstow Bury in Cornwall where the two circuits of rampart are not concentric to each other and may represent different phases of construction, but whether this site can be classified as a true multiple-enclosure fort is a moot point. Generally, however, such forts are indeed found predominantly in and towards the south-west of England and also in south Wales: Milber Down Camp in Devon, Scratchbury in Wiltshire and sites on Harding’s Down on the Gower of Glamorgan, for example. They do not occupy prominent topographical positions and characteristically have concentric defences, possibly with stock rearing in mind. Milber Down Camp is a fine example. Located between the Teign Estuary, Aller Brook and Babbacombe Bay at only 90–120 m OD it has three concentric enclosures each with bank and ditch. The inner measures only 116 m by 96 m, with the second and third enclosures just narrow strips. There was a much larger outer enclosure with low bank and ditch. A sunken hollow-way curves upslope to the entrance of the third enclosure and could have been used for stock watering at the Aller Brook at its base. The possible function of multiple-enclosure forts will be returned to later.

FIGURE 5. The partial contour hillfort of Earl’s Hill in Shropshire, with its significant ‘annexe’, appears tilted across the contours.

FIGURE 6. The ramparts of Crickley Hill inland promontory fort, Gloucestershire.

FIGURE 7. The Clawdd y Milwyr coastal promontory fort on St David’s Head in Pembrokeshire. The promontory, jutting out into the Irish Sea, is cut off by a massive dry-stone rampart, now about 100 m long, with two lesser stone walls outside and possibly shallow ditches between. Fine hut circles are still clearly visible.

FIGURE 8. Intact stone facework and stepped rampart behind extensive collapsed loose stone and walls extending downslope at Caer Drewyn, above Corwen and the River Dee in Merioneth. The hillfort appears deliberately sited to be seen from the Dee Valley below.

It follows that, with this great variability in location and type, not only is classification difficult and invariably unsatisfactory, but, as we have seen, the use of a single term is undoubtedly erroneous. As a result, it is a salutary fact that the old definitions of ‘camp’, ‘British camp’ or ‘encampment’, despite their Baden-Powellish connotations and intimations of a short-lived nature, are nearer the mark. However, for a variety of reasons, mostly non-academic, and of which convenience is one, ‘hillfort’, ‘rampart’ and ‘defences’ will be used in this book.

In later chapters the reasons for choice of location will be considered, but several basic points are pertinent here. Firstly, many hillforts are sited in significant and ‘commanding’ positions in the landscape, rather than being located at random (Brown 2002; 2004; Corney and Payne 2006a). As a result, severe exposure to the elements is common, as shown, for example, by Yeavering Bell in Northumberland, Mam Tor in Derbyshire, Titterstone Clee in Shropshire and South Barrule on the Isle of Man. Some sites are at a very high altitude indeed, as on the summit of Ingleborough in the Yorkshire Dales, where climatic conditions are extreme. All of this suggests that a great deal of strategic thinking went into their building in the first place. Secondly, there is considerable evidence that, at a large number of hillforts, there was a long-standing ‘interest’ in the location which pre-dated hillfort construction. Some sites saw activity as far back as the Mesolithic (see Chapter 3). Thirdly, although we have to be careful in the interpretation of the exceptional or unusual as evidence of some form of ‘ritual’, and especially so if we wish to use the term ‘religion’ (Dark 2002, 148), a significant feature of the limited number of extensive excavations undertaken has been evidence of widespread symbolic observance. The finding of human remains does not necessarily mean ritual sacrifice, of course – the person might just have been murdered and disposed of – but the context of human and animal bone in pits, four-poster structures, rampart burials and, in some cases, shrines, all suggest that propitiation was an essential part of hillfort life – an important consideration indeed, about which more will be said.

The dating of hillforts can also be problematical, and, purely as an example, in East Anglia in particular there are questions over a number of sites, notably Tasburgh in Norfolk and Clare Camp in Suffolk. The former was occupied in the middle and late Saxon period and has parallels with enclosures of tenth-century date at Witham and Maldon in Essex; but whether the banks were new at this time or modified from an earlier design is open to question (Davies et al. 1991, 71–2). In Wales, Bryn Euryn at Rhos-on-Sea in Denbighshire could also be earlier than its Dark Age credentials suggest.

Beyond doubt is that hillforts involved massive feats of civil engineering on the part of late Bronze Age and Iron Age communities and that the parameters of structure, location and timescale –some 1,000-plus years of development – are wide. A large number of true hillforts have been examined to some degree, but only at a comparative few has investigation been adequate. At Danebury around 57 per cent of the site was excavated and at Crickley Hill in Gloucestershire and Castell Henllys in Pembrokeshire over half of their interiors likewise. But at the 18.5 ha Maiden Castle the excavations of Sir Mortimer Wheeler (Wheeler 1943) and Niall Sharples (Sharples 1991a and b) covered less than 2 per cent of the hillfort’s area between them. Hard evidence is often lacking and much excavation, some of questionable value, has centred on the sectioning of ramparts, which has also contributed to the emphasis on the defensive nature of sites. Because of this, and as it is not possible to ‘get into the mind’ of prehistoric people, it is probable that the role of hillforts within society will never be truly understood.

It is inevitable that the very complexity and variety of hillforts has given rise to numerous models of function within specific regions (Collis 1981); one drawn up for one area may not be applicable to others. Indeed, it is most unlikely that all hillforts functioned in exactly the same way, either on a temporal or spatial basis, and evidence has to be considered extremely carefully.

As subsequent chapters will show, it is clear that hillforts were manifestly a social, economic and political phenomenon within which superstition and belief played an important part. Much work remains to be done on the definition of terms and, as further information on small settlements comes to light, the relationship between these and the larger sites may become clearer. It could be that there is no such thing as a ‘hillfort’ as such, but just a wide variety of different types of enclosure yet to be fully investigated and understood. The above problems will be returned to later, but first, to put things into further context, some potted history.

CHAPTER TWO

From antiquarian to modern

Interest in the hillforts of England and Wales and the Isle of Man has a long history. From around the sixteenth to nineteenth centuries all young ‘gentlemen’ were expected to have a classical education and appreciate the aesthetics of architecture and the natural sciences of geology and botany, which interacted to form a ‘romantic’ and ‘picturesque’ landscape. The classical landscapes in the seventeenth-century paintings of Claude Lorrain and Nicolas and Gaspard Poussin, and later in nineteenth-century England those of Constable and Turner, together with the writings of John Ruskin and poetry of Wordsworth and John Manley Hopkins, all contributed to this concentration on beauty and the effete. The continental ‘Grand Tour’ (in particular to the Mediterranean regions of Italy and Greece) was deemed essential to this progression.

In tandem with, and indeed spurred on by, the above, from around the beginning of the sixteenth century, in addition to the antiquities of Greece and Rome there appeared an embryonic interest in the antiquities of the British Isles; now generally referred to as ‘antiquarianism’, and descriptions of monuments and their landscape and regional settings began to appear. From this point, aristocratic gentlemen and ministers of the cloth tended to dominate the scene, and, gradually, about the mid nineteenth century and into the early twentieth century, tentative survey and excavation began on an increasingly rigorous basis, Glyn Daniel saying that archaeology [prehistoric] was the ‘creation of the Victorians’ between 1840 and 1900 (Daniel 1975; Chamberlain 2006).

Since then, a number of excavations, often, but not entirely, in the central southern zone of Cunliffe (1991, 344), which extends from Wessex and the South Downs through the Cotswolds and Welsh Marches to Wales outside the upland core, have shed light not only on hillfort structure and design but also on origin, function and their importance to both society and economy. Many of these investigations have just sectioned ramparts and opened trenches on a small scale, the large size of even the smallest hillfort making adequate excavation either very difficult or nearly impossible. Today, research is focused more on their status in the landscape and the ‘environs’ in which they appear to have assumed such a dominant role, and less on invasive techniques, with geophysics playing an important part. Let us begin at the beginning…

Early recordings and antiquarian study 1500–1800

Early interest in the ‘ancient camps’, as hillforts were often referred to, was often mixed up with folklore. There are a confusing number of hillforts with the name ‘Cadbury’ in England, but the excavated Cadbury Castle near South Cadbury in Somerset (called South Cadbury throughout this book), is the most famous and associated with Arthurian legend. John Leland (c. 1503–1552), considered by some to be the father of local history and bibliography, wrote: ‘At South Cadbyri standith Camallate [Camelot], sumtyme a famous toun or castelle.’ The site was also reputed to be the site where ‘the fairies stored grain grown on the surrounding plain’ (Alexander 2002, 41–2). This reference to grain storage through folklore is interesting when the possible function of the many pits found within hillfort interiors is considered, as we shall see in later chapters.

Between c. 1536 and 1542, Leland, then chaplain and librarian to Henry VIII and the first holder of the post of ‘King’s Antiquary’ in 1533, embarked on a tour of England and Wales visiting a wide variety of sites. At Badbury Rings, near Wimborne Minster in Dorset, for example, he observed that: ‘The Saxon Kinges had hard by the Toune a castelle now called Badbyri … The Ditches, Hilles and Site ther of be yet evidently seene.’

Following on from the work of Leland, the other great Elizabethan antiquary William Lambarde (1536–1601) and the Welsh cartographer and antiquary Humphrey Llwyd (1527–1568) who described the antiquities of Anglesey, it was the topographical and historical work ‘Britannia’ of William Camden (1551–1623) that was one of the most important works of its kind of the Renaissance. Headmaster of Westminster School and later Clarenceux King of Arms, his county-by-county description of Great Britain and Ireland covered antiquity in detail. The first edition of Britannia was published in Latin in 1586, proving so popular there were five further and enlarged editions up to 1607. The first English language edition was by Philemon Holland with Camden’s consultation in 1610, with a new edition by Holland’s son Henry in 1637. These translations became the standard works on British archaeology, going through many revisions; by Bishop Gibson in 1695, 1722, 1753 and 1772, and by Richard Gough in 1789 and 1806. Describing the features of Dorchester in Dorset, ‘the head town of the whole shire’, he described the hillforts in its vicinity: ‘The Danes, who raised as it is thought certaine trenches, whereof one is called Maumbury, being an acre inditched, and other Poundbury, somewhat greater, and the third a mile off is a camp with five trenches containing some ten acres, called Maiden-castle which a man may easily conjecture to have been a summer station or campe of the Romanes.’ The later inclusion of maps and plans added substantially to knowledge, as that of Maiden Castle by Francis Cary in the 1789 edition.

John Aubrey (1626–1697), philosopher and antiquary, is sometimes referred to as a pioneer archaeologist. His time at Trinity College, Oxford saw his interest in antiquities blossom, and he set about recording many megalithic and other field monuments in southern England, including Avebury, which he thought of as druid temples. Although publishing little, his manuscripts are held in the Bodleian Library; Volume II of his four volume ‘Monumenta Britannica’, the ‘Chorographia Antiquaria’ comprised surveys of monuments including hillforts, called ‘camps’ in the text.

Edward Lhuyd, or Llwyd, (1660–1709), naturalist, geologist, botanist, linguist and antiquarian among other specialities, is also important to our story. Finally becoming Keeper of the Ashmolean Museum in Oxford until his death, after 1687 he visited every county in Wales, then to Scotland, Ireland, Cornwall, Brittany and the Isle of Man recording features of interest. A friend of the Mostyns of Penbedw, Nannerch in Flintshire, in 1693–94 he described Moel Arthur hillfort in the Clwydian Range thus: ‘There is one fair one [camp] above my house [Penbedw] that looks over the Vale of Clwyd, with a treble trench of one side and a single one to the precipice. ’Tis call’d Moel Arthur. I suppose these were to secure cattle, etc. upon sudden inroads of enemies till the storm was over.’ Referring to Moel Fenlli, also in the Clwyds, he certainly appeared conscious of the possible defensive connotations of the ramparts: ‘Among them [the Clwyds], the highest is called Moel Enlhi: at the top whereof I observ’d a military fence or rampire, and a very clear Spring.’ Although not on the highest part of the Range, as is Moel Famau, Moel Fenlli, with its notable hillfort, can seem such from below. The spring is still there. Later, he was the first to describe the hillfort on Strumble Head in Pembrokeshire in Volume I of his ‘Archaeologia Britannica’, published shortly before his death in 1707. In calling the languages of Ireland, Wales, Cornwall, Brittany and Gaul ‘Celtic’ he started the Celtic ball rolling (see page 109) and sought to distinguish the inhabitants of this Atlantic zone from the ‘Britons’ of England. This ‘sparked off the wave of Celtomania’ that, to an extent, continues today (Cunliffe 2009, 57).

Whilst the rigour of rural life was all too apparent to those who had to endure it, to the urban effete the rural scene was all to do with the ‘picturesque’ and the ‘sublime’, and during the late eighteenth and into the nineteenth centuries ideas about the landscape and its place in the scheme of things began to take shape, and the ubiquitous hillforts were very much a part of this. The Reverend William Gilpin (1724–1804) was, with Richard Payne Knight and Uvedale Price, one of the principal theorists of the picturesque. He undertook nine tours of England, Wales and Scotland between 1768 and 1776, in imitation of the ‘Grand Tours’ of the antiquities and beauties of the Continent, looking at the natural landscape in terms of the created landscape of William Kent and Capability Brown. Throughout most of the eighteenth century, mountains were considered ‘unhospital’ places to the traveller, Gilpin in 1772 writing of ‘the mountains half-obscured by driving vapours; and mingling with the sky in awful obscurity’. Indeed most of all, the tours of the seventeenth and eighteenth centuries were undertaken by the worthy and curious for the purpose of naturalist and antiquarian enquiry, but the picturesque theories therein later became questioned by John Ruskin who saw the terrible conditions that the rural poor suffered at the hands of the ‘sublime’.

However, it was William Stukeley (1687–1765) who became the ‘central figure of early eighteenth century archaeology’ (Piggott 1985, 13) and forerunner in methodically measuring and documenting sites. He recognised the principle of stratigraphy and engaged in the excavation of sites and surrounds, becoming the first Secretary of the Society of Antiquaries of London, founded at the Bear Tavern in the Strand in 1707, and formalised in 1717. Between 1710–25 he went on tours of the countryside looking for archaeological sites and monuments and general features of interest to him, and like John Aubrey before him, seeing Avebury and Stonehenge as druid temples. The drawing by Stukeley of Ravensburgh Castle hillfort, on the Hertfordshire/Bedfordshire border, also possibly visited by Leland, is shown in Figure 9, and, although the site is now densely wooded, it appears remarkably accurate.

In the same vein, maps of the period began to show hillforts. Poston Camp in Herefordshire was recorded on Isaac Taylor’s map of 1754. The perimeter of Yeavering Bell in Northumberland was depicted on John Warburton’s map of 1716, and on a plan and view drawn by William Hutchinson in 1776, the whole being ‘of very remote antiquity’ (Oswald and Pearson 2005, 100–1). In 1791 the first account of Worlebury, above Weston-super-Mare, was published in Collinson’s ‘History of Somerset’, and in 1805 a plan of the site was produced by the artist George Cumberland.