Until very recently, the Iron Age was described as a time when Britain was subject to a succession of invasions, by Celts from mainland Europe.[*] These invasionist theories, however, and the notion that Britain was absorbed into a pan-European ‘Celtic civilisation’, have been found wanting:

There was no cross-European Celtic people. There was no broad-based Celtic art, society or religion. And there were never any Celts in Britain.

John Collis (‘Celtic Fantasy’, British Archaeological News 11, March 1994)

… our views of the Iron Age have seen radical change and controversy in a flourish of new research over the last ten years… No one is denying that people in Iron Age Britain spoke Celtic languages, or shared certain common cultural traditions with their contemporaries in mainland Europe, such as the use of La Tène ‘art’.[*] What has been shown to be untrue, however, is that there existed a single Celtic race whose members all had the same religion, psychological traits, and type of society, and who recognised themselves as ‘Celts’… While there were contacts, and shared cultural elements across Europe, it is the differences in all aspects of life between neighbouring areas that seem to have been more important than the similarities.

J.D. Hill (‘Weaving the strands of a new Iron Age’, British Archaeology 17, September 1996)

Archaeologists widely agree on two things about the British Iron Age: its many regional cultures grew out of the preceding local Bronze Age, and did not derive from waves of continental ‘Celtic’ invaders. And secondly, calling the British Iron Age ‘Celtic’ is so misleading that it is best abandoned.

Simon James (‘Peoples of Britain’, BBC History, 1998)

Iron making know-how was probably introduced into Britain, not by a hoard of invading Celtic warriors, but through normal, non-hostile, contacts with the Continent. Although iron ores were far more common (hence cheaper) than those of copper and, particularly, tin, wrought iron wasn’t necessarily better than bronze, and turning ore into usable object was a considerably more laborious procedure.[*] There is evidence to suggest that ironworking was taking place in Britain (at Hartshill Quarry, Berkshire) as early as the 10th century BC.[*] The first extant iron objects, however, are from a couple of centuries later – such as those from the, so called, Llyn Fawr Hoard.[*] As a consequence, the phrase ‘Llyn Fawr phase’ is sometimes used to describe the period, c.800 BC–c.600 BC, of the Bronze Age to Iron Age transition.[*] Iron did not completely displace bronze of course. Bronze production continued, though on a much smaller scale, being used for vessels, horse-harness fittings and decorative items.

Hillforts are defended enclosures, usually, as the name suggests, built on a hill-top. There are more than three thousand sites in Britain that fall into the broad category of hillfort. Most are found in central, southern and western England, Wales and south-eastern Scotland. They come in diverse shapes and a vast range of sizes. Some have just one circuit of defences (univallate), whilst others have more (multivallate). At most sites, the circuits comprise a rampart and a ditch (being defensive structures, the ditch is outside the rampart), but at some they are formed of ramparts only. The ramparts themselves were constructed using various methods and materials.

iron01

A Hillfort’s rampart could be built from material dug from a ditch on its outward side, from a quarry within the enclosure, or, sometimes, brought in from the surrounding area. Types of rampart construction include:

Dump.   An unsupported earth or rubble bank.

Timber revetment.   The earth or rubble core of the rampart is retained, either on just its outwards facing side or on both sides, by closely spaced timber uprights. Where the core is supported on both sides, this is often referred to as a ‘box rampart’

Stone revetment.   In this case, instead of timber, stone blocks, or drystone courses of smaller stones, retain the core.

Timber lacing.   Stone revetment, on both faces of the core, is strengthened by large numbers of transverse timber ties. In Scotland there are over sixty hillforts where the stonework of the ramparts has been fused by intense heat, created by burning of the timber lacing. These are termed ‘vitrified forts’, and it is the subject of debate whether this effect was intentional, accidental, or the result of hostilities.

Glacis.   Wood is not only subject to rot, but it is also susceptible to being set on fire. So in areas where it was expedient to build timber revetments, maintenance would have been an ongoing, difficult and laborious operation. Possibly, it was in answer to this problem that glacis ramparts appeared in the 4th century BC. In other forms of construction, there is a narrow berm between the rampart and the ditch, but glacis ramparts are constructed from dumps of earth and rubble positioned so that the outer face of the bank forms a continuous slope with the inner face of the ditch, at an angle of 30–45 degrees. They would not have been the grassy mounds seen today. Any assailant would have faced a very large barrier (at Maiden Castle, Dorset, for example, it was 25 metres from ditch bottom to rampart top) of slippery scree, surmounted by a breastwork.

Some sites categorised as hillforts are actually built on relatively flat ground. They are sometimes described as ‘plateau forts’, and, getting no advantage from natural features, rely entirely on man-made defences.[*] The entrances (there are usually one or two, which would have had timber gates) are the weak points in a hillfort’s defences, so much ingenuity and energy was often devoted to their design and construction. The entrance earthworks can be complex, and might incorporate long passages, bastions, barbicans, guardrooms, overlapping ramparts, hornworks and outworks.[*]

All this variety and complexity is, however, the product of some eight centuries of hillfort evolution. Monuments recognized as hillforts emerge towards the end of the Bronze Age, c.900 BC, but it was not for another century or two that the main phase of building began.[*] Generally, early hillforts were relatively simple, univallate, affairs. Around 400 BC many of them were abandoned. A few, however, not only continued in use, but received elaborate enhancements, such as multivallate defences and labyrinthine entrance passages. A possible explanation is that the abandoned hillforts belonged to tribal groups whose lands were annexed by stronger neighbours. These dominant groups developed their own hillforts as a demonstration of their strength. Indeed, like a modern corporate skyscraper, ‘developed hillforts’ appear designed as much for their impressive appearance, as for their actual purpose.

iron02

Julius Caesar campaigned in Gaul during the years 58–51 BC. He, himself, says (The Gallic War II, 6) that the usual method of attacking a stronghold employed by the Gauls and the Belgae (two of the three peoples who Caesar identifies as the inhabitants of Gaul[*]) was to surround it with men, and bombard the ramparts with stones to clear them of defenders, so that the attackers could get at the gates – which they would approach in a testudo formation, i.e. locking their shields together to provide protection, all around and overhead, like a tortoise shell. Presumably it was the same in Britain – hence the development of hillfort defences designed to counter that tactic. The greater depth of multivallate defences would impair the efficacy of attacking slingers. Constricted and convoluted entrance passageways meant that a party of men approaching the gate would have longer exposure to a barrage of stones, rained from the flanking ramparts. At the eastern entrance of Maiden Castle, Dorset, archaeologists discovered a pit containing over 22,000 pebbles – handily placed to be hurled onto any attacker.[*]

At Danebury, Hampshire, thousands of pebbles have also been found in the vicinity of the entrance – the largest single cache being a pit containing some 11,000.[*] Nevertheless, c.100 BC, the gate was destroyed by fire. It was never rebuilt, and intensive occupation of the hillfort ceased. Further, the bodies of some twenty-five men, women and children had been unceremoniously thrown into two pits near to the entrance – the open graves being left to silt-up naturally. It doesn’t seem unreasonable to conclude that Danebury had fallen to an attack.

Incidentally, since the great majority of hillforts contain no source of water, it would appear that prolonged sieges were not an expected military tactic.

EAST

WEST

Maiden Castle, near Dorchester, in Dorset, is the largest developed hillfort in Britain. In the Neolithic, there had been a ‘causewayed enclosure’ at the eastern end (bottom-left in the aerial photo) of the site, and then a large ‘bank barrow’ (about 545 metres by 19 metres) had been erected, running east–west across the centre of the hill.[*] Around 600 BC, more or less on top of the old causewayed enclosure, a univallate, ditch and rampart (utilising dump and timber-revetted box rampart techniques), hillfort was built.[*] Between c.400 BC and c.100 BC, the hillfort underwent several phases of development, to produce the massive glacis ramparts and complex entrances which are still impressive today. The enclosed area of the developed hillfort, over 17 hectares, is almost triple that of the original fortification, and the overall size of the site is in excess of 45 hectares. Sir Mortimer Wheeler excavated at Maiden Castle in the mid-1930s. Based on his findings, Sir Mortimer produced a famously dramatic description of Maiden Castle’s fall to the Romans, led by, the future emperor, Vespasian, in their westward advance after the invasion of AD 43. Further excavation, by Niall Sharples, took place in the mid-1980s, and a new appraisal of the evidence suggested that, by the time the Romans arrived, the hillfort had already outlived its usefulness – its defences were neglected and its population was much-reduced.[*] What seems clear, however, is that Maiden Castle was finally abandoned within a few decades of the Roman conquest. A small Romano-British temple was built within the interior in the late-4th century AD.[*]

In total contrast to Maiden Castle is Castell Henllys, in Pembrokeshire, which only encloses about ½ hectare.[*] Maiden Castle is classed as a ‘contour fort’ – its defensive circuits follow the contours of the hill on which it is superimposed. Castell Henllys, however, is a ‘promontory fort’. A promontory site provides a hillfort with natural defences for much of its perimeter, requiring extensive man-made defences only across the neck of land joining it to the adjacent area. (If the promontory in question juts out into the sea, then the fort is often called a ‘cliff castle’). Castell Henllys is built on a small spur overlooking the River Gwaun. There are precipices to the east, south and west, so most of its man-made defences are directed towards the north. These defences comprise a large inner bank, a ditch, a smaller outer bank and another ditch. The earthworks curve round to meet the precipice at the east, and at the west they merge at the place where the entrance was. Thirty metres north of the fort is an outer bank and ditch. Beneath the bank, archaeologists discovered a band of upright stones. The rows of stones, which are called a ‘chevaux-de-frise’, and are rare in Britain, would have impeded a mass attack, and provided a defence against chariots. Nevertheless, at some point in the hillfort’s evolution, they became superfluous, and the bank was thrown up on top of them. Castell Henllys was occupied from c.400 BC to c.100 BC. It was, seemingly, crowded with roundhouses, and may have had a population of 100–150 people. A lack of evidence for preliminary food processing, e.g. threshing, butchery, might imply that Castell Henllys was home to the local elite – prepared foodstuffs being brought into the hillfort by lower status families from the surrounding countryside.

The ‘roundhouse’ is the classic Iron Age dwelling – generally between 5 and 15 metres in diameter, with wattle and daub walls. It is assumed that the walls would have been at least chest height, to allow maximum use of the enclosed space. Since the walls were round, the roof was conical, with a timber framework and a covering of thatch (or possibly turf, which, being heavier, would require more support). Often, the house’s entrance, which might have a porch, perhaps for reasons of belief rather than any purely practical consideration, faces south-east, towards the midwinter sunrise, or east, towards sunrise at the equinox.

At the centre of the house was an open-hearth fire, which would have been maintained at all times.[*] A bronze cauldron might be suspended from a tripod, by a chain, over the fire. The fire would have also enabled food to be preserved by drying and smoking. (Salt was also available to preserve meat.[*]) Spinning and weaving were domestic activities.[*]

Glastonbury Lake Village was built, on an artificial island, in shallow water on the Somerset Levels. It was discovered, by Arthur Bulleid, in 1892. Excavations began in that year, and were completed in 1907.

Since then, interpretations of the layers of archaeology that were uncovered have varied. It seems possible, however, that development of the village started around 250 BC. It peaked about 100 BC (a, roughly, triangular site of approximately 1.4 hectares, housing perhaps fourteen or fifteen extended families, some 200 people[*]), but rising water levels caused its abandonment about 50 BC. As a result of the waterlogged conditions, structural woodwork and wooden objects were very well preserved. The island’s foundations were made of brushwood and logs, filled in with rushes, bracken, peat and clay – the whole being retained by a perimeter palisade. On top, individual circular clay floors were laid as bases for roundhouses. The wall of a house was made by driving a circle of posts through the floor into the foundations (leaving a gap for the entrance of course). The spaces between these uprights were then filled with wattle and daub, and a thatched roof was erected. Arthur Bulleid writes:

Dwellings were occasionally burnt down, but there was no evidence of a great conflagration, nor of the total destruction of the Village by fire. It was in the locality of the burnt huts that information was forthcoming regarding the construction of the walls and roofs. The floors were thickly covered with baked clay rubble, charcoal, and pieces of charred reed. The pieces of baked clay were frequently marked with the impressions of wattle work on one side, and with finger-prints on the other. The fragments of charred reed were presumably burnt thatch. The roofs of the dwellings were supported by a central post, the base of which was usually found near the margin of the hearth embedded in the clay floor. Some of the larger dwellings appear to have had two or even more central posts. Other details of the construction are left largely to our imagination… From the complete hurdles that were discovered we conclude the walls were about six feet high, and some dwellings had partitions. Remains of flooring boards and joists were found in a few huts.

The Lake-Villages of Somerset Fifth Edition (1958) §IV: E

As, over time, the foundations beneath a house compacted, so the floor would have to be built up again. In one instance a sequence of ten floors was found.

In dwelling-mounds with several floors, the superimposed layers of clay generally increased in diameter from below upwards. A new floor was usually accompanied by the re-erection of the dwelling, the diameter of the hut being increased.

The Lake-Villages of Somerset Fifth Edition (1958) §IV: E

The roundhouses were between 5½ metres and 8½ metres in diameter. The entrances were not on any particular alignment. It seems that some houses must have had solid wooden doors (though none remained), since an iron latch-lifter and “a stone with a pivot-hole” were among the finds. In fact, a vast array of everyday objects were found in and around the village – items of bronze, tin, lead and iron; amber, glass, jet and shale; wood, bone and antler; pottery, baked-clay, flint and stone. Bronze-working and iron-working were carried out on the site. Of the 274 bronze items found, most were “objects of personal adornment”, though the prize find was a bowl.[*] Of the 18 tin objects, the most intriguing was a:

… round and solid bar of pure tin; it was capped at each end by a bronze terminal, and measured 10¼ inches in length, and one inch in diameter.

The Lake-Villages of Somerset Fifth Edition (1958) Relics: I

Analysis of the bar’s surface suggested it had originally been gilded. There were 32 lead items, of which 13 were fishing-net sinkers. Of the 111 iron items, only 7 could be classed as weapons – “Daggers (including fragments) 3, spearheads 3, fragment of sword 1” – although:

That the inhabitants had iron weapons is certain, for numerous pieces of bronze bordering, and the terminals of three sheaths were discovered.

The Lake-Villages of Somerset Fifth Edition (1958) Relics: I

There were parts of 4 snaffle-bits; and 4 saws – the teeth were set so that the cutting action was on the backward rather than, as in a modern handsaw, the forward stroke. Some of the iron tools still had their wooden handles.[*] 2 ‘currency bars’ were found, but just a single coin – an incomplete ‘potin’, dating from round-about 50 BC.

Five objects of amber and twenty-seven of glass were found, and with the exception of three glass specimens, all the objects were beads. A lump of greenish-blue glass slag is of much interest and importance, as it strongly supports the supposition that glass was made at the Village… Only one object was found made of jet, a highly polished ring or bead …

The Lake-Villages of Somerset Fifth Edition (1958) Relics: III

The lathe was an Iron Age innovation, and objects of turned-wood and turned-shale were found at the Village.[*] There were 31 shale objects:

… these include two very fine lathe-turned and ornamented armlets. From the fact that a waste core was found, and several rings and other specimens in the process of making, we are led to believe that the shale was imported and worked at the village.

The Lake-Villages of Somerset Fifth Edition (1958) Relics: IV

Amongst the “large collection” of wooden items were:

Portions of fourteen tubs and cups, either cut from the solid or stave-made. In some the staves were dowelled together, in others fitted together like a modern barrel and hooped. Sixty-three pieces of frame-work, parts of looms or apparatus for making textile fabrics. Five ladles, parts of three lathe-turned wheel-hubs, several wheel-spokes, a ladder with four steps, a small door … [handles from various tools] … fragments of baskets, two small wood pins found in association with a bronze mirror, tweezers and some dark colouring-matter, a draughtsman, the tops of two spade-handles, three pieces of finely ornamented wood bands, two wooden mallets, three stoppers, an oak dish or trough, beams and planks with mortise holes, and several wicker-made hurdles.

The Lake-Villages of Somerset Fifth Edition (1958) Relics: V

A dugout canoe, of oak, was discovered by a labourer clearing the ditch of a nearby field. There were hundreds of bone and antler objects, including 89 weaving combs (counting fragments, and 5 which were not finished) and 5 dice. Three of the dice were found in association with 23 disc-shaped polished pebbles, which, presumably, were counters.

Although several tons weight of pot-sherds were unearthed only some half-a-dozen vessels were found intact… With reference to the production of the pottery, the clay and the other materials composing the various pastes were probably all products of the immediate neighbourhood, and there are reasons for thinking that the whole of the Glastonbury pottery was made at or near the village… With regard to the mechanism employed in making pottery, very few pots show any evidence of being wheel-made; the large bulk of the pottery was built up by hand.[*]

The Lake-Villages of Somerset Fifth Edition (1958) Relics: X

Among the “very numerous” objects made of baked-clay were loom weights and spindle whorls, pieces of two tuyères (used to conduct the blast of air from bellows into the furnace), and hundreds of sling pellets. Most of the (in excess of 230) spindle whorls found at the Village were of stone. Also of stone, either whole or fragmentary, were 18 saddle quern components and 38 rotary quern components.[*] A Neolithic polished stone axe-head, and part of a second, were also discovered. These would not have been brought to the Village for practical use, but as curios or objects of reverence. There were also a few flint artifacts.

There is no actual proof that the fabrication of flint implements took place at the Village, but as the collection includes 362 flakes and chips, some of the latter being very small, we are led to believe that a flint industry was carried on, if for no other purpose than for the making of scrapers.

The Lake-Villages of Somerset Fifth Edition (1958) Relics: XIV

Of course, living on an artificial island in a lake is not a concept unique to Somerset. In Scotland, loch-dwelling, on artificial islands called ‘crannogs’, has a history spanning thousands of years – from the 4th millennium BC, until at least the 17th century AD. It does appear, however, that the Iron Age and Roman period was the heyday of the crannog.[*] Typically, like at Glastonbury Lake Village, crannogs are true islands, having bases built up from the loch floor with organic material – brushwood, logs, peat, turf (perhaps augmented with stones) – retained by timber piling (a technique known as ‘packwerk’), but some loch-dwellers evidently lived in houses raised above the water on timber piles. These free-standing structures are also referred to as crannogs.

There are the remains of eighteen crannogs in Loch Tay alone. One of them, Oakbank Crannog, is a totally submerged mound, around 25 metres in diameter, that has been explored by underwater archaeologists since 1980. It is a multi-phase Iron Age site, but it seems it was originally constructed, around 500 BC, as a free-standing crannog – later, the collapsed structure becoming the basis of a packwerk-type mound, being overlaid with a layer of stones and a layer of large boulders.[*] Amongst the objects found in the debris of the collapsed house were: a wooden dish with butter sticking to it; a fragment of woven woolen cloth; a turned wooden disk (the waste removed from the base of a bowl once it has been taken off the lathe); an oak ard. Based on the findings at Oakbank, a reconstruction of a free-standing crannog was built in Loch Tay, some three miles east of Oakbank.

In the wild, rocky, landscape of ‘Atlantic Scotland’ (the north and west, including the Northern Isles and the Western Isles), drystone-walled roundhouses (today, called ‘Atlantic roundhouses’) began to be built around 800 BC. The early types, ‘simple Atlantic roundhouses’, had solid walls – sometimes quite massive: in the region of 4 metres thick. From about 400 BC, they began to acquire architectural features, such as galleries and stairways, within the thickness of the wall. These are classed as ‘complex Atlantic roundhouses’. Traditionally, Atlantic roundhouses are, in general, called ‘duns’ (the word dùn is Scottish Gaelic for ‘fortress’). Some are constructed on rocky-based artificial islands, and these are known as ‘island duns’. The imposing drystone towers, traditionally called ‘brochs’, fall under the heading of ‘complex Atlantic roundhouse’. Broch towers were probably fully developed by about 200 BC.

The towers are entered by a single, low, doorway and have no windows. This somewhat featureless exterior, combined with the slope of the wall, has led to them being likened to power station cooling towers. The entrance passage usually has a, so called, ‘guard-cell’ to one side, within the wall’s thickness. Presumably (though not proven), the towers would have had a number of wooden upper floors and a conical roof – certainly, typical features of ‘broch architecture’ are ‘scarcements’, i.e. stone ledges, around the inside of the tower which could have provided support for an internal timber structure. The name ‘broch’ is derived from the Old Norse borg, i.e. ‘fortress’, and these substantial towers would, no doubt, act as a stronghold when raiding parties attacked. They are, however, very prominent in the landscape. They were, clearly, built to be impressive. It does not seem unreasonable to see them as status symbols – blatantly advertising the wealth of the family who lived there. Their heyday would appear to span four hundred years, the last two centuries BC and the first two AD.

Broadly contemporary with the broch towers, and found in the Western Isles (mainly) and Shetland (but not Orkney), is a completely different class of circular drystone building: the ‘wheelhouse’ – named from its distinctive floorplan.[*] Arranged radially around the interior, like wheel spokes, are a number of stone piers. The central hub area is left open. The wall and piers supported corbelled stone roofs, which domed above each of the inter-spoke cells. At Cnip, on Lewis (Western Isles), the corbelled roofing of a wheelhouse is partially intact. The domes’ apexes would have been something like 6 metres above floor level. Whereas brochs probably made lavish use of timber, in an area where its supply was very limited, wheelhouses only required timber to construct the framework of a roof over the hub area. The example at Cnip had an interior diameter of 7 metres, eight spokes, and the hub diameter was in the region of only 3½ metres. An example at Sollas, on North Uist (Western Isles) was 11 metres in diameter, had 13 spokes, and a 7 metre hub diameter. Frequently, wheelhouses are sunk into machair (sand-based grassland), so don’t need massive walls to hold them up – Cnip and Sollas are both like this. Only the roofs would have projected above the land surface.