Celtic Iron Age hillforts are prominate features of the landscape of southern
Britain, especially the chalk uplands. Built during the five centuries of the
La Tène period, they may represent the princely fortified palaces of
the Hallstatt adapted to later Celtic warrior society. The spread in the use
of iron to agricultural tools in the La Tène increased productivity and
the accumulation of food surpluses. These supported a warrior elite who exercised
control by force of arms and were separate from agricultural workers. Only high
status burials, usually containing rich grave goods of this warrior elite, have
survived. Nevertheless, a complete picture of hierarchical Celtic society must
include humbler elements and these would have been present amongst the forces
available to attack or defend a hillfort.
Cunliffe emphasises that hillforts did not exist in isolation but were integral
parts of wider territory. He suggests, for example, that Cissbury hillfort in
the first century BC may have been the focal centre for a population of about
1,400 people farming the Sussex downland between the Arun and Adur rivers. Danebury
may have had a similar hinterland. Hillforts probably had several functions
such as display of power and defence. For example, Danebury, may have been used
for settlement, production and storage, perhaps in pits and the enigmatic four-poster
structures.
Hillforts were developed by a warrior-dominated agricultural society where wealth and power could be acquired by warfare. The warrior elite's well-being relied on agriculture so by mounting patrols and spying on neighbours it would have ensured that farmers could work in safety. Once the harvest was secure, part of the warrior elite, perhaps supported by the fittest workers, could attempt to seize their neighbours' harvest. Seasonal reiving, echoed in the much-later Irish sagas, would have been countered by hillforts. Seasonally occupied by the warrior elite and some of the farmers, hillforts would defend livestock and grain surpluses against raiders.
Hillfort design, especially the increasing complexity of entranceways, such
as Danebury's, seems to point to defence against surprise raids rather than
set-piece battles.
Collis discusses hillforts occupied in times of crisis and indicates that,
if occupation was continuous by some local farmers then agricultural production
and population would fall and consequently the size of the warrior elite and
their ability to defend their territory. Collis' model 3d shows some residents,
craftsmen or perhaps the warrior elite, with the surrounding farms in normal
production. Collis suggests that "the picture here is of a decentralised,
relatively egalitarian society, as opposed to the hierarchical one assumed by
Cunliffe". Perhaps this overstates the point. If hillforts were stores
defended by a warrior elite and focal points for farms which supported those
warriors, then they become more compelling as local centres of social organisation
and wealth, convenient manufacturing centres for craftsmen, trade centres for
their products and locations for inter-tribal gift exchange. "Fort Knox"
rather than "Fort Laramie".
This would make them attractive targets for neighbouring warrior elites and
provide reasons for their defence and destruction. Perhaps to this degree Collis
is correct. Southern Britain of the La Tène could have comprised a patchwork
of economic and social groupings each centred on their hillfort. As Collis states,
"we find no evidence … that the Durotriges were capable of acting
under a centralised leadership; rather Vespasian had to reduce each hill-fort
individually", because, of course, each hillfort represented a separate
community prepared to defend its own territory.
Evidence for Celtic weaponry is provided by excavation of graves, surviving
artwork on a variety of objects and written accounts of those who fought the
Celts in battle, such as the Greeks and Romans. Powell states that Celtic warriors
were equipped with a light thrusting iron sword, a dagger for close fighting
and spears. The cemetery at Nebringen, Baden-Wurttemberg, has warrior burials
with this standard set of weapons together with a large shield as does that
at Owlesbury in Hampshire. "Their long swords", according to Strabo
(IV, 4, 3) were "fastened on the right side" although the Parisii
of Yorkshire wore them on their back and drew them over their head.
Spears were of varied design, some had serrated, flame-shaped blades which,
according to Diodorus (Histories V.30), were designed to break "throughout
the entire length so that the blow not only cuts but also tears the flesh, and
the recovery of the spear rips open the wound". Each design had its specific
purpose. For example javelins would have been used as thrown missiles and pikes
for thrusting forward. Cavalry may have been equipped with lances. However the
exact purpose of each design remains elusive.
Diodorus (V, 30,2-3) continues: "Their armour includes man-sized shields
decorated in individual fashion. Some of them have projecting bronze animals
of fine workmanship …" Made principally of wood and leather with a
central umbo or boss to protect the warrior's fingers holding the central bar,
these shields occasionally survive. Grave 183 at Münsingen in Switzerland
is more typical though in that it contained only the iron boss while the wooden
parts had rotted. Metal and metal-faced shields, such as that from the Thames
at Chertsey, were also decorated but probably intended for parading or as votive
offerings rather than for warfare as they would have been too heavy in combat,
although they may have played a role in the pre-battle display. Fighting shields
appear to have been made of several thin layers of wood set crossways to create
a strong plywood material perhaps subsequently covered in leather and then decorated.
This provided a light, resilient and durable shield able to withstand cutting
sword strokes, thrusts from spears and protected the warrior from stones.
South Yorkshire provides evidence for chain-mail tunics, one being excavated
from a warrior's grave at Kirkburn. This is corroborated by the late first-century
BC statue of a Gaulish chief found at Vachères in France and Diodorus's
comment, when describing Celtic warriors, that "… Some have iron breastplates
of chain mail …".
To complete the Celtic warrior's accoutrements Diodorus concludes with: "On
their heads they wear bronze helmets which possess large projecting figures
lending the appearance of enormous stature to the wearer. In some cases horns
form one part with the helmet, while in other cases it is relief figures of
the foreparts of birds or quadrupeds …". Excavation evidence backs
up this written account. Although metal helmets would have been heavy, especially
with the leather lining cap that was undoubtedly fitted for comfort and additional
protection, their widespread use throughout the classical world indicates their
effectiveness, weight being the trade-off for increased security. Highly-decorated
helmets probably fulfilled ceremonial and display functions in Celtic pre-battle
display.
The elite infantryman of the Celtic army can now be visualised. Well-protected,
well-armed, well-fed, fit and mobile, he and his fellow warriors would form
a powerful force whether attacking or defending a hillfort. The most wealthy
warriors would have had a chariot to take them into battle and manoeuvre about
the field. Excavation of early La Tène high-status graves, for example
that at La Gorge Meillet, Marne, France, show warriors equipped with a two-wheeled
chariot which Powell suggests was drawn by a pair of small.
Cavalry gradually supplanted the importance of chariots in the Celtic army
and James shows that saddles introduced in the late Iron Age made mounted forces
effective despite the lack of stirrups. Celtic cavalry were rated as the best
horsemen in the Roman army of the late Republic, for example.
The Celtic warrior described would have represented the wealthiest and best
equipped element of the forces brought against a hillfort as his weapons would
have been relatively expensive, particularly a chariot and ponies or a cavalry
mount. Undoubtedly Celtic forces would also have included slingers which, although
lightly equipped, would have required a supply of stones for both defence and
attack, as shown by excavation evidence. Cunliffe, for example, states that
huge stockpiles of carefully selected stones were found near the gates of Maiden
Castle and Danebury in readiness to repel attackers. Slingers, cheap to equip,
were probably drawn from the lower echelons of society, especially from amongst
herdsmen where the sling was used to drive off predators. However, as Avery
points out, many of the excavated stones may have been for hand throwing, perhaps
by all sections of the army. The poorly-equipped majority of the Celtic army,
armed perhaps with simple wooden pikes and javelins and carrying woven shields
surely included stone throwing in their armoury. Also, once a fire had been
lit, any low-value wooden missiles could have been added as fuel.
A successful leader demonstrated his martial abilities by leading raids and
rewarding his followers with a share of the booty. Feasts, such as that described
by Athenaeus at which status was proclaimed and confirmed, appear to have been
the starting point for such raids.
Evidence for Celtic warfare comes from written sources, illustrations on artefacts
and archaeological finds. According to Cunliffe opposing forces drew up facing
each other, heroes from the warrior elite would exchange boastful abuse and
display their costly arms and armour. Perhaps this is where the elaborate metal
shields and helmets played their part. Single combat then occurred followed
either by a widespread mêlée, the retreat of the defeated hero's
force or the dispersal of both forces.
Celts employed psychological warfare to intimidate their enemies and give themselves
greater courage. Polybius' account of the battle of Telamon in 225 BC states
that "… the Romans … were terrified by … the Celtic host
and the dreadful din, for there were innumerable hornblowers and trumpeters
and, as the whole army were shouting their war-cries at the same time, there
was … a tumult of sound …". Diodorus (V, 30, 2-3) states that
the Celtic "trumpets are again of a peculiar barbaric kind … and produce
a harsh sound which suits the tumult of war." Examples of Celtic battle
horns, carynxes, have been discovered in Banffshire and Lincolnshire and a warrior
grave at Nebringen contained a rattle. No doubt Celtic warriors, like all soldiers,
also used alcohol to supplement their battle fever and deaden pain and exhaustion.
Two types of attack need to be distinguished. The most common would have been
seasonal raids to steal grain, livestock and other portable wealth. The less
common would be a full-scale attack to capture and destroy a hillfort and annex
the wider territory it defended.
The first method an attacker may have used is subterfuge, that is not revealing
the attack until the gate was secured. Using spies or buying traitors the gates
might be secretly opened so that, with a small number of elite warriors, this
crucial weak point in the defences could be captured. If it was a raid the band
would seize everything they could before retreating. If it was an attempt to
seize the hillfort itself the gate would be destroyed and the rest of the commander's
forces brought forward to disarm the defenders and occupy the site.
Rivet examines Caesar's description of the Gauls' attack on the Bibrax hillfort
in his Gallic Wars (Book II, 6) and convincingly argues that it is accurate.
In the specific example of Danebury, Cunliffe imagines an attack on the eastern
gate. His attackers batter their way through the defenders' fire. While this
might portray a raid, if an army attacked using the methods described by Caesar,
it would have the advantage of bringing more sling stones against the defenders
at the outer gate and either destroying them or forcing them back. Attacking
slingers could mount the outer hornworks and attack the "command post",
restricting its ability to counterattack with sling stones against attackers
burning the outer gate with impunity. Although there is no archaeological evidence
it seems likely that attacking slingers might be accompanied by other light
troops with woven hurdles for cover, together with additional ammunition.
Attacking slingers would be at a height disadvantage when clearing defenders
but this would be more than outweighed by the greater weight of sling stones
they could bring to bear. In addition, stonesmen, as described by Avery, might
supplement this attack by moving close in, protected by the attacking slingers'
barrage fired over their heads at the defenders up on the ramparts. The attacking
stonesmen could add their missiles to the assault at a range of 20 metres. With
attackers on the outer hornworks, the defenders' height advantage is reduced
and with the outer gate destroyed the east gate "barbican" is untenable
and the attack could develop as indicated in figure 20.
Cunliffe suggests that an attack on the east gate would have been extremely
hazardous and "that the gate was eventually burnt, presumably by enemy
action, is all the more remarkable." However, within this warrior-dominated
hierarchical society the techniques of warfare would have been tried, tested,
well understood and practised with precision. A combined stone and fire attack,
rather than a raid, leading to the penetration of the hillfort defences, would
surely be more hazardous for the defenders than the attackers.
Rather than being a "masterpiece of planning" the Danebury east entrance
seems to be a defenders' death trap if facing an army. Although slingers on
the "command post" may have been able to reach all ramparts, they
would have been exposed, inviting the enemy to concentrate fire upon them. The
key weakness is that, while slingers on the "command post" might disrupt
an attack, no supporting fire can be given from other ramparts. Defenders at
the east gate would have been on their own.
Rivet disentangles from the source material two additional native Celtic methods
of attack distinct from those subsequently learnt from the Romans, that of using
bundles of brushwood to fill up ditches and tunnelling so as to undermine banks.
Caesar's description is of an overwhelming attack by slingers who "surround
the whole circuit of the walls with a large number of men and shower it with
stones from all sides, so that the defences are denuded of men" (my emphasis)
and continues, "Then they form a testudo, set fire to the gates and undermine
the walls." This describes a stone and fire attack on a hillfort, not a
raid. Capturing a hillfort, as opposed to raiding for booty, would have meant
assembling a substantial force divided into specialist units comprising the
warrior elite, expert slingers, specialist firing teams, tunnellers, light-armed
troops and wagons with a plentiful supply of stones and fascines of brushwood.
If it was impossible to gain access by subterfuge this overwhelming and competent force might be displayed to demonstrate to the defenders that they had no chance of holding the fort. Taunts and challenges could be exchanged. Perhaps single-combat took place with possession of the hillfort resting upon the outcome. If a general engagement occurred then defenders in exposed positions, such as Danebury's east gate, would be crushed before an encircling line of slingers assailed the defenders.
As the ramparts were cleared the specialist firing parties could go in with
the warriors. Meanwhile, chariot-mounted warriors might threaten and launch
diversionary attacks at other points in the ramparts, especially if there was
a second gate. Fascines could fill ditches and tunnelling occur under cover
of the tumult. Once a breach had been secured either by burning a gate, demolishing
a rampart by undercutting or by assaulting the rampart across a filled ditch,
then the hillfort was doomed as the defensive ring of fighters imploded and
the action became a mêlée.
The first line of defence for a hillfort commander would be intelligence from
mounted warriors patrolling the defended territory. These warriors would watch
for surprise raids and spy on neighbouring communities that might pose a threat.
However, hillfort design was the primary means of defence and not merely a
boundary barring outsiders. Outside the hillfort itself, the military engineer
might restrict the approach road using lateral banks or install stone or wooden
chevaux de frise in the surrounding slopes to disorganise an attacking force.
The dump ramparts of La Tène hillforts were much wider than the earlier
wall-and-fill Hallstatt period structures and allowed defending slingers and
stonesmen to keep attacking warriors and firing parties at a distance from the
ramparts and vulnerable gates. For example, Danebury's dump ramparts were two
metres high and seven metres wide, while the ditch was six metres deep and between
11 and 12 metres wide. This would have put defenders at the end of the 20 metres
range of stonesmen. Attackers would be forced to rely on missiles such as sling
stones and javelins to clear the ramparts before assaulting the palisade and
gates. This could only be achieved by an army. Seasonal raiding parties without
slingers or stonesmen would have to rush the gate and face the defenders' barrage
of stones, as Cunliffe has suggested.
Entrances were weak points in the bank and ditch defence. Hillfort engineers devised entrances of increasing complexity to prevent "rush" attacks that relied on weight of numbers, a fast-moving disciplined group or a determined, localised attack after an initial barrage of missiles. Earlier wall-and-fill hillforts used extended entrance passageways, narrowing towards the gate to confine attackers within a vertical-sided, stone-lined killing zone overlooked from the ramparts and by a bridge or tower above the passageway. Such hazardous attacks must have been successfully mounted because the change of rampart design during the La Tène period indicates that warfare had also changed to counter them. Wider dump ramparts, manned by slingers and stonesmen, forced the fighting back away from the gates and acted as a deep defence against the range of the attackers' missiles.
Increasing complexity, such as at Maiden Castle, would have slowed and split
up "rush" attacks, caused disorganisation and made them static targets
for defenders' missiles. Surely the Celts would, like the Romans, have used
a planned and integrated attack utilising a variety of forces, such as outlined
above for an attack on Danebury.
The warrior elite alone would be insufficient to defend these wider and less
steep ramparts. To have, as Caesar says, the ramparts "lined with defenders"
all able-bodied males must have been pressed into service. For example, Cunliffe's
suggested population for the Cissbury area is 1,400. Using demographic analysis
of the Münsingen cemetery as a rough guide, this indicates 460 males, of
which perhaps 220 might be capable of fighting. If the warrior elite comprised,
say, 10% this gives only 22 well-armed men to defend Cissbury without using
the non-elite population. This is inadequate and does not reflect the supplies
of stones excavated at hillforts.
The final line of defence would be the main rampart, perhaps with a palisade
and extending over an inner gate by means of a wooden bridge or tower. However,
as Caesar's account suggests, even if a wooden palisade or stone wall surmounted
the rampart this was insufficient to prevent casualties from missiles. If the
defenders could maintain their position on the ramparts against a barrage of
stones and force the attackers to recklessly dash in to fire the gate, then
water and earth from the ramparts could be used to counter this. Perhaps, when
the enemy was first sighted, the gates were immediately closed and drenched
with water.
The development of complex entranceways suggests that the main threat was raiding. Complex entrances can be defended by a small group of warriors and slingers and do not draw men away from farming to man the ramparts. If raiders, by subterfuge or strength, got past an outer gate they would have to traverse a twisting, stone-lined corridor overlooked and controlled by defenders before reaching an inner gate. The corridor is a death trap for raiders. Previous knowledge of the entranceway gained by spying or bought from traitors would be vital. This is perhaps why entranceway design developed to conceal inner gates and their guards.
It seems, for example, that Danebury's east gate was designed to control anyone
entering the hillfort, prevent spies seeing the state and readiness of defences
and forestall surprise raids. Maiden Castle's interlocking defensive works,
in their final form, have no exposed position but still conceal the complexity
of the defence works. The gateways form an integral part of the ramparts and
were probably designed to withstand assaults by Celtic armies as well as defeat
raider bands.
Hillforts appear to have been used for a variety of purposes, many unconnected
with warfare and individual hillforts were perhaps tailored and developed to
counter particular threats. Maiden Castle, for example, grew with the needs
of the local tribe until, as the final redoubt of the Durotriges, it was assaulted
and captured by the Romans.
Danebury's design is a masterpiece of military engineering, as Cunliffe says.
It's gradual development mirrors the growth of the Celtic agricultural economy.
The earliest phases indicate defence against general attack while its latest,
most complex design, counters raids. It probably contained the portable wealth
of the local community, ruled and defended by its warrior elite. Its destruction
may represent the incorporation of that community, forcibly or otherwise, into
another tribe or the gates may have been ritually burnt when it was abandoned.
Celtic military engineering of hillforts was proactive and responded to new weapons, tactics and the changes within their society. Celtic methods of attack and defence illustrate their skill, ability, understanding and creativity. They provide an insight into a mature, developing and confident society that was only destroyed when confronted by the technological and organisational superiority of the Roman empire.
Works referred to
Avery, M., 1976: "Hillforts of the British Isles: A Student's Introduction", pp 1-58 in Harding (ed.), 1976.
Avery, M., 1986: " 'Stoning and Fire' at hillfort entrances of southern Britain", World Archaeology, Volume 18, No. 2, pp 216-230.
Avery, M., 1993: Hillfort Defences of Southern Britain, 3 vols., British Archaeological Reports, Oxford, Tempus Reparatum
Collis, J., 1981: A theoretical study of hill-forts, pp 66-76 in Guilbert (ed.), 1981
Collis, J., 1984: The European Iron Age, London, Batsford
Cunliffe, B., 1971: Some Aspects of Hill-forts and their Cultural Environments, pp 53-70, in Hill and Jesson (eds.) 1971
Cunliffe, B., 1983: Danebury, Anatomy of an Iron Age Hillfort, London, Batsford
Cunliffe, B., 1991: Iron Age Communities in Britain, London, Routledge
Cunliffe B., 1994: Iron Age Societies in Western Europe and Beyond, 800-140 BC, pp 336-372 in Cunliffe (ed.) 1994
Cunliffe, B. (ed.), 1994: The Oxford Illustrated Prehistory of Europe, Oxford, OUP
Cunliffe, B., 1995: Iron Age Britain, London, Batsford/English Heritage
Cunliffe, B., 1997: The Ancient Celts, Oxford, OUP
Guilbert, G., (ed.), 1981: Hill-Fort Studies, Leicester, Leicester University Press
Harding, D. W., (ed), 1976: Hillforts, Later Prehistoric Earthworks in Britain and Ireland, London, Academic Press.
Hill, D., and Jesson, M., 1971: The Iron-Age and its Hill-forts, Southampton, Southampton University Archaeology Society
James, S., 1993: Exploring the World of the Celts, London, Thames and Hudson
Powell, T. G. E., 1983: The Celts, London, Thames and Hudson
Rivet, A. L. F., 1971: Hill-forts in Action, pp 189-202 in Hill and Jesson (eds.) 1971