Iron Age Roundhouses – Danebury
Text by Barry Cunliffe.
All the necessities for making sound, comfortable buildings were readily to hand, and, in so far as we can judge from ground-plans and from tools, the constructional ability of the Iron Age inhabitants was considerable. There is no need to suppose that the structures were rustic hovels roughly lashed together. Efficient jointing and a complete mastery of materials would have meant stable structures finished to a high degree.
The principal house type at Danebury Ring was a circular, one-door structure ranging between 6 and 9m (20 and 30ft) in diameter. About 70 have been excavated. They fall into two basic types: those built with walls of vertical planks bedded in continuous wall slots; and those whose walls were of wattlework.
Building CS1 is the only example of the first type (images 01, 02). It had been built on an artificially created terrace cut into the chalk and for this reason had escaped the extensive erosion which has so denuded the other house plans, the only obscuring feature being a large tree which grew close to the doorway. The ground-plan consisted simply of a trench of circular plan, 20-30cm (9-12in) wide, cut into the chalk and continuous except at the door where two large post-pits were dug to contain the verticals of the door-frame. In the wall trench it was possible to distinguish two entirely different fillings: the soil-filled voids where the timbers had once been; and the chalk and flint packing on either side.
Careful planning and dissection showed the wall timbers to have been individual planks up to 40cm (16in) in breadth placed end to end. That some of them were distinctly wedge-shaped showed that they had been split radially from a large trunk. This kind of wall construction would not have been partic ularly weatherproof but tongue and grooving, with moss and resin caulking, or more simply the attachment of additional planks to cover the joints, would have overcome problems of draughts and driving rain, while the possibility that the conical roof may have been taken down to the ground, providing additional insulation, has much to commend it.
The door-frame was constructed of two pairs of roughly squared timbers, one pair in line with the wall, the other in front and at right angles, creating a shallow porch-like structure. Both would have been kept rigid by lintels. Nothing survives to indicate what kind of door was provided but one possibility is that it was a movable wattle or plank structure which could be slotted into place between the pairs of vertical posts or removed altogether and kept inside when the house was open.
How the roof was treated can only be guessed, but a conical arrangement of rafters thatched with reeds would have been simple and effective. The absence of internal supports might at first sight seem puzzling but so long as the wall top was suitably bound by a ring beam, the lateral thrust of the roof timbers would be contained and the whole structure would be quite stable.
The second type of house found at Danebury Ring, similar to the first except that the wall was made of wattlework, was the most common type (images 03, 04, 05, 06). A good example is provided by building CS20 where the details of the wattle wall are extremely well preserved (49). Here the individual vertical wattles, 2-4cm (%-1Yzin) across, were driven into the ground at intervals of about lScm (6in) around the circumference of the building. Driven, however, may not be quite accurate since it is probable that the holes were first made with a crowbar- a hard wood shaft with a pointed iron end – and the poles lightly hammered into place.
Once in position the horizontal members could be woven to create a tough and comparatively rigid wall, quite strong enough to support the weight of the conical roof or to allow some of the verticals to be taken up to form a roof. The door-frame was much the same as that of the plank-built CS1 with two exceptions: first, there was an additional pair of posts in front, but since these were sealed by the chalk-rubble door-sill they may have been something to do with the building process, to be removed when the superstructure was finished; second, a horizontal threshold beam had been laid between the posts of the inner door-frame. This is a particularly interesting feature to which we shall return again below.
Inside the building a rammed chalk floor survived but through it had been cut a shallow groove just inside the wall and not quite concentric with it. There was also an inner door-frame giving the impression, in plan, of being part of an inner vestibule. These features are unusual and not easy to explain, unless it is supposed that the inner slot took a slight wall of wattle to serve as an insulating inner skin and the vestibule was also part of an elaborate scheme to exclude the draughts. Another interesting feature was the two storage pits dug below the floor. Although they look rather hazardous on the plan, in their original form before erosion, the mouths would have been little more than 50cm (18in) in diameter and could thus simply, and safely, have been closed by a board or wicker cover.
CS1 Reconstruction at Butser Ancient Farm.
A number of buildings were discovered in which the only surviving traces of superstructure were the doorposts, the chalk door-sill and, where conditions of survival were good, the actual floor surface; there was no trace to show that wall timbers had been bedded in the subsoil. In other words the wall must have simply sat on the surface of the ground. One possible explanation for this is that the wattlework building had been constructed elsewhere and had been physically moved to the new location. If this were so there would have been considerable advantage in the door-frame being kept rigid by a threshold braced between the lower ends of the frame posts, matching the lintel at the top. To resite the structure would merely have required two new holes being dug for the projecting ends of the door-frame verticals.
Such an over-elaborate explanation needs to be further examined. In the first place why move a building? The answer to this must lie deep in the social structure of the community beyond archaeological reach; in the second, was it possible? In practice it was; a wattle wall could simply be uprooted or the stumps of the poles may already have rotted at ground level. If the thatch were removed a band of men could, without much effort, carry a small structure to a new location, where once sited, it could be rethatched. All this may seem far fetched but some such explanation is needed to contain the archaeological evidence. More to the point, the movement of timber structures is not unusual among surviving primitive communities.
The question of the superstructure of stakebuilt buildings presents interesting questions. Are we justified in interpreting the smaller stake-built structures in terms of wattle drum-like walls with conical raftered roofs? Another distinct possibility is that some of them may have been entirely of wattlework with selected wall poles bent inwards and bound in position with wattles to create a beehive style of roof continuous with the walls (image 04). The exterior would certainly need to be thatched to weatherproof it but such a structure would be quick to build, strong, and easy to move. Raising this possibility, without offering a conclusion, shows something of the limitations of the archaeological data.