32. Hadrian's Timber Wall

The latest research on this topic is presented here.
Between Hadrian’s Wall and the ditch to north, archaeologists have found three lines of double postholes running parallel to the Wall, which may represent an early timber 'Wall', albeit temporary, comprising a box rampart and the ditch. This was almost certainly the largest structure timber ever built in this country, its full extent is not known for certain, but it was quite probably 117 km long, and would have required an estimated 2.5 million trees.

The evidence for these double postholes, often referred to as ‘cippi pits’, had been picked up in several excavations, and was compiled by Paul Bidwell of Tyne and Wear Museums (TWM) Archaeology, who were responsible for several of the excavations, [1]. His paper sought to set the evidence in the wider context of other Roman frontiers, and drew on Julius Caesar’s Account of the Gallic war, [De Bello Gallico],[2], particularly the siege of Alésia, in reaching his conclusion that these postholes represented ‘obstacles’ on the berm, probably sharpened wooden entanglements, similar to the ‘cippi’ referred to by Caesar. We shall return to these arguments, and Caesar, later, but it is clear that I consider term 'obstacles’ to be somewhat underselling this remarkable structure.

The Wall, ditch, and 'cippi' pits at Shields Road, Byker, Newcastle [4]

The Evidence
The evidence was admirably described by Paul Bidwell,[1], so it will not be covered in great detail, but it comes principally from excavations at Buddle St, Wallsend, [3] Shields Road Byker, [4], Throckley, [5], and also from Melbourne St, Newcastle,[6]. As a structural archaeologist working with postholes, (my) best practice is consider each site on its own merits, and my analysis is based primarily on the site at Buddle St, where there is clearly defined block of pits, [below].

The Excavation and plan of Buddle Street, Wallsend

The double post pits occur in 3 rows, the inner and outer rows are aligned parallel to the wall (E-W), and the central row perpendicular to it (N-S). The pits have been robbed, and appear enlarged by this process, particularly at Buddle St; at Shields Road they were more rectangular and better defined. In most cases the traces of a pair of post pipes are evident at either end of the pit, representing posts of 0.2 – 0.3m in diameter.

Archaeologists spend a surprising amount of time watching men digging holes in the road, it's often ‘routine’, but just occasionally it really pays off, and the watching brief on a trench being dug by Northumbrian Water in Throckley yielded further important evidence. The trench ran for 2.2 km along the berm,parallel to wall, giving a long, but very narrow, glimpse of the post pits, and proving they extend as far as Throckley.

In addition, it also produced a most important piece of evidence; as the pits approached the position of Turret 11b on the wall, they changed direction towards it. What is significant here, is that the ditch does the same thing, curving to towards the turret. This suggests that the post pits, ditch, and turrets are part of the same system, at least at this point. It has puzzled archaeologist why the ditch curved in towards the turrets creating a narrowing of the berm, a phenomenon observed in several places, [1], but this is explained if the rampart represented by the post pits ran between, but in front of, the turrets,along with the ditch, as a temporary expedient, while the stretches of stone walling linking the turrets on a more direct line were under construction.

The deviation of the Postpits and Ditch toward the turret Throckley.

The theoretical model; a Hadrianic box rampart.

A theoretical model is a best fit design which conjectures the position of the horizontal timber based on the known position of the posts, the logic inherent in the structure, and the context of its construction. The Model is based on the use of a 10’ timber baulk as a standard component, although the posts are in some cases twice this length. Clearly, a variety of different lengths could have been used, and lengths in the range 10’ –12’ can readily be accommodated in the structure, and of coarse shorter lengths can also be used. As the 10’ baulk appears to be inherent in the design, and would greatly simplify the logistics and foraging, the model is illustrated on this basis.

Basic interpretation and simple theoretical model of the structure at Buddle St.

Despite the denuded state of the individual post pits the structure at Buddle St is laid out with a significant degree of accuracy and regularity. Each pit is centred at 4’, in Roman feet, (Pes - 0.296m), with the entire structure being 10’ wide edge to edge. The overall geometry is heavily triangulated, which would facilitate bracing, and begins with an offset, suggesting that this was also an important aspect of the design to help avoid ‘racking’. The front and back face, and perhaps other aspects of the structure could be offset, and this is shown in the model.

Theoretical Model; horizontal module plan and general section/elevation

The basic structure works by placing N-S baulk between the outer opposing pairs of posts, which project to support the baulks forming the E-W facings. The baulks running up the middle between the central set of posts support the centre of the N-S baulks. Laid out this way, natural gaps appear to accommodate diagonal bracing baulks to strengthen the structure. The box rampart is built up in alternate layers of E-W and N-S baulks, presumably offset and braced as required.

Box ramparts were widely used, particularly in the bronze age, and they have been discussed in an earlier articles, in essence it is a massive log built cavity wall, filled with material from a ditch which was usually dug in front.

Log built walls are fairly easy for an attacker to scale, [left A], so some form of parapet would be added to prevent this, [left B]. The ditch is an obstacle to direct assault on the rampart face, and making it more difficult to scale with ladders, [left c]. The spoil, in addition to filling the structure, can be built up on the berm to form a continuous slope with the ditch known as a glacis, giving the attacker no firm or level position for the attack the rampart from, [left d]. The glacis material also strengthens and conceals the base of the rampart. Spoil may be added to the outer edge of the ditch, where it is known as a counterscarp.

Putting some form roofing over the structure protects the defender and the structure from missiles and the elements. Similar features, built on medieval fortifications, often overhanging, (although they seldom survive), are known as hoardings, or even hoards, [7] ,(hiding and protecting the defenders).The model does not specify the distribution of spoil, but it must have gone somewhere, and filling the base of the rampart and creating a glacis is the most likely scenario, most excavations have found trace of a bank on this southern edge of the ditch, but its origin is unclear.

Theoretical model Construction Details

In the model, the top of the posts are tied together by horizontal bracing timbers and ties (N-S), jointed to the head of the post. Interestingly, the centre posts both align to only one of the outer posts leaving one post in each pair free to support a second set of braces and ties higher up the structure. The lower set is presumably the platform at the top of the rampart, and the higher supports a roof, or hoarding, or a second, higher gallery for defenders.

Building from logs necessitates little jointing, the baulks can be largely held in place by the weight of structure above, the gaps this creates between layers allows the use of wood that is not particularly straight. The basic model has a lot of space between the baulks, whether the facing is offset, [left A], or in blocks, [left B], these gaps can be filled with additional timbers, both N-S, [left c], and E-W, as well as further bracing in all planes.
One obvious variation is to have double thickness facing by placing baulks on both sides of the outer posts, [right]. The structure is very flexible and adaptable, the upper part can be lightly constructed, while the lower portion can be of much more massive, and part filled with spoil from the ditch, which can also be piled on the berm at the foot of the rampart to form a continuous glacis slope with the face of the ditch.
Timber logistics
"The line between disorder and order lies in logistics…” Sun Tzu The Art of War (c6th bce)
“Forget logistics, you lose"Lieutenant General Franks, 7th Corps Commander, Desert Storm [8]
The construction of the Hadrianic frontier, like any military operation, was foremost an exercise in logistics; in terms of the ‘Wall’, one obvious problem being the lack of existing sources of stone, aggregate, and lime, and the infrastructure necessary to exploit them, which would have contributed to the 6 years that it took to build it. The retreat from Southern Scotland, and the creation of the frontier in this form, suggests that the overriding concern in Northern England was ‘security’, without it, the job of constructing the wall would be much more difficult, thus creating a temporary defensible frontier could be seen as entirely logical, and almost a prerequisite. To be feasible, the project has to have realistic logistics, but since timber was the principle material used for the construction in the Iron Age, it should have been readily available, along with the roads and tracks, necessary to exploit it.

It is possible to estimate the quantity of timber required for the model, but what this represents, in terms of the area land needed to supply it, is dependent on 3 main factors;

1. The size of trees
2. The density of trees in woodland
3. The proportion of woodland in the countryside
As we have discussed before [9], the growth pattern of trees, in terms of size and density of woodland, can be modelled from standard forestry tables [10]. The amount woodland is not easily discerned for this period; however, Northern England was historically quite well forested, when a figure of 15% coverage was a rough average for the medieval countryside. Using this figure as a minimum, the model visualises from 15-50% woodland coverage in the area around the wall, in terms of the timber required to build it.
The model examines building the rampart from 10’ baulks taken from small, medium, and large trees; - you need fewer larger trees, but they grow in less dense woodland. The model assumes the use of timber in the round, rather than squared baulks split from larger timber. Each 4’ section of the basic model requires between 300’ and 470’ of timber, depending on thickness, which corresponds to about 20 – 47 trees, depending on their height; scaling this up into new money, each 100m stretch requires between 867 - 3953 trees, very roughly 3 ha worth.
What the model shows is that the larger more mature trees, [60 –80 year old oaks left], were the most efficient source of timber, with small younger trees being the least, despite growing more densely. As noted in earlier articles on this site, this is exactly the type of tree would have been grown for building in prehistory, [9].

A graphical representation of the timber required for the rampart, in terms of area of woodland

The model, though crude, and quite conservative, produces a reassuring answer; even with only 15% woodland cover, sufficient timber could probably be found within 1 km of the construction site. So even if we double the amount of timbers required, and then significantly increase its scarcity, the timber wall would appear logistically feasible, and quite doable - which is important.

Historical sources

Julius Caesar's Account of the Gallic War, is a must read far all archaeologists of the Iron Age, granted it's partial, biased, and in places, little short of Julian propaganda, but we don’t exactly have a plethora of first hand accounts of warfare in ‘Barbarian’ Europe to choose from. It tells the story of Caesar’s campaigns to subjugate Gaul between 58 – 51 bce, a quite extraordinary tale of total warfare Roman style.

One of the big set pieces at the climax of the war, and the book, is the siege of Alésia, a hillfort in France, where the Gaulish leader Vercongeterix, [below left] ,was holed up with most of his army. The Romans built a series of encircling siege works around the hillfort, including diverting rivers, then, in order to protect himself from any relieving forces, and the Gaulish cavalry, which had escaped, built a second set of defences to protect their siege works from attack from outside, [below [11]].

There is quite a detailed account of the building of these works, which is worth reproducing in full:

LXXII. Caesar, on learning these proceedings from the deserters and captives, adopted the 

following system of fortification; he dug a trench twenty feet deep, with perpendicular sides, in such a manner that the base of this trench should extend so far as the edges were apart at the top. He raised all his other works at a distance of four hundred feet from that ditch; [he did] that with this intention, lest (since he necessarily embraced so extensive an area, and the whole works could not be easily surrounded by a line of soldiers) a large number of the enemy should suddenly, or by night, sally against the fortifications; or lest they should by day cast weapons against our men while occupied with the works. Having left this interval, he drew two trenches fifteen feet broad, and of the same depth; the innermost of them, being in low and level ground, he filled with water conveyed from the river. Behind these he raised a rampart and wall twelve feet high: to this he added a parapet and battlements, with large stakes cut like stags' horns, projecting from the junction of the parapet and battlements, to prevent the enemy from scaling it, and surrounded the entire work with turrets, which were eighty feet distant from one another.

LXXIII.--It was necessary, at one and the same time, to procure timber [for the rampart], lay in supplies of corn, and raise also extensive fortifications, and the available troops were in consequence of this reduced in number, since they used to advance to some distance from the camp, and sometimes the Gauls endeavoured to attack our works, and to make a sally from the town by several gates and in great force. On which Caesar thought that further additions should be made to these works, in order that the fortifications might be defensible by a small number of soldiers.
Having, therefore, cut down the trunks of trees or very thick branches, and having stripped their tops of the bark, and sharpened them into a point, he drew a continued trench everywhere five feet deep.
These stakes being sunk into this trench, and fastened firmly at the bottom, to prevent the possibility of their being torn up, had their branches only projecting from the ground. There were five rows in connection with, and intersecting each other; and whoever entered within them were likely to impale themselves on very sharp stakes. The soldiers called these cippi.

Before these, which were arranged in oblique rows in the form of a quincunx, pits three feet deep were dug, which gradually diminished in depth to the bottom. In these pits tapering stakes, of the thickness of a man's thigh, sharpened at the top and hardened in the fire, were sunk in such a manner as to project from the ground not more than four inches; at the same time for the purpose of giving them strength and stability, they were each filled with trampled clay to the height of one foot from the bottom: the rest of the pit was covered over with osiers and twigs, to conceal the deceit.
Eight rows of this kind were dug, and were three feet distant from each other They called this a lily from its resemblance to that flower.
Stakes a foot long, with iron hooks attached to them, were entirely sunk in the ground before these, and were planted in every place at small intervals; these they called spurs.

De Bello Gallico LXXII - LXXIII

The term ‘cippi’ seems to apply to a trench filled with branches, and trunks with branches still attached, and then backfilled, presumably, wood remaining after the straight lengths of trunk had been used in the main defences. It is placed at a distance from the rampart so that the archers would be able to engage the enemy as they dealt with the entanglement, in addition, once this had been breached, the defenders could concentrate their fire on the breach.

It appears to me, that the term ‘cippi’ in this context is a bit of soldierly gallows humour Caesar is sharing with his readers. A cippi is a marker, a post, often marking a grave, [12], "so that the men call them cippi" – effectively ‘gravestones’, and can be seen in the same light as the expressions ‘lilies’ to describe the tapering pits, shaped like the flower of a lily, with a sharpened stake at the bottom, and the use of ‘spurs’ for the iron hooks designed to catch peoples feet. As Paul Bidwell noted the expressions only appear in this context and nowhere else.

There is another interesting passage about fort building from earlier in the campaign;

IX … … … … He ordered his camp to be fortified with a rampart twelve feet high, with breast-works built on it proportioned to its height; and two trenches, each fifteen feet broad, with perpendicular sides to be sunk: likewise several turrets, three stories high, to be raised, with a communication to each other by galleries laid across and covered over; which should be guarded in front by small parapets of osiers; that the enemy might be repulsed by two rows of soldiers. [osiers are small branches , like willow]The one of whom, being more secure from danger by their height, might throw their darts with more daring and to a greater distance; the other, which was nearer the enemy, being stationed on the rampart, would be protected by their galleries from darts falling on their heads. At the entrance he erected gates and turrets of a considerable height.

De Bello Gallico IX

This passage emphasises the importance of height, and the advantages of having more than one fighting platform, to concentrate firepower at the point of attack, as well as demonstrating the practice of protecting defenders from aerial bombardment. These structures were probably defencible from both sides.
Another important source of information on Roman military engineering is Trajan’s column in Rome, which depicts his two victorious campaigns against Dacians, (101/2 &105/6 ce), [above left]. One scene shows a ballista set on a structure made from layers of logs laid at right angles to each other, while another shows legionaries cutting and extracting timber baulks.
The Obstacle theory

Paul Bidwell’s theory that these post pits once held pointed pieces of wood is much easier to draw than to realise, and has several disadvantages as an explanation.
Generally, trees are grown to be straight, in conditions which discourage branching; in prehistory most postholes are 0.20-0.40m diameter, suggesting, if oak was used, trees were harvested at c. 50- 80 years. Close spaced groups of branches, may occur at the crown, [9], which is generally thrown away, however for the crown of a tree to be 0.20 –0.40m in diameter, the basal diameter would have to be quite significant greater, certainly larger than those that are represented in postholes of the period.

Another objection is practicality; once this entanglement was in place, how was it maintained?

It would rapidly become overgrown with weeds, and soon start to develop the characteristics of a hedge, and would end up providing cover for attackers.

An artists impression of the 'cippi' style wooden entanglements curving towards the turret at at Throckley , [13].

My third objection is on grounds of strategy; entanglements are designed to slow down attackers, and concentrate them into a field of fire in a battlefield context. As Caesar described, obstacles can be overcome by counter measures, (except by cavalry, which is unlikely to be attacking a wall). The most practical method of attacking a wall is with scaling ladders, which is one of the principle reasons for having a ditch(es), which is why the position of the Hadrian’s Wall Ditch is so odd, too far from the wall to be effective, especially with its normal, expected position, apparently usurped by a wooden entanglement.

The Buddle Street road mystery.
One of the most puzzling aspects of the Buddle St plan is that rampart starts on the east of the site, but is missing from the centre and west of the site, which is covered by a motley collection of postholes and other features. While I do not intend to pursue this idea very far, an obvious explanation occurs that explains most of the postholes to some degree, Which is what we do here!

An interpretive plan of the archaeology at Buddle St, Wallsend

There are a series of ditches and gullies, crossing the site roughly N-S, but about 85° to the line of the wall, the spacing of these gullies about 90-100’ [pes] apart with a central pair about 40’ apart, together with some areas of possible surfacing, is suggestive of a Roman road crossing the site. This being the case, then the road would have lead to the facilities on the riverbank prior to the construction of the wall, running north along the east side of Wallsend Golf Course, where its line may have been preserved in the landscape.

Sketch plan of the possible tower at Buddle Street

When the timber phase was constructed, the road was still in place, and the rampart starts from edge of the central carriageway. To the west is a series of postholes, mostly aligned on the road, but while respecting the line of back of the Timber Rampart. My initial interpretation of these features is that they represent a gateway structure constructed across the central carriageway with a tower and further barrier extending to the ditch at the west edge of the road. This ditch may have been re-cut, 4’ to the west at this stage, the ditch to the east appears to have been diverted through the timber wall, also moving it to the west by about 4’. The structures built across the road are quite complex, and very probably ad hoc, and require more detailed study in their own rite, and are on the long list for further study.

The approximate position of the possible road crossing the Buddle St site

If the road was still in use, and required a gateway, then the ditch must also had a gap in it at this point, and this may be reflected in the differing profile of the ditch at this point. By the time the wall was built here, the road would have been realigned on the Segudunum Fort just to the east.

Theoretical Structural Archaeology

Postholes, although often little regarded, are the key to understanding much of the archaeological built environment, and it’s best not to ignore them, or explain them away – they may be filled voids now, but they are called postholes because they once held posts, and therein lies their utility – as Lao Tzu might have said.
It is important to stress that as a researcher specializing in postholes, I am not necessarily that concerned with the broader implications of this, or any other, of my theoretical models. Evidence for posts can occur in a variety of periods and contexts, and I am in reality no more a Roman specialist than someone studying the carbonised remained of wood would be. I study the commonest of all features, postholes, and the structures they represent, which is not even regarded as a specialist subject, and I am probably the only archaeologist in the country studying these features from a structural as opposed to a structuralist perspective.

Conclusions

The construction of the Hadrian’s Wall begun in 122ce was certainly one of the more visible episodes in the Roman occupation, and a significant achievement, whose shear physical presence could easily obscure the more transitory aspects of frontier. The discoveryof the presence of the timber wall, though entirely logical, feasible, and consistent with the evidence, represents quite a big and surprising shift in the visual culture of the Roman Wall.
The model suggested for Buddle street works reasonable well for the other sites, but given the flexibility inherent in the design, and the need to adapt the structure to changes in slope, features such as streams, and differing ground conditions, combined with likelihood of towers, stairs, sally ports, and other ancillary features, we cannot be too prescriptive about the model, or its application. The detail of its construction, its height, number of galleries, disposition of towers was probably quite variable, and cannot be known with any degree of certainty.
I consider the general feasibility of a timber faced box ramparts built predominantly in alternating, and probably overlapping, layers of standard baulks to be proven as a good match for the evidence. The timber requirements are not excessive, materials for flooring, roofing, parapets, etc, would have come from the smaller trees of the woodland under story like hazel. In the context of establishing a secure military corridor across northern Britain, one would envisage the clearance of a wide strip of land, with a road, and defendable boundaries of both sides; several hundred feet to south of the Timber rampart, was the ditch and earth rampart system known as the Vallum. This defendable military corridor would help protect the workforce, materials, supplies, grazing animals, and stock, from the potentially hostile native population, whose former territory was being effectively split into two zones.

While the context of Caesar’s siege of Alésia was quite different, and the distinction between defender and attacker almost irrelevant, the military response was in many ways similar, in that a secure perimeter was established using ditch digging and rampart building to allow the main strategy to develop. The real skill in military engineering lies more in placement of defenses than in their actual building, which is usually simply a matter of logistics. Caesar claims that the first 18km was built in 3 weeks, he then had to do the same thing again, adding an addition 21km of contravallation.

How long would it take it build a rampart 117 km long? - about 20 weeks using Caesar's army - but of course it’s just not that simple, however, the Roman army were good at this sort of thing, it’s what they did for a living, and to some extent their lives depended on it, and creating a 117 km corridor was probably achievable within a year, especially since with static defensive structures, it’s all or nothing, you can’t have gaps, as the French learnt in 1939.

It is easy for the eye to drawn to the use of stone in Roman archaeology, but it must be remembered that virtually all military structures in Britain would have been initially built in timber, the fort at Vindolanda, where the evidence survives, had five timber phases,[15], before stone was used. Timber is quick, practical, available, and everybody can join in, since it requires very basic skills to construct this type of rampart. When the Rampart was dismantled it may have generated an estimated 5 million baulks, 10’ or so in length, so keep a lookout for those turning up in Roman structures of the period.

The existing ideas revolving around ’cippi pits’, seem to be derived from a very old joke lost in translation, but if we regard these post pits as foundations for a Timber rampart, other things fall into place; the width of the berm, and its apparent narrowing at the turrets, is explained by regarding the turrets, timber rampart, and ditch, as part of the same initial construction phase, built arround 122 ce, which, since it respects the ultimate line of the stone Wall, was clearly a temporary expedient from the outset.
When, after the probable six years or so, the Wall was complete, the timber wall, probably one of the largest wooden structures ever created, was dismantled, probably section by section, as the wall was completed; its timbers almost certainly recycled into the dozens of new forts, and, into the new stone wall, - it should not be forgotten, that without addition timber works, a simple wall and parapet, while impressive, was not a particularly defendable structure, (especially when your ditch is so far away).
So this enormous structure disappeared leaving only postholes, which, as so often happens, could have been easily overlooked, which would have been a shame, and rather a waste of about 2.5 million trees.
STOP PRESS
Latest article on Hadrian's Wall; Reverse Engineering the Vallum


Sources and further reading
[1] Bidwell, P T, 2005 'The system of obstacles on Hadrian's Wall; their extent, date and purpose', Arbeia J, 8, 53-76.
http://www.arbeiasociety.org.uk/journal.htm
[2] Caius Julius Caesar "De Bello Gallico" and Other Commentaries
English translation by W. A. MACDEVITT, introduction by THOMAS DE QUINCEY (1915) http://www.gutenberg.org/etext/10657
[3] Bidwell, Paul T.; Watson, Moira. 1989 'A Trial Excavation on Hadrian's Wall at Buddle Street, Wallsend'. Archaeologia Aeliana, 5th ser., 17 (1989), 21-28.,
[4]Grey literature: Shields Road, Newcastle, Phase 2b, archaeological excavation. TWM archaeology 10/2006
[5] T. Frain, J. McKelvey & P. Bidwell 2005 Excavations and watching brief along the berm of Hadrian’s Wall at Throckley, Newcastle upon Tyne, in 2001-2002. Arbeia J, 8
[6] Grey literature; Throckley, Newcastle upon Tyne, archaeological excavation and watching brief. TWM archaeology 12/2003
Grey literature: Newcastle, Melbourne Street, Archaeological excavations. Archaeological services, University of Durham.
http://csweb.bournemouth.ac.uk/aip/gaz2004/ene.pdf
[7] http://www.pitt.edu/~medart/menuglossary/INDEX.HTM
[8] http://www.gutenberg.org/etext/132
http://www.military-quotes.com/forum/logistics-quotes-t511.html
[9] See; http://structuralarchaeology.blogspot.com/2009/01/17-not-seeing-wood-or-trees.html
& http://structuralarchaeology.blogspot.com/2009/02/22-iron-age-graphs-important-discovery.html
[10] James. 1989, Forester's Companion -Cambridge University Press, ISBN 0631108114
[11] After; http://en.wikipedia.org/wiki/File:SiegeAlesia.png
[12] Cip´pus n. 1. A small, low pillar, square or round, commonly having an inscription, used by the ancients for various purposes, as for indicating the distances of places, for a landmark, for sepulchral inscriptions, etc.
Webster's Revised Unabridged Dictionary, published 1913 by C. & G. Merriam Co.
[13] After Fig 15 p73 Bidwell, P T, 2005 'The system of obstacles on Hadrian's Wall; their extent, date and purpose', Arbeia J, 8, 53-76.
[14] See Also: Carter, G. A. 1998: Excavations at the Orsett ‘Cock’ enclosure, Essex, 1976. East Anglian Archaeology Report No 86.
[15] Burley R 2009, Vindolanda A Roman fort on Hadrians Wall. Amberley ISBN978-1-84868-210-8