The construction of Hadrian's First Wall

It is now over two years since I published the analyses of the three lines of double postholes found on the berm north of Hadrian’s Wall.[1] I knew then that this was the foundation of a temporary timber rampart, and that these least visible of its features were the key to explaining Europe’s largest archaeological monument. However, more research has subsequently clarified the issue further, and understanding the engineering of the ‘Vallum’ behind the Wall has thrown further light on the sequence of construction.

Previous articles;
The Timber Wall
The Vallum
The Northern Frontier

So it is time to summarise these ideas, refine the models, and see how they fit into the early constructional history of Hadrian’s Wall.

The elephants in the room

After over a century of scholarship, hundreds of excavations, surveys, and the detail consideration of everything from individual inscriptions to thousands of sherds of pottery, there still remained one central, but subtle, conundrum: if you needed a continuous fortified frontier, and given it would only be effective once it was finished, how come it took about 10 years to complete?

Since there is now little doubt that this was a war zone[2], breaking your army into small work details and spreading them along a 120km front could be regarded as tactically unwise, if not downright reckless. Undertaking such a massive construction project might imply that the frontier was peaceful and secure, which tends to undermine the perceived motivation behind building the Wall.

The construction of a temporary Wall solves this problem, securing the frontier against infiltration in the first season. The Timber wall built in the East provides a context and explanation for the Turf Wall, this being the form of the temporary rampart in the west.

Similarly, the various explanations put forward to explain the Vallum, the earthwork behind the Wall, have also required a degree of special pleading; to say it is a boundary, as most accounts do, is to overlook the fact that this is not how Romans, or anyone else, ever built boundaries.[3]

It is, however, how they built roads, and as part of the initial plan for the frontier, it makes sense, as does the abandonment of the project as the Wall was scaled back following a ‘dislocation’.[4]

This is a key understanding; the dislocation[s] marks a distinct break in the Wall’s construction, and it is now thought that this was disrupted by serious warfare, perhaps on more than one occasion. When work resumes, the plans for the wall are scaled back. The ‘Broad’ Wall under construction is superseded by a ‘Narrow’ Wall, and there is a marked lowering of standards in the quality of the stonework.[4]

So we will briefly look at the model of the temporary rampart phase of the Wall, and at the nature of the Vallum, before seeing how they fit into the initial construction sequence.

The Timber Wall: Features of theoretical model

The theoretical model describes the general form of the rampart, and is based on the precise position of its posts, and on the principle that military engineering has to be credible and fit for purpose. In addition, we have some invaluable descriptions from Caesar’s account of his Gallic wars, which explain the general form and thinking behind this type of military engineering.

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….

De Bello Gallico LXXII

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 used for wattle]

De Bello Gallico IX [5]

When I first did the analysis, while it was obvious that that the posts were laid out to accommodate 60° ‘bracing’, I had not come across this reference to how the Roman army built ramparts in Polybius’s description of the Roman general Flamininus's campaign in Thessaly and the Battle of Cynoscephalae in 197 BC.

18. Flamininus . . . . ordered all his soldiers to cut stakes for a palisade to carry with them for use when required. This appears to be impossible when the Greek usage is followed, but on the Roman system it is easy to cut them. For the Greeks have difficulty in holding only their pikes when on the march and in supporting the fatigue caused by their weight, but the Romans, hanging their long shields from their shoulders by leather straps and only holding their javelins in their hands, can manage to carry the stakes besides.
Also the stakes are quite different. For the Greeks consider that stake the best which has the most and the stoutest offshoots all round the main stem, while the stakes of the Romans have but two or three, or at the most four strange lateral prongs, and these all on one side and not alternating. The result of this is that they are quite easy to carry — for one man can carry three or four, making a bundle of them, and when put to use they are much more secure.
For the Greek stakes, when planted round the camp, are in the first place easily pulled up; since when the portion of a stake that holds fast closely pressed by the earth is only one, and the offshoots from it are many and large, and when two or three men catch hold of the same stake by its lateral branches, it is easily pulled up.
Upon this an entrance is at once created owing to its size, and the ones next to it are loosened, because in such a palisade the stakes are intertwined and criss-crossed in few places.
With the Romans it is the reverse; for in planting them they so intertwine them that it is not easy to see to which of the branches, the lower ends of which are driven into the ground, the lateral prongs belong, nor to which prongs the branches belong. So, as these prongs are close together and adhere to each other, and as their points are carefully sharpened, it is not easy to pass one's hand through and grasp the stake, nor if one does get hold of it, is it easy to pull it up, as in the first place the power of resistance derived from the earth by all the portions open to attack is almost absolute, and next because a man who pulls at one prong is obliged to lift up numerous other stakes which give simultaneously under the strain owing to the way they are intertwined, and it is not at all probable that two or three men will get hold of the same stake.
But if by main force a man succeeds in pulling up one or two, the gap is scarcely observable. Therefore, as the advantages of this kind of palisade are very great, the stakes being easy to find and easy to carry and the whole being more secure and more durable when constructed, it is evident that if any Roman military contrivance is worthy of our imitation and adoption this one certainly is, in my own humble opinion at least.

Polybius Histories 18 [6]

This explains how the body of the rampart could be assembled, and accounts for the layout of the posts, where the 60° alignments appear to take precedence over 90° arrangements that might be expected in a traditional ‘box’ rampart. Surprising as it may seem, there is no reason in theory why some small gaps or loopholes could not be left in the body of the rampart, allowing for more concentrated fire from the defenders.

Theoretical model

The Timber Wall is formed by pairs of posts set holes, which:

• Form the foundations of the wall;
• Hold the horizontal components in position;
• Support the fighting platforms.
• Two sets of posts allows for two fighting platforms at different heights.

The body of the rampart is formed by horizontal timber baulks [valli]

• Thick end at the back, thin end pointed to the front;
• Made from standard lengths of timber 10 –12’;
• Stubs of branches help interlock the baulk.

Baulks may be placed:

• At an angle [60° to axis];
• Running across the structure [90° to axis];
• Parallel to the axis behind or in font the lines of posts.

Design Features:

• Simple design;
• Standard components;
• Simple to fabricate;
• Flexible, offering multiple configurations;
• Very strong interlocked and braced structure;
• Difficult to penetrate;
• Individual components difficult to remove;
• Can be dismantled and reused.

The Turf Wall

The frontier in the East had been moved north into previously unoccupied territory, while in the West it lay close to the Stanegate frontier that had existed for some forty years. It is reasonable to suppose that during this period most of the readily available timber had been used in the construction of timber forts like Vindolanda. Thus it was probably a lack of suitable timber that prompted the change of design.

The Turf wall survives only as a mound of earth, which appears to be built with turf stripped from the surrounding area. I would suggest that the structure might have been reinforced and stabilised with layers of brushwood or even timber.

No evidence of the nature of the upper part of the rampart has been found. The model speculates that the breastwork sat on top of the turf wall and would have been structured in a similar manner to the Timber Wall. There is still the requirement for credible fighting platforms and some form of parapet to prevent climbers.

We have to assume it, like all sections of this temporary wall, was built in the expectation that it would be attacked, and it has the same basic layout with turrets and milecastles as the stone Wall that superseded it. Unlike its Timber equivalent, the Turf Wall could not be easily dismantled, so its replacement in stone was not as straightforward.

The Vallum

The Vallum is simply the construction trench for a road that was never finished. This is the only explanation that fits the observable facts of the archaeology. It is comprised of a vertical or steep-sided, flat-bottomed trench where spoil has been moved 9m and piled in revetted spoil heaps to allow for the construction of a central metalled carriageway, with softer lanes on either side. This is the normal pattern for a road, and the only conceivable rationale for moving the spoil from the trench such a distance, and explains why a steep-sided, flat-bottomed cutting was created in soft and unstable ground.

In addition, the following aspects of its course confirm this conclusion:

• Laid out in straight lengths;
• Gentle corners;
• Avoids steep or uneven gradients;
• Avoids soft ground;
• Where it had to cross soft ground, the central trench is formed between earth banks;
• Follows close behind the wall and forts;
• Starts from a bridgehead at Newcastle;
• It has been suggested there were originally gaps in the northern spoil mound corresponding to milecastles.

Its abandonment may have saved up to c.40% in bulk materials, logistics, and labour. More importantly, the scaling back and change of emphasis evident in other aspects of the wall following the dislocation provides both an ideal context and an appropriate time frame.

Construction sequence

I have decided to use the word ‘step’ rather than ‘phase’, because I am describing one aspect of a more complex process. It is clear that, in accordance with normal logistical principles, the longest, more complex tasks, in particular those requiring the most skilled labour, were prioritised. In terms of the stonework, this was the milecastles. Thus, once construction of the Wall proper begins, things get quite complicated, with separate construction gangs working simultaneously on different aspects of the project.

We also have to be aware that there are forts to build, and all manner of sundry logistical tasks, not to mention the day to day garrisoning and guarding of the frontier.

However, it is possible to see logical steps in any major project of military engineering, and where the evidence is available, these are confirmed by the archaeology.

Establishing the linear frontier

Step 1 Plan/survey/lay out Route

Step 2 Clear route and establish construction track [Military Way]

Step 3 Build Temporary Timber and Turf Walls

Step 4 Dig Ditch

Main construction phase [working from East to West, and not completed]

Step 5 Dig Vallum Trench dug and build spoil mounds

Step 6 Dig Broad Wall foundation

Step 7 Build Broad Wall

[Step 8 Dismantle Timber Wall]

Dislocation: Warfare disrupts the process, perhaps more than once.

Post-Dislocation

• Vallum Road abandoned and in places backfilled;
• Wall completed as Narrow Wall.

Construction sequence - key points

• The presence of the temporary Wall explains why the majority of the spoil from the ditch was thrown north, with only a small glacis bank to the south.
• The Vallum was completed [Step 5], while foundation trench for the Wall [Step 6], was not.
• Digging of a road foundation suggests there was plenty of less-skilled labour available early in the project, and emphasises the importance of a proper road in the overall scheme.

Conclusions

... the Britons are unprotected by armour (?). There are very many cavalry. The cavalry do not use swords nor do the wretched Britons mount in order to throw javelins.'

Vindolanda Tablets [7]

As primarily an army of heavy infantry, the Romans used earth and timber ramparts to protect themselves, particularly from cavalry. Fixed installations ensure the enemy dismounts and fights on terms and ground chosen to suit the defenders.

The initial and temporary rampart represented by the Timber and Turf Walls provides a rational and secure basis for the much more complex and specialised task of building the Stone Wall. We can assume on the basis of the Turf Wall that it had some form of turrets and milecastles, and that therefore it was fully garrisoned and in every way fit for its purpose.

The rapid building of a fully functional frontier, probably within a year, was a strategic fait accompli that would have materially affected the military and political situation in the North of England. The arrangements and treaties that had ended the prior conflict, and prompted the building of a permanent frontier, may have been undermined by this unprecedented, and probably unexpected, move.

We cannot know if the appearance of this initial rampart was a causal factor in the warfare represented in the dislocation[s] in the construction, but its outbreak indicates that this cautionary approach was justified. Similarly, the building of additional forts along the Wall, when work resumed, also suggests that it had proved ineffective, emphasising the very real threat posed by incursion from the north.

While the scaling down of the size of the Wall could be regarded as an attempt to finish it quickly, the decline in the quality of the masonry may also indicate serious casualties among the skilled legionary workforce.

Once the Vallum is accepted as an unfinished road, it not only makes perfect sense of the archaeology, it also adds considerably to our understanding of the initial plan. Not going ahead to backfill the Vallum trench with suitable bedding material and cap it with a cut stone surface saved a huge amount of labour and material resources and is another indication of just how seriously warfare had impacted the project.

Hadrian’s Wall was soon replaced by the Antonine Wall, which suggests it wasn't a great strategic success. However, the building of Hadrian’s frontier was an extraordinary achievement, not least because in its initial form it was probably built in less than a year. In addition, the growing realisation that this period was punctuated by serious warfare emphasises both the difficulty of the task, as well as why it was felt necessary in the first place.

“... under the rule of your grandfather Hadrian what a number of soldiers were killed by the Jews, what a number by the Britons”

Marcus Cornelius Fronto, letter to Marcus Aurelius, AD162

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.
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.

Grey literature: Shields Road, Newcastle, Phase 2b, archaeological excavation. TWM archaeology 10/2006
[2] Breeze, D.J. 2003. "Warfare in Britain and the Building of Hadrian's Wall." Archaeologia Aeliana 32, 13 –16.

[3] For a recent account of the Vallum, see Wilmott, T. [ed]. 2009. Hadrian's Wall: Archaeological Research by English Heritage 1976-2000 http://www.english-heritageshop.org.uk/mall/productpage.cfm/EnglishHeritage/_51324/288647/Hadrian's%20Wall[Accessed 29/11/10]

[4] Hill, P. R. 2006. The construction of Hadrian's Wall. Tempus.

[5] [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/ebooks/10657 [Accessed 18/12/11]

[6] http://penelope.uchicago.edu/Thayer/E/Roman/Texts/Polybius/18*.html#18 [Accessed 25/03/2011]

[7] http://vindolanda.csad.ox.ac.uk/exhibition/army-2.shtml [Accessed 25/03/2011]