Hadrian’s First Wall [Part 2 of 3]

On Tyneside, between Hadrian’s Wall and the Ditch to the north, archaeologists have found three lines of double postholes, which it is argued, represent an early Timber Wall, which, along with the Ditch, formed part of a temporary frontier while Hadrian’s Wall was being built. Further, it is argued that the Turf Wall represents the continuation of this structure in the western sector of the Wall.  In addition, when the engineering and layout of the Vallum is examined, it appears to be an unfinished road, probably abandoned when warfare interrupted work on the Wall. These insights into the archaeology of Roman military engineering are the key to a new understanding how and why Hadrian’s Wall was built.

An updated  summery of a series of articles from this site on the timber and earth structures predating Hadrian’s stone Wall.

Presented in three parts:

1. The Timber wall

2. Reverse engineering the Vallum

3. The Construction of the First Wall

2. Reverse engineering the Vallum

The Vallum is a unique linear earthwork to the south of Hadrian’s Wall. When it was constructed, it ran in an unbroken line, following close to the course of the Wall, from Newcastle to Bowness-on-Solway, a distance of about 112km.

Schematic section of the main features of the Vallum and Wall

It comprises a flat-bottom ditch, roughly 6m wide and up to 3m deep, flanked by parallel mounds, set back about 9m from its edge. These north and south mounds were built of spoil from the ditch and were usually revetted with turf, and occasionally stones. There is often another small mound, known as the ‘marginal mound’, along the southern edge of the ditch.[19]

The Vallum Crossing at Condercum Fort, in Benwell, Newcastle, [20]

Where possible, the Vallum lies close to the rear of the Wall. However, unlike some sections of the Wall, it is laid out in very long straight sections, avoiding steep gradients, with a few very deliberate corners. In the famous central section where the Wall runs along the crags of the Whin Sill, the Vallum stays in the Valley.  Where it crosses very soft ground, as at White Moss in Cumbria, the ditch is not dug in, but created between mounds of material,[21] The ditch is uninterrupted except for causeways that have been found opposite the gates of the main Wall forts, and perhaps some milecastles.

Dating The Vallum

Work on Hadrian’s Wall probably began in AD 122, and the Vallum carefully skirts a series of forts, like Benwell and Birdoswald, which were added early in the construction process, indicating it was built at the same time or shortly after them.

The Vallum overlain by Fort at Carrawburgh

The defences at Birdoswald soon encroached on the Vallum, [22].  One of the later additions, the fort at Carrawbrough [Brocolitia], probably dating to AD 132-3, was built over the top of the Vallum.[22] This is consistent with a general pattern of backfilling and slighting of the ditch evident from this sort of period onwards. In places, the mounds are breached and the ditch backfilled at regular intervals of about 41m. [15]

The Vallum followed the Turf west of Birdoswald Fort closely. When this Wall was replaced in stone on a new course further north, probably late in Hadrian’s reign, there was no attempt to move the Vallum.

Plan of the Vallum overlain by the ditches of Birdoswald Fort [23]

Thus, the Vallum was maintained only for a short period, probably less than 10 years, after which it was no longer considered important, or worth preserving, in its original continuous form. Hadrian died in AD 138 and his successor, Antoninus Pius, had a new strategy. He moved the frontier north, and starting in A142, built the Antonine Wall across the Forth-Clyde isthmus.

Profiles and archaeological sections of the Vallum [24]

What was the Vallum for?
The Vallum is unique, which is always a problem to archaeology, where insights so often come from comparing things. While most commentators have draw attention to the road-like nature of its layout, and noted its possible function in communications, they have sought its true function elsewhere.  Little has changed in our understanding of this earthwork, as Tony Wilmott, a leading English Heritage archaeologist and Wall specialist, made clear recently:

“The straight lengths in which the Vallum is laid out are consistent with distances of the uninterrupted view of a surveyor. Changes in direction tend to occur where a new viewpoint is required to obtain another long straight view. This is the system used to lay out roads. It has been argued that as the Vallum was surveyed as a road would be, it must have functioned as a road. This does not follow, and if we accept that a road would require metalling, and metalling on the berms is sporadic, and sometimes doubtful, we have to discount all of the suggested variations of a function related to lateral communication.Richmond’s (1930) statement of the function of the Vallum remains valid:

‘the Vallum takes its place as a prohibited zone delimiting the south side of the military area, an unmistakable belt in which an obstacle is provided by the great ditch. Neither commerce nor interference with the soldiery could take place across it unchecked.’ ”

Tony Wilmott 2007 [25]

“ It was not a strictly military defensive system, for the ditch is not of military shape, nor are the symmetrical banks defensive; furthermore, the Vallum takes no account of commanding ground. Rather it is a barrier and line of demarcation defining the rear of the Wall-zone, and preventing entry except at fixed points.”

Sheppard Frere, 1974 [26]

“The Vallum allowed army units to segregate themselves from civilians and to use the zone between the two barriers for grazing military pack animals and cavalry horses.”

David Mattingly 2006 [27]

The Vallum at Down Hill on Google Earth

While this is all perfectly reasonable, it could be argued that these functions were a consequence of the Vallum’s construction, and not necessarily its intended purpose. Besides, all of this could have been achieved with significantly less effort. This is what introduces a note of an ambiguity; the Vallum is an odd and, above all, inefficient engineering solution. I have an expectation that Roman military engineering would be efficient, and built to last, like their roads.  Whatever potential benefits an almost-continuous earthwork may have initially provided, its short life and systematic backfilling suggest these were rapidly outweighed by the disadvantages.  However, there is a simple, and less ambiguous explanation for the creation of the Vallum: It is a construction trench for a road that was never completed.

The Vallum as an unfinished road

Quite simply, if a flat bottomed trench, set out like a road, is backfilled with appropriate material and capped with stone, you would create a carriage way. The space between this and the mounds on either side would be the lanes used by riders and pedestrians. It is entirely rational and appropriate that Hadrian would have specified a high quality road running along his new frontier, linking the main military establishments.

The Vallum West of Carrawbrough

Viewing the Vallum as an abandoned road project fits with a variety of observations have led to the suggestion that there was war in Britain during the 120s [18], this dislocation disrupted work on the Wall, resulting in downscaling of the project.

The Wall was slightly further north than the previous local frontier, a line of forts linked by a road known as the Stanegate. This had probably served as the northern limit of Roman territory before Agricola’s campaigns in Scotland. While the Stanegate was no doubt still used, and some of its forts, like Corbridge, were still important, the military geography had changed.

[left] Stanegate South East of Carvoran, south of Dintly wood

The road that was eventually built behind the Wall, called the 'Military Way', is not very substantial, and has been associated with the later reoccupation of the Wall. The Military Way runs between the Wall and the Vallum and, interestingly, in places, along the top of the Vallum’s north mound. Traces of metaled surfaces have been noted on the north berm and south berm of the ditch.

The practice of Roman road building

We have no technical sources for Roman road construction, despite a much-repeated idea that Vitruvius describes the process. [28] Our only literary source is a poem by Statius praising the Via Domitiana, built by Domitian in AD 95:

The first task here is to trace furrows, ripping up the maze of paths, and then excavate a deep trench in the ground.
The second comprises refilling the trench with other material to make a foundation for the road build-up.
The ground must not give way nor must bedrock or base be at all unreliable when the paving stones are trodden.
Next the road metalling is held in place on both sides by kerbing and numerous wedges.

Publius Papinius Statius c. AD 95 [29]

This poem was written to flatter the Emperor Domitian, which makes another important point: The quality of a road or other civil engineering project was a reflection of the person who initiated and sponsored it. A Roman politician’s prestige and standing was, in part, a reflection of the nature and quality of the engineering of the ‘good works’ that bore his name.

It was standard practice to place inscriptions on Roman buildings and other works indicating who had built them. This was particularly true of the army, and from their inscriptions, it is evident that the masonry aspects of the Wall system were primarily built by troops from three legions [legio II Augusta, legio XX Valeria Victrix and legio VI Victrix]. [16]

A total of ten fragments of inscribed stones have been found in the vicinity of the Vallum, and this has led to the suggestion that they relate to the construction of the Vallum. One group from near Denton mentions an auxiliary unit.[32] Unlike the Wall, where such inscriptions are more common, the Vallum, primarily a trench with spoil heaps, offers little opportunity for the placement of permanent inscriptions. Certainly, if the Vallum is an unfinished road, then the credit would have lain in finishing and surfacing it.

The basic form of a Roman road comprised a central carriageway with a hard surface, primarily for wheeled traffic, flanked by wide lanes of softer ground for horses and pedestrians. Ideally, the central carriageway would have a road surface of interlocking stone blocks set in mortar, with a convex profile to encourage drainage. This would normally be built on a roadbed made up from successive layers of graded material.  

A simplified section of a main urban street from Pompeii Via munita [30]

The roadbed was built up as a mound or agger, or set in a foundation trench or fossa, depending on ground conditions. Foundation trenches have flat bottoms to ensure even distribution of the load. They were usually dug with steep or vertical sides because long-term stability is not an issue in a trench dug to be rapidly backfilled. The Vallum’s form is consistent with its being a construction trench. Considerable effort has been taken to move the spoil away from the trench, placing the heaps to ensure wide margins on either side. The spoil heaps are neat and carefully constructed to stop their spreading onto these areas next to the trench.

A Roman road with agger [31]

In short, it is not the mounds themselves, but the areas between them and the trench, which are important. Similarly, the trench itself is a void that only becomes functionally significant when, and if, it is backfilled.

Digging the Vallum

I do not intend to look at the logistics in detail, and they can only be seen in the wider context of building the frontier; this has been well covered by Peter Hill.[16]  Using the classical model for the Vallum, [left] laid out in Roman feet, he suggests 1,465,313 m³ of material were excavated and moved to the North and South mounds. He points also out that more labour was required to move the spoil than was expended in digging the trench.

However, if the trench was going to ever to become a road, a similar volume of material was going to have to found for the backfill. Simplifying the maths, one and a half million m³ of backfill would weigh at least three million tonnes, probably more -- a perfectly good reason why the scheme was never completed.  To help visualise these figures; the large 8-wheeled lorries used to move this sort of material on our roads take about 20 tonnes; to finish the job you would require 150,000 lorry loads of hardcore, sand, gravel, lime and surfacing stones. In the absence of lorries, the Romans had the Army. However, digging a foundation trench, and even backfilling it with hardcore, requires considerably less skill and specialised resources than the construction of walls using lime mortar and dressed masonry.

The Wall complex at Bogle Hole

So in terms of the wider picture, the Vallum may have not have used legionary manpower, although this might be required for surfacing and finishing a road.  If that was the case, its construction, thus far, may have involved auxiliary troops. In addition, we cannot preclude the use of prisoners of war, convicts, levies, and slaves for some of these basic tasks, freeing the legionaries for more specialised construction tasks.

The ditch would be dug using mattocks, which were standard issue to the Roman army, but there are three basic possibilities for removing the spoil:

• Men with baskets;
• Animals, probably asses/donkeys with baskets [panniers];
• Carts or wagons, probably pulled by asses or oxen.

Although the latter is probably the most efficient, the actual method could be a combination of all three. While ard marks from ploughing, hoof prints, and even cart marks [16] have all been been found under and around the Vallum mounds on occasions, I am not aware of any archaeological work directed at this question.
Since the ground along the steep sides of the trench is potentially unstable, the spoil can only be extracted from either in front, or behind, the excavation, and thus, two possible methods of construction suggest themselves:

Two models for spoil extraction from the Vallum

Forward extraction: The trench is excavated with the spoil being lifted forward and upward to the carts or baskets, and from there to the spoil heaps being constructed in parallel with the trench. The disadvantage of this method is that all the spoil has to be lifted from the excavation to the bed of the cart, or into baskets, a task that becomes more difficult as the excavation gets deeper.
Reverse extraction: In this method, the spoil is loaded downwards into carts or baskets using the trench to take the spoil out backwards. If the trench face is reduced in spots, all but the lowest can be easily loaded into a cart or baskets, and potentially, the draught animals do the hard work of lifting the material by means of a ramp at the start of a section of the works. In addition, by dumping the material ahead of the trench, it is possible that as the excavation goes forward, the spoil heap moves backward, and so the distance between the two and the time taken to move the spoil remains fairly constant.
The first method relies on human power to lift the spoil from the trench, whereas the second uses draught power for part of this the task, but would move the spoil a greater distance and requires the transport to able to turn in the trench.
The Marginal Mound
The intermittent presence of this feature to the south of the Vallum ditch has probably prompted more debate than anything else. It was once the basis of the belief that the Vallum had been recut. A connection has also been suggested between the backfilling of the Vallum to create causeways, and the presence of the marginal mound.

In a road-building exercise, the trench would have had to be used to transport the layers of backfill to form the roadbed, since it is impractical to tip from the sides. Assuming that the spoil heaps are built from their ends, once completed, there would be nowhere to put additional spoil that might be created from a final cleaning of the of the trench, or from removing access ramps. This may account for the presence of the marginal mound; there was nowhere else to put excess spoil.

However, why this seems only to occur along the southern lip of the trench has also to be accounted for. The south side is often the down-slope side, but probably not consistently enough for that to be of significance. More reasonably, its presence may indicate that the north side, next to the Wall, was already being used for communication. This is also suggested by metalled surfaces that have been observed in places on this side, and by the later use of the North Mound as an agger for the Military Way.

If marginal mounds relate to the removal of ramps from the construction trench, then these may have been regularly spaced. Could the observation that, in places, the trench is backfilled at regular intervals of 45 yards /41m [140 pes ?] be related to this?

In other words, do the places where the ditch was backfilled to create crossings reflect the lengths in which the trench had been dug and the mounds constructed, and perhaps the positioning of temporary ramps used to extract spoil during construction?

This could explain any relationship observed between the marginal mound and the crossings. If the material comes from removing temporary works, and especially if it derives from a final cleaning of the trench, then work on the construction of the roadbed was expected to begin imminently.

Rethinking the Vallum

"When I see a bird that walks like a duck and swims like a duck and quacks like a duck, I call that bird a duck."

James Whitcomb Riley (1849–1916)

If the Romans had required a road along the frontier, digging a continuous flat-bottomed foundation trench, dumping the spoil far enough away to allow for two wide lanes on either side of the main carriageway, would be a good start, especially if their was plenty of unskilled labour.   Using the same quarries as the Wall, it would then be backfill with layers of carefully graded and packed material, and capped with a mortared stone surface of interlocking stones set in mortar on a bed of concrete.

The Vallum is laid out exactly like a road; it relentlessly maintains its course, and importantly, its levels. It punches through the hard rock of the dolerite Whin Sill at Limestone Corner and Black Carts, at a point where the Ditch in front of the Wall gives up. The base of the trench tends to be as wide as possible, which is why the sides are almost vertical when cut through solid rock, an unsustainable angle in softer material.  A key observation is that the form of the ‘trench’ is maintained by building up the edges rather than digging when it crossed soft ground at White Moss, Cumbria. This form of this earthwork, in the context of Roman engineering, can only be justified as the foundation for a roadbed.  [left]

Starting from the bridgehead at Newcastle [Pons Aelius], not Wallsend, for a lot of its course the Vallum runs quite close to the Wall, with little room for any kind of 'zone'. If a boundary is required, dig a ditch; it’s cheaper, more efficient and stable. This ‘function’ could equally well simply derive from the consequences of having a continuous trench with very few crossing places running for 112km across the country.  The continuous nature of the Vallum does not last very long, probably less than a decade, after which the disadvantages of this arrangement clearly outweighed any perceived utility. ‘Restricting access’ to the wall zone may have soon amounted to ‘hampering mobility’ for the Roman military.

While the digging the trench [Fossa] would represent a massive investment of manpower, moving the spoil 30’, or more, to the mounds, expended more resources than digging the trench. Clearly, the spoil has to go somewhere, but this is a wasteful inefficient solution if the objective is to create some form of boundary. Put simply, if one ditch and a mound would do the job, why build two mounds so far from the cut?

The Vallum cannot be understood in terms of other frontiers, or even other engineering projects, because it is unique, however, as a cut for a roadbed, it becomes explicable and comparable with other aspects of the Roman road network.  It becomes part of the pattern, and has lost its ambiguity. Furthermore, it becomes another unfinished aspect of the Wall, or rather, would add weight to the argument that the project was downgraded, and finished in a hurry.

That the Vallum was a construction trench for a road that was never finished is the only explanation that fits all the observable facts of the archaeology, explaining why a steep sided flat bottom trench was dug through soft ground, and why the spoil was carefully placed such a distance from the cut. 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.

That it was to remain unfinished, an abandoned project, and a lost aspiration, within a few years, also tells something else very interesting about the building of the frontier.  The decision to place an additional fort at Carrawbrough, over the backfilled Vallum, shows that the vision, and priorities, had changed.

The Vallum and Wall North of Haltwistle Golf club

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.  Regarding the Vallum as a road in-the-making fits into a wider picture that is beginning to emerge, suggesting that the period of the Wall’s construction was marked by political instability and warfare.

Part 3; The construction of Hadrian's First Wall here