British Archaeology banner

Cover of British Archaeology 120

Issue 120

Sep / Oct 2011



All the latest archaeology news from around the country

CBA Correspondent

Mike Heyworth plans for the future


Your views and responses


THE BIG DIG: Roman roads

Digging up the Picts

Sacred waters

Manga at the museum – NOT ONLINE

7 societies

Stones and names

Win or lose


ISSN 1357-4442

Editor Mike Pitts


THE BIG DIG/COVER STORY: The Roman road that was

Big Dig Trowel

"The rolling English drunkard", wrote GK Chesterton, "made the rolling English road". Then along came Roman engineers to straighten things out with martial efficiency. Thus one of the great premises of British history and popular culture. But it may be quite wrong. Tim Malim and Laurence Hayes explain.

For many archaeologists, "Is it Roman?" is an all too familiar question from an intrigued member of the public, seeing anything from a bronze age barrow to an Anglo-Saxon cemetery. Some of our most enduring and iconic remains are Roman. Roman roads epitomise an epoch in our history, with traces which can be seen in the landscape today – modern roads often still follow their alignments. There is therefore a common perception that one of the things that the Romans did for us, was to create our road system.

Roman roads are popularly imagined to head relentlessly straight to their destination. They typically consist of a consolidated embanked and cambered core of earth, puddled chalk or stones (the agger), surfaced with compacted stone or gravel. This roadway is set within a wider zone often defined by boundary ditches, and sometimes containing further drainage ditches, or trenches from which material for the agger was dug (borrow pits).

Quarry Map

From the air the Sharpstone Hill quarry shows as a grey stripe in the centre, due to expand into the area outlined in red. An old route called the Portway followed the north-east/south-west line of the hill, and a known Roman road crossed it from the south-east

There is no doubt that during the initial military campaigns the Roman army built a road infrastructure that serviced its bases and the administrative centres of the embryonic province. Prior to the conquest in AD43, however, there is also no dispute that there was a communications system of ancient routeways, of which the Icknield Way (Ridgeway) and Jurassic Way are two examples. These long distance routes would have been joined by a network of local tracks which linked settlements, farms and communal centres, and served economic needs such as droveways for moving sheep and cattle; whilst additional paths would have developed organically over the prehistoric landscape.

Archaeological evidence shows that well planned structures were also being built from neolithic times, through wet areas at least, such as the Sweet Track in the Somerset Levels, made with timbers cut down around 3807–3806BC and itself laid over an earlier track. Increasingly sophisticated structures of timber and imported stone are found during the bronze age, as at Eton Rowing Lake (Berkshire), Fiskerton (Lincolnshire), Fengate/Flag Fen (Cambridgeshire), and several examples in the Thames estuary. Even gridded and metalled (gravelled) streets have been found in the later iron age tribal centres at Danebury and Silchester (Hampshire) from around 400BC.

However, carefully surveyed and engineered all-weather rural roads have been considered the prerogative of Roman technology. We can now show that this is a misconception. In summer 2009 we had the opportunity to excavate and record a long stretch of a known Roman road 4km south of Shrewsbury, near the village of Bayston Hill in Shropshire. The investigations were required as a condition on consent for expansion of Tarmac's hard conglomerate, greywacke sandstone quarry at Sharpstone Hill – whose output, rather appropriately, is used for hardwearing surfaces for modern roads and Grand Prix tracks. Unusual site conditions and an intensive programme of scientific dating revealed the ancient road to have an unexpected history.

As well as this road, which crosses the hill, a routeway (sometimes called the Portway) is believed to have followed the ridge, perhaps connecting the hillforts at Haughmond Hill and Ebury to the north-east with the Burgs at Bayston, before continuing south-west to the Long Mynd plateau and beyond. From the ridge at Sharpstone long southward views allow other hillforts to be seen, such as Caer Caradoc at Church Stretton, and to the north the area covered by modern Shrewsbury is clearly visible. To the east the hulking mass of the Wrekin and its hillfort rises spectacularly out of the midland plain, and in this direction also lies Wroxeter Roman town, on the eastern side of the river Severn.

A Roman road has long been described as running north-west from the Severn at Wroxeter, before heading west and south-west to Caersws in Wales (Margary route 64 – see box on page 20). Three historic parishes met in the area where this road crossed the Portway at Sharpstone Hill. This was the focus of our investigations, where parts of a 400m length of road were visible as a ridge running through the fields.

Initial removal of the overburden in a 150m wide zone, showed that although some of the route had been badly eroded, other parts had been deeply buried beneath colluvium, soil washed down from higher levels. We selected six areas for detailed hand excavation. This revealed that gullies and rows of stakeholes defined the edges of some (but not all) road phases. We traced parallel shallow ditches on either side of the latest phase road, as well as pits and possible field ditches. The few artefacts (a scatter of coins and Roman pottery from the latest phase, high up in the sequence) seemingly confirmed the evidence from a second century AD settlement at nearby Meole Brace, north of Sharpstone Hill (to which the road was heading), that this was all Roman.

Amongst the areas excavated by hand, one particular section of road had survived remarkably intact within a natural depression. There we recorded four phases of metalled road surface lying beneath a post-Roman trackway. They were constructed on consolidated earth and pebble cores, and cambered for drainage purposes. Beneath them all was a brushwood foundation.

The quality of this preservation allowed us to design an ambitious plan. With scientific techniques rarely used in Roman archaeology, we hoped to increase understanding of the road's history and its environmental context, and to explain the curious and apparently illogical dog-legged route that the Roman road was believed to follow. Having previously used OSL dating successfully on the ditch and rampart of the early medieval Wat's Dyke at Gobowen in Shropshire, we thought that this approach might help us with the road. Perhaps we could distinguish between a first century military structure, and succeeding phases during the second to fourth centuries, and enhance interpretation of the Wroxeter hinterland and the development of the Meole Brace settlement.

Samples were therefore taken vertically throughout the sequence, so that OSL dating of the sediments could be attempted by Jean-Luc Schwenninger and David Peat. In addition samples of the brushwood foundations, and of charcoal from pits, were sent for radiocarbon dating. Ben Gearey investigated the pre-road environmental context, and we asked Richard Macphail to examine the micromorphology of the road sequence and the inter-related colluvial deposits. Peter Marshall has completed the scientific analyses through Bayesian modelling of the combined radiocarbon and OSL dating in relation to the stratigraphic sequence, refining the chronology of the main events.

All this has delivered an unambiguous iron age date for construction of the road's first three phases, starting around 200BC, with only the possibility of the uppermost metalled surface being Roman. However, even this is much more likely to have been constructed during the late iron age, and the route's origins might have been very much earlier.

The presence of charcoal, burnt sand and stones suggests that the original ground surface was cleared of vegetation by fire. Micromorphological analysis revealed dung beetles and churning of wet mud with calcitic dung (faecal spherulites), showing initial use of the route as a track for livestock. The area around the site was apparently open grassland that was probably created or at least maintained by grazing.

A 4.5m wide layer of elder brushwood was laid over this mud. It may have been an attempt to consolidate a route over a zone that became wet, and the unbroken state of the branches suggests that earth was quickly placed over the top. This deposit showed similar micromorphological characteristics to that below the brushwood, and we have therefore interpreted it as redeposited material used as the foundation (together with the brushwood) for the first road construction. The road surface was in two layers: gravel and small stones in a matrix of silty sand beneath, and river cobbles compacted into this above. This created an all-weather roadway of hard material, some 5m wide embanked about 50cm above the surrounding ground surface in the centre, with both deposits carefully cambered down on either side to help with drainage. The downhill southern side had been kerbed by a gully which contained a row of holes for stakes, the line of a possible hurdle fence.

This pattern of construction was followed in successive phases, so that in time the road grew to over a metre high and over 7m wide, a substantial carriageway to accommodate vehicle traffic. The final phase, of late iron age or possibly Roman conquest date, appears to represent repair for wheel ruts rather than a full road rebuilding. The river cobbles that had been used for each of the road surfaces were not of local origin, and must have been imported some distance, perhaps from the Severn itself over 3km away, presumably easier than quarrying the hard gritstone of the hill.

Marshall's Bayesian modelling suggests that at 95% probability the original iron age droveway was created around 200–5BC. Successive road constructions followed at 125BC–AD35, 110BC–AD70 and 105BC–AD105, with an 82% probability that this last phase was also iron age rather than Roman. The radiocarbon dating for one pit found beneath the line of the road, and for other pits surrounding it, are bronze age (three dates with maximum range of 1740–1120BC derived from oak, ash, birch, alder and hazel charcoal). The pit beneath the road was found at the point where the three historic parishes met, and had been dug to accommodate a substantial post about 70cm across. The interpretation we place upon this, combined with the microm orphological evidence for animal dung and trampling, is that it may have been a marker post. The road's origins might thus lie in a bronze age droveway that extended over the hill, within a landscape already identified as containing occupation and funerary remains from the period.

What are the implications of this analysis?

Evidence for well engineered and carefully surveyed roads reaching back through the iron age into the bronze age, naturally raises questions about the nature of the society that planned them. Why were such roads necessary? Clearly simple tracks are sufficient for humans and animals, so purpose-built roads in Britain would have been predominantly for the benefit of wheeled traffic. Who were the specialists with the skill and knowledge to design and project manage such enterprises? How was this knowledge invested in new generations, and how were such "professionals" maintained?

Was construction a communal activity or resourced by a powerful ruler? Can we suggest, for example, a parallel with medieval rulers and view the road at Sharpstone as a highway built to enable a peripatetic lifestyle for the king's household – the powerhouse of the local Cornovii tribe – as it migrated from one tribal centre (the Wrekin) to another (Old Oswestry)? What do such roads imply for the economic activity and long-distance exchange mechanisms for the communities who built them?

At Sharpstone perhaps we have a road built for movement of heavy goods and valuable livestock between the productive farmlands of the midlands plain and the mineral-rich resources of the Shropshire and Welsh uplands. This we cannot prove, but what we can challenge in future is the bland assumption that any road that is relatively straight, built with an agger and with a cambered, compacted stone surface, must be Roman. The work at Sharpstone Hill has shown that a pre-existing road was partially incorporated by Roman surveyors into the new network of roads they created, which has helped explain the rather illogical pattern from Watling Street west past Wroxeter towards Wales. There are other known examples of Roman roads that followed earlier routes, some with iron age dates for timber structures found beneath them. A fresh analysis of such roads, examining them from a prehistoric perspective, may significantly alter perceptions of the impact of Romanisation on Britain's infrastructure.

It is also important to avoid potential misinterpretation of terminology, so that we understand the nuances that distinguish a road from a route, a track, or a highway. Finally as archaeologists we ought to apply the principles of scientific dating, using multiple samples to corroborate results and a variety of techniques to help confirm the calculated dates, as standard procedure for the vast majority of investigations. We rely too much on serendipity in finding distinctive pottery in the fill of a feature in order to date a site. We need to mature as a discipline, and instead realise the full potential of other means of dating, so that we can establish the chronology of what might otherwise appear to be unexciting remains, whose true age could be concealed by lazy assumptions.

For example if such an approach was taken throughout the country and applied to the ubiquitous field system or enclosure ditch, how might the quantity of dates assembled help to refine our understanding of the development of the prehistoric and later landscape? Perhaps every regional research framework should have two overriding research aims before focusing on period-specific themes. These would be to establish the ancient communication system, and to apply scientific dating routinely and intelligently so that we can refine our knowledge of almost all archaeological evidence, rather than relying on techniques and preconceptions that were developed 50 years or more ago!

Tim Malim is technical discipline manager for archaeology and heritage at SLR Consulting; Laurence Hayes is an associate archaeologist at SLR. The archaeological programme was commissioned by Malcolm Lawer of Tarmac, and the fieldwork was completed by SLR Consulting with Gerry Martin Associates. The analysis programme included OSL dating by the University of Oxford's Research Laboratory for Archaeology & the History of Art, radiocarbon dating by Beta Analytic, palaeoenvironmental studies by the University of Birmingham's ArchaeoEnvironmental team, micromorphological study by UCL Institute of Archaeology, and Bayesian modelling by Chronologies. Full publication will comprise volume 85 (2011) of the Transactions of the Shropshire Archaeological and Historical Society.

Roman roads: more research needed

By John Poulter

Poulter book

All my life I have been intrigued by roads, railways and canals. I grew up in Leicester, near the old Fosse Way, so this interest naturally extended to Roman roads. Like so many before me, I was enormously impressed by the directness with which these routes strode across the countryside. How did the surveyors follow such straight lines? How did they know where they should be going? The roads were built nearly 2,000 years ago, yet here they were, still with us, and with many of their lines still in use.

My interest in Roman roads has been that of an engineer (albeit my degree was not in civil engineering). On business trips between England and Scotland I would use the A68, which, north of Hadrian's Wall, follows much of the course of the Roman Dere Street. I gained the impression that part of the route there had been planned from north to south, contrary to what might have been expected. After years of rumination, I developed a methodology for determining the directions in which Roman surveyors may have been working, which I explain in detail in my book.

Roman surveyors could follow their alignments with astonishing accuracy. For instance, the overall direction taken by Dere Street appears to have been underpinned by three dead straight alignments, each of more than 20 miles (32km). Similarly, the first 12.5 miles (20km) of Stane Street, connecting London and Chichester, is exactly aligned on Roman Chichester's east gate, 55 miles away (89km).

To achieve such precision, surveyors probably marked out the line from horizon to horizon, and this explains, I believe, the tendency of Roman roads to change direction on high ground. The best views for determining the new direction would be obtained on the far brow of a hill, or (in a valley) the near bank of a river. To identify such features now, means working on the ground, not just with maps. Applying these principles to Dere Street, and also to the alignment of Hadrian's Wall, produced coherent results, seeming to confirm the validity of my methodology. It appears that Roman engineers began by setting out a long distance framework for a road's broad direction. Local deviations (such as to avoid a marsh or to connect with a town) were then set out from this framework, sometimes many years later.

In Britain at least, Roman surveyors do not seem to have applied the principle of a ruling or maximum gradient to their roads. This technique, which would have enabled wheeled vehicles to complete a journey without assistance en route, was adopted for many of the turnpike roads in the 18th century, and then by many of the railways, forcing their lines to twist and turn through hilly country. Instead, by travelling along their much straighter roads, but in convoy, Roman vehicles would always have had help available on steep hills – and the security of company.

More research is needed. The most comprehensive study of Britain's Roman network is Ivan Margary's Roman Roads in Britain, first published in 1955 and now out of print. He described the course of every road then thought to be Roman. His numbering system is still used: Watling Street entering Wroxeter from the east is Margary 1H.

Many have proposed additions to Margary's routes, but research has suffered from too much enthusiasm and too little scholarship, leading in the past to spurious roads being identified. When people find the remains of an old, solid road, they nearly always attribute it to the Romans. The most famous example of a false attribution is the road over Blackstone Edge, north-east of Manchester. Once acclaimed as one of the finest surviving Roman roads in Britain, it was shown recently by the Greater Manchester Archaeological Unit as likely to have been built in the 18th century. Roman roads may have been less common than thought.

This is an edited extract from The Planning of Roman Roads and Walls in Northern Britain, by J Poulter (Amberley July 2010), £18.99 pp176 PB, ISBN 9781848685482

<< prev | index | next >>

CBA web:

British Archaeology

Jan/Feb 2005
Mar/Apr 2005
May/Jun 2005
Jul/Aug 2005
Sep/Oct 2005
Nov/Dec 2005
Jan/Feb 2006
Mar/Apr 2006
May/Jun 2006
Jul/Aug 2006
Sep/Oct 2006
Nov/Dec 2006
Jan/Feb 2007
Mar/Apr 2007
May/Jun 2007
Jul/Aug 2007
Sep/Oct 2007
Nov/Dec 2007
Jan/Feb 2008
Mar/Apr 2008
May/Jun 2008
Jul/Aug 2008
Sep/Oct 2008
Nov/Dec 2008
Jan/Feb 2009
Mar/Apr 2009
May/Jun 2009
Jul/Aug 2009
Sep/Oct 2009
Nov/Dec 2009
Jan/Feb 2010
Mar/Apr 2010
May/Jun 2010
Jul/Aug 2010
Sep/Oct 2010
Nov/Dec 2010
Jan/Feb 2011
Mar/Apr 2011
May/Jun 2011
Jul/Aug 2011
Sep/Oct 2011
Nov/Dec 2011
Jan/Feb 2012
Mar/Apr 2012

CBA Briefing

Courses & lectures
CBA Network
Grants & awards

CBA homepage