Wansdyke home I What's New I Sitemap I Bibliography I Vortigern Studies l Vortigern I POLLS I LINKS l Sitemaster I FAQs
search l about Vortigern Studies l Games I Arthurian Collection I View Guestbook I Sign Guestbook

  Vortigern Studies > Wansdyke > Articles > Roman Road

Vortigern Studies Index


.Wansdyke Project 21
is part of
Vortigern Studies



Articles about Wansdyke
click here

  Wansdyke and the Roman road
Robert Vermaat

Introduction - Problems

When Albany Major wrote his book about Wansdyke[1], he chose a fanciful route for Wansdyke after it had crossed Odd Down, and made it go north, across Bathampton Down and its hillfort, which many at the time believed to have been the stage for the Arthurian battle of Badonis Mons. However, later and more thorough research established that Wansdyke reaches the Horsecombe Vale and ends there, apparently using the deep valley of the Midford Brook to reach the Avon. However, what course it follows next was and is still a matter for debate. Fox & Fox, in their thorough study of Wansdyke declared that nothing connected the eastern and western sections of Wansdyke; they could even be of a different age!

However, this need not be the case. Though dating any earthwork is very difficult, both East[2] and West[3] Wansdyke have so far relegated Roman or post-Roman dates, and looking at them as contemporary or even belonging to the same scheme is still very likely. That brings back the question about how to look at the middle section. It is very clear that, whatever the use of the Roman road, it was never used as a defensive linear earthwork. The absence of a ditch, which distinguishes the road from both East- and West Wansdyke, would make it impossible to defend. This has led to the argument that it was not part of Wansdyke, and that the loop of the Avon would fill that role. Furthermore, any defended area north of the Avon was seen as indefensible in case of an attack, the defenders likely to be trapped north of the river.

This argument, though, is weakened by two counter-arguments. The first is that the Roman road, though indeed not added with a ditch to the north, could have functioned very well as a demarcation boundary. Two charters may indicate that at least during Medieval times, this section was known as Wansdyke. A section of the Roman road through Neston Park once showed that it had a great pile of earth piled on top of it. However, an 1819 sketch (below) by Skinner[4] of the Roman Road across Neston House shows the apparently heightened bank of the agger, and it seems quite easy to mistake the agger or bank of the road for a linear earthwork. This has happened elsewhere in Wiltshire (the Ackling Dyke). Today, this feature is much reduced because of landscaping during the early 19th century. The Roman Road, which has been proven to fall out of use on this particular section, would only have been a territorial demarcation. I will discuss the charters below.

Skinner's 1819 sketch of the Roman Road across Neston House
Skinner's 1819 sketch of the Roman Road across Neston House. Click the image to enlarge.

The second argument is that linear boundaries were constructed in front of river lines during Roman times, even when the river was only a short distance to the rear. One would only have to take one look at the Roman Limes in Germany to see that when it starts, just east of the River Rhine, it runs for some distance very close to the river. The builders of Wansdyke may have used the agger of the road as a ready demarcation line, but it is possible that this part was meant to be strengthened by a ditch at a later date. As recently has been argued by Peter Fowler in a study of the earthwork through West Woods[5], it seems that parts of East Wansdyke were never finished beyond the first attempts to lay out the course. A supposed interruption may be responcible for this, as West Wansdyke also has some gaps in its course. We may therefore assume on this evidence that there is a possibility that the builders of Wansdyke made use of the Roman road where it crosses the low-lying plain of the Avon Valley between Bath and Morgan’s Hill, but that it was never strengthened with a defensive ditch.

Therefore, it is imperative that we should look at the area between both earthworks, for evidence that it was considered to belong to East and West Wansdyke.

The Evidence

There is evidence of the identification of the Roman road with Wansdyke by the local population, who were after all responcible for naming both earthworks with the same name. Contrary to common assumption, there is in fact proof that during the Middle Ages, this stretch of road was also known as part of Wansdyke. There are in fact two charters in which Wansdyke is mentioned, and which are not connected to what are now called East and West Wansdyke, but to the connecting section of Roman Road. We have to explain away why these points were named after Wansdyke in the first place, or acknowledge that, at least the time when these charters were made, people actually believed that this section was part of Wansdyke. The charters will be discussed below.

The second problem is to find a possible connection in the area between the river Avon where the Midford Brook reaches it, and the point at which the Roman road possibly starts its role as a boundary. So, to appreciate how the Roman road might have been used to link the eastern and western sections of Wansdyke, it is sensible to examine the significance of a range of archaeological features which are obviously present in locality. These features, even if they are not related to Wansdyke, are nevertheless there and should be examined if only to disprove their significance. As the evidence stands today, there is a gap on the western slope of the Avon west of Bath, where no evidence of the Roman road can be traced, nor any earthwork between the Avon and the point where the road veered north towards the Foss Way. There are, however, some general points that are worth making. Grundy[6] presents evidence that the Melksham and Chippenham forests engulfed large areas of this territory, so the argument that a ditch and bank would have been unnecessary to dissuade intrusion from the northeast is strong. Grundy derives his boundaries from early place names, most of which are early English (and were Wiltshire place names). Note Grundy's reference to 'Wodenes Ditch' (see below), which he uses to site the route of the perambulation. Fox & Fox already drew attention[7] to the view that the loop of the Avon needed no defence because of the geology: the low-lying ground, probably with heavy forest cover, would have been sparsely populated even in Roman times. Although such a woodland cover would not automatically exclude the necessity for a linear boundary, it has been argued that such a gap between two sections of earthwork can be explained by its existence.


Fox & Fox reported that the section of the Roman road between Bath and Morgan’s Hill produced no evidence for any identification with Wansdyke. More so, they reported a 12th-century Monkton Farleigh charter that would produce only negative evidence[8]. However, this is not correct. In fact, two charters have produced names that identify the Roman road with Wansdyke. The charters have already been briefly touched upon above.

The first I'd like to discuss is from the medieval Cartulary of Shaftesbury Abbey, and is described by Robert Harvey[9]. This Cartulary contained a number of 12th-century rentals or surveys of the abbey’s estates, though dating them is very difficult, due to the fact that most documents are 15th-century copies, made because previous ones, dating from the 9th to 15th century, had become illegible. The reference to Wansdyke comes from Survey A, one of three surveys, probably compiled after 1122. Each of the three rentals gives the names of the tenants for each of the abbey’s vills, usually stating the seize of their holdings, rents and services owed.

The largest tenant in Survey A is a man called Passat, who held lands in various parts of the Bradford-on-Avon estate. The relevant part of his entry A2 is: "PASSAT holds half a hide of Tortelee for 4s. rent and 1 virgate adjoining his house for 6s. and 1 virgate of Wadenesdich for 6s. and" etc. Although the form ‘Wadenesdich’ itself is unique, it is very similar to other forms of the name (below). The author is unsure[10], but it is clear that, with Wadenesdich, only Wansdyke can be meant. Where the piece of land actually was situated is unknown. Tortelee is unidentified, and probably not related to the modern village of Turleigh, which was written as Turlyn or Tyrlyng in the 13th century. Besides, the man’s lands at Tortelee, his house and the piece of land in question may have been far apart, so it could very well have been situated in the parishes of either Wraxall or Atworth, which both belonged to the estate of Bradford-on-Avon, and whose northern borders are marked by the Roman road.

The second charter was described by G.B. Grundy, who came up with a reference to Wansdyke in his description of the bounds of Melksham Forest[11]. This time though, we notice a Wansdyke-name for the Roman road east of the river Avon. In the Perambulation of Melksham Forest, AD 1300, he notes that under point 18: "To Wodenes Ditch (the Wansdyke): This brings the By. (boundary) to a definite determinable point, the Wansdyke at the place where the Chittoe parish By. Coming from the S. meets it at the W. Edge of Spye Park (OM1) (26 SE)" and under point 19: "Descending by the Wodenes Ditch to the Avon: The Pn. Runs W. along the Wansdyke to meet the Avon (36 SW) a little more than m. about due S. of Lacock village".

Grundy, writing in 1939, did not yet realise that what he identified as Wansdyke, was in fact the Roman road. Interestingly enough, that difference was also not known before 1300, when the charter was written down. Spye Park lies north of the village of Bromham, only a few miles west of Morgan’s Hill, where East Wansdyke begins its spectacular crossing of the Downs.


What do these two identifications of the Roman road with Wansdyke suggest? I would not presume to go so far as to make a suggestion that these charters prove that the Roman road was in fact used by the builders of Wansdyke. However, we cannot ignore the fact that the local inhabitants had some idea about Wansdyke as a single concept, and did not hesitate to use the name even for the Roman road. It goes too far the subject of this article to go into the problem of how Wansdyke actually received its name, but we must recognise that somehow down the line, this stretch of Roman road[12] acquired it as well.

On a side-note, Grundy also traced the line of a very early Saxon military road[13] using very early land charters. This led from Kilmington (close to Longleat), via Cley Hill, Beckington, Bradford on Avon, Monkton Farleigh, Kingsdown, Corsham, Biddlestone, and possibly also to Malmesbury. Interestingly the road, which can be traced precisely, winds around the plateau at Monkton Farleigh, providing views across to Bath but making no effort to link to Bath. The County boundary achieves the same territorial objective (observation without intrusion). And, the identification of Wansdyke with the Roman road, heightens interest in the area between Bath and Bradford on Avon, and the features described in the article by Neil McDougall (at this site).


[1] Major, Albany F. and Burrow, Edward J. 1926: The Mystery of Wansdyke, (Burrow & Kingsway, London).
[2] Eagles, Bruce 1994: The Archaeological Evidence for Settlement in the Fifth to Seventh Centuries AD, in: Aston, Michael and Carenza Lewis, eds: The Medieval Landscape of Wessex, Oxbow Monograph 46, pp.13-32.
[3] Iles, Rob 1988: West Wansdyke: Recent Archaeological research and Future Prospects, in: Bristol & Avon Archaeology 7, pp. 6-10, and Gardner, Keith S. 1998: The Wansdyke Diktat: A Discussion paper, in: Bristol and Avon Archaeology 15, pp. 57-65.
[4] British Museum No. Add. 33, 654.
[5] Fowler, Peter J. 2001: Wansdyke in the Woods: An Unfinished Roman Military Earthwork for a Non-event, in: Peter Ellis (ed.), Roman Wiltshire and after, Papers in Honour of Ken Annable, pp. 179-198.
[6] Grundy, G.B. 1939; 'The Ancient Woodland of Wiltshire', WAM vol. XXXLVIII, pp. 540.
[7] Fox, Aileen and Cyril Fox 1960: Wansdyke reconsidered, in: Archaeological Journal 115 for 1958, p.3.
[8] ibid., p 9-10: Fox & Fox noticed a probably 12th-century charter for Monkton Farleigh, describing an 11th-century grant on both sides of the Roman road. "In doing so it must have crossed the Roman road twice and on neither occasion is it mentioned, let alone referred to as Wodnesdic."
[9] Harvey, Robert B. 1998: Shaftesbury Abbey's 12th-century Rentals for Bradford-on-Avon, in: Wiltshire Archaeological and Natural History Magazine 91, pp. 76-89.
[10] ibid., p. 79: Harvey asks himself: "can this be Wansdyke?"
[11] Grundy, G.B. 1939; p. 576-9.
[12] It must be noted that no other piece of the London-Bath Roman road is so far known to be called Wansdyke, which must be proof that the stretch between Bath and Morgan’s Hill was indeed seen as separate.
[13] Grundy, G.B. 1918; 'The Ancient Highways and Tracks of Wiltshire, Berkshire and Hampshire', in: Archaeological Journal vol. LXXV.

Wansdyke and the Roman Road is 2001 Robert Vermaat

Comments to: Robert Vermaat

VortigernStudies and Wansdyke Project 21 are copyright Robert Vermaat 1999-2007.
All rights reserved