Blog PostHadrian's Wall

Gigantic Prehistoric Canals (Dykes) – Wansdyke

This is an extract from the forthcoming book ‘Gigantic Prehistoric Canals of Britain – a gazetteer of all known Dykes’. This essay is a detailed analysis of Wansdyke and shows why the Dyke was constructed and disproving the current ‘expert’ analysis as being frivolous nonsense as a scientific explanation.

CASE STUDY – Wansdyke

According to Wikipedia, “Wansdyke consists of two sections of 14 and 19 kilometres (9 and 12 mi) long with some gaps in between. East Wansdyke is an impressive linear earthwork, consisting of a ditch and bank running approximately east-west, between Savernake Forest and Morgan’s Hill. West Wansdyke is also a linear earthwork, running from Monkton Combe south of Bath to Maes Knoll south of Bristol, but less impressive than its eastern counterpart. The middle section, 22 kilometres (14 mi) long, is sometimes referred to as ‘Mid Wansdyke’ but is formed by the remains of the London to Bath Roman road. It used to be thought that these sections were all part of one continuous undertaking, especially during the Middle Ages when the pagan name Wansdyke was applied to all three parts.

East Wansdyke in Wiltshire, on the south of the Marlborough Downs, has been less disturbed by later agriculture and building and remains more clearly traceable on the ground than the western part. Here the bank is up to 4 m (13 ft) high with a ditch up to 2.5 m (8.2 ft) deep. Wansdyke’s origins are unclear, but archaeological data shows that the eastern part was probably built during the 5th or 6th century. That is after the withdrawal of the Romans and before the takeover by Anglo-Saxons. The ditch is on the north side, so presumably it was used by the British as a defence against West Saxons encroaching from the upper Thames Valley westward into what is now the West Country.

West Wansdyke, although the antiquarians like John Collinson considered West Wansdyke to stretch from south East of Bath to the west of Maes Knoll, a review in 1960 considered that there was no evidence of its existence to the west of Maes Knoll.   Keith Gardner refuted this with newly discovered documentary evidence.  In 2007 a series of sections were dug across the earthwork which showed that it had existed where there are no longer visible surface remains.

It was shown that the earthwork had a consistent design, with stone or timber revetment. There was little dating evidence, but it was consistent with either a late Roman or post-Roman date. A paper in “The Last of the Britons” conference in 2007 suggests that the West Wansdyke continues from Maes Knoll to the hill forts above the Avon Gorge and controls the crossings of the river at Saltford and Bristol as well as at Bath.

As there is little archaeological evidence to date the western Wansdyke, it may have marked a division between British Celtic kingdoms or have been a boundary with the Saxons. The evidence for its western extension is earthworks along the north side of Dundry Hill, its mention in a charter and a road name.

Gigantic Prehistoric Canals (Dykes) - Wansdyke
The location of Wansdyke in prehistoric times – in comparison to the Victorian canal – Prehistoric canals (dykes)

Figure 10- Map comparing Wansdyke (green) to the Kennet and Avon canal
(Gigantic Prehistoric Canals (Dykes) – Wansdyke)

The area of the western Wansdyke became the border between the Romano-British Celts and the West Saxons following the Battle of Deorham in 577 AD.  According to the Anglo-Saxon Chronicle, the ‘Saxon’ Cenwalh achieved a breakthrough against the British Celtic tribes, with victories at Bradford on Avon (in the Avon Gap in the Wansdyke) in 652 AD,  and further south at the Battle of Peonnum (at Penselwood) in 658 AD,  followed by an advance west through the Polden Hills to the River Parrett.  It is, however significant to note that the names of the early Wessex kings appear to have a Brythonic (British) rather than Germanic (Saxon) etymology”.

Later scholars have often questioned the accuracy of the descriptions given by these antiquaries such as assertions that Wansdyke reached the Bristol Channel (Fox and Fox 1958) and even Asser’s statement that Offa’s Dyke ran from sea to sea (Hill and Worthington 2003 106). While some antiquarians were probably exaggerating the size of earthworks, we must be cautious of dismissing descriptions of the dykes from before they suffered the ravages of the Agricultural Revolution.

Some scholars went beyond merely describing the dykes and tried, often erroneously (with hindsight), to link them with known historical events like the Belgic invasions mentioned by Caesar or Caesar’s own invasion (Warne 1872 4-10; Guest 1883; Handford 1951 119-40). Among these early, rather speculative descriptions, the work of the Wiltshire historian Sir Richard Colt Hoare (1758-1838) stands out, not only in terms of the quality of his survey work but also his ability to differentiate between features of different dates, for example by realising that the central section of Wansdyke was actually a Roman road (Hoare 1812; Hoare 1821).1

Boundary Markers & Defensive Structures

The archaeologist Mortimer Wheeler produced an analysis of the dykes of southeast England and, like Fox, suggested that they were not primarily military structures but political boundary markers facing post-Roman Britons centred on London (Wheeler 1934 261).

Osbert Crawford’s 1953 book on Wansdyke analysis of the dykes as either military-political (with no clarification of what that meant in practice) or in respect of the coastal dykes (like Dane’s Dyke at Flamborough Head or the Cornish dykes) calling them beach-heads He also made no attempt to group what he termed defensive linear earthworks by period; in his list of them given as an appendix to his field archaeology guide, he includes prehistoric and Anglo-Saxon dykes together with undated earthworks (Crawford 1936 240-53). These shortcomings are easy to criticise now, though, at the time, Crawford’s work was exceptional.

The problem with Crawford’s work is that Wansdyke has a gap in the middle of the defence/boundary and no explanation for the lack of Dyke?

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
The ‘so called’ warring tribes of the Medieval period – Prehistoric canals (dykes)

Figure 11 – Crawford Political Boundaries

One of the fundamental weaknesses of this earthwork as a defensive structure is the fact it just stops in a middle of a field on the Eastern side of the Dyke.

A series of excavations carried out by H. Stephen Green on Wansdyke in the late 1960s provided the first opportunity to apply both snail and pollen analysis to dyke studies (Green 1971). While the evidence for snails was largely inconclusive, the pollen samples (analysed by GW Dimbleby) suggested that central parts of the eastern half of Wansdyke passed through a pasture. The pollen evidence from the eastern end of Wansdyke suggested the presence of nearby woodland (Savernake Forest), though it did not prove or disprove the hypothesis that it was an impassable barrier that protected the eastern flank of the dyke, as Fox had postulated (Fox and Fox 1958).

There have sadly been very few excavations or thorough investigations of these dykes (remembering there are over 1000 dykes in Britain), one of the earliest excavations was overseen by Lieutenant-General Pitt Rivers, who excavated Wansdyke at Shepherd’s Shore, Devizes.

Although most of the findings and accounts lack scientific accuracy, we find some aspects that support the dating of dykes earlier than the current ‘expert’ theories of Saxon Britain.  In one of his cuttings through the ditch and bank he noted that “At Wand’s house, there is a break in the line of the Dyke, which is occupied by the site of the Roman station of Verlucio, where quantities of Roman pottery are scattered on the soil”.

Pitt-Rivers also found vast amounts of pottery and iron nails in the bank of Wansdyke further down the earthwork. Isn’t it more likely, that the Romans placed their settlement actually on the Dyke as it may have still held water two thousand years ago and, looking at the positions of the findings, actually cleaned it out, which would allow them to take and receive supplies by boat to and from the river Kennet? Moreover, he noticed that at Sandy Lane and the Roman settlement in Spye Park – the river was still running inside the Dyke!

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
Stukeley’s map of Morgans Hill – Prehistoric canals (dykes)
Figure 12- Stukeley’s Map of 1724 showing the Roman road ‘sitting on top’ of the Wansdyke ditch

In fact, within his excavation of the dyke we noted that:

“Very little ‘silting’ had accumulated on the escarp, but in the ditch, it had collected to a depth of about 3 feet in the centre” – showing that the main purpose for the dyke was to take water like a canal.

The final proof of the fact that the Dykes predate the Roman invasion can be found strangely with empirical evidence from a drawing produced by the first British archaeologist William Stukeley in 1724.  The Roman road is clearly made on and above the Wansdyke ditch and then cutting through the Dyke’s bank – this is impossible to do if the bank and ditch were built AFTER the road (archaeologist and historians, please take note!!).

If we look at Wansdyke on an OS map, to the East, the ditch ends in the middle of nowhere, just before the forest of Savernlake.  The question that archaeologists fail to answer is ‘why stop there’?

If you continued another 6km, you would have reached water – a natural boundary (remembering you have already cut 19km), or just turn south, and that’s a mere 3km.  To the west, it’s even worse, you could go south again and connect to the river, using that as a natural defensive boundary, but no – it just stops dead.

If you were to attack the ditch, you would be mad – as you just walk around it as the German’s did on the Siegfried line at the outbreak of World War II.

The final proof of that the ditch is not defensive can be found in its method of construction.  Going back to the work of Pitt-Rivers, he produced a series of detailed cross-sectional drawings to support his investigation.  These drawings showed two significant findings: firstly, the ditches fill was spread over both sides of the ditch (one side slightly more than the other) – you would not do this if it’s a defensive feature as you would want all the ‘high ground’ on your side to help defend.  This kind of slightly even distribution we see on Iron Hill ditches and other featured sites such as Durrington, Stonehenge, and particularly Old Sarum.  In fact, the only site we don’t see this happen is at Avebury.

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
Savernake Forest – Prehistoric canals (dykes)
Figure 13 – Wansdyke East section just stops – the forest would not have been there during Roman times and hence the roman road going through the centre

Secondly, and most importantly, the bottom of the ditch is flat, again just like Stonehenge, Avebury and Durrington – it could also be Old Sarum, but sadly nobody has done an excavation of the ditches to date.

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
Pitt-Rivers cross section – Prehistoric canals (dykes)
Figure 14- Pitt-Rivers cross section shows ditch fill spread on BOTH sides – so not defensive

You do not make the bottom of a ditch flat if used as a defensive ditch or a land marker (you’re doubling the time it takes to make a ditch).  On Pitt-River’s excavation plan, the flat bottom is a third of the width of the ditch.

The Kennet and Avon Canal (Fig. 93) is a waterway in southern England with an overall length of 87 miles (140 km), made up of two lengths of navigable river linked by a canal.  The name is commonly used to refer to the entire length of the navigation rather than solely to the central canal section. From Bristol to Bath, the waterway follows the natural course of the River Avon before the canal links it to the River Kennet at Newbury, and from there to Reading on the River Thames. Quite remarkably, Wansdyke was constructed just 3 km north of the Kennet and Avon canal. If it was a prehistoric waterway, it would have achieved the same purpose of joining the River Thames to the Bristol Channel but some 6-8K years before the Victorian’s great canal system.

Now let us consider the human resources it needed to create such a structure.  At 33 km in length (33,000m), its volume can be calculated as about 618,750 cubic metres of chalk (if it is 2.5m deep and 7.5m wide as an average). This is the approximate volume of material removed from the ditch.  This is five times more than what was excavated at Avebury and two and a half times larger than Silbury hill. So, according to Atkinson’s calculations at Silbury Hill, we are looking at 45 million working hours, which equates to 100 people working for 12-hours every day 365 days a year for 102 years.

It took the Victorians one hundred years to build the Avon & Kennet Canal using metal tools and steam engines. Moreover, the Victorian canal was only 1.3m deep and 6m wide – the ancient canal ditch is in places twice as big as the Victorian counterparts. So, whoever built it must have spent many years completing this task – which suggested it was essential at the time of construction.

If you study a geological map of this area, you will be struck by the endless twisting rivers that once flowed on and around the Wansdyke earthwork.  These are called ‘superficial deposits’ and consist of sand, silt and pebbles.  This is evidence that ‘once upon a time’ rivers formerly flowed in these what we call today’ dry valleys’ (Palaeochannels), and the clue just like the name for these earthworks (Dykes) is in the title.

If we add the known prehistoric groundwater rivers (based on our hypothesis) to the map of this great ditch, we find that springs would feed the ditch naturally, flooding it – and there you have yourself a canal.   This 21-mile canal links (just like the Avon & Kennet) the river Thames to the Bristol Channel.   And there can be only one reason for this massive undertaking (boat travel for the purpose of trade) just like the Victorian ancestors would achieve some six to eight thousand years later.

However, the most miraculous revelation about these Dykes is the fact that the builders had a greater knowledge of hydrology than even today’s engineers and geologists, for the weakness and final demise of the canal system in Britain was the fact they were very slow (the  Victorians had to build at least one lock per mile of the canal to keep in the water high) the Kennet and Avon have over 100 locks. The original prehistoric channel had none as it relied on the natural groundwater table levels to fill the canal. Therefore, it was straighter and shorter in length than the Avon & Kennet, which would have allowed them to travel from the Thames to the Bristol Channel almost a hundred times faster than the Victorian barges.

Moreover, if you find this difficult to believe, you should bear in mind that there are over 1,000 dykes in Britain; a majority, 90%, are straight and less than 1km in length.   To date, every single dyke we have investigated has underground links to these prehistoric rivers created from the ‘post-glacial flooding’ after the last ice age. 

Furthermore, some of these Dykes are linked to ancient sites directly, as shown in this case study confirming the age of these earthworks. For if we add the known prehistoric groundwater rivers, we find out why this remarkable canal has two dead ends.  For West side of the dyke is in the middle of a prehistoric island, which perfectly cuts the island into two and would allow boats to sail from one end of the island to the other in the Mesolithic. At a later date, an extension was added (as the water table levels dropped) to keep it connected to the Bristol Channel in the Central and West Section.

This explains why the Central section is different and straighter than the Eastern section of Wansdyke and is often referred to as a ‘Roman Road’ on OS maps.  Moreover, it proves beyond doubt that it was not built as a boundary marker of a defensive ditch as it would have had a 6-mile gap in the middle of the earthwork.

This Dyke was built over hundreds if not thousands of years, which reflects not only the engineering and organisational skills of this advanced civilisation but, moreover, the sophistication and complexity of this trading society.

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
Gigantic Prehistoric Canals (Dykes) - Wansdyke 12

Figure 15 – East Wansdyke was an island in the Prehistoric Period with section analysis to follow

East Wansdyke Maps in Detail (from East to West)

Section 1 – North Savernake to Gore Copse – 1,995 metres long (6,650 working days – 20 men, 333 days).

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
East Wansdyke (section 1) with Cross Section -Prehistoric canals (dykes)

Figure 16 – Wansdyke NW of Wernham – with Mesolithic Water Levels

No information is available from Historic Monuments England – no survey or excavation has been undertaken.

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
East Wansdyke OS series – Prehistoric canals (dykes)

Figure 17 – Current OS Map

(Gigantic Prehistoric Canals (Dykes) - Wansdyke)
East Wansdyke 1800 map series – Prehistoric canals (dykes)

Figure 18 – 1800 map

page 34.1
LiDAR Map – Prehistoric canals (dykes)

The maps tell us very little, sadly, except that the subsoil of clay with flints is ideal for canals and water retention – so even if the water table fell, the clay would and could be used like a dew pond as a liner for rainwater retention.  We estimate at a rate of 0.3m per day this section of the Dyke would have taken 6,650 working hours to construct or 20 men working 333 days – this is based on one man cutting 0.3m per day (Ditch down to 2m and bank) although Fowler used 0.1423 metres per hour and claimed 1,000 men built the ENTIRE East Dyke in 30 days working 10-hours a day 6 days a week??

This ridiculous claim I have discounted as even with modern tools, it would be too demanding, and the logistics of feeding, mending tools and housing 1,000 men was not taken into account.  Therefore, we have used a more reasonable number of men (20) ‘digging’ as you would probably need double that number to feed and maintain and clear (it was mostly woods and forests in the past) in front of the diggers.

ment as the barrows were of Bronze age or earlier and as we have seen from the carbon dating evidence at Blackrock Lane much earlier than its 1500 BCE date.


Further Reading

For information about British Prehistory, visit for the most extensive archaeology blogs and investigations collection, including modern LiDAR reports.  This site also includes extracts and articles from the Robert John Langdon Trilogy about Britain in the Prehistoric period, including titles such as The Stonehenge Enigma, Dawn of the Lost Civilisation and the ultimate proof of Post Glacial Flooding and the landscape we see today.

Robert John Langdon has also created a YouTube web channel with over 100 investigations and video documentaries to support his classic trilogy (Prehistoric Britain). He has also released a collection of strange coincidences that he calls ‘13 Things that Don’t Make Sense in History’ and his recent discovery of a lost Stone Avenue at Avebury in Wiltshire called ‘Silbury Avenue – the Lost Stone Avenue’.

Langdon has also produced a series of ‘shorts’, which are extracts from his main body of books:

The Ancient Mariners

Stonehenge Built 8300 BCE

Old Sarum

Prehistoric Rivers

Dykes ditches and Earthworks

Echoes of Atlantis

Homo Superior

For active discussions on the findings of the TRILOGY and recent LiDAR investigations that are published on our WEBSITE, you can join our and leave a message or join the debate on our Facebook Group.

2 thoughts on “Gigantic Prehistoric Canals (Dykes) – Wansdyke

Comments are closed.