feature: British Archaeology 121, November / December 2011

On Google Earth the linear feature in the river is just visible, running from the island (bottom left) to the modern path opposite leading to the castle mound. To access the image, type +52 08 03.76 -0 27 50.60 into the "Fly to" or search box. That will take you to the island. Move the clock menu time slider to 2006, to a photo taken in the summer with low, clear water. Best views are seen from eye altitudes of 150m and 300m.

We regularly hear stories about new archaeological discoveries made by scanning Google maps. Matt Edgeworth has seen something unusual: a huge "underwater cropmark" in a river.

When people in medieval times looked at urban rivers, they saw something different from what we see today. The picturesque scene with pleasant walks along embankment promenades, past riverside gardens and ornamental bridges was a creation of the Victorian and Edwardian eras. No, what the medieval eye saw was more basic, more elemental. The river was seen and experienced, amongst other things, as a great flow of energy moving slowly and inexorably through the town.

It was a flow of energy, moreover, that could be harnessed and put to use. In many proto-urban towns throughout Britain, from at least the seventh or eighth centuries AD (the date of Saxon mills at Old Windsor and Tamworth), mills were built next to the river – larger than those on small streams out in the country. Artificial watercourses were cut to serve as headraces and tailraces, taking flow of water from the river and back again. In some cases, massive stone weirs were placed across the river to create a head of water to power the mills. These were such important features – so vital to the urban economy – that towns like Warwick, Ware and Wareham were actually named after their weirs.

Surprisingly few urban dwellers today know in which direction their local river flows. Which way is upstream and which downstream? Such knowledge, once so crucial, no longer has much significance for town inhabitants. Other flows – of electricity, gas and oil – now provide most of the energy that the town needs. But that does not mean that we should forget how important the flow of water was to people in the past. The river environment is full of clues as to how they once interacted with the energy of water – how they channelled, shaped and diverted it for their own purposes and projects.

A band of weeds

View of the river weeds from a boat, looking straight down. They are known by anglers as cabbage lilies, more formally as the submerged leaves of Nuphar lutea (yellow water-lily).

Take Bedford, for example, where a huge archaeological feature has come to light. A straight band of riverbed vegetation crosses the river Great Ouse diagonally from the north bank to a small island on the south side. This "underwater cropmark" is 40m long and 3m wide, visible on Google Earth on a 2006 aerial photo from 1km up (see main image).

Normally the feature can not be seen from the ground, with the water being too high and too murky. But in 2011 the exceptionally dry spring lowered the river, and the short rainbursts in June – when I drafted this article – significantly improved visibility. As never before in my lifetime, for about a month, the feature was clearly visible from the river bank.

This is not the first time that it has been noted. It was described and mapped by local historians in the late 19th and early 20th centuries, and variously identified as a former causeway, barrage, weir or dam. The earliest reference is by a former mayor of Bedford called George Hurst.

He wrote in 1859 that part of the foundation of a stone wall, ten feet wide (3m), could be seen crossing the river at low water. He mapped its position for us, and said the wall was the remains of a stone barrage in existence up to 1774, when it was demolished and the stone re-used for building works. Over 60 years later the reverend Farrar opted for the same idea. When the water is clear, he wrote in 1926, it is possible to trace a broad band of weeds at the bottom of the river, marking the course of the barrage which crossed the river obliquely in a south-westerly direction.

It seems, then, that the broad band of weeds visible today marks the course of a former large stone structure which crossed the river to the small island, already submerged when it was removed in the late 18th century as part of river navigation works. Quite why lily pads should find such favourable rooting conditions on the feature is not fully understood. But it is well known that growth of vegetation on dry land can be influenced by archaeological features: the same must surely be true underwater.

Liquid flow – a force driven by gravity yet intensively modified, shaped and utilised.

The idea that it was a former weir is supported by its diagonal orientation relative to the current, reminiscent of many other weirs both ancient and modern, such as those at Durham. Weirs of Norman date excavated by the University of Leicester Archaeological Services on former channels of the river Trent at Hemington, Leicestershire, give some indication of how they were built. There, parallel rows of timber piles were driven into the riverbed diagonally across the direction of flow, with intervening spaces filled with large stones.

Further supporting evidence for identifying Bedford's feature as a weir is provided by the small island at its southern end. This was formed by the digging of a semicircular channel around it – presumably in order to bypass the obstruction presented by the weir. This (later) feature enabled boats to get round the weir and negotiate the step from upper to lower river levels or vice versa, perhaps by means of a pound lock.

Other river features nearby may be related to the weir. A side channel diverts flow from the river on its south side from just upstream of the weir, directing water through a head-race to the site of Duck Mill – formerly Joel's Mill – one of two principal mills serving the town in late Saxon and medieval times. This is a multi-period riverscape, its fluid continuities and changes through time as open to archaeological interpretation as any landscape or deeply stratified site.

The word "weir" has its roots in the Old English werian, meaning to defend (as in "beware" or "to be wary"). Towns on rivers may once have incorporated weirs into their systems of defence: after all, it makes little sense to surround a town with earthen ramparts and deep ditches if river approaches were left open. An associated role of weirs may have been to create the head of water to direct flow through defensive earthworks. This is an important but neglected area of study. Urban defences are usually discussed in terms of their solid materials – the walls or earthen ramparts which presented physical barriers to attack. Water flowing through ditches outside of those ramparts receives little attention. In the rest of this article, I will argue that a flowing linear earthwork in southern Bedford, commonly supposed to be of late Anglo-Saxon construction, could only have functioned if there was an associated weir across the river.

The king's flowing ditch

Drawing

The King's Ditch and rampart today, and a similar stretch as it was in 1920, drawn by Beauchamp Wadmore.

Around the edge of the middle Saxon burh of northern Bedford, water was diverted into boundary ditches from nearby streams. A slope down to the river made it impossible to use river water for this purpose. But in the absence of such a gradient the diversion of river water around a town circuit was a real possibility. When Edward the Elder fortified the southern part of Bedford in AD 914, effectively turning the town into a "double-burh", the terrain was so flat on that side that the river was the obvious source of water to use. The Anglo-Saxon Chronicle records that King Edward's army stayed in the town for four weeks in order to build the defensive earthworks. A kilometre-long semicircular earthwork known as King's Ditch was constructed. This survives today in places as a broad ditch alongside a sizeable inner bank, covered in dense undergrowth. Other stretches have been levelled and built over, so that it runs through culverts beneath roads, offices and schools.

The most interesting feature about the King's Ditch is that it flows. This is no stagnant ditch. Nor does it have a languid current like the Great Ouse from which its times is provided by the existence of fishponds just inside the ditch on its south-eastern side. These fishponds originally belonged to St John's Hospital, a medieval religious institution. Maps show that there were channels bringing water in and out of the fishponds, which presupposes a strong flow in the ditch itself. Although later in date than the King's Ditch itself, the fishponds tapped into and made use of an already existing flow of water.

By rights the water should be stagnant, or should back up into the river the way it came, given the flat terrain of southern Bedford. The fact that it flows suggests that this was an important element of the initial design. The monument was built with a constant gradient and a contrived fall of water of about 2m between inlet and outlet – much greater than the natural fall of the river between those two points, with gravity enlisted as the principal force impelling water to flow through the ditch.

What was the weir for?

Why would the builders of the earthwork go to the trouble of making it flow? Consider for a moment why a town boundary ditch full of stagnant water would not work. Imagine how quickly such a ditch would start to silt up, fill with rubbish, get polluted and clog with vegetation. Carrying out periodic maintenance would be difficult. Thick slimy mud thrown up from the ditch would slip down into the water again, and the earthwork bank alongside would soon fall into disrepair. Above all, the water – since it has no momentum or energy – cannot be exploited for industrial purposes. Since it has no flow, it cannot be moved around or managed in any way at all.

Now consider the many benefits of a flowing ditch. The water itself does some of the work of maintenance, carrying away smaller particles of silt and leaving the gravel behind, making the ditch easier to clean out. It is self-scouring. The ditch takes away rubbish and pollutants, acts as a sewer and a drain, while replenishing itself with fresh water. It also serves as a flood defence. Unclogged by vegetation and mud, the stream can be used for bathing and washing of clothes, or for transport of goods by boat, with direct access to and from the river.

Flowing water is a vibrant material which can be controlled and moved around as required. It can be diverted through sluices into side channels – to fill further water features or flush them out, to cool forges, to serve potteries and perform innumerable other industrial functions. And that is not to mention the symbolic significance of encircling towns with flowing water.

How, then, was the artificially contrived fall and the flow of water through the ditch accomplished? There was one obvious way to do it, and that was to construct a substantial weir across the river somewhere between inlet and outlet of the King's Ditch. Its purpose was to create the necessary artificial step in water levels. Such a weir on the river is actually a logical necessity for the King's Ditch to function as a flowing earthwork, and its former existence might be deduced even if there were no material trace surviving of the structure itself. The technology of stepping the river through the building of weirs, in order to create a head of water to drive mills, was already well established. It made sense to apply the same principals to ditches of defensive earthworks too, especially if a large workforce was available to carry out the work.

Weirs rarely figure in archaeological discussions, and their impact on urban development has yet to be properly explored. They affected the depth and flow of water both upstream and downstream. One major effect was to deepen the river upstream, making fording difficult. In drowning fords, weirs helped create conditions that led to the widespread replacement of fords by bridges in the tenth–12th centuries. Another important effect may have been on the location of Norman castles, of which Bedford's is a possible example. It is often assumed that, when castles and weirs are found in close proximity, the castle came first and the weir was a subsidiary or later feature. In fact the temporal sequence may be the other way round. These were such key locations for control of river flow and boat traffic, as well as for exploiting the energy and resources of the river, that in many cases castles must have been placed next to pre-existing weirs.

As for the broad band of waterweeds that started this discussion, this is surely not an isolated example. Similar underwater "cropmarks", indicating the position of ancient structures, await discovery in other urban riverscapes, potentially visible on aerial photos (at least those taken when the water was low and clear). Detailed viewing of aerial images – zooming in and out to view at different scales and angles while panning along the course of rivers – has recently been made easy and accessible to all by Google Earth. There is the real prospect of a surge in the number of new sites, transforming the archaeological understanding of rivers and river towns.

Matt Edgeworth is honorary research fellow at the School of Archaeology & Ancient History, University of Leicester, and author of 'Fluid Pasts: Archaeology of Flow' (Bloomsbury Academic, October 2011).

Issue 121

Nov / Dec 2011

Contents

making tracks

news

letters

Mick's Travels

CBA Correspondent

spoilheap

features

Premiere at Wadi Faynan

The weir and the flowing earthworks of Bedford

THE BIG DIG/COVER STORY: Bouldnor Cliff

The Living Stones of Brittany

Onsite and Online at Craig Phadrig

Archaeology at the Festival of Britain

feature

The weir and the flowing earthworks of Bedford

A band of weeds

The king's flowing ditch

What was the weir for?

CBA web:

British Archaeology

CBA Briefing

CBA homepage