Bridge over troubled water: Roman finds from the Tees at Piercebridge and beyond

Ritual or rubbish – how should we interpret objects that ended their days in rivers during the Roman period? Hella Eckardt and Philippa Walton consider this question by exploring the more than 3,600 artefacts that have been recovered from the waters of the Tees at Piercebridge.


Roman bridges are wonders of engineering, and played a vital role in the Empire’s transport network. They were also hugely symbolic structures, embodying the might of imperial power and conquest. Famous examples include Caesar’s bridge across the Rhine, Caligula’s bridge across the Bay of Baiae, and Trajan’s stone bridge across the Danube. When carefully excavated, we can learn much about the ways in which they were constructed, and dendrochronological dating can even provide insights into the frequency of repairs: for example, the late Roman bridge at Cuijk in the Netherlands had three phases of construction, each approximately 20 years apart.

The sun sets over the River Tees at Piercebridge, near Darlington.

In antiquity, rivers were dangerous, and crossing them required rites in the form of both prayers and sacrifices to appease the river god. Occasionally a historical source records such rituals, as for Crassus at the Euphrates or Caesar at the Rubicon. We hear, too, about the special status of particular bridges. For example, the Pons Sublicius – the first bridge across the Tiber – was surrounded by strong taboos forbidding the use of metal in its construction and maintenance. During religious festivals, effigies made of bundles of straw were thrown into the river from the bridge. We know much less about the day-to-day offerings people made from bridges, though, as any altars do not normally survive, and finds from riverine contexts are usually very poorly published. Frequently only the coins or unusual objects are catalogued, and even those are usually completely divorced from their fluvial and archaeological context.

Thousands of Roman artefacts have been recovered from these waters in the last four decades, including this copper-alloy plaque, depicting a young male bust wearing a lion-skin headdress. It might represent Hercules or Alexander the Great, or possibly Commodus or Caracalla. It was folded when found, and its function is unknown.

This will change, however, with the publication of the entire assemblage of Roman finds from the River Tees at Piercebridge, near Darlington, following a two-year project funded by the Leverhulme Trust. Between the mid-1980s and 2018, more than 3,600 objects were retrieved from the riverbed there by two divers, Bob Middlemass and Rolfe Mitchinson – and, as well as providing a spotlight on riverine deposition at the edge of the Roman Empire, studying this assemblage has offered an opportunity to think about river finds more widely throughout the Roman world. In the past, even large hoards – such as those recovered from the Rhine at Neupotz – were explained as accidental losses or shipwrecks, while more humble finds were often seen simply as rubbish. While these interpretations are of value, they stand in stark contrast to arguments made by archaeologists working on prehistoric sites, who tend to see the plethora of axes, weapons, and other objects from rivers primarily as ritual offerings.

Piercebridge in context

Piercebridge is located in a part of northern Britain that is thought to have been occupied by the Brigantes, but which was conquered by the Roman army in the AD 70s. It sits at the point where Dere Street, the main Roman road north, crosses the River Tees. As a result, it must have seen a great deal of military traffic passing backwards and forwards from the legionary fortress at York to the northern frontier. It is also just 6km north of Stanwick, a major oppidum long associated with Cartimandua and now known to have been occupied between c.80/70 BC and AD 65/75 (see CA 325). Another important site nearby is Scotch Corner, a location where prehistoric (and then Roman) routes running east–west over the Pennines and north–south converged. As reported in CA 365, this was an important Late Iron Age settlement, which received very high-status imports and was involved in the manufacture of coin pellets, before possibly evolving into a Roman trading or supply centre.

Piercebridge in context: this map shows north-east England in the Roman period. The area around Piercebridge is thought to have been part of the territory of the Brigantes, which was conquered by Rome in the AD 70s. IMAGE: Mark Hoyle.

Today, the Roman fort at Piercebridge is largely obscured by the modern village, while the vicus is preserved underneath Tofts Field to the north of the River Tees; there are also a villa and small roadside settlement located south of the river. All were examined in a series of excavations throughout the 20th century. It has long been thought that a fort was built at Piercebridge in the Flavian period, but while this would make sense strategically, as yet excavated evidence is lacking. Instead, it is clear that an extensive settlement, perhaps even a small town, developed north of the river by the end of the 1st century AD and that there was a Roman military presence at the site from about AD 170/180. The surviving fort defences were constructed at some time in the 3rd century AD, and stone inscriptions record the presence of soldiers from three legions (Legio II Augusta, Legio VI Victrix, and Legio XXII Primigenia). Beyond the 4th century, though, only small-scale settlement continued.

An extensive programme of excavation and survey was carried out at Piercebridge throughout the 20th century. This means that it is possible to locate the finds from the River Tees within both a landscape and archaeological context, as shown on this plan. IMAGE: Mark Hoyle.

There were three ancient bridges across the Tees, all situated to the east of the early 16th-century bridge which still stands today. The oldest of these bridges was investigated by Channel 4’s Time Team in 2009, and is described as two parallel rows of timbers, one of which yielded a calibrated radiocarbon date of 40 BC-AD 85 (94.3% probability). This bridge may even pre-date the conquest, as it aligns with a footpath leading from Stanwick to the Tees, which is thought to echo a Late Iron Age route.

A substantial bridge was then constructed in c.AD 90 on the line of Dere Street, initially in timber, but in its second phase quite possibly with stone piers and abutments. Oak piles and timbers possibly associated with the bridge’s northern abutment were explored by the divers and also radiocarbon-dated by Time Team. The bulk of the Roman finds were found just downstream from this bridge. Finally, in the late 2nd or early 3rd century AD, a stone bridge was constructed c.200m downstream, for which Dere Street had to be realigned.

Divers Rolfe Mitchinson and Bob Middlemass have dedicated the past 30 years to finding out more about the Roman bridges at Piercebridge. Here they are holding Bob’s reconstruction of the wooden structures they have recorded on the bed of the River Tees.

The Tees is a fast-flowing river, with dangerous floods moving substantial amounts of gravel and sediment, but it appears that the Roman objects were recovered close to their original point of deposition in an area of relatively still water, possibly protected from being washed away by the bridge foundation timbers as well as by a later weir.

Both copper-alloy and pipe-clay figurines were recovered from the river. Here are some of the more complete examples, which include Mercury and his attributes the ram and the tortoise, as well as two Cupids, the companions of Venus.
Finds of precious metal are rare in Roman Britain outside hoard deposits, but this image depicts a few of the items of gold jewellery from the river. They include finger-rings with garnet and glass settings, elements of a variety of necklaces, earrings, and appliqués.

Ritual versus rubbish

Divers Bob Middlemass and Rolfe Mitchinson have recovered just over 3,600 artefacts from the bed of the Tees. The vast majority are Roman in date, although there are also a few late Iron Age, medieval, and even modern objects. As with all river finds, there are various ways these objects could have ended up on the riverbed. For example, they might represent deliberate deposits, most likely made from a bridge. However, the river would have provided a convenient place to dump rubbish, too, and middens on the riverbanks could have been washed away during flooding. Finally, the assemblage could be a mixture of all these depositional processes.

In order to get to grips with whether we were dealing with ritual practice or rubbish disposal, it was critical to explore the composition of the whole assemblage and publish it in its entirety rather than to focus on its most distinctive objects. There are, of course, some objects which could be considered religious or votive in nature. They include copper-alloy and pipe-clay figurines associated with Mercury and Venus, as well as a silver finger-ring with a dedication to the god Mars. There are also large quantities of precious metal objects, including more than 600 silver denarii and 60 fragments of gold jewellery. These are unlikely to have all been lost accidentally. However, there are numerous finds which could be categorised as casual losses or rubbish, such as iron nails, lead fishing weights, broken pottery, and animal bones with butchery marks.

This striking reconstruction of the bridge over the River Tees at Piercebridge emphasises the religious and magical aspects of people’s interactions with water.

As so few assemblages of objects from rivers have been published in their entirety, it is quite difficult to characterise what a Roman riverine deposit of ‘votive’ material or ‘rubbish’ might look like. Certainly, the assemblage has few similarities with riverine deposits on the Continent, which are dominated by vessels, weaponry, and armour. We therefore decided to compare the riverine assemblage with finds excavated in nearby settlements. This showed that valuable items such as silver coins and gold jewellery were more common in the river, while bone objects were under-represented. While the lack of bone objects is most likely the result of taphonomic processes, with lighter materials simply floating away, the presence of so many high-value items can only really be explained by deliberate selection.

The argument in favour of some deliberate selection and deposition is also supported by the coins from the river. Their chronological profile, which peaks in the late 2nd and early 3rd century AD, is very different from the late Roman emphasis of the coins recovered from the fort and vicus at Piercebridge. This indicates that the finds in the Tees were not simply washed out from settlement rubbish deposits, as they do not mirror the assemblages excavated on land. While we may never know exactly how the assemblage formed, and we do not discount the presence of some rubbish, we argue that many of the finds were deliberately deposited in the River Tees. The stunning reconstruction of the Piercebridge site by the archaeological illustrator Aaron Watson highlights the ritual and magical aspects of people’s interactions with the river.

The coins recovered from the River Tees span the entire Roman period, although issues dating to the late 2nd and 3rd century AD are most common. The anaerobic conditions on the riverbed mean that many are exceptionally well preserved.
Fourteen belt-mounts employing opposing trumpet motifs were recovered from the river. They are of a type found throughout the Roman world, but with particular concentrations along the Danubian limes.

The reasons for votive deposition, however, remain opaque. As we have already noted, the crossing of a dangerous river in the Roman world required offerings. The Tees may also have formed an important boundary, possibly even the northern boundary of Brigantian territory, making it a particularly appropriate focus for votive deposition. Interestingly, just a little way downstream and near the confluence of the Tees and a stream known as Piercebridge Beck, an altar to Mars Condates (Mars of the Confluence) was found in the 18th century. Five others are known from northern Britain, including a newly discovered altar at Scurragh House (CA 359). It is quite possible that the main shrine was at Piercebridge and that the riverine assemblage is in some way connected with the worship of Mars Condates.

An identity-focused approach

Moving away from the debate about ritual versus rubbish, another way of thinking about the finds from the River Tees is to consider what they reveal about the people who used and deposited them. Such an identity-focused approach is rarely taken for Roman objects from watery contexts, but it has the potential to give many new insights.

Over 190 items of equine equipment were found in the River Tees, some of which are shown in the lower image. Although not all horses were owned by soldiers in the Roman period, the sheer number of fittings recovered does suggest a military connection. The items illustrated include strap slides, distributors, and terrets. Other cavalry-related finds include studs intended for attachment to leather, which are usually interpreted as coming from horse harnesses. Many portray phalluses or vulvas, powerful apotropaic symbols intended to deflect the evil eye, and a number of examples can be seen in the top image.

Several of the objects dating to the 1st and 2nd centuries AD attest to the involvement of the ‘native’ population in depositional activity focused on the river. They include objects such as brooches and cosmetic mortars which are intimately associated with the creation and maintenance of ‘native’ identity. Interestingly, while these objects are ‘native’ in origin, they are not local, as they are all of types found more commonly in southern Britain. The owners of these objects had clearly travelled long distances before depositing them in the river, or the objects were traded from the south.

Despite the presence of these ‘native’ objects, though, the majority of the assemblage is overwhelmingly associated with the Roman military machine. While relatively little weaponry or armour has been recovered, there are large quantities of military belt-fittings and horse harness, suggesting a strong cavalry presence at Piercebridge. In fact, the river has yielded the largest collection of openwork belt-mounts known from the province, and has produced more harness-fittings than all of the cavalry installations on Hadrian’s Wall put together. Both categories of material were potent symbols of military identity and status, with some only having parallels on the German and Danubian limes. This is likely to say something about the ethnic identity of the soldiers who wore and used them, although their presence provides few clues as to whether they were legionaries or auxiliaries. Fortunately, more than 50 lead sealings recovered from the river do provide unequivocal evidence for soldiers of different status, with 13 originating with the Legio VI Victrix at York, three from auxiliary units based at Binchester, and two at Vindolanda. These address labels of consignments of goods and official documents allude to the considerable network of military communication that Piercebridge was part of.

Eighty pins were recovered from the river, with the majority being made from copper alloy, although silver, lead, and bone examples were also present. Particularly unusual are pins with perforated stems, some with glass heads. It is possible that streamers or beads were threaded through these holes.

Contemporary with the 2nd- and 3rd-century militaria are a range of objects securely associated with women and children, including hairpins and bracelets. Given their dating, it seems likely that the individuals represented by these artefacts were connected with the military community and included wives and relations, as well as tradeswomen and slaves. The unparalleled quantity of gold jewellery indicates the involvement of high-status women from the upper echelons of military society, and the finger-ring assemblage is dominated by small examples most likely worn by women and children. Whatever their status and reason for being in Piercebridge, these women appear – like the soldiers – to have had links with the provinces of Germany, Moesia, and Thrace, as at least two items of jewellery only have parallels there.

Future work

One major question remains to be answered: how common was the deposition of objects in rivers in Roman Britain? Inspired by the Piercebridge assemblage, an ongoing project at the University of Reading is now creating a database of finds from across the province, and analysing patterns both among the finds and in their riverine settings. Initial results show how important it is to really understand the fluvial context – was the find made in the riverbed, is it from the riverbank, or has the river changed course, eroding ancient settlement layers? We are also trying to establish whether there is any patterning in the types of objects found in Romano-British rivers. In some cases, it appears that finds were deposited singly, like the bronze head of Claudius or Nero from Saxmundham which is now in the British Museum (see CA 376). In others, large assemblages of objects have never been studied in detail. Watch this space for the results of our survey in the next few years.

You can listen to Hella Eckardt and Philippa Walton discuss Piercebridge further on this episode of the PastCast.
Further reading
Hella and Philippa’s book Bridge over Troubled Water: the Roman finds from the River Tees at Piercebridge in context was published by the Roman Society in August 2021. You can purchase a copy (, or consult an electronic Open Access version at
All of the finds from the River Tees at Piercebridge are catalogued on the Portable Antiquities Scheme database, see
ALL photos: Aaron Watson